The wound epidermis is a transient secretory epithelium that apposes the mesenchymal blastema of a regenerating urodele limb, and is required for regeneration. Previous studies have shown that the positional identity of the blastema is respecified by retinoic acid (RA; Maden, M. (1982),Nature 295, 672-675), that the blastema contains RA (Scadding, S. R. and Maden, M. (1994) Dev. Biol. 162, 608617), and that an RA-reporter gene introduced into the blastema is differentially activated along the proximodistal axis (Brockes, J. P. (1992) Proc. Natl. Acad. Sci. USA 89, 11386-11390). The newt limb wound epidermis has been explanted with minimal mesenchymal contamination and cultured under conditions where it retains expression and inducibility of marker antigens. We have assayed for the release of retinoids from the wound epidermis by coculture with cells transfected with an RA-responsive reporter gene. The reporter was activated to a level corresponding to stimulation by 0.1–1 nM RA, and this activation was substantially conferred by medium conditioned by the wound epidermis. No significant activation was observed for cells transfected with mutated reporter plasmids and analysed in parallel co-cultures. Wound epidermis from contralateral proximal and distal blastemas were compared for reporter activation, and gave a P/D activation ratio significantly greater than 1.Wound epidermis explants were cultured in the presence of tritiated retinol, and extracts were analysed by HPLC on three different columns. Radioactivity was detected in peaks corresponding to didehydroretinol, 9-cis RA and other unidentified metabolites. Analysis of conditioned media samples, some after pulse chase experiments, detected significant release of retinol, 9-cis RA and other metabolites. Although all-trans RA was detectable, the predominant acidic metabolite was 9-cis RA. These experiments establish the wound epidermis as a source of RA for local cellular interactions in the blastema.

Retinoic acid (RA) has the notable ability to respecify positional identity in the vertebrate limb, either along the anteroposterior axis during development (Tickle et al., 1982; Summerbell, 1983) or along the proximodistal (PD) axis during amphibian regeneration (Niazi and Saxena, 1978; Maden, 1982). This activity of exogenous retinoids could reflect an important endogenous role for these molecules in axial patterning. If this interpretation is correct, it should be possible to identify endogenous sources of active retinoids for the limb as one of the criteria for establishing such a role. Studies on chick and mouse limb development have gone some way to identifying potential sources in the limb bud and flank mesoderm (Thaller and Eichele, 1987; Wagner et al., 1990, 1992), but compelling evidence that such sources actually release active retinoids in an appropriate concentration is not yet available. This account is concerned with this issue in the context of limb regeneration.

Urodele amphibians such as the newt and axolotl are the only adult vertebrates that are able to regenerate their limbs. When the limb of an adult newt is amputated at any level along the PD axis, the stump tissues in the vicinity of the amputation plane are induced to de-differentiate, proliferate and form a growth zone or blastema (Wallace, 1981). Blastemal cells retain a memory of their level of origin along the PD axis such that only the distal structures are regenerated (Stocum, 1984). This positional identity is respecified in a dose-dependent manner after exposure to RA or precursor retinoids such as retinol, such that a distal blastema will regenerate additional proximal structures, giving rise to a serial duplication of part of the limb (Maden, 1982; Stocum and Crawford, 1987).

Several observations considered together with the capacity of RA to proximalise the regenerate during a particular stage of its development, suggest that the RA response pathway may be activated during normal limb regeneration. First, newt limb and limb blastemal cells express several different RA receptors (RARs) (Ragsdale et al., 1989, 1993). The RARs, together with a related family of nuclear receptors, the retinoid X receptors (RXRs) act as ligand dependent transcription factors which mediate the biological activities of RA (for review see Leid et al., 1992; Mangelsdorf et al., 1993). Second, analysis by HPLC of extracts of both epidermal and mesenchymal compartments of the axolotl blastema has detected retinoids in significant quantities (Scadding and Maden, 1994). Third, when an RAresponsive reporter gene was introduced into the newt limb blastema, it was found to be differentially activated on the P/D axis (Brockes, 1992). Although the reporter approach could not directly identify the species responsible for activation, the results suggest the possibility that local sources of endogenous retinoids may exist in the blastema. One obvious candidate for such a source is the wound epidermis.

The wound epidermis is formed within nine hours after amputation when the basal cells of the epidermis migrate distally to cover the amputation plane (Hay and Fishman, 1961), and is maintained at the distal tip of the blastema until the onset of digit formation (Salpeter and Singer, 1960). Contact between the wound epidermis and the underlying mesenchymal blastemal cells must be maintained throughout regeneration; replacement of the epidermis with a skin graft during the early stages arrests blastemal formation and removal at later stages results in regeneration of truncated limbs (Stocum, 1985). This latter result suggests that in addition to its importance for blastemal growth (Singer and Salpeter, 1961), the wound epidermis may act as a distal boundary for regenerative outgrowth (Tank and Holder, 1981). The wound epidermis seemed a candidate for a source of retinoids because it is directly apposed to blastemal cells during the period when they are responsive to exogenous RA, and in addition it displays ultrastructural and antigenic characteristics of a secretory epithelium (Singer and Salpeter, 1961; Chapron, 1995; Goldhamer et al., 1989; Ferretti and Brockes, 1988), a tissue type that is well established as requiring retinoids for its maintenance (Darmon, 1991; Lohnes et al., 1993).

In order to investigate this possibility, we have established conditions for culturing the newt wound epidermis without significant mesenchymal contamination and have shown that these explants retain expression of antigenic markers. The cultured tissue releases sufficient retinoid into the medium to activate reporter expression in co-cultured blastemal cells. Direct analysis of wound epidermis tissue and medium after labelling with tritiated retinol has demonstrated active metabolism and release of metabolites. It is intriguing that the principal acidic metabolite is 9-cis RA, a ligand which is able to activate both RARs and RXRs (Heyman et al., 1992; Levin et al., 1992) and which has been shown to be a more potent morphogen in the chick wing bud than all-trans RA (Thaller et al., 1993).

Animals

Newts (Notophthalmus viridescens) were purchased from Charles Sullivan Co., Inc. (Nashville Tennessee, USA). Methods for animal maintenance and surgery were as described previously (Kintner and Brockes, 1985).

Reagents

Synthetic retinoid standards were purchased from Sigma, except for all-trans didehydroretinol and 9-cis retinoic acid which were donated by Hoffmann La Roche. 3H-labelled all-trans retinol (specific activity approx. 40 Ci/mmol) was purchased from NEN. All solvents used in sample analysis were HPLC grade.

Plasmids

The plasmids RSV-RGR, GRE2tkCAT and PBL4 were a gift from D. Mangelsdorf and R. Evans. RSV-RGR encodes a chimaeric receptor protein containing the DNA binding domain of the glucocorticoid receptor and the ligand binding domain of the retinoic acid receptor downstream of the RSV promoter. The GRE2tkCAT plasmid carries the chloramphenicol acetyltransferase (CAT) coding sequence under the control of a tk promoter which had been cloned downstream of two copies of the glucocorticoid response element (Schule et al., 1988). The GREs have been deleted from the control PBL4 construct. The plasmid pSV2A-luc contains the luciferase gene downstream of the SV40 promoter (De Wet et al., 1987) and was obtained from D. Loskutoff.

Preparation of wound epidermis cultures

Six days after hindlimb amputation animals were subjected to terminal anesthesia; the amputated limbs were removed, sterilised in a 0.08% solution of sodium hypochlorite for 10 seconds and rinsed several times in sterile Dulbecco’s phosphate-buffered saline (PBS) which had been adjusted to urodele osmolarity (APBS). Manipulations of tissue in sterile APBS were carried out under aseptic conditions in a laminar flow cabinet and using a dissecting microscope. Iridectomy scissors were used to cut around the circumference of the wound epidermis at the skin/wound epidermis boundary; the wound epidermis was carefully removed from the limb, inverted, trimmed, mechanically stripped of any adherent mesenchymal cells or tissue fragments and rinsed several times in APBS. The tissue was transferred ‘right side up’ to the center well of an IVF culture dish (Costar) which had been coated with 0.5% gelatin or which had been seeded with newt limb blastemal (BlHl) cells (Ferretti and Brockes, 1988) and which contained only a small volume of medium in order to prevent the wound epidermis from floating; the outer well of the culture dish was filled with sterile water to prevent evaporation. Cultures were incubated at 24°C, 2% CO2, as described previously (Fekete and Brockes, 1988), in 250 μl Mimimal Essential Medium with Eagle’s salt (MEM; 63%) supplemented with 10% delipidated foetal calf serum (Rothblat et al., 1976). In some cases gentamicin (35 μg/ml) and neomycin (1 mg/ml) were added to the media as a precaution against bacterial infection which was observed with less healthy batches of animals. The wound epidermis attached to the coated well or cell layer in the first few hours of the culture period; epidermal cell migration generally began within 12 hours. Cultures could be readily maintained for 2 weeks, at which time significant keratinization became apparent. For studies on reporter activation, the wound epidermis was dissected as above, cultured for 24-48 hours on a monolayer of BlHl cells which had been transfected with reporter or control constructs as described below, and processed for enzyme assays. An analysis of the dose response of transfected cells to alltrans RA was done in parallel for each experiment: transfected cells were cultured for an equivalent period of time in media containing 0.1-10 nM RA and processed accordingly. For metabolic studies, several wound epidermis explants (15-20) were cultured together in a cryotube (Nunc) for 24 hours at room temperature in 200-250 μl Leibovitz’s L15 medium (60%) containing 500 nM 3H-labelled alltrans retinol. Tissue pieces were subsequently removed from the culture media and washed several times in PBS. Both tissue and media fractions were collected and stored at −70°C. In pulse-chase experiments the explants were rinsed in APBS, transferred to media containing 500 nM cold retinol and incubated for an additional 24 hours, prior to processing.

Blastemal cell culture and transfection

Newt BlHl cells were cultured as described previously (Ferretti and Brockes, 1988) in 63% MEM supplemented with 10% delipidated foetal calf serum (Rothblat et al., 1976). BlHl cells were transfected with plasmid DNA using a DNA particle delivery system (Dupont PDS-1000H biolistics machine) as described previously (Schilthuis et al., 1993). Typically, 20 μl gold particles were mixed with 3.6 μg GRE2tkCAT or PBL4, 1.6 μg RSV-RGR and 2.6 μg pSV2A-luc DNA prior to CaCl2/spermidine precipitation as described; these particles were used to transfect two 6-cm dishes of approx. 105 cells each.

Enzyme assays

Cells were washed twice in APBS prior to extraction with 125 μl Promega lysis reagent (E1531, diluted 1 in 10) and scraping with a rubber policeman. Extracts were centrifuged at 10,000 g for 3 minutes at 4°C prior to assay. CAT and luciferase assays were performed as described by Brockes (1992). CAT activity was generally normalised with respect to luciferase activity after correcting the latter for the square root relationship between concentration and signal in the assay (see Brockes, 1992).

Immunocytochemistry

The wound epidermis-specific monoclonal antibodies WE3 and WE6 were donated by Dr R. Tassava (Tassava et al., 1986). Wound epidermis cultures were prepared as above and in some cases grown in the presence of 0.1-100 nM RA. Explants were fixed, at various time points, in 100% methanol at −20°C for 10 minutes, rinsed several times with Dulbecco’s PBS containing 10% goat serum (PBS/goat), incubated for 45 minutes at room temperature with primary antibody diluted 1/100 in PBS/goat and rinsed several times with PBS/goat prior to incubation with the secondary antibody. Rhodamine-conjugated rabbit anti-mouse antibody (Dako) was diluted 1/100 in PBS/goat and reacted with the tissue for 30 minutes at room temperature in the dark. Explants were rinsed with PBS/goat followed by PBS and viewed, unmounted, using a Zeiss (Axiophot) microscope.

Column chromatography

Retinoids were extracted from tissue samples as previously described, with minor modifications (Thaller and Eichele, 1987). After sonication in 0.5 ml stabilising buffer, 20 μl aliquots were removed for protein analysis and 200 ng (each) of the appropriate synthetic standards were added prior to extraction with two volumes extraction solvent. The solvent phases were pooled after separation by low speed microcentrifugation, dried under nitrogen, resuspended in 100 μl methanol, centrifuged at high speed and transferred to HPLC autosampler vials. Extracts were analysed by C18 reverse phase and normal phase chromatography using Beckman System Gold Hardware with a uv detector (351 nm) in series with a solid scintillant radioisotope detector. For reverse-phase chromatography, 50 μl of extract was autoinjected onto a 5 μm encapped C18 LiChrospher 100 column (Merck), with an equivalent pre-column and eluted at 1 ml/minute under gradient conditions as follows: 40% mobile phase A (1% acetic acid), 60% mobile phase B (acetonitrile:methanol, 3:1) rising linearly to 100% mobile phase B over 25 minutes. For normal phase chromatography, the method was according to Kawamura et al. (1993), with modifications as described: 50 μl of extract was autoinjected onto a 5 μm LiChrospher 100 NH2 column (Merck), with an equivalent pre-column and eluted at 1 ml/minute, for 5 minutes initially, with 100% mobile phase C (chloroform: methanol, 9:1) changing over 1 minute to 100% mobile phase D (chloroform: methanol: acetic acid, 9.0:0.9:0.1) for a further 20 minutes. The eluant was monitored as above. The fractions corresponding to the RA peak were pooled, dried down, resuspended in 100 μl methanol and reanalysed by reverse-phase chromatography.

HPLC on a Partisil ODS column (25×0.46 cm) used solvent elution conditions (0.5 ml/minute) modified from a previous report (Cullum and Zile, 1986) with 70% methanol for 5.5 minutes, 80% methanol for 5.5 minutes, and 88% methanol for 29 minutes. Tissue or medium samples were extracted as described above and resuspended in 500 μl 50% methanol prior to injection onto the column. Fractions were collected and analysed for radioactivity by scintillation counting.

Gel filtration was performed directly on medium samples or tissue cytosol extracts on a calibrated Superose 12 column (Pharmacia) run in 50 mM Tris-HCl, pH 7.4, 0.15 M NaCl. Fractions were collected and analysed as above.

Culture of wound epidermis explants that retain cell identity

The wound epidermis was dissected from newt limb blastemas at 5 or 6 days post-amputation, rinsed in PBS to remove loosely adherent cells, and transferred to gelatin-coated culture dishes. Keratinocytes began to migrate radially from the tissue explant within 24 hours (Fig. 1). Several cell type-specific antibodies were used to characterise the cultures. The WE6 antigen is first detected in vivo at approximately 1–2 days post-amputation and continues to be expressed by all cells in the wound epidermis until the late stages of regeneration (Tassava et al., 1986). The explant cultures also showed widespread expression of WE6, both by migrating cells and those remaining in the explant (not shown). The WE3 antigen is an RA-inducible cytoskeletal-associated protein expressed in the basal layer of the newt wound epidermis and is first detected from 7-8 days after amputation (Goldhamer et al., 1989). Although the wound epidermis was removed from the blastema prior to the onset of WE3 expression, the antigen was detected in the migratory population after about 3 days in culture confirming their identity as the basal cells (Fig. 2A).

Fig. 1.

A typical wound epidermis explant after 48 hours in culture. The basal cells (b) have migrated radially approximately 0.7 mm from the edge of the explant tissue (e) which is approx. 2 mm in diameter. This explant was co-cultured with BlHl blastemal cells, which can be seen at the periphery (m). Scale bar, 100 μm.

Fig. 1.

A typical wound epidermis explant after 48 hours in culture. The basal cells (b) have migrated radially approximately 0.7 mm from the edge of the explant tissue (e) which is approx. 2 mm in diameter. This explant was co-cultured with BlHl blastemal cells, which can be seen at the periphery (m). Scale bar, 100 μm.

Fig. 2.

Expression of the wound epidermis-specific antigen WE3 in cultured wound epidermis. The WE3 antigen has been detected with a monoclonal antibody (mAb) and visualised with a rhodamine-labelled secondary antibody. The edge of the explant is indicated with a broken white line. (A) WE3 expression as it first appears (after 72 hours); reactivity is restricted to the basal cell population. The expression of the antigen does not depend on co-culture with blastemal cells. (B) WE3 expression in a 24 hour culture treated with 100 nM RA. The WE3 antigen is expressed earlier than in the untreated culture (A,D). Here the expression is not restricted to the basal cells; rhodamine-labelled cells can be seen within the explant tissue (out of the plane of focus). (C) A similar 24 hour wound epidermis culture which has not been treated with RA. (D) The same culture as in C after indirect immunofluorescence to detect the WE3 antigen; essentially no reactivity with the mAb is seen in untreated explants at this time. Scale bar, 100 μm.

Fig. 2.

Expression of the wound epidermis-specific antigen WE3 in cultured wound epidermis. The WE3 antigen has been detected with a monoclonal antibody (mAb) and visualised with a rhodamine-labelled secondary antibody. The edge of the explant is indicated with a broken white line. (A) WE3 expression as it first appears (after 72 hours); reactivity is restricted to the basal cell population. The expression of the antigen does not depend on co-culture with blastemal cells. (B) WE3 expression in a 24 hour culture treated with 100 nM RA. The WE3 antigen is expressed earlier than in the untreated culture (A,D). Here the expression is not restricted to the basal cells; rhodamine-labelled cells can be seen within the explant tissue (out of the plane of focus). (C) A similar 24 hour wound epidermis culture which has not been treated with RA. (D) The same culture as in C after indirect immunofluorescence to detect the WE3 antigen; essentially no reactivity with the mAb is seen in untreated explants at this time. Scale bar, 100 μm.

The expression of WE3 is strongly induced in newts treated with retinol or retinoic acid at various times post-amputation (Tassava, 1992). When explants were cultured in the presence of all-trans RA (10-100 nM), WE3 was detected by immunofluorescence after only 1 day of culture, and the antigen was no longer restricted to the basal cell population (Fig. 2B-D). Thus the explants express characteristic antigens in vitro and retain the ability to respond to RA. Wound epidermis explants that were given an 8-hour pulse of tritiated thymidine showed no labelled nuclei after autoradiography. This result parallels the early in vivo observations of Hay and Fischman (1961).

In order to assess the degree of mesenchymal contamination in the cultures, we used a blastemal cell-specific antibody 22/18 which reacts with the mesenchymal population underlying the wound epidermis (Gordon and Brockes, 1988). Several 1-3 day cultures were stained with the 22/18 antibody and were found to contain, on average, fewer than 30 and often less than 10 positive cells (n=27) out of approximately five thousand in the explant. This low level of contamination was a direct result of the fact that the wound epidermis was removed relatively early post-amputation when there were few adherent blastemal cells, most of which could be removed prior to culture with fine forceps and by gentle agitation (see Materials and Methods section).

RA-dependent reporter activity in cultured blastemal cells

Mesenchymal blastemal cells were co-transfected with three plasmids that provide a normalised reporter system which was used to monitor the level of retinoid in culture. One expresses a chimaeric receptor (the RGR) in which the ligand-binding domain of the glucocorticoid receptor has been replaced with the corresponding domain of RAR α1 (Fig. 3A). The second is a chloramphenicol acetyltransferase (CAT) reporter with regulatory sequences that include two glucocorticoid response elements (GREs; Fig. 3B). In the presence of RA, the chimaeric receptor activates CAT expression through the glucocorticoid response pathway. The third plasmid is a constitutively active luciferase construct that allows CAT expression to be normalised for differences in transfection efficiency or recovery of activity in extracts. The dose response of the transfected blastemal cells to the presence of 0.1 –100 nM all-trans and 9-cis isomers of RA in the culture medium is shown in Fig. 4. Transfected cells responded comparably to all-trans and 9-cis RA, but were much less sensitive to retinol (4A-C). No significant response was detected with control cell populations transfected with vector in place of the RGR plasmid (4A), or transfected with a reporter lacking the GREs (4B). In addition, no significant stimulation of basal reporter activity was observed when transfected cells were exposed to 0.1–100 nM dexamethasone, indicating a lack of responsiveness to glucocorticoids (not shown).

Fig. 3.

Schematic outline of the retinoid reporter system. (A) The chimaeric receptor construct is a fusion of the RAR ligand binding domain and the glucocorticoid receptor DNA binding domain, downstream of the RSV promoter. (B) The reporter construct: the CAT gene is located downstream of two copies of the glucocorticoid response element and the thymidine kinase promoter. RA stimulates the reporter by activating the chimaeric receptor at the GREs.

Fig. 3.

Schematic outline of the retinoid reporter system. (A) The chimaeric receptor construct is a fusion of the RAR ligand binding domain and the glucocorticoid receptor DNA binding domain, downstream of the RSV promoter. (B) The reporter construct: the CAT gene is located downstream of two copies of the glucocorticoid response element and the thymidine kinase promoter. RA stimulates the reporter by activating the chimaeric receptor at the GREs.

Fig. 4.

Reporter gene activation in response to RA. In addition to pSV2A, BlHl cells were tranfected with the GRE2tkCAT reporter plasmid and either RSV-RGR or the Puc 18 control (A) or with RSV-RGR and either GRE2tkCAT or the PBL4 control (B). Transfected cells were exposed to varying doses of RA from 0.1–100 nM. A clear dose response was observed when both the chimaeric receptor and the reporter were present in the cells (open diamonds). No activity above background was seen when the reporter was co-transfected with PUC18, or when the chimaeric receptor was cotransfected with PBL4 (filled diamonds). (C) The response seen with 9-cis RA (closed diamonds) was comparable to that seen with all-trans RA (open diamonds) and significantly greater than the response to retinol at the same concentration (dotted line).

Fig. 4.

Reporter gene activation in response to RA. In addition to pSV2A, BlHl cells were tranfected with the GRE2tkCAT reporter plasmid and either RSV-RGR or the Puc 18 control (A) or with RSV-RGR and either GRE2tkCAT or the PBL4 control (B). Transfected cells were exposed to varying doses of RA from 0.1–100 nM. A clear dose response was observed when both the chimaeric receptor and the reporter were present in the cells (open diamonds). No activity above background was seen when the reporter was co-transfected with PUC18, or when the chimaeric receptor was cotransfected with PBL4 (filled diamonds). (C) The response seen with 9-cis RA (closed diamonds) was comparable to that seen with all-trans RA (open diamonds) and significantly greater than the response to retinol at the same concentration (dotted line).

Activation of reporter cells by co-culture with wound epidermis

In order to assay for retinoid release from the wound epidermis, reporter cells were co-cultured with wound epidermis explants in 250 μl of medium, or alone in varying concentrations of RA. Cells that were co-cultured with the wound epidermis for 48 hours showed a stimulation of the CAT reporter which consistently fell into the range observed between 0.1–1 nM RA as determined in the parallel dose response analysis (Fig. 5). This stimulation was not observed in paired controls in which the wound epidermis from the contralateral limb of the same animal was co-cultured with control cells lacking the RGR, or in other cases where the GREs were absent from the reporter. Comparable results were obtained in 104 such culture pairs.

Fig. 5.

Induction of the RA-dependent CAT reporter gene in cells cocultured with the wound epidermis. The y-axis represents relative activation of the CAT reporter gene. Each bar on the x-axis represents the results from experimental (light grey) and control (dark grey) cultures which were done using wound epidermis tissue from the contralateral limbs of a single animal; the results have been superimposed for ease of comparison. (A) Results from a typical experiment in which cells were co-transfected with RSV-RGR (chimaeric receptor), GRE2tkCAT (or PUC18, as a control) and luciferase constructs and subsequently cultured with wound epidermis tissue. (C) The level of normalised CAT activity detected in the experimental co-cultures fell into the range of activity detected between 0.1–1.0 nM RA in the dose response. In control co-cultures, where only the reporter construct was transfected, CAT activity was at background. (B) Results from a similar experiment using PBL4 (the CAT reporter without the upstream GREs) as the control. (D) Results from the corresponding dose response analysis showing activation of the reporter between 0.1–1.0 nM RA.

Fig. 5.

Induction of the RA-dependent CAT reporter gene in cells cocultured with the wound epidermis. The y-axis represents relative activation of the CAT reporter gene. Each bar on the x-axis represents the results from experimental (light grey) and control (dark grey) cultures which were done using wound epidermis tissue from the contralateral limbs of a single animal; the results have been superimposed for ease of comparison. (A) Results from a typical experiment in which cells were co-transfected with RSV-RGR (chimaeric receptor), GRE2tkCAT (or PUC18, as a control) and luciferase constructs and subsequently cultured with wound epidermis tissue. (C) The level of normalised CAT activity detected in the experimental co-cultures fell into the range of activity detected between 0.1–1.0 nM RA in the dose response. In control co-cultures, where only the reporter construct was transfected, CAT activity was at background. (B) Results from a similar experiment using PBL4 (the CAT reporter without the upstream GREs) as the control. (D) Results from the corresponding dose response analysis showing activation of the reporter between 0.1–1.0 nM RA.

Conditioned media from several wound epidermis cultures was collected and assayed on reporter cells in order to determine if reporter activation was dependent on contact between the two cell types. In direct comparisons, conditioned media from 30 separate explant cultures was able to stimulate reporter cells at approximately 70% of the level of activation observed in coculture. We conclude that this property of the cultured wound epidermis is substantially independent of cell contact.

Comparison of stimulation by proximal and distal wound epidermis

We have compared reporter activation in transfected cells cocultured with proximal or distal explants to examine the possibility that wound epidermis originating at different axial levels might release different concentrations of retinoid. The wound epidermis explants were derived from contralateral newt hindlimb blastemas so that the axial comparison was made within a single animal. The ratio of normalised CAT reporter activities was determined for 33 paired co-cultures. The P/D ratio of activity was significantly greater than one, as determined by the sign test assuming a hypothesis of no difference (P<0.01). A similar comparison of values for luciferase, the co-transfected normalising activity, gave a mean P/D ratio of 1 and no significant P/D difference. Thus the results (Fig. 6) are consistent with a difference in the ability of the wound epidermis originating from proximal and distal blastemas to activate the reporter. However it should be noted that the comparison is made difficult because the response of the reporter is not linear with the concentration of RA in the range of 0.1 nM.

Fig. 6.

Reporter stimulation by proximal versus distal wound epidermis. The results of a comparison of the level of reporter gene activation in cells co-cultured with proximal (P) versus distal (D) wound epidermis. Each diamond represents the P:D ratio of normalised CAT activity from contralateral limbs of a single animal (n=33). This ratio was significantly greater than one (dotted line), as determined by the sign test (Colquhoun, 1971) (P<0.01). The mean P/D value = 1.3. A similar comparison of normalising luciferase activity in the same co-cultures gave a mean P/D value of 1 and thus no significant difference.

Fig. 6.

Reporter stimulation by proximal versus distal wound epidermis. The results of a comparison of the level of reporter gene activation in cells co-cultured with proximal (P) versus distal (D) wound epidermis. Each diamond represents the P:D ratio of normalised CAT activity from contralateral limbs of a single animal (n=33). This ratio was significantly greater than one (dotted line), as determined by the sign test (Colquhoun, 1971) (P<0.01). The mean P/D value = 1.3. A similar comparison of normalising luciferase activity in the same co-cultures gave a mean P/D value of 1 and thus no significant difference.

Metabolism of radiolabelled retinol by isolated wound epidermis

The wound epidermis was dissected from 15–20 hindlimb blastemas and cultured for 24 hours in 0.25 ml of medium containing 0.5 μM 3H-labelled retinol. The tissue and culture medium were separately extracted and processed for HPLC analysis. In some cases the labelled wound epidermis tissue was washed and incubated for a further 24 hours in medium containing unlabelled retinol, thus yielding ‘chased’ medium and tissue samples. Fractionation of the labelled tissue metabolites on a reverse phase C18 HPLC column revealed a major metabolite peak that co-eluted with 3,4-didehydroretinol and which constituted approximately 4–5 percent of the radioactivity in the retinol peak (Fig. 7A). RA and retinol were not clearly resolved on the C18 column, therefore the acidic metabolites were separated from unmetabolised retinol by analysis of tissue and media samples on a LiChrospher NH2 column (Fig. 7B). Both samples showed a large increase in NH2-elutable radioactivity over control media samples that were cultured in the absence of tissue. When the NH2 fractions corresponding to the acidic metabolites were collected and reanalysed by reverse phase C18 HPLC, two radioactive peaks were detected (Fig. 7C). One was an unidentified species (at 20 minutes, peak 7) which also appeared in control chromatograms from medium samples that had been incubated without tissue and which comigrated with the contaminant in the original reverse phase runs (Fig. 7A). The second radioactive peak, corresponding to approximately 40% of the radioactivity, coeluted with authentic 9-cis RA (Fig. 7C – peak 5, solid line). This peak was also detected after analysis of the medium conditioned by cultured wound epidermis, confirming that the 9-cis RA generated by the wound epidermis is released into the medium (Fig. 7D – peak 5, solid line). In the example of Fig. 7D, all the radioactivity migrated as 9-cis RA. We can clearly distinguish between 9-cis and all-trans RA with this methodology, as shown by the separation of the absorbance peaks of authentic 9-cis and all-trans RA in the examples of Fig. 7C,D. It should be noted in Fig. 7C that a minor peak of radioactivity coeluting with all-trans RA was at least 10-fold lower than the 9-cis peak, confirming the separation of the two isomers. In other unpublished studies using the same methods but embryonic tissues from other species, we can detect all-trans RA but not 9-cis RA.

Fig. 7.

Typical chromatograms illustrating the metabolic conversion of 3H-labelled all-trans retinol into 3H-labelled 3,4didehydroretinol and 3H-labelled 9-cis retinoic acid by wound epidermis tissue. Retention times of internal standards are indicated by arrowheads. Profiles were reproducible and comparable for both media and tissue samples unless otherwise indicated. All of the peaks discussed, with one exception (see A) appear only in the presence of the wound epidermis.(A) Reverse phase C18HPLC fractionation of 3H-labelled metabolites formed after incubation of the wound epidermis with 3H-labelled retinol. Peak 1 co-elutes with 3,4didehydroretinol and was notably only detected in the tissue fraction. Peak 2 corresponds to unmetabolised retinol. Peak 3 co-elutes with retinol (at approximately 20 minutes), but was often present in control media and thus was disregarded as a contaminant (see also C). (B) Normal phase NH2 HPLC of 3Hlabelled metabolites (as above) to separate the acidic metabolites from the non-acidic retinoids. Unmetabolised retinol elutes in the void volume (v). Peak 4 co-elutes with RA and represents all acid metabolites. (C) Fractions corresponding to peak 4 reanalysed by reverse phase C18HPLC. The solid line marks the radioactivity detected, the dotted line is the UV absorbance of authentic standards. Peak 5 is 9-cis RA and coelutes with a significant peak of radioactivity. Peak 6 is all-trans RA and a small peak of radioactivity is detectable. Peak 7 corresponds to the contaminant as it is present in control incubations without tissue. (D) Reverse phase C18 HPLC fractionation of the acidic metabolites from the medium of a culture of wound epidermis. The solid line marks radioactivity detected, and the dotted line the UV absorbance of authentic standards. The sole peak of radioactivity coelutes with 9-cis RA (peak 5). (E) Pulse chase media were analysed using a Partisil ODS column. The peaks labelled x may represent unidentified metabolites. Peak 4 co-elutes with 9-cis and all-trans RA. Peak 2 corresponds to unmetabolised retinol. It was not possible to resolve the various isomers of RA from each other, or 3,4-didehydroretinol from retinol using this system.

Fig. 7.

Typical chromatograms illustrating the metabolic conversion of 3H-labelled all-trans retinol into 3H-labelled 3,4didehydroretinol and 3H-labelled 9-cis retinoic acid by wound epidermis tissue. Retention times of internal standards are indicated by arrowheads. Profiles were reproducible and comparable for both media and tissue samples unless otherwise indicated. All of the peaks discussed, with one exception (see A) appear only in the presence of the wound epidermis.(A) Reverse phase C18HPLC fractionation of 3H-labelled metabolites formed after incubation of the wound epidermis with 3H-labelled retinol. Peak 1 co-elutes with 3,4didehydroretinol and was notably only detected in the tissue fraction. Peak 2 corresponds to unmetabolised retinol. Peak 3 co-elutes with retinol (at approximately 20 minutes), but was often present in control media and thus was disregarded as a contaminant (see also C). (B) Normal phase NH2 HPLC of 3Hlabelled metabolites (as above) to separate the acidic metabolites from the non-acidic retinoids. Unmetabolised retinol elutes in the void volume (v). Peak 4 co-elutes with RA and represents all acid metabolites. (C) Fractions corresponding to peak 4 reanalysed by reverse phase C18HPLC. The solid line marks the radioactivity detected, the dotted line is the UV absorbance of authentic standards. Peak 5 is 9-cis RA and coelutes with a significant peak of radioactivity. Peak 6 is all-trans RA and a small peak of radioactivity is detectable. Peak 7 corresponds to the contaminant as it is present in control incubations without tissue. (D) Reverse phase C18 HPLC fractionation of the acidic metabolites from the medium of a culture of wound epidermis. The solid line marks radioactivity detected, and the dotted line the UV absorbance of authentic standards. The sole peak of radioactivity coelutes with 9-cis RA (peak 5). (E) Pulse chase media were analysed using a Partisil ODS column. The peaks labelled x may represent unidentified metabolites. Peak 4 co-elutes with 9-cis and all-trans RA. Peak 2 corresponds to unmetabolised retinol. It was not possible to resolve the various isomers of RA from each other, or 3,4-didehydroretinol from retinol using this system.

Chased medium samples were analysed by HPLC on a silica column under conditions that resolved retinoic acid and retinol but did not separate 9-cis and all-trans isomers of retinoic acid. The elution profile revealed the presence of RA and retinol, as well as unidentified metabolites (fractions 6–12, Fig. 7D). The release of retinol and metabolites from the isolated wound epidermis was quantitatively significant, since subcellular fractionation of labelled tissue before and after the chase showed that approximately 60% of the cytosolic pool was released in 24 hours and appeared in the medium. When the radioactivity released into serum-free medium was analysed by gel filtration there was no evidence for significant association of the retinoids with proteins or other higher molecular mass components (data not shown).

Tissue sources of retinoids

As yet there is no compelling evidence that retinoids act as local autocrine or paracrine mediators of cellular interactions. A key component of such evidence would be the identification of cell sources that release active retinoids, in particular RA and its isomers and metabolites, in significant amounts. Because retinoids may associate with proteins that mask their ability to alter gene expression, it is important to assay both for chemical identity and for activation of reporter genes. Evidence has been presented that cells in Hensen’s node, the notochord and the floorplate may act as sources of RA, based on the two assay procedures (Wagner et al., 1990; Hogan et al., 1992; Wagner et al., 1992), but direct evidence on the release of retinoids from the isolated tissue fragments is not available. Furthermore the target cells for retinoid released from these tissues are unclear, although an interesting hypothesis has been suggested for Hensen’s node (Hogan et al., 1992). The present account shows that the adult newt wound epidermis can be cultured in numbers that permit a detailed evaluation of its activity as a source. Its location at the extremity of the regenerating limb suggests the mesenchymal cells of the blastema as a potential target.

Reporter stimulation by the isolated wound epidermis

The evidence from antigenic markers strongly indicates that the epidermis retains its identity when isolated in culture. These cultures are free of significant mesenchymal contamination and can be used to study aspects of wound epidermis function. We have used this culture system in conjunction with a sensitive reporter assay which depends on the activity of a chimaeric glucocorticoid/RA receptor. The host cells for the reporter system were derived originally from blastemal mesenchyme and express several markers of blastemal cells (Ferretti and Brockes, 1988). The results obtained from a comparison of the level of reporter activation in experimental and control cultures have shown consistent stimulation by the wound epidermis; this stimulation was in the range evoked by 0.1-1 nM RA. It is noteworthy that in vitro experiments comparing the capacity of chick limb bud mesoderm and ectoderm to stimulate a comparable RA-responsive reporter showed no activity with ectoderm (Wagner et al., 1992). The media conditioned by the wound epidermis was sufficient for reporter activation, albeit to a lesser extent than the tissue itself. This could reflect instability or degradation of retinoid after release, nevertheless the results clearly show that contact between the wound epidermis and the responding cells is not required. Reporter stimulation is not due to a general activation of transcription by the wound epidermis as there is no effect on the normalising activity of the luciferase reporter. Nor does stimulation reflect an increase in cell division in the cocultures, since we found no increase in labelling index after comparison of tritiated thymidine uptake in cells co-cultured with or without the wound epidermis.

The apparently low level of stimulation is more readily appreciated in the context of the blastema. The volume of culture medium containing a single explant is approximately 10 times larger than that of an adult limb blastema at a comparable stage (Fig. 8). Estimates for the levels of RA present in the amphibian blastema fall into the range 1 –50 nM (Scadding and Maden, 1994); this would translate to a concentration of 0.1–5 nM in the culture dish. It is likely that some of the RA extracted from the blastema is not normally available to activate gene expression, and thus the level of activation observed in the co-culture experiments appears reasonable. When a comparison was made of reporter activation by proximal and distal wound epidermis from single animals, the ratio of activation (P/D) was significantly greater than one. The results are limited in reproducibilty by the sensitivity of the reporter at the lower end of its concentration range, but raise the possibility that the wound epidermis could contribute to the differential activation of a reporter in implanted cells (Brockes, 1992), and more generally to the proximodistal identity of the blastema. A stronger interpretation awaits confirmation by a more sensitive assay.

Fig. 8.

Schematic comparison between the volume of a blastema and the wound epidermis culture, with respect to local retinoid concentration. The response of the RA-dependent reporter gene in wound epidermis cultures is comparable to that seen in cultures treated with 0.1 nM RA. However, the wound epidermis is cultured in 200 μl of medium; this is 10 times the estimated volume of the adult limb blastema (calculated as πr2h). Therefore, the concentration of RA in the culture dish extrapolates to a 10-fold higher concentration of RA in the limb.

Fig. 8.

Schematic comparison between the volume of a blastema and the wound epidermis culture, with respect to local retinoid concentration. The response of the RA-dependent reporter gene in wound epidermis cultures is comparable to that seen in cultures treated with 0.1 nM RA. However, the wound epidermis is cultured in 200 μl of medium; this is 10 times the estimated volume of the adult limb blastema (calculated as πr2h). Therefore, the concentration of RA in the culture dish extrapolates to a 10-fold higher concentration of RA in the limb.

Retinol metabolism in the isolated wound epidermis

The isolated wound epidermis actively metabolised tritiated retinol to didehydroretinol in short term incubations. It is of considerable interest that the primary acidic metabolite in analysis of both tissue and medium samples was 9-cis RA, rather than all-trans RA, which was detected at a much lower level. Previous studies of endogenous retinoids in the axolotl blastema have shown all-trans RA to be present in both the epidermal and mesodermal compartments of the blastema but would not have resolved the 9-cis RA from retinol (Scadding and Maden, 1994). The 9-cis isomer is a ligand for both RARs and RXRs, and has been detected as an endogenous retinoid in Xenopus embryos at neurulation and other stages (Creech Kraft et al., 1994), and as an isomerisation product of all-trans RA in mice (Kojima et al., 1994). It is not possible at present to identify a functional consequence that is particular to 9-cis RA in the wound epidermis, but this may be possible in future, for example in relation to the recently discovered activity of RXRs as dimerisation partners with the NGF1B and NURR1 receptors (Perlmann and Jansson, 1995).

The wound epidermis gave a number of other metabolites, some probably acidic, that were detected on the three HPLC columns but were not identified. These components, along with precursor retinol, were released into the medium by a process that resulted in rapid turnover of the cytosolic pool in pulsechase experiments (>50% in 24 hours at 25°C). It is noteworthy that the reporter co-culture experiments were performed in medium with lipid-extracted serum, in which release may be sustained from endogenous retinoid stores. It is clear that given the considerably lower activity of retinol towards the reporter, the activation observed in co-culture reflects the activity of 9cis RA and possibly other acidic metabolites.

It is interesting to compare these results with those recently obtained in comparable studies of retinol metabolism by human skin keratinocytes in culture. In one study the cells metabolised retinol to didehydroretinol and esterified forms, with little or no radioactivity detectable in retinoic acids (Rollman et al., 1993). A second study also showed a preponderance of didehydroretinol and esterified forms. However, small amounts of RA and didehydro RA were also detected in both cells and medium, although the fractional release into the medium was much less than that in the present study (Randolph and Simon, 1993). Some caution is warranted in this comparison of cultures from newt and man, but the metabolism in keratinocytes from the secretory epithelium of the wound epidermis seems to be biased towards acidic metabolites and release into the medium. It will be interesting in future to analyse retinol metabolism in a comparable mammalian epithelium.

RA as an endogenous ligand in regeneration

The present study has established the local synthesis and release of 9-cis RA by the wound epidermis, thus raising the question of what significance these properties might have in the context of the blastema. It is likely, as mentioned earlier, that RA is important for the maintenance and differentiation of the wound epidermis itself. The wound epidermis is distinct from the keratinising epithelium of the skin yet it is derived by migration of keratinocytes from the skin; it is interesting to speculate that the events of migration across the amputation plane provoke some change in retinol metabolism that results in enhanced production of RA. A second possible function is as a source for the mesenchymal cells of the blastema; RA from the wound epidermis could be implicated in promoting their formation by de-differentiation, or specifying their positional identity, possibly in conjunction with other modulators. It may not be the only source since we have preliminary evidence that cultured mesenchymal blastemal cells have some ability in isolation to metabolise retinol to RA. Taken in conjunction with the studies mentioned earlier, the present results strengthen the possibility that RA plays a significant role in the regeneration blastema. It will be interesting in future to study the consequences for regeneration of blocking the synthetic pathways and the receptor-based response mechanisms.

We are grateful to R. Evans and D. Mangelsdorf for kindly providing the constructs used in the reporter assay and to R. Tassava for donating the WE3 antibody. We thank G. Dunn for advice on statistical analysis, D. Wylie for technical assistance and O. Truong and G. Panayotou for help with column chromatography. We also thank K. Griffin for critical reading of the manuscript and constructive comments. This work was supported, in part, by a postdoctoral grant from the HFSPO. Financial assistance for HPLC analysis was provided by Action Research.

Brockes
,
J. P.
(
1992
).
Introduction of a retinoid reporter gene into the urodele limb blastema
.
Proc. Natl. Acad. Sci. USA
89
,
11386
11390
.
Chapron
,
C.
(
1974
).
Mise en évidence du role dans la régénération des amphibiens, d’une glycoprtéine secrétée par la cape apicale: étude cytochimique et autoradiographique en microscopie électronique
.
J. Embryol. Exp. Morphol
.
32
,
133
145
.
Colquhoun
,
D.
(
1971
).
Lectures on Biostatistics
Oxford
:
Oxford University Press
.
Creech Kraft
,
J.
,
Schuh
,
T.
,
Juchau
,
M.
and
Kimelman
,
D.
(
1994
).
The retinoid X receptor ligand, 9-cis-retinoic acid, is a potential regulator of early Xenopus development
.
Proc. Natl. Acad. Sci.USA
91
,
3067
3071
.
Cullum
,
M. E.
and
Zile
,
M. H.
(
1986
).
Quantitation of biological retinoids by high-pressure liquid chromatography: primary internal standardization using tritiated retinoids
.
Analyt. Biochem
.
153
,
23
32
.
Darmon
,
M.
(
1991
).
Retinoic acid in skin and epithelia
.
Sem. Dev. Biol
.
2
,
219228
.
De Wet
,
J. R.
,
Wood
,
K. V.
,
DeLuca
,
M.
,
Helinski
,
D. R.
and
Subramani
,
S.
(
1987
).
Firefly luciferase gene: structure and expression in mammalian cells
.
Mol. Cell. Biol
.
7
,
725
737
.
Fekete
,
D. M.
and
Brockes
,
J. P.
(
1988
).
Evidence that the nerve controls molecular identity of progenitor cells for limb regeneration
.
Development
103
,
567
573
.
Ferretti
,
P.
and
Brockes
,
J. P.
(
1988
).
Culture of newt cells from different tissues and their expression of a regeneration-associated antigen
.
J. Exp. Zool
.
247
,
77
91
.
Goldhamer
,
D. J.
,
Tomlinson
,
B. L.
and
Tassava
,
R. A.
(
1989
).
A developmentally regulated wound epithelial antigen of the newt limb regenerate is also present in a variety of secretory/transport cell types
.
Dev. Biol
.
135
,
392
404
.
Gordon
,
H.
and
Brockes
,
J. P.
(
1988
).
Appearance and regulation of an antigen associated with limb regeneration in N. viridescens. J. Exp. Zool
.
247
,
232
243
.
Hay
,
E. D.
and
Fishman
,
D. A.
(
1961
).
Origin of the blastema in regenerating limbs of the newt Triturus viridescens. An autoradiographic study using tritiated thymidine to follow cell proliferation and migration
.
Dev. Biol
.
3
,
49
77
.
Heyman
,
R. A.
,
Mangelsdorf
,
D. J.
,
Dyck
,
J. A.
,
Stein
,
R. B.
,
Eichele
,
G.
,
Evans
,
R. M.
,
Thaller
,
C.
(
1992
).
9 cis retinoic acid is a high affinity ligand for the retinoid X receptor
.
Cell
68
,
397
406
.
Hogan
,
B. L. M.
,
Thaller
,
C.
and
Eichele
,
G.
(
1992
).
Evidence that Hensen’s node is a site of retinoic acid synthesis
.
Nature
359
,
237
241
.
Kawamura
,
K.
,
Hara
,
K.
and
Fujuwara
,
S.
(
1993
).
Developmental role of endogenous retinoids in the determination of morphallactic field in budding tunicates
.
Development
117
,
835
845
.
Kintner
C. R.
and
Brockes
J. P.
(
1985
).
Monoclonal antibodies to the cells of a regenerating limb
.
J. Embryol. Exp. Morph
.
89
,
37
55
.
Kojima
,
R.
,
Fujimori
,
T.
,
Kiyota
,
N.
,
Toriya
,
Y.
,
Fukuda
,
T.
,
Ohashi
,
T.
,
Sato
,
T.
,
Yoshizawa
,
Y.
,
Takeyama
,
K.-I.
,
Mano
,
H.
,
Masushige
,
S.
and
Kato
,
S.
(
1994
).
In vivo isomerisation of retinoic acids
.
J. Biol. Chem
.
269
,
32700
32707
.
Leid
,
M.
,
Kastner
,
P.
and
Chambon
,
P.
(
1992
).
Multiplicity generates diversity in the retinoic acid signalling pathways
.
Trends Biochem. Sci
.
17
,
427
433
.
Levin
,
A. A.
,
Sturzenbecker
,
L. J.
,
Kazmer
,
S.
,
Bosakowski
,
T.
,
Huselton
,
C.
,
Allenby
,
G.
,
Speck
,
J.
,
Dratzeisen
,
C.
et al. 
(
1992
).
9 cis retinoic acid stereoisomer binds and activates the nuclear receptor RXR alpha
.
Nature
355
,
359
361
.
Lohnes
,
D.
,
Dastner
,
P.
,
Dierich
,
A.
,
Mark
,
M.
,
LeMeur
,
M.
and
Chambon
,
P.
(
1993
).
Function of retinoic acid receptor γ (RARγ) in the mouse
.
Cell
73
,
643
658
.
Maden
,
M.
(
1982
).
Vitamin A and pattern formation in the regenerating limb
.
Nature
295
,
672
675
.
Mangelsdorf
,
D. J.
,
Klewer
,
S. A.
,
Kakizuka
,
A.
,
Umesono
,
K.
and
Evans
,
R. M.
(
1993
)
Retinoid receptors
.
Recent Prog. Hormone Res
.
48
,
99
121
.
Niazi
,
I. A.
and
Saxena
,
S.
(
1978
).
Abnormal hind limb regeneration in tadpoles of the toad, Bufo andersoni, exposed to excess vitamin A
.
Folia Biol. (Krakow)
26
,
3
11
.
Perlmann
,
T.
and
Jansson
,
L.
(
1995
).
A novel pathway for vitamin A signalling mediated by RXR heterodimerization with NGFI-B and NURRI
.
Genes Dev
.
9
,
769
782
.
Ragsdale
C. W.
,
Gates
P. B.
,
Hill
D. S.
and
Brockes
J. P.
(
1993
).
Delta retinoic acid receptor isoform δ1 is distinguished by its N-terminal sequence and abundance in the limb regeneration blastema
.
Mech. Dev
.
40
,
99
112
.
Ragsdale
,
C. W.
,
Petkovich
,
M.
,
Gates
,
P. B.
,
Chambon
,
P.
and
Brockes
,
J. P.
(
1989
).
Identification of a novel retinoic acid receptor in regenerative tissues of the newt
.
Nature
341
,
654
657
.
Randolph
R. K.
and
Simon
M.
(
1993
).
Characterization of retinol metabolism in cultured human epidermal keratinocytes
.
J. Biol. Chem
.
268
,
9198
9205
.
Rollman
,
O.
,
Wood
,
E. J.
,
Olsson
,
M. J.
and
Cunliffe
,
W. J.
(
1993
).
Biosynthesis of 3,4-didehydroretinol from retinol by human skin keratinocytes in culture
.
Biochem. J
.
293
,
675
682
.
Rothblat
,
G. H.
,
Arbogast
,
L. Y.
,
Ouellette
,
L.
and
Howard
,
B. V.
(
1976
).
Preparation of delipidized serum protein for use in cell culture systems
.
In Vitro
12
,
554
557
.
Salpeter
,
M. M.
and
Singer
,
M.
(
1960
).
Differentiation of the Submicroscopic adepidermal membrane during limb regeneration in adult Triturus, including a note on the use of the term, basement membrane
.
Anat. Rec
.
136
,
27
32
.
Scadding
,
S. R.
and
Maden
,
M.
(
1994
).
Retinoic acid gradients during limb regeneration
.
Dev. Biol
.
162
,
608
617
.
Schilthuis
,
J. G.
,
Gann
,
A. A. F.
and
Brockes
,
J. P.
(
1993
).
Chimeric retinoic acid/thyroid hormone receptors implicate RAR-α1 as mediating growth inhibition by retinoic acid
.
EMBO J
.
12
,
3459
3466
.
Schule
,
R.
,
Muller
,
M.
,
Kaltschmidt
,
C.
and
Renkawitz
R.
(
1988
).
Many transcription factors interact synergistically with steroid receptors
.
Science
242
,
1418
1420
.
Singer
,
M.
and
Salpeter
,
M. M.
(
1961
).
Regeneration in Vertebrates: The Role of the Wound Epithelium
, pp.
277
-
311
.
New York
:
Basic Books, Inc
.
Stocum
,
D. L.
(
1984
).
The urodele limb regeneration blastema. Determination and organization of the morphogenetic field
.
Differentiation
27
,
13
28
.
Stocum
,
D. L.
(
1985
).
The role of the skin in urodele limb regeneration
. In
Regulation of Vertebrate Limb Regeneration
(ed.
Sicard
,
R. E.
), pp.
32
53
.
New York
:
Oxford University Press
.
Stocum
,
D. L.
and
Crawford
,
K.
(
1987
).
Use of retinoids to analyse the cellular basis of positional memory in regenerating amphibian limbs
.
Biochem. Cell Biol
.
65
,
750
761
.
Summerbell
,
D.
(
1983
).
The effect of local application of retinoic acid to the anterior margin of the developing chick limb
.
J. Embryol. Exp. Morph
.
78
,
269
289
.
Tank
,
P. W.
and
Holder
,
N.
(
1981
).
Pattern regulation in the regenerating limb of urodele amphibians
.
Q. Rev. Biol
.
56
,
113
142
.
Tassava
,
R. A.
(
1992
).
Retinoic acid enhances monoclonal antibody WE3 reactivity in the regenerate epithelium of the adult newt
.
J. Morph
.
213
,
159169
.
Tassava
,
R. A.
,
Johnson-Wint
,
B.
and
Gross
,
J.
(
1986
).
Regenerate epithelium and skin glands of the adult newt react to the same monoclonal antibody
.
J. Exp. Zool
.
239
,
229
240
.
Thaller
,
C.
and
Eichele
,
G.
(
1987
).
Identification and spatial distribution of retinoids in the developing chick limb bud
.
Nature
327
,
625
628
.
Thaller
,
C.
,
Hofmann
,
C.
and
Eichele
,
G.
(
1993
).
9-cis retinoic acid, a potent inducer of digit pattern duplications in the chick wing bud
.
Development
118
,
957
965
.
Tickle
,
C.
,
Alberts
,
B.
,
Wolpert
,
L.
and
Lee
,
J.
(
1982
).
Local application of retinoic acid to the limb bud mimics the action of the polarising region
.
Nature
296
,
564
565
.
Wagner
,
M.
,
Han
,
B.
and
Jessell
,
T. M.
(
1992
).
Regional differences in retinoid release from embryonic neural tissue detected by an in vitro reporter assay
.
Development
116
,
55
66
.
Wagner
,
M.
,
Thaller
,
C.
,
Jessell
,
T. M.
and
Eichele
,
G.
(
1990
).
Polarizing activity and retinoid synthesis in the floor plate of the neural tube
.
Nature
345
,
819
822
.
Wallace
,
H.
(
1981
).
Vertebrate Limb Regeneration, Chichester: J. Wiley
.