* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Shh signalling and cell death in limb development
Survey
Document related concepts
Endomembrane system wikipedia , lookup
Signal transduction wikipedia , lookup
Tissue engineering wikipedia , lookup
Extracellular matrix wikipedia , lookup
Cytokinesis wikipedia , lookup
Cell encapsulation wikipedia , lookup
Cell growth wikipedia , lookup
Programmed cell death wikipedia , lookup
Cell culture wikipedia , lookup
Cellular differentiation wikipedia , lookup
Hedgehog signaling pathway wikipedia , lookup
Organ-on-a-chip wikipedia , lookup
Transcript
4811 Development 127, 4811-4823 (2000) Printed in Great Britain © The Company of Biologists Limited 2000 DEV1593 Autoregulation of Shh expression and Shh induction of cell death suggest a mechanism for modulating polarising activity during chick limb development Juan Jose Sanz-Ezquerro*,‡ and Cheryll Tickle* Department of Anatomy and Physiology, Wellcome Trust Biocentre, University of Dundee, Dow Street, Dundee DD1 5EH, UK *This work was initiated at the authors’ previous address: Department of Anatomy and Developmental Biology, University College London, Gower Street, London WC1 6BT, UK ‡Author for correspondence (e-mail: [email protected]) Accepted 17 August; published on WWW 24 October 2000 SUMMARY The polarising region expresses the signalling molecule sonic hedgehog (Shh), and is an embryonic signalling centre essential for outgrowth and patterning of the vertebrate limb. Previous work has suggested that there is a buffering mechanism that regulates polarising activity. Little is known about how the number of Shh-expressing cells is controlled but, paradoxically, the polarising region appears to overlap with the posterior necrotic zone, a region of programmed cell death. We have investigated how Shh expression and cell death respond when levels of polarising activity are altered, and show an autoregulatory effect of Shh on Shh expression and that Shh affects cell death in the posterior necrotic zone. When we increased Shh signalling, by grafting polarising region cells or applying Shh protein beads, this led to a reduction in the endogenous Shh domain and an increase in posterior cell death. In contrast, cells in other necrotic regions of the limb bud, including the interdigital areas, were rescued from death by Shh protein. Application of Shh protein to late limb buds also caused alterations in digit morphogenesis. When we reduced the number of Shh-expressing cells in the polarising region by surgery or drug-induced killing, this led to an expansion of the Shh domain and a decrease in the number of dead cells. Furthermore, direct prevention of cell death using a retroviral vector expressing Bcl2 led to an increase in Shh expression. Finally, we provide evidence that the fate of some of the Shh-expressing cells in the polarising region is to undergo apoptosis and contribute to the posterior necrotic zone during normal limb development. Taken together, these results show that there is a buffering system that regulates the number of Shhexpressing cells and thus polarising activity during limb development. They also suggest that cell death induced by Shh could be the cellular mechanism involved. Such an autoregulatory process based on cell death could represent a general way for regulating patterning signals in embryos. INTRODUCTION retinoic acid (Tickle et al., 1985), or Shh protein (Yang et al., 1997) applied: more cells or higher concentrations of retinoic acid or Shh leading to specification of more posterior digit identity. Rather curiously, however, when polarising region cells (Tickle et al., 1975), retinoic acid (Tickle et al., 1985) or Shh (Chang et al., 1994; Riddle et al., 1993; Yang et al., 1997) are added to the posterior margin of chick limbs, normally patterned digits are obtained. This points to a possible buffering mechanism that compensates for an excess of polarising signalling. Very little is known about mechanisms controlling either number of cells in the polarising region or the temporal extent of its activity. It has been suggested that there is a positive feedback loop between Shh expression in the polarising region and Fgf4 expression in the apical ridge that links outgrowth and patterning (Laufer et al., 1994; Niswander et al., 1994) and in Shh knockout mice, limb-bud outgrowth is impaired, leading to distal truncations and absence of digits (Chiang et al., 1996). Wnt7a, which is expressed in dorsal ectoderm, also contributes to maintenance of Shh expression. In chick limbs in which During outgrowth of embryonic limb buds, a group of posterior mesenchymal cells, known as the polarising region, produces signal(s) that pattern the anteroposterior axis of the developing limb. The polarising region was discovered by grafting small pieces of tissue from the posterior margin of one chick limb bud to the anterior margin of another limb bud (Saunders and Gasseling, 1968). This operation leads to a mirror image duplication of the digits. Retinoic acid was the first molecule shown to reproduce this effect (Tickle et al., 1982) and now is known to induce expression of sonic hedgehog (Shh) (Riddle et al., 1993). Cells of the polarising region express Shh, which is a secreted signalling molecule that can also produce duplications (Riddle et al., 1993) and has been proposed to mediate polarising region activity, probably via bone morphogenetic proteins (BMPs) (Drossopoulou et al., 2000). Polarising region signalling is dose dependent, since the identity of duplicated digits depends on number of polarising region cells transplanted (Tickle, 1981), or concentration of Key words: limb, apoptosis, Shh, bcl-2, cell number, chick embryo 4812 J. J. Sanz-Ezquerro and C. Tickle dorsal ectoderm has been removed (Yang and Niswander, 1995), and in limbs of mice in which Wnt7a is functionally inactivated (Parr and McMahon, 1995), Shh expression is reduced and this can lead, in both cases, to loss of posterior structures. The polarising region, paradoxically, is associated with a major area of programmed cell death in developing limbs: the posterior necrotic zone. Programmed cell death is a welldocumented feature of normal embryonic development (Glücksmann, 1951) and the existence of an evolutionarily conserved genetic programme that controls cell death is now well accepted (Raff, 1998). This essential physiological cell death is known to play several general roles during development and morphogenesis, including the control of cell number, particularly in the immune and nervous systems (reviewed in Jacobson et al., 1997; Vaux and Korsmeyer, 1999. Some examples of the importance of cell death in specific developmental processes have been reported, for example, cavitation of the early embryo (Coucouvanis and Martin, 1995), inner ear development (Fekete et al., 1997), neural tube closure (Weil et al., 1997), tooth development (Vaahtokari et al., 1996), and ductal morphogenesis and lumen formation in the mammary gland (Humphreys et al., 1996). In limb development, the best known example of programmed cell death occurs in the interdigital areas that are involved in separation of the digits (Pautou, 1975; Saunders and Fallon, 1967). Three other areas of well-defined massive cell death have also been described in early chick limb buds (Hinchliffe, 1982): one located in the anterior margin, the anterior necrotic zone; another in central core mesenchyme, the opaque patch; and a third in the posterior margin, the posterior necrotic zone. The role of these areas of cell death is not clear. It has been suggested that anterior and posterior necrotic zones might control the number of mesenchymal cells available to form digits and thus their prominence in chick limb buds would be related to the decreased number of digits compared with other vertebrates. However, the effects on the anterior and posterior necrotic zones when parts of the limb are removed are completely different (Hinchliffe and Gumpel-Pinot, 1981), suggesting that each zone might have a specific function and be differently regulated. Because the posterior necrotic zone is associated with the polarising region, we tested the relationship between cell death and Shh signalling from the polarising region. Our results lead us to propose that apoptosis might play a part in regulating the number of Shh-expressing cells during chick limb development. MATERIALS AND METHODS Embryos and surgical manipulations Fertilised White Leghorn chicken eggs were obtained from Needle farm (UK). They were incubated at 38°C for three days, and then windowed and embryos staged according to Hamburger and Hamilton (1951). Limb buds were exposed and microsurgery was carried out using fine watchmaker forceps and electrolytically sharpened tungsten needles. For bead implantation, a small cut was made with a needle at the desired position and a bead was introduced into the mesenchyme with the aid of the needle and forceps. Polarising tissue was removed by reference to maps of polarising activity (MacCabe et al., 1973) and patterns of Shh expression (Riddle et al., 1993). A cut was made in the flank just posterior to the base of the limb bud (running anteriorly through the proximal part of the bud), a loop was made in the posterior apical ectodermal ridge and another cut was made in the middle of the limb. A piece of tissue was then taken away with the aid of forceps, leaving the apical ridge intact. For polarising region or control anterior grafts, wing buds were dissected from stage 21 embryos in growth medium (GM; MEM with Hepes plus 10% foetal calf serum, 1% glutamine, 1% penicillin-streptomycin solution, all from Gibco). After trypsinisation in 10× trypsin (Gibco) for 30 minutes at 4°C, the ectoderm was removed and the polarising region or anterior tissue dissected. These blocks of tissue were transferred to a host embryo with a Gilson pipette and grafted under a loop of apical ridge lifted away from the posterior mesenchyme of the host wing bud. Some grafts were labelled with DiI by placing them in a solution of 100 µg/ml DiI in GM for 15 minutes at room temperature, before grafting. Beads Shh (a gift from A. M. McMahon; Marti et al., 1995) was stored in 14 mg/ml aliquots at –70°C. Dilutions were made in storage buffer as described (Yang et al., 1997). Fgf4 (R&D) was used at 0.75 mg/ml. Staurosporin (Sigma) was used at 100 µM diluted in GM from a 10 mM stock in DMSO. Affi Gel-CM Blue beads (BioRad, 150-250 µm diameter) were used for Shh and staurosporin, whereas heparin beads (Sigma) were used for Fgf4. To soak the beads, 2 µl of the solutions were placed in a bacterial Petri dish; beads were transferred with forceps and soaked at room temperature for at least 1 hour. Soaked beads were stored at 4°C and used within two weeks. DiI labelling DiI (Molecular Probes) was used at 3 mg/ml concentration in DMSO. A Picospritzer (General Valve Corporation, N.J.) was used to inject the solution in the posterior margin of wing buds using glass capillary pipettes. Embryos were collected and analysed for fluorescent labelling using a Leica MZ-FLIII microscope with a rhodamine filter and subsequently stained for cell death with Nile Blue. After washing in PBS, embryos were photographed and fixed in 4% paraformaldehyde (PFA) for subsequent in situ hybridisation. Nile Blue staining and TUNEL For Nile Blue staining, embryos were dissected in PBS and incubated in a 1/5000 solution of Nile Blue A (Sigma) in PBS for 15 minutes at 37°C in a rolling incubator. Embryos were transferred to cold PBS and washed for 15 minutes in PBS at 4°C, photographed and fixed in 4% PFA for subsequent in situ hybridisation. Some specimens that were not analysed by in situ hybridisation were left to wash in PBS overnight to improve staining and then photographed. In all cases the operated limb was compared with the contralateral, which served as a control, so that even small changes in cell death or Shh expression could be detected. Quantitative analysis of the data was made by counting the number of Nile Blue-positive cells in freshly stained embryos or in pictures of stained embryos, or by measuring the areas of the Shh-expression domains in pictures of in situ hybridisation processed embryos using NIH Image software. Statistical analysis of data was carried out using paired Student’s t-test comparing experimental with contralateral limbs. For TUNEL staining, embryos were fixed in 4% PFA, transferred to a series of sucrose solutions in PBS (5%, 15% and 30%), embedded in OCT compound and quick frozen in isopentane. 10 µm cryosections of limb buds were cut and subjected to TUNEL staining using the In Situ Fluorescein Cell Death Detection Kit from Boehringer, following the manufacturer’s instructions. Fluorescein labelling was observed using a Zeiss microscope. In situ hybridisation and cartilage staining In situ hybridisation was performed according to standard protocols (Nieto et al., 1996). Probes used (Shh (Cohn et al., 1995), Gdf5 Shh signalling and cell death in limb development 4813 (Merino et al., 1999a) and RCAS p27gag, from C. Tabin (Goff and Tabin, 1997) have been described elsewhere. Embryos were photographed using a standard camera attached to a Zeiss microscope or a digital camera attached to a Leica MZ-8 microscope and images analysed using Photoshop software. Alcian Green was used to stain cartilage in stage 36 embryos as described (Drossopoulou et al., 2000). Double labelling For double-labelling experiments, embryos were subjected to wholemount in situ hybridisation to measure Shh expression and developed with NBT/BCIP. Cryosections (10µm) of stained limbs were then subjected to TUNEL staining as described above. Retrovirus production and injection Retroviruses were produced according to standard procedures (Morgan and Fekete, 1996). Chicken embryo fibroblasts were grown in DMEM supplemented with 10% FBS, 2% chicken serum and 1% penicillin-streptomycin solution (all from Gibco). Cells were transfected with RCAS(BP)B-hbcl2 plasmid (a gift from Prof. S. Hughes), which contains the human BCL2 gene (Givol et al., 1994) and has previously been reported to inhibit cell death in chick embryos (Fekete et al., 1997), using Lipofectamine (Gibco). RCAS(BP)B virus, with no insert, was used as control. Cells were passaged for a week to amplify the virus. Supernatants from those cultures were collected and concentrated by ultracentrifugation. Viruses were injected using a Picospritzer at several spots in the prospective wing regions (Morgan and Fekete, 1996) of stage 10-12 or stage 17 chick embryos. Embryos were returned to the incubator and collected at different times after injection for analysis of cell death and for in situ hybridisation. Fig. 1. Shh expression and cell death following grafts of polarising region cells (A-H) or Shh beads (I-L) to the posterior margin. (A) Graft of polarising region cells to posterior margin of a stage 20 wing bud. Embryo collected immediately after the operation and subjected to in situ hybridisation with a Shh specific probe. Note Shh-expressing cells in graft (arrow), close to endogenous Shh expression domain in host. (B) Embryo showing Shh expression 24 hours after a polarising region graft. Note reduction of the proximal part of the Shh expression domain in right wing bud (arrow) and that the graft does not express Shh. (C) Same embryo stained with Nile Blue for cell death. Note position of graft containing cell death (arrowhead) and an increased amount of dead cells in the posterior necrotic zone of host (arrow) in the same area where reduction of Shh expression was observed. (D) Another example of polarising region graft showing persistence of cell death in the posterior margin (arrow). (E) Ventral view of limbs showing Shh expression 24 hours after grafting anterior cells to the posterior margin. Note proximal expansion of Shh domain in operated wing (arrow). (F) Nile Blue staining of same embryo. Note reduction of cell death in the operated wing as compared with the contralateral limb (arrow). (G,H) Same operated limb as in D. This graft was labelled with DiI before implantation to follow the position of grafted cells. As can be seen by comparing G (DiI labelling) with H (cell death), Nile Blue-positive cells in the posterior-most margin (arrows) are host cells. (I) Posterior Shh bead. In situ hybridisation with a Shh-specific probe. Note reduction in number of Shh-expressing cells. (J) Posterior control bead. No change in Shh expression. (K) Embryo showing Shh expression after implantation of a Shh bead. (L) Same embryo stained with Nile Blue, showing increase in cell death (arrow). Note that the area where extra cell death is induced (arrow in L) corresponds to the region of reduced Shh expression (arrow in K). RESULTS Effects of posterior polarising region grafts and Shh beads on Shh expression and cell death It has previously been described that increasing polarising signalling in chick limb buds, by placing a polarising region graft (Tickle et al., 1975) or extra Shh (Chang et al., 1994; Riddle et al., 1993; Yang et al., 1997) at the posterior margin, where the polarising region itself is located, does not have any effect on digit patterning. This suggests that Shh expression and/or signalling is regulative. To test this possibility, both grafts of polarising region cells and Shh-soaked beads were implanted in the posterior margin of stage 20 chick wing buds and the effects on Shh expression analysed. Additional polarising region cells from stage 21 embryos were grafted. Some embryos (n=2) were fixed immediately and in situ hybridisation confirmed that Shh-expressing cells had been transplanted (Fig. 1A). 80% of embryos (n=15) fixed at 20-24 hours showed a marked reduction in endogenous Shh expression (Fig. 1B, arrow) and no Shh expression could be seen in the graft. As controls, anterior cells were grafted to the posterior margin of limb buds. In most of these embryos, there was no significant decrease in Shh expression (75%, n=16) and, in some of them, Shh expression actually increased (Fig. 1E, arrow). To test the effects of Shh directly, beads soaked in the N-terminal active peptide of Shh protein were implanted posteriorly in stage 20 wing buds. 24 hours later, Shh expression was reduced in 92% of embryos (n=26) (Fig. 1I). 4814 J. J. Sanz-Ezquerro and C. Tickle Fewer cells expressed Shh and the most proximal part of the domain was missing in the majority of manipulated limb buds. Control beads had no effect on Shh expression (n=12) (Fig. 1J). To investigate whether this effect was stage dependent we implanted Shh beads posteriorly in later stage embryos (stage 22-23). A reduction in endogenous expression of Shh could again be seen in 75% of treated embryos (n=4). These results indicate that high levels of Shh can repress its own expression. The posterior necrotic zone, an area of massive cell death present during chick wing development, seems to co-localize with the polarising region, and in other systems cell death has been shown to control cell number. Therefore, we examined cell death after the above manipulations, which we had found repressed Shh expression. 20-24 hours after grafts of polarising region cells to the posterior margin, the number of Nile Bluepositive dead cells in the posterior necrotic zone of the host limb had clearly increased (40-300% more dead cells) in 50% of the embryos (n=22) (Fig. 1C, arrow). Often this increase reflected a distally expanded posterior necrotic zone (Fig. 1C arrow) and/or cell death persisting for longer (Fig. 1D, arrow). Furthermore, extensive cell death was often seen in the graft (Fig. 1C, arrowhead). Evidence that cell death was induced in the host was obtained by grafting DiI labelled cells (n=3). As can be seen by comparing Fig. 1G with 1H, the cell death present in Fig. 1H occurs in cells that are not labelled with DiI. Moreover, as could be seen in some embryos which were first stained for cell death with Nile Blue and subsequently fixed and processed for Shh expression, the area where extra cell death was induced corresponded to the area where a decrease in Shh expression was observed (compare Fig. 1B and 1C, arrows). In most control grafts of anterior cells there was no increase in cell death (93% of cases, n=28) and in many cases cell death was actually decreased (68% of embryos) (Fig. 1F). When these latter limbs were analysed for Shh expression we found that, in some of them, Shh expression was increased proximally (19% of cases) (Fig. 1E, arrow). Application of Shh beads to the posterior margin also led to an increase in the number of Nile Blue-positive dead cells in the posterior necrotic zone (Fig. 1L). In 15/17 of embryos, the area of induced extra cell death coincided precisely with the area in which Shh expression was reduced (compare Fig. 1L with 1K, arrows). The reduction in Shh expression could be detected 10-12 hours after bead implantation, the same time at which a slight increase of cell death could be seen. When control beads soaked in buffer were implanted, no change in either cell death or Shh expression could be seen in most embryos (8/10). The observed reduction in endogenous Shh expression after increasing Shh signalling is consistent with a buffering mechanism that regulates the amount of Shh expression, and the effects of Shh on apoptosis in posterior mesenchyme suggest a possible cellular mechanism (see below). Characterisation of the effects of Shh protein on limb-bud programmed cell death In most previous work, Shh has been reported to be a survival factor rather than a cell death inducer, therefore, we carried out another series of experiments to examine in detail the effects of Shh protein on limb-bud cell death. We first implanted Shh beads (14 mg/ml concentration) at the posterior margin of stage 20-22 chick wing buds and followed the effects on cell death with time (Tables 1, 2). In 81% (n=16) of embryos examined 18-24 hours later, there was a substantial increase in the number of Nile Blue-positive cells in the posterior necrotic zone (Fig. 2A, arrow). More Nile Blue-positive cells were seen in the normal region of cell death and/or the cell death domain extended more dorsally and distally. In two cases, development of the posterior necrotic zone was accelerated (Fig. 2E, arrowhead). An increase in cell death in the posterior necrotic zone was first seen in some embryos after 12 hours, peaked by 20 hours (Table 2) and persisted until 40 hours. By 48 hours, the posterior necrotic zone appeared have returned to normal (n=3). Shh beads also increased cell death in the posterior necrotic zone when implanted into older limb buds (stages 24/25) (Table 2). We then examined the effects of different doses of Shh. An increase in cell death in the posterior margin Table 1. Effects on cell death when 14 mg/ml Shh beads are implanted to stage 20-22 chick limb buds at different positions (Nile Blue-positive cells 18-24 hours after bead implantation) Bead (limb) Position of bead Shh (wing) Posterior Control (wing) Posterior Shh (wing) Central Shh (wing) Anterior Control (wing) Anterior Shh (leg) Shh (leg) Control (leg) Posterior Anterior Anterior Effects on cell death (percentage of embryos affected) PNZ n* ++ or +++ (81%‡) = or – (19%) +++ (0%) = or – (100%) +++ (33%) = (67%) +++ (25%) = (75%) = (100%) 16 – – – (100%) 5§ 9 = (100%) 7§ ++ (80%) n.p. n.p. 3§ 8 ANZ – – or – – – (100%¶) – – – (100%) 2 = (100%) 5 – – or – – – (100%) – – or – – – (100%) = (100%) n* 4 4§ 2 4§ 13 6 OP – – or – – – 100% = (100%) – – or – – – (100%) n* 16 9 3§ – – – (100%) 8 = (100%) 2 – – or – – – (100%) – – or – – – (100%) = (100%) Key: Number of Nile Blue-positive cells. +++, marked (100-700%) increase; ++, moderate (50-100%) increase; +, slight (10-50%) increase; =, <10% difference; –, slight (10-50%) reduction; – –, moderate (50-90%) reduction; – – –, absence. *n=number of embryos. ‡Two cases: posterior necrotic zone appeared precociously. §In rest of the embryos, cell death not yet present in that area. ¶Two of these beads were implanted in stage 23-24 embryos. ANZ, anterior necrotic zone; OP opaque patch; PNZ, posterior necrotic zone; n.p., cell death not present at that time in that position. 5 9§ 6 Shh signalling and cell death in limb development 4815 Table 2. Effects on cell death at different times after application of 14 mg/ml Shh beads at various stages Position of bead Stage Posterior 20-22 24-25 Anterior 20-22 24-25 Time after bead implantation (hours) 6-8 12 14 16 18 20 20-30 4 6 8 20-30 Effects on cell death (percentage of embryos affected) PNZ n* = (100%) + (14%) + (50%) + (75%) ++ (43%) +++ (100%) + (100%) 5 7 4 4 7 4 2 ANZ/OP n* = (100%) – (82%) – – or – – – (100%) – – or – – – (100%) 3 11 12 9 Key: Number of Nile Blue-positive cells. +++, marked (100-700%) increase; ++, moderate (50-100%) increase; +, slight (10-50%) increase; =, <10% difference; –, slight (10-50%) reduction; – –, moderate (50-90%) reduction; – – –, absence. *n=number of embryos. ANZ, anterior necrotic zone; OP opaque patch; PNZ, posterior necrotic zone. Fig. 2. Effects of Shh protein on cell death in chick limb buds. (A-E) Nile Blue staining of limb buds showing patterns of cell death 20-24 hours after Shh bead application to different positions at stage 20. Left: contralateral limbs. Right: operated limbs. (A) Posterior Shh bead in wing bud. Note increase in cell death posterodistally (arrow). (B) Anterior Shh bead in leg bud. Note absence of cell death in anterior necrotic zone as compared to the anterior necrotic zone in contralateral limb (arrow). (C) Posterior control bead. Normal appearance of the posterior necrotic zone. (D) Anterior control bead. No change in cell death. (E) Central Shh bead. Opaque patch clearly present in the control limb (arrow) but absent in operated limb bud. Note the precocious appearance of an intense area of cell death posteriorly in operated limb bud (arrowhead). (F) Quantitative analysis of Shh or control beads effects on cell death in limb buds. Graph shows mean number±standard deviation of Nile Bluepositive cells of operated versus contralateral limbs for the conditions described. *** P<0.001 in a paired Student’s t-test. (G-J) TUNEL staining of cryosections from operated limb buds (G,I control beads; H,J Shh beads) collected 24 hours after bead application. Dead cells show fluorescent labelling. (G) Posterior control bead in wing. Labelled cells represent the normal posterior necrotic zone present at that time (arrow). (H) Posterior Shh bead in wing. Note marked increase in number of TUNEL-positive cells postero-distally (arrow). (I) Anterior control bead in leg. Note intense anterior necrotic zone present at this time (arrow). (J) Anterior Shh bead in leg. Arrow indicates the area with a complete absence of cell death where a anterior necrotic zone is normally located. was also seen after application of beads soaked in 1mg/ml Shh (17/19 embryos) but when beads soaked in 0.1 mg/ml Shh were used, only one out of five embryos showed a slight increase in Nile Blue-positive cells in the posterior necrotic zone. With control beads, the number of dead cells in the posterior necrotic zone was either normal (4/9) or slightly reduced (5/9) (Fig. 2C). Shh beads placed posteriorly also affected cell death in the opaque patch and anterior necrotic zone (Table 1) (Fig. 2A). In these regions, Nile Blue-positive cells were either completely absent or their number much reduced in all cases. Thus, Shh rescued cell death. This contrasts to the increase seen in the posterior necrotic zone. To investigate the rescue effects of Shh on cell death in the opaque patch and anterior necrotic zone further, Shh beads were applied anteriorly or centrally. At 18-24 hours, cell death in the anterior necrotic zone and in the opaque patch was again completely absent or much reduced (Table 1) (Fig. 2B compare with control bead in Fig. 2D). In 3/11 cases, we could also detect an increase in the size of the posterior necrotic zone (Table 1). A slight decrease in cell death in the anterior necrotic zone and the opaque patch could be detected as early as six hours after bead implantation (Table 2) and substantial rescue was maintained in all embryos up to 48 hours after bead 4816 J. J. Sanz-Ezquerro and C. Tickle application. Shh also rescued cell death in older embryos (stage 24/25). Beads soaked in 1 mg/ml Shh could still prevent cell death in these areas (11/11), but when 0.1 mg/ml Shh beads were used, only a partial rescue was observed in the anterior necrotic zone in 67% of embryos. Similar bead implantations were carried out in chick leg buds. Again, Shh increased cell death in the posterior margin but rescued cell death in the anterior necrotic zone and the opaque patch (Table 1). A quantitative analysis of data obtained in experiments using beads soaked in 14 mg/ml Shh is shown in Fig. 2F, where number of Nile Blue-positive cells is compared between contralateral and operated limbs after Shh bead application. After posterior application of Shh beads, the number of Nile Blue-positive cells in the posterior necrotic zone of the operated wings significantly increased (117.0±40.7 in the operated wings versus 71.4±29.7 in the contralateral wings, P<0.001 as analysed by paired Student’s t-test). However, application of control beads posteriorly did not have such an effect (62.4±24.6 Nile Blue-positive cells in the operated limbs versus 68.0±27.9 in the contralateral). Anterior application of Shh-soaked beads led to a significant decrease in the number of Nile Blue-positive cells in the anterior necrotic zone and opaque patch of wings and legs (10.0±19.0 Nile Blue-positive cells in the operated limb versus 76.2±46.7 in the contralateral, P<0.001 according to a paired Student’s t-test). Control beads applied anteriorly did not alter significantly the number of Nile Blue-positive cells (76.1±25.0 in the operated limbs versus 78.1±31.0 in the contralateral). Although Nile Blue staining has been shown to label apoptotic cells specifically (Abrams et al., 1993) we confirmed the apoptotic nature of the cell death regulated by Shh using TUNEL labelling on sections of operated limbs. A marked increase in the number of TUNEL-positive cells in the posterior necrotic zone was seen in wing buds after implantation of Shh beads posteriorly (Fig. 2H), while with a control bead the posterior necrotic zone was normal (Fig. 2G). A complete absence of TUNEL-positive cells in anterior mesenchyme was also seen after anterior Shh bead implantation in leg buds (Fig. 2J), while with a control bead there was still extensive cell death in the anterior necrotic zone (Fig. 2I). Shh also rescued interdigital cell death when applied to later limb buds. When Shh-soaked beads (14 mg/ml) were applied to the third interdigital space of chick leg buds at stage 27-28 and interdigital cell death analysed by Nile Blue staining 48 hours later at stage 32, at a time when cell death is normally well established, Nile Blue-positive cells were absent in the operated third interdigital region and also much reduced in number in the second interdigital region, when compared with the contralateral limb (n=6) (Fig. 3A). The interdigital region also appeared wider with no signs of membrane regression. However, this rescue effect was transient, and by 66-72 hours cell death started to be observed again distally (n=7) (Fig. 3B). This later resumption of cell death, in the majority of cases, did not restore the normal regression of the interdigital membranes and led to soft tissue syndactyly with high frequency (63% of cases, n=16) (Fig. 3C). In addition to this syndactyly, we noticed that application of Shh at late stages had remarkable effects on digit morphogenesis (see also a recent report by Dahn and Fallon (2000), who independently observed similar effects). When Shh-soaked beads (14 mg/ml) were applied to the third interdigit at stage 27-28, in 87% of cases (13/15) digit 2, anterior and far away from the bead, was Fig. 3. Effects on cell death and digit morphology of application of Shh beads to late limbs. Shh-soaked beads (14 mg/ml) were applied to the third interdigital space of stage 27-28 leg buds. (A,B) Nile Blue staining to show cell death. Note rescue of cell death at 48 hours after bead implantation in A but reappearence of apoptosis at 72 hours in B (arrow). Note also widening of the operated interdigital space in A,B, and noticeable lengthening of digits 2 and 4 in operated leg in B (asterisks). Numbers in control limb denote digit identity, from anterior 1 to posterior 4. (C) Whole-mount legs showing syndactyly in the third interdigital area (arrow) four days after the operation. (D,E) Alcian Green staining of leg buds four days after bead application. Note elongation of digit 2 in D and of digit 2 and 4 in E, with the formation of a new phalanx (asterisks) and a new joint (arrows), and truncation of digit 3. Arrowheads mark the position of the bead. Numbers denote digit identity as in B. (F) In situ hybridisation with the joint-specific marker Gdf5 to show the extra joints (arrows). A-C operated limbs to the left. D-F operated limbs to the right. Shh signalling and cell death in limb development 4817 longer (Fig. 3D,E), owing either to an increase in length of the penultimate phalangeal element or to the formation of an extra phalanx with a new joint, as confirmed by Gdf5 expression (Fig. 3F), thus giving a digit with four phalanges instead of the normal three. In most cases (80%), digit 3, close and anterior to the beads, was shorter and truncated, with a reduced number of phalangeal elements (Fig. 3D,E) while digit 4, close and posterior to the beads was not affected in the majority of cases (60%). In one case, digit 4 was elongated and had an extra phalanx and joint (Fig. 3E). The above results show that Shh has opposite effects on apoptosis in the different regions of programmed cell death in chick limb buds: Shh rescues cells in the anterior necrotic zone, opaque patch and interdigital areas, but increases cell death in the posterior necrotic zone. Relationship between cell death in the posterior necrotic zone and Shh expression in the polarising region after experimentally manipulating the number of Shh-expressing cells The results presented above show that addition of polarising region cells expressing Shh or implantation of Shh beads to the posterior margin leads to a decrease in endogenous Shh expression and that this decrease is accompanied by an increase in cell death. This suggests the possibility that the buffering mechanism that keeps Shh signalling constant is cell death. To test this idea further, we carried out additional experiments in which we either increased or decreased the number of Shh-expressing cells at the posterior margin of the limb and looked at the effects on Shh expression and cell death. To increase the number of Shh-expressing cells, we applied Fgf4-soaked beads, an operation that has been shown to increase the extent of the Shh expression domain (Yang and Niswander, 1995). Indeed, at 14 hours (1/1 cases), 18 hours (4/4 cases) or 24 hours (5/7 cases) after application of Fgf4-soaked beads to the posterior margin of stage 20 wing buds, the Shh expression domain expanded proximally along the posterior margin (Fig. 4A, arrow). However, by 28-40 hours, Shh expression was now found to be restricted distally and resembled the normal pattern of expression seen in the contralateral limb bud (Fig. 4C). These changes in Shh expression were accompanied by changes in the extent of cell death. At 18-24 hours, the number of Nile Blue-positive cells had decreased (2/2 cases at 18 hours, 3/5 cases at 24 hours) (Fig. 4B) but at 28 or 40 hours after Fgf4 bead application, cell death had increased in all embryos (4/4 at 28 hours; 8/8 at 38-42 hours) (Fig. 4D). These areas of extra cell death in the proximal part of the bud corresponded to regions where ectopic Shh expression had previously been seen. Indeed, analysis in the same embryos (2/2 at 28 hours; 8/8 at 38-42 hours) showed that areas of cell death and of Shh expression were now almost mutually exclusive, except for a small overlap between the distal-most area of cell death and the proximal-most area of Shh expression (compare Fig. 4C with 4D, arrows). Fig. 4. Effects on Shh expression and cell death after increasing or reducing the number of Shh-expressing cells in the polarising region. (A-D) implantation of Fgf4 beads. (A) Embryo showing Shh expression 18 hours after a Fgf4 bead was implanted to the posterior margin of a stage 21 wing bud. Note proximal expansion of Shh expression domain in operated limb bud (arrow). (B) Same embryo stained with Nile Blue. Note reduction in cell death seen in the posterior necrotic zone as compared with the contralateral wing (arrow). (C) Embryo processed for in situ hybridisation to reveal Shh expression 40 hours after Fgf4 bead implantation. Shh domain is now restricted distally in right limb bud and resembles normal expression found in contralateral limb. (D) Same embryo stained with Nile Blue. Note marked increase in cell death along posterior margin of right wing bud as compared with normal posterior necrotic zone in contralateral limb. Note also that the area undergoing cell death does not express Shh except for a small overlap (arrows in C,D). (E-J) Effects of reducing number of polarising region cells. (E-G) Surgical removal of Shh-expressing cells. (E) Embryo fixed immediately after the operation and subjected to in situ hybridisation to reveal Shh expression. Note that right experimental wing has a reduced number of Shh-expressing cells (arrow). (F,G) Embryo collected 24 hours after surgery. F shows Shh expression and G shows Nile Blue staining. Note that Shh expression is proximally expanded (arrow in F), while cell death is absent in operated right limb as compared with the contralateral left limb (arrow in G). (H-J) Staurosporin treatment. H shows embryo 4 hours after a staurosporin bead was inserted in posterior margin of a stage 20 wing bud. In situ hybridisation shows reduction in Shh expression (arrow). (I,J) Embryo collected 24 hours after treatment. I shows Shh expression and J shows cell death in the same embryo. Note increase in Shh expression in operated limb with some cells expressing Shh more proximally than in control (arrow in F) and reduction in cell death in operated limb as compared with normal posterior necrotic zone present in contralateral limb (arrow in E). 4818 J. J. Sanz-Ezquerro and C. Tickle To reduce the number of Shh-expressing cells, a small area of the posterior margin of stage 20-21 chick wing buds was surgically removed, leaving the apical ectodermal ridge in place. We confirmed that part of the Shh domain had been excised by fixing some embryos just after the operation. As shown in Fig. 4E, the number of Shh-expressing cells was reduced after this operation (5/5 of embryos). However, 24 hours later, cells expressing Shh were present in the posterior margin of the limb in all embryos and in many of these (50%, n=24) there was a proximal expansion of the Shh expression domain, when compared with the contralateral limbs (Fig. 4F, arrow). In most other embryos (48%), the size and shape of the Shh expression domain was similar to that in the contralateral limb. In just one case was the number of Shh-expressing cells reduced. Analysis of cell death in the same embryos showed a variable degree of reduction in number of dead cells in the posterior necrotic zone in 96% of embryos (n=24). Cell death was completely absent in most cases (75%, n=24) (Fig. 4G). This decrease in cell death coincided with the observed expanded domain of Shh expression. In another set of experiments, some cells expressing Shh at stage 20 were killed by applying beads soaked in 100 µM staurosporin (a protein kinase inhibitor known to induce apoptosis (Jacobson et al., 1996). 4 hours after staurosporin application, a reduction in number of Shh-expressing cells can be detected (n=2) (Fig. 4H). However, at 22 hours, in all cases the domain of Shh expression was restored, and in 2/3 cases some Shh-expressing cells were found more proximally than in the contralateral limb (Fig. 4I). Moreover, a clear reduction in the number of Nile Blue-positive cells in the posterior margin of the limb buds could be seen (3/3) (Fig. 4J). All these results obtained after experimental manipulation of the number of Shh-expressing cells indicate a regulatory ability of the Shh expression domain and show that this is related to cell death. There is a consistent inverse correlation between the extent of Shh expression domain and cell death in the posterior necrotic zone and altering the amount of Shh signalling can affect both these parameters. When Shh signalling is increased, i.e. by grafting polarising region cells or applying Shh beads, or after Fgf-4 bead application, cell death is increased and Shh expression reduced; when Shh signalling is reduced, i.e. by surgical removal or chemical killing of polarising region cells, cell death is reduced and Shh expression expanded. This suggests a buffering model in which any imbalance in the number of Shh-expressing cells will be sensed and adjusted through appropriate regulation of cell death. Effects of blocking cell death on the expression domain of Shh Our model predicts that blocking cell death will lead to an expansion of the Shh expression domain. We tested this hypothesis by directly preventing cell death. We used the RCAS retroviral vector to overexpress ectopically the proto-oncogene Bcl2, the product of which is a protein known to be able to block cell death in the chicken embryo (Fekete et al., 1997). RCAS-Bcl2 virus was injected into the right side lateral plate mesoderm (prospective right limb region) of chicken embryos at stages 9-12 or stage 17. Analysis of some embryos at 48 hours (n=7), 72-80 hours (n=7), 4 days (n=2) or 6 days (n=1) after injection by in situ hybridisation with a viral-specific probe confirmed spread of infection. A variable degree of infection was detected in the targeted right limbs and an example of extensive infection of the right wing is shown in Fig. 5A. Infected embryos were analysed at stage 23-25 for both cell death in the posterior necrotic zone and Shh expression. In 48% of cases (n=25), the right wing had a 10% or more reduction in the number of Nile Blue-positive cells in the posterior necrotic zone (Fig. 5B). In one case, no dead cells could be detected in the right wing bud, while in the contralateral left wing bud a posterior necrotic zone was already established (Fig. 5E). Most other embryos (11/25) showed less than a 10% difference in the number of Nile Blue-positive cells. Quantitation of these results showed that the decrease in the Fig. 5. Effects of Bcl2 expression on cell death and Shh expression. (A) In situ hybridisation with a viral message-specific probe. Embryo fixed 80 hours after RCAS-Bcl2 virus injection to right lateral plate mesoderm at stage 10. Note extensive infection of right wing bud (arrow). (B,D) Nile blue staining (B) and Shh + Fgf8 expression (D) in same embryo 72 hours after RCAS-Bcl2 injection. Note reduced number of Nile Blue-positive cells in right wing bud (arrow in B) and proximally expanded Shh expression domain (arrow in D). (C) Another example of expansion in Shh expression after RCAS-Bcl2 injection (arrow). (E) Another example of decreased cell death after RCAS-Bcl2 injection. Note absence of dead cells in targeted right wing (arrow). (F,G) Viral expression (F) and Nile Blue staining (G) of same wings, five days after injection with RCAS-Bcl2 virus of the nascent right wing bud at stage 18-19. Note extensive infection of right wing, including interdigit (arrow in F). Cell death is absent in that area, as can be seen by absence of Nile Blue-positive cells in the right wing interdigit (arrow in G) as compared to the left wing (arrowhead). Shh signalling and cell death in limb development 4819 number of Nile Blue-positive cells in the infected versus contralateral wings was significant (average number of cells 69.5±33.7 in the infected, 79.9±35.9 in the contralateral, P<0.05 in a paired Student’s t-test). In most cases (9/12), reduction in cell death was accompanied by an increase in number of Shh-expressing cells and/or extension of the Shh expression domain proximally (Fig. 5C,D). Quantitation of the Shh expression domains showed a significant increase of the Shh-expressing areas in infected wings versus control wings (average percentage increase 116.5±11.8 of control limbs, P<0.01 in a Student’s t-test). Thus, when cell death is blocked, this leads to an expansion of Shh expression, which is consistent with the model that cell death regulates the number of Shh-expressing cells. Controls were carried out by infecting stage 9-12 embryos in a similar way with a control virus (RCAS with no insert). Infection of the right wing was confirmed 72 hours after injection with an RCAS-specific probe, showing spread of infection (n=6). When cell death and Shh expression were analysed, no significant differences between the infected wings and the contralateral ones were observed either in the number of Nile Blue-positive cells (79.1±40.6% in the right wings versus 76.2±38.0% in the control) or the area of the Shh expression domains (mean=99.9±0.9% of the control). Functional activity of the delivered Bcl2 gene in infected limbs was confirmed by the decreased number of Nile Bluepositive cells in the anterior necrotic zone in 7/14 cases and in the interdigital region in 4/5 cases (compare viral message expression in Fig. 5F with cell death in Fig. 5G in the same embryo). Some embryos (n=50) were left to develop until day 10 and several defects were observed; delay in regression of interdigital membranes leading to partial soft tissue syndactyly (n=4), blips of tissue in the distal part of digits/toes (n=3), abnormal flexures at some joints leading to ‘straight’ limbs (n=3) and, most remarkably, nodules of ectopic cartilage associated with some long bones (radius/tibia) (n=2). All these defects point to a possible role of cell death in other processes of limb development such as joint formation and cartilage morphogenesis. Relationship between domain of Shh-expression and the posterior necrotic zone in normal limb development The polarising region was discovered during investigations of the properties of the posterior necrotic zone (Saunders and Gasseling, 1968) and expression of Shh at the posterior margin of the wing bud correlates with maps of polarising activity. Shh expression is initiated at Hamburger and Hamilton (1951) stage 17-18 in posterior mesenchyme of chick wing buds, maintained along the posterior bud margin as the bud grows out and then becomes restricted distally at stage 24, where it remains until stage 28 (Riddle et al., 1993). The posterior necrotic zone, on the other hand, first appears as a discrete patch of dead cells midway along the posterior margin of chick wing buds at stage 23; by stage 24, a massive area of cell death extends along the proximal two thirds of the posterior margin and then cell death continues in the distal part of the limb until stage 29-30 (Hinchliffe, 1981). The results from our experimental manipulations pointed to a direct link between Shh-expressing cells in the polarising region and cell death in the posterior necrotic zone. To visualise directly whether Shh expression and the posterior necrotic zone overlap in normal limb development, we performed sequential double labelling of stage 24 chick embryos. As can be seen in Fig. 6A, the posterior necrotic zone, a group of cells labelled with Nile Blue, extends along the proximal 2/3 of the posterior margin of the wing bud. In the same wing bud (Fig. 6B), Shh expression extends proximally from the distal tip, next to the posterior boundary of the apical ectodermal ridge, up to almost half way along the wing bud. Thus, the proximal part of the Shh expression domain clearly overlaps with the distal part of the posterior necrotic zone, as can be seen in Fig. 6C where a superimposed image of Fig. 6A,B is shown. To trace the cell lineage relationship between Shhexpressing cells and the posterior necrotic zone we injected DiI into the posterior margin of stage 20 wing buds, at somite levels 19-20, in a subapical position (Vargesson et al., 1997; Fig. 6F). Some embryos were fixed immediately and in situ hybridisation confirmed that labelled cells were indeed in the Shh-expressing domain (n=7) (Fig. 6G). Other embryos were collected at 24 hours and, by this time, some DiI-labelled cells had died and were present in the posterior necrotic zone as confirmed by Nile Blue staining (n=11) (Fig. 6H,I, arrows); Fig. 6J is a merged image showing double labelled DiI- and Nile Blue-positive cells (arrows). Other labelled cells were found proximal to the posterior necrotic zone (Fig. 6H,I, arrowheads). When embryos were collected 48 hours after labelling (n=7), Nile Blue-positive cells were still found in the posterior necrotic zone. The same result was obtained when a limb bud was labelled posterodistally at stage 22-23 and analysed at stage 25-26. Finally, to investigate at the single cell level whether Shhexpressing cells undergo apoptosis in the posterior necrotic zone, we cut cryosections of limbs that had undergone wholemount in situ hybridisation to visualise Shh expression and carried out TUNEL staining to reveal apoptosis. As can be seen by comparing Fig. 6D and 6E (arrows), TUNEL-positive cells are present in the Shh expression domain. Furthermore, we observed some individual cells that expressed Shh (Fig. 6L,O) and were also TUNEL positive (Fig. 6K,N, merged image in Fig. 6M,P), thus providing direct evidence that Shh-expressing cells die in the posterior necrotic zone. DISCUSSION Buffering model for regulating signalling in the limb Our results suggest a model for the regulation of polarising activity in chick limb development (Fig. 7). We have found that Shh can repress its own expression and lead to increased cell death, and both experimental manipulations and in vivo evidence suggest that there is a link between the two. We propose that cell death induced by Shh could act as a buffer to regulate the number of Shh-expressing cells. In normal limb development, there will be a balance between generation of new cells expressing Shh, under the influence of Fgf4 and other growth factors produced distally, and loss of cells expressing Shh proximally. This will keep the Shh signal at the appropriate level and restrict the Shh domain distally close to the apical ridge. The model is compatible with previous results showing that the removal of posterior apical ectodermal ridge, which 4820 J. J. Sanz-Ezquerro and C. Tickle Fig. 6. Relationship between cell death in the posterior necrotic zone and Shh expression domain. (A) Stage 24 chick wing bud stained with Nile Blue to reveal dead cells. Arrow indicates the distal part of the posterior necrotic zone. (B) Shh expression as revealed by in situ hybridisation in the same limb. Note the distal restriction of expression and the overlap between the fainter proximal area of Shh expression with the distal part of the cell death domain seen in A (arrow placed in the same position as in A). (C) Superimposed image of A and B showing the overlap. (D,E) Cryosections of a stage 24 wing bud firstly stained for Shh expression by whole-mount in situ hybridisation (E) and subsequently with TUNEL to reveal apoptosis (D). Note apoptotic cells (arrows) in the Shh expression domain. (F-J) Fate map of polarising region cells. (F) Stage 20 wing bud injected with DiI (fluorescent labelling) in the posterior margin (arrow). (G) Same bud showing Shh expression. Note that DiI labelled cells (arrow) are in the Shh expression domain. (H) Wing bud 24 hours after DiI injection. Note labelled cells in medial part of posterior margin (arrow) and more proximally, close to base of limb bud (arrowhead). (I) Same limb stained with Nile Blue. Note that medial labelled cells are in posterior necrotic zone (arrow) and that the more proximal labelled cells in H have passed through the necrotic zone (arrowhead). (J) High resolution merged image of H and I showing DiI labelled cells that are Nile Blue-positive (arrows). (K-P) High-magnification images of limb cryosections double stained for both Shh expression by whole-mount in situ hybridisation (L,O) and apoptosis by TUNEL (K,N). Note double labelled cells in the polarising region (arrows, merged images in M,P). Posterior is at the bottom, distal towards the right in all limbs. leads to a decrease in Shh expression (Laufer et al., 1994; Niswander et al., 1994), also results in reduction in cell death in the posterior necrotic zone (Brewton and MacCabe, 1988). A buffering system can account for several regulative features of polarising region signalling. It can explain why limbs with normal patterns develop after application of extra Shh (polarising region cells (Tickle et al., 1975), Shhexpressing cells (Riddle et al., 1993) or Shh beads (Yang et al., 1997) to the posterior margin of chick buds, and why normal patterned limbs also develop after most, but not all, of the polarising region is removed (Fallon and Crosby, 1975; Pagan et al., 1996). It is important to note that this buffering mechanism seems to operate only in posterior cells competent to express Shh and/or that have been primed by posterior polarising signals (supported by the finding that Shh expands the normal posterior necrotic zone but it does not induce cell death in other surrounding cells). This is why dose-dependent effects of Shh signalling on digit pattern can be observed anteriorly where no buffer exists and the effect of a given number of polarising region cells can be analysed (Tickle, 1981). Our model implies a positive-feedback loop between Shh and Fgf4 and/or other Fgfs distally to maintain the population of Shh-expressing cells. This is consistent with the finding that Fgf4 application results in an initial increase in Shh expression. The fact that this initial increase of Shh expression ultimately leads to extra cell death and reduction of Shh fits well with an autoregulatory buffering system. The model also implies a negative feedback loop proximally, away from the influence of the ridge, driven by Shh itself and involving apoptosis, to eliminate Shh-expressing cells. It is not clear how this negativefeedback loop could operate. One possibility is that apoptosis is induced in cells when the Shh signal in posterior mesenchyme limb bud reaches a certain threshold. Polarising region cells expressing Shh may also be particularly dependent on the presence of survival ridge signals and/or be especially sensitive to apoptotic stimuli, so that they initiate the cell death programme once they have moved away from the apical Shh signalling and cell death in limb development 4821 death in the developing limb can explain the absence of necrotic zones in limbs of talpid3, a polydactylous chicken mutant. In talpid3, a defect in the Shh signalling pathway, involving inability to induce high levels of patched expression, has been postulated to lead to a widespread diffusion of Shh protein (Lewis et al., 1999). Thus, ectopic Shh protein would be able to rescue cell death anteriorly and in the interdigital spaces, while failure to localise high concentrations of Shh protein posteriorly would also reduce apoptosis in this area. AER apoptosis growth and patterning Shh Fgfs OUT Shh expressing cells IN Fig. 7. Signal buffering model. Shh-expressing cells are added distally under influence of Fgfs produced in ridge. It should be noted that other growth factors produced distally are probably involved in the input. Some cells are left behind as more cells are added distally and are lost proximally through apoptosis, which is mediated by Shh either directly or through an indirect mechanism. Fluctuations in the number of Shh-expressing cells will be compensated by regulation of cell death. AER, apical ectodermal ridge. ectodermal ridge (AER). Alternatively, it can not be excluded that cells of the polarising region first reduce Shh expression as they become distant from the AER and only subsequently initiate apoptosis. BMPs, which have been shown to act as apoptotic signals, might mediate Shh-induced cell death and expression of BMPs and/or BMPs inhibitors, which has been shown to be regulated by Shh, could modulate the precise amount of cell death. Indeed, application of gremlin, an inhibitor of BMPs, to developing limb buds has been reported to lead to an expansion of the Shh expression domain (Capdevila et al., 1999; Merino et al., 1999b). This is postulated to be due to an increase in Fgf4 signalling, but apoptosis could be reduced and could also contribute to the observed increase in Shh expression. Different effects of Shh on cell death Although Shh acts as a cell death signal in posterior limb bud cells, Shh rescues anterior and interdigital limb cells from death. Shh is better known as a survival factor and Shh has also been shown to rescue cell death in other areas of the embryo such as somites (Teillet et al., 1998) and neurones (Miao et al., 1997). This highlights the fact that different cell populations can respond in a different way to the same signal. Our time and dose results suggest that the survival effect of Shh on anterior cells may be direct. The opposite effects of Shh on the different regions of cell Roles of cell death in signalling regions Cell death has been shown to play an important developmental role in the control of cell number, which is essential for proper organ size and tissue homeostasis (reviewed in Green (1998); Jacobson et al. (1997); Vaux and Korsmeyer (1999)). Here, we suggest that apoptosis may play a role in the regulation of the number of signalling cells in the polarising region of the limb, contributing to modulate its activity and extent. Regulation of the number of Shh-expressing cells in limb development could be important in the dose-dependent patterning properties of polarising region signalling. It could also be important in setting the signal level that gives the appropriate number of digits. In chick wing development, the posterior necrotic zone is a prominent feature, while in limbs with more digits (chick leg buds and mouse limbs) the posterior necrotic zone is still present but much reduced (Hinchliffe, 1981, 1982; Milaire and Rooze, 1983). In chick legs, the smaller necrotic zone is associated with a more-extensive Shh expression domain. There could be different thresholds of Shh that induce apoptosis, or the balance between growth and apoptotic signals controlled by Shh could be different in the different limbs, which would ensure the appropriate amount of Shh signalling to achieve the characteristic final number of digits and limb pattern. The ability to control polarising activity could also be an important safety mechanism to avoid excessive signal levels. Finally, apoptosis has been shown to mediate silencing of the enamel knot, which acts as a signalling centre in tooth development (Vaahtokari et al., 1996) and thus cell death might be also involved in the silencing of Shh expression, when outgrowth and patterning of the limb are completed. In this context, our finding that Shh application to late limb buds, when the endogenous signal has been switched off, can make digits grow longer, sometimes forming new joints and thus more phalangeal elements, could be very important in understanding how the final steps of digit formation are achieved. We have preliminary evidence that implicates both Fgf and Bmp signalling in this process (J. J. S.-E., unpublished). We suggest that autoregulation of signal production via cell death might be a general developmental mechanism. Dosedependent Shh signalling has been demonstrated in the neural tube (Ericson et al., 1996) and thus, according to our ideas, apoptosis might regulate the number of Shh-expressing cells here. Indeed, it has been reported recently that Shh induces cell death in the floor plate, which is the region of the neural tube that produces Shh (Oppenheim et al., 1999). It is also possible that this mechanism could act in other regions to regulate the levels of other signals, such as Fgfs. These are produced in the AER, where cell death also occurs (Ferrari et al., 1998; Todt and Fallon, 1984). 4822 J. J. Sanz-Ezquerro and C. Tickle We thank A.M. McMahon for Shh protein and S.H. Hughes for the RCAS-Bcl2 construct; Juan Hurle for the Gdf5 probe; and Litsa Drossopoulou for her advice on bead implantation. This research was supported by a long term EMBO postdoctoral fellowship to J. J. S.E. and a MRC programme grant to C. T. REFERENCES Abrams, J. M., White, K., Fessler, L. I. and Steller, H. (1993). Programmed cell death during Drosophila embryogenesis. Development 117, 29-43. Brewton, R. G. and MacCabe, J. A. (1988). Ectodermal influence on physiological cell death in the posterior necrotic zone of the chick wing bud. Dev. Biol. 126, 327-330. Capdevila, J., Tsukui, T., Rodriguez-Esteban, C., Zappavigna, V. and Izpisua-Belmonte, J. C. (1999). Control of vertebrate limb outgrowth by the proximal factor Meis2 and distal antagonism of BMPs by Gremlin. Mol. Cell 4, 839-849. Chang, D. T., Lopez, A., von Kessler, D. P., Chiang, C., Simandl, B. K., Zhao, R., Seldin, M. F., Fallon, J. F. and Beachy, P. A. (1994). Products, genetic linkage and limb patterning activity of a murine hedgehog gene. Development 120, 3339-3353. Chiang, C., Litingtung, Y., Lee, E., Young, K. E., Corden, J. L., Westphal, H. and Beachy, P. A. (1996). Cyclopia and defective axial patterning in mice lacking Sonic hedgehog gene function. Nature 383, 407-413. Cohn, M. J., Izpisua-Belmonte, J. C., Abud, H., Heath, J. K. and Tickle, C. (1995). Fibroblast growth factors induce additional limb development from the flank of chick embryos. Cell 80, 739-746. Coucouvanis, E. and Martin, G. R. (1995). Signals for death and survival: a two-step mechanism for cavitation in the vertebrate embryo. Cell 83, 279287. Dahn, R. D. and Fallon, J. F. (2000). Interdigital regulation of digit identity and homeotic transformation by modulated BMP signaling. Science 289, 438-441. Drossopoulou, G., Lewis, K. E., Sanz-Ezquerro, J. J., Nikbakht, N., McMahon, A. P., Hofmann, C. and Tickle, C. (2000). A model for anteroposterior patterning of the vertebrate limb based on sequential longand short-range Shh signalling and Bmp signalling. Development 127, 13371348. Ericson, J., Morton, S., Kawakami, A., Roelink, H. and Jessell, T. M. (1996). Two critical periods of Sonic Hedgehog signaling required for the specification of motor neuron identity. Cell 87, 661-673. Fallon, J. F. and Crosby, G. M. (1975). Normal development of the chick wing following removal of the polarizing zone. J. Exp. Zool. 193, 449455. Fekete, D. M., Homburger, S. A., Waring, M. T., Riedl, A. E. and Garcia, L. F. (1997). Involvement of programmed cell death in morphogenesis of the vertebrate inner ear. Development 124, 2451-2461. Ferrari, D., Lichtler, A. C., Pan, Z., Dealy, C. N., Upholt, W. B. and Kosher, R. A. (1998). Ectopic expression of Msx-2 in posterior limb bud mesoderm impairs limb morphogenesis while inducing BMP-4 expression, inhibiting cell proliferation and promoting apoptosis. Dev. Biol. 197, 12-24. Givol, I., Tsarfaty, I., Resau, J., Rulong, S., Silva, P. P. d., Nasioulas, G., DuHadaway, J., Hughes, S. H. and Ewert, D. (1994). Bcl-2 expressed using a retroviral vector is localised primarily in the nuclear membrane and the endoplasmic reticulum of chicken embryo fibroblasts. Cell Growth Differ. 5, 419-429. Glücksmann, A. (1951). Cell deaths in normal vertebrate ontogeny. Biol. Rev. Camb. Philos. Soc. 26, 59-86. Goff, D. J. and Tabin, C. J. (1997). Analysis of Hoxd-13 and Hoxd-11 misexpression in chick limb buds reveals that Hox genes affect both bone condensation and growth. Development 124, 627-636. Green, D. R. (1998). Apoptotic pathways: the roads to ruin. Cell 94, 695-698. Hamburger, V. and Hamilton, H. (1951). A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49-92. Hinchliffe, J. R. (1981). Cell death in embryogenesis. In Cell death in biology and pathology, (ed. I. D. Bowen and R. A. Lockshin), pp. 35-78. London: Chapman & Hall. Hinchliffe, J. R. (1982). Cell death in vertebrate limb morphogenesis. In Progress in Anatomy. Vol. 2 (ed. R. J. Harrison and V. Navaratman), pp. 119. Cambridge: Cambridge Universisty Press. Hinchliffe, J. R. and Gumpel-Pinot, M. (1981). Control of maintenance and anteroposterior skeletal differentiation of the anterior mesenchyme of the chick wing bud by its posterior margin (the ZPA). J. Embryol. Exp. Morphol. 62, 63-82. Humphreys, R. C., Krajewska, M., Krnacik, S., Jaeger, R., Weiher, H., Krajewski, S., Reed, J. C. and Rosen, J. M. (1996). Apoptosis in the terminal endbud of the murine mammary gland: a mechanism of ductal morphogenesis. Development 122, 4013-4022. Jacobson, M. D., Weil, M. and Raff, M. C. (1996). Role of Ced-3/ICE-family proteases in staurosporine-induced programmed cell death. J. Cell Biol. 133, 1041-1051. Jacobson, M. D., Weil, M. and Raff, M. C. (1997). Programmed cell death in animal development. Cell 88, 347-354. Laufer, E., Nelson, C. E., Johnson, R. L., Morgan, B. A. and Tabin, C. (1994). Sonic hedgehog and Fgf-4 act through a signaling cascade and feedback loop to integrate growth and patterning of the developing limb bud. Cell 79, 993-1003. Lewis, K. E., Drossopoulou, G., Paton, I. R., Morrice, D. R., Robertson, K. E., Burt, D. W., Ingham, P. W. and Tickle, C. (1999). Expression of ptc and gli genes in talpid3 suggests bifurcation in Shh pathway. Development 126, 2397-2407. MacCabe, A. B., Gasseling, M. T. and Saunders, J. W., Jr (1973). Spatiotemporal distribution of mechanisms that control outgrowth and anteroposterior polarization of the limb bud in the chick embryo. Mech. Ageing Dev. 2, 1-12. Marti, E., Bumcrot, D., Takada, R. and McMahon, A. P. (1995). Requirement of 19K form of Sonic hedgehog for induction of distinct ventral cell types in CNS explants. Nature 375, 322-325. Merino, R., Macias, D., Ganan, Y., Economides, A. N., Wang, X., Wu, Q., Stahl, N., Sampath, K. T., Varona, P. and Hurle, J. M. (1999a). Expression and function of Gdf-5 during digit skeletogenesis in the embryonic chick leg bud. Dev. Biol. 206, 33-45. Merino, R., Rodriguez-Leon, J., Macias, D., Ganan, Y., Economides, A. N. and Hurle, J. M. (1999b). The BMP antagonist Gremlin regulates outgrowth, chondrogenesis and programmed cell death in the developing limb. Development 126, 5515-5522. Miao, M., Wang, M., Ott, J. A., D’Alessandro, J. S., Woolf, T. M., Bumcrot, D. A., Mahanthappa, N. K. and Pang, K. (1997). Sonic Hedgehog promotes the survival of specific CNS neuron populations and protects these cells from toxic insult in vitro. J. Neurosci. 17, 5891-5899. Milaire, J. and Rooze, M. (1983). Hereditary and induced modifications of the normal necrotic patterns in the developing limb buds of the rat and mouse: facts and hypotheses. Arch. Biol. (Bruxelles) 94, 459-490. Morgan, B. A. and Fekete, D. M. (1996). Manipulating gene expression with replication-competent retroviruses. In Methods in Cell Biology. Vol. 51, pp. 185-218. San Diego: Academic Press. Nieto, M. A., Patel, K. and Wilkinson, D. G. (1996). In situ hybridization analysis of chick embryos in whole mount and tissue sections. In Methods in Cell Biology. Vol. 51, pp. 219-235. San Diego: Academic Press. Niswander, L., Jeffrey, S., Martin, G. R. and Tickle, C. (1994). A positive feedback loop coordinates growth and patterning in the vertebrate limb. Nature 371, 609-612. Oppenheim, R. W., Homma, S., Marti, E., Prevette, D., Wang, S., Yaginuma, H. and McMahon, A. P. (1999). Modulation of early but not later stages of programmed cell death in embryonic avian spinal cord by Sonic Hedgehog. Mol. Cell. Neurosci. 13, 348-361. Pagan, S. M., Ros, M. A., Tabin, C. and Fallon, J. F. (1996). Surgical removal of limb bud Sonic hedgehog results in posterior skeletal defects. Dev. Biol. 180, 35-40. Parr, B. A. and McMahon, A. P. (1995). Dorsalizing signal Wnt-7a required for normal polarity of D-V and A-P axes of mouse limb. Nature 374, 350353. Pautou, M. P. (1975). Morphogenese de l’autopode chez l’embryon de poulet. J. Embryol. Exp. Morphol. 34, 511-529. Raff, M. (1998). Cell suicide for beginners. Nature 396, 119-122. Riddle, R. D., Johnson, R. L., Laufer, E. and Tabin, C. (1993). Sonic hedgehog mediates the polarizing activity of the ZPA. Cell 75, 14011416. Saunders, J. W. and Fallon, J. F. (1967). Cell death in morphogenesis. In Major Problems in Developmental Biology, (ed. M. Locke), pp. 289-314. New York: Academic Press. Saunders, J. W. and Gasseling, M. T. (1968). Ectodermal-mesenchymal interactions in the origin of limb symmetry. In Epithelial-mesenchymal Interactions (ed. R. Fleischmeyer and R. E. Billingham), pp. 78-97. Baltimore: Williams & Wilkins. Teillet, M.-A., Watanabe, Y., Duprez, P. J. D., Lapointe, F. and Shh signalling and cell death in limb development 4823 Douarin, N. M. L. (1998). Sonic hedgehog is required for survival of both myogenic and chondrogenic somitic lineages. Development 125, 20192030. Tickle, C. (1981). The number of polarizing region cells required to specify additional digits in the developing chick wing. Nature 289, 295-298. Tickle, C., Alberts, B., Wolpert, L. and Lee, J. (1982). Local application of retinoic acid to the limb bond mimics the action of the polarizing region. Nature 296, 564-566. Tickle, C., Lee, J. and Eichele, G. (1985). A quantitative analysis of the effect of all-trans-retinoic acid on the pattern of chick wing development. Dev. Biol. 109, 82-95. Tickle, C., Summerbell, D. and Wolpert, L. (1975). Positional signalling and specification of digits in chick limb morphogenesis. Nature 254, 199202. Todt, W. L. and Fallon, J. F. (1984). Development of the apical ectodermal ridge in the chick wing bud. J. Embryol. Exp. Morphol. 80, 21-41. Vaahtokari, A., Aberg, T. and Thesleff, I. (1996). Apoptosis in the developing tooth: association with an embryonic signaling center and supression by EGF and FGF-4. Development 122, 121-129. Vargesson, N., Clarke, J. D., Vincent, K., Coles, C., Wolpert, L. and Tickle, C. (1997). Cell fate in the chick limb bud and relationship to gene expression. Development 124, 1909-1918. Vaux, D. L. and Korsmeyer, S. J. (1999). Cell death in development. Cell 96, 245-254. Weil, M., Jacobson, M. D. and Raff, M. C. (1997). Is programmed cell death required for neural tube closure? Curr. Biol. 7, 281-284. Yang, Y., Drossopoulou, G., Chuang, P. T., Duprez, D., Marti, E., Bumcrot, D., Vargesson, N., Clarke, J., Niswander, L., McMahon, A. et al. (1997). Relationship between dose, distance and time in Sonic Hedgehog-mediated regulation of anteroposterior polarity in the chick limb. Development 124, 4393-4404. Yang, Y. and Niswander, L. (1995). Interaction between the signaling molecules WNT7a and SHH during vertebrate limb development: dorsal signals regulate anteroposterior patterning. Cell 80, 939-947.