- Email: [email protected]
Limb Patterning: Reports of Model's Death Exaggerated
Current Biology, Vol. 12, R628–R630, September 17, 2002, ©2002 Elsevier Science Ltd. All rights reserved. PII S0960-9822(02)01137-5
Limb Patterning: Reports of Model’s Death Exaggerated Lewis Wolpert
The progress zone model for the specification of positional values for patterning the proximodistal axis of the vertebrate limb has been questioned, but the results can be largely reconciled with the old model.
One of the basic processes in development is pattern formation, that is the specification of the spatial organisation of cells so that well-defined patterns of cell differentiation develop [1]. For vertebrates the development of the limb is an important model for studying pattern formation as the pattern of cartilage differentiation is very clear and relatively simple, and in the chick the limb is accessible to experimental manipulation [2]. One model for patterning the limb is based on positional information, that is the cells acquire positional identities with respect to boundaries, typically due to a gradient in a morphogen. A rather different model for a growing system is based not on a morphogen gradient but on cells measuring the time they spend in a progress zone [3]. It is this model for the proximal–distal patterning of the limb proposed some thirty years ago that is currently being questioned [4]. I should declare here that I have an interest in the progress zone model, as one of its originators, but while the new data are provocative, in my view they do not provide a compelling reason to consign the progress zone model to the dustbin of history. And it seems rather ironic that the progress zone model is now being questioned, as at the same time a similar mechanism has recently been proposed for the specification of the patterning of Hox genes along the anterior–posterior axis of the embryo, the cells measuring the time they spend in the region linked to the regressing node [5–7]. In the limb bud cells acquire positional identities along each of the three axes of the limb — proximal–distal, anterior–posterior, and dorsal–ventral — and this determines how the cells will differentiate (Figure 1). For the anterior–posterior axis, thumb to little finger, there is good evidence for a graded signal from the polarising region at the posterior margin of the bud which is based on Sonic hedgehog; a high level of the gradient signals digit 4 while a low level digit 2 [8]. For the proximal–distal axis, the model suggests that the cells in a region at the tip of the limb, about 300 microns deep, are specified by the thickened apical ectoderm to be a progress zone. This is the region where the cells acquire their positional identities. For example it is only in the progress zone
Department of Anatomy and Developmental Biology, University College, Gower Street, London WC1E 6BT, UK.
Dispatch
that a signal from the polarising region can change anterior–posterior positional identities. The cells in the zone are dividing and so cells leave the zone continually; thus if time in the zone specifies position along the proximal–distal axis, the cells that remain in longest will form the digits while the cells that leave early will develop as proximal structures. Removal of the apical ridge has long been known to result in truncations and this fits with but does not establish the model’s validity. A very recent report by Dudley et al. [4] finds that ridge removal results in significant cell death in the underlying region and claim that this, rather than loss of the progress zone, is the the cause of the truncations. They do however admit that this explanation cannot account for truncations in the autopod, the distal segments of the limb, as at that later stage they saw no cell death. They nevertheless claim that their results of the effect of ridge removal fit with cell death of progenitor cells rather than with the failure to specify more distal structures. However all their results are interpreted in terms of there being but three sets of elements along the axis — zeugopod, stylopod and autopod — and this is in marked contradiction of the results of earlier papers that clearly identified seven elements: humerus, radius and ulna, two carpal elements and three elements in the third digit including a metacarpal [9,10]. Each of these elements is initially about the size of a progress zone and comes from one doubling of the zone, but the wrist elements grow hardly at all compared to the radius and ulna. The neglect of the wrist elements which are laid down between stages 21 and 24 as revealed by apical ectodermal ridge (AER) removal (Figure 2) also has severe implications for Dudley et al.’s interpretation of their results, particularly those relating to lineage. In order to show that all the elements are already specified in the early bud, Dudley et al. [4] labelled cells in the early bud, and then interpreted their observations as showing that labelled cells rarely ended up in more than one segment, meaning that the cells are already specified at this early stage. But they completely ignore the wrist elements and their results clearly show labelled cells ending up in three or four segments. Moreover other studies confirm this latter observation: cells labelled under the ridge end up in several more proximal elements as predicted by the progress zone model [11]. It is also important to realise — which the authors apparently fail to do — that if there is very early specification, then each cartilaginous element will be represented by around just four cells — seven elements in 300 microns. This is most implausible as the precision required is unknown in vertebrate patterning. Dudley et al. [4] do not discuss how this early pattern might be specified. The best evidence for the progress zone model comes from killing cells in an early limb bud by X-irradiation, which results in the loss of proximal elements
Current Biology R629
Anterior
Anterior Progress zone
Cartilage of developing humerus
Humerus Radius
Proximal
Distal Apical ectodermal ridge
digit 2 digit 3
Figure 1. Cells acquire their positional identity in the progress zone. After they have left the zone, cartilage elements begin to develop in a proximal-todistal sequence. The wrist elements are not shown.
digit 4 Polarizing region
Ulna Posterior
Posterior
Proximal
Distal Current Biology
whereas distal ones are normal [12]. In terms of the model, cell death results in few normal cells leaving the zone at early cell cycles until the remaining cells in the progress zone divide to repopulate it (Figure 3). Thus very few cells spend just a short time in the progress zone. Dudley et al. [4] simply dismiss the Xirradiation results as reflecting the degree of determination of the cells at the time of irradiation, though why distal cells should be more determined than proximal cells is not explained, and would be quite against their progressive determination model. When cells from an early bud, stage 20, are dissociated and then placed back in the ectodermal jacket all elements develop. Dudley et al. [4] did the same experiment with the pooled mesenchyme from the distalmost 100 microns. In terms of their model these cells are all fated to form distal elements yet they give the same result as total mesenchyme, which would seem clearly to contradict their model. They explain it away by somehow distinguishing between specification and determination. This result is just what the progress zone model predicts but they argue that, by stage 20, the cells will have been sufficiently long in the progress zone that the most proximal element should not develop. This is a reasonable objection but may merely reflect small differences in staging. However when they take the tissue from stage 22 only distal elements develop, as predicted by the progress zone model.
There is at present no reason to accept, as has been suggested, that the progress zone has fallen victim to progress, and in fact it still provides the only plausible model for proximal–distal patterning. The resolution of these issues in favour of early specification would come about if molecular differences corresponding to the proximal–distal elements were identified in the early bud. By contrast, if evidence for a timer such as an oscillator was found in at the tip of the bud the progress zone model would be strongly supported [13]. Time will tell. But rather like the letter from Mark Twain that the report of his death was an exaggeration, so it is with the progress zone. References 1. Wolpert. L. et al. (2002). Principles of Development, 2nd edn. (Oxford University Press). 2. Tickle, C. (2000). Limb development: an international model for vertebrate pattern formation. Int. J. Dev. Biol. 44, 101–108. 3. Summerbell, D., Lewis, J. and Wolpert, L. (1973). Positional information in chick limb morphogenesis. Nature 244, 492–496. 4. Dudley, A.T., Ros, M.A. and Tabin, C.T. (2002). A re-examination of proximodistal patterning during vertebrate limb development. Nature 418, 539–544. 5. Dale, K.A. and Pourquie, O. (2000). A clock-work somite model. Bioessays 22, 72–83. 6. Kerszberg, M. and Wolpert, L. (2000). A clock and trail model for somite formation, specialization and polarization. J. Theor. Biol. 205, 505–510. 7. Vasiliauskas, D. and Stern, C. (2001). Patterning the embryonic axis: FGF signaling and how vertebrate embryos measure time. Cell 106, 133–136. 8. Sanz-Ezquerro, J. and Tickle, C. (2001). ‘Fingering’ the vertebrate limb. Differentiation 69, 91–99.
A
1st cycle
PZ
B PZ
0•2
28 27 7
6
26 5
25–21
20 19
18
432
1
0
Stage Age, τ Current Biology
Figure 2. The level of truncation after the excision of the apical ectodermal ridge (AER). The time is measured in terms of cell doublings. Note the time spent in laying down the wrist, and that each element is about one progress zone doubling prior to growth.
1st cycle
PZ
0•36
PZ
1st cycle
2nd cycle
PZ
1st 2nd cycle cycle PZ
0•5
Current Biology
Figure 3. A model to illustrate the effect of cell killing in the progress zone. A cell’s proximal–distal positional value may depend on the time it spends in the progress zone and in (A) 25 cells leave each cell cycle. If cells are killed (B), open circles, then the number of normal cells leaving at each doubling (shown as a fraction beneath) is limited until the zone is repopulated, leading to the loss of proximal structures.
Dispatch R630
9. Summerbell, D. (1974). A quantitative analysis of the excision of the AER from the chick limb bud. J. Embryol. Exp. Morphol. 32, 651–660. 10. Lewis, J. (1975). Fate maps and the pattern of cell division. J. Embryol. Exp. Morphol. 33, 419–434. 11. 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. 12. Wolpert, L., Tickle, C. and Sampford, M. (Appendix by J. Lewis) (1979). The effect of cell killing by X-irradiation on pattern formation in the chick limb. J. Embryol. Exp. Morph. 50, 175–198. 13. Tickle. C. and Wolpert, L. (2002). Progress zone, dead or alive. Nat. Cell Biol., in press.
Limb Patterning: Reports of Model’s Death Exaggerated Lewis Wolpert
The progress zone model for the specification of positional values for patterning the proximodistal axis of the vertebrate limb has been questioned, but the results can be largely reconciled with the old model.
One of the basic processes in development is pattern formation, that is the specification of the spatial organisation of cells so that well-defined patterns of cell differentiation develop [1]. For vertebrates the development of the limb is an important model for studying pattern formation as the pattern of cartilage differentiation is very clear and relatively simple, and in the chick the limb is accessible to experimental manipulation [2]. One model for patterning the limb is based on positional information, that is the cells acquire positional identities with respect to boundaries, typically due to a gradient in a morphogen. A rather different model for a growing system is based not on a morphogen gradient but on cells measuring the time they spend in a progress zone [3]. It is this model for the proximal–distal patterning of the limb proposed some thirty years ago that is currently being questioned [4]. I should declare here that I have an interest in the progress zone model, as one of its originators, but while the new data are provocative, in my view they do not provide a compelling reason to consign the progress zone model to the dustbin of history. And it seems rather ironic that the progress zone model is now being questioned, as at the same time a similar mechanism has recently been proposed for the specification of the patterning of Hox genes along the anterior–posterior axis of the embryo, the cells measuring the time they spend in the region linked to the regressing node [5–7]. In the limb bud cells acquire positional identities along each of the three axes of the limb — proximal–distal, anterior–posterior, and dorsal–ventral — and this determines how the cells will differentiate (Figure 1). For the anterior–posterior axis, thumb to little finger, there is good evidence for a graded signal from the polarising region at the posterior margin of the bud which is based on Sonic hedgehog; a high level of the gradient signals digit 4 while a low level digit 2 [8]. For the proximal–distal axis, the model suggests that the cells in a region at the tip of the limb, about 300 microns deep, are specified by the thickened apical ectoderm to be a progress zone. This is the region where the cells acquire their positional identities. For example it is only in the progress zone
Department of Anatomy and Developmental Biology, University College, Gower Street, London WC1E 6BT, UK.
Dispatch
that a signal from the polarising region can change anterior–posterior positional identities. The cells in the zone are dividing and so cells leave the zone continually; thus if time in the zone specifies position along the proximal–distal axis, the cells that remain in longest will form the digits while the cells that leave early will develop as proximal structures. Removal of the apical ridge has long been known to result in truncations and this fits with but does not establish the model’s validity. A very recent report by Dudley et al. [4] finds that ridge removal results in significant cell death in the underlying region and claim that this, rather than loss of the progress zone, is the the cause of the truncations. They do however admit that this explanation cannot account for truncations in the autopod, the distal segments of the limb, as at that later stage they saw no cell death. They nevertheless claim that their results of the effect of ridge removal fit with cell death of progenitor cells rather than with the failure to specify more distal structures. However all their results are interpreted in terms of there being but three sets of elements along the axis — zeugopod, stylopod and autopod — and this is in marked contradiction of the results of earlier papers that clearly identified seven elements: humerus, radius and ulna, two carpal elements and three elements in the third digit including a metacarpal [9,10]. Each of these elements is initially about the size of a progress zone and comes from one doubling of the zone, but the wrist elements grow hardly at all compared to the radius and ulna. The neglect of the wrist elements which are laid down between stages 21 and 24 as revealed by apical ectodermal ridge (AER) removal (Figure 2) also has severe implications for Dudley et al.’s interpretation of their results, particularly those relating to lineage. In order to show that all the elements are already specified in the early bud, Dudley et al. [4] labelled cells in the early bud, and then interpreted their observations as showing that labelled cells rarely ended up in more than one segment, meaning that the cells are already specified at this early stage. But they completely ignore the wrist elements and their results clearly show labelled cells ending up in three or four segments. Moreover other studies confirm this latter observation: cells labelled under the ridge end up in several more proximal elements as predicted by the progress zone model [11]. It is also important to realise — which the authors apparently fail to do — that if there is very early specification, then each cartilaginous element will be represented by around just four cells — seven elements in 300 microns. This is most implausible as the precision required is unknown in vertebrate patterning. Dudley et al. [4] do not discuss how this early pattern might be specified. The best evidence for the progress zone model comes from killing cells in an early limb bud by X-irradiation, which results in the loss of proximal elements
Current Biology R629
Anterior
Anterior Progress zone
Cartilage of developing humerus
Humerus Radius
Proximal
Distal Apical ectodermal ridge
digit 2 digit 3
Figure 1. Cells acquire their positional identity in the progress zone. After they have left the zone, cartilage elements begin to develop in a proximal-todistal sequence. The wrist elements are not shown.
digit 4 Polarizing region
Ulna Posterior
Posterior
Proximal
Distal Current Biology
whereas distal ones are normal [12]. In terms of the model, cell death results in few normal cells leaving the zone at early cell cycles until the remaining cells in the progress zone divide to repopulate it (Figure 3). Thus very few cells spend just a short time in the progress zone. Dudley et al. [4] simply dismiss the Xirradiation results as reflecting the degree of determination of the cells at the time of irradiation, though why distal cells should be more determined than proximal cells is not explained, and would be quite against their progressive determination model. When cells from an early bud, stage 20, are dissociated and then placed back in the ectodermal jacket all elements develop. Dudley et al. [4] did the same experiment with the pooled mesenchyme from the distalmost 100 microns. In terms of their model these cells are all fated to form distal elements yet they give the same result as total mesenchyme, which would seem clearly to contradict their model. They explain it away by somehow distinguishing between specification and determination. This result is just what the progress zone model predicts but they argue that, by stage 20, the cells will have been sufficiently long in the progress zone that the most proximal element should not develop. This is a reasonable objection but may merely reflect small differences in staging. However when they take the tissue from stage 22 only distal elements develop, as predicted by the progress zone model.
There is at present no reason to accept, as has been suggested, that the progress zone has fallen victim to progress, and in fact it still provides the only plausible model for proximal–distal patterning. The resolution of these issues in favour of early specification would come about if molecular differences corresponding to the proximal–distal elements were identified in the early bud. By contrast, if evidence for a timer such as an oscillator was found in at the tip of the bud the progress zone model would be strongly supported [13]. Time will tell. But rather like the letter from Mark Twain that the report of his death was an exaggeration, so it is with the progress zone. References 1. Wolpert. L. et al. (2002). Principles of Development, 2nd edn. (Oxford University Press). 2. Tickle, C. (2000). Limb development: an international model for vertebrate pattern formation. Int. J. Dev. Biol. 44, 101–108. 3. Summerbell, D., Lewis, J. and Wolpert, L. (1973). Positional information in chick limb morphogenesis. Nature 244, 492–496. 4. Dudley, A.T., Ros, M.A. and Tabin, C.T. (2002). A re-examination of proximodistal patterning during vertebrate limb development. Nature 418, 539–544. 5. Dale, K.A. and Pourquie, O. (2000). A clock-work somite model. Bioessays 22, 72–83. 6. Kerszberg, M. and Wolpert, L. (2000). A clock and trail model for somite formation, specialization and polarization. J. Theor. Biol. 205, 505–510. 7. Vasiliauskas, D. and Stern, C. (2001). Patterning the embryonic axis: FGF signaling and how vertebrate embryos measure time. Cell 106, 133–136. 8. Sanz-Ezquerro, J. and Tickle, C. (2001). ‘Fingering’ the vertebrate limb. Differentiation 69, 91–99.
A
1st cycle
PZ
B PZ
0•2
28 27 7
6
26 5
25–21
20 19
18
432
1
0
Stage Age, τ Current Biology
Figure 2. The level of truncation after the excision of the apical ectodermal ridge (AER). The time is measured in terms of cell doublings. Note the time spent in laying down the wrist, and that each element is about one progress zone doubling prior to growth.
1st cycle
PZ
0•36
PZ
1st cycle
2nd cycle
PZ
1st 2nd cycle cycle PZ
0•5
Current Biology
Figure 3. A model to illustrate the effect of cell killing in the progress zone. A cell’s proximal–distal positional value may depend on the time it spends in the progress zone and in (A) 25 cells leave each cell cycle. If cells are killed (B), open circles, then the number of normal cells leaving at each doubling (shown as a fraction beneath) is limited until the zone is repopulated, leading to the loss of proximal structures.
Dispatch R630
9. Summerbell, D. (1974). A quantitative analysis of the excision of the AER from the chick limb bud. J. Embryol. Exp. Morphol. 32, 651–660. 10. Lewis, J. (1975). Fate maps and the pattern of cell division. J. Embryol. Exp. Morphol. 33, 419–434. 11. 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. 12. Wolpert, L., Tickle, C. and Sampford, M. (Appendix by J. Lewis) (1979). The effect of cell killing by X-irradiation on pattern formation in the chick limb. J. Embryol. Exp. Morph. 50, 175–198. 13. Tickle. C. and Wolpert, L. (2002). Progress zone, dead or alive. Nat. Cell Biol., in press.