DEVELOPMENTAL BIOLOGY 3230
Limb Development
Limbs have been studied as models for developmental pattern formation and regeneration for over a 100 years. The reasons for this are that limbs are easily observed during development and the limbs of many animals regenerate quite readily. In contrast developing and regenerating organ systems are generally hidden from view and more difficult to manipulate. Historically, the developing limbs of amphibians, chicks, and insects have been most intensively studied. The models for regeneration have been primarily the salamander and the cockroach. The advent of gene targeting has made it feasible to study limb development in the mouse. Historical studies in animals as diverse as insects and vertebrates gave amazingly similar results. The same set of simple cellular rules could be used to characterize cockroach and salamander regeneration experiments. Developing limbs arise from a limb field.
Transplantion and ablation studies showed that cells in a limb field are pluripotent and regulate their fate based on their position within the field. Organizing centers arise at the posterior margin and the distal AP margin that regulate limb pattern formation along the three body axis. The organizing centers establish morphogen gradients that specify the proximal-distal, anterior-posterior, and dorsal-ventral positional values of cells in the limb.

The illustration above shows a diagram of a chick forelimb. All vertebrate limbs are variations of the same basic plan. The most proximal bone is called the stylopod (humerus), the next is called the zeugopod (ulna and radius), and the most distal the autopod (metacarpals and digits). Thus, we can easily describe the proximal distal limb pattern by describing the pattern of bones. The anterior-posterior axis is defined by the pattern of digits. The most anterior is the thumb and the pinky the most posterior. The dorsal ventral axis is defined as the back of your hand (dorsal) to the palm (ventral). In each case the pattern that defines each axis is easily scored.
One of the first questions we need to ask is what determines the postion of the limb field? Remember that as the vertebrate anterior posterior and dorsal ventral axes are being established that there are gradients along the AP and DV axis of BMP, Wnt, Fgf, and retinoic acid. These morphogen gradients effect the regional expression of the vertebrate Hox genes along the body axes. We shall see how the retinoic acid, Fgf, and Hox gene expression patterns specify the position of the limb fields.
The illustration at right shows the limb field of a salamander. Notice the fate map assigned todifferent regions of the limb field. However, as with all embryonic fields, the cells within the field can be respecified by transplanting to new locations or by ablating cells and creating new neighbor relationships. For example, the entire free limb region can be extirpated and the remaining cells within the limb field will regulate their development to give a normally patterned limb. Soon after the formation of the limb field, you can experimentally determine that the DV and AP axes have formed.
The figure above shows the axes and position of the fore limb and hind limb fields in an embryonic chicken. The scanning electron micrographs in D and E show the underlying mesoderm that is involved in specifying the limb field and also contributing to the mesodermal limb precursors. The dorsal mesoderm (somites), intermediate mesoderm, and lateral plate mesoderm all contribute inductive signals to specify the limb field. Mesodermal cells from the somite (hypaxial myotome bud) contribute to the limb bud and will give rise to the muscle components of the limb. The lateral plate mesoderm contributes mesodermal cells that will form the bones and connective tissue of the limb. Notice in (E) a unique feature of the developing limb bud, the apical ectodermal ridge or AER. This is a thickened ridge of ectodermal cells that is important for limb patterning. The dorsal mesoderm induces radical fringe rFng and Wnt-7a in the dorsal limb bud ectoderm, while the ventral mesoderm of the lateral plate induces engrailed-1(en-1) in the ventral limb bud ectoderm.
The functional roles of the limb mesoderm and the AER have been determined by ablation and transplantion studies. In the figure at right you can see the results of many of these experiments. Removing the AER causes limb development to cease. Transplanting an extra AER to a limb bud causes a partial duplication of limb pattern. Removing the limb mesoderm causes limb development to cease. Replacing the fore limb mesoderm with hind limb mesoderm causes the limb to differentiate a hind limb pattern. Putting non-limb mesoderm in the place of limb mesoderm causes limb development to cease. Conclusions from these experiments suggest that the limb mesoderm both supports limb development and also specifies the identity of the limb (fore limb vs. hind limb). The AER is important both for supporting limb development and also in patterning the limb. Notice in the last experiment that a bead of Fgf can support limb development in the absence of an AER. This strongly suggests that the AER secretes Fgf to maintain proliferation of the limb mesoderm.

Transplantation of early and late AER, and the proliferating mesoderm just under the AER called the progress zone or PZ, led to further insights into the specific roles of the ectodermal and mesodermal components of the limb bud. Transplanting early and late AERs to young limb buds results in normal development.
Thus the patterning information in AER is not "instructive", but more likely permissive. However, the result of transplanting early and late PZ cells is very different. Early PZ supports the development nearly all proximal to distal structures so you see a duplication of proximal structures in (A), but transplantation of late PZ supports development of only the most distal (late developing) structures so you see a limb that is missing the proximal structures but with normal distal sturctures (B).
These, and other experiments have lead to a model where Fgf supplied by the AER specifies proximal distal fate in PZ cells. The longer cells remain in the PZ exposed to Fgf, the more distal their identity.
The PZ model for proximal distal specification is illustrated at right. You can imagine the exposure to Fgf from the AER is the morphogen gradient. The longer cells stay in the PZ the higher the exposure to Fgf. When cells leave the PZ their proxmal-distal positional value is fixed and they differentiate appropriately. So, the first mesodermal cells derived from the lateral plate mesoderm that exit the progress zone would condense to form the humerus, the most proximal bone in the limb. The last to exit give rise to the cartilage of the digits, the most distal bones.
How does the limb mesoderm specify limb identity? The answer is not completely understood, but two transcription factors, Tbox 4 and Tbox 5, are important. These genes are regionally expressed along the body axis and are correlated with fore limb and hind limb identity. The expression of Tbx5 and Tbx4 are probably regulated directly or indirectly by the retinoic acid gradient or regional expression of Hox genes. Notice that when a bead soaked in Fgf is applied to the flank between the normal limb buds, it will induce an ectopic bud. The limb that develops exhibits patterns specific to both the fore limb and hind limb.
Normally, we see Tbx 5 always associated with fore limb identity and Tbx4 associated with hind limb identity. We can experimentally cause the ectopic expression of Tbx4 and show we can convert the identity of the limb bud from fore limb to hind limb. Thus, from the very first, the limb mesoderm is specified to be either fore limb or hind limb based on its axial postion, and expression of Tbx5 or Tbx4.
What patterns the anterior-posterior limb axis? Transplantation studies in the late 60s by Saunders and Gasseling suggested an hypothesis. Transplanting the posterior mesodermal tissue from a donor to the anterior margin of a host leads to the development of a mirror image AP pattern. Many transplantions and ablation studies suggested that the posterior margin mesoderm was behaving like an "organizing" center that regulated the AP patterning of the limb. It was named the Zone of Polarizing Activity or ZPA. The model from these studies suggested that the ZPA was the source of a diffusible morphogen that specified the AP positional identity of cells. An intensive search ensued for the ZPA morphogen! In 1982 Tickle presented evidence that retinoic acid could mimic the effects of the ZPA. Several years of experiments supported the role of retinoic acid as the ZPA morphogen. It was found in the limb bud and present at a higher concentration in the posterior margin. Retinoic Acid Receptors were also found in the limb bud. Then in 1987 experiments by Bryant and a Japanese group showed that RA acted by inducing ZPA tissue!
The next real breakthrough came when it was realized that many of the molecular pathways regulating fly development were also important in vertebrate development. Hedgehog, a segment polarity gene that we studied earlier, is involved in wing patterning and when ectopically expressed in the anterior wing compartment causes wing pattern duplications. In 1993 Riddle showed that a vertebrate homolog of fly hh, called sonic hh, is expressed by the ZPA!
The pictures at right show the expression pattern of sonic hh. You can clearly see that it is expressed at the posterior margin of the each developing limb bud. The picture at far right shows a chick embryo in which a new limb bud has been induced by application of a Fgf soaked bead.
The landmark dates are from Tickle's historical review of limb development. Experiments like the one at right were performed to prove that Shh is the ZPA morphogen. Cells transfected with a DNA construct that expresses Shh at a high level are transplanted into the anterior margin of a limb bud. The ectopic Shh seems to mimic all aspects of the ZPA function. Shh was also found to induce BMP-2 expression by the ZPA. Current evidence suggests that both Shh and BMP morphogen gradients act together to pattern the limb. Evidence is also emerging suggesting an anterior signal that antagonizes the Shh gradient. The more closely patterning in the limb is looked at the more "homologies" are found with fly DV patterning signals
The model for limb anterior posterior patterning by the ZPA is illustrated below. This is a figure from Wolpert's book, Principles of Developmental Biology. Wolpert has been a major figure in the study of limb pattern formation and developmental biology. The model proposes that a simple morphogen gradient (Shh and BMP-2) determines digit identity by threshold values and can explain the results of many experimental manipulations of the ZPA.
The following figures summarize what is known about limb development and patterning. Remember, it this was forelimb then Tbx5 would be expressed to specify fore limb identity (Tbx4, hindlimb). The anterior-posterior pattern of Hox genes probably specifies the location of the limb bud (Hox 9 genes are candidates). Intermediate mesoderm secretes Fgf8 and induces lateral plate mesoderm to secrete Fgf10. Lateral plate mesoderm Fgf10 induces the overlying ectoderm to form the apical ectodermal ridge (AER). The AER now secretes Fgf8 into the progress zone mesoderm to maintain proliferation. Fgf8 from the AER induces the formation of the ZPA in the posterior most limb mesoderm. This is the only mesoderm competent to form ZPA due to the RA gradient and the expression of Hox 8. Shh from the ZPA feeds back on the AER to regulate Fgf expression.
Dorso-ventral patterning is also established very early. Dorsal mesoderm induces rFng and Wnt7a in the dorsal limb bud. Ventral (lat.Plate) mesoderm induces engrailed-1 expression in the ventral limb bud. Notice how the major factors regulating DV (Wnt), AP (Shh), and PD (Fgf) pattern interact with one another. It should be obvious that you need coordination among these axial patterning systems so that the right struture develops the right DV, AP, and PD pattern.
Wnt-7a is expressed by dorsal ectoderm. It induces Lmx-1 in the dorsal mesoderm and specifies dorsal pattern. rFng is also expressed in the dorsal ectoderm. The inhibitory interaction between en-1 expressing cells and rFng expressing cells determines the position of the AER at the equatorial margin. En-1 expressing cells antagonize Wnt-7a and defines the ventral region of the limb bud.
Transverse section through progress zone of limb bud. Wnt-7a in dorsal ectoderm induces the Lmx-1 transcription factor in the dorsal limb mesoderm. Lmx-1 specifies dorsal mesoderm. Engrailed is expressed in the ventral ectoderm and antogonizes the affects of Wnt7a to "ventralize" the limb bud.
A summary of the important interactions among the major axial patterning systems. This predicts that mutantions affecting primarily one axial system will have significant affects on the others. This has been demonstrated experimentally by making mouse knock outs of many of these important molecules. The Wnt7a KO mouse has a ventralized limb that is also missing posterior pattern elements and truncated.
The affects on AP and PD patterning are thought to be due to Wnt7a normal function as a positive regulator of Shh. Reduced Shh would lead to AP patterning defects and also PD defects because of Shh importance as a positive regulator of Fgf. The en-1 KO mouse has a dorsalized limb and the AER forms at a more ventral location.
The next major question is how are the morphogen gradients of these axial patterning systems interpreted? What genes respond to the threshold values to actually specify pattern elements? The answer seems to again be the Hox genes. Remember, the Hox genes are transcription factors that act as molecular switches so there is no reason they can't be used at different times and places to specify different regional identities. First it was noted that the Hox genes are expressed in a complex PD and AP pattern. At right you can see a figure from one of Capecchi's early papers where he describes the pattern of Hox d-10 through d-13 expression during limb development in the mouse. Notice in A that there is a PD progression and in B an AP progression as the digits begin to be specified. The proof of Hox gene importance in limb development is shown in C. KO mice were made that knocked out the functions of Hox a-11 and d-11. In these double KO mice you can see that the radius and ulna are almost completely lacking.
The figure at right summaries the results of many experiments looking at the expression patterns of the Hox genes durig limb development and the effects of making mouse targeted gene KO mutations in many of the Hox genes. Notice that you can generally order the expression of the Hox genes along the PD axis and that order correlates with the 3' to 5' linear arrangement of the Hox genes on the chromosomes.
This figure summaries the finding that the more 3' Hox genes function to specify the patterned identity of the more proximal limb structures while the distal limb structures are specified by the more 5' Hox genes. This is obviously similar to the expression of these genes along the anterior posterior axis of the animal.
Limb malformations are relatively common in humans. Anterior posterior patterning defects are seen in about 1% of births. The hand at right shows a relatively mild defect, polydactyly. Many of the human genetic defects causing limb malformations are the result of mutations affecting the Shh signaling pathway or the Hox genes. If you are interested in genes affecting human limb formation you should check out the OMIM and searh their database for "limb bud". This is a great site for anyone interested in the genetic basis of human diseases.
The pictures below show a much more severe limb phenotype that is associated with mutations of the HoxD13 gene. Notice the difference between the heterozygous on the left and a the homozygous condition on the right. Again we can see the important contributions to understanding human diseases that have come from studying the lowly fruit fly!
The figure below summaries the similarities between wing development in the fly and chick. Truly amazing that structures that appear so different and that occur on animals separated by more than 600 million years of evolution share the fundamental molecular mechanisms patterning.
This is a summary figure from Wolpert's book, the Principles of Development, but I think you will find it nicely summaries what I have told you.