Limb development

Limb development in tetrapods — animals with four limbs — is an area of active research in developmental biology. Limb formation begins in the limb field, as a limb "bud." Fibroblast growth factor (FGF) induces formation of an organizer, called the apical ectodermal ridge (AER), which guides further development and controls cell death. Apoptosis — programmed cell death — is necessary to eliminate webbing between digits.

The limb field is a region specified by expression of certain homeobox (Hox) genes and Tbx genes; Tbx5 for forelimb and Tbx4 for hindlimb. Establishment of the limb field requires retinoic acid signaling in the developing trunk of the embryo from which the limb emerges.

Limb formation results from series of epithelial-mesenchymal inductions between the limb bud mesenchymal cells and the overlying ectodermal cells. Cells from the lateral plate mesoderm and the myotome migrate to the limb field and proliferate to create the limb bud. The lateral plate cells produce the cartilaginous and skeletal portions of the limb while the myotome cells produce the muscle components. The lateral plate mesodermal cells secrete a fibroblast growth factor (FGF7 and FGF10, presumably) to induce the overlying ectoderm to form an important organizing structure called the apical ectodermal ridge (AER). . The AER reciprocatively secretes FGF8 and FGF4 which maintains the FGF10 signal and induces proliferation in the mesoderm. The position of FGF10 expression is regulated by Wnt8c in the hindlimb and Wnt2b in the forelimb. The forelimb and the hindlimb are specified by their position along the anterior/posterior axis and possibly by two T-box containing transcription factors: Tbx5 and Tbx4, respectively.

Programmed cell death removes the "webbing" between the digits and joints. BMP signaling induces cell death and Noggin blocks cell death in the digits.

Precartilage condensations
The limb's skeletal elements are prefigured by tight aggregates of mesenchymal cells called precartilage condensations. Mesenchymal condensation is mediated by extracellular matrix and cell adhesion molecules. Cartilage differentiates from the condensations, giving rise to the skeletal primordia. In most tetrapod limb skeletons (though not in some amphibians), the cartilage skeleton is replaced by bone later in development.

Periodicities of the limb pattern
The limb is organized into three regions: stylopod, zeugopod, and autopod (in order from proximal to distal). Within the distal two of these regions the skeleton contains a number of periodic and quasi-periodic pattern motifs. The zeugopod consists of two parallel elements along the anteroposterior (AP) axis and the autopod contains 3-5 (in most cases) elements along the same axis. The digits also have a quasi-periodic arrangement along the proximodistal (PD) axis, consisting of tandem chains of skeletal elements. The generation of the basic limb plan during development results from the patterning of the mesenchyme by an interplay of factors that promote precartilage condensation and factors that inhibit it.

The development of the basic limb plan is accompanied by generation of local differences between the elements. For example, the radius and ulna of the forelimb zeugopod are distinct from one another, as are the different fingers in the autopod. These differences can be treated schematically by considering how they are reflected in each of the limb's three main axes.

Axial patterning and related issues
Many investigations into the development of the limb skeletal pattern have been influenced by the "positional information" concept proposed by Lewis Wolpert in 1971. In tune with this idea, efforts have been made to identify diffusive signaling molecules (morphogens) that traverse orthogonal axes of developing limbs and determine locations and identities of skeletal elements in a concentration-dependent fashion.

Proximal/distal patterning
Hox genes contribute to the specification of the stylopod, zeugopod and autopod. Mutations in Hox genes lead to proximal/distal losses or abnormalities. Three different models have been advanced for explaining the patterning of these regions.

Progress Zone model
The AER creates a zone of cell proliferation and lays down the limb from proximal to distal. The time cells leave the AER determines their positional value. Proximal structures are formed earlier than distal structures.

The Progress Zone model was proposed 30 years ago but recent evidence has conflicted with this model.

Experimental evidence:
 * Removing the AER at a later period of development results in less disruption of distal structures than if the AER was removed early in development.
 * Grafting a new bud tip on top of an old bud tip results in a deletion and duplication of structures.

Early allocation and progenitor expansion model (or prespecification model)
Cells are specified for each segment in the early limb bud and this population of cells expand out as the limb bud grows. This model is consistent with the following observations. Cell division is seen throughout the limb bud. Cell death occurs within a 200 μm zone subjacent to the AER when it is removed; cell death forecloses some patterning. FGF-releasing beads are able to rescue limb development when the AER is removed by preventing this cell death.

Experimental evidence:
 * Labeled cells in different position of an early limb bud were restricted to single segments of the limb.
 * Limbs lacking expression of required FGF4 & FGF8 showed all structures of the limb and not just the proximal parts.

More recently, however, the investigators primarily responsible for both the Progress Zone and Prespecification models have acknowledged that neither of these models accounts adequately for the available experimental data.

Activator-inhibitor (or reaction-diffusion) model
This model is based on the self-organizing properties of excitable media described by Alan Turing in 1952. The excitable medium is the limb bud mesenchyme, in which cells interact by positively autoregulatory morphogens such as TGF-β and inhibitory siginaling pathways involving FGF and Notch. Proximodistal and anteroposterior axes are not considered to be independently specified, but instead emerge by transitions in the number of parallel elements as the undifferentiated apical zone of the growing limb bud undergoes reshaping. This model only specifies a "bare bones" pattern. Other factors like Sonic hedgehog (Shh) and Hox proteins, primary informational molecules in the other models, are proposed instead to play a fine-tuning role.

Experimental evidence:
 * Limb mesenchymal cells, when dissociated and grown in culture or reintroduced within ectodermal "hulls" can recapitulate essential aspects of pattern formation, morphogenesis and differentiation.
 * Peculiarities of the limb skeletal pattern in the mouse Doublefoot mutant are predicted outcomes of a Turing-type mechanism.

Anterior/posterior patterning
The Zone of Polarizing Activity (ZPA) in the limb bud has pattern-organizing activity by action of a morphogen gradient of Sonic hedgehog (Shh). Shh is both sufficient and necessary to create the ZPA and specify the anterior/posterior pattern in the distal limb (Shh is not necessary for the polarity of the stylopod). Shh is turned on in the posterior through the early expression of Hoxd genes, the expression of Hoxb8, and the expression dHAND. Shh is maintained in the posterior though a feedback loop between the ZPA and the AER. Shh induces the AER to produce FGF4 and FGF8 which maintains the expression of Shh.

Digits 3,4 and 5 are specified by a temporal gradient of Shh. Digit 2 is specified by a long-range diffusible form of Shh and Digit 1 does not require Shh. Shh cleaves the Ci/Gli3 transcriptional repressor complex to convert the transcription factor Gli3 to an activator which activates the transcription of HoxD genes along the anterior/posterior axis. Loss of the Gli3 repressor leads to the formation of generic (non-individualized) digits in extra quantities.

Dorsal/ventral patterning
Dorsal/Ventral patterning arises from Wnt7a signals in the overlying ectoderm not the mesoderm. Wnt7a is both necessary and sufficient to dorsalize the limb. Wnt7a also influences the anterior/posterior axis and loss of Wnt7a causes the dorsal side of limbs to become ventral sides and causes missing posterior digits. Replacing Wnt7a signals rescues this defect. Wnt7a is also required to maintain expression of Shh.

Wnt-7a also causes Lmx-1, a Lim Hox gene (and thus a transcription factor), to be expressed. Lmx-1 is involved in dorsalisation of the limb, which was shown by knocking out the Lmx-1 gene in mice. The mice lacking the Lmx-1 produced ventral skin on both sides of their paws. There are other factors thought to control the DV patterning; Engrailed-1 represses the dorsalizing effect of Wnt-7a on the ventral side of the limbs.