How limb size and shape is controlled is a long-standing question in developmental biology. Leonardo da Vinci's famous painting of the Vitruvian man was made on the basis of a study of the proportions of the human male body by Vitruvius (Vinci, 1487). According to his study, the length of a man's outspread arms is equal to his height. The distance from the elbow to the tip of the hand is a quarter of a man's height. The distance from the elbow to the armpit is one-eighth of a man's height. These proportions of segment lengths are regulated probably by intrinsic growth control. In some animals like newts or a cricket nymph (Fig. 1A), after leg amputation, regeneration occurs to obtain a normal leg length by recovering the missing portion (Fig. 1B) (Brockes and Kumar, 2008), implying that leg segments possess intrinsic information about their final length. Recently, the advent of large-scale genome-wide association studies has led to dramatic progress in identifying the specific genetic variants that explain normal variation in human height (Lango et al., 2010). Hundreds of genetic variants in at least 180 loci influence adult height. The 180 loci are enriched with genes that underlie growth plate development, including hedgehog (Hh) and TGFβ signaling (Weedon and Frayling, 2008). However, the relationship between Hh or TGFβ signaling and human height or the intrinsic information of limb length remains to be elucidated.
The other long-standing question is how shape of each limb segment is determined. Each cell in the limb segment is likely to proliferate and differentiate according to its positional identity. However, we still do not know how cells memorize positional values (PVs) (for recent reviews, see Brockes and Kumar, 2008; Nakamura et al., 2008a). In limb regeneration, cells at the amputated position know their position in the limb segment and restore the missing portion (Fig. 1B). Molecular mechanisms that provide PVs to cells and mechanisms underlying cellular positional memory are poorly understood. PVs in the limb may be somehow related to mechanisms that stop cell growth when limb length reaches its final size during development or regeneration.
Recently, the Dachsous-Fat (Ds-Ft) and Warts-Hippo (Wts-Hpo) signaling pathways (Fig. 1C) are thought to be the key to answering these questions. Ft is a large atypical cadherin molecule (>5,000 amino acids) as a transmembrane receptor with 34 cadherin domains in its extracellular region (Mahoney et al., 1991). In Drosophila, null alleles of ft are lethal, and mutants have overgrown imaginal discs. ft mutants with weak viable alleles exhibit broadening of the abdomen (hence the name) and thick legs (Fig. 2) (for a review, see Eaton, 2003). Ds is also a large (>3,000 amino acids) atypical cadherin with 27 cadherin domains (Clark et al., 1995). The efficiency of their heterodimer formation is modulated by a Golgi-localized protein kinase Four-jointed (Fj) (Villano and Katz, 1995; Matakatsu and Blair, 2004; Ishikawa et al., 2008), which phosphorylates Ft and Ds (Brittle et al., 2010; Simon et al., 2010). Ds-Ft are thought to be a ligand-receptor pair that acts via the atypical myosin Dachs (D) (Mao et al., 2006), whose localization in a cell reflects the Ft activity (Matakatsu and Blair, 2008; Rogulja et al., 2008). Both fj and ds are expressed in gradients in developing imaginal discs (Clark et al., 1995; Villano and Katz, 1995; Brodsky and Steller, 1996; Buckles et al., 2001).
Ft is known to be a Drosophila tumor suppressor that suppresses the growth of the imaginal discs. This function has been linked to the Wts-Hpo signaling pathway (Fig. 1C) (Huang et al., 2005; Cho et al., 2006; Edgar, 2006; Reddy and Irvine, 2008; Pan, 2010), including Lowfat (Lft), Warts (Wts), Hpo kinases, Salvador, Mob-as-tumor-suppressor, and the 4.1, ezrin, radixin, moesin (FERM) domain proteins Expanded (Ex), Merlin (Mer), and Kibra (Fig. 1C) (for a recent review, see Halder and Johnson, 2010). These proteins suppress growth by the phosphorylation or cytosolic retention of the transcriptional coactivator Yorkie (Yki), which cannot bind to DNA directly (Huang et al., 2005). Yki interacts with a transcription factor Scalloped (Sd), which induces tissue overgrowth (Goulev et al., 2008; Wu et al., 2008; Zhang et al., 2008). The Wts-Hpo signaling factors are now considered as general regulators of organ size in both Drosophila and mammals (Reddy and Irvine, 2008; Halder and Johnson, 2010; Pan, 2010; Zhao et al., 2010). Thus, it is reasonable to speculate that the Ds-Ft and Wts-Hpo signaling pathways are involved in controlling the size of various organs. Since several excellent reviews have been published on these signaling pathways (for recent reviews, see Kango-Singh and Singh, 2009; L. Zhang et al., 2009; Grusche et al., 2010; Halder and Johnson, 2010; Oh and Irvine, 2010; Pan, 2010; Yi and Kissil, 2010; Zhao et al., 2010), we have not described these pathways in detail in this review. Rather, we aim to review recent advances in determining leg segment size and PVs in the leg segment, focusing on the Ds-Ft and Wts-Hpo signaling pathways.
GROWTH AND PLANAR CELL POLARITY ARE REGULATED BY A GRADIENT OF POSITIONAL VALUES
The legs of insect nymphs of the cricket or cockroach can regenerate after amputation (Fig. 1B) (for reviews, see Wolpert et al., 2007; Brockes and Kumar, 2008; Nakamura et al., 2008a). Interestingly, when a distally amputated tibia, for example, is grafted onto a host tibia that has been cut at a more proximal position, growth occurs at the junction between graft and host, and then the missing central regions of the tibia are intercalated, referred to as normal intercalary regeneration (see Fig. 8G). On the other hand, when a tibia amputated proximally is grafted onto a more distal position, intercalary regeneration occurs with the reverse orientation to the rest of the leg (see Fig. 8I), as indicated by the direction in which the bristles point (reverse intercalary regeneration). These phenomena suggest two important points: (1) when cells with non-continuous PVs are placed next to each other, the missing PVs are intercalated by growth to provide a set of continuous PVs; (2) the gradient in PVs also specifies planar cell polarity (PCP). It is known that intercalary regeneration also occurs in a circumferential direction (French, 1978). Thus, the gradient in PVs appears closely related to both growth control and establishment of PCP. Lawrence and colleagues (for a recent review, see Lawrence et al., 2007, 2008) proposed a hypothesis that if the direction of slope or vector of the gradient normally determines polarity, then the steepness could measure dimension and regulate growth, referred to as “Steepness model” (Fig. 3A). An important aspect of this model is that dimension sensing depends on the linear gradient of PVs established between the boundaries of a defined and growing population of cells whose maximum and minimum scalar values are constant (Fig. 3A); consequently, as the organ grows, the gradient becomes less steep (Fig. 3A). Thus, the steepness of the gradient is effectively a measure of dimension in one axis that could be extrapolated to every cell. The growth would cease when the slope of the linear gradient declines to a certain threshold value (Fig. 3A). This may be the simplest model to explain how a gradient of positional values is correlated to size and shape of organs.
INVOLVEMENT OF THE DS-FT SIGNALING SYSTEM IN PLANAR CELL POLARITY
Until approximately 2000, it had been speculated that a gradient of a morphogen would regulate PCP, growth, and dimension somehow. However, it turned out that morphogens such as Hedgehog (Hh), Decapentaplegic (Dpp), and Wingless (Wg) operate upstream of the systems controlling PCP and growth (Struhl et al., 1997; for a review, see Schwank and Basler, 2010). The Ds-Ft signaling system including Fj has been gradually figured out to be an upstream system of PCP regulation in the Drosophila wing (Adler et al., 1998; Zeidler et al., 2000) and eye (Yang et al., 2002; for a recent review, see Strutt, 2009). Yang et al. (2002) showed evidence that positional information regulating Ft action is provided by graded expression of Ds across the eye and the action of Fj, which is expressed in an opposing expression gradient and appears to modulate Ds function. These lines of evidence supported that Ft, Ds, and Fj are involved in establishing PCP in Drosophila organs. Recently, Lawrence and colleagues (Lawrence et al., 2007, 2008) proposed “the Ds/Ft model” to explain how the Ds-Ft system generates PCP (Fig. 3B). In the Drosophila adult abdomen, the dorsal epidermis is segmented and divided into a chain of anterior (A) and posterior (P) compartments. Most of the A cells make cuticular hairs or bristles that point posteriorly. As shown in Figure 3B (Casal et al., 2006), a gradient of Ds increases from anterior to posterior, while there is an opposing gradient of Fj. Although there is no gradient of Ft proteins, there should be a gradient of Ft activity, driven by the action of Fj on Ft. Active Ft could become stabilized in the membrane of one cell so that it can form trans-heterodimers with Ds in the next cell. Only those molecules of Ft and Ds that form trans-heterodimers are shown in Figure 3B. Thus, although the Ds protein gradient peaks posteriorly, the gradient of Ds molecules engaged in trans-heterodimers peaks anteriorly. The polarity of a cell might depend on a difference between the number of Ds molecules that are engaged in trans-heterodimers on the anterior and posterior faces of the cell, with the polarity of that cell pointing down the differential (from high to low, as shown in Fig. 3B). This model is supported by Repiso et al. (2010), showing that gradients of the Ds-Ft system are the main determinant of the PCP of larval denticles. Lawrence et al. (2008) extended the steepness model in combination with the Ds-Ft model to explain the mechanisms underlying the determination of organ size and PCP, which includes the Wts-Hpo signaling pathway as the mechanism for controlling growth (Rogulja et al., 2008; Willecke et al., 2008). However, Lawrence et al. (2008) also pointed out problems with the extended model: (1) The precise nature of the Ds-Ft gradient is unknown; although their Ds-Ft hypothesis posits that the numbers of Ds-Ft trans-heterodimers are the key variable, this is not proven. In their model, a difference in the number of heterodimers between the two faces of a cell may be the cue for planar polarity and the amount of that difference could represent the steepness, but there is no direct evidence. (2) It seems important for the steepness hypothesis that the gradient is more or less linear, as only a linear gradient could convey consistent information of dimension to all cells. But it is not known whether the gradient is linear. So far, however, this may be the simplest model to interpret experimental results related to the D-Ft signaling pathway.
RELATIONSHIP BETWEEN MORPHOGEN GRADIENTS AND THE DS-FT SIGNALING PATHWAY IN THE WING IMAGINAL DISC
In general, morphogens such as dpp/BMP, Wg/Wnt, Hh/Sonic hh (Shh), epidermal growth factor (EGF), and fibroblast growth factor play a fundamental role in limb patterning and growth in vertebrate and invertebrate limbs (for a recent review, see Schwank and Basler, 2010). Signaling cascades of such morphogens and interactions among them have been elucidated in limb development (for a recent review, see Butterfield et al., 2010). However, their functions in regulating growth and limb size remain elusive.
The Drosophila wing imaginal discs are patterned along the PD and AP axes by gradients of the morphogens Wg and Dpp, which are induced by Hh expressed in the posterior compartment (Fig. 4) (for a recent review, see Neto-Silva et al., 2009). Dpp and Wg gradients specify cell fate within the disc by activating the expression of target genes in a concentration-dependent manner. If Dpp and Wg also function as mitogens in a concentration-dependent manner, proliferation of disc cells is expected to depend on the distance to the sources, which are localized along the AP and dorsoventral (DV) boundaries (Neto-Silva et al., 2009; Schwank and Basler, 2010). Thus, higher cell proliferation would be observed near the boundaries. However, unexpectedly, Dpp and Wg cause uniform proliferation in the disc, suggesting the presence of other mechanisms that modify their gradients to produce uniform proliferation.
Several models have been proposed to interpret this phenomenon (Neto-Silva et al., 2009; Schwank and Basler, 2010). In this review, we focus on the Ds-Ft signaling pathway in the disc, because functions of Ft, Ds, and Fj have been extensively investigated using the Drosophila wing imaginal disc. In the wing disc, ds is highly expressed in proximal regions, and fj is expressed distally, whereas ft is expressed uniformly throughout the disc (Fig. 4) (Garoia et al., 2000; Mao et al., 2006; Reddy and Irvine, 2008). Since uniform misexpression of Fj and Ds in the wing disc inhibits cell proliferation and growth, graded expression of Ft regulators such as Fj and Ds could modulate Ft signaling by polarizing Ft activity within the cell (Rogulja et al., 2008; Willecke et al., 2008). Ft, Ds, and Fj function together during the earlier stages of wing and leg development to control growth along the PD axis (Mao et al., 2006). Viable mutations in each of these three genes give rise to flies with defects in the ratio of PD to AP growth of the wing and leg (Bryant et al., 1988; Clark et al., 1995; Villano and Katz, 1995; Adler et al., 1998; Zeidler et al., 2000). Wings and legs of flies with viable ft alleles display overgrowth mostly in the AP direction (Fig. 2) (Garoia et al., 2000; Mao et al., 2006). The defects caused by ft, ds, and fj mutations suggest that they are involved in regulating growth differences along both AP and PD axes.
In 2008, Rogulja and coworkers (Rogulja et al., 2008) demonstrated that Dpp signaling regulates the expression and localization of components of the Ft signaling pathway, and Ft signaling is required for the Dpp gradient to affect cell proliferation. Because the influence of Dpp on growth may be a function of slope or steepness of the Ds-Ft gradient, if the Ds-Ft gradient is constant, i.e., linear gradient, uniform cell proliferation would be observed. Since the Dpp gradient in the wing disc may provide PVs, positional information may be linked to cell proliferation via the activity of the Ds-Ft signaling pathway. Furthermore, since Ft and Ds influence planar cell polarity (PCP), the vector of their expression gradients may determine PCP along the PD axis (Casal et al., 2006; Aigouy et al., 2010; Harumoto et al., 2010).
Recently, Zecca and Struhl (2010) demonstrated the role of Wg in wing disc growth, which depends on the capacity of wing cells to recruit neighboring, non-wing cells into the wing primordium. Wg along the DV boundary induces and sustains the expression of vestigial (vg) (Fig. 4), which encodes a transcription factor that specifies the wing fate. vg-expressing boundary cells send a “feed-forward” signal that acts along with Wg to induce vg expression in neighboring cells. They found that vg generates the feed-forward signal by creating a steep difference in Ds-Ft signaling between wing and non-wing cells. This difference downregulates the activity of the Wts-Hpo signaling pathway in non-wing cells, leading to a burst of Yki activity and vg induction in response to Wg. The feed-forward signal entails the generation of opposing Ft (wing) and Ds (non-wing) signals. Zecca and Struhl (2010) also found that D acts by repressing the Wts-Hpo signaling pathway, leading to a burst of Yki activity. In turn, Yki forms a complex with Sd, a DNA-binding protein that directly activates Vg. Thus, they posited that steepness, as conferred by the local difference in Ds-Ft signaling across each cell, is not a direct reflection of the Wg slope, but rather an indirect response governed by Vg activity. Their results provided an alternative interpretation for growth control of the wing imaginal disc: “steepness,” as conferred by the local differential of Ft-Ds signaling across each cell, promotes wing growth not by functioning as a relatively constant parameter to set a given level of Wts-Hpo pathway activity in all cells (Zecca and Struhl, 2010). Further investigation is necessary to elucidate how Dpp and Wg regulate the size of the wing imaginal disc via Ds-Ft signaling.
RELATIONSHIP BETWEEN MORPHOGEN GRADIENTS AND THE DS-FT SIGNALING PATHWAY IN THE CRICKET LIMB BUD
As well as in the Drosophila leg imaginal disc, Hh, Wg, Dpp, and Egfr ligand are expressed as morphogens in the cricket Gryllus limb bud (for a review, see Nakamura et al., 2008a). hh is expressed in the posterior compartment, wg is in the ventral side of the AP boundary, and dpp is in the dorsal side of the AP boundary (for a review, see Kojima, 2004). The most distal cells that are influenced by both Wg and Dpp produce Egfr ligand at the tip of the limb bud, inducing growth along the PD axis (Campbell, 2002; Galindo et al., 2002). In Drosophila, the Egfr ligands are known to be Vein (Vn), Spitz, and Gurken, among which Vn acts as a morphogenetic signal to establish new distal fates at the tip of the developing leg that is located at the center of the leg imaginal disc (Campbell, 2002; Galindo et al., 2002, 2005). In Gryllus limb bud, Egfr is also likely to be involved in establishment of the PD axis (Nakamura et al., 2008b). However, since no homolog of vn has been identified in other insect orders, an Egfr ligand might be a Tgfα-like protein (Lynch et al., 2010).
Gryllus ft (Gb'ft) is highly expressed in the proximal region of each primordium of leg segments, femur, tibia, and tarsus, and shows a PD gradient (Fig. 5A) (Bando et al., 2009). Gb'ds is highly expressed in the distal region of each leg segment primordium (Fig. 5A). Gb'fj is highly expressed in the proximal region of each segment primordium (Bando et al., 2009). Putative linear distributions of their proteins in a leg segment are schematically illustrated in Figure 5B. Expression patterns of the three genes in the cricket leg bud are different from those in the wing imaginal disc. Although no intensive study has investigated their functions in the leg imaginal disc, these may be similar because phenotypes of their mutants are similar to the wing phenotypes (Mao et al., 2006). Although there is no evidence for a Ft-Ft or Ds-Ds trans-homodimer formation in Drosophila cells, we hypothesized the presence of the trans-homophilic interactions of Ft or Ds in their higher concentration region in addition to the heterophilic interaction, as shown in Figure 5C, based on their expression patterns and their structures implying the presence of both homophilic and heterophilic interactions (Halbleib and Nelson, 2006). In Gryllus, RNA interference (RNAi) has been used to analyze gene functions, but no significant phenotype was observed in limb buds analyzed for the three genes. Since the Egfr ligand acts as a morphogen providing the positional information along the PD axis in the leg imaginal disc, some relation between EGF and the Ds-Ft signaling pathway may exist. This speculation has been supported by studies on tumorigenesis by Hamaratoglu et al. (2006).
Mer, the protein product of the Neurofibromatosis type-2 (NF2), acts as a tumor suppressor in mice and humans (Hamaratoglu et al., 2006). Mer is an adaptor protein with a FERM domain. In Drosophila, it functions with Ex (Hamaratoglu et al., 2006). Hamaratoglu et al. (2006) showed that Drosophila Mer and Ex are required for proliferation arrest and apoptosis in developing imaginal discs and components of the Wts-Hpo signaling pathway. Recently, Benhamouche et al. (2010) demonstrated that the expansion phenotype of facultative bipotential liver progenitor cells (oval cells) observed in Nf2 (Mer) liver mouse models might be mediated through Mer's regulation of Egfr signaling, indicating that Mer can control the abundance and signaling of membrane receptors such as Egfr, thus suggesting a possible link between Egfr and Wts-Hpo signaling (for a recent review, see Yi and Kissil, 2010). Mer regulates the internalization and signaling of Egfr in response to the formation of cell–cell contacts (Yi and Kissil, 2010). Furthermore, there is evidence of modulation of Egfr signaling by the Wts-Hpo signaling pathway. Ectopic expression of YAPS127A (YAP is a mammalian homolog of Yki.), which is the constitutively active form of human YAP in Drosophila, specifically induces vn expression, leading to an increase in Egfr signaling (J. Zhang et al., 2009). The intertwined nature of the relationship between the Egfr and Wts-Hpo signaling pathways was corroborated further by genetic studies in flies (Yi and Kissil, 2010). The overgrowth phenotype caused by increased activity of yki can be partially suppressed by loss-of-function mutants of Egfr or vn, whereas reduced Yki activity resulting from overexpression of its negative regulators, Wts and Hpo, is enhanced by Egfr or vn mutants (J. Zhang et al., 2009). These results in Drosophila imply the existence of a loop relationship starting from the Egfr ligand (Vn in Drosophila) to the EGF ligand via Egfr, Mer-Ex, Wts-Hpo signaling pathway, and Yki. These signaling cascades may play important roles in limb development and regeneration (Yi and Kissil, 2010).
ROLES OF THE DS-FT AND WTS-HPO SIGNALING PATHWAYS DURING LEG REGENERATION
Recently, the Ds-Ft and Wts-Hpo signaling pathways were demonstrated to be actually involved in intercalary leg regeneration as well as normal regeneration after leg amputation using the cricket G. bimaculatus (Bando et al., 2009). The greatest advantage of the cricket system is that regeneration-dependent RNAi (rdRNAi) can be used for loss-of-function analyses. rdRNAi is a type of RNAi that occurs specifically after leg amputation in cricket nymphs that have been injected with double-stranded RNA for a target gene (Nakamura et al., 2008a). In this system, when the metathoracic (T3) tibia of the third instar nymph is amputated (Fig. 1B), adult leg restoration occurs in approximately 40 days (6 ecdyses). As observed in the limb bud, an Egfr ligand may be then induced by Gb'Dpp and Gb′Wg in a blastema composed of epithelial stem cells, which begins to undergo rapid proliferation to restore the lost portion in the fourth instar (Figs. 1B; see Fig. 7) (Mito et al., 2002; Nakamura et al., 2008b). In the fifth instar, the tibiae, tibial spurs, tarsi, and tarsal claws are restored in a miniature form (Fig. 1B). In the seventh instar, the amputated legs restore the missing portion to regain a nearly normal appearance (Figs. 1B and 6). Since no leg regeneration was observed after amputation in the case of rdRNAi against Gb'armadillo (Gb'arm) (see Fig. 8C), the canonical Wnt pathway is likely involved in the initiation of regeneration (Nakamura et al., 2007). Using the rdRNAi method, Bando et al. (2009) examined the function of genes involved in the Ds-Ft signaling pathway. On the basis of the data in Drosophila (Reddy and Irvine, 2008), they selected a set of 23 genes in the Ds-Ft signaling pathway (Fig. 1C) and found at least 15 genes are involved in leg regeneration, indicated by yellow in Figure 1C (Bando et al., 2009). In this review, we focus mainly on the functions of the Gryllus homolog of ft (Gb'ft), Gb'ds, Gb'ex, or Gb'Mer.
Most typical phenotypes induced by rdRNAi against Gb'ft or Gb'ds are short and thick legs (Fig. 6) (Bando et al., 2009), while the regeneration legs of the control and the contralateral legs appear normal (Fig. 6A–D). Interestingly, although the regenerated leg becomes very short and thick, the distal structures, tarsi, and claws are well formed (Fig. 6). These phenotypes are consistent with the Drosophila ft and ds mutants (Fig. 2) (Mao et al., 2006). These typical phenotypes suggest that Gb'ft and Gb'ds are involved in (1) formation of the Ds-Ft gradient (Fig. 5B and C), (2) regulation of growth through the Wts-Hpo signaling pathway (Fig. 7B), (3) regulation of proliferation ratio along the PD and circumferential axes, (4) epimorphosis, and (5) determination of leg segment length.
Formation of the Ds-Ft Gradient
Since Gb'ft expresses in the proximal side of the regenerating leg segment, while expression of Gb'ds is localized in the distal side, as observed in the limb bud (Fig. 5B), it was reasonable to speculate that phenotypes obtained by RNAi would be opposite to each other, depending on amputated position. However, their phenotypes (short and thick) were similar to each other and did not depend on the amount of Gb'ds or Gb'ft transcripts (Fig. 6). This is consistent with the fact that Drosophila mutant phenotypes of both ft and ds have short and thick adult legs, despite the fact that ft and ds have distinct expression patterns in Drosophila imaginal discs (Garoia et al., 2000; Ma et al., 2003). These results were not expected if Ds and Ft acted as a traditional ligand and receptor. Thus, Ds and Ft are likely to act together as a pair and form a gradient of the activity of the Ds-Ft signaling pathway (Bando et al., 2009) (Fig. 5B and C). This is consistent with the “boundary model” (Willecke et al., 2008) in which cells respond to disparities in the levels of Ds-Fj between cells. It is interesting to note that, differing from the Gb'ft expression pattern, expression of ft in the Drosophila wing imaginal disc is rather uniform. These observations imply that there are several mechanisms to provide positional information for cells with different combinations of Ds, Ft, and Fj. So far, however, there is no direct evidence showing that the Ds-Ft gradient is linear and provides positional information.
Regulation of Growth Through the Wts-Hpo Signaling Pathway
To analyze gene functions during Gryllus leg regeneration, rdRNAi experiments were performed against genes involved in the Wts-Hpo signaling pathway (Bando et al., 2009) (yellow and gray in Fig. 1C). Although no RNAi phenotype was observed for Sav-Hpo-Mats (gray in Fig. 1C), phenotypes exhibiting abnormal regeneration were observed by RNAi against target genes in the Wts-Yki signaling pathway (yellow in Fig. 1C). Since Hpo transcripts were actually depleted by RNAi (Bando et al., 2009), the Sav-Hpo-Mats pathway might be dispensable during leg regeneration. The Ds-Ft signaling pathway may regulate leg size through the Wts signaling pathway, as shown in Figure 7 (Bando et al., 2009). This is consistent with the observation that the Wts-Hpo signaling pathway is involved in an intrinsic mechanism that restricts organ size (Edgar, 2006; Dong et al., 2007; Pan, 2007, 2010; Yin and Pan, 2007; Lawrence et al., 2008; Zhao et al., 2010). Thus, when neighboring cells express different amounts of Ds or Ft during regeneration (Fig. 5C), Wts-Hpo target genes appear up-regulated, thereby contributing to the control of leg size, which is consistent with a model that differences in Ds-Ft levels between cells regulate the Hippo pathway (Cho et al., 2006; Lawrence et al., 2008; Rogulja et al., 2008; Willecke et al., 2008).
On the other hand, the most typical phenotypes induced by rdRNAi against Gb'ex, or Gb'Mer are longer legs than control legs (Fig. 6) (Bando et al., 2009). The regenerated legs of adults subjected to Gb'ex and Gb'Mer rdRNAi are longer than normal control legs (Figs. 6 and 8F), and Gb'ex and Gb'Mer regulate cell proliferation induced by the presence of positional disparities (Fig. 7). These results suggest that Gb'ex and Gb'Mer are also involved in the allometric growth of the leg segment (Fig. 7). In Drosophila, Ex and Mer negatively regulate cell growth and proliferation through the Wts-Hpo signaling pathway (Hamaratoglu et al., 2006; Pellock et al., 2007). In mammalian cells, Nf2/Mer is known to be a critical regulator of contact-dependent inhibition of proliferation (Curto and McClatchey, 2008). Thus, Ex and Mer activities may regulate contact-dependent inhibition of proliferation via the Wts signaling pathway to restore the proper leg segment size during regeneration (Fig. 7). As previously mentioned, since these factors are involved in the regulation of Egfr signaling, Egfr as well as its ligand induced by Dpp and Wg should be regulated through Ex and Mer during regeneration. Recently, another upstream component of the Hpo signaling pathway, Kibra, was identified (Yu et al., 2010). Kibra functions together with Mer and Ex in a protein complex localized to the apical domain of epithelial cells, and this protein complex regulates the Wts-Hpo kinase cascade via direct binding to Hpo and Sav (Yu et al., 2010). Furthermore, as an upstream component of Ex, the transmembrane protein Crumbs (Crb), a determinant of epithelial apical-basal polarity in Drosophila embryos, was also identified (Ling et al., 2010). These new factors are likely to function as apical components of the Wrs-Hpo signaling pathway. Since they may be involved in Gryllus leg regeneration (our unpublished data), the Mer-Ex-Kibra complex functions through Crb and Ds-Ft, which may respond to positional disparities in regeneration.
Regulation of Proliferation Ratio Along the PD and Circumferential Axes
On using rdRNAi against Gb'ft and Gb'ds, legs become thick because of overproliferation along the circumferential axis and low proliferation along the PD axis. This phenotype suggests that Ds-Ft may be involved in determining the ratio of PD to circumferential proliferation, i.e., leg shape. Since legs subjected to Gb'd rdRNAi exhibit short and not thick legs (Bando et al., 2009), and Gb'd is epistatic to Gb'ft and Gb'ds, Gb'D may mediate the components of the Ds-Ft signaling pathway along the PD axis, thereby controlling leg size (Fig. 7). The enlarged phenotype of nymphs subjected to Gb'wts rdRNAi is suppressed by rdRNAi against Gb'd in Gryllus, indicating that Gb'd is downstream of Gb'wts (Fig. 7). On the other hand, Baena-Lopez et al. (2005) demonstrated that a relationship exists between oriented cell division and morphology of the organs by showing that oriented cell division is controlled by Ds-Ft signaling in wing imaginal discs. Thus, the proliferation ratio may depend on the oriented cell division controlled by the Ds-Ft signaling system.
To describe the two major types of regeneration, Morgan (1901) coined the terms morphallaxis and epimorphosis. Morphallaxis refers to regeneration without proliferation at the cut surfaces, while epimorphosis refers to regeneration in which proliferation of a material precedes the development of the new part. However, the molecular mechanisms underlying the regulation of cellular proliferation remain elusive. Insect leg regeneration is known as epimorphosis, i.e., the generation of a new leg at the amputated site via blastemal cell proliferation (Bohn, 1965). Bando et al. (2009) showed that the Ds-Ft signaling pathway is essential for epimorphosis. The short phenotype indicates that remodeling of a pre-existing stump takes place through a morphallaxis-like process, making the regenerated legs shorter than normal (Figs. 6–8D).
Determination of Leg Segment Length and Shape
The size of each leg segment normally scales with the overall body size, a phenomenon known as allometry (Shingleton et al., 2007). The size of regenerated legs in the phenotypes does not scale with the overall body size (Fig. 6). Furthermore, the size of the regenerated legs depends upon the site of tibial amputation (Bando et al., 2009). These results indicated that Ds-Ft signaling pathway including Mer-Ex-Kbr and Wts-Hpo pathways is involved in intrinsic determination of the leg segment size and shape.
ROLES OF THE DS-FT SIGNALING PATHWAY DURING INTERCALARY LEG REGENERATION
In intercalary regeneration (Fig. 8G, I), most of the regenerated cells are derived from the donor or host with a more distal positional identity (Bohn, 1976; French et al., 1976), suggesting that regenerating cells inhibit the proliferation of more proximal cells at the junction, a phenomenon known as “distal preponderance” (Brockes and Kumar, 2008). The extent and origin of a regenerated portion can be determined by observing its surface morphology. In case of reverse intercalation, PCP of regenerates along the PD axis can be determined by the orientation of bristles (Fig. 8I). Bando et al. (2009) performed transplantation experiments by grafting an amputated T2 piece to a T3 host in the same animal. In nymphs subjected to Gb'ft or Gb'ds rdRNAi, although host-graft jointed regions enlarge, neither normal nor reverse intercalary regeneration occurs to restore the missing region. However, in nymphs subjected to Gb'fj rdRNAi, intercalary regeneration occurs normally in both grafting experiments. These results indicate that Gb'ft and Gb'ds are essential for intercalary regeneration (Fig. 8H, J). In nymphs subjected to Gb'ex or Gb'Mer rdRNAi, both normal and reverse intercalary regeneration took place, but regenerated cells were derived from both distal and proximal pieces, supporting the hypothesis that Gb'ex and Gb'Mer are involved in the directional contact-dependent inhibition of proliferation, leading to a proximal respecification (Bando et al., 2009). These results strongly indicated that the Ds-Ft signaling pathway plays important roles in intercalary as well as normal regeneration.
THE Ds-Ft SIGNALING PATHWAY LINKS POSITIONAL AND ALLOMETRIC INFORMATION TO DETERMINE REGENERATED LEG SIZE AND SHAPE: INTERPRETATION ACCORDING TO THE Ds-Ft STEEPNESS MODEL
A widely accepted model for leg regeneration is the intercalation model, which is based on positional information (for a review, see Brockes and Kumar, 2008). This model is based on the intercalation of new structures so as to re-establish the continuity of PVs during regeneration. However, explaining the changes in leg size based on this model is difficult. Thus, the intercalation model should be extended to include the control of growth and tissue size. Several models have been proposed to explain regulation of organ size (for a recent review, see Schwank and Basler, 2010). The Ds-Ft steepness model was employed to interpret phenomena observed in leg development and regeneration (Bando et al., 2009).
A modified Ds-Ft steepness model for leg regeneration acts as follows (Bando et al., 2009). Nymphal leg regeneration depends on the following two major processes: proliferation and differentiation of blastemal cells (indicated by yellow in Fig. 8B) and growth of the pre-existing stump (indicated by orange in Fig. 8A). In each of these processes, new positional identities are specified in relation to new segment boundaries. According to the Ds-Ft steepness model, in normal regeneration, a very steep gradient should be formed in the regenerating blastema (Fig. 8B). The regenerate may grow so as to restore the normal pre-existing steepness. Reassignment of positional identities after amputation will correlate with a similar resetting of the minimum Ds-Ft scalar value, and, thus, the results are consistent with the steepness hypothesis.
Growth of the pre-existing stump is a normal component of leg growth in which the pre-existing stump cells proliferate according to some allometric signals, which may be related to the maximum scalar value and the slope of the gradient, whilst keeping their original positional information similar to that in the truncated leg of adults subjected to Gb'arm rdRNAi (Fig. 8C) (Nakamura et al., 2007).
In the absence of proliferation and differentiation of blastemal cells, as observed in the legs subjected to Gb'ft rdRNAi, the minimum scalar value, which is the most distal PV, would be established at the site of amputation, and the Ds-Ft gradient would be expected, in turn, to shift down with the same slope as the pre-existing one (indicated by pink, Fig. 8D and E). The Ds-Ft steepness model provides an explanation for the observation that the final leg size depends on the amputated position, if we assume that the gradient shifts down with the same slope, resulting in that cells at an amputated position have the minimum scalar value (indicated by pink, Fig. 8D and E). Thus, the observed respecification of regenerated legs induced in legs treated with rdRNAi against Gb'ft or Gb'ds is as would be predicted by the Ds-Ft steepness model. Thus, the Ds-Ft gradient likely functions to link positional and allometric information to regulate leg segment growth, and Ex-Mer activity is related to a threshold value of the slope of the gradient that determines the time for growth cessation (Fig. 8F). All rdRNAi phenotypes observed in the cricket leg regeneration, including intercalary regeneration (for details, see Fig. 8G–J), are consistent with the Ds-Ft steepness model for regeneration (Fig. 8) (Bando et al., 2009).
RELATIONSHIP OF THE DS-FT SIGNALING PATHWAY WITH VERTEBRATE LEG REGENERATION
In mammals, four types of Ft proteins, namely Ft1, Ft2, Ft3, and Ft4, have been identified (Rock et al., 2005; Tanoue and Takeichi, 2005). Among them, Ft4 is the vertebrate counterpart of Drosophila Ft based on their sequence similarity (Rock et al., 2005; Matakatsu and Blair, 2006). Recently, Saburi et al. (2008) reported phenotypes of Ft4-null mice. Loss of Ft4 leads to defects in the polarity of the hair cells of the inner ear as well as defects in tissue elongation in the cochlea and neural tube, indicating that Ft4 functions to regulate PCP in vertebrates. Furthermore, loss of Ft4 leads to defects in oriented cell division in the developing kidney, indicating that controlling oriented cell division is a conserved function of the Ft family of cadherins and suggesting a mechanism for the formation of tubular dilations and cysts observed in Ft4 mutants. Saburi et al. (2008) demonstrated that the Ft-Ds-Fj PCP signaling cassette is conserved in vertebrates and provides experimental evidence to support the hypothesis that polycystic kidney disease can be caused by defective PCP signaling. On the other hand, two types of Ds, Ds1 and Ds2, have been identified in mammals (Rock et al., 2005). At the mRNA level, the former is broadly expressed in mouse embryos, whereas the latter exhibits restricted expression (Rock et al., 2005). Until now, Ft1 is expressed in rat limb buds (Ponassi et al., 1999), and expression of other Ft proteins in mouse limb buds has not yet been described.
Recently, da Silva et al. (2002) found that a cell surface protein, a glycosylphosphatidylinositol-anchored three-finger protein (TFP) Prod1, is implicated in the local cell-to-cell interactions mediating PD positional identity in leg regeneration of the newt (for a review, see Brockes and Kumar, 2008). Recently, a salamander Prod1 was identified and it turned out that the salamander Prod1 is anchorless (Blassberg et al., 2011). Since the salamander Prod1 interacts with the newt Egfr (Blassberg et al., 2011), the positional information in cells may be provided through phosphorylation of extracellular regulated kinase 1 and 2.
Prod1 interacts with the anterior gradient protein (nAG) that is widely expressed in many animal species. nAG has been identified as a secreted ligand for Prod1 in a yeast two-hybrid screen, and recombinant proteins have been subsequently demonstrated to form a complex with Prod1 (Kumar et al., 2007). nAG appears to play a key role in the nerve dependence of limb regeneration. The division of limb blastemal cells depends on the regeneration of axons in the major peripheral nerve branches following amputation (Singer, 1952). These axons upregulate nAG expression in the Schwann cells of the distal nerve sheath, and, subsequently, in dermal gland cells underlying the wound epithelium. A secreted nAG binding to Prod1 is likely to stimulate Egfr activity. It is interesting to note that in regeneration of Zebrafish fins, the oncogenes ErbB2 and ErbB3, two members of the Egfr family, are essential for mounting a successful regeneration response (Rojas-MuÒoz et al., 2009). As previously mentioned, a component of the Wts-Hpo pathway, Mer, may interact with Egfr to regulate cell proliferation during regeneration. In this case, although the ligand of Egfr may depend on species, proliferation of blastemal cells might be commonly regulated through the Ds-Ft and Wts-Hpo signaling pathways. Until now, an insect homolog of Prod1 has not been identified. However, since the Ds-Ft signaling pathway is conserved in vertebrates (for recent reviews, see Grusche et al., 2010; Halder and Johnson, 2010; Pan, 2010), this pathway might be involved in vertebrate leg regeneration, probably interacting with the Prod1 system.
Studies with the cricket have indicated that leg size and shape are regulated through the Ds-Ft signaling pathway, including the Ex-Mer and Wts-Hpo signaling pathways, during regeneration and have provided cues for understanding the molecular mechanisms underlying regeneration that was first described a century ago. Many questions such as what ultimately limits leg size and whether steepness of the Ds-Ft gradient promotes the growth of other organs and determines organ sizes remain unanswered. Of course, Drosophila offers an ideal model system for such studies: The wide range of genetic and genomic approaches available for use in flies has helped in elucidation of the mechanisms underlying determination of leg size and leg or wing disc regeneration (Bergantiños et al., 2010; Sun and Irvine, 2010). Recent evidence encouragingly suggests that the Wts-Hpo signaling pathway and Yki may act through the regulation of cell fate to control organ size in vertebrate systems (Dong et al., 2007; Lu et al., 2010; Pan, 2010; Zhao et al., 2010). Further work is needed, however, to identify the upstream signals and downstream master regulatory genes of the Ds-Ft signaling pathway in limb development and regeneration.
We thank anonymous reviewers for helpful comments that greatly improved this review. We are grateful to the members of our lab for providing data for crickets. This review is partly supported by a grant from the Ministry of Education, Culture, Sports, Science and Technology of Japan (22124003 and 22370080) to T. Bando, T. Mito, H. Ohuchi, and S. Noji.