JOURNAL TRANSCRIPT

---

Developmental Cell, Vol. 4, 231–240, February, 2003, Copyright 2003 by Cell Press

EGF Signaling Patterns the Feather Array by Promoting the Interbud Fate Radhika Atit,1 Ronald A. Conlon,2 and Lee Niswander1,* 1 Howard Hughes Medical Institute and Molecular Biology Program Memorial Sloan-Kettering Cancer Center New York, New York 10021 2 Department of Genetics Case Western Reserve University University Hospitals Cleveland Cleveland, Ohio 44106

Summary Feather buds form sequentially in a hexagonal array. Bone morphogenetic protein (BMP) signaling from the feather bud inhibits bud formation in the adjacent interbud tissue, but whether interbud fate and patterning is actively promoted by BMP or other factors is unclear. We show that epidermal growth factor (EGF) signaling acts positively to establish interbud identity. EGF and the active EGF receptor (EGFR) are expressed in the interbud regions. Exogenous EGF stimulates epidermal proliferation and expands interbud gene expression, with a concurrent loss of feather bud gene expression and morphology. Conversely, EGFR inhibitors result in the loss of interbud fate and increased acquisition of feather bud fate. EGF signaling acts directly on the epidermis and is independent of BMP signaling. The timing of competence to interpret interbud-promoting signals occurs at an earlier developmental stage than previously anticipated. These data demonstrate that EGFR signaling actively promotes interbud identity.

Introduction The developing avian feather bud array is an established model system that permits the study of two major processes in development, cell fate determination, and pattern formation. In each tract, the hexagonally positioned feather bud cells are surrounded by an intervening region of interbud cells. As the feather array develops, there is an exchange of inductive, repressive, and permissive signals that guide a cascade of developmental events culminating in the formation of feather bud and interbud regions. Perturbation in cell fate decisions, proliferation, migration, or cell-cell communication results in changes in the growth and patterning of the feather bud array. The mechanisms that guide establishment of the hexagonal feather array are not well understood. Although the overall sequence of interactions necessary for bud formation has been outlined, only some of the molecules *Correspondence: [email protected]

that regulate the various parameters in these models are defined. The prevailing model integrates the mechanisms of reaction-diffusion, lateral inhibition, and the propagation of a morphogenetic wave that progresses across the field (Jung and Chuong, 2000). It has been proposed that an early inducing signal originates from the embryonic midline in the dermis and moves as a morphogenetic wave to progressively stimulate the overlying epidermis to form placodes, or thickened patches of epithelium (Tanda et al., 1995). Studies on time determination of feather bud position have revealed that the morphogenetic wave precedes placode formation by one row (Davidson, 1983). Consequently, feather bud development occurs sequentially from the midline of the dorsal tract, as the temporal development of competence closely precedes the morphogenetic wave and enables the skin to interpret initial bud-promoting signals mediated by the FGF pathway (Davidson, 1983; Jung et al., 1998). Epidermal signals from the placode stimulate the dermal cells to migrate and condense under it (Desbiens et al., 1991). Reciprocal and coordinated interactions between the epidermis and dermis lead to the morphogenesis of individual feather buds. In addition to promoting signals, each bud produces an inhibitory signal that acts to inhibit bud formation in surrounding cells such that each bud is surrounded by a region of interbud. This inhibition appears to be mediated by bone morphogenetic proteins (BMP) and refined by Notch signaling (Crowe et al., 1998; Noramly and Morgan, 1998). Bmp-2 is expressed within the bud and is speculated to transmit an inhibitory signal across several cells. New buds can form beyond the influence of the inhibitory signal emanating from the bud. As the primary inducing bud signal within the dermis moves laterally past the inhibitory zone, the next row of buds forms. BMP-2 acts early in the process of lateral inhibition but then is no longer required (Noramly and Morgan, 1998). It has been suggested that the bud fate is restricted by another, later mechanism that acts on nonplacode and placode cells (Noramly and Morgan, 1998). Many genes that are expressed within the bud have been characterized to guide the growth and patterning of the feather bud. However, genes that reside within the interbud and actively guide interbud identity and development remain elusive. EGF is widely known for its role in cell proliferation, but it is also involved in fate determination and survival. In this study, we have examined the role of EGF signaling in cell fate determination and patterning in the feather array. EGF is a ligand for the tyrosine kinase EGF receptor (EGFR) encoded by the erbB1 gene (Olofsson et al., 1986). During vertebrate development, EGF and its receptor are expressed and functionally active at sites of epithelial-mesenchymal interactions such as the developing otic vesicle, craniofacial primordia, kidney, lung, tooth, and limb (Canoun et al., 1993; Diaz-Ruiz et al., 1993; Kronmillier et al., 1991; Partanen and Thesleff, 1987). During development of mouse skin, EGF is ex-

Developmental Cell 232

pressed in the basal keratinocytes of the epidermis and the outer sheath of hair follicles (Partanen and Thesleff, 1987). Transgenic mice that have diminished EGFR signaling in the skin have aberrant hair follicles (Fowler et al., 1995). Studies of EGF transgenic and knockout mice have only examined the hair follicles at later stages in development. EGF has been reported to stimulate epithelial cell proliferation and perturb hair and feather bud growth in explants (Cohen, 1965; Kashiwagi et al., 1997). These various studies provide some insight into the complicated regulation of hair follicle development, but the role of EGF signaling in the developing skin still remains largely undefined. In the present study, we found that EGF and activated EGFR are expressed in skin prior to formation of the epithelial placodes, and then are downregulated in the epidermal placode and maintained in the interbud through later stages of feather array formation. This suggests a role in interbud development and in defining boundaries between interbud and bud. We tested the role of EGF signaling during early feather bud development in skin explants (Jung et al., 1998). Treatment with exogenous EGF was found to suppress bud formation and expand interbud marker expression, whereas EGFR inhibitors caused bud fusions. Localized and shorter treatment with EGF and EGFR inhibitor to the lateralmost region of the skin perturbs interbud development, indicating the lateral tissue is receptive to the interbudpromoting signal. This study suggests that the zone of competence to interpret interbud-promoting signals is established significantly prior to the advancement of the morphogenetic wave. This indicates that the first step in the acquisition of interbud fate occurs before the known production of inhibitory signals from the placode. Our data strongly suggest that EGF signaling plays a key role in promoting the interbud fate. Results EGF Signaling Is Localized to the Endogenous Interbud In order to determine the role of EGF signaling in early feather bud development, the expression of EGF and the phosphorylated, active form of EGFR was examined. Immunostaining of stage 32 skin (Hamburger and Hamilton, 1951) with an antibody against EGF revealed the presence of EGF in the interbud epidermis and dermis during early bud development (Figure 1A). EGF expression remained in the interbud throughout early bud development (Figure 1B; stage 37 was the latest examined). An antibody to the phosphorylated form of the EGFR was used to determine where EGF signaling was active. Activated EGFR and EGF were detected in the lateral developmentally younger preplacode skin of stage 30–36 skin (Figure 1C and data not shown). As the epithelial placodes formed, activated EGFR expression was decreased in the placode compared to the region adjacent to the placode (Figure 1D). During early bud stages, activated EGFR was expressed throughout the interbud region but not in the bud (Figure 1E). In the short bud stage (stage 36), active EGFR was most strongly expressed at the junction of interbud and feather bud (Figure 1F). A similar expression pattern was seen in stage 30 skin cultured in control conditions for 3 days

(Figure 1G). If skin was cultured in the presence of 300 ng/ml EGF, activated EGFR was detectable across the entire thickened epidermis (Figure 1H). In skin treated with AG1478, an inhibitor of EGFR (see below), phosphorylated EGFR was undetectable (Figure 1I). Nonspecific staining was minimal with either the phosphorylated EGFR antibody or secondary goat anti-mouse antibody, as seen in control immunostainings (Figure 1J; no secondary antibody). These data imply that EGFR signaling plays an important role in the interbud. EGF Signaling Promotes Interbud Fate At HH stage 30, three rows of nascent buds are visible in the dorsal feather tract. Dorsal skin derived from HH stage 30 was cultured for 3 days in the presence of varying amounts of purified EGF protein (Figure 2). Skin cultured in the absence of EGF (n ⫽ 90) recapitulated the spatial aspects of in vivo feather development (Figure 2A). With the finite size of the explant limiting further progression, 10–14 rows of feather buds emerge in each explant within 3 days. A 10 ng/ml dose of EGF (n ⫽ 14) did not significantly perturb bud patterning or outgrowth, but the spacing between buds at the lateral edge was slightly increased (Figures 2B and 2J). At 50 ng/ml of EGF, increased interbud spacing was readily observed in the lateral edges of the explant (Figures 2C and 2J). Higher concentrations of EGF progressively inhibited feather bud formation and outgrowth lateral to the preexisting rows from stage 30 skin. At 100 ng/ml EGF (n ⫽ 15), the number of buds was decreased, and spacing between buds was significantly increased in the lateral-most rows (Figures 2D and 2J). Interbud spacing was analyzed by calculating the percent interbud area, and an increase in interbud area correlated with an increase in the dose of EGF (Figure 2J). At 200 ng/ml (n ⫽ 15), the two to three rows of preexisting feather buds were visible, but all other future bud development was inhibited. Severe inhibition of outgrowth of existing buds was also observed (Figure 2E). At 300 ng/ml EGF (n ⫽ 45), the highest concentration used, no new feather buds were visible, and preexisting feather bud primordia normally present on stage 30 skin were no longer apparent (Figure 2F). Tissue recombination experiments with stage 31 skin clearly demonstrate that the effect of EGF is directly on the epidermis. Epidermis treated for 14 hr with EGF recombined with untreated dermis led to a loss of new bud formation or scattered feather buds with large intervening interbud areas (Figure 2I and data not shown; n ⫽ 11). In contrast, EGF-treated dermis recombined with untreated epidermis (n ⫽ 8) produced buds comparable to the control cultures (n ⫽ 6; Figures 2H and 2G, respectively). Previous studies have also shown that the effect of EGF on the epidermis is direct and not due to secondary consequences of an initial effect on the dermis (Cohen, 1965). Histological analysis of control skin cultures showed normal epidermal and dermal arrangement of the feather bud and interbud (Figure 3A). In contrast, 300 ng/ml EGF led to inhibition of feather bud formation and outgrowth (Figure 3B), epidermal thickening with invaginations (Figure 3B, arrow), and keratinization of subperidermal layers (Figure 3B, arrowheads). The regular periodic arrangement of bud and interbud was completely absent.

EGF Signaling Promotes Interbud Fate 233

Figure 1. EGF and Phosphorylated EGFR Are Localized in the Interbud Region EGF protein expression is in the interbud epidermis and dermis (arrows) in the cross-section of stage 32 (A) and stage 36 skin (B). The active form of the EGFR is detectable with a phosphorylated EGFR-specific antibody (C–I). Phosphorylated EGFR (Y-EGFR) is detectable in skin epidermis prior to placode formation (C) and then is downregulated in the placode and maintained in the adjacent epidermis ([D]; arrows mark higher level expression in the nonplacode epidermis). During dermal condensation, EGFR is present in the interbud epidermis (arrows) of stage 31 skin (E). In the short bud stage, phosphorylated EGFR is detectable in the interbud and near the interbud-bud junction in stage 36 skin ([F], arrows) and in skin cultured under control conditions ([G], arrows). Some staining for EGF and Y-EGFR is seen in the periderm of older buds (A, B, F, and G). EGFR is detectable throughout the greatly thickened epidermis of EGF-treated skin (H) and not visible in the AG1478-treated skin after 3 days in culture (I). In controls (no secondary or primary antibody), no nonspecific staining is observed (J). All pictures were taken at the same magnification; hatched black lines demarcate the epidermis (e) and dermis (d).

Our data with purified recombinant EGF at much lower doses confirm results from a previously reported study using partially purified EGF from the mouse submaxillary gland (Cohen, 1965). Skin treated with 300 ng/ml EGF was analyzed further for bud identity by examining the expression of molecular markers. The bud markers Bmp-2, Bmp-4, Dermo-1, Fgf-4, and Shh were expressed in control-cultured skin (Figure 3C and data not shown). In EGF-exposed skin, Bmp-2, Bmp-4, Dermo-1, Fgf-4, and Shh expression patterns were severely downregulated in the skin within 18 hr (Figure 3D and data not shown). Expression of an interbud marker was also examined. Collagen I is expressed in the interbud dermis of developing avian skin (Chuong et al., 1996) and control culture explant skin (Figure 3E). In EGF-treated skin, collagen I expression was continuous throughout the dermis (Figure 3F).

Proliferation profiles are distinctive between the control and EGF-treated skin. Proliferation in skin was examined by incorporation of BrdU. Bud dermis proliferates significantly more than the interbud dermis (Jung and Chuong, 2000). In control cultures, numerous BrdUpositive proliferating cells were visible in the bud dermis and epidermis, and relatively less proliferation was seen in the interbud dermis (n ⫽ 6; Figure 3G). In skin treated with EGF for 3 days, the basal keratinocytes in the thickened epidermis and in the invaginations were actively proliferating (Figure 3H). The notable decrease in the dermal proliferation in EGF-treated cultures resembled the interbud proliferation profile (n ⫽ 7). Skin cultures treated with EGF for shorter time periods yielded similar results (data not shown). These results demonstrate that increased and ubiquitous EGF treatment leads to loss of bud markers and an expansion of interbud marker.

Developmental Cell 234

Figure 2. EGF Treatment Affects Epidermis and Promotes Interbud Fate Stage 30 embryonic chick dorsal skin was cultured for 3 days without exogenous EGF (A) or with 10 ng/ml (B), 50 ng/ml (C), 100 ng/ml (D), 200 ng/ml (E), and 300 ng/ml (F) EGF. (B–D) In 10–100 ng/ml EGF, spacing between buds (rows ⬎3) that form in culture is increased. (E) In 200 ng/ml EGF, only the specified rows at stage 30 skin are present and stunted. In 300 ng/ml (F), all feather buds are absent. Control tissue recombination with untreated skin (G) produced buds comparable to dermis treated with EGF (H). Epidermis treated with EGF recombined with untreated dermis failed to support formation of new feather buds (I). Percent interbud area was calculated (see Experimental Procedures) from control and EGF-treated explants (J). Pictures were taken at the same magnification.

EGF Signaling Promotes Interbud Fate 235

Figure 3. Characterization of EGF-Induced Interbud Cross-sections of skin cultured without (A, C, E, and G) or with (B, D, F, and G) 300 ng/ml EGF for 3 days. Sections were stained with hematoxylin/eosin (A and B). Epidermal invaginations (arrow) and hyperkeratinization (arrowheads) are visible in EGF-treated skin. Loss of expression of the feather bud marker Shh is evident in EGF-treated skin (D) in comparison to control (C). Interbud marker, collagen I (white arrows), is present in the interbud dermis of the control skin (E) and in the entire dermis of the EGF-treated skin (F). Proliferating cells were visualized with antibody against BrdU (white arrows in [G] and [H]). Note the decrease in dermal cell proliferation in EGF-treated skin (H). Hatched white lines outline the epidermis.

Inhibition of EGFR Signaling Leads to a Loss of Interbud Fate In order to determine whether EGFR mediates the promotion of interbud fate, explant cultures were incubated with pharmacological inhibitors of the EGFR signaling pathways. The EGFR inhibitors AG1478 and ZD1839 block the substrate site of the EGF receptor kinase and prevent autophosphorylation (Gazit et al., 1991). AG1478 is selective for EGFR at concentrations up to 100 ␮M, and several studies have previously demonstrated the specificity and effectiveness of AG1478 in the micromolar range in organ cultures (Ben-Bassat et al., 1999; Ellis et al., 2001; Zieske et al., 2000). We performed doseresponse studies to determine the effect of blocking EGF signaling (Figures 1F and 4). Two days after exposure to AG1478, fusions of feather buds were visible (Figures 4B and 4C, arrow). Fused buds did not have any histologically detectable interbud (Figure 4G), and Shh transcripts were detected in a single continuous band across the fused buds (Figure 4H). The fusions

most often occurred along the medial-lateral axis, and each row of fused buds was separated by an area of bud-less skin. Fusions along the anterior-posterior axis were observed, but at a lower frequency. Preexisting buds did not fuse but their outgrowth was stunted (Figure 4C). Similar but less severe effects were obtained with the lower affinity EGFR inhibitor ZD1839 (data not shown). By blocking one of the downstream EGFR signaling mediators, the phosphoinositide 3-kinase (PI3K) pathway, we reproduced feather bud stunting in the absence of bud fusions. LY 294002 is a potent and selective inhibitor of PI3K at concentrations up to 50 ␮M, and it acts on the ATP binding site of the enzyme (Vlahos et al., 1994). Incubation of skin explants with LY 294002 at 10 ␮M and 50 ␮M (data not shown and Figure 4D, respectively) resulted in stunting of outgrowth without fusions of the bud. Similar results were obtained with rapamycin, an inhibitor of mTOR kinase downstream of PI3 kinase (data not shown). The studies are consistent with a role for EGF in promoting interbud fate. These

Developmental Cell 236

Table 1. Time Course of Ubiquitous Application of EGF and EGFR Inhibitor Shows the Temporal Competence of Skin to Interpret EGFR Signaling Bud rows

1

2

3

4

5

6

EGF 6 hr 12 hr 24 hr 3 days

⫹ ⫹ x ⫺

⫹ ⫹ x –

⫹ x ⫺ ⫺

x x ⫺ ⫺

⫺ ⫺ ⫺ ⫺

⫺ ⫺ ⫺ ⫺

AG1478 6 hr 12 hr 24 hr 3 days

x x x x

x x x x

x x x x

x x x fused

x fused fused fused

fused fused fused fused

⫹, ⫺, and x indicate the presence, absence, and stunting of bud rows, respectively.

Figure 4. EGFR Signaling Inhibitors Promote Bud Fate Skin cultured without inhibitors (A), 200 nM AG1478 (B), 50 ␮M AG1478 (C), and 50 ␮M LY 294002 (D) for 3 days. Note the fusions (arrows) in skin cultured with AG1478 (B and C). Bud outgrowth also appears stunted. Cross-sections of control skin and AG1478 skin were stained with hematoxylin/eosin (E–G). In the region of bud fusion there is no morphologically distinguishable interbud (G). AG1478 (B, C, and F) and LY 294002 (D) both affect bud outgrowth. Whole-mount in situ hybridization with the feather bud marker Shh shows a continuous band of expression in the fused buds (H).

results indicate that the PI3K pathway mediates outgrowth but the interbud-promoting activity is guided by the non-PI3K arm(s) of the EGFR signaling pathway. Interbud Competence Resides in Regions of the Skin Ahead of the Morphogenetic Wave The tissue recombination experiments and the increase in interbud area relative to the increase in dose of EGF suggested that EGF exposure was altering the competence of the skin. To explore this further, we examined the competence of various parts of the skin explant to interpret the interbud-promoting signal from the EGFR pathway by exposing the skin to ubiquitous or localized sources of EGF or the EGFR inhibitor AG1478. These

studies were also conducted to determine the timing and reversibility of EGFR signaling effects on interbud fate and patterning. Skin explants were treated with 300 ng/ml EGF or 50 ␮M EGFR inhibitor AG1478 for various durations, and then incubated with control media for a total of 3 days. The results are summarized in Table 1 (n ⫽ 6–8/condition). With a 6 hr exposure to EGF, preexisting rows 1 and 2 of buds grew normally (Figure 5B). Row 3, which may be specified but is not visibly distinguishable at stage 30, remained unperturbed by a brief exposure to EGF in vitro. However, row 4 was stunted and patterning was abnormal. Buds in row 5 and lateral to it were inhibited from forming. In a 12 hr exposure to EGF, rows 1 and 2 grow relatively normally but rows 3, 4, and 5 are stunted and widely spaced (Figure 5C). Further lateral rows were inhibited. A 24 hr exposure to EGF followed by 2 days of recovery in control media dramatically perturbed patterning and outgrowth of preexisting and any future buds (Figure 5D). Three days of continuous exposure to EGF led to inhibition of preexisting and new feather bud development (Figure 5E). Blocking EGFR signaling with AG1478 also perturbed the proper patterning of bud and interbud regions. A brief 6 or 12 hr exposure led to fusion of the two most lateral rows of buds (Figures 5G and 5H, respectively). After 24 hr of exposure, more medial rows fused together but the preexisting rows did not fuse in culture (Figure 5I). Even a 3 day exposure failed to perturb the budinterbud identities of the preexisting rows of stage 30 skin (Figure 5J). The stunting of feather buds occurred and remained irreversible after a short exposure of 1–3 hr to AG1478 (data not shown). These pulse-chase studies demonstrate the early and irreversible effects of EGF and EGFR inhibitor. They also demonstrate that the competence to interpret an interbud signal and affect fate is highest in the lateral regions and is gradually repressed toward the midline. In order to specifically test the regional competence of the skin explants, a localized source of EGF or AG1478 was placed at two different positions on the skin. A localized source of 50 ng/␮l EGF was delivered for 4 hr in the lateral-most region of a stage 30 skin explant (n ⫽ 8). After a total of 3 days in culture, buds form in rows

EGF Signaling Promotes Interbud Fate 237

Figure 5. Localized or Brief Application of EGF and EGFR Inhibitor Demonstrates an Extended Area of Interbud Competence Skin was cultured for various times with 300 ng/ml EGF (B–E) or 50 ␮M EGFR inhibitor AG1478 (G–J), and then in control media for up to 3 days. Representative bud formation and growth in control cultures (A). EGF treatment over time (B–E) causes progressive loss of new bud formation, diminished bud outgrowth, and increased spacing between the buds. AG1478 treatment over time (G–J) causes bud fusions, including the lateral edges of the explants after only a 6 hr exposure to AG1478 (more details are in the Discussion). Twelve hour exposure to a localized source of 10 mM AG1478 (previous location of source indicated by a dark spot of charcoal) in the lateral region of the explant leads to bud fusions in the nearby vicinity ([F], arrows). The results of this figure are summarized in Table 1.

1–4 (the prespecified region at the time of treatment), but more lateral buds do not form (data not shown). A localized source of 10 mM AG1478 was delivered for 12 hr in the lateral-most region of the explant (n ⫽ 12). After an additional 2.5 days of culture under control conditions, bud fusions occurred around the localized source (Figure 5F; the source is marked with a piece of charcoal). In addition, spacing between the buds was decreased and bud outgrowth inhibited in a larger region around the localized source. When AG1478 was localized along the midline, the treated explants resembled control cultures (data not shown). Thus, at stage 30, the lateral region of the skin far from the midline can be altered as to its bud/interbud fate, indicating that it is competent to interpret interbud-promoting signals via EGF signaling. Together, the time-dependent and localized source experiments demonstrate the early and irreversible effects of EGF and EGFR inhibitor. Most importantly, these studies indicate that interbud competence resides in regions of the skin previously thought to be naı¨ve and that interbud fates can be specified at an earlier time than previously realized. EGF and BMP Act Independently in the Regulation of Interbud Development Finally, we addressed whether EGF may act in concert with the other described feather bud inhibitors, BMPs. Noramly and Morgan (1998) have proposed that BMP2 in the early bud primordia plays a role in inhibition of bud formation. Their model proposes that BMP diffuses

from the bud, inhibiting new bud formation within a distinct radius of cells. Because EGF and BMP-2 both mediate bud inhibition, it is possible that one signal is the upstream activator of the second. In the femoral tract, Bmp-2 is expressed at low levels in the ectoderm lateral to the most recently formed row of buds. It then becomes restricted within sites of future buds, prior to the appearance of the epidermal placodes (Noramly and Morgan, 1998). However, in the dorsal tract, Bmp-2 is first visible only within the placode (Jung et al., 1998). We examined whether EGF treatment for 16, 20, or 24 hr induced Bmp-2 expression in skin explants. We found that EGF suppresses the expression of Bmp-2 in the preexisting buds and that there was no indication of Bmp induction in regions of bud formation (data not shown). A localized source of EGF also did not induce Bmp-2 expression (data not shown). Thus, EGF does not seem to be involved in the endogenous activation of Bmp-2 expression. These data suggest that EGF acts downstream or independently of BMP in the inhibition of bud formation. In the converse experiment, whole-mount immunohistochemistry with EGF antibody on skin cultured with BMP-4 beads did not reveal EGF induction (data not shown). As another test, BMP-4 beads were added to explants cultured in the presence of AG1478 in the media (Figures 6A and 6B, n ⫽ 9 each). In this experiment, BMP was capable of producing a zone of bud inhibition around the bead in the region of AG1478-induced bud fusions (Figure 6B, arrow). These data indicate that

Developmental Cell 238

Figure 6. BMP-Mediated Bud Inhibition Can Occur in the Absence of EGFR Signaling A BMP-4-soaked bead (1 ␮g/␮l) was placed on the lateral area of a stage 30 skin explant without EGFR inhibitors (A) and with 50 ␮m AG1478 (B) for 3 days. Note that the zone of inhibition occurs amid AG1478-induced fusions (arrow).

BMPs do not require EGF signaling to exert its inhibitory effect on bud formation and suggest that EGF and BMP act independently to regulate interbud development. Discussion Chick embryonic feather buds and interbuds arise in a distinct pattern. However, factors implicated in specifying and/or guiding interbud development have not been reported. The experimental data in this study support a role for EGFR signaling in promoting interbud development and in establishing interbud fate. Prior to feather bud development, phosphorylated EGFR (active form of the receptor) and EGF are expressed in the preplacode skin. Throughout early feather bud development, EGF and active EGFR are expressed in the interbud but not the bud. EGF treatment acts directly on the epidermis (our data and Cohen, 1965) and prevents new bud formation, which leads to diminished feather bud marker expression and expansion of interbud marker. Conversely, inhibitors of EGF signaling lead to bud fusion and loss of interbud tissue. The expression pattern data together with the other experiments in this report reveal a novel role of the EGF ligand/receptor system as an important mediator of interbud identity. Our data also indicate that, during later aspects of bud formation, EGF signaling promotes bud growth. EGFR signaling appears to actively promote interbud fate. Our results indicate that EGF is expressed in the interbud throughout early bud development and that it is functionally important for interbud development. Increasing doses of EGF increase the area of the interbud, and at higher doses leads to bud loss with a corresponding gain of interbud fate. Moreover, within 18 hr of EGF treatment, the bud-promoting factors including SHH and FGF are negatively regulated. Our studies are also supported by those of Dohrmann et al. (2002), in which they applied an EGF-soaked bead to skin explants. Figure 6A in their paper indicates increased spacing between buds, bud loss, and stunting of buds. Although their results were interpreted differently, their results are consistent with ours and provide additional evidence for a role for EGF in the specification of interbud fate. Prior to our studies, very few distinct markers for the interbud of the short feather bud stage

(HH stage 35–37) were known. Most early interbud markers are expressed in the bud after early specification. Collagen I is expressed in the interbud dermis throughout early bud development. EGF treatment leads to the expansion of collagen I expression throughout the entire dermis and an absence of feather bud markers, indicating a shift in fate to interbud. EGF treatment causes a decrease in proliferation of bud dermal cells assayed by BrdU labeling. However, alteration in dermal proliferation does not affect the competence of the dermis to produce buds as tested in tissue recombination experiments. Interestingly, epidermal cells proliferate rapidly in the EGF-treated cultures, resulting in a thickened epidermis with irregularly spaced invaginations. We speculate that these invaginations arise from excess epidermal proliferation in the absence of supporting dermal proliferation. EGF signaling plays an additional and later role in promoting bud growth. Inhibition of EGF signaling causes the preexisting buds to be stunted due to decreased outgrowth. EGF signaling is present at the bud/ interbud border during this later aspect of bud growth, suggesting that the EGF signal for bud growth arises at the border of the bud. The importance of localized EGF signaling is supported by the disrupted outgrowth of preexisting buds following ubiquitous treatment with either EGF or EGF inhibitor. The distinct phenotypes obtained by exposing skin to the various pharmacological inhibitors of EGFR signaling also enable us to separate the contribution of different branches of the EGFR signaling pathway to interbud development and bud growth. Inhibition of EGFR signaling at the level of the EGFR kinase leads to a dose-dependent fusion of the buds and significant stunting of the buds due to decreased outgrowth. Dermal proliferation in the established buds of rows 1, 2, and 3 is significantly decreased, whereas newly forming buds in the lateral regions had a similar proliferation profile as control skin buds (data not shown). A subset of the phenotype is observed upon inhibition of the pathway downstream at the level of PI3 kinase using LY 294002. This inhibitor leads to stunting of the buds. Thus, the PI3K arm of the EGFR signaling cascade mediates bud outgrowth, whereas other signaling cascades downstream of the EGFR mediate interbud development. Both EGF and BMP inhibit bud development. This suggests that these two factors act in concert to regulate bud formation. EGF is expressed in the interbud and actively promotes interbud development, whereas Bmp-2 and Bmp-4 are expressed in the bud and inhibit bud formation in neighboring cells. We tested the relationship between these factors and found that EGF does not induce Bmp expression and that BMP does not stimulate EGF expression. Therefore, although these factors both regulate interbud development, they appear to function independently. Experiments in which the skin was treated for short periods demonstrate the early and irreversible effects of EGF and EGFR inhibitor. More interestingly, they suggest that the zone of competence is broader than previously defined. The current understanding of feather bud induction relies on a reaction-diffusion mechanism of bud activators and inhibitors combined with a positive morphogenetic signal that progresses laterally with time

EGF Signaling Promotes Interbud Fate 239

(reviewed in Jung and Chuong, 2000). Studies on explant cultures have shown that the morphogenetic wave precedes placode formation by approximately one row (Davidson, 1983). Fgf-4 and Shh are expressed early in the placode and promote bud identity. In explant cultures of skin from HH stage 29-30, an FGF-soaked bead placed near the primary row can induce bud fusions (Jung et al., 1998). FGF beads placed in the lateral unspecified regions fail to elicit fused buds. This suggested that unlike the medial regions, the lateral region of the HH stage 30 skin is not competent to respond to bud activators. During normal bud development, Bmp-2 and Bmp-4 are expressed in the bud placode and mediate inhibition of bud in a zone surrounding each bud. BMP4 beads can suppress feather bud formation along the midline or in the lateral regions (Jung et al., 1998). We have shown in this study that BMP and EGFR signaling do not share a common pathway leading to bud inhibition. Moreover, we show that EGFR signaling is an interbud-promoting signal. Our study shows that the skin is regionally competent to respond to an interbud-promoting signal. When stage 30 skin was exposed to AG1478 for a short duration, bud fusions were observed in the extreme lateral regions of the explant (Figures 5F and 5G), significantly ahead of the proposed morphogenetic wave and the area of placode formation. Thus, the zone of competence to interpret an interbud-promoting cue by EGFR signaling resides in a larger area than previously realized. In contrast, bud-promoting signals can be processed only as the bud-inducing morphogenetic wave moves laterally. EGF activity as assayed by the presence of phosphorylated EGFR is observed in the lateral regions (areas in which the placodes have not yet formed). Based on AG1478 inhibitor studies, EGF signaling is required for interbud fate in preplacodal skin. This indicates that the capacity to respond to EGF is present and necessary for interbud fate to be established. Moreover, it indicates that interbud fate is specified prior to the specification of bud fate. As the bud forms, bud-promoting and bud-inhibiting signals are produced, which serves to establish the hexagonal array of bud and interbud. Thus, the response to the interbud-promoting signal must be overcome in areas of bud formation. We find that active EGF signaling occurs in the preplacode skin and in the interbud region during and after placode formation. This suggests that EGF signaling is actively repressed in the bud and maintained in the interbud regions. Many issues remain to be resolved, such as the identity of molecules that regulate EGF expression in the preplacodal skin and interbud; the arm of the EGFR signaling pathway that guards interbud identity; and how proliferation of the epidermis and dermis are normally tightly coordinated. Our study advances an understanding of EGFR signaling in specification of interbud identity. This report also expands the framework within which we understand determination of fate in the developing skin. EGFR signaling components are frequently upregulated in a variety of human tumors. Results from this study contribute to our understanding of the consequences of abnormal EGFR signaling.

Experimental Procedures Explant Culture White Leghorn chicken eggs were obtained from SPAFAS. Explant cultures were performed according to Jung et al. (1998). At the start of culture, EGF (R and D Systems), AG1478 (Calbiochem), ZD1839 (kind gift from Neal Rosen), and LY 294002 (kind gift from Eric Holland) were added once, directly to the culture media. BMP-4 (1 ␮g/␮l) beads were prepared as previously described and placed in the lateral areas of the skin explant (Niswander et al., 1993). A local source of 10 mM AG1478 or 50 ng/␮l EGF was delivered in a piece of filter paper placed on the dermal side of the explant for 12 hr. The filter paper was replaced with a piece of activated charcoal to mark the site of application. Proliferation was assessed by incorporation of BrdU (Sigma). BrdU labeling was accomplished by adding 50 ng BrdU/ml of media during the last 2 hr of incubation at 37⬚C. Buds were visualized by contrast staining with 0.5% Nile blue for 30 s at room temperature and rinsing in PBS. Each experiment was repeated at least three times with three to four explants per condition. Tissue recombinations were done on stage 31-32 skin. The explant skin was incubated in DMEM plus 2 mg/ml dispase plus 20 mM HEPES (pH 7.6) for 10 min on ice. Epidermis and dermis were cultured with EGF (300 ng/ml) for 14 hr. The dermis and epidermis were carefully disassociated with watchmaker forceps. EGFtreated epidermis or dermis was recombined with untreated dermis and epidermis, respectively, and cultured for another 3 days in culture conditions described in Jung et al. (1998). Immunohistochemistry Skin was fixed flat on polyethylene terephthalate trace-etched membrane in 4% paraformaldehyde in PBS at 4⬚C for 60 min, washed with several changes of PBS, and equilibrated in 30% sucrose at 4⬚C before embedding for cryosectioning in OCT medium (Fisher). Immunofluorescent detection of proteins on 10 ␮ cryosections was performed using standard methods (Yamada et al., 1993). For BrdU detection, sections were placed in blocking buffer (PBS plus 0.1% Triton and 1% heat-treated goat serum) for 15 min and 1 N HCl for 2 min at 50⬚C before incubating with 6 ␮g/ml monoclonal BrdU antibody (Roche). Collagen I was detected by using a 1:50 dilution of M-38 monoclonal antibody (Developmental Studies Hybridoma Bank) in blocking buffer (PBS plus 0.1% of heat-treated goat serum and 1.0% BSA; Chuong et al., 1996). Species-appropriate secondary antibodies conjugated to Cy3 (Jackson Immuno Research Labs) were used at 1:200 dilution in blocking buffer. EGF expression was determined on paraffin sections of skin using mouse monoclonal anti-EGF (15 ␮g/ml) and phosphorylated EGFR was detected on cryosections with phospho-Y-1173 anti-EGFR antibodies (1:250; Upstate Biotechnology). Sections were quenched in 3% H2O2 in methanol for 10 min and blocked in 1.0% BSA plus 2% goat serum before incubating with primary antibody overnight at 4⬚C. Secondary biotinylated goat anti-mouse (Jackson Immuno Research Labs) was used at 1:250 for 60 min at room temperature. Signal was amplified using the ABC kit (Vectastain) and visualized using DAB. In Situ Hybridization Whole-mount in situ hybridization was performed as described previously (Henrique et al., 1995), with the exception that treatment with 5 ␮g/ml proteinase K for 5 min at room temperature was employed and BM purple substrate (Roche) was used in place of BCIP/ NBT. In situ hybridization to sections was performed as described elsewhere (Holmes and Niswander, 2001). Quantitative Measurements of Interbud Area Average interbud area was calculated from enlarged photographs of seven explants for each dose of EGF. Analysis was done on a defined area (42,500 pixels) on either side of the midline lateral to row 2. Contours of buds were traced and the area was calculated using Scion Image 1.6a (NIH image). Interbud area was determined by subtracting total bud area from the defined area. Percent interbud area and standard deviation of percent interbud area between individual defined areas was calculated.

Developmental Cell 240

Acknowledgments We are thankful to Eric Holland, Neal Rosen, and Mark Messaro for providing guidance with EGFR inhibitors. We thank Heidi Wong for conducting pilot experiments and Steve Fisher for the use of his facilities. Many thanks go to Paul Wilson, Julie Zikherman, Scott Weatherbee, John Timmer, and Julian Halliday for their insightful comments on the manuscript. The M-38 collagen I antibody was developed by Dr. John McDonald and was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by the Department of Biological Sciences, University of Iowa, Iowa City. This work was supported by the Howard Hughes Medical Institute, the NIH, and a Cancer Center Support grant, and by an NIH-NRSA fellowship to R.A.

Henrique, D., Adam, J., Myat, A., Chitnis, A., Lewis, J., and IshHorowicz, D. (1995). Expression of a Delta homologue in prospective neurons in the chick. Nature 375, 787–790. Holmes, G., and Niswander, L. (2001). Expression of slit-2 and slit-3 during chick development. Dev. Dyn. 222, 301–307. Jung, H.S., and Chuong, C.M. (2000). Periodic pattern in formation of feathers. In Molecular Basis of Epithelial Appendage Morphogenesis, C.M. Chuong, ed. (Austin, TX: R.G. Landes Company), pp. 359–366. Jung, H.S., Francis-West, P.H., Widelitz, R.B., Jiang, T.X., Ting-Berreth, S., Tickle, C., Wolpert, L., and Chuong, C.M. (1998). Local inhibitory action of BMPs and their relationships with activators in feather formation: implications for periodic patterning. Dev. Biol. 196, 11–23.

Received: August 15, 2002 Revised: December 26, 2002

Kashiwagi, M., Kuroki, T., and Huh, N. (1997). Specific inhibition of hair follicle formation by epidermal growth factor in an organ culture of developing mouse skin. Dev. Biol. 189, 22–32.

References

Kronmillier, J.E., Upholt, W.B., and Kollar, E.J. (1991). Expression of epidermal growth factor mRNA in the developing mouse mandibular process. Arch. Oral Biol. 36, 405–410.

Ben-Bassat, H., Rosenbaum-Mitrani, S., Hartzstark, Z., Levitzki, R., Chaouat, M., Shlomai, Z., Klein, B.Y., Kleinberger-Doron, N., Gazit, A., Tsvieli, R., and Levitzki, A. (1999). Tyrphostins that suppress the growth of human papilloma virus 16-immortalized human keratinocytes. J. Pharmacol. Exp. Ther. 290, 1442–1457. Canoun, C., Ma, C., Halpern, D., Shum, L., Bringas, P., Jr., Sank, A., and Slavkin, H.C. (1993). Endogenous epidermal growth factor regulates limb development. J. Surg. Res. 54, 638–647. Chuong, C.M., Widelitz, R.B., Ting-Berreth, S., and Jiang, T. (1996). Early events during avian skin appendage regeneration: dependence on epithelial-mesenchymal interaction and order of molecular reappearance. J. Invest. Dermatol. 107, 639–646.

Niswander, L., Tickle, C., Vogel, A., Booth, I., and Martin, G.M. (1993). FGF-4 replaces the apical ectodermal ridge and directs outgrowth and patterning of the limb. Cell 75, 579–587. Noramly, S., and Morgan, B. (1998). BMPs mediate lateral inhibition at successive stages of feather tract development. Development 125, 3775–3787. Olofsson, B., Pizon, V., Zahraoui, A., Tavitian, A., and Therwath, A. (1986). Structure and expression of the chicken epidermal growth factor locus. Eur. J. Biochem. 160, 261–266.

Cohen, S. (1965). The stimulation of epidermal proliferation by a specific protein (EGF). Dev. Biol. 12, 394–407.

Partanen, A.M., and Thesleff, I. (1987). Localization and quantitation of 125I-epidermal growth factor binding in mouse embryonic tooth and other embryonic tissues at different developmental stages. Dev. Biol. 120, 186–197.

Crowe, R., Henrique, D., Ish-Horowicz, D., and Niswander, L. (1998). A new role for Notch and Delta in cell fate decisions: patterning the feather array. Development 125, 767–775.

Tanda, N., Ohuchi, N., Yoshioka, H., Noji, S., and Nohno, T. (1995). A chicken Wnt gene, Wnt-11, is involved in dermal development. Biochem. Biophys. Res. Commun. 211, 123–129.

Davidson, D. (1983). The mechanism of feather pattern development in the chick. II. Control of the sequence of pattern formation. J. Embryol. Exp. Morphol. 74, 261–273.

Vlahos, C.J., Matter, W.F., Hui, K.Y., and Brown, R.F. (1994). A specific inhibitor of phosphatidylinositol 3-kinase, 2-(4-morpholinyl)-8phenyl-4H-1-benzopyran-4-one (LY294002). J. Biol. Chem. 269, 5241–5248.

Desbiens, C.N., Queva, C., Jaffredo, T., Stehelin, D., and Vandenbunder, B. (1991). The relationship between cell proliferation and the transcription of the nuclear oncogenes c-myc, c-myb, c-ets-1 during feather morphogenesis of the chick embryo. Development 111, 699–713. Diaz-Ruiz, C., Perez-Tomais, R., Cullere, X., and Domingo, J. (1993). Immunohistochemical localization of transforming growth factor␣ and epidermal growth factor-receptor in the mesonephros and metanephros of the chicken. Cell Tissue Res. 271, 3–8. Dohrmann, C.E., Noramly, S., Raftery, L.A., and Morgan, B.A. (2002). Opposing effects on TSC-22 expression by BMP and receptor tyrosine kinase signals in the developing feather tract. Dev. Dyn. 223, 85–95. Ellis, A.G., Nice, E.C., Weinstock, J., Levitzki, A., Burgess, A.W., and Webster, L.K. (2001). High-performance liquid chromatographic analysis of the tyrphostin AG1478, a specific inhibitor of the epidermal growth factor receptor tyrosine kinase, in mouse plasma. J. Chromatogr. B Biomed. Sci. Appl. 754, 193–199. Fowler, K.J., Walker, F., Alexander, W., Hibbs, M.L., Nice, E.C., Bohmer, R.M., Mann, G.B., Thumwood, C., Maglitto, R., Danks, J.A., et al. (1995). A mutation in the epidermal growth factor receptor in waved-2 mice has a profound effect on receptor biochemistry that results in impaired lactation. Proc. Natl. Acad. Sci. USA 92, 1465– 1469. Gazit, A., Osherov, N., Posner, I., Yaish, P., Poradosu, E., Gilon, C., and Levitzki, A. (1991). Tyrphostins. 2. Heterocyclic and ␣-substituted benzylidenemalononitrile tyrphostins as potent inhibitors of EGF receptor and ErbB2/neu tyrosine kinases. J. Med. Chem. 34, 1896– 1907. Hamburger, V., and Hamilton, H. (1951). A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49–92.

Yamada, T., Pfaff, S., Edlund, T., and Jessell, T. (1993). Control of cell pattern in the neural tube: motor neuron induction by diffusible factors from notochord and floor plate. Cell 73, 673–686. Zieske, J., Takahashi, H., Hutcheon, A., and Dalbone, A. (2000). Activation of epidermal growth factor receptor during corneal epithelial migration. Invest. Ophthalmol. Vis. Sci. 41, 1346–1355.