In vivo sonic hedgehog pathway antagonism temporarily results in ancestral proto-feather-like structures in the chicken

Feather morphogenesis in the chicken embryo

The feathers of hatched chickens can be categorized into two groups, including simple plumulaceous feathers (i.e., down-type feathers) that lack a central ridge known as a rachis, and the more ordered pennaceous feathers that do exhibit a rachis (i.e., flight and contour feathers) (Fig 1A). Using volumetric LSFM, we describe the development of feathers in the embryonic chicken from E9 to E14 (Fig 1B and 1C). By E9, feather placodes cover the chicken’s dorsum and exhibit local expression of Shh (Fig 1B, left panel) [21]. By E10, we observe significant feather bud outgrowth across the body (Fig 1B, right panel), with Shh expression becoming localized to the distal tips of developing units [9,37]. Feather buds subsequently propagate across the wings of the embryonic chicken (Fig 1C). At E10, the first placodes can be seen on the posterior edge of the wing. By E11, feather bud coverage across the wings is complete, with a developmental gradient observable from posterior, with the most advanced buds, to anterior, with the least advanced buds. Feather bud outgrowth continues throughout E12, at which stage the longitudinal domains of increased cell density defining the position of future barb ridges become visible. Additionally, optical LSFM sections reveal epidermal SHH immunofluorescence in feather buds on the wing from E10 to E12. Keratinization of feather buds restricts the use of SHH antibodies beyond E12. Invagination of wing feather buds and formation of the follicle are observable from E13 onwards. Overall, feather morphogenesis requires outgrowth, branching, and invagination, and previous research has demonstrated that the Shh pathway is a key regulator of each of these three developmental processes [7–9,34,37].

Sonidegib treatment temporarily changes Shh expression pattern and arrests feather morphogenesis

To experimentally investigate the role of the Shh pathway in mediating feather bud morphogenesis, we employ our recently published protocol for the rapid intravenous injection of developing amniote embryos [16,35,39]. This technique facilitates drug delivery directly into the vascular system of embryos developing within a hard-shelled egg. To precisely antagonize the Shh pathway, we injected chicken embryos with sonidegib (also called NVP-LDE225), a specific and potent inhibitor of Smoothened (Smo) [32,40]. Embryos were treated with a single injection of sonidegib at E9 to perturb feather bud development on the body and wings (Fig 1B and 1C), before fixation and analysis at subsequent embryonic and post-embryonic stages (see Materials and methods for further details).

Control samples treated with DMSO exhibit normal outgrowth of feather buds across the body and the wings from E10 to E14, producing elongated and filamentous feather buds (S1A, S2–S6 Figs). Conversely, sonidegib-treated embryos exhibit a dose-dependent reduction in feather bud outgrowth (S1B–S1D, S2–S6 Figs), with higher sonidegib doses resulting in increasingly restricted outgrowth. Indeed, samples treated with 300 µg sonidegib (the highest dose) exhibit small and nodular feather buds at E14 (S1D, S2–S6 Figs).

To investigate how our sonidegib treatment impacts the local gene expression of Shh during feather bud development on the wings, we use whole mount in situ hybridization (WMISH) (Fig 2). In control samples at E10, we observe the local expression of Shh in immature feather buds partially covering the wings (Fig 2A). By E11, Shh expression becomes focused in the distal tip of outgrowing units [9,37]. From E12 to E13, feather buds become dramatically elongated whilst Shh becomes restricted to the characteristic longitudinal expression domains that determine the formation of individual feather barbs [7,8]. In samples treated with 100 µg sonidegib, spatial expression patterns of Shh remain similar to those observed in control samples (Fig 2B). However, feather bud outgrowth becomes restricted, with units appearing slightly smaller across comparable embryonic stages. Remarkably, samples treated with either 200 or 300 µg sonidegib exhibit dramatically modified Shh expression patterns at E10, with large striped domains of Shh appearing instead of the characteristic spots associated with individual feather placodes (Fig 2C and 2D). Note that similar results have been reported ex vivo with feather-forming tissue explants treated with cyclopamine, a drug with multiple effects, including antagonism of the Shh pathway [37]. Our results indicate that a reduction of Shh signaling, an activator of early feather patterning [9,18], modifies the underlying reaction-diffusion dynamics to produce a pattern of stripes instead of spots. Remarkably, by E11 Shh expression recovers its spotted pattern associated with individual feather domains (Fig 2C and 2D), demonstrating that reaction-diffusion patterning robustly resists temporary pharmacological down-regulation of Shh signaling. However, although feather bud patterning recovers by E11, subsequent morphogenesis remains perturbed: from E12 to E13 these feather buds undergo substantially less outgrowth (Fig 2C and 2D) than observed in control samples. At E13, in samples treated with 300 µg sonidegib, this results in a dramatic reduction in the longitudinal Shh expression domains associated with branching morphogenesis (Fig 2D). At E14, as keratinization of feathers restricts the use of WMISH, we instead show the morphological impact of sonidegib treatment on the wings (Fig 2B–2D, bottom row). Again, these data reveal a dose-dependent reduction in feather bud outgrowth as samples treated with 300 µg sonidegib exhibit greatly shortened and unbranched feather buds.

Fig 2. WMISH reveals that sonidegib treatment arrests feather branching morphogenesis.

(A) The feathers of DMSO-treated control samples undergo normal patterning and morphogenesis [9], including the emergence of longitudinal Shh expression domains associated with branching morphogenesis from E12 to E13 [7,8]. (B) Samples treated with 100 µg sonidegib exhibit similar feather patterning to control samples but with slightly perturbed feather bud outgrowth (units are shorter across comparable embryonic stages). (C, D) Samples treated with either 200 or 300 µg sonidegib exhibit large striped expression domains of Shh at E10, demonstrating that these treatments temporarily modify the underlying reaction-diffusion dynamics. Normal spotted patterning recovers by E11 with Shh expression restricted to individual feather domains. However, subsequent morphogenesis from E11 to E13 remains perturbed as feather buds undergo substantially less outgrowth than in control samples. (D) At E13, samples treated with 300 µg sonidegib also reveal a reduction in longitudinal Shh expression domains associated with feather branching morphogenesis. (A–D, bottom row) As Shh WMISH becomes problematic after E13, only the morphological effect of sonidegib-induced dose-dependent reduction in feather bud outgrowth is shown at E14 (also shown in S1 Fig).

https://doi.org/10.1371/journal.pbio.3003061.g002

Next, we use LSFM of nuclear-stained (TO-PRO-3) chicken wings to examine the morphological effect of sonidegib treatment at E14 (Fig 3). Again, we observe a dose-dependent effect of treatment upon feather bud outgrowth, with higher doses resulting in shorter feather buds (Fig 3A–3D). In addition to a dramatic reduction in outgrowth, we also observe substantial variation in the shapes and sizes of feather buds in samples treated with 300 µg sonidegib, including the fusion of individual units (Fig 3D, oval outline). Importantly, feather branching morphogenesis is arrested at E14 by sonidegib treatment. Indeed, although feather barbs are present in control and lower-dose treatments (Fig 3A–3C, third row of panels, white arrows), samples treated with 300 µg sonidegib reveal no branching at E14 (Fig 3D, two lowest panels). Hence, although some longitudinal Shh expression domains are visible at E13 (Fig 2D, lowest panel), this gene expression is not translated into a corresponding morphological pattern of cell densities by E14. Furthermore, LSFM reveals that the invagination (inward growth) of feather follicles is perturbed by sonidegib treatment, as samples treated with higher doses exhibit less advanced follicles (Fig 3A–3C). Remarkably, feather bud invagination is entirely absent in samples treated with 300 µg sonidegib (Fig 3D, two lowest panels, blue arrow). Therefore, at high doses, sonidegib treatment strongly perturbs feather bud morphogenesis, restricting feather outgrowth, follicle invagination and branching morphogenesis (Fig 3D).

Fig 3. Sonidegib treatment arrests feather bud outgrowth and invagination.

We use LSFM of nuclear stained (TO-PRO-3) chicken wings at E14 to investigate the morphological effect of sonidegib treatment. (A) Control samples exhibit normal development of elongated feather buds with well-advanced follicles and visible barb ridges (white arrow). (B, C) We observe a dose-dependent effect of sonidegib treatment, with higher doses resulting in shorter feather buds and less advanced follicle development. (D) Samples treated with 300 µg sonidegib exhibit minimal feather bud outgrowth, the fusion of feather buds (white oval dotted outline), absent branching morphogenesis, and no follicle invagination (blue arrow).

https://doi.org/10.1371/journal.pbio.3003061.g003

To investigate whether the morphological effects of sonidegib treatment constitute permanent modifications to feather morphology, we next examined treated embryonic samples fixed at later developmental stages. By E17, we observe a partial recovery of feather morphogenesis, with sonidegib-treated embryos exhibiting long, filamentous, and keratinized feathers (S7 Fig). However, these feathers remain shorter in samples treated with higher sonidegib doses and their coverage remains sparse in embryos treated with 300 µg sonidegib. At E21, we observe this same dose-dependent effect of decreased feather elongation and coverage with increasing strengths of sonidegib treatment (S8 Fig). These results demonstrate that, although treatment with 300 µg sonidegib strongly perturbs feather morphogenesis until E14 (Figs 2–3), this developmental process subsequently undergoes a partial recovery, presumably as the inhibitory effect of sonidegib declines (S7 and S8 Figs).

Samples treated with high doses of sonidegib also exhibit naked patches of skin at late embryonic stages (S7 and S8 Figs), including in the anterior region of the wing at E21 (Fig 4, top panels). We next investigate these samples using LSFM of nuclear staining (TO-PRO-3). The wings of E21 control samples exhibit normal fully developed, deeply embedded feather follicles, accompanied by highly keratinized and filamentous feather shafts (Fig 4A). Unexpectedly, naked skin regions of samples treated with 300 µg sonidegib do exhibit invaginated feather follicles; however, these structures are deformed and are not associated with feather outgrowths (Fig 4B). Furthermore, these follicles are substantially less developed than those of control samples and exhibit less differentiated and less keratinized tissues. Therefore, rather than abolishing or modifying the spatial patterning of feather follicles in naked regions of the skin at E21, treatment with 300 µg sonidegib at E9 both aborts feather bud outgrowth and partially reduces follicle development. However, these perturbed feather follicles remain present beneath the skin.

Fig 4. Naked skin regions in sonidegib-treated embryos contain perturbed follicles at E21.

(A) Control samples exhibit normal coverage of keratinized and filamentous feathers. LSFM of nuclear stained (TO-PRO-3) control samples reveals deeply embedded feather follicles from which highly keratinized feather shafts emerge, consisting of multiple individual barbs (white arrows). (B) The anterior wings of samples treated with 300 µg sonidegib exhibit naked regions of skin adorned with perturbed sub-epidermal feather follicles (blue arrows) that completely lack the outgrowth of feather buds (white arrows), and exhibit less tissue differentiation than observed in control samples. Therefore, sonidegib treatment does not abolish the patterning of feather follicles, but instead dramatically perturbs feather bud and follicle morphogenesis in these regions, resulting in naked areas of the skin.

https://doi.org/10.1371/journal.pbio.3003061.g004

Sonidegib treatment specifically down-regulates Shh pathway activity

We next sought to elucidate the molecular mechanisms underpinning the effect of sonidegib treatment using bulk RNA-seq analysis, to confirm the specific inhibitory effect of this drug on the Shh pathway [32,40]. Embryos were injected at E9 with either DMSO as a control or 300 µg sonidegib, and fixed daily at four subsequent developmental stages from E10 to E13. RNA was extracted for sequencing from dissected embryonic wing tissue from four biological replicates at each of the four developmental time points (Fig 5A) (see Materials and methods for details).

Fig 5. Bulk RNA sequencing reveals that sonidegib treatment down-regulates Shh pathway signaling.

(A) Samples were injected with either DMSO (controls) or 300 µg sonidegib at E9. Wings were then dissected for RNA sequencing from E10 to E13, with four biological replicates used for each treatment at each stage. (B) From E11 onwards, PCA reveals a clear separation between control and treated samples. (C) Differentially expressed genes (DEGs) (filtered by a fold change of ≥1.4 and a false-discovery rate (FDR) adjusted P-value of ≤0.05) are shown in volcano plots for four embryonic stages. At E10, Shh pathway members Ptch1, Ptch2 and Gli1 are down-regulated. From E11 to E13, both Ptch2 and Shh are persistently down-regulated in treated samples relative to controls. See file S1 Data for the data underlying the graphs shown in the figure.

https://doi.org/10.1371/journal.pbio.3003061.g005

First, to assess the reliability of treatment across individual biological replicates, we use Principal Component Analysis of our RNA-seq dataset. This reveals a clear separation of sonidegib-treated (Fig 5B, red dots) versus DMSO control replicates (green dots) at each embryonic stage. Second, using a cluster analysis, we confirm that both sonidegib-treated and control samples are clearly distinguishable by their transcriptomic profiles (S9 Fig). Indeed, after filtering differentially expressed genes (DEGs) with a false-discovery rate (FDR) adjusted P-value of ≤0.05, the corresponding heat maps show that, at each time point, individual replicates cluster together according to their treatment type. Third, we identify key DEGs between sonidegib-treated and control samples (Fig 5C). To define DEGs, we filtered our dataset using a fold change of ≥1.4 and an FDR adjusted P-value of ≤0.05 (see Supporting information file S1 for both filtered and unfiltered DEGs).

At E10, i.e., 1 day after sonidegib treatment, we detect 56 DEGs (37 down-regulated and 19 up-regulated) and we already observe strong down-regulation of the SHH receptors Ptch1 and Ptch2, as well as the downstream target Gli1. Note that we also observe a small increase in the expression of Shh itself, which may be indicative of an initial compensatory response following Shh pathway inhibition. We also observe strong up-regulation of fibroblast growth factor 20 (Fgf20), a key regulator of early feather bud development [26], which again may constitute a compensatory response. Our results also reveal weak down-regulation of SRY-related HMG-box 18 (Sox18), which has previously been observed in developing feather buds [41] and can induce the development of ectopic feathers when mis-expressed [42].

At E11, we detect 47 DEGs (40 down-regulated and 7 up-regulated). This includes the down-regulation of Ptch2 and Gli-like transcription factor 1 (Glis1), demonstrating the continuation of decreased Shh pathway activity. Note that Shh itself is now also down-regulated, suggesting that the initial compensatory response to the treatment observed at E10 is now arrested. We also observe the down-regulation of Sox18 at this stage. By E12, we identify 311 DEGs (245 down-regulated and 66 up-regulated), which again includes suppressed expression of both Ptch2 and Shh itself. This sustained repression of Shh indicates that sonidegib treatment triggers negative feedback of intrinsic Shh signaling. At E13, we detect 749 DEGs (567 down-regulated and 182 up-regulated). Again, this includes the down-regulation of both Ptch2 and Shh. Importantly, these strong and temporally consistent responses of both Ptch2 and Shh conform to our previous RNA-seq analysis of chicken embryos treated with SAG [35], which has an inverse effect on Shh signaling [31]. Note that the much larger numbers of DEGs at E12 and E13 (in comparison to E10 and E11) are expected as the perturbation of Shh signaling impacts numerous downstream targets of the transcription factors included in the Shh pathway. Overall, our analysis reveals that Shh pathway members are consistently and negatively differentially expressed in response to sonidegib treatment at E9 (Fig 5C).

Next, we present the temporal expression dynamics for genes associated with skin appendage development in mean transcripts per million (TPM) (S10 Fig). These results reveal the down-regulation of Shh pathway members, including Shh, Ptch1, Ptch2, and Gli1, in response to sonidegib treatment (S10A Fig, top row of graphs). We also observe strong down-regulation of hedgehog interacting protein Hhip (S10A Fig, second graph in bottom row). In line with the above DEG analyses, our TPM analyses (S10 Fig) show (i) no significant differential expression of Smo itself in response to the pharmacological targeting of this receptor [35] and (ii) temporary up-regulation of Fgf20 at E10, which may constitute a compensatory effect of Shh pathway inhibition. Note that the fibroblast growth factor binding protein 1 (Fgfbp1) and EDAR-Associated Via Death Domain (Edaradd) are both significantly down-regulated at E12 or E13 (S10B Fig), similar to Sox21, and wingless/integrated (Wnt) family members Wnt6, Wnt16 and Wnt10a (S10C Fig). This is indicative of their roles in late feather morphogenesis, which are suppressed in sonidegib-treated samples (Figs 2 and 3). Finally, we observe a dramatic down-regulation of genes associated with Keratin (Krt) production, including Krt14, Krt6a, Krt23, Krt40 and feather keratin 1-like (Fk1-like) (S10D Fig). Again, we attribute this to the arrested morphogenesis of feather buds in sonidegib-treated samples. Overall, these results confirm that Shh pathway activity is specifically down-regulated [32,40] in developing feather buds, in response to a single intravenous injection with sonidegib at E9.

To further investigate our transcriptomic data, we next present a gene ontology (GO) enrichment analysis of “biological processes” [43] associated with our DEG sets (fold change of ≥1.4 and FDR-adjusted P ≤ 0.05) from each time point. The 10 most significant GO terms are reported for each stage, along with the number and names of the detected DEGs (S11 Fig). At both E10 and E11, Shh is associated with all of the 10 most significant GO terms, including “regulation of cell communication” and “regulation of signaling” (S11A and S11B Fig). Shh pathway members, including Ptch1, Ptch2, Gli1 and Glis1, also frequently contribute to significant GO terms throughout these stages. At E12, Shh is associated with nine of the 10 most significant GO terms, the most significant of which is “regulation of cell differentiation” (S11C Fig). At E13, we observe Shh in 5 of the 10 most significant GO terms, including “regulation of gene expression” (S11D Fig). Overall, our GO enrichment analysis reveals that sonidegib treatment is significantly associated with the regulation of Shh pathway-associated biological processes.

Finally, we present a Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of DEG sets from each time point (S12A Fig) [44]. At E10, the most significant detected KEGG pathway is “Hedgehog signaling”, with numerous Shh pathway members differentially expressed, including Shh, Ptch, Gli, and Hhip (S12B Fig, DEGs shown in red). Across all subsequent stages, the most consistently detected significant KEGG pathway is “Basal cell carcinoma” which involves Shh pathway members, including Shh and Ptch (S12C–S12E Fig, DEGs shown in red). Overall, our analysis demonstrates that Shh-associated KEGG pathways are significantly differentially regulated following sonidegib treatment.

Reactivation of the Shh pathway rescues feather morphogenesis

We next sought to investigate the specificity of the effect by which sonidegib perturbs embryonic feather development. Therefore, we repeated our sonidegib treatments in combination with the delivery of SAG (Figs 6 and S13). Contrary to sonidegib, which inhibits Shh pathway activity, SAG is a Shh pathway activator [35]. This rescue experiment allows us to test the specificity of sonidegib on SMO because both sonidegib and SAG are considered to act directly, albeit with opposing effects, on this key transducer of the Shh pathway [31,32]. Embryos were treated at different embryonic stages with SAG, together or following 300 µg sonidegib treatment at E9. The precise dosage of SAG was adjusted according to approximate embryo weight, with more developed embryos receiving larger doses. All embryos were fixed at E14 to examine the effects of these treatments on feather morphogenesis (see S13 Fig for all replicates).

Fig 6. SAG treatment rescues sonidegib-perturbed feather morphogenesis at E14.

Embryos were treated with SAG, a SMO agonist, together or following sonidegib delivery at E9. (A) Control embryos injected with DMSO at E9 exhibit normal elongated, filamentous feathers at E14. (B) Embryos treated at E9 with 300 µg sonidegib exhibit at E14 substantially reduced units that lack both branching and follicle invagination (blue arrow). (C) However, the combined injection of 300 µg sonidegib and 50 µg SAG at E9, prevents abnormal morphogenesis of feather buds. (D) Treatment with 100 µg SAG at E10 rescues feather morphogenesis in samples treated with 300 µg sonidegib at E9. (E) Treatment with 150 µg SAG at E11 only partially rescues feather morphogenesis in samples treated with 300 µg sonidegib at E9 (i.e., feather buds remain shorter at E14) probably because of a shorter recovery time (i.e., E11–E14). LSFM sections in the bottom panels reveal normal feather branching (white arrows) and follicle morphogenesis in the control and in all samples treated with both sonidegib and SAG, although samples treated with SAG at E11 exhibit reduced follicle invagination (E, bottom panel).

https://doi.org/10.1371/journal.pbio.3003061.g006

Control embryos treated with DMSO at E9 develop normal filamentous feather buds by E14 (Fig 6A, see also Fig 3), whilst embryos treated with 300 µg sonidegib exhibit substantially reduced units that lack both branching and follicle invagination (Fig 6B). However, samples treated with a combined injection of 300 µg sonidegib and 50 µg SAG at E9 display normal elongated feather buds (Fig 6C). Samples treated with 300 µg sonidegib at E9 followed by 100 µg SAG at E10 exhibit similarly elongated feather buds (Fig 6D). When the two injections are separated by a longer interval (300 µg sonidegib at E9 followed by 150 µg SAG at E11) the treated embryos exhibit a less complete rescue of the feather phenotype (i.e., with less elongated feather buds; Fig 6E). This latter result could be due to a reduced receptivity to SAG at E11 or, more simply, to the reduced recovery time between SAG treatment at E11 and fixation at E14. LSFM sections reveal that all samples treated with both sonidegib and SAG undergo normal feather branching and follicle invagination (Fig 6C–6E, bottom row), although follicle development is slightly less advanced in samples treated with sonidegib at E9 and SAG at E11 (Fig 6E, bottom panel). Overall, these experiments demonstrate that the intravenous delivery of SAG is sufficient to robustly rescue feather morphogenesis following sonidegib treatment. Hence, this recovery of feather morphogenesis highlights the specificity of both drugs on the Shh pathway.

Feather morphogenesis robustly recovers post-hatching

We next sought to uncover the effect of sonidegib treatment on feather development in post-embryonic chickens. Therefore, we allowed chickens treated with sonidegib at E9 to hatch and monitored their development until 49 days post-hatching (dph).

At 1 dph, we observe a dose-dependent effect of sonidegib treatment (Fig 7, first row). Control chickens exhibit normal coverage of feathers (Fig 7A). However, feather coverage in samples treated with 100 µg sonidegib appears more sparse, with some patches of skin visible through the plumage (Fig 7B). Samples treated with either 200 or 300 µg sonidegib exhibit large areas of naked skin, particularly on the back (Fig 7C–7D). Upon closer inspection of these naked regions, we observe regularly patterned elevations of the skin (Fig 7D, top-right panel). As observed on naked regions of the wing at E21 (Fig 4B, top-right panel), these low protrusions likely correspond to dormant feather follicles. Next, we removed feathers from the dorsal midline to compare their morphologies (S14 Fig). Interestingly, we observe a dose-dependent treatment effect, with higher sonidegib doses resulting in smaller down-type feathers. All these results indicate that Shh pathway antagonism at E9 perturbs feather morphogenesis in newly hatched chickens.

Fig 7. Post-embryonic effect of sonidegib treatment.

(A) Control samples exhibit complete coverage of (yellow) plumulaceous down-type feathers from 1 dph (top row) until 7 dph (second row). By 28 dph (third row), down-type feathers are observed transitioning into (white) pennaceous contour feathers. This process is complete by 49 dph (bottom row). (B) Samples treated with 100 µg sonidegib exhibit slightly more sparsely patterned feathers, with some patches of skin visible through their plumage. However, by 49 dph, their feather coverage appears indistinguishable from controls. (C, D) Samples treated with either 200 or 300 µg sonidegib exhibit large regions of naked skin on their backs, from 1 dph to 28 dph. Upon closer inspection, these naked regions exhibit regularly patterned skin elevations, which likely correspond to dormant subepidermal follicles (as shown in Fig 4B). By 49 dph, the feather coverage of samples treated with 200 or 300 µg sonidegib recovers and appears comparable to both controls and samples treated with 100 µg sonidegib.

https://doi.org/10.1371/journal.pbio.3003061.g007

At 7 dph, we observe this same dose-dependent effect of sonidegib treatment on feather coverage (Fig 7, second row). By 28 dph, we observe the widespread transition from down-type feathers into pennaceous adult contour feathers in all hatched chickens (Fig 7, third row), and naked skin regions persist on the lateral back surface of samples treated with higher doses of sonidegib (Fig 7C–7D, third row). However, by 49 dph, these naked regions are no longer visible, and the feather coverage of different treatments is now difficult to distinguish (Fig 7, fourth row). By removing the feathers from these samples at 49 dph, we reveal the patterning of feather follicles across the backs and wings of each treatment group (S15 Fig). Follicle patterning appears comparable between control chickens and those injected with different doses of sonidegib. Therefore, although sonidegib treatment at E9 temporarily arrests feather morphogenesis at earlier stages, it does not perturb the spatial distribution of follicles at 49 dph.

Overall, our results reveal that Shh pathway antagonism via intravenous sonidegib injection gives rise to dormant feather buds at the hatching stage, corresponding to naked regions of the skin (Fig 7). These naked regions persist until 28 dph. However, dormant follicles are subsequently re-activated by 49 dph, perhaps during the first natural molting cycle of the chicken, resulting in normal feather coverage.

Previous research has shown that the relative level of Shh signaling within the posterior edge of the wing is involved in determining the location of specialized flight feathers [38]. Flight feathers are large pennaceous feathers that emerge in precisely patterned rows from the posterior region of either wing (Fig 1A). In accordance with previous work, our sonidegib treatments also perturb flight feather specification (S16 Fig) [38]. Flight feathers are clearly visible on the wings of control samples at 7 dph (S16A Fig, second row). However, these units become smaller in samples treated with increasing doses of sonidegib and are almost entirely absent in birds treated with the highest dose (S16D Fig). This effect persists at both 28 dph (S16 Fig, third row) and 49 dph (S16 Fig, fourth row), with smaller contour-type feathers instead adorning the wing. Therefore, this modification to flight feather identity constitutes a permanent modification to the patterning of the integument.