Ectodysplasin overexpression reveals spatiotemporally dynamic tooth formation competency in stickleback and zebrafish

ABSTRACT

Organ initiation is often driven by extracellular signaling molecules that activate precursor cells competent to receive and respond to a given signal, yet little is known about the dynamics of competency in space and time during development. Teeth are excellent model organs for studying cellular competency because they can be activated with the addition of a single signaling ligand, Ectodysplasin (Eda). To investigate the role of Eda in tooth specification, we generated transgenic sticklebacks and zebrafish with heat shock-inducible Eda overexpression. We found that stickleback Eda can drive ectopic tooth formation in at least eight distinct morphological domains. Both zebrafish and stickleback exhibit maximal responsiveness to Eda overexpression during pioneer tooth initiation. Analysis of candidate receptor expression in sticklebacks reveals that ectopic tooth formation in the pharynx correlates with Edar and Troy expression, while only Troy expression was detected in regions of the face where teeth can form, suggesting that competency may involve spatially restricted receptor expression. These findings underscore the latent developmental potential, i.e. competency, of the vertebrate dentition and provide insights into organ competency during embryonic and post-embryonic development.

Highlighted Article: Competency to respond to ectopic Eda by initiating tooth organs shifts anatomically within the head and temporally across development, suggesting that differential receptor expression could be a molecular mechanism underlying this dynamic competency.

INTRODUCTION

The localization of organ initiation is a highly regulated process that ultimately shapes a given body plan. Many organs are specified by secreted signaling ligands, which allows for some coordination between tissues during organ initiation (Perrimon et al., 2012). For cells to demonstrate ‘competency’ to respond to a given signal, they must present a proper complement of receptors and other signal transduction machinery. The extent to which cells are competent to initiate the formation of a given organ type, and how this competency changes throughout development, remains largely unknown. In vertebrates, internal organs typically exist as single units or in pairs that are specified with a conserved connectivity to other organs and tissues. Conversely, epithelial organs like teeth, hair, feathers and scales tend to be more labile in their position and number both within and between species. Epithelial organs thus demonstrate a unique window into understanding how the morphological location of organ initiation can be altered.

Teeth can vary especially widely in morphological location between species (Berkovitz and Shellis, 2023). Similarly, the overall number of teeth present in a given tooth field can also vary at homologous morphological locations between species, or sometimes even within species (Ellis et al., 2015; Miller et al., 2014; Oeschger et al., 2022). Meanwhile, the cell types that constitute tooth organs demonstrate strikingly conserved developmental genetic programs across vertebrates (Gruenhagen et al., 2022; Rasch et al., 2016; Rostampour et al., 2019). Dental mesenchyme and epithelium coordinate through various extracellular signaling pathways to differentiate into odontoblasts and ameloblasts, respectively, which secrete bony tissues known as dentine and enamel (or enameloid), respectively. Thus, conserved tooth units are specified using genetic cascades that must be localized in a comparatively plastic fashion, allowing teeth to ‘jump’ around body plans with relative ease both experimentally and over evolutionary time. In particular, ray-finned fishes have accrued an impressively diverse array of dental morphologies. Across this group, at least 16 different oropharyngeal bones naturally sport teeth in at least one lineage (Stock, 2007; Vandewalle et al., 2000), while a few species of fish also specify small teeth externally (known as denticles) (Greenwood and Greenwood, 1968; Mori and Nakamura, 2022).

Ectodysplasin (Eda) is a tumor necrosis factor (TNF) which is known to bind at least one class of tumor necrosis factor receptor (TNFR), Ectodysplasin Receptor (Edar) (Elomaa et al., 2001). Once bound by Eda, Edar activates the Nuclear Factor Kappa B (NF-κB) pathway, which induces transcription of genes involved in epithelial organ specification, including teeth (Liang et al., 2019; Ohazama and Sharpe, 2004; Sadier et al., 2014; Schmidt-Ullrich et al., 2006). Eda has been shown to be both a necessary factor for proper dental development, as well as sufficient to drive the differentiation of ectopic and supernumerary teeth. Human EDA mutants present a disease known as Ectodermal Dysplasia (for which EDA was named), typically characterized by dental abnormalities ranging from one or more missing teeth (tooth agenesis) to complete tooth loss (Liu et al., 2022; Zeng et al., 2017). Fish models, including sticklebacks and zebrafish, also demonstrate stark losses of teeth when Eda is mutated, resulting in significantly reduced tooth numbers (Harris et al., 2008; Wucherpfennig et al., 2019). Human clinical trials have shown that prenatal exposure to a recombinant human EDA (Fc-EDA) is associated with a less-severe reduction in primary tooth number than in untreated siblings with the same EDA genotype (Schneider et al., 2018). Previous overexpression experiments used transgenes that drove the zebrafish eda gene under the control of the ubiquitous ef1a promoter in zebrafish and Astyanax models, finding that both species were capable of differentiating ectopic teeth (Aigler et al., 2014; Jandzik and Stock, 2021). Intriguingly, the ectopic teeth that arose in these experiments were often morphologically discontiguous with endogenous tooth fields, meaning they did not represent simple expansions of existing tooth fields, but rather de novo deployment of a tooth differentiation program at distinct morphological locations. Furthermore, the ectopic tooth fields that arose in these experiments occur naturally in other fish species, suggesting that these experimental phenotypes phenocopy existing natural variation across fish groups. Notably, some of the eda-induced teeth in Astyanax were shown to have been repeatedly lost and regained in parallel within their order (Characiformes). This response to exogenous Eda suggests that a latent developmental potential to form teeth has persisted for 100 million years or more, having been reactivated in separate lineages multiple times in order to reinvoke teeth at certain oropharyngeal bones (Aigler et al., 2014; Jandzik and Stock, 2021). However, these previous studies used ubiquitous overexpression of eda, and thus lacked temporal control of Eda overexpression, raising the question of how temporally labile tooth formation competency is during embryonic and post-embryonic development.

To test the spatial and temporal windows of tooth formation competency, we developed transgenic lines of sticklebacks and zebrafish capable of heat shock-inducible Eda expression. By using a comparative approach in two fish species with divergent dental arcades (Square et al., 2021), we can determine whether tooth formation responsiveness demonstrates similar or disparate characteristics among divergent groups. Our transgenes use the zebrafish hsp70l promoter to overexpress an mCherry-P2A-Eda polycistronic gene in each zebrafish and stickleback, where the Eda coding region supplied to each species encodes their endogenous protein. Overall, these transgenic fish lines allow us to test: (1) whether competency is spatially restricted in two different species; (2) whether competency is temporally dynamic; and (3) whether ectopic teeth are retained or are capable of replacement after exogenous Eda overexpression is discontinued. We find that Eda overexpression indeed elicits ectopic tooth formation in sticklebacks, which can specify teeth in at least eight unique dental domains, while zebrafish are limited to the previously reported single (paired) ectopic tooth domain in the dorsal pharynx. We further find that zebrafish and sticklebacks both demonstrate the highest potential to form ectopic pharyngeal teeth during the window wherein their first endogenous teeth (known as ‘pioneer teeth’) are specified, which are hypothesized to possess special characteristics as compared to other endogenous primary teeth (Gibert et al., 2019; Sadier et al., 2020). We also observed tooth growth and remodeling consistent with tooth regeneration at ectopic tooth fields multiple months after heat shock activation of Eda during embryonic stages. Finally, cross-referencing ectopic tooth formation profiles with TNFR gene expression in sticklebacks provides evidence that Edar and Troy, but not Relt, are expressed in regions coincident with ectopic tooth formation competency during periods where ectopic teeth are most receptive to initiation, providing a possible mechanism involving receptor expression for the dynamic spatial and temporal competency.

RESULTS

Eda gene duplication and differential paralog retention

Some fish genomic assemblies house more than one annotated Eda gene. A previous synteny analysis found that some teleost species possess two copies of Eda, while the spotted gar, a non-teleost, possesses only one copy, suggesting that these duplicates arose via the teleost whole-genome duplication (TGD) (Braasch et al., 2009). To further understand the evolutionary relationships among Eda loci, we generated a phylogeny using Eda amino acid sequences from a range of diverse fish genomes (Fig. S1). Overall, this phylogeny provides additional evidence that Eda gene duplicates originate from the TGD. After duplication, both resulting Eda paralogs were retained by many, but not all teleost lineages, as each paralog group formed separate clades that generally recapitulate widely-accepted relationships among major fish groups (Betancur-R et al., 2017). As previously hypothesized (Braasch et al., 2009), our analysis supports a scenario where zebrafish (and other Cypriniforms) have retained the opposite Eda paralog to sticklebacks (and other Neoteleosts); following the previous designations, here we refer to the paralog retained by sticklebacks and zebrafish as the ‘A’ and ‘B’ paralogs, respectively.

Eda-driven ectopic tooth formation in stickleback can be triggered during or after endogenous tooth specification

To test whether a type A Eda can cause ectopic tooth development in sticklebacks, we generated a stickleback cross using the Eda overexpression (OE) transgene and a dlx2b:eGFP reporter transgene (Jackman et al., 2022 preprint; Jackman and Stock, 2006). As in zebrafish, this dlx2b reporter transgene is expressed in a tooth germ-specific manner in sticklebacks, allowing us to detect nascent teeth in living fish. We split this cross into two groups for two non-overlapping OE treatments to test whether ectopic tooth specification required exogenous Eda during the same window wherein endogenous teeth were specified. In the first half of the cross, we overexpressed stickleback Eda during embryonic stages from 1-9 days post fertilization (dpf) by administering ∼1 h heat shocks (see Materials and Methods) every 12 h (16 total heat shocks) followed by a 6-day recovery period (Fig. 1A). This treatment encompassed the establishment of all endogenous tooth fields, lasting from neural crest cell migration to pioneer tooth differentiation. We found that 100% (n=8/8) of GFP⁺, Eda OE sticklebacks grew multiple ectopic teeth, whereas 0% of control fish had ectopic teeth (Fig. 1B-D, brackets). All heat shocked Eda OE fish demonstrated a higher number of oral tooth germs, and in 5/8 individuals the dentary tooth organs were arranged in a cluster as opposed to the single row seen in control individuals (Fig. 1E-H, brackets). In the pharynx, ectopic teeth formed along the ventral midline (Fig. 1I-L). At ceratobranchial 4 (cb4) bones, clusters of three to six ectopic teeth were detected in 8/8 individuals. At hypobranchial 3 (hb3), clusters of two to four ectopic teeth were observed in 8/8 fish. At hb2 and hb1 we observed one or two ectopic teeth in 3/8 fish and 7/8 fish, respectively. We additionally detected basihyal (bh) tooth organs in 2/8 individuals (Fig. 1F, arrow). At cb5, Alizarin Red staining revealed additional supporting bone (bone of attachment) at cb4 and cb5 in 8/8 fish, sometimes in regions without ectopic teeth (Fig. 1L, black arrow). In 3/8 fish, the left and right endogenous cb5 pharyngeal tooth fields were fused across the midline (Fig. 1L, arrowhead).

Fig. 1.

Embryonic Eda overexpression in sticklebacks. (A) Outline of the embryonic treatment. Fish were heat shocked twice per day from 1 to 9 dpf, from late gastrulation to hatching. Drawings modified from Swarup (1958). (B-D) Ventral views (anterior to top) of fish carrying dlx2b:eGFP, marking nascent tooth germs (green). Black bracket marks the anterior-posterior extent of endogenous pharyngeal tooth germs in B-D. White bracket in C indicates ectopic tooth germs. B shows a heat shocked control fish that does not carry the Eda overexpression (OE) transgene, n=7 fish. C shows a heat shocked treatment fish that does carry the Eda OE transgene (mCherry), n=8 fish. D shows an un-heat shocked control fish that carries the Eda OE transgene (mCherry in lens only), n=8 fish. The individuals in B and C are imaged after heart removal, during pharyngeal extraction; the individual in D is a future treatment fish for the experiment depicted by Fig. 2, and is thus imaged under light anesthesia without heart removal. Note the magenta mCherry signal in panels C and D relative to B. In panel D, the green and red channel fluorescence gain settings were higher than in B,C to highlight the pharyngeal tooth distribution and red lens. (E,F) Mandibular tooth germs marked by the dlx2b:eGFP transgene in control (E) and Eda OE fish (F). Black brackets indicate a single row of teeth in E,F. White brackets in F indicate supernumerary tooth rows. Arrow in F indicates a tooth on the anterior basihyal (bh). (G,H) Alizarin Red oral skeleton preparations from control (G) and Eda OE (H) individuals. (I,J) Ventral pharyngeal tooth germs marked by the dlx2b transgene in control (I) and Eda OE fish (J). White arrows in J mark ectopic tooth germs. (K,L) Alizarin Red ventral pharyngeal skeleton preparations from control (K) and Eda OE (L) fish. Arrows in L mark ectopic teeth. Black arrow in L marks ectopic bone of attachment. Arrowhead in L indicates an instance of induced pharyngognathy, i.e. midline fusion of ventral tooth fields. Scale bars: 200 μm (B-D); 50 μm (E-H); 100 μm (I-L).

To test whether ectopic teeth could be activated at developmental stages after endogenous pioneer tooth differentiation, we overexpressed Eda during larval stages in the remaining individuals of the cross described above. We administered heat shocks once every 12 h from 17-25 dpf (16 total heat shocks), again followed by a 6-day recovery period (Fig. 2A), i.e. the same duration and cadence of heat shocks and the same recovery period as the first group but shifted to a later stage. We assessed GFP fluorescence in fish carrying dlx2b:eGFP, and later stained all individuals with Alizarin Red. We found that only 3/23 fish formed ectopic pharyngeal teeth: 2/23 demonstrated one or two cb4 teeth, and 1/23 demonstrated hb1 and bh teeth (Fig. S2). Thus, the potential to form ectopic pharyngeal teeth upon exposure to exogenous Eda is still maintained, albeit at a lower efficacy, long after endogenous tooth field establishment. Interestingly, a distinct class of ectopic teeth was additionally prompted by this treatment: we found that 20/23 of these Eda OE fish specified tooth organs on their face (Fig. 2B-E). These face teeth were typically located superficially to the anteriormost two infraorbitals (io1 or io2), as found in 20/23 Eda OE fish (Fig. 2C, arrows). Face teeth were also observed on the ventrolateral sides of the dentary, articular and/or angular bones in 17/23 Eda OE fish (Fig. 2C, arrowhead). Alizarin Red staining revealed a general lack of bone of attachment surrounding the bases of most face teeth, unlike the observed effect in the pharynx (Fig. 2E-E″). Notably, when bone of attachment was observed, it was always on or near the ventral dentary.

Fig. 2.

Larval Eda overexpression in sticklebacks. (A) Outline of the late larval treatment. Fish were heat shocked twice per day from 17 to 25 dpf. Drawings modified from Swarup, 1958. (B,C) Larval Eda OE causes face tooth germ initiation in Eda OE (C) but not in control fish (B). n=23 OE fish and n=19 control fish. Arrows in C mark the ventralmost and dorsalmost infraorbital (io) teeth. Arrowhead marks a cluster of ventral dentary (v.dent) teeth. Gray outlines indicate the position of io1, which is shown dissected and stained in panels D and E. (D,E) Dissection of the first io from an Eda OE fish (E) reveals ossified teeth on the face, which were not present in control fish (D). Insets indicated in E are shown enlarged in E′ and E″. Scale bars: 100 μm (B,C); 50 μm (D,E).

Tooth differentiation on the stickleback face is negatively correlated with neuromast differentiation

To test whether face tooth differentiation was spatially biased, we performed another late larval treatment by administering two heat shocks per day to overexpress Eda from 20-36 dpf (32 total heat shocks). By counterstaining with DASPEI, which marks neuromasts (yellow in Fig. 3A-B′), we first documented the spatial distribution of face teeth by using neuromast positions as reference points (Fig. 3C). We also counted a subset of neuromasts (those that form nearest the ectopic teeth, referred to hereafter as ‘anterior neuromasts’) to test whether neuromast number demonstrated a relationship with face tooth number. DASPEI staining alongside the dlx2b:eGFP reporter revealed that these face teeth were highly spatially biased, usually situated near neuromasts (Fig. 3B′, arrowheads), and occasionally they appeared to form in place of neuromasts (Fig. 3B′, arrow), suggesting copatterning between neuromasts and Eda-driven ectopic teeth. We reasoned that if face teeth sometimes arose from repurposed neuromast progenitors, we would expect to see a negative correlation between anterior neuromast number and the presence of face teeth. We indeed found a significant negative linear relationship between the number of face teeth and the number of anterior neuromasts (P=0.006; Fig. 3D). Overall anterior neuromast number was also significantly reduced in the Eda OE condition (P=0.001; Fig. 3E). These results are consistent with the hypothesis that the neuromast cell linage donates cells to face tooth formation, though there are other possible explanations for this negative correlation, e.g. that Eda negatively affects neuromast differentiation directly and separately from its role in promoting tooth differentiation. In situ hybridization for Pitx2 on this same treatment condition revealed ectopic expression of this transcription factor (Fig. 3F,G), suggesting that indeed these are bona fide teeth/odontodes rather than generic ectopic bone that is coincidentally conical.

Fig. 3.

Face tooth location and negative relationship with anterior neuromasts in sticklebacks. (A,B) Left lateral views of control (A) and Eda OE (B) fish that underwent a larval heat shock treatment from 20-36 dpf. n=29 OE and n=12 control fish. DASPEI shown in yellow, eGFP from the dlx2b tooth reporter shown in green, mCherry shown in magenta. A′ and B′ show the insets indicated in A and B, without the brightfield overlay. Arrowheads in B′ mark ectopic face teeth growing nearby neuromasts. Arrow in B′ marks a face tooth growing in the position of a neuromast. (C) Diagram showing neuromast (black dot) position relative to face tooth (green dot) position at 36 dpf. (D) Scatterplot of anterior neuromasts (pictured as black dots in panel C) as a function of the number of face teeth. A statistically significant negative linear correlation was observed (P=0.006, linear regression). Gray band shows 95% confidence surrounding the line of best fit (blue). (E) A box and whisker plot showing a significant reduction in anterior neuromast counts for Eda OE versus the control condition (mock heat shock). Boxes represent the 25th-75th percentiles, the median is shown as a black bar, whiskers indicate minimum and maximum values. ***P=0.001 (Wilcoxon Rank Sum test). n=12 control versus n=29 Eda OE fish. (F,G) Pitx2 in situ hybridization revealed ectopic expression coincident with the region where ectopic teeth form in the Eda OE condition but not the control condition. Scale bars: 100 μm.

Eda overexpression causes accelerated differentiation of endogenous tooth germs

To ask whether endogenous tooth organ development could be accelerated by overexpression of a type A Eda, we again overexpressed Eda in sticklebacks on a dlx2b:eGFP background. We administered heat shocks once per day from 2-9 dpf (8× heat shocks), preceding endogenous oral tooth differentiation, and checked daily for eGFP signal. We observed accelerated pioneer tooth differentiation in the oral jaws (Fig. 4A-B′, arrowheads): all treatment animals (n=9/9) had four oral tooth germs at 9 dpf, whereas no control fish had any detectable GFP signal (n=0/10). Notably, the four GFP⁺ tooth germs observed in this arrangement – a bilateral pair of pioneer tooth germs on each the premaxillary and dentary tooth fields – is reminiscent of normal expression of the dlx2b enhancer at 12 dpf in wild-type (WT) fish (Fig. 4C), suggesting that endogenous tooth development is accelerated by Eda OE.

Fig. 4.

Eda overexpression causes precocious tooth differentiation in sticklebacks. (A,B) At 9 dpf, Eda OE fish (n=9) demonstrated precocious oral tooth germs compared with control fish (n=10). Images show eGFP from the dlx2b tooth reporter and mCherry from the OE transgene. A′ and B′ show the fish indicated in A and B in greater detail. Black arrowheads indicate pioneer tooth germs. (C) At 12 dpf in wild type (WT), pairs of germs are observed on the premaxilla and dentary (black arrowheads), similar to the condition seen in the 9 dpf Eda OE fish, suggesting these represent accelerated endogenous teeth. Scale bars: 100 μm.

Tooth formation competency in sticklebacks shows distinct dose-dependence and temporal requirements across different regions of the head

We next aimed to understand the anatomical and temporal windows of Eda-responsive tooth formation competency in sticklebacks by further varying the timing and number of heat shock treatments (i.e. ‘doses’ of Eda). We thus conducted a series of unique Eda OE treatments followed by an analysis of tooth ossification by Alizarin Red staining at 24 dpf for each treatment. See Fig. 5A for a heatmap showing the summary of results of each treatment, Fig. 5B for an illustrated map of tooth formation competency in the pharynx, and Table S1 for a description of major developmental events during this period. These experiments revealed that not only are different bones ‘primed’ for a tooth formation response during different developmental intervals, but also that different regions of the head skeleton demonstrate different propensities for ectopic tooth formation under a given heat shock regiment. For example, ectopic teeth were often observed anterior to the first epibranchial (eb1), positioned upon the medial-most gill raker that emanates from this element; these teeth displayed the widest interval during which they could be specified by a single heat shock, lasting from 4 to 10 dpf, compared to regions like hb1 and cb4 where the induction window was narrower, lasting from 5 dpf to 7 or 8 dpf, respectively. Some regions appeared to be generally more recalcitrant to tooth induction via Eda, for example the bh never produced any teeth following a single heat shock, whereas multiple heat shocks could produce teeth in this region, though always at a lower rate compared to regions like cb4 or eb1 in the same treatment. Similarly, face teeth were only observed in multi-heat shock treatments, and the infraorbital teeth were always induced at a higher rate compared to those on the ventral dentary.

Fig. 5.

Stickleback tooth formation competency through ontogeny. (A) Heatmap showing the percentage occurrence of ectopic tooth formation at different morphological locations. The leftmost column describes each treatment, each other column represents a bone or morphological region where ectopic teeth were observed in the percentage of fish shown. Treatments below the black bar consisted of multiple heat shocks, those above are single heat shock treatments. Ectopic teeth were not detected in any negative control fish [n=0/169 mCherry-negative, heat shocked fish (not listed); n=0/27 mCherry-positive, un-heat shocked fish]. See Figs 2 and 3 for io position in context. bh, basihyal; hb1-3, hypobranchial 1-3; cb4, ceratobranchial 4; eb1, epibranchial 1; io, infraorbital; v. dent, ventral dentary. (B) Alizarin Red-stained pharyngeal skeleton (left) and an illustration of the skeletal elements indicating whether each element ever bore teeth under at least one Eda OE treatment. Dorsal tooth plate 1 and 2, DTP1 and DTP2; ventral tooth plate, VTP. Other abbreviations same as panel A.

Tooth formation competency in zebrafish shows distinct dose-dependence and temporal requirements

To define the Eda dose and temporal requirements in zebrafish ectopic tooth formation, we performed an array of heat shock experiments using the zebrafish eda coding sequence (a type ‘B’ Eda) and analyzed ectopic dorsal pharyngeal tooth formation using Alizarin Red at 10 dpf (Fig. 6A,B). Similar to sticklebacks, we found that the rates of dorsal pharyngeal tooth formation in zebrafish upon eda OE were also dosage and time dependent (Fig. 6C). Zebrafish ectopic tooth formation competency also appears to peak surrounding endogenous pioneer tooth initiation, which occurs at ∼2 dpf (Van der Heyden and Huysseune, 2000): 24% of fish heat shocked at 2 dpf demonstrated at least one ectopic tooth. Conversely, only 6% and 3% of fish demonstrated at least one ectopic tooth when heat shocked at 1 or 3 dpf, respectively. Like in sticklebacks, multi-heat shock treatments usually demonstrated a higher rate of ectopic tooth formation, to the exclusion of a double heat shock at 3 and 3.25 dpf.

Fig. 6.

Zebrafish tooth formation competency though ontogeny. (A,B) Alizarin Red-stained pharyngeal tooth fields dissected from the 24+48+72 h experiment showing a pair of bilateral ectopic teeth (arrows). Oral views, dorsal to top. (C) Results of different heat shock treatments on ectopic tooth formation in zebrafish. Ectopic teeth were not detected in any negative control fish [n=0/36 mCherry-negative, heat shocked fish (not listed); n=0/9 mCherry-positive, un-heat shocked fish]. Red dots indicate the timing of heat shocks relative to the timeline shown at the top of C; treatments occurred at 24 h, 30 h, 48 h, 54 h, 72 h and/or 78 h post fertilization. Scale bars: 25 μm.

Ectopic pharyngeal teeth, but not face teeth, are renewed without additional Eda overexpression

Previous Eda overexpression experiments in zebrafish and Astyanax used a Xenopus elongation factor 1a (ef1a) promoter that drives transcription in most zebrafish cell types at all life stages, including in adults (Amsterdam et al., 1996; Johnson and Krieg, 1994; Moon et al., 2013). As this promoter remains active during adult stages, it remained unknown whether ectopic teeth could be maintained if Eda OE was discontinued. To test whether ectopic teeth could be retained in zebrafish after the cessation of Eda OE, we overexpressed zebrafish eda at 24, 30 and 48 hpf, followed by a 2-month recovery period, at which point we assayed upper pharyngeal histology using Hematoxylin and Eosin (H&E) staining on sections or by Alizarin Red wholemount staining (Fig. 7A,B). We found ectopic teeth were present in n=4/10 fish with the OE construct, and in all cases at least one of those teeth was an unerupted tooth germ associated with a nearby erupted tooth (Fig. 7B, arrows), indicative of ongoing new tooth growth 2 months after the last dose of Eda was administered.

Fig. 7.

Ectopic pharyngeal tooth fields undergo new tooth specification for months following discontinued Eda overexpression. (A,B) H&E-stained sagittal sections (anterior to left) from control (A) and eda OE (B) zebrafish at 2 months of age following three embryonic heat shocks. Tooth germs are indicated with arrows, a later-stage tooth is indicated by an arrowhead. (C-D′) Alizarin Red-stained pharyngeal skeletons from control (C) and Eda OE (D) sticklebacks 2.5 months after six heat shocks from 6-11 dpf. Anterior to top. Arrow in D′ indicates an ectopic tooth germ with evidence of active bone remodeling. (E) Live bone staining pulse-chase assay was used to observe new tooth production and tooth retention following embryonic Eda OE. E shows Alizarin Red, E′ shows Calcein, E″ shows an overlay. Teeth with any Alizarin signal contain retained bone and were thus present at the pulse (‘retained’). Teeth with only Calcein signal contain only newly formed bone and thus were not present at the time of the pulse, but only the chase (‘new’). Arrowhead in E″ indicates an active tooth replacement event. Scale bars: 10 μm (A,B); 200 μm (C,D); 100 μm (E).

We next asked whether stickleback ectopic pharyngeal teeth demonstrated a similar phenomenon. First, we overexpressed Eda in sticklebacks once per day from 6-11 dpf and allowed a 2.5 month recovery period before analysis of skeletal morphology by Alizarin Red (Fig. 7C,D). We found that n=11/13 Eda OE fish had ectopic teeth in at least one location within their pharynx, with most fish (n=9/13) demonstrating four or more distinct ectopic tooth fields. In each case, evidence of ongoing bone remodeling beneath unerupted tooth germs was present, typified by punctured bone on the underlying skeletal element (Fig. 7D′, arrow). Thereafter we performed a follow-up experiment where we overexpressed Eda at 5 dpf (1× heat shock) and allowed the fish to recover until 4 months of age, at which point we used pulse-chase live bone staining to assay tooth turnover (see Materials and Methods). We found that all ectopic pharyngeal tooth fields were actively undergoing tooth turnover, as evidenced by the presence of Alizarin-negative, Calcein-positive teeth (‘new’ teeth) in 5/5 fish with ectopic tooth fields (Fig. 7E-E″), in some cases with clear signs of retained tooth resorption adjacent to a new tooth germ (Fig. 7E″, arrowhead).

We next asked whether face teeth in sticklebacks were capable of being maintained or renewed. We subjected fish to a twice daily Eda OE treatment from 17-27 dpf (20× heat shocks), keeping only those fish that grew at least one face tooth for later sampling, as evidenced by dlx2b reporter expression at 28 dpf. Subsets of fish were thereafter stained with Alizarin Red at different timepoints. At least one face tooth was observed in the following fractions of fish at the specified recovery timepoints: 6 days, 92% (n=11/12); 12 days, 50% (n=4/8); 18 days, 8% (n=1/13). The single ectopic tooth observed at the final collection point was ankylosed to the dentary and no longer demonstrated GFP fluorescence from dlx2b:eGFP. Throughout all collection points, we observed no clear morphological evidence of tooth replacement, nor any bell-stage dlx2b reporter-positive tooth germs at the 12 or 18 day collection points. Thus, Eda-induced face teeth appear to be incapable of renewal after Eda OE treatments are discontinued.

Expression domains of Edar and Troy correlate with tooth formation competency

We hypothesized that Eda-driven ectopic tooth formation competency would likely be coincident with the endogenous expression of one or more TNFRs. Using hybridization chain reaction (HCR), we first analyzed the expression of Edar, Troy (also known as TNFRSF19) and Relt (TNFRSF19L), three closely related TNFR genes all previously implicated in epithelial organogenesis (Charles et al., 2009; Harris et al., 2008; Kim et al., 2019; Kojima et al., 2000; Ohazama and Sharpe, 2004). We first assayed the expression of these three TNFR genes at 6, 8 and 12 dpf (Fig. S3A-C). Edar and Troy were expressed in tooth germs as well as some naïve or non-dental cell types (detailed below), rendering them as plausible candidate genes for involvement with ectopic tooth formation. Relt expression, on the other hand, was detected in a highly specific manner only in ameloblasts (mature inner dental epithelium) in mid- to late-bell-stage tooth germs. At 6 dpf, during the most sensitive interval for ectopic tooth formation in the ventral pharynx, Relt was not detected anywhere in the ventral pharynx, rendering it a poor candidate for involvement with ectopic tooth specification.

We thus expanded our expression survey to 6, 10 and 14 dpf, and assayed the expression of Edar and Troy alongside Eda (Fig. 8). Notably, at 6 dpf we observed Troy- and Edar-positive cell populations along the ventral pharynx, approximately coincident with the medial regions of hb1-3 and cb4 where ectopic teeth were observed following Eda OE (Fig. 8A,B, gray boxes). Edar and Troy were both expressed in all observed cap-stage tooth germs, including the pioneer teeth on cb5 (Fig. 8A′,A″,B, black arrowhead). Edar was expressed more robustly in bell-stage tooth germs compared to Troy (Fig. 8C′,C″,D′,D″, white and gray arrowheads). Troy expression was additionally detected in tastebuds at 10 and 14 dpf (Fig. 8C″, white arrow), as well as gills and their presumed precursor domains (Fig. 8A″,C″, carets). Edar transcripts were also detected in gill rakers (Fig. 8D′, double black arrow). Eda expression was detected in mesenchyme in the tooth field, particularly in the cells surrounding tooth germs on the medial side of the cb5 tooth fields (Fig. 8A‴,C‴,D‴, black arrows). Eda expression was also weakly detected in mesenchyme medial to cb4 and hb3 (Fig. 8C‴,D‴, brackets), as well as in epithelium flanking the gill rakers at 10 and 14 dpf (Fig. 8D‴, double white arrow). At higher resolution, Eda, Edar and Troy were detected in dynamic patterns in both dental epithelia and mesenchyme (Fig. S4).

Fig. 8.

Edar, Troy and Eda expression in the stickleback ventral pharynx and face. (A-D) Ventral pharyngeal preparations of wild-type (WT) sticklebacks showing Edar, Troy and Eda expression at 6, 10 and 14 dpf. Anterior to top in all panels, except for B, where anterior is to the left. A, C and D show maximum intensity projections of z-stacked optical slices, B shows a sagittal plane orthogonal reconstruction as indicated in A⁗. Gray boxes indicate the competency region along the ventral pharynx. Black arrowheads in A′, A″ and B mark the right-side pioneer tooth on cb5. White arrowheads in C′ and D′ mark bell-stage tooth germs. Double black arrow in D′ indicates a gill raker. Carets in A″ and C″ indicate gill precursors and differentiating gills, respectively. Gray arrowheads in C″ and D″ mark bell-stage tooth germs with little detectable expression. White arrow in C″ marks a taste bud. Black arrows in A‴, C‴ and D‴ indicate medial mesenchyme at cb5. Brackets in C‴ and D‴ indicate medial mesenchyme near hb3 and cb4. (E) Edar and Troy expression in a flattened facial prep at 8 dpf. Anterior to left, the eye is demarcated with a dashed line and the mouth opening is labeled ‘mo’. White arrow in E indicates ventral pharyngeal muscles. Black arrows in E′ indicate neuromast lines. Scale bars: 100 μm (A,C,D,E); 25 μm (B).

To ask whether zebrafish also express edar and troy in their ventral pharynx despite lacking a tooth formation response in these regions under eda OE, we also assayed the expression of these receptors via HCR at 2 dpf, the timepoint we found to have the highest rate of ectopic tooth formation. We found that troy was expressed throughout the ventral pharynx, similar to our observations of stickleback Troy at 6 dpf, but edar was not detected in this same region as it was in sticklebacks (Fig. S3D). We additionally detected both TNFRs in the pioneer tooth germ at this stage (black arrowhead in Fig. S3D).

To test whether Edar, Troy or Relt were expressed in a manner that might presage face teeth, we again tested for the expression of these three TNFRs at 8 dpf, representing the midpoint of an Eda OE treatment, whereby we found that >50% of fish formed facial teeth in the infraorbital region. Relt expression was not detected. Edar expression was observed in ventral pharyngeal muscles but was not detected in any region in which face teeth form (Fig. 8E). Troy, on the other hand, was found to mark neuromast primordia as well as interneuromast cells (Fig. 8E′). Thus, Troy but not Edar or Relt expression marks the facial regions where teeth are competent to form.

To assess the response of stickleback TNFR expression to Eda OE, we again performed HCR to detect Edar, Troy and Relt on control and Eda OE larvae. As expected, we found that Edar, Troy and Relt, which all mark endogenous teeth (Fig. S3), were expressed in nascent ectopic tooth germs on both the face and in the pharynx (Fig. S5). Additionally, we noted that neuromasts in the Eda OE condition (marked by Troy) displayed some disrupted organization relative to the control condition, especially those neuromasts with neighboring early tooth germs (marked by Relt expression).

DISCUSSION

Ectopic tooth specification via Eda overexpression is most efficient during pioneer tooth formation stages

Using different sets of heat shock treatments to induce Eda expression, we measured tooth formation competency through time for different regions capable of this response in sticklebacks and zebrafish. In both species, ectopic pharyngeal teeth can be activated via a lower number of Eda doses during stages where endogenous tooth field activation is taking place compared to earlier or later stages (Fig. S6). However, in sticklebacks, the competency to form teeth at some of these locations is maintained through late larval stages, though more Eda OE doses were required, which still yielded relatively low tooth formation rates. We thus infer that the regions of the head skeleton that retain this latent developmental potential undergo a short period during which they are most receptive to tooth induction via Eda, thereafter becoming more recalcitrant to tooth induction. Experiments using Eda mutants combined with Eda OE could resolve whether endogenous tooth fields follow a similar ‘competency arc’ whereby their disposition to tooth induction changes through ontogeny. If endogenous tooth fields do exhibit temporal changes in competency through time, this could have implications for human health applications: perhaps in utero treatment with Fc-EDA during pioneer tooth initiation stages at 6-8 weeks would restore a greater proportion of the dentition than did the treatment beginning at 26 weeks, which was more efficient at restoring sweat glands (Schneider et al., 2018), which initiate at 20 weeks of gestation and beyond.

One unexpected detail revealed by Eda in situ hybridization in sticklebacks is that Eda expression is present along the posterior midline of the pharynx, medial to the cb4 and hb3 competency domains (Fig. 8C‴,D‴, brackets). This endogenous expression domain may partly explain why cb4 is particularly inclined to undergo tooth formation, as there is already a baseline level of Eda present in this region, lowering the amount of exogenous Eda required to elicit the tooth formation response. However, the generally less competent hb3 is within this domain, challenging this explanation. We observed no unique epithelial appendages in this region of cb4, leaving the potential function (if any) of this Eda expression domain a mystery.

Pharyngeal tooth fields, once activated, appear to sustain themselves via regeneration

A temporally inducible promoter allowed us to control the number and timing of Eda doses. By initiating teeth at early stages and then rearing fish for several months afterwards, we were able to determine that ectopic pharyngeal tooth fields often persist long-term. The ectopic teeth present at these later stages were thus unlikely to have been exposed to exogenous Eda from the OE transgene. Instead, we hypothesize that these later-observed ectopic teeth arise from regeneration events at ectopic tooth fields. Our pulse-chase data in sticklebacks supports this hypothesis, whereby in some cases the arrangement of new and retained teeth was highly consistent with tooth replacement (Fig. 7E″, arrowhead). Further experiments using cell lineage tracing could determine the origin of these late ectopic teeth.

Interestingly, face teeth do not appear to have the capacity to replace themselves by engaging in tooth regeneration. Only 1/13 individuals with confirmed face tooth initiation maintained a tooth 18 days after the treatment ended, and this single tooth represented a rare case where it had become ankylosed to the underlying dentary. Two possible explanations for this lack of regenerative potential are: (1) ectopic face teeth are inherently different from pharyngeal teeth, and fail to specify successional epithelia and/or mesenchymal progenitors that are required for the maintenance of a tooth field, or (2) the lack of ankylosis on the face leads to the rapid loss of most face teeth due to epidermal sloughing or mechanical disturbances, causing them to simply fall off before successional cell types can initiate a replacement tooth organ. Thus, the single retained face tooth we observed may have been poised to undergo regeneration. Further analyses of gene expression associated with successional dental cell types, such as Sox2, Lgr6 and Nfatc1 (Juuri et al., 2012; Martin et al., 2016; Rasch et al., 2016; Square et al., 2021) and additional experiments with longer observation windows could resolve these possibilities.

The evolution of tooth organ specification

Reptiles, amphibians and all major groups of fishes have member species that specify teeth on at least one bone of the head skeleton besides the dentary or maxilla (Berkovitz and Shellis, 2023). Given that individual teeth are widely accepted as homologous organ units (Sire et al., 2009), it is apparent that the relocalization of tooth initiation is not only developmentally feasible, but has occurred often. This lability lends support to a model where, at least in some species, regions of the head skeleton are predisposed to be competent to execute a tooth formation genetic program, having maintained a ‘latent developmental potential’ over the course of tens or even hundreds of millions of years of evolution (Aigler et al., 2014; Jandzik and Stock, 2021). Such developmental potential could ostensibly require only one or a few mutations to reactivate. For example, we predict that gene regulatory changes that cause an expansion of Eda expression domains should be sufficient to reactivate tooth fields in species like sticklebacks and zebrafish.

At least three experimental conditions have been shown to increase primary tooth number or cause ectopic teeth to form: exogenous Eda and retinoic acid (RA) have been shown to cause ectopic tooth fields in fish models (Aigler et al., 2014; Gibert et al., 2019; Jandzik and Stock, 2021; Seritrakul et al., 2012), while Osr2 loss-of-function has been shown to expand the endogenous molar tooth field in mice (Gibert et al., 2019; Zhang et al., 2009). However, published RA treatments and Osr2 loss-of-function that yield to additional tooth formation are also lethal. This leaves Eda as a strong candidate gene for potential involvement in dental rearrangements over evolutionary time because, experimentally and clinically, it is tolerated in excess, or as a null allele. The widespread tolerability of changes to Eda titer suggests that most mutations that affect Eda expression would at least not be lethal, making Eda genes excellent targets for selectable mutations.

As previously hypothesized, here we provide additional evidence that Eda was duplicated at the TGD, creating type A and B paralogs, followed by differential retention and loss of Eda paralogs in various teleost groups. Here, we show that stickleback fish subjected to overexpression of their own Eda (type A) are also capable of prompting ectopic tooth development, notably in many regions where other related fish species specify teeth. For example, some bass species (Beloniformes), which as Neoteleosts, like sticklebacks, we infer possess only type A Eda, are known to demonstrate teeth on hb2, hb3 and the bh (Gosline, 1985), all of which are regions where sticklebacks are capable of specifying ectopic teeth under Eda OE conditions. Whether differences in the amino acid sequence of Eda proteins underlie these dental competency differences is unknown, which could be tested by overexpressing Eda in a cross-species manner (e.g. overexpression of stickleback Eda in zebrafish). Notably, overexpression of zebrafish eda in Astyanax caused ectopic teeth to form in the ventral pharynx (Jandzik and Stock, 2021), suggesting that not all observed differences in dental competency across species are due to Eda amino acid sequence. Regardless, both type A and type B teleost Eda duplicates, which diverged ∼350 million years ago during early teleost evolution, have retained the capability to prompt ectopic teeth.

Contrary to all observed ectopic pharyngeal tooth fields, the face teeth observed in sticklebacks do not correspond to teeth widely observed in other teleosts. Notably, there is at least one species that does exhibit such teeth: the denticle herring (Denticeps clupeoides). This species indeed specifies small teeth (denticles) on its infraorbitals, though notably it similarly adorns additional regions of its face with teeth (Greenwood and Greenwood, 1968), including the braincase and branchiostegal rays, where we never observed ectopic teeth in our experiments. Taking a broader view, most Chondrichthyan species specify denticles across their exterior (Nicklin et al., 2024), leaving open the possibility that their natural presence in Denticeps and their experimental appearance in sticklebacks could be a harkening back to a rarely-used, but persistent, developmental competency. Eda OE experiments in additional fish species will help resolve whether this competency is present beyond a select few species. On the other hand, it remains possible that face teeth in sticklebacks represent a fundamentally different kind of cellular transformation than is observed in the pharynx: rather than activating a latent developmental potential to form bona fide tooth fields, perhaps what occurs on the face realizes only enough of the tooth development program to form individual teeth that otherwise lack some crucial surrounding components that comprise a true dental arcade. This alternative is supported by the apparent inability of face teeth to be maintained, unlike the pharyngeal teeth, and the rare occurrence of any bone of attachment. With this in mind, we speculate that ectopic tooth initiation on the face may not necessarily use the same pathways or downstream genetic mechanisms as occur inside the pharynx, despite their common entry point (Eda activation). Further experiments with various TNFR mutants will help resolve the pathway used to activate ectopic teeth in the pharynx and on the face.

Expression patterns of TNFR genes give insights into potential receptor utilization during ectopic tooth formation

We performed expression analyses of the gene encoding the canonical Eda receptor, Edar, and two of its most closely related receptor genes, Troy and Relt, to determine which of these TNFRs, if any, are expressed in a manner that presages ectopic teeth. These analyses revealed that Relt expression was detected in a strikingly specific manner in mature ameloblasts and is thus a poor candidate gene with respect to the initial ectopic tooth formation response. On the other hand, stickleback Edar and Troy expression both marked regions of the ventral pharynx where ectopic teeth were capable of forming, while in zebrafish these regions only exhibited troy expression. In the infraorbital and mandibular regions, where face teeth most typically form, we found only Troy expression in nascent neuromasts and interneuromast cells. Given our observation that face tooth formation and neuromast counts are anticorrelated, we speculate that Eda can signal through Troy in certain contexts, and that this interaction can cause cells in the neuromast lineage to take on a fate of the dental epithelium. Notably, neuromasts are not present inside the pharyngeal cavity, thus ectopic pharyngeal tooth epithelium is highly unlikely to follow this same developmental route.

Despite the associations we observed between TNFR expression and ectopic tooth specification, other factors are clearly required for the ectopic tooth formation response. First, we detected Edar or Troy expression in stickleback tissues that never formed ectopic teeth in our experiments (e.g. gill and gill raker precursors). Second, publicly available single cell RNA-seq data suggests that zebrafish neuromasts likely express troy at 5 dpf (Lange et al., 2023), suggesting that troy-positive neuromast cells are not necessarily competent to form teeth under Eda OE in zebrafish. Third, zebrafish edar expression has been previously shown at the dentary and premaxilla, where ectopic tooth formation is not triggered by Eda OE (Aigler et al., 2014). Fourth, zebrafish edar expression in the dorsal pharynx is promoted by Eda OE, suggesting that Eda can somehow upregulate the expression of its own receptor in a region where it was not detected in WT (Aigler et al., 2014). Additional factors are thus likely required beyond one or more TNFRs to allow ectopic tooth formation in response to Eda. Further tests using Troy and Edar mutants with concurrent Eda overexpression could test whether either of these receptors are necessary for conferring ectopic tooth formation competency. These and other experiments, such as those examining the expression and function of other TNF signaling pathway components, like Edaradd, could additionally test whether the molecular basis of organ formation competence in this system is defined by differential or combinatorial expression of receptors and associated cellular machinery.

Conclusion

Here, we show that the competency to respond to ectopic Eda by initiating tooth organs shifts anatomically within the head and temporally across development. In both zebrafish and sticklebacks, ectopic pharyngeal tooth fields can maintain themselves through the sustained addition of new teeth in the absence of additional heat shock-driven Eda expression. However, face teeth in sticklebacks do not appear to regularly engage in renewal or replacement and instead are usually lost within 18 days following the cessation of Eda OE. Finally, expression analyses of three TNFR genes in sticklebacks revealed that Edar and Troy, but not Relt, are expressed in a manner consistent with a potential role in conferring the ectopic tooth initiation response. Together these data reveal that during development, the competency to respond to a secreted signal and initiate organogenesis is surprisingly dynamic and suggest that differential receptor expression could be a molecular mechanism underlying this dynamic competency.

MATERIALS AND METHODS

Amino acid alignment and phylogenetic analysis

Eda amino acid sequences were collected from the National Center for Biotechnology Information (NCBI) GenBank and Ensembl. Accession numbers for each sequence are listed in Table S2. Genomicus versions 99-106 were additionally used to browse synteny data and predicted orthology relationships (Nguyen et al., 2022, 2018). Amino acid sequences were aligned with MUSCLE (Edgar, 2004) in MEGA7 (Kumar et al., 2016). The tree was inferred by using the Maximum Likelihood method and Whelan and Goldman (WAG) model (Whelan and Goldman, 2001) in MEGA11 (Tamura et al., 2021). The tree with the highest log likelihood (−10,634.31) is shown in Fig. S1. Initial tree(s) for the heuristic search were obtained by applying the Neighbor-Joining method to a matrix of pairwise distances estimated using the Jones-Taylor-Thornton (JTT) model (Jones et al., 1992). This analysis was performed on 37 amino acid sequences from 26 species. All positions in the alignment with less than 80% site coverage were eliminated, leaving a total of 337 positions used to build the tree. A bootstrap analysis was conducted with 100 replicates; the percentage of trees in which the associated taxa clustered together is shown near each node in Fig. S1. The resulting tree was rearranged and some formatting was performed in FigTree version 1.3.1 (http://tree.bio.ed.ac.uk/software/figtree/).

Transgene plasmids

Plasmids for transgenesis were created as previously described (Square et al., 2023). The zebrafish eda coding sequence (NCBI reference sequence NM_001115065.1) with an antisense T3 site after the stop codon on the 3′ end was synthesized by Gene Universal and restriction cloned into the pT2overCherry construct using XbaI and XhoI. Stickleback Eda was codon optimized and synthesized by Integrated DNA Technologies. The codon optimized gene was designed to output a protein product that matches NCBI reference sequence XP_040030617.1. Codon optimized stickleback Eda was then restriction cloned into pT2overCherry using XbaI and XhoI.

The 4122 bp zebrafish dlx2b promoter and upstream regulatory region was amplified from zebrafish genomic DNA using the primers 5′-GAGTCATTTTGATCTGGAGAAAGCTGATG-3′ and 5′-TTCGCAGGAAGAAGAGACTACTCAACG-3′. The product was reamplified with the primers 5′-gccgggatccGAGTCATTTTGATCTGGAGAAAGCTGATG-3′ and 5′-ggtggtgtcgacCGCAGGAAGAAGAGACTACTCAACG-3′ to add BamHI and SalI to the 5′ and 3′ ends of the product, respectively (lowercase letters indicate spacers and restriction enzyme cut sites). The reamplified product was digested with BamHI and SalI. The pT2He plasmid (Howes et al., 2017) was digested with BamHI and SalI in parallel, and the smaller insert containing the heat shock promoter was discarded. The digested PCR product and remainder of the pT2He plasmid backbone (without the hsp70l promoter) were ligated together, creating pT2_Dr_d2b, which drives eGFP expression directly under the control of the 4122 bp zebrafish promoter and upstream regulatory region, recreating the main components of the previously published transgene cassette (Jackman and Stock, 2006). Once sequence verified with Sanger traces from each end of the new inserts, plasmids were midiprepped (QIAGEN), phenol-chloroform extracted, precipitated and resuspended in DEPC-treated water per standard methods.

Fish husbandry and transgenic line establishment

All husbandry and experiments were performed with approval of the Institutional Animal Care and Use Committee of the University of Florida (protocol IACUC202300000692) or the University of California-Berkeley (protocol AUP-2015-01-7117). Sticklebacks were raised in 110 l aquaria at 18°C in 3.5 g/l Instant Ocean salt and 39.4 mg/l sodium bicarbonate. Sticklebacks [Cerrito Creek (CERC)] were fed a common diet of live Artemia as young fry, live Artemia and frozen Daphnia as juveniles and frozen bloodworms and Mysis shrimp as subadults and adults. Zebrafish (AB strain) were raised according to standard methods (Sprague et al., 2001). At the end of each experiment, fish were euthanized with MS-222. Reporter transgene fluorophores and DASPEI-stained fish were imaged without fixation or alcohol exposure. Alizarin Red skeletal stains, pulse-chase labeling and in situ hybridization material were then fixed in 4% formaldehyde overnight at 4°C or for 4 h at room temperature (RT). Fixative was washed out with one rinse and two 10 min phosphate buffered saline with 0.1% Tween-20 (PBST) washes. Live pulse-chase bone labeling was imaged without alcohol exposure, all other assays were followed by rinsing and storage in methanol (in situ material) or ethanol (skeletal stains or H&E).

Transgenesis was accomplished in sticklebacks and zebrafish by coinjecting Tol2 mRNA and prepared plasmids described above. In the case of the Eda OE and dlx2b reporter transgenes in sticklebacks, fish were injected (Erickson et al., 2016), outcrossed and screened as previously described (Square et al., 2023) to achieve and maintain a single insertion transgenic background for all experiments herein. The zebrafish eda OE line (ZFIN accession bk407tg) was uniquely and intentionally maintained as multi-insertion in order to increase the heat shock dosage due to the generally weaker effects of this transgene. For the experiments shown in Fig. 7C, all assessed individuals resulted from outcrosses of the same male fish, who we infer carried three insertions of the transgene (201/237 mCherry⁺, ∼84.8%).

Heat shock transgene activation

Heat shocks were administered essentially as previously described, with some modification (Square et al., 2023). Zebrafish were confined to a 50 ml tube in 25 ml of embryo medium and placed in a 38°C water bath for 65 min (it takes ∼5 min for 25 ml of fish water to increase ∼10°C). For sticklebacks, single heat shocks before 10 dpf were performed as described for zebrafish using 50 ml tubes and a water bath. For multiple heat shocks in sticklebacks, or for heat shocks occurring between 10 and 30 dpf, stickleback fry were housed in a 2 l tank with light aeration and one 50 W aquarium heater set to 29°C, which would be powered on in two or more 2 h intervals (it takes ∼50 min for 2 l of fish water to increase ∼10°C). When necessary, fish were lightly anesthetized with MS-222 and sorted based on their fluorescence profile. ‘Control’ fish refer to heat-shocked siblings that did not inherit the OE transgene and experienced the same heat shock regiment as their OE transgene-carrying siblings, except those shown or described in Fig. 1D ‘Control (HS-)’, Fig. 5A ‘0 shocks’ and Fig. 6B ‘0 shocks’, which indicate OE transgene carriers that were not heat shocked.

Alizarin Red skeletal staining

Alizarin Red S was used for skeletal staining essentially as previously described (Ellis and Miller, 2016). Samples were stained in a 0.008% Alizarin Red S solution in 1% KOH for at least 24 h, replacing the solution if necessary (if the solution turned pale). Large samples from subadults were sometimes left in Alizarin Red staining solution for up to 1 week. Once adequately stained, samples were transferred to 0.5% or 1% KOH for 1-3 days to continue clearing, if necessary. Tooth plates were then dissected and scored and/or imaged by further clearing in 50% glycerol for 1-5 days before flat mounting in 50% or 90% glycerol in 1× PBS.

Live pulse-chase bone labeling

Live pulse-chase bone labeling on sticklebacks was performed essentially as previously described (Square et al., 2023). Following heat shock treatments and long recovery windows as described above, fish were placed into a tank containing an Alizarin Red live staining solution (0.1 g/l Alizarin Red S with 1 mM HEPES) made in otherwise normal tank water (Ellis et al., 2015). Sticklebacks were pulsed with Alizarin Red for 24 h in 2 l of Alizarin Red live staining solution. Fish were then rinsed twice and washed twice for 20 min in normal tank water before returning to their standard 110 l tank in the vivarium. After 18 days, fish were chased with Calcein live staining solution (0.05 g/l Calcein with 1 mM sodium phosphate in normal tank water) for 16 h in 2 l. Fish were then rinsed twice and washed four times for 20 min in normal tank water before being euthanized and fixed. The next day, samples were rinsed once and washed twice for 20 min in tap water, and pharyngeal skeletons were dissected out and treated with 1% KOH overnight at RT (without a storage/ethanol step). Skeletons were then rinsed in tap water and washed twice for 5 min in PBST and prepared for mounting and imaging by stepping the samples through 30, 60 and 90% glycerol in 1× PBS. Samples were then flat-mounted and imaged essentially as previously described (Ellis and Miller, 2016; Square et al., 2023).

In situ hybridization

To visualize mRNA distributions, in situ hybridization was carried out either by a traditional wholemount colorimetric method or HCR. Colorimetric wholemount ISHs were performed essentially as previously described (Square et al., 2021). The stickleback Pitx2 riboprobe was previously published (Ellis et al., 2016; Square et al., 2021). HCR was carried out using proprietary probe sets developed by Molecular Instruments and by following a modified version of the manufacturer's protocol, namely using a higher concentration and volume of probe and probe hybridization solution (16 pmol in 1.8 ml), and material fixation/preparation was performed as outlined above.

DASPEI staining

DASPEI{2-[4-(Dimethylamino)styryl]-N-ethylpyridinium iodide} was used to visualize neuromast position on live stickleback fry (Wark and Peichel, 2010). Sticklebacks were anesthetized and live stained with 0.02% DASPEI in normal tank water plus MS-222 for 5 min before washing in normal tank water plus MS-222 and imaging.

Sectioning and H&E staining

H&E staining was used to observe ectopic teeth in zebrafish. Following fixation, zebrafish heads were sectioned on the sagittal plane at a thickness of 7 µm. Slides were subjected to a series of staining and destaining washes, coverslipped and imaged as described previously (Square et al., 2021).

Supplementary Material

Table S2. Accession numbers and genome assembly information for sequences usedto reconstruct Eda amino acid sequence similarity in Fig. S1.

Common names, Latin names, protein names, accession numbers, and the assembly from which each coding region was derived are listed in Fig. S1 (See electronic spreadsheet file downloadable from the Development website).

Peer review history

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.204907.reviewer-comments.pdf