Getting out of a mammalian egg: the egg tooth and caruncle of the echidna

In the short-beaked echidna, Tachyglossus aculeatus, after an initial period of in utero development, the egg is laid in the pouch and incubated for 10 days. During this time, the fetuses develop an egg tooth and caruncle to help them hatch. However, there are only a few historical references that describe the development of the monotreme egg tooth. Using unprecedented access to echidna pre- and post-hatching tissues, the egg tooth and caruncle were assessed by micro-CT, histology and immunofluorescence, to map the changes at the morphological and molecular level. Unlike mammalian tooth germs that develop by invagination of a placode, the echidna egg tooth developed by evagination, similar to that of the first teeth in some reptiles. The egg tooth ankylosed to the premaxilla, rather than forming a mammalian thecodont attachment, with loss of the egg tooth post-hatching associated with high levels of odontoclasts, and apoptosis. The caruncle formed as a separate mineralisation from the adjacent nasal capsule, and as observed in birds and turtles, the nasal region epithelium expressed markers of cornification. Together, this highlights that the monotreme egg tooth shares many similarities with reptilian teeth, suggesting that this tooth is conserved from a common ancestor of mammals and reptiles.


16
The short-beaked echidna, Tachyglossus aculeatus, is one of five species in the mammalian 17 subgroup Monotremata. The other living members of the order are the three long-beaked 18 echidnas: Zaglossus attenboroughi, Zaglossus bartoni and Zaglossus bruijni, as well as the platypus, 19 Ornithorhynchus anatinus. It is estimated that monotremes diverged from therian mammals 20 ( have only a few teeth, showing the evolution away from tooth development and towards hardened 28 pads (Archer et al., 1985;Pascual et al., 1992). In contrast, all echidna fossils identified to date are 29 edentulous (Pascual et al., 1992). Since modern platypus nestlings still possess 3 cusped molars 30 which are lost by adulthood, the echidna is assumed to have lost cheek teeth after diverging from 31 the platypus. Monotremes have also lost some of the enamel forming genes over time and echidnas 32 have lost additional enamel genes compared to platypuses, which still have a thin enamel layer on 33 their cheek teeth (Luckett & Zeller, 1989;Zhou et al., 2021). This is consistent with other animals 34 that evolved edentulism, such as pangolins, birds and baleen whales which have lost one or more 35 enamel producing genes over time (Meredith et al., 2013;Meredith et al., 2014). 36 The only tooth which is retained in modern echidnas is the egg tooth, which is lost shortly after 37 hatching. An ancestral amniote characteristic conserved in monotremes and reptiles that 38 distinguishes them from therian mammals is oviparity. At the time the echidna egg is laid it is 15-39 17 mm in diameter and the embryo is at an early somite stage (Griffiths, 1968(Griffiths, , 1989Hughes & Hall, 40 1998; Semon, 1894). During the short incubation period, the embryo develops into a fetus and the 41 young hatches from the egg after only 10-10.5 days. The monotreme eggshell is leathery, porous 42 and consists of loosely wound keratinous fibres (Griffiths, 1968). Notably, monotremes develop 43 both an egg tooth and a caruncle to escape from their egg (Griffiths, 1968;Hill & De Beer, 1950;44 Hughes & Hall, 1998). 45 Reptiles and birds also possess either a caruncle or an egg tooth. Reptiles, turtles, Rhynchocephalia 46 and crocodilians have a caruncle, a thickened keratinised epithelium positioned above the nasal 47 cartilages (Alibardi, 2020). In contrast, squamates have a true tooth that can be either single or 48 paired (Fons et al., 20220). Even viviparous reptiles have an egg tooth, although it is smaller and 49 hidden under a layer of connective tissue (Hermyt et al., 2017). The vast majority of birds also use 50 an egg tooth to hatch out of their egg. However, the avian egg tooth is very similar to the caruncle 51 of turtles and crocodiles, consisting of a sharp, keratinized 'horn-like projection' rather than being 52 an actual tooth (Clark, 1961; Kingsbury et al., 1953;Wang et al., 2017). The structure and 53 appearance of the avian egg tooth is incredibly uniform across most of the one hundred avian 54 species examined so far (Wetherbee, 1959). Interestingly, the caruncle in monotremes is supported 55 by a bony protrusion known as an os caruncle, and therefore, is very different from that observed 56 in birds and reptiles (De Beer, 1949;Hill & De Beer, 1950). The relationship of the os caruncle to the 57 premaxilla has been debated as to whether it is a just an extension of the premaxilla or is an 58 independent ossification which fused with the premaxilla (Hill & De Beer, 1950 First the oral epithelium thickens, and sinks into the neural crest derived mesenchyme, which 64 condenses around the epithelium and creates the tooth germ; this process relies on both sonic 65 hedgehog (SHH) and fibroblast growth factor (FGF) signals for proliferation and stratification of the 66 dental placode (Li et al., 2016). The epithelium then extends into the mesenchyme, wrapping around 67 the condensing mesenchyme to form the tooth cap (Tucker & Sharpe, 2004). The primary enamel 68 knot forms at the tip of the tooth bud in the cap stage of development and acts as a signalling centre. 69 The bell stage then follows this, where cyto-differentiation occurs and it is here that odontoblasts 70 and ameloblasts form, which produce dentine and enamel respectively (Tucker & Sharpe, 2004). 71 Odontoblasts are columnar cells that form a uniform layer around the dental pulp cavity. They are 72 differentiated from the dental pulp and odontoblastic secretions form the dentine layer on their 73 outer surface (Balic & Thesleff, 2015). There are different types of dentine involved in tooth 74 development including primary, secondary and tertiary dentine. Osteodentine is a type of dentine, 75 named for its resemblance to bone, and is observed when tertiary dentine develops rapidly, and in 76 doing so, traps odontoblasts and other nearby cells (Avery, 1994). After the development of the 77 dentine layer, ameloblasts mature on the outer surface of the dentine and secrete proteins such as 78 amelogenin leading to enamel formation between the ameloblasts and the dentine layer (Sire et al., 79 2007). Enamel is the hardest substance in the body and covers the tooth crown. After the 80 development of this enamel layer the root of the tooth develops and the tooth erupts through the 81 gums (Thesleff & Juuri, 2012). 82 The mechanism of attachment for most mammals is thecodonty, which refers to the presence of a 83 tooth root being secured in a socket by a periodontal ligament (Spellerberg, 1982 (Manger et al., 1998). In echidnas, the egg tooth is first observed as a median conical 104 papilla protruding from the anterior end of the snout. Similar to the platypus, the echidna egg tooth 105 is described as being attached to the premaxilla and containing a dental pulp covered in dentine 106 (Hill & De Beer, 1950;Seydel, 1899). Here we have, therefore, focused on the echidna egg tooth and 107 caruncle, to understand the potential homology of these structures with their functional equivalents 108 across birds and reptiles, and to address issues such as method of attachment and mechanisms of 109 loss. 110

111
The fetal echidna egg tooth forms by evagination of the epithelium 112 To follow egg tooth development in the echidna post-oviposition fetal samples were imaged and 113 sectioned. Externally, no egg tooth was visible in the day 4 fetus but by day 6 post-oviposition 114 onwards, the egg tooth was clearly visible ( Figure 1A-C). This was confirmed with the histology of 115 the echidna fetal heads which showed substantial changes in growth, particularly in the tooth 116 region, between day 4 and day 6 ( Figure 1D, G). In the day 4 fetus, the epithelium was thickened 117 and expressed SHH, confirming the egg tooth was at the placode stage of development ( Figure 1F). 118 No evidence of invagination of the placode into the underlying mesenchyme was evident but the 119 mesenchyme under the placode had started to condense. At day 6 the egg tooth projected outwards 120 forming a tooth by evagination, in contrast to the invagination observed during eutherian 121 mammalian dental development ( Figure 1G-H). SHH remained expressed in the evaginating 122 epithelium, with an additional positive domain in the oral epithelium further back in the mouth 123 ( Figure 1I). 124 The fetal echidna egg tooth forms a dentine layer continuous with the premaxilla bone 125 In the day 6 fetus, a mineralised layer had begun to form but the dental pulp had only just started 126 to differentiate, and the cells were very densely packed ( Figure 1H). By day 7.5, the dentine layer 127 was clearly distinct with odontoblasts lining the inner surface, and the cells of the dental papilla 128 appeared much more loosely packed (Figure 2A, E, I). The epithelium around the egg tooth was 129 thickened at the tip with the formation of polarised cells that had a morphology similar to 130 ameloblasts. However, an enamel layer was not evident at any of the stages examined ( Figure 2K, 131 L). The smooth layer of dentine at the tip of the egg tooth gave way to osteodentine at the base, 132 with mesenchymal cells trapped within the dentine matrix ( Figure 2I, J). This region appeared 133 continuous with the premaxillary bone, as confirmed by Alizarin red staining ( Figure 2D, H). No root 134 was observed at any of the stages examined. The egg tooth is therefore, anchored directly to the 135 bone. 136 The premaxilla unites the egg tooth and caruncle 137 The caruncle was first detected in the day 7.5 fetus and was clear in the day 8 fetus as an ossified 138 circle of mesenchyme pressed against the nasal capsule ( Figure 3A, B). Unlike the nasal capsule, 139 which expresses high levels of Type II collagen, the caruncle did not express this cartilage marker, 140 and stained with alizarin red, suggesting it is indeed a bone ( Figure 3B, D, E, I). The caruncle could 141 be seen to attach directly with the premaxillary bone, which extended upwards over the nasal 142 cartilages, thereby, uniting the egg tooth and caruncle ( Figure 3G, H). To confirm the arrangement 143 of these skeletal elements, micro-CT was used to show the tissue in 3D ( Figure 3C, F Supplementary 144 Figure 3). The dentine of the egg tooth could be observed at day 8 connecting to the trabecular 145 bone of the premaxilla, with two arms of bone extending up from the premaxilla to contact the 146 caruncle ( Figure 3F). The development of the premaxilla was more advanced than the other bones 147 of the head, highlighting the need for this region to develop in advance of the rest of the structures 148 to support the egg tooth and caruncle. 149 The epithelium overlying the osseous caruncle was thickened compared to the epithelium over the 150 rest of the head as has been previously reported ( Figure 4A) (Hill & De Beer, 1950). To observe the 151 pattern of differentiation in more detail, markers of the different layers of epithelium were 152 investigated. Keratin 14 is often used as a marker of the basal epithelium and is expressed in 153 mammalian skin from early stages of development as the cells commit to stratification (Koster & 154 Roop, 2004). Around the head, Keratin 14 was expressed in the basal layer of the epithelium, 155 however, in the caruncle it was expressed in the intermediate layer ( Figure 4B, C). Keratin 8, another 156 marker of uncommitted epithelium (Liu et al., 2013), was also expressed in the head epithelium but 157 was absent from the caruncle ( Figure 4D, E). Finally, Loricrin, a marker of terminal differentiation, 158 was only observed in the caruncle epithelium, in the top layer ( Figure 4B), confirming the caruncle 159 epithelial cells had formed a cornified layer. 160

Post-hatching loss of the egg tooth by apoptosis and odontoclast activity 161
Loss of the egg tooth was followed by micro-CT post-hatching ( Figure 5A-C). The egg tooth, clearly 162 evident on the day of hatching was lost in the day 4 pouch young but was evident in some specimens 163 at day 3 ( Figure 5B-C). The mechanism of loss was therefore followed by TRAP staining to identify 164 clast cells and by TUNEL for apoptosis. At the day of hatching very few TRAP positive cells were 165 evident in the egg tooth and surrounding tissue ( Figure 5D). In contrast at day 2, just before loss of 166 the tooth, high numbers of TRAP positive cells were evident, particularly within the tooth against 167 the dentine layer ( Figure 5E). Thinning and breaks in the dentine were evident, suggesting that the 168 tooth was removed by odontoclasts. High levels of apoptotic cells, as labelled with TUNEL, were 169 evident in the dental pulp, particularly at the border with the premaxilla, highlighting loss of this 170 tissue after hatching ( Figure 5F). A combination of detachment of the dentine and death of the cells 171 of the pulp therefore led to its loss. This mode of loss was also observed in the platypus, where 172 breaks in the osteodentine were evident, flanked by multinucleated clast cells (Supplementary 173 Figure 5). Multinucleate clasts cells were also identified in the dental papilla up against the dentine, 174 suggesting that this mode of removal from the inside is conserved in monotremes (Supplementary 175 Figure 5C). In contrast to the egg tooth, no TRAP staining was evident in the echidna caruncle up 176 until day 4 pouch young (data not shown). In keeping with this, micro-CT analysis indicated that the 177 caruncle was still present in a day 11 pouch young but was no longer evident by approximately day 178 50 ( The mature echidna egg tooth at hatching was comprised of layers of dental pulp, odontoblasts and 203 dentine. However, whilst the odontoblasts were comprised of a clearly defined layer, they were 204 more irregular in shape compared to the uniformly columnar odontoblasts typically identified in 205 mammalian teeth. These more rudimentary odontoblasts are likely due to the temporary nature of 206 the tooth. The dentine layer developed rapidly. Whilst the egg tooth was not fully mineralised at 207 day 6, it was at its maximum thickness and mineralised only 1.5. days later. Histological examination 208 of the dentine also revealed that it contained trapped cells at the base of the tooth, which 209 resembled trabecular bone making it difficult to distinguish where the tooth ended and the 210 premaxilla started. We identified this layer as a type of dentine known as osteodentine which can 211 superficially resemble bone due to its rapid formation that can cause cells to be trapped inside it 212 (Avery, 1994). This is consistent with the rapid formation of the echidna egg tooth and the 213 rudimentary development of the odontoblasts. In further contrast to other mammals, the trabecular bone of the echidna and platypus premaxilla 227 was fused to the egg tooth (ankylosed) rather than being held in a socket by a periodontal ligament 228 as occurs in all other mammals examined to date. The rapid projection of the tooth between days 4 229 and 6 and the absence of any invagination of the epithelium, means that a root was not evident as 230 roots form from invaginations of the epithelium (Thesleff & Juuri, 2012). In the absence of 231 invagination in the echidna, the dentine made direct contact with the bone for attachment. A fused 232 tooth is sturdier than a tooth attached via soft tissue (Jenkins & Shaw, 2020). An ankylosed 233 connection can also likely be established faster than a thecodont attachment. This would be an 234 advantage for the echidna since they have only a brief period of time to develop a structure strong 235 enough to allow them to hatch from their egg. Acrodont attachments are also evident in squamate 236 egg teeth, with the dentine attached to the bone via an attachment tissue (Hermyt et al., 2020). 237 Before loss of the egg tooth, large numbers of TRAP positive clast cells were observed within the 238 egg tooth, lining the dentine at the base of the tooth. These cells appeared to be odontoclasts, as 239 they were observed adjacent to regions of thinned dentine, and thereby could break the connection 240 between the tooth and premaxilla. Similar multinucleated cells were observed in the platypus, 241 highlighting a conserved mode of tooth loss in monotremes. Removal of the dental pulp was then 242 completed by programmed cell death. 243 The premaxilla extended up and appeared to fuse with the os caruncle, thereby connecting the egg 244 tooth and caruncle into one functional unit. The premaxilla was one of the first bones in the head 245 to ossify, confirming its important support role for both the egg tooth and caruncle. In addition to 246 the formation of the os caruncle in the mesenchyme, the epithelium overlying this structure was 247 cornified, with a more complex stratified epithelium, with similarities to that observed in birds and 248 turtles (Alibardi, 2020). This suggests that while the mesenchymal os caruncle may be a novelty in 249 monotremes, the epithelial caruncle may be homologous to that of reptiles, suggesting that the 250 common amniote ancestor of reptiles and mammals had both of these characteristics. This would 251 suggest that reptiles without caruncles have lost these as they came to rely more heavily on the egg 252 tooth itself.

Micro-CT Scanning and 3-D reconstructions 276
One of the fetuses had already been collected and processed for histology before the 277 commencement of this project so was unable to be micro-CT (micro computed tomography) 278 scanned. All remaining echidna fetuses and pouch young samples were scanned by micro-CT. Micro-279 CT scanning of the echidna fetuses and the d11 py was performed with a phoenix nanotom m 280 (Waygate Technologies, Huerth, Germany) operated using xs control and phoenix datos|x 281 acquisition software (Waygate Technologies). An X-ray energy of 35-40 kV and 300 µA was used. 282 Scans were conducted using a molybdenum target to maximize contrast from the soft tissue 283 specimens. The voxel resolution was optimised to the size of the specimen and varied from 3.1 to 284 11.1 μm. The number of X-ray projections collected through a full 360-degree rotation was also 285 optimized to the size of the specimen, varying from 599 to 1798 projections leading to scan times 286 of 5 to 15 min in a fast scan mode (Table 1). The data was exported as 16-bit volume files for imaging 287 and analysis. All remaining pouch young samples were scanned using a Scanco micro-CT scanner 288 and images analysed using Amira (ThermoFisher Scientific, Waltham, MA, USA sections were submitted to the same procedures except that the first antibody was replaced by the 338 relative isotype control at the same concentration as the primary antibody. The specificity of all 339 antibodies was confirmed by staining in positive and negative control tissues and lack of staining in 340 the isotype controls. Images were acquired on a Nikon (Minato City, Japan) A1R spectral confocal 341 microscope housed at the University of Melbourne, or a Zeiss Apotome fluorescence microscope 342 housed at King's College London. Visualisation settings were initially optimised to eliminate 343 background using the isotype control on each slide. These settings were then used to visualise the 344 remaining positive sections on the slide. 345

346
The authors declare no conflict of interests. 347    Frontal views of egg tooth. A,B: The egg tooth has started to detach from the premaxilla with a break 569 evident on the LHS. C,D: Multinucleated cells are evident inside the dental pulp lining the dentine 570 (C), and at the border between the tooth and premaxilla in the regions of bone loss (D). 571