Tooth attachment and pleurodont implantation in lizards: Histology, development, and evolution

Abstract

Squamates present a unique challenge to the homology and evolution of tooth attachment tissues. Their stereotypically pleurodont teeth are fused in place by a single “bone of attachment”, with seemingly dubious homology to the three‐part tooth attachment system of mammals and crocodilians. Despite extensive debate over the interpretations of squamate pleurodonty, its phylogenetic significance, and the growing evidence from fossil amniotes for the homology of tooth attachment tissues, few studies have defined pleurodonty on histological grounds. Using a sample of extant squamate teeth that we organize into three broad categories of implantation, we investigate the histological and developmental properties of their dental tissues in multiple planes of section. We use these data to demonstrate the specific soft‐ and hard‐tissue features of squamate teeth that produce their disparate tooth implantation modes. In addition, we describe cementum, periodontal ligaments, and alveolar bone in pleurodont squamates, dental tissues that were historically thought to be restricted to extant mammals and crocodilians. Moreover, we show how the differences between pleurodonty and thecodonty do not relate to the identity of the tooth attachment tissues, but rather the arrangements of homologous tissues around the teeth.

A look into the teeth and jaws of modern lizards reveal similar tooth attachment tissues to those in mammals. This challenges over a century of assumptions about dental evolution.

1. INTRODUCTION

Tooth attachment and implantation in extant and fossil amniotes have enjoyed a research resurgence in recent years due to a revival of histology‐based methods of analysis (Bertin et al., 2018; Bramble et al., 2017; Brink et al., 2014; Budney et al., 2006; Caldwell et al., 2003; Chen et al., 2018; Dumont et al., 2016; Fong et al., 2016; García & Zurriaguz, 2016; Haridy, 2018; LeBlanc et al., 2017a, 2017b; Luan et al., 2009; Maxwell et al., 2011a; Snyder et al., 2020). These studies have challenged previous hypotheses of homology of the tissues that anchor teeth to the jaws (tooth attachment), the geometry of tooth implantation (i.e., thecodonty, pleurodonty, and acrodonty), and the phylogenetic significance of both dental features.

The early comparative works of Owen (1840), Tomes (1874, 1882), and Peyer (1968) partitioned the diversity of vertebrate dentitions into a series of seemingly discrete anatomical categories: teeth that are implanted into sockets are thecodont, teeth that are attached to the crest of the jaw are acrodont, and teeth that are attached to the lingual surface of the labial wall of the jaw are pleurodont. Thecodonty, i.e, where teeth appear to be set in discrete sockets, was thought to characterize only the teeth of mammals and crocodilians. This concept of thecodonty, which is still applied broadly today, was characterized histologically by a three‐part tissue attachment system, consisting of root cementum, a periodontal ligament, and the alveolar bone that forms the socket. These three periodontal tissues perform specific functions in mammals and crocodilians (Bosshardt et al., 2008; Lin et al., 2017; McIntosh et al., 2002; Miller, 1968). The cementum consists of acellular and cellular regions that serve as the anchoring points for the collagen fiber bundles of the periodontal ligament, which form Sharpey's fibers (Nanci, 2013). In mammals, the periodontal ligament is an unmineralized network of collagen fiber bundles, fibroblasts, sensory receptors, and other cellular components that suspends the tooth within its socket, helps it withstand the forces of occlusion, and serves a mechanosensory role (Nanci, 2013). The collagen fibers are arranged into groups around the tooth roots, serving specific roles in supporting each tooth. The periodontal ligament forms a similar supportive sling for each tooth in crocodilians, but shows more evidence of mineralization compared to its mammalian counterpart (Berkovitz & Sloan, 1979; McIntosh et al., 2002; Miller, 1968). The alveolar bone forms the other anchoring point for the fibers of the periodontal ligament and is a distinct bone tissue formed by the dental follicle rather than the bone of the jaw (LeBlanc et al., 2017a; Nanci, 2013; Cate & Mills, 1972).

In contrast, all other vertebrate teeth were characterized under a wide variety of types and subtypes of implantation geometries: pleurodont, subpleurodont, acrodont, subacrodont, and even subthecodont. All of these non‐thecodont implantation modes were thought to possess a single dental attachment tissue, “bone of attachment”, that fused or ankylosed the teeth to the jaws in the absence of histologically diagnosable cementum, a periodontal ligament, and alveolar bone (Howes, 1979; Peyer, 1968; Tomes, 1874; Zaher & Rieppel, 1999).

But through detailed histological work beginning with Caldwell et al. (2003), cementum, alveolar bone, and the periodontal ligament have been identified even in traditionally non‐thecodont amniotes (Budney et al., 2006; LeBlanc et al., 2017a, 2018; LeBlanc & Reisz, 2013; Maxwell et al., 2011a; Pretto et al., 2014). Moreover, the supposedly discrete tooth implantation categories are now seen merely as descriptors of continuous variables of jaw and tooth root architecture, making it difficult to partition tooth implantation into phylogenetically or biologically meaningful categories (Bertin et al., 2018; Caldwell et al., 2003; LeBlanc et al., 2017b).

Despite extensive research on the development and histology of mammal and crocodilian teeth (LeBlanc et al., 2018; Luan et al., 2009; Osborn, 1984; Ten Cate, 1997; Xiong et al., 2013), one extant group that has still received remarkably little study are the pleurodont squamates. On gross visual inspection, most squamate groups exhibit a stereotypical pleurodont mode of tooth implantation (Figure 1) (Zaher & Rieppel, 1999), which supposedly poses a conceptual challenge to the homology of amniote tooth attachment tissues: pleurodont‐implanted teeth are supposed to be attached to the jaws via “bone of attachment” (Howes, 1979; Luan et al., 2009; Zaher & Rieppel, 1999). For the remainder of this study we refer to this stereotypical, but surprisingly diverse squamate mode of tooth implantation as “pleurodont” for the sake of brevity. This is meant to be a general term, referring to the mode of implantation and not to any of the subtle variations in geometry that others have reported in squamates (Lessmann, 1952; Zaher & Rieppel, 1999).

FIGURE 1.

Pleurodont teeth in skeletonized specimens and in histological sections. (a) Lingual view of a skeletonized maxilla of the cordylid Cordylus cordylus (ROM R40) showing pleurodont teeth. (b) Isolated (but not shed) teeth from the marine iguana, Amblyrhynchus cristatus (AMNH 114492) showing the open labial surfaces of two pleurodont teeth (a functional and a replacement tooth). These teeth separated from the jaws after the specimen was skeletonized. (c) Coronal section of a pleurodont dentary tooth of the anguid Elgaria (Gerrhonotus) principis (UAMZ unnumbered). (d) Coronal section of a developing tooth in a Cordylus cordylus dentary (CMNAR 15418) showing the asymmetrical labial and lingual extensions of the developing root. Abbreviations: eg, enamel gap; La, labial; lw, lingual wall of the jaw; pl, pleura of the jaw; ss, subdental shelf. Black arrows in (d) indicate positions of the labial (right) and lingual (left) extensions of the developing root

The unusual shapes of the jaws and teeth of pleurodont squamates have also led many to conclude that they do not have true tooth sockets made of alveolar bone (Luan et al., 2009; Rieppel, 1978; Zaher & Rieppel, 1999). Instead, some researchers interpret squamate tooth attachment as anchorage to “apparent sockets” made from a distinct tissue known as an interdental ridge along the mesial and distal margins of each tooth (Luan et al., 2009; Zaher & Rieppel, 1999). These differences in turn have led to extensive debate regarding the nature of tooth attachment and implantation in squamates more broadly and have featured prominently in the debates over their evolutionary relationships (Budney et al., 2006; Caldwell et al., 2003; Kearney et al., 2006; LeBlanc et al., 2017a; Lee, 1997; Luan et al., 2009; Rieppel & Kearney, 2005; Zaher & Rieppel, 1999).

The aim of the present study is to describe and compare the development, histology, and ontogeny of teeth in a sample of extant squamates with varying degrees of pleurodont tooth implantation (we do not discuss acrodont squamates here as they are the subject of continuing study, but see Buchtová et al. (2013); Dosedělová et al. (2016); Haridy et al. (2018)). Through serial histology and Micro‐Computed Tomography (μCT) analyses of several squamate species, we describe the development and microanatomical properties of the attachment tissues and address whether or not pleurodont squamates produce cementum, periodontal ligament, and alveolar bone similar to other extant and extinct amniotes. With a sample of modern Iguana iguana, we also record the ontogenetic changes in tooth attachment and the development of the interdental ridges to determine their origins. We then compare Iguana‐type tooth implantation and attachment to two radically different groups (varanids and tupinambine teiids) to document variation in squamate tooth tissue development and evolution. Lastly, we use discrete differences in tooth and jaw anatomy to define types of squamate implantation and distinguish them from other implantation modes.

2. MATERIALS AND METHODS

For the present study, most of the sections were stained using Masson's trichrome (Suppl. Info), which was helpful in differentiating epithelial tissues (light pink) from the collagen‐rich connective and attachment tissues (blue). Other sections were stained using a standard hematoxylin and eosin stain (H&E) (Suppl. Info). Previously made sections were also re‐examined. These were already stained for a Master's thesis project (Budney, 2004) with H&E, Gomori's trichrome, or Masson's trichrome, or were stained using Masson's trichrome for a previous study of squamate tooth attachment tissues (Maxwell et al., 2011c) (see Table 1 for details).

TABLE 1.

Sections and specimens personally examined for this study. Specimens donated by L. W. Kline are denoted with “K”. Institutional abbreviations: AMNH, American Museum of Natural History, New York, USA; CMNAR, Canadian Museum of Nature; FMNH, Field Museum of Natural History, Chicago, USA; ROM, Royal Ontario Museum, Toronto, Canada; UAMZ, University of Alberta, Edmonton, Canada

Taxon	Material (Cat. No.)	Method of study	Plane of section	Source
Iguana iguana	Partial dentary (“K1”)	Decal sectioning, Masson's trichrome stain	Transverse	This study
Iguana iguana	Partial dentary (“K2”)	Decal sectioning, Masson's trichrome stain	Transverse	This study
Iguana iguana	Partial dentary (“K3”)	Decal sectioning, Masson's trichrome stain	Coronal	This study
Iguana iguana	Partial dentary (“K4”)	Decal sectioning, Masson's trichrome stain	Transverse	This study
Iguana iguana	Partial dentary (UAMZ R951)	Decal sectioning, Masson's trichrome stain	Transverse	This study
Iguana iguana	Dentary (ROM R 7716)	Images	N/A	This study
Amblyrhynchus cristatus	Isolated teeth (AMNH 114492)	Images	N/A	This study
Amblyrhynchus cristatus	Isolated tooth (AMNH 29937)	Images	N/A	This study
Amblyrhynchus cristatus	Skull (AMNH 76197)	Images	N/A	This study
Cordylus cordylus	Partial dentary (CMNAR 15418)	Decal sectioning, H&E stain	Coronal	Budney, 2004 (MSc thesis)
Cordylus cordylus	Skeletonized skull and mandible (ROM R40)	Images	N/A	This study
Elgaria (Gerrhonotus) principis	Partial dentary (UAMZ unnumbered)	Decal sectioning, Masson's trichrome staining	Coronal	Budney, 2004 (MSc thesis)
Lacerta vivipara	Partial dentary (CMNAR 4372)	Decal sectioning, Masson's trichrome staining	Coronal	Budney, 2004 (MSc thesis)
Sauromalus atus	Partial dentary (CMNAR 25719)	Decal sectioning, Masson's trichrome staining	Coronal	Budney, 2004 (MSc thesis)
Scincus scincus	Partial dentary (CMNAR 29529‐2)	Decal sectioning, Gomori's trichrome staining	Coronal	Budney, 2004 (MSc thesis)
Varanus niloticus	Partial dentary (CMNAR 13884)	Decal sectioning, Masson's trichrome staining	Coronal	Budney, 2004 (MSc thesis)
Varanus sp.	Partial dentary (UAMZ unnumbered)	Decal sectioning, H&E staining	Coronal + Transverse	This study
Varanus rudicollis	Partial dentary (UALVP 53485)	Decal sectioning, Masson's trichrome staining	Transverse	Maxwell et al. (2011)
Dracaena guianensis	Complete skeleton (ROM R0377)	Images	N/A	This study
Dracaena guianensis	Skull and mandibles (FMNH 207657)	Images +μCT scan	N/A	This study
Tupinambis teguixin	Complete skeleton (ROM R8380)	Images	N/A	This study
Caiman sclerops	Partial dentary (CMNAR 25747‐4)	Decal sectioning, H&E staining	Coronal	Budney, 2004 (MSc thesis)

For the sections made specifically for this study, lower jaws were dissected out of formalin‐fixed specimens using scalpels and a Buehler Isomet 1000 wafer‐blade saw. These specimens were then decalcified using Cal‐Ex™ (Fischer Scientific™, 5.5% HCl, 0.12% EDTA) for up to 28 days, Cal‐Rite™ decalcifying/fixation solution for 48–72 hours (Richard‐Allan Scientific™, 10% formic acid, 3–4% formaldehyde), or Shandon™ TBD‐1™ decalcifier for 4–8 hours (Thermo Scientific™, 35% HCl). For specimens left in decalcifier solution for longer than 24 hours, the solution was replaced with fresh decalcifier every 24 hours. The dentaries were left in the decalcifier solutions until they could be manually bent, indicating that the bone of the jaw had been sufficiently decalcified. Specimens were then placed into a tissue processor overnight for clearing and paraffin infiltration. These were then embedded in paraffin wax and sectioned to between 5 and 8 μm thickness using a rotary microtome (Leica 2025). Sections were then mounted to charged glass slides (Fisherbrand™, Superfrost™ Plus). Stained sections were imaged using a Nikon DS‐Fi3 microscope camera mounted to a Nikon E6000 POL cross‐polarizing microscope, and Nikon NIS Elements microscopy imaging software.

μCT scans of a skeletonized dentary of the tupinambine teiid Dracaena guianensis were done using a Bruker‐SkyScan 1076 μCT scanner at the Pharmaceutical Orthopaedic Research Lab (University of Alberta). Scans were completed at 18 μm resolution, with an x‐ray tube voltage of 100 kV and a current of 100 μA, a 1.0 mm aluminum filter, and an exposure time of 1770 ms. Three scan projections were averaged at each step through the 180° of rotation, with 0.5° rotation between each step. Raw image projections were then reconstructed using NRecon software (ver. 1.7.0.4). Virtual slices through the dentary were generated using Dragonfly [ver. 2.0 for Windows, Object Research Systems (ORS)].

All of the specimens sectioned for this study were either from preserved museum collections, or were donated to the authors by L. W. Kline, who had used Iguana iguana specimens for previous studies of tooth replacement (Kline & Cullum, 1984).

3. RESULTS

3.1. Iguana‐type implantation

3.1.1. General morphology

Iguana‐type implantation occurs in several squamate lineages. We use the term “Iguana‐type” only because our largest and most thorough histological sample comes from Iguana iguana, but we found this implantation mode in other iguanids, a cordylid, scincomorph, and a lacertid (Figures 1, 2, 3; Suppl. Figure S1). In this type, tooth implantation and attachment are strongly asymmetric and this creates dramatically different perceptions of the attachment tissues when the tooth is reduced to a two‐dimensional slide. Because of this, we provide a description of our histological samples in two planes of section: coronal and transverse. The coronal sections are best suited for documenting the interaction between the jaw and an individual tooth, whereas the transverse sections reveal details about the tooth‐to‐tooth attachment and important structures conventionally referred to as interdental ridges.

FIGURE 2.

Coronal (standard) sections of an Iguana iguana dentary. (a) coronal section through the middle of a fully developed tooth (“K3”). (b) close‐up image of the epithelial tissue (the combined HERS and dental lamina) contacting the lingual surface of the tooth in (a). (c) close‐up image of the labial side of the tooth in (a) showing the lack of epithelial‐tooth contact and the development of periodontal tissues. (d) coronal section in the interdental space between two dentary teeth (“K3”). (e) close‐up of the connective and epithelial tissues in the interdental space in (d). Note that the dental lamina forms a wall of epithelial tissue lingual to the jaw. Arrowheads follow the directions of the principle fiber groups. (f) close‐up of fibroblasts within the Interdental Connective Tissue (ICT) in (e). Arrowheads follow the direction of the principle fiber group (g) close‐up of osteoblasts within the ICT. Insets are images of a skeletonized dentary of I. iguana (ROM R 7716) showing the planes of section. Abbreviations: ac, acellular cementum; cc, cellular cementum; ct, connective tissue; de, dentine; dl, dental lamina; ep, epithelial tissue; fb, fibroblasts; ICT, Interdental Connective Tissue; Mc, Meckel's cartilage; odb, odontoblasts; osb, osteoblasts. Asterisks indicate tearing and splitting of the epithelium in the sections

FIGURE 3.

The periodontal tissues in the cordylid lizard Cordylus cordylus (CMNAR 15418) in coronal section. (a) overview image of an ankylosed dentary tooth, taken near the midline of the functional tooth. (b) close‐up of the lingual surface of the tooth in (a) showing a very thin layer of cementum and a Hertwig's Epithelial Root Sheath (HERS) that nearly contacts the lingual surface of the tooth. (c) close‐up of the epithelium that apposes the lingual surface of a functional tooth, which is divided into a dental lamina and HERS. (d) overview image of an ankylosed dentary tooth, taken slightly off of the midline of the tooth in (a). (e) close‐up of the lingual surface of the tooth in (d) showing thick layers of cementum and collagen fiber bundles of a periodontal ligament along the tooth surface. Note the considerable distance between HERS and the dentine surface of the tooth root. (f) same image as (e) under cross‐polarized light showing embedded Sharpey's fibers in the cellular cementum. Insets are images of a skeletonized dentary of C. cordylus (ROM R 40) showing the planes of section. Abbreviations: ac, acellular cementum; cc, cellular cementum; cfb, collagen fiber bundles; de, dental lamina; dl, dental lamina; eg, enamel gap; ft, functional tooth; HERS, Hertwig's Epithelial Root Sheath; lw, lingual wall of the jaw; oe, oral epithelium; pl, pleura; sf, Sharpey's fibers. Asterisks indicate tearing and separation of the epithelium from the dentine in the sections

3.1.2. Iguana‐type implantation in coronal histological sections (the standard view)

Nearly all published sections of pleurodont implanted teeth are in a coronal view at a single tooth position (e.g., Gaengler & Metzler, 1992; Peyer, 1968; Zaher & Rieppel, 1999), with very few exceptions (Fuenzalida et al., 1999; Lessmann, 1952; Luan et al., 2006, 2009). Our sections in this plane yielded very similar results to those of previous researchers: the dentine of each tooth is asymmetrically distributed, such that the labial surface of the tooth is shorter and fuses to the crest of the pleura (the lingual side of the labial wall of the jaw) (Berkovitz & Shellis, 2017; Gaengler, 2000). The lingual side of the tooth extends farther towards the base of the alveolar margin, anchoring to a lingual extension of the jawbone, or to a slightly raised lingual wall of the jaw (Figures 1, 2, 3). The dental pulp mostly consists of loose connective tissue, several blood vessels, and nerves. The dentine‐producing odontoblasts line the inner margins of the dentine and send long thin cell processes through it, describing the dentine as a typical, tubular orthodentine (Figure 2a,b).

In this plane of section, there are virtually no attachment tissues along the lingual surfaces of the teeth. This is because the dental lamina and an additional epithelial layer form the lingual wall to which the dentine and enamel of each tooth is directly apposed (Figure 2a,b; Figure 3c). This second invagination is typically called Hertwig's Epithelial Root Sheath (HERS) (Luan et al., 2006; McIntosh et al., 2002) when it coats the tooth root and is readily distinguishable from the epithelial cells of the dental lamina. However, this was only the case in our coronal sections of some lizards (e.g., Cordylus, Figure 3c; Suppl. Figure S3). In our Iguana sections, the dental lamina and HERS are connected together as a single band of epithelium that extends all the way from the oral epithelium to the base of the tooth (Figure 2a,b). Though the epithelium surrounding the lingual side of the tooth appears to be separated from the dentine of the tooth (Figure 2a) creating the impression of a large gap for the enamel (which has been demineralized prior to sectioning), the enamel would only extend a short distance down the tooth (Suppl. Figure S4). Furthermore, in some regions of these sections, the epithelium is still attached to the lingual surfaces of the teeth (Figure 2b; Suppl. Figure S4). This suggests that the epithelium covering the lingual surfaces of these teeth would normally adhere to the tooth, and the gaps in our sections formed from tearing of the epithelium.

The dental lamina separates from HERS at the very base of the tooth in coronal sections. At higher magnifications, the epithelial cells of the combined HERS and dental lamina are slightly different: the more lingual cells – those of the dental lamina –stain slightly darker with Masson's trichrome staining in our Iguana sample (Figure 2a,b; Suppl. Figure S4). HERS cells contact the dentine surface of the root for nearly its entire length, as they do in several other tetrapods (Luan et al., 2006; McIntosh et al., 2002).

All of the attachment of an Iguana‐type tooth in coronal section appears to be between the base of the labial wall of the tooth to the crest of the pleura. In this region, HERS does not contact the surface of the tooth, leaving room for mineralized attachment tissues derived from the underlying connective tissue (Figure 2c). These mineralized attachment tissues appear to be deposited centrifugally off of the dentine surface and have many cell inclusions. We refer to this tissue as cellular cementum, given its topological and developmental similarity to the cellular cementum in crocodilians, mammals, and previous reports of cementum in Iguana (Berkovitz & Sloan, 1979; Diekwisch, 2001; LeBlanc et al., 2017a; Luan et al., 2006, 2009; McIntosh et al., 2002).

Replacement teeth develop from the tip of the dental lamina at the bases of the functional teeth (Figure 3a,d), but the rest of the epithelial tissue (HERS and dental lamina together) dictates the arrangements of the attachment tissues on the lingual sides of pleurodont teeth. In coronal sections through the midlines of these teeth, the epithelium contacts the dentine of the tooth (Figure 2A, B; Figure 3a‐c), but in more mesial or distal sections the epithelium separates from the dentine of the tooth, leaving an intervening layer of fibrous connective tissue. In these regions, the lingual surfaces of the teeth are coated in a thin band of acellular tissue, which is covered by a much thicker, cellular layer that we interpret as acellular and cellular cementum respectively (LeBlanc et al., 2017b, 2018; LeBlanc & Reisz, 2013; Maxwell et al., 2011a) (Figure 3d–f). The cellular cementum also contains partially mineralized collagen fiber bundles (Sharpey's fibers) embedded at right angles to the long axis of the tooth (Figure 3e,f). These Sharpey's fibers extend outwards from the cementum into a band of collagen fiber bundles within the overlying connective tissue, between the tooth and the dental lamina. These fibers seemingly disappear in coronal sections into the fibrous connective tissue, but their insertion points are clear in transverse sections (see below).

We also investigated the arrangements of tissues at the positions of the interdental ridges in coronal section. In these regions, the dental lamina is still present, forming a wall of epithelium lingual to the tooth row (Figure 2d,e). However, unlike in the coronal sections taken at the positions of the teeth, the interdental regions have large pockets filled with collagen‐rich connective tissue between the jaw and the dental lamina. For the remainder of this study we refer to these tissue masses in between tooth positions as the Interdental Connective Tissue (ICT). The ICT primarily consists of collagen fiber bundles extending from the bone to the dental lamina and, in areas closer to the dental lamina, in dorsoventral directions (Figure 2e). The ICT also contains several elongate cell bodies associated with the collagen fiber bundles, which we interpret as fibroblasts (Figure 2e,f). Closer to the bone tissue, the cells are larger, round, and form the newest layers of bone on the lingual surface of the pleura (Figure 2e,g). We interpret these cells as osteoblasts that are frequently trapped in their mineralizing osteoid matrix, forming osteocytes.

3.1.3. Iguana‐type implantation in transverse histological sections: crest of the jaw

Transverse sections through juvenile and adult I. iguana dentaries reveal a much more complex arrangement of dental tissues, which is completely missed in the standard coronal view. In order to properly document these tissues, we describe their development from successive stages of tooth development in serial transverse sections through the jaw.

At the early stages of dental development, a developing, unerupted tooth has not yet reached its functional position (Figure 4a). Osteoclasts are still actively resorbing the attachment tissues from the previous tooth generation (Figure 4b). The osteoclasts leave telltale scalloping (Howship's lacunae) along the hard tissues, which show the farthest extent of osteoclast activity at any given stage. This extensive resorption indicates that the attachment tissues that develop at the subsequent developmental stages are derived from the dental follicle of the new tooth and are not recycled between tooth generations.

FIGURE 4.

The resorptive stage of tooth development in Iguana iguana (“K1”) in transverse section. (a) an unerupted tooth approaching the pleura of the jaw, which is preceded by a population of bone and dentine‐resorbing osteoclasts that make room for the developing tooth and its future attachment tissues. (b) close‐up of the surface of the pleura in (a) showing osteoclasts and Howship's lacunae. Abbreviations: am, ameloblasts; eg, enamel gap; Hl, Howship's lacunae; jb, bone of the jaw; oscl, osteoclasts; oe, outer enamel epithelium; rp, resorption pit; rt, replacement tooth; sr, stellate reticulum

Once resorption is complete and nearly all of the old dental tissues have been removed, the developing tooth begins to attach to the jaw. At the crest of the pleura, each erupting tooth is suspended within collagen‐rich connective tissue (Figure 5a,b). The lingual, mesial and distal surfaces of the developing tooth are covered by epithelial cells, which impede the development of any attachment tissue. As was the case in the coronal sections, the epithelial covering usually tore away from the dentine surface in our transverse sections, giving the appearance of a gap around a developing tooth crown. Normally, this gap would be interpreted as the former position of the fully demineralized enamel (Nanci, 2013), however, this was only the case in Figure 4a, where ameloblasts clearly fringed the gap, indicating the formation of enamel. In all of the other sections, the epithelial covering is almost entirely made from HERS (there are no ameloblasts lining the “gaps” in the sections and pockets of epithelial cells are still directly attached to the dentine in some areas [Suppl. Figure S4]).

FIGURE 5.

Transverse sections of developing dentary teeth in Iguana iguana (“K1”, “K2”) taken near the crests of the jaws. (a) an erupting tooth with a developing periodontium. (b) close‐up of periodontium along one side of the tooth in (a) showing initial alveolar bone and periodontal ligament formation. Arrowheads highlight the reversal line marking the onset of alveolar bone deposition. (c) a fully erupted, but not yet ankylosed tooth, showing ligamentous tooth attachment. (d) close‐up of periodontium of tooth in (c) showing cementum, periodontal ligament, and alveolar bone, but no ankylosis. Arrowheads highlight the reversal line marking the onset of alveolar bone deposition. (e) a nearly completely ankylosed tooth with a nearly completely mineralized periodontium. (f) close‐up of the remaining unmineralized portion of the periodontal ligament around the tooth in (e). Arrowheads highlight the reversal line marking the onset of alveolar bone deposition. Insets are images of a skeletonized dentary of I. iguana (ROM R 7716) showing the planes of section. Abbreviations: ab, alveolar bone; ab, alveolar bone; ac, acellular cementum; cb, cementoblasts; cc, cellular cementum; cfb, collagen fiber bundles; de, dentine; HERS, Hertwig's Epithelial Root Sheath; ICT, Interdental Connective Tissue; jb, jawbone; oab, alveolar bone from an old tooth; ode, dentine from an old tooth; osb, osteoblasts; PDL, periodontal ligament; rl, reversal line. Asterisks indicate tearing and separation of the epithelium from the dentine in the sections

The HERS covering the tooth root is also directly connected to the dental lamina. We observed this combined HERS and dental lamina along the lingual surfaces of the teeth in coronal sections as well (Figure 2b). This differs from the condition in some amphibians, where the dental lamina and the HERS are separate structures, with the HERS exclusively covering the lingual surface of the tooth (Luan et al., 2006). As a result, the epithelium covering the lingual poles of the tooth roots is unusually thick in I. iguana, which has been noted in this species before, as well as other squamates (Luan et al., 2006; McIntosh et al., 2002).

Despite the absence of attachment tissues along the lingual, mesial, and distal surfaces of the teeth in transverse section, at these early stages of tooth eruption, the periodontium in I. iguana is already differentiated into two discrete tissues. Each erupting tooth forms a thin layer of bone along a partially resorbed labio‐mesial or labio‐distal surface of the neighboring teeth (Figure 5a,b). This bone tissue is formed by osteoblasts associated with the new tooth, which deposit bone along this resorptive surface, forming a reversal line in thin section (Figure 5b). This bone also contains the mineralized ends of thick collagen fiber bundles extending into the ICT. We refer to this bone as the alveolar bone, given that it (1) defines the outer boundaries of the developing periodontium to which the tooth is attached via collagen fiber bundles; and (2) mineralizes centripetally towards the developing tooth, unlike the cellular cementum (LeBlanc et al., 2018; Ten Cate, 1997; Ten Cate & Mills, 1972).

As the tooth continues to develop, it begins to form the first periodontal tissue attachment to the crest of the pleura (Figure 5c,d). At this stage, three periodontal tissues are visible: (1) a thin layer of alveolar bone forming along the pleura and neighboring teeth, separated from the other hard tissues by a reversal line; (2) a collagen fiber‐ and fibroblast‐rich periodontal ligament attaching the tooth to the alveolar bone; and (3) a centrifugal mineralization of cementum forming along the surface of the tooth root, which envelops fibers of the periodontal ligament, as well as several cementoblasts (Figure 5d). This stage of periodontal tissue formation is virtually identical to that of newly erupted teeth in crocodilians (LeBlanc et al., 2017a), but unlike the thecodont teeth of crocodilians and mammals, the developing pleurodont teeth only show this attachment to the crest of the pleura and to neighboring teeth.

The alveolar bone and cementum of a fully erupted tooth gradually fuse it to the pleura (Figure 5e,f). However, the periodontal ligament is still present throughout this process. Eventually, it is completely mineralized once the mineralization fronts of the alveolar bone and cementum meet. In teeth that have not yet fused to the pleura, the periodontal ligament is still visible, forming organized networks of collagen fiber bundles that anchor the tooth to the jaw (Figure 5f).

3.1.4. Iguana‐type implantation in transverse histological sections: middle of the tooth

The extreme asymmetry in the development of Iguana‐type teeth means that the tooth attachment architecture is very different between transverse sections taken at the crest of the jaw compared to deeper sections. Below the crest of the jaw, for example, each tooth root is U‐shaped in transverse section, with an opening directed toward the labial wall of the dentary (pleura) (Figure 6). This opening receives vasculature from within the bone of the jaw (via branches from the internal mandibular artery [cf. Oelrich, 1956]). In thecodont teeth, these openings are located at the apex of the root. In pleurodont teeth, the pulp is open along most of the labial side of the tooth, limiting its attachment just to the apex of the pleura, with each tooth primarily anchored to neighboring teeth by a thick periodontium (Figures 1b, 6). A similar tooth‐to‐tooth attachment has been described in pleurodont amphibians (Luan et al., 2006).

FIGURE 6.

Transverse sections of dentary teeth in Iguana iguana (“K1”, “K2”) taken below the crests of the jaws. (a) an erupting tooth with a fibrous attachment system. (b) close‐up of the developing periodontium along the tooth in (a) showing the fibroblasts of a developing periodontal ligament and osteoblasts of the alveolar bone. Arrowheads highlight the reversal line marking the onset of alveolar bone deposition. (c) tooth attachment tissue mineralization in a fully erupted tooth. Arrowheads highlight the reversal line marking the onset of alveolar bone deposition. (d) close‐up of mineralizing attachment tissues in (d) under cross‐polarized light, showing Sharpey's fibers of the mineralizing periodontal ligament within the cementum and alveolar bone. (e) a fully ankylosed tooth showing extensive attachment tissues in between teeth. (f) close‐up of mineralized attachment tissues in (e). Insets are images of a skeletonized dentary of I. iguana (ROM R 7716) showing the planes of section. Abbreviations: ab, alveolar bone; ac, acellular cementum; cc, cellular cementum; ce, cementocyte; de, dentine; fb, fibroblasts; gd, globular dentine; odb, odontoblasts; osb, osteoblasts; osc, osteocytes; PDL, periodontal ligament; sf, Sharpey's fibers. Asterisks indicate tearing and separation of the epithelium from the dentine in the sections

At the earliest stages of periodontal tissue formation, each tooth in I. iguana is attached mesiodistally to neighboring tooth positions through ICT‐derived attachment tissues. The teeth are not yet fused in place but show two discrete cell populations and tissues in the developing periodontium. The first cell type is concentrated around the mesial and distal edges of the developing tooth root. These cells are elongate and are associated with long thin collagen fibers (Figure 6a,b). These appear to be the fibroblasts of a developing periodontal ligament. As in the higher coronal sections, the second cell population is a cluster of rounded cells that form bone along a resorbed surface of the neighboring teeth (Figure 6b). These cells are osteoblasts that form the first layers of alveolar bone to which the fibers of the periodontal ligament are attached. As this tissue continues to mineralize centripetally, it entombs several of the osteoblasts, which in turn form osteocytes.

At later stages of dental ontogeny, osteoblasts along the edges of the developing alveolar bone continue to deposit bone matrix, gradually enclosing the periodontal ligament in mineralized tissue (Figure 6C, D). Meanwhile, acellular and cellular cementum develop from the surfaces of the tooth roots, growing into thick layers that entomb the other ends of the ligament (Figure 6d). We interpret these mineralized tissues as acellular and cellular cementum based on their topology and growth direction (LeBlanc et al., 2016a, 2017b, 2017a). The acellular cementum is adjacent to the outermost layer of dentine, which is more globular in appearance in deeper coronal sections. The acellular cementum is a thin band of homogenous tissue that incorporates large fiber bundles from the periodontal ligament during mineralization (Figure 6f). The more external layers of cementum contain cementocytes, as well as long strands of Sharpey's fibers from the periodontal ligament (Figure 6d,f). Each tooth becomes ankylosed to the adjacent teeth once the cementum and alveolar bone completely mineralize the intervening ligament and surround the associated vasculature (Figure 6e). At this stage, each tooth is seemingly attached to its neighbor via a single mass of mineralized tissue, previously referred to as “bone of attachment” (Gaengler & Metzler, 1992; Luan et al., 2009; Peyer, 1968; Tomes, 1882). However, this tissue cannot be a single bone mass, given that it has multiple mineralization points and directions. Furthermore, the acellular cementum and the Sharpey's fibers of the periodontal ligament within the mineralized tissues are still visible even in fully ankylosed teeth (Figure 5f; 6d,f).

3.1.5. Interdental ridges: histology and development

Several researchers have noted the presence of a stand‐alone layer of bone between teeth in squamates, which is often referred to as an interdental ridge (Kley, 2006; Luan et al., 2009; Rieppel & Zaher, 2001; Scanlon & Lee, 2002; Zaher & Rieppel, 1999). Our histological sample of juvenile and adult I. iguana allowed us to determine the origins of this structure in a pleurodont squamate.

In juvenile I. iguana, the structure in between teeth is composed entirely of the periodontal tissues (alveolar bone, periodontal ligament, and cementum) from a single tooth (Figure 7a–c). These periodontal tissues always belong to the youngest tooth at any given point along the jaw. This is because, as each new tooth forms, it first causes resorption of the lingual surface of its functional predecessor. This lingual resorption continues until the functional tooth root is resorbed and the tooth crown is shed. Resorption continues until nearly all of the dentine and attachment tissues of the previous tooth are resorbed (Figure 4). However, in some cases, small fragments of the attachment tissues – and even dentine – may remain. These remnants are always located in the mesial and distal corners of the resorption pit, where osteoclasts may never reach (Figure 5e–f). These remnants are extremely small in juvenile specimens and are completely buried by the development of the attachment tissues of the newly erupted tooth.

FIGURE 7.

Development of the interdental ridges in Iguana iguana. (a) transverse section through a dentary of a small Iguana iguana (“K2”). Mesial is to the right, lingual is toward the top. (b) close‐up of the interdental tissues between two teeth in (a). The mineralized tissue in between the teeth is composed of alveolar bone and cementum. Black arrowheads highlight the reversal line marking the onset of alveolar bone deposition. (c) same image as (b) under cross‐polarized light. (d) transverse section through a dentary of a large Iguana sp. (UAMZ R951). (e) close‐up of an interdental ridge between two teeth in (d). (f) same image as (e) under cross‐polarized light showing fragments of dentine from pervious tooth generations embedded within the interdental ridge. (g) tooth migration as indicated by the fragments of dentine preserved within the interdental ridges (white arrows track tooth migration from each fragment of old teeth embedded within the interdental ridges). Abbreviations: ab, alveolar bone; cc, cellular cementum; ode, dentine from an old tooth; idr, interdental ridge

Over repeated bouts of resorption and re‐deposition of new dental tissues, these remnants of old dentine, old alveolar bone, and old cementum accumulate. In larger I. iguana, these accumulations are so extensive that they form wedge‐shaped structures in between successive teeth, which physically separate neighboring tooth positions (Figure 7d–f). These structures are visible in skeletonized specimens and are often called interdental ridges (e.g., Luan et al., 2009; Zaher & Rieppel, 1999). Close inspection shows that these are not distinct dental tissues, but the remnants from previous tooth generations at each locus. These remnants include large fragments of dentine and attachment tissues (alveolar bone and cementum) that accumulated during the growth of the animal and can be used to track how the positions of successive tooth generations have migrated as the jaws grew. In one specimen, we were able to use the dentine fragments within the interdental ridge to determine that some of the teeth had slightly migrated posteriorly over successive replacement cycles (Figure 7g). Kline and Cullum (1984) observed the same posterior migration of the tooth positions in their ontogenetic study of I. iguana, but by non‐histological means. The histology and development of the interdental ridges in pleurodont squamates is identical to the interdental ridges, plates, and bone in other amniotes (LeBlanc & Reisz, 2013; LeBlanc et al., 2017b; Haridy et al., 2018; He et al., 2018).

3.2. Histology of Varanus‐type implantation

Varanid tooth replacement and attachment have been studied extensively (Maxwell et al., 2011c; Rieppel, 1978; Zaher & Rieppel, 1999). As such, we focus on the three‐dimensional arrangements of the hard and soft tissues surrounding the teeth of Varanus sp. and compare them to those of Iguana and other pleurodont squamates. Unlike in many other squamates, Varanus, Heloderma, and Lanthanotus teeth have fluted roots that form a honeycomb‐like structure in transverse sections (Figures 8, 9) (Rieppel, 1978). These structures are made of plicidentine: centripetal infoldings of the developing root that form prior to dentinogenesis (Kearney et al., 2006; Maxwell et al., 2011b, 2011c). We refer to these types of implantation as “Varanus‐type”, because the presence of plicidentine has a dramatic effect on tooth implantation and attachment, which is typified by the dentition of Varanus and its closest relatives.

FIGURE 8.

Coronal sections through dentary teeth of Varanus. (a) section through a dentary tooth of Varanus sp. (b) close‐up of the labial surface of the root in (a) showing the development of cementum. (c) coronal section of a dentary tooth of Varanus niloticus (CMNAR 13884). (d) close‐up image of a root of V. niloticus (CMNAR 13884) showing plicidentine and accumulations of attachment tissues. Abbreviations: ab, alveolar bone; ac, acellular cementum; cb, cementoblasts; cc, cellular cementum; de, dentine; del, dentine lamella (plicidentine fold); oat, old attachment tissues; pl, pleura; rl, reversal line. Asterisks indicate tearing and separation of the epithelium from the dentine in the sections

FIGURE 9.

Transverse sections of dentary teeth in Varanus rudicollis (UALVP 53485). (a) transverse section taken near the base of the teeth. Note the convoluted cross‐sections of the teeth, caused by plicidentine. (b) close‐up of the unmineralized interdental space between two teeth in (a) showing the dental lamina and the connective tissue. (c) transverse section taken closer to the root. (d) close‐up of plicidentine and thin attachment tissues from a tooth in (c). Abbreviations: ab, alveolar bone; cc, cellular cementum; del, dentine lamella (plicidentine fold); dl, dental lamina; HERS, Hertwig's Epithelial Root Sheath; ICT, Interdental Connective Tissue; jb, jawbone; PDL, periodontal ligament; rl, reversal line. Asterisks indicate tearing and separation of the epithelium from the dentine in the sections

Varanus‐type teeth are very shallowly implanted in jaws that have no subdental shelf or lingual wall and have a tall pleura (Figure 8). As in I. iguana and other squamates, the dental lamina and HERS in Varanus adhere to the lingual surfaces of the teeth and form a continuous lingual wall of epithelial tissue in transverse sections (Figure 9). As such, only very thin layers of attachment tissues form along the base, labial, and lingual surfaces of the teeth. Close inspection of these regions of the teeth revealed thin layers of cellular bone‐like tissue that develops centrifugally from the dentine of the tooth (Figure 8b). This tissue is consistent with cellular cementum (Maxwell et al., 2011c). Varanid teeth differ from those of other pleurodont squamates in the primary sites of attachment between the roots and the jaws. In Iguana‐type attachment, the teeth have a large labial opening and thus can only attach to the crest of the pleura and to neighboring teeth. In Varanus, there is a labial wall of plicidentine for each tooth and the teeth are anchored solely to the pleura of the jaws (Figures 8, 9).

In both planes of section, Varanus teeth have extensive infoldings of dentine that partition the pulp into a series of small chambers. At first glance, the dentine lamellae along the labial portions of the root appear to directly contact the bone of the jaw, but higher magnifications reveal thin layers of mineralized, cellular tissue that mediate the contact between dentine and jawbone (Figures 8d, 9d). This cellular layer is usually very thin. In one tooth section, the root measures nearly 900 μm in labiolingual diameter, whereas the attachment tissue fastening the tooth to the jaws is only 45 μm thick in the same dimension. Although it is thin, each dentine lamella is associated with a small amount of attachment tissue that forms an extensive surface area for tooth attachment along the base and labial surface of each tooth. This attachment tissue has two origins: a small amount of alveolar bone that forms along the surface of the jawbone, and a small layer of cellular cementum that develops along the surface of the tooth (Maxwell et al., 2011c). These layers form around each developing tooth, and remnants can be left behind from successive tooth replacement events, as in other squamates (Figure 8d).

The ICT in Varanus consists of a large fibrous mass of tissue in between neighboring teeth (Figure 9b). The ICT is bounded lingually by a continuous dental lamina (Figure 9b). Unlike in Iguana, there are no prominent layers of alveolar bone or accumulations of interdental bone between teeth. Instead, an unmineralized ICT forms the only intervening tissue. The collagen fibers of the ICT are fairly loose and disorganized, but in the labial‐most regions, the collagen fibers form a bridging mass that connects neighboring teeth (Figure 9b). The collagen fibers here are more organized, all of which extend mesiodistally across the interdental space. In Iguana, these fibrous masses are usually completely mineralized within the cementum and alveolar bone matrices of the neighboring teeth, but in Varanus, they remain unmineralized either as periodontal or transdental ligament fibers that connect neighboring teeth (Nanci, 2013).

3.3. Dracaena‐type implantation

Dracaena and Tupinambis display an unusual form of tooth attachment and implantation compared to the other two modes mentioned above (Figure 10). These teiids have relatively bulbous teeth as adults, although only Dracaena is thought to consistently eat hard‐shelled prey (Dalrymple, 1979; Dessem, 1985). The jaw geometry at the middle and distal ends of the tooth rows is superficially similar to that of pleurodont squamates, but µCT scans of a Dracaena dentary reveal dramatic differences in tooth implantation along the jaw (Figure 10a–e). At the mesial end of the tooth row, each tooth is held in separate U‐shaped troughs along the jaws, and the surrounding attachment tissues are symmetrical on the labial and lingual sides (Figure 10a,b). This is akin to thecodonty in other extant mammals and crocodiles (Bertin et al., 2018). Virtual sections through more distal tooth positions show that the lingual and labial walls of the jaws become progressively shorter and the teeth are more shallowly implanted (Figure 10c,d). However, at no point do the teeth of Dracaena approach Iguana‐ or Varanus‐type attachment and implantation.

FIGURE 10.

Tooth implantation and attachment in two tupinambine teiids. (a) virtual coronal section through an anterior dentary tooth of Dracaena guianensis (FMNH 207657). Asterisk indicates the opening of the pulp chamber. (b) virtual coronal section through a slightly more posterior tooth. (c) virtual coronal section through a middle tooth. Asterisk indicates the opening of the pulp chamber. (d) virtual coronal section through a posterior tooth. (e) lingual view of a left dentary of Dracaena guianensis (ROM R0377) showing relative positions of virtual sections taken from FMNH 207657 in (a–d). (f) close‐up image of the 9th tooth in (e) showing extensive cementum deposition prior to ankylosis. Note the presence of an apical wear facets, indicating that this tooth was functional and suspended by soft tissues. (g) close‐up image of the 11th tooth in (e) showing complete ankylosis, once the cementum fuses the tooth in place. (h) virtual transverse section of a D. guianensis dentary (FMNH 207657) showing the circular dentine roots and extensive cementum surrounding each tooth. (i) lingual view of a Tupinambis teguixin dentary (ROM R8380) showing ankylosed teeth and one unfused tooth with cementum coating the tooth root. (j) lingual view of the anterior teeth in the same dentary in (i) showing ankylosed teeth and interdental ridges. Abbreviations: c, cementum; de, dentine; dl, dental lamina; idr, interdental ridge; jb, jawbone; lw, lingual wall of the jaw; pl, pleura of the jaw; rp, resorption pit; rt, replacement tooth; wf, wear facet

Skeletonized jaws of Dracaena and Tupinambis can preserve teeth at multiple developmental stages. In both genera, the replacement teeth develop within a large resorption pit (Figure 10h–j). The teeth are then suspended by soft tissues (probably a periodontal ligament) for some time prior to complete ankylosis. In several skeletonized specimens, the teeth seemingly float within a discrete alveolus and the teeth are coated in thick layers of cementum, which mineralizes centrifugally from the surface of the tooth root (Figure 10f,g,i,j). These teeth are fully erupted and clearly functional in our Dracaena sample, given the presence of wear facets on the unfused teeth (Figure 10f). The teeth eventually fuse to the jaw via continued mineralization of the cementum of the tooth. In the end, the cementum in these teiids is the dominant mineralized attachment tissue, forming along all sides of the tooth roots (Figure 10h). The cementum is thick and well‐vascularized, similar to the osteocementum in mosasauroid squamates and ichthyosaurs (Caldwell, 2007; Caldwell et al., 2003; LeBlanc et al., 2017a; Maxwell et al., 2011a).

This form of tooth implantation differs from the attachment and implantation of other squamates in several key ways (Table 2). First, tooth attachment in Dracaena and Tupinambis is far more symmetrical along the labial and lingual surfaces of the tooth root, which suggests that the dental lamina and the HERS do not directly adhere to the lingual surfaces of the tooth roots. Second, the surrounding jaw geometry is more symmetrical compared to that of pleurodont jaws. The labial and lingual walls are almost equal in height along several portions of the jaw, although occasionally the pleura is slightly higher (Figure 10a–d). Third, each tooth is coated with extensive layers of vascularized cementum that form the dominant mineralized component of the periodontium. In pleurodont squamates, these tissues are either extremely reduced (Figure 8) or are roughly equally composed of cementum and alveolar bone (Figure 6). Lastly, each tooth in Dracaena and Tupinambis has an apical opening at the bottom of the root for the entry of blood vessels and nerves into the pulp. The dentine of these teeth is circular in all transverse sections along the teeth, indicating that the labial and lingual walls of the developing tooth roots form symmetrically, unlike the condition in pleurodont squamates. In all of these respects, Dracaena‐type tooth implantation is more similar to the thecodont condition in the extinct mosasauroid squamates (Caldwell, 2007; Caldwell et al., 2003; LeBlanc et al., 2017a).

TABLE 2.

Variations in squamate tooth implantation, based on the study specimens

Implantation mode	Attachment mode in fully developed teeth	Dental lamina‐tooth contact	Attachment to pleura?	Main site of tooth attachment	Tooth root cross‐section	Extent of mineralized attachment tissues
Iguana‐type	Ankylosis	Yes‐ attached to lingual surface of tooth	Yes‐ Only at crest of pleura	Neighbouring teeth or interdental ridges	U‐shaped or C‐shaped	Alveolar bone and cementum along mesial and distal surfaces. Absent on lingual surface
Varanus‐type	Ankylosis	Yes‐attached to lingual surface of tooth	Yes‐ throughout	Pleura	Infolded (plicidentine)	Alveolar bone and cementum extremely thin. Restricted to labial surface and base of tooth
Dracaena‐type	Ankylosis	No‐ space for attachment tissue development on all sides of tooth	Yes‐ throughout	All directions	Circular	Cementum extremely thick and vascularized around the whole tooth root

4. DISCUSSION AND CONCLUSIONS

4.1. “Pleurodonty” and thecodonty: differences in topology, not tissue types

Our coronal and transverse sections through squamate jaws revealed the same periodontal tissues – cementum, periodontal ligament, and alveolar bone – as in thecodont mammals and crocodilians (Figures 2, 3, 5, 6, 7, 8, 9, 10; Suppl. Figure S2). Squamates show a diversity of tooth implantation modes (Berkovitz & Shellis, 2017; Caldwell, 2007; Edmund, 1960; Haridy, 2018; LeBlanc et al., 2017a; Peyer, 1968; Zaher & Rieppel, 1999) (Table 2), but where pleurodonty fundamentally differs from other implantation modes is not in the identity of the attachment tissues, but in two major topological features (Table 2).

4.1.1. HERS and the dental lamina restrict tooth attachment in squamates

The first feature is the arrangement and size of the dental lamina and HERS along the lingual side of the tooth. In modern thecodont taxa (crocodilians and mammals), the dental lamina is a thin strand of epithelial tissue that is disconnected from the oral epithelium (LeBlanc et al., 2017b; Whitlock & Richman, 2013) (Suppl. Figure S2). The HERS is a separate structure from the dental lamina, the former of which eventually breaks apart around the developing tooth, allowing dental follicle‐derived tooth attachment tissues to adhere to the tooth root surface (Luan et al., 2006; McIntosh et al., 2002). By comparison, the HERS and dental lamina in Iguana, Sauromalus, Cordylus, Scincus, Lacerta, Varanus, and via phylogenetic bracketing, presumably in other pleurodont squamates, form a continuous wall of epithelial tissue that is directly attached to the lingual surface of the tooth (Delgado et al., 2005; Richman & Handrigan, 2011) (Figures 2, 9, 11a–d).

FIGURE 11.

Interpretations of tooth attachment and implantation in Iguana‐, Varanus‐, and Dracaena‐type dentitions. (a) coronal view of Iguana‐type implantation and attachment. Note the extensive epithelium covering the lingual surface of the dentine. (b) transverse view of Iguana‐type implantation and attachment. Note the partially mineralized interdental spaces. (c) coronal view of Varanus‐type implantation and attachment. Note the extensive covering of epithelium over most of the dentine. (d) transverse view of Varanus‐type implantation and attachment. Note the nearly completely unmineralized interdental spaces (e) coronal view of Dracaena‐type attachment and implantation. Note the interpreted absence of epithelium‐dentine contact. (f) transverse view of Dracaena‐type attachment and implantation. Note the completely mineralized interdental spaces. Abbreviations: dl, dental lamina; HERS, Hertwig's Epithelial Root Sheath; ICT, Interdental Connective Tissue

Furthermore, unlike in thecodont taxa, HERS is a permanent structure and it impedes the formation of attachment tissues wherever it covers the root surface (Luan et al., 2006; McIntosh et al., 2002). This has an important effect on the arrangement of the connective tissues associated with each tooth; periodontal tissues can only form at the very base of the lingual surface of the tooth, or along the mesial and distal surfaces of the teeth, which are free from the epithelial covering (Figures 2, 3, 5, 6). Our sample of tupinambine teiids demonstrates that this may not be an insurmountable constraint in squamates. Dracaena and Tupinambis have probably lost this direct connection between the epithelium and the tooth root and can form tooth attachment tissues along the entire lingual surfaces of the tooth roots (Figure 10). As far as we are aware, deposition of such extensive layers of cementum along the lingual surfaces of the tooth roots is impossible without severing the connection between the developing root and HERS (Luan et al., 2006; McIntosh et al., 2002). Therefore, these types of dentitions would not qualify as pleurodont, despite having superficially similar jaw geometries along some regions of the jaw (Figure 11e,f).

4.1.2. Asymmetrical tooth development in squamate teeth

The second major difference is in the developmental symmetry of the tooth root and the entry point for the pulp vasculature and innervation. During root formation in thecodont teeth, the labial and lingual extensions of HERS are symmetrical, eventually forming a conical or cylindrical root. Once root formation is complete, the root apex remains open to receive the blood vessels and nerves for the pulp. In pleurodont teeth, the labial extension of HERS in developing teeth is truncated compared to the lingual wall (Figure 1; Suppl. Figure S1, S3). This truncation leaves the mature tooth with a long lingual root wall, a short labial wall, and a large opening along the labial surface of the tooth, exposing the pulp, and providing an entry point for the pulp vasculature and innervation (Figure 1b,d). Dental transplant experiments in Iguana have shown that even when tooth buds are transplanted to the corneal tissue, the developing root forms asymmetrically in the absence of jawbone, suggesting that it is an intrinsic feature of pleurodont teeth (Howes, 1979). This asymmetry means that there are virtually no attachment tissues along the labial surface of the tooth root, except for the contact between the tooth and the crest of the pleura. Again, Dracaena and Tupinambis demonstrate that some squamates have shifted away from this mode of tooth development. Dracaena and Tupinambis have symmetrical tooth roots and apical openings for the passage of the pulp vasculature, similar to other thecodont amniotes (LeBlanc et al., 2017b; LeBlanc & Reisz, 2013; Snyder et al., 2020) (Figures 10, 11). The same condition is found in many snakes and the extinct mosasaurs (Caldwell, 2007; Caldwell et al., 2003; LeBlanc et al., 2017a).

As a result of the persistence of epithelial tissue along the lingual surfaces of the teeth, and the lack of attachment sites along the labial surfaces, most pleurodont teeth mainly anchor to the surfaces of the neighboring teeth mesially and distally via mineralized attachment tissues (Figures 6, 7, 11a–d). This greatly differs from the condition in thecodont teeth, where there is attachment all around the tooth root (Berkovitz & Sloan, 1979; LeBlanc et al., 2017b) (Suppl. Figure S2). However, this should not be conflated with the difference in tooth attachment tissues between pleurodont squamates, mammals, and crocodilians.

4.2. Squamates have tooth sockets and tooth roots

According to some authors, only the roots of the thecodont teeth of mammals and crocodilians are housed in true sockets (Luan et al., 2009; Peyer, 1968; Zaher & Rieppel, 1999). The rationale is that true sockets are discrete units of bone surrounding a tooth, and the connecting tissue is a periodontal ligament. This supposedly differs from the case in most other amniotes, which have “bone of attachment” fusing teeth directly to the jaws, or to interdental ridges of bone (Luan et al., 2009). This dichotomy between “socketed” and “non‐socketed” tooth implantation modes started with Tomes' (1874) description of “bone of attachment” and depends on the non‐homology of attachment tissues between mammals, crocodilians, and all other amniotes (see below). As we have shown here and previously (Caldwell, 2007; Caldwell et al., 2003; LeBlanc & Reisz, 2013; LeBlanc et al., 2016a, 2017b, 2017a, 2018), cementum, periodontal ligament, and the alveolar bone that define a tooth socket (Ten Cate & Mills, 1972) are not restricted to mammals and crocodilians, but are present across extinct and extant members of Amniota.

Our ontogenetic analysis of I. iguana dentaries in particular shows that the alveolar bone forms a groove for the tooth root in transverse sections (Figures 3, 4, 5, 6). This is similar to what occurs in hadrosaurid dinosaurs, where the topological shift in the enamel and cementum‐bearing surfaces of the tooth roots led to an evolutionary shift toward a one‐sided alveolar groove for each tooth (Bramble et al., 2017; LeBlanc et al., 2016b). In squamates, this asymmetry in alveolar shape is associated with the extent and topology of the dental lamina and HERS, the asymmetrical development of the tooth root, and the shape of the jaws.

We also reject the hypothesis that squamates differ from thecodont amniotes in attaching their teeth to an “apparent socket” made from a standalone interdental ridge of bone (contra Luan et al., 2009; Zaher & Rieppel, 1999). The interdental ridges in pleurodont squamates are accumulations of old dental tissues from repeated bouts of tooth resorption, migration, and replacement. This has not only been demonstrated in extant pleurodont squamates, but in mosasaurs (Caldwell, 2007; Caldwell et al., 2003), snakes (Budney et al., 2006; LeBlanc et al., 2017a), dinosaurs (LeBlanc et al., 2017b), Paleozoic reptiles (Haridy et al., 2018), non‐mammalian synapsids (LeBlanc et al., 2018), and even in stem amniotes (LeBlanc & Reisz, 2013). These structures form as a consequence of being polyphyodont, where successive generations of tooth replacement never fully resorb and remove remnants of previous generations of tooth attachment tissues.

A similar argument can be made for the nature of the tooth root in squamates. Some have argued that only mammals and crocodilians have true tooth roots (Peyer, 1968; Zahradnicek et al., 2012). This is difficult to reconcile with the observations of cementum (a root tissue) along the tooth roots of squamates presented here and elsewhere (Caldwell et al., 2003; Luan et al., 2006; McIntosh et al., 2002), and the presence of HERS. In mammals and crocodilians, HERS aids in the development of the tooth root, eventually disintegrates into small clusters of epithelial cells, and allows the surrounding periodontal tissues to attach around the whole tooth root. In squamates, HERS is a permanent structure (Luan et al., 2006) that impedes tooth attachment tissue development. However, this does not mean that squamates do not have a true tooth root. Hertwig's original discovery of the root sheath was not in mammals, but in amphibians (Luan et al., 2006). This tissue is responsible for the development of the tooth root and it is present in all tetrapods. However, HERS does show some differences across mammals, crocodilians, and squamates. In crocodilians and mammals, HERS is only two cells thick along developing tooth roots (Luan et al., 2006; McIntosh et al., 2002; Nanci, 2013; Zahradnicek et al., 2012). This is the case in the developing teeth of some squamates (Suppl. Figure S3), but we also observed an epithelial covering along the lingual and labial sides of developing teeth that was much thicker in many sections. A thick HERS has been observed in developing gecko teeth (McIntosh et al., 2002: Figure 7c; Zahradnicek et al., 2012). The significance of this difference is unclear (Zahradnicek et al., 2012), but this does not preclude squamates from having tooth roots.

4.3. Squamates do not have “bone of attachment”

The classical hypothesis for the evolution of tooth implantation in amniotes posits that thecodonty evolved from an ancestral tooth implantation mode with a primitive complement of “bone of attachment” (Berkovitz & Shellis, 2017; Gaengler & Metzler, 1992; Osborn, 1984; Peyer, 1968; Tomes, 1882; Zaher & Rieppel, 1999). This evolutionary transformation would have required not only a shift in the architecture of the teeth and jaws, but a developmental shift in the nature of the periodontium to accommodate teeth within a socket, via a periodontal ligament (Osborn, 1984). This purely hypothetical model relies on descriptions of reptilian tooth attachment tissues by Charles Tomes (1874, 1882) who asserted that nearly all reptiles fuse their teeth to the jaws via a single “bone of attachment” tissue. Tomes observed a single bone mass, independent of the bone of the jaw that seemed to form a strong connection with each tooth:

…the lacunae are large and very irregular in form, and it may very fairly be called a rough and imperfect osseous structure, with which, at its upper parts, dentinal tubules blend. It is this irregular bone which at times breaks away and remains attached to the base of a tooth, and it may be conveniently designated as ‘bone of attachment’… (Tomes, 1874: 45)

Tomes first described the process of ankylosis by “bone of attachment” in a modern snake:

So soon as the tooth with its capsule has reached its future position, a development of new bone takes place beneath it, upon the surface of the jaw, and progresses with very great rapidity, until it reaches the base of the dentine pulp (Tomes, 1874: 49)

The ripple effect from this description was far‐reaching, and continues to affect our interpretations of tooth tissue homology in modern amniotes (Berkovitz & Shellis, 2017; Buchtová et al., 2013; Caldwell et al., 2003; Gaengler & Metzler, 1992; Huysseune & Sire, 1998; LeBlanc et al., 2017a; Luan et al., 2006, 2009; Smith & Hall, 1993; Zaher & Rieppel, 1999). This in turn has affected previous characterizations of pleurodonty in squamates, the homology of the attachment tissues within Squamata, and its bearing on the phylogenetic relationships of several major groups (Caldwell et al., 2003; LeBlanc et al., 2017a; Lee, 1997; Luan et al., 2009; Rieppel & Kearney, 2005; Zaher & Rieppel, 1999).

Tomes’ descriptions provide some histological properties for “bone of attachment”: (1) it is resorbed and reformed with each individual tooth; (2) it contains large cell lacunae, (3) it is more irregular than the surrounding bone of the jaw, and (4) it extends from the surface of the jaw toward the base of a tooth (Tomes, 1874, 1882). Whereas some authors would argue that these distinguish “bone of attachment” from alveolar bone or cementum in crocodilians and mammals (e.g., Luan et al., 2009), they actually show no difference from the development and histology of alveolar bone in polyphyodont amniotes (Budney et al., 2006; Caldwell, 2007; Caldwell et al., 2003; LeBlanc & Reisz, 2013; LeBlanc et al., 2016a, 2017b, 2017a, 2018; Maxwell et al., 2011a; Pretto et al., 2014). Alveolar bone is always resorbed and rapidly formed anew with each tooth replacement event, forming an irregular mass of bone with abundant (and often large) cell lacunae that extends from the jaw to the tooth root. In some groups it continues to mineralize until it fuses the tooth in place, whereas in others it stops, leaving space for the unmineralized periodontal ligament to suspend the tooth (LeBlanc et al., 2016a, 2017a, 2018).

The “bone of attachment” paradigm hinges on the presence of a single attachment tissue in ankylosed teeth (Tomes, 1874). However, in addition to identifying its developmental and histological similarity to alveolar bone, we, and others (Caldwell et al., 2003; Luan et al., 2006, 2009; Maxwell et al., 2011c) have demonstrated the presence of a second mineralized periodontal tissue in squamates: cementum. In Dracaena, Tupinambis (Figure 10), and the extinct mosasaurs (Caldwell, 2007; Caldwell et al., 2003; LeBlanc et al., 2017a), it is the cementum – growing in the opposite direction of supposed “bone of attachment” – that forms the principle mineralized portion of the periodontium. In other squamates it can form a slightly smaller portion of the periodontium, as in Iguana (Figures 2, 5, 6), or it may be significantly reduced, as in Varanus (Figures 8, 9).

Furthermore, the presence of Sharpey's fibers from a periodontal ligament – and even the periodontal ligament itself – in several squamate lineages is inexplicable under the “bone of attachment” paradigm. Conventionally, the periodontal ligament is a specific network of partially mineralized collagen fiber bundles that surrounds and supports a thecodont tooth root (Luan et al., 2009; McIntosh et al., 2002; Miller, 1968; Nanci, 2013). The orthodontic and physiologic roles of this ligament are well‐known, and the orientations of the collagen fibers around the tooth reflect its various functions (Bosshardt et al., 2008; Lin et al., 2017; Naveh et al., 2018). For most squamates, this is clearly not the case, given that the teeth are fused to the jaws once fully erupted. However, there is an unmineralized stage of tooth attachment in squamates. A fibrous tissue anchors into the tooth cementum and surrounding alveolar bone of newly erupted teeth (Figures 5, 6). We have demonstrated here that the squamate periodontal ligament is a transient tissue – containing an abundance of fibroblasts, cementoblasts, and osteoblasts –that is eventually completely encapsulated within the mineralized cementum and alveolar bone. The collagen fiber bundles of the squamate periodontal ligament are not normally oriented as a tensile sling around the tooth like in mammals or crocodilians (LeBlanc et al., 2017a); instead they presumably orient and fix an erupting tooth in place prior to ankylosis.

Given that the squamate periodontal ligament clearly does not perform the same functions as it does in mammals and archosaurs (Bosshardt et al., 2008; LeBlanc et al., 2017b; McIntosh et al., 2002; Miller, 1968), we suggest a much broader definition of this dental tissue in order to communicate its homology across Amniota: the amniote periodontal ligament is a network of collagen fiber bundles, prone to partial or complete mineralization (Bosshardt et al., 2008; LeBlanc et al., 2016a; McIntosh et al., 2002; Nanci, 2013), that initially anchors a tooth to the jaw through its connections to cementum and alveolar bone, all of which are derivatives from the dental follicle. This definition emphasizes the developmental and topological similarities of this fibrous connective tissue ‐ as well as its histological variability ‐ across an astonishing diversity of living and extinct species (Budney et al., 2006; Caldwell et al., 2003; LeBlanc et al., 2017b, 2018; LeBlanc & Reisz, 2013; Maxwell et al., 2011a).

Despite attaching to the crest of the pleura and to neighboring teeth via mineralized tissue, we conclude here that modern pleurodont squamates have an equally complex and modular periodontium, with the same periodontal tissue complement as thecodont crocodilians and mammals (Figures 2, 3, 4, 5, 6, 7, 8, 9, 10). Where we disagree with Tomes’ original distinction between the teeth of reptiles and those of mammals is that ankylosis superficially masks the tissue complexity underlying reptilian tooth attachment (Caldwell et al., 2003; LeBlanc et al., 2016a, 2017a; Pretto et al., 2014). Proper histological characterization of reptilian tooth attachment requires multiple samples, documenting dental ontogeny in multiple planes of section (we observed that the coronal plane of section, commonly adopted in previous studies of squamate tooth histology, was the least informative in this regard [Figure 2A]). The historical absence of the latter in descriptive works has greatly limited our understanding of tissue complexity in all squamates (but see Buchtová et al., 2013; Dosedělová et al., 2016; Haridy, 2018; Luan et al., 2006).

We recognize that squamate pleurodonty represents a different arrangement of oral tissues compared to other forms of tooth implantation. However, we also conclude that it is imperative to understand that the oral tissues associated with pleurodonty are histologically indistinguishable in their details from the oral tissues of mammals and crocodiles, a.k.a., the thecodont taxa. With respect to Tomes’ (1874, 1882) “bone of attachment”, we conclude that it is histologically identical to alveolar bone in most cases (LeBlanc & Reisz, 2013; LeBlanc et al., 2016a), to a mineralized periodontal ligament in others (LeBlanc et al., 2017a; Luan et al., 2009), and sometimes confused with cementum (Rieppel & Kearney, 2005). Because of this confusion arising from Tomes (1874, 1882), the “bone of attachment” as characterized by subsequent authors has included alveolar bone, mineralized periodontal ligament, cementum, and the resorption remnants of previous dental generations in such structures as interdental plates (LeBlanc et al., 2017b) (Figure 7). In other words, “bone of attachment” cannot be characterized histologically as a distinct tissue, consistent with arguments first made by Caldwell et al. (2003).

AUTHOR CONTRIBUTIONS

A.R.H.L. conceived the study, A.R.H.L. and I.P. performed the histological analyses, M.R.D., A.R.H.L., and I.P. collected the CT‐scan data, A.R.H.L., I.P., D.O.L., and M.W.C. interpreted the data, all authors wrote and edited the manuscript.

Supporting information

Supplementary Material

DATA AVAILABILITY STATEMENT

Raw histology and CT images and files used in this study are available at: http://morphobank.org/permalink/?P3846, project #3846.