Paleobiology, 45(4), 2019, pp. 676–697 DOI: 10.1017/pab.2019.31

Article

Bony lesions in early tetrapods and the evolution of mineralized tissue repair

Eva C. Herbst , Michael Doube , Timothy R. Smithson, Jennifer A. Clack, and John R. Hutchinson

Abstract.—Bone healing is an important survival mechanism, allowing vertebrates to recover from injury and disease. Here we describe newly recognized paleopathologies in the hindlimbs of the early tetrapods Crassi- gyrinus scoticus and Eoherpeton watsoni from the early Carboniferous of Cowdenbeath, Scotland. These path- ologies are among the oldest known instances of bone healing in tetrapod limb bones in the fossil record (about 325 Ma). X-ray microtomographic imaging of the internal bone structure of these lesions shows that they are characterized by a mass of trabecular bone separated from the shaft’s trabeculae by a layer of cortical bone. We frame these paleopathologies in an evolutionary context, including additional data on bone healing and its pathways across extinct and extant sarcopterygians. These data allowed us to synthesize information on cell-mediated repair of bone and other mineralized tissues in all vertebrates, to reconstruct the evolutionary history of skeletal tissue repair mechanisms. We conclude that bone healing is ancestral for sar- copterygians. Furthermore, other mineralized tissues (aspidin and dentine) were also capable of healing and remodeling early in vertebrate evolution, suggesting that these repair mechanisms are synapomorphies of vertebrate mineralized tissues. The evidence for remodeling and healing in all of these tissues appears con- currently, so in addition to healing, these early vertebrates had the capacity to restore structure and strength by remodeling their skeletons. Healing appears to be an inherent property of these mineralized tissues, and its linkage to their remodeling capacity has previously been underappreciated.

Eva C. Herbst* and John R. Hutchinson. Structure and Motion Laboratory, Royal Veterinary College, Hawkshead Lane, Hatfield, Hertfordshire AL9 7TA, United Kingdom. E-mail: [email protected], [email protected] Michael Doube. Department of Infectious Diseases and Public Health, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong. E-mail: [email protected] Timothy R. Smithson and Jennifer A Clack. University Museum of Zoology, Cambridge, Downing Street, Cambridge, CB2 3EJ, United Kingdom. E-mail: [email protected], [email protected]

Accepted: 13 August 2019 First published online: 27 September 2019 *Correspondence to: Eva C. Herbst, [email protected]. Data available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.30kp1qg; and Figshare Digital Repository: https://doi.org/10.6084/m9.figshare.9211643

Introduction are to regain competitive function. Despite its Bones have a remarkable ability to heal after broad distribution in extant taxa and clear traumatic injury and in the face of infectious, role in enhancing evolutionary fitness within neoplastic, and other insults. Up to half of and across species, an evolutionary context juvenile humans experience and survive bone for bone healing has not yet been elucidated. fracture (Jenkins et al. 2018). Wild animals The broad category of bone repair may be such as hawks, bottle-nosed dolphins, urban subdivided into several different categories: foxes, and serrasalmid fish (the group includ- regeneration, nonregenerative large-scale heal- ing piranhas) have high rates of fracture inci- ing, remodeling, and biophysical annealing dence and survival (de Smet 1977; Harris (Fig. 1). In regeneration (Fig. 1A), whole limb, 1978; Roth et al. 2002; Kolmann et al. 2018). In tail, and fin segments regrow via the formation human and veterinary medicine, fracture repair of a blastema, which comprises dedifferen- may be enhanced by supporting bones with tiated cells (Egawa et al. 2014; see Akimenko splints, casts, slings, surgical implants, and et al. [2003] and Akimenko and Smith [2007] medication; wild animals are not afforded for fin regeneration). We divide nonregenera- such care and must heal by themselves if they tive large-scale healing into three subcategories:

© 2019 The Paleontological Society. All rights reserved. This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited. 0094-8373/19 Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 677

FIGURE 1. Generalized schematic of bone repair mechanisms: A, regeneration; B, bone healing; C, remodeling; and D, biophysical annealing. *For healing of dermal bone in various taxa (via cartilaginous or fibrous callus formation), see Moss (1962), Irwin and Ferguson (1986), and Geurtzen et al. (2014). For endochondral healing mechanisms with surgical fixation, see Shapiro (2008).

fracture repair, bony collar formation (sub- collar of subperiosteal woven bone (Uthgen- periosteal reactive bone and exostoses), and nant et al. 2007; Shapiro 2008; Marsell and modeling and remodeling around space-filled Einhorn 2011). We do not include healing lesions (such as cysts, abscesses, and neo- of stress fractures or microfractures in the plasms), which for brevity we refer to collect- category of fracture repair, because the healing ively as bone healing (Fig. 1B). Bony collar process of stress and microfractures omits fibro- formation is modeling “outside” the bone cartilaginous callus formation. organ, by the formation of subperiosteal react- Bone healing usually incorporates remodel- ive bone and exostoses. An example of bony ing. During fracture repair, woven bone in the collar formation is in equine “bucked shins,” callus is removed and replaced by mature where bone forms in response to high-strain lamellar bone, restoring the strength of the cyclic injury (Nunamaker 2002). Modeling bone organ (Shapiro 2008). In our terminology and remodeling around space-filling lesions (following Frost 1991), remodeling refers to the help heal bone in response to disease and infec- process of bone resorption by osteoclasts and tion. Here we use the term fracture repair only to deposition by osteoblasts without an overall refer to the healing processes that occur after change in organ shape (i.e., formation follows gross organ breakage into separate fragments. resorption on the same surface), and modeling This “secondary” or “indirect” fracture repair refers to unbalanced formation and resorption, process is characterized by rapid fibrous/cartil- leading to a net change in organ shape (e.g., cal- aginous callus formation that closes the fracture lus formation) or size (e.g., during growth) gap and stabilizes interfragmentary movement (Fig. 1B,C). Modeling and remodeling function (Shapiro 2008; Marsell and Einhorn 2011). In during normal bone growth; modeling contrast, incomplete and fatigue (“stress”)frac- increases bone size during development, and tures that are stable and lack a fracture gap heal modeling and remodeling enable growth to by secondary osteonal remodeling that fixes the occur while keeping dynamic strains consistent intracortical cracks and/or by deposition of a (Biewener et al. 1986;Frost1991). Furthermore,

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 678 EVA C. HERBST ET AL.

bone remodeling can function in mineral 2018). Paleopathologies can provide insight homeostasis (Doherty et al. 2015). Remodeling into the evolution of bone repair, because repairs microfractures in cortical bone (Burr paleopathologies are direct evidence of bone’s 1993; Mori and Burr 1993; Bentolila et al. response to insult. 1998). Bone’s healing response can consist of Although it is often difficult to determine the tissue formation only (i.e., modeling), for exact etiology of a paleopathology (Moodie example, subperiosteal bone deposition form- 1917), bone healing, regeneration, and remod- ing a bony collar. Remodeling is not always eling can be distinguished from one another necessary to reconnect two fractured segments in the fossil record. For example, the presence or to infill a lesion, but is advantageous in the of a fracture callus in a fossil indicates that the healing pathway, because it re-establishes full animal survived the fracture and started the strength in the healed structure (Shapiro fracture repair process. Bone healing may also 2008). Due to its tissue maintenance function manifest as an unusual bone formation in and ability to repair microfracture and stress response to nonfracture insult or disease; some- fracture independently, we place remodeling in times it is not possible to specify the type of its own category within bone repair, while rec- bone healing (response to fracture vs. other dis- ognizing its contribution to the later stages of ease). Remodeling can be identified by the pres- fracture repair. Modeling around space-filling ence of secondary osteons. Microcracks (and lesions, and bony collar formation, can also be hence, their repair) are difficult to identify in followed by remodeling. However, the subcat- the fossil record; however, there is evidence of egories of bone healing are united by the char- remodeling as part of growth (Giles et al. acteristic that they are modeling processes, with 2013) and as a late phase of fracture repair in a net change in organ shape (e.g., formation of a fossils (Lingham-Soliar 2004;ZammitandKear callus or collar or infilling of a lesion). There- 2011; Pardo-Pérez et al. 2018). fore, we do not include remodeling as a subcat- Paleopathological evidence of regeneration egory of bone healing. consists of unique limb patterning associated Microfractures may repair through biophys- with abnormal regeneration. Fröbisch et al. ical annealing (Fig. 1D). In biophysical anneal- (2014) described limb regeneration in the fossil ing, mineralized matrix accretes within record in the 300-million-year-old temnospon- microcracks (Boyde 2003), presumably by pas- dyl amphibian Micromelerpeton. Tail regener- sive precipitation of hydroxyapatite and other ation in the microsaurs Hyloplesion and moieties from bone tissue fluid onto exposed Microbrachis resembles that of extant salaman- matrix surfaces, as occurs in acellular dental ders, and these early tetrapods may also have enamel. The exact mechanism of biophysical been able to regenerate limbs (Fröbisch et al. annealing is not known, but it does not involve 2015; van der Vos et al. 2018). Nogueira et al. remodeling or osteocyte apoptosis. Osteocytes (2016) reported that gene expression is similar might have an indirect role by modulating in extant salamander limb and lungfish fin fluid composition in the surrounding lacunar regeneration. All these studies analyzed their canalicular space (Herman et al. 2010; Seref- findings in a broader phylogenetic context, Ferlengez et al. 2014). inferring that regeneration of both tails and We refer to regeneration, bone healing, and appendages (fins/limbs) is ancestral for tetra- remodeling as being cell-mediated, because pods, perhaps even a synapomorphy of all sar- they involve the resorption of bone by osteo- copterygians or even all osteichthyans. clasts and deposition of bone by osteoblasts, Here our goal is to answer the question of whereas biophysical annealing lacks obvious when bone healing mechanisms evolved by cellular control. All three cell-mediated bone synthesizing available data, similar to the stud- repair processes, but not biophysical annealing, ies of regeneration noted. We describe new evi- have been discovered in fossils (Capasso et al. dence of bone healing in the hindlimbs of the 1996; Tanke and Currie 1998; Anné et al. 2014; early tetrapods Crassigyrinus scoticus and Eoher- Fröbisch et al. 2014, 2015; Stilson et al. 2016; peton watsoni from the early Carboniferous. Pardo-Pérez et al. 2018; van der Vos et al. Along with the slightly older Ossinodus (Warren

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 679

and Ptasznik 2002; Bishop et al. 2015), these are Therefore, the earliest exoskeletons may have the earliest known instances of healing in tetra- been derived from both neural crest cells and pods. We considered these pathologies in a mesoderm (Shimada et al. 2013). These discov- phylogenetic context of extinct and extant verte- eries support the link between dermal bones brates to reconstruct the evolution of nonregen- and endochondral bones. Although tetrapod erative repair in bone. limb bones are endochondral and endoskeletal, Our investigation began with bone healing in and fish fin rays are dermal bones and are part limb (endochondral) bones of tetrapods and of the exoskeleton, both share a mesodermal expanded to include dermal bones. While der- origin. Embryonic origin therefore does not mal and endochondral bones are characterized separate the endoskeleton and the exoskeleton by different ossification pathways (endochon- (Hall 2014; Hirasawa and Kuratani 2015). In dral bones form via a cartilage precursor), light of the overlapping characteristics of ossifi- both of these types of bone arise from the activ- cation pattern and embryonic origin in the ity of osteoblasts, and the bone formed does not endoskeleton and exoskeleton, we investigated differ between the two mechanisms (Shapiro endochondral and dermal bones of both the 2008). Furthermore, endochondral bones such endo- and exoskeleton. as limb bones undergo subperiosteal intra- We compared bone healing with regeneration membranous ossification in the diaphysis and to examine the evolution of these different repair part of the metaphysis during development, mechanisms. Furthermore, we investigated and intramembranous ossification also occurs remodeling capacities to determine whether during fracture repair of these bones (Hall remodeling and healing evolved concurrently. 2005; Shapiro 2008). In birds and mammals, In addition to our first aim of reconstructing intramembranous (dermal) bones form second- how bone healing evolved and how it relates ary cartilage during fracture repair (Irwin and to regeneration and remodeling, our second Ferguson 1986). The loading environment aim was to reconstruct the evolution of healing seems to have a strong influence on the type in other skeletal tissues. We surveyed the fossil of bone ossification, during both development record of mineralized vertebrate skeletal tissues and fracture repair (Shapiro 2008 and refer- with evidence of healing and remodeling ences therein). Evolutionarily, ancestrally endo- (bone, aspidin, dentine), including conodont chondral bones can become membrane bones mineralized tissues as an outgroup comparison in some taxa, such as the orbitosphenoid of with unusual convergent evolution of minera- Leposternon (Hall 2005). lized tissue repair. The early vertebrate exoskel- Bone can also be categorized as belonging to eton evolved before the endoskeleton, and the endoskeleton or exoskeleton. While all consisted of both bone (aspidin in heterostra- bones in the exoskeleton are dermal and ossify cans) and dentine (Donoghue et al. 2006; Hall intramembranously, the endoskeleton ossifies and Witten 2007). Aspidin has recently been through both endochondral ossification (e.g., identified as acellular dermal bone based on in limb bones) and intramembranous ossifica- spaces that were probably left behind by intrin- tion (in membrane bones such as many cranial sic collagen bundles, much like those found in elements). It was traditionally thought that the teleosts (Keating et al. 2018). There is still debate exoskeleton developed from neural crest cells about the evolution of mineralized tissues and the endoskeleton developed from meso- (Smith and Hall 1990; Donoghue and Sansom derm (Hirasawa and Kuratani 2015). However, 2002; Donoghue et al. 2006; Hall and Witten recent research has shown that the endoskel- 2007; Doherty et al. 2015;Hall2015), but a recent eton and exoskeleton cannot be distinguished evolutionary analysis showed that cellular bone by ossification type or embryonic origins. For evolved from aspidin at least twice in vertebrate example, mouse parietals (part of the exoskel- evolution (Keating et al. 2018). Furthermore, eton) have a mesodermal origin (Jiang et al. Hall and Witten (2007) discussed the overlap 2002). Furthermore, in the medaka (Oryzias), in various characteristics of the different minera- the mesoderm is the embryonic origin of the lized vertebrate skeletal tissues and the presence trunk exoskeleton (scales and fin rays). of intermediate tissues, concluding that a plastic

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 680 EVA C. HERBST ET AL.

skeletogenic cell, along with a skeletal resorbing morphology with which we could compare cell, evolved with the first vertebrate minera- the potentially pathological features. lized skeletal tissues. We investigated whether We placed our results into a phylogenetic healing of vertebrate skeletal tissues evolved framework to examine the evolution of bone early on in vertebrate evolution and whether evi- healing. Sarcopterygian evolutionary relation- dence of healing and remodeling abilities coin- ships were based on recent phylogenetic ana- cide with one another in time and taxa. lyses (Zhu and Yu 2002; Ruta et al. 2003; Institutional Abbreviations.—NHMUK R, Pyron and Wiens 2011; Chiari et al. 2012; Natural History Museum, London, United Clack et al. 2016). We mapped all occurrences Kingdom; NMS G, National Museums Scot- of bone healing in the fossil record onto the sar- land, Edinburgh, United Kingdom; QMF, copterygian phylogeny, recording fracture Queensland Museum, Brisbane, Australia. repair in extant animals as a proxy for bone healing ability and comparing these healing pathways and regeneration pathways in repre- Materials and Methods sentatives of all major sarcopterygian taxa. We investigated unusual bony features in the We also investigated bone healing and regen- early tetrapods Crassigyrinus and Eoherpeton. eration in extant actinopterygians (including All specimens are from the Dora Bonebed in acellular bone in teleosts), and then broadened the Limestone Coal Formation of the Clackman- our phylogenetic survey to all vertebrates to nan Group, Ammonoid Zone E1a of the Pende- map the occurrence of healing in all minera- lian substage, Serpukhovian, Mississippian lized tissues. The vertebrate phylogenetic tree (Browne et al. 1999). The Dora site is near Cow- was based on Gess et al. (2006), whose analysis denbeath, Fife, Scotland. We studied several found “agnathans” to be polyphyletic and specimens (Supplementary Table 1) with one placed osteostracans as the sister taxon to or more unusual features that we identified as gnathostomes. The phylogenetic placement of likely pathologies. Only the Crassigyrinus ribs acanthodians is contentious; phylogenetic ana- were previously described as pathological. lyses have placed Acanthodes and most other We imaged the bones with X-ray microtomo- acanthodians either as stem osteichthyans or graphy (XMT; Nikon Metrology XT H and stem chondrichthyans, the latter being sup- HMX 225 ST, and Bruker Skyscan 1172) to ported by more recent analyses and reanalysis investigate the internal structure, noting the of Acanthodes anatomy (Brazeau 2009; Brazeau qualitative density and distribution of cortical and de Winter 2015; Giles et al. 2015). Regard- and trabecular bone (scan data available from less, all of these analyses placed Doliodus within the Figshare Digital Repository: https://doi. Chondrichthyes, so we classified it as such for org/10.6084/m9.figshare.9211643). The result- our phylogenetic analysis. We omitted hagfish ing scans were then segmented in Mimics 19.0 and lampreys from our phylogeny, because (Materialise, Leuven, Belgium) to remove the they do not have mineralized tissues. matrix, view cross sections of the internal anat- omy, and create a 3D model of each bone. In Results one case, further image processing in ImageJ (Schneider et al. 2012) was used to highlight First, we present new observations of paleo- the trabecular architecture (see Supplementary pathologies in two early tetrapods, Crassigyri- Text 3 for information on the macro). nus and Eoherpeton (“Pathologies in Early Where possible (i.e., for the Crassigyrinus Tetrapods” section). Next, we place these paleo- metatarsal and Eoherpeton right fibula), the pathologies in the context of a synthesis of pub- external and internal morphology of the pre- lished data on bone healing, regeneration, and sumably pathological bone was compared remodeling in extinct and extant taxa (“Bone with a normal bone from the same species Healing, Remodeling, and Regeneration” sec- and site. Where no comparative limb material tion). For bone healing in extant taxa, we used existed, we used features associated with mus- fracture repair as a proxy for bone healing (frac- cle attachment as an example of normal ture repair is a type of bone healing, and

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 681

therefore indicates a taxon’s ability to respond which have been reported as the earliest verte- to damage by blastic activity to repair bone tis- brate mineralized tissues in the fossil record sue). In our synthesis, we included remodeling and are known from the late Cambrian onward related to both growth and healing, because (Sansom et al. 1992). Unfortunately, the phylo- they involve the same mechanism, and our genetic relationship of conodonts remains aim was to determine whether occurrences of uncertain, and a recent study disputed the clas- bone healing and remodeling are linked. sification of conodonts as vertebrates (Turner We then conducted a similar synthesis of et al. 2010). Conodont mineralized tissue is healing and remodeling in aspidin (“Aspidin not homologous to enamel (Kemp 2002b; Healing and Remodeling” section) and dentine Turner et al. 2010; Murdock et al. 2013), despite (“Dentine Healing and Remodeling” section) to prior reports (Sansom et al. 1992; Donoghue examine whether there is a link between the and Sansom 2002). Here we classify conodonts evolution of these two processes in other as an outgroup to vertebrates. Even recent stud- mineralized tissues. We do not discuss regener- ies that considered conodonts as vertebrates ation for aspidin and dentine, because in our agreed that the tissues constituting dental ele- terminology regeneration is distinct from ments evolved convergently with vertebrate bone healing and refers to the formation of a mineralized tissues (Shirley et al. 2018). Investi- body segment via the dedifferentiation of cells gating healing and remodeling in conodont and the recapitulation of developmental pro- dental elements can tell us whether certain cesses. In the fossil record, regeneration has mineralized tissue repair mechanisms are been reported only for sarcopterygian bone, unique to vertebrate mineralized tissues. and the identification in fossils is based on com- Overall results of the phylogenetic syntheses parison with salamander regeneration abnor- are presented in “Phylogenetic Synthesis.” malities. While Halstead (who also published under the name “Tarlo”) reported “regener- Pathologies in Early Tetrapods ation” in heterostracan dentine, this term refers XMT images revealed that the unusual to healing via replacement of dentine tubercles growths on the Crassigyrinus hindlimb and (Halstead 1969, 1973). Similarly, Lebedev et al. ribs and Eoherpeton right fibula are pathological (2009) used “regeneration” to mean regrowth of (Figs. 2–6). There is not enough evidence to dis- tissue in response to injury, the process that we cern exactly which disease or injury caused refer to as “bone healing.” these lesions in Crassigyrinus and Eoherpeton, Enamel and cartilage are not included in our but the lesions are evidence of bone healing. synthesis; although they are also skeletal tis- Crassigyrinus.—Two unusual protuberances sues, we chose to omit them for the following are present on the hindlimb bones of Crassigyr- reasons. Enamel is unable to remodel (Good- inus. They are on the right tibia of block NMS G man and Rose 1994; Maas and Dumont 1999), 1984.15.3 and on the left femur of block NMS G and no fracture or caries repair has been 1984.15.1 (see Supplementary Text 1 for more reported outside clinical intervention. Unminer- specimen information). Panchen and Smithson alized cartilage does not preserve well in the fos- (1990) described these features as being unique sil record, and although there is evidence of to Crassigyrinus and interpreted them as pos- calcified cartilage in early vertebrates (Janvier sibly being the attachment sites for a ligament et al. 2004), no evidence of healing or remodel- that prevented dislocation of the knee during ing has been reported. The capacity for extant swimming. tetrapod cartilage to remodel is contentious, Crassigyrinus femur.—The protuberance on but it appears that neither complete remodeling the Crassigyrinus femur is on the distal end of nor full healing are possible (Jackson et al. 2001; the ventral surface, near the knee joint Sharma et al. 2013). Even the cartilaginous skel- (Fig. 2). It is a pitted protrusion that is rounded eton of elasmobranchs does not remodel or heal proximally and widens slightly where it meets (Ashhurst 2004; Johansson et al. 2004). the distal surface of the bone. The interior of In the section on “Conodont Mineralized Tis- this growth is composed of trabecular bone, sues,” we discuss conodont dental elements, which is separated from medullary trabecular

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 682 EVA C. HERBST ET AL.

FIGURE 2. Crassigyrinus left femur in ventral (A) and anterior (B) views. Internal trochanter in transverse (C) and long-axis (D) sections; protuberance in transverse (E, G) and long-axis (F, H) sections. it, internal trochanter; pr, protuberance. Scale bars: 10 mm.

bone by cortical bone. On the proximal end of same bone. The internal trochanter is present the growth, some cortical bone modeling is evi- in other early tetrapods and has been inferred dent on the exterior of the trabecular bone. The to be a key muscle insertion site (Godfrey location of trabecular bone superficial to the 1989; Molnar et al. 2018). Comparing the cortex indicates that the growth is a pathology. unusual growth with the internal trochanter To confirm that this bone distribution is allowed us to control for intraspecific and onto- pathological and not characteristic of normal genetic variation in bone structure, because the bony features in Crassigyrinus, we investigated two features are on the same bone from the the internal structure of the well-developed same individual. The cross sections revealed internal trochanter on the anterior side of the that while there is some anteroposterior change

FIGURE 3. Crassigyrinus right tibia in medial (A) and posterior (B) views; protuberance in long-axis (C) and transverse (D) sections. oc, original cortex; pr, protuberance. Scale bars: 10 mm (A, B); 3 mm (C, D).

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 683

FIGURE 4. Crassigyrinus metatarsal in extensor (A) and medial or lateral (B) views; protuberance in long-axis (C) and trans- verse (D) cross sections. dc, drainage channel; pr, protuberance. Scale bars: 10 mm (A, B); 3 mm (C, D).

in the density of the trabecular bony tissue, the endochondral ossification of the bone are traits trabecular bone of the internal trochanter is of an osteochondroma, we reject this diagnosis, continuous with that of the femoral shaft because in osteochondromas the trabecular (Fig. 2C,D). This is different from the protuber- bone of the shaft is continuous with the tra- ance, where there is a clear separation of the tra- becular bone of the growth (Khurana et al. becular bone of the protuberance from the bone 2002). The presence of the original cortex shaft, demarcated by cortical bone (Fig. 2E,F). beneath the protuberance (Fig. 2E,F) suggests This difference between the growth and the that the pathology could be an exostosis internal trochanter confirms our identification formed after damage to the periosteum. Note of the unusual femoral growth as a pathology that at the distal end of the protuberance and not a muscle attachment site. Although (Fig. 2G,H), the original cortex is absent. How- the location on the distal femur and the ever, this is also the case for the proximal

FIGURE 5. Rib pathologies in Crassigyrinus NHMUK VP R10000. A, B, Photos of bony calluses in the healed ribs; C, 3D model of a bony callus on a healed rib fragment [A shows where fragment C broke off from the slab]; D, long-axis cross section; E, transverse cross section of nonpathological end of fragment; F, transverse cross section showing normal structure of the rib shaft; G, transverse cross section showing irregular bone formation external to the cortical bone; H, transverse cross section showing irregular bone in callus; I, transverse cross section showing sinus with a drainage channel, probably a site of infection. dc, drainage channel. Black lines in C indicate locations of cross sections D–I. Scale bars: 5 mm.

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 684 EVA C. HERBST ET AL.

FIGURE 6. Eoherpeton right fibula in lateral (A) and anterior (B) views; protuberance in long axis (C, E) and transverse (D, F) cross sections. The fossil has infill between trabecular spaces in E and F, this hyperdense matrix in intertrabecular spaces was removed in C and D (see Supplementary Text 3 for image-processing macro information). pr, protuberance. Scale bars: 10 mm (A, B); 3 mm (C–F).

(nonpathological) end (Fig. 2D). The articular conclude that this growth is evidence of bone surfaces of the limb elements of the Crassigyri- healing. nus fossil are not ossified, the bones probably Crassigyrinus metatarsal.—We also discov- had cartilaginous joint surfaces, and there ered a similar lesion on the extensor surface of may have been taphonomic distortion in this a Crassigyrinus metatarsal, NMS G 1984.15.2 element. Therefore, the extent of the cortex (Fig. 4A,B). As in the protuberances in the does not provide conclusive evidence to iden- tibia and femur, cortical bone separates the pro- tify the subcategory of bone healing present in tuberance from the trabecular bone in the shaft this element. (Fig. 4C,D). While we could not compare the Crassigyrinus tibia.—The protuberance on femoral and tibial lesions with normal long the Crassigyrinus tibia is on the proximal end bones (only one Crassigyrinus femur and one of the medial surface (Fig. 3). Like the femur, Crassigyrinus tibia have been found), there it is on the flexor surface of the bone, close to are six metatarsals attributed to Crassigyrinus the knee joint. The tuberosity on the tibia is tri- (Herbst and Hutchinson 2018). Only one meta- angular, and its surface is rugose, although not tarsal has this growth, which supports its iden- as pitted as the femoral protuberance. The tification as a pathology. There is a channel internal structure is very similar to that of the connecting a cavity within the bone protuber- femoral protuberance, with cortical bone separ- ance to the outside (“dc” in Fig. 4C), which ating the growth and tibial shaft (Fig. 3C,D). As may represent a draining sinus tract resulting described for the femur, the presence of the ori- from infection of the bone. Such infection of ginal bone cortex (Fig. 3C) is characteristic of an the cortical bone is called “osteitis” (although exostosis. However, this cortex becomes irregu- osteitis can be caused by osteomyelitis, infection lar (Fig. 3D), perhaps due to healing of a frac- of the bone marrow) (Pineda et al. 2009;Tie- ture, although no clear fracture line is visible. mann and Hofmann 2009). The pathology Unlike the femoral protuberance, there is no could also be a bony cyst, in which the fluid- evidence of cortical bone deposition on the filled cavity has started to heal by new bone for- external surface of the protuberance. The exter- mation (Jacobs 1955). Infections and cysts can be nal and internal morphology leads us to caused by various factors, such as trauma, but it

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 685

is not possible in this case to narrow down the Eoherpeton.—The right fibula of Eoherpeton etiology of this pathology. Regardless, all of has a protuberance similar to the Crassigyrinus these possibilities indicate net changes in bone lesions, in that it is a raised, button-shaped geometry as a healing response. growth on the anterior margin of the lateral Comparison with Ossinodus.—A strikingly side of the distal fibula (Fig. 6). Smithson similar pathology to those in the Crassigyrinus (1985) reported this feature and noted that is hindlimb bones, in terms of external and not present in Archeria, whose overall fibular internal structure, was described on the prox- morphology is similar to that of Eoherpeton. imal right radius of Ossinodus QMF 37451 We identified this feature as a possible path- (Queensland Museum, Brisbane), an early ology, and therefore investigated the internal tetrapod from the mid-Viséan of Australia structure. Our XMT imaging supports this (Warren and Ptasznik 2002). The Ossinodus interpretation. The bone is much less dense in radial pathology is especially similar to that of the presumably pathological area compared the Crassigyrinus tibia, because in both cases with the ridge on the posterior edge of the med- the pathology is at the proximal end of the zeu- ial side. The latter ridge is probably part of the gopod, near the elbow and knee joints, respect- normal morphology of the fibula, because it is ively (Supplementary Fig. 1). The Ossinodus also described in Archeria (Smithson 1985), as pathology is the oldest known evidence of well as most other early tetrapods, and is an bone healing in a tetrapod. Based on the path- osteological correlate of the origin of the flexor ology and other features of the internal bone accessorius lateralis (Molnar et al. 2018). The dif- structure, Bishop et al. (2015) inferred that Ossi- ference in density between the ridge and the nodus moved on land to some degree. The pres- growth is not just a taphonomic artifact of the ence of similar pathologies in the aquatic presence of infill (white in Fig. 6E,F) versus no Crassigyrinus and in the well-ossified and infill (Fig. 6C,D). Furthermore, the protuber- more terrestrial Eoherpeton (see “Eoherpeton” ance is not found on the left fibula associated section) demonstrates that early tetrapods with Eoherpeton (see Supplementary Text 2 for with a variety of locomotor modes experienced specimen info). such pathologies. Crassigyrinus ribs.—The Crassigyrinus ribs exhibit the only pathology in Crassigyrinus that Bone Healing, Remodeling, and Regeneration has been previously described. The NHMUK Extinct Taxa.— VP R10000 specimen shows bony callus forma- Stem tetrapods.—The earliest known case of tion, which is evidence of a healed fracture, in bone healing in a tetrapod is in a radius of Ossi- four ribs (Panchen 1985). The surfaces of these nodus from the mid-Viséan (∼333 Ma) of Aus- calluses are rugose, like the surfaces of the hind- tralia (Warren and Ptasznik 2002; Bishop et al. limb pathologies (Fig. 5A–C). Unlike in the 2015). This pathology has a similar appearance hindlimb, the rib pathologies involve the entire to the Crassigyrinus tibial protuberance. God- cross section of the rib at the site of the callus. frey (1988) described fracture calluses in the Whereas the normal bone structure in the rib ribs of a tetrapod from the late Viséan of West shaft is very dense (Fig. 5F), the callus is com- Virginia, and attributed these ribs to Crassigyr- posed of loose unorganized trabecular bone inus. There is not enough evidence to support and lacks cortical bone (Fig. 5H). This corrobo- this taxonomic referral, but nevertheless it is rates Panchen’sinterpretationofthesecalluses another case of fractured and healed ribs in a as evidence of healed rib fractures. The XMT tetrapod from the Carboniferous. Cases of scans also revealed subperiosteal primary bone bone healing are even more common after the formation, which may have formed in reaction Carboniferous and have been described for to the nearby fracture or as a result of a separate various taxa. We end our synthesis of bone subcritical fracture event (Fig. 5G). There is also healing data at the Carboniferous for fossil tet- a large sinus in the bone with a channel leading rapods, because subsequent cases of healing do to the outside, which might be a site of infection not add any new information on ancestral with a drainage channel (Fig. 5I). patterns.

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 686 EVA C. HERBST ET AL.

Other sarcopterygians (lobe-finned fish): placoderm Plourdosteus shows healing of the porolepiforms, lungfish, Rhizodontida.—The compact bone over the spongy bone at the oldest known instance of bone repair in Sarcop- injury margins. Another placoderm, Bothriole- tergygii is fracture repair with callus formation pis, exhibits lesions and healing (including cal- in response to bite injury on a scale of the por- lus formation) on both the ventrolateral plate olepiform Glyptolepis from the Eifelian stage, and pectoral fin (in different individuals). Sec- Middle Devonian. A scale that probably ondary osteons in another Bothriolepis species belongs to the porolepiform Holoptychius also also indicate remodeling capacity (Downs and displays healed bone (Lebedev et al. 2009). Donoghue 2009). Indeed, remodeling during The first evidence of endoskeletal bone healing growth is often found in various placoderms, is in the Carboniferous, and there are many including Dunkleosteus (Giles et al. 2013). examples in non-tetrapod sarcopterygians of Osteostracans.—Although healing has not this period. Several healed ribs (which are been reported in osteostracans, resorption of endochondral bone) are reported in lungfish bone has been observed in Tremataspis (Denison from Scotland (Sharp and Clack 2013), Kansas 1952). (Rothschild and Martin 2006), and England Extant Taxa.— (Barkas 1873). The healed bone at the sites of Sarcopterygians (including Tetrapoda).— these fractures appears very similar to that in Fracture repair of endochondral bones is the healed Crassigyrinus ribs, forming a callus similar across extant tetrapods. Axolotls on the rib shaft. There are also two fractured (Ambystoma), which are capable of regener- lepidotrichia (fin rays composed of dermal ation, repair fractures in the same manner as bone) with callus formation (indicating a heal- mammals, and in tetrapods (regardless of ing response) in the rhizodontid (extinct tetra- regeneration capacity) there is a critical gap podomorph fish) Barameda from the early size beyond which a fracture cannot heal Carboniferous of Australia (Garvey et al. (Hutchison et al. 2007). Fracture repair of der- 2005). We found no studies with evidence of mal bones differs between extant tetrapod bone healing in fossil Actinistia (coelacanths). taxa. Mammals and birds heal dermal bones Actinopterygians (ray-finned fish).—Evi- by forming a cartilaginous callus in a process dence of bone healing is rarely reported in the called secondary chondrogenesis (Irwin and fossil record of actinopterygians. However, Ferguson 1986). Secondary chondrogenesis is there are several cases from the Eocene Messel the formation of cartilage from the periostea pits. Fracture repair, identified by callus forma- of dermal bones after intramembranous ossifi- tion, was present in the bony caudal finraysof cation has occurred (Hall 2005). Salamanders Cyclurus (Amiidae), Amphiperca (Serranidae), and newts do not form cartilage during the Paleoperca (Serranidae), and Rhenaoperca (Mick- healing of dermal bones (Goss and Stagg lich and Mentges 2012). 1958; Hall and Hanken 1985). There is no evi- Placoderms.—There is evidence of healing dence of secondary chondrogenesis in snakes and remodeling around a trauma-induced or lizards (Irwin and Ferguson 1986). No mech- lesion in the dorsal plate (composed of dermal anistic studies of fracture repair, to our knowl- bone) of the placoderm Dunkleosteus from the edge, have been conducted on crocodiles or Givetian stage, Middle Devonian (Capasso turtles. In dermal bone fractures without sec- et al. 1996). The outer layer (which exhibits ondary chondrogenesis, a callus still forms, the most extensive healing and remodeling) but it is composed of fibrous tissue (Irwin and was described as “dentine-like.” However, we Ferguson 1986). The African lungfish Proto- determine the identity of the healed tissue to pterus heals dermal bone in its lower jaw by be bone, not dentine. The study noted the pres- forming a fracture callus, and bone remodeling ence of bone cells and lamellae in this layer, and is associated with the injury (Kemp 2001). Bone histological analysis of Dunkleosteus showed no healing in extant coelacanths has not been evidence of dentine (Giles et al. 2013). Lebedev described. Fracture repair in extant tetrapods et al. (2009) reported several instances of bone is usually followed by remodeling (see healing in Late Devonian taxa, as follows. The “Introduction”). In mammalian endochondral

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 687

fracture repair, there are two stages of osteo- bone fractures, but unlike in sarcopterygians, clast activity: resorption of calcified cartilage their fracture repair also involves tissue dedif- after soft callus formation and remodeling of ferentiation; a blastema forms in both crush bone in the hard callus (Gerstenfeld et al. fractures (Sousa et al. 2012) and normal frac- 2003). In calcified cartilage resorption, osteo- tures of fin rays (lepidotrichia, which are der- clasts are sometimes referred to as “chondro- mal bones), as well as in fractures of skull clasts,” due to the substrate on which they are bones (Geurtzen et al. 2014). In crush fractures working. Chondroclasts and osteoclasts appear in zebrafish, the bone repair process takes to be the same cell type based on shared fea- longer than in amputation regeneration tures and regulatory mechanisms and osteo- (Sousa et al. 2012). In medaka (Oryzias), fin clasts’ capacity for resorbing mineralized regeneration also proceeds more quickly than cartilage (Helfrich 2003;Włodarski et al. 2014). fracture repair (Takeyama et al. 2014). No sec- Some extant sarcopterygians can repair ondary cartilage forms during the healing of damaged skeletal tissue not only through frac- dermal bones in fish (Irwin and Ferguson ture repair but also through regeneration. 1986). The fracture repair process differs from the Glass knifefish (Eigenmannia) can regenerate limb regeneration process. Limb regeneration endochondral bone in their tail. Amputated involves cell dedifferentiation, and the forma- vertebrae do not reform, but are replaced by a tion of a blastema, whereas fracture repair long rod with the same histological structure does not. Axolotls can fully regenerate ampu- as the vertebrae (Kirschbaum and Meunier tated limbs (Hutchison et al. 2007). The frog 1981). Polypterus (bichirs) regenerate their pec- Xenopus forms a blastema but only grows a car- toral fins (composed of both endoskeletal and tilaginous spike in response to amputation dermal bones) by forming a blastema, similar (Egawa et al. 2014). Lizards can regenerate to amphibian regeneration. These studies their tail, but a cartilaginous rod replaces demonstrated that actinopterygian fins can bony vertebrae (Jacyniak et al. 2017). The lung- regenerate in response to amputation of the fish Protopterus can regenerate endochondral endoskeleton (endochondral fin bases), instead and dermal bone in its fins and tail, and the of regeneration being restricted to only the regeneration mechanism is similar to that of exoskeleton (composed of the finrays; urodeles (Conant 1970). dermal bone) (contrary to Akimenko and Bone remodeling is widespread in tetrapods, Smith 2007). Gene expression is also similar in enabling growth, repair, and mineral homeo- Polypterus, lungfish, and salamander regener- stasis (Witten and Huysseune 2009; Doherty ation, suggesting a common evolutionary ori- et al. 2015). Although remodeling is much gin of regeneration in these taxa (Lu et al. reduced in small birds and mammals (Currey 2019). Tilapia can regenerate lepidotrichia of et al. 2017), remodeling is still possible in the anal fin, and these regenerated structures these taxa; for example, it can be induced in segment and branch during growth, as do lepi- rats by increased loading (Bentolila et al. dotrichia formed during normal growth (Kemp 1998). Extant lungfish (Neoceratodus) show evi- and Park 1970). dence of bone resorption in and below their In the zebrafish crush-injury model, a soft dental plates during normal growth (Kemp callus formed from an epithelial thickening, 2001, 2002a). and bone deposition at the crush site was Actinopterygians.—Regeneration in acti- rounded, which is morphologically similar to nopterygians proceeds like urodele regener- bony callus formation in tetrapod fracture ation, involving a blastema formed of repair. Gene expression indicated that in the dedifferentiated cells (Kemp and Park 1970; crush model, repair was prolonged and prob- Knopf et al. 2011; Sousa et al. 2011). Zebrafish ably involved remodeling (Sousa et al. 2012). (Danio, osteocyte-bearing teleosts) can regener- Similarly, in zebrafish fin ray fracture, a callus ate their caudal and pectoral fins in response to (composed of collagen fibers) formed and fin ray amputation (Akimenko et al. 1995; then mineralized to form bony callus. The Sousa et al. 2011, 2012). They can also heal subsequent remodeling appears similar to

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 688 EVA C. HERBST ET AL.

remodeling of a fracture callus in humans (Peignoux-Deville et al. 1982, 1989; Bordat (Geurtzen et al. 2014). 1987). No healing of this bone or chondrichth- Moss (1962) investigated fracture repair of yan cartilage has been reported. dermal bones (opercular and lower jaw) in freshwater fish, the osteocytic Tilapia and the Aspidin Healing and Remodeling anosteocytic Carassius. Both species formed Extinct Taxa.— fracture calluses in the lower jaw, although Car- Heterostracans.—Aspidin healing has been assius appeared to show earlier and more abun- found in Psammolepis (Paleontological Museum dant bone formation in the fracture callus. of St. Petersburg University 46-1), which shows Carassius was able to calcify and ossify the frac- a projection with both spongy and lamellar ture callus even in acalcemic water, unlike aspidin, the former of which is much denser Tilapia. Moss (1962) also observed fracture cal- than the spongy layer found in uninjured dor- lus formation in the anosteocytic marine fish sal plates (Lebedev et al. 2009). Resorption of Fundulus, but the fracture callus was smaller aspidin and deposition of new aspidin has than in the freshwater fish examined and was been reported in Ganosteus (Tarlo 1964; Halstead mostly cartilaginous, with some calcification 1969, 1973). Furthermore, growth series of Pter- and some bone formation. Hematoma forma- aspis reveal that remodeling must have occurred tion was not as pronounced in the fracture during fusion of head and trunk plates; there is a repair of these teleost fishes as in tetrapods seamless continuation in the cancellous aspidin (Moss 1962); in mice, inflammation and hema- between some adult trunk plates, whereas toma formation plays an important role in the juvenile plates are separate and rimmed with healing process (Bahney et al. 2018). Although spongy aspidin (Halstead 1969). There is also these cartilage calluses in fish formed in dermal evidence of remodeling in Tesseraspis, Corvaspis, bone (which develops through intramembra- Phialaspis, Amphiaspis,andPsammosteus (Keat- nous rather than endochondral ossification), ing et al. 2015). they formed only upon damage to primary car- Extant Taxa.—Aspidin is not found in any tilage; therefore, the jaw callus cartilage is not extant taxa. Most teleosts have acellular bone “secondary cartilage.” The opercula did not (see section on Bone Healing, Remodeling, form cartilaginous calluses and did not heal and Regeneration: Extant Taxa), but this acellu- well (Moss 1962). larity is a derived evolutionary trait (Meunier The anosteocytic medaka (Oryzias) can heal and Huysseune 1991; Moss 1961). the caudal lepidotrichia, forming a fracture cal- lus that ossifies and remodels (Takeyama et al. Dentine Healing and Remodeling 2014). Osteoclasts are first induced to resorb Extinct Taxa.— bone fragments and then to remodel the callus Sarcopterygians (lobe-finned fish): porolepi- (Takeyama et al. 2014). Remodeling in response forms and lungfish.—The porolepiform Glypto- to increased load has been reported in the lepis from the Eifelian stage of the Middle opercula of tilapia, Oreochromis (Atkins et al. Devonian shows denticle healing during frac- 2015). In the anosteocytic amphibious fish ture repair of a bite injury on a scale (Lebedev Kryptolebias, gravitational loading increased et al. 2009). The lungfish Mioceratodus from gill arch stiffness, and proteomic analysis the mid-Tertiary shows a lesion without den- showed that the mechanisms were similar to tine healing or remodeling (although normal those of tetrapod bone remodeling (Turko tissue growth was observed near the lesion). et al. 2017). Kemp (2001) concluded that fossil lungfish Chondrichthyans.—The skeleton of extant were unable to remodel or repair dentine. chondrichthyans contains only small amounts Chondrichthyans and placoderms.—The of bone, for example, in the neural arches of acanthodian Doliodus (see “Materials and dogfish (Scyliorhinus). This bone shows evi- Methods” for information about chondrichth- dence of resorption by mononucleated cells, yan affinity) from the Early Devonian shows and mono- and multinucleated cells can resorb dentine healing of fin spine fractures without bone implanted in the dorsal musculature visible resorption (Burrow et al. 2017).

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 689

The Middle Devonian arthrodire placoderms Extant Taxa.— Eastmanosteus and Actinolepis show irregular Tetrapods.—Dentine in humans can repair in tubercle formation in response to injury on response to insults such as disease, chemicals, or their dermal plates (Lebedev et al. 2009). trauma by depositing tertiary dentine; this den- Heterostracans and osteostracans.—The old- tine is called either reparative or reactive dentine est clear case of cell-mediated healing of a based on the cell history of the odontoblasts mineralized tissue in a vertebrate is dentine involved in its secretion (Smith et al. 1995). healing on the dorsal shield of the heterostracan Molecules involved in dentinogenesis during Larnovaspis from the Lochkovian stage, Early development are re-expressed during tooth repair Devonian (Lebedev et al. 2009). Johanson (Mitsiadis and Rahiotis 2004). Remodeling of et al. (2013) found the majority of healing of a dentine in human teeth occurs both during devel- Psammolepis injury to be dentine infilling opment (via formation of secondary dentine) and (rather than aspidin healing), and argued that in pathological reactions (Mitsiadis et al. 2008). such dentine infilling to repair bone evolved Lungfish.—Dentine in the extant lungfish before bone repair. However, aspidin also Neoceratodus is unable to remodel during healed in heterostracans (see “Aspidin Healing growth. Partial dental ridge fracture in Proto- and Remodeling” section). Another specimen pterus showed that the damaged dentine was of Psammolepis shows healing of dentine by for- not remodeled or repaired, although new den- mation of unusual, small dentine tubercles (in tine grew near the fracture (Kemp 2001, 2002a). addition to the aspidin healing in the same spe- cimen, described in “Aspidin Healing and Conodont Mineralized Tissues Remodeling”) (Lebedev et al. 2009). The Mid- The phosphate-based dental elements of con- dle Devonian psammosteid Pycnosteus grew odonts from the Silurian were also capable of secondary dentine tubercles on a healing scale healing fractures and scratches (Shirley et al. (Lebedev et al. 2009). Another Psammosteus spe- 2018). This repair was systematic, but a patho- cies shows healing of a fracture via dentine logical specimen is also known, in which tubercle deposition (Halstead 1969). Keating abnormal growth and/or repair led to abnor- et al. (2018) reported remodeling in the superfi- mal morphology. No remodeling of the cial layer (composed on dentine) of the dorsal damaged areas occurred. shield of the heterostracan Psammosteus. Resorption of dentine tubercles followed by Phylogenetic Synthesis subsequent growth of new tubercles has been Here we consider the previously discussed observed in the heterostracan Ganosteus from observations in a phylogenetic context to recon- the Middle Devonian (Halstead 1973). Second- struct how bone healing evolved. No major ary dentine formation in a heterostracan variation in bone healing patterns of dermal resembled that of secondary dentine formed and endochondral bone is evident in extinct or in human teeth in response to caries, which extant tetrapods (Fig. 7), except for secondary indicates a conserved healing mechanism chondrogenesis in the healing of dermal bone (Tarlo 1964; Halstead 1973). in birds and mammals. However, the absence Lebedev et al. (2009) described several of secondary chondrogenesis during healing other cases of healing in heterostracans, for of dermal bones in nonavian reptiles, amphi- example, in the dorsal plates of several Psam- bians, and several teleost fish, and differences molepis species from the Givetian, Middle in the secondary cartilage of birds and mam- Devonian. Both injuries damaged the aspidin mals, indicate that this process evolved conver- layer in the shield, which subsequently gently in birds and mammals (Irwin and healed, but the identification of the healed tis- Ferguson 1986;Hall2005). Secondary cartilage sue was not reported; it could be either aspi- also forms during the development of the max- dinordentine.Denison(1952) reported illa, dentary, and cleithrum (after intramembra- remodeling of dentine tubercles (resorption nous ossification) in the teleost Poecilia sphenops and subsequent replacement) in the osteostra- (Benjamin 1989). This indicates a third inde- can Cephalaspis. pendent origin of secondary chondrogenesis

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 690 EVA C. HERBST ET AL.

FIGURE 7. Distribution of bone healing and regeneration in sarcopterygians, showing key data discussed in “Bone Healing, Remodeling, and Regeneration,” which establish that these fundamental healing mechanisms are ancestral for the clade. See “Materials and Methods” for detail on phylogeny.

(Hall 2005). However, to our knowledge, sec- inflammation pathways play a larger role in ondary chondrogenesis during dermal fracture fracture repair of tetrapods than in fracture repair has only been reported in birds and repair of osteocytic freshwater and anosteocytic mammals, and its absence has been reported marine and freshwater teleost fish. It is unclear in other teleost fish such as Fundulus (Moss whether this difference in pathways indicates 1962). It would be interesting to investigate a convergent evolution of specific healing pro- whether dermal bone fracture repair in Peocilia cesses in actinopterygian and sarcopterygian also involves secondary chondrogenesis. bone or modifications to an ancestral healing Fossil porolepiform and rhizodontid fish pathway. However, several other similarities show evidence of dermal bone healing, and fos- of the fracture repair pathways suggest the lat- sil lungfish show endochondral bone healing ter. The formation of a fracture callus occurs in with calluses like those found in fracture repair sarcopterygians as well as actinopterygians, of extinct and extant tetrapods. Extant lungfish and in most cases this callus is subsequently form and remodel a callus in response to der- remodeled. In both medaka (Oryzias) lepidotri- mal bone injury. Therefore, we infer that the chial fracture and mammalian long bone frac- same fundamental mechanisms of healing of tures, two stages of osteoclast activity have dermal and endochondral bone were ancestral been reported. The second stage of osteoclast not only for tetrapods but for Sarcopterygii. activity remodels the bony callus in both groups, There are interesting differences in sarcoptery- and blocking the induction of these osteoclasts gian and actinopterygian fracture repair path- prevents callus remodeling (Schindeler et al. ways. For example, zebrafish dedifferentiate 2008; Takeyama et al. 2014). cells to form a blastema during both fracture The presence of a cartilaginous versus a repair and regeneration, but tetrapods only dedif- fibrous callus stage varies within actinoptery- ferentiate cells during regeneration. Additionally, gians and sarcopterygians, depending on

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 691

FIGURE 8. Cell-mediated healing and remodeling of mineralized tissue in vertebrates, indicating the ancestral nature of skeletal repair capacity. Certain species have been added as examples, but the fossil record extends beyond these examples. See “Materials and Methods” for detail on phylogeny. Artwork by Nobu Tamura, CC BY-SA 3.0 license, http://spinops. blogspot.com, https://creativecommons.org/licenses/by-sa/3.0.

bone type and the specific taxon. However, in perhaps all osteichthyans (Fröbisch et al. 2015; all taxa and both dermal and endochondral Nogueira et al. 2016; van der Vos et al. 2018; bones, fracture of the bone at the organ level Lu et al. 2019). However, there does not appear results in callus formation. Bone healing to be a direct link between healing and regener- evolved early in vertebrates, as is evident in ation processes; regeneration is not an exagger- heterostracans (aspidin healing) and placo- ated form of bone healing, as evidenced by derms (cellular bone healing). A recent evolu- differences in pathways and the inability to tionary analysis suggested that aspidin was heal a fracture gap in both regenerating and the ancestral bone type, with cellular bone nonregenerating taxa (Roy and Lévesque 2006). evolving from it at least two times in vertebrate We then considered how the healing of evolution (Keating et al. 2018). Cellular bone mineralized tissues evolved within vertebrates. first evolved in osteostracans, and its presence The evolutionary homologies of these tissues in placoderms, acanthodians, and osteichth- remain contentious, but our synthesis (Fig. 8) yans suggests that cellular bone is a synapo- infers that the capacity for cell-mediated repair morphy of gnathostomes and osteostracans of skeletal tissues was ancestral for vertebrates. (Brazeau and Friedman 2014; Davesne et al. Bone remodeling is widespread in extinct and 2019). Because a healing mechanism is found extant vertebrates, and bone remodeling in both aspidin and cellular bone, and cellular mechanisms are similar in extant tetrapods bone evolved from aspidin, it is most parsimo- and anosteocytic fish (Turko et al. 2017), nious that the healing mechanism evolved in which supports our interpretation that bone early vertebrate bone (regardless of its cellular- remodeling is also ancestral for vertebrates. ity). This, taken together with the widespread Furthermore, we found that remodeling distribution of bone healing in extant verte- co-occurred with healing in bone (including brates, leads us to conclude that dermal bone aspidin) and dentine in the same taxa (hetero- healing evolved early on in vertebrates, rather stracans, placoderms) early on in vertebrate than convergently evolving in all these taxa. evolution. Not only are the same taxa capable Based on evidence from fossil and extant ani- of remodeling and healing, but remodeling mals, bone regeneration in the appendages and has been reported as a part of the healing pro- tail is also ancestral for sarcopterygians and cess, for example, in Dunkleosteus (Capasso

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 692 EVA C. HERBST ET AL.

et al. 1996). Although healing has not been early tetrapods (Fröbisch et al. 2015; Nogueira reported for osteostracans, they were able to et al. 2016; van der Vos et al. 2018; Lu et al. resorb bone, and based on their phylogenetic 2019), anatomical analysis of growth series of placement, we conclude that the absence of early tetrapods suggests that lissamphibian- healing evidence may be due to lack of preser- like water to land metamorphosis was not vation. Conodonts (an outgroup to vertebrates) ancestral for early tetrapods, and evolved in a were able to heal their dental elements, but this lissamphiban ancestor (Schoch 2001). Branchio- healing does not show evidence of remodeling saurids (temnospondyls) from the late Carbon- (Shirley et al. 2018). The convergent evolution iferous show the earliest reported evidence of of mineralized tissues in conodonts provides lissamphibian-like metamorphosis (Schoch and an example of a healing mechanism with for- Fröbisch 2006). On the other hand, although mation only. no fossil evidence exists, it is possible that early tetrapods metamorphosed in water (Schoch 2001). Alibardi’s(2018)argumentis Discussion that genetic pathways important for metamor- The paleopathologies we described in Eoher- phosis also enable regeneration, and if meta- peton and Crassigyrinus are some of the earliest morphosis is lost, then these genes are also reported cases of bone healing for tetrapods. lost, preventing regeneration. If a fish-like meta- Our evolutionary synthesis led us to discover morphosis between larval and adult stages is that, like regeneration, bone healing is ancestral discovered in early tetrapods, this would recon- for tetrapods and probably for sarcopterygians. cile Alibardi’s(2018) argument of the link However, the pathways of these two repair pro- between metamorphosis and regeneration cesses (bone healing and regeneration) are dis- with fossil evidence (Schoch 2001;Schochand tinct, and most sarcopterygians have lost Fröbisch 2006) that lissamphibian-like meta- regeneration capacities, whereas bone healing morphosis evolved after regeneration evolved. is ubiquitous among all sarcopterygians stud- Another hypothesis concerns the role of dev- ied. Furthermore, our phylogenetic synthesis elopmental pathways in regeneration; in of vertebrates suggests that healing is ancestral amniotes, development is more dependent on for vertebrate bone and other early vertebrate transient structures such as somites, and as mineralized tissues. Bone, aspidin (acellular their existence is limited to early embryo stages, bone), and dentine were all able to undergo this might prevent development of new struc- both healing and remodeling early in verte- tures during the regeneration process. Amphi- brate evolution. This evidence led us to infer bians, on the other hand, develop limbs in a that there is an evolutionary and mechanistic more “self-organized” manner (Galis et al. 2003). link between healing and remodeling capacity. There does not seem to be a link between It is interesting to note that, although both regeneration and healing, but can the fossil regeneration and repair could be advantageous record reveal anything more about whether to survival, most extant sarcopterygians cannot remodeling and healing are linked? Minera- regenerate, but all sarcopterygians seem to lized tissues can heal by formation only have retained the ability to heal bone. There (i.e., without resorption), but healing is usually are several hypotheses about the evolution, followed by remodeling in extant vertebrates. maintenance, and loss of regeneration. For In bone (including aspidin) and dentine, there example, some taxa might have reduced likeli- is evidence of both remodeling and healing hood of structure loss, and therefore regener- emerging in the same taxa early on in verte- ation might have been lost because it was brate evolution. Ashhurst (2004) suggested neutral or costly to maintain (Bely and Nyberg that bone’s capacity to remodel and integrate 2010). Alibardi (2018) suggested that loss of old and new matrix might explain why bone regeneration is linked to the loss of metamor- can heal well and shark cartilage (which does phosis during development. However, this cor- not remodel) cannot. The connection between relation is not supported by the fossil record of frequent remodeling and full healing ability tetrapods. While regeneration was ancestral for appears to exist not only for bone and cartilage,

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 693

but for other vertebrate skeletal tissues. remodeling and healing evolved early on and Our phylogenetic analysis reveals that dentine have been retained in extant animals, suggest- and aspidin (acellular bone) both have a long ing that the benefits of these two processes out- evolutionary history of remodeling and healing weigh the metabolic and mechanical costs. capabilities (evident in early fossil vertebrates), Healing is advantageous, because it restores and enamel has neither. the strength of bone through modeling. Injuries Remodeling did not necessarily evolve as a and insults from biological (intraspecific, part of the healing pathway. Indeed, there has predator–prey, disease) and environmental been much discussion about whether bone interactions (fatigue damage, trauma) would remodeling initially evolved for homeostasis, prove detrimental to extinct and extant verte- growth, and/or removal of microdamage (Mar- brates without healing. Healing through blastic tin 2003; Doherty et al. 2015; Currey et al. 2017). formation only is possible; such a repair mech- It is debated whether heterostracan dentine anism was present in mineralized tissues of remodeling (via tubercle turnover) evolved as conodonts (Shirley et al. 2018). However, it is a healing or growth mechanism (Halstead only through remodeling that healing without 1969). It is difficult to test why remodeling mass gain (and associated metabolic costs) is evolved, and there may have been a combin- possible (Martin 2003). Although the ultimate ation of selective pressures. Regardless, our evo- evolutionary cause (sensu Mayr 1961) for the lutionary synthesis reveals that there is a link evolution of remodeling remains unclear, our between healing and remodeling processes. results elucidate the timing and consequences Future studies comparing cell interactions in of the evolution of remodeling in relation to bone, dentine, cartilage, and enamel may give the evolution of healing. insight into the specific tissue properties that enable active remodeling and healing. Perhaps Conclusions exceptions in a tissue’s ability to heal and remodel can be traced to differences in the cell Both the aquatic Crassigyrinus and the more and matrix interactions in the tissues of these terrestrial Eoherpeton were able to heal bone, specific taxa. Such exceptions include the showing healing in response to injury or disease increased capacity of healing of specific types in early tetrapods with a range of locomotor of cartilage, such as the auricular cartilage in capacities and habitats. Bone healing is found rabbits (Hall 2015), and the inability of lungfish in extant and extinct Osteichthyes, as well as pla- to remodel or heal their dental plates. Indeed, coderms (early gnathostome vertebrates). Aspi- lungfish tooth plates differ from teeth of most din (acellular bone) healing is found in other vertebrates, because they continuously heterostracans (early agnathan vertebrates). grow and have denteons, so perhaps there is Dentine healing is also found in heterostracans some physiological link between these two and placoderms. Based on our analysis of extinct traits and the inability to remodel or repair. and extant taxa, we conclude that ancestral ver- On the other hand, lungfish might have lost tebrate skeletons were able to heal after insult the ability to remodel and heal dentine because and injury. Our analysis reveals an early, con- this trait was no longer adaptive, and continu- current evolutionary origin of both healing ous growth enabled sufficient restoration of and remodeling of dentine and bone (including function after injury. In any case, the inability aspidin). This shows that early vertebrates were of lungfish to remodel and heal dentine further not only capable of healing but could recover supports a link between these processes. fully by reaping the benefits of subsequent From an evolutionary standpoint, both heal- remodeling (which restores mature bone tissue ing and remodeling have metabolic costs. structure and strength) in these tissues. Remodeling also has a mechanical cost, because there is a period of structural weakness Acknowledgments when resorption has occurred but deposition of new bone has not yet re-established bone mass We thank S. Chapman (Natural History and strength (Felder et al. 2017). However, both Museum London) and S. Walsh (National

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 694 EVA C. HERBST ET AL.

Museums Scotland) for specimen access, Bentolila, V., T. M. Boyce, D. P. Fyhrie, R. Drumb, T. M. Skerry, and M. B. Schaffler. 1998. Intracortical remodeling in adult rat long K. Smithson (University Museum of Zoology, bones after fatigue loading. Bone 23:275–281. Cambridge) for scanning of specimens, Biewener, A. A., S. M. Swartz, and J. E. A. Bertram. 1986. Bone P. Bishop for discussions, and the Queensland modeling during growth: dynamic strain equilibrium in the chick tibiotarsus. Calcified Tissue International 39:390–395. Museum, Brisbane, and P. Bishop for Ossinodus Bishop, P. J., C. W. Walmsley, M. J. Phillips, M. R. Quayle, photographs. We thank two anonymous C. A. Boisvert, and C. R. McHenry. 2015. Oldest pathology in a reviewers for their constructive feedback, Tetrapod bone illuminates the origin of terrestrial vertebrates. PLoS ONE 10:e0125723. which greatly improved this article. We thank Bordat, C. 1987. Étude Ultrastructurale de l’os des Vertèbres du the Natural Environment Research Council Sélacien Scyliorhinus canicula L. Canadian Journal of Zoology for supportive funding under grant numbers 65:1435–1444. Boyde, A. 2003. The real response of bone to exercise. Journal of NE/K004751/1 (J.R.H.) and NE/J022713/1 Anatomy 203:173–189. (J.A.C., T.R.S.) and the Wellcome Trust for Brazeau, M. D. 2009. The braincase and jaws of a Devonian “Acan- funding under grant number 093234 (X-ray thodian” and modern gnathostome origins. Nature 457:305–308. Brazeau, M. D., and V. de Winter. 2015. The hyoid arch and brain- microtomography scanner at the Royal Veterin- case anatomy of acanthodes support chondrichthyan affinity of ary College). Artwork by N. Tamura, CC “Acanthodians.” Proceedings of the Royal Society of London B BY-SA 3.0 license, http://spinops.blogspot. 282:20152210. Brazeau, M. D., and M. Friedman. 2014. The characters of Palaeo- com, https://creativecommons.org/licenses/ zoic jawed vertebrates. Zoological Journal of the Linnean Society by-sa/3.0. 170:779–821. Browne, M. A. E., M. T. Dean, I. H. S. Hall, A. D. McAdam, Monro., S. K., and J. I. Chisholm. 1999. A lithostratigraphical Literature Cited framework for the Carboniferous rocks of the Midland Valley Akimenko, M. A., S. L. Johnson, M. Westerfield, and M. Ekker. of Scotland. British Geological Survey Research Report RR/99/ 1995. Differential induction of four Msx homeobox genes during 07. British Geological Survey, Keyworth, Nottingham, U.K. fin development and regeneration in zebrafish. Development Burr, D. B. 1993. Remodeling and the repair of fatigue damage. 121:347–357. Calcified Tissue International 53(Suppl. 1):S75–S81. Akimenko, M., M. Marí-Beffa, J. Becerra, and J. Géraudie. 2003. Old Burrow, C. J, S. Turner, J. G. Maisey, S. Desbiens, and R. F Miller. questions, new tools, and some answers to the mystery of fin 2017. Spines of the stem chondrichthyan Doliodus latispinosus. regeneration. Developmental Dynamics 226:190–201. Canadian Journal of Earth Sciences 54:1248–1262. Akimenko, M., and A. Smith. 2007. Paired fin repair and regener- Capasso, L. L., F. Bacchia, N. Rabottini, B. M. Rothschild, and ation. Pp. 152–162 in B. K. Hall, ed. Fins into limbs: evolution, R. Mariani-Costantini. 1996. Fossil evidence of intraspecific development and transformation. University of Chicago Press, aggressive behaviour of Devonian giant fishes (Arthrodira, Chicago. Dinichthyidae). Journal of Paleopathology 8:153–160. Alibardi, L. 2018. Organ regeneration evolved in fish and amphi- Chiari, Y., V Cahais, N. Galtier, and F. Delsuc. 2012. Phylogenomic bians in relation to metamorphosis: speculations on a post- analyses support the position of turtles as the sister group of birds embryonic developmental process lost in amniotes after the and crocodiles (Archosauria). BMC Biology 10:1–14. water to land transition. Annals of Anatomy/Anatomischer Clack, J. A., C. E. Bennett, D. K. Carpenter, S. J. Davies, N. C. Fraser, Anzeiger 222:114–119. T. I. Kearsey, J. E. A. Marshall, D. Millward, B. K. A. Otoo, E. Anné, J., N. P. Edwards, R. A. Wogelius, A. R. Tumarkin-Deratzian, J. Reeves, A. J. Ross, M. Ruta, K. Z. Smithson, T. R. Smithson, S. W. I. Sellers, A. van Veelen, U. Bergmann, D. Sokaras, A. Walsh. 2016. Phylogenetic and environmental context of a R. Alonso-Mori, K. Ignatyev, V. M. Egerton, P. L. Manning. Tournaisian tetrapod fauna. Nature Ecology and Evolution 2014. Synchrotron imaging reveals bone healing and remodel- 1:0002. ling strategies in extinct and extant vertebrates. Journal of the Conant, E. B. 1970. Regeneration in the African lungfish, Proto- Royal Society Interface 11:20140277. pterus. Journal of Experimental Zoology 174:15–32. Ashhurst, D. E. 2004. The cartilaginous skeleton of an elasmo- Currey, J. D., M. N. Dean, and R. Shahar. 2017. Revisiting the links branch fish does not heal. Matrix Biology 23:15–22. between bone remodelling and osteocytes: insights from across Atkins, A., J. Milgram, S. Weiner, and R. Shahar. 2015. The phyla. Biological Reviews 92:1702–1719. response of anosteocytic bone to controlled loading. Journal of Davesne, D., F. J. Meunier, A. D. Schmitt, M. Friedman, O. Otero, Experimental Biology 218:3559–3569. and R. B. J. Benson. 2019. The phylogenetic origin and evolution Bahney, C. S., R. L. Zondervan, P. Allison, A. Theologis, J. Ashley, of acellular bone in teleost fishes: insights into osteocyte function J. Ahn, T. Miclau, R. Marcucio, and K. D. Hankenson. 2018. in bone metabolism. Biological Reviews 94:1338–1363. The cellular biology of fracture healing. Journal of Orthopaedic Denison, R. H. 1952. Early Devonian fishes from Utah, Part 1. Research 37:35–50. Osteostraci. Fieldiana (Geology), 11:265–287. Barkas, T. P. 1873. Illustrated guide to the fish, amphibian, reptil- De Smet, W. M.A. 1977. The fate of old bottle-nosed dolphins, ian, and supposed mammalian remains of the Northumberland Tursiops truncatus, in nature as revealed by the condition of Carboniferous Strata. London: W.M. Hutchings. their skeletons. Aquatic Mammals 5:78–86. Bely, A. E., and K.G. Nyberg. 2010. Evolution of animal regener- Doherty, A. H., C. K. Ghalambor, and S. W. Donahue. 2015. Evolu- ation: re-emergence of a field. Trends in Ecology and Evolution tionary physiology of bone: bone metabolism in changing 25:161–170. environments. Physiology 30:17–29. Benjamin, M. 1989. The development of hyaline-cell cartilage in the Donoghue, P. C. J., and I. J. Sansom. 2002. Origin and early head of the black molly, Poecilia sphenops. Evidence for secondary evolution of vertebrate skeletonization. Microscopy Research cartilage in a teleost. Journal of Anatomy 164:145–154. and Technique 59:352–372.

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 695

Donoghue, P. C. J., I. J. Sansom, and J. P. Downs. 2006. Early evolu- Hall, B. K, and J. Hanken. 1985. Repair of fractures jaws in the spot- tion of vertebrate skeletal tissues and cellular interactions, and the ted salamander: do amphibians form secondary cartilage? Jour- canalization of skeletal development. Journal of Experimental nal of Experimental Zoology 233:359–368. Biology (Molecular and Developmental Evolution) 306B:278–294. Hall, B. K., and E. P. Witten. 2007. Plasticity of and transitions Downs, J. P., and P. C. J. Donoghue. 2009. Skeletal histology of between skeletal tissues in vertebrate evolution and develop- Bothriolepis canadensis (Placodermi, Antiarchi) and evolution of ment. Pp. 13–56 in J. S. Anderson and H.-D. Sues, eds. Major tran- the skeleton at the origin of jawed vertebrates. Journal of Morph- sitions in vertebrate evolution. Indiana University Press, ology 270:1364–1380. Bloomington. Egawa, S., S. Miura, H. Yokoyama, T. Endo, and K. Tamura. 2014. Halstead, L. B. 1969. Calcified tissues in the earliest vertebrates. Growth and differentiation of a long bone in limb development, Calcified Tissue Research 3:107–124. repair and regeneration. Development, Growth and Differenti- Halstead, L. B. 1973. The heterostracan fishes. Biological Reviews ation 56:410–424. 48:279–332. Felder, A. A., C. Phillips, H. Cornish, M. Cooke, J. R. Hutchinson, and Harris, S. 1978. Injuries to foxes (Vulpes vulpes) living in suburban M. Doube. 2017. Secondary osteons scale allometrically in mam- london. Journal of Zoology 186:567–572. malian humerus and femur. Royal Society Open Science 4:170431. Helfrich, M. H. 2003. Osteoclast diseases. Microscopy Research and Fröbisch, N. B., C. Bickelmann, and F. Witzmann. 2014. Early evo- Technique 61:514–532. lution of limb regeneration in tetrapods: evidence from a Herbst, E. C., and J. R. Hutchinson. 2018. New insights into the 300-million-year-old amphibian. Proceedings of the Royal Society morphology of the Carboniferous tetrapod Crassigyrinus scoticus of London B 281:20141550. from computed tomography. Earth and Environmental Science Fröbisch, N. B., C. Bickelmann, J. C. Olori, and F. Witzmann. 2015. Transactions of the Royal Society of Edinburgh 109:157–175. Deep-time evolution of regeneration and preaxial polarity in Herman, B. C., L. Cardoso, R. J. Majeska, K. J. Jepsen, and M. tetrapod limb development. Nature 527:231–234. B. Schaffler. 2010. Activation of bone remodeling after fatigue: Frost, H. M. 1991. Some ABC’s of skeletal pathophysiology. 6. The differential response to linear microcracks and diffuse damage. growth/modeling/remodeling distinction. Calcified Tissue Bone 47:766–772. International 49:301–302. Hirasawa, T., and S. Kuratani. 2015. Evolution of the vertebrate Galis, F., G. P. Wagner, and E. L. Jockusch. 2003. Why is limb regen- skeleton: morphology, embryology, and development. Zoo- eration possible in amphibians but not in reptiles, birds, and logical Letters 1:1–17. mammals? Evolution and Development 5:208–220. Hutchison, C., M. Pilote, and S. Roy. 2007. The axolotl limb: a Garvey, J. M., Z. Johanson, and A. Warren. 2005. Redescription of model for bone development, regeneration and fracture healing. the pectoral fin and vertebral column of the rhizodontid fish Bar- Bone 40:45–56. ameda decipiens from the lower Carboniferous of Australia. Jour- Irwin, C. R., and M. W. J. Ferguson. 1986. Fracture repair of reptil- nal of Vertebrate Paleontology 25:8–18. ian dermal bones: can reptiles form secondary cartilage? Journal Gerstenfeld, L. C., D. M. Cullinane, G. L. Barnes, D. T. Graves, and T. of Anatomy 146:53–64. A. Einhorn. 2003. Fracture healing as a post-natal developmental Jackson, D. W., P. A. Lalor, H. M. Aberman, and T. M. Simon. 2001. process: molecular, spatial, and temporal aspects of its regulation. Spontaneous repair of full-thickness defects of articular cartilage Journal of Cellular Biochemistry 88:873–884. in a goat model. Journal of Bone and Joint Surgery 83A:53–64. Gess, R. W., M. I. Coates, and B. S. Rubidge. 2006. A lamprey from Jacobs, M. H. 1955. The traumatic bone cyst. Oral Surgery, Oral the Devonian period of South Africa. Nature 443:981–984. Medicine, Oral Pathology 8:940–949. Geurtzen, K., F. Knopf, D. Wehner, L. F. A. Huitema, Jacyniak, K., R. P. McDonald, and M. K. Vickaryous. 2017. Tail regen- S. Schulte-Merker, and G. Weidinger. 2014. Mature osteoblasts eration and other phenomena of wound healing and tissue restor- dedifferentiate in response to traumatic bone injury in the zebra- ation in lizards. Journal of Experimental Biology 220:2858–2869. fish fin and skull. Development 141:2225–2234. Janvier, P., M. Arsenault, and S. Desbiens. 2004. Calcified cartilage Giles, S., M. Rücklin, and P. C. J. Donoghue. 2013. Histology of in the paired fins of the osteostracan Escuminaspis laticeps (Tra- “placoderm” dermal skeletons: implications for the nature of quair 1880), from the Late Devonian of Miguasha (Québec, Can- the ancestral gnathostome. Journal of Morphology 274:627–644. ada), with a consideration of the early evolution of the pectoral fin Giles, S. M., Friedman, and M. D. Brazeau. 2015. Osteichthyan-like endoskeleton in vertebrates. Journal of Vertebrate Paleontology cranial conditions in an Early Devonian stem gnathostome. 24:773–779. Nature 520:82–85. Jenkins, M., S. Nimphius, N. H. Hart, P. Chivers, T. Rantalainen, Godfrey, S. J. 1988. Isolated tetrapod remains from the Carbonifer- K. Rueter, M. L. Borland, F. McIntyre, K. Stannage, and ous of West Virginia. Kirtlandia 43:27–36. A. Siafarikas. 2018. Appendicular fracture epidemiology of chil- Godfrey, S. J. 1989. The postcranial skeletal anatomy of the dren and adolescents: a 10-year case review in Western Australia Carboniferous tetrapod Greererpeton burkemorani Romer, (2005 to 2015). Archives of Osteoporosis 13(63):1–10. 1969. Philosophical Transactions of the Royal Society of Lon- Jiang, X., S. Iseki, R. E. Maxson, H. M. Sucov, and don B 323:75–133. G. M. Morriss-Kay. 2002. Tissue origins and interactions in the Goodman, A. H., and J. C. Rose. 1994. Assessment of systematic mammalian skull vault. Developmental Biology 241:106–116. physiological perturbations from dental enamel hypoplasias Johanson, Z., M. Smith, A. Kearsley, P. Pilecki, E. Mark-Kurik, and and associated histological structures. Yearbook of Phyiscal C. Howard. 2013. Origins of bone repair in the armour of fossil Anthropology 33:59–110. fish: response to a deep wound by cells depositing dentine Goss, R. J., and M. W. Stagg. 1958. Regeneration of lower jaws in instead of dermal bone. Biology Letters 9:20130144. adult newts. Journal of Morphology 102:298–309. Johansson, P. K. E., T. G. Douglass, and C. G. Lowe. 2004. Caudal Hall, B. K. 2005. Bones and cartilage, 1st ed. Elsevier Academic spine replacement and histogenesis in the round stingray, Uroba- Press, San Diego. tis halleri. Bulletin of the Southern California Academy of Sciences Hall, B. K. 2014. Endoskeleton/exo (dermal) skeleton—meso- 103:115–124. derm/neural crest: two pair of problems and a shifting paradigm. Keating, J. N., C. L. Marquart, and P. C. J. Donoghue. 2015. Hist- Journal of Applied Ichthyology 30:608–615. ology of the heterostracan dermal skeleton: insight into the origin Hall, B. K. 2015. Bones and cartilage: developmental and evolution- of the vertebrate mineralised skeleton. Journal of Morphology ary skeletal biology, 2nd ed. Elsevier, London. 276:657–680.

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 696 EVA C. HERBST ET AL.

Keating, J. N., C. L. Marquart, F. Marone, and P. C. J. Donoghue. Mori, S., and D. B. Burr. 1993. Increased intracortical remodeling 2018. The nature of aspidin and the evolutionary origin of following fatigue damage. Bone 14:103–109. bone. Nature Ecology and Evolution 2:1501–1506. Moss, M. L. 1961. Studies of the acellular bone of teleost Kemp, A. 2001. Consequences of traumatic injury in fossil and fish. I. Morphological and systematic variations. Acta Anatomica recent dipnoan dentitions. Journal of Vertebrate Paleontology 46:343–362. 21(1):13–23. Moss, M. L. 1962. Studies of the acellular bone of teleost fish. II. Kemp, A. 2002a. Growth and hard tissue remodelling in the denti- Response to fracture under normal and alcemic conditions. tion of the Australian lungfish, Neoceratodus forsteri (Osteichthyes: Acta Anatomica 48:46–60. Dipnoi). Journal of Zoology 257:219–235. Murdock, D. J. E., X. P. Dong, J. E. Repetski, F. Marone, Kemp, A. 2002b. Hyaline tissue of thermally unaltered conodont M. Stampanoni, and P. C. J. Donoghue. 2013. The origin of cono- elements and the enamel of vertebrates. Alcheringa 26:23–36. donts and of vertebrate mineralized skeletons. Nature 502:546–549. Kemp, N. E., and J. H. Park. 1970. Regeneration of Lepidotrichia Nogueira, A. F., C. M. Costa, J. Lorena, R. N. Moreira, and Actinotrichia in the tailfin of the teleost Tilapia mossambica. G. N. Frota-Lima, C. Furtado, M. Robinson, C. T. Amemiya, Developmental Biology 22:321–342. S. Darnet, and I. Schneider. 2016. Tetrapod limb and sarcoptery- Khurana, J., F. Abdul-Karim, and J. V. M. G. Bovee. 2002. Osteo- gian fin regeneration share a core genetic programme. Nature chondroma. Pp. 234–236 in C. D. M. Fletcher, K. K. Unni, and Communications 7:13364. F. Mertens, eds. World Health Organisation classification of Nunamaker, D. M. 2002. On bucked shins. AAEP Proceedings tumours: pathology and genetics of tumours of soft tissue and 48:76–89. bone. IARC Press, Lyon. Panchen, A L. 1985. On the Amphibian Crassigyrinus scoticus Wat- Kirschbaum, F., and F. J. Meunier. 1981. Experimental regeneration son from the Carboniferous of Scotland. Philosophical Transac- of the caudal skeleton of the glass knifefish, Eigenmannia virescens tions of the Royal Society of London B 309:505–68. (Rhamphichthydae, Gymnotoidei). Journal of Morphology Panchen, A. L., and T. R. Smithson. 1990. The pelvic girdle and 168:121–135. hind limb of Crassigyrinus scoticus (Lydekker) from the Scottish Knopf, F., C. Hammond, A. Chekuru, T.s Kurth, S. Hans, Carboniferous and the origin of the tetrapod pelvic skeleton. C. W. Weber, G. Mahatma, et al. 2011. Bone regenerates via dedif- Transactions of the Royal Society of Edinburgh (Earth Sciences) ferentiation of osteoblasts in the zebrafish fin. Developmental Cell 81:31–44. 20:713–724. Pardo-Pérez, J. M., B. P. Kear, M. Gómez, M. Moroni, and E. Kolmann, M. A., P. Urban, and A. P. Summers. 2018. Structure and E. Maxwell. 2018. Ichthyosaurian palaeopathology: evidence of function of the armored keel in piranhas, pacus, and their allies. injury and disease in fossil “fish lizards.” Journal of Zoology Anatomical Record. doi: 10.1002/ar.23986. 304:21–33. Lebedev, O. A., E. Mark-Kurik, V. N. Karatajūtė-Talimaa, Peignoux-Deville, J., C. Bordat, and B. Vidal. 1989. Demonstration E. Luksevics, and A. Ivanov. 2009. Bite marks as evidence of pre- of bone resorbing cells in elasmobranchs: comparison with osteo- dation in early vertebrates. Acta Zoologica 90:344–356. clasts. Tissue and Cell 21:925–933. Lingham-Soliar, T. 2004. Palaeopathology and injury in the extinct Peignoux-Deville, J., F. Lallier, and B. Vidal. 1982. Evidence for the mosasaurs (Lepidosauromorpha, Squamata) and implications for presence of osseous tissue in dogfish vertebrae. Cell and Tissue modern reptiles. Lethaia 37:255–262. Research 222:605–614. Lu, S., L. Yang, H. Jiang, J. Chen, G. Yu, Z. Chen, X. Xia, and S. He. Pineda, C., R. Espinosa, and A. Pena. 2009. Radiographic imaging 2019. Bichirs employ similar genetic pathways for limb regener- in osteomyelitis: the role of plain radiography, computed tomog- ation as are used in lungfish and salamanders. Gene 690:68–74. raphy, ultrasonography, magnetic resonance imaging, and scin- Maas, M. C., and E. R. Dumont. 1999. Built to last: the structure, tigraphy. Seminars in Plastic Surgery 23:80–89. function, and evolution of primate dental enamel. Evolutionary Pyron, R. A., and J. J. Wiens. 2011. A large-scale phylogeny of Anthropology 8:133–152. Amphibia including over 2800 species, and a revised classifica- Marsell, R., and A. Einhorn. 2011. The biology of fracture healing. tion of extant frogs, salamanders, and caecilians. Molecular Injury 42:551–555. Phylogenetics and Evolution 61:543–583. Martin, R. B. 2003. Fatigue microdamage as an essential element of Roth, A. J., G. S. Jones, and T. W. French. 2002. Incidence of bone mechanics and biology. Calcified Tissue International naturally-healed fractures in the pectoral bones of North Ameri- 73:101–107. can accipiters. Journal of Raptor Research 36:229–230. Mayr, E. 1961. Cause and effect in biology. Science 134:1501–1506. Rothschild, B. M., and L. D. Martin. 2006. Skeletal impact of disease. Meunier, F. J., and A. Huysseune. 1991. The concept of bone tissue New Mexico Museum of Natural History and Science Bulletin 33. in Osteichthyes. Netherlands Journal of Zoology 42:445–458. Roy, S., and M. Lévesque. 2006. Limb regeneration in axolotl: is it Micklich, N., and S. Mentges. 2012. Fin ray fractures in messel superhealing? Scientific World Journal 6(S1):12–25. fishes. Kaupia: Darmstädter Beiträge Zur Naturgeschichte Ruta, M., M. I. Coates, and D. L. J. Quicke. 2003. Early tetrapod rela- 18:19–27. tionships revisited. Biological Reviews 78:251–345. Mitsiadis, T. A., C. De Bari, and I. About. 2008. Apoptosis in devel- Sansom, I. J., M. P. Smith, H. Armstrong, and M. M. Smith. 1992. opmental and repair-related human tooth remodeling: a view Presence of the earliest vertebrate hard tissue in conodonts. Sci- from the inside. Experimental Cell Research 314:869–877. ence 256:1308–1311. Mitsiadis, T. A., and C. Rahiotis. 2004. Parallels between tooth Schindeler, A., McDonald, M. M., Bokko, P., and Little, D. G. 2008. development and repair: conserved molecular mechanisms fol- Bone remodeling during fracture repair: the cellular picture. lowing carious and dental injury. Journal of Dental Research Seminars in Cell and Developmental Biology 19:459–466. 83:896–902. Schneider, C. A., W. S. Rasband, and K. Eliceiri. 2012. NIH Image to Molnar, J. L., Rui, D., J. R. Hutchinson, and S. E. Pierce. 2018. Evo- ImageJ: 25 years of image analysis. Nature Methods 9:671–675. lution of hindlimb muscle anatomy across the tetrapod Schoch, R. R. 2001. Can metamorphosis be recognized in Palaeo- water-to-land transition, including comparisons with forelimb zoic amphibians? Neues Jahrbuch für Geologie und Paläontolo- anatomy. Anatomical Record. doi: 10.1002/ar.23997. gie–Abhandlungen 220:335–367. Moodie, R. L. 1917. Studies in paleopathology. I. General consid- Schoch, R. R., and N. B. Fröbisch. 2006. Metamorphosis and neot- erations of the evidences of pathological conditions found eny: alternative pathways in an extinct amphibian clade. Evolu- among fossil animals. Annals of Medical History 1:374–393. tion 60:1467–1475.

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021 EVOLUTION OF SKELETAL REPAIR 697

Seref-Ferlengez, Z., J. Basta-Pljakic, O. D. Kennedy, C. J. Philemon, Takeyama, K., M. Chatani, Y. Takano, and A. Kudo. 2014. In-vivo and M. B. Schaffler. 2014. Structural and mechanical repair of dif- imaging of the fracture healing in medaka revealed two types fuse damage in cortical bone in vivo. Journal of Bone and Mineral of osteoclasts before and after the callus formation by osteoblasts. Research 29:2537–2544. Developmental Biology 394:292–304. Shapiro, F. 2008. Bone development and its relation to fracture Tanke, D. H., and P. J. Currie. 1998. Head-biting behavior in repair. The role of mesenchymal osteoblasts and surface osteo- theropod dinosaurs: paleopathological evidence. Gaia 15: blasts. European Cells and Materials 15:53–76. 167–184. Sharma, A. R., S. Jagga, S. Lee, and J. Nam. 2013. Interplay Tarlo, L. B. H. 1964. Psammosteiformes (Agnatha)—a review with between cartilage and subchondral bone contributing to patho- descriptions of new material from the Lower Devonian of Poland: genesis of osteoarthritis. International Journal of Molecular I. General part. Palaeontological Polonica 13:1–135. Sciences 14:19805–19830. Tiemann, A. H., and G. O. Hofmann. 2009. Principles of the therapy Sharp, E. L., and J. A. Clack. 2013. A review of the Carboniferous of bone infections in adult extremities: are there any new develop- lungfish genus Ctenodus Agassiz, 1838 from the United Kingdom, ments? Strategies in Trauma and Limb Reconstruction 4:57–64. with new data from an articulated specimen of Ctenodus interrup- Turko, A. J., D. Kültz, D. Fudge, R. P. Croll, F. M. Smith, tus Barkas, 1869. Earth and Environmental Science Transactions M. R. Stoyek, and P. A. Wright. 2017. Skeletal stiffening in an of the Royal Society of Edinburgh 104:169–204. amphibious fish out of water is a response to increased body Shimada, A., T. Kawanishi, T. Kaneko, H. Yoshihara, T. Yano, weight. Journal of Experimental Biology 220:3621–3631. K. Inohaya, M. Kinoshita, Y.O. Kamei, K. Tamura, and Turner, S., C. J. Burrow, H. P. Schultze, A. Blieck, W. E. Reif, H. Takeda. 2013. Trunk exoskeleton in teleosts is mesodermal C. B. Rexroad, P. Bultynck, and G. S. Nowlan. 2010. False teeth: in origin. Nature Communications 4:1638–1639. conodont–vertebrate phylogenetic relationships revisited. Geodi- Shirley, B., M. Grohganz, M. Bestmann, and E. Jarochowska. 2018. versitas 32:545–594. Wear, tear and systematic repair: testing models of growth Uthgennant, B. A., M. H. Kramer, J. A. Hwu, B. Wopenka, and M. dynamics in conodonts with high-resolution imaging. Proceed- J. Silva. 2007. Skeletal self-repair: stress fracture healing by ings of the Royal Society of London B 285:20181614. rapid formation and densification of woven bone. Journal of Smith, A. J., N. Cassidy, H. Perry, C. Begue-Kirn, J. V. Ruch, and Bone and Mineral Research 22:1548–1556. H. Lesot. 1995. Reactionary dentinogenesis. International Jour- van der Vos, W., F. Witzmann, and N. B. Fröbisch. 2018. Tail regen- nal of Developmental Biology 39:273–280. eration in the Paleozoic tetrapod Microbrachis pelikani and com- Smith, M. M., and B. K. Hall. 1990. Development and evolutionary parison with extant salamanders and squamates. Journal of origins of vertebrate skeletogenic and odontogenic tissues. Zoology 304:34–44. Biological Reviews of the Cambridge Philosophical Society Warren, A., and R. Ptasznik. 2002. The earliest fractured tetrapod 65:277–373. bone. Alcheringa: An Australasian Journal of Palaeontology Smithson, T. R. 1985. The morphology and relationships of the Car- 26:459–463. boniferous amphibian Eoherpeton watsoni Panchen. Zoological Witten, P. E., and A. Huysseune. 2009. Comparative view on Journal of the Linnean Society 85:317–410. mechanisms and functions of skeletal remodelling in teleost Sousa, S., N. Afonso, A. Bensimon-Brito, M. Fonseca, M. Simoes, fish, with special emphasis on osteoclasts and their function. J. Leon, H. Roehl, M. L. Cancela, and A. Jacinto. 2011. Differen- Biological Reviews 84:315–346. tiated skeletal cells contribute to blastema formation during zeb- Włodarski, K. H., A. Brodzikowska, and B. Kuzaka. 2014. Are rafish fin regeneration. Development 138:3897–3905. chondroclasts and osteoclasts the same? Folia Biologica (Kraków) Sousa, S., F. Valerio, and A. Jacinto. 2012. A new zebrafish bone 62:135–142. crush injury model. Biology Open 1:915–921. Zammit, M., and B. P. Kear. 2011. Healed bite marks on a Cret- Stilson, K. T., S. S. B. Hopkins, and E. B. Davis. 2016. Osteopathol- aceous ichthyosaur. Acta Palaeontologica Polonica 56:859–863. ogy in Rhinocerotidae from 50 million years to the present. PLoS Zhu, M., and X. Yu. 2002. A primitive fish close to the common ONE 11:e0146221. ancestor of tetrapods and lungfish. Nature 418:767–770.

Downloaded from http://pubs.geoscienceworld.org/paleobiol/article-pdf/45/4/676/4955869/s0094837319000319a.pdf by guest on 29 September 2021