Tyrannosaurus rex, Osborn, 1905

Mary Higby Schweitzer, Jennifer L. Wittmeyer & John R. Horner, 2008, One pretty amazing T. rex, Tyrannosaurus rex, the tyrant king, lndiana University Press, pp. 92-100 : 93-98

Determining sex in extinct animals is difficult because most features commonly used to assign sex are lost in the process of fossilization. Despite this difficulty, many bonv features of dinosaurs have been interpreted to be evidence of sexual dimorphism, including degree of robustness in sauropods and their close relatives ( Weishampel and Chapman 1990; Galton 1997; Benton etal. 2000), theropods ( Carpenter 1990; Larson 1994; Smith 1998) and protoceratopsids ( Tereschenko and Alifanov 2003); horn core size in ceratopsids ( Godfrey and Holmes 1995); or presence or absence of the first caudal chevron (Larson and Frey 1992; Larson 1994), to name a few. However, even if such features could definitively be shown to be products of sexual differentiation, it remains impossible to assign a particular feature unambiguously to a specific sex (e.g., the robust morph being female; Carpenter 1990; Larson 1994). At best, assigning sex to a specific morphotvpe of dinosaurs has fallen within the realm of speculation. What is needed is an unambiguous means of assigning a particular sex to male and female morphs. One possibility is the identification of medullary bone in dinosaurs.

Medullary bone is an ephemeral reproductive tissue that is present in living taxa and that is found exclusively in female, actively reproducing birds. This bonv tissue lines the medullary cavities of the long bones of extant birds; it is chemically and morphologically distinct from other bone types. Special characteristics of composition and structure contribute to the high metabolic rates of medullary bone. In fact, it is capable of being metabolized 10 to 15 times faster than cortical bone ( Simkiss 1961; Dacke et al. 1993), and it serves as an easily mobilized calcium storage tissue for the production of calcareous eggshell ( Sugiyama and Kusuhara 2001). Its presence in dinosaurs would indicate sex, support phylogenetic proximity, suggest shared reproductive physiological strategies w ith extant birds, and indicate reproductive phase at the time of death.

Comparison of Medullary and Cortical Bone Characteristics

In addition to protection and support ofvital internal organs, bone plays an important role in calcium metabolism in vertebrates, including all avian taxa ( Miller and Bowman 1981). Long bone formation in extant birds procedes much the same as in other vertebrate taxa through endochondral ossification of preexisting cartilage models ( Whitehead 2004; Tavlor et al. 1971). Bone elongation involves periosteal deposition, and concurrent endosteal osteoclastic resorption at the metaphyseal region, resulting in overall maintenance ofbone morphology and thickness during longitudinal growth ( Taylor et al. 1971).

In both formation and elongation, bone production involves 2 phases, which reflect the composite nature of bone material. In the first, the boneforming cells (osteoblasts) secrete organic matrix (osteoid) ( Taylor et al. 1971; McKee et al. 1995). This matrix primarily consists of the fibrous helical protein collagen I and the accessory collagen V; the noncollagenous proteins osteocalcin, osteopontin, and osteonectin ( Bonucci and Gherardi 1975; McKee et al. 1993; Gerstenfeld et al. 1994; Sugiyama and Kasuhura 2001; Wang et al. 2005), and bone sialoprotein ( Gerstenfeld et al. 1994; Robey 1996 and references therein); serum proteins, including hemoglobin and albumin ( McKee et al. 1993); and various glycosaminoglycans ( Bonucci and Gherardi 1975; Dacke et al. 1993; Arias and Fernandez 2001; Wang et al. 2005). Therefore, cortical and trabecular bone have specific, characteristic, and defineable chemical and molecular profiles.

However, in female birds, a unique bone type is formed as the result of a surge in blood estrogen levels at the onset of sexual maturity ( Bonucci and Gherardi 1975; Knott and Bailey 1999; Dacke et al. 1993; Whitehead 2004).

Medullary bone does not occur naturally in any other taxon ( Elsey and Wink 1986; Dacke et al. 1993), and it is present only during the reproductive period in all living female birds, filling the marrow cavities of many skeletal elements (Wilson and Thorpe 1998; Van Neer et al. 2002). It is produced byspecialized osteoblasts that lie within the endosteum, a thin connective tissue layer that lines the marrow surfaces of the bones ( Van Neer et al. 2002).

Medullary bone exists only to offset the effects of bone resorption during shelling by serving as an easily mobilized source of calcium, and it has no direct biomechanical function ( Bonucci and Gherardi 1975; Wilson and Thorp 1998). It is chemically and morphologically distinct from other bone types. Although medullary bone has been assumed to be present in extant paleognaths, it has not been previously imaged or studied, and no data exist regarding the morphology or chemistry of this bone type in ratites.

The mineral phase of both medullary and cortical bone is primarily hydroxyapatite (Ca10(PO4)6(OH)2), but the ratio of mineral to organics is measurably higher in medullary bone ( Ascenzi et al. 1963; Taylor et al. 1971; Dacke et al. 1993), and medullary bone incorporates a higher proportion of calcium carbonate ( Pelligrino and Blitz 1970) than other bone types. Medullary bone is not only more highly mineralized than cortical bone, but also the distribution of minerals is different between the 2 bone types. In cortical bone, the mineral crystals are regularly distributed at the head of the A bands of collagen molecules ( Taylor et al. 1971), but in medullary bone, mineral distribution and orientation is much more random, with mineral crystals additionally deposited in intrafibrillar spaces ( Ascenzi et al. 1963; Taylor et al. 1971). In addition, medullary bone does not exhibit birefringence because of the random arrangement of both collagen fibrils and mineral, whereas other bone types are anisotropic in polarized light ( Miller and Bowman 1981; Wilson and Thorp 1998). Finally, the mineral crystals incorporated into medullary bone are somewhat larger than the microcrystalline apatite of other bone types ( Ascenzi et al. 1963), producing a greater crystallinity index.

The organic phase of medullary bone differs significantly from that of cortical and trabecular bone. Collagen makes up a greater proportion of the organic matrix of cortical bone, whereas the percentage of noncollagenous proteins to collagen is far greater in medullary bone, comprising approximately 40% of the total organics ( Knott and Bailey 1999). The concentration of various glyclosaminoglycans is greater in medullary than cortical bone, and it incorporates different amino sugars ( Bonucci and Gherardi 1975). Hexosamine and keratan sulfate are much more prevalent in medullary than cortical bone ( Taylor et al. 1971; Wang et al. 2005), which incorporates chondroitin sulfate instead. In addition, relatively high concentrations of tartrateresistant acid phosphatase (TRAP), an enzyme involved in digestion of bone ( Sugiyama and Kusuhara 2001), are found in medullary bone. These chemical differences are reflected in the differential response of the 2 bone types to various histochemical stains ( Fig. 7.1 ; Taylor et al. 1971; Sugiyama and Kusuhara 2001; Wang et al. 2005).

Function of Medullary Bone

Unlike other bone types, medullary bone has no biomechanical or other supportive function. It exists solely as a calcium storage tissue that aids in mineral mobilization to the shell gland during lay ( Dacke et al. 1993; Wilson and Thorp 1998; Whitehead 2004). As mentioned previously, medullary-bone formation in birds is triggered by increased levels of both estrogen and androgens that accompany ovulation, activating osteoblasts to begin secretion of osteoid while inhibiting osteoclast activity ( Dacke et al. 1993; Whitehead 2004). The formation of medullary bone begins approximately 1 or 2 weeks before lay. It is maintained during the full laying cycle, and it may persist up to 1 week after lay before resorption is complete ( Reynolds 2003). Medullary bone osteoclasts in female birds are specialized to contain estrogen receptors in their cell membranes, which, when triggered by rising reproductive hormones, increases the efficiency of mobilizing stored calcium ( Miller 1981). Although evidence of medullary bone may be found in virtually all skeletal elements of extant birds, it is most abundant in the femur and tibiotarsus of most birds studied ( Reynolds 2003), and, consistent with its function as a source of rapid calcium mobilization, it is infused with abundant vessels and blood sinuses. In fact, up to 40% of the calcium used in eggshell formation comes directly from the resorption of medullary bone ( Mueller et al. 1969; Dacke et al. 1993). Although it is not known to serve a direct mechanical function, in reducing the resorption of cortical and trabecular bone, it may aid in maintaining integrity and strength of structurally important bone ( Whitehead 2004), and indeed, the presence of medullary bone in long bones of laying birds has been shown to increase fracture resistance of these elements ( Fleming et al. 1998).

Like birds, most reptiles, including crocodiles and alligators, also produce calcareous eggshell, but apparently do not produce medullary bone ( Elsey and Wink 1986; Dacke et al. 1993). This may be because of different mechanisms of shelling ( Jackson et al. 2002) and overall greater bone density that can offset the calcium draw without requiring additional bone storage sources. Thus, extant nonavian archosaurs undergo bone resorption during lay, but the structural integrity and biomechanical function of these organisms is apparently not compromised during shelling.

Although medullary bone has not been previously observed or described in dinosaurs, it was proposed that reproducing dinosaurs, at least in the theropod lineage most closely related to avian dinosaurs, would possess this ephemeral tissue ( Martill et al. 1996; Chinsamy and Barrett 1997). The failure to observe or identify these fragile reproductive tissues in dinosaurs previously may be due to a number of taphonomic and/or biological factors, or observational bias. First, we do not have anv way of estimating the length of the reproductive cvcle in theropods. There is a wide range of reproductive strategies among living birds, and the extent and distribution of medullary bone in these taxa differ correspondingly ( Schraer and Hunter 1985). If theropods reproduce seasonally, they may only possess the tissue for a maximum of a month or less. Second, because of the relativ ely thick, dense cortex, the need for medullary bone may be less in these animals, so the medullary layer may be quite thin. Third, in extant birds, the tissue is quite fragile, and separates easily from the overlying cortex ( Fig. 7.1 B, C). It may be that the tissues are lost, either during fossilization or during subsequent recovery and preparation. Third, it may be that medullary bone differs sufficiently from that of extant derived birds so that it is not recognized. A fourth factor may be the failure to examine bones for its presence because of collection techniques requiring bone to be conserved, and not broken to expose interior fragments.

Medullary Bone in T. Rex

At the end of field season in 2002, a well-preserved specimen of Tyrannosaurus rex ( MOR 1125 ) was found as an association of disarticulated elements. The site was located at the base of the Hell Creek Formation (Lancian), about 8 m above the Fox Hills Sandstone. Soft, well-sorted sandstones derived from an estuarine or fluvial setting surrounded the skeletal elements. Some of the elements evidenced slight crushing, but overall preservation was excellent. MOR 1125 , nicknamed Bob-rex after its discoverer, Bob Harmon, is a relatively small but fully adult T. rex . In comparison with another T. rex , FMNH PR2081 , with a femur length of about 131 cm, the femur of MOR 1125 is only' 107 cm in length. By use of lines of arrested growth, MOR 1125 was calculated to be about 18 years old at the time of death ( Horner and Padian 2004).

The remote region where MOR 1125 was recovered had no roads into the site, so a helicopter was required to transport field jackets to the MOR laboratories. However, the jacket containing the femur and other elements was too heavy to be airlifted out, and the jacket and bones they contained were broken and rejacketed for removal. In the process, many internal fragments that were visually free of preservative or consolidants were collected for analyses.

When these fragments were examined in hand sample, a bony tissue lining the endosteal surface of the bone could be seen that was distinct in texture, appearance, and distribution from other described dinosaur bone tvpes. The morphological similarity of the new tissues to avian medullary bone was immediately apparent (Schweitzer et al. 2005b). Figure 7.2 shows fresh-fracture images of Tyrannosaurus rex endosteal tissues ( Fig. 7.2 A, B), compared with medullary bone tissues in reproducing ostrich ( Fig. 7.2 C) and emu ( Fig. 7.2 D). The hallmark traits of medullary bone—dense vascularity and random, woven bone pattern—are clearly visible in all samples. Large erosion cavities are visible in all medullary tissues (indicated with an asterisk in figures), indicating that calcium mobilization has occurred.

Demineralization of extant bony tissues is commonly used to more clearly observe microstructural characteristics, such as fibril orientation; and when mineral is removed, the primarily collagenous protein matrix is exposed. Conventional wisdom has held that when fossilized dinosaur bone is subjected to the same treatment, the bone would dissolve completely because no proteinaceous material would persist over the course of geological time ( Hoss 2000).

In order to determine characteristics of presumed medullary tissues, we prepared a partial demineralization designed to etch mineral enough to expose underlying patterns. At this point, we discovered an unexpected and novel characteristic to this bony tissue. As minerals were dissolved from the medullary bone, the sample did not disintegrate, but, similar to extant bone, tissues remained ( Schweitzer et al. 2005a). Furthermore, these dinosaur tissues exhibited apparent original flexibility, comparable to that seen in extant ratites. However, these characteristics are not germane to this chapter and will be discussed elsewhere. The retention of a pliable and fibrous matrix after demineralization speaks to unusual preservation in this dinosaur material and suggests that perhaps theorized modes of fossilization may need to be reevaluated. Figure 7.3 demonstrates the persistence of fibrous tissues after demineralization. Small fragments of emu and dinosaur demineralized medullary bone tissues show random fiber orientation, and large open spaces for vessels and vascular sinuses permeate the tissues. The morphological similarity between extant and fossil samples is clearly visible and supports the hypothesis of a common origin to the tissues.

Conclusion

The endosteally derived bone tissues observed in this specimen of Tyrannosaurus rex ( MOR 1125 ) have all of the morphological characteristics of medullary bone, a distinctive avian reproductive tissue. Although not identical to published accounts of extant neognaths, the dinosaur tissues fall within the range of morphological variation observed in ratites. This bone tissue is derived from the endosteum, is highly vascular, and exhibits the random, woven bone arrangement consistent with very rapidly deposited bone. In addition, it has been identified on the endosteal surfaces of both femora and one tibia, the only bones examined for the presence of this tissue. The distribution is consistent with that seen in extant birds and suggests an organismal, rather than pathological, response.

Pathologies of the endosteum are relatively rare and localized, and they are usually accompanied bv cortical bone anomalies in the affected regions. No anomalies were observed, either grossly or microscopically, in MOR 1125 . In light of the fact that the relationship between theropod dinosaurs and birds is robustly supported (e.g., Gauthier 1986; Sereno 1997; Holtz 2004), it is most parsimonious to conclude that this novel tissue seen in MOR 1125 is medullary bone, and its presence in theropods not only adds independent support to the robustly demonstrated relationship between theropods and birds, but also suggests that these organisms had similar reproductive physiological strategies. In addition, its presence provides a means for unambiguous assignment of sex in dinosaurs. With careful examination, other, less ephemeral morphological traits may be identified in this specimen that can be applied to differentiate nonreproducing females in this lineage.

Figures

Figure 7.1. Medullary bone in extant laying hen. (A) Gross cross section of femur of actively laying hen shows extensive medullary bone formation. New bone is randomly oriented and much more porous than overlying cortical bone. (B) Low magnification and (C) high magnification of histological section of demineralized bone from laying hen. Chemical differences between cortical and medullary bone are indicated by differential response of each bone type to hematoxylin and eosin staining. In (C), separation of the medullary bone from cortical bone is seen as sectioning artifact. Large, multinucleated osteoclasts are visible around bone spicules, and small osteoblasts align along preexisting bone spicules, active in deposition of new bone. Abbreviations: CB, cortical bone; MB, medullary bone; ELB, endosteal laminar bone; OCL, osteocyte lacunae; OC, osteoclast; OB, osteoblast. Scales as indicated. See color version of this figure in the accompanying CD-ROM

Figure 7.2. Fresh fracture of tibiae of MOR 1125 (A, B), ostrich (C), and emu (D). Morphology and microstructure of medullary bone is observed in all cases as distinct from overlying cortical bone. Medullary bone is less organized and more vascular. Large vascular sinuses can be seen in the medullary bone, and in some cases, large erosion cavities (*) are visible at the interface between medullary and cortical bone and in the medullary bone itself, indicating some resorption of bone has occurred. In emu (D), a large elongate erosion room is infilled with new medullary bone with characteristic corrugated texture. Abbreviations as in Figure 7.1. T, trabecular spicule. Scales as indicated. See color version of this figure in the accompanying CD-ROM.

Figure 7.3. Demineralized fragments of medullary bone from emu (A) and MOR 1125 (B). The fibrous, woven pattern of bone matrix is visible in both cases, and the relatively lacy appearance results from penetration of the bone by blood vessels. Scales as indicated. For methods on demineralization, see Schweitzer et al. (2005a), supplemental online information. See color version of this figure in the accompanying CD-ROM