The mammalian circumorbital region is characterized by considerable morphological variation. The primitive condition for mammals is to possess a ligament running along the lateral aspect of the orbit between the frontal bone and zygomatic arch. In primates, all taxa exhibit a postorbital bar with haplorhines exhibiting the derived condition of postorbital septum. Although a postorbital bar has evolved in multiple clades (e.g., primates, herpestids, felids, pteropodids, and artiodactyls), the postorbital septum is unique among mammals (Rosenberger, 1986; Cartmill, 1970, 1972, 1980, 1992; Ross, 1995; Ross and Hylander, 1996; Ravosa et al., 2000a, b, 2006; Menegaz and Kirk, 2009).
Generally, there are two types of functional models for the development of bony circumorbital features. The first proposes that bony circumorbital structures provide insulation to the lateral orbital wall or postorbital region, an explanation derived from the Visual Predation Hypothesis (VPH). The VPH posits that the evolution of a postorbital bar in basal primates with large, convergent, and frontated orbits acts to stabilize the postorbital margins against ocular distortions that may occur during routine chewing or biting (Cartmill, 1970, 1972, 1974, 1992; Ravosa et al., 2000b, 2006). A similar functional argument has been extended to the haplorhine postorbital septum, where orbital closure is posited to insulate the orbital contents from the contractions of the anterior temporalis muscle (Cartmill, 1980; Ross, 1995).
Alternatively, others have argued that bony circumorbital structures are designed to resist masticatory stresses (e.g., Greaves, 1985; Rosenberger, 1986; Tattersall, 1995; Wolpoff, 1996; Bookstein et al., 1999; Cox, 2008). For instance, Greaves (1985, 1991, 1995) proposed that the mammalian postorbital bar functions to resist torsion of the facial skull relative to the braincase during unilateral mastication. According to this model, torsion about the long axis of the skull results in the compression of the working-side postorbital bar and tension along the balancing-side postorbital bar. A similar argument has been posited regarding the function of the haplorhine postorbital septum (Rosenberger, 1986).
Subsequent studies have used strain gauges to measure in vivo stress and strain distribution in various craniofacial regions during normal chewing and biting. Work in the two major groups of living primates, haplorhines, and strepsirhines demonstrates that the primate circumorbital region exhibits much lower peak strains during biting and chewing than along the mandible and midface (Hylander et al., 1991a, b; Ross and Hylander, 1996; Ravosa et al., 2000a, b, c, 2006). Thus, in contrast to the inferior craniofacial region, there is no evidence that the development of bony circumorbital structures such as the browridge, postorbital bar, and postorbital septum are adaptations to resist masticatory loads during routine feeding behaviors.
Recent experimental research in pigs, which retain the primitive mammalian condition of a postorbital ligament, provides additional insight into the function of the circumorbital region (Lemme et al., 2005). Using transducers implanted along the lateral orbital wall, the postorbital ligament in pigs was found to exhibit average elongations of about 1% on both the working and balancing sides during alert mastication. This elongation appears differentially related to the activity of the working-side masseter muscle (Lemme et al., 2005). As a 5% elongation results in the failure of tendon, which is another highly collagenous connective tissue but less elastic than ligament (cf., Wainwright et al., 1976), it is possible that the postorbital ligament exhibits safety factors comparable to some masticatory elements (cf., Hylander et al., 1991a, b; Hylander and Johnson, 1997; Ravosa et al., 2000a, b, c, 2010a).
This novel information helps to frame an outstanding issue regarding the evolution of the mammalian circumorbital region. Currently, it is unknown if the transformation from the primate ancestral soft-tissue condition to a bony postorbital bar in basal primates is related to variation in masticatory stresses, perhaps of a dietary nature, or other unrelated factors (Ravosa et al., 2000a, b, c, 2006). The absence of such information hinders a complete understanding of morphological variation characterizing primate and mammalian phylogeny. As described earlier, the experimental data for primates with a postorbital bar and a postorbital septum indicate that cranial changes during the evolution of the haplorhine postorbital septum are unrelated to shifts in diet-related masticatory stresses. On the other hand, preliminary analyses of circumorbital soft tissues in pigs do not negate the possibility that the postorbital ligament may experience significant loads, with the evolutionary transformation (versus current function) of the postorbital bar potentially related to variation in masticatory stresses.
To fill this significant gap in our understanding of the evolutionary morphology of the mammalian circumorbital region, this experimental study examines the long-term effects of diet-induced variation in mechanical loading during mastication on the microanatomy of postorbital soft tissues in growing rabbits. The New Zealand domestic white rabbit (Oryctolagus cuniculus) is a particularly unique experimental model, in that it exhibits the primitive mammalian condition, where the lateral orbital wall consists of soft tissues and a number of important similarities in the form and function of the masticatory apparatus shared with primates. For instance, considerable in vivo data exist for rabbits regarding jaw-adductor muscle activity, jaw-kinematic and jaw-loading patterns, masticatory function during ontogeny, intracortical remodeling, and diet-related feeding behaviors (Weijs and de Jongh, 1977; Weijs and Dantuma, 1981; Weijs et al., 1987, 1989; Langenbach et al., 1991, 1992, 2001; Hirano et al., 2000; Langenbach and van Eijden, 2001; Taylor et al., 2006; Menegaz et al., 2009, 2010; Ravosa et al., 2007, 2008a, b, 2010b). Similar to a number of mammalian taxa including primates, rabbits demonstrate postnatal plasticity of masticatory soft and hard tissues in response to diet-induced variation in loading patterns (Beecher and Corruccini, 1981; Bouvier and Hylander, 1981, 1982, 1984, 1996a, b; Beecher et al., 1983; Kiliaridis et al., 1985; Bouvier, 1987, 1988; Yamada and Kimmel, 1991; Ravosa et al., 2007, 2008a, b; Menegaz et al., 2009, 2010). Like anthropoids, rabbits have a deep face and a high jaw joint capable of both rotation and translation (Weijs and Dantuma, 1981; Crompton et al., 2006). Furthermore, both anthropoids and rabbits show delayed activity of the balancing-side deep masseter muscle underlying transverse mandibular movements during unilateral mastication (Hylander et al., 1987, 2000; Weijs et al., 1989).
Although functional changes due to masticatory loading are documented for bony elements in the craniofacial skeleton (e.g., mandible; Ravosa et al., 2007), little is known regarding the plasticity responses of postorbital soft tissues to variation in masticatory stresses. Although dietary shifts have not been posited for primate origins (cf., Cartmill, 1992), the derivation of a postorbital bar and postorbital septum led us to test for a circumorbital response to variation in loading conditions. Clarifying the long term, natural influence of altered masticatory stresses on postorbital soft tissues is fundamental for interpreting the behavioral and ecological correlates of morphological variation in extant and extinct mammals as well as for understanding the biomechanics and function of routinely loaded cranial structures.
Here, we extend what is currently known about hard tissue biomechanics in order to test the hypothesis that variation in masticatory stress produced by differences in dietary properties is correlated with variation in postorbital soft-tissue morphology. In the facial skull and feeding apparatus of the same animal model, increased masticatory loading has been observed to result in larger jaw-joint proportions, greater cortical bone thickness, and elevated hard-tissue biomineralization (Ravosa et al., 2007, 2008a, b; Menegaz et al., 2009, 2010). If masticatory stress plays a role in the growth and form of the circumorbital region, similar adaptive responses should be observed in the soft-tissue structures of the lateral orbital wall. Thus, plasticity studies of this kind can be used to inform our notions regarding function, biomechanics, and adaptation (sensu Gotthard and Nylin, 1995). On the one hand, differences in postorbital soft-tissue structures between dietary loading cohorts would suggest that this circumorbital structure is influenced by variation in chewing stresses, thus contrasting with data for bony elements of the circumorbital region (Hylander et al., 1991a, b; Ross and Hylander, 1996; Ravosa et al., 2000a, b, c, 2006). On the other hand, a lack of differences in postorbital soft-tissue structures between dietary loading cohorts would contrast with differences observed for masticatory structures (Ravosa et al., 2007; Menegaz et al., 2009, 2010), and instead suggest that postorbital soft tissues are unaffected by variation in chewing stresses much like bony circumorbital structures (Ravosa et al., 2000c).
MATERIALS AND METHODS
The sample consisted of 30 genetically similar New Zealand domestic white rabbits (O. cuniculus) that were obtained as weanlings (4 weeks old) and raised for 26 weeks until attaining adult status at 30 weeks old (Sorensen et al., 1968; Yardin, 1974). In accordance with an ACUC-approved protocol, three dietary cohorts of 10 rabbits each were established to induce postweaning variation in jaw-muscle forces and masticatory loads (cf., Ravosa et al., 2007, 2010b; Menegaz et al., 2009, 2010). Weaning was chosen as the starting point for dietary manipulation because plasticity decreases with age (Hinton and McNamara, 1984; Meyer, 1987; Bouvier, 1988; Rubin et al., 1992; Ravosa et al., 2008b), because this approximates shifts in masticatory function in the wild (cf., Ravosa et al., 2007, 2008a, b), and because it minimizes the confounding influence of diets other than those used in our protocol.
Cohorts were separated into an “under-use” cohort fed only powdered pellets (n = 10), a “control” cohort fed intact Harlan TekLad rabbit pellets (n = 10), and an “over-use” cohort fed whole rabbit pellets supplemented daily with two 2.5-cm cube hay blocks (n = 10). Table 1 summarizes the dietary mechanical properties of the experimental food. In rabbits and other mammals, fracture-resistant foods such as pellets and hay require absolutely larger jaw-muscle activity and greater cyclical loading during food processing, which in turn increases masticatory stresses (Weijs and de Jongh, 1977; Bouvier and Hylander, 1981; Weijs and Dantuma, 1981; Weijs et al., 1987, 1989; Hylander et al., 1992, 1998, 2000; Ravosa et al., 2007, 2008a, b). The inclusion of pellets in the diet of all rabbits facilitated adequate nutrition for normal postweaning growth.
Table 1. Dietary mechanical properties of rabbit experimental foods measured with a portable food tester
Food items (sample size)
Young's modulus: E in MPa (mean, range)
Toughness: R in Jm−2 (mean, range)
Hardness: H in MPa (mean, range)
Ground pellets require minimal oral preparation, which reduces the amount of cyclical loading during unilateral mastication (under-use). The control diet consisted of intact pellets. Hay requires greater forces to process, which, in addition to the greater processing required for intact pellets, results in increased peak loads and increased cyclical loading during molar biting and chewing (over-use). The mechanical properties of wet hay model the exposure of hay to saliva. Missing values are due to the inability to perform that particular test on the item (due to its size, shape, or properties).
Pellets (n = 10)
Wet hay (n = 15)
Dry hay (n = 15)
Histology and Immunohistochemistry
Once euthanized at 30 weeks, rabbit crania were fixed in 10% buffered formalin. The eyes were enucleated by removing the superficial connective tissues then transecting the optic nerve. The postorbital ligament spans the bony gap along the lateral orbital margin between the temporal process of the frontal bone and the frontal process of the temporal (squamosal) bone (see Fig. 1). Once enucleated, surrounding connective tissues were further removed to expose lateral orbital wall tissues. Only right-sided postorbital soft tissues were dissected away for histology. These were fixed in 10% buffered formalin for 48 hr, followed by a 24-hr decalcification period using Cal-Ex II (Fisher Scientific). Decalcified soft tissues from each specimen were segmented into inferior, intermediate, and superior sections (see Fig. 1) and then dehydrated and paraffin embedded at the University of Missouri Veterinary Medical Diagnostics Laboratory (VMDL). Data for superior and inferior regions are not presented here, because the inferior and superior tissue segments were largely uniform in microanatomy regardless of loading regime. Thus, only data for the intermediate region of the postorbital soft tissues are presented and discussed herein.
To provide a comparison between postorbital soft tissues and a cranial structure known to be influenced by masticatory stresses (Ravosa et al., 2007), the left-sided temporomandibular joint (TMJ) from each of the above specimens was fixed, decalcified, embedded, and sectioned. Coronal sections (6 μm) from the intermediate joint region were used for histological and immunohistochemical analyses of extracellular matrix (ECM) composition and properties of TMJ articular cartilage. Such assays were performed following similar analyses of postorbital soft tissues.
Several histological and immunohistochemical stains were used to analyze changes in the protein composition of the postorbital ligament as a function of dietary properties (Kiernan, 1999; Ravosa et al., 2007, 2008a, b). H&E staining was used to identify general soft-tissue architecture. Collagen orientation and intensity were defined through picrosirius-red staining (Junqueira et al., 1979). Safranin-O/fast green immunohistochemistry (Kiernan, 1999) identified proteoglycan content in the ECM of the soft tissue, and thus variation in tissue viscoelasticity. Immunostaining (Kiernan, 1999) was used to identify variation in Type-II collagen expression among loading cohorts, which in turn tracks variation in tissue viscoelasticity. The latter two staining protocols were similarly applied to TMJ articular cartilage for comparison with the postorbital ligament.
H&E highlights general variation in tissue organization among loading cohorts and postorbital soft-tissue sections (Fig. 2). H&E staining shows that the postorbital ligament is composed of elastic fibrocartilage, rather than dense connective tissue. Identifying postorbital soft-tissue type requires consideration of cell type. Dense connective tissues are composed of ligament with fibroblasts at the matrix interface, whereas a fibrocartilaginous ligament contains chondrocytes. The inset in Fig. 2 illustrates morphology typical of fibrocartilage, which is commonly described as dense fibrous tissue arranged in bundles with chondrocytes embedded among the bundles. Fibrocartilage appears designed to resist compressive loads and to protect blood vessels from compression (Benjamin and Ralphs, 1998). Similar to other connective tissues, fibrocartilage in the postorbital region may be a precursor to osteogenesis and ossification (Carter, 1987). Indeed, the intermediate section of “over-use” cohort shows increased disorganization and degradation, which may be indicative of tissue turnover and osteogenesis. However, tissues stained for Type-II collagen expression (see later) gave little indication of being cleaved, which suggests that osteogenesis in this region may be unlikely.
The postorbital ligament in “under-use,” control, and “over-use” cohorts indicate low levels of safranin-O staining for proteoglycans, which suggests a lower viscoelastic capacity of these soft tissues (Fig. 3). Further, as there is little variation across dietary groups, there is a minimal postorbital response to masticatory loading, and thus minor variation in tissue viscoelasticity across dietary cohorts. These findings are in marked contrast to findings for soft tissues in the rabbit masticatory complex (Fig. 4). In particular, the TMJ shows considerable variation across dietary cohorts, with “under-use,” control and then “over-use” groups exhibiting progressively lower levels of safranin-O staining, indicative of decreased articular cartilage viscoelasticity (see also Ravosa et al., 2007).
The intermediate aspect of “under-use,” control, and “over-use” dietary cohorts exhibit low levels of postorbital soft-tissue immunostaining (Fig. 5). For the most part, there is little variation across groups in Type-II collagen staining intensity, suggestive of the absence of masticatory-related variation in the viscoelasticity of postorbital tissues. However, in the “over-use” diet group, there is greater expression of Type-II collagen localized to sites where collagen fibers appear to be degraded. As noted earlier for diet-induced patterns of variation in proteoglycan content, TMJ articular cartilage exhibits similar variation in the expression of Type-II collagen (Fig. 4), with “under-use,” control and then “over-use” groups having progressively lower levels of staining, indicative of decreased TMJ articular cartilage viscoelasticity (see also Ravosa et al., 2007).
Picrosirius-red staining indicates a significantly altered collagen fibril organization in the “over-use” cohort relative to “under-use” rabbits (Fig. 6). Although “under-use” rabbits show parallel bundles of collagen fibers, “over-use” collagen exhibits a disorganized and less complex fiber structure. These data suggest that collagen fiber organization of the postorbital ligament is altered by masticatory loading.
This experimental study evaluated the plasticity and function of circumorbital soft-tissue structures during masticatory function in a rabbit model of the primitive condition of the mammalian lateral orbital wall, a condition resembling that for the ancestors of primates (Cartmill, 1992; Ravosa et al., 2000a, b, 2006; Bloch and Boyer, 2002). A series of histological and immunohistochemical techniques were used to identify variation in postorbital soft-tissue architecture, collagen fiber organization, and ECM composition as a function of diet-induced masticatory stresses.
To an extent, previous descriptions of the mammalian circumorbital region failed to adequately describe the structure of soft tissues along the postorbital region. The rabbit postorbital region is composed of dense fibrous tissue bundles with chondrocytes among the bundles. Fibrocartilage is considered a “transitional tissue,” as it can differentiate from both dense fibrous connective tissue and hyaline cartilage (Benjamin and Ralphs, 1998). Fibrocartilage is generally avascular, and the lack of blood vessels allows it to resist routine compressive loads. Consistent with the general microstructure of fibrocartilage, rabbit postorbital tissues were characterized by variation in their internal organization (see earlier). The more complex combination of a basket-weave pattern and parallel arrangement of collagen fibrils was unique to the intermediate aspect of the postorbital ligament, independent of cohort. When cohorts are analyzed separately, the intermediate section of the postorbital soft tissues in control and “under-use” rabbits indicates a similar arrangement, whereas a more irregular fibril orientation is observed in “over-use” rabbits. This finding suggests that long term increases in masticatory loading results in a degradative response, whereby postorbital fibrocartilage exhibits decreases in multidirectional collagen fiber orientation and subsequent reductions in the ability to resist multidirectional stresses.
Aggrecan, a hydrophilic molecule in fibrocartilage is confined in the ECM, when the tissue is compressed by the basket-weave arrangement of collagen fibrils (Benjamin and Ralphs, 2004). In vivo loading models indicated that aggrecan gene expression was both load and frequency dependent in mouse intervertebral discs with higher-stress cohorts exhibiting short-term elevations in aggrecan gene expression (MacLean et al., 2003). Long-term exposure to elevated compressive loads downregulates aggrecan gene expression, which may lead to a myriad of biomechanical effects including cell death, loss of meshwork arrangement, and subsequent degeneration of collagen fibers. This would suggest that the exposure to long-term loading may decrease tissue cohesion, and thus overall structural integrity, a pattern characteristic of postorbital cartilage and jaw-joint cartilage (Ravosa et al., 2007, 2008a, b).
The extent to which such variation in ECM expression is observed in other mammals with a primitive postorbital region remains to be determined. Unlike inferior craniofacial regions such as the symphysis and TMJ (Ravosa et al., 2007, 2008a, b), there was minimal variation across cohorts in proteoglycan expression in the rabbit postorbital ligament. As tissue viscoelasticity is likewise presumably low across such groups, this suggests that postorbital soft tissues are unlikely to be designed for resisting significant stresses during chewing and biting. Results of Type-II collagen staining across dietary cohorts offer further support for this interpretation.
The avascular nature of the rabbit postorbital fibrocartilage prevents the tissue from being able to respond (or adapt) to long-term loading patterns. Conversely, fibrocartilage exposed to low and normal loading regimes maintains multidirectional collagen fiber orientation (e.g., Provenzano and Vanderby, 2006), which may demonstrate a nonelastic response in the “under-use” and normal cohorts (Figs. 3 and 4). Picrosirius-red staining identified variation across rabbit dietary groups, where the complexity of collagen fiber orientation appears to be inversely related to masticatory stresses. Whether this represents permanent cartilage degeneration or collagen degradation as a precursor to tissue biomineralization is currently unknown, although preliminary analyses suggest little support for the latter interpretation.
What are the implications of the earlier evidence regarding postorbital soft-tissue plasticity for understanding circumorbital adaptation in primate and nonprimate mammals? Most importantly, there is no evidence that long-term elevations in masticatory stress can induce the epigenetic transformation of soft tissues of the lateral orbital wall into a bony postorbital bar. Rather, ECM components related to tissue viscoelasticity show a greater potential for elevated load-resisting abilities and a larger plasticity response in TMJ articular cartilage versus the postorbital ligament. Although dietary shifts has not been posited for primate origins, the data presented here have broader implications for mammals. In particular, as noted for the postorbital bar and septum, it is unlikely that circumorbital soft tissues show adaptive responses to increases in routine masticatory forces. Indeed, contrary to predictions of masticatory-stress models of the circumorbital region, elevated loading results in the degradation of circumorbital soft-tissue structures. Although masticatory elements such as the mandibular symphysis and TMJ vary dynamically in response to functional requirements, the postnatal plasticity of circumorbital soft tissues is much less pronounced under the loading regimes tested here. These results provide no support for the supposition that evolution of the postorbital bar in various mammal clades (felids, herpestids, pteropodids, etc.) is due to elevated masticatory stresses. Instead, the best evidence to date suggests that increased orbital convergence and/or orbital frontation may require a stiff lateral orbital margin to maintain high levels of stereoscopic nocturnal visual acuity during predatory behaviors (Cartmill, 1970, 1972, 1974, 1992; Noble et al., 2000; Ravosa et al., 2000b, 2006; Menegaz and Kirk, 2009). Obviously, this is not the final word regarding the evolution of the postorbital bar as there are many clades in which high levels of visual acuity remain to be implicated as selective pressures underlying morphological differences in the mammalian circumorbital region.
To further address circumorbital function and evolution, we examined the long-term plasticity of circumorbital soft tissues in a rabbit model of the primitive mammalian condition. Masticatory forces related to long-term processing of a tougher or more resistant diet appear to result in a reduced organizational complexity of circumorbital soft tissues, particularly collagen fibers indicating a decreased ability of the postorbital ligament to resist multidirectional loads. However, although diet-related variation in masticatory stress may affect ECM composition and structural properties in certain craniomandibular elements (i.e., TMJ or symphysis), current experimental data provide minimal evidence that circumorbital tissues are similarly influenced by masticatory forces. Indeed, the postorbital ligament only maintains an ability to resist multidirectional loads when such stresses are moderate-to-low in magnitude, and there are few indicators that ECM composition confers much tissue viscoelasticity in any of the treatment groups. Rather, long-term exposure to elevated cyclical mechanical loading may be maladaptive, such that the soft- to hard-tissue transformation in the postorbital region of primates occurred due to different selective pressures almost certainly unrelated to mechanical loading.
In sum, an “osteocentric” focus of research on circumorbital biomechanics has hindered our understanding of craniofacial evolution in primates and other mammals. As this is the first study aimed at identifying the microanatomical function and plasticity of the postorbital region of a species retaining the basal mammalian condition, future investigations would benefit from similar work on the development of the circumorbital region in other species. Another point for further consideration is the role of masticatory loading on tissue turnover, particularly addressing whether long term, elevated masticatory stresses result in permanent tissue degradation (i.e., degeneration) or if tissue degradation is a precursor to osteogenesis.
J. Organ, T. Smith, V. DeLeon, and Q. Wang are thanked for inviting us to contribute to their special issue on experimental approaches to morphology. B. Wright kindly performed the analyses of rabbit food properties. W. Phillips is thanked for helping with initial phases of this project. We are also indebted to L. Coussens for the Picrosirius-red protocol and D. Miller for access to and assistance with a polarizing microscope. E.J. and A.N.D. were supported by MU LS UROP Fellowships. R.A.M. was supported by MU Life Sciences and NSF Graduate Research Fellowships. Lastly, this manuscript benefited from the comments of V. DeLeon and two anonymous reviewers.