Understanding Entheses: Bridging the Gap Between Clinical and Anthropological Perspectives
Abstract
An enthesis is the interface where tendon meets bone, providing both muscle anchorage and stress dissipation. Previous anthropological research suggests size and complexity of entheses observable in osteological material, are indicative of the strain magnitude resulting from repetitive muscle contractions during the performance of daily routines. These proposed “musculoskeletal stress markers” are routinely incorporated into bioarcheological studies as evidence of general activity patterns past human populations participated in. However, much of how this complex osteotendinous interface develops and responds to mechanical strains is poorly understood. The following review seeks to shed light on this structural enigma by synthesizing current findings in both the clinical and anthropological literature, in the interest of generating new conversations in how entheses respond to contractual forces and the systemic influences (i.e., genetics, hormones), that surround their morphological development. Only once we truly understand the etiology of tendon insertion sites, will the value of enthesial research in reconstructing human behavior be determined. Anat Rec, 2012. © 2012 Wiley Periodicals, Inc.
Anthropologists regularly implement bone remodeling principles in response to mechanical loads to infer a causal relationship between muscle tendon insertions and activity levels. Previous research suggests size and complexity of tendon insertions are indicative of strain magnitude resulting from habitual physical activity (Lane, 1887; Kennedy, 1989; Hawkey and Merbs, 1995; Churchill and Morris, 1998; Hawkey, 1998; Steen and Lane, 1998; Stirland, 1998; Wilczak and Kennedy, 1998; Capasso et al., 1999; Weiss, 2003, 2004, 2007; al-Oumaoui et al., 2004; Molnar, 2006; Cardoso and Henderson, 2010; Thara et al., 2010; Havelková et al., 2011). Using both qualitative and quantitative data to assess the expression of insertion morphology along long bone diaphyses, inferences are frequently made regarding both generalized and specific activities past humans participated in throughout their daily lives. Recognition of this potential relationship between mechanical strain and the osteotendinous interface has led many to refer to tendon insertions sites as musculoskeletal stress markers (MSM). However, more empirical analyses exploring the factors responsible for these proposed MSM are needed before we can confidently begin to assess individual behavior and draw population-based inferences from series of skeletal remains. With this lack of understanding in the etiology of periosteal tendon insertion morphology, particularly in the mechanical and structural relationship between soft and hard tissues, this interface should be referred to by the clinical term, enthesis.
In 1997, a symposium entitled “Activity Patterns and Musculoskeletal Stress Markers: An Integrative Approach to Bioarcheological Questions” was held at the 66th annual American Association of Physical Anthropologists meeting. The primary focus of this symposium was to develop a standardized method for the collection and analysis of osteological indicators presumed to have a direct correlation with physical activity. Selected papers given at this meeting, along with later contributions to the topic, were assembled into a volume of the International Journal of Osteoarcheology (1998) devoted entirely to MSM. Contributors assessed the role of entheses in reconstructing behavior as well as their methodological and statistical limitations.
The organizers of the symposium and the publication to follow recognized a lack of understanding in how hormones, bone remodeling rates, and various biomechanical agents affect enthesial development. However, they continued to advocate the value of muscle tendon insertions in interpreting gender-specific activities at both the individual and population-level, as long as contextual support from archeological and ethnographic sources is available. The need for an established biological and social context when conducting behavioral interpretations is nothing new, especially since the last decade of bioarcheology centered on how to account for this osteological paradox, or inherent mortality bias present when assessing past populations (Wood et al., 1992; Larsen, 1997). Still, one cannot help but question what the predictive value of MSM oriented research is when conclusions can only safely be drawn if additional contextual support is present; which is a variable factor when assessing archeological populations.
These concerns notwithstanding, the decade to follow would include dozens of publications incorporating “musculoskeletal stress markers” into their research model, and many graduate students eager to apply the evolving methodology in both archeological and forensic settings. As momentum continued to progress forward, many of those applying the methods increasingly made concessions to the unknown parameters hypothesized to impact enthesial morphology alongside presumed mechanical forces.
To resolve some of these concerns and continue to diminish methodology subjectivity, a practicum entitled “Workshop in Musculoskeletal Stress Markers (MSM): Limitations and Achievements in the Reconstruction of Past Activity Patterns,” was held in 2009 at the University of Coimbra, Portugal. Including some of the most published practitioners of MSM assessment, the workshop succeeded in generating a standardized terminology, as well as highlighting some of the methodological and interpretative contributions and restrictions associated with the use of tendon insertion morphology in behavioral reconstructions. The most important contribution to come from this seminar was the progress made in qualitatively assessing enthesial markers via 3D scanning, limiting the subjective nature of ranked morphological indicators. This technological progression in the assessment of entheses may allow for statistical correlations between skeletal morphology and activity, potentially allowing for the delineation of more meaningful macroscopic methodology in accounting for morphological development. There continues to be a lack of sound research exploring the etiology of tendon insertion scarification, and whether it is truly indicative, or at least primarily representative, of habitually applied mechanical loads (Robb, 1998; Wilczak, 1998; Jurmain, 1999; Schlecht, 2004; Pearson and Buikstra, 2006; Zumwalt, 2006; Schlecht, 2008; Weiss et al., 2012).
In attempting to understand this unique tissue interface and what truly drives its development, there are cost factors and technological limitations that must be considered, notwithstanding the constraints inherent in longitudinal development studies on living subadult populations. However, before anthropologists can continue to incorporate MSMs in their biological profiles, researchers must have a better grasp on how entheses respond to both internal and external environmental factors. Though anthropologists have made strides in attempting to methodologically and statistically control variables such as age, sex and body mass (Villotte et al., 2010a, b; Weiss et al., 2012), there are still many other unexplored factors that may alter our perspective on enthesial response to load. Assuming enthesial development is reflective of skeletal loading regimes, then methodology that is macroscopically reproducible may be devised.
That repetitive muscular contractile forces primarily influence enthesis development is a rational conclusion. However, few investigators attempting behavior reconstructions question the lack of direct evidence for such a relationship. Several researchers note that most activity-related studies fail to consider other mechanical and systemic influences on osteotendinous responsiveness to load, such as body mass or genetics or both (Robb, 1998; Wilczak, 1998; Weiss, 2003; Schlecht, 2004; Zumwalt, 2006; Schlecht, 2008). Additionally, previous studies rarely account for muscle activity intensity, an individual's skeletal maturity, or the time frame in which a proposed activity may have taken place during one's lifetime. Moreover, investigators often justify their use of MSM data by citing research conducted by Chamay and Tschantz (1972) and Woo et al. (1981) as confirmation that activity increases periosteal apposition in regions associated with muscle tendon insertions. Both of these early exploratory studies in bone remodeling specifically concentrated on long bone diaphyses, referencing its “hypertrophic” state at the periosteum following loading regimes. The respective authors were not concerned with local enthesial sites but the diaphysis as a whole. Additionally, our understanding of both mechanical and systemic influences in regional and global skeletal remodeling is much better understood since the pioneering days of these early implementers of “Wolff's law” (see Frost, 2003; Bergmann et al., 2011).
On the basis of the findings of Chamay and Tschantz (1972) and Woo et al. (1981), many anthropologists propose that repetitive muscular activity strengthens the tendon-anchoring mechanism as bone cell proliferation increases via capillary expansion, stimulating periosteal modeling and remodeling activity (Torg et al., 1972; Hawkey and Merbs, 1995; Zumwalt, 2006). Weiss (2003, 2004, 2007) has found statistically significant correlations between the robusticity, geometrical properties, and MSMs of long bones, suggesting that tendon insertion development is at least partially dependent on mechanical influences, since it is well established that cross-sectional structure is indicative of such a relationship (Larsen and Ruff, 1990; Ruff et al., 2006). Nevertheless, these analyses fail to identify a definitive affiliation between mechanical strain and insertion morphology since the qualitatively ranked MSM variables are statistically incompatible with the quantitatively calculated geometrical measurements. Additionally, nondemographic systemic influences and potentially impactful characteristics, such as tendon fiber volume independent of contractual strain, remain unaccounted for in these studies.
If nothing else, previous research described above underscores the potential value in quantifying enthesis morphology. However, the degree to which insertion sites respond to load is still inadequately understood. More than a decade since the AAPA symposium, little has been done to address the initial concerns raised by Peterson and Hawkey (1998). Pearson and Buikstra (2006) attribute this to a lack of interest among clinicians since osteophyte formation rarely has a detrimental effect in living humans. To date, Ann Zumwalt (2006) has conducted the most thorough study assessing longitudinal development of entheses in exercised sheep. However, there remains a need for additional well designed experiments assessing enthesial development and what, if any, contribution these markers have in occupation-based research. Despite these shortcomings, there is a wealth of research centered on bone's global and localized response to both mechanical and systemic influences, and a large body of research examining the functional roles of entheses and how they adapt to a changing mechanical environment (Frost, 2003; Thomopoulos et al., 2010; Thomopoulos et al., 2011). This article seeks to synthesize available clinical information within an anthropological context, providing a theoretical foundation from which future research can be grounded.
The following is a review of previous clinical research, emphasizing tendon anatomy, the biomechanical properties of collagen, and the unique osteotendinous interface where stress concentrates and dissipates following muscle contraction. Additionally, the few studies that have directly attempted to address concerns in applying enthesial data to behavior-related anthropological studies are discussed.
Tendon Anatomy and Biomechanics
Muscle tendons are crucial in enabling proper bone function. Tendons transmit tensile loads generated through muscle contraction to bone surfaces, allowing joint motion and posture maintenance. This often requires tendons to distance muscle bellies from joints by transmitting force around corners, thus using bone as a pulley mechanism (Gelberman et al., 1988; Benjamin and Ralphs, 1998; Currey, 2002; Benjamin et al., 2004). Muscles and their associated tendons form muscle-tendon units, serving to restrain dynamic loading (Nordin and Frankel, 2001). Additionally, they have dynamic characteristics enabling repair following injury, and response to variable exercise through tensile strength modifications (Frank et al., 1988; Woo et al., 1988; Benjamin and Ralphs, 1998; Bloebaum and Kopp, 2004). This behavior is comparable to mechanotransduction in bone, wherein osteocytes are recognized to sense and respond to changes in mechanical load and make any necessary microstructural repairs in the tissue.
Microscopically, tendons consist of sparsely spaced fibroblasts surrounded by an abundance of extracellular matrix (ECM). The ECM accounts for ∼80% of connective tissue, which is primarily water (70%) with roughly one-third (30%) consisting of collagen, ground substance, and some elastin. Fibroblasts in tendons arrange themselves in longitudinal rows separated by collagen fibers and other organic minerals (Benjamin and Ralphs, 1998). Collectively, these cells communicate with one another via gap junctions, allowing for the uninhibited diffusion of various ions and molecules (McNeilly et al., 1996). This may provide the method in which tendons detect and respond to stress/strain accordingly, mimicking lacunocanalicular networks between osteocytes and bone lining cells within bone's intracortical matrix (Skerry et al., 1989; Lanyon, 1993; Aarden et al., 1994; Marotti, 1996; Martin, 2000; Cowin and Moss, 2001; Martin, 2003b).
Within the ECM, dense, parallel collagen fibers provide strength and flexibility to tendons. Similar to bone's organic matrix, the predominant component of these tissues is Type I collagen. Collagen is the strongest fibrous protein, providing tendons high tensile strength that surpasses most soft tissues in the human body (Gelberman et al., 1988). This strength is primarily attributable to how tropocollagen, the basic collagen unit, arranges itself. Each tropocollagen molecule is composed of three alpha chains that individually coil in a left-handed helix, and then collectively coil in a right-handed one. This forms a molecule consisting of glycine, praline, and hydroxyproline amino acid sequences, vital to forming proper helixes (Ramachandran, 1963). Additionally, these amino acids aid in crosslinking and staggering collagen molecules, while aggregating them into fibrils. The complex molecular union provides connective tissues with their strength, enabling them to function under stress (Nordin and Frankel, 2001). Aggregation of collagenous materials continues as fibrils cooperatively form fibers. Fiber bundles aggregate forming fascicles ensheathed in loose connective tissue known as endotenon. Fascicles unify forming the tendon, which is similarly ensheathed in epitenon (Fig. 1). Fascicles and tendons slide across one another, thus their associated sheaths are imperative for reducing friction. This movement presumably minimizes failure, allowing structural adaptation to both compressive and shearing forces (Clark and Sidles, 1990; Benjamin and Ralphs, 1998).

Schematic diagram of tendon microstructure demonstrating the intricate bundling of collagen fibrils. This allows for slippage within their respective bundle sheaths to minimize friction during elongation.
As mentioned earlier, the two remaining components of tendons organic matrix are ground substance and elastin. Ground substance primarily consists of highly aggregated proteoglycans that bind extracellular fluid, essentially forming a gelled matrix (Nordin and Frankel, 2001). This amassed material may provide stability to collagenous fibers, contributing to their general strength. The presence of elastin in tendons is minimal, accounting for only about 2% of the matrix, but provides some flexibility to the structure.
The biomechanical properties of tendons are in many respects analogous to bone, as both are viscoelastic materials. Using a load-elongation curve (Fig. 2), tensile deformation occurring in tendons following strain is measurable to the point of tissue avulsion. The connective tissue stretches freely as relaxed wavy collagen fibers straighten, or shear past one another, under applied load (Hirsch, 1974; Viidik et al., 1982; Woo et al., 1994; Currey, 2002). Once fibers have elastically stiffened, the tissue linearly progresses with stress and strain proportional to one another. As cyclic loading continues with elongation surpassing the yield point, tissue deformation becomes plastically nonrecoverable. Here fiber bundles begin to unpredictably fail as the tendon deforms, eventually causing avulsion.

Stress/strain curve for tendon tested to failure. (A) Toe region where wavy collagen fibers begin straightening under initial stress. (B) Linear region where tissue deformation is proportionate to applied load. (C) Yield point where subsequent loading results in small force reductions (dips) as collagen fibers begin failing. (D) Point of maximum loading, followed by rapid failure and avulsion.
Tendons have very high tensile strength with a modulus of around 200 megapascals (Thomopoulos et al., 2011). However, their compressive strength is much lower as they lose shape and collapse. This is attributable to immeasurable fibril beams having small second moments of area, or axial resistance to bending, whereby shear stiffness is low (Currey, 2002). In comparison, bone has considerably more shear stiffness with roughly an equal tensile strength of ∼20 gigapascals (Hems and Tillmann, 2000; Thomopoulos et al., 2011). However, tendon fibers have an elastic modulus ten times smaller than that of bone (Hems and Tillmann, 2000). This variability between soft and hard tissues requires a complex structure, the enthesis, which both balances differing elastic moduli and secures tendons to bone.
Enthesial Design
An enthesis is the interface where tendon meets bone. Entheses, in engineering terms, are sites of stress concentration at the hard and soft tissue junction where mechanical properties differ (Benjamin et al., 2002, 2006). Opposing elastic moduli balance tensile loads, dissipating stress away from the osteotendinous interface and into bone or tendon or both (Biermann, 1957; Knese and Biermann, 1958; Benjamin et al., 2002). In addition to force transmission, another crucial role of entheses is to anchor tendons, enabling static and dynamic load resistance. To achieve this, tendon fibers splay, forming a plexus at the insertion point that provides a firm anchor, equally resistive to insertion angle change in response to variable directional loads occurring during joint movement (Benjamin et al., 2006; Thomopoulos et al., 2006). This enthesial design is commonly related to that of tree roots, noting that both plants and tendons require a relatively small proportion of material for anchorage (Ennos et al., 1993; Benjamin et al., 2006). Often entheses intermingle with one another (e.g., the insertion of vastus lateralis, vastus intermedius, adductor magnus, and adductor brevis mm. along the lateral lip of the linea aspera) overlapping attachment sites for greater tendon security (Benjamin et al., 2004). Additionally, Knese and Biermann (1958) have proposed that the splaying of entheses is not only vital for anchorage, but also in limiting the degree to which a tendon stretches. As tendons stretch they narrow, increasing their vulnerability to rupture.
Entheses may be described as one of two categorical units—fibrous and fibrocartilaginous—depending on tissue type present at the osteotendinous junction. Fibrocartilaginous entheses are only present on the epiphyseal or apophyseal long bone ends, whereas fibrous entheses attach to long bone diaphyses. This distinction between enthesis type and location corresponds to bone origin, either intramembranous or endochondral ossification (Biermann, 1957; Knese and Biermann, 1958; Francois et al., 2001; Doschak et al., 2005). Entheses rooted in thick layers of cortical bone are fibrous, ossifying intramebranously, while those attaching to thin cortical layers are fibrocartilaginous, ossifying endochondrally. Benjamin et al. (2002) suggest this may relate to nutrient foramen access. However, our understanding of early postnatal enthesial development is limited.
Research on osteotendinous development in animal models has yielded some information on the presence of early enthesial signatures and the intricate relationship between the contrasting tissues during long bone maturity. For example, Hurov (1986) described the periosteal surface of diaphysial entheses in fetal rabbit long bones to be coarse-fibered, compositely falling between fibrocartilage and lamellar bone. Additionally, in their research of intermediate filament presence in human enthesial development, Abe et al. (2010) found both the proteins vimentin and desmin to potentially play significant roles in the remodeling of the periosteal surface as mechanical strain is applied both in utero and postnatally to the emerging skeletal scaffold. The periosteum has a separate pathway for growth compared to the associated bone. As the bone elongates at the epiphyseal plates, the periosteum interstially expands in conjunction with the underlying bone (Muhl and Gedak, 1986). As this envelope migrates, so to do the attached tendons, forming a complex union that occurs once the emerging muscle tendons seek out nearby primary ossification centers in bone at approximately the 9th week of fetal development (Moore and Persaud, 2003).
Fibrous entheses, characterized by “fleshy fibers,” attach either directly (Fig. 3) to bone or indirectly via the periosteum (Fig. 4) (Hems and Tillmann, 2000; Benjamin et al., 2002; Benjamin et al., 2006). These entheses are associated with large, powerful muscle bodies, such as the quadriceps group and deltoideus m. (Biermann, 1957). Periosteal attachments dissipate stress over a large expanse of bone, limiting their ability to stretch (Biermann, 1957; Benjamin et al., 2002). Fibrous entheses that lack tendons typically insert dense connective tissue fibers directly into the periosteum, equally allowing stress transmission over a large area. With age, many periosteal fibrous entheses become bony attachments as the periosteum disintegrates over time (Matyas et al., 1990; Benjamin et al., 2002). Mechanical strain transmitted from a tendon to the outer periosteal envelope of bone may be responsible for the morphological indicators used in behavioral reconstructions, such as pit features forming via compressive forces or small protuberances or both developing via tensile forces (Rogers et al., 1987). However, there is little evidence to support this assumption clinically (Kumai and Benjamin, 2002).

Fibrous bony insertion of temporal muscle along inferior temporal line. Note the perforating fibers (arrows) embedded within the interstitial lamellae, but not crossing the cement line of the secondary osteon. Adapted from Hems and Tillmann (2000).

Fibrous periosteal insertion of masseter m. to the medial plate of the zygomatic arch. Note the short tendons interlaced with collagenous periosteal fibrils at the point of insertion. Adapted from Hems and Tillmann (2000).
Fibrocartilaginous entheses typically attach tendons to small, localized regions of bone lacking a thick cortical layer and periosteum. This allows for more precise limb movements about the joint and may potentially dissipate stress as the thin cortical shell deforms under loading (Benjamin et al., 2002, 2006). All of these entheses indirectly insert into bone through a structure consisting of four distinct zones that progressively shift between structural materials (Fig. 5), (Benjamin et al., 2002, 2006; Thomopoulos et al., 2006, 2011). Moving distally towards the insertion site, the first zone is a dense fibrous connective tissue containing Type I collagen and proteoglycans, forming the tendon proper. The second zone is uncalcified fibrocartilage (UF), containing multiple collagen types, with Types II and III most prevalent. The third zone contains calcified fibrocartilage (CF), which is predominantly Type II collagen. This zone serves to anchor tendon to bone, forming a highly irregular junction between collagen fibers and lamellae (Benjamin and Ralphs, 1998; Clark and Stechschulte, 1998). Cartilage that anchors tendons during endochondral ossification remains and calcifies via metaplasia; thus CF is the functional equivalent of collagenous fibers present in fibrous entheses that calcify within interstitial bone (Haines and Mohuiddin, 1968; Matyas et al., 1990; Ishikawa et al., 2001; Benjamin et al., 2006). Separating the UF and CF zones is an avascular calcification front, or tidemark (TM), that serves as a boundary between soft and hard tissues (Benjamin et al., 1986, 2002, 2006). Mineralization at the TM produces a straight, flat surface, minimizing damage as soft tissue insertion angles change from their natural perpendicular approach during movement (Benjamin et al., 1986, 2002; Evans et al., 1990; Thomopoulos et al., 2006). Therefore, the TM reduces tendon wear and tear by promoting a gradual bend in collagen, reducing fiber splay within bone's immediate vicinity (Schneider, 1956; Benjamin et al., 1986; Evans et al., 1990; Benjamin and Ralphs, 1998). Lastly, the fourth zone is bone, containing mostly Type I collagen.

Tracing of entire fibrocartilaginous insertion of biceps brachii m. onto the radial tuberosity. Fibrocartilaginous zones are labeled as the following: compressed fibrocartilage (CFC), uncalcified fibrocartilage (UF), tidemark (TM), and calcified fibrocartilage (CF) embedded in the intracortical bone matrix (B). Adapted from Benjamin et al. (1992).
It appears that a gradual shift from soft to hard tissue may enable efficient load transfer, and reduce stress concentrations (Benjamin et al., 2002, 2006; Thomopoulos et al., 2003, 2011); essentially balancing tensile force between two materials of widely varying elasticity (Hems and Tillmann, 2000). Doschak and Zernicke (2005) suggest these four regions correspond to the transitional zones of increased stiffness outlined on stress/strain curves for tendon. Additionally, tendon failure studies consistently demonstrate the biomechanical efficiency of fibrocartilaginous entheses, with avulsion fractures often occurring within adjacent subchondral bone (Lieber et al., 1992; Lam et al., 1995; Gao et al., 1996; Chu et al., 2003; Thomopoulos et al., 2003). There also may be an association between subchondral avulsion fractures and accumulated microdamage adjacent to entheses (Benjamin et al., 2002). For example, in a study administering bisphosphonates for osteoporosis treatment, a high incidence of avulsion fractures along the long vertebral spines where soft tissue attaches was noted (Hirano et al., 2000). Benjamin et al. (2004) propose a reduction in tendon rupture where an “enthesis organ” is present. They define an enthesis organ as an interface where a subtendinous bursa is present for friction reduction as the insertion angle of collagen fibers changes during joint movement. This allows direct tendon–bone contact (pulley mechanisms), dissipating stress away from the enthesis (Benjamin et al., 2002, 2006), similar to mechanics involved in fibrous attachments.
Osteological Evidence of Entheses
Despite the biomechanical efficiency of entheses in dissipating stress away from the tendon insertion point, wear and tear is inevitable. In dry skeletal material, tendon attachment sites are visible along external bone surfaces. Tissue type present at insertion defines the morphology of an enthesis. Bony fibrous attachments along the diaphyses leave rugous landmarks characterized as raised ridges and roughened bone (Fig. 6). In contrast, periosteal fibrous entheses appear osteologically as smooth markings (Hems and Tillmann, 2000). Fibrocartilaginous attachments are also smooth, often slightly depressed and better circumscribed (Fig. 7), resulting from the TM boundary separating UF from penetrating CF (Benjamin et al., 2002, 2006).

Pronator teres m. insertion along the lateral diaphysis of the radius. This is an example of moderate fibrous enthesis development.

Radial tuberosity. This is a fibrocartilaginous enthesis with a thin cortical shell that presumably flexes under strain generated via biceps brachii m. contraction.
Hawkey and Merbs (1995) fail to distinguish between tissue type in their ordinal ranking method of MSM expression. This is a fundamental flaw considering the histological and biomechanical differences between fibrous and fibrocartilaginous entheses noted above. In their approach, they collectively characterize enthesial morphology as robust, pitted, or ossified. Serially ranking varying degrees of enthesial expression, they proposed that robust markers, or rugged markings (e.g., sharp ridges and crests), are produced through normal activity, resulting from an increase in attachment area to avoid tendon avulsion. Stress lesions or pitting, are stated to result from regular microtrauma at the insertion site that induce periosteal resorption and/or necrosis due to an interruption in the blood supply as tendon fibers avulse and reattach. Lastly, exostosis is attributed to macrotrauma, or complete tendon avulsion, that results in the ossification of tendon tissue during healing. Although this method does not account for varying types of enthesial morphology, it remains the most employed method by bioarcheologists conducting behavioral reconstructions. Many have made small revisions to the methodology, such as removing indicators of exostosis demonstrating less frequently occurring tendon avulsions, or combining the ranked scores of robusticity and pitting to calculate total muscle use (Molnar, 2006; Weiss, 2007). Additionally, Weiss (2003) noted that muscle groups work in tandem when performing activities, and thus enthesial scores should be aggregated to enhance statistical predictability and subsequently assess entheses collectively as biological units. However the core statistical and observational limitations with the method remain, since populational variation among enthesial insertions continues to be accounted for using serial ranking methodology.
The 2008 MSM Workshop in Portugal, following the lead of earlier research conducted by Havelková and Villotte (2007), corrected for the lack of distinction between enthesis type in Hawkey and Merb's method, with the development of methodology focused upon fibrocartilaginous entheses. In their simplified method they distinguish between the stages of fibrocartilaginous enthesis development—those that appear healthy, those with slight indications of enthesopathy development, and those with a large presence of enthesopathies. Since this publication, Villotte et al. (2010a, b) have continued to employ and revise this method, ultimately reducing the intra- and interobserver error to 2 and 4%, respectively. This recording technique distinguishes between varying expanses of tendon fiber insertion along the attachment site, along with morphological signatures found within the distinct smooth attachment surface of fibrocartilaginous entheses.
Interpreting Activity
Anthropologists and clinicians often presume morphological variability of long bone entheses result from strenuous, habitual activity, where muscle contractions are frequently recruited for limb movement (Kennedy, 1983; Hawkey, 1988; Lai and Lovell, 1992; Hawkey and Merbs, 1995; Chapman, 1997; Nagy, 1998; Peterson, 1998; Steen and Lane, 1998; Stirland, 1998; Cook and Dougherty, 2001; Benjamin et al., 2002, 2006; Eshed et al., 2004). However, the relationship between soft and hard tissues, and its effect on enthesis morphology, lacks sufficient scientific testing. Activity reconstructions using entheses profiles are frequently conducted using subjective macroscopic methods that may be irrelevant and/or inaccurate in assessing enthesis morphology (Bryant and Seymour, 1990; Robb, 1998; Wilczak, 1998; Knüsel, 2002). Further concerns arise with the regular failure to account for potential nuances that may affect entheses response to applied load, such as body mass, age, genetic precursors, skeletal maturity, and muscle contractile rates (Stirland, 1998; Wilczak, 1998; Zumwalt, 2006). Moreover, previous studies frequently lack explicit biological justification for why particular parameters are initially chosen for profile assessment (Robb, 1998; Wilczak, 1998; Zumwalt, 2005, 2006).
That larger, more active muscles induce skeletal hypertrophy in relation to tendon insertion sites appears to be a reasonable statement. Therefore it is suggested that enlarged entheses are advantageous, allowing stress applied to the periosteal surface to be proportional within each square unit of surface area, ultimately reducing the effect of contractile forces at the osteotendinous junction (Biewener, 1992; Zumwalt, 2006). This may occur as blood flow increases within the periosteum in response to forceful muscle contractions, initiating the proliferation of bone cells, and ultimately increasing skeletal hypertrophy (Chamay and Tschantz, 1972; Woo et al., 1981; Herring, 1994; Weiss, 2003; Zumwalt, 2006). Support for this hypothesis is often drawn from Dysart et al., (1989), wherein denervation of the deltoideus muscle within a sample rat population is followed by bone resorption at the associated tendon insertion site (deltoid tuberosity). This is an example of subnormal loading, or disuse, which occurs when an applied load is reduced (Frost, 1987; Martin, 2003b). However, remodeling initiated due to the loss of muscle innervation may be more a result of decreased muscle weight via atrophy rather than diminished contractile force. More recent studies using both myostatin-null and dystrophin-null mice, wherein muscle growth continues unregulated, challenge the validity of this relationship between contractile forces and enthesial development (Hamrick et al., 2000; Montgomery et al., 2005). Both studies found enthesis size to correlate with an increase in muscle size, and subsequently body mass, thus their expanse is not necessarily dependent on muscle activity. These findings suggest enthesial development is more attributable to tendon fiber volume rather than tensile forces (Elkasrawy and Hamrick, 2010). Therefore differences in activity patterns among males and females, which are often reported in enthesial-based studies, may be more attributable to sexually dimorphic dissimilarities, rather than behavior related to divisions of labor.
Additionally, if one considers the biomechanical principles of entheses and their role in balancing variable elastic moduli, it would appear this interface between soft and hard tissues is reasonably protected. Therefore, stress magnitudes at the osteotendinous junction resulting from habitual activity may not exceed remodeling thresholds, and thus not require additional morphological changes within the adjacent periosteum. In fact, according to Frost's (1987, 2003) mechanostat—a negative feedback loop that guides bone's mechanical competence by providing a system that detects and corrects for strain stimuli—if entheses were to significantly respond to applied strains via increased periosteal remodeling, the mechanical loads would have to be outside the realm of normal activity levels and so infrequent as to not shift the threshold range to accommodate the new strain levels. This begs the question of whether rugous enthesial morphology is indicative of accumulated microtrauma resulting from habitual activity, or is it more reflective of sporadic strenuous mechanical loads that occasionally exceed remodeling thresholds, inducing a more drastic remodeling response?
Similar to bone remodeling thresholds, age has been shown to compound the concerns associated with uncovering enthesis etiology (Hawkey and Merbs, 1995; Weiss, 2003; Zumwalt, 2006). If normal daily activity falls within an individual's predetermined remodeling threshold, thus serving to maintain bone's mechanical competence, then the degree of insertion site development may be more indicative of periosteal degradation following regular stress accumulated over one's lifetime via normal long bone movement, rather than elevated strains generated through physically taxing exercises. It is well known that subadults demonstrate more periosteal apposition during skeletal development compared to that of adults who have achieved skeletal maturity (Frost, 1997; Ruff et al., 2006). However, as stress accumulates intracortically in long bones via physical activity, some bone modeling may occur with new bone formation along the periosteal envelope in conjunction with bone resorption endosteally. In this event bone apposition is minimal, and thus significant periosteal changes via high strain activity would not be expected. This is consistent with Hawkey and Merbs (1995) observation that in their Thule Eskimo skeletal collection enthesial insertions become more rugous with age, yet statistically these differences are insignificant.
Conversely, a number of studies have found age to significantly correlate with enthesis size (Weiss, 2003, 2007; Belcastro et al., 2006; Cardoso and Henderson, 2010) suggesting that morphological development may be partially indicative of life-long periosteal wear and tear following routine activity. However, this conclusion is confounded by the issue of assessing age, since most of these studies consisted of archeologically excavated individuals of unknown age and sex. In the interest of discovering etiological markers of enthesial development, initial studies should consist of skeletal populations with known demographics (i.e., recorded cemetery plots or donated medical cadavers) to control these factors and potentially discover age-related correlations applicable in use with archeologically assessed populations.
Further complications in interpreting activity patterns from osteological markers arise when considering the affects systemic influences have on bone remodeling. The role of genetic factors in enthesial development is poorly understood. Frost (2003) suggests that genetics predetermine the baseline conditions in bone in utero. These baseline conditions are predominately driven by Hox gene expression, which current research postulates is responsible for canalized limb patterning, basic neuromuscular and physiologic anatomy, and the biologic machinery necessary for increasing bone strength following birth (Shubin et al., 1997; Capdevila and Belmonte, 2001; Chiu and Hamrick, 2002). Lovejoy (2002) also suggests that genetic precursors play a more significant role than environmental factors in the primary development of morphological indicators on bone. Earlier prenatal observations support these findings with the appearance of distinguishable hypotrochanteric fossa on the femoral diaphysis where a portion of the gluteus maximus inserts (Hrdlicka, 1934). In fact, Enlow (1990) noted that enthesial development is insignificant during adolescence, suggesting that mechanical strain during growth and development merely provides threshold levels for bone growth. Therefore, threshold levels set during adolescence preserve throughout adult life with mechanical strain maintaining bone strength; thus the focus is on bone retention not deposition (Lovejoy et al., 2003). Hypothetically, these observations account for deposition and resorption reduction observed in adult long bones when exposed to increased strain levels, since mechanical mechanisms are bound to predetermined threshold levels set during initial growth and development. Accordingly, proponents of this position suggest that genetics dominate long bone morphogenesis, and that environmental factors have little impact on skeletal development.
Regardless of whether mechanical loading or genetic regulation is predominately responsible for bone development and maintenance, it is apparent that a complex interaction exists between these two factors that cannot be ignored when attributing morphological signatures to physical activity. Especially since it is unlikely that individuals or populations osteologically respond to similar physical activities in the same manner.
Equally of concern in interpreting enthesial data is the influence hormones have on bone remodeling in localized regions associated with tendon insertion sites. Although direct links between hormone levels and entheses are unknown, there is a wealth of research demonstrating the global effects hormones have on bone remodeling. Osteoblasts, which are responsible for producing the proteins necessary for formation of bone's organic matrix and controlling the mineralization of bone, provide a good example for why the contribution of hormones in terms of skeletal maintenance cannot be ignored when reconstructing behavior from skeletal elements. Influencing osteoblastic functions is a series of hormones, for which these cells have receptors. For example, a normal level of estrogen in pre-menopausal women appears to maintain bone formation by upregulating osteoblasts and thus increasing endosteal bone formation (Frost, 1992; Martin, 2003b; Pearson and Lieberman, 2004). However when estrogen is depleted, primarily in conjunction with age, osteoblasts are downregulated resulting in a reduction of new bone forming teams; thus endosteal bone formation is halted (Weaver, 1998). Similarly in men, a normal level of testosterone allows for periosteal bone formation, with a decrease in level ultimately leading to lower bone mass (Martin, 2003b).
Additional hormones, such as vitamin D and those associated with the parathyroid, also influence osteoblasts and may act independently or in tandem with estrogen. Parathyroid hormone indirectly stimulates bone formation when it is coupled with bone resorption, and vitamin D inhibits osteoblasts by directly influencing Cbfa1 expression; which is responsible for mesenchymal cell differentiation into mature osteoblasts and the rate of bone secretion (Pearson and Lieberman, 2004). Thus the roles of hormones in bone remodeling is complex, and must be considered when reconstructing past behaviors, especially those presumed to have differed between the sexes.
Recent research has sought to resolve some of these uncertainties in enthesial interpretations using objective, metrically quantified insertion profile data. Weiss (2003) correlated aggregated ranked scores of enthesial expression from seven muscles in the upper limbs of 91 Native British Columbians with body mass, age, and long bone robusticity (composite of cross-sectional geometric properties). She found all three variables to significantly correlate with insertion development; age being the best predictor. These findings suggest muscle insertion development is intricately linked with age, skeletal build, and overall body mass. Schlecht's (2004, 2008) analysis of an Anglo-Saxon and Norman cemetery population revealed a significant correlation between the length and width of the linea aspera and associated femoral geometric properties when controlling for age and body mass. Additionally, a recent study conducted by Weiss (2012) found that among a pre-European (2180–250 BP) contact population in California, sex differences interpreted from standardized aggregate MSM scores were more indicative of variable body mass than perceived gender role differences.
Zumwalt (2006) has conducted the most direct investigation exploring the presumed relationship between enthesial development and normal physical activity. Using 20 adult female sheep at least 4 years of age, she trained 10 to run on treadmills 1 hr a day, 5 days a week for 90 days. The 10 remaining sheep were used as a control and were only exposed to normal enclosure activity. At the end of the experiment, all 20 sheep were euthanized and six entheses of the fore and hind limbs were dissected. Using a sophisticated method of capturing three-dimensional enthesis profiles followed by fractal analysis quantification, Zumwalt found little difference in insertion site development within the exercised population and the control group. This suggests that within her population of adult sheep enthesial morphology is not indicative of muscle size or activity. However, future experiments are needed to explore skeletal development amongst growing subadults, wherein the macro-structural properties of bone are much more malleable (Ruff et al., 2006).
Animal models lend themselves favorably to experimental research into how entheses develop in response to variable loading. Not only do they allow for skeletal exploration via dissection, but they also provide necessary controls such as age, body mass, diet, genetics and activity regime. Various animal species have regularly been incorporated into research centered on bone's response to environmental influences since certain mammalian species, particularly primates, mice and canine, demonstrate similar biological mechanisms present in human skeletal tissue (Currey, 2002). However, there are some key differences that must be considered. First, humans have a much longer period of skeletal maturation, potentially impacting the degree to which bone responds to environmental forces. Second, the human skeleton is subjected to a larger degree of fatigue damage due to our flexible limb structure and considerably long life spans. Third, smaller mammals demonstrate little to no Haversian remodeling during normal loading regimes (e.g., murine models).
In light of the variable results derived from these pioneering studies seeking to find direct correlations between tendon insertion sites and activity, incorporation of enthesial data in activity-related studies should be used cautiously, if at all. This concern is particularly heightened since whether enthesial morphology significantly remodels following the cessation of skeletal growth remains unknown.
Future Directions
To further investigate the relationship between applied mechanical strain and enthesial development, histomorphometric approaches examining bone turnover history should initially be employed. Histomorphometry, a method yet to be incorporated into insertion development studies, may conclusively resolve the controversy surrounding enthesis etiology. If tensile forces are responsible for enthesis development, then a statistically significant relationship with osteon population densities (OPD), or the number of intact and fragmentary secondary osteons per unit area, and secondary osteon area, should be exemplified.
OPD reflects the visible remodeling history of compacta in bone with strains that exceed one's remodeling threshold generating an increase in targeted remodeling. This is based on the premise that if an osteonal remodeling event is initiated by the emergence of microdamage, then the activation frequency of these events should be proportional to the rate of fatigue damage in individuals, once normalized for any metabolic and/or hormonal effects (Frost, 1969; Robling and Stout, 2000; Parfitt, 2002). Therefore as microdamage increases, basic multicellular unit activation also increases (Martin et al., 1998), maintaining bone's mechanical competence while extending its fatigue life (Frost, 2003; Martin, 2003a). Measuring the mean activation frequency, once adjusted for age, should then be indicative of one's physical activity level, or mechanical loading history. Thus, higher activation frequencies suggest greater physical activity.
Osteon area reveals strain level, thus smaller osteons should be abundant in regions of tendon insertions for increased intracortical fatigue resistance. Completed osteons in humans range from 150 to 350 μm in diameter, which is dependent upon the size of the osteoclast cutting cone (van Oers et al., 2008). Frost (1990) originally proposed an inverse relationship between osteon diameter and the magnitude of applied strains, with larger osteons concentrated within the endosteal envelope where strain levels are lower, and smaller osteons periosteally where strain magnitudes increase. Research conducted by Skedros et al. (2001) support this hypothesis, concluding that in artiodactyl (even-toed ungulates) calcanei larger osteons are more prevalent in regions undergoing tension opposed to more heavily strained compressed areas.
Van Oers et al. (2008) employed computational models to demonstrate that osteoclastic bone resorption is inhibited by osteocyte signals generated via strain, resulting in osteons with smaller diameters in cortical regions subjected to high strain magnitudes. The authors' propose three explanations for this inverse relationship between loading magnitude and osteon size. First, generation of new osteons will produce a smaller cutting cone, minimizing the effects of porosity, and thus the overall strength of the intracortical matrix. Second, dense concentrations of smaller osteons will absorb energy more efficiently than larger ones as the effects resulting from accumulated microdamage are reduced, with cement lines serving to limit crack propagation (Sobelman et al., 2004; O'Brien et al., 2005). Third, smaller osteons in regions of high strain may enhance the resistance towards osteon pullout or debonding of the osteon along the cement line surface (Skedros et al., 2007).
Following the conclusions of the above studies, regions of bone with an elevated OPD should have osteons with smaller diameters. Therefore, if enthesial morphology is indicative of physical activity, then a pattern should emerge intracortically, wherein bone turnover is elevated in response to strain dissipated through the osteotendinous interface. This may occur in direct relation with the enthesis, or away from the point of insertion along the diaphysis. The latter scenario is more likely since the enthesis mechanism is very complex, and presumably designed to dissipate generated forces away from the osteotendinous junction. If a significant correlation is demonstrated between remodeling rate and tendon insertion site, then use of macroscopically quantified enthesial data for behavioral reconstruction of past living populations is one step closer to being validated. This would allow further, more direct investigation of the relationship between intracortical remodeling and quantified enthesial morphology.
CONCLUSIONS
Enthesial morphology has increasingly gained prominence in studies investigating skeletal signatures of occupational stress in both modern and past human populations. However, the relationship tendons have with bone, especially within the enthesial interface, is not as straightforward as has been generally assumed. Our knowledge of these mechanisms has steadily increased in recent years, yet there are still many gaps in understanding how entheses develop and maintain themselves throughout their functional life. While there is an extensive body of literature documenting skeletal responses to mechanical loads, there are fewer comprehensive studies addressing this relationship within discrete tendon insertion sites frequently incorporated into occupation-based studies.
This lack of scientific validation directly applied to the osteologic portion of the enthesial mechanism leaves a number of questions unanswered. Do tensile loads primarily influence enthesial morphology, or is their form more indicative of muscle size, and subsequently body mass? Do fibrocartilaginous and fibrous entheses, which differ in their biologic construction, respond to mechanical loads in a similar manner? Are contractile strains dissipated along the diaphysis or concentrated at the point of tendon insertion, generating an adaptive response to reinforce the osteotendinous interface? Is habitual repetitive activity, which is presumably within bone's normal modeling/remodeling threshold, sufficient to initiate an adaptive response? If entheses abide by baseline threshold levels outlined in the mechanostat, then is their morphology more indicative of activity levels prior to skeletal maturity or infrequent loads that exceed normal threshold levels after maturity or both? Is enthesial morphology reflective of microtrauma and periosteal degradation accumulated with age, rather than of differing activity regimes within a population? Histomorphologic investigations into bone turnover rates of regions directly associated with tendon insertion sites would be a novel starting point for addressing the role mechanical loading plays in the development and maintenance of this complex tissue interface.
With so many uncertainties concerning potential mechanical, systemic, and genetic influences in tendon insertion site development and maintenance, inclusion of enthesial size and morphological expression in occupationally related research must be conducted cautiously. Before anthropologists can reliably interpret activity levels from varying degrees of enthesial expression, we must first account for other plausible influences in their development. Moreover, we must understand the underlying biologic mechanism responsible for bone functional adaptation, while recognizing the biomechanical relationship tendons share with bone in dissipating stress. Therefore, innovative investigations into factors clouding this issue are needed before the value of enthesial data in activity reconstructions can be fairly assessed. Finally, signatures of repetitively applied strain are not restricted to the periosteal envelope. Thus, to truly assess the relationship between entheses and mechanical load, bones' compacta, where the real scars of physical activity may be found, must not be ignored in future examinations.
Acknowledgements
The author thanks Drs Sam Stout, Clark Larsen, Britney Kyle McIlvaine, and Laurie Reitsema for their invaluable advice on earlier versions of this manuscript.