Characterization of the Anchoring Morphology and Mineral Content of the Anterior Cruciate and Medial Collateral Ligaments of the Knee

Authors


Abstract

The manner in which ligament connects to bone remains an area of interest for researchers, bioengineers, and clinicians. Stable fixation of an anterior cruciate ligament (ACL) graft has been shown to be paramount to preventing excess anterior tibial translation and to restoring the normal kinematics of the knee joint. In this study, the surface area of attachment and the mineral characteristics of the ACL and medial collateral ligament (MCL) attachment sites were characterized to determine the factors that contributed to ligament attachment strength. Findings from this study indicated that the area of attachment of the ACL's insertion was significantly greater than the ligament's origin (95.8 mm2 ± 21.5 vs. 73.2 mm2 ± 16.2, P = 0.009). Additionally, the ACL was measured to have a greater surface area of attachment when compared with the MCL (84.5 mm2 ± 18.8 vs. 58.2 mm2 ± 23.8, P = 0.005); although, the MCL was observed to have a greater region of calcified fibrocartilage (CFC) than the ACL (533.0 μm ± 116.9 vs. 195.5 μm ± 36.6, P = 0.0003). No significant correlation was observed between the ligament's area of attachment and the thickness of the CFC region. Measurements of ash percent suggested that the boundary region, between the CFC and host bone, possessed the least mineral content for the three regions of interest. These data suggest that ligament attachment strength can be attributed to several factors, including the ligament's area of attachment, regional thickness, and mineral content of the CFC. Anat Rec,, 2011. © 2011 Wiley-Liss, Inc.

Anterior cruciate ligament (ACL) injuries exceed 300,000 incidences per year, with nearly half requiring surgical reconstruction (Gotlin and Huie,2000; Pennisi,2002; Vunjak-Novakovic et al.,2004). Multiligament injuries, involving the ACL and the medial collateral ligament (MCL), have been estimated to account for 11–38% of these ACL reconstructions (Duncan et al.,1995; Kaeding et al.,2005). Although most MCL injuries demonstrate the capacity to self-heal, the ACL lacks this intrinsic property and often requires surgical intervention. Current reconstructive practices utilize autograft and allograft ligament sources. Although these approaches have proven to be initially successful, neither approach has currently been shown to be capable of preventing anterior tibial translation, restoring normal kinematics, and regenerating the ligament to bone insertion site without concomitant complications in long-term clinical follow-up studies (Meunier et al.,2007; Kessler et al.,2008; Sinclair et al.,2008,2010; Oiestad et al.,2010; Pernin et al.,2010; Sutherland et al.,2010). Because of the limited ability of the reconstructed ligament to integrate with the host bone, there is a need to understand the microanatomy of these ligaments and their insertions.

Table 1. P values for corresponding comparisons between the ACL, MCL, and regions of interest
LigamentAttachmentRegionAsh percentP value
  1. The statistical analyses were done using paired and unpaired t-tests (Stata Corp). The symbol “*” indicates a statistically significant difference between comparison regions. The table illustrates that there was no significant difference in ash percent between CFC and host bone regions. The table also illustrates that the boundary region possessed significantly lower ash percent values than the CFC and host bone regions, the only exception being in the ACL insertion site.

ACLOriginCFC65.60.012*
Boundary63.9 
CFC65.60.399
Host bone65.4 
Boundary63.90.021*
Host bone65.4 
InsertionCFC66.00.021*
Boundary63.5 
CFC66.00.869
Host bone65.6 
Boundary63.50.080
Host bone65.6 
MCLOriginCFC65.70.021*
Boundary63.3 
CFC65.70.585
Host bone66.1 
Boundary63.30.012*
Host bone66.1 
InsertionCFC66.90.034*
Boundary64.9 
CFC66.90.315
Host bone67.9 
Boundary64.90.021*
Host bone67.9 

To restore normal joint kinematics and ensure the long-term viability of an implanted graft, it is crucial to understand the physical properties, mineral content, and microanatomy of the intact ligament attachment site. Differences in morphology and microstructure of ligaments and tendons have been investigated extensively using various techniques such as light microscopy, magnetic resonance imaging, scanning electron microscopy (SEM), and transmission electron microscopy (Cooper and Misol,1970; Benjamin et al.,1986,1991; Evans et al.,1991; Gao and Messner,1996; Gao et al.,1996; Rufai et al.,1996). Numerous studies have described the presence of four distinct zones at the ligament to bone interface, commonly referred to as the enthesis (Schneider,1956; Cooper and Misol,1970; Benjamin et al.,1991; Gao and Messner,1996; Benjamin and Ralphs,1998; Moffat et al.,2008). These zones originate from mid-substance and are composed of the ligament, uncalcified fibrocartilage (UFC), calcified fibrocartilage (CFC), and host bone. The fibrocartilage zones function to dissipate perceived stresses (Gao and Messner,1996), to reduce ligament taper (Schneider,1956), and to improve resistance to failure by providing a gradual transition in chemical and mechanical properties from flexible ligament to rigid bone (Schneider,1956; Benjamin et al.,1986).

Apart from the evident differences in mineral content, UFC and CFC differ in many other aspects. Collagen fibers observed in the CFC region are reported to lack the crimp pattern that was observed in the ligamentous and UFC regions (Cooper and Misol,1970). Furthermore, these collagen fibers do not cross the CFC to bone boundary, also known as the cement line (Gao et al.,1996; Benjamin et al.,2002). It has been observed that collagen fibers do cross the UFC to CFC tidemark (Subit et al.,2008).

The ACL and MCL do indeed vary significantly in mechanical properties (Gao et al.,1996). Mechanical testing of human ligaments, performed en bloc (Woo et al.,2000), has demonstrated ultimate strength values on the order of approximately 2,100 N and 1,600 N for the ACL and MCL, respectively. However, differences in CFC thickness at the ACL and MCL entheses may help to explain variations in attachment strength recorded at these sites. Differences in other properties of the ligament attachment site, such as surface area and mineral content, may also contribute to the ligament's ultimate strength. To date, no studies have investigated associations between the surface area of ACL and MCL attachments and CFC thickness in either the human or animal model knee joint.

Although numerous studies have investigated ligament attachment in human and animal models (Cooper and Misol,1970; Evans,1991; Gao and Messner,1996; Rufai et al.,1996; Benjamin and Ralphs,1998; Moffat et al.,2008; Subit et al.,2008), none have applied backscattered electron microscopy (BSE) technology with weighted mean gray level (WMGL) measurements to quantitatively analyze (Bloebaum et al.,1990,1992,1994,1997a,b,1998; Moffat et al.,2008) the anatomical microstructure and mineral content of the attachment site. The goal of this investigation was to determine if there were differences in structure and mineral content between the origin and insertion sites of ACL and MCL through the analysis of three facets of ligament attachment: (1) surface area of the attachment site, (2) CFC thickness, and (3) mineral content using BSE and WMGL techniques previously published by Bloebaum et al. (1997c; see also Skedros et al.,1993a,b; Vajda et al.,1996,1995).

To investigate the ligament attachment sites of the ACL and MCL, using the commonly adopted sheep model for knee ligament characterization and ACL reconstructions (Seitz et al.,1997; An and Friedman,1999; Shea et al.,2002; Hunt et al.,2005; Burger et al.,2007; Rumian et al.,2007; Meller et al.,2008a,b; Scheffler et al.,2008; Tapper et al.,2009), this study tested three hypotheses to better understand the nature of ligament attachment within the knee. The first hypothesis challenged that the thickness of the CFC layer would vary significantly between the origin and insertion sites in the ACL and MCL. The second hypothesis tested was that the area of attachment would show a positive correlation between the thickness of the CFC layer in the origin and insertion sites of both the ACL and MCL. This was based on the rationale that these two variables may work together to enhance the ligament's attachment strength. The third hypothesis stated that there would be no significant difference between the mineral content in the three different regions (CFC, boundary region, and host bone; Fig. 1) of the origin and insertion sites in both the ACL and MCL.

Figure 1.

BSE image illustrating varying gray levels (parasagittal, ×200). The regions in white represent higher amounts of mineral content, and the darker gray and black regions represent areas of lower mineral content. The image also illustrates the difference in lacunae structure of resident chondrocytes in the CFC (white arrows) compared to the smaller osteocyte lacunae (black arrows) of the host bone.

MATERIALS AND METHODS

Specimen Preparation

Twelve sheep knees from postmortem skeletally mature ewes were obtained for ligament analysis (seven left knees and five right knees). The sheep model was selected for this work because of its known similarities to human bone, fibrocartilage, and tendon attachment (Seitz et al.,1997; An and Friedman,1999; Shea et al.,2002; Hunt et al.,2005; Burger et al.,2007; Rumian et al.,2007; Meller et al.,2008a,b; Scheffler et al.,2008; Tapper et al.,2009). It should be noted that the specimens came from previously sacrificed animals and, therefore, were considered exempt by the University's Institutional Animal Care and Use Committee. The specimens were dissected to remove the soft tissues around the joint including the menisci. Following dissection, only the ACL and MCL remained intact. The MCL and ACL were then excised along with the bony attachments using an oscillating bone saw. Wide margins of bone were excised at the attachment sites to ensure that the entire attachment area was left undisturbed. Ligaments were then cut mid-substance to avoid disturbing the chondral attachment zones. Additionally, the ligaments were cleaned of all synovial tissue to allow for unobstructed observations of the ligament structure. The specimens were then fixed in 70% ethanol, dehydrated in ascending grades of ethanol (70%–100%), and cleared in xylenes.

Surface Attachment Area Analysis

Each specimen was examined using a macroscope (Nikon SM 2800; Nikon, Melville, NY), and images were digitally recorded using computerized image analysis software (Image Pro Plus; Media Cybernetics, Bethesda, MD). The origin and insertion sites were identified, and the perimeter was outlined. From these data, the area of ligament attachment (mm2) was quantified.

CFC Thickness Analysis

Following the measurements of area of attachment, each of the specimen were embedded in poly(methyl methacrylate) (PMMA) using standard techniques (Emmanual et al.,1987; Emmanual,1988; Sanderson and Kitabayashi,1994; Sanderson,1995). The specimen were then cut into blocks using a vertical band saw, and the blocks were then cut down to 2-mm-thick parasagittal sections using a diamond disk (Leco Corporation, St. Joseph, MI) and Lapidary Slab Saw (Lortone, Mukilteo, WA) to allow visualization of the enthesis at intervals of ∼2 mm. The blocks were manually ground and polished to an optical finish using a variable-speed rotary grinding wheel (Buehler Incorporated, Lake Bluff, IL) and standard techniques (Bloebaum et al.,1989). The samples were sputter coated with a conductive layer of gold for 90 sec (Hummer VI-A; Anatech, Alexandria, VA). Each specimen was examined by a SEM (JSM 6100; JEOL Incorporated, Peabody, MA) equipped with a BSE detector (Tetra; Oxford Instruments, Cambridge, UK) and associated image capture software (ISIS; Oxford Instruments). Digital BSE images were recorded of the attachment sites for the MCL and the ACL using a Kalman frame-averaging technique. These images allowed for clear distinction between areas composed of UFC, CFC, and host bone (Fig. 1). Images were then analyzed with image analysis software (Image Pro Plus; Media Cybernetics, Bethesda, MD). The CFC thickness was obtained by tracing the distal and proximal boundaries (tidemark and cement line) and measuring the distance (μm) between each.

Mineral Content Analysis

Following BSE analysis for CFC thickness, the specimens were repolished to remove the high atomic number (Z) conductive gold (Z = 79) coating. Each block was then carbon (Z = 6) coated, within 2 days of mineral content imaging, for 15 sec at 4.5 V (Cressington 208C; Ted Pella, Redding, CA). The lower atomic number of carbon, unlike the high atomic number of gold, ensures the BSE detection of the lower atomic number mineral levels in bone (Zavg ≈ 10–12; Bloebaum et al.,1989,1997c; Boyce et al.,1990; Skedros et al., 1993). BSE imaging was conducted using a JEOL 6100 SEM and Oxford ISIS 300 imaging software. Operating conditions included an accelerating voltage of 20 kV, working distance of 15 mm, and probe current of 0.75 nA. Brightness and contrast were kept constant throughout the entire study. Backscattered electrons were captured by a solid-state lithium BSE detector (Tetra, Oxford Instruments, Buckinghamshire, UK). Digital BSE images taken at 200× magnification with a resolution of 512 × 416 pixels and 8 bits/pixel (256 distinct gray levels) were captured using a Kalman frame-averaging technique. Calibration was performed at 20-min intervals using pure aluminum and carbon standards (Tousimis, Rockville, MD) and two bone standards consisting of whale tympanic bulla and deer antler (Vajda et al.,1996). Pure element standards (carbon and aluminum) were used for calibration because they have been shown to correct for fluctuations in the electron beam of the microscope (Skedros et al.,1993a,b; Vajda et al.,1995,1996; Bloebaum et al.,1997c).

BSE has been shown to measure mineral content accurately (to 2.4%) in bone (Boyce et al.,1990; Vajda et al.,1995). For mineral analysis, backscattered electron images were captured, in parasagittal orientation, from the ligament attachment regions immediately proximal and distal to the interface between CFC and bone, dividing the specimens into three regions: CFC region, boundary region, and host bone region (Fig. 2). Five images were taken from random sites in each region (CFC, boundary region, and host bone) and from each specimen.

Figure 2.

Parasagittal sections of the medial collateral ligament attachment site viewed using scanning electron microscopy techniques. (A) Secondary image illustrating the interdigitation of the MCL with bone at the tibial insertion: (1) MCL; (2) region of interdigitation; and (3) host bone. (B) Backscatter image of the same region: (1) calcified fibrocartilage, cement line (white arrows) and (2) host bone. The host bone was cancellous structured in morphology.

Images of the element standards and the bone standards were captured along with the 15 images of each specimen. Images were later analyzed and the weighted mean gray level was measured from five random sites on each image (Scion Image; Scion Corp, Frederick, MD; Boyce et al.,1990; Skedros et al.,1993a,1993b; Vajda et al.,1995,1996,1998). Mineral content, reported as ash percent, was calculated from calibrated WMGL measurements using a linear regression model of bones (antler and bulla) with a known ash content as previously published (Boyce et al.,1990; Vajda et al.,1996; Bloebaum et al.,1997c). Ash percent was expressed as ashed bone weight (WAB) divided by dry, defatted bone weight (WDB) multiplied by 100 [(WAB/WDB) × 100] for BSE imaging (Bloebaum et al.,1997c).

Statistical Methods

The area (mm2), CFC thickness (μm), and calibrated WMGL measurements from ACL and MCL attachments were averaged to obtain a single regional value per sheep (n = 8). Because the area percent, depth, and mineral content values were normally distributed, differences between the means were compared using independent sample t-tests. The P-values were adjusted for multiple comparisons using Finner's multiple procedure (Finner,1993), which controls the Type I error without the need to first test the global hypothesis with ANOVA. All reported P-values for these comparisons represented multiple comparison adjusted P-values. Differences were considered significant with a P-value of <0.05. Correlations were performed using Pearson's coefficient of correlation tests. All statistical calculations were performed using Intercooled Stata v. 8.0 (Stata Corporation, College Station, TX).

RESULTS

Surface Attachment Area

The surface area of attachment for the ACL origin (73.2 mm2 ± 16.2) was significantly less (P = 0.031) when compared with the surface area of the ACL insertion (95.8 mm2 ± 21.5; Fig. 3). There was no significant difference (P = 0.344) when comparing the area of attachment for the MCL origin (54.4 mm2 ± 28.3) with the surface area of the MCL insertion (62.0 mm2 ± 19.3). The surface area of attachment for the ACL origin (73.2 mm2 ± 16.2) was significantly larger (P = 0.049) when compared with the surface area of the MCL origin (54.4 mm2 ± 28.3). The surface area of attachment for the ACL insertion (95.8 mm2 ± 21.5) was significantly greater (P = 0.031) when compared with the surface area of the MCL insertion (62.0 mm2 ± 19.3). In summary, the surface area of attachment values recorded for the ACL (origin 73.2 mm2, insertion 95.8 mm2) were significantly greater (P < 0.05) than those recorded for the MCL (origin 54.4 mm2, insertion 62.0 mm2).

Figure 3.

Graph illustrating the attachment area for the origin and insertion sites of the ACL and MCL. The surface areas measured for ACL attachment, both origin and insertion, were greater in value than those of the MCL (“*” indicates significant difference).

CFC Thickness

BSE showed that the thickness of the CFC at the ACL origin (187.3 μm ± 35.7) was similar (P = 0.539) to that of the ACL insertion (197.6 μm ± 37.4; Fig. 4). The thickness of the CFC at the MCL origin (623.8 μm ± 116.2) was significantly greater (P = 0.023) when compared with the thickness of the CFC at the MCL insertion (442.2 μm ± 117.6). The thickness of the CFC at the ACL origin (187.3 μm ± 35.7) was significantly less (P < 0.0001) when compared with the thickness of the CFC at the MCL origin (623.8 μm ± 116.2). The thickness of the CFC at the ACL insertion (197.6 μm ± 37.4) was significantly less (P = 0.001) when compared with the thickness of the CFC at the ACL insertion (197.6 μm ± 37.4). In summary, the CFC thickness of the MCL attachments (origin 623.8 μm, insertion 442.2 μm) was significantly greater (P < 0.05) than the CFC thickness of the ACL (origin 187.3 μm, insertion 197.6 μm).

Figure 4.

Graph illustrating the thickness of the CFC region at the origin and insertions sites of the ACL and MCL. The figure shows that the CFC thickness was greater at the MCL attachment sites than the ACL attachment sites (“*” indicates significant difference).

Mineral Content

The mineral content was reported as ash percentage (Table 1). There were three regions measured (CFC region, boundary region, and host bone region) for the ACL and MCL origin and insertion sites. The origin site of the ACL had ash percent values of 65.6% ± 1.8, 63.9% ± 2.4, and 65.4% ± 2.4 for the CFC, boundary region, and host bone regions, respectively (P > 0.05; Fig. 5). The insertion site of the ACL had ash percent values of 66.0% ± 2.9, 63.5% ± 2.4, and 65.6% ± 2.2 for the CFC, boundary region, and host bone regions, respectively (P > 0.05). The origin site of the MCL had ash percent values of 65.7% ± 1.3, 63.3% ± 2.2, and 66.1% ± 2.7 for the CFC, boundary region, and host bone regions, respectively (P > 0.05). The insertion site of the MCL had ash percent values of 66.9% ± 2.2, 64.9% ± 2.0, and 67.9% ± 2.2 for the CFC, boundary region, and host bone regions respectively (P > 0.05). When comparing the mineral content between regions, the boundary regions had a lower mineral content when compared with their respective CFC and host bone regions. The CFC region in the ACL attachment sites had a higher mineral content than the host bone, whereas the CFC region in the MCL attachment sites had a lower mineral content than the host bone. Furthermore, these data served to demonstrate the composite nature of the ligament's CFC region, its intrinsic ability to resist fatigue failure, and ability to provide toughness at the attachment site.

Figure 5.

Graph illustrating ash percents for the three regions (CFC, boundary, host bone) for the ACL and MCL origin and insertion sites. Results suggest that the boundary region had the lowest ash percent when compared with the CFC and host bone regions for all origin and insertion sites. The ACL host bone appeared to have a lower ash percent than ACL CFC region, although there was no statistical difference. MCL host bone data suggest higher ash percent than the MCL CFC region.

DISCUSSION

This work demonstrated that differences in structure and composition exist between the origin and insertion sites of the ACL and MCL of the sheep knee. These data also indicated that there were differences in how the ACL and MCL used different biological processes to obtain skeletal attachment. These structural and compositional differences at the attachment sites likely result from regionally dependent loads experienced by the ligaments of the knee throughout its normal range of motion. The complex, multidirectional loads of the ACL require different anatomical attachment than the MCL that is subject to significantly lower and more simplistic loading conditions (Wijdicks et al.,2010; Yang et al.,2010).

Initial characterization of the ACL indicated a greater surface area of attachment at the ligament's site of insertion when compared with its origin in the sheep model used for this work (Fig. 3). These findings were comparable with those of Butler et al. (2004) who observed the same trends in the surface areas of attachment within the human knee. They surmised that less strength was required at the ACL's site of origin owing to the fact that the majority of loads acting on the ligament were dispersed at the tibial insertion site.

The MCL, in contrast to the ACL, possessed a similar surface area of attachment at its origin and insertion sites. Furthermore, the surface areas of attachment for the MCL's origin and insertion sites were significantly less than those of the ACL. These findings, consistent with results reported by Gao and Messner (1996) and Gao et al. (1994,1996), suggest that the larger surface area of the ACL attachment imparts the ligament with a greater ability to dissipate loads in vivo. The large surface area of the ACL insertion site allows the ligament to accommodate nonuniform loading conditions through the broad and varied attachment of the ACL to the subchondral bone (Duthon,2006).

Characterization of the CFC thickness demonstrated that the MCL's site of origin possessed significantly greater CFC thickness than that was observed at the ligament's insertion site. The same significance was not observed with the ACL. There was no statistical difference between the CFC thickness of the ligament's origin and insertion sites in the ACL. The increased thickness of the MCL's CFC region was attributed to the need for a decrease in localized stresses due to the reduction in surface area of ligament attachment (Gao and Messner,1996). Furthermore, it has been noted that the greater thickness of the MCL's CFC region functions to provide the ligament with effective load transfer from the ligament to the bone under normal loading conditions, without concomitant disturbance to the attachment site (Schneider,1956; Benjamin et al.,1986).

Numerous studies have offered theories to explain the role of CFC morphology at the ligaments' insertion sites. Variations in CFC properties, consistent with those observed in this study, have been associated with the forces acting on the ligament (Evans et al.,1991; Benjamin and Ralphs,1998), the amount of movement occurring at the insertion site (Schneider,1956; Gao and Messner,1996), the mechanical strength of the ligament (Gao and Messner,1996; Gao et al.,1996), and loading conditions that occur at different stages of maturity. Studies of the human knee by Evans et al. (1991) and Benjamin et al. (1991) suggested that the amount of CFC at a ligament enthesis was positively related to the force exerted on the ligament.

No significant correlation was observed between the thickness of the CFC region and the surface area of attachment at the enthesis of the ACL and MCL origin and insertion sites. Thus, the data did not support the second hypothesis, where it was stated that there would be a positive correlation between the area of ligament attachment and the thickness of the CFC region at the origin and insertion sites of the ACL and MCL. Although data from the ACL suggest a possible correlation between CFC thickness and the ligament's area of attachment, it would be necessary to examine more samples before it could be determined whether or not this finding is significant. In contrast, an inverse relationship was observed between the surface area of attachment and fibrocartilage thickness of the MCL.

The final hypothesis examined whether significant differences would be observed in the percent ash content of the three regions of interest (CFC, boundary region, and host bone) at the ligaments' attachment sites. Evaluation of the ACL and MCL revealed that both possessed significantly lower mineral content within the boundary region when compared with the CFC and host bone regions (Fig. 1). This finding reinforced the fact that distinct regions and variations in mineral content were present as the ligament transitioned to the bony attachment. Furthermore, it is likely that the high mineral content and resultant stiffness associated with these regions are associated with the need for preservation of ligament attachment morphology as it transitions to bone and during loading cycles (Schneider,1956; Benjamin et al.,1986).

Ligaments possess different mechanisms for attachment that are dependent on how the ligament is loaded. Skeletal attachment of ligaments can be varied by altering the surface area of attachment, thickness of the CFC region, and mineral content of the fibrocartilage region. This variation in structure and composition was exemplified by the adaptational mode for accommodating variations in loading between the ACL and MCL. These differences appear to be indicative of “Nature” adapting the ligament's enthesis to its function within the joint. The ACL, which serves as the primary stabilizer of the knee, possesses a greater surface area of attachment than the MCL. The increased area of attachment appears to allow the ACL to accommodate the greatest loads encountered throughout the knee's normal range of motion. The MCL, a supporting ligament, has a lower surface area of attachment, but appears to have adapted to loading and maintaining attachment with increased thickness of the CFC region. Further investigation at various ligament sites in composition studies in human and other animal studies may help in better understanding how the enthesis region adapts to the variations in ligament attachment surface area and bone mineral and CFC depth.

In conclusion, characterization of the ACL and MCL attachment site was performed in a translational animal model that was commonly used for investigating and improving ligament reconstruction procedures (Seitz et al.,1997; An and Friedman,1999; Shea et al.,2002; Hunt et al.,2005; Burger et al.,2007; Rumian et al.,2007; Meller et al.,2008a,b; Scheffler et al.,2008; Tapper et al.,2009). The structural characterization described in this work hopefully contributes to the body of knowledge that attempts to address how the composition of the ligament attachment occurs in this translational model. The significance of the differences in CFC mineral content and its potential impact on the mechanical properties of the ligament attachment sites will require further investigation. Subsequent work should be performed to examine the significance of these mineral content differences and their effects on the ligament's mechanical attachment properties. Furthermore, although the ligaments of the ovine knee have been reported to be good models for studying ligament insertion (Seitz et al.,1997; An and Friedman,1999; Shea et al.,2002; Hunt et al.,2005; Burger et al.,2007; Rumian et al.,2007; Meller et al.,2008a,b; Scheffler et al.,2008; Tapper et al.,2009), evaluation of the human ACL and MCL will be necessary to better characterize the ligament attachment sites and to determine whether those findings are consistent with the results presented in this study.

Acknowledgements

The authors thank Gregory Stoddard for assistance with and interpretation of statistical analysis.

Ancillary