INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MICROTRAUMA OR SUBFAILURE INJURY in tendon and ligament may occur either as the result of overuse or as a single traumatic event (5). Pathologies of the musculoskeletal system, in which microtrauma is thought to play a role, include tendinitis and tendinosis of the upper extremity (20), including tennis elbow (8) and carpal tunnel syndrome (31), as well as lower limb pathologies such as Achilles tendinitis (8) and posterior tibial tendinitis (4), among others. Microtrauma diminishes mechanical properties in knee ligaments (19) and is hypothesized to be a source of lower back pain through microtrauma of either the iliolumbar ligament (35) or the intervertebral disk (42). Fruensgaard and Johannsen (10) reported that partial tears of the anterior cruciate ligament often lead to reduced levels of activity and performance. Although conservative treatment is often successful, tendon and ligament microtrauma and partial tears may accumulate damage to the point that load bearing is compromised and complete rupture or secondary damage occurs (12). In addition, ligament microtrauma may result in increased laxity, which in turn is associated with degenerative joint disease and osteoarthritis (9). These studies suggest the importance of studying subfailure injury and intrinsic healing of ligaments.

A sprain is defined as an acute injury to a ligament or joint capsule without dislocation. Sprains are classified by severity on the basis of clinical examination or imaging. Grade I sprains are mild stretches with no discontinuity of the ligament and no clinically detectable increase in joint laxity. Grade II sprains are moderate stretches in which some fibers are torn. Enough fibers remain intact so that the damaged ligament has not failed. These grade II sprains produce detectable abnormal laxity at the joint compared with the uninjured, contralateral side. Grade III sprains are severe and consist of a complete or nearly complete ligament disruption and result in significant joint laxity. Severe sprains involving complete disruption of the ligament and resulting in significant joint laxity (grade III) constitute <15% of all ligament sprains (2). This leaves >85% of the sprains in which subfailure damage is the dominant issue.

Early studies of mechanical properties of tendons and ligaments disclosed that irreversible mechanical behavior (and presumably damage) occurs at relatively low levels of strain [2.5-4.5% beyond preloaded reference levels (32, 38)]. In one study of rabbit anterior cruciate ligaments, subfailure injury (~80% of failure stretch) altered the shape of loading curves and thereby increased joint laxity (28). In another study, dramatic reductions in peak force were observed in human cadaveric inferior glenohumeral ligament after repeated subfailure loadings (39). Laws and Walton (23) used a sheep model to investigate a grade II injury to the medial collateral ligament (MCL). They reported an initial 13% reduction in the tensile strength and an increase in laxity but full recovery at 6 wk. They also reported progressive healing by infiltrating fibroblasts with less of the typical inflammatory response seen in complete failure models. The first time interval for histological investigation was after 1 wk, so the early healing response was not described. The injury was not well controlled, and the molecular responses in the MCL were not described. Although changes in ligaments have been observed at subfailure loadings, the mechanisms and effects of damage accumulation are still not known.

Cells in the body are subjected to complex mechanical loadings, consisting of tension, compression, shear, or combinations of the three types of load and deformation. Researchers have studied the mechanical stimulation and behavior of cells in vitro (1, 11, 13, 14, 40). Hsieh et al. (14) studied human fibroblasts from the anterior cruciate and medial collateral ligaments in vitro under equibiaxial strains of 5 and 7.5%, showing that cell strain induced expression of types I and III collagen. They suggested that remodeling of ligament tissue may take place by a continuous microhealing process whereby scar tissue (predominately type III collagen) is formed and later matures into remodeled tissue. Arnozczky et al. (3) showed that cell deformation increased with tissue strain in situ but that when the relationship is nonlinear and wide, a variation was observed. None of these studies examined cell necrosis in situ or the mechanism by which mechanically induced cell necrosis occurs in soft tissue.

This study examines structural and cellular damage in ligaments after a subfailure damage injury. Sprains are not currently well quantified in this subfailure region. This study 1) quantifies the onset of structural damage in the medial collateral ligament as characterized by nonrecoverable change in tissue length (from preload) after a subfailure stretch, and 2) quantifies regions of cellular damage (not the number of necrotic cells) in ligament as a function of subfailure ligament strain. The tissue's mechanical properties after a subfailure stretch were also evaluated. It should be noted that for this study structural damage is considered nonrecoverable changes in the entire ligament structure and does not reflect examination of the extracellular matrix (ECM) (i.e., collagen fiber or fibril integrity) at the microstructural level.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental studies. This study was approved by the institutional animal use and care committee and meets National Institutes of Health (NIH) guidelines for animal welfare. Sprague-Dawley male rats (weight 250 ± 25 g) were used as our animal model. Immediately after death, all extraneous tissue was carefully dissected to expose each MCL. Intact MCLs, including femoral and tibial bone segments, were carefully harvested with care taken not to disturb the ligament insertion sites. The tissues were kept hydrated in Hanks' physiological solution (pH = 7.4). Structural damage (n = 25 ligaments) and cellular damage (n = 24 ligaments) were evaluated in this study. Twelve additional ligaments (n = 6 pairs) were used to evaluate changes in the ligaments' mechanical properties after a subfailure stretch.

Statistical analysis. Statistical analysis was performed to determine the strain at which the onset of damage occurs from structural and cellular standpoints and to determine whether the two were different from one another. An observation of cellular damage (cellular damage per ligament) was defined as D_ci. The general expression D_ci ~ N(µ_ci, ) indicates that D_ci was expected to have the mean value µ_ci, a normal distribution about this mean value with variance . The mean value was modeled as a function of strain _ci as

(1)

with

where _c was the strain threshold below which cellular damage is zero and above which cellular damage increases linearly with strain with slope _c. The kink at _c was obtained by the indicator function I_c(_ci), which equals 0 when _ci < _c and equals 1 otherwise. A similar development was applied for structural damage D_si, modeling the mean value (µ_si) as a function of strain (_si), hence

(2)

with

To obtain the "best" estimates of the parameters _c, _c, _s, and _s (estimates of the nuisance parameters _c and _s were also needed), maximum likelihood estimates (MLEs) were used (e.g., Ref. 24). MLEs are those values of the parameters that maximize the probability of observed data. Thus, to obtain MLEs, the probability density function of the observed data was constructed; this probability density function is referred to as the likelihood equation. For mathematical convenience, rather than maximizing the likelihood directly, the log likelihood was maximized (the position of maxima are invariant to log transformation).

(3)

(4)

(5)

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Results indicate that tissue stretch-induced structural damage and cellular damage are fundamentally different in their behaviors when analyzed as a function of applied strain. Qualitatively, regions of cellular damage can be seen to increase with strain by examination of the confocal microscopic images (Fig. 1). In addition, the strain produced extensive damage to individual cells as seen in TEM images at 3.2% ligament strain (Fig. 2). These images show a typical cell for which the membrane is intact and the cellular contents are undisturbed in a ligament with 0% strain. A typical mechanically damaged cell in the strained ligament displays a cell membrane that has been ruptured and the cellular contents that have been released into the ECM. Data for structural damage (D_S) and cellular damage (D_C) seen in Figs. 3 and 4, respectively, clearly show the different strain-dependent behaviors for structural and cellular damage. Statistically, the onset of structural damage (_s) was found to be at a strain of 5.14% from preload, and this level of strain can be seen to be well within the linear region of the stress-strain curve for the rat ligaments used in this study (Figs. 5). The 95% profile likelihood confidence intervals for structural damage are 4.50-5.69. The onset of changes in cellular damage (_c) was found to be 0 ( = 0%). That is, statistically, changes in cellular damage in rat MCLs begin with the application of ligament strain from preload, and structural damage occurs at strains >5.14% from preload. Some regions of cellular damage can be seen at 0% strain (i.e., preloaded treatment), yet statistical analysis could not reveal any other onset because of scatter in the data. No consistent localization of regions of cellular damage could be identified in the ligaments. The slope values () for structural and cellular damage are 1.42 and 4.00, respectively. From these values of  and , Eqs. 1 and 2 are

(1)

(2)

with the indicator function I(_i) equal to 0 when _i <  and equal to 1 otherwise (Fig. 3). The null hypothesis of identical thresholds has H₀: _c = _s, and the alternative hypothesis is H_A: _{c s}. Performing a line search finds the maximum likelihood under the null hypothesis to be L⁰ = 127.152 ( estimated to be 4.79). Performing a grid search yields L^A = 109.5 (_c estimated at 0, _s estimated at 5.14). The equation ² = 2(L^A  L⁰) would have a ² distribution with one degree of freedom under the null distribution. With ² = 35.3, the null hypothesis is soundly rejected (P < 0.0001) and the alternate hypothesis accepted, indicating that the onset of structural and cellular damage occur at significantly different levels of strain. Stress-strain curves demonstrate no change in shape when the subfailure stretch is below the damage threshold (Fig. 5, A-C). A subfailure stretch above the damage threshold elongates the toe region of the stress-strain curve and decreases tissue stiffness and ultimate stress (Fig. 5, D-F).

View larger version (37K): [in this window] [in a new window]   Fig. 1.   Rat medial collateral ligaments at strains of 0 (A), 6 (B), and 11% (C). Red and yellow regions indicate cell necrosis, and green regions are healthy live cells. Note the extent of necrosis in the strained ligaments and the increase in necrosis with higher strains.

View larger version (114K): [in this window] [in a new window]   Fig. 2.   Fibroblasts in a normal unstrained ligament (left) and fibroblasts in a strained ( = 3.2%) ligament (right). Normal ligaments have intact cellular membranes in a longitudinal section at 24,000× (top left), transverse section at 9,100× (middle left), and longitudinal section at 24,000× (bottom left). Ligament that has been strained to 3.2% (right) shows cellular damage in the form of ruptured cell membranes and the release of cellular contents in the extracellular matrix in a longitudinal section at 24,000× (top right), transverse section at 9,100× (middle right), and longitudinal section at 24,000× (bottom right). Note interdigitation of collagen fibrils with normal ligaments in left column.

View larger version (7K): [in this window] [in a new window]   Fig. 3.   Structural damage (DS) vs. strain () in rat medial collateral ligament. Structural damage represents tissue laxity. The onset of damage is at 5.14% strain, as can be seen by Eq. 2. Note that this level of strain is well into the linear region of the stress-strain curve for this tissue, as can be seen from Fig. 5.

View larger version (8K): [in this window] [in a new window]   Fig. 4.   Cellular damage (DC) vs. strain () in the rat medial collateral ligament. Necrotic area is normalized by the area of the ligament. Equation 1 demonstrates that cellular damage statistically begins with the onset of loading (from a preloaded state) and occurs at lower strain levels than structural damage.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Stress-strain curves for a control rat medial collateral ligament (), subfailure stretch (), and stretched ligament () that received a subfailure stretch. The subfailure stretch reached a peak strain of 0% (A), 4.7% (B), 5.1% (C), 6% (D), 7% (E), and 9% (F). Subfailure strains () below 5.14% (A-C) did not change the toe region or tangential modulus on a subsequent stretch (), and the toe region, tangential modulus, and ultimate stress of the ligament () were not changed compared with the contralateral normal ligament (). Subfailure strains () above 5.14% (D-F) elongated the toe region and lowered the tangential modulus () compared with the initial stretch (). The toe region was elongated and the tangential modulus and ultimate stress decreased () compared with the contralateral normal ligament (). Note the relative symmetry in contralateral ligaments in 0% subfailure strain (A).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subfailure damage has been shown to alter the mechanical properties of the anterior cruciate ligament and lengthen the toe region of the force-displacement curve, which results in increased joint laxity (28). Results from this study indicate that rat medial collateral ligaments strained above 5.14% from preload do not regain their original length after significant recovery time (300× the time of test). This recovery is considerably longer than other studies showing that ligaments stretched below 5% completely recover in <10× the time of test (30). Hence, we conclude that ligaments stretched beyond this threshold remain "stretched." These findings are valuable to researchers performing multiple tests on the same ligament specimen because no change in length or properties is evident in the tissue below ~5% when testing under the methods described in this study for rat.

We speculate that the increase in elongation after ~5% strain and change in mechanical properties are the result of fiber damage arising from two possible mechanisms. One mechanism would be torn or plastically deformed fibers. Torn fibers would be consistent with the fiber failure mode of Hurschler's micromechanical model for ligament behavior (16), and plastically deformed fibers could be supported by experimental observations by Sasaki et al. (33, 34) and Kukreti and Belkoff (22) who observed that collagen fibrils, which make up collagen fibers, elongate during tendon loading. In addition, Yahia et al. (41), using scanning electron microscopy, reported damage to collagen fibers in subfailure strained ligaments. Another possible mechanism for the observed ligament laxity could be biochemical degradation of the ECM from protease release associated with the observed cellular necrosis. Regardless of the mechanism, the resulting increase in tissue length represents tissue laxity and can be hypothesized to increase joint laxity.

The statistical threshold for structural damage is well into the linear region of the stress-strain curve, a region unlikely to be regularly reached in normal activity. The upper half of the linear region of stress-strain curve has been associated with partial tearing of anterior cruciate ligaments (26). A microstructural study of the rat MCL in situ using scanning electron microscopy has shown that during knee flexion normal ligament has a crimped pattern in knee flexion that is only removed near full extension (17). Our preload removes some crimping and hence results in our preloaded reference length being longer than typical reference lengths in vivo.

The statistical threshold of cellular damage was found to be at 0% strain from preload in the rat MCL. That is, statistically, cellular damage begins with the application of tissue strain. It should be noted that physically one would not expect an increase in cellular damage at small strains as our statistical analysis implies. However, necrotic cells are present in the control tissues ( = 0, i.e., preload) and are present after very small strains (above reference preload). This behavior did not allow the authors to identify any statistical threshold other than the preloaded value. One would not expect high levels of cell necrosis during normal activity; therefore, as mentioned in the previous paragraph, the strains represented when this method is used are likely to be higher than those typically seen in the rat in vivo. The ratio of the area of fluorescence to the actual cell dimensions is not known, and therefore the number of necrotic cells is not counted nor is the percent area of necrotic cells in the tissue quantified. In addition, cellular changes may take place during the processing of the tissues because of processing time and handling, elevating the magnitude of cellular damage. However, all tissues were prepared in the same controlled manner, and changes in cellular damage were seen between preloaded zero strain controls and stretched ligaments, with increasing cellular damage with increasing strain indicating changes relative to the reference values. This difference in damage properties indicates fundamentally different behaviors between cellular and structural damage.

This study quantifies regions of cellular damage in the ligament as a function of strain. As stated in the previous paragraph, this work does not count the number of necrotic cells in the ligament. To the authors' knowledge, no other study has quantified cellular damage in ligament as a function of strain. However, mechanical stimulation of cells in vitro has been studied (1, 11, 13, 14, 40). Strain in fibroblasts during in vitro equibiaxial testing on membranes are often higher than the tissue strains at which we are reporting cell damage ( > 2% from a preloaded state). However, the relationship between reference (initial) strains used in these studies compared with the complex cell loading in an in vivo state is unknown. Furthermore, differences between the reference strain in our study and previously published in vitro cell deformation studies is also unknown.

We propose that microstructural irregularities in ECM organization create local distortions in fibroblasts during ligament strain that result in the cellular damage reported in this study. Fibroblasts in connective tissue are found to be in columns between fibril bundles (18, 36), and fibrils cross and interweave (29). Using confocal microscopy to track cells as a measure of microstructural strain during tendon tissue elongation, Hurschler et al. (15) reported that microstructural strains in ligament can be both larger and smaller than macroscopic ligament strain (± ~2% of the ligament strain) at different locations in the same tissue and that, at the microstructural level, even negative strains can exist. Because fibroblast cells adhere to ECM collagen fibrils, cell deformation is related to fibril displacement and deformation. Squier and Bausch (36) reported junctional attachments between processes of different and same fibroblasts in rat tail tendon. In addition, Squier and Bausch (36) showed that fibroblasts exhibited invaginations of single or groups of collagen fibrils and stated that "this arrangement may be indicative of fibril elongation or serve to transmit tension between the fibroblast and the collagen fibrils." This result is supported by the TEM images in Fig. 2, which show fibrils interdigitating with the cell membrane. Because fibril orientation around the cell is not completely uniform, the fibrils will stretch independently, creating nonlinear cell strains and nonuniform cellular distortions as seen in Arnozczky et al. (3). The nonuniform fibril strain patterns seen must produce local cellular distortions far beyond what occurs on monolayers of cultured cells and must contribute to mechanically induced cell death reported herein. These large local distortions from nonuniform fibril loading, fibrils slipping past one another, and complex cell-matrix interactions help explain the results of this study where changes in cellular damage occur before the onset of structural damage. In contrast to the complex loadings and large distortions of fibroblast in vivo, fibroblasts in vitro on stretched membranes undergo less complex and less distortional loadings. Although in vitro stretched membrane cell deformation studies provide extremely valuable information about biological mechanisms, they may not accurately simulate in situ cells during ligament deformation.

Considering the organization of the fibroblast-ECM interaction, one may expect to see cellular damage in longitudinal lines along the axis of the ligament. However, it is reasonable that cellular damage is not seen in longitudinal lines in this study. By using the above techniques, slices through the tissue are compiled and overlaid, and because all fibers and fibrils are not perfectly aligned and do sit skewed from one another through the depth of the tissue one would not expect to see clear longitudinal distributions. By overlaying slices, the slight mismatching of collagen fibers or fibrils, size effects, and fibril interweaving and overlap limit the ability to see longitudinal columns because all the slices are "compressed" together. It is entirely possible that many semilongitudinal scans are adjacent, skewed, or slightly angled as a result of complex fibril organization, resulting in the distributions seen in Fig. 1. It should be noted that, because only regions of necrotic cells are being quantified (overlaid yellow regions are counted as damaged) and green (viable) regions are not being quantified, the overlaying process does not induce an error in the cellular damage measure used in this study because cellular damage only considers necrotic cells normalized by tissue area.

Limitations exist in this study and should be noted. Structural damage was measured at only one rate of loading from a preloaded state, and the effect of multiple loadings was not investigated. In addition, the ligament microstructure was not examined for collagen fiber or fibril integrity (tearing or plastic deformation), and the mechanical properties were evaluated with only one tissue per applied subfailure stretch. Cellular damage was examined only at one time point and, therefore, only early cellular damage is quantified. In addition, processing times may have increased the magnitude of cell death in the tissue, yet changes in cellular damage are seen to increase with strain in tissues processed in the same controlled manner. The time-dependent aspects of cellular damage and the possibility of further necrosis resulting from apoptosis remain unexplored. In addition, in vivo loads in rat MCLs are not known and, hence, this behavior must be studied in a better model with more elucidation into the in vivo reference states regarding tissue microstructure and deformation during normal activity. The rat model used herein leaves questions open regarding effects of ligament size and species. Further studies in human ligaments are recommended.

In summary, structural and cellular damage occur at different levels of tissue strain in a rat MCL. Subfailure strain above the damage threshold changed the mechanical properties of the ligament. Further investigations into loading rate, multiple loadings, ligament microstructural displacements and deformations, cell deformation, and cell biology need to be performed to understand the subfailure behavior of ligament and the role of cell death in the healing process.