Journal of the American Academy of Orthopaedic Surgeons 364 The incidence of knee ligament injuries resulting from sports activi- ties has stimulated interest in liga- ment healing. Clinical experience has shown that an isolated ruptured medial collateral ligament (MCL) of the knee can heal with nonoperative care, but that a midsubstance tear of the anterior cruciate ligament (ACL) usually does not. An ACL deficiency can result in chronic knee instability with a 2-degree injury to other soft- tissue structures and eventually progressive osteoarthritis. Primary surgical repair of the ACL has not been successful; as a result, recon- struction with the use of replace- ment grafts is usually performed to restore knee stability. Studies involving animal models have also demonstrated that isolated MCL injuries can heal spontane- ously, with excellent knee function. These results have supported the clinical decision that grade III MCL injuries should continue to be treated nonoperatively. The ability of the MCL to heal primarily offers an op- portunity for the examination of the mechanism of ligament healing. These experiments have provided useful information on the outcome of isolated MCL injuries as well as the effect of intrinsic and extrinsic factors on that outcome. The asso- ciation of an ACL disruption with an MCL tear completely changes the prognosis, however. Tissue engi- neering, including the use of growth factors, cell therapy, and gene transfer techniques, has shown some potential for enhancing the healing process. Although most of the experimentation has dealt with the ACL and MCL in the knee, the principles elucidated and the knowl- edge gained can serve as the basis for further studies to improve un- derstanding of the healing of other ligaments and tendons as well. Healing of Grade III MCL Tears Although methods of treating liga- mentous injuries have seen sub- stantial improvements in recent years, there remain many questions about enhancing the rate, quality, and completeness of extra-articular ligament healing. To adequately address the intricacies of this process and improve on techniques used in clinical practice, it is essen- tial to increase our knowledge Dr. Woo is Ferguson Professor of Orthopaedic Surgery and Director, Musculoskeletal Research Center, University of Pittsburgh. Ms. Vogrin is Research Engineer, Musculo- skeletal Research Center. Mr. Abramowitch is Graduate Research Assistant, Musculoskeletal Research Center. One or more of the authors or the department with which they are affiliated have received something of value from a commercial or other party related directly or indirectly to the sub- ject of this article. Reprint requests: Dr. Woo, Musculoskeletal Research Center, PO Box 71199, Pittsburgh, PA 15213. Copyright 2000 by the American Academy of Orthopaedic Surgeons. Abstract Although methods of treating ligamentous injuries have continually improved, many questions remain about enhancing the rate, quality, and completeness of ligament healing. It is known that the ability of a torn ligament to heal depends on a variety of factors, including anatomic location, presence of associated injuries, and selected treatment modality. A grade III injury of the medial collat- eral ligament (MCL) of the knee usually heals spontaneously. Surgical repair followed by immobilization of an isolated MCL tear does not enhance the healing process. In contrast, tears of the anterior cruciate ligament (ACL) and the poste- rior cruciate ligament often require surgical reconstruction. The MCL compo- nent of a combined ACL-MCL injury has a worse prognosis than an isolated MCL injury. The results of animal studies suggest that nonoperative treatment of an MCL injury is effective if combined with operative reconstruction of the ACL. Experimentation using animal models has helped to define the effects of ligament location, associated injuries, intrinsic factors, surgical repair, recon- struction, and exercise on ligament healing. New techniques utilizing growth factors and cell and gene therapies may offer the potential to enhance the rate and quality of healing of ligaments of the knee, as well as other ligaments in the body. J Am Acad Orthop Surg 2000;8:364-372 Healing and Repair of Ligament Injuries in the Knee Savio L Y. Woo, PhD, DSc, Tracy M. Vogrin, MS, and Steven D. Abramowitch Savio L Y. Woo, PhD, DSc, et al Vol 8, No 6, November/December 2000 365 about the basic science of ligament healing. Due to the accessibility, frequency of injury, and healing properties of the MCL, it has become a primary model for scientific research, the results of which can be transferred to many of the other ligaments in the body. The healing process in an isolated MCL tear is affected by various systemic and local factors and is somewhat similar to that in vascular tissues. 1 Clinically, grade I and grade II MCL injuries heal well within 11 to 20 days after injury. 2 In contrast, healing of a grade III MCL tear may continue for years after the initial injury. The healing process can be roughly divided into four overlap- ping phases: hemorrhage, inflam- mation, repair, and remodeling. Its description can be divided into his- tologic, biochemical, and biome- chanical events. Histology After a midsubstance tear of the MCL (which is characterized by the mop-end appearance of its torn ends), the hemorrhage phase begins with blood flowing into the gap cre- ated by the retracting ligament to form a hematoma. In response to the increased vascular and cellular reactions resulting from the injury, inflammatory and monocytic cells migrate into the injury site, convert the clot into granulation tissue, and phagocytize the necrotic tissue. This marks the beginning of the in- flammatory stage. Within approxi- mately 2 weeks, a continuous net- work of immature, parallel collagen fibers replaces the granulation tis- sue. The inflammatory phase con- cludes with the formation of extra- cellular matrix in a random pattern in the central region of the ligament. The formation of extracellular matrix by fibroblasts also marks the beginning of the reparative phase, in which the ligament superficially resembles its preinjury appearance. The torn ends of the ligament are no longer visible, and the granula- tion tissue has been replaced by immature, parallel collagen fibers. New blood vessels begin to form, and fibroblasts continue to actively produce extracellular matrix. This phase begins within 5 to 7 days after injury and concludes after sev- eral weeks. Overlapping with the reparative phase, the remodeling phase be- gins several weeks after injury and continues for months or even years. In this phase, collagen fibers con- tinue to align along the long axis of the ligament, resulting in increased maturation of collagen matrix. The conversion from a random pattern to one demonstrating alignment of the fibers has been shown to corre- late directly with an improvement in the biomechanical properties of the ligament. Long-term animal studies have demonstrated that the histologic and morphologic appearance of healed ligaments is different from that of injured ligaments. When tis- sue is viewed by using transmission electron microscopy after 2 years of healing, the number of collagen fibrils is increased compared with the noninjured ligament, but their diameters and masses are actually smaller. 3 Additionally, “crimping” patterns within the healing liga- ment remain abnormal for up to 1 year, and collagen fiber alignment remains poor. 1,4 Biochemistry Studies conducted on animal models have shown that the extra- cellular matrix of the healing liga- ment exhibits a number of important changes, particularly in glycosami- noglycans, elastin, and other glyco- proteins. Early in the process, there are changes in collagen fiber type and distribution, with a greater pro- portion of type III fibers than in nor- mal ligaments. This ratio returns to normal after approximately 1 year. 1 Although the healing ligament shows increases in the number of collagen fibers, the number of ma- ture collagen cross-links is only 45% of the normal value after 1 year. 5 If the joint is immobilized during the healing process, significant changes in collagen synthesis and degrada- tion can occur. There is a direct rela- tionship between the decrease in the biomechanical properties of the healing MCL and the number of col- lagen cross-links, as well as a de- crease in the mass and diameter of the collagen fibers. 5 The collagen types at the bone-MCL interface have also been studied; types II, IX, X, and XIV have been identified. 6 Further studies are needed to assess whether ligament injury and healing affect the presence and amount of these types of collagen. Biomechanics Biomechanical characterization of a healing ligament is based on two elements. Functional testing involves determination of the con- tribution of the ligament to knee kinematics as well as the in situ forces in the ligament in response to external loading conditions. Tensile testing provides an assess- ment of the structural properties of the bone-ligament-bone complex and the mechanical properties of the ligament substance. 4 To analyze the functional capa- bilities of the healing MCL, the effect of surgical repair and immo- bilization on the varus-valgus laxity of the knee was studied. Canine MCLs with grade III injuries that were treated nonoperatively with early mobilization and full weight bearing demonstrated restoration of normal stability by 48 weeks post- operatively. Ligaments that were surgically repaired and then immo- bilized tended to have more valgus laxity than control ligaments. It is important to note, however, that the number of degrees of freedom al- lowed during testing of the knee can have a considerable impact on Ligament Injuries in the Knee Journal of the American Academy of Orthopaedic Surgeons 366 the results obtained. 4,7 After sec- tioning of the MCL, only small in- creases in valgus laxity (21%) were observed if knee motion was al- lowed in all directions. However, when anterior-posterior translation and internal-external rotation were constrained, valgus laxity increased significantly (171% [P<0.05]). These results suggest that with normal knee joint motion, other structures, especially the ACL, compensate for the absence of the MCL during val- gus rotation. 4 A robotic universal force-moment sensor testing system provides a method for making multiple determi- nations of knee kinematics as well as the in situ force or tension in liga- ments in response to external loading conditions. 8 With this technology, the ability exists to evaluate various knee conditions and reconstruction techniques with comparison to the intact knee, thus minimizing intra- specimen variability. The potential also exists to characterize knee kine- matics in vivo and to determine the in situ forces in knee ligaments for in vivo loading conditions. In one study using this testing system in a goat model, 9 the healing MCL showed increased valgus rotation compared with control ligaments at both 6 and 12 weeks. However, knee stability did not improve from 6 to 12 weeks. Future studies using this system will examine the effects of ACL recon- struction in a combined ACL-MCL injury model. Tensile testing can provide valu- able information on the strength and quality of healing tissue and can allow comparison with the intact lig- ament. Two sets of data can be obtained from a uniaxial tensile test: the load-elongation curve illustrates the structural properties of the bone- ligament-bone complex, and the stress-strain curve demonstrates the mechanical properties of the liga- ment substance. Figure 1 shows a typical load-elongation curve. The curve is nonlinear and consists of a toe region, a linear region (where the slope reflects the stiffness of the femur-MCL-tibia complex), and a failure region. The variables repre- senting the structural properties of the femur-MCL-tibia complex in- clude linear stiffness, ultimate load, and energy absorbed at failure. Figure 2 is a typical stress-strain curve of the MCL midsubstance. By normalizing for the cross-sectional area of the tissue, the stress (defined as force per unit area) can be calcu- lated. The strain is calculated as the change in length of the tissue under the tensile load, divided by its origi- nal length. A nonlinear stress-strain relationship can therefore be illus- trated, with toe, linear, and failure regions similar to those of the load- elongation curve. Properties that can be obtained from this graph include modulus of elasticity (the proportional constant between stress and strain), tensile strength, ultimate strain, and strain energy density. In tensile testing studies using animal models, Weiss et al 10 dem- onstrated that the biomechanical properties of the healing MCL remain inferior to those of the intact ligament for as long as 1 year after injury. After 52 weeks of healing, only the stiffness of the femur- MCL-tibia complex returned to near-normal levels, while the ulti- mate load was still significantly (P<0.05) lower than the control value. Furthermore, the mechanical properties of the midsubstance of healing MCLs remained significantly (P<0.05) inferior, even though the cross-sectional area of the healing MCL was much larger than that of the intact MCL. Thus, the healing pro- cess involves a larger quantity of lesser- quality ligamentous tissue. Factors Influencing Ligament Healing There are numerous factors that af- fect the healing response of an in- jured ligament. These include the site and severity of the injury, vari- ous intrinsic factors (e.g., circulation and infection), the type of treat- ment, and the degree of mobiliza- tion after injury. Isolated Ligament Injuries In laboratory studies using ani- mal models, the MCL has been Ultimate load Linear stiffness Ultimate elongation Load, N Elongation, mm 400 300 200 100 0 2 4 6 8 10 Energy absorbed at failure Toe Region Linear Region Failure Region Figure 1 Typical load-elongation curve of the bone-MCL-bone complex. Savio L Y. Woo, PhD, DSc, et al Vol 8, No 6, November/December 2000 367 shown to heal well compared with the ACL and the posterior cruciate ligament (PCL). These findings have been supported by clinical ob- servations. 11,12 Variations in healing ability may be attributable to differ- ences in the blood supply and in the articular environment (i.e., intra- articular or extra-articular). The more “hostile” synovial environ- ment surrounding the ACL may also be a factor; however, some studies have suggested positive effects of synovial fluid on ligament healing. 13 Structural differences (e.g., fiber orientation and crimp pattern) and cellular differences (e.g., fibroblast shape) may con- tribute to these variations. 14 Biomechanical factors may also play a role in the ability to heal. For example, the ACL contributes to knee stability in multiple directions (i.e., anterior, internal, and valgus), while the MCL primarily restrains valgus knee rotation. Thus, a rup- tured MCL receives some protection from other structures, such as the ACL and the joint capsule, and is not subject to the same forces that may impede healing. In contrast, the soft- tissue structures surrounding the ACL may not be able to accommo- date the multidirectional demands so as to allow healing. 15 Therefore, information about the in vivo loads in ligaments is crucial to determina- tion of the optimal load and amount of stretch that will optimize ligament healing. Combined Ligamentous Injuries Some knee injuries involve multi- ple ligaments. The prognosis for these combined injuries is generally worse regardless of which type of treatment is selected. Some authors have reported satisfactory results with nonoperative treatment of com- bined ACL-MCL injuries. 16 Others advocate surgical reconstruction of the ACL with repair of the MCL to adequately restore knee function. 17 Still others choose to surgically re- construct the ACL without address- ing the MCL. 11 Clinical studies have generally been inconclusive. Hillard-Sembell et al 18 reported on 66 patients treated with MCL repair and ACL reconstruction (n = 11), ACL reconstruction only (n = 33), or nonoperative treatment for both lig- aments (n = 22). No differences in valgus instability or in knee func- tion during activities were observed between the three groups studied. Further clinical studies are needed to determine the optimal treatment for these combined injuries, but some evidence may be obtained from animal models. The effects of ACL deficiency on the healing of the injured MCL have been studied by using both rabbit and canine models. 15 Biomechanical evaluation indicated that knees with untreated combined ACL-MCL in- juries showed significantly (P<0.05) increased valgus laxity (Fig. 3) and a reduction in tissue quality of the healed MCL. Considerable degener- ation of the joint was also observed. Laboratory studies in a rabbit model have suggested that nonoper- ative treatment with full weight bearing and mobilization of the injured MCL with reconstruction of the ACL can result in successful healing of the MCL. 19 Although other animal studies also suggested that MCL repair combined with ACL reconstruction reduced valgus laxity and improved the structural properties of the femur-MCL-tibia complex better in the short term (12 weeks), after 52 weeks no differ- ences in biomechanical or biochemi- cal properties were observed. Intrinsic Factors A number of intrinsic factors may also contribute to the healing response of the injured ligament. Any disease that affects endocrine or metabolic homeostasis may affect ligament healing. In a study on the healing MCL of hypophysectomized rats, interstitial cell–stimulating hor- mone and testosterone replacement markedly affected the ultimate load of the repaired ligament and the rates of collagen and glycosamino- glycan synthesis or degradation. 20 Diabetes mellitus results in circula- Figure 2 Typical stress-strain curve describing the mechanical properties of the MCL midsubstance. Tensile strength Modulus of elasticity Ultimate strain Stress, MPa Strain, % 80 60 40 20 0 2 4 6 8 10 Strain energy density Toe Region Linear Region Failure Region Ligament Injuries in the Knee Journal of the American Academy of Orthopaedic Surgeons 368 tory abnormalities, and insulin defi- ciency alters collagen synthesis and cross-linking; therefore, both condi- tions may negatively affect ligament healing. Ligament healing can also be affected by local conditions, such as poor circulation and infection, that hinder the proliferation of cells, thereby prolonging the inflammatory phase of healing. Type of Treatment Studies have shown that treat- ment selection can also have an impact on the process of ligament healing. Several animal studies of isolated MCL injuries have shown better results with nonoperative treatment than with surgical repair followed by immobilization. In one study, 4 nonoperative treatment with- out immobilization was compared with surgical repair and 6 weeks of immobilization in a transected MCL canine model. Histologic sections revealed that the alignment of the fibroblasts was more longitudinal in the repaired ligaments at 12 weeks, but at 48 weeks both the repaired and the nonrepaired tissues were similar but neither resembled the normal MCL. Biomechanical data indicated that valgus rotation, stiff- ness, and ultimate load of the femur- MCL-tibia complex for the nonre- paired group were closer to control values throughout the 48-week study than those for the repaired and immobilized group. However, the quality of the healed tissue of both the repaired and nonrepaired MCLs was significantly (P<0.05) different from control values. Similar results were obtained in a study using a rabbit model. 10 A “mop-end” tear of the MCL sub- stance was created by placing a stainless steel rod beneath the MCL and pulling it medially, rupturing the MCL in tension and causing a midsubstance tear and damage to the insertion sites. Treatment was either nonoperative with no immo- bilization or surgical repair. No sta- tistically significant differences could be demonstrated between the repaired and nonrepaired groups at 6 or 12 weeks for any biomechani- cal property, including structural properties of the femur-MCL-tibia complex (Fig. 4), the mechanical properties of the MCL midsub- stance (Fig. 5), and varus-valgus knee rotation. These findings are in agreement with clinical reports of positive outcomes with nonopera- tive treatment followed by early motion and functional rehabilita- tion. 21 This is now generally con- sidered to be the preferred method of treatment for isolated grade III injuries of the MCL. 11 Surgical repair of midsubstance tears of the ACL and PCL has been inadequate and appears to fail over time. 12 As a result, direct repair is generally not performed, and liga- ment reconstruction is the preferred treatment for most cruciate liga- ment injuries. For the ACL, autolo- gous tendons of the knee, particu- larly the medial hamstring tendon and the central third of the patellar tendon–bone complex, are the most commonly used graft sources. The reported biomechanical prop- erties of cadaveric grafts are some- what variable due to differences in testing methods and graft sizes. A 14-mm-wide patellar tendon graft is much less stiff than the ACL of a young adult (27.4 ± 3.0 N/mm vs 242 ± 28 N/mm) but is stronger (2,900 ± 260 N vs 2,160 ± 157 N). 22,23 However, a recent study has demon- strated that a 10-mm patellar tendon graft, which is most often used for ACL reconstruction surgery, has a stiffness of 210 ± 66 N/mm and an ultimate load of 1,784 ± 580 N. 24 The corresponding values for a braided quadrupled semitendinosus-gracilis graft were 238 ± 71 N/mm and 2,421 ± 538 N, respectively. 24 Bone–patellar tendon–bone grafts offer the advantage of bone-to-bone healing, which may occur more quickly than tendon-to-bone heal- ing. Patellar tendon grafts have the disadvantages of increased graft-site morbidity, a decrease in quadriceps strength, and a greater prevalence of anterior knee pain. Hamstring ten- don grafts reduce graft-site morbid- 6 4° 3° 2° 1° 1° 2° 3° 4° 5° 6° 7° 8° 4 2 .2 .4 .6 ValgusVarus Intact Varus-valgus moment, N-m MCL transected MCL + ACL transected Figure 3 A typical plot demonstrating the nonlinear relationship between varus-valgus rotation of the knee (horizontal axis) and applied varus-valgus moment (vertical axis) for a time-zero specimen. The graph also illustrates the increase in rotation of the knee after transection of the MCL and after transection of both the MCL and the ACL. (Reproduced with permission from Woo SLY, Young EP, Ohland KJ, Marcin JP, Horibe S, Lin HC: The effects of transection of the anterior cruciate ligament on healing of the medial collateral lig- ament: A biomechanical study of the knee in dogs. J Bone Joint Surg Am 1990;72:382-392.) Savio L Y. Woo, PhD, DSc, et al Vol 8, No 6, November/December 2000 369 ity, but their disadvantages include slower tendon-to-bone healing and diminished hamstring function. Clinically, the choice of graft ma- terial remains a subject of debate. However, no single reconstructive technique has proved superior in terms of functional stability, preven- tion of osteoarthritis, or complica- tion rate. Allografts are also a valid alternative to autografts, particularly for revision ACL reconstructions. The optimal method of graft fixa- tion has been a subject of recent research, especially because of recent clinical reports of bone-tunnel en- largement after ACL reconstruction. It has been suggested that the con- struct should be stiff enough that there is minimal motion between the graft and the bone tunnel, so as to facilitate healing between the tunnel and the graft. A strong, stiff con- struct has also been considered important to withstand the stresses during early rehabilitation. Interfer- ence screws made of either titanium or a biodegradable material are the fixation devices of choice for patellar tendon grafts. However, a wide va- riety of devices for hamstring fixation have been used, including staples, washers, suture and post, titanium buttons, cross-pins, and interference screws. 24 The ultimate load and stiff- ness of numerous fixation devices 250 300 350 150 100 50 1 2 3 4 5 6 7 200 Sham 6 weeks 12 weeks Nonrepaired Elongation, mm Load, N 0 250 300 350 150 100 50 1 2 3 4 5 6 7 200 Sham 6 weeks 12 weeks Repaired Elongation, mm Load, N 0 Figure 4 A, Load-elongation curves representing the structural properties of the femur-MCL-tibia complex for sham-operated and nonre- paired groups. B, Load-elongation curves representing the structural properties of the femur-MCL-tibia complex for sham-operated and surgically repaired groups. (Reproduced with permission from Weiss JA, Woo SLY, Ohland KJ, Horibe S, Newton PO: Evaluation of a new injury model to study medial collateral ligament healing: Primary repair versus nonoperative treatment. J Orthop Res 1991;9:516-528.) 30 20 10 1 2 3 4 5 6 Nonrepaired Strain, % Stress, MPa Sham 6 weeks 12 weeks 0 30 20 10 1 2 3 4 5 6 Repaired Strain, % Stress, MPa Sham 6 weeks 12 weeks 0 Figure 5 A, Stress-strain curves representing the mechanical properties of the MCL substance for sham-operated controls and nonre- paired groups. B, Stress-strain curves representing the mechanical properties of the MCL substance for sham-operated controls and surgi- cally repaired groups. (Reproduced with permission from Weiss JA, Woo SLY, Ohland KJ, Horibe S, Newton PO: Evaluation of a new injury model to study medial collateral ligament healing: Primary repair versus nonoperative treatment. J Orthop Res 1991;9:516-528.) A B A B Ligament Injuries in the Knee Journal of the American Academy of Orthopaedic Surgeons 370 have been quantified, but additional biomechanical studies are needed to assess the effect of cyclic loading (used to simulate the low-intensity, repetitive loading of rehabilitation) on their stability. In some circumstances, patients with an isolated PCL injury do well with nonoperative treatment and early mobilization. Reconstruction of the PCL is generally performed only on patients with multiple liga- mentous injuries and those who are high-performance athletes. 25 Be- cause patients with grade III PCL tears are known to develop medial compartment and patellofemoral chondrosis, many surgeons opt to treat only the larger and stronger anterolateral bundle of the PCL, using an autologous patellar ten- don graft. However, it has been shown in laboratory testing that Achilles tendon allograft offers a large amount of collagen, which can fill the bone tunnels completely, allowing bone fixation on the fem- oral side with a calcaneal bone plug and on the tibial side with a soft- tissue washer. 26 Clinically, no one graft has been proved superior to the other op- tions, and residual knee laxity and early arthritis have been reported after PCL reconstruction surgery. A recently proposed double-bundle PCL reconstruction appears to have some biomechanical benefits in ca- daveric knees. 27 Immobilization Versus Controlled Motion and Exercise In the past, immobilization after ligament injury was believed to be necessary to protect the healing lig- ament from stress. However, it has been shown in the laboratory that immobilization results in disorgani- zation of collagen fibrils, decreases in the structural properties of the bone-ligament-bone complex, re- sorption of bone at ligament inser- tion sites, and other detrimental effects on the knee joint. 28 In con- trast, controlled motion has been shown to be beneficial to the heal- ing ligament. Intermittent passive motion has been reported to im- prove the longitudinal alignment of cells and collagen at 6 weeks, as well as matrix organization and collagen concentration, and also to increase the ultimate load of the femur-MCL-tibia complex by as much as four times. 28 Tipton et al 29 reported that in a canine model ex- ercise positively affected the heal- ing MCL as evidenced by the in- creased ultimate load of the healing tissue. Follow-up studies in a rat model revealed that exercise en- hanced ligament healing, as evi- denced by a more rapid return of DNA and collagen synthesis. There are also some clinical data that demonstrate the advantages of motion after ligament injury. Reider et al 21 have reported favorable clini- cal results 5 years after treatment of isolated MCL injuries with early motion and functional rehabilitation. Current clinical recommendations after MCL injuries include early con- trolled range-of-motion exercises as soon as pain subsides. 11,21 However, in an unstable joint, motion too early or applied too aggressively may be detrimental to the healing process. The effects of motion after treat- ment of ACL and PCL injuries are hotly debated, although it is gener- ally agreed that rehabilitation is criti- cal to prevent arthrofibrosis and re- store knee function. The appropriate timing of rehabilitation has not yet been established. Early accelerated rehabilitation has been advocated by some to minimize knee stiffness and ensure complete knee extension. 30 However, others have supported less aggressive rehabilitation to allow vascularization and incorporation of the graft. Additional laboratory studies evaluating the in situ forces in the ligament and joint kinematics during rehabilitation are needed in conjunction with randomized, pro- spective clinical studies. Tissue Engineering and Ligament Healing Because the biomechanical and bio- chemical properties of the healing MCL fail to return to normal, and the quantity of healing tissue appar- ently increases to make up for the deficiency, researchers are now exploring other modalities that can improve the quality of healing tis- sues, as well as accelerate the rate of healing. Advances in the fields of molecular biology and biochemistry may have applications in the liga- ment healing process. Although the results are still preliminary, new techniques utilizing growth factors and gene transfer and cell therapies may prove useful in accomplishing these goals in the ligaments of the knee, as well as other ligaments throughout the body. Growth Factors Growth factors are small polypep- tides that bind to specific receptors on the surfaces of cells, activating pathways for complex intracellular signal transduction. By modulating cellular behavior, growth factors have shown the ability to affect cell proliferation and migration, matrix synthesis, and the secretion of addi- tional growth factors. Because ex- pressions of various growth factors and their receptors have been dem- onstrated during various phases of the healing process, it is essential to understand their significance. Before growth factors can be used in vivo, their effects on fibro- blast proliferation, matrix synthesis, and cell migration must be exten- sively evaluated in an in vitro set- ting. To date, studies using various animal models have shown that while transforming growth factor-β is a good promoter of matrix syn- thesis, platelet-derived growth fac- tor, basic fibroblast growth factor, and epidermal growth factor are positive mitogens on fibroblasts of the ACL and MCL. Savio L Y. Woo, PhD, DSc, et al Vol 8, No 6, November/December 2000 371 In vivo studies based on the re- sults from in vitro experiments have shown that high doses of platelet- derived growth factor applied with fibrin sealant can have positive effects on the injured ligaments of rabbits, causing significant (P<0.05) increases in the structural proper- ties of the femur-MCL-tibia com- plex. 31 However, studies in other animal models, with varying dosage and method and timing of applica- tion, have also shown dramatic results when growth factors are ad- ministered independently. Collec- tively, the results of these studies show variations that may be species-, dosage-, and treatment-specific, thus demonstrating the complex nature of ligament healing. Addi- tional studies should be directed at defining the variables that affect the specificity of growth factors. Gene Transfer Technology Growth factors have half-life periods of a few days in vivo. Therefore, their use as a treatment modality necessitates repetitive applications in order to maintain potency. For this reason, focus has been placed on designing and en- hancing delivery vehicles for growth factors. As gene transfer technology develops, the ability to control the expression and regula- tion of proteins in a host cell will enable researchers and clinicians to administer treatment over extended periods without the need for repet- itive applications. Although using gene transfer as a method for treating ligament in- juries is still in its infancy, this tech- nology has shown some promise. Indirect methods of gene transfer involving the transplantation of genetically altered tissues via a ret- roviral vector have resulted in the expression of lacZ marker gene for as long as 6 weeks in the ACLs and MCLs of animals. 32 Methods of direct gene transfer using HVJ- liposome viruses and adenoviral vectors have shown similar poten- tial. 32,33 Studies of gene transfer as a therapeutic method for manipu- lating ligament healing have also shown positive effects on collagen fibril diameter and distribution, as well as significant (P<0.05) increases in the mechanical properties. 33 Cell Therapy Another area of research with potential applications in ligament healing is cell therapy. The concept is that implantation of genetically manipulated cells can enhance the repair of ligaments as those cells become constituents of the healing tissue. In vitro and in vivo studies with mesenchymal stem cells have shown their ability to differentiate into various cell types involved in many of the phases of ligament healing. In one study involving the transplantation of nucleated cells, including mesenchymal stem cells, from bone marrow into a pocket around the transected MCL of in- bred rats, donor cells could be identified in the midsubstance of the ligament after 7 days, demon- strating the potential for migration of transplanted cells. This study highlights the possibility that cell therapy using nucleated cells may lead to new methods of treatment for ligament injuries. 34 Summary A great deal of progress has been made in elucidating the biology and biomechanics of ligament healing, which has influenced the clinical management of ligament injuries. Studying the effects of stress and motion on healing ligaments is criti- cal to understanding the healing process and the role played by mol- ecules and cells. This knowledge will be furthered by studying load- ing conditions that closely simulate the stresses and motions that occur in vivo. The future contributions of basic science research appear to lie in the engineering of ligament healing so as to accelerate the rate of healing and improve the quality of healing tissue. Potential applications in- clude the use of growth factors and the development of vehicles for their delivery. The use of scaffolds and biomatrices on which cells can be seeded may provide a better means of replacing injured soft tissues. Al- though much of the current focus has been on the ligaments of the knee, particularly the MCL and ACL, the knowledge gained may be applicable to other ligaments. Mul- tidisciplinary collaboration between molecular biologists, morphologists, bioengineers, and clinicians will en- hance the understanding of ligament healing and ultimately improve clin- ical outcomes. Acknowledgments: The assistance of Nobuyoshi Watanabe, MD, and the sup- port of the Musculoskeletal Research Center and NIH Grant #AR41820 are grate- fully acknowledged. References 1. Frank C, Woo SLY, Amiel D, Harwood F, Gomez M, Akeson W: Medial col- lateral ligament healing: A multidisci- plinary assessment in rabbits. Am J Sports Med 1983;11:379-389. 2. Derscheid GL, Garrick JG: Medial col- lateral ligament injuries in football: Nonoperative management of grade I and grade II sprains. Am J Sports Med 1981;9:365-368. 3. Frank C, McDonald D, Shrive N: Col- lagen fibril diameters in the rabbit medial collateral ligament scar: A longer term assessment. Connect Tissue Res 1997;36:261-269. Ligament Injuries in the Knee Journal of the American Academy of Orthopaedic Surgeons 372 4. Woo SLY, An KN, Arnoczky SP, Wayne JS, Fithian DC, Myers BS: Anatomy, biology, and biomechanics of tendon, ligament, and meniscus, in SR Simon (ed): Orthopaedic Basic Science. Rose- mont, Ill: American Academy of Ortho- paedic Surgeons, 1994, pp 45-87. 5. Frank C, McDonald D, Wilson J, Eyre D, Shrive N: Rabbit medial collateral ligament scar weakness is associated with decreased collagen pyridinoline crosslink density. J Orthop Res 1995;13: 157-165. 6. Niyibizi C, Visconti CS, Kavalkovich K, Woo SLY: Collagens in an adult bo- vine medial collateral ligament: Immu- nofluorescence localization by confo- cal microscopy reveals that type XIV collagen predominates at the liga- ment-bone junction. Matrix Biol 1994; 14:743-751. 7. Inoue M, McGurk-Burleson E, Hollis JM, Woo SLY: Treatment of the medi- al collateral ligament injury: I. The im- portance of anterior cruciate ligament on the varus-valgus knee laxity. Am J Sports Med 1987;15:15-21. 8. Rudy TW, Livesay GA, Woo SLY, Fu FH: A combined robotic/universal force sensor approach to determine in situ forces of knee ligaments. J Biomech 1996;29:1357-1360. 9. Withrow J, Clineff T, Abramowitch S, Papageorgiou C, Woo S: Biomechanical properties of healing goat medial collat- eral ligaments. Transactions of the 1999 Meeting of the American Society of Bio- mechanics. Pittsburgh: American Society of Biomechanics, 1999, pp 202-203. 10. Weiss JA, Woo SLY, Ohland KJ, Horibe S, Newton PO: Evaluation of a new injury model to study medial col- lateral ligament healing: Primary repair versus nonoperative treatment. J Orthop Res 1991;9:516-528. 11. Indelicato PA: Isolated medial collat- eral ligament injuries in the knee. J Am Acad Orthop Surg 1995;3:9-14. 12. Johnson RJ, Beynnon BD, Nichols CE, Renstrom PAFH: The treatment of inju- ries of the anterior cruciate ligament. J Bone Joint Surg Am 1992;74:140-151. 13. Nickerson DA, Joshi R, Williams S, Ross SM, Frank C: Synovial fluid stimulates the proliferation of rabbit ligament: Fibroblasts in vitro. Clin Orthop 1992;274:294-299. 14. Hart RA, Woo SLY, Newton PO: Ul- trastructural morphometry of anterior cruciate and medial collateral liga- ments: An experimental study in rab- bits. J Orthop Res 1992;10:96-103. 15. Woo SLY, Young EP, Ohland KJ, Marcin JP, Horibe S, Lin HC: The effects of transection of the anterior cruciate ligament on healing of the medial collateral ligament: A biome- chanical study of the knee in dogs. J Bone Joint Surg Am 1990;72:382-392. 16. Jokl P, Kaplan N, Stovell P, Keggi K: Non-operative treatment of severe injuries to the medial and anterior cru- ciate ligaments of the knee. J Bone Joint Surg Am 1984;66:741-744. 17. Frolke JP, Oskam J, Vierhout PA: Primary reconstruction of the medial collateral ligament in combined injury of the medial collateral and anterior cruciate ligaments: Short-term results. Knee Surg Sports Traumatol Arthrosc 1998;6:103-106. 18. Hillard-Sembell D, Daniel DM, Stone ML, Dobson BE, Fithian DC: Com- bined injuries of the anterior cruciate and medial collateral ligaments of the knee: Effect of treatment on stability and function of the joint. J Bone Joint Surg Am 1996;78:169-176. 19. Yamaji T, Levine RE, Woo SLY, Niyibizi C, Kavalkovich KW, Weaver- Green CM: Medial collateral ligament healing one year after a concurrent medial collateral ligament and anterior cruciate ligament injury: An interdis- ciplinary study in rabbits. J Orthop Res 1996;14:223-227. 20. Tipton CM, Matthes RD, Maynard JA, Carey RA: The influence of physical activity on ligaments and tendons. Med Sci Sports 1975;7:165-175. 21. Reider B, Sathy MR, Talkington J, Blyznak N, Kollias S: Treatment of iso- lated medial collateral ligament inju- ries in athletes with early functional rehabilitation: A five-year follow-up study. Am J Sports Med 1994;22:470-477. 22. Noyes FR, Butler DL, Grood ES, Zer- nicke RF, Hefzy MS: Biomechanical analysis of human ligament grafts used in knee-ligament repairs and reconstructions. J Bone Joint Surg Am 1984;66:344-352. 23. Woo SLY, Hollis JM, Adams DJ, Lyon RM, Takai S: Tensile properties of the human femur-anterior cruciate liga- ment-tibia complex: The effects of specimen age and orientation. Am J Sports Med 1991;19:217-225. 24. Wilson TW, Zafuta MP, Zobitz M: A biomechanical analysis of matched bone-patellar tendon-bone and double- looped semitendinosus and gracilis tendon grafts. Am J Sports Med 1999; 27:202-207. 25. Covey DC, Sapega AA: Injuries of the posterior cruciate ligament. J Bone Joint Surg Am 1993;75:1376-1386. 26. Harner CD, Höher J: Evaluation and treatment of posterior cruciate liga- ment injuries. Am J Sports Med 1998; 26:471-482. 27. Harner CD, Janaushek MA, Kanamori A, Yagi M, Vogrin TM, Woo SLY: Biomechanical analysis of a double- bundle posterior cruciate ligament reconstruction. Am J Sports Med 2000; 28:144-151. 28. Woo SLY, Horibe S, Ohland KJ, Amiel D: The response of ligaments to injury: Healing of the collateral ligaments, in Daniel DM, Akeson WH, O’Connor JJ (eds): Knee Ligaments: Structure, Func- tion, Injury, and Repair. New York: Raven Press, 1990, pp 351-364. 29. Tipton CM, James SL, Mergner W, Tcheng TK: Influence of exercise on strength of medial collateral knee liga- ments of dogs. Am J Physiol 1970;218: 894-902. 30. Shelbourne KD, Nitz P: Accelerated rehabilitation after anterior cruciate lig- ament reconstruction. Am J Sports Med 1990;18:292-299. 31. Hildebrand KA, Woo SLY, Smith DW, et al: The effects of platelet-derived growth factor-BB on healing of the rabbit medial collateral ligament: An in vivo study. Am J Sports Med 1998; 26:549-554. 32. Hildebrand KA, Deie M, Allen CR, et al: Early expression of marker genes in the rabbit medial collateral and anterior cruciate ligaments: The use of different viral vectors and the effects of injury. J Orthop Res 1999;17:37-42. 33. Nakamura N, Timmermann SA, Hart DA, et al: A comparison of in vivo gene delivery methods for antisense therapy in ligament healing. Gene Ther 1998;5:1455-1461. 34. Watanabe N, Takai S, Morita N, Kawata M, Hirasawa Y: A new method of dis- tinguishing between intrinsic cells in situ and extrinsic cells supplied by auto- geneic transplantation employing trans- genic rats. Trans Orthop Res Soc 1998; 23:1035.