Small laboratory animal models of anterior cruciate ligament reconstruction

2.1 Methods of inducing ACL injury in small laboratory animals

Some key considerations in modeling ACL reconstruction amongst various small animal species for preclinical research include the nature of the ACL injury, type and timing of treatment, and clinical relevance of the outcome measures. Many preclinical models of ACL injury that are combined with reconstruction employ surgical transection of the ligament followed by immediate treatment via reconstruction or repair. This approach has been described for mice, rats, and rabbits, but not guinea pigs.^34-39 In addition to studying ACL reconstruction, ACL complete or partial surgical transection is also used to induce posttraumatic OA to study the effects of joint degeneration, which has been described in rodents, guinea pigs, and rabbits.^{13, 40-44} While surgical-induced injury is an ethical, cost-effective, and biomechanically relevant approach, an important limitation of an ACL surgical transection injury model when compared to an induced ACL rupture model is the lack of whole-joint injury that occurs after an acute rupture (i.e., injury to the articular cartilage, meniscus, subchondral bone, and joint capsule) and the associated intraarticular responses (i.e., hemarthrosis, inflammation, and edema). The immediate reconstruction at the time of transection also does not recapitulate what occurs in human scenarios.

Alternatively, mechanically induced ACL rupture models in small animal species have been described with mice and rats, but not in rabbits or guinea pigs.^{11, 12} Noninvasive means of inducing ACL injury produce some of the associated intraarticular injuries and responses associated with ACL injury. Mechanical rupture models have been used to evaluate a range of topics including the effects of ACL injury on muscle atrophy and progression of posttraumatic OA.^{45, 46} While noninvasive means of inducing ACL injury permit a timing to reconstruction that may better replicate the human condition (i.e., surgical reconstruction that occurs separately from the ACL injury), to the best of our knowledge, there has not been a preclinical study that combined a noninvasive method of inducing ACL rupture followed by ACL reconstruction in small laboratory animals.

2.2 ACL graft ligamentization in small laboratory animals

Successful ACL graft remodeling, incorporation, and function depend on a complex interplay among factors including inflammation, joint comorbidities, graft biologic and material properties, tunnel placement and fixation method, and postoperative management protocol.⁴⁷ Animal studies can engender widely different durations and magnitudes for functional graft healing based on each of these primary factors. In general, early healing and proliferation phases for tendon grafts used for ACL reconstruction in animals exhibit a greater extent of initial degeneration compared to what is reported for human patients, such that mechanical properties of the ACL reconstruction may be more compromised initially.⁴⁸ In contrast, graft integration and remodeling are typically completed within 6–12 months in animal models, encompassing a much shorter period compared to the 3-year or longer period reported for humans.⁴⁹

There should therefore be a general recognition that much of what is known on the timing of various phases of graft ligamentization after ACL reconstruction is established in larger animal species (i.e., rabbits, sheep, and dogs), and not in smallest laboratory species such as mice and rats. Hence, there is a need for additional information on the timing of various phases of graft ligamentization amongst the smallest laboratory animal species. This has resulted in the lack of standardization on defined postoperative endpoint time points amongst mice and rat ACL reconstruction studies. For example, common study endpoints for rodents are 2 and 4 weeks whereas others have used 3 and 6 weeks.^{36, 37, 50} It is unclear what phases of graft ligamentization this time point corresponds to amongst mice and rats. The lack of long-term ligamentization studies among mice and rats suggests that current endpoint analyses may be due to established laboratory practice and study constraints rather than defined ligamentization healing phases.

In contrast to mice and rats, there are longer-term ligamentization studies in rabbits that may guide investigators on the relative timing of graft maturation phases in this species. Giordano et al.³⁸ histologically evaluated the ligamentization of semitendinosus (ST) autograft amongst New Zealand white rabbits.³⁸ One month after ACL reconstruction, numerous disorganized fibroblasts and collagenous fibers were identified during the early phases of ligamentization. A marked cellular necrosis stage was noted by 3–4 weeks after graft placement. This was then followed by a reduction of cellular necrosis that was observed in the early phase of the neo-ligament healing process. By 8 weeks, the collagen fibers had become aligned in parallel with newly formed capillaries and highly differentiated fibroblasts. At 40 days after surgery, the tissue necrosis stage was no longer present, and vascularization of the graft was readily seen. At 24 and 48 weeks the transplanted tendon still differed histologically from both semitendinosus tendon (ST) and the native ACL. Given there were no further significant histologic changes seen after 24 weeks when compared to 48 weeks, the authors concluded that the ligamentization process in rabbits may be complete by 24 weeks with semitendinosus autografts in New Zealand White rabbits.

Similar to semitendinosus autograft, a long-term rabbit study on patellar tendon (PT) autograft has also been performed. Ballock et al.³⁹ evaluated ACL reconstruction using PT autograft over 52 weeks in New Zealand white rabbits.³⁹ At 6 weeks, the graft was hypercellular with rounded or ovoid cells with some central area of hypocellularity. The graft collagen orientation was haphazard and there were no vascular channels noted at 6 weeks. At 30 weeks, the graft had areas of longitudinal fiber orientation with variable crimp patterns and occasional vascular channels along the periphery of the graft. At 52 weeks, the graft had an overall parallel alignment with vascular segments in the periphery and mid-substance of the graft. A notable finding in this study was that the anteroposterior (AP) laxity of the reconstructed knees at 52 weeks among the rabbits was greater than ACL deficient knees (5.2 ± 0.7 mm vs. 3.4 ± 0.3 mm, respectively). These findings may suggest the grafts evaluated in Ballock et al. may be clinically failed grafts that were under significant strain and stretched over the course of the study rather than successful stable grafts. In summary, investigators may need to consider what is known in terms of the timing of tendon graft healing amongst different small laboratory animal species when choosing the appropriate species for their study.

2.3 Tendon graft considerations in small animal ACL reconstruction models

The intrinsic characteristics of the tendon graft used to reconstruct the injured ligament along with the variation among the small animal species should be another primary consideration for preclinical study design.¹⁸ The superficial digital flexor (SDF)—also known as long digital flexor (LDF) in bipedal animals—and the semitendinosus (ST) tendon autografts are frequently the tendon graft used for ACL reconstruction in mice, rats, and rabbits. Comparatively, tendon grafts are more commonly used for human patients including patellar (PT), hamstring (HT), and quadriceps (QT) tendons. Traditional tendon graft choices for human ACL reconstructions are commonly more effectively employed and clinically relevant for large animal models of ACL reconstruction.^{51, 52} While time-zero tissue biomechanical properties of human tendon graft used for ACL reconstruction are widely available and typically exceeds native ACL properties,⁵³ such time-zero data regarding the material properties of tendon grafts used for small animal ACL reconstruction relative to the species' native ACL are largely lacking. It should also be recognized that the biological properties of tissue chosen for animal ACL reconstruction may vary, and therefore, need to be considered during the analysis.⁵⁴

It has been suggested that all tissue grafts used for ACL reconstruction experience a decrease in strength of more than half during the first stages of healing and only recover between 50% and 60% of the native ACL's material properties after remodeling is complete.⁴⁸ Importantly, differences between autografts and allografts in terms of differences in material properties, incorporation, and healing noted in human patients hold true in animal models as long as harvest, processing, and storage methods are similar.^{51, 55, 56} In addition, species-specific differences in knee morphology, ACL graft geometry, and composition affect biomechanical properties, increasing the importance of graft choice in each species for comparisons with human grafts. One anatomic variation amongst species worth noting that may be relevant to studying the ACL is the functional bundles concept. It has been a long-held belief that the human ACL has two main functional bundles (anteromedial and posterolateral bundles).⁵⁷ This however has been recently revisited where some have described additional bundles or contend that the ACL is actually a single ribbon-like structure.^58-61 Regardless of the number of bundles, it is likely that tendon graft reconstruction of the ACL likely does not recapture all of the exquisite detail of the native ACL. The main comparative morphological variations amongst mice, rats, and rabbits relative to humans are summarized in Table 1.

2.4 Desired outcome analyses in small animal ACL reconstruction

The final consideration in designing a preclinical study is the available study outcome analyses that may be used to answer the study question. Serial clinical exams to assess ACL graft laxity (i.e., Lachman exam or anterior drawer), which is possible in larger animal species,^{25, 51, 52, 64} are not as easily performed in the smaller laboratory animal species, such as mice and rats. Similarly, functional assessment, such as gait analysis, is achievable but may be comparably less established after ACL reconstruction in small animal laboratory species when compared to certain larger animal models, particularly dogs where it is commonly used in canine clinical care.^{51, 64} Biomechanical endpoint analyses of small animal femoral-ACL-tibia-complexes however are commonly performed including time-zero and/or initial endpoint measures of load-to-failure and stiffness.^{18, 34-37, 65} Less commonly, Young's modulus, absorbed energy, and displacement are reported. High-resolution imaging including microcomputed tomography (μCT), magnetic resonance imaging (MRI; ≥7 T), and micro-polyethylene terephthalate (PET) have all been used to evaluate graft healing after small animal models of ACL reconstruction.^66-68 A variety of histological endpoint assessments including the intraarticular ACL graft, graft tunnels, and chondral surfaces are also readily performed in small animal models. The main characteristics of animal studies involving ACL reconstruction in small animal species, including type of procedure, graft used, postoperative management, and outcome measures are summarized in Supporting Information: Table S1.

3.1 Rodent ACL reconstruction models

Mice and rats can serve as important models for biomedical research, particularly mechanistic studies targeting ACL injury and reconstruction.^{69, 70} The anatomical, physiological, and genetic similarities to humans in combination with their small size, ease of handling, capabilities for genetic manipulation, and abundant availability foster the use of rodents in ACL research.⁶⁹ Further, rodents present economic advantages based on costs for animals, housing, and maintenance as well as shorter timeframes to reach skeletal maturity. Mice have been recently used for ACL reconstruction studies,^{71, 72} and its inherent advantages relative to other small animal species include several transgenic species of mice that are available, which permits evaluation of impact of specific genes in healing after ACL reconstruction via knockout models. Based on its larger relatively larger size and associated technical advantages, rats have traditionally served as a more attractive rodent model with the Sprague–Dawley strain being the most popular.^{17, 37, 73, 74}

3.2 Tendon graft choices in rodent ACL reconstruction

The SDF tendon autograft is the primary choice for ACL reconstruction in both mice and rats based on access, harvest technique, and size (Figures 1 and 2).^{17, 35-37, 50, 72, 75-77} None of the tendons typically used for human ACL autografts nor any type of allografts have been commonly used in both rodent models, which hinders direct comparison of graft properties across species.

3.3 Outcome measures in rodent ACL reconstruction models

The “grimace scale” provides a standardized, semiautomated method to evaluate pain severity in rodents.^{78, 79} To assess function, gait analysis methods for rodent models measure stride width, stride length, and velocity by applying stamp-pad ink to the hind paws before the animal walking through a tunnel lined with white paper.⁸⁰ Semiautomated methodologies for kinematic gait analysis that incorporate video analysis, custom arthrometers, and fluoroscopic analysis to evaluate flexion-extension and anterior-posterior translation have been specifically designed for small laboratory animals.^81-87 Additionally, controlled loading after surgery while the animal is chemically restrained has been performed to make comparisons among physical therapy protocols and activity regimes.^{35, 72} Other functional outcome measures used in rodents, such as lameness scores, have also being adapted from other larger species and applied in several studies.^81-90

Imaging outcomes in rodents usually include the evaluation of tendon-to-bone integration assessed using μCT. Bone volume fraction, bone density, and area metrics for bone tunnels provide valuable information in regards to the tendon graft integration with the bone.^{35, 37, 72, 75, 91} The recent advancement and availability of high-field MRI scanners along with improved designs in radiofrequency and more efficient pulse sequences have made high-quality images feasible for acquisition in small rodent knees.⁶⁷ High-strength (≥7 T) MRI can detect cartilage thickness changes and arthritic features that are challenging at lower strengths.^{67, 68} High magnetic fields with specifically adapted measurement strategies permit direct or indirect measures for rodent cartilage pathobiology.⁹² Goebel et al.⁹³ used 7 T MRI after ACL transection (ACLT) to evaluate cartilage thickness in weight-bearing areas of the knee to compare control and OA-induced rats. In addition, techniques such as ultra-short-echo time along with contrast agents (gadolinium) facilitate the assessment of hydration status and estimation of glycosaminoglycan content that could indicate pathological processes within the joint.⁹⁴ However, due to specific anatomic and histologic variations in rodents with respect to humans, such as persistent epiphyseal cartilage, increased posterior curvature of the femoral condyles, and a distally fused tibiofibular joint, direct extrapolation of findings may be limited.^{71, 95, 96}

In addition to imaging analysis, other data related to cartilage and joint health can be obtained from rodent ACL models. While serum, urine, and synovial fluid (SF) biomarkers may be useful to assess joint inflammation and cartilage degradation in humans, rodents are often limited by the volume of body fluids that can be collected. As an example, only 0.2–0.4 and 1–2 ml of blood can be collected from mice and rats, respectively, using jugular vein access under anesthesia, and only 0.5–1 ml of blood can be collected from mice, or 5–10 ml from rats, when using cardiac puncture at study termination.^{97, 98} SF collection is much more complicated and only ~1 and ~100 µl can be recovered after joint lavage in mice and rats respectively, after adjusting for dilution.^{83, 84, 89, 90, 97, 98} Alternative methods of SF recovery have been described including Whatman Paper Recovery, Polyacrylate Bead recovery, and calcium sodium alginate compound recovery, which may facilitate extracting small amounts of SF for biomarker analysis.⁸⁹

Biomechanical testing is a common procedure that evaluates the material properties of grafts before and after surgery. AP displacement of the tibia with respect to the femur, load to failure, and stiffness are the main outcome measures assessed. However, lack of consistency in testing methods has resulted in some degree of variability in reported measures. Fu et al.³⁶ used a 70° flexion and an anterior pulling force (2 N) applied to the tibia at a displacement rate of 20 mm/min to test AP displacement, and tested load to failure at 50 N with a displacement rate of 40 mm/min. Kawakami et al.⁵⁰ tested load to failure at a 0.25 mm/s (15 mm/min) displacement rate, while the Rodeo et al. uses a 0.167 mm/s (10 mm/min) rate.^{37, 75} Hence, such discrepancies may affect interpretation of biomechanical results, especially when added to other confounding variables such as animal sex, age, or timepoint differences among studies. The biomechanical results of mice and rat ACL reconstructions are listed in Table 2.

Graft healing within the bone tunnel is commonly evaluated using histological scores that have been adapted from other species and are now well established for rodent tissues.^{36, 99} A graft degeneration scale from 0 to 3 was used in the study by Fu et al. under bright field and polarized illumination.³⁶ In addition, this score evaluates healing in the bone-tendon interface and adverse peri-graft bone changes (defined as the presence of bone intrusion into the tendon graft, peri-graft bone erosion, or the presence of cysts in peri-graft interface).³⁶ Another histologic score developed for rabbits has also been used to assess ACL graft incorporation in rats.^{99, 100} Histologic scores for tendon-bone healing assessment in rodents are summarized in Supporting Information: Table S2.

3.4 Clinical questions evaluated using rodent ACL models

The rat model has been particularly valuable for improving ACL reconstruction techniques by assessing the effects of tunnel location, ACL tendon graft pretension magnitude, and postoperative activity protocols on graft strain, tendon-to-bone integration, graft remodeling, and graft survival.^{17, 18, 37, 75, 101-103} Similarly, murine models of ACL reconstruction have been used to study the tendon-to-bone formation and the impact of postoperative mechanical load on graft tunnel incorporation after ACL reconstruction.^{34, 35} Although these studies have significantly enhanced our understanding regarding pretension magnitude, graft positioning, and the importance of controlling the postoperative mechanical load on reconstructed rodent knees, consensus in experimental design still needs to be reached. Moreover, the use of high graft pretension magnitudes relative to body weight poses an additional challenge in how to determine a correct comparison method that allows for valid translation to clinical application. Augmentation techniques to enhance graft-bone integration and to mitigate bone tunnel diameter expansion around the graft along with the improved histological appearance and higher tensile load to failure are focuses of current work using rodent models.⁵⁰

3.5 Rabbit ACL reconstruction models

Rabbit models have been effectively used with reported advantages over rodent models, including animal size, graft selection, and sampling capabilities.^{18, 65, 104-107} While numerous variations of lapin ACL reconstruction have been described, a New Zealand White rabbit model using long digital extensor (LDE) tendon autograft is the most used model.¹⁰⁸

3.6 Graft choices for rabbit ACL reconstruction

Common grafts used to reconstruct the ACL in rabbits include PT, ST, and LDE autografts.^{18, 65, 104-106, 108, 109} However, due to the variation in size, especially cross-sectional area, the mechanical properties of rabbit tendon grafts vary significantly (Table 3). While the PT graft can be harvested similar to human patients, it is not as commonly used as ST and LDE grafts. While the ST can be harvested without complications, the graft is significantly smaller and shorter when compared to the LDE graft, making the surgical procedure more challenging (Figure 3B). On the other hand, LDE grafts are more easily harvested due to their increased length and more superficial distribution along the distal hindlimb (Figure 3A). Although the tendon autografts used in rabbits may be similar to those used in humans, the need for open arthrotomy and lack of specific instrumentation for implantation and fixation used to reconstruct the ACL, could limit the translational potential of this model (Figure 4).

3.7 Outcome measures in rabbit ACL reconstruction models

Like rodents, the grimace scale is often used as an assessment of pain severity in rabbits.^{110, 111} Due to the natural position and structural biology of rabbit hindlimbs, which ambulate through hopping instead of walking, forces applied to the ACL graft might not be comparable to other species. Kinematic gait analysis is therefore only accurate when using complex methods such as infrared emitting diodes implanted through intracortical pins in the tibia and the femur in rabbit ACL models.¹¹² Functional evaluation in the rabbit is also assessed by adapting the Lachman test, range of motion, and lameness score from humans and larger species.^112-114

High-resolution MRI along with adapted grading has been used in rabbit models of ACLT to evaluate cartilage integrity.¹¹⁵ Kajabi et al.¹¹⁶ used multiparametric 9.4 T MRI to demonstrate early degenerative changes in the superficial cartilage of an ACLT rabbit model. Furthermore, Chai et al.¹¹⁷ used 7.1 T MRI to evaluate ligament-bone integration, tunnel diameter, and signal-to-noise ratio in a rabbit model of ACL reconstruction using a synthetic PET ligament.¹¹⁷ Additional imaging techniques, such as µCT scans, have also been implemented in the rabbit model to assess tendon-bone integration. For example, Bi et al.⁶⁵ utilized µCT to demonstrate that reconstructing the rabbit ACL with silk-collagen scaffold decreased the average femoral bone tunnel area compared to the semitendinosus autograft at 4 weeks postoperatively.⁶⁵ Quantitative computed tomography was used to determine that a 3-week COX-2 inhibitor (celecoxib) treatment decreased area of trabecular bone, cortical density, and new bone formation.¹⁰⁴ Moreover, µCT was also used to evaluate new bone formation between an additive manufactured (AM) porous-titanium interference screw implant and a conventional interference screw.¹⁰⁵

In contrast to rodents, evaluation of ACL healing and OA biomarkers is more feasible in the lapin model due to larger volumes of blood and SF.^{118, 119} Moreover, Luo et al.¹²⁰ demonstrated significant differences in SF proteome between ACL transected and control rabbits, including adiponectin, pyruvate kinase, fibrinogen, paraoxonase-1, and carboxilesterase-2. Further, the concentrations of inflammatory mediators (IL-1β and IL-17) and gene expression of the degradative enzyme, MMP-13, were also significantly higher in antero-medial bundle and posterolateral bundle resection groups compared to controls.¹³

With regard to biomechanical testing, Woo et al.¹²¹ described a detailed protocol to evaluate the structural properties of the rabbit femur-anterior cruciate ligament-tibia complex (FATC). However, there is not a consensus on the most appropriate biomechanical testing protocol for the rabbit FATC, with published reports using different knee flexion angles ranging from 45° to 90° flexion during testing and different displacement rates ranging from 0.5, 5, 10, 20, 25, to 60 mm/min (1 mm/s).^{18, 65, 104-106, 122} Because of the variation in testing technique, it is difficult to compare biomechanical outcomes among studies.

In terms of quantitative histological assessment, a four-point histological grading score has been developed for ACL reconstruction in the rabbit.¹²³ Demirag et al.¹²³ used a semiquantitative histological score to assess the healing process at the bone-tendon interface based on the presence of fibrovascular tissue consisting of fibroblasts and inflammatory cells and newly formed Sharpey-like collagen fibers (Supporting Information: Table S2).¹²³

3.8 Clinical questions evaluated with rabbit ACL models

Like rodents, lapin models have been used to evaluate the effects of surgical technique on ACL healing. Graft tension magnitudes in rabbits range from 1 to 17 N, the latter being associated with initial (2-week postoperative) occurrence of necrosis and lower cellularity.¹⁸ However, careful interpretation of results is recommended, especially when applying high pretension to the graft, which could potentially lead to cartilage damage as seen in humans when magnitudes above 40 N are implemented.¹²⁴ Therefore, adequate scaling of tension magnitudes to each animal model is needed.

In addition, rabbit model studies have been able to examine the effects of inflammatory inhibition, growth factors, and bio-enhancement on tendon graft-bone integration.^104-106 Based on their relative size and knee anatomy, clinically relevant fixation devices have been evaluated and SF biomarkers have been assessed in rabbits.^{104-106, 122} Consequently, the rabbit model holds promising potential for cost- and time-effective preclinical ACL reconstruction studies aimed at screening methods for improving graft tendon healing and incorporation.