Treatment of tendon and ligament injuries remains a challenge in human and veterinary orthopedics. Following injury, natural tendon healing results in scar tissue formation that has biomechanical characteristics inferior to those of healthy, uninjured tendon.1 Tendon healing is further hindered by the low vascularity and cellularity of tendons, low number of progenitor cells, and mechanical stress inherent to the local tissue environment.2,3 Ligament or tendon injury that is extensive, chronic, or resistant to treatment may warrant replacement.
Tendon and ligaments may be replaced with synthetic and natural substances. Tissue-engineered scaffold biomaterials such as synthetic polymers (eg, polyglycolic acid) and natural ECM components (eg, collagen derivatives) have been developed in an attempt to improve and accelerate healing or reconstruct tendons altogether. Limitations of these scaffolds include their limited biocompatibility, rapid or unpredictable degradation rates, and biomechanical weakening caused by degradation.4 Ultimately, tendon healing that occurs naturally or through tissue-engineered constructs is comprised predominantly of type III collagen, which is structurally weaker than the type I collagen found primarily in healthy tendons.5
Autografts and allografts provide an alternative method for tendon and ligament reconstruction to circumvent some of the limitations of tissue-engineered constructs. Both types of grafts represent a natural ECM scaffold that initially provides biomechanical properties similar to those of the native tendon as well as a biochemical and structural composition to guide cell growth. In addition, growth factors may be trapped within the natural ECM, providing a source of tenogenic stimuli that promotes cell differentiation.6 However, the potential for considerable donor site morbidity associated with autografts sometimes favors the use of allografts.7,8 Because allografts are antigenic and induce an immune response when transplanted, development of strategies to remove cellular antigens has become an increasingly popular area of research.9
Decellularization refers to the process of removing cells from tissue specimens while preserving the structural and functional proteins that constitute the ECM. As a result, cell-associated immunogenic antigens are eliminated from the tissue, enhancing its biocompatibility. Decellularized tissue scaffolds provide a large source of readily available, biocompatible graft material that can resemble native tissue structurally, biochemically, and biomechanically. Such properties facilitate graft incorporation and survival.10
Many decellularization protocols have been described for various tissues with different results. Adverse effects of some decellularization processes include impairment of the viability of colonizing cells, loss of growth factors bound by the ECM, washout of GAGs, loss of ECM structure, and induction of sterile inflammation.11 Although milder protocols may preserve architecture, they result in a greater amount of DNA remaining than is achieved with harsher protocols. Harsher protocols, on the other hand, may facilitate removal of residual DNA but do so at the expense of eliminating desirable ECM. Consequently, there is a need to balance cell- and antigen-removal properties while preserving tissue composition and mechanical properties.
A described technique for evaluating the effectiveness of immunologic sterilization of tendons is DNA quantification.10 A previous study12 involving tendon decellularization identified a mean percentage of DNA reduction from 44% to 83% for more effective protocols. In addition to verifying removal of antigenic debris, one must also confirm that desirable components of the ECM (ie, GAGs, collagen, and protein) have been retained. These components are required for matrix infiltration during graft incorporation and to maintain biological and biomechanical properties of the tendon. Therefore, mechanical testing of the decellularized ECM may be used to evaluate the integrity of the structural proteins.9
Tendon allografts have used been used clinically in canine medicine, yet few studies13–16 have been conducted to investigate the effects of decellularization on canine tendons. The objective of the study reported here was to characterize the biochemical, histologic, and biomechanical properties of commercially available native and decellularized SDFTs and DDFTs from the pelvic limbs of orthopedically normal dogs. We adopted the null hypothesis that decellularized tendons would have biochemical and biomechanical characteristics similar to those of native tendons, with minimal cellular debris (DNA content).
Materials and Methods
Specimen preparation
Pelvic limb flexor tendons (SDFTs and DDFTs) that had been harvested from cadavers of previously healthy adult dogs were obtained from a clinical donor program.a Owners of donor dogs had signed a form confirming that their dog died of nonorthopedic causes and authorizing the collection of tissues for transplantation to benefit other animals or for use in research related to improving transplantation medicine. No dogs were euthanized for the purpose of tissue donation. Within a few hours after euthanasia, tissue samples were aseptically harvested in a clean room facilitya and quarantined at −70°C until processed.
A swab specimen was collected from each tendon specimen and submitted for culture. Tendon specimens were excluded if microorganisms were detected or if the duration of frozen storage exceeded 5 years. The 13 donor dogs (9 males and 4 females) ranged in age from 6 months to 4 years and in body weight from 37.3 to 100 kg. When pairs of tendons from the same donor were available, one was processed without decellularization and the other was processed with decellularization. For SDFT specimens, 4 pairs of tendons were collected from 4 dogs and 3 individual tendons were harvested from 3 dogs. For 4 pairs of large-diameter SDFTs, tendons were split in half along the long axis and subsequently processed. For DDFT specimens, 6 pairs of tendons were collected from 6 dogs. Processing yielded 10 paired SDFT specimens (5 native and 5 decellularized) and 12 paired DDFT specimens (6 native and 6 decellularized).
Processing without decellularization
Frozen tendon specimens were thawed in warm PBS solution, tagged with an identifying marker, and cleaned of connective tissue and muscle. Specimens were then left to soak overnight in PBS solution to further release any residual hemoglobin, lipids, and other cellular elements from cells ruptured by the freeze-thaw process. The following day, specimens were measured (length and circumference) and the dimensions were recorded. Specimens were wrapped in gauze, covered with PBS solution, sealed into double-layer peel pouches, and frozen at −70°C until shipped on dry ice for testing.
Processing with decellularization
Frozen tendon specimens were tagged, cleaned, and measured exactly as described for native (processed without decellularization) tendons. Whole specimens were then decellularized by use of a protocol based on a published method.17 Briefly, specimens were treated with a series of solutions, including a hypotonic buffer solution followed by a hypertonic buffered solution containing a nonionic detergent to rupture cells and facilitate removal of lipid membranes. Afterward, specimens were processed by deoxyribonuclease enzymatic digestion in a buffered saline (0.9% NaCl) solution followed by treatment with an anionic detergent at mild alkaline pH to facilitate removal of intracellular elements of the ruptured cells. Final steps included an overnight soak in PBS solution and brief water washes to remove residual cellular debris. Specimen dimensions were again measured after decellularization and recorded. Tendon specimens were then packaged and stored as previously described.
All tendon specimens were thawed at room temperature (approx 21°C) for 24 hours prior to evaluation and mechanical testing. After the thaw stage, the proximal musculotendinous portion of each specimen was excised and discarded. The adjacent distal 2-cm portion was then collected for biochemical and histologic analysis. The remaining specimen was used for mechanical testing. Blinding of investigators to tendon group assignment was performed by numeric labeling of each tendon sample at the time of preparation. Testing order randomization was accomplished by drawing of specimen numbers from a box.
Biochemical characterization
Native and decellularized tendon specimens were solubilized prior to testing. First, specimens were coarsely minced (cut into portions measuring 8 to 27 mm3), weighed, placed in 1.8-mL sterile, low-protein–binding polypropylene microcentrifuge tubes, and dehydrated in an oven for 18 hours at 80°C. Dry weight of specimens was measured, and portions were allocated for papain or pepsin digestion. Prior to digestion, the dehydrated specimens were placed individually into a cryogenic mortar and pestle system that had been previously cooled in liquid nitrogen for 20 minutes. Specimens were pulverized until a powder consistency was achieved. The mortar and pestle were cleaned thoroughly with 70% ethanol between specimens.
Papain digestion was achieved by adding 1 mL of 0.1% papain solutionb to dehydrated, pulverized tendon tissue (mean dry weight, 4.7 mg) and incubating in a 65°C water bath for 96 hours prior to biochemical testing. The DNA content of the papain digests was quantified with a fluorescent nucleic acid stain kitc in accordance with the manufacturer's protocol. Briefly, a DNA standard curve was generated by measuring the fluorescence of double-stranded DNA solutions of various known concentrations. Sample analysis was performed by evaluating sample fluorescence following serial dilution. The DNA concentration was determined by comparison of specimens to the standard curve.
Protein concentration in papain digests was measured by use of a colorimetric protein assay kitd in accordance with the manufacturer's standard microplate protocol. This assay yields an estimate of protein concentration by spectrophotometric analysis of color changes produced when protein binds to a triphenylmethane dye. Protein concentration in dry tissue was determined by evaluating specimen absorption relative to a standard curve.
Sulfated GAG concentration was assessed by use of a dye-binding proteoglycan detection kite in accordance with the manufacturer's protocol. The kit involves a metachromatic dye (1,9-dimethylene blue) that binds sulfated GAGs. Binding induces a shift in the absorption spectrum, which is directly proportional to the amount of sulfated GAGs. A standard curve was generated, and proteoglycan concentration in dry tissue was measured for each specimen.
Pepsin digestion was achieved by placing 2.7 mg of each pulverized tendon specimen in a 1.8-mL sterile, low-protein–binding polypropylene microcentrifuge tube. One milliliter of cold acid-pepsin solution (0.1 mg of pepsinb/1 mL of 0.5M acetic acid) was added to each tube. All tubes were agitated by vortex device for 10 seconds and incubated at 4°C for 96 hours.
Solubilized collagen concentration in pepsin digests was measured with a collagen assay kitf in accordance with the manufacturer's protocol. In the assay, dye binds to a repeating tripeptide sequence (gly-X-Y) conserved among the fibrillar collagens (types I, II, III, V, and IX). Dye binding affects spectrophotometric absorbance, which was read at 544 nm. A standard curve was generated by use of linear regression mathematic software,g and collagen concentration in dry tissue was extrapolated.
Biochemical tests for all specimens were run in duplicate, and all data points yielded a coefficient of variance of < 10%.
Histologic analysis
Tendon tissue sections (2-cm portions) were processed by fixation in 10% formalin for 3 months and then transferred to 70% ethanol for long-term storage. Sections were submitted for routine histologic staining with H&E to evaluate fiber pattern and cellularity. Presence of tissue DNA was further evaluated by hybridization with EthD-1b and examination with a digital inverted fluorescence microscopeh as previously described.12 The EthD-1 is a high-affinity nucleic acid stain that is weakly fluorescent until bound to DNA and emits red fluorescence. Brightness and exposure settings were standardized for all sections. Each section stained with H&E or EthD-1 was evaluated for the number of nuclei per hpf. The same pathologist reviewed 3 hpfs/sample to determine the frequency of nuclear content.
Biomechanical testing
Twenty-two constructs were tested, consisting of native tendon specimens (5 SDFT and 6 DDFT) and decellularized tendon specimens (5 SDFT and 6 DDFT). Tendon specimens were thawed and then maintained in saline solution–soaked gauze throughout the testing procedure until immediately before mechanical testing. Anatomic dimensions of each specimen were recorded by use of CTi prior to testing. Images were reviewed for evidence of irregularities or defects. After specimens were sectioned for biomechanical testing as described previously, cross-sectional areas were measured on CT images at the midpoint of each specimen as well as at 1 cm proximal and distal to the midpoint. The cross-sectional area used for calculation of biomechanical properties was the mean of these 3 measurements.
Specimens were mounted onto a servohydraulic testing machinej by passing the tendon through a closed ring and securing the tendon ends with alligator-tooth clamps (Figure 1). Clamp-to-ring distance was 1 cm for all constructs. Tendon constructs were exposed to static tensile preloading of 20 N for 60 seconds, followed by preconditioning consisting of a cyclic tensile load modulating between 20 N and 150 N at 1 Hz for 150 cycles. After the preconditioning stage, tension was applied to failure at a rate of 0.5mm/s. Tensile force and displacement were recorded at a sampling rate of 20 Hz. Mechanical testing was videorecorded from the start of static tensile loading through to failure. Force resulting in 3 mm of displacement as well as force and displacement at ultimate failure were recorded. Mode of failure (midsubstance vs clamp-tendon interface) for each construct was recorded. Stiffness was calculated as the slope of the central third of the stress-strain curve between the beginning of testing and ultimate failure. The Young modulus was also calculated for each construct.
Photograph of the testing apparatus used for evaluation of biomechanical properties of SDFT and DDFT specimens harvested from canine cadavers, with a tendon secured in closed-ring configuration. Specimens were mounted onto a servohydraulic testing machine by passing each tendon through a closed ring and securing the tendon ends with alligator-tooth clamps. After a preconditioning stage, tension was applied to failure at a rate of 0.5 mm/s.
Citation: American Journal of Veterinary Research 77, 4; 10.2460/ajvr.77.4.388
Statistical analysis
Statistical analysis was completed in 3 steps: data quality check, assessment of within-subject independence, and pairwise group comparisons with the conservative Wilcoxon rank sum test. Data quality and distributional assumptions were checked with summary statistics and data plots.
In pairwise group comparisons, tendon specimens from the same dog appeared in both groups (ie, native SDFT and native DDFT), violating the independent groups assumption of the Wilcoxon rank sum test. The effect of this contravention was assessed by inspection of scatterplots to determine Pearson correlations and assess dependence. The second step was a sensitivity analysis to determine the effect of dependence on the analysis. Tendons were omitted to make independent groups and the analysis was rerun, with no change in the findings.
Because data dependence had no effect on final inferences, native DDFT specimens were compared with native SDFT specimens, decellularized DDFT specimens were compared with decellularized SDFT specimens, native DDFT specimens were compared with decellularized DDFT specimens, and native SDFT specimens were compared with decellularized SDFT specimens by use of the 2-sample normal approximation Wilcoxon rank sum test. Threshold for significance for all tests was a value of P < 0.05.
Results
Biochemical characterization
Mean DNA concentration in decellularized SDFT specimens harvested from canine cadavers (3.22 ng/mg; range, 1.0 to 4.89 ng/mg) was significantly (P = 0.003) less than that of the native SDFT specimens (12.56 ng/mg; range, 6.52 to 30.41 ng/mg). No significant differences in GAG, collagen, or protein concentrations were identified between decellularized and native SDFT specimens (Table 1).
Mean ± SD DNA, GAG, collagen, and protein concentrations (μg/mg of dry tissue) in commercially obtained native and decellularized SDFT and DDFT specimens harvested from the pelvic limbs of 13 canine cadavers.
| Specimen type | DNA | GAG | Collagen | Protein |
|---|---|---|---|---|
| Native SDFT (n = 5) | 12.56 ± 8.48 | 2.0 ± 0.61 | 4.04 ± 0.82 | 50.46 ± 9.63 |
| Decellularized SDFT (n = 5) | 3.22 ± 1.48 | 4.99 ± 0.92 | 5.39 ± 0.54 | 49.26 ± 3.31 |
| Native DDFT (n = 6) | 13.45 ± 9.82 | 3.52 ± 2.13 | 4.05 ± 0.68 | 48.45 ± 18.71 |
| Decellularized DDFT (n = 6) | 2.45 ± 3.00 | 5.68 ± 2.37 | 4.04 ± 0.81 | 49.90 ± 14.39 |
When pairs of tendons from the same donor were available, one was processed without decellularization and the other was processed with decellularization. For SDFT specimens, 4 pairs of tendons were collected from 4 dogs and 3 individual tendons were harvested from 3 dogs. For 4 pairs of large-diameter SDFTs, tendons were split in half along the long axis and subsequently processed. For DDFT specimens, 6 pairs of tendons were collected from 6 dogs. Processing yielded 10 paired SDFT specimens (5 native and 5 decellularized) and 12 paired DDFT specimens (6 native and 6 decellularized).
Mean DNA concentration of decellularized DDFT specimens (2.45 ng/mg; range, 0 to 6.94 ng/mg) was significantly (P = 0.007) less than that of native DDFT specimens (16.24 ng/mg; range, 3.75 to 33.35 ng/mg). No significant differences in GAG, collagen, or protein concentration were identified between native and decellularized DDFT specimens. Although not statistically different, mean reduction in DNA content was greater for DDFT specimens (80.4%) than for SDFT specimens (72.6%).
Compared with each other, decellularized SDFT specimens were comprised of significantly (P = 0.005) more collagen than were decellularized DDFT specimens. For protein, GAG, and DNA concentrations, no statistical differences were identified between the 2 tendon groups.
Histologic analysis
Both H&E staining and red fluorescence EthD-1 hybridization of tissue sections from native tendon specimens revealed tenocyte nuclei with the typical elongated shape, arranged in parallel with collagen fibers (Figure 2). In contrast, decellularized tendon specimens had similar fiber patterns with considerably fewer tenocyte nuclei.
Representative photomicrographs of histologic sections of tendon specimens harvested from the same dog and stained with H&E (A and B) or EthD-1 (C and D), showing native (A and C) and decellularized (B and D) tissue. Considerable loss of DNA content in decellularized tissue is characterized by lack of nuclei (loss of blue staining in H&E sections and loss of red fluorescence in EthD-1 treated samples). Bar = 50 μm.
Citation: American Journal of Veterinary Research 77, 4; 10.2460/ajvr.77.4.388
Biomechanical testing
No significant differences in biomechanical properties were identified among the 4 tendon groups tested (Table 2). Mode of failure was at the tendon-clamp interface for all DDFT constructs (Figure 3). In contrast, 4 of 10 SDFT constructs failed at midsubstance in the region adjacent to the anatomic cap, where the tendon would have crossed the tuber calcanei. The remaining 6 SDFT constructs failed at the tendon-clamp interface.
Representative photographs of the 2 types of failure in canine tendon specimens that occurred during the biomechanical testing shown in Figure 1. A—Failure occurred at the clamp-tendon interface for this native DDFT specimen. B—Failure occurred midsubstance for this native DDFT specimen.
Citation: American Journal of Veterinary Research 77, 4; 10.2460/ajvr.77.4.388
Mean ± SD biomechanical values for the canine tendon specimens in Table 1.
| Specimen type | Load at 3 mm (N) | Displacement at failure (mm) | Load at ultimate failure (N) | Stiffness (N/m) | Young modulus (N/mm2) |
|---|---|---|---|---|---|
| Native DDFT (n = 5) | 1524.8 ± 397.1 | 5.1 ± 1.9 | 2014.3 ± 229.5 | 473.7 ± 146.3 | 136.4 ± 52.9 |
| Decellularized DDFT (n = 5) | 1307.7 ± 398.9 | 5.4 ± 0.9 | 1954.5 ± 620.1 | 445.8 ± 124.3 | 114.8 ± 37.2 |
| Native SDFT (n = 6) | 1257.3 ± 282.2 | 4.6 ± 1.2 | 1721.3 ± 729.9 | 420.9 ± 110.4 | 101.3 ± 24.0 |
| Decellularized SDFT (n = 6) | 1288.3 ± 381.0 | 4.3 ± 0.6 | 1,594.0 ± 368.7 | 413.2 ± 154.7 | 129.7 ± 49.3 |
See Table 1 for key.
Discussion
Decellularized canine SDFT and DDFT specimens in the present study had biomechanical properties similar to their native counterparts and to each other. However, decellularization of both types of tendon specimens yielded significant decreases in DNA content relative to that of native tendon specimens. We therefore rejected our null hypothesis that decellularized tendons would have biochemical characteristics similar to those of native tendons. On the other hand, we failed to reject the null hypothesis that decellularized tendons would have biomechanical properties similar to those of native tendons. These findings suggested that canine tendon specimens that were decellularized in accordance with the described protocol may provide a reasonable tendon scaffold for tendon or ligament replacement in dogs because it had mechanical properties similar to those of the tissue it would be used to replace with less foreign antigen content than in native tendon specimens.
The purpose of DNA quantification in the present study was to evaluate the effectiveness of the decellularization process for removal of cellular antigens. Both decellularized SDFT and DDFT specimens contained significantly less DNA than did paired native tendons via biochemical testing and histologic evaluation. These findings were similar to those achieved with previously published protocols.12 Reduction in DNA content was greater for DDFT specimens than for SDFT specimens. This apparent difference between tendon types may have been attributable to the anatomy of the SDFT, variations in the decellularization process, and natural variation in tissue composition. Gross examination revealed that SDFT specimens had a dense cap at the point where they would have crossed over the calcaneal bursa in the pelvic limb from which they had been harvested. It is recognized that the effects of decellularization on the ECM are dependent on the source, composition, and density of the tissue.11 Therefore, the structural composition of this cap may have mitigated the effects of decellularization on DNA removal from SDFT specimens.
Despite the significant decrease in DNA content following decellularization of tendon specimens, small amounts of nuclear material or cytoplasmic debris remained in the scaffold. The biological consequences of such remnants remain unclear; however, there is reportedly no direct cause-and-effect relationship between these remnants and an adverse host response.11 Regardless, minimizing DNA content without disrupting desirable tissue qualities remains the goal of decellularized-graft creation.12
Protein, collagen, and GAGs are components of the ECM important for maintaining the biological and mechanical integrity of tendons.9 The goal of decellularization is to eliminate DNA content while preserving the native ECM. In this respect, the decellularization protocol used in the present study was successful given that no significant reduction in protein, collagen, or GAG content was appreciated in decellularized versus native tendons. Interestingly, decellularized SDFT specimens had significantly greater collagen content than did decellularized DDFT specimens. This could be most easily explained by differences in tendon anatomy and natural composition.
Biochemical findings in the present study were supported by the biomechanical data. Results of mechanical testing revealed no significant differences in biomechanical properties among any of the tendon groups. Indirectly, these results suggested that the decellularization method used did not dramatically affect the components of the ECM contents and structure. These results also corroborated the findings of other studies,10,12,18 in which the biomechanical properties of mammalian tendons were evaluated by use of various decellularization protocols. In the study reported here, we elected to pass the tendon over a bar and secure the 2 tendon ends in a clamp. This approach was chosen because tendons are passed over a fixation pin for reconstruction of the anterior cruciate ligament in humans.19–21 In addition, in previous mechanical testing studies22,23 in which graft performance was evaluated for reconstruction of the anterior cruciate ligament (humans) and cranial cruciate ligament (dogs), grafts had superior mechanical properties when the tendon was passed over a transfemoral pin and tendon fixation to the bone was limited to a single site.
Mode of failure between SDFT and DDFT specimens was often different in the present study. Grossly, DDFT specimens appeared uniform throughout their length. However, when secured in the testing apparatus, the tissue was crushed at the tendon-clamp interface. Crushing compromises the structural integrity of the tendon, rendering it biomechanically weaker and thus predisposing to failure at this junction. The fairly high frequency of midsubstance failure for SDFT specimens was likely attributable to a natural region of elastic-modulus mismatch adjacent to the dense anatomic SDFT cap. The abrupt change in material properties in this region predisposes the tendon to this particular mode of failure.
In the study reported here, decellularized canine flexor tendons had mechanical properties similar to those of native tendons, with decreased antigen (DNA) content. As such, they may provide suitable, biocompatible graft scaffolds for bioengineering applications such as tendon or ligament repair or replacement, with less risk of immunologic rejection than nondecellularized allograft tendons. Deep digital flexor tendons may provide superior graft characteristics because of their anatomic uniformity, which was lacking in SDFT specimens. Long-term biocompatibility of decellularized canine DDFTs and SDFTs remains to be assessed in future in vivo studies involving dogs.
Acknowledgments
The authors thank Qingshan Chen for assistance with mechanical testing, Josh Parker for digital imaging and figure preparation, and Paula Overn for histologic preparations and EthD-1 staining of samples. Tendons were provided by Veterinary Transplant Services.
ABBREVIATIONS
| DDFT | Deep digital flexor tendon |
| ECM | Extracellular matrix |
| EthD-1 | Ethidium homodimer-1 |
| GAG | Glycosaminoglycan |
| SDFT | Superficial digital flexor tendon |
Footnotes
Veterinary Transplant Services, Kent, Wash.
Sigma-Aldrich, St Louis, Mo.
Quant-iT PicoGreen dsDNA reagent kit, Invitrogen, Grand Island, NY.
Coomassie Plus Bradford assay kit, Thermo Scientific, Waltham, Mass.
Rheumera proteoglycan detection kit, Astarte Biologics, Redmond, Wash.
Sircol soluble collagen assay, Biocolor, Carrickfergus, Antrim, Northern Ireland.
MARS data analysis software, BMG Labtech, Cary, NC.
Inverted fluorescent microscope model, Fisher Scientific, Pittsburg, Pa.
Aquilion 64-slice CT scanner, Toshiba American Medical Systems Corp, Tustin, Calif.
858 MiniBionix, MTS Systems, Eden Prairie, Minn.
References
1. Jung HJ, Fisher MB, Woo SL. Role of biomechanics in the understanding of normal, injured, and healing ligaments and tendons. Sports Med Arthrosc Rehabil Ther Technol 2009;1: 9.
2. Sharma P, Maffulli N. Tendon injury and tendinopathy: healing and repair. J Bone Joint Surg Am 2005; 87: 187–202.
3. Stewart AA, Barrett JG, Byron CR, et al. Comparison of equine tendon-, muscle-, and bone marrow-derived cells cultured on tendon matrix. Am J Vet Res 2009; 70: 750–757.
4. Liu Y, Ramanath HS, Wang DA. Tendon tissue engineering using scaffold enhancing strategies. Trends Biotechnol 2008; 26: 201–209.
5. Maffulli N, Ewen SW, Waterston SW, et al. Tenocytes from ruptured and tendinopathic Achilles tendons produce greater quantities of type III collagen than tenocytes from normal Achilles tendons. An in vitro model of human tendon healing. Am J Sports Med 2000; 28: 499–505.
6. Hoganson DM, O'Doherty EM, Owens GE, et al. The retention of extracellular matrix proteins and angiogenic and mitogenic cytokines in a decellularized porcine dermis. Biomaterials 2010; 31: 6730–6737.
7. Linder LH, Sukin DL, Burks RT, et al. Biomechanical and histological properties of the canine patellar tendon after removal of its medial third. Am J Sports Med 1994; 22: 136–142.
8. Burks RT, Haut RC, Lancaster RL. Biomechanical and histological observations of the dog patellar tendon after removal of its central one-third. Am J Sports Med 1990; 18: 146–153.
9. Gilbert TW, Sellaro TL, Badylak SF. Review: decellularization of tissues and organs. Biomaterials 2006; 27: 3675–3683.
10. Pridgen BC, Woon CY, Kim M, et al. Flexor tendon tissue engineering: acellularization of human flexor tendons with preservation of biomechanical properties and biocompatibility. Tissue Eng Part C Methods 2011; 17: 819–828.
11. Schulze-Tanzil G, Al-Sadi O, Ertel W, et al. Decellularized tendon extracellular matrix—a valuable approach for tendon reconstruction? Cells 2012; 1: 1010–1028.
12. Youngstrom DW, Barrett JG, Jose RR, et al. Functional characterization of detergent-decellularized equine tendon extracellular matrix for tissue engineering applications. PLoS ONE 2013;8:e64151.
13. Shino K, Kawasaki T, Hirose H, et al. Replacement of the anterior cruciate ligament by an allogeneic tendon graft: an experimental study in the dog. J Bone Joint Surg Br 1984; 66: 672–681.
14. Ozasa Y, Amadio PC, Thorasen AR, et al. The effect of surface modification on gliding ability of decellularized flexor tendon in a canine model in vitro. J Hand Surg Am 2013; 38: 1698–1704.
15. Omae H, Sun YL, An KN, et al. Engineered tendon with decellularized xenotendon slices and bone marrow stromal cells: an in vivo animal study. J Tissue Eng Regen Med 2012; 6: 238–244
16. Ning LJ, Zhang Y, Chen XH, et al. Preparation and characterization of decellularized tendon slices for tendon tissue engineering. J Biomed Mater Res A 2012; 100: 1448–1456.
17. Klement P, Wilson GJ, Yeger H. Inventors; HSC Research Development Corporation, assignee. Process for preparing biological mammalian implants. US patent 4,776,853. October 11, 1988.
18. Cissell DD, Hu JC, Griffiths LG, et al. Antigen removal for the production of biomechanically functional, xenogeneic tissue grafts. J Biomech 2014; 47: 1987–1996.
19. Price R, Stoney J, Brown G. Prospective randomized comparison of endobutton versus cross-pin femoral fixation in hamstring anterior cruciate ligament reconstruction with 2-year follow-up. ANZ J Surg 2010; 80: 162–165.
20. Mirzatolooei F. Comparison of short-term clinical outcomes between transtibial and transportal TransFix femoral fixation in hamstring ACL reconstruction. Acta Orthop Traumatol Turc 2012; 46: 361–366.
21. Hapa O, Barber FA. ACL fixation devices. Sports Med Arthrosc 2009; 17: 217–223.
22. Milano G, Mulas PD, Ziranu F, et al. Comparison between different femoral fixation devices for ACL reconstruction with doubled hamstring tendon graft: a biomechanical analysis. Arthroscopy 2006; 22: 660–668.
23. Biskup JJ, Balogh DG, Haynes KH, et al. Mechanical strength of four allograft fixation techniques for ruptured cranial cruciate ligament repair in dogs. Am J Vet Res 2015; 76: 411–419.
