See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/44900631 Engineering an In Vitro Model of a Functional Ligament from Bone to Bone Article in Tissue Engineering Part A · November 2010 DOI: 10.1089/ten.TEA.2010.0039 · Source: PubMed CITATIONS READS 28 293 3 authors: Jennifer Z Paxton Liam Grover 29 PUBLICATIONS 228 CITATIONS 152 PUBLICATIONS 2,318 CITATIONS The University of Edinburgh SEE PROFILE University of Birmingham SEE PROFILE Keith Baar University of California, Davis 139 PUBLICATIONS 4,779 CITATIONS SEE PROFILE All content following this page was uploaded by Jennifer Z Paxton on 08 December 2016. The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately. TISSUE ENGINEERING: Part A Volume 16, Number 11, 2010 ª Mary Ann Liebert, Inc. DOI: 10.1089/ten.tea.2010.0039 Engineering an In Vitro Model of a Functional Ligament from Bone to Bone Jennifer Z. Paxton, Ph.D.,1,2 Liam M. Grover, Ph.D.,3 and Keith Baar, Ph.D.1,* For musculoskeletal tissues that transmit loads during movement, the interfaces between tissues are essential to minimizing injury. Therefore, the reproduction of functional interfaces within engineered musculoskeletal tissues is critical to the successful transfer of the technology to the clinic. The goal of this work was to rapidly engineer ligament equivalents in vitro that contained both the soft tissue sinew and a hard tissue bone mimetic. This goal was achieved using cast brushite (CaHPO42H2O) anchors to mimic bone and a fibrin gel embedded with fibroblasts to create the sinew. The constructs formed within 7 days. Fourteen days after seeding, the interface between the brushite and sinew could withstand a stress of 9.51  1.7 kPa before failure and the sinew reached a Young’s modulus value of 0.16  0.03 MPa. Treatment with ascorbic acid and proline increased the collagen content of the sinew (from 1.34%  0.2% to 8.34%  0.37%), strength of the interface (29.24  6 kPa), and modulus of the sinew (2.69  0.25 MPa). Adding transforming growth factor-b resulted in a further increase in collagen (11.25%  0.39%), interface strength (42  8 kPa), and sinew modulus (5.46  0.68 MPa). Both scanning electron and Raman microscopy suggested that the interface between the brushite and sinew mimics the in vivo tidemark at the enthesis. This work describes a major step toward the development of tissue-engineered ligaments for the repair of ligament ruptures in humans. Introduction A nterior cruciate ligament (ACL) rupture is one of the most common musculoskeletal injuries in the developed world, with *37 ACL tears per 100,000 people each year.1 This equates to over 110,000 per year in the United States alone. As a result of poor innate regeneration and repair of these tissues, return of full function will only occur with a suitable replacement. The most common methods of surgical reconstruction use either the semitendinosus and gracilis tendons or the middle third of the patellar tendon to replace the failed ACL.2 To accelerate recovery, these grafts are often performed with a small portion of bone so that the repair extends from bone to bone.3 In the absence of these bone plugs, graft strength and recovery can be impaired.4 Even though autografting has a high success rate, serious complications at the site of tissue harvest often occur, including rupture of the donor tendon,5 chronic pain,2 and decreased muscle strength,6,7 all of which compromise normal activity.8 As a result of this donor-site morbidity, novel sources of ligaments for reconstruction are needed. Spalazzi et al.9,10 have developed a triphasic scaffold to mimic the fibrocartilaginous transition seen at the insertion site of tendons/ligaments to bone. However, the attachment potential of these scaffolds with either the hard or soft tissues has yet to be reported. Phillips et al.11 have produced progressively calcified and stiffer scaffolds using a gradient of the transcription factor runt-related transcription factor 2 to control osteoblast differentiation and mineralized matrix deposition. Several groups are also attempting to enhance repair at the tendon/ligament insertion site in vivo. Local application of specific growth factors important in endochondral ossification such as bone morphogenetic protein (BMP)12 and granulocyte colony-stimulating factor13 has increased the ultimate tensile stress of grafts at the insertion. Materials such as calcium phosphate14 and magnesium phosphate bone cements,15 polyglycolic acid sheets,16 and demineralized bone matrix17,18 are also being employed to improve repair at tendon/ligament insertion sites and have shown promising results regarding the formation of a fibrocartilaginous interface in vivo. A ligament by definition is a collagenous soft tissue (sinew) that connects bone to bone. Although the sinew component has been engineered by a number of groups,19–25 there is only one previous attempt to engineer not only the Divisions of 1Molecular Physiology and 2Mechanical Engineering and Mechatronics, University of Dundee, Dundee, United Kingdom. 3 School of Chemical Engineering, College of Physical Science and Engineering, University of Birmingham, Edgbaston, Birmingham, United Kingdom. *Present address: Functional Molecular Biology Lab, Department of Neurobiology, Physiology, and Behavior, University of California, Davis, California. 3515 3516 soft tissue component but also the interface between the soft tissue and a hard tissue analogue.24 What was clear from that report was that the presence of hydroxyapatite (HA) within the polymer clearly enhanced the strength of the interface between the soft and hard tissues. Indeed, HA is well known to form intimate attachments with both hard and soft tissues—a feature that has been exploited to enhance the attachment of metallic prostheses.26,27 Even though HA is the calcium phosphate ceramic most frequently used as a coating and synthetic bone replacement, it is virtually insoluble under physiological conditions, meaning that degradation of the ceramic and replacement by native tissue is slow (*5 vol % per year). Complete mechanical incorporation of tissueengineered ligaments will only occur when the ceramic material is replaced with endogenous bone. Brushite (CaHPO42H2O), by contrast to HA, is almost entirely resorbed both in vitro and in vivo since it is several orders of magnitude more soluble than HA in physiological conditions.28,29 The aim of this study was to determine whether monoliths formed from brushite cement alone could be used to engineer a functional soft-to-hard tissue interface. These interfaces are required for connecting engineered sinews to machines in vitro and for replacement of diseased and damaged ligaments if engineered ligaments are ever to become a clinical option. Materials and Methods Brushite cement manufacture The brushite cement was made by incrementally combining b-tricalcium phosphate [Ca3(PO4)2] with orthophosphoric acid (Sigma-Aldrich). The b-tricalcium phosphate was manufactured by reactive sintering of a powder containing CaHPO4 (MallinckdrodtBaker) and CaCO3 (Merck) with a theoretical calcium to phosphate molar ratio of 1:5. The powder mixture was suspended in absolute ethanol and mixed for 12 h. The suspension was then filtered and the resulting cake heated in an alumina crucible to 14008C for 12 h and 10008C for 6 h before quenching in a desiccator in ambient conditions. The resulting sinter cake was then crushed using a pestle and mortar and was passed through a 125 mm sieve. Fabricating brushite anchors Individual anchors were designed using Solidworks software, and casting frames containing 20 anchors were produced using a Thermojet Solid Object Printer (3D SYSTEMS). The Solid Object Printer produced casting frames of a waxlike hardened thermoplastic material (combination of hydrocarbons, urethanes amides, and esters). The thermoplastic frames were then filled with silicone glue (Dow Corning) and allowed to set for 48 h. Once set, the silicone mold was removed from the thermoplastic frame and used to produce individual anchors. Individual anchors were formed as described previously.30 Briefly, brushite paste was spread into the silicone mold and centrifuged at 3700 rpm for 15 s (Eppendorf) before minutien insect pins (Fine Science Tools) were inserted into each anchor. The anchors were then left to set at room temperature overnight. Individual anchors, trapezoidal in shape with approximate dimensions 3.53.733 mm at their PAXTON ET AL. widest points, were removed from the mold the following day and stored at room temperature until used. Soft tissue formation The soft tissue of the ligament was engineered as described previously with some modification.23,30–32 Briefly, cement anchors were pinned onto Sylgard-coated 35 mm plates, and the plates and cement anchors were sterilized by soaking in 70% ethanol for 20 min. Five hundred microliters of growth media (Dulbecco’s Modified Eagle Medium [DMEM] supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin) containing 10 U/mL thrombin (Calbiochem), 2 mL/mL aminohexanoic acid (200 mM; Sigma-Aldrich), and 2 mL/mL aprotinin (10 mg/mL; Roche) solution was added to each dish and agitated to cover the surface of the plate. Two hundred microliters of fibrinogen (20 mg/mL; Sigma-Aldrich) was added drop wise, and the resulting fibrin gel was left to polymerize at 378C for 1 h. Embryonic chick tendon fibroblasts of passages between 2 and 5 were seeded on top of the gel at a concentration of 100103/mL. Chick tendon fibroblast cells were chosen since previous work has identified the presence of fibropositor cells, important for aligned collagen fiber production, when cultured under tension in this system.23 Constructs were fed every 2–3 days with DMEM supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. Constructs were supplemented with ascorbic acid (AA; 50 mM), proline (P; 50 mM), and/or transforming growth factor (TGF-b; 2.5 ng/mL) on day 7 after plating. Soft–hard tissue attachment The attachment of the cement anchor to the soft tissue was assessed by manually removing one anchor from the plate, every 3 days, and observing whether the soft tissue remained attached. Attachment was scored on a binary scale as attached/not attached as described previously (Fig. 1D).30 Tensile testing Individual constructs were mechanically loaded to determine the strength of the soft–hard tissue interface. Tensile tests were conducted in a custom-built tensile testing machine, adapted from a design described in Larkin et al.33 Briefly, grips were manufactured as the inverse of the cement anchor using rapid prototyping (Spectrum Z510; Z Corporation). The grips were designed so that the specimen was immersed in saline during the test. To test the mechanical interface, the cement anchors were inverted and inserted into the grips. Care was taken to assure that the soft tissue was not in contact with any part of the grips, so the recorded values represent the interface stress and not the soft tissue mechanical properties. One grip was attached to a custombuilt force transducer,34 whereas the other was attached to a stepper motor. Using LabVIEW (National Instruments), the sample was loaded at a constant rate of elongation of 0.4 mm/s and the resulting force was measured. Ultimate tensile stress at the interface was calculated using the surface area available for attachment for each anchor. Collagen content The collagen content of the ligament constructs was determined using a hydroxyproline assay.35 Briefly, ligament TISSUE-ENGINEERED LIGAMENT 3517 FIG. 1. Tissue attachment time of three different anchor materials. Ligaments engineered using (A) polyethylene glycol diacrylate-hydroxyapatite (PEG-HA) hydrogels, (B) silk sutures, and (C) brushite cement anchors 1 week after seeding. (D) Attachment testing of one of the brushite anchors. (E) Quantitative assessment of attachment of the anchors and their tissueengineered ligaments. Brushite cements attach to ligament constructs significantly longer than PEG-HA hydrogels and similar, if not better, than silk sutures. Results are mean  standard error of the mean of PEG-HA (n ¼ 6, suture n ¼ 8, brushite n ¼ 24). *Compared to PEG-HA hydrogel group ( p < 0.05). constructs (n ¼ 6 in each group) were removed from their cement anchors and dried in an oven for 30 min at 1108C. Each sample was then weighed and hydrolyzed in 200 mL of 6 N HCl at 1308C for 3 h. The liquid was removed by allowing the HCl to evaporate for 30 min in a fume hood at 1308C. The resulting pellet was resuspended in 200 mL of hydroxyproline buffer. Samples were further diluted 1:8 in hydroxyproline buffer. One hundred and fifty microliters of chloramine T solution was added to each sample, vortexed, and left at room temperature for 20 min. One hundred and fifty microliters of aldehyde-perchloric acid solution was then added to each tube before the tubes were vortexed and incubated in a preheated water bath at 608C for 15 min. After incubation, tubes were left to cool for 10 min and then samples/standards were read at 550 nm on a mQuant Microplate Spectrophotometer (Bio Tek Instruments Limited). Hydroyproline was converted to collagen using a factor of 13.8% as reported previously.36 Scanning electron microscopy Ligament constructs were formed for 1 week before being supplemented with AA and proline for a further 1 week in culture. At the 2-week time point, samples were fixed in 2.5% glutaraldehyde in PIPES buffer (pH 7.2) for 24 h. Samples were then prepared for scanning electron microscopy (SEM), mounted on aluminum stubs using carbon adhesive tabs, and coated with *15 nm Au/Pd using a Cressington 208HR sputter coater. Samples were examined using a Philips XL30 ESEM operating at an accelerating voltage of 15 kV. Raman microscopy For imaging the interface, after 14 days of culture, the constructs were immersed in TissueTekÒ (VWR) and then frozen in liquid-nitrogen-cooled isopentane. Eight-micrometerthick longitudinal sections were cut using a cryostat (Leica). The sections were then mounted onto a microscope slide 3518 PAXTON ET AL. and examined using a confocal Raman microscope (Alpha 300R; WITEC) equipped with a 785 nm laser. A selected area of 175 mm2 at the interface was mapped at a resolution of 175 points per line each requiring 19.64 s at an appropriate wave number for the phosphate molecule. The fibrin was observed by exciting the sample at <300 cm1, which causes autofluorescence. Masson’s trichrome Ligament constructs were soaked overnight in a 30% sucrose solution and then frozen in Tissue Tek OCT (OCT) compound in liquid-nitrogen-cooled isopentane. Eightmicrometer-thick longitudinal sections were cut on a cryostat (Leica CM3050S Cryostat), and the sections were stained using an AccustainÒ Trichrome (Masson) Staining Kit (Sigma-Aldrich) as per the manufacturer’s instructions. DNA content Ligament constructs (n ¼ 10 in each group) were detached from their anchors, snap frozen in liquid nitrogen, powdered on dry ice, and digested overnight in proteinase K lysis buffer (50 mM tris, 100 mM ethylenediaminetetraacetic acid, 0.5% sodium dodecyl sulfate, and 400 mg/mL proteinase K) at 558C. DNA was isolated by phenol:cholorofrom:isoamyl precipitation and quantified on a Nanodrop Spectrophotometer (Thermo Scientific). Glycosaminoglycan content Dried tissues (n ¼ 4 in each group) were weighed and digested in papain buffer (2 U/mL papain, 5 mM cysteine, and 5 mM ethylenediaminetetraacetic acid) for 24 h at 608C. After papain digestion, 100 mL of each sample was added in triplicate to individual wells of a 24-well plate, and dimethylmethylene dye (45 mM 1,2 dimethylmethylene blue, 40 mM glycine, 40 mM NaCl, and 10 mM HCl) was then added to each well, and the plate was read immediately at 525 nm on a mQuant Microplate Spectrophotometer (Bio Tek Instruments Limited). Glycosaminoglycan (GAG) content of each sample was determined using a chondroitin sulfate standard curve. Statistics Data are presented as means  standard error of the mean. Differences in mean values were compared within groups and significant differences were determined by analysis of variance with post hoc Tukey-Kramer HSD test using BrightStat (www.brightstat.com). The significance level was set at p < 0.05. Results Anchor attachment time The previous gold standard for the attachment of engineered ligaments/tendons in vitro is the woven silk suture; therefore, we sought to determine how the brushite cement attachment performs compared to this standard. The sinew remained attached to woven sutures for a period of 21  4 days (Fig. 1). A polyethylene glycol diacrylate-hydroxyapatite (PEG-HA) composite remained attached for a much shorter period as previously described (4  1 days of attachment; Fig. 1).24 Optimized brushite-based anchors remained attached FIG. 2. Functional analysis of the cement anchor–sinew interface. (A) A custom-built tensile testing machine was used for conducting tensile tests on the ligament constructs. (a) Force transducer, (b) shaped grip, (c) moving grip, (d) stepper motor, and (e) microcontroller. (B) Ligament construct placed upside down into grips before testing. (C) Ligament construct undergoing tensile test. Tissue can be seen detaching from the right-hand-side cement anchor. (D) Mean ultimate tensile stress values after 1 or 4 weeks in culture. (E) Seven days of treatment with ascorbic acid (AA), proline (P), and/or transforming growth factor (TGF)-b. *Significantly greater then corresponding control ( p < 0.05). for a period of 26  2 days (Fig. 1) before failure of the interface, exceeding the attachment time of both the PEG-HA and the woven silk suture. Mechanics of the interface The plastic nature of the cement paste enabled the manufacture of a trapezoidal-shaped monolith that could easily TISSUE-ENGINEERED LIGAMENT 3519 Table 1. Collagen Content Over Time With and Without Supplementation With Ascorbic Acid, Proline, and Transforming Growth Factor-b Treatment Construct age Collagen content (%) GAG content (mg/mg) Untreated AAþP (1 week) AAþPþTGF-b (1 week) 2 weeks 4 weeks 2 weeks 2 weeks 1.45  0.32 4.78  0.32 2.92  0.18 8.51  0.59 8.34  0.90 10.52  0.82 11.25  0.95 ND Each group is significantly different from all others ( p < 0.05). AA, ascorbic acid; GAG, glycosaminoglycan; ND, not determined; TGF, transforming growth factor. be gripped for tensile testing. When inverted, the test measured the interface strength, whereas the sinew mechanics were tested with the anchor upright. Constructs that received no media supplements were able to withstand interface stresses of 10.7  0.8 kPa (Fig. 2) before failure of the interface. To determine whether the collagen content of the soft tissue affected the interface between the ligament and the calcium phosphate, constructs were maintained for 4 weeks to allow the accumulation of more collagen (Table 1), or treated for 7 days with AA and proline, or AA, praline, and TGF-b. Increasing the time in culture (1 week, 7.8  0.75 kPa; 4 week, 17.6  0.88 kPa) or supplementing the media with AA, praline, and TGF-b (CTL, 9.5  1.7 kPa; AAþP, 29.2  6.0 kPa; AAþPþTGF-b, 41.8  8.4 kPa) significantly improved the tensile strength of the interface. microscopy (Fig. 5). Light micrographs of the tissue sections demonstrated that the interface between the brushite cement was not obviously demarcated. Instead, the edge of the brushite in contact with the sinew was uneven with a tidemark similar to the enthesis in vivo (Fig. 5A). The interface was then imaged using confocal Raman microscopy (Fig. 5B) and mapped according to the intensity of a Raman peak indicative of the P–O vibration in brushite and other calcium phosphate salts (Fig. 5C). Interestingly, the image map demonstrated the presence of P–O-rich areas within the fibrin matrix, which was located using the tendency of Mechanics of the sinew Adding AA and proline to the culture media also led to an increase in Young’s modulus for the sinew (Fig. 3). A 22-fold increase was observed with 50 mM AA and proline compared with the untreated controls (CTL, 0.16  0.03 MPa; 50 mM AAþP, 2.97  0.25 MPa). The addition of TGF-b in combination with AA and proline lead to a further twofold increase in modulus compared to AA and proline alone (5.46  0.68 MPa) and a 34-fold increase in when compared with the untreated control group (Fig. 3). Ultrastructure of the interface To examine the microstructure of the brushite–sinew interface, scanning electron micrographs were obtained from anchors before integration with the sinew and the constructs following 7 days of formation and a further 7 days of treatment with AA and proline (Fig. 4). The SEM images show that the monolith of brushite begins as an ordered crystalline structure that becomes completely surrounded by the fibrin with regions where the sinew becomes invaginated within the anchors (Fig. 4B). Further, like ligaments in vivo,37 the collagen in the midsection of the sinew (Fig. 4E) has a different orientation to that near the bone (Fig. 4F). In the midsection, the collagen is aligned along the line of tension (from top left to bottom right), whereas near the brushite anchor the collagen becomes somewhat disorganized (Fig. 4E, F). Since the SEM images of the interface showed that the sinew infiltrated the brushite anchor and there were areas of calcium phosphate that appeared in the sinew, the interface was further characterized using light and confocal Raman FIG. 3. Functional analysis of the sinews. (A) Representative stress–strain curves for control sinews and sinews treated for 1 week with AA and proline or AA, proline, and TGF-b. (B) Average Young’s modulus of the sinews. *Significantly different from control; {significantly different from AAþP ( p < 0.05). 3520 PAXTON ET AL. FIG. 4. Scanning electron microscope images of the sinew constructs. (A) Cement anchor showing the brushite crystal structure. (B) Close-up of the cement anchor–sinew interface. (C) Side view of cement anchor showing the sinew wrapping around the bottom surface. (D) Further magnified image of the cement anchor (left) brushite (right) sinew. (E) Collagen fibers within the mid portion of the sinew (the line of force is between the top left and bottom right corners). (F) Collagen fibers in within the tissue portion of the ligament near the cement anchor. biological molecules to autofluoresce at lower than 300 cm1 (Fig. 5D). The presence of calcium phosphate crystals within the fibrin gel and the unevenly demarcated edge between cement and fibrin matrix may indicate active reprecipitation of calcium phosphate near the interface resulting in the formation of a graded interface. Ultrastructure of the sinew Staining of ligament constructs with Masson’s trichrome demonstrated that the fibrin matrix (stained red) was de- graded over time and replaced by endogenously produced collagen (stained blue) only in the presence of AAþP or AAþPþTGF-b (Fig. 6, panels 1 and 2 vs. 3–6). Collagen content continues to increase and fibrin continues to decrease over time. Four weeks of supplementation resulted in greater collagen staining than observed after 1 week (Fig. 6, panels 3–6 vs. 9–12). Treatment with AAþP and AAþPþTGF-b also appeared to increase the number of cells in the sinew (stained black). The DNA content of the AAþP-treated grafts increased 2.4-fold, whereas the AAþPþTGF-b constructs had no significant increase in cell number compared to TISSUE-ENGINEERED LIGAMENT 3521 FIG. 5. A sinew–brushite interface after 14 days of culture. (A) A light micrograph showing the sinew interface with the brushite cement anchor (black). Notice the irregularity of the interface between the cement material and the cell-seeded fibrin (gray), which suggests reprecipitation of mineral within the fibrin matrix (arrows). (B) Confocal Raman microscopic image of the interface. (C) Phosphate peaks indicative of brushite away from the periphery of the cement. (D) The organic regions of the interface were mapped using the autofluorescence of the biological samples when excited at wave numbers of <300 cm1. Color images available online at www.liebertonline.com/ten. controls (CTL, 27.0  0.82 mg; AAþP, 64.1  0.80 mg; AAþPþ TGF-b, 31.6  0.71 mg). As with collagen, the GAG content of the sinew increases over time, and with AAþP treatment, GAG content increases from 4.78  0.32 mg/mg at 2 weeks to 8.51  0.59 mg/mg at 4 weeks. Supplementation with AAþP increased GAG content to 10.52  0.82 mg/mg (Table 1). Ligament longevity During the 4-week experiment, it was noted that even though testing the constructs every third day resulted in failure within 26 days, when the constructs were left undisturbed in culture they could be cultured for longer periods. Two separate experiments to determine how long the interface could remain intact when left undisturbed demonstrated that the constructs routinely remained intact for over 12 weeks. At the end of the 12 weeks, the sinew had a white appearance, contained high levels of collagen, and grossly resembled a ligament in vivo (Fig. 7). Discussion Using a bioresorbable calcium phosphate, brushite, and a fibrin cast soft tissue construct, we have produced ligaments that have a functional hard to soft tissue interface. Although this interface is still not as strong as the soft tissue itself (i.e., the ligament fails at the interface and not in the midsubstance of the construct), the 41.8  8.4 kPa ultimate tensile strength is in the same order of magnitude as embryonic ligaments. We had previously observed that increasing the calcium phosphate content in a composite material resulted in increasing mechanical stability of a musculoskeletal interface.24 This suggested that phase-pure calcium phosphate monoliths would be the optimal material for engineering the interface between hard and soft musculoskeletal tissues. Using pure calcium phosphate, in the form of brushite cement, the ligamentous interface can be maintained for over 25 days when tested every third day. When left untested, constructs routinely remain attached for >12 weeks. We have shown that 3522 PAXTON ET AL. FIG. 6. Histology of the sinew at 2 and 5 weeks of culture. Representative images showing the collagen content of the sinews using Masson’s trichrome. Fibrin stains red, collagen stains blue, and black indicates cell nuclei. Fibrin can be seen to reduce over time (red), whereas collagen (blue) increases over time and noticeably with AAþP and AAþPþTGF-b supplementation. Panels 1, 3, and 5 show 2 week old constructs at 10magnification. Panels 2, 4, and 6 show the same constructs at 20magnification. Panels 7, 9, and 11 show 5 week old constructs at 10magnification. Panels 8, 10, and 12 show the same constructs at 20  magnification. the attachment and the strength of the interface are determined by the surface area of the anchor and stress/strain concentrations.30 Therefore, creating high surface area anchors with minimal regions of stress/strain concentrations can further increase the strength of the interface. The fibrin-based sinews have been engineered from embryonic chick,23 primary rat and human adult tendon/ ligament fibroblasts,30,31 mesenchymal stem cells (Baar and Vidal, unpublished), and primary muscle cells (Baar, unpublished). Therefore, autologous sinews can be engineered either from stem cells isolated from the patient’s marrow or from primary cells isolated from a biopsy and expanded in culture. Within 5 weeks in vitro, the fibrin used to cast the sinew was largely digested and replaced by collagen produced by the cells (Fig. 6). From work using cells isolated from human ACLs, we have seen that large amounts of type I collagen are found throughout the sinew, whereas type III collagen is found in lower amounts within the core of the graft.31 Treating the grafts with AA, proline, and TGF-b increased the collagen and GAG content of the grafts and this resulted in improved mechanical strength and stiffness. The collagen content of the sinew in the presence of AAþPþTGFb increased to 11.25%  0.39%. This represents 20% of the collagen content of the adult ACL.38 Similarly, the GAG content of the sinew is within the physiological range for ligaments. The GAG content that was achieved with 1 week of AAþP treatment (10.52  0.82 mg/mg) compares favorably with the GAG content of the adult ACL (7.4 mg/mg). The rapid acquisition of in vivo levels of GAGs in the sinew indicates that these components mature faster than the collagen matrix. The high GAG content might also play an important role in the development of the sinew and its lower collagen content since GAGs, especially small leucine-rich proteoglycans, are important for collagen fibrillogenesis and can prevent lateral fusion of collagen fibrils.39,40 Overall, these data suggest that with AAþPþTGF-b treatment the sinew rapidly develops both structurally and functionally. In vivo, the osteoligamentous junction or enthesis is designed to transmit force from ligaments to bone37,41,42 with minimal stress/strain concentrations. As tendons and ligaments develop from fibrous outgrowths of the cartilaginous primordial bone before its ossification,43 the transition from tendon/ligament to bone is a zonal arrangement, comprising four separate regions: (1) tendon/ligament, (2) fibrocartilage, (3) mineralized fibrocartilage, and (4) mineralized bone.42,43 The outer limit of calcification is demarcated by a tidemark, signifying the transition between the calcified and noncalcified regions of the enthesis.37,41 The zone of calcified fibrocartilage interdigitates with the bone, greatly increasing the surface area for attachment of tendon/ligament to bone, resulting in a decrease in stress/strain concentrations.37,41,43 This complex structure has been elegantly produced in vitro using a triphasic scaffold.9,10 We report here that the tidemark and interdigitated calcium phosphate–sinew interface can be reproduced in vitro using brushite monoliths and a fibrin-based sinew. As the ligament enters the enthesis, the collagen orientation changes from a parallel alignment in the midsection to a more random alignment near the bone.37 This property is reproduced in the grafts reported here. In the midsection of the sinew the collagen is aligned in parallel along the line of force (Fig. 4). Similar to the enthesis in vivo, near the anchor the orientation of the collagen becomes more TISSUE-ENGINEERED LIGAMENT 3523 ture plate, these constructs are indeed scalable to the anatomy of a patient. In fact, we have already produced ligament constructs 3.5 cm long, approximately the length of the average human ACL (Fig. 7). Conclusions Using brushite and a fibrin cast sinew, we have produced ligaments that have a functional hard to soft tissue interface. The interface (41.8  8.4 kPa) is not as strong as the soft tissue (5.46  0.68 MPa), but the ultimate tensile strength is in the same order of magnitude as embryonic ligaments. Further, increasing the collagen content and decreasing regions of stress/strain concentrations within the brushite anchors should continue to improve the interface and move tissue engineering closer to addressing the clinical need for ACL reconstruction. Acknowledgments FIG. 7. Gross morphology of ligament constructs after 12 weeks in culture. (A) A 2 cm ligament held upright by the cement anchor with metal forceps; (B) the construct from above; and (C) the construct from the side. Note that the pins have been removed, the cement–ligament interface is intact, and the construct is self-supporting. (D) A construct with a gauge length of 3.5 cm, equivalent to the average gauge length of a human anterior cruciate ligament. Color images available online at www.liebertonline.com/ten. The authors acknowledge the assistance of Paul Maher in creating the brushite anchors. The work was supported by grants from the Engineering and Physical Sciences Research Council (EPSRC–EP/E008925/1) and the Biotechnology and Biological Sciences Research Council (BBSRC–BB/F002084/ 1). The Raman microscope used in this research was obtained through Birmingham Science City: Innovative Uses for Advanced Materials in the Modern World (West Midlands Centre for Advanced Materials Project 2), with support from Advantage West Midlands (AWM) and part funded by the European Regional Development Fund (ERDF). Disclosure Statement No competing financial interests exist. random. Since this property in vivo is thought to give the ligament graded mechanics,44 this might be an important aspect of the functional interface. In vivo, the enthesis is so well developed that tendon and ligament failure tends to occur at the subchondral bone or the midsubstance of the sinew and not at the interface.45 The engineered ligaments described here fail at the interface when tested while inverted. The failure at the interface could be the result of the interface only containing two of the four zones of the enthesis (ligament and bone). The absence of fibrocartilage may play a key role in the strength of the interface. Therefore, inducing fibrocartilage formation could be important in further development of this model. Fibrocartilaginous regions have been ectopically produced in vivo using the TGF-b family member BMP-2. Injection of BMP-2 into a rabbit flexor tendon results in the progressive development of a fibrocartilaginous enthesis.12 Since brushite can absorb growth factors,46 adding BMP-2 to the brushite anchors may locally convert cells to a fibrocartilaginous fate. These cells may continue to remodel the brushite at the interface and produce a four-zone transition. The experiments described here have used ligament constructs that are 2 cm long. For these ligaments to be clinically viable in humans, they need to be scalable. Since the size and shape of the anchors can be altered using different molds and the length of the graft can be adjusted by increasing the distance between the anchors and the size of the tissue cul- References 1. Gianotti, S.M., Marshall, S.W., Hume, P.A., and Bunt, L. Incidence of anterior cruciate ligament injury and other knee ligament injuries: a national population-based study. J Sci Med Sport 12, 622, 2009. 2. Spindler, K.P., Kuhn, J.E., Freedman, K.B., Matthews, C.E., Dittus, R.S., and Harrell, F.E., Jr. Anterior cruciate ligament reconstruction autograft choice: bone-tendon-bone versus hamstring: does it really matter? A systematic review. Am J Sports Med 32, 1986, 2004. 3. Brand, J., Jr., Hamilton, D., Selby, J., Pienkowski, D., Caborn, D.N., and Johnson, D.L. Biomechanical comparison of quadriceps tendon fixation with patellar tendon bone plug interference fixation in cruciate ligament reconstruction. Arthroscopy 16, 805, 2000. 4. Nebelung, W., Becker, R., Urbach, D., Ropke, M., and Roessner, A. Histological findings of tendon-bone healing following anterior cruciate ligament reconstruction with hamstring grafts. Arch Orthop Trauma Surg 123, 158, 2003. 5. Mastrokalos, D.S., Springer, J., Siebold, R., and Paessler, H.H. Donor site morbidity and return to the preinjury activity level after anterior cruciate ligament reconstruction using ipsilateral and contralateral patellar tendon autograft: a retrospective, nonrandomized study. Am J Sports Med 33, 85, 2005. 6. Aune, A.K., Holm, I., Risberg, M.A., Jensen, H.K., and Steen, H. Four-strand hamstring tendon autograft compared with 3524 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. patellar tendon-bone autograft for anterior cruciate ligament reconstruction. A randomized study with two-year followup. Am J Sports Med 29, 722, 2001. Beynnon, B.D., Johnson, R.J., Fleming, B.C., Kannus, P., Kaplan, M., Samani, J., et al. Anterior cruciate ligament replacement: comparison of bone-patellar tendon-bone grafts with two-strand hamstring grafts. A prospective, randomized study. J Bone Joint Surg Am 84-A, 1503, 2002. O’Neill, D.B. Arthroscopically assisted reconstruction of the anterior cruciate ligament. A prospective randomized analysis of three techniques. J Bone Joint Surg Am 78, 803, 1996. Spalazzi, J.P., Dagher, E., Doty, S.B., Guo, X.E., Rodeo, S.A., and Lu, H.H. In vivo evaluation of a multiphased scaffold designed for orthopaedic interface tissue engineering and soft tissue-to-bone integration. J Biomed Mater Res A 86, 1, 2008. Spalazzi, J.P., Doty, S.B., Moffat, K.L., Levine, W.N., and Lu, H.H. Development of controlled matrix heterogeneity on a triphasic scaffold for orthopedic interface tissue engineering. Tissue Eng 12, 3497, 2006. Phillips, J.E., Burns, K.L., Le Doux, J.M., Guldberg, R.E., and Garcia, A.J. Engineering graded tissue interfaces. Proc Natl Acad Sci U S A 105, 12170, 2008. Hashimoto, Y., Yoshida, G., Toyoda, H., and Takaoka, K. Generation of tendon-to-bone interface ‘‘enthesis’’ with use of recombinant BMP-2 in a rabbit model. J Orthop Res 25, 1415, 2007. Sasaki, K., Kuroda, R., Ishida, K., Kubo, S., Matsumoto, T., Mifune, Y., et al. Enhancement of tendon-bone osteointegration of anterior cruciate ligament graft using granulocyte colony-stimulating factor. Am J Sports Med 36, 1519, 2008. Tien, Y.C., Chih, T.T., Lin, J.H., Ju, C.P., and Lin, S.D. Augmentation of tendon-bone healing by the use of calciumphosphate cement. J Bone Joint Surg Br 86, 1072, 2004. Gulotta, L.V., Kovacevic, D., Ying, L., Ehteshami, J.R., Montgomery, S., and Rodeo, S.A. Augmentation of tendonto-bone healing with a magnesium-based bone adhesive. Am J Sports Med 36, 1290, 2008. Yokoya, S., Mochizuki, Y., Nagata, Y., Deie, M., and Ochi, M. Tendon-bone insertion repair and regeneration using polyglycolic acid sheet in the rabbit rotator cuff injury model. Am J Sports Med 36, 1298, 2008. Sundar, S., Pendegrass, C.J., Oddy, M.J., and Blunn, G.W. Tendon re-attachment to metal prostheses in an in vivo animal model using demineralised bone matrix. J Bone Joint Surg Br 91, 1257, 2009. Sundar, S., Pendegrass, C.J., and Blunn, G.W. Tendon bone healing can be enhanced by demineralized bone matrix: a functional and histological study. J Biomed Mater Res B Appl Biomater 88, 115, 2009. Banes, A.J., Gilbert, J., Taylor, D., and Monbureau, O. A new vacuum-operated stress-providing instrument that applies static or variable duration cyclic tension or compression to cells in vitro. J Cell Sci 75, 35, 1985. Calve, S., Dennis, R.G., Kosnik, P.E., 2nd, Baar, K., Grosh, K., and Arruda, E.M. Engineering of functional tendon. Tissue Eng 10, 755, 2004. Garvin, J., Qi, J., Maloney, M., and Banes, A.J. Novel system for engineering bioartificial tendons and application of mechanical load. Tissue Eng 9, 967, 2003. Hairfield-Stein, M., England, C., Paek, H.J., Gilbraith, K.B., Dennis, R., Boland, E., et al. Development of self-assembled, tissue-engineered ligament from bone marrow stromal cells. Tissue Eng 13, 703, 2007. PAXTON ET AL. 23. Kapacee, Z., Richardson, S.H., Lu, Y., Starborg, T., Holmes, D.F., Baar, K., et al. Tension is required for fibripositor formation. Matrix Biol 27, 371, 2008. 24. Paxton, J.Z., Donnelly, K., Keatch, R.P., and Baar, K. Engineering the bone-ligament interface using polyethylene glycol diacrylate incorporated with hydroxyapatite. Tissue Eng Part A 15, 1201, 2009. 25. Vunjak-Novakovic, G., Altman, G., Horan, R., and Kaplan, D.L. Tissue engineering of ligaments. Annu Rev Biomed Eng 6, 131, 2004. 26. Gross, K.A., Ray, N., and Rokkum, M. The contribution of coating microstructure to degradation and particle release in hydroxyapatite coated prostheses. J Biomed Mater Res 63, 106, 2002. 27. Wie, H., Hero, H., Solheim, T., Kleven, E., Rorvik, A.M., and Haanaes, H.R. Bonding capacity in bone of HIP-processed HA-coated titanium: mechanical and histological investigations. J Biomed Mater Res 29, 1443, 1995. 28. Xia, Z., Grover, L.M., Huang, Y., Adamopoulos, I.E., Gbureck, U., Triffitt, J.T., et al. In vitro biodegradation of three brushite calcium phosphate cements by a macrophage cellline. Biomaterials 27, 4557, 2006. 29. Grover, L.M., Gbureck, U., Wright, A.J., Tremayne, M., and Barralet, J.E. Biologically mediated resorption of brushite cement in vitro. Biomaterials 27, 2178, 2006. 30. Paxton, J.Z., Donnelly, K., Keatch, R.P., Baar, K., and Grover, L.M. Factors affecting the longevity and strength in an in vitro model of the bone-ligament interface. Ann Biomed Eng 38, 2155, 2010. 31. Bayer, M.L., Yeung, C.Y., Kadler, K.E., Qvortrup, K., Baar, K., Svensson, R.B., et al. The initiation of embryoniclike collagen fibrillogenesis by adult human tendon fibroblasts when cultured under tension. Biomaterials 31, 4889, 2010. 32. Paxton, J.Z., Donnelly, K., Keatch, R.P., and Baar, K. Engineering the bone-ligament interface using polyethylene glycol diacrylate incorporated with hydroxyapatite. Tissue Eng Part A 15, 1201, 2009. 33. Larkin, L.M., Calve, S., Kostrominova, T.Y., and Arruda, E.M. Structure and functional evaluation of tendon-skeletal muscle constructs engineered in vitro. Tissue Eng 12, 3149, 2006. 34. Dennis, R.G., and Kosnik, P.E. Mesenchymal cell culture: instrumentation and methods for evaluating engineered muscle. In: Atala, A., and Lanza, R., eds. Methods in Tissue Engineering. San Diego: Harcourt, Academic Press, 2002, pp. 307–16. 35. Woessner, J.F., Jr. The determination of hydroxyproline in tissue and protein samples containing small proportions of this imino acid. Arch Biochem Biophys 93, 440, 1961. 36. Arruda, E.M., Calve, S., Dennis, R.G., Mundy, K., and Baar, K. Regional variation of tibialis anterior tendon mechanics is lost following denervation. J Appl Physiol 101, 1113, 2006. 37. Benjamin, M., and Ralphs, J.R. Entheses—the bony attachments of tendons and ligaments. Ital J Anat Embryol 106, 151, 2001. 38. Rumian, A.P., Wallace, A.L., and Birch, H.L. Tendons and ligaments are anatomically distinct but overlap in molecular and morphological features—a comparative study in an ovine model. J Orthop Res 25, 458, 2007. 39. Raspanti, M., Viola, M., Sonaggere, M., Tira, M.E., and Tenni, R. Collagen fibril structure is affected by collagen concentration and decorin. Biomacromolecules 8, 2087, 2007. TISSUE-ENGINEERED LIGAMENT 40. Danielson, K.G., Baribault, H., Holmes, D.F., Graham, H., Kadler, K.E., and Iozzo, R.V. Targeted disruption of decorin leads to abnormal collagen fibril morphology and skin fragility. J Cell Biol 136, 729, 1997. 41. Benjamin, M., Toumi, H., Ralphs, J.R., Bydder, G., Best, T.M., and Milz, S. Where tendons and ligaments meet bone: attachment sites (‘‘entheses’’) in relation to exercise and/or mechanical load. J Anat 208, 471, 2006. 42. Benjamin, M., Kumai, T., Milz, S., Boszczyk, B.M., Boszczyk, A.A., and Ralphs, J.R. The skeletal attachment of tendons— tendon ‘‘enthesis.’’ Comp Biochem Physiol A Mol Integr Physiol 133, 931, 2002. 43. Doschak, M.R., and Zernicke, R.F. Structure, function and adaptation of bone-tendon and bone-ligament complexes. J Musculoskelet Neuronal Interact 5, 35, 2005. 44. Woo, S.L., Gomez, M.A., Seguchi, Y., Endo, C.M., and Akeson, W.H. Measurement of mechanical properties of ligament substance from a bone-ligament-bone preparation. J Orthop Res 1, 22, 1983. 3525 45. Benjamin, M., and Ralphs, J.R. Fibrocartilage in tendons and ligaments—an adaptation to compressive load. J Anat 193 (Pt 4), 481, 1998. 46. Le Nihouannen, D., Hacking, S.A., Gbureck, U., Komarova, S.V., and Barralet, J.E. The use of RANKL-coated brushite cement to stimulate bone remodelling. Biomaterials 29, 3253, 2008. Address correspondence to: Keith Baar, Ph.D. Functional Molecular Biology Lab Department of Neurobiology, Physiology, and Behavior University of California Davis, CA 95616 E-mail: fmblab@googlemail.com Received: January 21, 2010 Accepted: June 29, 2010 Online Publication Date: August 18, 2010