Tissue-engineered intervertebral discs produce new matrix, maintain disc height, and restore biomechanical function to the rodent spine

Abstract

Lower back and neck pain are leading physical conditions for which patients see their doctors in the United States. The organ commonly implicated in this condition is the intervertebral disc (IVD), which frequently herniates, ruptures, or tears, often causing pain and limiting spinal mobility. To date, approaches for replacement of diseased IVD have been confined to purely mechanical devices designed to either eliminate or enable flexibility of the diseased motion segment. Here we present the evaluation of a living, tissue-engineered IVD composed of a gelatinous nucleus pulposus surrounded by an aligned collagenous annulus fibrosus in the caudal spine of athymic rats for up to 6 mo. When implanted into the rat caudal spine, tissue-engineered IVD maintained disc space height, produced de novo extracellular matrix, and integrated into the spine, yielding an intact motion segment with dynamic mechanical properties similar to that of native IVD. These studies demonstrate the feasibility of engineering a functional spinal motion segment and represent a critical step in developing biological therapies for degenerative disc disease.

Among the most common physical conditions for which patients see their doctors are back and neck pain, which carry an estimated annual cost to society up to $100 billion (1). Current conservative and operative treatment options are mostly palliative in nature and fail to restore function to the spine. The most common target for treatment of back and neck pain is the intervertebral disc (IVD) (2–5). IVD degeneration is characterized by loss of proteoglycan, loss of disc height, annulus fibrosus (AF) damage and tears, spondylolisthesis, spinal stenosis, herniated discs, neoinnervation, hypermobility, and inflammation (6–8). Conservative treatments including medication and physiotherapy are the first line of defense in treating these disorders. Despite these treatments, it is estimated that between 1.5 and 4 million patients in the United States await surgical intervention (9).

Such surgical interventions may involve the removal of herniated tissue or the entire IVD and replacing it with a mechanical device designed to either fuse the adjacent vertebrae or to preserve some motion. Regardless of approach, motion segment mobility is altered, often precipitating degeneration in adjacent motion segments (10). Nonbiological total disc replacement implants were developed to avoid this loss of motion at the operated level, and as a result, reduce the incidence of adjacent segment disease. The efficacy of such implants is a matter of much debate (11–13); however, it is clear that nonbiological total disc replacement implants suffer from failure modes commonly associated with traditional metal/polyethylene arthroplasty, such as mechanical failure, dislodgement, polyethylene wear, and associated osteolysis and implant loosening. More recently, increasing attention has been turned toward creating tissue engineering strategies to restore function to the diseased or injured IVD.

The intervertebral disc is composed of two distinct regions, the AF and the nucleus pulposus (NP). The NP is a gelatinous tissue, rich in proteoglycan and type II collagen that is surrounded by the AF, a highly organized fibrocartilage predominantly made of type I collagen and proteoglycans. A number of tissue engineering strategies have focused on creating either the AF or NP separately (14), but much interest has recently been focused on creating a composite tissue-engineered total disc replacement (TE-TDR) implant that contains both AF and NP tissue (15–19). The successful replication of spinal motion and function in human IVD allograft transplantation (20) suggests that a properly engineered IVD implant could be an important tool for clinical spine care.

Despite the promise of tissue engineering approaches for design of IVD implants (15–19) and preliminary reports of their transplantation into the spine (21), to date none have demonstrated the ability to restore the structure and function of a motion segment in vivo. Here we show that our previously documented approach to producing TE-TDR implants with circumferentially aligned collagen fibrils in the AF (15), combined with image-based design techniques to reproduce precise anatomy (22–26) yielded implants that integrated with the rat caudal spine, reproduced appropriate tissue structure, and generated a mechanically functional motion segment in the rat caudal spine.

Results

Engineered IVDs Reproduce Native Shape and Composite Structure over 6_ Mo of Implantation.

To study the function of tissue-engineered IVD implants, we chose to replace healthy IVD in the rat caudal spine. The rat caudal spine was chosen for ease of surgical access, repeatability of the surgery, and the levels of stress and strain imposed on caudal discs (27–29). We constructed composite TE-TDR implants seeded with ovine AF and NP cells and inserted them into the caudal 3/4 disc space of athymic rats. Samples were harvested at 6 wk and 6 mo along with separate groups containing no implant (discectomy) and native discs reimplanted into the caudal spine, analogous to an allograft (30).

Microcomputed tomography (μCT) and MRI images of the caudal 3/4 disc space were obtained and used to produce a TE-TDR implant with native dimensions (Fig. 1A). μCT images allowed the outer boundaries of the IVD to be determined from the bony ends of the neighboring vertebrae and the NP dimensions from the MRI data. Using these clinically relevant imaging techniques to design the engineered disc and combining them with the collagen contracted-AF/alginate-NP engineered disc construction we reported previously (15), we constructed anatomically shaped discs (Fig. 1E) that were surgically implanted into the native caudal 3/4 disc space (Fig. 1C).

Fig. 1.

Anatomical TE-TDR, designed from MRI and CT, survives in disc space for 6 mo. (A) CT and MRI design procedure for obtaining TE-TDR dimensions. (B) Fabrication of TE-TDR. (i) NP dimensions used to design injection molds via computer-aided design. (ii) Injection mold 3-D printed out of acrylonitrile butadiene styrene plastic. (iii) Cell-seeded alginate was injected into mold, removed, (iv) placed in center of 24 well plate, and cell-seeded collagen was poured around alginate NP. (v) After 2 wk of culture, cell-seeded collagen contracts around the NP to form composite TE-TDR. (C) Intraoperative images showing exposed caudal 3/4 disc space and implanted TE-TDR. (D) T2-weighted MRI of implanted disc space (marked by yellow arrows) and adjacent native levels immediately postoperative, at 1 mo, and 6 mo after implantation. (E) History of TE-TDR in native disc space. Intraoperative photo showing explanted native IVD next to the TE-TDR (day 0) that was implanted in its place and TE-TDR after being implanted into native disc space for 6 mo. (F) Size of engineered IVD compared to native IVD. Measurements were taken along the lateral and ventral-dorsal planes of the engineered and native IVD. Engineered IVD measurements were taken at day 0 prior to implantation (n = 12) and compared to explanted native discs (n = 12). Engineered IVD measurements were also taken after 6 mo of implantation (n = 12).

The imaging and fabrication technique produced discs that mimicked the native morphology (Fig. 1 E and F). These anatomically shaped engineered discs were implanted into the native caudal disc space and the MRI demonstrated hydrated tissue in the disc space post operatively at 6 wk and at 6 mo that were similar to the native adjacent levels (Fig. 1D). Upon explantation at 6 mo, the tissue had a distinct composite cartilaginous appearance and IVD shape upon explantation (Fig. 1 E and F).

Engineered IVD Integrated with Neighboring Vertebrae and Produced Extracellular Matrix with Native Levels of Collagen and Proteoglycan in the AF and NP.

It was hypothesized that our composite engineered IVD would produce an ECM in the native caudal disc space that was rich in collagen and proteoglycans, similar to the native IVD, and would integrate with the vertebral bodies. To test this hypothesis, we performed histological staining for collagen and proteoglycans in the disc space, as well as quantitative biochemical analysis on the explanted tissue at 6 mo for collagen, proteoglycan, and DNA (31–33). In the discectomy group, in which the native disc had been removed and nothing implanted in its place, the disc space collapsed with no production of collagen and proteoglycans (Fig. 2 C and D). In contrast, engineered discs contained properly distributed collagen and proteoglycan in the AF and NP by 6 mo (Fig. 2 E, F, and I). Similar to the native IVD, increased collagen was observed in the AF compared to the NP at 6 mo as indicated by both the hydroxyproline content and picrosirius red staining (Fig. 2 E and I). Collagen type I was distributed throughout the disc whereas type II collagen demonstrated some localization in the NP region (Fig. S1). Proteoglycans were seen in both the AF and NP regions, but with increased proteoglycan content in the NP region compared to the AF region, as seen by histology and biochemical analysis (Fig. 2 F and I). Overall, biochemical analysis showed no significant differences in collagen or proteoglycan content between the engineered and native AF and NP after 6 mo in the disc space. In contrast, the reimplanted IVD showed significantly more collagen (p < 0.05) in the AF and significantly less proteoglycan (p < 0.05) in the NP after 6 mo. DNA content indicated that robust cell proliferation occurred within the disc space in both the engineered and reimplanted groups. Collectively the analysis of the ECM produced by TE-TDR indicates phenotypically appropriate spatial deposition of collagen and proteoglycans, which resulted in the composite structure observed in the engineered tissue (Fig. 1E).

Fig. 2.

TE-TDR produces integrated tissue with IVD-like collagen and proteoglycan content in the native disc space. Picrosirus red collagen staining for (A) native IVD, (C) discectomy group, (E) TE-IVD at 6 mo, and (G) reimplanted IVD at 6 mo. Alcian blue staining for (B) native IVD, (D) discectomy group, (F) TE-IVD at 6 mo, and (H) reimplanted IVD at 6 mo. (I) Biochemical analysis of glycosaminoglycan (proteoglycans), hydroxyproline (collagen), and DNA (cells) content for the native disc (n = 6), engineered TE-TDR (n = 6), and reimplanted IVD (n = 6) broken down by region of the disc (∗= p < .05 compared to native; # = p < .05 compared to all groups). Histology of TE-TDR and native tissue interface (VB, vertebral bodies; EP, end plate) at (J, K) 6 wk and (L, M) 6 mo demonstrates progressive integration of TE-TDR with native tissue (arrows point to small disruptions of integration at 6 wk, absent at 6 mo).

In addition to appropriate composition and arrangement of ECM, a critical requirement for a tissue-engineered IVD is integration with neighboring vertebrae.

Analysis of the implant-end plate interface by histology demonstrated progressive integration over 6 mo. At 6 wk integration was apparent, but small discontinuities were observed at the end plate and engineered tissue boundaries (Fig. 2 J and K). At 6 mo, the boundary was integrated completely with no large discontinuities observed at this interface (Fig. 2 L and M). Under polarized light microscopy, collagen bundle networks can be observed crossing the boundary between the end plate/bone interface and the engineered tissue (Fig. S2). This integration suggests that the proteoglycan and collagen-rich matrix produced by the engineered disc can function as a unit with the native spine.

Engineered IVD Produced Functional Tissue that Maintained Disc Height and had Similar Mechanical Properties to Native IVD.

To determine whether tissue-engineered IVD implants generated a functional motion segment, we analyzed the disc height and the dynamic mechanical properties of the motion segments. Maintenance of vertebral spacing is a critical role of the IVD and is a primary clinical indicator of IVD health. As such, maintenance of disc height is a critical indicator of the performance of a tissue-engineered IVD. Tissue-engineered IVD implants maintained 79 ± 18% of the disc height at 6 wk and 83 ± 13% at 6 mo, whereas the discectomy group collapsed (37 ± 3%) (Fig. 3A). At 6 wk and at 6 mo the engineered disc was comparable to the reimplanted native disc at maintaining caudal disc space (Fig. 3A).

Fig. 3.

TE-TDR produces a mechanically functional tissue in the native disc space. (A) Percentage of disc space height maintained at 6 wk and 6 mo for discectomy group (n = 12), reimplant group (n = 12), and TE-TDR group (n = 12). (B) Dynamic mechanical properties at 6 mo for TE-TDR implanted motion segments (n = 6), intact native motion segments (n = 6), and reimplant motion segments (n = 6) showing representative stress-strain curves at 1 Hz and apparent moduli from 0.01–10 Hz. (C) Equilibrium modulus, hydraulic permeability, and percent hysteresis (energy dissipation) for intact native motion segments (n = 6), TE-TDR implanted motion segment (n = 6), and reimplant motion segments (# = p < 0.05 compared to discectomy) (∗= p < 0.05 compared to native).

The ability to properly sustain axial loads is a key function of the IVD (34). To investigate how the engineered discs restored this function of the spine, we analyzed the dynamic mechanical properties of motion segments of native caudal spines and those receiving implants by imposing sinusoidal strains of 3% amplitude (3% tensile strain to 3% compressive strain) at frequencies ranging from 0.01 to 10 Hz. Intact motion segments were tested to assess effects of integration of the IVD tissue with the end plate and vertebrae, and the mechanical performance of tissue-engineered IVD was compared to that of native caudal motion segments and of spines that received reimplanted healthy discs (Fig. 3 B and C). Motion segments with engineered tissue had a moderately larger modulus than the native motion segment (p = 0.020) over the full range of frequencies tested. This modulus was maintained over both the compressive and tensile strain regions of the test, demonstrating the integration observed within the histology was mechanically functional at the levels of tensile strain imposed here. The capacity of the engineered IVD to dissipate mechanical energy, as indicated by the hysteresis in the stress-strain curve, was similar to that of the native motion segment (32% compared to 41%) (Fig. 3C). The reimplanted disc showed no significant difference in apparent modulus compared to the native disc but dissipated significantly less energy (p < 0.05) than both the engineered and intact native motion segment. The ability to dissipate mechanical energy is a critical function of the IVD, and the lack of this function may contribute to adjacent segment disease seen after fusion or total disc arthroplasty (12, 13). In this way, tissue-engineered IVD may restore the energy damping capacity of the spine in a way that is not available with current treatment options.

In addition to dynamic mechanical testing, the motion segments were subjected to stress-relaxation testing to determine the static compressive equilibrium modulus and the hydraulic permeability of the tissue (35). The equilibrium modulus and hydraulic permeability of the engineered disc were not significantly different from the intact native disc (Fig. 3C) or reimplanted disc. Overall, both the dynamic and quasi-static compressive data indicates the engineered discs restored compressive mechanical function to the caudal spine by producing a tissue with similar properties to the native IVD.

Discussion

The production of a TE-TDR implant for use in the treatment of discogenic back and neck disorders, as well as for use in studying IVD pathology and structure-function relationships has been a recent area of scientific interest. Yet, the successful creation of a functional engineered IVD in the native disc space had yet to be achieved (15–19). The ability to replace an IVD with a living and mechanically functional engineered IVD provides great promise in the treatment and understanding of spinal disease. However, the implantation of a TE-TDR implant into the native disc space provides a number of challenges to the disc that had yet to be addressed because of the lack of in situ studies.

The main challenges in developing and delivery of a tissue-engineered IVD implant are thought to be (i) generating functional tissue in the disc space (36), (ii) securing the implants in the spine to ensure that they will integrate with the neighboring vertebrae (15–18, 37), and (iii) developing an implant that can withstand the complex mechanical loading of the disc space (38). Here we demonstrated that our tissue-engineered IVD implants were able to meet all three of these challenges in the rat caudal spine by producing a collagen and proteoglycan-rich, well-integrated, and mechanically functional tissue in the native disc space. This study provides unique evidence that a tissue-engineered IVD implant can replace the native IVD in the spine.

A critical piece of enabling technology involved the use of clinically relevant imaging modalities, MRI and CT, to design tissue-engineered IVD. This technique allowed the creation of a TE-TDR implant in the size and shape of the native IVD (Fig. 1) and was integral to producing a disc that fit properly into the native disc space (Fig. 1C). MRI and CT were used to design the disc because it is easily translatable to clinical practice and makes patient-specific TE-TDR implant design possible. Patient-specific design is likely to be important for TE-TDR, as proper size matching is essential for IVD transplantation success (30). After 6 mo of implantation, a hydrated tissue was produced within the disc space (Fig. 1D) that maintained the overall shape of the implanted TE-TDR implant and disc space (Fig. 1 E and F).

The production of proteoglycans and collagen in quantities similar to the native disc demonstrated that the cells survived in the caudal disc space and successfully produced tissue de novo. In previous work in vitro, a proteoglycan content of only approximately 25% was obtained in vitro for both AF and NP (18) whereas a proteoglycan content on the order of the native NP and approximately 33% of the AF was obtained in subcutaneous implantations (17). In general, in vitro studies have not yielded IVD tissue with collagen and proteoglycan similar to that of native tissue (14, 17–19). Together with the current data, these results suggest that the caudal intervertebral disc space provides sufficient nutrient supply in our model to enable the production and maintenance of robust ECM by IVD implants.

There are likely a number of contributing factors to the success of this TE-TDR implant in producing a robust ECM in the disc space. These factors may include the increased cytokine signaling likely experienced in an in vivo environment and the anabolic effect that mechanical stimulation, which would be present in the native disc space, has been shown to have on IVD cells (39, 40). However, the disc space is known to have limited nutrient availability (i.e., glucose and oxygen) and poor waste product transport (e.g., lactate) out of the disc (41). The high permeability of the collagen/alginate TE-TDR construct at implantation may increase the availability of nutrients throughout the engineered IVD in the disc space and promote ECM production. At this time, it is unclear which of these effects, or others, are promoting such a robust ECM development, but it will be important in future work to further elucidate these mechanisms, as they will be important for the future design principles of TE-TDR implants. It is important to note that despite the similarities in ECM content between the native and engineered disc, a number of differences did exist. These included a lack of lamellar structure in the AF, differences in collagen I and II distribution, and larger variances in ECM content in the engineered discs than observed in the native discs. It is currently unclear how these differences will affect the long-term viability and function of TE-TDR, but will become important to understand as work in TE-TDR moves forward.

In addition to the ECM production, the ability of the TE-TDR implant to integrate with the end plates of the neighboring vertebral bodies and produce a functional motion segment is a key finding of this work. Although the arrangement of lamellar collagen in TE-IVD implants is not the same as native tissue, polarized light microscopy shows large, organized fiber bundles from the implants inserting into neighboring vertebrae (Fig. S2c). Furthermore, this integrated ECM produced by the TE-TDR implant is mechanically functional within the motion segment, as indicated by the ability of the motion segment to withstand tensile stresses imposed during dynamic mechanical tests (Fig. 3B). This mechanical data is consistent with the histology that shows progressive integration of the implants into the spine over the course of 6 mo (Fig. 2 L and M).

Despite the successful replacement of the native IVD with an engineered IVD composite in the caudal rat model, it is important to note the challenges that remain before TE-TDR can become a reality in the clinic. The engineered IVD demonstrated an ability to produce functional tissue in the caudal disc space of the rat. However, a number of important differences exist between the rat caudal disc space and the human cervical and lumbar disc space that may provide additional hurdles to clinical translation. First, the larger size of the human IVD and decreased permeability of the degenerated end plate will provide decreased nutritional availability in clinical patients compared to the rat caudal disc space (36). It will be important to investigate how the challenge of the nutrient-deprived human disc space environment will affect tissue development and integration of engineered IVD with native tissues. In addition, although there is evidence that the mechanical loading environment of the rodent tail shares some similarities with the human spine (27–29), the loading environments are undoubtedly different and will likely provide additional mechanical challenges to an engineered disc beyond that seen in the caudal disc space. Furthermore, only axial mechanical properties were investigated within this study. It will be important in future work to investigate how such composites perform in the full 6° of freedom experienced in the human spine, including bending and torsion.. Finally, ovine AF and NP cells were used in this study, but because of senescence observed in aged and degenerated AF and NP cells (7), an alternative cell source such as mesenchymal stem cells will likely be necessary in the clinic and may perform differently than the cells chosen for this study. Furthermore, it is unclear how these cells will respond when implanted into the inflammatory environment of the degenerated disc space (42), which is not replicated in the current model. Despite these challenges, this work demonstrates important and promising advancements in our understanding of how engineered IVD develops, integrates, and functions in the native disc space. Furthermore, this model of TE-TDR has immediate applications in the investigation of how various IVD properties (e.g., composition, architecture, permeability, cell type, stiffness, energy dissipation, etc.) affect structure-function relationships and pathology in vivo. The ability to introduce tissues with controllable AF and NP properties into the native disc space, in vivo, allows relationships to be studied that are not currently possible in traditional animal or in vitro models.

In summary, this data demonstrates that a tissue-engineered IVD can be implanted into the caudal spine, remain in place, withstand the mechanical loads, and survive and produce an integrated and mechanically functional ECM similar to the native IVD. These findings provide support for the development of tissue-engineered IVD technologies and evidence that the challenges associated with TE-TDR may be overcome. Translating the success of studies in rats to larger animal models will be a key step in moving this technology closer to clinical applications, whereas exploiting the rat model can provide valuable insight into in vivo IVD structure-function relationships and pathology. Overall, this study demonstrates promising advancements toward the therapeutic application of TE-TDR and develops a platform to better investigate and understand the IVD and its associated pathology.

Methods

Cell Preparation.

Cell preparation was based on previously described techniques (15, 16). IVDs were dissected out of lumbar region of skeletally mature (∼14 months) Finn/Dorset cross male sheep (Cornell University Sheep Program). Tissue was washed in PBS (Dulbecco’s PBS; Gibco BRL) and then separated into AF and NP region. Tissue was dissected into small pieces and digested in 200 mL of 0.3% wt/vol collagenase type II at 37 °C for 9 h for AF tissue and 6 h for NP tissue. Digested tissue was filtered through 100 μm nylon mesh (BD Biosciences) and centrifuged at 936 g for 7 min. Cells were counted and seeded at 2,500 cells/cm² in culture flasks with Ham’s F-12 media (Gibco BRL) that contained 10% fetal bovine serum, penicillin (Gemini Bio Products), (100 units/mL), streptomycin (100 μg/mL), amphotericin B (250 ng/mL), and ascorbic acid (25 μg/mL). Cells were cultured at 37 °C, 5% CO₂, and normoxia to confluence with media changes every 3 d. At confluence, cells were removed from flasks with 0.05% trypsin (Gibco BRL) and counted with a hemocytometer. Cells were then seeded into TE-TDRs.

IVD Fabrication.

T2 weighted MRI images and μCT images were obtained of caudal 3/4 disc level (imaging specifics in imaging section). T2 weighted MRI images were imported in DICOM format to slicOmatic v4.3 (TomoVision) and the NP was manually segmented and converted to point cloud images of the NP. Point cloud images were converted to surface and solid models in Studio 4.0 (Geomagic Inc.). This process resulted in a model containing the dimensions and shape of the NP (Fig. 1A).

μCT images were converted to DICOM format and imported into slicOmatic v4.3 (TomoVision) where the boney surfaces of the vertebral bodies were segmented to obtain the overall shape and dimensions of the caudal 3/4 disc space. The μCT-derived dimensions of the disc space were then combined with the MRI-derived NP model to obtain the target dimensions of the TE-TDR implant (Fig. 1A).

TE-TDR implant of target dimensions was created using contracted collagen (AF)/alginate (NP) technique (Fig. 1B) (15). MRI-derived NP surface and solid model was transferred into SolidWorks to create an injection mold of NP. The Injection mold was 3-D printed of acrylonitrile butadiene styrene plastic on FDM 3,000 machine (Stratasys).

Three percent (wt/vol) low viscosity grade alginate (FMC BioPolymer) seeded with 25 × 10⁶ NP cells/mL was mixed with 0.02 g/mL CaSO₄ (Sigma-Aldrich) to crosslink the alginate, and injected into the NP mold. Cell-seeded alginate NP was then removed from molds and placed in the center of a well of a 24 well plate. Collagen type I was obtained from rat-tail tendon (Sprague Dawley, 7–8-wk old) (Pel-Freez Biologicals) using established protocols (15, 43). One or two milligrams per millileter collagen gel solution seeded with 1 × 10⁶ AF cells/mL was subsequently poured and gelled around the alginate NP. Constructs were cultured for 2 wk in previously described media while collagen gel contracted around aliginate NP to the proper AF dimensions.

Surgery.

After 2 wk of in vitro culture, TE-TDR constructs were implanted into the caudal spine of athymic rats (n = 18). All animal procedures were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of the Hospital for Special Surgery. Rats were anesthetized using ketamine (Ketaset, 100 mg/mL) 80–90 mg/kg, and xylazine (Rompun, 20 mg/mL) 5 mg/kg, which were mixed together and administered intraperitoneally. If necessary, anesthesia was prolonged by administration of isoflurane via nose cone. A unique method was established to remove the native disc and to prepare the disc space for implant insertion in the tail. The vertebral column was exposed and the native IVD (caudal 3/4) removed. Upon removal, the caudal 3 and caudal 4 vertebral bodies were minimally retracted to allow the insertion of the engineered disc into the disc space. The disc space was released to press-fit the implant in place and wound closure was performed in layers. An initial dose of 0.01–0.05 mg/kg buprenorphine (Buprenex) was administered intraoperatively or immediately postoperatively prior to anesthetic recovery. Buprenorphine treatments were performed for 2 d postoperatively.

In addition to TE-TDR implants, two control groups were studied. The discectomy group followed the above procedure and removed the native disc but implanted nothing back into the disc space (n = 12). The reimplantation group followed the above procedure, removed the native IVD and then reimplanted that native IVD back into the disc space (n = 12).

After implantation, rats were maintained up to 6 mo. MRI images were taken of all three groups at 6 wk (n = 12/group) and 6 mo (n = 12/group) and analyzed for disc height. In the TE-TDR group, 12 animals were sacrificed at 6 mo and analyzed for histology (n = 6), mechanics (n = 6), and biochemistry (n = 6). In addition, a smaller cohort of six animals was sacrificed at 6 wk for histology. The two control groups were taken out to 6 mo. The reimplantation group was analyzed for histology (n = 6), mechanics (n = 6), and biochemistry (n = 6). The discectomy group was analyzed for histology (n = 6) but not biochemistry and mechanics due to the lack of tissue in the disc space.

Imaging.

μCT images of rat caudal spine were obtained from caudal 3 to caudal 5 vertebrae on a Scanco μCT 35 system (Scanco Medical) with an isotropic resolution of 30 μm.

MRI images were obtained on a 7.0 Tesla Bruker 70/30 Magnetic Resonance Imaging (Bruker Biospin) system. Rats were anesthetized with 1.5% Isoflurane during imaging procedures. A high-resolution T1-weighted flash sequence (resolution: 78.1 um × 58.6 um × 1 mm) was obtained for disc space measurements at 6 wk and 6 mo and a T2-weighted sequence (resolution: 104.2 um × 78.1 um × 1 mm) was obtained for implant design and at 6 wk and 6 mo post implantation.

Histology.

Spines and bone samples were cleaned of muscle and preserved in 10% phosphate-buffered formalin and were fixed at room temperature for 2 d. After an overnight running water rinse, samples were decalcified in 10% EDTA in 0.05 M Tris buffer, pH 7.4, until bone was soft and flexible. An overnight running water rinse was conducted in a VIP tissue processor to paraffin. Embedded samples were sectioned at 5 μm thickness and subsequently stained with alcian blue for proteoglycans, picrosirius red staining for collagen, and hemotoxylin and eosin.

Biomechanics.

TE-TDR motion segments and intact native motion segments were both cleaned of surrounding tissue to result in bone-disc-bone motion segments after sacrificing the animals at 6 mo. Prepared motion segments were mounted on ELF 3,200 mechanical testing frame (EnduraTech) using modified microvices (McMaster-Carr) (27). Unconfined stress-relaxation tests were performed at 5% strain incremental steps to a total of 20% strain. Equilibrium modulus and permeability were calculated by fitting resulting stresses to a poroelastic model (35). In addition, motion segments were subjected to a sinusoidal dynamic frequency sweep from 0.01–10 Hz at ± 3% strain around zero strain. Apparent modulus and percent hysteresis were calculated from dynamic data. All data analysis was performed using Matlab software.

Biochemistry.

Native IVD and TE-TDR tissue were both dissected out of disc space using a scalpel. For the native IVD, NP and AF tissue were separated and tested individually. For the TE-TDR tissue, the representative NP region at the center of the tissue was removed using a 2-mm biopsy punch and tested as NP and the surrounding AF region was tested as AF. Tissues were analyzed for glycosaminoglycan content using a modified dimethylmethylene blue assay (31), total collagen using the hydroxyproline assay (33), and DNA using the Hoechst dye assay (32). All data were normalized to wet weight.

Disc Measurements.

Images were taken of all engineered IVD immediately prior to implantation along with pictures of the fully intact explanted native IVD being replaced. Size measurements along the lateral and ventral-dorsal plane of both the engineered and native IVD were calculated using Image J software (National Institutes of Health). Measurements of engineered IVD at 6 mo after implantation were calculated from MRI images.

Statistical Analysis.

All statistical analysis was performed using two-factor ANOVA and Tukey post hoc test. Data represented as mean ± standard deviation. P values < 0.05 are considered statistically significant.