|Home | About | Journals | Submit | Contact Us | Français|
Degeneration of the intervertebral disc (IVD) represents a significant musculoskeletal disease burden. Although spinal fusion has some efficacy in pain management, spine biomechanics is ultimately compromised. In addition, there is inherent limitation of hardware-based IVD replacement prostheses, which underscores the importance of biological approaches to disc repair. In this study, we have seeded multipotent, adult human mesenchymal stem cells (MSCs) into a novel biomaterial amalgam to develop a biphasic construct that consisted of electrospun, biodegradable nanofibrous scaffold (NFS) enveloping a hyaluronic acid (HA) hydrogel center. The seeded MSCs were induced to undergo chondrogenesis in vitro in the presence of transforming growth factor-β for up to 28 days. The cartilaginous hyaluronic acid–nanofibrous scaffold (HANFS) construct architecturally resembled a native IVD, with an outer annulus fibrosus–like region and inner nucleus pulposus–like region. Histological and biochemical analyses, immunohistochemistry, and gene expression profiling revealed the time-dependent development of chondrocytic phenotype of the seeded cells. The cells also maintain the microarchitecture of a native IVD. Taken together, these findings suggest the prototypic potential of MSC-seeded HANFS constructs for the tissue engineering of biological replacements of degenerated IVD.
Degeneration of the intervertebral disc (IVD) is a common and significant source of morbidity in our society. From 25% to 80% of adults over the course of their life will experience an episode of significant low back pain,1 and as many as 90% of these people will improve without formal treatment.2 However, for those with low back pain that does not resolve spontaneously, surgical management may be indicated.3 Historically, surgical interventions have focused on fusion of the involved IVD levels, which may eliminate pain but does not attempt to restore disc functions. Over 200,000 spinal fusions were performed in the United States in 2003 in an attempt to address the pain associated with lumbar disc degeneration, up 134% from 1993.4 Spinal fusion is not a benign procedure; it significantly alters the biomechanics of the spine, which can cause adjacent level degeneration.5,6 As a result, there has been increasing interest in the concept of IVD replacement in the past decade.7–10 The replacement of the IVD holds tremendous potential as an alternative to spinal fusion for the treatment of degenerative disc disease (DDD) by offering an effective, motion-preserving alternative.
At present, two lumbar total disc replacement (TDR) implants (Charite, DePuy Spine; Prodisc-L, Synthes Spine) are approved for human use in the United States, and another two implants (Maverick, Medtronic; Flexi-core, Stryker Spine) are far along in the Food and Drug Administration approval process.5–8 These disc replacement technologies are all designed to preserve flexion, extension, and lateral bending motions and restore disc space height; however, they do little to address compressive forces and their longevity is unknown, but inherently limited due to their inability to biointegrate and subsequently remodel. Furthermore, metal on polyethylene and, to a lesser degree, metal on metal implants generate wear debris that can cause cellular reactions that lead to osteolysis and other potentially deleterious effects.11 In light of these drawbacks, a cell-based tissue-engineered replacement disc offers a potentially promising treatment alternative for DDD, by combining the height restoration and motion preserving attributes of metallic implants with the ability for permanent biointegration.
Cell-based tissue engineering is an emerging field that involves seeding cells into a biomaterial scaffold to fabricate functional biological substitutes for the replacement of lost or damaged tissues.12 For successful cell-based tissue engineering, cells should ideally interact with an appropriate scaffolding material that closely mimics the structural, biological, and mechanical functions of native extracellular matrix (ECM).13 To achieve this goal, a scaffold material must be structurally, biochemically, and biomechanically biocompatible with the seeded cells. The electrospinning method has been used to fabricate three-dimensional, porous, nano-scale fiber–based scaffolds for various tissue engineering applications.13–16 Electrospun nanofibrous scaffolds (NFS) have recently been recognized as a novel biomaterial that closely mimics the architectural scale and morphology of native fibrillar collagen, including that in the native IVD ECM. The combined characteristics of high porosity (90%), favorable mechanical properties, high surface area-to-volume ratio, and a wide range of possible pore sizes (5–475μm)17 make NFS a promising and effective component for functional tissue engineering. Over the last several years, we have engineered a number of different bioengineered tissue constructs using biodegradable polymers seeded with various cell types, including adult human multipotent mesenchymal stem cells (MSCs).18–23
To develop a tissue engineering solution to IVD degeneration, the engineered construct must exhibit characteristics of structural support and discrete tissue architecture that mimics that of the IVD, that is, annulus fibrosus (AF) and nucleus pulposus (NP). For this purpose, in this study, we have combined poly(L-lactic acid) (PLLA) NFS and hyaluronic acid (HA) hydrogel to produce an amalgam (hyaluronic acid–nanofibrous scaffold, HANFS). The addition of HA is intended to improve structural support and to enhance biocompatibility during cell differentiation and proliferation.24–28
We report here the use of the novel HANFS seeded with human bone marrow–derived MSCs to construct an engineered IVD composed of both AF- and NP-like components.
The materials and reagents used in this study and their sources are as follows: PLLA (MW=50,000) and poly (2-hydroxyethyl methacrylate) (poly-HEMA), Polysciences (Warrington, PA); chloroform and N,N-dimethylformamide (DMF), Fisher Scientific (Pittsburgh, PA); Dulbecco's modified Eagle's medium (DMEM, high glucose), phosphate buffered saline (PBS), and penicillin–streptomycin, Gibco BRL Life Technologies (Grand Island, NY); fetal bovine serum (FBS; selected lots), Atlanta Biologicals (Atlanta, GA); Hank's balanced salt solution (HBSS), BioSource International (Camarillo, CA); ITS-plus Premix, BD Biosciences (Bedford, MA); recombinant human transforming growth factor-β1 (TGF-β1), R&D Systems (Minneapolis, MN); Trizol reagent, glycogen, and SuperScript First-Strand Synthesis System kit for RT-PCR, Invitrogen (Carlsbad, CA); collagen type I (SP1.D8) and collagen type II antibody (II-6B3), aggrecan antibody (1-C-6), and cartilage proteoglycan link protein antibody (8-A-4), Developmental Studies Hybridoma Bank (Iowa City, IA); Broad Spectrum Histostain-SP Kit, Zymed Laboratories (San Francisco, CA); Blyscan™ sulfated glycosaminoglycan (sGAG) assay kit, Accurate Chemical & Scientific (Westbury, NY); RediPlateTM 96 PicoGreen dsDNA Assay, Molecular Probes (Eugene, OR); and all other reagents, Sigma Chemical (St. Louis, MO).
Human bone marrow–derived MSCs were obtained from patients undergoing lower extremity reconstructive surgery as previously described,19,29,30 and with approval by Institutional Review Board (Walter Reed Army Medical Center, Washington, DC). Briefly, bone marrow was harvested from the interior of the femoral neck and head or the intramedullary canal of long bones using a bone curet, and transferred to 50mL conical tubes containing DMEM supplemented with 10% FBS and antibiotics (50μg/mL of streptomycin, 50 IU of penicillin/mL). The marrow-containing tube was vigorously vortexed, and a 10 cc syringe fitted with an 18 G needle was used to aspirate the homogenized bone marrow solution while leaving large debris such as bone chips to settle to the bottom of the tube.
The collected bone marrow was centrifuged at 1000rpm for 5min, and the resultant cell pellet reconstituted in culture medium and plated in 150cm2 cell culture flasks. Nonadherent hematopoietic and red blood cells were removed during medium changes leaving adherent MSCs attached to the culture flask. Cells were grown in a medium composed of DMEM supplemented with 10% FBS and antibiotics, and maintained at 37°C in a humidified, 5% CO2 atmosphere. Cells were consistently removed using 0.05% trypsin/ethylene diamine tetraacetic acid (EDTA) solution when they reached approximately 80% confluency, and the culture medium was replaced every 3 days. Cells obtained at passages 1–3 were used in this study.
NFSs were fabricated by the electrospinning process as described previously.16 Briefly, a solution of the biodegradable polymer, PLLA, was prepared at 0.145g/mL by dissolving 1.6g of PLLA in the organic solvent mixture composed of 10mL of chloroform and 1mL of DMF followed by vortex-mixing overnight at room temperature. The polymer solution was placed in a vertically fixed 20mL glass syringe fitted with a 10cm, 18 G needle. A 16kV electric field was applied at a distance of 20cm between an aluminum foil sheet covering a copper plate and the needle tip resulting in a 0.8kV/cm charge density (voltage/distance) on the polymer solution. After 11mL of polymer solution was totally consumed at the rate of 1.8mL/h, an electrospun PLLA mat measuring 144cm2 with a thickness of approximately 1mm is formed homogeneously on the aluminum foil. The mat was then removed, placed in a vacuum chamber for at least 48h to remove organic solvent residue, and then stored in a desiccator. The physical properties of these electrospun nanofibers have been previously discussed.17,23
The electrospun mat was cut into 1cm2 units, and both sides of the scaffold were sterilized by ultraviolet irradiation in a laminar flow hood for 30min. To provide a hydrophilic surface conducive for efficient cell attachment, scaffolds were prewetted by immersion in HBSS for 24h. MSCs grown in 150cm2 cell culture flasks were trypsinized, counted, and plated at a density of 400,000 cells/cm2 onto the surface of prewetted scaffolds that were placed in 24-well culture plates precoated with 0.3% poly-HEMA to prevent cell attachment to the tissue culture polystyrene surface. Cells were seeded onto one side of NFS and incubated at 37°C for 2h to allow MSCs to diffuse into and adhere to the scaffold. Two hours later, the other side was seeded with cells and incubated for additional 2h. During the 4h of incubation, 20μL of serum-containing culture medium was applied to each cellular scaffold every 30min to prevent desiccation of the constructs.
The HANFS amalgam was created by injecting approximately 250μL of a 2×107 cells/mL HA slurry into the center of the NFS square using a 23 G needle (Fig. 1). This created a pressurized pouch, cuboidal in shape, with an inner core of HA and nanofibrous elements surrounded by a sheath of dense NFS.
To induce an IVD cell-like phenotype, all cell-seeded cultures were maintained in culture medium composed of DMEM, 10ng/mL of TGF-β1, 100nM dexamethasone, 50μg/mL ascorbate 2-phosphate, 100μg/mL sodium pyruvate, 40μg/mL proline, antibiotics (50μg/mL of streptomycin and 50 IU of penicillin/mL), and ITS-plus Premix diluted 1:100 for final concentrations of 6.25μg/mL insulin, 6.25μg/mL transferrin, 6.25μg/mL selenious acid, 5.33μg/mL linoleic acid, and 1.25mg/mL bovine serum albumin. Cell culture medium was replaced every 3 days. It is important to note that the intent of the study was not to replicate native IVD cells, rather to utilize cells that have the capacity to produce a similar ECM.
Engineered constructs were washed in PBS, fixed in 4% phosphate-buffered paraformaldehyde at 4°C for 30min, dehydrated through a graded series of ethanol, infiltrated with Histo-Clear (American Mastertech Scientific), embedded in paraffin, and sectioned at 10μm thickness. For histological analysis, sections were deparaffinized in Histo-Clear, rehydrated using a graded series of ethanol, and stained with hematoxylin and eosin (H&E) and Alcian blue (pH 1.0).
Immunohistochemical analysis was used to detect aggrecan; collagen types I, II, and IX; and cartilage proteoglycan link protein. To detect collagen types I, II, and IX, sections were predigested with 300U/mL of hyaluronidase at 37°C for 15min before incubation in 15μg/mL of the respective antibodies. For the detection of aggrecan and link protein, sections were predigested with 1.5U/mL of chondroitinase for 15min at 37°C, and then incubated in 10μg/mL of aggrecan antibody at 37°C for 1h or in 6μg/mL of link protein antibody at 4°C overnight. Broad Spectrum alkaline phosphatase–conjugated secondary antibodies were used for immunodetection and developed using BCIP-NBT (Broad Spectrum Histostain-SP kit). Tissue sections without treatment with primary antibodies served as controls.
Cultured cell-seeded HANFS constructs were harvested, washed in PBS, fixed in 2.5% glutaraldehyde for 20min, dehydrated in a series of graded concentrations of ethanol, and vacuum-dried. Dehydrated constructs were cut, sputter-coated with gold using a sputter coating unit (MED010; BAL-TEC, Liechtenstein). Top and cross-sectional views were imaged at an accelerating voltage of 20kV using a Hitachi Model-4500 scanning electron microscope (Japan).
Total RNA was isolated from HANFS constructs with 800μL Trizol reagent. After addition of 160μL chloroform to the homogenized samples, RNA was precipitated using 400μL of isopropanol. RNA pellets were dissolved in 20μL of RNase- and DNase-free water, and RNA yields were estimated based on A260. First-strand cDNA was reverse transcribed from 3μg of total RNA using the SuperScript First-Strand Synthesis System kit, and gene-specific amplicons were produced by PCR using the oligonucleotide primers shown in Table 1, for collagen types I, II, IX, X, and XI; cartilage oligomeric matrix protein (COMP); and glyceraldehydes-3-phosphate dehydrogenase (GAPDH).
GAPDH gene expression was used as an internal control for mRNA loading. Thirty-two cycles were used to amplify all gene sequences, and the PCR products were electrophoretically analyzed with ethidium bromide staining.
sGAG content was quantified using a commercially available assay kit (Blyscan). sGAGs were extracted from HANFS cultures following a method modified from a previous report.31 Briefly, samples were harvested, washed with PBS, and digested with 300μg/mL papain in 20mM sodium phosphate (pH 6.8), 5mM EDTA, 2mM dithiothreitol (DTT) at 60°C for 12h. The extract was cleared by centrifugation and analyzed for sGAG and DNA. For sGAG, the sample was reacted with the Blyscan dye reagent, 1,9-dimethylmethylene blue (DMMB), for 30min. Unbound dye was removed by centrifugation, and bound dye was released from the insoluble sGAG–dye complex, and quantified spectrophotometrically on the basis of A656. The total amount of sGAG was estimated from a standard curve generated using chondroitin 4-sulfate as a standard. Results were expressed as the ratio of sGAG amount to DNA amount. For DNA, the extract was assayed using the RediPlate 96 PicoGreen dsDNA methods (Invitrogen).
Data collected from quadruplicate samples are expressed as mean±SD, and analyzed statistically with a two-tailed Student's t-test, with significance determined at p<0.05.
The HANFS construct fabricated here approximated the two regions of the native IVD, the NP and AF. The exterior or cortical region consists of nonwoven electrospun nanofibers that became tensioned after HA injection (Fig. 1). MSCs were loaded into this region prior to HA injection to ensure uniform distribution of cells. Histological and immunohistochemical staining was performed at 7, 14, and 28 days and confirmed both uniform distribution of cell loading as well as cell morphology changes during the experimental period. H&E staining demonstrated uniform cell loading in the AF region at the early time points. At later time points, the cells began to elongate and became layered in a concentric fashion, similar to the microarchitecture of the native AF, which is organized in a series of concentric fibrous rings that impart much of the tensile strength to the IVD.32 Increased cartilaginous ECM deposition was also detected with alcian blue staining in the histological sections, with complete filling of the pores within the NFS in the AF region by day 28 (Fig. 2). Increased sulfated proteoglycan content in the ECM was detected over the culture period as seen by increasing alcian blue staining on histological sections, and as noted by significant increases in sGAG content of the constructs as determined using the DMMB assay (see below).
Initially, the NP region appeared to be sparsely populated with cells and showed little ECM deposition (Fig. 2). Later in the culture period, after deposition of more ECM, cells were detected in greater numbers and they appeared rounded and encapsulated, notably different in morphology from the layered cells in the cortical AF region of the construct. Over time in culture, the adjacent AF and NP regions assumed the different histological features described above, with the AF completely encapsulating the NP region.
Alcian blue staining, which revealed a sulfated proteoglycan–rich ECM, showed increasing intensity in the HANFS construct through the 28-day culture period, with the most intense staining localized to a ring-like zone in the AF region (Fig. 2). Alcian blue staining of the NP appeared amorphous without distinct organization. This staining pattern correlated with the intended architectural design of the construct, that is, a cross section consisting of an organized ring-like structural barrier enveloping a relatively amorphous center. It was noteworthy that an integrated transition formed between the two regions in these constructs, and approximated that for the AF and NP regions in native human IVD, where there is no distinct division between the two regions.
Immunohistochemical staining for known IVD ECM components showed the presence of collagen types I and II, collagen type IX (not shown), aggrecan, and link protein. The staining pattern was similar to that seen with alcian blue staining, with intensity increasing over the 28-day culture period (Fig. 3). Substantial deposition of collagen types I, II, and IX; aggrecan; and link protein was first noted in the immediate pericellular area, with increasing signal in the intercellular matrix over 28 days. The presence of these ECM components was consistent with the cartilage phenotype of a native IVD. Specifically, collagen types II and IX indicated the presence of a fibrillar collagen network, in conjunction with a proteoglycan-rich matrix as shown by the intense staining for aggrecan and link protein.
Scanning electron microscopy revealed uniform cell distribution in the NFS, similar to that observed in our previous cartilage tissue engineering studies.18,19 Matrix accumulation was apparent in both the AF and NP regions and continued to increase over time, filling the small pores between the nanofibers and the larger void within the NP region (Fig. 4). Nanofiber architecture remained intact for the duration of the experiment and ultimately became intimately associated with the ECM produced by the seeded cells.
RT-PCR was performed to assess the expression of ECM genes characteristic for the cartilage phenotype of the IVD. Specifically, collagen types I, II, IX, X, and XI; aggrecan; and COMP were all found to be expressed by day 14, with collagen type I and COMP expression occurring as early as day 7 (Fig. 5). Of particular significance was the initiation and maintenance of collagen type II and IX expression. Expression of collagen types II and IX generally required higher-density cell culture in a three-dimensional microenvironment.19,33 These conditions were thus apparently achieved in the HANFS constructs.
Production of sGAG was also detected in the engineered constructs. The DMMB assay showed sGAG accumulation as early as day 7, and the levels increased over the 28-day culture period (Fig. 6).
In this study, we report the successful production of a biphasic cartilaginous tissue construct, using a novel HANFS biomaterial amalgam seeded with adult human MSCs. The biphasic architecture and ECM distribution and composition of the constructs resemble those of native IVD, suggesting that these constructs have the potential to develop and mature into IVD-like tissues. While cell-based tissue engineering has been considered a promising approach to tissue repair and regeneration, applications to IVD have been limited. While other investigators have engineered portions of the IVD,34 this is the first construct to be developed as a complete unit, possessing both AF and NP regions. Since the IVD fails as a unit, an ideal biological solution must address the entire disc, not just one component. Furthermore, the fundamental challenge for partial IVD constructs has been their inability to restore the complete IVD as a functionally integrated unit.34,35 In this study, we have demonstrated in vitro that MSCs loaded on a HANFS scaffold can produce a bioengineered cartilaginous tissue that exhibits the distinct anatomical architecture and biosynthetic features of the native human IVD.
While multiple biological approaches have been considered for the treatment of DDD, including biologics- and gene-based therapies, we will focus our discussion on tissue-engineered solutions and their comparison to current TDR options.36–38 The current tissue engineering approaches for IVD replacement primarily focus on regenerating a specific component of the IVD.39–43 Specifically, the NP was the first and continues to be the most frequent target for cell-based tissue-engineered solutions for IVD pathology.44,45 The aim of the approach reported here is the introduction of in vitro–cultured IVD using an MSC-based tissue engineering approach to restore function of the degenerated IVDs. Sato et al. have developed a true tissue-engineered construct to regenerate injured portions of the IVD, in that it is composed of allogeneic AF cells loaded on a three-dimensional collagen scaffold, rather than simply injecting a slurry of cultured cells.46 The biocompatibility and successful integration of this experimental construct provide optimism for the more challenging endeavor presented here to regenerate an entire IVD, rather than restoring regions resected or lost during treatment of herniated NP. Meisel et al. have taken cell-based therapies to the next step using autologous disc chondrocyte transplantation (ADCT), reporting results at 2-year follow-up in the first 28 patients (12 received ADCT) enrolled in an unblinded randomized (no control) multi-center clinical trial (EuroDisc Study).44 This study has suggested that transplanted ADCTs remain viable and restore proteoglycan and collagen type II production, thus correlating to the findings from their previous canine model.47 While the results of this clinical study may not be dramatic, the fact that cell-based IVD therapies are being clinically tested shows promise for future cell-based therapies.
Bioengineered disc replacements should outperform traditional metal and synthetic devices currently used in spine surgery. Unlike metal and synthetic polymer devices, a bioengineered replacement disc will not release exogenous wear particles over the life of the implant. Although newer TDRs have adopted a metal-on-metal design with significantly lower wear rates than those for metal-on-polyethylene prostheses, the long-term biological effects, both local and systemic, of metal ions created when this type of prosthesis wears are still unknown. Additionally, metal ions from spinal implants, while to a significantly lower degree, have been demonstrated to incite wear-induced osteolysis.48,49 In contrast, bioengineered IVDs have the potential for minimal to no antigenicity and, most importantly, can biointegrate. Thus, unlike metal prosthesis that begin to wear from the time they are implanted, a biointegrated disc may hold the potential to maintain and improve its function with time.
Bioengineering an IVD requires the combination of an ideal cell source, biomaterial scaffold, and enabling biological signals. As noted above, previous IVD tissue engineering studies have investigated the proliferative and regenerative potential of mature NP and AF cells.41 These cell types have the disadvantage of being scarce and difficult to safely harvest.34 The amount of harvestable NP cells is greatly reduced in degenerated IVDs, and harvesting requires injury to the annulus (needle puncture, typically). Most pragmatic of concerns is that autologous harvesting may not be financially and commercially sustainable. A very viable option is thus the use of MSCs. A number of studies have demonstrated the potential use of MSCs in IVD tissue engineering with promising results.50–53 MSCs may provide a more ideal cell source for the regeneration of the two distinct regions of the IVD for the following reasons: (1) they are easily accessible for harvest, with bone marrow being a frequently used source; (2) they possess extensive self-renewal and expansion capability; (3) they possess little to no immunogenicity; and (4) most importantly, as demonstrated in this study, they are capable of differentiating into cells phenotypically and biosynthetically similar to those found in AF and NP regions of the IVD. These characteristics make MSCs well suited for cell-based tissue engineering of an IVD.
In this study, the MSCs were seeded into a novel HANFS construct; the cells proliferated, differentiated, and produced a proteoglycan-rich ECM with a protein expression profile similar to that of a native IVD. While there are many biocompatible polymers used in IVD tissue engineering,46,54,55 we have found that electro-spun PLLA nanofibers, with the unique characteristic of nanofiber size that exponentially increases the surface area of the scaffold, allow for efficient cell loading and generation of an environment highly conducive to chondrogenesis. The efficient production of a proteoglycan-rich matrix within the NFS is critical for IVD tissue engineering, since diminished proteoglycan production and disc dehydration are central to DDD.
We performed RT-PCR for several genes that are thought to be important for IVD ECM composition. Full interpretation of the temporal expression of these genes listed is not possible at this time as there is no clear order that has been established in the developing IVD. Additionally, we have found that there is asymmetric loading of our construct that may result in differences in the quantity of ECM deposition in various regions of the disc. For example, while Col II is expected to be seen in both the AF and NP of a native disc, its concentration in the NP should be greater. Our construct initially demonstrates higher Col II expression in the AF, but progressive increases in NP and AF are seen throughout the culture period and equivalent production in both regions is seen by 28 days. We are currently investigating new cell loading techniques that may improve the distribution of cells and associated ECM production.
The biphasic HANFS scaffold and IVD growth medium directed region-specific differentiation of MSCs into cells with phenotypes and biosynthetic activities resembling those of the two distinct regions in the native IVD. Our preliminary studies reported here thus demonstrate the potential for this specific construct to be a prototype for further research into the creation of an ideal tissue-engineered IVD. Future studies will evaluate the mechanical properties of the MSC-seeded HANFS constructs and to optimize the culture conditions. It is likely that exposure of the constructs to mechanical stimulation during culture will improve ECM production, metabolism, and the mechanical characteristics of the tissue-engineered IVD.56,57 In addition, a number of growth factors have been shown to play a role in IVD development, maintenance, and repair,58–61 and the importance of these growth factors in IVD tissue engineering will be studied to optimize culture conditions and construct function.
DDD is a widespread problem with no biologic solution, and for which current surgical options have produced variable results. The search for a biologic solution for DDD is in its infancy. Currently, research is focused on proof-of-concept investigations, such as the study reported here. We have demonstrated here that a tissue construct that resembles the overall gross, histological, biochemical, and biosynthetic properties of a native IVD can be engineered by implanting human MSCs into a novel HANFS scaffold.
This work was supported by the Intramural Research Program of the NIAMS, NIH (ZO1 AR41131).