Disk cell engineering

The IVD is composed of multiple subunits that integrate seamlessly to form sophisticated and complex mechanical function. It has 3 main structures: a gelatinous nucleus pulposus (NP) core wrapped around and confined by a fibrous lamella structure, the annular fibrosus (AF), and cartilaginous endplates (CEPs) of the vertebrae sandwiching the NP and AF (see article elsewhere in this issue by Grunhagen and colleagues). The 3 compartments are different in mechanical properties and are functionally dependent on each other. These compartments are made up of various matrix components, predominantly collagens and proteoglycans, and comprise different cells that are thought to play roles in maintaining the matrix integrity. From a developmental point of view, IVD formation involves the diversification of cells from a primitive anlage consisting of the notochord and its surrounding mesenchymal cells, a process that involves vigorous cell differentiation during the embryonic stage and continual postnatal remodeling of its microenvironment, including the extracellular matrices.

The primary role of the IVD is to provide mechanical support and motion, which is largely attributed to the viscoelastic properties (the viscous and elastic behavior under deformation) of the IVD. IVD-like properties and function may possibly be obtained by simply using acellular scaffolds or devices. Materials with desirable viscoelastic properties pertaining to IVD are often biodegradable, however, and, therefore, their function may not be sustained in the long term. Cell-containing constructs are advantageous with cells enable to remodel the scaffold template, thereby maintaining or enhancing matrix integrity. More importantly, IVD transplantation studies have demonstrated the integration of disk allograft to recipient vertebrae by biologic remodeling at the endplates, including that of misaligned disks, which were also found to self-correct postsurgery. Because the self-correction is considered to eventually contribute to motion segment kinematics and stability in the long term, cellular IVD constructs would theoretically outperform acellular prosthetic devices in function.

The loss of integrity and viscoelasticity of the NP is one of the earliest observable events in disk degeneration, suggesting that the engineering of the NP is crucial to the success of a functional IVD construct. Considerable effort has been invested in understanding the NP, in particular, delineating the NP cell phenotype so as to facilitate NP cell engineering for the creation of cellular IVD constructs. Two cell populations are thought to exist in the NP: small nonvacuolated cells with a chondrocyte-like phenotype, and large vacuolated cells, often referred to as notochordal cells. The large vacuolated NP cells have been shown to originate from the notochord, whereas controversy still surrounds the origin of the chondrocyte-like cells, particularly in human NP. Nevertheless, recent studies have provided some insights to the molecular identity of the latter type of cells based on microarray-based gene expression profiles of NP cells from adult human, bovine, and chondrodystrophoid dog. These studies did not yield common genetic markers that are translatable among different species but rather indicated the presence of species-specific markers, for example PAX1 and FOXF1 in human, A2m and Anxa4 in canine, and Snap25 and Krt8 in bovine. Expression profile studies of rodent NP cells, which predominantly consist of large vacuolated notochordal cells, also suggested that markers may not be shared with human NP cells, and indicated that notochordal and chondrocyte-like NP cells may be distinct in phenotypes. This is supported by cell sorting studies in which the two populations were separately extracted from the same animal for comparative analyses. Nevertheless, recent studies suggest that a cross talk may possibly exist between the two cell populations, complicating the search of the true NP cell identity.

IVD and articular cartilage have different mechanical properties. The transplantation of chondrocyte-like cells or the use of such cells in the engineering of NP constructs may not produce a disk with ideal function. Alternatively, future IVD engineering will likely use stem cells and other progenitor cell types to differentiate into NP cells and generate them in large quantity. Unless the phenotype and the functional characteristics of the notochordal and chondrocyte-like NP cell types are clearly defined, the generation of authentic NP cells from stem/progenitor cells and hence bioengineering of NP with native mechanical properties may not be achieved. In vitro studies have suggested that stimulation by co-culturing with NP cells, induction of chondrogenic transcription factor SOX9, or stimulation by chondrogenic growth factor transforming growth factor β1 (TGF-β1) are attractive strategies to drive differentiation of adult stem cells, such as mesenchymal stem cells (MSCs), into NP-like cells. Whether or not the differentiated cells have attained an authentic NP cell or chondrocyte-like phenotype, however, remains elusive. Engineering the NP with chondrocytes is not desirable because it is likely that the construct will become hyaline cartilage instead of NP tissue and hence possess inappropriate viscoelastic properties.

Annulus fibrosus cells are generally referred to as fibrochondrocytes. This phenotype is based on the ability of AF cells to produce collagen I and III, in addition to collagen II and aggrecan, which are produced to a lower extent. Compared with NP cells, the molecular phenotype of AF cells and their changes in disk degeneration are less clear. Although NP and AF are morphologically different and supposedly have different cell phenotypes, recent transcription profiling studies indicated that AF cells express a large number of nonhousekeeping genes at similar level to that of NP cells. Nevertheless, other potential markers are suggested to be differentially expressed at higher levels in the AF relative to the NP, for example, VCAM1 in human. Fibromodulin has been shown to be a specific marker of the AF in rodent ; however, its expression pattern in humans or large animal models is unclear. Alternatively, although the CEP is assumed to be analogous to hyaline cartilage and consists of chondrocytes, the molecular phenotype of CEP cells is also not clear, and there is a lack of evidence that demonstrates their similarities. Histologic findings have suggested that there is a difference in the glycosaminoglycan composition and collagen VI and X expression between CEPs and growth plate cartilage. It is not clear whether or not the use of chondrocytes in bioengineering can fully fulfill the function of the CEP.

Engineering the disk microenvironment

The bulk of the IVD is composed of extracellular matrix, which plays prominent roles in the regulation of the disk cell environment and providing anchorage to disk cells. The extracellular matrix of the IVD, like cartilage, is comprised of mostly collagens, which provide tensile strength of the disk, and proteoglycans, which function to reduce the internal friction in the disk matrix and to distribute load. They also account for the viscoelastic behavior of the IVD, contributing to the shock-absorbing property. Based on the relationship between function and form, materials that can mimic the anatomic architecture and mechanical properties of native IVD (reviewed by Nerurkar and colleagues ) are of interest to disk bioengineers.

Like hyaline cartilage, aggrecan and collagen II are the two main extracellular matrix components of mature NP. Although the NP in mature human IVD is thought to be analogous to articular cartilage, the nature of the matrix and hence their mechanical properties are not exactly the same. NP in young individuals has a high proteoglycan to collagen content, with a suggested glycosaminoglycan-to-hydroxyproline ratio of 27:1, in comparison with a 2:1 ratio in hyaline cartilage. The high proteoglycan content in the NP matrix facilitates the retainment of water, which attributes to the high hydrostatic pressure exhibited in the NP. Hydrogel scaffolds have been commonly used with the intention of simulating the NP microenvironment and entrapment of the newly deposited proteoglycan to facilitate the establishment of hydrostatic pressure. The effectiveness of various hydrogel-based scaffolds, either made from natural hydrophilic biomolecules or synthetic polymers, for NP engineering or repair has been documented. To date, alginate is one of most commonly adopted hydrogel scaffolds for NP cell culturing due to its ease of manipulation, biodegradability, and inert bioactivities. Hyaluronic acid (HA) has been used for the treatment of osteoarthritic knee joints via the direct application into the synovial cavity, and in vivo studies propose that HA or HA-derived hydrogel may facilitate NP function and promote motion segment mechanics. Because the NP plays an important role in withstanding the compressive load so as to maintain disk height and range of motion of spinal segment, however, pure hydrogel scaffolds, which lack confined compressive strength, may not be adequate for NP engineering. Collagen I microspheres and calcium polyphosphate may, alternatively, provide good tensile and compressive strength to support stem cells or NP cells in NP engineering. Recent studies have focused on the generation of collagen-incorporated or polymer-linked hydrogels. These hybrid scaffolds mimic the native microenvironment of the NP to reproduce its viscoelastic and load distribution behavior within the IVD. Although swelling is known to provide the main load-bearing mechanism in the NP, the extensive collagen network inside the disk has also been suggested as supporting a considerable portion of load because the collagen fibril meshwork contributes to the compressive modulus to the tissue. Moreover, collagens have been shown to act as a reservoir of signaling ligands, such as TGF-β1 and bone morphogenetic protein 2 for collagen II, and are able to transduce mechanical signals, therefore serving as important regulators of cell function and homeostasis.

In addition to NP engineering, current research has attempted to use injectable scaffolds as carriers to deliver cells with the aim of salvaging disk degeneration or on its own, as fillers for NP replacement using a minimally invasive approach. Studies have also developed injectable materials that have the ability to self-assemble into a higher-order network, resulting in a solution-to-gel transition. For example, atelocollagen (pepsin-digested collagen) can self–cross-link to form a fibrous meshwork, and chitosan and synthetic peptides have been reported to self-assemble into a nanofiber network. Other natural biomolecules, including hyaluronan and chitosan, when modified with cross-linkable moieties are capable of chain polymerization through photochemical reactions. These materials, as injectable media, may deliver cells of interest by providing a transient framework that prevents leakage of implants and allows for the accumulation of the extracellular matrix deposited by the introduced cells.

Although intradiscal pressure exerted by the NP plays an important role in disk function, it critically depends on the integrity of the AF. In addition, the trans-AF delivery system remains the commonly used way to manipulate the NP in clinical and experimental settings. The construction of AF may facilitate the repair of prolapsed disks and supplement nucleus replacement. Annulus closure devices may possibly provide an effective means to treat prolapsed disks; however, they may restrict segment motion and possibly modify the load distribution in the IVD. AF construction is understandably indispensable to IVD engineering, but effective bioengineering of AF may not be implemented without proper understanding of its microstructure and mechanics.

The AF lamellae are mainly composed of collagen I fiber bundles and have anisotropic mechanical behavior. These bundles are approximately concentric to the lamellae around the NP, where the direction of alignment in one lamella differs to the next by 30°. This angle-ply architecture of lamellae is thought to be designed to resist shear resulting from complex physiologic stresses, such as a combination of axial loading and torsion. Annulus fibers are interconnected via intralamellar crossbridges and interlamellar bridges. In vitro studies showed that excess circumferential constraint may have a negative impact on NP cell metabolic activities, which suggests that tissue rigidity needs to be carefully controlled during AF construction. The rigidity of the AF largely depends on the mechanical properties of the materials used during fabrication. Various materials have been tested for AF tissue construction, including porous silk, polymer nanofibers, polylactide/Bioglass composite, and alginate/chitosan composite. These scaffolds provide a framework of desirable mechanical and bioinductive properties for future AF engineering. An alternative is collagen gel or collagen-glycosaminoglycan composite, although fabrication into a specific geometry (such as fibers or lamella) or characteristic topographic template may be limited. Because AF cells are normally aligned with the lamella fibers, it may be ideal that AF constructs can be engineered through simultaneous controlled placement of AF cells and orientation of the matrix they interact with, such as using scaffolds with specifically designed microgrooves.

Collagen fiber structure determines the mechanical strength and elasticity of the annulus. Recent studies in AF engineering have shed light on some important aspects of its structural properties at the molecular level. Nerurkar and colleagues showed that, through electrospun nanofiber fabrication, a bilamellar tissue model with AF-like angle-ply architecture can be generated. By mechanical testing and modeling, they demonstrated that the bonding between the angle-ply lamellae is crucial to the resistance of interlamellar matrix to local deformations and, therefore, functions to reinforce the overall tensile response of AF architecture. Moreover, an in vitro study indicated that fibronectin can play a pivotal role in facilitating AF cell attachment and alignment on nanofibers, implying that a synergy between collagen and other noncollagen matrix components may be required to provide AF cells a niche to attain appropriate activities. Altogether, these findings indicate that the AF structure is not just a multilayered fibrous tissue but built with a sophisticated hierarchy of intralamellar and interlamellar supramolecular interactions.

The CEP is involved in attachment of both AF and NP fibers. Using a triphasic model, Hamilton and colleagues also suggested that the CEP is a critical interface for bone-disk integration by providing an adhesive force that is resilient to shear loading. In addition, they reported that CEPs may secrete factors to stimulate proteoglycan and inhibit tumor necrosis factor α production in NP cells, suggesting CEPs have a role in regulating NP homeostasis. CEPs are thought to be similar to articular cartilage and can be artificially engineered by plating and incubating chondrocytes at a high density on the target interface where they secrete matrices to model the interface into a cartilage layer.

Engineering the disk microenvironment

The bulk of the IVD is composed of extracellular matrix, which plays prominent roles in the regulation of the disk cell environment and providing anchorage to disk cells. The extracellular matrix of the IVD, like cartilage, is comprised of mostly collagens, which provide tensile strength of the disk, and proteoglycans, which function to reduce the internal friction in the disk matrix and to distribute load. They also account for the viscoelastic behavior of the IVD, contributing to the shock-absorbing property. Based on the relationship between function and form, materials that can mimic the anatomic architecture and mechanical properties of native IVD (reviewed by Nerurkar and colleagues ) are of interest to disk bioengineers.

Like hyaline cartilage, aggrecan and collagen II are the two main extracellular matrix components of mature NP. Although the NP in mature human IVD is thought to be analogous to articular cartilage, the nature of the matrix and hence their mechanical properties are not exactly the same. NP in young individuals has a high proteoglycan to collagen content, with a suggested glycosaminoglycan-to-hydroxyproline ratio of 27:1, in comparison with a 2:1 ratio in hyaline cartilage. The high proteoglycan content in the NP matrix facilitates the retainment of water, which attributes to the high hydrostatic pressure exhibited in the NP. Hydrogel scaffolds have been commonly used with the intention of simulating the NP microenvironment and entrapment of the newly deposited proteoglycan to facilitate the establishment of hydrostatic pressure. The effectiveness of various hydrogel-based scaffolds, either made from natural hydrophilic biomolecules or synthetic polymers, for NP engineering or repair has been documented. To date, alginate is one of most commonly adopted hydrogel scaffolds for NP cell culturing due to its ease of manipulation, biodegradability, and inert bioactivities. Hyaluronic acid (HA) has been used for the treatment of osteoarthritic knee joints via the direct application into the synovial cavity, and in vivo studies propose that HA or HA-derived hydrogel may facilitate NP function and promote motion segment mechanics. Because the NP plays an important role in withstanding the compressive load so as to maintain disk height and range of motion of spinal segment, however, pure hydrogel scaffolds, which lack confined compressive strength, may not be adequate for NP engineering. Collagen I microspheres and calcium polyphosphate may, alternatively, provide good tensile and compressive strength to support stem cells or NP cells in NP engineering. Recent studies have focused on the generation of collagen-incorporated or polymer-linked hydrogels. These hybrid scaffolds mimic the native microenvironment of the NP to reproduce its viscoelastic and load distribution behavior within the IVD. Although swelling is known to provide the main load-bearing mechanism in the NP, the extensive collagen network inside the disk has also been suggested as supporting a considerable portion of load because the collagen fibril meshwork contributes to the compressive modulus to the tissue. Moreover, collagens have been shown to act as a reservoir of signaling ligands, such as TGF-β1 and bone morphogenetic protein 2 for collagen II, and are able to transduce mechanical signals, therefore serving as important regulators of cell function and homeostasis.

In addition to NP engineering, current research has attempted to use injectable scaffolds as carriers to deliver cells with the aim of salvaging disk degeneration or on its own, as fillers for NP replacement using a minimally invasive approach. Studies have also developed injectable materials that have the ability to self-assemble into a higher-order network, resulting in a solution-to-gel transition. For example, atelocollagen (pepsin-digested collagen) can self–cross-link to form a fibrous meshwork, and chitosan and synthetic peptides have been reported to self-assemble into a nanofiber network. Other natural biomolecules, including hyaluronan and chitosan, when modified with cross-linkable moieties are capable of chain polymerization through photochemical reactions. These materials, as injectable media, may deliver cells of interest by providing a transient framework that prevents leakage of implants and allows for the accumulation of the extracellular matrix deposited by the introduced cells.

Although intradiscal pressure exerted by the NP plays an important role in disk function, it critically depends on the integrity of the AF. In addition, the trans-AF delivery system remains the commonly used way to manipulate the NP in clinical and experimental settings. The construction of AF may facilitate the repair of prolapsed disks and supplement nucleus replacement. Annulus closure devices may possibly provide an effective means to treat prolapsed disks; however, they may restrict segment motion and possibly modify the load distribution in the IVD. AF construction is understandably indispensable to IVD engineering, but effective bioengineering of AF may not be implemented without proper understanding of its microstructure and mechanics.

The AF lamellae are mainly composed of collagen I fiber bundles and have anisotropic mechanical behavior. These bundles are approximately concentric to the lamellae around the NP, where the direction of alignment in one lamella differs to the next by 30°. This angle-ply architecture of lamellae is thought to be designed to resist shear resulting from complex physiologic stresses, such as a combination of axial loading and torsion. Annulus fibers are interconnected via intralamellar crossbridges and interlamellar bridges. In vitro studies showed that excess circumferential constraint may have a negative impact on NP cell metabolic activities, which suggests that tissue rigidity needs to be carefully controlled during AF construction. The rigidity of the AF largely depends on the mechanical properties of the materials used during fabrication. Various materials have been tested for AF tissue construction, including porous silk, polymer nanofibers, polylactide/Bioglass composite, and alginate/chitosan composite. These scaffolds provide a framework of desirable mechanical and bioinductive properties for future AF engineering. An alternative is collagen gel or collagen-glycosaminoglycan composite, although fabrication into a specific geometry (such as fibers or lamella) or characteristic topographic template may be limited. Because AF cells are normally aligned with the lamella fibers, it may be ideal that AF constructs can be engineered through simultaneous controlled placement of AF cells and orientation of the matrix they interact with, such as using scaffolds with specifically designed microgrooves.

Collagen fiber structure determines the mechanical strength and elasticity of the annulus. Recent studies in AF engineering have shed light on some important aspects of its structural properties at the molecular level. Nerurkar and colleagues showed that, through electrospun nanofiber fabrication, a bilamellar tissue model with AF-like angle-ply architecture can be generated. By mechanical testing and modeling, they demonstrated that the bonding between the angle-ply lamellae is crucial to the resistance of interlamellar matrix to local deformations and, therefore, functions to reinforce the overall tensile response of AF architecture. Moreover, an in vitro study indicated that fibronectin can play a pivotal role in facilitating AF cell attachment and alignment on nanofibers, implying that a synergy between collagen and other noncollagen matrix components may be required to provide AF cells a niche to attain appropriate activities. Altogether, these findings indicate that the AF structure is not just a multilayered fibrous tissue but built with a sophisticated hierarchy of intralamellar and interlamellar supramolecular interactions.

The CEP is involved in attachment of both AF and NP fibers. Using a triphasic model, Hamilton and colleagues also suggested that the CEP is a critical interface for bone-disk integration by providing an adhesive force that is resilient to shear loading. In addition, they reported that CEPs may secrete factors to stimulate proteoglycan and inhibit tumor necrosis factor α production in NP cells, suggesting CEPs have a role in regulating NP homeostasis. CEPs are thought to be similar to articular cartilage and can be artificially engineered by plating and incubating chondrocytes at a high density on the target interface where they secrete matrices to model the interface into a cartilage layer.