6
Biology, Mechanics, and Genetics of the Disk: State of the Art
Mauro Alini, Sibylle Grad, Hans-Joachim Wilke, Fabio Galbusera, and Alessandra Colombini
■ Introduction
The human spine has a challenging mechanical role in the musculoskeletal system. Being located in the center of the human body and articulated with the head and the four limbs, the spine is subjected to high loads in all anatomic planes. At the same time, it must ensure a certain degree of flexibility to enable physiological movement and it must adequately protect the spinal cord. One of the key elements to fulfill these mechanical functions are the intervertebral disks (IVD), which account for 15 to 20% of the total length of the human spine, depending on age, degeneration, work load, and diurnal variation. Each IVD acts like a soft pad connecting adjacent vertebral bodies and homogeneously distributing the stresses on the vertebral end plates. Furthermore, together with the spinal ligaments, it governs spine flexibility by enabling sufficient mobility to perform physiological tasks but avoiding excessive movement that would put the spinal cord at risk. These capabilities are made possible by its highly optimized structure, both macroscopically and microscopically.
To fulfill these functions as a joint connecting the vertebral bodies, the IVD is composed of three morphologically distinct tissues (Fig. 6.1). The central part, the nucleus pulposus (NP), is a highly hydrated tissue containing proteoglycan (PG)-rich extracellular matrix (ECM). The NP core is circumferentially constrained by the annulus fibrosus (AF), a fibrous tissue consisting of collagen fibrils arranged in concentric lamellae. In the transition zone or the inner AF, a gradual change in ECM composition ensures the proper integration of the tissues. Located cranially and caudally from the disk, thin layers of hyaline cartilage, the cartilaginous end plates (EPs), function to prevent NP bulging into the vertebrae and to enable the nutrition of the disk cells. The adult human IVD is an avascular organ, with blood capillaries terminating at the outer surface of the AF and the vertebral bodies; diffusion and fluid flow are thus essential for transport of nutrients and waste products into and out of the tissues.
The complex regulation of cellular and extracellular components in the IVD can be disturbed as a consequence of aging, genetic predisposition, and epigenetic or environmental factors. The degenerative cascade, which involves cell senescence, eventually cell death, matrix breakdown, tissue dehydration, and fibrosis, leads to a loss of function of the disk and spinal instability.1 Fissures and tears in the AF, bulging or herniation of the NP, inflammatory processes, and pathological ingrowth of nerves and blood vessels can cause debilitating pain and a severe reduction in quality of life. Genetic and hereditary factors are considered to play a central role as susceptibility factors for developing disk degeneration and back pain. This chapter discusses the fundamentals of intervertebral disk biology and mechanics, the changes occurring during disk degeneration, and the genetic variants that may be involved in the initiation and progression of the degenerative process.
Fig. 6.1 Schematic illustration of the mature intervertebral disk. Midsagittal cross section shows the anatomic regions with matrix composition and representative cell types. CEP, cartilaginous end plate; NP, nucleus pulposus; AF, annulus fibrosus.
■ Biology of the Intervertebral Disk
The template for the development of the spinal segments is provided by the notochord. The early notochord is a rod-like axial structure derived from the mesoderm. Cells within the notochord synthesize PGs that increase the osmotic pressure within cell vacuoles, raising the pressure within the notochord and causing it to elongate and straighten. This forms the basis of the vertebral column. Sclerotome-derived mesenchymal cells condense to a perinotochordal sheath to give rise to the AF. Noncondensed regions of the sheath form the cartilaginous primordial of the vertebrae and the cartilaginous end plate; finally the retained notochord begins to condense to form the NP.2
Nucleus Pulposus
The ECM of the NP primarily consists of PGs, in particular aggrecan and versican, which are large aggregating PGs that contain high numbers of negatively charged glycosaminoglycan (GAG) side chains.3 Specifically, aggrecan contains chondroitin sulfate and keratan sulfate side chains and binds to hyaluronan, forming large molecules entrapped in a collagen network. The high concentration of fixed negative charge provides the NP with a substantial osmotic potential that, by pulling in water, confers the high hydration of the NP. In a young healthy disk, the water content reaches 90%, whereas this proportion generally drops to 70% in older adults. Besides aggrecan and versican, smaller PGs are also found in the NP, including biglycan, decorin, and fibromodulin that function to organize the ECM structure, facilitate cell signaling, and bind growth factors. About 20% of the ECM in the NP consists of collagenous molecules, primarily collagen type II, although collagen types III, V, VI, IX, and XI are also present. Different from articular cartilage, the collagen fibrils appear distributed in a random way, and the ratio of PG to collagens, measured as GAG to hydroxyproline, has been documented as 27:1 in the young NP, as compared with 2:1 in hyaline cartilage end plate.4
The cell population of the early postnatal NP mainly consists of large (30-40 μm) notochordal cells that contain intracellular vacuoles and are commonly arranged in clusters. In humans, notochordal cells are lost within the first decade of life and are replaced by mature NP cells. Although the cells of the mature NP have originally been described as chondrocyte-like due to their similar morphology and ECM synthesis, large-scale gene expression studies have provided more in-depth insight into their phenotype.5 Phenotypical features of mature NP cells that may distinguish them from other cell types reflect their adaptation to the hypoxic, acidic, nutrition-deprived microenvironment, such as expression of hypoxia inducible factor (HIF)-1α , carbonic anhydrase 12, and glucose transporter proteins, their development from the notochord, such as expression of brachyury, cytokeratins 8, 18, and 19, and PGs; and their secretion of a specific ECM molecule composition with a high PG/collagen ratio. In addition, CD24 has been suggested as a marker of mature NP cells.6
Both the density (around 4,000 cells/mm3) and the metabolic activity of NP cells are low compared with other cartilaginous tissues,7 which may be one reason for the limited regenerative potential of the NP. However, several recent reports have described the presence of cells with stem or progenitor characteristics in the human adult NP.8 These cells were shown to express typical surface markers of mesenchymal stem cells and to have the ability to differentiate into osteogenic, chondrogenic, and adipogenic lineages. In particular, a progenitor cell population positive for the angiopoietin receptor Tie2 and disialoganglioside 2 (GD2) has been identified in the NP.9 These cells form spheroid colonies, are clonally multipotent, and maintain their growth and differentiation potential after in vivo transplantation. A significant decrease in the number of these NP progenitor cells was observed with aging and degeneration of the human disk, indicating impaired capacity for regeneration.
Annulus Fibrosus
During intervertebral development, while the notochord differentiates to generate the NP, it pushes against the surrounding annular condensations, thereby inducing the formation of the inner and outer AF. The outer AF is composed of dense concentric lamellae with type I collagen fibers lying parallel within lamellae, whereas the inner part of the AF is made of more widely spaced layers with higher amounts of type II collagen and PG. A network of elastin fibers is also present between the layers. The large aggregating PGs aggrecan and versican, and the small interstitial PGs decorin, biglycan, fibromodulin, and lumican compose 10 to 20% of the AF dry weight.5 Similar to the NP, the roles of the GAG substituted PGs are tissue hydration enabling rapid reversible deformation, whereas binding of PG to collagens, growth factors, and other matrix components plays a role in the ECM assembly and in repair processes.
With ~ 9,000 cells/mm3, the cellularity of the human AF is higher compared with the NP.7 AF cells appear fibroblast-like with spindle-shaped morphology,10 although spatial variations exist depending on the local mechanical situation. Similarly, regional differences exist with respect to the AF cell phenotype.5 Outer AF cells primarily synthesize type I collagen, whereas the more chondrocyte-like cells toward the inner AF increasingly produce type II collagen. Without distinguishing regional variations, the GAG-to-hydroxyproline ratio in human lumbar disk AF is around 1.6:1, with little variation by age or by extent of degeneration.4
Expression ratios of matrix proteins are effective to distinguish between cellular phenotypes in IVD and cartilage tissues. Collagen II/aggrecan and collagen II/collagen I ratios have been reported highest in cartilage, with collagen II/aggrecan higher in AF than in NP, and collagen II/collagen I higher in NP than in AF cells. In addition, the greatest expression of collagen V was found in AF cells, suggesting that this collagen might be considered as an AF cell marker. Microarray studies have revealed expression profiles of AF cells (reviewed by Pattappa et al5 and Guterl et al11), although challenges remain to characterize the AF phenotype due to high local and species-dependent variations. Ten genes with AF/NP intensity ratios ≥ 10, including the PG decorin, were reported in rat cells. In canine, 77 genes with a NP/AF signal log ratio of ≤ -1 were recognized, among which were collagen XIV, cell adhesion molecules, and integrin precursors. Tenomodulin, a member of the small PG family, was found to be increased in bovine and human AF compared with NP and cartilage cells. Yet the difficulty of determining clear cutoffs between AF and NP cells has given rise to the assignment of certain molecules as general IVD markers.
The presence of cells with stem or progenitor cell characteristics has also been described in the AF of different species, including human healthy and degenerative tissue.8 Although this finding confirms the general intrinsic healing potential, different circumstances may limit the effectiveness of a repair response. These factors include the decline in the number of progenitor cells and increased cell senescence with aging, potential traumatic lesions, systemic diseases, and genetic associations.8
Cartilaginous End Plate
The end plate consists of a thin layer of hyaline cartilage that has its maximal thickness at birth and becomes thinner with age. In the adult, the EP width is 0.5 to 1 mm. Different functions can be assigned to this cartilaginous structure; it is a mechanical barrier that prevents the IVD from applying pressure directly onto the bone, contributes to load distribution toward the vertebrae, and plays a role in preserving the nutrition and vitality of the NP cells. As with articular cartilage, the main component is water (close to 80% at birth, < 70% in adults), and the chondrocytic cells are embedded in an aggrecan and type II collagen-rich ECM.5 The cell density is around 15,000 per mm3, and the PG to collagen (expressed as GAG to hydroxyproline) ratio has been measured as 2:1,4 which also approximates the levels of articular cartilage. The EP transitions into bone through a region of calcified cartilage.
The EP represents the main route by which nutrients diffuse into the NP. The exchange of solutes is ensured by capillaries present in the calcified part of the EP. It has been reported that the central region enables the highest diffusion of small molecules, whereas at the tissue periphery the cartilage is less permeable. However, the molecule size and charge play an important role, such that small molecules like glucose and oxygen can migrate through the disk more easily than large molecules (e.g., proteins).7,12
Cells with stem cell characteristics similar to bone marrow stromal cells were also identified in the end plate of degenerative human disks. In addition, stem cell niches have been described in tissues surrounding the IVD.8 In particular, populations of slow cycling cells were detected in the AF border to ligament zone and the perichondrium of rabbit, porcine, and human IVDs; moreover, cell migration routes from these niches toward the IVD were described.8,13
Intervertebral Disk Microenvironment
One striking aspect of IVD biology is that the cells of the NP and the inner part of the AF are removed from the vascular system. Capillaries of blood vessels supplying the vertebral bone terminate at the end plate, traversing just its superficial region. Similarly, only small numbers of capillaries are found at the very outer surface of the AF.14 As a consequence, biochemical analyses and modeling studies have reported that the oxygen tension within the IVD is substantially reduced, and the cell metabolism is essentially anaerobic. It has been found that disk cells adapt to this hypoxic environment by limiting the consumption of oxygen and by consecutive stabilization of the transcription factor HIF-1α.15 The anaerobic metabolism also brings elevated lactic acid and low pH conditions, posing further challenges for cell survival.
Another important feature that characterizes the disk specific environment is the increased osmolarity. The high osmotic pressure, reaching values up to 200 mOsm/kg above the norm, is important to resist the axial loads acting on the spine. Specific factors, such as the transcription factor tonicity enhancer binding protein (TonEBP), are expressed by the disk cells to regulate the levels of nonionic osmolytes for the maintenance of the osmotic properties of the cytosol, which is critical for cell survival.16 Due to the avascular nature of the IVD, residual fragments of the ECM turnover tend to accumulate in the disk matrix rather than diffusing into the circulation.17 With aging, the levels of these fragments gradually increase. However, in the case of aggrecan fragments, the negatively charged GAG chains continue to contribute to the swelling pressure as long as they are retained in the NP.
Degenerative Processes
Intervertebral disk degeneration (IDD) is a complex multifactorial process that is determined by the interaction of genetic and environmental factors. Hallmarks include altered matrix composition, overall degradation of matrix components, and changes in cell numbers, cell phenotype, and metabolic activity. Although similar changes are observed during the normal aging process, their pathological acceleration and augmentation can lead to painful debilitating conditions. Morphologically, degenerative changes are evident as a reduction in disk height, disk bulging, and loss of NP/AF demarcation. These changes result from an increased collagen I/collagen II synthesis ratio of the NP cells, loss of sulfated GAGs, subsequent dehydration of the NP, and disruption of collagen orientation in the AF. There is also an increase in collagen cross-linking, rendering the tissue stiffer and more prone to rupture.11 The PG composition is changing, showing an overall loss of aggrecan, combined with a shift toward versican, biglycan, and decorin production. An increase in nonenzymatic glycosylation can lead to the production of advanced glycation end products (AGEs) that can cause further tissue stiffness. Enhanced fibronectin production and fibronectin fragment accumulation has also been described during degeneration, further accelerating matrix breakdown.1,3,18
Matrix degradation is primarily mediated by a variety of proteolytic enzymes, in particular members of the matrix metalloproteinase (MMP) and a disintegrin and metalloprotease with thrombospondin motifs (ADAMTS) families.18 Several MMPs, such as MMP1, 3, 7, and 13, have been shown to be upregulated in degenerative disk tissue. Similarly, the expression of the aggrecanases ADAMTS1, 4, 5, 9, and 15 is increased with degeneration. In addition, cathepsins and high temperature requirement serine protease A1 (HTRA1) are involved in IVD matrix turnover. All these enzymes are regulated by soluble mediators, such as anabolic, catabolic, and inflammatory factors, and by tissue inhibitors of metalloproteinases (TIMPs), which function to inhibit MMPs and ADAMTSs.
Both nerve and vessel ingrowths into the IVD occur during the degenerative process. This innervation and vascularization may be triggered by the loss of PG, annular clefts and tears, or soluble angiogenic and neurotrophic factors, such as pleiotrophin, vascular endothelial growth factor (VEGF), nerve growth factor (NGF), and brain-derived neurotrophic factor (BDGF), which have been identified in degenerative and painful disks.19
A range of proinflammatory factors are induced or upregulated during disk degeneration.20 These include interleukins (ILs) (predominantly IL-1), tumor necrosis factor-α (TNF-α), prostaglandin E2 (PGE2), nitric oxide, and interferon-γ (IFN-γ), although other pathways such as IFN-α signaling may also contribute to the process. Inflammatory mediators can trigger the expression of the above-described proteases, angiogenic and neurogenic molecules, thereby accelerating the degradation. In addition, chemotactic and anabolic factors may induce cell attraction and differentiation, initiating a repair response. However, the diminished capacity renders this response insufficient in the adult IVD.8
■ Mechanics of the Intervertebral Disk
The combination of the three anatomic structures, the NP, the AF, and the vertebral end plates, determines a highly uniform stress distribution inside the disk and vertebral bodies, thus minimizing the risk of failures and disruptions even under the action of high loads. The disk nucleus is a gel-like material with high water content, ranging up to 90% in the lumbar spine. Its PG content gives it the capability to attract water molecules and therefore to create an osmotic pressure gradient with respect to the external environment. Previous studies found that the intradiskal pressure is almost hydrostatic, that is, it acts with equal magnitude in all directions,21 and is therefore orthogonally and homogeneously transmitted to the inner surface of the AF. The annulus has also a rather high water content (50-70% in the lumbar spine), but mainly differs from the nucleus in its highly organized network of collagen fibers rather than the unstructured, jelly appearance of the NP. Similar to fiber-reinforced tanks designed to contain fluids under pressure, the concentric layers of collagen fibers of the annulus limit the nucleus bulging that would result from the intradiskal pressure. In addition, the criss-cross pattern of the fibers in combination with a complex three-dimensional network connecting the adjacent fiber layers enable a certain degree of disk bending without risking excessive local strains and possible disruptions. Finally, the vertebral end plates ensure a strong yet elastic connection between the intervertebral disk and the adjacent vertebral bodies. End plates consist of a thin layer of hyaline cartilage adjacent to a layer of semiporous subchondral bone, which exhibits several small marrow contact holes enabling nutrient supply and waste removal.22
Structural Integration of the Nucleus, Annulus, and End Plates
A recent series of reports highlighted a complex structural integration between the components of the intervertebral disk. Indeed, a complex three-dimensional fiber network exists both at the nucleus-annulus interface23 as well as at the frontier between the disk and the vertebral end plates.24,25 Microstructural analyses of ovine specimens found that the nucleus has a distinct fibrosity, mostly vertically oriented, which extends into the inner annulus and is connected to its collagen fiber structure.25 This fibrosity is capable of sustaining nonnegligible tensile loads (up to 30 N for the whole nucleus) and covers the whole structure, from end plate to end plate. Despite its disordered appearance, mechanical tests revealed that the nucleus fibrosity is highly structured. Fibers appear highly convoluted and folded in the physiological state, and progressively unfold with increasing tensile load. In addition to the vertical fibrosity integrating the nucleus with the end plate, horizontally aligned fibers connecting the nucleus with the inner annulus were also revealed by microstructural analysis and mechanical testing.23
The integration between the annular fibers and the end plates was also found to exhibit a high degree of complexity.24 To withstand high loads, both axial and shear ones, and taking into account the thinness of the cartilage end-plate layer and therefore the difficulties in achieving a strong connection, fiber bundles branch into multiple sub-bundles, named “leaves” by the authors, which have the function of increasing the area of interaction between the annular fibers and the end plate. By exploiting this optimized structure, the annulus-end-plate interface proved to be able to sustain loads with magnitude comparable to those that may disrupt the annular fibers themselves.
Spinal Loads
The organized and optimized structure of the intervertebral disk evolved as a strategy to minimize the risk of mechanical failure while supporting the high loads acting in the spine. As a matter of fact, the exact magnitudes and directions of the loads active in the spine are technically challenging to measure and remain only limitedly known. In vitro and computational studies hypothesized a compressive load ranging from 400 to 500 N acting in the lumbar spine in the standing posture. The load was supposed to be aligned with the longitudinal axis of the spine and to follow its curvature, and was therefore defined as a follower load. This hypothesis was tested by means of numerical simulations and proved to be a simple but rather reliable estimation of the spinal loads in standing.26 However, more sophisticated computational studies explicitly simulating the action of the trunk muscles, predicted a more complex loading environment even for the standing posture, in which shear loads and bending moments should not be neglected especially in the lower lumbar spine.27
Higher loads are acting in the spine in other postures and movements, such as in forward bending, and are aggravated if external loads are also present, for example by lifting a weight. By combining in vitro tests and measurements conducted on patients implanted with telemeterized spinal fixators, Wilke et al28 estimated a force generated by the erector spinae of 520 N for a forward flexion of 30 degrees, and of 130 N for an extension of 15 degrees. Regarding axial rotation, the best match with in vivo data was predicted for a compressive load of 720 N combined with a moment of 5.5 Nm (newton meters).29 As a matter of fact, the loads mentioned above are for the most part sustained by the intervertebral disk. Despite spinal ligaments and facet joints also being subjected to remarkable loads, it was found that 80 to 90% of the compressive load acting in standing is supported by the intervertebral disk. In vitro studies based on stepwise reduction of a lumbar functional spine unit demonstrated that the disk is the structure responsible for most of the motion restriction also in flexion-extension and axial rotation,30 thus arguably supporting for the most part the spinal loads also in other postures and during daily activities.
Stresses in the Intervertebral Disk
In vivo evaluation of the stresses acting in the intervertebral disks has been performed by means of pressure transducers implanted in living subjects (Table 6.1). The technique was pioneered by Nachemson and coworkers21,31,32 and used in several later studies.33–35 Wilke et al34 analyzed the intradiskal pressures in the lumbar spine during a wide range of daily activities and physical exercises (Fig. 6.2). High values up to 2.3 MPa (megapascals) were recorded during physically demanding tasks such as weight lifting. Similar results were obtained using the same measuring technique in the thoracic spine.35
Adams and coworkers36 measured the stress profile in cadaveric spine specimens subjected to various loads by means of linear transducer elements mounted on thin needles. With this technique, the whole stress profile in the direction of the needle can be estimated, thus providing richer data with respect to the punctual measurements obtained with traditional pressure transducers. Stress profilometry conducted on specimens with no signs of disk degeneration revealed a uniform distribution of the stresses in the whole disk, and confirmed the hydrostatic nature of the stress in the NP.
