CHAPTER 270 Concepts of Disk Degeneration and Regeneration
Conceptually, disc degeneration is a product of lifelong degradation with synchronized remodeling of discs and neighboring vertebrae, including simultaneous adaptation of the disc structures to changes in physical loading and responses to the occasional injury.1
Early degenerative changes should refer to accelerated age-related changes.…Degenerative disc disease should be applied to a degenerate disc that is also painful.2
Much of the published literature concentrates on degeneration in the lumbar region, and it remains to be seen whether the same concepts apply to other regions. It should be emphasized that disk degeneration and disk herniation are closely related but not synonymous. Conceptually, degeneration is required to allow herniation to occur. It is well accepted that a herniated disk may cause pain and neurologic deficit. The role of disk degeneration in spine symptomatology is less clear. Although it is believed that disk degeneration may contribute to low back complaints, accurate identification of an abnormally or symptomatic degenerative disk has been difficult.3 The ability to do so will be essential for the successful application of potential regenerative technologies.
The financial burden of spinal disorders is considerable. The annual spine-related health care expenditure in the United States for 2005 was estimated to be $86 billion, in the same realm as cancer care ($89 billion) and accounting for 9% of all health care spending.4 However, the indirect costs associated with spine disorders may be much greater.5 If disk regeneration is possible and proves safe and efficacious, the potential impact on population health would therefore be significant.
Composition of a “Normal” Disk
The intervertebral disk consists of three components. The central, gelatinous nucleus pulposus is contained by the multilaminated annulus fibrosus. The superior and inferior aspects of the disk are covered by cartilaginous vertebral end plates (Fig. 270-1). Significant changes in the anatomic structure and biochemical composition of these elements occur in development and continue throughout life. It is therefore difficult to define an archetypal normal disk. The generally accepted description of a healthy young adult disk is as follows.6 The nucleus pulposus consists of scattered cells and thin collagen fibrils dispersed in a proteoglycan-rich ground substance. The nucleus has a very high water content because of the hydrophilic nature of the proteoglycans. The annulus fibrosus consists of concentric lamellae of collagen fibers and a small amount of elastic fibers bound together by a lesser amount of proteoglycan gel. Within a lamella, the obliquely oriented collagen fibers are parallel to each other, but adjacent lamellae alternate the orientation of their fibers. The fibroblast-like cells of the annulus fibrosus are spindle-shaped and extend along the collagen fibrils, whereas the cells of the nucleus pulposus appear rounded and have been likened to chondrocytes.7 The outer lamellae of the annulus fibrosus insert into the circumferential bony ring apophysis of the upper and lower vertebral surfaces. The fibers of the inner lamellae are continuous with those of the vertebral end plate and form an envelope around the nucleus pulposus. The vertebral end plates are mostly fibrocartilage, with lesser amounts of hyaline cartilage found nearer the vertebral bodies.
Nerve fibers have been identified in the outer layers of the annulus fibrosus. The sources of these fibers in the anterior and lateral aspect of the annulus are direct branches from the ventral rami of the spinal nerve roots and the gray rami communicantes of the sympathetic trunk. The posterior annulus receives fibers from the sinuvertebral nerve, which forms from the union of an autonomic root from the gray rami communicantes and a somatic root from the ventral rami.8–10
The adult disk is avascular and relies on nutrients from adjacent vertebrae to maintain its living cellular constituents. Hence, the metabolic needs of the disk are met passively by diffusion through the adjacent bony and cartilaginous end plates, as well as through its own matrix. This leads to a marked concentration gradient in the disk, with lowest oxygen tension, glucose concentration, and pH in the center. Disk chondrocytes are well adapted to this environment, with maximal proteoglycan production occurring at 5% oxygen and a pH of 7.0.11
The dominant proteoglycan in disk matrix is aggrecan, which consists of a core protein conjugated with many keratan sulfate and chondroitin sulfate molecules. Keratin sulfate and chondroitin sulfate provide the high anionic charge densities needed for the osmotic properties of aggrecan. Through interaction with hyaluronan and link protein, aggrecan forms large aggregates and functions to retain water in the nucleus.12 Nonaggregating proteoglycans are also found in the disk matrix and include small leucine-rich repeat proteoglycans, fibromodulin, lumican, decorin, and biglycan. They may have a role in the mechanical behavior of the disk through their interaction with fibrillar collagen.13
Collagen in the disk is predominantly of the type I and type II varieties. Both are found in close to equal amounts in the annulus, whereas type II predominates in the nucleus.14 In progressing from the outer annulus inward, the ratio of type I to type II decreases. Multiple other collagen types are also found in disks, including types V and XI (as hybrid fibrils), types IX, XII, and XIV (noncovalently bonded to fibril surfaces), and types VI and III (pericellular).15
Disk Degeneration
Morphology
The macroscopic appearance of lumbar disk degeneration has been classified into stages defined by the three components of the disk (nucleus, annulus, and end plates) from the perspective of a midsagittal section by Thompson and coworkers.16 The mean grade of disk degeneration increases with age, but a wide variation in grades can be appreciated within any given age group. The stages are described in general terms as I, normal juvenile disk; II, normal adult disk; III, mild disk degeneration; IV, moderate disk degeneration; and V, severe disk degeneration (Table 270-1).17
The temporospatial course of degeneration has been further detailed in cadaveric studies of lumbar spine specimens at the macroscopic18 and histologic17 levels (Fig. 270-2). The degenerative process begins in the first 2 decades of life, plateaus in the third and fourth decades, and then accelerates in the fifth to seventh decades and beyond; however, within the gross categories of age and Thompson macroscopic grade, substantial variation in the degree of histologic changes can be seen. The foci of abundant notochordal cells seen in fetal specimens disappear by the end of the first decade and are replaced by chondrocytes, whose proliferation rates, cell density, and death increase further with advancing age. True degeneration begins in the nucleus with fibrous transformation, mucous degeneration, and granular changes. The subsequent development of jagged empty spaces in the nucleus (called nuclear clefts) precedes the three types of annular tears—concentric, radial, and rim tears. In specimens of advanced age, the appearance of scar-like tissue formation and the development of large tissue defects impede appreciation of the disk’s morphology.
Underlying these macroscopic and histologic changes is deterioration of the biochemical composition of the disk. Three phases of matrix turnover have been described.19 The first phase (growth) is characterized by active synthesis of type II collagen and aggrecan. The second phase (maturation) consists of a reduction in the synthesis of both these constituents. In the third phase (degeneration and fibrosis), denaturation of type II collagen occurs and synthesis of type I collagen is increased. No change occurs in the low levels of aggrecan synthesis. Overall, the total proteoglycan and water content of the disk decreases with increasing age and morphologic grade,19 and the ratio of aggregating proteoglycans to total proteoglycans decreases. The relative increase in nonaggregating proteoglycan molecules is due to an accumulation of aggrecan fragments as a result of proteolysis.20 The proteolytic aggrecanases found in the intervertebral disk are produced by enzymes from both the matrix metalloproteinase (MMP) and ADAMTS (a disintegrin and metalloproteinase with thrombospondin motifs) families, each with specific sites of action.21
There is evidence that aggrecan breakdown is under the influence of a variety of signaling molecules, including growth factors—insulin-like growth factor I (IGF-I), basic fibroblast growth factor (bFGF), platelet-derived growth factor (PDGF), transforming growth factor-β (TGF-β)—some bone morphogenic proteins (BMPs), and the cytokines interleukin-1 (IL-1) and IL-6, and tumor necrosis factor-α (TNF-α).21 Central to the interplay between growth factors and aggrecanases are the cells of the intervertebral disk. They control the balance between anabolic and catabolic processes in the matrix. As mentioned earlier, cells of the intervertebral disk rely on diffusion for delivery of nutrients and removal of metabolic waste. Interference or degradation of these pathways, particularly at the subchondral bone–disk junction, may lead to a decline in cell function, loss of growth factor production, and activation of aggrecanase, all likely to be key events in the pathogenesis of disk degeneration.22
Imaging
Macroscopic, histologic, or biochemical examination of the disk is limited to cadaveric specimens and is therefore not useful in the clinical or epidemiologic study of disk disease. Instead, medical imaging studies, including both plain radiography and magnetic resonance imaging (MRI), have been used as a surrogate measure of disk degeneration. On plain radiography, the intervertebral disk itself is radiolucent, and assessment of degeneration depends on measurement of loss of disk height, as well as the effect on the adjacent vertebrae in terms of osteophyte formation and end plate sclerosis. MRI allows direct assessment of the disk itself, including hydration and the degree of desiccation. Such factors have been combined and validated in several proposed grading scales.23,24
The most accepted MRI classification is the system described by Pfirrmann and colleagues (Table 270-2).25 It is based on the macroscopic Thompson grade and relies on the T2-weighted appearance of the nucleus and annulus, as well as disk space height. T2 disk signal characteristics on MRI have been correlated with both proteoglycan and water content.26,27 Changes in the appearance of the vertebral end plates were previously described by Modic’s group and are now synonymous with his name.28 Type I changes are hypointense on T1 sequences and hyperintense on T2. They correlate with histologic evidence of fibrovascular replacement of the subchondral marrow. Type II changes are T1 and T2 hyperintense and correlate with fatty deposition in the end plate zones. The later described type III changes are T1 and T2 hypointense and correlate with end plate sclerosis on plain radiographs.29
A combined imaging score incorporating elements from radiographic and MRI examination has been developed and correlated with morphologic and biochemical variables (Fig. 270-3).27 The score correlates well with macroscopic grade in advanced degeneration but fails to discriminate well between lower grades, thus suggesting that routine imaging techniques lack the sensitivity needed for early detection of changes. This lack of sensitivity has led to investigation of specialized MRI techniques. End plate diffusion patterns, assessed by the change in signal intensity of the disk space after gadolinium enhancement, have been used to identify end plate changes, but these sequences are time-consuming to obtain.30 High-resolution magnetic resonance spectroscopy has been used to examine the biochemical spectra of intervertebral disks from cadavers but has not been reported in the clinical setting.31,32 Showing great clinical promise is the quantitative MRI spin-lock technique T1ρ, which has been correlated with proteoglycan content and macroscopic grade in cadaveric studies33,34 and has been shown to be feasible for in vivo studies (Fig. 270-4).35,36
