Genetic Risk Factors for Lumbar Disk Disease

Musculoskeletal disorders such as lumbar disk disease are among the most common conditions for which patients seek medical care. Although clinical studies have provided insights into disease prevalence and clinical management, recent research advances have yielded understanding about the molecular mechanisms that may be involved and the genetic mutations that may contribute to a variety of musculoskeletal and connective tissue disorders.

Members of the collagen protein family are the most abundant structural components of the extracellular matrix of connective tissues.¹ In bone, cartilage, tendon, and skin the collagens are organized into ropelike fibrils, called heterotypic because they are composed of multiple collagen types. Collagen types differ in length and chain composition, but all have in common a triple-helical structure and a primary sequence composed of uninterrupted repeats of the sequence Gly-X-Y, where gly is glycine, X is often proline, and Y is often hydroxyproline. Type II collagen is the most abundant collagen of cartilagenous tissues and is referred to as the major collagen. It forms heterotypic fibrils with the successively less abundant (hence "minor") collagens, types IX and XI. The collagenous fibrils provide the strength necessary for tissues to resist tensile forces. In cartilage, where resistance to compressive forces is also a major issue, the tissue also has a significant proportion of proteoglycan.

The structural components of the extracellular matrix are good candidates for the causative molecule in a variety of relatively rare genetic disorders, as well as more common musculoskeletal conditions that have a genetic component. Disease-causing mutations in types II and XI collagen have been demonstrated in a number of chondrodystrophies.^{1 - 2} Most mutations involve substitutions of another amino acid for one of the glycine residues that occur at every third position along the collagen helix. These glycine residues are crucial for helix formation because their small side chain can be accommodated in the constricted space inside the helix. Other causative mutations result in exon skipping defects or premature chain termination.

In this issue of THE JOURNAL, Paassilta and colleagues³ focus on the other minor collagen of cartilagenous tissues, type IX. Type IX is a heterotrimer composed of 3 distinct chains: α1, α2, and α3. Unlike collagen types I, II, III, V, and XI, which have a long uninterrupted collagenous domain, type IX collagen is a FACIT collagen—that is, a fibril-associated collagen with interrupted triple helix.⁴ Type IX chains have 3 collagenous regions, labelled COL1 to COL3 from carboxyl to amino end of the chain, interspaced with 4 short nonhelical domains. The interrupted structure of type IX collagen makes possible its role in matrix. Type IX collagen does not form fibrils on its own but anchors the interstitial collagen to other components of matrix. Most of each type IX collagen molecule, including the COL1 and COL2 domains, lies flat on the surface of the type II collagen fibril, to which it is covalently cross-linked. But one of the noncollagenous domains functions as a molecular hinge and allows a globular domain and the helical COL3 domain at the amino end to protrude from the surface of the heterotypic fibrils and be available for binding to other matrix proteins.

A small number of disease-causing mutations in type IX collagen have been described in murine models and in humans. A murine strain lacking the α1 chain of type IX collagen developed progressive degenerative osteoarthritis.⁵ Transgenic mice expressing type IX collagen with a large deletion covering part of the COL3 domain and most of the central COL2 domain had osteoarthritis and progressive degeneration of intervertebral disks.⁶ In patients, the chondrodystrophy multiple epiphyseal dysplasia is characterized by pain and stiffness in large joints, sometimes including osteoarthritis, mild short stature, and stubby fingers. Investigators have demonstrated a splicing defect involving skipping of a single exon (exon 3) coding for 12 amino acids in the protruding COL3 domain in either α2(IX)^{7 - 8} or α3(IX)^{9 - 11} in a small number of families.

The research team led by Ala-Kokko has engaged in an intensive effort to find other significant changes in the sequences of the type IX collagen chains in patients with disorders involving cartilagenous tissues, such as chondrodystrophies, osteoarthritis, rheumatoid arthritis, and lumbar disk disease (LDD). For each patient, the investigators conducted an exhaustive exon-by-exon examination of the genes coding for the 3 chains of type IX collagen using conformation-sensitive gel electrophoresis (CSGE) followed by sequencing of products with altered gel migration pattern in CSGE. The CSGE technique is very sensitive for detection of sequence variation and has proved its worth in mutation detection in other collagen genes.¹² No changes typical of those known to be disease-causing mutations in collagen genes were found in this convincingly thorough search.

However, the search did reveal 2 amino acid substitutions that are significantly more prevalent in patients with lumbar disk disease than in healthy controls. The first substitution, already reported by this team, occurs at residue 326 of the α2(IX) chain and causes the substitution of a tryptophan residue for the usual glutamine at that position.¹³ This Trp2 allele was found in 7 patients with LDD but not in any healthy individuals. Linkage studies showed that family members who had inherited the Trp2 allele had magnetic resonance imaging (MRI) or computed tomography scans with intervertebral disk abnormalities. The second substitution, reported in this issue of THE JOURNAL, occurs at position 103 of the α3(IX) chain and is also a substitution to a tryptophan, in this instance from the usual arginine residue.³ The Trp3 allele had a frequency of 12.2% in 171 LDD patients and 4.1% in 186 healthy controls. Unfortunately, no pedigree linkage studies were reported with the Trp3 allele in LDD patients or controls. Still, the significantly higher frequency of the Trp3 allele in LDD patients vs controls makes it a credible genetic factor for a 3-fold increase in the risk of symptomatic LDD in the Finnish population.

These steps toward understanding the genetic modifiers of LDD are significant because intervertebral disk disease is among the most common musculoskeletal disorders. Its origins are multifactorial, including occupational physical stress, smoking, taller height, and obesity.¹⁴ There is also a clear genetic component as observed in twin studies and with familial aggregation of lumbar disk herniation.^{15 - 16} However, it is difficult to exactly define the LDD patient group, because more than a third of asymptomatic people have bulges or protrusions at more than 1 disk level on spine MRI.¹⁷ The study by Paassilta et al has the advantage of being performed using a relatively homogeneous Finnish population. Patients with LDD were defined by having at least 1 month of sciatica as well as by findings on physical examination and MRI. The healthy control population was defined simply on the basis of being asymptomatic, without a physical examination or MRI demonstration of normal disks at all levels. If the 15 asymptomatic controls with the Trp3 allele have multilevel disk protrusions on MRI, then the role of the Trp3 allele as a risk factor for LDD may be even greater than proposed. Conversely, if these 15 controls have normal disks at all levels, then a less directly causal risk is implied.

For the 2 substitutions in type IX collagen, it is possible to speculate as to how a residue that is not normally found in type IX chains might modify function. In the collagenous Gly-X-Y triplet pattern, the Trp2 substitution is in an X position and the Trp3 substitution is in a Y position. A handful of patients have been described with changes in X and Y position residues in types I, II, and III collagen, which have been postulated to modify phenotype.¹⁸ Residues in the Y position are more exposed to the solvent surrounding the helix than to the center of the collagen helix. When the X and Y residues are not proline or hydroxyproline but have a charged side chain, they play an important role in helix self-association and ligand binding by forming ionic bonds mediated by water molecules.¹⁹ In the Trp3 allele, the substitution occurs in the COL3 region, which protrudes out from the heterotypic fibrils and is available for ligand binding. So the substitution of a less soluble tryptophan, with its aromatic side chain, for the usual charged arginine residue could interfere with binding to an as-yet-unidentified matrix molecule and weaken the resistance of the intervertebral disk to compressive forces. The substitution identified in the Trp2 allele is also a tryptophan for a charged residue. Furthermore, it occurs near the region of the type IX helix involved in cross-linking to type II collagen in the heterotypic fibrils. Paassilta et al³ cite unpublished data suggesting altered binding between type II collagen and type IX collagen with the Trp2 substitution. This could also weaken the matrix of the intervertebral disks.

This thorough molecular study of type IX collagen is deserving of both clinical and biochemical follow-up. The prevalence of the Trp alleles can be determined in other ethnic populations, including comparisons of prevalence among asymptomatic persons with normal findings on spine MRI, asymptomatic persons with abnormal findings on spine MRI, and symptomatic patients. Family linkage studies in symptomatic and asymptomatic persons with the Trp3 allele will aid in understanding the risk level involved. From the laboratory, knock-in mice with each of the Trp alleles would demonstrate whether the locus was directly contributory to LDD. Moreover, while the Trp alleles together occurred in about 16% of Finnish LDD patients, other loci are almost certainly involved. A prime candidate among them encodes aggrecan, a highly abundant proteoglycan in cartilage whose extensive hydration contributes to resistance to tissue deformation. Knock-out mice for aggrecan have a high prevalence of vertebral disk herniation and degeneration.²⁰

Although there has been much progress in understanding the role of collagen and mutations in the genes coding for collagen in common musculoskeletal disorders, further investigation of the interaction between genetic and environmental factors such as physical stress are needed. A more complete understanding of risk factors and their mechanisms will be essential to rational preventive and symptomatic care.