An association between the GDF5 SNP rs143383 and LDD in women is an important finding and is consistent with the known association of the T allele with OA (17
). LDD is common and known to be a cause of low back pain, a condition that results in great social costs. It has been established that LDD is highly heritable, and some genetic variants have been identified (for review, see ref.1
), but much of the variation remains to be accounted for. The finding of an association with the combination of disc space narrowing and osteophytes, but only in women, may reflect the sample size or may indicate that the effect is detectable only in the most severely affected. Sex-specificity of the effects of genetic variants is well recognized in OA.
GDF5 is a proven variant in peripheral joint OA; this has been demonstrated in a number of populations. Furthermore, it is associated with skeletal height, suggesting mediation by either disc or vertebral body height or long bone length (8
). GDF5 was selected as an OA candidate gene initially because phalangoepiphyseal dysplasia patients develop severe, early-onset hip OA. Miyamoto et al (5
) performed a case–control association study of a number of SNPs around GDF5 and found the most significant to be rs143383. This association has been shown for both hip and knee OA in both Japanese and Han Chinese populations and, more recently, confirmed in European populations in 2 published meta-analyses (6
). Since the T to C mutation at position +104 lies in the promoter region of GDF5, the hypothesis is that the variant influences GDF5 transcription, with individuals with OA having reduced transcriptional ability.
This study has a number of strengths and weaknesses. First, it combines a number of well-characterized cohorts that have been subjected to LDD phenotyping, albeit using different methods. The study of degenerative disease of the disc lags behind the study of hip, knee, and hand OA. In addition, LDD suffers from a lack of standardized epidemiologic definitions, which hampers attempts to study it in a systematic way. The use of twins and any relatedness within the other cohorts was taken into account. This study had 80% power to detect an OR for the T allele of 1.19 or higher for narrowing and 1.15 for osteophytes in the total sample size with P < 0.05 (assuming an additive genetic model and T allele frequency of 60%, using Quanto software version 1.2.4). This is further supported by the results of the false-positive report probability, suggesting that the results are more likely to be true positive than false positive.
We investigated whether the 5′-untranslated region SNP rs143383 GDF5 polymorphism was associated with LDD using 5 cohorts of European descent and found evidence of association in women. This SNP is thought to result in reduced transcription of GDF5, a prochondrogenic growth factor which has been shown to increase fiber size in the ligament complex model of knee instability in rats. One might postulate a number of mechanisms by which the SNP exerts its influence on LDD risk, including direct effects on the disc itself or by influence on periarticular structures such as the longitudinal spinal ligaments. Further work is needed to establish its mechanism of action, and this may reveal cellular pathways leading to a greater understanding of the pathophysiology of lumbar disc degeneration.