Two gene polymorphisms were identified as potential genetic modifiers of DMD in both the Padova and CINRG cohorts: SPP1/osteopontin and CELSR2. However, when adjusting for multiple testing, only SPP1/osteopontin remained statistically significant.
The effect of SPP1 genotype on progression was similar in magnitude to the pharmacologic use of steroids (about 1 year difference in the loss of ambulation). The Padova patients with the less common G allele were weaker using both upper and lower limbs MRC (p = 0.001) and also showed earlier loss of ambulation. At age 14 years, 20% of patients with DMD with TT genotype were ambulatory, whereas none with GT/GG genotype were ambulatory. In the CINRG cohort, patients with the G allele showed lower grip strength (p = 0.001), with the effect size largest within the nonambulatory, steroid-treated patients (). By grip strength, GT/GG patients showed a 12% decrease in strength relative to TT across the age range, and this increased to 19% difference in nonambulatory patients.
The rs28357094 polymorphism is 66 bp upstream of the transcriptional start site, and has been shown to alter SPP1 binding to the promoter. The rare G allele has been shown to decrease promoter strength, leading to lower SPP1 mRNA production from the gene, by in vitro cell transfection assays in nonmuscle cells.23
SPP1 has been intensively studied as a marker for tumor progression, and its role in inflammation and tissue remodeling in many pathologies and disorders.24,–26
The goal of most clinical trials in DMD is to slow the progression of the disease. However, since the rate of progression is variable, large numbers are required for appropriate statistical powering. We used the CINRG cross-sectional data to carry out power calculations on genotype-stratified subcohorts. Assuming that a 15% deflection of strength loss is clinically significant, fewer patients with DMD with the G genotype are needed to detect a slowing of progression (n = 106), compared to those homozygous for the ancestral T genotype (n = 1,520 patients required).
The identification of genetic modifiers also provides insights into disease pathogenesis.27
Osteopontin is elevated in dystrophin-deficient muscle, both in humans14,28
and mdx mice.29
However, elevations of SPP1 are not specific to dystrophin deficiency, and are seen in multiple dystrophies, including Becker muscular dystrophy, calpain 3 (LGMD2A), and dysferlin deficiency (LGMD2B),18
and mouse models of muscular dystrophy.30
SPP1 is considered an inflammatory marker; however, inflammatory muscle diseases do not show striking OPN levels.18,31
Thus, SPP1 elevations may correspond with muscle remodeling rather than inflammation. In the mdx mouse, SPP1 was shown to be expressed by both muscle fibers and a subset of muscle inflammatory cells.18
Double knock-out (SPP1-null, dystrophin-null) mice showed less fibrosis and less functional deficits, suggesting that SPP1 expression is deleterious to dystrophic muscle.18
The mdx/SPP1 data and the DMD data presented here may seem contradictory; in mdx mice less SPP1 leads to improvement, whereas in patients with DMD less SPP1 leads to greater weakness. However, SPP1 is a multifunctional molecule that has positive roles in muscle regeneration32
and negative roles in muscle inflammation, and the mouse and human data may simply highlight these different roles for SPP1.
Taking into account all the findings presented here, the less common G allele of the rs28357094 SPP1/osteopontin promoter polymorphism is associated in DMD with greater weakness and younger age at loss of ambulation. Further research is needed to define the role of the rs28357094 promoter polymorphism on DMD pathophysiology. Regardless of the molecular action, it is important to note that stratification of patients with DMD by genotype has the potential to increase the statistical power and sensitivity of clinical trials. Given the data presented, the clinical trial planner would have 2 choices: either open the trial to all comers or restrict it to the G subgroup, in which case the savings (reduced numbers of patients needed for the trial) is roughly ¼. Conversely, the number of patients with the more severe G genotype is less common, necessitating pretrial screening of larger numbers of patients with DMD.