LTBP4 genotype predicts age of ambulatory loss in Duchenne muscular dystrophy

Abstract

Objective: Duchenne muscular dystrophy (DMD) displays a clinical range that is not fully explained by the primary DMD mutations. Ltbp4, encoding latent transforming growth factor-β binding protein 4, was previously discovered in a genome-wide scan as a modifier of murine muscular dystrophy. We sought to determine whether LTBP4 genotype influenced DMD severity in a large patient cohort.

Methods: We analyzed nonsynonymous single nucleotide polymorphisms (SNPs) from human LTBP4 in 254 nonambulatory subjects with known DMD mutations. These SNPs, V194I, T787A, T820A, and T1140M, form the VTTT and IAAM LTBP4 haplotypes.

Results: Individuals homozygous for the IAAM LTBP4 haplotype remained ambulatory significantly longer than those heterozygous or homozygous for the VTTT haplotype. Glucocorticoid-treated patients who were IAAM homozygotes lost ambulation at 12.5 ± 3.3 years compared to 10.7 ± 2.1 years for treated VTTT heterozygotes or homozygotes. IAAM fibroblasts exposed to transforming growth factor (TGF) β displayed reduced phospho-SMAD signaling compared to VTTT fibroblasts, consistent with LTBP4' role as a regulator of TGFβ.

Interpretation: LTBP4 haplotype influences age at loss of ambulation, and should be considered in the management of DMD patients.

Figure 1

LTBP4 protein and position of amino acid substitutions induced by single nucleotide polymorphisms (SNPs). A) SNPs rs2303729, rs1131620, rs1051303 and rs10880 alter LTBP4; the amino acid substitutions are shown in red. The amino terminus of LTBP4 binds to the extracellular matrix (ECM). The protein is composed of epidermal growth factor (EGF)-like repeats (light blue boxes). B) LTBP4 protein haplotypes and frequencies estimated from 381 individuals found in the 1000 Genomes interim Phase I data from populations with European ancestry. The rs1051303 nsSNP alters the protein at amino acid 820 and is in perfect linkage disequilibrium (r² = 1) with rs1131620. The pairwise linkage disequilibrium (r²) between the other three nsSNPs in our DMD cohort versus 1000 Genomes was: rs2303729:rs1131620 (0.67 vs. 0.68), rs2303729:rs10880 (0.50 vs. 0.45), and rs1131620:rs10880 (0.62 vs. 0.65).

Figure 2

LTBP4 genotypes are associated with age of loss of ambulation in patients with Duchenne muscular dystrophy. A) Patients who have lost ambulation by 20 yrs of age are classified by their nonsynonymous LTBP4 genotype. For each genotypic class, the survival curve to age of loss of ambulation and the Wald test asymptotic p-value (recessive) using a linear model applied to age of ambulatory loss as a quantitative trait is shown. B) DMD mutation class does not differ for those with the protective LTBP4 allele. Distribution of age at ambulatory loss grouped by LTBP4 rs10880 genotype (individuals per genotype; Thr:Thr = 109, Thr:Met = 114, Met:Met = 31), with dystrophin mutation class indicated by color: green = truncating mutations, gold = other mutations (see Tables S3 and S4 for definitions of truncating DMD mutation versus other DMD mutation classes). Detailed dystrophin mutation nomenclature is shown for patients A through J from the upper ambulatory range in the rs10880 Thr:Met and Met:Met groupings.

Figure 3

The LTBP4 rs10880 minor homozygote (TT = Met:Met) and IAAM/IAAM diplotype is associated with extended ambulation in glucocorticoid-treated Duchenne muscular dystrophy patients. A) Age of ambulatory loss classified by rs10880 genotype and color-coded for patients by their steroid treatment. Two pairs of brothers discordant for their rs10880 genotypes are shown connected by the dotted lines. B) Survival curves for DMD patients with the LTBP4 IAAM/IAAM versus other haplotype pairs, with steroid treated versus naïve individuals plotted separately. The IAAM haplotype consists of SNPS: rs2303729 (I=Ile), rs1131620 (A=Ala), rs1051303 (A=Ala) and rs10880 (M=Met) respectively. There were 16 IAAM homozygotes in the steroid treated group and 10 IAAM homozygotes in the naïve group.

Figure 4

Fibroblasts homozygous for the IAAM or VTTT LTBP4 alleles or heterozygous (IAAM/VTTT) were cultured and latent TGFβ was added. LTBP4 protein binds and sequesters latent TGFβ, and it is predicted that IAAM binds less latent TGFβ in the matrix leading to reduced TGFβ signaling, seen as phosphorylated SMAD (pSMAD). IAAM fibroblasts had less pSMAD/total SMAD at baseline and at two different doses of latent TGFβ (p =0.02, ANOVA with repeated measures).