An X-linked myopathy with postural muscle atrophy and generalized hypertrophy, termed XMPMA, is caused by mutations in FHL1

Abstract

We have identified a large multigenerational Austrian family displaying a novel form of X-linked recessive myopathy. Affected individuals develop an adult-onset scapulo-axio-peroneal myopathy with bent-spine syndrome characterized by specific atrophy of postural muscles along with pseudoathleticism or hypertrophy and cardiac involvement. Known X-linked myopathies were excluded by simple-tandem-repeat polymorphism (STRP) and single-nucleotide polymorphism (SNP) analysis, direct gene sequencing, and immunohistochemical analysis. STRP analysis revealed significant linkage at Xq25-q27.1. Haplotype analysis based on SNP microarray data from selected family members confirmed this linkage region on the distal arm of the X chromosome, thereby narrowing down the critical interval to 12 Mb. Sequencing of functional candidate genes led to the identification of a missense mutation within the four and a half LIM domain 1 gene (FHL1), which putatively disrupts the fourth LIM domain of the protein. Mutation screening of FHL1 in a myopathy family from the UK exhibiting an almost identical phenotype revealed a 3 bp insertion mutation within the second LIM domain. FHL1 on Xq26.3 is highly expressed in skeletal and cardiac muscles. Western-blot analysis of muscle biopsies showed a marked decrease in protein expression of FHL1 in patients, in concordance with the genetic data. In summary, we have to our knowledge characterized a new disorder, X-linked myopathy with postural muscle atrophy (XMPMA), and identified FHL1 as the causative gene. This is the first FHL protein to be identified in conjunction with a human genetic disorder and further supports the role of FHL proteins in the development and maintenance of muscle tissue. Mutation screening of FHL1 should be considered for patients with uncharacterized myopathies and cardiomyopathies.

Figure 1

Pedigree of the X-Linked Myopathy with Postural Muscle Atrophy Families (A) Only the core pedigree is shown here. Affected males are indicated by filled squares. The patient from the Schwarzmeier et al. study is indicated with an arrow. (B) The second XMPMA family, from the UK, is shown. The index patient is marked with an arrow.

Figure 2

Phenotype and Histological Findings of Patient 20 The clinical phenotype is composed of a scapulo-axio-peroneal syndrome with an athletic appearance. Predominant shoulder girdle atrophy with scapular winging and axial muscular atrophy of postural muscles is found in the patients (A and B). Muscle biopsies show moderate degenerative myopathy. In addition, a few round autophagic vacuolar changes are apparent. H & E stain is shown in (C) and (D). ATPase (pH. 9.4) is shown in (E). Increase in fiber-size variation with hypertrophy of type 2 fibers (arrow in [E]) and atrophy of both fiber types is seen (dark stained fibers correspond to type 2 fibers). In the NADH staining (F), some core-like NADH-negative central zones are seen (arrow in [F]). Immunohistology with monoclonal antibodies against desmin (G) of specimens from patient 50 showed positive staining in core-like lesions (arrow and insert in [G]), and some subsarcolemmal desmin accumulation. Electron microscopy of the muscle from patient 1 shows an autophagic vacuole with autophagic debris including myelin-like figures (H). White scale bars in (C)–(G) adjusted to 30 μm. Magnification × 3300 is shown in (H).

Figure 3

Two-Point Parametric and Multipoint Nonparametric Linkage Analysis for the XMPMA Family Linkage data for X chromosome STRPs, analyzed with the SuperLink program version 1.5 from easyLINKAGE plus v5.02. Allele frequencies were assumed to be equal, and disease frequency was set to 0.0001. Multipoint LOD scores were calculated with ALLEGRO version 1.2c. Positions of markers are indicated in centimorgan (cM), as provided by Marshfield. Maximum LOD scores are shown.

Figure 4

Ideogrammatic Representation of the XMPMA Locus on the Distal Arm of Chromosome X, Genomic Organization, Mutation Identification, and Protein Sequence of the FHL1 Gene (A) The XMPMA locus on Xq26.3 is indicated, along with the intron/exon structure of FHL1 isoforms A, B, and C. Protein sizes (kDa) for the different isoforms were determined by western hybridization by Brown et al. for FHL1a and by Ng et al. for FHL1c. (B) Electropherograms indicating the wild-type and mutation sequence for the Austrian and UK XMPMA families are shown, as well as the secondary structure of FHL1, indicating the position of the resulting amino acid substitution, C224W, relative to structural features in the protein. (C) Amino acid sequence for FHL1a, indicating LIM domains that were underlined, and Zn²⁺-binding cysteine residues were indicated in red type. (D) Comparative analysis: Analysis of the amino acid sequence for fourth LIM domain of FHL1 across vertebrate species with BOXSHADE. Zn²⁺-binding cysteine residues (^∗) and the position of mutation Cys224Trp (C224W) (^∗) are indicated. The domain and cysteine residues are highly conserved across vertebrates, from primates through to amphibians and primitive fish.

Figure 5

Isoforms of FHL1 The organization of tandem LIM domains, as well as nuclear localization signals, nuclear export signals, and protein binding domains for FHL1 isoforms A, B, and C are shown. The cDNA and protein for each isoform are displayed alongside each other, with cDNA size and the number of amino acids present in each isoform being indicated on the right-hand side. The positions of the mutations in the Austrian family (p.C224W) and UK family (p.F127_T128insI) are shown relative to the three different FHL1 isoforms.

Figure 6

Western Blot and FHL1/Myosin Slow Immunofluorescence (A) FHL1 western blotting: In specimens from patients 50 and 11 (P1 and P2), a normal control (Co), and controls with either Becker MD (BMD) or limb girdle MD (LGMD), the specific FHL1 band at 29 kDa is severely reduced compared to normal and disease controls. The 50 kDa band corresponds to alpha-sarcoglycan (adhalin) showing equal quantity of protein loading in all samples. (B) Anti-FHL1 and anti-myosin slow-type immunofluorescence of human muscle from two patients with FHL1 mutation. The top row shows control samples: Overlay shows that FHL1 (red) does not colocalize to myosin slow type (green). Shown in the middle row: An overlay in a patient with FHL1 mutation reveals nearly complete loss of FHL1 staining. Shown in the bottom row: Overlay in a patient with FHL1 mutation shows a very few remaining anti-FHL1-positive fibers.

Figure 7

Expression Analysis of FHL1 in Primary Human Myoblasts Anti-FHL1 staining in primary myoblasts from control (Co) and from patient 50 (P1). The patient's myoblasts show lower expression of FHL1, but subcellular distribution is not different to control. Both patient and control show primarily cytoplasmic but also nuclear FHL1 distribution.