Constitutive activation of the calcium sensor STIM1 causes tubular-aggregate myopathy

Abstract

Tubular aggregates are regular arrays of membrane tubules accumulating in muscle with age. They are found as secondary features in several muscle disorders, including alcohol- and drug-induced myopathies, exercise-induced cramps, and inherited myasthenia, but also exist as a pure genetic form characterized by slowly progressive muscle weakness. We identified dominant STIM1 mutations as a genetic cause of tubular-aggregate myopathy (TAM). Stromal interaction molecule 1 (STIM1) is the main Ca(2+) sensor in the endoplasmic reticulum, and all mutations were found in the highly conserved intraluminal Ca(2+)-binding EF hands. Ca(2+) stores are refilled through a process called store-operated Ca(2+) entry (SOCE). Upon Ca(2+)-store depletion, wild-type STIM1 oligomerizes and thereby triggers extracellular Ca(2+) entry. In contrast, the missense mutations found in our four TAM-affected families induced constitutive STIM1 clustering, indicating that Ca(2+) sensing was impaired. By monitoring the calcium response of TAM myoblasts to SOCE, we found a significantly higher basal Ca(2+) level in TAM cells and a dysregulation of intracellular Ca(2+) homeostasis. Because recessive STIM1 loss-of-function mutations were associated with immunodeficiency, we conclude that the tissue-specific impact of STIM1 loss or constitutive activation is different and that a tight regulation of STIM1-dependent SOCE is fundamental for normal skeletal-muscle structure and function.

Figure 1

Characterization of Tubular Aggregates in Skeletal Muscle Histology, immunofluorescence, and electron microscopy of muscle biopsies from affected individuals II.1 (family 1, p.Asp84Gly) and III.1 (family 2, p.His109Asn). (A) Histological analysis of transverse sections (10 μm) revealed aggregations on modified Gomori trichrome and NADH-TR staining, but not on succinic-dehydrogenase (SDH) staining (II.1, family 1). (B) Ultrastructural analysis demonstrated prominent tubular aggregation with single- or double-walled membranes on transversal (II.1, family 1) and longitudinal (III.1, family 2) sections. (C) Immunofluorescence showed colocalization of STIM1 and RYR1 in the periphery of the aggregates, whereas SERCA1 homogeneously labeled the aggregates (II.1, family 1). The upper panel shows muscle sections of a healthy individual.

Figure 2

Identification of STIM1 Mutations in Autosomal-Dominant TAM (A) Pedigrees indicate dominant inheritance of TAM, and sequence analysis confirmed the segregation of the heterozygous mutations with the disease. (B) Schematic representation of the STIM1 domains. Arrows indicate the position of the substitutions in the EF-hand domains (amino acids 63–128). The following abbreviations are used: TM, transmembrane domain; CC, coiled-coil domain; SP, serine- and proline-rich domain; and K, polylysine. (C) Protein-sequence conservation of the STIM1 EF hands. The affected amino acids are highly conserved. (D) The resolved protein structure demonstrates a close proximity of the affected amino acids to the Ca²⁺-binding domain. (E) Immunoblot of endogenous STIM1 in individual II.1 from family 1 and control myoblasts shows a comparable STIM1 level. β-Tubulin was used as a loading control.

Figure 3

STIM1 Clustering in C2C12 Cells in Dependence of Thapsigargin and Quantification (A) C2C12 myoblasts were transfected with wild-type or altered STIM1 YFP constructs with the use of Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) and were incubated with or without thapsigargin (2 μM; Sigma-Aldrich, St. Louis, MO, USA). Point mutations c.216C>G (p.His72Gln), c.251A>G (p.Asp84Gly), c.325C>A (p.His109Asn), and c.326A>G (p.His109Arg) were introduced by site-directed mutagenesis with the Pfu DNA polymerase (Stratagene, La Jolla, CA, USA). Thapsigargin (TG) induced wild-type STIM1 oligomerization and clustering. A similar clustering was seen for all altered constructs independently of TG treatment. (B) Quantification of STIM1 YFP clustering in C2C12 cells in dependence of TG. At least 100 cells per transfection from three independent experiments were counted and assessed for STIM1 YFP clustering, i.e. for the presence of massive and area-wide clusters. The total number of counted cells was set to 100%, and the percentage of cells (y axis) containing clusters is shown as a light gray column. Error bars represent the SD. Wild-type STIM1 YFP barely clustered without TG addition. In contrast, all altered constructs showed massive clustering independently of TG. The human wild-type and p.Asp76/78Ala STIM1 YPF constructs were a kind gift from Nicolas Demaurex (University of Geneva, Switzerland). p.Asp76/78Ala is an artificial alteration reported to constitutively activate STIM1.

Figure 4

Impact of the STIM1 p.Asp84Gly Substitution on Ca²⁺ Homeostasis in TAM Myoblasts Myoblasts (1.5 × 10⁶ cells) were harvested and incubated in 1 ml medium containing Indo-1 (Invitrogen) and physiological Ca²⁺ concentrations and were then incubated in nominally Ca²⁺-free Ringer solution. Data analysis was done with FlowJo software. Cell viability was verified by TO-PRO-3 staining (Invitrogen). The y axes show the Indo-1 ratio, and the x axes show the time in seconds. (A) Cytometric analysis indicated a higher basal [Ca²⁺] and an increased Ca²⁺ entry after store depletion with TG in myoblasts from individual II.1 (family 1, p.Asp84Gly). (B) An increased Ca²⁺ influx in TAM cells was also seen without TG treatment. (C) To challenge the mechanisms regulating Ca²⁺ influx, we directly added 20 mM Ca²⁺ without previous Ca²⁺-store depletion, resulting in a significantly higher Ca²⁺ influx in the TAM myoblasts.