SEPN1-related myopathy depends on the oxidoreductase ERO1A and is druggable with the chemical chaperone TUDCA

Abstract

Selenoprotein N (SEPN1) is a protein of the endoplasmic reticulum (ER) whose inherited defects originate SEPN1-related myopathy (SEPN1-RM). Here, we identify an interaction between SEPN1 and the ER-stress-induced oxidoreductase ERO1A. SEPN1 and ERO1A, both enriched in mitochondria-associated membranes (MAMs), are involved in the redox regulation of proteins. ERO1A depletion in SEPN1 knockout cells restores ER redox, re-equilibrates short-range MAMs, and rescues mitochondrial bioenergetics. ERO1A knockout in a mouse background of SEPN1 loss blunts ER stress and improves multiple MAM functions, including Ca²⁺ levels and bioenergetics, thus reversing diaphragmatic weakness. The treatment of SEPN1 knockout mice with the ER stress inhibitor tauroursodeoxycholic acid (TUDCA) mirrors the results of ERO1A loss. Importantly, muscle biopsies from patients with SEPN1-RM exhibit ERO1A overexpression, and TUDCA-treated SEPN1-RM patient-derived primary myoblasts show improvement in bioenergetics. These findings point to ERO1A as a biomarker and a viable target for intervention and to TUDCA as a pharmacological treatment for SEPN1-RM.

Keywords: ER stress; ERO1; SEPN1; TUDCA; core myopathy; multi mini-core disease.

Graphical abstract

Figure 1

ERO1 a biomarker of SEPN1-RM (A) Venn diagram of intersecting target gene sets from ER and oxidative stress responses. (B) Top 10 processes, per the 2023 update, from GO analysis of common genes in (A). (C) Manhattan plot illustrating enrichment of input gene set in the 2023 GOBP gene set, with each point representing a single term. (D) Search tool for the retrieval of interacting genes/proteins (STRING) network analysis illustrates SEPN1’s connections with the nine specified genes (from A), revealing protein-protein functional and physical interactions based on published evidence compiled by STRING database (v.12). Each colored bubble signifies a functionally related gene cluster, with gray lines denoting intracluster connections and orange lines representing intercluster connections. The strength of the connections is indicated by the thickness of the connecting lines. (E) Top 5 diseases from the Rare Diseases AutoRIF database linked with the common gene subset from (A). (F) Phenotype of the SEPN1-RM patient P5 at age 15 years, showing amyotrophy of the upper limb proximal muscles (particularly deltoids) and the typical SEPN1-RM scoliosis, with dorsal and lumbar hyperlordosis and lateral trunk deviation. (G) Real-time RT-qPCR analysis of human ERO1 expression in cDNA samples derived from paravertebral muscle biopsies of three healthy controls and three patients with SEPN1-RM. Error bars denote SD; ∗p < 0.05 by Mann-Whitney test.

Figure 2

Interaction between SEPN1 and ERO1 impinges on short-distance MAMs and bioenergetics (A) FLAG and ERO1 immunoblots of FLAG-tagged SEPN1 immunopurified with FLAG-M2 antibody from lysate of cells that were untransfected or transfected with expression plasmids of the indicated proteins. The immunoprecipitates were resolved on reducing and non-reducing SDS-PAGE and Coomassie. On the right, a bar graph indicates the relative levels of ERO1 associated with its bait, FLAG-SEPN1 (n = 3, unpaired t test). The bottom images represent the 5% of the total input protein lysate immunopurified and resolved on reducing SDS-PAGE. (B) Scheme of the mass spectrometry analysis of FLAG-SEPN1 immunoprecipitates from HEK293T cells. The Coomassie band around 175 kDa was cut, digested, and analyzed by mass spectrometry. Spectra counts of ERO1 on those of SEPN1 in the same band are expressed as a ratio (low, high). (C) Detection of SEPN1 oligomers in WT and ERO1 KO HEK293T cells. FLAG immunoprecipitates were analyzed on non-reducing immunoblot. ERO1 immunoblot indicates the amount of ERO1 in the indicated samples. On the bottom is a bar graph indicating the ratio SEPN1 oligomers on total SEPN1 n = 3, one-way ANOVA). (D) ERO1 and SEPN1 immunoblot in WT, ERO1 KO, SEPN1 KO, and DKO HeLa cells. Immunoblot of KDEL containing proteins and Ponceau indicate protein loading. (E) Bar plots indicating the baseline fluorescence excitation ratio of roGFP2, reflecting the redox state of roGFP2 localized in the ER. On the right are traces of time-dependent changes in the fluorescence excitation ratio of roGFP2. Cells were exposed to a dithiothreitol (DTT; 1 mM) pulse of 20 min followed by a washout of the reductant and to 1 μM thapsigargin (Tg) treatment. Each data point represents the mean ± SEM of the fluorescence excitation ratio of roGFP2 (N = 8 fields of view per well obtained from two independent experiments, one-way ANOVA for repeated measures followed by Tukey’s multiple comparisons). (F) Images and quantification of SPLICS_L^ER-MT in WT and SEPN1 KO HeLa cells (n = 3, unpaired t test). (G) Images and quantification of SPLICS_S^ER-MT in WT and SEPN1 KO treated for 3 h with 1 μM Tg (n = 3, uncorrected Fisher’s least significant difference [LSD] test two-way ANOVA). (H) Images and quantification of SPLICS_S^ER-MT in WT and SEPN1 KO treated for 24 h with 1 mM TUDCA (n = 3, uncorrected Fisher’s LSD test two-way ANOVA). (I) Images and quantification of short SPLICS in WT, SEPN1 KO, ERO1 KO, and DKO cells (scale bars: 25 μm). (J) Trace and quantification of ATP production in WT, SEPN1 KO, ERO1 KO, and DKO cells after stimulation with histamine. CPS indicates counts per second (Bonferroni multiple comparison after two-way ANOVA).

Figure 3

UPR induction in SEPN1 KO diaphragms is rescued in DKO counterparts (A) Graphical representation of the cross between SEPN1 KO and ERO1 KO mice to get (SEPN1, ERO1 KO) DKO mice. (B) Dot plots in hallmark gene sets indicating the upregulation of UPR in SEPN1 KO diaphragms (pink/red dots) and the downregulation in DKO (light blue/blue dots) from 9-month-old mice (n = 4). (C) Bar plots indicating downregulation (blue bars) of PERK-, IRE1-, and ATF6-mediated UPR pathways by GOBP in the DKO when compared to SEPN1 KO. FDR indicates false discovery rate. (D) Real-time qPCR on cDNA from diaphragms of mice of the genotypes indicated (n = 5, one-way ANOVA). (E) ERO1 and SEPN1 immunoblot from 3-month-old mice. GAPDH immunoblot indicates equal protein loading. (F) ERO1 immunoblot from 9-month-old mice. GAPDH immunoblot indicates equal protein loading. On the right is a bar graph indicating ERO1 levels in arbitrary units (a.u.) (n = 4, unpaired t test).

Figure 4

Altered OXPHOS in diaphragms of SEPN1 KO mice is regularized in DKO mice (A) Bar graphs indicating the top ten most perturbed gene sets (hallmark) of SEPN1 KO and DKO diaphragms. Enrichment and their FDR-adjusted p values were computed using a camera (preranked) and were determined on the hallmark gene sets collection. The x axis reports the logarithmically transformed FDR in the form of −10×log10 (FDR), with a bold intercept (x = 13.01) indicating the FDR threshold of 0.05. Red bars: upregulated; blue bars: downregulated. (B) Dot plots in hallmark gene sets indicating the down- (blue dots) and upregulation (red dots) of OXPHOS in SEPN1 KO and DKO diaphragms, respectively (n = 4). (C) Heatmap of OXPHOS genes from the hallmark gene sets collection differently regulated in SEPN1 KO and DKO diaphragms. On the right are bar graphs indicating results from real-time qPCR on cDNA from diaphragms of mice of the genotypes indicated (n = 5, one-way ANOVA). (D) Bar graphs indicating mitochondrial protein content of the diaphragms after isolation of pure mitochondria (Mito;n = 3, one-way ANOVA). (E) Immunoblot of the different complexes (I–V) of OXPHOS in diaphragms of the indicated genotypes. On the right are quantifications of the complexes. (F) ATP levels in diaphragm, soleus, and EDL muscles (n = 5, one-way ANOVA).

Figure 5

Defects in CRU-to-Mito apposition of SEPN1 KO mice are rescued by ERO1 loss and TUDCA (A) Hematoxylin and eosin (H&E), nicotinamide adenine dinucleotide tetrazolium reductase (NADH-TR), and wheat germ agglutinin (WGA) staining of representative transverse frozen sections of diaphragms from 6-month-old mice. On the bottom is a histogram of the relative frequency of the minimal Feret diameter of the diaphragm fibers. (B) Sirius red staining of transverse frozen sections of diaphragms. On the bottom are dot plots indicating Sirius red-positive areas as percentages. (C) Representative images of fast (IIA and IIX)- and slow (I)-twitch muscle fibers in diaphragm muscle sections using immunostaining with anti-fast and anti-slow MyHC antibodies. Below, the bar graph shows the percentages of fiber types (n = 5 mice/genotype, one-way ANOVA). (D) Representative EM images from WT (A and B), SEPN1 KO (C and D), DKO (E and F), and SEPN1 KO, TUDCA (G and H) diaphragms, respectively. Labeling: large arrows point to the Z line; small arrows point to CRUs or triads; asterisk is for longitudinal A-band; Mito and M is for mitochondria. Scale bars: (A, E, C, and G) 1 μm and (B, F, D, and H) 0.5 μm. On the bottom is quantification of CRU, mito (mitochondria), and CRU/mito pairs. Data are expressed as average number per 100 μm² and shown as mean ± SEM (n = 3, One-way ANOVA followed by post-hoc Tukey test for multiple comparisons) (E) Time to 50% basal measurement of calcium uptake after single-pulse stimulation in primary culture of mouse FDB fibers. (F) Effect of recovery of diaphragm single-fiber tension (n = 5 mice/genotype, non-parametric one-way ANOVA, Kruskall-Wallis multiple comparison test).

Figure 6

TUDCA-mediated improvement of SEPN1 KO muscle phenotype (A) Treatment scheme of TUDCA in WT and SEPN1 KO mice. (B) Weekly weights of mice during treatment. (C) Dot plots indicating the TUDCA levels in diaphragms of WT and SEPN1 KO mice after TUDCA treatment or placebo (n = 6). (D) Real-time qRT-PCR analysis of ERO1 and CHOP from mRNA from diaphragms (n = 5, one-way ANOVA). (E) Time to 50% basal measurement of calcium uptake after single-pulse stimulation in primary culture of mouse FDB fibers. (F) Effect of recovery of diaphragm single-fiber tension (n = 10 mice/genotype, one-way ANOVA, Kruskall-Wallis multiple comparison test).

Figure 7

ER stress associated with ATP deficit in human SEPN1-RM primary myoblasts rescued by TUDCA treatment (A) Graphical representation of myoblast isolation in patients with SEPN1-RM. (B) Real-time RT-qPCR analysis of ERO1 expression in SEPN1-RM myoblasts relative to age- and passage-paired healthy myoblast controls. (C) Level of ATF4-mScarlet fluorescent signal with respect to each maximum (594 nm) 24 h post-H₂O₂ treatment (positive signal control). (D) ATP level (relative luminescence unit, RLU) in human myoblast cultures treated with increasing doses of Tg (0–400 nM) and TUDCA (0, 0.25, 0.5, 1, and 2 mM). Line plot shows the running average with the overlaid scatterplot indicating the range of ATP levels observed in each condition (n = 5, two-way ANOVA with Dunnett post hoc correction).