All patients had a vacuolar myopathy due to lipid storage (Fig. 2a–b). The amount of lipid varied from moderately increased to massive accumulation as seen by lipid staining with Sudan black (Fig. 2c, Table 2). In general, the type 1 muscle fibers were more affected with larger vacuoles (Fig. 2d). Enzyme histochemical staining for oxidative enzymes (Complex II; succinate dehydrogenase, SDH, and Complex IV; cytochrome c oxidase, COX) showed a reduced staining intensity in most patients compared to controls (Fig. 2e–f, Table 2). Gomori trichrome (GT) staining (Fig. 2b) and electron microscopy (see below) showed a marked increase in the number of mitochondria. Typical ragged red muscle fibers were infrequent. In some of the cases occasional scattered necrotic fibers were observed (Table 2). There was no apparent increase in interstitial connective tissue.

Muscle pathology in patient P10. Insets are normal controls. a Multiple large and small vacuoles are present in the muscle fibers (hematoxylin and eosin, H&E) b In Gomori trichrome (GT) staining, mitochondrial proliferation is seen as red granular material, mainly in fibers with vacuoles. c Lipid staining (Sudan black, SB) showing the storage material in the vacuolated fibers to be composed of lipids. d Muscle fiber typing by immunofluorescence analysis of myosin heavy-chain isoforms showing that type 1 (blue) muscle fibers are more vacuolated that type 2A (green) muscle fibers. Type 2B (red) muscle fibers are only present in the control (inset). e Enzyme histochemical staining of cytochrome c oxidase (COX) showing variable and in general pale staining of muscle fibers indicating a partial COX deficiency. f Enzyme histochemical staining of succinate dehydrogenase (SDH) showing pale (and partially artefactual staining due to lipid accumulation) indicating SDH deficiency. Bars = 50 µm

Electron microscopy was performed in seven of the 11 patients. All patients showed mitochondrial proliferation, which was extensive in some patients (Fig. 3, Table 2). Most of the patients displayed numerous small mitochondria with dark matrix resulting in an overall dark appearance of the mitochondria. Although many of the mitochondrial profiles appeared unusually small, many were elongated (Fig. 3 and Supplementary Fig. 2 and 3). In several patients, some enlarged mitochondria measuring more than 1 µm in diameter were observed (Fig. 3b and Supplementary Fig. 2 and 3). The arrangements of mitochondrial cristae were in general not deranged, but in most cases small electron-dense inclusions, frequently round in shape, were observed (Supplementary Fig. 2 and 3). “Parking lot” paracrystalline inclusions were not observed.

Electron micrograph of muscle in patient P10. a Abnormal lipid storage in the muscle fibers is seen as increased size of lipid droplets, which in some fibers show clustering (arrow). b There is an abnormally large number of pleomorphic mitochondria, which in this image cluster in a subsarcolemmal region. c In this fiber the pleomorphic, dark mitochondria are accumulated in the intermyofibrillar compartment (arrows). d At high magnification the dense matrix of the mitochondria is seen, as well as their close connection to lipid droplets (arrow)

Genetic analysis was performed in all 11 patients (1 WES and 10 WGS). No variants that fulfilled the criteria to be likely pathogenic or pathogenic according to the American College of Medical Genetics and Genomics (ACMG) that could explain a metabolic disorder with an autosomal recessive inheritance were identified [22]. In one patient, a single heterozygous variant of uncertain significance (VUS) was identified in ETFDH.

Detailed characterization of mtDNA was performed in 10 patients after WGS of skeletal muscle DNA, which showed a mean read depth of mtDNA of 157,500x. Bioinformatic analysis did not reveal any increase of large scale mtDNA deletions or duplications in any of the lipid storage myopathy cases compared to controls (Fig. 4a). The mtDNA copy number was in general higher in the patients with lipid storage myopathy with a mean of 7701 mtDNA copies per 1 nuclear DNA copy (range 3314–12,178) compared to controls muscle (mean 4029; range 2552–5054) (Fig. 4b).

Investigation of mtDNA rearrangements and mtDNA copy number. a No increased number of mtDNA deletions or duplications were identified in patients with lipid storage myopathy associated with sertraline treatment. The circles illustrate mtDNA. The red (duplications) and blue (deletions) lines represent large-scale rearrangements of mtDNA. The intensity of each line corresponds to the number of specific deletions/duplications. For comparison, a typical example of patients with inclusion body myositis (IBM) with multiple deletions and duplications and a normal control is illustrated. b The mtDNA copy number were in general increased in the patients with lipid storage myopathy associated with sertraline treatment, which may reflect the increased number of mitochondria. For comparison, normal controls and a representative group of patients with IBM that frequently show reduced mtDNA copy numbers in spite of mitochondrial proliferation are illustrated

On the background of lipid accumulation, mitochondrial proliferation, abnormal mitochondrial structure, apparently reduced activity of Complex II and IV as seen on enzyme histochemistry (COX and SDH), and accumulation of acylcarnitines of various lengths in blood, all pointing toward a mitochondrial dysfunction, we studied the expression of mitochondrial proteins in the proteome of the muscle-biopsy specimens. The analysis included eight patients and eight age-matched controls. The quantitative analysis was based on nanoscale liquid chromatography coupled to TMT based LC-MS3. From the basic analysis of the proteomic data, around 4,300 proteins were identified. Of these, 3,600 were identified in all samples and quantified. The principal component analysis (PCA) showed that the controls were distributed as a group separate from the patients (Fig. 5a). 1,926 proteins were significantly different (adjusted p value (FDR) < 0.05) between the control group and the sertraline group (Fig. 5b), the majority being upregulated in the sertraline group (Fig. 5b, c). The downregulated proteins were mainly associated with mitochondria, especially the respiratory chain in spite of the mitochondrial proliferation (Fig. 5d, the maps are generated at the Proteomap website (www. https://www.proteomaps.net/; [16]). In the heat map as well as in the PCA analysis based on the overall protein profiling the controls are more homogeneous than the patients. In the heat map the patients form two clusters, both separate from the control cluster. It is not clear from clinical or pathological features why the patients fall into two clusters.

Basic proteomic data. a Principal component analysis (PCA) shows that the controls are clustered together and no apparent outliers. b Volcano plot including all approximately 3600 quantified proteins illustrating sertraline group versus controls. The proteins with an adjusted p value (false discovery rate, FDR) <0.05 and a fold change of <0.5 (log2FC < −1) (blue) or a fold change of >2 (log2FC > 1) (red) are indicated. c Heatmap displays the expression of differentially expressed proteins identified from the proteomics analysis. Each column corresponds to one sample (P, patient; C, control). Red indicates high expression level; green indicates low expression level. Hierarchical clustering shows the dendrogram based on the differences in protein profile for patients and controls, and shows difference in the clustering between patients and controls. d Proteomap shows that many significantly downregulated proteins with a fold change of less than 0.5 are involved in oxidative phosphorylation

To study the most important pathways involved in the fatty acid energy metabolism we analyzed the five individual complexes of the respiratory chain (CI-CV), the ß-oxidation (FAO) and the citric acid cycle (TCA-cycle) (Figs. 6, 7 and Supplementary Table 4). Analysis of the subunits of the five enzyme complexes of the respiratory chain showed that Complex I was markedly downregulated. Most of all significantly downregulated proteins with an average fold change of less than 0.5 were Complex I subunit proteins (NADH dehydrogenase) (Fig. 6a). The subunits of Complex II (succinate dehydrogenase) were all downregulated but with a lesser fold change than Complex I (Fig. 6a). The vast majority of Complex IV (cytochrome c oxidase) subunits were also downregulated (Fig. 6a). On the other hand, subunits of Complex III and V were generally unchanged or upregulated (Fig. 6a). Including assembly proteins in the analysis revealed that the assembly factors of Complex I were generally upregulated, which was different from the subunits (Fig. 6b). No assembly factors of Complex II were identified in all samples and therefore not quantified (Fig. 6c). In contrast, most of Complex III, IV and V assembly factors were upregulated (Fig. 6d–f).

Volcano plots of subunits and assembly factors involved in Complex I–V in the respiratory chain (oxidative phosphorylation) a Volcano plots of subunits of Complex I–V showing downregulation of mainly Complex I but also of Complex II and IV, whereas Complex II and V are upregulated and possibly unchanged in relation to the increased number of mitochondria. b–f Volcano plats of the individual complexes I–V including also assembly factors. While the subunits of Complex I, II and IV are all statistically downregulated most assembly factors are up regulated and possibly unchanged in relation to the increased number of mitochondria

Volcano plots of proteins involved in fatty acid ß-oxidation (a) and citric acid cycle (b). Most of these enzymes appear to be up-regulated and possibly unchanged in relation to the increased number of mitochondria. One exception is ETFDH encoding the ETF-QO complex which is significantly downregulated more than twofold (red symbol in a). Citrate synthase (CF), a common biochemical marker for overall mitochondrial volume, is significantly upregulated in line with the increased number of mitochondria (red arrow). For explanation of gene symbols see Supplementary Table 5

For explanation of gene symbols see Supplementary Table 4.

The enzymes involved in FAO were in general upregulated (Fig. 7a). An interesting exception being ETF-coenzyme Q oxidoreductase (ETF:CQ) encoded by ETFDH, which was significantly downregulated with a fold change of less than 0.5.

Likewise, the enzymes of the TCA cycle were generally upregulated including citrate synthase (Fig. 7b). This upregulation probably reflects the increased number of mitochondria.

The proteomic results indicated downregulation mainly of respiratory chain subunits of Complex I and to some extent of Complex II and IV. We performed immunofluorescence analysis to study the cellular distribution of these complexes taking the mitochondrial proliferation in individual fibers into account by simultaneous staining of a mitochondrial membrane protein (porin, VDAC1). There was a profound deficiency of Complex I (NDUFB8) in most of the patients and only occasional muscle fibers expressed normal levels in relation to the increased number of mitochondria (Figs. 8 and 9, Table 2, Supplementary Fig. 4). For Complex II (SDHB) and IV (MT-CO1), the deficiency was less pronounced compared to the Complex I deficiency. The level of complex II and IV varied between different adjacent muscle fibers in a mosaic pattern (Figs. 8 and 9, Table 2). There was evidence of an increased number of mitochondria in all patients (Figs. 8 and 9, Table 2).

Quadruple immunofluorescence assay of Complex I and IV of the respiratory chain in Patient P10 (a–d) and a simultaneously stained control (e–h). In merged illustrations (d and h) yellow fibers are normal, red fibers are Complex I deficient, green fibers are Complex IV deficient and blue fibers are deficient of both Complex I and IV. There is profound deficiency of Complex I (a) and partial deficiency of Complex IV (b) in P10. Only occasional fibers show preserved Complex I and IV (arrow). The increased intensity of VDAC1 staining in many muscle fibers in P10 indicates increased number of mitochondria. Bars = 50 µm

Quadruple immunofluorescence assay of Complex I and II of the respiratory chain in Patient P10 (a–d) and a simultaneously stained control (e–h). In merged illustrations (d and h) yellow fibers are normal, red fibers are Complex I deficient, green fibers are Complex II deficient and blue fibers are deficient of both Complex I and IV. There is profound deficiency of Complex I (a) and partial deficiency of Complex II (b) in P10. Only occasional fibers show preserved Complex I and IV (arrow). The increased intensity of VDAC1 staining in many muscle fibers in P10 indicates increased number of mitochondria. Bars = 50 µm

Western blot analysis was performed with the same antibodies as the immunofluorescence assay in addition to Complex III (UQCRC2) and Complex V (ATPB). The overall pattern showed a reduced amount of Complex I, II and IV but no reduction of Complex III and V (Table 2 and Supplementary Fig. 5).