Bi-allelic LETM1 variants perturb mitochondrial ion homeostasis leading to a clinical spectrum with predominant nervous system involvement

Abstract

Leucine zipper-EF-hand containing transmembrane protein 1 (LETM1) encodes an inner mitochondrial membrane protein with an osmoregulatory function controlling mitochondrial volume and ion homeostasis. The putative association of LETM1 with a human disease was initially suggested in Wolf-Hirschhorn syndrome, a disorder that results from de novo monoallelic deletion of chromosome 4p16.3, a region encompassing LETM1. Utilizing exome sequencing and international gene-matching efforts, we have identified 18 affected individuals from 11 unrelated families harboring ultra-rare bi-allelic missense and loss-of-function LETM1 variants and clinical presentations highly suggestive of mitochondrial disease. These manifested as a spectrum of predominantly infantile-onset (14/18, 78%) and variably progressive neurological, metabolic, and dysmorphic symptoms, plus multiple organ dysfunction associated with neurodegeneration. The common features included respiratory chain complex deficiencies (100%), global developmental delay (94%), optic atrophy (83%), sensorineural hearing loss (78%), and cerebellar ataxia (78%) followed by epilepsy (67%), spasticity (53%), and myopathy (50%). Other features included bilateral cataracts (42%), cardiomyopathy (36%), and diabetes (27%). To better understand the pathogenic mechanism of the identified LETM1 variants, we performed biochemical and morphological studies on mitochondrial K⁺/H⁺ exchange activity, proteins, and shape in proband-derived fibroblasts and muscles and in Saccharomyces cerevisiae, which is an important model organism for mitochondrial osmotic regulation. Our results demonstrate that bi-allelic LETM1 variants are associated with defective mitochondrial K⁺ efflux, swollen mitochondrial matrix structures, and loss of important mitochondrial oxidative phosphorylation protein components, thus highlighting the implication of perturbed mitochondrial osmoregulation caused by LETM1 variants in neurological and mitochondrial pathologies.

Keywords: LETM1; Wolf-Hirschhorn syndrome; genetics; mitochondria; mitochondrial diseases; neurodegeneration; neurology; oxidative phosphorylation; potassium transport; volume homeostasis.

Graphical abstract

Figure 1

Clinical features and neuroimaging findings of the individuals with bi-allelic LETM1 variants (A) From left to right, facial photos of the affected individuals F1:S1, F1:S2, F5:S1, and F8:S1. All persons wear glasses due to bilateral optic atrophy. All persons have prominent noses. F1:S1 and F1:S2 show long thin faces, low-set ears, and teeth abnormalities. (B) In (i) (F1:S1), severe cerebellar atrophy (arrows) and pontine hypoplasia (arrowheads) are shown, while in (ii) (F6:S1), only mild vermian hypoplasia is noted. In (iii) arrowheads point at the severe optic nerve and chiasm atrophy in 2 different individuals (F5:S1 and F7:S2). Mild ventricular dilatation is present in (iv) (F3:S1). (C) Clinical features of the affected individuals with bi-allelic LETM1 variants. GDD, global developmental delay; ID, intellectual disability; MRI, magnetic resonance imaging; MRC, mitochondrial respiratory chain.

Figure 2

Pedigrees with the segregations of the LETM1 variants and LETM1 protein architecture with a partial sequence alignment of the variants (A) Family trees of the individuals with bi-allelic LETM1 variants. Square, male; circle, female; black symbols, affected individuals; white symbols, unaffected individuals. (B) Schematic representation of the human LETM1 organization in introns, shown as a line, and exons, shown as boxes, and of LETM1 domains as indicated by the residue numbers and the color code: coiled-coil motifs, light yellow; transmembrane helices, blue; LETM/ribosomal-binding like domain, lavender; and putative EF-hands, green. All identified missense variants in the affected individuals (black) and non-pathogenic variants (blue) are mapped according to their positions. The amino acid sequence of human LETM1 was aligned with LETM1 orthologs using Clustal Omega and alignments with LETM1 from other species are shown for all segments that contain missense variants, indicated in bold red letter. Residue conservation is shown below the alignment as fully conserved (^∗), highly conserved (:), or partially conserved (.). UniProt accession numbers for H.s. (H. sapiens), M.m. (M. musculus), S.c. (S. cerevisiae), D.r. (D. rerio), C.e. (C. elegans), D.m. (D. melanogaster), and A.t. (A. thaliana) LETM1 used in this alignment are O95202, Q9Z2I0, Q08179, Q1LY46, Q9XVM0, P91927, and F4J9G6, respectively.

Figure 3

Effects of LETM1 variants on mitochondrial morphology and proliferation in fibroblasts (A) LETM1 variants perturb the mitochondrial network. Confocal images of fibroblasts stained with Mitotracker Red. Shown is a representative overview of the cells (bars 5 μm, except F10 10 μm) and details magnified from the box (bars 5 μm). C1 and C2, healthy donors; F1:S2, c.878T>A (p.Ile293Asn) and c.2094del (p.Asp699Metfs^∗13); F2, c.2220G>C (p.^∗740Tyrext26); F5, c.1072G>A (p.Asp358Asn); F10, c.2071−9C>G (p.Val691fs4^∗). Arrow indicates representative fragmented mitochondria. For statistics, see Figure S1C. (B) LETM1 variants cause swollen mitochondria and loss of cristae. The ultrastructure of control (C1) and affected individual (F5 and F10) fibroblasts was investigated by transmission electron microscopy and images show overviews (left panels, bars 2 μm) and details (right panels, bar 500 nm). Arrow indicates swollen mitochondria. (C and D) Variants differently affect LETM1 stability and OXPHOS proteins in fibroblasts samples. Total lysates of fibroblasts were analyzed by immunoblotting using the indicated antibodies, GAPDH, or β-actin as loading control: C2 and C3, healthy donors; F1:S1 and F1:S2, c.[878T>A; 2094del], p.[Ile293Asn; Asp699Metfs^∗13]; F2, c.2220G>C (p.^∗740Tyrext26); F10, c.2071−9C>G (p.Val691fs4^∗); F11, c.898C>T (p.Pro300Ser) (C). Quantitative graphs from independent experiments representing the protein bands, normalized to the housekeeping proteins, and calculated as a percentage of controls; data are expressed as mean ± SEM (n ≥ 3 independent experiments) (D).

Figure 4

LETM1 variants affect the stability of LETM1 and OXPHOS components in muscle samples (A and B) Western blot analysis of LETM1 and components of the OXPHOS complexes I, II, III; and IV in muscle samples from F11 and quantitative graphs. Total lysates of muscle samples from healthy donors (C4, C5) and F11 c.898C>T (p.Pro300Ser) (S1, S2) were analyzed by immunoblotting using the indicated antibodies; VDAC served as a loading control (A). Quantitative graphs represent the protein levels relative to controls and normalized to VDAC. Data are expressed as mean ± SEM; n ≥ 3 independent experiments (B). (C) Immunohistochemical staining of OXPHOS subunits and VDAC of the muscle of F5 and control subjects. Muscle samples from healthy donors (C6, C7) and F5 c.1072G>A (p.Asp358Asn) were stained for each of the five OXPHOS subunits using the indicated antibodies; VDAC served as a control. Magnification 400×. (D and E) Western blot analysis of subunits of the OXPHOS complexes, citrate synthase, and GAPDH of the muscle of F5 and control subjects. Total lysates of muscle samples were analyzed by immunoblotting using the indicated antibodies; VDAC, GAPDH, and CS served as loading controls. C6, healthy donor; F5, c.1072G>A (p.Asp358Asn) (D). Quantitative graphs representing the protein levels percentage relative to controls (normalized to GAPDH). n ≥ 3 independent experiments.

Figure 5

Functional implication of LETM1 variants on yeast mitochondria (A) LETM1 variants fail to restore KHE activity of yeast letm1Δ. Isolated and de-energized mitochondria were subjected to KOAc and changes of optical density at OD₅₄₀ immediately measured. Left upper panel: representative traces of KOAc-induced swelling in S. cerevisiae LETM1 WT mitochondria (WT, blue) or S. cerevisiae letm1Δ mitochondria overexpressing the empty plasmid (e, yellow) or the plasmid carrying human LETM1 WT untreated (w, green) or treated (wq, gray) with quinine or the human LETM1 variants; color code as indicated in the inserted table: c.754–756del (p.Lys252del) (1, red), c.878T>A (p.Ile293Asn) (2, bottle green), c.881G>A (p.Arg294Gln) (3, aqua), c.898C>T (p.Pro300Ser) (4, dark green), c.913A>C (p.Ile305Leu) (5, lavender), c.1072G>A (p.Asp358Asn) (6, violet), c.1139G>C (p.Arg380Pro) (7, beige), c.1178G>A (p.Arg393His) (8, turquois), c.1760A>G (p.Lys587Arg) (9, mauve), c.2071−9C>G (p.Val691fs4^∗) (10, purple), c.2094del (p.Asp699Metfs^∗13) (11, dark blue), compound (12, lilac), c.2220G>C (p.^∗740Tyrext26) (13, olive). Right upper panel: quantified rates of KOAc-induced swelling amplitudes (t = 60 s) from 3 independent experiments. An overview of the swelling rate is given in Figure S5. One-way ANOVA with Dunnett’s multiple comparisons test performed against S. cerevisiae letm1Δ transformed with empty pVT-103U plasmid ^∗p = 0.0426, ^∗∗p = 0.0026, ^∗∗∗p = 0.0006, ^∗∗∗∗p < 0.0001. And for p.Ile305Leu and p.Lys587Arg relatively to S. cerevisiae letm1Δ transformed with WT, ns > 0.05, ^∗p = 0.0169. (B) Ectopic expression of LETM1 variants in S. cerevisiae letm1 Δ. Isolated mitochondria (upper panel) and total protein lysates (left lower panel) from the same strains as in (A). Subcellular fractions (T, total; SN, post-mitochondrial supernatant; M, mitochondria) (right lower panel) were immunoblotted using the indicated antibodies; Por1p and Act1p served as mitochondrial and total (and SN) loading control, respectively.