Phenotypic and genetic spectrum of ATP6V1A encephalopathy: a disorder of lysosomal homeostasis

Abstract

Vacuolar-type H+-ATPase (V-ATPase) is a multimeric complex present in a variety of cellular membranes that acts as an ATP-dependent proton pump and plays a key role in pH homeostasis and intracellular signalling pathways. In humans, 22 autosomal genes encode for a redundant set of subunits allowing the composition of diverse V-ATPase complexes with specific properties and expression. Sixteen subunits have been linked to human disease. Here we describe 26 patients harbouring 20 distinct pathogenic de novo missense ATP6V1A variants, mainly clustering within the ATP synthase α/β family-nucleotide-binding domain. At a mean age of 7 years (extremes: 6 weeks, youngest deceased patient to 22 years, oldest patient) clinical pictures included early lethal encephalopathies with rapidly progressive massive brain atrophy, severe developmental epileptic encephalopathies and static intellectual disability with epilepsy. The first clinical manifestation was early hypotonia, in 70%; 81% developed epilepsy, manifested as developmental epileptic encephalopathies in 58% of the cohort and with infantile spasms in 62%; 63% of developmental epileptic encephalopathies failed to achieve any developmental, communicative or motor skills. Less severe outcomes were observed in 23% of patients who, at a mean age of 10 years and 6 months, exhibited moderate intellectual disability, with independent walking and variable epilepsy. None of the patients developed communicative language. Microcephaly (38%) and amelogenesis imperfecta/enamel dysplasia (42%) were additional clinical features. Brain MRI demonstrated hypomyelination and generalized atrophy in 68%. Atrophy was progressive in all eight individuals undergoing repeated MRIs. Fibroblasts of two patients with developmental epileptic encephalopathies showed decreased LAMP1 expression, Lysotracker staining and increased organelle pH, consistent with lysosomal impairment and loss of V-ATPase function. Fibroblasts of two patients with milder disease, exhibited a different phenotype with increased Lysotracker staining, decreased organelle pH and no significant modification in LAMP1 expression. Quantification of substrates for lysosomal enzymes in cellular extracts from four patients revealed discrete accumulation. Transmission electron microscopy of fibroblasts of four patients with variable severity and of induced pluripotent stem cell-derived neurons from two patients with developmental epileptic encephalopathies showed electron-dense inclusions, lipid droplets, osmiophilic material and lamellated membrane structures resembling phospholipids. Quantitative assessment in induced pluripotent stem cell-derived neurons identified significantly smaller lysosomes. ATP6V1A-related encephalopathy represents a new paradigm among lysosomal disorders. It results from a dysfunctional endo-lysosomal membrane protein causing altered pH homeostasis. Its pathophysiology implies intracellular accumulation of substrates whose composition remains unclear, and a combination of developmental brain abnormalities and neurodegenerative changes established during prenatal and early postanal development, whose severity is variably determined by specific pathogenic variants.

Keywords: ATP6V1A; developmental delay; epileptic encephalopathy; lysosomal disorder; progressive brain atrophy.

Figure 1

Brain MRI in patients with ATP6V1A pathogenic variants. MRIs of patients who were imaged at least twice are shown in A–G. Images were taken from the initial and last follow-up investigations at 1.5 to 3 T and include T₁- or T₂-weighted and FLAIR sequences. Structural abnormalities include a combination of cerebellar and brainstem atrophy, dilated ventricles and subarachnoid spaces, thinning of the corpus callosum and hypomyelination. Comparison of initial and follow-up images demonstrates various rates of progression and anatomic involvement. In Patient (Pat.) 8 (A), at age 1 year 6 months, hypomyelination was the only feature, while cerebellar atrophy was present at 7 years 6 months. In Patients 9 (B) and 13 (E), MRI scan at age 6 months and 3 months, respectively, showed severe signs of atrophy that involved the cerebral cortex and white matter in Patient 9 and all brain, brainstem and cerebellar structures in Patient 13. Six months later, in Patient 9, at 12 months of age, atrophy had remarkably progressed, yet with minimal cerebellar involvement; in Patient 13, at 9 months, dramatic generalized shrinking of all brain structures had occurred. In Patient 15 (F), severe hypomyelination was present 1 week after birth followed, 4 months later, by severe brain atrophy. In Patient 24 (G) moderate brain atrophy with hypomyelination was present at 10 months, which became more severe at 2 years 9 months. In Patients 10 (C) and 11 (D), mild atrophy became apparent from age 3 to 5 years (Patient 10) and 11 months to 2 years 3 months (Patient 11). Coronal sections of the brain of six patients who had only one MRI scan are presented in H. These images, taken at different ages, show hypomyelination in all patients and minor atrophic changes in two (Patients 6 and 7).

Figure 2

Schematic representation of the ATP6V1A protein. The structure of ATP6V1A includes the ATP-synt_ab_N (light blue), ATP_synt_ab_Xtn (purple) and ATP-synt_ab (light orange) domains. The red lollipops show the location of the pathogenic variants identified in the patients in this series (patients’ identifiers are reported in grey between brackets), whereas the blue lollipops represent the missense variants reported in the gnomAD database.

Figure 3

Cryo-EM structure of the Atp6v1a. (A) The cryo-EM structure of the Atp6v1a subunit from the mammalian V-ATPase (PDB code 6VQ9). Pathogenic variant sites are drawn as red, yellow, green and orange spheres if they are near the phosphate binding loop (p-loop, blue spheres), A/B interface, A/D or non-catalytic A/B interface and uncharacterized locations, respectively. The pink sphere is the magnesium ion, and the bound ADP molecule is drawn as sticks. (B and C) Structure of the V1 domain from upper (B) and lateral (C) views. One A subunit is shown as grey tube and the other two as grey surface, and B and D subunits are shown as ochre and blue surfaces, respectively. Pathogenic variant sites are drawn as in A.

Figure 4

ATP6V1A/LAMP1 expression and LysoTracker labelling in patient-derived fibroblasts. (A) Left: Representative western blots from fibroblasts lysates of patients (Patients 1, 2, 16 and 18) and the respective controls (Ctrl 1, 2 16 and 18). Right: Quantification of ATP6V1A expression levels normalized on GAPDH signal and expressed, for each patient, as percentage of the respective control. Data are means ± SEM from five independent experiments. *P < 0.05; Kruskal–Wallis/Dunn’s tests. (B) Left: Representative images of fibroblasts incubated with Lysotracker (red) and stained with phalloidin (white). Higher magnifications of the fields labelled by dotted squares are shown in the insets. Scale bar = 20 µm. Right: Quantification of Lysotracker intensity. The fluorescence signal was measured in the cell body identified by phalloidin staining. Each dot represents the mean fluorescence intensity of a single cell. Data are from 26/17 cells for Ctrl 1/Patient 1, 27/36 cells for Ctrl 2/Patient 2, 39/38 cells for Ctrl 16/Patient 16, 22/20 cells for Ctrl 18/Patient 18. (C) Left: Representative images of fibroblasts double stained with LAMP1 (green) and phalloidin (white). Higher magnifications of the fields labelled by dotted squares are shown in the insets. Scale bar = 20 µm. Right: Quantification of LAMP1 intensity. Immunoreactivity was measured in the cell body identified by phalloidin labelling. Each dot represents the mean fluorescence intensity of a single cell. Data are from 26/19 cells for Ctrl 1/Patient 1, 13/20 cells for Ctrl 2/Patient 2, 19/15 cells for Ctrl 16/Patient 16, 22/18 cells for Ctrl 18/Patient 18. *P < 0.05, **P < 0.01, ***P < 0.01; unpaired Student’s t-test/Mann–Whitney U-test.

Figure 5

Evaluation of lysosomal pH in patient-derived fibroblasts. (A) Representative images of fibroblasts incubated with LysoSensor yellow/blue dextran and visualized in live at 340 and 380 nm excitation. White lines represent cells detected in bright field and used as regions of interest for intensity measurement. Scale bar = 10 µm. (B) The calibration curve, obtained by plotting the fluorescence intensity 380/340 ratios as a function of pH, was fitted with linear regression. Data are means ± SEM from the five cells shown in A. (C) pH value derived from 340/380 ratio and relative calibration curve. Each dot represents the mean pH from a single coverslip. 5–14 coverslips have been analysed for each experimental group with an average of 14 cells analysed for each coverslip. *P ≤ 0.001; **P ≤ 0.0001 unpaired Student’s t-test/Mann–Whitney U-test.

Figure 6

Ultrastructural analysis of patient-derived fibroblasts. (A–C) Control fibroblasts. (D–Q) Fibroblasts from Patient 1 (D and E), Patient 2 (G–I), Patient 16 (L–N) and Patient 18 (O–Q). Fibroblasts from patients bearing ATP6V1A pathogenic variants showed several cytoplasmic single membrane-bounded vacuoles filled with heterogeneous substances, resembling autolysosomes (arrows in D, E, G, H, L, M, O and P). Vacuoles from patients’ fibroblasts were filled with various substances, such as lamellated membrane structures (arrowheads in F and N), osmiophilic material (asterisks in F, I and Q), electron-dense granular material (g in I) and substances with different electron-density (Q). N = nucleus; n = nucleolus; rer = rough endoplasmic reticulum; m: mitochondria; ly = lysosome; L = lipid droplets; arrows = cytoplasmic single membrane-bounded vacuoles; arrowheads = lamellated membrane structures; asterisks = osmiophilic material; g = electron-dense granular material. Scale bars = 2 µm (A, D, G, L and O), 1 µm (B and C), 500 nm (E, H, M and P), 200 nm (F, I, N and Q).

Figure 7

Ultrastructural analysis of patient-derived iNs. (A–D) Representative electron micrographs showing somata and proximal dendrites of two iPSC-derived controls (Ctrls 1 and 18; A and C), and Patient 1 (B) and Patient 18 (D) iNs (highlighted in red). [A(i and ii) and C(i and ii)] High magnification images of the boxed regions in A and C, respectively. Asterisks indicate lysosomes in control iNs. [B(i and ii) and D(i and ii)] High magnification images of the boxed regions in B and D, respectively. Note the abundance of lysomes in Patient 1 (B) and Patient 18 (D) iN somata compared to control iNs. Asterisks indicate lysosomes with dense concentric lamellae, a number sign (#) points to irregular shaped lysosomes filled with heterogeneous material and lipid droplets, and arrows point to small round-shaped lysosomes filled with electron-dense material. (E) Large lysosome in a control iN. Note the homogenous and round structure of the organelle. (F) Lysosomes (Lys) in the iN soma of Patient 1. (G) Lysosomes with heterogeneous material and lipid droplets in the iN soma of Patient 18. Note the irregular shape of lysosomes filled with concentric lamellae. (H, left) Lysosomal areas in iNs from four control subjects [n = 2 independent iPSC-derived iN subclones for Ctrl 1 (-c1/c2) and Ctrl 18 (-c1/c2) and one subclone from two independent hESC-derived iNs (ESC_Ctrl 1 and 2)] and two patients [n = 2 independent iPSC-derived iN subclones per each subject (Patient 1−c1/c2 and Patient 18-c1/c2)]. Data are presented as medians with 95% confidence intervals (CIs). (H, right) Pooled area of lysosomes for control and patient samples (n_control = 2429 lysosomes from six samples; n_patient = 2918 lysosomes from four samples; P < 0.001, Mann–Whitney’s U-test). Data are presented as medians with 95% CIs. (I, left) Density of lysosomes in iNs from four control subjects [n = 2 independent iPSC-derived iN subclones for Ctrl 1 (-c1/c2) and Ctrl 18 (-c1/c2) and one subclone from two independent hESC-derived iNs (ESC_Ctrl 1-c1 and ESC_Ctrl 2-c1) and two patients (n = 2 independent iPSC-derived iN subclones per each subject (Patient 1-c1/c2 and Patient 18-c1/c2)]. Data are presented as medians with 95% CIs. (I, right) Pooled densities of lysosomes in control and patient samples (n_control = 58 cells from six samples; n_patient = 29 cells from four samples; P = 0.702, Mann–Whitney’s U-test). Data are presented as medians with 95% CIs. Lys = lysosome, Ctrl = control; Pat = patient; ns = not significant, *** P < 0.001. Scale bars = 5 µm (A–D), 500 nm (Ai–Di, Aii–Dii, G), 200 nm (E and F).