1 
De novo VPS4A mutations cause multisystem disease with abnormal neurodevelopment 
Catherine Rodger,1,2,25 Elisabetta Flex,3,25 Rachel J. Allison,1,2,25 Alba Sanchis-Juan,4,5,25 
Marcia A Hasenahuer,2,6 Serena Cecchetti,7 Courtney E. French,2 James R. Edgar,1,8 
Giovanna Carpentieri,3,9 Andrea Ciolfi,9 Francesca Pantaleoni,9 Alessandro Bruselles,3 
Genomics England Research Consortium,10 Roberta Onesimo,11 Giuseppe Zampino,11,12 
Francesca Marcon,13 Ester Siniscalchi,13 Melissa Lees,14 Deepa Krishnakumar,15 Emma 
McCann,16 Dragana Yosifova,17 Joanna Jarvis,18 Michael C. Kruer,19 Warren Marks,20 
Jonathan Campbell,21 Louise E. Allen,22 Stefano Gustincich,23,24 F Lucy Raymond,1,2,26 
Marco Tartaglia,9,26* and Evan Reid1,2,26** 
 
1. Cambridge Institute for Medical Research, University of Cambridge, Cambridge, 
CB2 0XY, UK. 
2. Department of Medical Genetics, University of Cambridge, Cambridge, CB2 0QQ, 
UK.  
3. Department of Oncology and Molecular Medicine, Istituto Superiore di Sanità, Rome, 
00161, Italy 
4. Department of Haematology, NHS Blood and Transplant Centre, University of 
Cambridge, Cambridge, CB2 0XY, UK 
5. NIHR BioResource, Cambridge University Hospitals NHS Foundation Trust, 
Cambridge Biomedical Campus, CB2 0QQ, Cambridge, UK 
6. European Molecular Biology Laboratory – European Bioinformatics Institute 
(EMBL-EBI), Wellcome Genome Campus, Hinxton, Cambridgeshire, CB10 1SA, 
UK 
7. Microscopy Area, Core Facilities, Istituto Superiore di Sanità, Rome, 00161, Italy 
8. Department of Pathology, University of Cambridge, CB2 1QP, UK 
 2 
9. Genetics and Rare Diseases Research Division, Ospedale Pediatrico Bambino Gesù, 
IRCCS, Rome, 00146, Italy  
10. Genomics England, London UK 
11. Fondazione Policlinico Universitario A. Gemelli -IRCCS, Rome, 00168, Italy  
12. Università Cattolica del Sacro Cuore, Rome, 00168, Italy 
13. Unit of Mechanisms, Biomarkers and Models, Department of Environment and 
Health, Istituto Superiore di Sanità, Rome, 00161, Italy 
14. Department of Clinical Genetics, Great Ormond Street Hospital, London, WC1N 3JH, 
UK  
15. Department of Paediatric Neurology, Cambridge University Hospitals NHS 
Foundation Trust, Cambridge, CB2 0QQ, UK 
16. Department of Clinical Genetics, Liverpool Women’s Hospital, Liverpool, L8 7SS, 
UK 
17. Department of Medical Genetics, Guys’ and St Thomas’ NHS Foundation Trust, 
London, SE1 9RT, UK 
18. Clinical Genetics, Birmingham Women’s and Children’s NHS Foundation Trust, 
Birmingham, B15 2TG, UK. 
19. Phoenix Children's Hospital, Phoenix, Arizona, 76109, USA 
20. Cook Children’s Medical Centre, Fort Worth, Texas, 76104, USA 
21. Colchester Hospital, East Suffolk and North Essex NHS Foundation Trust, Essex, 
CO4 5JL, UK 
22. Ophthalmology Department, Cambridge University Hospitals NHS Foundation Trust, 
Cambridge, CB2 0QQ, UK 
23. Department of Neuroscience and Brain Technologies, Istituto Italiano di Tecnologia, 
Genova, 16163, Italy. 
 3 
24. Area of Neuroscience, SISSA, Trieste, 34136, Italy. 
 
25.  These authors contributed equally to this work  
26. These authors contributed equally to this work 
 
 
 
Correspondence: 
* Email: marco.tartaglia@opbg.net  
** Email: ealr4@cam.ac.uk 
  
 4 
Abstract 
The endosomal sorting complexes required for transport (ESCRTs) are essential for multiple 
membrane modelling and membrane-independent cellular processes. Here we describe 6 
unrelated individuals with de novo missense variants affecting the ATPase domain of 
VPS4A, a critical enzyme regulating ESCRT function. Probands had structural brain 
abnormalities, severe neurodevelopmental delay, cataracts, growth impairment and anaemia. 
In cultured cells, over-expression of VPS4A mutants caused enlarged endosomal vacuoles 
resembling those induced by expression of known dominant-negative ATPase-defective 
forms of VPS4A. Proband-derived fibroblasts had enlarged endosomal structures with 
abnormal accumulation of the ESCRT protein IST1 on the limiting membrane. VPS4A 
function was also required for normal endosomal morphology and IST1 localisation in iPSC-
derived human neurons. Mutations affected other ESCRT-dependent cellular processes, 
including regulation of centrosome number, primary cilium morphology, nuclear membrane 
morphology, chromosome segregation, mitotic spindle formation and cell cycle progression. 
We thus characterize a distinct multisystem disorder caused by mutations affecting VPS4A, 
and demonstrate that its normal function is required for multiple human developmental and 
cellular processes.  
 5 
Introduction 
The endosomal sorting complexes required for transport (ESCRTs) are multifunctional 
membrane modelling machineries that drive membrane fission or constriction in cellular 
processes that involve “inside out” membrane topology.1-3 These are exemplified by fission 
reactions that cause vesicle budding away from the cytoplasm, in which ESCRT-III 
complexes assemble on the inner cytosolic face of a vesicle neck and promote membrane 
constriction from the inside. Modification of ESCRT-III complexes drives fission, and this is 
performed by the catalytic activity of members of the VPS4 ATPase family (which in 
vertebrates comprises two paralogues, VPS4A and VPS4B) – thus VPS4 is an indispensable 
component of all ESCRT-related membrane modelling.4 Processes that involve this type of 
membrane topology include formation of the late endosomal multivesicular body (MVB), 
nuclear envelope reformation and the abscission stage of cell division, amongst others.1-3 In 
addition, certain ESCRT-III-associated proteins are active in more conventional “outside in” 
fission, notably in endosomal tubule fission, where atypical ESCRT-III proteins constrict 
from the outside to promote fission of sorting tubules from the endosomal body.5; 6 
Study of the ESCRT complexes and VPS4 in the endocytic pathway has informed 
mechanistic understanding of their role in membrane modelling.2; 3; 7-9 Endocytosis regulates 
the cell surface concentration of plasma membrane proteins and so controls multiple critical 
cellular processes. After endocytosis from the cell surface, membrane proteins are trafficked 
to the early sorting endosome, from where they may be sorted away from the endosomal 
system (e.g. to the plasma membrane) or retained for degradation in the late endosome-
lysosome pathway.10 Membrane proteins to be degraded are exposed to the lumenal 
degradative compartment of the late endosome and lysosome by a process involving inward 
budding of the endosomal limiting membrane to form the intralumenal vesicles (ILVs) of the 
late endosome or multivesicular body (MVB). Concentration and sorting of cargoes into, and 
 6 
formation of, the ILVs is accomplished by the action of the ESCRT- 0, I, II and III 
complexes.3; 7-9 In the final stage of this process, ILVs are released into the MVB lumen by 
the ESCRT-III complex, comprising a number of charged multivesicular body proteins 
(CHMPs) that are recruited from monomeric cytosolic pools to form a filamentous structure 
inside the ILV neck. Complex formation is accompanied by a conformational change in the 
CHMP proteins that exposes C-terminal motifs that bind to MIT (Microtubule-interacting and 
trafficking) domains in interacting proteins such as VPS4, promoting their endosomal 
recruitment.11-14 VPS4 functions as a hexameric ring which, using energy from ATP-
hydrolysis, modifies the ESCRT-III complex filaments by unfolding subunits through its 
central pore – this subunit removal has been proposed to constrict ESCRT-III filaments and 
tighten the ILV neck.1; 4  
ESCRT-III proteins and VPS4 also have roles unrelated to membrane modelling, participating 
in the dynamic control of mitotic spindle morphology and mitotic spindle checkpoint 
signalling, as well as in multiple aspects of centrosome biology and primary cilium formation. 
Cells lacking many different individual ESCRT-III or VPS4 proteins develop aberrant nuclei 
composed of fragmented or interconnected micronuclei, an increased number of centrosomes, 
multipolar spindles and abnormal chromosome alignment during metaphase.15 In addition, 
VPS4 dynamically localises to centrosomes and regulates centrosome function, position, 
number and morphology.15; 16 Related to this, loss of VPS4 or over-expression of a dominant 
negative ATPase-defective VPS4 mutant has also been linked to reduced primary cilium 
formation independent of ESCRT-III, via a mechanism proposed to involved disrupted 
centriolar satellite assembly at the centrosome.16 
Although mutations in the ESCRT-III proteins CHMP1A, CHMP2B and CHMP4B cause 
autosomal recessive pontocerebellar hypoplasia (MIM 614961), autosomal dominant 
amyotrophic lateral sclerosis - frontotemporal dementia (MIM 614696) and autosomal 
 7 
recessive cataract (MIM 605387) respectively, somewhat surprisingly other ESCRT-III-
related proteins have not been linked to genetic disease.17-19 In this study, we have identified 
and functionally characterised multiple de novo heterozygous missense mutations in VPS4A 
(MIM 609982), which cause a severe neurodevelopmental disorder characterised by severe 
hypotonia and developmental delay (DD), intellectual disability (ID), structural brain 
abnormalities including thin corpus callosum and ponto-cerebellar hypoplasia, extrapyramidal 
neurological dysfunction, congenital cataracts with visual dysfunction, sensorineural deafness 
and haematological abnormalities, providing evidence of an essential function of this ATPase 
in multiple cellular and developmental processes in humans.  
 8 
Materials and Methods 
Subjects 
Clinical data and DNA specimens were collected, stored and used following procedures in 
accordance with the ethical standards of the declaration of Helsinki protocols, with signed 
informed consents from the participating subjects or families. The study was approved by the 
local Institutional Ethical Committee of the Ospedale Pediatrico Bambino Gesù, Rome 
(1702_OPBG_2018) and the Cambridge South Research Ethics Committee (13/EE/0325). 
All probands except proband 5 were analysed in the context of dedicated research projects 
focused on undiagnosed disorders, while proband 5 was referred for diagnostic genetic 
testing. Explicit permission was obtained to publish the photographs of the subjects shown in 
Fig. 1.  
Genomic analyses 
WES and WGS was performed using DNA samples obtained from leukocytes. A trio-based 
strategy was used in all cases. WES and WGS data processing, and variant filtering and 
prioritization by allele frequency, predicted functional impact, and inheritance models were 
performed as previously reported.20-22 The de novo origin of the VPS4A mutations was 
confirmed by Sanger sequencing (primer sequences available on request). 
Protein sequence conservation and mapping of variants to homologous VPS4 protein 
structures 
Sequence conservation of VPS4 proteins was analysed across orthologous and paralogous 
protein sequences for different model species as previously described.23 Constrained coding 
regions model was run for gnomAD 2.1.1 exomes.24 
There are protein structures homologous to human VPS4A in hexameric or monomeric forms 
in Protein Data Bank (PDB).25 Because the hexamer is the functional form of VPS4A, first 
 9 
we mapped all the de novo missense variants to the homologous yeast VPS4, since it is the 
most similar structure to human available in this oligomeric state. This structure has ADP 
bound and a cyclic peptide in the central pore (PDB ID 6OO2, cryo-EM with resolution 
4.4Å).26 Then, we compared our observations to the monomeric forms using the homologous 
human and mouse VPS4B, all X-ray crystal structures. These were in apo forms (human PDB 
ID 1XWI, with resolution 2.8Å; mouse PDB ID 2ZAM, with resolution 3.5Å) and in ATP or 
ADP bound forms (mouse PDB ID 2ZAN and 2ZAO, with resolution 2.8Å and 3.2 Å 
respectively).27; 28 All structures were visualized and aligned with PyMOL version 2.0 (The 
PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC). 
Antibodies  
Antibidies used: rabbit polyclonal anti-beta III tubulin (ab18207), rabbit polyclonal anti-
CHMP2B (ab33174), mouse monoclonal anti-M6PR (ab2733), mouse monoclonal anti-
MAP2 (ab11268), rabbit polyclonal anti-pericentrin (ab4448), rabbit polyclonal anti-Tau 
(ab64193) from Abcam; mouse monoclonal anti-CD63 (clone H5C6; Developmental Studies 
Hybridoma Bank, University of Iowa); rabbit polyclonal anti-Cathepsin D (219361), rabbit 
polyclonal anti-Histone H2A.X (07-627), mouse monoclonal anti-myc (clone 4A6, 05-724) 
from EMD Millipore; mouse monoclonal anti-TfnR (13-6800), mouse monoclonal Anti-
BrdU (B35130) from Invitrogen; IRDye-conjugated secondary antibodies from LICOR; 
rabbit polyclonal anti IST1 (51002-1-AP) from Proteintech Group; mouse monoclonal anti-
SNX1 (611582), anti-EEA1 (610456), anti-Rab5 (610742) from BD transduction 
laboratories; mouse monoclonal anti-CHMP6 (clone B-3; sc-398963), mouse monoclonal 
anti-EGFR (clone A-10, sc-373746), mouse monoclonal anti-LAMP1 (H4A3), mouse 
monoclonal anti-lamin A/C (E-1: sc-376248), mouse monoclonal anti-acetylated α-tubulin 
(sc-23950), mouse monoclonal VPS4A (clone A-11, sc-393428) from Santa Cruz 
Biotechnology, Inc; horseradish peroxidase (HRP)-conjugated secondary antibodies, mouse 
 10 
monoclonal anti-β-tubulin (T4026) from Sigma-Aldrich; rabbit polyclonal anti-GAPDH 
(2118), mouse monoclonal anti-Nanog (clone 1E6C4, 4893), rabbit polyclonal anti-oct-4 
(2750), rabbit polyclonal anti-Sox2 (2748) from Cell Signalling Technology; Alexa Fluor 
488-, 594- and 568-labelled secondary antibodies for immunofluorescence from Molecular 
Probes. 
Constructs 
Lentiviral plasmids used the A62 backbone and the packaging plasmids pMD VSV-G and 
pCMV Δ8.91. A62-myc-VPS4A was generated by cloning VPS4A from pLNCX2-mCherry-
VPS4A into the A62 vector (NheI-EcoRI) with the addition of an N-terminal myc tag. Mutant 
versions of A62-myc-VPS4A (p.R284G, p.R284W, p.E228Q, p.E206K, p.I337V and 
p.P168S) were generated by site-directed mutagenesis. sgRNA sequences targeting the 
transcription start site of VPS4A were selected from the Weissman CRISPRi-v2 library.29 
Sense and antisense sgRNA oligonucleotides were designed with 5’CACC and 3’CAAA 
overhangs, respectively and cloned into pKLV-U6gRNA-EF(BbsI)-PGKpuro2ABFP (BpiI) 
for lentivirus production. The following sgRNA were used: 
Scrambled: GGGACGCGAAAGAAACCAGT 
VPS4A G1: GGCAGGGCGGCCGCTCGCAG 
VPS4A G2: GACTCGGCTCCCGCTGCGAG 
VPS4A G3: GGGAGCCGAGGTACTGGGTC 
pLNCX2-mCherry-VPS4A was a gift from Sanford Simon (Addgene plasmid # 115334; 
RRID:Addgene_115334). pKLV-U6gRNA-EF(BbsI)-PGKpuro2ABFP was a gift from 
Kosuke Yusa (Addgene plasmid # 50946; RRID:Addgene_50946). A62 was a gift from 
Michael Fernandopulle (University of Cambridge, UK).  
Cell Culture 
 11 
Proband fibroblasts, HeLaM and HEK-293T cells were grown in complete Dulbecco’s 
Modified Eagle’s Medium (DMEM, Sigma-Aldrich) supplemented with 10% foetal bovine 
serum (FBS, Sigma-Aldrich), 100 U/mL penicillin and 100 µg/mL streptomycin (Sigma-
Aldrich) and 2 mM L-Glutamine (Sigma-Aldrich). Human i3N and CRISPRi-i3N iPSCs 
(generated in a WTC11 iPSC background line) were a gift from Michael Ward (NIH, 
Bethesda, USA). iPSCs were cultured in TESR-E8 (STEMCELL Technologies) on dishes 
coated with Matrigel Matrix (Corning). TESR-E8 was replaced daily and cells were passaged 
at 80%-90% confluency with 0.5 mM EDTA to maintain colony growth and with the ROCK 
inhibitor Y-27632 (10 µM, Tocris). CRISPRi-i3N iPSCs stably expressing pKLV were 
additionally cultured in the presence of 2 µg/mL puromycin (Sigma-Aldrich). All cell lines 
were cultured with 5% CO2 at 37°C and were regularly tested for mycoplasma contamination.  
HeLa Cell transfection 
HeLa cells were transfected with purified plasmid using polyethylenimine (PEI, Sigma-
Aldrich). In brief, a mixture of 50 µL optiMEM and 1.69 µg of DNA was prepared and 
incubated at room temperature for 5 minutes. Another mixture containing 150 µL optiMEM 
and 3.38 µL PEI was prepared and incubated for 5 minutes. Both solutions were then mixed 
together and incubated for 20 minutes. The total volume was then added to one well of a 6-
well plate, already containing 1.5 mL of DMEM supplemented with 10% FBS and 2 mM L-
Glutamine. Cells were typically transfected 24 h after plating and incubated with transfection 
reagents for 29 h. 
I3Neuron differentiation 
Differentiation into i3Neurons was as previously described, with slight modifications.30 
Briefly, on Day 0 iPSCs were dissociated into single cells using StemPro Accutase (Thermo 
Fisher Scientific) and seeded at a density of 150,000 cells/cm2 on Matrigel-coated culture 
 12 
dishes in Induction Medium (IM) composed of DMEM/F-12, 1X N-2 Supplement, 1X MEM 
Non-Essential Amino Acids Solution, 1X GlutaMAX Supplement (all Thermo Fisher 
Scientific), 10 µM Y-27632 and 2 µg/mL doxycycline hydrochloride (Sigma-Aldrich). Pre-
differentiated cells were maintained in IM for 3 days with daily medium changes. After the 3-
day differentiation period, cells were dissociated with StemPro Accutase and seeded at 5 x 
104 cells/cm2 onto culture plates coated with 0.1 mg/mL poly-L-ornithine (Sigma-Aldrich). 
Cells were maintained in Cortical Neuron Culture Medium, composed of BrainPhys 
Neuronal Medium (STEMCELL Technologies), 1X B-27 Supplement (Thermo Fisher 
Scientific), 10 ng/mL BDNF (PeproTech), 10 ng/mL NT-3 (PeproTech) and 1 µg/mL mouse 
Laminin (Thermo Fisher Scientific) with half media changes carried out every 3-4 days. 
Stable cell lines 
Stable cell lines were generated by lentiviral transduction of iPSCs with the VPS4A and 
sgRNA lentivectors described earlier. Briefly, HEK-293T cells were co-transfected with a 
lentiviral expression construct and the packaging vectors pCMVΔ8.91 and pMD VSV-G at a 
ratio of 1:0.7:0.3 using TransIT-293 (Mirus Bio) as per the manufacturer’s instructions. The 
viral supernatant was collected 48 h post-transfection, passed through a 0.45 µm filter, and 
added to target cells in the presence of 10 µg/mL polybrene (Sigma-Aldrich). Typically, 
following spinoculation at 1800 rpm for 1h at 32°C, cells were transduced for 16 h. 
Transduced cells were selected by adding puromycin at a final concentration of 1 µg/mL 
from 24 h if required. 
Immunoblotting 
Cells were washed twice on ice with PBS and subsequently scraped with ice-cold Triton X-
100 lysis buffer (1% Triton X-100, 150 mM NaCl, 50 mM HEPES pH 7.4, 1 mM EDTA, 
10% (v/v) glycerol and protease inhibitors). Samples were centrifuged at 20,000 g for 10 
 13 
minutes at 4°C. Sample buffer was added to supernatant and samples were heated at 95°C for 
5 minutes. Proteins were resolved by SDS-PAGE and transferred to a PVDF membrane. 
Membranes were blocked in 5% (w/v) skimmed milk powder in PBS containing 0.1% Tween 
20 for 30 minutes at room temperature before being probed with primary and secondary 
antibodies. Membranes were visualised using an ECL Western Blotting Detection Kit (GE 
Healthcare) for HRP-conjugated antibodies or, for IRDye-conjugated secondary antibodies, 
imaged directly for infrared fluorescence signal detection on an Odyssey Infrared Imaging 
System using LICOR Image Studio software (LICOR, US). Western blots were quantified by 
densitometry using ImageJ. 
Analysis of centrosome and mitotic spindle number, morphology and chromosome 
segregation 
After 24 h of culture in complete medium, fibroblasts were treated with 2 mM thymidine 
(Sigma-Aldrich) for 24 h, washed with PBS 1X, recovered with complete medium for 3 h and 
then treated with 100 ng/ml nocodazole (Sigma-Aldrich) for 12 h. Afterwards, fresh drug-free 
medium was added and recovery was allowed for the different time points (15 to 120 minutes) 
by fixing cells every 15 minutes using PHEMO buffer for 10 min at room temperature.  
Primary cilium analysis 
Cells were plated onto coverslips, maintained for 24 h in low serum medium to promote 
emission of cilia and then fixed in absolute chilled methanol for 10 minutes at -20°C.  
BrdU assay 
Assessment of cells in the different cell cycle phases was performed by dual flow cytometry 
analysis of cells incorporating BrdU and stained with the fluorescent DNA probe propidium 
iodide (PI). Briefly, cells were incubated for 1 h with BrdU (Sigma-Aldrich) at a final 
concentration of 30 µM. Then, BrdU was removed, cells were rinsed with PBS prior 
 14 
harvesting, and permeabilized using ice-cold 100% ethanol. Cells were incubated with HCl 
3N to denature DNA, and 0.1 M sodium tetraborate to stop this reaction. Finally, fibroblasts 
were incubated with an anti-BrdU antibody followed by goat anti-mouse Alexa Fluor 488 
secondary antibody. Cells were then re-suspended in a buffer containing 10 µg/ml RNase A 
and 20 µg/ml PI and immediately analysed by FACS. 
Cytokinesis-block micronucleus and chromosome aberration assays 
The cytokinesis-block micronucleus assay was conducted following a previous protocol.31 In 
actively dividing cells, cytokinesis was blocked with 4.5 µg/ml cytochalasin B (Sigma-
Aldrich), an inhibitor of actin polymerization. Twenty four hours later, cells were collected by 
cytospin centrifugation (Shandon Cytospin 3, Thermo Fisher Scientific) at 600 r.p.m for 5 
min and fixed in absolute methanol at −20°C for 10 min. Slides were stained with 3% Giemsa 
(Sigma-Aldrich) in Sorensen buffer, pH 6.8, and the analysis was performed by using an 
optical microscope. Cells were analysed following previously reported criteria.31  
Chromosome aberrations were analysed in mitotic cells obtained from actively dividing cells 
treated for 2 hours with colcemid (0.1 µg/ml, Sigma-Aldrich). Cells were harvested by 
standard procedures. Briefly, after 10 min incubation at 37°C in 0.075 M KCl, fibroblasts 
were fixed three times with cold methanol/acetic acid (3:1). Slides were prepared by a 
conventional air-drying technique and stained with 5% Giemsa in Sorensen buffer pH 6.8.  
Immunofluorescence microscopy on fixed cells 
Cells were fixed at room temperature in 3.7% (v/v) formaldehyde in PBS and permeabilised 
in PBS containing 0.1% (v/v) saponin (Sigma-Aldrich) or 0.1% (v/v) Triton X-100 (Sigma-
Aldrich). Coverslips were labelled with primary and secondary antibodies as previously 
described.32 Slides were analysed with a LSM980 confocal microscope (63x NA 1.40 oil 
immersion objective, 37°C), LSM880 confocal microscope (100x or 63x NA 1.40 oil 
 15 
immersion objective, 37°C), LSM780 confocal microscope (63x NA 1.40 oil immersion 
objective, 37°C), Leica TCS SP2 AOBS confocal microscope (63x NA 1.40 oil immersion 
objective, 37°C) or an AxioImager Z2 Motorized Upright Microscope (100x or 63x NA 1.40 
oil immersion objective, room temperature, Axiocam 506; ZEISS). Images were 
subsequently processed using Huygens Professional software for deconvolution, ImageJ, 
Adobe Photoshop and Adobe Illustrator.  
Image analysis and quantitation 
To determine the percentage of cells with large marker-positive endosomes, the largest 
organelle per cell was measured using ImageJ, with at least 100 cells recorded per 
experimental condition. For quantification of the mean number of puncta per cell, images of 
≥20 cells per condition were analysed using the ImageJ ‘analyse particles’ command. Co-
localisation analysis was performed by a) individual puncta were first delineated by intensity 
thresholding in ImageJ and the number of co-localised puncta were then counted manually for 
5 cells (proband fibroblasts) or ≥ 20 cells per condition per experimental replicate 
(i3Neurons), or b), the extent of co-localisation was determined by calculating the Pearson’s 
correlation coefficient for red and green pixels in each cell using the Coloc2 ImageJ plugin for 
10 cells per condition. To assess the percentage of cells with aberrant centrosome number, 
≥25 cells were analysed in each of the 6 biological repeats for each experimental condition 
(200 cells/line). To assess the percentage of cells with aberrantly shaped nucleus, ≥40 cells 
were analysed in each of the 4 biological repeats for each experimental condition (200 
cells/line). To assess the percentage of cells with micronuclei and nucleoplasmic bridges, 200 
cells (micronuclei) or 250 (nucleoplasmic bridges) were analysed in each of the 5 
(micronuclei) or 4 (nucleoplasmic bridges) biological repeats for each experimental condition 
(1000 cells/line). For the analysis of chromosome aberrations, ≥30 well-spread metaphases 
were analysed in each of the 3 biological repeats for each experimental condition (100 
 16 
cells/line). Since fixation procedures may often result chromosome loss, the analysis was 
restricted to metaphases containing 45-46 chromosomes. To assess DNA damage, ≥40 cells 
were analysed in each of the 4 biological repeats for each experimental condition (200 
cells/line), counting only cells showing more than 20 foci positive to γ-H2AX staining. 
Finally, for determining the amount of cells with aberrant primary cilium, a total of 100 cells 
were analysed for each cell line over two experiments. 
Statistical analysis  
Statistical analysis and post hoc tests were carried out as described in figure legends using 
GraphPad Prism 8. The statistical significance is denoted on graphs by asterisks (*), where 
*P<0.05, **P<0.01, ***P<0.001 and n.s = not significant. 
  
 17 
Results 
Genetic Analysis 
Using data from GeneMatcher, the UK National Institute for Health Research Bioresource 
and Genomics England Research Consortium,33 or repositories linked to diagnostic testing 
(GeneDx Laboratory), 6 unrelated individuals with de novo variants in VPS4A 
(NM_013245.3) were identified. All had a common and distinctive phenotype including 
microcephaly, profound neonatal onset of hypotonia and global developmental delay, with 
similar structural brain abnormalities and cataracts in the majority. 
Five of the probands had de novo heterozygous missense variants at amino acid position 284. 
These included four (Probands 1, 3, 4 and 5) who had a c.850A>T (p.Arg284Trp) substitution 
and a single case (Proband 2) with a c.850A>G (p.Arg284Gly) change. Three of these cases 
were identified from a trio-based whole genome sequencing (WGS) approach (Probands 1, 3 
and 4),34; 35 whilst the others were identified using trio-based whole exome sequencing (WES) 
in the context of the Undiagnosed Patients Program at the Ospedale Pediatrico Bambino Gesù 
(Proband 2) or within routine care by GeneDx Laboratory (Proband 5). In addition, trio WGS 
in a further case identified a de novo variant at c.616G>A (p.Glu206Lys; Proband 6). In each 
case, there was a single plausible de novo variant in the absence of any pathogenic variant in 
genes previously associated with Mendelian diseases. Subsequent WGS in Proband 2 further 
excluded the occurrence of other clinically relevant variants. The DNA variants causing the 
p.Arg284 and p.Glu206 alterations were not present in control population databases 
(gnomAD, ExAC, TOPMed), fell in regions highly constrained for variation in control 
populations (Fig. 1A) and affected conserved amino acids of the AAA ATPase domain of 
VPS4A that are invariable across VPS4 family members from multiple species (Fig. S1). 
They had strong computational evidence for pathogenicity (Table S1).  
 18 
Two DNA variants c.502C>T (p.Pro168Ser) and c.1009A>G (p.Ile337Val), were also 
observed in one and three unrelated families respectively from the Genomics England 
Research Consortium. These were also absent from gnomAD. However, the mode of 
inheritance of the variants was unavailable and the pathogenicity of these substitutions was 
less plausible computationally, with lower CADD (25 and 23) and REVEL scores (0.911 and 
0.336) respectively. Additionally, the clinical phenotype of the probands was of non-specific 
ID, without the distinctive features that were prominent in the other subjects, so these variants 
were assessed as having uncertain significance.  
Mapping of putative pathogenic missense variants in VPS4A to homologous VPS4 
protein structure 
Protein structural mapping of the putative pathogenic variants to the yeast VPS4 homologous 
protein structure revealed their proximity to the ATP catalytic site and the pore lining loops. 
Arg284 was a hot-spot of de novo mutations in our subjects. It is one of two arginine residues 
known as the arginine fingers, a motif that is part of the interface between chains of the 
hexamerised VPS4, where the catalytic active site is assembled (Fig. 1B-D). Mutations of 
equivalent arginine finger residues have been studied in multiple ATPases, and cause 
complete loss of in vitro or in vivo catalytic activity of the protein, indicating that they are 
necessary for ATP hydrolysis.36 The arginine fingers may also be important for 
oligomerisation of ATPase protein complexes.36 Thus, existing functional data strongly 
support a deleterious effect of mutations affecting Arg284. 
In contrast, the Glu206 residue is functionally uncharacterised. It is located by the N-terminal 
of the α3-helix in the boundary with the Pore loop 1 (Fig. 1B-D). The lateral chain of Glu206 
(Glu211 in yeast) points to the α5-helix that is on top and connected to Pore loop 2 in all the 
chains of the hexamer (Fig. 1B-D, Fig. S2A). Pore loops 1 and 2 are very flexible and shape 
the central pore of the hexamer in ATPases. In the yeast structure both interact with a 
 19 
synthetic peptide that mimics ESCRT-III proteins, highlighting their role on the translocation 
of these proteins through the pore (Fig. 1D).26 The replacement of Glu206 by lysine 
introduces a change from a negative to a positively charged and slightly bigger lateral chain. 
Additionally, mapping of this position to the monomeric human and mouse VPS4B shows 
that the Glu213 (equivalent to Glu206) adopts different orientations, and the N-terminal of the 
α3-helix is slightly unfolded with the coordinates of Pore loop 1 unresolved (Fig. S2B), 
suggesting a possible conformational change in this region between the monomeric and 
hexameric forms. Therefore, we suggest that this variant is likely to affect the fold of the N-
terminal of α3-helix, the flexibility of the Pore 1 loop and the lateral interactions between α3 
and α5 helices, hence resulting in alterations in the recognition and translocation of ESCRT-
III proteins. 
Clinical profile of probands 
The six probands with de novo substitutions affecting Glu206 and Arg284 had a consistent 
phenotype characterized by severe DD, profound ID and dystonia (Figs. 1E, S3, Table S1, see 
Supplemental Note). Children were very delayed in establishing head control and none 
achieved independent walking. Other common findings were cerebellar hypoplasia (five 
individuals out of six, the other showing uncharacterized severe cerebral atrophy) with a 
variable degree of corpus callosum hypoplasia. One individual also had bilateral 
polymicrogyria. Epilepsy was present in three and dystonia in five. Eye involvement was also 
a common finding, including congenital cataract, retinal dystrophy, and in one case congenital 
Leber amaurosis. Four individuals were diagnosed with hepatosplenomegaly and/or steatosis. 
Three subjects had anaemia, which was characterised as dyserythropoietic in two. Severe 
feeding difficulties were present in four individuals, requiring assisted feeding in three. Two 
had sensorineural deafness. Severe growth retardation, generally for all parameters, was 
present in most cases. Notably, severe microcephaly (typically with Z scores < -5) was 
 20 
universal. Overall, the disorder seems to have a poor prognosis as two affected individuals 
died in childhood or early adult life (Table S1).  
In general, the affected individuals presented with a complex and severe phenotype with some 
features reminiscent of a ciliopathy-related disorder (cerebellar hypoplasia, retinal dystrophy 
or Leber amaurosis, sensorineural deafness), a significant neurodevelopmental condition 
(severe microcephaly and ID), and other features including cataracts, hepatosplenomegaly and 
congenital anaemia, giving rise to a distinct syndrome.  
VPS4A disease-associated variants have a dominant negative effect on endosomal 
morphology 
To investigate the molecular pathological mechanism of the disease-associated VPS4A 
alterations, we first examined the cellular expression of VPS4A, and found no alteration in 
protein abundance in fibroblasts from individuals with de novo p.Arg284Trp or p.Arg284Gly 
substitutions versus healthy parental controls, suggesting that mutant protein stability is 
unaltered (Fig. 2A,B). 
The published, rationally designed VPS4A-Glu228Gln mutant and the equivalent VPS4B 
mutant (p.Glu235Gln) are ATP hydrolysis-defective. Cultured cells over-expressing either of 
these mutants develop significantly enlarged endosomal vacuoles, caused by a dominant 
negative effect of incorporation of the ATPase-defective protein into VPS4 hexamers and 
subsequent failure of disassembly of the ESCRT complexes, which in turn makes ESCRT 
proteins unavailable for subsequent rounds of ILV formation.37-39  
 We used this phenotype to interrogate whether the identified disease-associated variants have 
a dominant negative effect on VPS4A function. We first confirmed that, as expected, 
expression of VPS4A-Glu228Gln in HeLa cells caused an enlarged endosomal vacuolar 
phenotype, in contrast to wild-type VPS4A, which exhibited a pan-cytosolic distribution 
 21 
consistent with its known expression pattern (Fig. 2C, D). Expression of VPS4A-Arg284Trp, 
VPS4A-Arg284Gly or VPS4A-Glu206Lys caused the development of vacuolar endosomal 
structures identical to those generated by VPS4A-Glu228Gln expression (Fig. 2E-G). In 
contrast, the expression pattern of VPS4A-Pro168Ser and VPS4A-Ile337Val resembled that 
of wild-type VPS4A (Fig. 2H, I). We concluded that the p.Arg284Trp, p.Arg284Gly and 
p.Glu206Lys amino acid substitutions exert a dominant negative effect on VPS4 hexamer 
function, while the p.Pro168Ser and p.Ile337Val sequence changes do not; considered with 
the clinical and bioinformatics data above, this suggests that the latter two variants are 
unlikely to be pathogenic. 
Physiological expression of disease-associated VPS4A causes abnormal endolysosomal 
morphology in proband fibroblasts 
We examined whether physiological expression of heterozygous p.Arg284Trp or 
p.Arg284Gly mutants in proband-derived fibroblasts caused altered endosomal morphology. 
We did not observe extremely large vacuolar structures of the type observed in cells over-
expressing exogenous ATPase-defective VPS4. However, the proband-derived cell lines had 
an increase in the percentage of cells with larger vesicles labelled by EEA1 (an early 
endosome marker), CD63 (which is typically enriched in the ILVs of the late endosome) and 
LAMP1 (enriched in lysosomes), in the absence of significant alterations in the total number 
of endosomal structures of each compartment (Fig. 2J-L). We did not observe any increased 
co-localisation between early endosomal, late endosomal or lysosomal markers in proband 
cells (Fig. S4). Thus heterozygous expression of mutant VPS4A at endogenous levels causes 
significant enlargement of multiple endosomal compartments, without apparent content 
mixing between them. 
Localisation and known functions of the core ESCRT-III complex at endosomes are not 
defective in proband fibroblasts 
 22 
As cells over-expressing dominant negative ATPase-defective VPS4 develop accumulations 
of ESCRT proteins on enlarged endosomal structures, we investigated whether the proband 
fibroblasts showed similar aberrant endosomal ESCRT-III localisation. We examined the 
ESCRT-III protein CHMP2B, a core ESCRT-III component that is recruited to the endosomal 
membrane during ILV formation. Surprisingly, there was no alteration in the number of 
cellular CHMP2B puncta with an area greater than 0.1 µm2, or in the percentage of EEA1- or 
CD63-positive endosomes that were associated with CHMP2B puncta (Fig. S5A,B).  
We examined late endosomal ultrastructure in the proband cell lines. Consistent with the light 
microscopy findings, large endosomes appeared to be more prominent. However, endosomes 
were still competent to make ILVs, a key function of ESCRT-III, as ILVs within MVBs were 
readily observed in the proband cells (Fig. S5C). 
Efficient degradation of the epidermal growth factor receptor (EGFR) involves sorting to the 
ILVs of the MVB, via a mechanism that requires core ESCRT-III components.40 EGFR 
degradation is inhibited by VPS4B depletion, either alone or combined with VPS4A 
depletion, or by expression of dominant negative ATPase-defective VPS4B-Glu235Gln.38; 41-
43 To our knowledge the specific role of VPS4A in this process has not been investigated. We 
examined EGFR degradation in proband fibroblasts carrying the VPS4A-Arg284Trp or 
VPS4A-Arg284Gly mutants and, consistent with the retained competence of these cells to 
make ILVs and regulate endosomal localisation of core ESCRT-III components, we found 
that it was not inhibited; indeed EGFR degradation was increased at 180 minutes post 
internalisation (Fig. S5D).  
We concluded that heterozygous expression of mutant VPS4A in proband cells does not affect 
the cellular distribution of a core ESCRT-III complex member, prevent formation of ILVs, or 
adversely affect the degradation of EGFR. 
 23 
The atypical ESCRT-III protein IST1 accumulates on endosomes in proband fibroblasts 
ESCRT-III proteins also play a role in fission of endosomal sorting tubules from the parent 
endosome. Rather than involving the core ESCRT-III complex, this activity is mediated by a 
complex formed of two atypical ESCRT-III proteins, IST1 and CHMP1B.5; 6; 44 Suitable 
antibodies are available to visualise IST1 by immunofluorescence. At steady state IST1 is 
localised in juxtaposition with early endosomal markers, and it is recruited to endosomal 
membranes by dominant negative VPS4A-Glu228Gln.45 We examined the appearance and 
localisation of IST1 and observed an increased number of IST1 puncta and increased 
percentage of EEA1-positive endosomes associated with IST1 puncta in proband-derived 
fibroblasts with the p.Arg284Trp or p.Arg284Gly substitutions, consistent with the idea that 
VPS4A ATPase activity is required to regulate the association of IST1 with endosomes (Fig. 
3A). No increased recruitment of IST1 to late endosomes or lysosomes was observed (Fig. 
S6). HeLa cells lacking IST1 show increased tubulation of the endosomal tubular marker 
SNX1 caused by a failure of endosomal tubule fission,6 but we did not consistently observe 
this phenotype in the proband fibroblasts (Fig. 3B).  
We concluded that heterozygous VPS4A mutations cause aberrant accumulations of the 
atypical ESCRT-III protein IST1 on endosomal membranes.  
Human neurons lacking VPS4A exhibit similar endosomal phenotypes to proband 
fibroblasts 
In view of the prominent neurodevelopmental phenotype observed in probands affected by 
VPS4A mutants, we attempted to model their effect in human neurons. We first generated i3 
iPSC lines expressing the VPS4A-Arg284Trp or -Arg284Gly mutants, using lentiviral 
transduction of appropriate constructs. In contrast to lines expressing wild-type VPS4A, these 
lines showed no detectable exogenous VPS4A expression by immunoblotting, and only very 
 24 
sparse expression (<1% of cells) was observed by immunofluorescence. Those cells that did 
express the mutant VPS4A exhibited enlarged, vacuolar endosomal structures typical of 
dominant negative VPS4 mutants (Fig. S7).37-39 In i3 iPSCs, the neurogenic transcription 
factor NGN2 is integrated under a doxycycline-responsive promoter at a safe harbour locus in 
the WTC11 iPSC line.46 This experimental system allows simple and rapid generation of 
glutamatergic cortical neurons (i3Ns) upon brief culture of the iPSCs in the presence of 
doxycycline; the cells have morphological and biochemical properties of neurons 14 days 
post-induction, and are electrically active after 21 days (Fig. S8).46 However, no VPS4A-
mutant expressing cells were identified upon differentiation of the i3 iPSCs to neurons and we 
concluded that over-expression of these altered proteins is incompatible with neuronal 
survival. 
In light of the apparent dominant negative effect of the over-expressed VPS4A mutants, we 
reasoned that lack of VPS4A may have similar cellular consequences to heterozygous 
physiological expression of VPS4A mutants that are capable of blocking the function of the 
wild-type protein, and so may provide useful insights into the potential functional effects of 
the mutants in neurons. In addition, analysis of neurons lacking VPS4A will elucidate the 
physiological role of VPS4A in these cells. We therefore employed a modified i3 iPSC 
system, in which CRISPR-inhibition (CRISPRi) machinery is integrated into a safe harbour 
locus.47 In CRISPRi, an enzymatically dead Cas9 fused to a KRAB transcriptional repressor is 
targeted close to the transcriptional start site of a target gene by a single guide RNA (sgRNA), 
thereby inhibiting expression of the gene. This system has advantages over standard CRISPR-
based knock-out systems, including high specificity with strikingly few off-target effects and 
low toxicity.48 We targeted VPS4A for repression in CRISPRi-i3 iPSCs, using two 
independent sgRNAs, confirmed cellular depletion of VPS4A in the iPSCs, then differentiated 
each line to i3Ns (Fig. 4A, B). We examined endosomal morphology in these neurons at 14 
 25 
days differentiation. While there was no significant increase in the percentage of neurons that 
had enlarged EEA1-positive puncta (Fig. 4C), we observed a significant increase in the 
percentage of cells with enlarged structures marked by CD63, LAMP1 or the lysosomal 
enzyme cathepsin D (Fig. 4D, E). In addition, there was an increase in the number of IST1 
puncta per cell and of IST1 localisation on early and late endosomal structures (Fig. 5A, B). 
No difference in the number of puncta of the core ESCRT-III component CHMP6 was 
observed (Fig. 5C). 
We concluded that VPS4A regulates endosomal size and endosomal membrane localisation of 
the atypical ESCRT-III protein in human neurons, and that loss of VPS4A in neurons largely 
recapitulates phenotypes that are observed in proband cells expressing dominant negative 
VPS4A at physiological heterozygous levels. 
VPS4A mutants affect centrosome and mitotic spindle organization, and are associated 
with aberrant chromosomal segregation and G2/M cell cycle arrest  
Centrosomes serve as solid-state signaling platforms to dynamically regulate a wide array of 
cellular structures and processes. The ESCRT-III complex and VPS4 proteins are required to 
maintain normal centrosome morphology and function, and their silencing alters centrosome 
and spindle pole numbers, frequently producing multipolar spindles and defects in 
chromosome segregation and nuclear morphology.15 To further validate the functional 
relevance of the identified VSP4A mutations we analysed centrosome and mitotic spindle 
organization in synchronized proband-derived fibroblasts carrying the heterozygous 
p.Arg284Trp or p.Arg284Gly substitutions. As expected, in interphase, control cells typically 
had two discernible centrosomes (Fig. 6A). Similarly, during mitosis, these cells formed 
normal bipolar spindles with two centrosomes. In metaphase, canonical mitotic spindles with 
properly aligned chromosomes were observed in the vast majority of cases (Fig. 6B). In 
contrast, proband fibroblasts showed an anomalous centrosome number and morphology in 
 26 
interphase (Fig. 6A); similarly, multipolar spindles were observed during mitosis, resulting in 
a high frequency of aberrant chromosome alignment during metaphase (Fig. 6B). Aberrant 
chromosome segregation was documented by the presence of both lagging and bridging 
chromosomes during anaphase and telophase (Fig. 6B), and was associated with polyploidy 
and production of micronuclei, i.e. encapsulated lagging chromosomes or damaged 
chromosome fragments not incorporated in the main nucleus after anaphase (Fig. 7A). 
Consistent with these findings, a high proportion of proband fibroblasts were observed at 
G2/M, as measured by BrdU incorporation flow cytometry analysis (Fig. 7B), which possibly 
results from altered G2/M and abscission checkpoint activation and/or a faulty progression 
towards cell division.49-51 
Proband fibroblasts have abnormal nuclear envelope morphology and increased DNA 
damage 
During anaphase, when chromosome separation has been achieved, the nuclear envelope is 
reassembled around the forming nuclei, to coordinate proper segregation of the nuclear 
content in daughter cells and assure the structural integrity and functionality of the nuclear 
compartment. The ESCRT-III complex and VPS4A and B contribute to nuclear envelope 
sealing and spindle disassembly at the nuclear envelope-microtubule intersection sites during 
mitotic exit,52; 53 and defective ESCRT function causes abnormal nuclear membrane sealing 
and altered morphology.54; 55 Moreover, the ESCRT-III complex contributes to repair of 
nuclear envelope ruptures during interphase, and expression of a dominant negative VPS4A 
protein delayed repair.56; 57 Based on these considerations, we explored possible changes in 
nuclear morphology and architecture in proband fibroblasts expressing the VPS4A-
Arg284Trp or -Arg284Gly mutants. Immunofluorescence microscopy with the nuclear 
envelope marker lamin A/C labelling demonstrated a significant increase in the proportion of 
proband-derived cells with irregular nuclear morphology compared to control cells (Fig. 8A). 
 27 
It has been established that nuclear deformation may result in a breaking of the nuclear 
envelope, which in turn exposes chromosomal DNA to the cytoplasmic environment, thus 
promoting DNA damage. Consistent with this, immunofluorescence microscopic 
identification of γH2AX foci, a marker of damaged DNA, showed an increased number of 
positive cells and foci per cell among those carrying the VSP4A mutants compared to control 
cells (Fig. 8B), indicating increased spontaneous DNA damage. Thus nuclear envelope 
morphology and integrity are altered in fibroblasts from individuals with heterozygous 
VPS4A mutations. 
Defective VPS4A function affects primary cilium morphology 
Vesicular trafficking plays an essential role in cilium biogenesis and function. Of note, VPS4 
has been identified as a dynamic component of the ciliary transition zone, a region in which 
the mother centriole, tethering to plasma membrane by the transition fibres, becomes the 
basal body for primary cilium formation. Defective VPS4A ATP hydrolysis causes a block of 
ciliogenesis after formation of the ciliary vesicle and this function appears to be ESCRT-III-
independent.16 We therefore analysed primary cilium biogenesis and morphology in 
fibroblasts from affected subjects. Assessment of cilium structure in starved fibroblasts 
revealed the presence of aberrant primary cilium formation in both fibroblast lines with 
mutated VPS4A alleles. Specifically, normal cilia were absent in fibroblasts expressing the 
VPS4A-Arg284Gly mutant, which instead showed a visible basal body (dot cilium); 
similarly, a dot cilium was documented in most fibroblasts heterozygous for the p.Arg284Trp 
substitution, although a small number of elongated cilia were also observed in cells from this 
proband (Fig. 8C). Thus defective VPS4A has pleiotropic consequences on diverse cellular 
processes, including perturbation of a variety of centrosome-dependent structures.  
 28 
Discussion 
The ESCRT-III complex and VPS4 together form a multifunctional membrane modelling 
machinery. Against the background of the multiplicity of functions ascribed to ESCRT-III, it 
is perhaps surprising that only three complex members, CHMP1A, CHMP2B and CHMP4B, 
have been implicated in Mendelian genetic disease thus far.17-19 This may be explained by 
functional redundancy between ESCRT-III complex components, or because loss of 
components might lead to an embryonic lethal phenotype, as has been observed in some, but 
not all, mouse models lacking specific CHMP proteins.58; 59 However, we now report a 
distinct syndromic neurodevelopmental disorder caused by dominantly acting amino acid 
substitutions in VPS4A, a key enzyme that regulates ESCRT-III function. We propose the 
acronym CIMDAG (Cerebellar hypoplasia and Cataracts, Intellectual disability, congenital 
Microcephaly, Dystonia and Dyserythropoeitic Anemia, Growth retardation) to highlight the 
main clinical features of this syndrome, which may also include other structural brain 
abnormalities, retinal dystrophy, hepatosplenomegaly and sensorineural deafness. The 
haematological features of this condition are thoroughly characterised in the accompanying 
paper.60 
The pleotropic clinical effects observed in CIMDAG likely reflect the multitude of cellular 
functions in which ESCRT-III and VPS4A participate, including those demonstrated to be 
affected by our studies. In the future it will be important to unravel which ESCRT-III and 
VPS4A functions underlie the pathology of the different clinical phenotypes we observed. 
The centrosomal and mitotic defects we observed are strong candidates to underlie 
microcephaly and other growth impairments in CIMDAG. These processes play a crucial role 
during brain development, including in neurogenesis, neuronal migration and polarity, and 
defects in them commonly underlie neurodevelopmental diseases.61 Defects in mitosis have 
also been linked to dyserythropoietic anaemia and are a feature of the anaemia seen in 
 29 
CIMDAG.60; 62 In contrast, congenital cataract is a recognised feature of lysosomal storage 
diseases and so it may be related to the endolysosomal dysfunction that we observed, which 
may also be relevant for the health of mature post-mitotic neurons, as it has been implicated in 
several forms of neurodegeneration.63 Finally, it is striking that many of the clinical features 
found in our probands (including cerebellar hypoplasia, retinal dystrophy, Leber amaurosis, 
DD, ID, cataract, sensorineural deafness and hypogonadism) occur in ciliopathies, and our 
data support the proposed role of VPS4A in controlling primary cilium morphogenesis.16; 64 
Thus it is possible that abnormal primary cilium function contributes to the CIMDAG 
phenotype. 
Multiple heterozygous VPS4A loss-of-function mutations are present in general population 
databases, indicating that a haploinsufficiency mechanism is unlikely to cause the type of 
severe early childhood condition that we describe.65 In contrast, our data and published 
evidence point to the p.Glu206Lys, p.Arg284Trp and p.Arg284Gly mutants having a 
dominant negative effect. Over-expression of each of these mutants in immortalised cells 
caused development of the highly characteristic enlarged endosomal structures that are 
induced by expression of known dominant-negative forms of VPS4A. Similar, although less 
marked, enlarged endosomal phenotypes were also observed in proband cells. As VPS4A 
protein expression was not altered in proband cells, assuming equal expression of wild-type 
and mutant VPS4A we expect that a large majority of VPS4A hexamers will have impaired 
function as they will contain at least one mutant subunit. This may explain the similar 
endosomal phenotypes we observed in proband fibroblasts and iPSC-derived neurons lacking 
VPS4A.  
The archetypal function for VPS4 proteins is in endosomal sorting and our observations 
elucidate details of the physiological role of VPS4A in this process. Over-expression of 
dominant negative VPS4A in cultured cells affects ILV formation, causes trapping of ESCRT 
 30 
proteins on the endosomal membrane and inhibits EGFR degradation.38; 39 However, a 
surprising observation in our study was that proband fibroblasts expressing dominant negative 
VPS4A mutations at heterozygous physiological levels showed no obvious effect on ILV 
formation or EGFR degradation. Consistent with this, we did not observe accumulation of the 
“core” ESCRT-III complex member CHMP2B on the endosomal membrane, suggesting that 
functional VPS4A is not required for removal of the core ESCRT-III complex from the 
endosomal membrane. This may be because of redundancy between VPS4A and VPS4B – 
previous studies have shown that double knock-down of these proteins is required for 
formation of a “VPS4-dominant-negative”- type endosomal compartment. An ESCRT-
independent ILV formation pathway has also been described, which could also provide an 
explanation for retained ILV formation.66; 67 In contrast, the endosomal localisation of IST1 
was increased in both proband fibroblasts and iPSC-derived neurons lacking VPS4A, so it 
appears that VPS4A is absolutely required for recycling of this protein, the best-characterised 
function of which is in promoting endosomal tubule fission.5; 6 Further studies will be required 
to elucidate the functional consequences of this accumulation on endosomal tubule fission 
dynamics and endosomal receptor traffic, and its relationship to pathogenesis in our probands, 
but we speculate that the enhanced EGFR degradation we observed in patient fibroblasts may 
be explained by defective tubular sorting of this receptor away from endosomal degradation. 
In summary, we have identified de novo missense mutations affecting the key ESCRT-
regulation enzyme VPS4A in probands with a distinct multisystem neurodevelopmental 
condition. Study of the functional effects of the mutations demonstrated that they act by a 
dominant negative mechanism to cause effects on multiple ESCRT-dependent cellular 
pathways, and indicate an absolute requirement for proper VPS4A function in 
neurodevelopment and other physiological developmental processes in humans.  
  
 31 
Supplemental Data includes supplemental clinical reports, author list for genomics England 
research consortium, supplemental methods, eight figures and one table. 
Declaration of Interests 
The authors declare no competing interests. 
Acknowledgements 
We thank the families who graciously agreed to participate in this study. We thank Michael 
Ward (NIH, Bethesda) for the gift of i3 iPSC lines. This work was supported by: UK Medical 
Research Council Project Grants [MR/M00046X/1], [MR/R026440/1] and Project grant from 
National Institute of Health Research Biomedical Research Centre at Addenbrooke's Hospital 
(to E.R.), Fondazione Bambino Gesù (Vite Coraggiose) and Italian Ministry of Health (CCR-
2017-23669081) (to M.T.), National Institute for Health Research (NIHR) for the Cambridge 
Biomedical Research Centre and NIHR BioResource (Grant Number RG65966) (to F.L.R.), 
and a Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society 
(Grant Number 216370/Z/19/Z) (to J.E.). CIMR was supported by a Wellcome Trust Strategic 
Award [100140] and Equipment Grant [093026].  
This research was made possible through access to the data and findings generated by the 
100,000 Genomes Project. The 100,000 Genomes Project is managed by Genomics England 
Limited (a wholly owned company of the Department of Health and Social Care). The 
100,000 Genomes Project is funded by the National Institute for Health Research and NHS 
England. The Wellcome Trust, Cancer Research UK and the Medical Research Council have 
also funded research infrastructure. The 100,000 Genomes Project uses data provided by 
patients and collected by the National Health Service as part of their care and support. 
 
 
 32 
Web Resources 
Online Mendelian Inheritance in Man (OMIM): http://www.omim.org 
GnomAD, http://gnomad.broadinstitute.org/ 
ExAC database, http://exac.broadinstitute.org/ 
NHLBI Trans-Omics for Precision Medicine (TOPMed), https://www.nhlbiwgs.org/ 
dbNSFP, https://sites.google.com/site/jpopgen/dbNSFP 
ClinVar, https://www.ncbi.nlm.nih.gov/clinvar/ 
Pfam, https://pfam.xfam.org/ 
UniProtKB, https://www.uniprot.org/ 
Protein Data Bank (PDB), https://www.rcsb.org/ 
Addgene, http://www.addgene.org/ 
 
Data and Code Availability 
The pathogenic variants identified in this work have been submitted to ClinVar (submission 
ID: SUB7923752). WES and WGS datasets have not been deposited in a public repository 
due to privacy and ethical restrictions but are available from the corresponding authors on 
request. The 100,000 Genomes Project data used in this work are available from the 
Genomics England Research Environment, with restrictions to ensure the confidentiality and 
appropriate use of the data. 
  
 33 
References 
1. McCullough, J., Frost, A., and Sundquist, W.I. (2018). Structures, Functions, and 
Dynamics of ESCRT-III/Vps4 Membrane Remodeling and Fission Complexes. Annual 
review of cell and developmental biology 34, 85-109. 
2. Schöneberg, J., Lee, I.-H., Iwasa, J.H., and Hurley, J.H. (2017). Reverse-topology 
membrane scission by the ESCRT proteins. Nature reviews Molecular cell biology 18, 5-17. 
3. Vietri, M., Radulovic, M., and Stenmark, H. (2020). The many functions of ESCRTs. 
Nature reviews Molecular cell biology 21, 25-42. 
4. Caillat, C., Maity, S., Miguet, N., Roos, W.H., and Weissenhorn, W. (2019). The role of 
VPS4 in ESCRT-III polymer remodeling. Biochemical Society transactions 47, 441-448. 
5. McCullough, J., Clippinger, A.K., Talledge, N., Skowyra, M.L., Saunders, M.G., 
Naismith, T.V., Colf, L.A., Afonine, P., Arthur, C., Sundquist, W.I., et al. (2015). Structure 
and membrane remodeling activity of ESCRT-III helical polymers. Science 350, 1548-1551. 
6. Allison, R., Lumb, J.H., Fassier, C., Connell, J.W., Ten Martin, D., Seaman, M.N., Hazan, 
J., and Reid, E. (2013). An ESCRT-spastin interaction promotes fission of recycling tubules 
from the endosome. J Cell Biol 202, 527-543. 
7. Katzmann, D.J., Babst, M., and Emr, S.D. (2001). Ubiquitin-dependent sorting into the 
multivesicular body pathway requires the function of a conserved endosomal protein sorting 
complex, ESCRT-I. Cell 106, 145-155. 
8. Babst, M., Katzmann, D.J., Snyder, W.B., Wendland, B., and Emr, S.D. (2002). 
Endosome-associated complex, ESCRT-II, recruits transport machinery for protein sorting at 
the multivesicular body. Dev Cell 3, 283-289. 
9. Babst, M., Katzmann, D.J., Estepa-Sabal, E.J., Meerloo, T., and Emr, S.D. (2002). Escrt-
III: an endosome-associated heterooligomeric protein complex required for mvb sorting. Dev 
Cell 3, 271-282. 
 34 
10. Maxfield, F.R., and McGraw, T.E. (2004). Endocytic recycling. Nat Rev Mol Cell Biol 5, 
121-132. 
11. Yang, D., Rismanchi, N., Renvoise, B., Lippincott-Schwartz, J., Blackstone, C., and 
Hurley, J.H. (2008). Structural basis for midbody targeting of spastin by the ESCRT-III 
protein CHMP1B. Nat Struct Mol Biol. 
12. Stuchell-Brereton, M.D., Skalicky, J.J., Kieffer, C., Karren, M.A., Ghaffarian, S., and 
Sundquist, W.I. (2007). ESCRT-III recognition by VPS4 ATPases. Nature 449, 740-744. 
13. Obita, T., Saksena, S., Ghazi-Tabatabai, S., Gill, D.J., Perisic, O., Emr, S.D., and 
Williams, R.L. (2007). Structural basis for selective recognition of ESCRT-III by the AAA 
ATPase Vps4. Nature 449, 735-739. 
14. Bajorek, M., Schubert, H.L., McCullough, J., Langelier, C., Eckert, D.M., Stubblefield, 
W.M., Uter, N.T., Myszka, D.G., Hill, C.P., and Sundquist, W.I. (2009). Structural basis for 
ESCRT-III protein autoinhibition. Nature structural & molecular biology 16, 754-762. 
15. Morita, E., Colf, L.A., Karren, M.A., Sandrin, V., Rodesch, C.K., and Sundquist, W.I. 
(2010). Human ESCRT-III and VPS4 proteins are required for centrosome and spindle 
maintenance. Proceedings of the National Academy of Sciences 107, 12889. 
16. Ott, C., Nachmias, D., Adar, S., Jarnik, M., Sherman, S., Birnbaum, R.Y., Lippincott-
Schwartz, J., and Elia, N. (2018). VPS4 is a dynamic component of the centrosome that 
regulates centrosome localization of γ-tubulin, centriolar satellite stability and ciliogenesis. 
Scientific Reports 8, 3353. 
17. Skibinski, G., Parkinson, N.J., Brown, J.M., Chakrabarti, L., Lloyd, S.L., Hummerich, H., 
Nielsen, J.E., Hodges, J.R., Spillantini, M.G., Thusgaard, T., et al. (2005). Mutations in the 
endosomal ESCRTIII-complex subunit CHMP2B in frontotemporal dementia. Nature 
genetics 37, 806-808. 
 35 
18. Shiels, A., Bennett, T.M., Knopf, H.L.S., Yamada, K., Yoshiura, K.-i., Niikawa, N., 
Shim, S., and Hanson, P.I. (2007). CHMP4B, a novel gene for autosomal dominant cataracts 
linked to chromosome 20q. American journal of human genetics 81, 596-606. 
19. Mochida, G.H., Ganesh, V.S., de Michelena, M.I., Dias, H., Atabay, K.D., Kathrein, 
K.L., Huang, H.T., Hill, R.S., Felie, J.M., Rakiec, D., et al. (2012). CHMP1A encodes an 
essential regulator of BMI1-INK4A in cerebellar development. Nat Genet 44, 1260-1264. 
20. Flex, E., Martinelli, S., Van Dijck, A., Ciolfi, A., Cecchetti, S., Coluzzi, E., Pannone, L., 
Andreoli, C., Radio, F.C., Pizzi, S., et al. (2019). Aberrant Function of the C-Terminal Tail of 
HIST1H1E Accelerates Cellular Senescence and Causes Premature Aging. Am J Hum Genet 
105, 493-508. 
21. Bauer, C.K., Calligari, P., Radio, F.C., Caputo, V., Dentici, M.L., Falah, N., High, F., 
Pantaleoni, F., Barresi, S., Ciolfi, A., et al. (2018). Mutations in KCNK4 that Affect Gating 
Cause a Recognizable Neurodevelopmental Syndrome. Am J Hum Genet 103, 621-630. 
22. Turro, E., Astle, W.J., Megy, K., Gräf, S., Greene, D., Shamardina, O., Allen, H.L., 
Sanchis-Juan, A., Frontini, M., Thys, C., et al. (2020). Whole-genome sequencing of patients 
with rare diseases in a national health system. Nature 583, 96-102. 
23. Sanchis-Juan, A., Hasenahuer, M.A., Baker, J.A., McTague, A., Barwick, K., Kurian, 
M.A., Duarte, S.T., BioResource, N., Carss, K.J., Thornton, J., et al. (2020). Structural 
analysis of pathogenic missense mutations in GABRA2 and identification of a novel de novo 
variant in the desensitization gate. Molecular Genetics & Genomic Medicine n/a, e1106. 
24. Havrilla, J.M., Pedersen, B.S., Layer, R.M., and Quinlan, A.R. (2019). A map of 
constrained coding regions in the human genome. Nature Genetics 51, 88-95. 
25. ww, P.D.B.c. (2018). Protein Data Bank: the single global archive for 3D 
macromolecular structure data. Nucleic Acids Research 47, D520-D528. 
 36 
26. Han, H., Fulcher, J.M., Dandey, V.P., Iwasa, J.H., Sundquist, W.I., Kay, M.S., Shen, P.S., 
and Hill, C.P. (2019). Structure of Vps4 with circular peptides and implications for 
translocation of two polypeptide chains by AAA+ ATPases. Elife 8. 
27. Scott, A., Chung, H.-Y., Gonciarz-Swiatek, M., Hill, G.C., Whitby, F.G., Gaspar, J., 
Holton, J.M., Viswanathan, R., Ghaffarian, S., Hill, C.P., et al. (2005). Structural and 
mechanistic studies of VPS4 proteins. The EMBO Journal 24, 3658-3669. 
28. Inoue, M., Kamikubo, H., Kataoka, M., Kato, R., Yoshimori, T., Wakatsuki, S., and 
Kawasaki, M. (2008). Nucleotide-dependent conformational changes and assembly of the 
AAA ATPase SKD1/VPS4B. Traffic 9, 2180-2189. 
29. Horlbeck, M.A., Gilbert, L.A., Villalta, J.E., Adamson, B., Pak, R.A., Chen, Y., Fields, 
A.P., Park, C.Y., Corn, J.E., Kampmann, M., et al. (2016). Compact and highly active next-
generation libraries for CRISPR-mediated gene repression and activation. Elife 5. 
30. Fernandopulle, M.S., Prestil, R., Grunseich, C., Wang, C., Gan, L., and Ward, M.E. 
(2018). Transcription Factor-Mediated Differentiation of Human iPSCs into Neurons. 
Current protocols in cell biology 79, e51. 
31. Fenech, M. (2007). Cytokinesis-block micronucleus cytome assay. Nat Protoc 2, 1084-
1104. 
32. Connell, J.W., Lindon, C., Luzio, J.P., and Reid, E. (2009). Spastin couples microtubule 
severing to membrane traffic in completion of cytokinesis and secretion. Traffic 10, 42-56. 
33. Sobreira, N., Schiettecatte, F., Valle, D., and Hamosh, A. (2015). GeneMatcher: a 
matching tool for connecting investigators with an interest in the same gene. Hum Mutat 36, 
928-930. 
34. French, C.E., Delon, I., Dolling, H., Sanchis-Juan, A., Shamardina, O., Mégy, K., Abbs, 
S., Austin, T., Bowdin, S., Branco, R.G., et al. (2019). Whole genome sequencing reveals 
 37 
that genetic conditions are frequent in intensively ill children. Intensive care medicine 45, 
627-636. 
35. Turro E, A.W., Megy K, Gräf S, Greene D, Shamardina O, et al. (2020). Whole-genome 
sequencing of patients with rare diseases in a national healthcare system Nature 583, 96-102. 
36. Wendler, P., Ciniawsky, S., Kock, M., and Kube, S. (2012). Structure and function of the 
AAA+ nucleotide binding pocket. Biochimica et Biophysica Acta (BBA) - Molecular Cell 
Research 1823, 2-14. 
37. Scheuring, S., Röhricht, R.A., Schöning-Burkhardt, B., Beyer, A., Müller, S., Abts, H.F., 
and Köhrer, K. (2001). Mammalian Cells Express Two VPS4 Proteins Both of Which are 
Involved in Intracellular Protein Trafficking. Journal of Molecular Biology 312, 469-480. 
38. Fujita, H., Yamanaka, M., Imamura, K., Tanaka, Y., Nara, A., Yoshimori, T., Yokota, S., 
and Himeno, M. (2003). A dominant negative form of the AAA ATPase SKD1/VPS4 impairs 
membrane trafficking out of endosomal/lysosomal compartments: class E vps phenotype in 
mammalian cells. Journal of cell science 116, 401-414. 
39. Bishop, N., and Woodman, P. (2000). ATPase-defective Mammalian VPS4 Localizes to 
Aberrant Endosomes and Impairs Cholesterol Trafficking. Molecular Biology of the Cell 11, 
227-239. 
40. Bache, K.G., Stuffers, S., Malerod, L., Slagsvold, T., Raiborg, C., Lechardeur, D., 
Walchli, S., Lukacs, G.L., Brech, A., and Stenmark, H. (2006). The ESCRT-III subunit 
hVps24 is required for degradation but not silencing of the epidermal growth factor receptor. 
Mol Biol Cell 17, 2513-2523. 
41. Lin, H.H., Li, X., Chen, J.-L., Sun, X., Cooper, F.N., Chen, Y.-R., Zhang, W., Chung, Y., 
Li, A., Cheng, C.-T., et al. (2012). Identification of an AAA ATPase VPS4B-dependent 
pathway that modulates epidermal growth factor receptor abundance and signaling during 
hypoxia. Molecular and cellular biology 32, 1124-1138. 
 38 
42. Yoshimori, T., Yamagata, F., Yamamoto, A., Mizushima, N., Kabeya, Y., Nara, A., 
Miwako, I., Ohashi, M., Ohsumi, M., and Ohsumi, Y. (2000). The mouse SKD1, a 
homologue of yeast Vps4p, is required for normal endosomal trafficking and morphology in 
mammalian cells. Molecular biology of the cell 11, 747-763. 
43. Fraile-Ramos, A., Pelchen-Matthews, A., Risco, C., Rejas, M.T., Emery, V.C., Hassan-
Walker, A.F., Esteban, M., and Marsh, M. (2007). The ESCRT machinery is not required for 
human cytomegalovirus envelopment. Cell Microbiol 9, 2955-2967. 
44. Allison, R., Edgar, J.R., Pearson, G., Rizo, T., Newton, T., Günther, S., Berner, F., 
Hague, J., Connell, J.W., Winkler, J., et al. (2017). Defects in ER–endosome contacts impact 
lysosome function in hereditary spastic paraplegia. J Cell Biol 216, 1337-1355. 
45. Agromayor, M., Carlton, J.G., Phelan, J.P., Matthews, D.R., Carlin, L.M., Ameer-Beg, S., 
Bowers, K., and Martin-Serrano, J. (2009). Essential role of hIST1 in cytokinesis. Mol Biol 
Cell 20, 1374-1387. 
46. Wang, C., Ward, M.E., Chen, R., Liu, K., Tracy, T.E., Chen, X., Xie, M., Sohn, P.D., 
Ludwig, C., Meyer-Franke, A., et al. (2017). Scalable Production of iPSC-Derived Human 
Neurons to Identify Tau-Lowering Compounds by High-Content Screening. Stem Cell 
Reports 9, 1221-1233. 
47. Tian, R., Gachechiladze, M.A., Ludwig, C.H., Laurie, M.T., Hong, J.Y., Nathaniel, D., 
Prabhu, A.V., Fernandopulle, M.S., Patel, R., Abshari, M., et al. (2019). CRISPR 
Interference-Based Platform for Multimodal Genetic Screens in Human iPSC-Derived 
Neurons. Neuron. 
48. Kampmann, M. (2018). CRISPRi and CRISPRa Screens in Mammalian Cells for 
Precision Biology and Medicine. ACS Chem Biol 13, 406-416. 
49. Kastan, M.B., and Bartek, J. (2004). Cell-cycle checkpoints and cancer. Nature 432, 316-
323. 
 39 
50. Bartek, J., and Lukas, J. (2007). DNA damage checkpoints: from initiation to recovery or 
adaptation. Curr Opin Cell Biol 19, 238-245. 
51. Caballe, A., Wenzel, D.M., Agromayor, M., Alam, S.L., Skalicky, J.J., Kloc, M., Carlton, 
J.G., Labrador, L., Sundquist, W.I., and Martin-Serrano, J. (2015). ULK3 regulates 
cytokinetic abscission by phosphorylating ESCRT-III proteins. Elife 4, e06547. 
52. Vietri, M., Schink, K.O., Campsteijn, C., Wegner, C.S., Schultz, S.W., Christ, L., 
Thoresen, S.B., Brech, A., Raiborg, C., and Stenmark, H. (2015). Spastin and ESCRT-III 
coordinate mitotic spindle disassembly and nuclear envelope sealing. Nature 522, 231-235. 
53. Olmos, Y., Hodgson, L., Mantell, J., Verkade, P., and Carlton, J.G. (2015). ESCRT-III 
controls nuclear envelope reformation. Nature 522, 236-239. 
54. Ventimiglia, L.N., Cuesta-Geijo, M.A., Martinelli, N., Caballe, A., Macheboeuf, P., 
Miguet, N., Parnham, I.M., Olmos, Y., Carlton, J.G., Weissenhorn, W., et al. (2018). 
CC2D1B Coordinates ESCRT-III Activity during the Mitotic Reformation of the Nuclear 
Envelope. Dev Cell 47, 547-563.e546. 
55. Asencio, C., Davidson, I.F., Santarella-Mellwig, R., Ly-Hartig, T.B., Mall, M., 
Wallenfang, M.R., Mattaj, I.W., and Gorjánácz, M. (2012). Coordination of kinase and 
phosphatase activities by Lem4 enables nuclear envelope reassembly during mitosis. Cell 
150, 122-135. 
56. Raab, M., Gentili, M., de Belly, H., Thiam, H.R., Vargas, P., Jimenez, A.J., 
Lautenschlaeger, F., Voituriez, R., Lennon-Duménil, A.M., Manel, N., et al. (2016). ESCRT 
III repairs nuclear envelope ruptures during cell migration to limit DNA damage and cell 
death. Science 352, 359-362. 
57. Arii, J., Watanabe, M., Maeda, F., Tokai-Nishizumi, N., Chihara, T., Miura, M., 
Maruzuru, Y., Koyanagi, N., Kato, A., and Kawaguchi, Y. (2018). ESCRT-III mediates 
 40 
budding across the inner nuclear membrane and regulates its integrity. Nature 
communications 9, 3379. 
58. Shim, J.H., Xiao, C., Hayden, M.S., Lee, K.Y., Trombetta, E.S., Pypaert, M., Nara, A., 
Yoshimori, T., Wilm, B., Erdjument-Bromage, H., et al. (2006). CHMP5 is essential for late 
endosome function and down-regulation of receptor signaling during mouse embryogenesis. J 
Cell Biol 172, 1045-1056. 
59. Ghazi-Noori, S., Froud, K.E., Mizielinska, S., Powell, C., Smidak, M., Fernandez de 
Marco, M., O'Malley, C., Farmer, M., Parkinson, N., Fisher, E.M., et al. (2012). Progressive 
neuronal inclusion formation and axonal degeneration in CHMP2B mutant transgenic mice. 
Brain 135, 819-832. 
60. Seu, K.G., Trump, L.R., Emberesh, S., Lorsbach, R.B., Johnson, C., Meznarich, J., 
Underhill, H.R., Chou, S.T., Sakthivel, H., Nassar, N., et al. VPS4A mutations in humans 
cause syndromic congenital dyserythropoietic anemia due to cytokinesis and trafficking 
defects. American Journal of Human Genetics in press. 
61. Woods, C.G. (2004). Human microcephaly. Curr Opin Neurobiol 14, 112-117. 
62. Paw, B.H., Davidson, A.J., Zhou, Y., Li, R., Pratt, S.J., Lee, C., Trede, N.S., Brownlie, 
A., Donovan, A., Liao, E.C., et al. (2003). Cell-specific mitotic defect and dyserythropoiesis 
associated with erythroid band 3 deficiency. Nat Genet 34, 59-64. 
63. Hu, Y.B., Dammer, E.B., Ren, R.J., and Wang, G. (2015). The endosomal-lysosomal 
system: from acidification and cargo sorting to neurodegeneration. Translational 
neurodegeneration 4, 18. 
64. Reiter, J.F., and Leroux, M.R. (2017). Genes and molecular pathways underpinning 
ciliopathies. Nat Rev Mol Cell Biol 18, 533-547. 
 41 
65. Karczewski, K.J., Francioli, L.C., Tiao, G., Cummings, B.B., Alföldi, J., Wang, Q., 
Collins, R.L., Laricchia, K.M., Ganna, A., Birnbaum, D.P., et al. (2020). The mutational 
constraint spectrum quantified from variation in 141,456 humans. Nature 581, 434-443. 
66. Theos, A.C., Truschel, S.T., Tenza, D., Hurbain, I., Harper, D.C., Berson, J.F., Thomas, 
P.C., Raposo, G., and Marks, M.S. (2006). A lumenal domain-dependent pathway for sorting 
to intralumenal vesicles of multivesicular endosomes involved in organelle morphogenesis. 
Dev Cell 10, 343-354. 
67. Edgar, J.R., Eden, E.R., and Futter, C.E. (2014). Hrs- and CD63-dependent competing 
mechanisms make different sized endosomal intraluminal vesicles. Traffic 15, 197-211. 
 
  
 42 
Figure Legends 
Figure 1. De novo missense variants in VPS4A mapped to the schematic protein diagram 
and the homologous yeast structure. A) Allele count of missense variants in gnomAD and 
the constrained coding regions’ (CCRs) percentiles are represented for human VPS4A 
(NP_037377) and are aligned with protein domains. The de novo missense variants cluster in 
constrained regions of the large ATPase domain. Coordinates of the protein domains were 
from Pfam (UniProtKB entry Q9UN37). MIT = microtubule interacting and trafficking; AAA 
= ATPase family associated with various cellular activities; AAA_lid = AAA+ lid domain; 
Vps4_C = Vps4 C-terminal oligomerization domain; LC = low complexity region. Disordered 
regions are shaded in translucent grey. B-D) De novo missense variants in VPS4A are mapped 
to the cryo-EM structure of the ATPase domain of the homologous yeast VPS4 in 
homohexameric form (PDB ID: 6OO2). The approximate locations of the active sites are 
shaded in orange, with the ADP nucleotides represented in sticks, when present. The 
conserved motifs that define the ATP binding site and pore loops 1 and 2 are shown in dark 
blue or grey. B) Structure of the homohexamer, with the six chains alternately coloured in 
blue and white. C) Structure of a single chain. Both p.R284W and p.R284G are observed to 
affect the R-finger of the active site and p.E206K affects the intra-chain interface between α3 
and α5-helices, located after the pore loops 1 and 2. Only the name of these two helices is 
shown for clarity. p.E228Q, a rationally-designed mutant that produces dominant-negative 
ATPase-defective VPS4A, falls in the Walker B motif of the active site. D) Zoom in of the 
inter-chain interface. The pore loops 1 and 2 shape the pore and interact with the synthetic 
peptide (in yellow) that shows how the ESCRT-III protein would translocate through the pore. 
E) Images of probands at representative ages.  
Figure 2. VPS4A disease-associated variants have a dominant negative effect on 
endosomal morphology. A) Representative VPS4A immunoblots of fibroblast cell lysates 
 43 
from a proband (Proband 1) with the p.R284W sequence change and her parents (control-1 
and control-2), and from a proband (Proband 2) with the p.R284G sequence change. B) 
Immunoblot band intensities from 3 such experiments were quantified, normalised to GAPDH 
loading control values, and plotted in the corresponding graph. C-I) HeLa cells were 
transfected with constructs expressing wild-type myc-VPS4A, myc-VPS4A containing the 
rationally-designed ATPase-defective p.E228Q mutant, or myc-VPS4A harbouring the 
sequence changes identified in probands. Cells were fixed, labelled with anti-Myc and anti-
RAB5 antibodies and visualised with confocal microscopy. The inset panels show higher 
magnification views of the boxed regions- examples of large vacuolar endosomal structures 
are shown. J-L) Cultured fibroblasts from the control subjects and probands indicated were 
fixed, labelled with EEA1 (early endosomes) (J), CD63 (preferentially labels late endosomes) 
(K) and LAMP1 (predominantly labels lysosomes) (L), then visualised by widefield 
immunofluorescence microscopy. The percentage of cells with an endosomal organelle over a 
nominal cut-off size and the number of labelled organelles per cell was counted in n=3 
biological repeats for each marker (in 100 cells per experimental condition in each repeat), 
then quantified in the corresponding charts. Bars in all plots show mean ± S.E.M., p-values 
calculated by one-way ANOVA with Tukey’s post-hoc test for repeated measures. Scale bars 
= 10 µm. 
Figure 3. The atypical ESCRT-III protein IST1 accumulates on endosomes in proband 
fibroblasts. A) Cultured fibroblasts from control subjects or the probands indicated were 
fixed, labelled against EEA1 and IST1, then visualised with confocal immunofluorescence 
microscopy. The number of IST1 puncta per cell and the percentage of EEA1-positive 
endosomes associated with an IST1 punctum was quantified in 3 such experiments (in 5 cells 
per experimental condition in each biological repeat) and plotted in the corresponding charts. 
Arrows indicate juxtaposed or co-localised puncta. B) Cultured fibroblasts from control 
 44 
subjects or the probands indicated were fixed and labelled for the endosomal tubular marker 
SNX1, then visualised with widefield immunofluorescence microscopy. The percentage of 
cells with at least one SNX-1-positive tubular structure >1.2µm in length was quantified in 
100 cells per sample, and the results for 3 such experiments were plotted in the corresponding 
chart. Bars indicate mean ± S.E.M., p-values calculated by one-way ANOVA with Tukey’s 
post-hoc test for repeated measures. Micrograph scale bars = 10 µm. 
Figure 4. Human neurons lacking VPS4A exhibit similar endosomal phenotypes to 
proband fibroblasts. A) CRISPRi-i3N iPSCs with were transduced with a scrambled sgRNA 
and three separate sgRNAs (G1-G3) directed against VPS4A. Cell lysates were 
immunoblotted against VPS4A. B) Selected lines were treated with doxycycline to induce 
neuronal differentiation, then blotted against VPS4A 14 days later. GAPDH signal validates 
equal lane loading. C) CRISPRi-i3N iPSCs expressing the guides indicated were 
differentiated to neurons for 14 days, fixed and labelled for EEA1. The percentage of cells 
with an EEA1-positive organelle over a nominal cut-off size was visualised by widefield 
microscopy and quantified in 3 experiments (≥ 100 cells per experimental condition in each 
repeat). The number of EEA1-positive endosomes per cell was visualised using confocal 
microscopy and quantified in 3 experiments (≥ 20 cells per experimental condition in each 
repeat). Quantifications are plotted in the corresponding graphs. D) i3Neurons expressing the 
sgRNAs indicated were fixed and labelled for CD63, then visualised by widefield 
microscopy. The percentage of cells with a CD63 organelle over a nominal cut-off size was 
quantified as described for EEA1. E) i3Neurons expressing the sgRNAs indicated were fixed 
and labelled for LAMP1 and cathepsin D, then visualised by widefield microscopy. The 
percentage of cells that had a LAMP1 or cathepsin D organelle over a nominal cut-off size 
was quantified as described for EEA1. Bars show mean ± S.E.M., p-values generated by one-
way ANOVA with Tukey’s correction for multiple testing. Scale bars = 10 µm. 
 45 
Figure 5. The atypical ESCRT-III protein IST1 accumulates on endosomes in human 
neurons lacking VPS4A. i3Neurons expressing the guides indicated were fixed and labelled 
for IST1 and EEA1 (A) or IST1 and CD63 (B), then visualised by confocal 
immunofluorescence microscopy. The number of IST1 puncta per cell and the percentage of 
EEA1- or CD63- positive endosomes associated with an IST1 punctum was quantified in 3 
experiments per marker (in ≥20 cells per experimental condition in each repeat) and plotted in 
the corresponding charts. Arrows indicate juxtaposed or co-localised puncta. C) i3Neurons 
expressing the guides indicated were fixed and labelled for CHMP6, then visualised by 
confocal immunofluorescence microscopy. The number of CHMP6 puncta per cell was 
quantified in 3 biological repeats (≥20 cells per experimental condition in each repeat). Bars 
indicate mean ± S.E.M., p-values calculated by one-way ANOVA with Tukey’s correction for 
multiple testing. Micrograph scale bar = 10 µm. 
Figure 6. Defective VPS4A function affects centrosome numbers and mitotic spindle 
organization. Confocal microscopy analysis was performed in synchronized skin fibroblasts 
from subjects with VPS4A mutations and control cells. Images are representative of three 
stages of the cell cycle (A), interphase, (B), metaphase and anaphase-telophase. Cells were 
stained using antibodies against pericentrin (centrosome marker) and a-tubulin (microtubules 
and mitotic spindle); chromosomes with DAPI. Scale bars represent 15 µm. The 
corresponding graphs show the mean +/- SEM of 6 separate counts (≥25 cells/line each) for a 
total of 200 cells/line scored.  P-values were calculated by one-way ANOVA with Tukey’s 
correction for multiple testing. 
Figure 7. VPS4A mutations cause aberrant chromosome segregation and alter cell cycle 
progression. A) Staining performed using a fluorescent probe (anti-lamin A/C green) or 
Giemsa show a significant increase in micronuclei (arrow), chromosome bridges (arrow) and 
aneuploidy in proband cells compared to control cells. In experiments to assess micronuclei 
 46 
and chromosome bridges, graphs show the mean +/- SEM of 5 (micronuclei) or 4 
(chromosome bridges) separate counts (200 cells/line each, micronuclei; 250 cells/line each, 
chromosome bridges) for a total of 1000 cells/line scored. In experiments to assess 
aneuploidy, graphs show mean +/- SEM of 3 separate counts (≥30 cells/line each) for a total 
of 100 cells/line scored. B) Cell cycle phases of control (top) and probands’ (bottom) 
fibroblasts as measured by BrdU incorporation and propidium iodide (PI) flow cytometry 
analysis. The upper box identifies cells incorporating BrdU (S phase), the lower left box 
identifies G0/G1 cells, and the lower right box represents G2/M cells. A representative of 
three independent experiments is shown. In all experiments, p-values were calculated by one-
way ANOVA with Tukey’s correction for multiple testing. Graph bars show mean +/- SEM.  
Figure 8. Proband fibroblasts have abnormal nuclear envelope morphology, increased 
DNA damage and abnormal primary cilium morphology. A) An increased number of 
aberrantly shaped nuclei in fibroblasts carrying the VPS4A mutations was observed versus 
control cells. Staining was performed using anti-lamin A/C and DAPI. Scale bar is 10 µm. B) 
Representative images showing an increase number of γ-H2AX foci in probands’ fibroblasts 
carrying compared to control cells. The staining was performed using γ-H2AX antibody and 
DAPI. Scale bar is 2.5 µm. In all experiments, mean +/- SEM of 4 separate counts (≥40 
cells/line each) for a total of 200 cells/line scored. p-values were calculated by one-way 
ANOVA with Tukey’s correction for multiple testing. Graph bars show mean +/- SEM. C) 
Confocal images showing altered primary cilium morphology in proband fibroblasts 
compared to control cells. Cells heterozygous for the p.R284G amino acid change show 
absent cilia with only a visible basal body (dot cilium, zoomed image), whereas cells with the 
p.R284W substitution show either a dot cilium or occasionally an elongated or normal cilium 
(zoomed image). Primary cilia are labelled with acetylated α-tubulin, basal bodies and nuclei 
 47 
are labelled with g-tubulin and DAPI respectively. Scale bar is 5 µm. 100 cells were analysed 
for each line over two independent experiments, bars represent the mean. 
 








 Supplemental Note: Case Reports for representative probands. 
Proband 1 (c.850A>T, p.Arg284Trp).   
Proband 1 is the 1st child of healthy unrelated parents. Her father was 35 years old and 
mother 33 years old at the time of her birth.  Routine antenatal testing showed a low PAPP-
A; a Harmony trisomy screening test was normal.  No abnormalities were noted on 
antenatal ultrasound scan, but in view of the low PAPP-A, serial ultrasound screening was 
undertaken; the appearances were normal until around the 36 weeks scan, when growth 
was noted to be small. A semi-urgent caesarean section was undertaken in view of the 
growth and breech presentation; she was born with birthweight of 2310g (Z -2.8) with a 
head circumference of 31.4cm (Z -2.6) at 37 weeks. 
She cried at birth, and did not need any active resuscitation. She would not breast-feed but 
sucked reasonably well by bottle with expressed breast milk. She failed her newborn 
hearing test twice, but further testing subsequently confirmed normal hearing.   
Parents were concerned about her visual behaviour and that she was not fixing or following 
at 8 weeks of age.  Ophthalmological examination showed significant central and lamellar 
lens opacities but no microphthalmia or anterior segment abnormalities.  Her father at that 
time was found to have very mild symptomatic bilateral sutural lens opacities.  Following 
cataract removal, fundal examination on the proband showed an abnormal retinal 
appearance with abnormal atrophic changes at the macula in keeping with a retinal 
dystrophy, which was confirmed with very poor ERG and VEP amplitudes. 
She developed mild symptoms of cow's milk protein intolerance and gastro-oesophageal 
reflux, which was treated with ranitidine and omeprazole, and started on Nutramigen. 
Feeding has continued to be a problem; at the age of 2 years she takes no solids orally, and 
milk is mild given via NGT. She has persistent vomiting and poor weight gain. 
Developmental progress has been slow. She started smiling from around 6 months of age, 
around which time head control was improving; she was beginning to move her head to 
sounds at 7 months of age. At the age of 2 years, she is able to roll from back to front, and 
pivot on her back using her legs, but is not able to support her head on her hands when 
prone. She is not sitting independently. She laughs responsively and is able to attract 
 parents’ attention by laughing. She recognises her parents’ voices, and will respond to 
familiar stories and songs. She has no words, but has some vocalisations. 
Sleep pattern is poor. She developed obstructive sleep apnoea, and underwent a 
tonsillectomy. She takes melatonin and tends to fall asleep easily, but will often wake after 
30-60 minutes and then repeatedly throughout the night, sometimes staying awake for 
several hours. 
Examination showed central hypotonia, with hands held in a clenched position. She tends to 
hold her mouth open with a protruding tongue. Her movements are dyskinetic. She has 
small hands and feet. 
Array CGH, neurometabolic investigations and a TORCH screen were normal. An 
echocardiogram was reported as normal. NGS of a panel of genes associated with cataracts 
and retinal dystrophy did not identify any pathogenic changes.  She has had an anaemia 
with haemoglobin of 88 g/l in infancy, then later 83 g/l, but possibly dietary owing to her 
intake. The blood film stated that the red cells showed anisocytosis. 
MRI: Sulci are mildly prominent which in conjunction with microcephaly suggest cerebral 
volume loss. The corpus callosum is present; it is thin suggesting white matter volume loss. 
Sylvian fissures are widened anteriorly. Cortical gyration is likely to be within normal limits. 
Cerebellar vermis and hemispheres are hypoplastic and there is a small posterior fossa. 
Brain stem is within normal limits.  
ERG showed no consistent retinal responses above the level of noise to a range of stimulus 
intensities. 
VEPs showed no consistent pattern reversal cortical responses evident above the level of 
noise confirming a very degraded pattern. 
Proband 2 (c.850A>G, p.Arg284Gly).   
An 8-month-old male presented with a phenotype characterized by microcephaly (-5 SD), 
congenital cataract, large ears, long palpebral fissures, strabismus, retrognathia with wide 
mandibular angle, single transverse palmar crease, bilateral clubfoot, hypertonia, global 
developmental delay. The pregnancy was uncomplicated. He was born to partners who were 
cousins, at 39 weeks of gestation, by elective secondary caesarean section. Auxologic 
 parameters were normal, Apgar score was 3 at 1 minute and 9 at 5 minutes after birth, 
orotracheal intubation was not performed. 
Four days after birth he developed severe haemolytic anaemia requiring transfusion. 
Haemolytic episodes recurred leading to liver and spleen enlargement, common causes of 
anaemia and haemolisis were ruled out. Afterwards the proband’s medical history was 
significant for severe oromotor disability and failure to thrive with weight and length < -5 SD, 
parents declined artificial nutrition. He presented with hypovision and moderate bilateral 
sensorineural deafness. 
At 18 months psychomotor development delay was confirmed, cerebral MRI revealed corpus 
callosum hypoplasia and severe cerebellar hypoplasia. Electroencephalography detected 
abnormal brain electrical activity; the proband didn’t present with clinical seizures. The 
neurological pattern got worse and he developed severe intellectual disability and total 
dependence from caregiver. 
At the last clinical evaluation, 3 months before death, he presented with severe malnutrition, 
microcephaly, large ears, long palpebral fissures, strabismus, long oval face, long philtrum, 
gingival hypertrophy, exposed upper incisors, retrognathia with wide mandibular angle, single 
transverse palmar crease, flexion deformity of wrist, elbow, knee with popliteal pterigium, 
bilateral clubfoot, scoliosis of the spine and hypogenitalism. He also showed yellow sclera, 
suggestive of liver involvement. He died at the age of 29 years from respiratory failure 
secondary to pneumonia. 
Proband 3 (c.850A>T, p.Arg284Trp).   
The proband was the second child to unrelated Caucasian parents. He was born at term 
following an uncomplicated pregnancy and discharged home at 6 hours of age. He was 
readmitted within 24 hours with vomiting and drowsiness. He was investigated for sepsis 
and treated with antibiotics. Microcephaly and hypotonia were noted on that admission. He 
continued to vomit and had failure to thrive. Feeding became a significant issue and never 
was successfully established – he had a PEG inserted but vomiting remained problematic. 
He subsequently had a fundoplication and a jejunal tube placed – this helped his symptoms 
but he continued to retch. He made very little in the way of developmental progress – he 
could hold his head in the midline and responded to his parents voice; he never fixed and 
 followed. From the beginning, he had episodes of arching and irritability – initially thought 
to be due to significant gastro-oesophageal reflux.  With time, these evolved into dystonic 
and dyskinetic movements. As his condition progressed, he was diagnosed with ‘salt and 
pepper’ retinitis, cataract and liver dysfunction. He had progressively enlarging liver and 
mild conjugated hyperbilirubinaemia and abnormal liver function tests. He did not develop 
liver failure. Prior to his death he had short episodes of gazing in one direction associated 
with limb movements but was never formally diagnosed with seizures. He died at 26 months 
of age. 
  
He was extensively investigated from a neurometabolic point of view and his positive 
findings included 
1.       MRI - bilateral frontal polymicrogyria and Pontocerebellar hypoplasia 
2.       CK – raised on 2 occasions to 1500, normal on other occasions 
3.       Muscle Biopsy – said to show features of myopathy 
4.       Liver Biopsy – microvesicular steatosis and haemosiderosis 
5.       Respiratory chain analysis – normal 
6.       Microarray 17p11.2 del - VUS paternally inherited 
7.       Cardiff Cortical malformation gene panel– no significant findings 
 
Additional findings at Post mortem included 
1. Skeletal survey showed ‘hair on end’ appearance to outer table of skull  
2. Moderate abdominal ascites 
3. Malrotation and volvulus - Appendix, caecum and proximal ascending colon in left 
upper quadrant 
   
The clinical impression was always of a rare neuro metabolic disorder. 
 
  
 Author List for Genomics England Research Consortium 
Ambrose J. C. 1 , Arumugam P.1, Baple E. L. 1, Bleda M. 1, Boardman-Pretty F. 1,2, Boissiere J. 
M. 1 , Boustred C. R. 1 , Brittain H.1, Caulfield M. J.1,2 , Chan G. C. 1 , Craig C. E. H. 1, Daugherty 
L. C. 1, de Burca A. 1 , Devereau, A. 1 , Elgar G. 1,2, Foulger R. E. 1 , Fowler T. 1, Furió-Tarí P. 1, 
Hackett J. M. 1 , Halai D. 1 , Hamblin A.1, Henderson S.1,2, Holman J. E. 1 , Hubbard T. J.  P. 1, 
Ibáñez K.1,2, Jackson R. 1, Jones L. J. 1,2, Kasperaviciute D. 1,2 , Kayikci M. 1, Lahnstein L. 1 , 
Lawson K. 1, Leigh S. E. A. 1, Leong I. U. S. 1, Lopez F. J.1 , Maleady-Crowe F. 1 , Mason J. 1, 
McDonagh E. M. 1,2, Moutsianas L. 1,2, Mueller M. 1,2 , Murugaesu N. 1, Need A. C. 1,2, Odhams 
C. A. 1, Patch C. 1,2, Perez-Gil D. 1 , Polychronopoulos D. 1, Pullinger J. 1, Rahim T. 1, Rendon A. 
1, Riesgo-Ferreiro P.1, Rogers T. 1 , Ryten M. 1 , Savage K. 1 , Sawant K. 1, Scott R. H. 1 , Siddiq A. 
1, Sieghart A. 1, Smedley D. 1,2 , Smith K. R. 1,2, Sosinsky A. 1,2, Spooner W. 1, Stevens H. E. 1, 
Stuckey A. 1, Sultana R. 1, Thomas E. R. A. 1,2, Thompson S. R. 1, Tregidgo C. 1 , Tucci A. 1,2, 
Walsh E. 1, Watters, S. A. 1, Welland M. J. 1 , Williams E. 1, Witkowska K. 1,2 , Wood S. M. 1,2, 
Zarowiecki M.1 
 
1. Genomics England, London, UK 
2. William Harvey Research Institute, Queen Mary University of London, London, EC1M 6BQ, 
UK. 
 
  
 Supplemental Methods  
 
EGFR degradation assay 
EGFR degradation assays were performed as described previously.32 Briefly, proband 
fibroblasts were serum starved overnight before the addition of 100 ng/mL EGF 
(Calbiochem) in the presence of 10 µg/mL cyclohexamide (Sigma-Aldrich). Cells were lysed 
at 0, 30, 90 and 180 minute time-points, then analysed by western blotting. 
Electron Microscopy 
Fibroblasts were seeded to Thermanox (Thermo Fisher Scientific) plastic coverslips and fixed 
with 2% PFA, 2.5% glutaraldehyde, and 0.1 M cacodylate buffer (pH 7.2). Cells were post-
fixed with 1% osmium tetroxide:1.5% potassium ferricyanide before being incubated with 
1% tannic acid to enhance contrast. Cells were dehydrated using increasing percentages of 
ethanol before being embedded onto EPON stubs or beam capsules. Resin was cured 
overnight at 65°C, and coverslips were removed using a heat-block. Ultrathin (50- to 70-nm) 
conventional sections were cut using a diamond knife mounted to a Reichart ultracut S 
ultramicrotome. Sections were collected onto copper grids stained using lead citrate. 
Sections were viewed on a FEI Tecnai transmission electron microscope at a working voltage 
of 80 kV. 
  

 Figure S1. Multiple sequence alignment of VPS4 proteins from representative species.  
The novel de novo missense variants in VPS4A are represented in the corresponding columns 
of the alignment, using the same symbols as in Figure 1. The top track depicts CCRs 
percentiles for human VPS4A (hVPS4A). The protein domains depicted in the second track 
are based on hVPS4A (UniProtKB entry Q9UN37). The third track shows secondary structural 
elements based on PDB structures for different regions of hVPS4A or human VPS4B 
(hVPS4B). In the multiple sequence alignment residues are coloured according to their 
physicochemical properties (Jalview color scheme). PONAB= orangutan; MOUSE= mouse; 
CANLF= dog; CHICK= chicken; XENTR= frog; DANRE= zebrafish; DROME= drosophila; CAEEL= 
worm; ARATH= arabidopsis; 9FUNG= fungus; DICDI= slime mold; SCHPO= yeast; CANAL= 
Candida albicans; 9ARCH= heimdallarchaeota. The bottom track shows amino acid 
conservation. 
  

 Figure S2. Protein structural superposition of the ATPase domain of A) the 6 chains of the 
VPS4 hexamer from yeast (PDB ID: 6OO2) and B) the monomeric VPS4B from human (PDB 
ID: 1XWI) and mouse (PDB ID: 2ZAM, 2ZAO, 2ZAN). In (B), chains in white correspond to 
apo structures and chains in blue have ADP or ATP bound. The superposition of all chains 
was done using the human structure as reference. In yellow is the synthetic peptide that 
shows how the ESCRT-III proteins would translocate through the pore. Residues with missing 
coordinates are represented with dashed lines. The flexible linker connector is missing in 
human and mouse structures (B) but can be partially observed as α-helix in four chains of 
yeast VPS4 structure (A). 
  

 Figure S3. Additional clinical images of probands.  A) Photographs showing probands 1, 2 
and 5.  Note common features of microcephaly, long palperal fissures, strabismus, long 
relatively smooth philtrum and broad and arched eyebrows. In the older probands, large 
ears, long oval face, gingival hypertrophy, exposed upper incisors and retrognathia with wide 
mandibular angle are common features. Proband 2 also had jaundice, a single transverse 
palmar crease, flexion of wrist, elbow, knee with popliteal pterigium, bilateral clubfoot and 
scoliosis. B) MRI images of probands 1 and 5. Proband 1: Sagittal and axial T2 MRI images, 
Proband 5, left is sagittal T1 MRI image and right is T2 flair axial image. Abnormal features in 
both include thinning of the corpus callosum, atrophy or under-development of the 
cerebellum, pontine hypoplasia and white matter volume loss. In proband 1 there is 
extensive dysgyria and in proband 5 there is periventricular gliosis and scalloping of the 
lateral ventricles. 
  

CD63 LAMP1 
1.0 
0.8 
0.6 
0.4 
0.2 
Control-1 Control-2 R284W R284G 
Figure S4 Part 2 
Figure S4. VPS4A proband fibroblasts show no endosomal compartment content mixing. 
Cultured fibroblasts from control subjects and the probands indicated were fixed, labelled 
with the early endosomal marker RAB5 and CD63 (A), EEA1 and LAMP1 (B) and CD63 and 
LAMP1 (C), then visualised by confocal immunofluorescence microscopy. Colocalisation was 
measured using Pearson’s co-efficient (in 20 cells in each repeat), and the result of n=3 
biological repeats for each set of markers is quantified in the corresponding charts. Bars 
indicate mean ± S.E.M., p-values were calculated by one-way ANOVA with Tukey’s post-hoc 
test for repeated measures. Micrograph scale bars = 10 µm. 

 Figure S5. Proband fibroblasts show no defects in endosomal functions of the core ESCRT-
III complex. A-B) Cultured fibroblasts from the control subjects or the probands indicated 
were fixed, labelled against EEA1 (A) or CD63 (B) and the core ESCRT-III complex member 
CHMP2B, then visualised by confocal immunofluorescence microscopy. In (A) the number of 
CHMP2B puncta and the percentage of EEA1-positive endosomes associated with a CHMP2B 
punctum was quantified in 3 experiments (5 cells per experimental condition in each repeat) 
and plotted in the graphs beneath the images. In (B) the percentage of CD63-positive 
endosomes associated with a CHMP2B punctum was quantified in 3 experiments (in 5 cells 
per experimental condition in each repeat) and plotted beneath the images. Micrograph 
scale bar = 10 µm. Arrows indicate juxtaposed or co-localised puncta. C) Fibroblasts derived 
from the subjects indicated were imaged for transmission electron microscopy. MVBs 
containing ILVs were evident in all cases (indicated by *).  Scale bar = 500 nm. D) Fibroblasts 
from probands and controls were stimulated with EGF, lysed at the times indicated, then 
immunoblotted to monitor EGFR degradation. The line chart on the left shows EGFR 
quantification normalised to the initial amount of EGFR for a representative experiment, 
while the plot on the right shows the percentage of EGFR remaining at the 180 min time 
point for n=3 such experiments. In all plots, bars indicate mean ± S.E.M., p-values calculated 
by one-way ANOVA with Tukey’s post-hoc test for repeated measures. 
 
  

 Figure S6. No increased recruitment of IST1 to late endosomes or lysosomes in proband 
fibroblasts. Cultured fibroblasts from the control subjects and the probands indicated were 
fixed and labelled for A) IST1 and CD63 or B) IST1 and LAMP1. The percentage of CD63 or 
LAMP1-positive endosomes associated with an IST1 punctum was quantified in 3 
experiments per marker (in 5 cells per experimental condition in each repeat) and plotted in 
the corresponding charts. Arrows indicate juxtaposed or co-localised puncta. Bars indicate 
mean ± S.E.M., p-values were calculated by one-way ANOVA with Tukey’s post-hoc test for 
repeated measures. Micrograph scale bar = 10 µm. 
 
  

 Figure S7. Expression of VPS4A containing disease-associated sequence changes causes the 
development of vacuolar endosomal structures in iPSCs. A) I3 iPSCs were lentivirally 
transduced with expression constructs for myc-tagged wild-type VPS4A, or forms of VPS4A 
containing the p.R284G or p.R284W disease-associated changes. Cell lysates were then 
immunoblotted with the antibodies indicated. GAPDH immunoblotting serves as a control to 
verify equal protein loading in each lane. B) Transduced cells were also fixed and labelled 
with antibodies against the myc epitope and RAB5, then visualised by confocal 
immunofluorescence microscopy. Examples of rare cells containing large vacuolar 
endosomal structures are shown in the inset higher magnification boxes. Micrograph scale 
bar = 10 µm. 
  

  
Figure S8. Time course experiment demonstrating loss of pluripotency and gain of 
neuronal characteristics in i3N iPSCs and neurons. A) i3N iPSCs were cultured in the 
presence of doxycycline for 3 days. Cell lysates were then made at the time points indicated 
and blotted against the neuronal differentiation markers MAP2, Tau and bIII-tubulin, as well 
as the pluripotency markers Oct-4 and NANOG. GAPDH immunoblotting serves as a control 
to validate equal protein loading across lanes. B-E) i3N iPSCs (B and C) and i3Neurons 14 days 
post induction of differentiation (D and E) were fixed and processed for confocal 
immunofluorescence microscopy with the pluripotency (Oct-4, NANOG and SOX2) and 
neuronal markers (TAU and bIII-tubulin) indicated. 
 
 
 
 
 
 
Patient number Patient 1 Patient 2 Patient 3 Patient 4 Patient 5 Patient 6
Gender (M/F) F M M F F M
Ethnicity Caucasian (White British) Caucasian Caucasian (White British) Caucasian Caucasian Caucasian (White British)
Genetic Infirmation
(Genome Build) 38 37 38 38 38
gDNA (Genome Build) 16:69320768 A>T chr16:69354671 A>G 16:69320768 A>T 16:69320768 A>T 16:69320768 A>T 16:69319539G.A 
cDNA c.850A>T c.850A>G c.850A>T c.850A>T c.850A>T c.616G>A
Protein (NM_013245.2) p.Arg284Trp p.Arg284Gly p.Arg284Trp p.Arg284Trp p.Arg284Trp p.Glu206Lys
Mutation type missense missense missense missense missense missense
REVEL score 0.97 0.983 0.97 0.97 0.97 0.917
CADD score 35 34 35 35 35 34
Inheritance de novo de novo de novo de novo de novo de novo
 Analysis method WGS WES/WGS WGS WGS WES WGS
 Clinical Information 
Age at last examination (y or m) 2 y 29 y 2 y 6 m 6 y
Deceased (age) no yes (29 y) yes (26 m) no no
Cause of death respiratory failure complicating pneumonia bowel ischaemia and volvulus
Neurologic
Intellectual disability (IQ or severity) severe severe severe severe severe severe
Developmental delay severe severe severe severe poor head control severe severe
Motor delay yes severe severe yes yes
Speech delay yes severe absent yes absent
Epilepsy no yes no yes no yes
Hypotonia yes yes yes yes yes no
Spasticity mild in upper limbs yes yes no yes
Ataxia yes not assessed - severe motor delay no non-ambulatory; dysmetria no
Dystonia yes profound yes yes yes yes no
Other abnormal movements no no chorea no no
Sleep disturbances yes yes unknown yes; central sleep apnea yes
Brain imaging
Major features hypoplastic cerebellar vermis and hemispheres cerebellar hypoplasia
bilateral polymicrogyria and 
pontocerebellar hypoplasia pontocerebellar hypoplasia
progressive pontocerebellar atrophy 
involving vermis and hemispheres; 
severe cerebral atrophy on CT 
at 1 year
Other MRI findings thin corpus callosum corpus callosum hypolasia follow - up MRI parenchymal loss and possible dysmyelination
decereased cerebral white matter 
volume with preiventriculat gliosis
Growth Parameters
Delivery (wks+days) 37 39 40 38+6 33 38+4
Weight at birth (gr; SD or centile) 2310 g (Z -2.8) 3120  g (25th) 2450 g (Z -2.4) 2083 g (50th-75th) 2073 g (<0.04th)
Length at birth (cm; centile) 48 cm (9-25th) 45 cm (75th)
Head circumference at birth (cm; SD or 
centile) 31.4 cm (Z -2.6) 35 cm (50th) 30.5 cm (<0.04th, Z< -3) 25 cm (<0.04th, Z<-3) 31 cm (0.04th-2nd)
Microcephaly yes yes yes yes yes yes
Height (cm; SD) 74.5 cm (Z -1.87) 115 cm (Z -5.5) 78 cm (at 26m) (Z -3.0 ) 124 cm (Z 1.4)
Head circumference (cm; SD or centile) 46 cm (Z -6.8) 42 cm (at 26m)  (Z -4.5) 39 cm (at 6 m) (Z -5) 42.3 cm (<< 0.04th) 46 cm (Z -5.0)
Weight (kg; SD) 7.645 kg (Z -2.31) 15 kg (Z -6.0) 12.45 kg (at 26m) (Z -0.41) 6.56 kg (at 6 m) (Z -1.1) 21.3 kg (Z 0.17) 2nd-9th centiles
Feeding difficluties yes - severe gastroesophageal reflux yes yes yes - gastroeosophageal reflux no
Nasogastric (NG) feeding or percutaneous 
endoscopic gastrostomy (PEG) NG/PEG-J no (declined)
PEG/ fundoplication and jejunal 
tube yes no
Eye phenotype
Congenital Cataract yes yes yes no yes
Retinal Dystrophy yes no salt and pepper retinitis no yes
Leber congenital amaurosis no no no yes no
Vision no fixing or following poor  fixing and no following no fix and following poor  fixing and following at 6 m diminished diminished
Staphyloma yes no no
Aphakia yes (post-operative) no no no yes
Liver function
Hepatomegaly yes and conjugated bilirubinaemia at 1 yr mild hepatosplenomegaly
progressive hepatomegaly from 1 
year of age congenital hepatosplenomegaly no
Other raised AFP no haemosiderosis mild cholestasis,  iron overload secondary to CDA-1
Gallstones no yes microvesicular steatosis no no
Musculoskeletal anomalies
Lipodystrophy no yes no no no
Abnormal CPK no yes mild on 3 occasions yes on 2 occasions not done no
Muscle biopsy no no myopathic no no
Scoliosis no yes, severe no no yes
Toe abnormalities II toes hypoplasia no short toes no
Hip dysplasia no yes no coxa valga yes
Single palmar creases no yes no no no
Talipes (bilateral) no yes no yes no yes
Haematological anomalies
Anemia
yes- raised reticulocyte count and 
platelets 
blood film - anisopoikilocytosis
macrocytic anemia with 
anisopoikilocytosis no yes no
Hemolytic crisis no yes no no no
Bone marrow biopsy/aspiration no dyserythropoiesis no - normal at post mortem no
Congenital dyserythropoitic anaemia yes no yes no
Other features
Renal defects no no no no
Liver fibrosis no no no no
Sensorineural deafness no yes no yes no
Hypogonadism no yes no no
Stipsis no yes yes yes no
Stomatitis no yes no no no
Dental anomalies no yes no no (delayed eruption)
Recurrent infections no yes no no no
Aphakia post operative no no no yes
Other comments left testicular torsion and right undescended testis
Table S1. Genetic and clinical features of patients with de novo heterozygous VPS4A sequence alterations.