IntroductionWD repeat and FYVE domain-containing protein 3 (WDFY3) encodes an autophagosome scaffold protein involved in the selective recruitment and degradation of macromolecular cellular components, such as aggregated proteins [1]. WDFY3 is part of the beige and CHS (BEACH) domain-containing protein family and includes a BEACH domain, a pleckstrin homology (PH) domain, five WD40 domains, and a C-terminal Fab1/YOTB/Vac1/EEA1 (FYVE) domain that facilitates integration into vesicular membranes [2]. It is involved in key cellular processes including autophagy and endocytosis, is evolutionarily conserved from invertebrates, and is uwidely expressed [3], suggesting its functional relevance in maintaining normal cellular and organismal functions.In our previous work, we described 13 probands, mostly with de novo WDFY3 variants, who presented with intellectual and/or developmental deficits. Truncating variants were associated with increased brain size, while missense variants in the PH domain were linked to microcephaly and developmental delay. Wdfy3-haploinsufficient mice also exhibited macrocephaly and learning deficits, mirroring the human phenotype. We thus postulated that WDFY3 haploinsufficiency leads to increased brain size and mild to moderate neurodevelopmental delay (NDD) [4].Prior studies show WDFY3 controls neuronal progenitor division and neuronal migration [5]. Using mosaic analysis, it was found that homozygous neuronal Wdfy3 loss causes cortical lamination defects and reduced dendritic arborization [2]. Yet, the effects of WDFY3 haploinsufficiency in human neurons remain underexplored.Here, we review 9 previously reported individuals [4] and describe 23 new cases with protein-truncating WDFY3 variants to refine the clinical phenotype. In SH-SY5Y neuroblastoma cells, WDFY3 loss-of-function (LoF) disrupted genes involved in nervous system development, MAPK/ERK1/2, and WNT/β-catenin signaling. Functional assays showed WDFY3 knockdown (KD) increases neuronal division and decreases neuronal differentiation, downregulates WNT/β-catenin, and upregulates MAPK/ERK1/2. Importantly, LiCl treatment restored WDFY3 expression and improved neuronal outgrowth and arborization in heterozygous Wdfy3-LoF cells.MethodsGenetic diagnosisMost probands underwent exome sequencing, processed at various centers using in-house or commercial tools. WDFY3 variants not confidently identified via trio sequencing or low-quality data were confirmed via Sanger sequencing. Sequencing methods are detailed in prior publications [4, 6] and Supplemental Text 1.SH-SY5Y cell cultureSH-SY5Y cells were purchased from the Leibniz Institute DSMZ − German Collection of Microorganisms and Cell Cultures. Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM)/F12 medium supplemented with 15% fetal calf serum (FCS) and glutamine (2 mM) in a humidified atmosphere containing 5% CO2 at 37 °C. Cells were passaged at ~75% confluency and discarded after 15 passages to prevent senescence.Animal husbandryThe Wdfy3lacZ mouse line was generated by targeted insertion of a lacZ reporter cassette as part of the Knockout Mouse Project (KOMP) and maintained on C57BL/6NJ background as previously described [5]. Mice were housed in Plexiglas cages (55 × 33 × 19 cm) and maintained in facilities approved by Association for Assessment and Accreditation of Laboratory Animal Care International under standard laboratory conditions (21 ± 2 °C; 55 ± 5% humidity) on a 12-h light/dark cycle, with ad libitum access to both water and standard rodent chow. In this study, both male and female mice were utilized similarly to ensure a balanced and comprehensive investigation. No sex differences were noted in the results.Cortical neuron culture from P0 pupsMatings were set up between wild type (+/+) and heterozygous (+/-) Wdfy3lacZ mice [5]. Newborn pups (P0) were euthanized by decapitation, and cortices were dissected in HBSS under sterile conditions. Cortical hemispheres were transferred to sterile tubes, excess HBSS removed, and incubated with 1 ml trypsin at 37 °C, 300 rcf for 15 min. After removing trypsin, 1 ml fetal bovine serum (FBS) was added and mixed gently. The tissue was washed twice with HBSS, then triturated in 1 ml HBSS to dissociate cells. The suspension was filtered through a 40 μm strainer and centrifuged. The pellet was then resuspended in warm 1 ml full plating medium which consisted of Neurobasal medium supplemented with 2% B-27 Supplement, 250 μM L-glutamax, 50 μg/ml penicillin-Streptomycin, and 50 μg/ml gentamicin.Cells were counted using trypan blue and a hemocytometer. They were plated on poly-D-lysine-coated 100 × 20 mm dishes at 1.2 × 10⁶ cells/dish. Right and left cortices were processed separately, with one hemisphere treated with 5 mM LiCl and the other left non-treated, enabling within-brain comparisons. Medium was changed on day 4, and cultures were maintained for 8 days to allow attachment and neurite outgrowth.For harvesting, cells were trypsinized at 37 °C for 3 min, trypsin was inactivated with FBS, and cells were centrifuged at 300 rcf for 5 min. The pellet was collected for RNA isolation.Cortical neuron culture from E13.5 embryosDay 13.5 embryos were collected from timed pregnancies of Wdfy3+/lacZ dams crossed with Wdfy3+/lacZ males. Dams were euthanized with Isoflurane, and embryos harvested in cold PBS. Amnion and yolk sacs were collected for genotyping, and brains were dissected and placed in PBS. After removing meninges, cortical hemispheres were isolated into HBSS under a cell culture hood. Cell isolation then followed the P0 protocol. After counting, cells were plated on poly-D-lysine-coated coverslips in 24-well plates at 20,000 cells per well. Treatment was again divided between hemispheres, and the cultivation process followed the P0 protocol.WDFY3 siRNA-mediated knockdown in SH-SY5Y cellsTo ensure optimal growth, cells were re-seeded at ~ 3 500 cells/cm² one day before transfection. Cells were trypsinized, counted, divided into experimental groups, and washed once with DPBS. Transfection was performed using the Neon™ Transfection System (Invitrogen, Waltham, Massachusetts, USA) per the manufacturer’s protocol. Either scrambled control siRNA or a pool of four WDFY3 siRNAs (10 nM final concentration; see Supplemental Table 1) were added to the cell pellet. Post-transfection, cells were transferred to pre-warmed medium and seeded into multiwell plates for further analysis or treatment.Plasmid and siRNA co-transfection of SH-SY5Y cellsThe Exon1Htt103Q plasmid was a kind gift from Prof. Anne Simonsen, University of Oslo. This plasmid was co-transfected with siRNAs using 0.5 µg of plasmid per transfection using the electroporation protocol as described above.RNA sequencing and differential gene expressionWe followed previously described methods [7, 8]. Briefly, RNA was extracted as outlined below (see qPCR), and quality assessed using the Agilent Bioanalyzer with the RNA 6000 Pico kit (Agilent, Santa Clara, California, USA). RNA-seq libraries were prepared from 500 ng total RNA using the TruSeq RNA Library Prep Kit v2 (Illumina, San Diego, California, USA) and sequenced on an Illumina NovaSeq 6000 with 151 bp paired-end reads.Reads were aligned to the hg38 human genome using STAR (version 2.6.1d [9]), and transcript levels quantified with htseq-count (v0.6.0). Genes with G, p.(Ser2834*); c.8313T > G, p.(Tyr2771*); c.6073 + 2T > C, p.?) [22,23,24].The truncating variants were distributed across the entire protein length (Fig. 1), but none were located within the last coding exon. According to ACMG criteria, all truncating variants were classified as pathogenic (PVS1, PM2, and PS2 if de novo or PP1 if inherited from an affected parent) [25]. We propose classifying variants where PVS1 and PM2 apply as likely pathogenic.Fig. 1: Identified de novo truncating WDFY3 variants in human subjects with developmental delay.Full size imageVariants in bold have not been published, the others were published in [4] or [22,23,24]. The c.8098-1 G > C variant is present in a pair of monozygous twins. Variants marked in blue are paternally inherited, and variants marked in red are maternally inherited.All probands exhibit signs of NDD, i.e., a delay in the development of motor and language skills or intellectual disability (ID). Formal IQ data was available for 24 probands. Mild ID was defined as an IQ of 50–69, and moderate ID was defined as an IQ of 35–49 [26]. Sixteen probands (67%) had mild to moderate ID. Head circumference measurements were available for 24 probands, 21 (88%) of whom showed macrocephaly or a normal-high occipito-frontal circumference (OFC > 80th percentile). Most probands also displayed neurobehavioral disorders: 23/32 (72%) were diagnosed with autism spectrum disorder (ASD), 9/32 (28%) showed signs of attention deficit hyperactivity disorder (ADHD), and 2/32 (6%) had obsessive-compulsive disorder (OCD).A fetal case (c.349_350del, p.(Ser117*)) with left heart dysplasia, mitral atresia, aortic valve stenosis, ventricular septal defect, and normal head circumference at 25 weeks of gestation has been reported in the literature [27]. This finding parallels observations in Wdfy3-haploinsufficient mice presenting with multiple congenital heart defects [28]. Yet, none of our probands exhibit heart defects. For nine patients cerebral MRIs were conducted, out of whom four were reported with unspecific abnormalities.Two monozygotic twins (c.8098-1G > C, p.?) show a nearly identical clinical phenotype. One proband (c.7016del, p.(Tyr2339Serfs*)) had a low-normal OFC (11th percentile), associated with growth restriction (height and weight T, p.(Arg2673*)) where the father exhibits macrocephaly (+4.8 SD), mild NDD, and autism.3)A maternally inherited variant (c.3257_3258del, p.(Gly1086Aspfs*5)) where the mother is described as having a mood disorder, though no further clinical details were available.4)Another maternally inherited variant (c.1906C > T, p.(Arg636*)) where the mother shows no specific phenotype apart from minor dysmorphic facial features.These cases highlight the reduced penetrance and variable clinical phenotype of WDFY3 LoF variants. A detailed clinical description and photos of three individuals are provided in the supplementary material, with a summary of proband data in Fig. 1 and Table 1.Table 1 Main clinical features of probands, who are newly described in the current study.Full size tableWDFY3 knockdown impairs selective autophagyWDFY3 is involved in selective autophagy, acting as a scaffold protein [32, 33]. To investigate whether WDFY3 KD alters the clearance of aggregated proteins, we co-transfected neuronal SH-SY5Y cells with WDFY3 or control siRNA and a plasmid encoding exon 1 of the huntingtin gene followed by a polyglutamine expansion (Exon1Htt103Q) and tagged with green fluorescent protein (GFP). This protein forms aggregates that are broken down through selective autophagy [34]. We detected increased levels of the Exon1Htt103Q protein in WDFY3 KD cells both through immunostaining and Western blotting (Fig. 2A−C), indicating impaired autophagic substrate flux and accumulation of aggregated protein.Fig. 2: In vitro WDFY3 KD in neuronal cells impairs selective macroautophagy.Full size imageA Immunostaining of control and WDFY3 KD cells that were co-transfected with a GFP-tagged Exon1Htt103Q plasmid. Cells were stained for the cytoskeletal protein β-actin (red) and nuclei were counterstained with Hoechst 33342 (blue). Scale bar 125 μm. Cells with a WDFY3 KD show more GFP aggregates (green). B Immunostaining of control and WDFY3 KD cells: Cells were stained for p62/SQSTM1 (green) and nuclei were counterstained with Hoechst 33342 (blue). Scale bar 75 μm. WDFY3 KD cells show fewer p62 puncta when compared to cells transfected with control siRNA. C Representative Western blots of control and WDFY3 siRNA- and Exon1Htt103Q-cotransfected SH-SY5Y cells: Expression of GFP-tagged Exon1Htt103Q protein was increased by 1.1992 ± 1.328-fold. Additionally, we detected a trend towards lower protein expression of the ubiquitin-binding protein p62 in WDFY3 KD cells (reduction by 0.616 ± 0.319-fold, n = 5, p = 0.095).Furthermore, we observed decreased protein expression of p62 (also known as sequestosome-1, SQSTM1) and fewer p62 puncta in WDFY3 KD cells (Fig. 2B, C and Supplemental Fig. 1A). Since WDFY3 and p62 co-localize with poly-ubiquitinated proteins targeted for autophagic degradation [32], these findings further support that WDFY3 is crucial for the proper processing of poly-ubiquitinated proteins. Moreover, these results demonstrate that WDFY3 KD results in the expected impairment of selective macroautophagy and a loss of protein function.Transcriptomics of SH-SY5Y cells with WDFY3 knockdownTo understand the molecular mechanisms underlying WDFY3 LoF in the nervous system, we transiently knocked down WDFY3 in SH-SY5Y cells and performed genome-wide RNA-seq to identify transcriptomic changes in non-differentiated neuronal cells. KD efficiency was evaluated by transcriptome sequencing (log2FoldChange = −1.85, padj < 0.001) and qPCR (Fig. 3A). Principal component analysis revealed a clear separation of WDFY3 KD and control groups, allowing us to proceed with the differential expression analysis. We identified 130 dysregulated genes (padj < 0.05).Fig. 3: WDFY3 KD alters WNT/β-catenin and MAPK/ERK1/2 signaling pathways.Full size imageA WDFY3 expression in control and WDFY3 KD SH-SY5Y cells: WDFY3 expression is reduced to 0.271 ± 0.042-fold (normalized to hypoxanthine phosphoribosyltransferase (HPRT) and TATA-box–binding protein (TBP) expression, n = 5, p = 0.002). B Western blots of control and WDFY3 KD SH-SY5Y cells 24 h after transfection: β-catenin protein decreases to 0.794 ± 0.112-fold (normalized to α-tubulin, n = 7, p = 0.023). C GSK3B expression in control and WDFY3 KD SH-SY5Y cells: GSK3B was upregulated 1.182 ± 0.1-fold (normalized to the means of HPRT and TBP expression, n = 5, p = 0.008). D Western blots of control and WDFY3 KD SH-SY5Y cells 5 days after transfection: ERK1/2 increased 1.67 ± 0.336-fold (normalized to α-tubulin, n = 4, p = 0.023).Gene ontology overrepresentation analysis revealed that pathways related to “autonomic nervous system development”, “axonogenesis”, “axon development”, and “cell growth” were affected. To understand the underlying molecular mechanisms, we explored signaling pathways enriched in up- or down-regulated genes. We found that the dysregulated genes impact downregulated GO terms such as the WNT/β-catenin, MAPK/ERK1/2, and NOTCH signaling pathways (Supplemental Fig. 3A, B). To validate these findings, we proceeded with functional testing.Dysregulated signaling of WNT/β-catenin pathway and MAPK/ERK1/2 pathwayOne of the downregulated signaling pathways identified in the RNA-Seq is the WNT/β-catenin pathway, which is crucial for brain development, including neuronal stem cell polarity, division, and maturation [35]. We previously reported that WNT/β-catenin signaling is downregulated in Wdfy3-haploinsufficient mice [4]. Additionally, WDFY3 regulates dishevelled-3, an intermediary in the WNT/β-catenin pathway, by promoting its breakdown via selective autophagy [33].WDFY3 KD cells show lower levels of endogenous β-catenin (Fig. 3B), likely due to WNT/β-catenin pathway downregulation. However, this effect was only significant in cells analyzed 24 h after transfection, suggesting a possible counterreaction. It is possible that in response to this, GSK3B expression is upregulated (Fig. 3C). Furthermore, our transcriptomic analysis of cells five days after transfection revealed downregulation of CTNNB1, encoding for β-catenin, and an upregulation of AXIN2, a downstream target of WNT/β-catenin signaling. Interestingly, both WNT5a and WNT5b, which activate non-canonical WNT/calcium signaling, were also downregulated (Supplemental Table 2) [36]. We, however, found no other conclusive effects of WDFY3 KD on other parts of canonical WNT/β-catenin signaling in cells analyzed 24 h after transfection: neither was the protein expression of GSK-3β altered, likely due to the long half-life of the protein [37], nor were the gene expressions of AXIN2 or CCND1 encoding Cyclin D1 changed (Supplemental Fig. 1B-D). There was also no effect on the protein expression of any dishevelled isoforms (Supplemental Fig. 1E). We further confirmed these results on cortical neurons of Wdfy3-haploinsufficient mice (Supplemental Fig. 2A−D).Previous studies have shown that ERK1/2 inactivation leads to increased differentiation and decreased proliferation of both neuronal stem cells [38] and in the spinal cord [39]. Our transcriptomics analysis identified an upregulation of MAPK3, which encodes ERK1 (Supplemental Table 2). We confirmed this finding at the protein level by detecting an upregulation of total ERK1/2 in WDFY3 KD vs. control cells (Fig. 3D). However, this effect was significant only in cells analyzed five days after transfection, not 24 h after transfection.WDFY3 knockdown leads to increased proliferation and decreased neuronal cell differentiationSince both WNT/β-catenin and MAPK/ERK1/2 signaling majorly influence proliferation and neuronal cell differentiation, we were prompted to investigate the effects of WDFY3 KD in our in vitro model next.To determine the effects of WDFY3 KD on proliferation, SH-SY5Y cells were cultured for four days following electroporation. At the end of the experiment, the number of cells was significantly greater in the WDFY3 KD group (Fig. 4A). This finding was further supported by increased expression of the proliferation marker Proliferating Cell Nuclear Antigen (PCNA) (Fig. 4B) and increased Ki-67 staining of proliferating cells (Fig. 4C) in WDFY3 KD cells.Fig. 4: Knockdown of WDFY3 leads to increased proliferation and decreased differentiation of neuronal cells.Full size imageA Hoechst nuclei staining of control and WDFY3 KD SH-SY5Y cells: cell count significantly increased 1.438 ± 0.255-fold after 4 days of culture in WDFY3 KD cells (n = 5, p = 0.01). B PCNA gene expression in control and WDFY3 KD SH-SY5Y cells 24 h after transfection: PCNA shows a 1.17 ± 0.179-fold trend towards a higher expression in WDFY3 KD SH-SY5Y cells (normalized to the means of HPRT and TBP expression, n = 6, p = 0.065). C Ki-67 immunofluorescence staining of control and WDFY3 KD SH-SY5Y cells one day after transfection: WDFY3 KD cells presented a 1.129 ± 0.053-fold (n = 4, p = 0.011) greater percentage of Ki-67 positive cells. D Bright field images of control and WDFY3 KD SH-SY5Y cells differentiated for five days, scale bar 200 µm: Neurite protrusions in WDFY3 KD cells were 0.736 ± 0.101-fold shorter than those in control cells by an average of 30.9 µm (n = 5, p = 0.005).To determine the effects on neuronal cell differentiation, transfected SH-SY5Y cells were morphologically assessed after five days of treatment with 10 µM retinoic acid [40, 41]. We observed significantly shorter and fewer neurite-like protrusions in the WDFY3 KD group. Additionally, visually fewer cells in the WDFY3 KD group appeared to survive the differentiation process (Fig. 4D).WDFY3 is upregulated during SH-SY5Y differentiation and by in vitro lithium chloride treatment in SH-SY5Y and heterozygous cortical neuronsSince WDFY3 KD alters neural differentiation, we were prompted to investigate whether the expression of WDFY3 differs between differentiated and non-differentiated neural cells. Our analysis revealed elevated WDFY3 expression in differentiated SH-SY5Y cells compared to non-differentiated cells (Fig. 5A), suggesting that WDFY3 plays a role in neural cell maturation. Given the potential therapeutic implications of the relationship between WDFY3 and neural differentiation, we sought to identify substances that could enhance WDFY3 expression.Fig. 5: WDFY3 expression is increased in differentiated SH-SY5Y cells; treatment with lithium chloride leads to increased WDFY3 expression and improves the dendritic arborization of Wdfy3lacZ cortical neurons.Full size imageA WDFY3 gene expression in control siRNA-transfected nondifferentiated and differentiated SH-SY5Y cells: WDFY3 was 2.672 ± 0.866-fold higher expressed in differentiated control siRNA-transfected cells compared to non-differentiated cells. (n = 4, p = 0.018). Gene expression in control siRNA-transfected SH-SY5Y cells cultured in medium containing 0.1% FCS and 0 or 5 mM LiCl for 24 h: B WDFY3 increased by 1.448 ± 0.367-fold (n = 5, p = 0.045). C AXIN2 increased by 1.213 ± 0.216-fold (n = 5, p = 0.024). D GSK3B increased by 1.512 ± 0.326-fold (n = 5, p = 0.022). E Western blots of SH-SY5Y cells that were cultured in medium containing 0.1% FCS and 0 or 5 mM LiCl for 15 min. Phosphorylated GSK-3β (Ser9) was significantly increased in 5 mM compared to 0 mM treated SY5Y cells. F Wdfy3 gene expression in cortical neurons of wild type and heterozygous Wdfy3lacZ pups (P0) cultured in medium containing 0 or 5 mM LiCl for 8 days: Wdfy3 increased by 9.027 ± 4.403-fold (n = 3, p = 0.024) in wild type cells and by 100.515 ± 182.895-fold (n = 4, p = 0.049) in heterozygous cells. G Tuj1 staining and Sholl profile analysis of LiCl effects on cultured cortical neurons from Wdfy3lacZ embryos (E13.5). Representative Tuj1 (green) staining of cultured cortical neurons (E13.5) from Wdfy3lacZ mice treated with 5 mM LiCl for eight days. Nuclei were stained with DAPI (blue). Dendritic arborization of wild type (+/+) cortical neurons did not significantly differ with or without LiCl treatment. Compared with no LiCl treatment, LiCl treatment increased the number of intersections in heterozygous (+/−) cortical neurons. Scale bar, 20 µm. Tracings are shown in Supplemental Fig. 4. H Sholl profile analysis using a mixed-model ANOVA revealed increased dendritic arborization in heterozygous cortical neurons treated with LiCl (5 mM) than in nontreated heterozygous cortical neurons. Compared with nontreated wild type cells, nontreated heterozygous cortical neuronal cells presented fewer intersections. No significant difference was observed between the WT (5 mM) and non-treated (0 mM) cells. *p