SHANK2 mutations impair apoptosis, proliferation and neurite outgrowth during early neuronal differentiation in SH-SY5Y cells

SHANK2 mutations have been identified in individuals with neurodevelopmental disorders, including intellectual disability and autism spectrum disorders (ASD). Using CRISPR/Cas9 genome editing, we obtained SH-SY5Y cell lines with frameshift mutations on one or both SHANK2 alleles. We investigated the effects of the different SHANK2 mutations on cell morphology, cell proliferation and differentiation potential during early neuronal differentiation. All mutant cell lines showed impaired neuronal differentiation marker expression. Cells with bi-allelic SHANK2 mutations revealed diminished apoptosis and increased proliferation, as well as decreased neurite outgrowth during early neuronal differentiation. Bi-allelic SHANK2 mutations resulted in an increase in p-AKT levels, suggesting that SHANK2 mutations impair downstream signaling of tyrosine kinase receptors. Additionally, cells with bi-allelic SHANK2 mutations had lower amyloid precursor protein (APP) expression compared to controls, suggesting a molecular link between SHANK2 and APP. Together, we can show that frameshift mutations on one or both SHANK2 alleles lead to an alteration of neuronal differentiation in SH-SY5Y cells, characterized by changes in cell growth and pre- and postsynaptic protein expression. We also provide first evidence that downstream signaling of tyrosine kinase receptors and amyloid precursor protein expression are affected.

Neurodevelopmental disorders have a high prevalence (1-3%) and are often caused by genetic factors. Genes of the SHANK (SH3 and multiple ankyrin repeat domains protein) family (comprising SHANK1, SHANK2 and SHANK3) have been linked to a spectrum of neurodevelopmental disorders 1 . Deleterious de novo SHANK2 mutations have been identified in individuals with intellectual disability (ID), autism spectrum disorders (ASD), developmental delay, and attention deficit and hyperactivity disorder [2][3][4][5] . In the SFARI gene database 6 SHANK2 is categorized as a high confidence autism risk gene. In addition, SHANK2 has been linked to the pathology of neuropsychiatric (schizophrenia, bipolar disorder) 7-10 and neurodegenerative disorders 11 .
SHANK2 encodes for a postsynaptic scaffolding protein at glutamatergic synapses in the brain, essential for proper synapse formation, development and plasticity 12,13 . As organizer of a large protein complex, the so-called postsynaptic density, it connects the different types of glutamate receptors (α-amino-3-hydroxy-5-methyl-4isoxazolepropionic acid receptor (AMPA), N-methyl-d-aspartat receptor (NMDA), metabotropic glutamate receptors) to molecules of many signaling pathways and to the actin cytoskeleton 14 . Impaired insulin signaling in the brain was postulated to contribute to the development of autism in genetically susceptible individuals 15 . As SHANK2 directly interacts with IRSp53 (insulin receptor substrate p53) 16,17 , it may be involved in insulin signaling in the brain, making this pathway susceptible to SHANK2 mutations. In addition, studies in mice and human iPSC-derived neurons provided first evidence that a loss of SHANK could be partially rescued by administration of insulin like growth factor 1 [18][19][20] . Moreover, SHANK dysregulation may also contribute to the molecular pathology of Alzheimer´s disease (AD) 21 . In synaptosomes isolated from the middle frontal gyrus from patients with AD increased SHANK2 levels were reported 11 . Mutations in the gene coding for amyloid precursor protein (APP) have been identified in several individuals with AD 22,23 . Both SHANK2 and APP are expressed during neuronal differentiation and are involved in synapse formation and neural plasticity [23][24][25] , and thus could potentially interact in the course of ASD and neurodegenerative disorders.

Results
Genome editing of SH-SY5Y cells to generate SHANK2 mutant lines. To investigate the functional consequences of SHANK2 mutations that have been identified in patients with ASD and intellectual disability, we introduced SHANK2 frameshift mutations in SH-SY5Y cells using CRISPR/Cas9-based genome editing. A guide RNA was designed to target the genomic region adjacent to the position of the SHANK2-R841X variant, which had previously been identified in a patient with ASD 4 . The non-homologous end-joining repair mechanism of the cells led to the introduction of a heterozygous adenosine insertion, which caused a frameshift and premature stop codon (6H6, p.L837Sfs*49). Additionally, we generated a compound heterozygous mutation line (6A4) with an adenosine deletion on one allele and an adenosine insertion on the other, resulting in two different premature stop codons (p.L837Sfs*49, p.L837Wfs*14) ( Table 1). As a control, we selected cells that underwent the complete process of genome-editing without being edited (1F11) as well as the original SH-SY5Y line (WT). Mutations were confirmed after several passages ( Supplementary Fig. 1) and putative predicted offtarget sites were ruled out by Sanger sequencing of the respective genomic regions (Supplementary Table 1).
Western blot experiments revealed a complete loss of SHANK2 wildtype expression for the compound heterozygous cell line and a reduction of SHANK2 wildtype protein by one half for the heterozygous SHANK2 mutant lines (Fig. 1, Supplementary Fig. 2). The SHANK2 antibody targets amino acids 849-1029, a region downstream of the introduced frameshift mutation (6A4: p.L837Sfs*49, p.L837Wfs*14, 6H6:p.L837Sfs*49), therefore the reduction in protein levels reflects the diminished amount of wildtype SHANK2 protein, whereas the presence of a truncated version of the protein cannot be excluded. Treatment with cycloheximide, which blocks nonsensemediated mRNA decay (NMD), did not reveal any evidence for NMD in both cell lines ( Supplementary Fig. 3) even though quantitative real-time PCR (qPCR) analysis of SHANK2 mRNA expression revealed significantly reduced levels in the cell line with compound heterozygous SHANK2 mutations ( Supplementary Fig. 4A).
Cells with bi-allelic SHANK2 mutations show increased proliferation and lower apoptosis rates compared to controls during neuronal differentiation. To assess basic properties of the SHANK2mutant cell lines, we analyzed cell proliferation at the undifferentiated stage. Immunofluorescence microscopy with Ki67, a cell proliferation marker, showed similar proliferation rates for all lines. Apoptosis was analyzed with the TUNEL assay, which revealed low apoptosis rates under basic conditions without any differences between the cell lines (Fig. 2).
Next, we differentiated the SH-SY5Y lines according to an improved neuronal differentiation protocol published by Chiocchetti et al. 32 . With this protocol, SH-SY5Y cells can be differentiated into neurons that express genes relevant in early cortical development and associated to neurodevelopmental/-psychiatric disorders 32 .
To validate the process of neuronal differentiation, we quantified the expression of selected marker proteins by Western blot analysis in the control cell line (WT) under basic culturing conditions and after 7 days of differentiation. A significant increase of the neuronal marker β3-Tubulin (TUBB3) together with decreased levels of the neuronal stem cell protein Nestin (NES) after 7 days of differentiation indicate that the SH-SY5Y cells are starting to differentiate into neurons (Supplementary Figs. 5 and 6). This finding is also supported by an increased expression of the presynaptic marker protein Synaptophysin (SYP) and the postsynaptic density protein 95 (PSD-95) (Supplementary Figs. 5 and 6). Expression analysis of AKT Serine/Threonine Kinase 1 (AKT) was also included, as it is an important mediator of growth factor-induced neuronal survival in the developing nervous system. No significant expression differences could be found for AKT, SHANK2 and the ionotropic glutamate receptor NMDA type subunit 2B (GRIN2B) (Supplementary Figs. 5 and 6). As we could detect elevated protein levels of glutamatergic (SYP, PSD-95) and pan-neuronal (TUBB3) proteins in parallel to the reduction of the neuronal stem cell marker Nestin after 7 days of differentiation, we selected this stage for further analysis to investigate impairments in early neuronal differentiation. We analyzed cell proliferation after 7 days of differentiation when wildtype cells still proliferate with low apoptosis rates. The cell line 6A4 with bi-allelic SHANK2 mutation showed significantly more Ki67-positive cells, indicating a higher rate of cell proliferation compared to controls, whereas the heterozygous line 6H6 behaved similarly to the controls (Fig. 2). However, reduced numbers of TUNEL-positive cells were identified for both cell lines with SHANK2 mutations (6A4 and 6H6), indicating a lower apoptosis rate compared to controls (Fig. 2).
Bi-allelic SHANK2 mutation impairs neurite outgrowth. To identify morphological impairments in early neuronal differentiation processes, we analyzed neurite outgrowth after 4 days of differentiation in the different cell lines by measuring the length of β3-Tubulin-positive cellular extensions in FiJi 36 . The control cells (WT, 1F11) extended long neurites and started to form neuronal networks (Fig. 3). The heterozygous SHANK2 mutant line 6H6 developed similarly to the controls, whereas the compound heterozygous SHANK2 mutant line www.nature.com/scientificreports/ (6A4) showed an obvious impairment regarding network formation and neurite extension. The 6A4 cell line had significantly shorter neurites compared to controls and 6H6 (Fig. 3).

Changes of gene expression in SHANK2 mutant cells. To further elucidate the influence of SHANK2
mutations on basic cellular and neuronal differentiation properties, we analyzed mRNA expression levels of several genes involved in these processes by qPCR (n = 3 experiments) ( Supplementary Fig. 4). The bi-allelic mutation of SHANK2 in SH-SY5Y cells initially led to strongly reduced SHANK2 mRNA levels compared to the control in the undifferentiated state. After 7 days of differentiation, however, SHANK2 mRNA levels were equal between the cell lines and after 16 days of differentiation even an increase of SHANK2 mRNA expression was found for the compound heterozygous cell line ( Supplementary Fig. 4). We also analyzed the gene expression of microtubule-associated protein 2 (MAP2), the presynaptic gene synaptophysin (SYP) and the postsynaptic www.nature.com/scientificreports/ gene discs large MAGUK scaffold protein 4 (DLG4, encoding for PSD-95) to determine if neuronal, presynaptic and postsynaptic genes are differentially expressed. Furthermore, the expression of all SHANK family members (SHANK1, SHANK2, SHANK3) was analyzed as well as several glutamate receptor genes, as SHANKs are known to interact with glutamate receptors: AMPA receptor subunit 2 (GRIA2), metabotropic glutamate receptor 1 (GRM1) and the ionotropic glutamate receptor kainate type 2 (GRIK2). To find out if GABAergic neurons or aminergic neurons are affected we included GABA receptor beta 3 (GABRB3) and thyroxin hydroxylase (TH) expression analysis. The neuronal stem cell gene Nestin (NES) has been included to determine the presence of very immature neuronal cells.
In undifferentiated cells, a significantly increased expression of NES and a nominally significant reduction in the expression of both SHANK1 and SHANK2 was observed in the compound heterozygous line 6A4 (Supplementary Fig. 4A). Furthermore, expression of GRIA2 and GABRB3 were reduced. A gene predominantly expressed in aminergic neurons, tyrosine hydroxylase, was expressed at low levels with slightly elevated levels in the heterozygous cell line 6H6. Nestin levels remained robustly elevated in the cell line with a bi-allelic SHANK2 mutation (6A4) from the undifferentiated state throughout neuronal differentiation after 7 and 16 days (Supplementary Fig. 4A-C). The results reached statistical significance after correction for multiple testing. Additionally, we found reduced levels of tyrosine hydroxylase in the compound heterozygous cell line (6A4) after 7 and 16 days of differentiation ( Supplementary Fig. 4B,C).
After 16 days of differentiation, increased expression levels of the presynaptic gene SYP, all three members of the SHANK family as well as GRIA2 and GABRB3 were identified in the compound heterozygous SHANK2 mutant line ( Supplementary Fig. 4). No mRNA expression differences in SHANK2 mutant lines were detected, neither for DLG4, GRM1, GRIK2, nor MAP2 ( Supplementary Fig. 4C).

Changes of protein expression in SHANK2 mutant cells.
Next, we assessed the expression of several proteins by Western blot to elucidate the consequences of the loss of SHANK2 regarding neuronal differentiation. SHANK2 expression was reduced in the undifferentiated state and after 7 days of differentiation in both SHANK2 mutant lines (6A4, 6H6) compared to the control 1F11 ( Fig. 1). We included the non-edited wildtype SH-SY5Y cells (WT) in the Western blot analysis as a second control cell line, and compared expression levels to 1F11. The comparison between the two control lines revealed similar expression levels for WT and 1F11 for all analyzed proteins (Fig. 4). A comparable expression of all tested genes was also shown on mRNA level (Supplementary Fig. 7).
To elucidate the influence of SHANK2 mutations on differentiated cells, we compared protein expression levels between the control 1F11 and the two mutant cell lines after 7 days of differentiation. The heterozygous SHANK2 mutant line 6H6 showed significantly reduced amounts of the pre-and postsynaptic proteins SYP and PSD-95, whereas other proteins were not affected (Fig. 4, Supplementary Fig. 8). The compound heterozygous SHANK2 mutant line 6A4 had highly increased Nestin levels and significantly reduced levels of SYP and PSD-95, indicating impaired neuronal differentiation (Fig. 4, Supplementary Fig. 8). No differences in protein expression could be detected for β3-Tubulin, GRIN2B and AKT (Fig. 4, Supplementary Fig. 8). With β3-Tubulin immunocytofluorescence microscopy staining, we detected long dendritic protrusions starting to form networks after 7 days of differentiation in the controls (WT, 1F11) and the heterozygous SHANK2-mutant cell line 6H6 (Fig. 5a). The compound heterozygous SHANK2 cell line 6A4 lost this ability, showing less and shorter dendritic protrusions (Fig. 5a). As Nestin expression showed the most prominent change on transcriptome level in this cell line, we further investigated its expression by immunofluorescence microscopy and performed a co-staining of Nestin and β3-Tubulin during neuronal differentiation (Fig. 5b). We compared cell line 6A4 to the control 1F11 in the undifferentiated stage and after 7 and 16 days of differentiation. Both cell lines started to differentiate into neurons and Nestin-positive cells disappeared in the control cell line 1F11, whereas Nestin-positive cells could still be detected throughout differentiation in cell line 6A4. The Nestin-positive cells showed a different morphology with less cellular protrusions compared to β3-Tubulin-positive cells, even at an advanced stage of differentiation (Fig. 5b).
SHANK2 mutations affect tyrosine kinase receptor downstream signaling pathways. It has been hypothesized that insulin signaling contributes to the development of autism in genetically susceptible individuals 15 . As SHANK2 (as well as SHANK1 and SHANK3) directly interacts with IRSp53 (insulin receptor substrate p53), it may also play a role in insulin signaling in the brain 17,37 . The serine/threonine kinase AKT acts downstream of the tyrosine kinase receptor signaling pathways. Therefore, we analyzed the phosphorylation levels of AKT in relation to overall AKT levels by Western blot analysis to get an insight into a putative impairment of insulin or, more generally, into tyrosine kinase receptor downstream signaling during early neuronal differentiation. Interestingly, the cells with bi-allelic SHANK2 mutation showed an increase of phosphorylated AKT, (Fig. 6a, Supplementary Fig. 9), suggesting that a loss of SHANK2 impairs downstream signaling of tyrosine kinase receptors such as insulin receptors during early neuronal differentiation.
Bi-allelic SHANK2 mutations influence amyloid precursor protein expression. APP regulates both synapse formation and neuronal plasticity and is expressed during neuronal differentiation [23][24][25] . Mutations in APP have been linked to AD 22,23 , in which SHANK expression was also found to be altered 11 . SHANK2 and APP function might be linked, as they are both implicated in synaptic signal transduction, and might interact during the course of neurodegenerative disorders. Therefore, we asked if SHANK2 mutations can influence APP expression. The quantification of APP protein expression by Western blot analysis revealed a significant decrease of APP levels in the compound heterozygous SHANK2 mutant line, whereas cells with heterozygous SHANK2 mutation showed no significant difference compared to controls (Fig. 6a, Supplementary Fig. 9).

Discussion
In this study, the molecular and cellular consequences of mono-allelic and bi-allelic SHANK2 frameshift mutations were investigated in the neuroblastoma cell line SH-SY5Y. The mutations were introduced by CRISPR/Cas9 genome editing in exon 23-encoding for part of the proline-rich region-in proximity to a previously identified nonsense mutation from a patient with ASD (R841X) 4 . Deleterious de novo SHANK2 mutations have always been identified in a heterozygous state in patients with ASD, intellectual disability, attention deficit and hyperactivity disorder as well as language impairment, with one allele unaffected 2-5 . Therefore, a heterozygous SHANK2 mutant line is the appropriate model to identify impairments that contribute to the pathology in patients. Nevertheless, the compound heterozygous line is useful to explore SHANK2 function in human neurons. Our data showed that both mono-and bi-allelic mutations of SHANK2 led to reduced wildtype SHANK2 protein levels in SH-SY5Y cells. The bi-allelic mutation of SHANK2 in SH-SY5Y cells initially resulted in strongly reduced SHANK2 mRNA levels compared to the control in the undifferentiated state which is in line with reduced protein levels. After 7 days of differentiation SHANK2 mRNA levels were equal between the cell lines even though protein levels were strongly reduced and after 16 days of differentiation even an increase of SHANK2 mRNA expression was found for the compound heterozygous cell line ( Supplementary Fig. 4). This might be caused by a compensatory mechanism taking place during neuronal differentiation leading to an upregulation of gene transcription because the protein is missing. The treatment of the cells during differentiation with retinoic acid and BDNF might also stimulate SHANK2 transcription. www.nature.com/scientificreports/ The parallel analysis of both the compound heterozygous and the heterozygous lines revealed shared and distinct alterations in cell function. Similar consequences in both cell lines were identified concerning lower apoptosis rates, and reduced pre-and postsynaptic protein levels.
At the beginning of neuronal differentiation, cells have to switch from proliferation to differentiation 38 . Both SHANK2 mutant lines had lower apoptosis rates after 7 days of differentiation, and the compound heterozygous line also showed higher proliferation compared to controls. This may point to a role of SHANK2 regarding the switch from proliferation to differentiation during the early neuronal differentiation process in neuroblastoma cells. Both cell lines showed diminished pre-(SYP) and postsynaptic (PSD95) protein expression after 7 days of differentiation. SHANK2 is known to interact with PSD95 through the guanylate kinase-associated protein (GKAP) to anchor the AMPA and NMDA receptors at the postsynaptic density 39 . Glutamatergic markers were affected by SHANK2 mutations with reduced PSD95 and GRIA2 expression. So far, the role of SHANK2 has been mainly investigated in mature synapses 30 , whereas our results suggest that it may already have important functions during early neuronal differentiation, even prior to synapse formation. Interestingly, despite being a predominantly postsynaptic protein, the SHANK2 mutations also affect presynaptic protein expression, as shown by diminished SYP expression.
In the cell line with bi-allelic SHANK2 mutations expression levels of other autism-related genes and proteins (according to the SFARI Gene database 6 ) were affected, such as SHANK1, SHANK3, GABRB3, GRIA2 and PSD95. This may point to synergistic effects and also emphasizes the importance of these genes and proteins-as well as the signaling cascades they belong to-in the pathology of autism. Cells with bi-allelic SHANK2 mutations remained in a neuronal precursor cell state, as shown by high Nestin levels and impaired neurite outgrowth. They lost their ability to extend long neurite protrusions and to form networks, which is in line with diminished levels of the pre-and postsynaptic proteins SYP and PSD95. All these observations indicate a severe impairment of neuronal differentiation. The strongest effects of Shank2 deletions were only seen in the cell line with bi-allelic SHANK2 mutation, which compares to the rat and mouse models that show the strongest phenotype only when both Shank2 alleles are deleted [27][28][29] . In humans bi-allelic mutations have not yet been described and may lead to embryonic death. Human individuals with heterozygous deleterious de novo SHANK2 mutations show a general developmental delay with delayed development of speech as well as learning impairments and mild to severe intellectual disability 4,5 . This may be a long term consequence of impaired neuronal differentiation which cannot be fully compensated. In the homozygous Shank2 knockout mice, cognitive impairments as well as a disturbed synaptic signal transmission have been observed 28,29 . The very early neuronal development in utero or during the first postnatal days has not been studied in these animals, therefore we do not know if an impaired neuronal differentiation may precede impaired synaptic signaling.
In a recent publication, the SHANK2-R841X mutation was investigated in human iPSC-derived cortical neurons 31 . The authors also showed that SHANK2-deficiency had an impact on dendritic growth, resulting in increased dendritic length and complexity, and led to a perturbation of neurodevelopmental gene set expression 31 . This is in line with some of our findings, even though the data between the two studies cannot be directly compared. The work on human cortical neurons (HCN2) by Luo et al. 40 showed that reduced SHANK2 expression decreased neurite numbers and length in human neurons, which is comparable to our findings of reduced neurite outgrowth in the compound heterozygous SHANK2 mutant line 40 . They also report an impact on proliferation and apoptosis, but in their study, low SHANK2 expression inhibited proliferation and promoted apoptosis, whereas increased proliferation and decreased apoptosis was found in our study. The differences between our work and those two studies can be explained by two facts: (a) the cell models used are notably different, despite the fact that all models are human cell lines; (b) the focus of our research was on the onset of neuronal differentiation, whereas Zaslavsky et al. 31 and Luo et al. 40 used differentiated cortical neurons.
The severe impairment of neuronal differentiation might be partly explained by an increase of tyrosine kinase receptor signal activity. Interestingly, 6A4 cells showed elevated p-AKT levels, indicating more active tyrosine kinase receptor downstream signaling. During the process of neuronal differentiation, cells were constantly treated with retinoic acid (RA) and brain-derived neurotropic factor (BDNF). RA is used to activate the transcription of gene sets to induce differentiation and inhibit cell proliferation 38 . BDNF binds to tyrosine receptor kinase beta and activates several signaling cascades including the phosphatidyl-inositol-3-kinase (PI3K)-AKT pathway, which, in turn, inhibits apoptosis, regulates protein synthesis and remodels the cytoskeleton 41 . In parallel, other growth factors like insulin are present in the culture medium, which activate further tyrosine kinase receptors like the insulin receptor. At synapses, the insulin receptor was found co-expressed with insulin receptor substrate p53 (IRSp53), which is phosphorylated upon stimulation with insulin and acts as key factor in cytoskeleton reorganization, hence mediating neurite outgrowth 42 . IRSp53 is a direct interaction partner of the SHANK proteins 16 , linking SHANK2 to insulin signaling. As we identified increased p-AKT levels in the 6A4 cell line, which occurred together with reduced apoptosis rates and increased proliferation, we concluded that bi-allelic SHANK2 mutations might affect BDNF and/or insulin downstream signaling during early neuronal differentiation (Fig. 6b, Table 2). This is in line with other findings that describe the dysregulation of AKT signaling, which is part of the PI3K-AKT-mTor cascade downstream of tyrosine kinase receptor signaling, as a cause of ASD and other neurodevelopmental disorders 43,44 .
In addition, we could also show that SHANK2 mutations have an impact on APP expression, a protein extensively studied in the context of AD, although its exact biological function is not yet well understood 23 . The reduced APP expression in cells with bi-allelic SHANK2 mutations may also contribute to the diminished neurite outgrowth and some of the neuronal differentiation impairments. These results point to a functional link between APP and SHANK2 during neuronal development. Low APP expression occurred together with high p-AKT levels in the cell line 6A4. APP expression is known to be regulated by growth factors like insulin and BDNF which activate PI3K and AKT. AKT further phosphorylates glycogen synthase kinase 3 (GSK3) which acts as an inhibitor and results in less translation of APP protein 45 (Fig. 6b). www.nature.com/scientificreports/ There are functional aspects of APP which make it interesting to look at in neurodevelopmental disorders: APP regulates synapse formation, neurite outgrowth and neuronal plasticity. Furthermore, it is expressed during early neuronal differentiation at sites where SHANK2 is also expressed [23][24][25]46 . In addition, there is increasing evidence showing that molecular and cellular processes involved in neurodevelopment are also relevant to the pathology of neurodegenerative diseases 47,48 . One connection could be the pathological reactivation during neurodegeneration of genetic programs physiologically relevant during neurodevelopment, especially in Alzheimer's disease 49 . There is also a genetic overlap between intellectual disability and cognitive aging 48 . Since SHANK2 is involved in fundamental molecular and cellular processes in neurons, it is plausible that its role might also be relevant in neurodegeneration.
The results of our study should be considered with prudence, as they were obtained in vitro in SH-SY5Y cells. As these neuroblastoma cells were tumor-derived, it is not possible to investigate all aspects of neuronal properties. The intrinsic biological variation of the neuroblastoma cell population is likely to mask low effects resulting from SHANK2 mutations, especially in the heterozygous state. Our marker gene expression analysis revealed significant expression differences, mostly in the compound heterozygous SHANK2 mutant line. The cell line with a heterozygous SHANK2 mutation may still show small changes that could not be identified due to the given limitations of our experimental setup. Despite the fact that three independent experiments were carried out, this may not allow the detection of alterations with very low effect sizes.
p-AKT levels were investigated as a general indicator of tyrosine receptor downstream signaling. Additional work is necessary to further specify which tyrosine kinase receptor signaling pathway is mostly influenced (e.g. BDNF, insulin, IGF1), which downstream signaling path of PI3K-AKT is mainly altered (e.g. mTOR, GSK3beta, ERK), and how this affects both the transcriptome and proteome.
In summary, we have shown that mono-and bi-allelic SHANK2 frameshift mutations resulted in an impairment of early neuronal differentiation in SH-SY5Y cells, characterized by changes in cell growth properties and reduced pre-and postsynaptic protein expression. We also provided the first link between SHANK2 mutations and impaired receptor tyrosine kinase downstream signaling, and revealed an impact of SHANK2 mutations on APP expression. Further investigation is warranted, preferably in human iPSC-derived neuronal cells from patients with SHANK2 mutations, to follow-up these novel insights into SHANK2-dependent pathways.

Methods
Cell culture. Human neuroblastoma cells (SH-SY5Y) were obtained from the Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures. The cells were grown on 75 cm flasks in Dulbecco's modified Eagle medium (DMEM, Thermo Fisher Scientific), supplemented with 15% fetal calf serum, 1% non-essential amino acids and 1% Penicillin-streptomycin at 37 °C in a humidified environment with 5% CO 2 . Cells were split at 80-90% confluency and 2 × 10 5 cells were plated per well on a 6-well cell culture plate for RNA and protein isolation. Undifferentiated cells were harvested 3 days after seeding. Cells were differentiated according to the protocol published by Chiocchetti et al. 32 and harvested after 7 or 16 days of differentiation, with medium changes 3 times per week. For neurite tracing and immunocytochemistry, 7.5 × 10 4 cells were seeded per well on a 24-well plate on glass coverslips coated with collagen. The cells were tested against mycoplasma contamination with the Venor GeM Classic mycoplasma detection kit for conventional PCR (Minerva Biolabs) and all used cell lines were free of mycoplasma. Genome edited cells were used for experiments with passages ranging from 15 to 20.
Genome editing. The SHANK2-gRNA was cloned with two primers: SHANK2_G2FCAC CGT CGA GGG ATG CCC AGA AAC G and SHANK2_G2R AAA CCG TTT CTG GGC ATC CCT CGA C into the pSpCas9(BB)-2A-GFP vector (Addgene, 48138) using the restriction site BBsI. The plasmid was then introduced into the SH-SY5Y cells using lipofectamine (Invitrogen). Four days after transfection, GFP-positive living cells were sorted by fluorescence-activated cell sorting (FACS) as single cells into 96 well plates. The cell sorter used was a BD Figure 6. Analysis of the influence of SHANK2 mutations on AKT serine/threonine kinase (AKT) and amyloid precursor protein (APP) expression (7 days of differentiation) (a) and Overview of the putative influence of SHANK2 mutations on tyrosine kinase receptor downstream signaling (b). (a) Western blot analysis was carried out to identify p-AKT levels, as an indicator for tyrosine kinase receptor signaling pathway activity, and APP expression levels. P-AKT and AKT levels were normalized against GAPDH and the ratio between the normalized p-AKT/AKT expression calculated. APP expression was normalized against GAPDH. The mean for both controls (WT and 1F11) was taken for comparison with the SHANK2 mutant lines. The cell line with compound heterozygous SHANK2 mutation (6A4) showed impaired p-AKT levels and APP expression levels. Cropped western blot membrane images are shown and full-length blots are presented in Supplementary Fig. 9. n = 5 experiments, one-way ANOVA (ANOVA results: p-AKT F = 16.65, P = 0.0003; APP F = 25.83, P < 0.0001). Each mutant line was compared against the mean of both controls. ***P ≤ 0.001, correction for multiple testing with Dunnett's test. Data were presented as box-plots. (b) Scheme illustrating the potential regulation of AKT signaling in the differentiating SH-SY5Y cells downstream of BDNF/TRKB or insulin/insulin receptor signaling. Increased p-AKT levels probably led to reduced apoptosis, increased proliferation and reduced amounts of amyloid precursor protein (APP). Further investigation is needed regarding the ERK and mTOR signaling downstream of AKT, to elucidate an effect from SHANK2 mutations. These pathways have been already linked to ASD pathology 61 . The part of the signaling cascade which was analyzed in this study is illustrated in green. In this simplified scheme not all different factors which influence AKT and APP expression can be considered. BDNF brain derived neurotropic factor, TRKB tyrosine receptor kinase beta, ERK extracellular-signal regulated kinase, mTOR mammalian target of rapamycin. Nonsense mediated mRNA decay. 2 × 10 5 cells were seeded per well on a 6-well plate and cultured for 24 h. Then, culturing medium was changed to either mock treatment (2 ml medium + 50 µl water) or cycloheximide (CHX) treatment (100 µg/ml, CHX was dissolved in water) and cultured for 6 h. Then, cells were harvested, RNA was isolated and SHANK2 expression quantified with the primers SHANK2_F and SHANK2_R (details in Supplementary For the analysis of neurite length, cells were fixed after 4 days of differentiation. Images were taken in the channel corresponding to the β3-Tubulin signal and neurites were traced using the "Simple neurite tracer" tool of the FiJi software 36 to determine their length. Three independent experiments were carried out and cells from six images were analyzed for each cell line per experiment. Ki67 and TUNEL signals were quantified by using the FiJi 36 software and by normalizing the number of signals to the total number of nuclei (as determined by Hoechst 33342 staining). Three independent experiments were carried out and cells of six images were analyzed for each cell line per experiment.
Protein expression was normalized to the amount of GAPDH and the values obtained for the reference line 1F11 were set to 100. P-AKT and AKT levels were first normalized against GAPDH and then the ratio of the normalized data P-AKT/AKT was calculated. The normalized values were presented in box-plots.  Immunofluorescence images in Fig. 2 were taken with the Leica DMI4000B microscope with the camera Leica-DFC3000G-0001062113 and the objective N PLAN L 20×/0.35 DRY using the Leica Application Suite X software in the red, green and blue channel. The acquisition was performed with the time and space resolution XY = 0.959 µm and Z = 4.490 µm, 12 bit images were taken with a resolution of 1296 × 966 pixels. All pictures were taken with the same settings. In the illustrated TUNEL and Ki67 pictures brightness was increased by 30% for all conditions in the Power Point software. Images in Fig. 3 were taken with the same microscope and the same settings only for the green channel, brightness was increased by 20% for all conditions in Power Point software. Images in Fig. 5 were taken with the same microscope and camera and the objective HC PL APO 40×/0.85 DRY. The acquisition was performed with the time and space resolution XY = 0.395 µm and Z = 0.761 µm, 12 bit images were taken with a resolution of 1296 × 966 pixel. All pictures were taken with the same settings. Colour saturation was increased to 200% for all conditions in Power Point software.

Equipment and settings. Westernblot images in
Statistical analysis. Prism 8 (GraphPad Software) software was used for data analysis. The Shapiro Wilk test was performed to test for normal distribution when sample sizes were ≥ 5. In the TUNEL assay, Ki67 quantification, analysis of neurite length and Western blot experiments, a one-way ANOVA test was used for the comparison of each SHANK2 mutant line (6A4, 6H6) against the reference 1F11. In addition, the WT line was compared to 1F11. A P-value of ≤ 0.05 was considered nominally significant. The Dunnett's test was used to correct for multiple testing (n = 3 tests). A two-tailed unpaired Student´s T-Test was performed for the pairwise comparison of protein expression in undifferentiated versus the 7 days differentiated WT cell line.
Two-way ANOVA was performed for the comparison of RNA expression levels between the reference 1F11 and the two mutant lines (6A4, 6H6), with gene and cell line as the two factors; the Dunnett's test was used to correct for multiple testing for each time point.

Data availability
The datasets supporting the conclusions of this article are included within the article and its additional files. www.nature.com/scientificreports/