All animal procedures were conducted in accordance with European, national and institutional guidelines and protocols, and were approved by the responsible government authority: Landesamt für Natur, Umwelt und Verbraucherschutz NRW (Animal Licence: LANUV NRW under the reference number 84-02.04.2014.A 126). Homozygous Ncald^KO/KO and heterozygous Ncald^KO/WT [B6N(Cg)-Ncald^{tm1.1(KOMP)Vlcg}/J, Stock Number: 018575] animals were acquired from Jackson Laboratory. Animals used for all experiments were provided food and water ad libitum and were caged in small groups on a 12 h light/ dark cycle. These animals were genotyped using following primers: mmu NcaldWTfw: 5′-TTTCCCTTACGGGGATGCT-3′; mmu NcaldWTrev: 5′-AGCATTTCTGCCTTGCTGAT-3′; mmu NcaldKOfw: 5′-CGGTCGCTACCATTAC-3′; mmu NcaldKOrev: 5′-GCATGTGTGACAACAG-3′.

Tissues were lysed in ice cold RIPA buffer (Sigma) together with protease (Complete Mini, Roche) and phosphatase inhibitors (Thermo scientific, Waltham, MA, USA). For further analysis, the following primary antibodies were used; anti-NCALD (1:1,000, 12925-1-AP, Proteintech), anti-beta-actin (1:2,500, A5316, Sigma), anti-GAPDH (1:5,000 G-9295, Sigma), anti-myelin basic protein (anti-MBP; 1:1,000 SMI94, Covance), anti- MAP3K10 (1:500, NBP1-87737, Novus Biologicals), anti-pJNK (1:500, sc-6254, Santa Cruz), anti-JNK (1:1,000, #9252, Cell signaling). Chemiluminescence signal was detected with HRP conjugated-secondary antibodies and Chemiluminescence reagent (Thermo Scientific, Waltham, MA, USA) according to manufacturer’s protocol.

A freezing microtome was used for cutting 40 μm thick consecutive horizontal brain sections. 0.2% gelatin solution in 250 mM Tris-HCl was used for mounting the sections and left overnight at 40°C heating plate for drying. For further staining the sections were first rehydrated for 1 min in water and then stained in 0.1% cresyl violet solution for approximately 8 min. Following this, the sections were washed three times in water (2 min each) and an ascending ethanol series (50%, 70%, 80%, 90%) was used for dehydration. Sections were finally destained with 96% ethanol and 0.5% acetic acid solution and washed twice in 100% ethanol (2 min each). Subsequently, the sections were incubated in xylene for at least 2 min or until they were mounted using Entellan.

Mice were sacrificed at P14, P30 or 4 months by ketamine/ xylazine overdose followed by transcardial perfusion with saline solution (0.85% NaCl, 0.025% KCl, 0.02% NaHCO₃, pH 6.9, 0.01% heparin, body temperature) and ice cold 4% paraformaldehyde (PFA) freshly depolymerized in 1×phosphate-buffered saline (PBS), pH 7.4. The fixed brains were carefully isolated from the skull and were further stored overnight in the same ice-cold 4% PFA solution as used for transcardial perfusion. For further storage and cryoprotection, the brains were transferred to a mixture of 20% glycerol and 2% dimethylsulfoxide in 0.1 M phosphate buffer. Consecutive horizontal sections (40 μm) were collected in six series using a freezing microtome. Corresponding brain sections from wildtype (WT) and Ncald^KO/KO littermates (gender matched) were stained simultaneously for further immunohistochemical analysis as previously described (Kononenko et al., 2017). The following antibodies were used: anti-NCALD (1:100, 12925-1-AP, Proteintech), anti-NeuN (1:500, EPR 12763, Abcam), anti-TBR1 (1:500, ab31940), anti-CUX1 (1:200 sc-13024, Santa Cruz), anti-glial fibrillary acidic protein (anti-GFAP; 1:500, G3893, Sigma), anti-Ki-67 (1:500, ab15580 Abcam), anti-DCX (1:500, AB2253, Merck), anti-adenomatous polyposis coli (anti-APC; 1:500, OP80, Merck), anti-MBP (1:1,000 SMI94, Covance).

Hippocampi were isolated from P1–P5 postnatal mice and the neurons were cultured as described previously (Kononenko et al., 2013). Calcium phosphate transfection procedure was used to transfect the cultured hippocampal neurons at DIV 6–8 with eGFP plasmid as previously described (Threadgill et al., 1997). Neurons were fixed using 4% PFA on DIV 14–15 and were additionally immunostained with anti-GFP (1:5,000, ab13970, Abcam) antibody. Neurons were then imaged with AxioImager M2 fluorescence microscope (Zeiss) appended with ApoTome.2 which allowed virtual confocality. Single cells captured with soma in the center of the image were then subjected to Sholl analysis using ImageJ Sholl Analysis Macro (Gensel et al., 2010). Dendritic branching was determined by adding the total number of intersections within 220 μm from the cell body.

E13.5 mouse embryos were used for dissecting spinal cords (Hosseinibarkooie et al., 2016). Trypsin (Worthington) and DNAse (Sigma) mixture was used for isolating neurons. These singularized neurons were sieved and plated on poly-D-lysine/laminin (Sigma) coated coverslips. Neurobasal medium along with B27 supplement, 2 mM L-glutamine, 1× pen-strep (Invitrogen, Carlsbad, CA, USA) containing 50 ng/μl BDNF, 50 ng/μl GCNF and 50 ng/μl CNTF (Peprotech) was used as culture medium at 37°C in a humidified incubator with 5% CO₂ add space after 5%). Along with axonal length (longest neurite), we have also analyzed primary branching and secondary branching (branches from the longest neurite were considered primary whereas branches from the primary branches were considered secondary) of these MNs.

On DIV 14 the neurons on coverslips were fixed using 4% PFA in PBS at RT for 15 min and washed three times with 1×PBS at room temperature. After blocking with PBS containing 5% normal goat serum (NGS) and 0.3% Saponin for 1 h, neurons were incubated for 1 h with following primary antibodies anti-NCALD (1:600, 12925-1-AP), anti-VGLUT1 (1:300, 131004, Synaptic Systems), anti-VGAT (1:300, 135011, Synaptic Systems), anti-TAU (1:800, sc-390476, Santa Cruz) and anti-Choline Acetyltransferase (anti-CHAT; 1:500, AB144P, Millipore, Burlington MA, USA) diluted in blocking solution. Coverslips with neurons were then rinsed three times with PBS and incubated for 30 min with corresponding secondary antibodies (diluted 1:1,000 in PBS containing 0.3% Saponin and 5% NGS). Subsequently, coverslips with neurons were washed three times in 1×PBS and mounted using Immumount. A random stretch of neurites with certain observable puncta with co-localization (yellow) signal was chosen. A line was drawn on this stretch and ImageJ plot profile function was used for each channel individually to calculate the intensity plot along the line. Intensity values were plotted against the XY value on the line for each channel in GraphPad Prism 6 software. Subsequently, the plots for NCALD and each synaptic marker were superimposed. The asterisks represent overlapping peaks of each channel, thus showing the co-localization.

Confocal imaging was performed using a commercial Leica SP8 TCS microscope (Leica Microsystems) equipped with four laser lines 405 nm, 488 nm, 552 nm and 638 nm. Samples within each independent experiment were acquired with equal settings. Images were acquired with an HC PL APO 20×/0.75 CS2 or HC PL APO 63×/1.40 CS2 (oil) objectives (Leica Microsystems), a scanning format of 1,024 × 1,024, eight-bit sampling, and 1 zoom, yielding a pixel dimension of 567.62 × 567.62 or 90.09 nm × 90.09 nm in the x and y dimensions, respectively.

The brain and spinal cord samples were collected at P30 and P14, respectively (From both WT and Ncald^KO/KO mice). The tissue samples were homogenized and lysed in NP40-based lysis buffer (50 mM Tris-HCl, 1% NP40, 100 mM NaCl, 2 mM MgCl2 including protease inhibitor). Ten microliter of Control rabbit IgG (SantaCruz) and NCALD polyclonal antibody were used for immunoprecipitation using protein A paramagnetic MicroBeads (Miltenyi) following the manufacturer’s instruction. Finally, the IP columns were washed at least six times with lysis buffer. The bound fraction of proteins was directly used for further mass spectrometry analysis.

All samples were analyzed on a Q-Exactive Plus (Thermo Scientific, Waltham, MA, USA) mass spectrometer coupled to an EASY nLC 1,000 UPLC (Thermo Scientific, Waltham, MA, USA). Peptides were loaded with solvent A (0.1% formic acid in water) onto an in-house packed analytical column (50 cm × 75 μm I.D., filled with 2.7 μm Poroshell EC120 C18, Agilent). Peptides were chromatographically separated at a constant flow rate of 250 nL/min using the following gradient: 5–30% solvent B (0.1% formic acid in 80% acetonitrile) within 66 min, 30–50% solvent B within 13 min, followed by washing and column equilibration. The mass spectrometer was operated in data-dependent acquisition mode. The MS1 survey scan was acquired from 300 to 1,750 m/z at a resolution of 70,000. The top 10 most abundant peptides were isolated within a 1.8 Th window and subjected to HCD fragmentation at a normalized collision energy of 27%. The AGC target was set to 5e5 charges, allowing a maximum injection time of 120 ms. Product ions were detected in the Orbitrap at a resolution of 35,000. Precursors were dynamically excluded for 20 s.

All mass spectrometric raw data were processed with Maxquant (version 1.5.3.8) using default parameters. Briefly, MS2 spectra were searched against the Uniprot MOUSE.fasta (downloaded at: 18.47.2017) database, including a list of common contaminants. False discovery rates on protein and PSM level were estimated by the target-decoy approach to 1% (Protein FDR) and 1% (PSM FDR) respectively. The minimal peptide length was set to seven amino acids and carbamidomethylation at cysteine residues was considered as a fixed modification. Oxidation (M) and Acetyl (Protein N-term) were included as variable modifications. The match-between runs option was enabled. LFQ quantification was enabled using default settings.

Statistical significance was calculated from independent experiments (n) for analysis of all experiments. A two-tailed unpaired student’s t-test was used to evaluate the statistical significance between two groups for all normally distributed raw data. The statistical significance between more than two groups for all normally distributed raw data was evaluated using one-way analyses of variance (ANOVA; Tukey post hoc test was used to determine the statistical significance between the groups). Significant differences were accepted at p < 0.05. For box plots the median divides the box, while the upper boundary of the box corresponds to the third quartile and the lower boundary corresponds to the first quartile. The minimum and the maximum values extend as bars from the bottom and top of the box.

In order to gain an in-depth understanding of NCALD function in the brain, we acquired the heterozygous Ncald knockout (Ncald^KO/WT) animals from the Jackson Laboratories [Bl6N(Cg)-Ncald^{tm1.1(KOMP)Vlcg}/J, Stock Number: 018575]. After crossing the heterozygous parents, we obtained 25% Ncald^KO/KO, 50% Ncald^KO/WT and 25% Ncald^WT/WT mice, according to the Mendelian inheritance. Ncald^KO/KO mice were fertile (with lower fertility rate) and showed normal survival (data not shown). We observed more than a half of NCALD reduction in Ncald^KO/WT and a complete absence of NCALD in Ncald^KO/KO animals in the brain (Figure 1A). Furthermore, Ncald^KO/KO animals revealed significantly reduced body weight compared to their WT littermates (Figure 1B). Since NCALD is highly abundant in neurons, we next analyzed the gross morphology of NCALD deficient brains (Figure 1B). We found that Ncald^KO/KO animals have significantly smaller brains compared to WT littermates, however when normalized to reduced body weight the reduction in brain weight was not significant (Figure 1B). Interestingly, neither the body weight nor the brain weight or morphology were altered by the heterozygous Ncald knockout (Supplementary Figures S1A–C).

Next, we used the Nissl staining on consecutive 40 μm thin sections to characterize the brain morphology of Ncald^KO/KO mice at 2 weeks (P14), 1 month (P30) and 4-months of age (adult). We found that although no significant changes in brain weight and gross morphology could be detected at P14 (Supplementary Figures S2A,B) and P30 (Supplementary Figures S2C,D), adult Ncald^KO/KO brains exhibited significantly enlarged lateral ventricles and disturbed hippocampal morphology, which was accompanied by a significantly reduced subgranular zone length (SGZ; Figures 2A–D). Interestingly, this SGZ reduction was proportional to the hippocampal volume, indicating that Ncald^KO/KO brains possess smaller hippocampi in general. These morphological changes in Ncald^KO/KO animals could either indicate a brain maturation defect occurring during the adolescence or be a sign of a progressive neurodegeneration, which would get more severe later in the adulthood. To answer this question we analyzed the brain morphology of 1.5 years Ncald^KO/KO animals and older. However, we did not detect any severe exacerbation of the phenotype in old Ncald^KO/KO animals compared to the adult mice (data not shown). Ncald^KO/KO brains also revealed no signs of astrogliosis, marked by immunostaining for GFAP (Pekny and Pekna, 2014; Supplementary Figure S3A). Furthermore, neither overall neuronal cell density nor neuronal complexity per se were altered by the homozygous Ncald knockout in mice (Supplementary Figures S3B,C). Taken together, our data point to brain maturation defects in the adult Ncald^KO/KO mice and suggest that NCALD has a specific role during postnatal brain maturation.

To address the role of NCALD in brain maturation, we first analyzed NCALD levels in the WT mouse brain at various developmental time points (E16, P1 and P10–P14). We found that NCALD is present at very low levels during the embryonic stages and increases dramatically at P10–P14 (Figure 3A and Supplementary Figure S1D). In the adult brain, NCALD was found to be present throughout the brain, but was particularly enriched at multiple sites including hippocampal regions, dentate gyrus (DG) and CA3 as well as in the presubiculum (PreS; Figure 3B). Immunostaining of Ncald^KO/KO brain with NCALD antibody failed to show any signal (data not shown), demonstrating the specificity of the antibody. In cultured hippocampal neurons NCALD was co-localized with the excitatory vesicular glutamate transporter 1 (VGLUT1) and inhibitory vesicular GABA transporter (VGAT) presynaptic markers, indicating the presence of NCALD at presynaptic terminals (Figure 3C).

Since NCALD is strongly enriched in the hippocampus (Figure 3B) and homozygous Ncald knockout brains reveal defects in the architecture of the DG (Figure 2), we speculated that NCALD might be specifically involved in the regulation of the hippocampal morphology in the mouse brain. Since, defects in the hippocampal morphology in the adult Ncald^KO/KO animals can be associated with impaired postnatal neurogenesis, we first examined the overall cell proliferation in the DG using Ki-67, a protein only expressed in actively dividing cells (Scholzen and Gerdes, 2000), and found no general defect in cell proliferation (Supplementary Figure S5B). Following this, we analyzed newly generated neurons in the DG of Ncald^KO/KO animals and their control littermates by using Doublecortin (DCX) antibody, which labels newly generated granule cells undergoing migration (Couillard-Despres et al., 2005). We observed that the DCX intensity was significantly increased in Ncald^KO/KO mice at P14 whereas it was unaltered by homozygous Ncald deletion in the DG of P30 animals (Figures 4A,B). In contrast, in the DG of adult Ncald^KO/KO animals, we observed a significant reduction of DCX intensity as well as DCX⁺ neuronal density when compared to WT littermates (Figures 4C,D). Moreover, those few DCX⁺ neurons, which were present in the Ncald^KO/KO DG showed a tangential orientation instead of integration in the granule cell layer pointing towards a possible migration defect (Figure 4C).

To investigate this possibility, we analyzed the cortical layering in the 4-month-old Ncald^KO/KO animals. Improper organization of six cortical layers might be a consequence of the defects in neuronal migration, originating prenatally or early postnatally (Molyneaux et al., 2007). Immunohistochemical analysis of superficial and deep cortical layers stained with T-Box, Brain 1 (TBR1) and Cut Like Homeobox 1 (CUX1) antibodies revealed no alteration of cortical structure in Ncald^KO/KO animals (Supplementary Figures S4A,B), which suggests that NCALD does not regulate cortical layering. Taking into consideration a significant loss of DCX⁺ neurons in the adult Ncald^KO/KO animals, we also analyzed the heterozygous Ncald^KO/WT brains. We found that the DCX⁺ neuronal density as well as the DCX intensity was not significantly altered in Ncald^KO/WT brains, indicating that heterozygous Ncald knockout does not affect adult neurogenesis (Supplementary Figure S4C). Considering the possibility of degeneration of the DCX⁺ neurons, we immunostained Ncald^KO/KO brains for cleaved caspase 3 (an apoptotic marker), but we did not detect any apoptotic cell deaths in neurons lacking NCALD (data not shown).

To further investigate at which stage of adult neurogenesis NCALD regulates the DG granule cell function, we first quantified all the proliferating cells in the DG using the Ki-67 antibody as a cell-proliferating marker (Supplementary Figure S5A). However, we did not find any significant difference in the number of proliferating cells in adult Ncald^KO/KO compared to WT mice (Supplementary Figure S5B). Next, to label neural stem cells type 1 and type 2 we used GFAP and Nestin antibodies in combination with Ki-67, while type 3 neuronal stem cells were identified as being positive for Ki-67 along with DCX (Supplementary Figure S5A, and Sibbe et al., 2015). However, we did not observe any significant differences in the percentages of type1, 2 or 3 population of proliferating neuroblasts in Ncald^KO/KO compared to WT mice (Supplementary Figures S5C,D). Taken together, these data indicate that NCALD is not involved in neuroblast proliferation and likely regulates the DG granule cell function at the stage of neuronal maturation.

Myelination is another important postnatal developmental event in the mouse brain (O’Rourke et al., 2014). Therefore, we next examined the myelination upon NCALD depletion by analyzing the levels of MBP in the 1-month-old Ncald^KO/KO animals. By using both immunohistochemistry and Western blotting, we found that MBP levels were significantly lower in brains lacking NCALD (Supplementary Figure S6A). Since the defects in the myelination could originate from a direct role of NCALD in oligodendrocytes, we analyzed the NCALD expression levels in oligodendrocytes by immunostaining with anti-APC antibody, an oligodendrocyte-specific marker (Lang et al., 2013). Co-localization analysis showed that NCALD is absent from these cells (Supplementary Figure S6B), indicating that NCALD does not function directly in oligodendrocytes.

To analyze if MBP reduction observed at P30 is a result of delayed myelination, we quantified the levels of MBP also at P14 and adult brains of Ncald^KO/KO mice (Supplementary Figures S6C–E). Indeed, we found no significant reduction in the MBP levels in adult brains, pointing towards a possible myelination delay in the 1-month-old Ncald^KO/KO animals. Since, it is known that the decreased myelination can be a consequence of reduced number of axons, we also analyzed the levels of neurofilament, an axonal marker in Ncald^KO/KO brain (Supplementary Figures S6C–E). We found no significant difference in the neurofilament levels ruling out decreased number of axons as the cause for the myelination defect observed in the Ncald^KO/KO brain.

In order to unveil the mechanism underlying the morphological defects observed in the brain upon Ncald knockout, the interactome of NCALD was investigated using mass spectrometry (LC-MS) analysis of WT and Ncald^KO/KO brain samples at P30. Three different IPs were performed in triplicates with the following groups: (1) three IPs with NCALD antibody using WT brain lysate; (2) three IPs with NCALD antibody using Ncald^KO/KO brain lysates; and (3) three independent negative IPs (beats only) with WT lysates. The mass spectrometry result confirmed the absence of NCALD in the brain of Ncald knockout samples. The final list of identified proteins was based on the fact that the candidate was present only in the WT NCALD IPs (group 1) but absent from all IPs of groups 2 and 3. Only three proteins met these criteria (Figure 5A). Due to the highest number of identified peptides for MAP3K10 this protein was selected for further analysis. We first confirmed the interaction of MAP3K10 and NCALD in the brain by co-immunoprecipitation using the MAP3K10 antibody (Figure 5A).

MAP3K10 functions as a part of the MAP kinase pathway, upstream of two major MAPKs, namely JNK and P38 (Hirai et al., 1997). JNK is known to regulate various cellular processes, including apoptosis, neuronal differentiation, axonal growth and branching. Additionally, JNK has recently been reported to regulate the adult neurogenesis (Coffey, 2014; Mohammad et al., 2018; Figure 5B). Taking into account the impaired adult neurogenesis in Ncald^KO/KO animals, we evaluated the JNK and pJNK level in adult NCALD-depleted brains by Western blotting. We found that in Ncald^KO/KO brains the phosphorylation of JNK (pThr 183 and pTyr 185) was significantly upregulated compared to wildtype littermates (Figure 5C). At the same time, the phosphorylation of JNK in heterozygous Ncald knockout brains was not altered (Supplementary Figure S7A). These data suggest a specific role for NCALD in the MAP3K10-regulated JNK activation. These findings, together with the fact that alterations in the JNK pathway have already been reported in SMA (Genabai et al., 2015), prompted us to subject the spinal cord lysates from Ncald^KO/KO and WT animals to LC/MS analysis. Our results from the spinal cord confirmed the data obtained from the brain and revealed the MAP3K10 as one of the top interaction partners of NCALD (data not shown). However, we did not find any significant differences in p-JNK in the spinal cord from Ncald^KO/KO animals, suggesting a brain-confined modulation of JNK pathway by NCALD (Supplementary Figure S7B).

Finally, since NCALD reduction has been shown to act protective in MNs derived from the spinal cord of SMA transgenic mice (Riessland et al., 2017), we analysed the effect of NCALD deficiency on MN morphology in Ncald^KO/WT and Ncald^KO/KO mice. We examined the morphology of cultured MNs with a MN specific marker ChAT as well as microtubule stabilising protein TAU. In accordance with our previous findings (Riessland et al., 2017), we found that both 50% reduction and a complete deletion of NCALD significantly increased the length of MN axons compared to controls (Figure 6). Interestingly, Ncald^KO/KO MNs had significantly shorter axons when compared to MNs isolated from Ncald^KO/WT mice. Since this phenotype in Ncald^KO/KO mice was accompanied by a strong, but not significant increase in a number of secondary branches, we reason that decreased axonal length in Ncald^KO/KO neurons results from their increased secondary branching (Figure 6). On other hand, we found that in Ncald^KO/WT and Ncald^KO/KO hippocampal neurons axonal length was not altered (Supplementary Figure S7C).

We found that NCALD levels are steadily increasing at postnatal stages until P14. Moreover, we found that homozygous Ncald knockout significantly increases the intensity of DCX (a marker for newly differentiated and immature cells) in the early adolescent (P14) brain. These data are in agreement with our previous finding, where NCALD reduction in cultured MN-like cells has been shown to promote their early differentiation (Riessland et al., 2017). In contrast to the P14, we found that the loss of NCALD in the adult DG significantly decreases the DCX intensity as well as the DCX⁺ cell number. The transition of immature DCX positive cells to mature granule cells is most pronounced during P7–P28, hence it is a crucial stage for the formation of the SGZ (Radic et al., 2017). Furthermore, the full electrophysiological maturation of new granule cells progresses over the period of ca. 3–7 weeks after cell division (Overstreet-Wadiche and Westbrook, 2006; Zhao et al., 2006). Therefore, potential effects of NCALD loss on brain morphology would arise earliest at 7 week-of age. This is strongly in line with our observations, where no significant changes in the hippocampus could be detected in P30 animals, however they were prominent in adult Ncald knockout mice. Additionally, we found that NCALD depletion does not affect the proliferation of various types of neuroblasts in the adult DG. However, future studies using the BrdU incorporation method are required to fine-tune the onset of these aberrations and reinforce our findings. Moreover, it has been shown that adult neurogenesis can directly regulate the volume and morphology of the hippocampus (Fuss et al., 2014; Baptista and Andrade, 2018). Therefore, the loss of adult neurogenesis observed in adult Ncald knockout animals could be an accumulated loss of DCX⁺ immature neurons, which were not able to integrate into the granule layer. Interestingly, in a recent database of RNA expression profiles during adult neurogenesis, Ncald expression has been shown to specifically increase at the immature granule cell stage (Hochgerner et al., 2018). These data corroborate our finding that adult Ncald^KO/KO animals show a significant reduction specifically in immature granule cell population of neurons (Figures 4C,D) and points towards a stage specific role of NCALD in the adult neurogenesis.

We show that the changes in adult neurogenesis observed in Ncald^KO/KO mice are accompanied by hyperactivation of the JNK pathway in the brain of NCALD depleted animals. These data are in agreement with the recent study showing that JNK acts as a negative regulator of adult neurogenesis (Mohammad et al., 2018). Considering that the gradient regulation of JNK pathway during the embryonic development of the brain has been implicated in the migration and maturation of neurons (Hirai et al., 2002), JNK activation could well be the mechanism responsible for the aberrant migration of newborn neurons in adult Ncald^KO/KO animals. Therefore, further rescue studies of adult neurogenesis defects in Ncald^KO/KO animals treated with blood-brain barrier permeable JNK inhibitor (Mohammad et al., 2018) as well as by targeted overexpression of Ncald can deepen our understanding of NCALD function.

Most of the NCALD-enriched brain regions described in the current study have already been reported (Girard et al., 2015), except for the PreS. PreS is a part of the parahippocampal region (Witter et al., 2000) which has been strongly implicated in spatial navigation (Boccara et al., 2010). Thus, we suggest that spatial navigation might be disturbed by NCALD loss in the adult mice, however this hypothesis needs to be tested in future experiments.

Although NCALD has been previously implicated in synaptic function (Kedracka-Krok et al., 2015), our study provides the first direct evidence of NCALD localization at synaptic boutons of hippocampal neurons via colocalization of NCALD with VGLUT1, an excitatory synaptic marker and VGAT, an inhibitory synaptic marker.

Multiple studies indicate the importance of hippocampal shape and morphology in cognitive functions (Smith et al., 2012; Voineskos et al., 2015). Smaller hippocampi, as well as severe morphological disturbances in hippocampal shape observed in Ncald^KO/KO animals implicate NCALD function in cognition. Indeed, NCALD has been associated with schizophrenia (Vercauteren et al., 2007) and autism (Ben-David et al., 2011), both representing neurodevelopmental disorders associated with cognitive loss and characterized by enlargement of lateral ventricles both in human patients (Movsas et al., 2013) and genetic mouse models (Pletnikov et al., 2008). A similar ventricle enlargement phenotype, as well as impairment in postnatal development has been observed for Ncald^KO/KO mice in the current study. Moreover, phenotypic data available at the International Mouse Phenotyping Consortium platform indicate that Ncald^KO/KO mice are hyperactive and anxious and exhibit severe reduction of body mass,¹ similar to what we observed in our Ncald^KO/KO animals. Henceforth, a detailed study of NCALD function in the pathophysiology of schizophrenia, autism, depression, stress-related disorders like bipolar disorder and anxiety, as well as a comprehensive analysis cognitive behavior of Ncald^KO/KO animals have a potential to reveal a disease mouse model for neurodevelopmental disorders with subtle behavioral symptoms, thereby improving our understanding of such disorders.

SMA therapeutics reached a very important milestone in the last 2 years, with the FDA and EMA approval of Nusinersen for treatment of SMA patients outside the clinical trial (Scoto et al., 2017). Nusinersen is a modified ASO which binds to the SMN2 pre-mRNA at an intronic splice silencer site, whereby it disrupts the interaction with negative splicing factors. This in turn promotes the inclusion of exon 7 in the SMN2 mRNA thereby enhancing the SMN protein levels (Rigo et al., 2014). Nusinersen has shown significant benefits in the progress of SMA patients, however it could not fully revert the normal motor functions in patients (Talbot and Tizzano, 2017). Specifically, in case of type I SMA patients, who usually have only two SMN2 copy and some even only one (Feldkötter et al., 2002), increasing the SMN level via one or two SMN2 copies may not be sufficient to fully counteract SMA pathology. Therefore, there is a need for SMN-independent therapies, which can work in combination with Nusinersen or other SMN-dependent therapies (Finkel et al., 2017; Mendell et al., 2017).

We have already published the therapeutic importance of NCALD reduction in the context of SMA (Riessland et al., 2017). Supporting our previous studies here, we show that reduction of NCALD even by 50%, such as in Ncald^KO/WT mice has no systemic consequences for the animal physiology. Furthermore, we found that NCALD reduction in Ncald^KO/WT MNs increases the axonal length independent of SMA.

We observed that Ncald^KO/KO can alter JNK phosphorylation in the brain but not in the spinal cord (Figure 5C and Supplementary Figure S7B). Although both the spinal cord and the brain are the part of the CNS, there are significant variations in their metabolic, functional and defense mechanisms (Panov et al., 2011). These variations could underlie the neuronal type-specific response of JNK signaling pathway to the NCALD depletion in either brain or spinal cord. Moreover, we found that while Ncald^KO/KO significantly upregulated the JNK phosphorylation (Figure 5C), half reduction (Ncald^KO/WT) has no effect on the JNK activation (Supplementary Figure S7A), suggesting its safe use in SMA therapy.

Interestingly, increased JNK phosphorylation has also been detected in the spinal cord of SMA mice, which however is linked to reduced SMN levels and cellular stress response in spinal cord MNs (Genabai et al., 2015). Consequently, inhibition of JNK in SMA mice by using the D-JNKI1 inhibitor ameliorated the SMA pathology (Schellino et al., 2018). We speculate that different signaling pathways operate to regulate JNK signaling in SMA disease mouse model and in Ncald knockout condition. In this scenario, cellular stress response causes JNK activation in SMA mice, whereas Ncald knockout leads to the loss of MAP3K10–NCALD interaction, which causes the increased JNK phosphorylation in the brain.

Taken together, our study identifies a novel link between NCALD and adult neurogenesis in the hippocampus, possibly via MAP3K10-JNK pathway and establishes the safety of NCALD reduction as viable option for a combinatorial therapy to treat SMA.