Mice were housed, bred, and treated in an authorized facility (agreement number A751319). All experimental procedures involving mice have been approved by the French Ministry of Research and Higher Education (project authorization number 3572-201601 0817294743 v5). C57BL/6J (Charles River laboratories) and Id4−/− (Yun et al., 2004) mice were used at 2–3 months of age. Glast:CreERT2 (Mori et al., 2006) mice bred to Id4fl mice (Best et al., 2014) to obtain Glast:CreERT2;Id4fl mice. To induce Id4 deletion, a solution of 60 mg/kg of tamoxifen and 20 mg/kg of progesterone diluted in corn oil was administered by oral gavage to pregnant females harvesting embryos at embryonic day 15 (E15). Pups were then obtained at P0 or P10 for wholemount dissection.

Animals were sacrificed by cervical dislocation. Then, brain was dissected and the whole ventricular-subventricular zone (V-SVZ) was microdissected as described in Mirzadeh et al. (2010). Fresh tissue was either fixed with 4% paraformaldehyde (PFA; Electron Microscopy Sciences, EMS) and incubated with anti- ZO-1 (1:200, Thermo Fisher Scientific ref. 402200); acetylated tubulin (6-11B, 1:200, Sigma Aldrich ref. T6793) primary antibodies; or fixed with cold 70% ethanol for 10 min and incubated with the anti-gamma tubulin (GTU88, 1:200, Abcam ref. ab11361) primary antibody. Samples were incubated with secondary antibodies Alexa Fluor^TM 488 and 596 (1:1000, Life Science Technologies). Then wholemount sections were microdissected and mounted with fluoromount (Sigma, ref. F4680). Planar cell polarity was determined by the altered orientation of the BB with respect of the cell wall as described in Mirzadeh et al. (2010).

Animals were anesthetized with 1 g/Kg sodium pentobarbital (EUTHASOL) and intracardially perfused with 4% PFA in NaCl 0.9% solution. Brains were dissected and post-fixed in the same solution for 24 h. Fifty micron coronal sections were obtained using a vibratome (Microm). Floating sections were permeabilized with 0.01 M phosphate buffer saline (PBS) containing 0.1% Triton X-100 for 5 min, blocked for 1 h with 10% Normal Goat Serum (Eurobio Ingen, cat. CAECHV00-0U) in PBS–Triton 0.1% (blocking buffer) at room temperature RT and incubated overnight at 4°C with rabbit anti-ID4 (1:1000, Biocheck ref. BCH-9/82-12) and mouse anti-FOXJ1 (1:500, eBiosc 14-9965-82) antibodies diluted in blocking buffer. Then samples were incubated with anti-mouse Alexa Fluor^TM 488 and anti-rabbit Alexa Fluor^TM 596 secondary antibodies (1:1000, Life Science Technologies) diluted in blocking buffer. Finally, sections were incubated in DAPI solution for nuclear staining (Invitrogen, cat D3571) and mounted on glass-slides with Fluoromount (Sigma, cat. F4680).

For scanning electron microscopy, whole-mount preparations of the lateral wall of lateral ventricles of four animals per group were dissected and fixed with 2% PFA + 2.5% glutaraldehyde (EMS) in 0.1 M phosphate buffer (PB) and post-fixed with 1% osmium tetroxide (EMS) in PB for 2 h, rinsed with deionized water, and dehydrated first in ethanol then with CO₂ by critical point drying method. The samples were coated with gold/palladium alloy by sputter coating. The surface of the lateral wall was studied under a Hitachi S-4800 scanning electron microscope using Quantax 400 software (Bruker Corporation) for image acquisition.

For pre-embedding immunogold staining, mice were perfused with 4% PFA in 0.1 M PB. Brains were postfixed in the same fixative solution overnight at 4°C and sectioned into 50 μm transversal sections using a vibratome. Pre-embedding immunogold staining with rabbit anti-ID4 antibody (1:500; Biocheck) were carried out as previously described (Sirerol-Piquer et al., 2012). Sections were contrasted with 1% osmium tetroxide, 7% glucose in 0.1 M PB and embedded in Durcupan epoxy resin. Subsequently, 1.5 μm semithin sections were prepared, lightly stained with 1% toluidine blue and selected at the light microscope level. Selected levels were cut into 60–80 nm ultrathin sections. These sections were placed on Formvar-coated single-slot copper grids (EMS) stained with lead citrate and examined at 80 kV on a FEI Tecnai G² Spirit (FEI Company) transmission electron microscope equipped with a Morada CCD digital camera (Olympus).

Fluorescent images were obtained using a Zeiss ApoTome 2 Microscope or Olympus Confocal microscope FV1000. Images were analyzed using ImageJ software.

Statistical analysis was performed using GraphPad Prism 6. Unless otherwise indicated in the figure legends, non-parametric Mann–Whitney test was used to compare experimental and control groups. Values are expressed as mean ± standard deviation.

We performed ID4 immunolabeling on wholemount preparations of the LV of adult C57BL/6J mice (Figure 1A). We detected the expression of ID4 protein in ependymal cells as can be observed by co-localization with FOXJ1 protein (Figure 1B). To confirm this observation, we performed immunogold labeling for the ID4 protein in ultrathin sections from the V-SVZ. Several cell types, such as progenitor stem cells and neuroblasts were positive for ID4-immunogold (Figure 1C indicated by distinct pseudo-colors). In addition, ID4 labeling was detected in ependymal cells lining the LV confirming our immunofluorescence results (arrowheads in Figure 1C).

In order to investigate the role of ID4 in ependymal cells, we analyzed the V-SVZ of Id4−/− mice (Id4KO) (Yun et al., 2004). Id4KO mice consistently displayed enlarged ventricles (Figures 2A,B). This was associated with thinning of the ventricular wall and stretching of the ependymal cells, as observed by light microscopy in toluidine blue-stained semithin sections (Figure 2C). By electron microscopy, the V-SVZ is characterized by a high cell density organized in four cell layers (Doetsch et al., 1997) (delimited by dotted line in Figure 2C, right panel) compared to the adjacent brain parenchyma displaying lower cell density and abundance of myelin sheets (asterisks in Figure 2C, right panels). Ependymal cells were more elongated and thinner in Id4KO animals, as shown by quantifications on electron microscopy images (Figure 2D). Scanning electron microscopy of wholemount preparations revealed the absence of adhesion point –the area where lateral and medial ventricle walls adhere to each other in Id4−/− mice (Figure 2E). In addition, we observed that ependymal cell density was decreased in all three rostral, central and caudal areas of the ventricle wall but we did not observe obvious differences in cilia shape or length.

To confirm a decrease in ependymal cell density, we performed immunofluorescent staining on wholemount preparations of the ventricle wall (labeling tight junctions with ZO-1 antibody) and cilia (acetylated tubulin, 6-11B) in WT and Id4KO mice (Figure 3A). The density of ependymal cells was significantly decreased in Id4KO mice together with an increase in the cell surface when compared the same regions (Figures 3B,C). Despite higher intensity of ZO-1 staining, we found no notable differences in the number of pinwheels or the number of monociliated cells (corresponding to B1 cells) in the center of the pinwheels (Figures 3D,E). Several reports have linked PCP disruption with CSF flow alterations (Mirzadeh et al., 2010; Jiang et al., 2020). To investigate whether PCP of ependymal cells could be altered in Id4KO brains, we measured cilia BB orientation (Mirzadeh et al., 2010). Basal bodies were labeled with anti-γ-tubulin (GTU88) antibody and their orientation was determined relative to the ependymal cell wall labeled with anti-ZO1 antibody. We noticed that the organization of BB patches was altered in the Id4KO mice (indicated by arrowheads in Figure 3F), with a significant decrease of the median orientation (Figure 3G).

Ependymal cells differentiate from radial glia cells at E14-16 and maturation occurs from caudal to rostral orientation during the first postnatal weeks (Spassky et al., 2005), during which the primary cilium is replaced by multiple motile cilia (9 + 2 microtubule doublets) (Mirzadeh et al., 2010). To evaluate whether ID4 was involved in ependymal cell differentiation and/or maturation we induced ID4 deletion in GlastCreERT2-Id4flox (Id4cKO) mice at E15 and analyzed their phenotype at P0 to evaluate differentiation and at P10 to analyze maturation (Figure 4A). We performed immunofluorescence for γ-tubulin (GTU88) to identify the BB and for Z01 to identify the cell wall in wholemount preparations at both P0 and P10 in rostral and caudal regions (Figure 4B). At P0, the number of ependymal cells was significantly decreased in the Id4cKO notably in the rostral area (Figure 4C), suggesting that ID4 may be involved in differentiation of ependymal cells from radial glial cells at late embryonic stages. We next evaluated ependymal cell maturation by measuring the proportion of ependymal cells with multiple motile cilia. The presence of multiciliated cells is very low at P0 when maturation starts. In contrast, at P10, ependymal cells display a characteristic BB pattern indicative of maturation (Spassky et al., 2005). We observe decreased fraction of multiciliated cells in rostral regions of the ventricle wall in Id4cKO compared to control animals. Together, our data suggest that ID4 may be important for ependymal cell maturation and correct cilia development.