CROCCP2 acts as a human-specific modifier of cilia dynamics and mTOR signaling to promote expansion of cortical progenitors

Summary

The primary cilium is a central signaling component during embryonic development. Here we focus on CROCCP2, a hominid-specific gene duplicate from ciliary rootlet coiled coil (CROCC), also known as rootletin, that encodes the major component of the ciliary rootlet. We find that CROCCP2 is highly expressed in the human fetal brain and not in other primate species. CROCCP2 gain of function in the mouse embryonic cortex and human cortical cells and organoids results in decreased ciliogenesis and increased cortical progenitor amplification, particularly basal progenitors. CROCCP2 decreases ciliary dynamics by inhibition of the IFT20 ciliary trafficking protein, which then impacts neurogenesis through increased mTOR signaling. Loss of function of CROCCP2 in human cortical cells and organoids leads to increased ciliogenesis, decreased mTOR signaling, and impaired basal progenitor amplification. These data identify CROCCP2 as a human-specific modifier of cortical neurogenesis that acts through modulation of ciliary dynamics and mTOR signaling.

Graphical abstract

Highlights

• •CROCCP2: a hominid-specific gene duplicate uniquely expressed in human fetal cortex
• •CROCCP2 overexpression in the mouse embryo leads to cortical progenitor amplification
• •CROCCP2 is required for basal progenitor amplification in human cortical organoids
• •CROCCP2 acts by decreasing ciliogenesis and enhancing the mTOR pathway

van Heurck et al. identify CROCCP2, a hominid-specific gene duplicate, as a human-specific modifier of neurogenesis in the developing cerebral cortex. They find that CROCCP2 is necessary and sufficient to enhance human cortical progenitor amplification and acts by decreasing primary cilia dynamics and enhancing the mTOR pathway.

Introduction

The primary cilium is a sensory organelle present in most cells, regulating developmental processes and signaling (Goetz and Anderson, 2010). In neural development, it is essential for spatial patterning, neurogenesis, neuronal migration, and differentiation (Louvi and Grove, 2011). Moreover, ciliopathy diseases caused by mutations in genes critical for ciliary dynamics or function display neurodevelopmental abnormalities (Lee and Gleeson, 2011; Louvi and Grove, 2011). As no protein translation occurs in the cilia, trafficking is a major controller of ciliary dynamics and function. This process is under the control of intraflagellar transport (IFT) proteins, responsible for trafficking toward (IFTB) or from the cilia (IFTA), which promote the dynamic growth or resorption of the cilia (Nachury and Mick, 2019).

The disruption of cilia-related genes can lead to diverse and sometimes opposite phenotypes (Foerster et al., 2017; Guo et al., 2015; Tong et al., 2014; Wilson et al., 2012). This reflects the pleiotropy of cilia function depending on cell and developmental context, as well as non-ciliary functions of IFT proteins. Several ciliary genes have been shown to be crucial for neurogenesis (Guo et al., 2015). Ciliary dynamics has also been linked to proper apico-basal polarity (Higginbotham et al., 2013), delamination of neural progenitors (Das and Storey, 2014; Wilsch-Bräuninger et al., 2012), and the regulation of asymmetric divisions (Paridaen et al., 2013). While loss of function of ciliary genes often results in impaired neurogenesis, decreased ciliogenesis can also result in the amplification of neural progenitors, in particular basal progenitors in the developing cerebral cortex (Foerster et al., 2017; Wilson et al., 2012). In these cases, the impact of cilia on neurogenesis is thought to result in part from the modification of cilia-dependent signaling pathways, including mTOR and SHH (Foerster et al., 2017; Tong et al., 2014; Wilson et al., 2012). Ciliary dynamics can also influence cell-cycle speed in a direct fashion: ciliary resorption is required for cell-cycle re-entry (Pruski et al., 2016; Spalluto et al., 2012), thereby influencing the balance of self-renewal versus differentiation of neural progenitors (Kim et al., 2011; Li et al., 2011).

The mechanisms of neurogenesis are largely conserved among all vertebrates, but divergence in neurogenic patterns is thought to play a major role in brain evolution (Libé-Philippot and Vanderhaeghen, 2021). This is most striking in the human cerebral cortex, which is characterized by a massive increase in size and complexity (Amadio and Walsh, 2006; Lui et al., 2011; Rakic, 2009). Cortical progenitors constitute a diverse set of cells located in the proliferative zones lining the lateral ventricles of the dorsal telencephalon, from which all cortical projection neurons are generated (Bonnefont and Vanderhaeghen, 2021; Dimou and Götz, 2014; Kriegstein and Alvarez-Buylla, 2009). The main cortical progenitors are radial glial cells (RGCs), which constitute the ventricular zone (VZ) and divide at its apical surface. RGCs undergo multiple rounds of asymmetric cell divisions, enabling the generation of diverse types of neurons while maintaining a pool of progenitors (Bonnefont and Vanderhaeghen, 2021). In primates, RGCs display stemness for a much longer period (for several months for human RGCs instead of a few days in the mouse) and thus generate more neurons throughout development (Van den Ameele et al., 2014; Astick and Vanderhaeghen, 2018). Species differences are also linked to other types of progenitors, which are more prominent in species with a larger cortex. These are the intermediate/basal progenitor cells (IPCs) and the outer/basal RGCs (oRGCs), which both divide at more basal levels in the cortex, thereby contributing to generating additional neurogenic niches (Borrell and Götz, 2014; Lui et al., 2011; Sun and Hevner, 2014). The IPCs lack any apico-basal processes and divide basally to create the subventricular zone (SVZ). The oRGCs, particularly those expanded in the human cortex, share epithelial features of RGCs and divide basally to create the outer SVZ (oSVZ) and can undergo multiple rounds of self-renewing divisions, thus providing an important additional source of increased neuronal output.

Divergent neurogenic patterns between human and non-human primates are likely to be linked mostly to genomic changes in transcriptional regulatory sequences (Mitchell and Silver, 2018; Pollen et al., 2019). However, recent work has uncovered the role in neurogenesis of “hominid-specific” (HS) genes, which result from recent gene duplications in hominid/human genomes (Dennis and Eichler, 2016). HS genes include >20 gene families that are dynamically expressed during human corticogenesis (Charrier et al., 2012; Suzuki, 2020; Suzuki et al., 2018). These include human-specific NOTCH2NL genes, which increase the self-renewal potential of human cortical progenitors (Fiddes et al., 2018; Florio et al., 2018; Suzuki et al., 2018), and TBC1D3 and ARGHAP11B, which are involved in basal progenitor amplification (Florio et al., 2015; Ju et al., 2016). However, the function during neurogenesis of the other HS genes, if any, remains unknown.

Here, we study CROCCP2, a HS gene duplicate highly expressed during human corticogenesis that results from a partial duplication of the ciliary rootlet coiled-coil (CROCC) gene, which encodes a large coiled-coil protein, called Rootletin, that constitutes the ciliary rootlet (Yang et al., 2002). CROCC is mostly known to be required for the maintenance of sensory cilia (Chen et al., 2015; Mohan et al., 2013; Styczynska-Soczka and Jarman, 2015; Yang et al., 2002, 2005) and centrosome cohesion (Au et al., 2017; Conroy et al., 2012), but its function is otherwise largely unknown.

Here, we find that CROCCP2 is expressed in the human developing cortex but not in other species. CROCCP2 gain of function in the mouse cortex leads to decreased ciliogenesis, increased mTOR signaling, and thereby increased amplification of basal cortical progenitors, while CROCCP2 loss of function in human cortical progenitors leads to the opposite effects. These effects require the inhibitory interaction of CROCCP2 with the ciliary trafficking protein IFT20 (Follit et al., 2006). Our data identify CROCCP2 as a novel uniquely required for human cortical progenitor amplification and potentially link ciliary dynamics to mTOR signaling and human brain evolution.

Results

CROCCP2 emerged through partial duplication of CROCC and is highly expressed during human corticogenesis

We previously identified >20 gene families resulting from HS duplication and highly expressed during human corticogenesis, including paralogs of the CROCC gene (Suzuki et al., 2018). Inspection of the human genome revealed the presence of 7 CROCC-like duplicated loci, including ancestral CROCC and two annotated pseudogenes (CROCCP2 and 3). Only CROCC and CROCCP2 showed abundant expression during human corticogenesis, and hence we focused on these genes (Figure 1). CROCCP2 is the result of a partial duplication of exons 13–21 of CROCC (Figure 1A). CROCC encodes a large protein (2,017 AA) comprising two main domains, a head domain located at the N terminus, followed by a large coiled-coil repetitive domain (Yang et al., 2002; Figure 1B). The CROCCP2 protein is predicted to be 111 AA long and corresponds to the middle part of the CROCC coiled-coil domain (Figure 1B). Importantly, specific peptide sequences corresponding only to CROCCP2 (and not to CROCC nor to CROCCP3) have been identified in several human proteome studies, supporting its existence as a protein (Huttlin et al., 2017; Rolland et al., 2014).

Figure 1.

CROCCP2 is an HS duplicate highly expressed in human corticogenesis

(A) Schematic representation of CROCC and human-specific paralog CROCCP2 gene structure. Conserved exons are highlighted in light gray. Protein-coding and noncoding exons are depicted in blue and gray, respectively. CROCCP2 corresponds to duplication of exons 13–21 and is composed of three coding exons.

(B) Putative protein structure of CROCC and CROCCP2. CROCC is composed of a small head domain and a large coiled-coil repetitive domain. CROCCP2 corresponds to partial duplication of coiled-coil domain.

(C) RNA-seq profile of CROCC HS gene family during human corticogenesis.

(D) RNA in situ hybridization using specific probes for CROCC and CROCCP2 at nine gestational weeks (GWs). VZ, ventricular zone; SVZ, subventricular zone; IZ, intermediate zone; CP, cortical plate. Scale bar, 100 μm.

(E) CROCC protein expression on whole human embryo at GW7, scale bar, 5 mm.

(F) CROCC expression study during human corticogenesis (GW 12 cortical wall). (F′) Inset at apical surface of ventricular zone, note elongated bundles located underneath basal body stained by gamma-tub. Scale bar, 100–5 μm in inset.

(G) CROCC expression during mouse corticogenesis (E14 cortex). (G′) Inset of one primary cilia. Scale bar, 50–3 μm in inset.

(H) Schematic representation of interphase apical progenitors.

See also Figures S1 and S2.

Inspection of available primate genomes revealed multiple CROCC-like coding sequences in the Catarrhini (old world monkeys), suggesting extensive species-specific genomic reorganizations. The CROCCP2 locus is particularly divergent compared with the paralogous loci, but importantly, the CROCCP2 protein studied here was uniquely found in humans and chimpanzees. However, contrary to a single copy of CROCCP2 in humans, the chimpanzee genome (panTro6) contains four tandemly duplicated regions encoding gene products similar to human CROCCP2.

These data indicate that the CROCC gene was duplicated on multiple occasions in the Catarrhini lineages, leading to a gene duplicate encoding the CROCCP2 protein uniquely in humans and chimpanzees.

Using our previously reported human fetal cortex RNA sequencing (RNA-seq) datasets tailored to distinguish human-specific paralogs (Suzuki et al., 2018), we detected strong expression of CROCCP2 at higher levels than CROCC and barely detectable levels of CROCCP3 (Figure 1C). Both genes displayed comparable but distinct patterns, with CROCC expression decreasing after 12 gestational weeks (GWs) while high expression of CROCCP2 was maintained throughout corticogenesis. Moreover, analysis of single cell transcriptome data of human, chimpanzee, and macaque expression in fetal cortex and organoids (Nowakowski et al., 2017; Pollen et al., 2019) confirmed the expression of CROCCP2 in human fetal cortex, both in progenitors and neurons, while interestingly, its expression was undetectable in the chimpanzee cortical organoids (Figure S1). CROCCP2 is thus characterized by prominent and species-specific expression in the human fetal cortex.

To characterize the spatial expression of CROCC and CROCCP2, we performed in situ hybridization on human fetal brain from 9 to 21 GW (Figure 1D, and data not shown). CROCC was found to be expressed throughout the cortical wall, with the highest expression at the apical surface of the VZ. In contrast, CROCCP2 was expressed uniformly throughout the cortical wall (Figure 1D).

We next examined the ancestral CROCC protein localization during human and mouse corticogenesis using a CROCC antibody that does not recognize CROCCP2. This revealed prominent expression throughout the central nervous system at early stages (Figure 1E). In the mouse and human developing cerebral cortex, CROCC immunoreactivity was found to be highest at the apical surface of RGCs (Figures 1F–1H), where it forms a large elongated structure underneath the basal body and the primary cilium. A less intense, punctate labeling was also found throughout the cortical wall, consistent with centrosomal localization. We examined in more detail the expression of CROCC by STED microscopy in a 3D in vitro model system of human corticogenesis (Anja Hasche and P.V., unpublished data). This revealed that CROCC formed large (up to several microns) striated substructures underneath the primary cilum (Figures S2A–S2C). Finally, examination of CROCC in cycling RGCs in mouse and human cortical progenitors revealed that CROCC is present in mitotic centrosomes (Figures S2D–S2G), often in an asymmetric fashion, suggesting association with the mother centriole, as reported previously (Chen et al., 2015).

CROCCP2 gain of function in the mouse embryonic cerebral cortex leads to increased amplification of cortical progenitors

The expression data suggest that CROCC/CROCCP2 could play a role during cortical neurogenesis. To test this, we first performed gain of function of CROCCP2 during mouse mid-corticogenesis (E13) using in utero electroporation (IUE) of plasmids encoding tagged versions of human CROCCP2. The CROCCP2 gain of function did not noticeably alter the overall cell distribution throughout the cortical wall (Figures 2A–2C). However, it led to an increase in the proportion of Tbr2+ basal progenitors (Figures 2G–2I), while the proportion of Pax6+ RGC or Neurod2+ neurons remained unchanged (Figures 2D–2F and 2J–2L). Staining for mitotic marker phosphohistone H3 (pH3) revealed a strong (>3-fold) increase in the proportion of mitotic cells located in basal compartments (abventricular or basal mitoses), together with a more modest increase in apical mitoses (Figures 2M–2O), leading to an increase in the number of total mitotic cells (Figure 2O). Almost all basally dividing cells were Tbr2+ (97.9% versus 100% in controls [n = 9] and CROCCP2 [n = 12] embryos, respectively) and very few Sox2+ (median 0% for both control and CROCCP2 overexpression; n = 9 CONTROL/11 CROCCP2 embryos) (Figures S3A and S3B). These data suggested increased generation and proliferation of IPCs, which was further tested using EdU nuclear labeling to examine cell-cycle dynamics following CROCCP2 gain of function. This revealed an increase in proliferation in VZ/SVZ (Figures 2P–2R) together with an increased proliferative index of Tbr2+ cells (Figure 2S), while Sox2+ cells remained unchanged (Figures S3C–S3E).

Figure 2.

CROCCP2 gain of function leads to basal cortical progenitor amplification

IUE of pCIG (CONTROL) or Pcig-MYC-CROCCP2 (CROCCP2) at E13.5.

(A and B) Immunofluorescence analysis of coronal sections of E14.5 brains stained for DAPI and GFP analyzing the distribution of transfected cells in CONTROL (A) and CROCCP2 (B).

(C) Histograms showing the percentage of GFP⁺ cells in ventricular zone (VZ), subventricular zone and intermediate zone (SVZ/IZ) and cortical plate (CP). n = 19 CONTROL embryos, 16 CROCCP2 embryos, p = 0.1977 (VZ), p = 0.1631 (SVZ).

(D, E, G, H, J, and K) Immunofluorescence analysis of coronal sections of E14.5 brains stained for GFP and Pax6 (D and E) or GFP and Tbr2 (G and H) or GFP and Neurod2 (J and K) for CONTROL (D, G, and J) and CROCCP2 (E, H, and K). Arrows point to double-positive cells.

(F, I, and L) Scatter plots showing the percentage of GFP⁺ cells co-expressing either Pax6 (F), Tbr2 (I), or Neurod2 (L). n = 12 CONTROL embryos, 26 CROCCP2 embryos, p = 0.6356 (PAX6); n = 27 CONTROL embryos, 26 CROCCP2 embryos, ^∗∗p = 0.0082 (TBR2), n = 14 CONTROL embryos, 11 CROCCP2 embryos, p = 0.787 (Neurod2).

(M and N) Immunofluorescence analysis of coronal sections of E14.5 brains stained for GFP and phosphohistone H3 (PH3) in (M) CONTROL and (N) CROCCP2. (N′, N″, and N‴). Inset showing GFP, PH3, and Tbr2 immunofluorescence of basally dividing cells. Vast majority of cells are Tbr2 positive.

(O) Scatter plot showing the percentage and distribution of PH3⁺/GFP⁺ cells. n = 19 CONTROL embryos, 16 CROCCP2 embryos, ^∗∗∗p = 0.0003 (ALL mitosis), ^∗∗p = 0.0073 (APICAL), ^∗∗∗p = 0.0002 (BASAL).

(P and Q) Immunofluorescence analysis of coronal sections of E14.5 brains stained for GFP, EDU, and Tbr2 in CONTROL (P) and CROCCP2 (Q). (P′ and Q′) Inset focusing on triple-positive cells pointed by arrows.

(R) Scatter plot showing the proportion and distribution of EDU⁺/GFP⁺ cells. n = 23 CONTROL embryos, 19 CROCCP2 embryos, ^∗p = 0.0367 (ALL cells), ^∗p = 0.0183 (VZ), ^∗∗∗p < 0.001 (SVZ/IZ).

(S) Scatter plot showing proportion of triple-positive cells EDU⁺, Tbr2⁺, GFP⁺ n = 6 CONTROL embryos, 7 CROCCP2 embryos, ^∗∗p = 0.007. (A, B, D, E, G, H, J, K, M, N, P, and Q) Scale bar, 25 μm.

(T–W) Immunofluorescence analysis of coronal sections of E15.5 brains stained for GFP and DAPI (T and V) or GFP and Neurod2 (U and W) in CONTROL (T and U) or CROCCP2 (V and W). Inset showing double positive (U′ and W′) and GFP-positive only (U″ and W″) cells. Scale bar, 50 μm.

(X) Histograms showing the percentage of GFP⁺ cells in ventricular zone (VZ), subventricular zone (SVZ), intermediate zone (IZ), and cortical plate (CP). n = 16 CONTROL embryos, 13 CROCCP2 embryos, p = 0.3422 (VZ), p = 0.6590 (SVZ), ^∗p = 0.050 (IZ), p = 0.3511 (CP).

(Y) Scatter plot showing proportion of GFP⁺ Neurod2⁺ cells. (n = 12 CONTROL embryos, 7 CROCCP2 embryos, ^∗∗∗∗p < 0.0001). (C, F, I, L, O, R, S, X, and Y) Data are represented as mean ± SEM. Each dot represents an embryo, p values by Student’s t test.

See also Figure S3.

We next examined the consequences of CROCCP2 gain of function at a later time point (48 h post electroporation). We observed an increased cell proportion in the intermediate zone (IZ) that contains newly generated neurons, together with an increased proportion of NeuroD2-positive neurons (Figures 2T–2Y). RGCs and IPCs appeared to be unchanged in number and proliferation (Figures S3F–S3Z), and there was no increase in apoptosis in CROCCP2-expressing cells (Figure S3AA).

Overall, these data suggest that the CROCCP2 gain of function in the mouse cortex in vivo leads to an increase in the expansion of cortical progenitors, most prominently IPC basal progenitors, leading to increased neuronal output.

CROCCP2 decreases ciliary length of cortical progenitors

We next sought to determine the mechanism underlying CROCCP2 effects on neurogenesis. Given the localization of CROCC at the level of the cilia and the importance of this organelle in corticogenesis, we examined the impact of CROCCP2 on ciliary morphology in apical and basal progenitors. To this aim, we performed co-electroporation of CONTROL/CROCCP2 expression vectors together with ciliary marker ARL13B fused with RFP-expressing vector. This revealed that the primary cilia of RGCs, located at the apical surface, were greatly shortened following CROCCP2 gain of function (Figures 3A–3C), while there was no detectable effect on ciliary length in IPCs or newborn neurons (Figures S4A–S4F). We further examined cilia in ex vivo cultures of mouse embryonic cortex to maximize cellular resolution. This confirmed a strong shortening of the primary cilium in Sox2+ RGCs and even an increase in RGCs devoid of any detectable cilium (Figures 3D–3J).

Figure 3.

CROCCP2 gain of function leads to reduced ciliary length in apical cortical progenitors

(A and B) Mouse in utero co-electroporation of ARL13B-RFP fused vector with pCIG (CONTROL) (A) or pCIG-MYC-CROCCP2 (CROCCP2) (B) at E13.5, followed by analysis at E14.5 through immunofluorescence staining of RFP and GFP. Scale bar, 10 μm. (A′ and B′) Inset illustrates primary cilia at the VZ apical surface. Scale bar, 5 μm.

(C) Scatter plot showing primary cilia size at the VZ apical surface. n = 210 cilia, 5 CONTROL embryos, 221 cilia, 4 CROCCP2 embryos, ^∗∗∗p < 0.001.

(D, E, H, and I) Immunofluorescence of primary culture of mouse embryonic cortex infected with either pLenti-CIG-CONTROL (CONTROL) (D and H) or pLenti-CIG-MYC-CROCCP2 (CROCCP2) (E and I) expressing lentivirus at E13 followed by analysis 72 h later. Scale bar, 10 μm. (D and E) Immunostaining of DAPI, GFP, Sox2, and Arl13b. (H and I) Immunostaining of DAPI, GFP, acetylated tubulin (AC TUB), and Sox2.

(F and G) (F) Scatter plots showing ciliary length in Sox2⁺ cells using Arl13b (n = 4 experiments, 322 cilia CONTROL, 326 cilia CROCCP2, ^∗∗∗∗p < 0.0001) and (G) the percentage of nonciliated Sox2⁺ cells (n = 4; ^∗p = 0.286).

(J) Scatter plot showing ciliary length in Sox2⁺ cells using acetylated tubulin (n = 2, 22 CONTROL cilia, 24 CROCCP2 cilia, ^∗∗p = 0.0029). (C, F, and J) Data are presented as mean ± SEM, one dot represents one cilium, p values by Student’s t test. (G) Data are presented as median ± IQ, each dot represents mean percentage of nonciliated cells per experiment, p values by Mann-Whitney test.

See also Figure S4.

Overall, these data indicate that CROCCP2 overexpression leads to a strong reduction in ciliary size in mouse RGCs.

CROCCP2 interacts with IFT20 regulator of ciliary trafficking

We then sought to identify the molecular link(s) between CROCCP2 and primary cilia shortening. A previous proteomic study described direct interactions between CROCCP2 and IFT20 (Rolland et al., 2014), an IFTB protein that regulates the transit from Golgi to the base of the primary cilia (Follit et al., 2006; Pampliega et al., 2013). Remarkably, co-transfection of CROCCP2 with IFT20 in HEK cells revealed their co-localization at a perinuclear localization (Figures 4A and 4B), partially overlapping with Golgi marker GM130 (Figures 4C and 4D). This finding was confirmed in human cortical progenitors in vitro, in which overexpressed CROCCP2 was co-localized with endogenous IFT20 (Figures 4E and 4F). Similar Golgi-like localization was found in mouse cortical progenitors overexpressing CROCCP2 (Figures S4G and S4H).

Figure 4.

CROCCP2 co-localizes with IFT20 at the Golgi level

(A and B) Immunofluorescence staining of HEK cells co-transfected with pCIG-HA-CROCCP2 (HA (CROCCP2) (B) or pCIG (CONTROL) (A) & p3xFLAG-IFT20 (FLAG(IFT20)). Note co-localization of CROCCP2 and IFT20 tagged proteins in CROCCP2 expressing cells.

(C and D) Immunofluorescence staining with GM130 on HEK cells transfected either with p3xFLAG-IFT20 (C) or pCIG-HA-CROCCP2 (D) expressing vector. Note partial co-localization of IFT20 with GM130 (C) and CROCCP2 with GM130 (D).

(E and F) Human 2D in vitro corticogenesis (D32), infected with pLenti-CIG-CONTROL (CONTROL) (E) or pLenti-CIG-MYC-CROCCP2 (MYC(CROCCP2) (F) at D25 and analyzed 6 days after infection. Note Myc signal co-localization with Ift20 endogenous signal (F).

(G–J) Immunofluorescence staining of DAPI, GFP, and Arl13b on mouse cortex primary culture after ex utero at E13.5 using scramble shRNA (G), Ift20 shRNA (shIFT20) (H), pCIG (CONTROL) (I), PCIG-MYC-CROCCP2 (CROCCP2) (J) vectors. Data analyzed 72 h after transfection. Arrows indicate short primary cilia.

(K) Scatter plot showing ciliary length measured through Arl13b immunostainings. Data are represented as mean ± SEM. Each dot represents one cilium. n = 2, 45 cilia CONTROL, 90 cilia scramble shRNA, 49 cilia CROCCP2, 72 cilia shIFT20, p = 0.8053 (CONTROL versus scramble shRNA), ^∗∗∗p = 0.0005 (CONTROL versus CROCCP2), ^∗∗∗p = 0.0008 (CONTROL versus shIFT20), ^∗∗p = 0.0020 (scramble shRNA versus shIFT20), ^∗∗p = 0.0015 (scramble shRNA versus CROCCP2), and p = 0.7801 (CROCCP2 versus shIFT20). p values by one-way ANOVA followed by post hoc Tukey test.

(A–J) Scale bar, 10 μm.

See also Figure S4.

CROCCP2 impacts cortical neurogenesis through IFT20 inhibition

Co-localization of CROCCP2 and IFT20 suggests that CROCCP2 could impact ciliogenesis by inhibiting IFT20 trafficking function required for ciliogenesis. To test this hypothesis, we first examined the effect of IFT20 loss of function in the mouse embryonic cortex, using IUE of IFT20-targeting shRNAs (Figures S5A–S5D). Remarkably, the knockdown (KD) of IFT20 led to the shortening of primary cilia in vitro (Figures 4G–4K). Moreover, the downregulation of IFT20 in vivo led to a striking increase in the proportion of mitotic cells (Figures 5A, 5B, and 5E) and IPCs (Figures 5C, 5D, and 5F).

Figure 5.

CROCCP2 affects cilia and neurogenesis through IFT20 downregulation

(A–D, G–J, and L–O) IUE of scramble shRNA, shRNA against Ift20 (shIFT20), pCIG (CONTROL), PCIG+p3XFLAG-IFT20 (CONTROL+IFT20), pCIG-MYC-CROCCP2 (CROCCP2), PCIG-MYC-CROCCP2+p3XFLAG-IFT20 (CROCCP2+IFT20) at E13.5. (A, B, and G–J) Immunofluorescence of coronal sections of E14.5 brains stained with GFP and phosphohistone H3 (PH3) or (C, D, and L–O) stained with GPF and Tbr2. Scale bar, 25 μm.

(E) Scatter plot showing the percentage of PH3+/GFP+ cells and their localization. n = 9 scramble shRNA embryos, 7 shIFT20 embryos, ^∗p = 0.0399 (ALL mitosis), p = 0.05482 (APICAL), ^∗∗∗∗p < 0.0001 (BASAL).

(F) Quantification of Tbr2⁺/GFP⁺ progenitors (n = 10 scramble shRNA embryos, 7 shIFT20 embryos, ^∗∗∗∗p < 0.0001).

(K) Scatter plot showing the percentage of PH3+/GFP+ positive cell and their distribution. n = 30 CONTROL embryos, 5 CONTROL+IFT20 embryos, 34 CROCCP2 embryos, 13 CROCCP2+IFT20 embryos; ALL mitosis: p = 0.8110 (CONTROL versus CONTROL+IFT20), ^∗∗∗∗p < 0.0001 (CONTROL versus CROCCP2), p = 0.7257 (CONTROL versus CROCCP2+IFT20), p = 0.4684 (CROCCP2 versus CONTROL+IFT20), p = 0.4280 (CONTROL+IFT20 versus CROCCP2+IFT20), ^∗∗∗p = 0.0002 (CROCCP2 versus CROCCP2+IFT20). APICAL mitosis: p = 0.5865 (CONTROL versus CONTROL+IFT20), ^∗∗p = 0.0059 (CONTROL versus CROCCP2), p = 0.9964 (CONTROL versus CROCCP2+IFT20), p = 0.9763 (CONTROL+IFT20 versus CROCCP2), p = 0.5271 (CONTROL+IFT20 versus CROCCP2+IFT20), ^∗p = 0.0434 (CROCCP2 versus CROCCP2+IFT20). BASAL mitosis: p = 0.9929 (CONTROL versus CONTROL+IFT20), ^∗∗∗∗p < 0.0001 (CONTROL versus CROCCP2), p = 0.9990 (CONTROL versus CROCCP2+IFT20), p = 0.0531 (CONTROL+IFT20 versus CROCCP2), p = 0.9982 (CONTROL+IFT20 versus CROCCP2+IFT20), ^∗∗p = 0.0033 (CROCCP2 versus CROCCP2+IFT20).

(P) Scatter plot showing the percentage of TBR2⁺/GFP⁺. n = 7 CONTROL embryos, 3 CONTROL+IFT20 embryos, 7 CROCCP2 embryos, 8 CROCCP2+IFT20 embryos: p = 0.9835 (CONTROL versus CONTROL+IFT20), ^∗∗∗p = 0.0002 (CONTROL versus CROCCP2), p = 0.9409 (CONTROL versus CROCCP2+IFT20), ^∗∗p = 0.0013 (CROCCP2 versus CONTROL+IFT20), p = 0.9999 (CONTROL+IFT20 versus CROCCP2+IFT20), ^∗∗∗∗p < 0.0001 (CROCCP2 versus CROCCP2+IFT20). (E, F, K, and P) Data are presented as mean ± SEM; each dot represents an embryo. (E and F) p values by Student’s t test. (K and P) p values by one-way ANOVA followed by Tukey post hoc test.

See also Figure S5.

These data indicate that IFT20 loss of function leads to decreased ciliary length and increased progenitor amplification, as does the gain of function of CROCCP2. We next tested whether IFT20 gain of function could block the effects of CROCCP2, by performing their combined gain of function (Figures 5G–5P). This revealed that IFT20 gain of function could completely block the effects of CROCCP2 on ciliary length (Figures S5E–S5I), apical and basal mitoses (Figures 5G–5K), and IPC expansion (Figures 5L–5P).

These data indicate that CROCCP2 affects mouse cortical neurogenesis at least in part through the inhibition of IFT20.

CROCCP2 effects on neurogenesis are mediated by increased mTOR signaling

What could be the downstream mechanism linking CROCCP2, ciliary dynamics, and neurogenesis? The primary cilium can regulate many signaling pathways, including the mTOR pathway (Boehlke et al., 2010; Foerster et al., 2017), which interestingly is hyperactivated in human cortical progenitors compared with other primate species (Pollen et al., 2019).

We therefore examined mTOR activity following CROCCP2 gain of function (Figure 6). We first looked at cell size, focusing on the surface of the apical endfoot of RGCs in interphase (Figures 6A–6E), as it is known to be increased by mTOR (Foerster et al., 2017). This revealed that CROCCP2 gain of function in RGCs leads to an increase in RGC apical endfoot surface (Figures 6A–6C and S6). We next examined the levels of phospho-S6 protein, a classical readout of the mTOR pathway activity, on cortical progenitors transfected with CROCCP2. This revealed an important increase in phospho-S6 levels in CROCCP2-overexpressing cortical progenitors (Figures 6D–6F). These data indicate that CROCCP2 gain of function leads to increased mTOR signaling. Could this be directly linked to the effects of CROCCP2 on neurogenesis? To test this hypothesis, we performed combined gain of function of CROCCP2 and KD of mTOR in vivo. Remarkably, this revealed that the effects of CROCCP2 on basal progenitor amplification could be entirely blocked by mTOR loss of function (Figures 6G–6R).

Figure 6.

CROCCP2 affects cilia and neurogenesis through mTOR signaling upregulation

(A and B) Flat-mount analysis of IUE using pCIG (CONTROL) (A) or pCIG-MYC-CROCCP2 (CROCCP2) (B) performed at E13.5 and fixed 24 h later (A, B, D, and E). Whole-mount staining of en-face view of the ventricular zone using DAPI, GFP and ZO1 to stain apical endfoot. (A and B) Arrows point to interphase progenitor apical endfoot—stars mark mitotic progenitors. Scale bar, 10 μm.

(C) Scatter plot of of apical endfoot surface measurement. Data are represented as mean ± SEM. One dot represents one apical endfoot. n = 136 apical endfeet CONTROL—3 embryos,116 apical endfeet CROCCP2—4 embryos, 152 apical endfeet of non-electroporated (NOT ELEC) cortex sides—2 embryos, ^∗∗∗∗p < 0.0001 (CONTROL versus CROCCP2), ^∗∗∗∗p < 0.0001 (NOT ELEC versus CROCCP2), p = 0.8261 (NOT ELEC versus CONTROL). p values by one-way ANOVA followed by Tukey post hoc test.

(D and E) Immunofluorescence analysis of mouse embryonic cortex primary culture infected with either plenti-CIG (CONTROL) (D) or plenty-CIG-MYC-CROCCP2 (CROCCP2) (E) at E13.5 and analyzed 72 h later. DAPI, GFP, and pS6 immunostainings as a readout of the MTOR pathway activity. Scale bar, 25 μm.

(F) Scatter plot showing the percentage of pS6⁺/GFP⁺ cells. Data are represented as mean ± SEM, one dot represents one microscopic field (n = 4, 22 microscopic fields CONTROL, 20 microscopic fields CROCCP2, 22,631 cells CONTROL, 20,382 cells CROCCP2) ^∗∗∗p = 0.004 using Student’s t test.

(G–J and N–Q) IUE of pCIG + scramble shRNA (CO+SHSCR), pCIG-MYC-CROCCP2 + shRNA (CRP2+SHSCR), pCIG-MYC-CROCCP2 + mix of shRNAs against mTOR (CRP2+SHMTOR), pCIG + mix of shRNAs against mTOR (CO+SHMTOR) at E13.5. (G–J) Immunofluorescence of coronal sections of E14.5 brains stained with GFP and phosphohistone H3 (PH3) or (N–Q) stained with GPF and TBR2. Scale bar, 25 μm.

(K–M) Scatter plots showing the percentage and distribution of PH3+/GFP+ cells. BASAL: ^∗∗p = 0.0091 (CRP2+SHSCR versus CRP2+Shmtor), ^∗∗p = 0.0035 (CRP2+SHSCR versus CO+Shmtor), ^∗∗p = 0.0017 (CRP2+SHSCR versus CO+SHSCR), p = 0.7267 (CRP2+Shmtor versus CO+Shmtor), p = 0.8762 (CRP2+Shmtor versus CO+SHSCR), p = 0.9779 (CO+Shmtor versus CO+SHSCR) using one-way ANOVA followed by Tukey post hoc test. APICAL: nonsignificant differences among means, F = 1.503, p = 0.2372 using one-way ANOVA. ALL: nonsignificant differences among means, F = 2.955, p = 0.0511 using one-way ANOVA. (n = 7 CO+SHSCR embryos, n = 11 CRP2+SHSCR embryos, n = 8 CRP2+Shmtor embryos, n = 4 CO+Shmtor embryos). Data are presented as mean ± SEM; each dot represents an embryo.

(R) Scatter plot showing the percentage of TBR2+/GFP+ cells. ^∗∗∗∗p < 0.0001 (CRP2+SHSCR versus CRP2+SHmtor), ^∗∗∗∗p < 0.0001 (CRP2+SHSCR versus CO+SHMTOR), ^∗∗∗∗p < 0.0001 (CRP2+SHSCR versus CO+SHSCR), p = 0.4249 (CRP2+Shmtor versus CO+Shmtor), p = 0.8844 (CRP2+Shmtor versus CO+SHSCR), p = 0.1556 (CO+Shmtor versus CO+SHSCR) using one-way ANOVA followed by Tukey post hoc test (n = 8 CO+SHSCR embryos, n = 7 CRP2+SHSCR embryos, n = 6 CRP2+Shmtor embryos, n = 3 CO+Shmtor embryos). Data are presented as mean ± SEM; each dot represents an embryo.

See also Figure S6.

Overall, these data indicate that CROCCP2 activates the mTOR pathway, which is required for its effect on basal progenitor amplification.

CROCCP2 is required for basal progenitor amplification in human corticogenesis in vitro

We next explored the potential functions of CROCCP2 in the more physiological context of human corticogenesis. To address this crucial point, we first performed CROCCP2 gain-of-function experiments using lentiviral transduction of CROCCP2 in in vitro embryonic stem cell-based models of human corticogenesis (Figures 7 and S7). We first examined the potential effects of CROCCP2 on the cilia of human cortical progenitors in 2D cultures (Espuny-camacho et al., 2013; Figures S7A–S7C). Interestingly, we found that the length of control human cortical progenitors is slightly longer than their mouse counterparts (compared with Figure 3), but CROCCP2 overexpression led to a strong decrease in ciliary length, as observed in the mouse (Figures S7A–S7C). These data indicate that CROCCP2 negatively regulates ciliary dynamics in human cortical progenitors but that additional mechanisms control the final ciliary length in mouse versus human cells. We next examined the impact of CROCCP2 gain of function on mTOR signaling, which revealed increased phopho-S6 levels in human cortical progenitors (Figures S7D–S7F).

Figure 7.

CROCCP2 is required for amplification of basal progenitors in human cortical organoids

(A–H) Cytoarchitecture and cellular identity (FOXG1, SOX2, TBR2, TBR1) of human cortical organoids at day 30 following infection with (A–D) pLenti-CIG-GFP (CONTROL) or (E–H) pLenti-CIG-MYC-CROCCP2 (CROCCP2). Scale bar, 100 μm in (A and E) and 20 μm in (B)–(D) and (F)–(H).

(I and J) Immunofluorescence of Hoechst, GFP, and PH3 performed on cryosections of hES-cell derived organoids at day 30. Arrows point to GFP+ PH3+ double-positive cells. Scale bar, 25 μm.

(K) Scatter plot with mean ± SEM showing the percentage of phospho-histone H3 population among GFP-positive cells (CO versus CRP2 total PH3: ^∗∗∗∗p < 0.0001, CO versus CRP2 apical PH3: ^∗∗∗∗p < 0.0001, CO versus CRP2 basal PH3: ^∗∗∗p = 0.0003, n = 4, pCIG-control [CO]: 4,179 cells from 56 VZs of 12 organoids, pCIG-CROCCP2 [CRP2]: 2,094 cells from 29 VZs of 9 organoids).

(L and M) Immunofluorescence of GFP and TBR2 performed on cryosections of cortical organoids at day 30. Arrows point to GFP+ TBR2+ double-positive cells. Scale bar, 20 μm.

(N) Scatter plot with mean ± SEM showing the percentage of TBR2 (^∗p = 0.0149) populations among GFP-positive cells (L and N: n = 3, pCIG-control [CO]: 581 cells from 6 VZs of 6 organoids, pCIG-CROCCP2 [CRP2]: 681 cells from 6 VZs of 6 organoids; M: pCIG-control [CO]: n = 3, 2,011 cells from 10 VZs of 4 organoids, pCIG-CROCCP2 [CRP2]: n = 4, 1,983 cells from 12 VZs of 5 organoids).

(O and P) Immunofluorescence of Hoechst, GFP, TBR2, and TUJ1 performed on cryosections of cortical organoids at day 60. Red and yellow arrowheads point to GFP+ TBR2+ and GFP+ TUJ1+ double-positive cells, respectively. Scale bar, 20 μm.

(Q–S) Scatter plots with mean ± SEM showing the percentage of (Q) SOX2 (^∗∗p = 0.0036), (R) TBR2 (^∗∗∗p = 0.0006) and (S) TUJ1 (^∗∗p = 0.0013) populations among GFP-positive cells (R and S: n = 3, scramble shRNA [SCR]: 1,567 cells from nine VZs of six organoids, CROCCP2 shRNA [SHCRP2]: 1,301 cells from nine VZs of six organoids; Q: n = 3, scramble shRNA [SCR]: 2,163 cells from 13 VZs of 6 organoids, CROCCP2 shRNA [SHCRP2]: 2,620 cells from 18 VZs of 7 organoids).

(T and U) Immunofluorescence of Hoechst, GFP, and SOX2 performed on cryosections of cortical organoids at day 60. Arrows point to GFP+ SOX2+ double-positive cells located outside of VZ structures delimited by dotted lines. Scale bar, 20 μm.

(V) Scatter plot with mean ± SEM showing the percentage of basal SOX2 cells (^∗p = 0.0198) among total SOX2-positive GFP-positive cells (n = 3, scramble shRNA [SCR]: 668 cells from 12 VZs of 6 organoids, CROCCP2 shRNA [SHCRP2]: 794 cells from 12 VZs of 6 organoids).

See also Figure S7.

Next, we performed gain of function of CROCCP2 in human embryonic stem cell (hESC)-derived cortical organoids (Figures 7A–7H; Cederquist et al., 2019) to assess its effects on neurogenesis. Remarkably, this led to an increase in the proportion of mitotic cells located in basal and apical compartments (Figures 7I–7K) together with an increased proportion of Tbr2+ IPC (Figures 7L–7N), indicating increased amplification of cortical progenitors, most strikingly IPCs, as observed in the mouse in vivo.

Next, we performed loss-of-function experiments in human cortical cells, using shRNAs targeting a CROCCP2-specific noncoding sequence not found in CROCC, leading to efficient downregulation of CROCCP2 (Figure S7G). We first found that cortical cells, following KD of CROCCP2, display increased ciliary length and decreased mTOR signaling, indicating the physiological requirement of CROCCP2 for ciliary dynamics and mTOR signaling in human cortical cells (Figures S7H–S7L).

We then used the KD approach in human cortical organoids at day 60 (Figures 7O–7V), a stage during which basal progenitor amplification starts to be very active. Remarkably, the KD of CROCCP2 resulted in a decrease in the proportion of TBR2+ IPC (Figures 7O, 7Q, and 7R) but also SOX2+ cells located outside the VZ-like structures (Figures 7T–7V), thus corresponding likely to oRGCs. We also found an increased number of neurons following CROCCP2 loss of function (Figures 7P and 7S), suggesting that direct neurogenesis was increased, perhaps reflecting reduced indirect neurogenesis through basal progenitors (Figure 7).

Collectively, these data indicate that CROCCP2 is required for human cortical progenitor amplification, in particular of IPCs and oRGC-like basal progenitors.

Discussion

Here, we identify CROCCP2, a HS gene duplicate that is selectively expressed in human cortical progenitors, where it is required for their amplification. CROCCP2 appears to exert effects on apical and basal progenitors, but the most striking effect of CROCCP2 is the amplification of basal progenitors. Human brain evolution is characterized by an expansion in basal progenitor compartiments, associated with increased brain size, increased neuronal number, and increased brain folding (Lui et al., 2011). Impairment of IPC generation and amplification by EOMES/TBR2 gene disruption leads to severe microcephaly in humans (Baala et al., 2007), and amplification of IPCs has been linked to increased cortical size in amniotes (Cárdenas et al., 2018). We also found that in models of human corticogenesis but not in the mouse, CROCCP2 is linked to the amplification of oRGC-like progenitors. This may have important evolutionary significance as oRGCs are particularly amplified in the human fetal cortex where they display extensive self-renewing capacities (Pollen et al., 2015). As CROCCP2 overexpression in the mouse only leads to IPC amplification, this suggests that the impact of CROCCP2 on progenitor amplification depends on a species-specific context. It will be interesting in the future to further study the function of CROCCP2 in human cortical progenitors and its potential interactions with other HS genes regulating their amplification (Fiddes et al., 2018; Florio et al., 2015; Ju et al., 2016; Suzuki et al., 2018).

CROCCP2 emerged in evolution from a partial duplication of CROCC. CROCC is required for the maintenance of specialized cilia of sensory cells, but its implication in nonspecialized primary cilium remains unclear (Chen et al., 2015; Mohan et al., 2013; Yang et al., 2005). CROCC was previously proposed to play a role in intracellular trafficking (Yang and Li, 2005), as well as centrosomal cohesion by interacting with C-Nap1 (Bahe et al., 2005), the disruption of which leads to dwarfism and microcephaly (Floriot et al., 2015). These data suggest that CROCC may have an ancestral function in cortical neurogenesis, while CROCCP2 displays species-specific effects on cilia and basal progenitor expansion. Future work should determine whether CROCC can function in relation with CROCCP2 in the context of human neurogenesis.

Our data suggest that CROCCP2 acts on ciliary dynamics by interacting with IFT20 at the level of the Golgi apparatus. IFT20 is a well-known IFTB protein, underlining the important role of primary cilia during corticogenesis. Interestingly, IFT20 was previously involved in epithelial-mesenchymal transition (Han et al., 2018), a process mechanistically linked to delamination of RGCs and their conversion into basal progenitors (Itoh et al., 2013), and was also found to be required for normal ciliogenesis and proliferation in mouse adult neural stem cells (Amador-Arjona et al., 2011). While our data provide a direct link between CROCCP2 and ciliogenesis, it is interesting to note that the overall length of the primary cilium in cortical progenitors appears to be similar in mouse and human species, indicating that additional mechanisms control the final length of the cilium in each species. In any case, our data point to the importance of ciliary dynamics, in addition to length per se, on the biology of cortical progenitors, which will be important to study further during neurogenesis.

The primary cilium is involved in complex interplays with the mTOR pathway and autophagy (Boehlke et al., 2010; Pampliega and Cuervo, 2016). Hyperactivation of the mTOR pathway was previously linked to the absence of primary cilia (Boehlke et al., 2010; Foerster et al., 2017), and increased mTOR activity was proposed as a mechanism linking decreased ciliary dynamics and increased generation of basal progenitors (Foerster et al., 2017; Wilson et al., 2012), consistent with our results. Interestingly, the mTOR pathway was shown recently to be hyperactivated in the human fetal cortex compared with macaques and chimpanzees, most strikingly in oRGCs (Pollen et al., 2019), where it was shown to be required for normal cell morphogenesis but not proliferation (Andrews et al., 2020). While we found that CROCCP2 is required for oRGC proliferation, it will be important to explore further the relationships between CROCCP2 and mTOR in human oRGCs, given the many implications of this pathway in neurodevelopmental diseases (Lipton and Sahin, 2014). Other potential signaling pathways that could link ciliary function and neurogenesis should be explored, such as SHH and Wnt (Kostic et al., 2019; Park et al., 2018), but also YAP, recently linked to the control of cortical progenitor generation and amplification (Jabali et al., 2022; Shao et al., 2020). Finally, while our data are focused on cortical progenitors, we found that CROCCP2 is also expressed in neurons, in which case it could contribute to species-specific properties as well.

To sum up, our data identify CROCCP2 as a new human-specific modifier of cortical neurogenesis that links ciliary dynamics, mTOR and basal progenitor amplification, further pointing to recent gene duplications as important molecular effectors of the increase in human brain size and complexity.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti-Pax6	Covance	CAT#PRB-278P; RRID: AB_291612
Rat monoclonal anti-Phospho-Histone H3	Abcam	CAT#Ab10543; RRID: AB_2295065
Goat polyclonal anti-SOX2	Santa Cruz	CAT#sc-17320; RRID: AB_2286684
Rabbit polyclonal anti-SOX2	Abcam	CAT#Ab92494; RRID:AB_10585428
Rabbit polyclonal anti-RFP	Rockland	CAT#600-401-379; RRID: AB_2209751
RAT monoclonal anti-RFP 5F8	Chromotek	CAT#5f8-100; RRID:AB_2336064
Rabbit polyclonal anti-Tbr2 (Eomes)	Abcam	CAT#Ab23345; RRID:AB_778267
Chicken polyclonal anti-EGFP	Abcam	CAT#Ab13970; RRID:AB_300798
Rat monoclonal anti-EGFP (GF090R)	Nakalai Tesque	CAT#04404-84; RRID:AB_10013361
Mouse polyclonal anti-Myc clone 9E10	Roche	CAT#11667149001; RRID:AB_390912
Goat polyclonal anti-Myc	Abcam	CAT#Ab9132; RRID:AB_307033
Rabbit polyclonal anti-rootletin	Sigma-Aldrich	CAT#ABN1714; RRID:AB_2923540
Rabbit polyclonal anti-CROCC	Abcam	CAT#Ab121653; RRID:AB_11129547
Goat polyclonal anti-rootletin	Santa Cruz	CAT#sc-67824; RRID:AB_2229966
Mouse polyclonal anti-ZO1	Invitrogen	CAT#33-9100; RRID:AB_87181
Mouse monoclonal anti-gamma tubulin GTU88	Sigma-Aldrich	CAT#T6557; RRID: AB_477584
Rabbit polyclonal anti-ARL13B	Proteintech	CAT#17711-1-AP; RRID:AB_2060867
Mouse monoclonal anti-Acetylated tub	Sigma-Aldrich	CAT#T6793; RRID:AB_477585
Rabbit polyclonal anti-Neurod2	Abcam	CAT#Ab104430; RRID: AB_10975628
Mouse monoclonal anti-Ha	Sigma-Aldrich	CAT#H3663; RRID:AB_262051
Rabbit polyclonal anti-Flag	Sigma-Aldrich	CAT#F7425; RRID:AB_439687
Mouse monoclonal anti-Flag M2	Sigma-Aldrich	CAT#F1804; RRID:AB_262044
Rabbit polyclonal anti-IFT20	Proteintech	CAT#13615-1-AP; RRID:AB_2280001
Mouse polyclonal anti-GM130	BD Biosciences	CAT#6108223; RRID:AB_398142
Rabbit polyclonal anti-pS6	Cell Signaling Technology	CAT#2211; RRID:AB_331679
Rabbit monoconal anti-Actin	Sigma-Aldrich	CAT#A2066; RRID:AB_476693
Rabbit polyclonal anti-pMtor	Cell Signaling Technology	CAT#2971; RRID:AB_330970
Rabbit polyclonal anti-Ofd1	Abcam	CAT#Ab222837; RRID:AB_2827156
Rabbit polyclonal anti-active Caspase 3	R&D Systems	CAT#AF835; RRID:AB_2243952
AlexaFluor488 donkey anti-mouse	Thermo Fisher Scientific	CAT#A-21202; RRID:AB_141607
AlexaFluor488 donkey anti-rabbit	Thermo Fisher Scientific	CAT#A-21206; RRID:AB_2535792
AlexaFluor488 donkey anti-goat	Thermo Fisher Scientific	CAT#A-11055; RRID:AB_2534102
AlexaFluor488 donkey anti-rat	Thermo Fisher Scientific	CAT#A-21208; RRID:AB_141709
AlexaFluor405 donkey anti-rabbit	Abcam	CAT#Ab175651; RRID:AB_2923541
AlexaFluor647 donkey anti-goat	Jackson ImmunoResearch	CAT#705-605-147; RRID:AB_2340437
AlexaFluor647 donkey anti-mouse	Jackson ImmunoResearch	CAT#715-605-150; RRID:AB_2340862
Cyanin3 donkey anti-mouse	Jackson ImmunoResearch	CAT#715-165-150; RRID:AB_2340813
Cyanin3 donkey anti-rabbit	Jackson ImmunoResearch	CAT#711-165-152; RRID:AB_2307443
Cyanin3 donkey anti-rat	Jackson ImmunoResearch	CAT#712-165-150; RRID:AB_2340666
STAR RED 635 P goat anti rabbit IgG	ABBERIOR	CAT#ST635-1002-500UG; RRID: AB_2893229
STAR RED 580 goat anti mouse IgG	ABBERIOR	CAT#ST580-1001-500UG; RRID:AB_2923543
Biological samples
Human fetal brains	This study	N/A
Chemicals, peptides, and recombinant proteins
Knockout DMEM	Thermo Fisher Scientific	CAT#10829018
Non-essential Amino Acids	Thermo Fisher Scientific	CAT#11140035
Penicillin/Streptomycin	Thermo Fisher Scientific	CAT#15070063
2-BetaMercaptoethanol	Merck	CAT#M3148
200mM L-glutamine	Thermo Fisher Scientific	CAT#25030024
PluriSTEM Dispase-II Solution	Merck	CAT#SCM133
Collagenase type IV	Thermo Fisher Scientific	CAT#17104-019
Matrigel Growth factor reduced	Corning	CAT#354230
Matrigel hESC-qualified	Corning	CAT#354277
E8 medium Flex	Thermo Fisher Scientific	CAT#2A2858501
mTESR1	Stem Cell Technologies	CAT#85850
Stem-Pro Accutase	Thermo Fisher Scientific	CAT#A1110501
ROCK inhibitor Y-27632	Abcam	CAT#Ab120129
B27 supplement	Thermo Fisher Scientific	CAT#17504044
B27 supplement, minus vitamin A	Thermo Fisher Scientific	CAT#12587010
N-2 supplement	Thermo Fisher Scientific	CAT#17502048
Noggin	R&D systems	CAT#1967-NG
mFreSR	Stem Cell Technologies	CAT#05855
HEPES	Thermo Fisher Scientific	CAT#15630080
LDN193189	Stem Cell Technologies	CAT#72147
SB431542	Stem Cell Technologies	CAT#72234
XAV939	Stem Cell Technologies	CAT#72674
DnaseI from bovine pancreas	Sigma-Aldrich	CAT#D4527-40KU
NeuroBasal medium	Thermo Fisher Scientific	CAT#21103049
Horse Serum	Thermo Fisher Scientific	CAT#16050-122
Glutamax 1%	Thermo Fisher Scientific	CAT#35050038
100mM Sodium Pyruvate	Thermo Fisher Scientific	CAT#11360039
N-Acetyl-L-Cysteine	Sigma-Aldrich	CAT#A7250
HBSS	Thermo Fisher Scientific	CAT#14025-050
Human FGF-basic	Peprotech	CAT#100-18B
Tissue TEK	SAKURA	CAT#4583
Triton X-100	Sigma-Aldrich	CAT#T8787
Nycodenz	Axis-Shield Density Gradient Media	CAT#1002424
Coverslips thickness 1.5 diameter 18 mm	VWR	CAT#631-0153
FLUORODISH 60 u 35 mm high	IBIDI	CAT#81166
Glycerol mounting medium	DAKO	CAT#C0563
DABKO anti fading agent	Sigma-Aldrich	CAT#D27802
Critical commercial assays
Prime Star	TAKARA	CAT#R050Q
In Fusion cloning	Clontech	CAT#638909
Nucleofector 2B	LONZA	CAT#AAB-1001
Nucleofector Lonza kit for HES	LONZA	CAT#VPH-5022
5-Ethynyl-20-deoxyuridine (EdU)	Thermo Fisher Scientific	CAT#E10187
Click-iT EdU Alexa Fluor 555 Imaging Kit	Thermo Fisher Scientific	CAT#C10338
X-tremeGENE HP DNA Transfection Reagent	Merck	CAT#6366244001
Experimental models: Cell lines
Human embryonic stem cells H9	WiCell	CAT#NIHhESC-10-0062
Experimental models: Organisms/strains
Mouse: ICR(CD1)	Charles River laboratory	Strain code 022
Oligonucleotides
Primers for in situ hybridization probe preparation	This study	Table S1
Primers for DNA construction	This study	Table S2
Recombinant DNA
psPAX2	Addgene	CAT#12260
pMD2.G	Addgene	CAT#12259
pCAG::myc-tagged CROCCP2-IRES-eGFP	This study	N/A
pLENTI-CAG ::myc-tagged CROCCP2-IRES-eGFP	This study	N/A
pCAG::3xHA-tagged CROCCP2-IRES-eGFP	This study	N/A
pLENTI-CAG::3xHA-tagged CROCCP2-IRES-eGFP	This study	N/A
pCAG::IRES-eGFP	This study	N/A
pLENTI-CAG::IRES-eGFP	This study	N/A
pSilencer2.1-CAG-Venus	Tiberi et al., 2012	N/A
pSilencer2.1-CAG-Venus SH1 IFT20	This study	N/A
pSilencer2.1-CAG-Venus SH4 IFT20	This study	N/A
pCMV::GFP-tagged IFT20 (mouse)	Addgene	CAT#45608
pCMV::3xflag-tagged IFT20 (human)	Addgene	CAT#118033
pCAG::3xflag-tagged IFT20 (human)-IRES-RFP	This study	N/A
pLENTI-Silencer2.1-CAG-Venus SH CROCCP2	This study	N/A
pLENTI-Silencer2.1-CAG-Venus SH MTOR	This study	N/A
pLKO-RFP-shCntrl	Addgene	CAT #69040
pCMV::RFP-tagged-ARL13B-P2A	Kinzel et al., 2010	N/A
pPGK Puro	Addgene	CAT#11349
Software and algorithms
Fiji/ImageJ (version 1.0)	Shindeline et al.,2012	https://fiji.sc/
Prism 8 for macOS (version 8.4.4)	N/A	www.graphpad.com
Other
Zeiss LSM 780 confocal microscope	ZEISS	N/A
LEICA DMi8 inverted confocal microscope equipped with STED	LEICA	N/A
Human embryonic cortex transcriptome	Suzuki et al., 2018	#EGAD00001003915

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to Pierre Vanderhaeghen (pierre.vanderhaeghen@kuleuven.be).

Materials availability

All plasmids generated in this study are available without restriction.

Experimental model and subject details

Human fetal tissue collection and preparation

The study was approved by three relevant Ethics Committees (Erasme Hospital, Université Libre de Bruxelles, and Belgian National Fund for Scientific Research FRS/FNRS) on research involving human subjects. Written informed consent was given by the parents in each case.

Human fetuses were obtained following medical pregnancy termination. Fetuses aged 7 gestational weeks (GW), (9 GW, 12 GW, 15 GW, and 21 GW) were used for RNA sequencing and in situ hybridization of cortical tissue. All cases were examined with standard feto-pathological procedures and none displayed clinical or neuropathological evidence of brain malformation. As soon as possible after expulsion (less than 6 hours), the brain was removed using standard fetal autopsy procedure, frozen in liquid nitrogen for RNA extraction and embedded as a whole in OCT compound (Tissue-Tek Sakura), then snap-frozen in a 2-methylbutane on dry-ice bath for histological studies.

Mice

All mouse experiments were performed with the approval of the Université Libre de Bruxelles Committee for animal welfare. Mouse housing, breeding and experimental handling were performed according to the ethical guidelines of the Belgian Ministry of Agriculture in agreement with European community Laboratory Animal Care and Use Regulations. Animals were housed under standard conditions (12 h light:12 h dark cycles) with food and water ad libitum. For IUE experiments, timed-pregnant mice were obtained by mating adult RjOrl:SWISS CD1 mice (Janvier, France). The plug date was defined as embryonic day E0.5 and the day of birth as P0.

In utero electroporation (IUE)

IUE was performed as previously described (Bonnefont et al., 2019). Briefly, timed-pregnant mice were anesthesized with a ketamine/xylazine mixture at E13.5, and each uterus was exposed under sterile conditions. Plasmid solutions containing 1-2 mg/ml of DNA were injected into the lateral ventricles of the embryos using a heat-pulled capillary. Electroporation was performed using tweezers electrodes (Nepa Gene) connected to a BTX830 electroporator (five pulses of 25 V for 100 ms with an interval of 1s). Embryos were placed back into the abdominal cavity, and mice were sutured and placed on a heating plate until recovery. Embryos were collected 24 or 48 hours after IUE and perfused transcardiacally with ice-cold 4% paraformaldehyde. Brains were dissected and soaked in 4% paraformaldehyde overnight at 4 °C. They were then washed in PBS for 24 h then soaked in 30% sucrose solution overnight for cryopreservation. Brains were included in Tissue-TEK OCT compound (Sakura) at -20 °C overnight then stored at -80 °C. After equilibration the blocs were sectioned with 20-μm thickness with a cryostat (Leica).

Cell lines

Human H9 embryonic stem cells (WiCell) were routinely propagated on mitotically inactivated mouse embryonic fibroblasts in Knockout DMEM (Thermo Fisher Scientific) supplemented with 20% Knockout Serum Replacement (Thermo Fisher Scientific), 1X Non-essential Amino Acids (Thermo Fisher Scientific), 1X Penicillin/Streptomycin (Thermo Fisher Scientific), 1X 2-Mercaptoethanol (Merck), 2mM L-glutamine (Thermo Fisher Scientific).

Method details

DNA constructs

Coding sequence of CROCCP2 was amplified by PCR using Prime Star DNA polymerase (TAKARA) from GW9 human fetal cortex cDNA. Primers were designed on the basis of the sequence of the GRCh38/hg38 reference human genome. The PCR fragment was then subcloned into the pCAG::myc-IRES-eGFP expression plasmid (Dimidschstein et al., 2013; Tiberi et al., 2012) by In Fusion cloning (Clontech). DNA fragment of pCAG::myc-tagged CROCCP2-IRES-eGFP was then amplified and cloned into lentiviral plasmid backbone (gift from Cecile Charrier) using In Fusion cloning. A pCAG::3xHA-tagged CROCCP2-IRES-eGFP expression vector and lentiviral expression vector were also generated. All constructs were verified by DNA sequencing and western blot. The Flag-IFT20 sequence obtained from Addgene plasmid #118033 was cloned into pCAG::IRES-RFP expression plasmid. The RFP-tagged ARL13B vector is a kind gift of Prof Heiko Lickert (Kinzel et al., 2010). Primers used to create the constructs are summarized in Table S1.

The shRNA target sequences for IFT20 (5’-CCAAAGAAGCAGAGAACGA-3’ and 5’-TTTATTGACCAATTTATA-3’), MTOR (5'-AATGTTGACCAATGCTATGGA-3') and CROCCP2 (5'- TGCCATAAATGCTCAATAATTGT-3') were cloned downstream of the U6 promoter into the pSilencer2.1-CAG-Venus (pSCV2)-plasmid or pSCV3 plasmid for CROCCP2 as previously described (Tiberi et al., 2012). The efficiency of the IFT20 shRNAs were validated following immunostaining of primary culture of mouse cortical progenitors performed after ex utero electroporation using shRNA targeting mouse and that of the CROCCP2 shRNA by western blot using protein lysates of P19 cells overexpressing myc-tagged CROCCP2.

Transcriptome analyses

Bulk RNAseq analysis was as described in Suzuki et al. (2018). Publicly available primary and organoid samples from Pollen et al. (2019) were downloaded from GSE124299. CPM normalized cells were scaled after top 2000 variable genes were selected. Principle component analysis were performed and first 20 principal components were selected for non-linear dimensional reduction with t-stochastic neighbour embedding. Plotting tools of Scanpy were used for visual purposes (Wolf et al., 2018).

In situ hybridization

In situ hybridization was performed using digoxigenin-labeled RNA probes (DIG RNA labeling kit, Roche) and alkaline phosphatase revelation (NBT/BCIP kit #SK-5400; Vector) using PCR-amplified or above-mentioned plasmid templates as previously described. Sense probes were used as a negative control for each gene tested. Images were acquired with a Zeiss Axioplan 2 microscope and a Spot RT3 camera using the Spot 5.2 software. The probes used in this study are summarized in table S2.

2D Cortical differentiation of human ESC

Cortical differentiation from human ESC was performed using previously described protocol (Espuny-Camacho et al., 2013) with slight modifications. At differentiation day -8, the hES cells were dissociated using PluriSTEM Dispase-II (Merck)/ Collagenase (Thermo Fisher Scientific) solution and plated on Growth factor-reduced matrigel (Corning) in Essential 8 Flex medium (Thermo Fisher Scientific). At differentiation day -2, the cells were dissociated using Stem-Pro Accutase (Thermo Fisher Scientific) and plated at low confluency (5,000–10,000 cells/cm²) on hES qualified matrigel (Corning) using Essential 8 Flex medium supplemented with 10 μM ROCK inhibitor (Y-27632; Abcam). At differentiation day 0, the medium was changed to DDM (Gaspard et al., 2008) supplemented with B27 (Thermo Fisher Scientific) and 100 ng/mL Noggin (R&D systems), and the medium was replenished every day. After 16 days of differentiation, the medium was changed to DDM, supplemented with B27 (DDM/B27), and changed every day. At day 25, the progenitors were dissociated using Accutase and frozen using mFreSR (Stem Cell technologies). Cortical progenitors were thawed on matrigel-coated coverslips 12-well using DDM/B27 supplemented with 10 μM ROCK inhibitor and fixed after 6 days using 4% paraformaldehyde for 20 min at 4° or ice-cold methanol for 6 min and then rinsed 3 times with PBS before immunostaining.

3D (organoid) cortical differentiation of human ESC

The differentiation of human ESC into cortical organoids was performed using previously described protocol (Cederquist et al., 2019) with slight modifications. Briefly, feeder-free H9 human ES cells were dissociated to single cells and 12,000 cells/well were aggregated in low-attachment conical bottom 96-well plate in mTESR1 medium (Stem Cell technologies) with supplement, P/S, 10 μM ROCK inhibitor (Y-27632; Abcam), as well as the different lentiviruses at the appropriate dilutions. Cortical induction of the spheroids was started 2 days later (differentiation day 0) using E6 medium containing 100 nM LDN (Stem Cell technologies), 10 μM SB431542 (Stem Cell technologies), and 5 μM XAV939 (Stem Cell technologies). At day 7, embryonic bodies of each condition were pooled and embedded in Matrigel drops (16 organoids/cookie) and cultured in an organoid-based N2/B27 without vitamin A medium. Organoids were then removed from the Matrigel drops 7 days after embedding and placed on an orbital shaker in N2/B27 with vitamin A medium. Half medium was changed every other day until fixation of the organoids at day 30 or 60 using ice-cold 4 % paraformaldehyde for 24 h.

Lentiviral preparation and infection

HEK293T cells were transfected by packaging plasmids, psPAX2 (Addgene) and pMD2.G (Addgene), and the transfer vector using Xtreme gene 9 transfection reagent (Sigma-Aldrich). Two days after transfection, culture medium was collected and viral particles were enriched by filtration (Amicon Ultra-15 Centrifuge Filters, Merck). Titration of every batch of lentiviral preparation was estimated following transduction of HEK293T cells. For 2D cortical differentiations, the hES cell-derived cortical cells were infected by the lentiviral constructs at differentiation day 25. The culture medium was changed the next day and phenotypes were analyzed 6 days after infection for CROCC & CROCCP2 subcellular localisation. For infection of cortical organoids, lentiviral preparations were ultracentrifugated at 100,000 x g, 4°C for 2 h and pellets resuspended in sterile PBS in order to prevent cell toxicity. Mouse primary cortex cultures were infected on the day of infection and phenotypes were analyzed 72h later.

Cell-cycle labeling assay

Cell-cycle kinetic differences were assessed by labelling cortical progenitor cells using a nucleotide analog 5-ethynyl-2'-deoxyuridine (EdU; Thermo Fisher Scientific) in vivo following 150 μg EdU injection into the peritoneal cavity of pregnant mice 1 hour before the sacrifice of the embryos. Detection of EdU was performed using Click-iT EdU Alexa Fluor 555 Imaging Kit (Thermo Fisher Scientific).

Mouse cortex primary culture

Mouse embryonic dorsal cortices were dissected under binocular at E13.5 in ice-cold HBSS (Thermo Fisher) containing with 1% Penicillin/Streptomycin (Thermo Fisher), 0.1 M HEPES (Thermo Fisher). They were then dissociated in Trypsin (Thermo Fisher) supplemented with 25mM EDTA (Merck) and 50 mg DNASE I (Sigma-Aldrich) for 30 min at 37°. Tubes were shaken every 10 min. Cortices were centrifuged and neutralized twice with NeuroBasal medium supplemented with 1% Penicillin/Streptomycin, 1% Glutamax, 1% Sodium Pyruvate, 1x B27 with vitamin A, 1x N-2 supplement (all from Thermo Fisher) and 1 mM N-acetyl cysteine (Sigma-Aldrich). Cells were passed through 0.4 micron cell strainer (Falcon), and then counted. Cells were then infected with adequate lentivirus and plated at 400,000 cells/ 3.5 cm² on hES qualified-matrigel (Corning)-coated coverslips (VWR). For shRNA experiments, brain cortices were electroporated ex utero before being isolated and dissected as described above. In all experiments cells were cultivated for 72 h in culture medium as described above with addition of 10ng/mL human b-FGF (Peprotech) and culture medium was changed everyday. The cells were then rinsed with PBS and fixed either with 4% paraformaldehyde for 20 min at 4 °C or using a Glyoxal solution as previously described (Richter et al., 2018) before being washed 3 times in PBS.

En-face ventricular zone wholemount

Wholemount staining of en-face view of the ventricular zone was performed as described in Mirzadeh et al. (2010). Briefly, 24h after IUE, mouse embryos were collected and perfused transcardiacally with ice-cold 4% paraformaldehyde. Electroporated cortices were then dissected to completely expose the lateral wall and flatened out with the ventricle side up. Cortices were soaked in 4% paraformaldehyde overnight and then rinsed 3 times in PBS before overnight incubation in RIMS solution (Yang et al., 2014).

Immunofluorescence staining

The tissues were washed with PBS for 10 min, then with PBS/0.3% Triton X-100 for 30 min. They were then quenched in 4 mM glycine solution for 30 min, except for whole mount staings, then blocked in PBS containing 0.3% Triton X-100 and 3% horse serum, for at least 1 hour. The tissues were incubated overnight at 4°C with primary antibodies in blocking solution. After three washes in PBS/0.3% Triton X-100, slides were incubated during 2 hours at room temperature with secondary antibodies in PBS/0.3% Triton X-100. After washing in PBS/0.3% Triton X-100, the tissues were mounted on a slide glass with DAKO glycergel mounting medium (DAKO) complemented with the DABCO anti-fading agent.

Confocal microscopy

Confocal imaging was performed on a LSM780NLO confocal system fitted on an Observer Z1 inverted microscope (Zeiss) equipped with a Chameleon Vision II 690-1064 nm multiphoton laser (Coherent Europe). Fluorochromes were separated by linear unmixing using ZEN 2012 software (Zeiss). IUE were imaged using a Plan Apochromat 20x/0,8 dry objective, in vivo cilia imaging was done using LD C Apochromat 40x/1.1 water immersion objective and in vitro cilia imaging was done using LD C Apochromat 63x/1.4 oil immersion objective (Zeiss). Wholemount cortices were flattened into fluorodish (Ibidi) and covered with glass coverslip before image acquisition using the 63x objective on 4-5 Z-stacks spaced at 0.3 μm to cover the whole tight-junction region.Images were then processed using the Fiji/ImageJ software (Schindelin et al., 2012).

Sted images were acquired using a DMi8 inverted microscope (Leica) equipped with STED with 592nm (CW), 660nm (CW) and 775nm (pulsed) depletion lasers. HC PL APO 100x/1.4 oil objective was used. Acquisition was done with LAS X and deconvolution using Huygens Professional Version 17.1. For STED imaging, ABBERIOR secondary antibodies were used.

Quantification and statistical analysis

Statistical analysis

Results are shown as mean^±standard error (S.E.M.) of at least three biologically independent experiments or median with interquartile range when mean of experiments is presented. Student’s unpaired t-test was used for two group comparisons. Analyses of multiple groups were performed by a one-way analysis of variance (ANOVA) followed by post hoc Tukey’s test. For all tests, a P-value inferior to 0.05 was taken as statistically significant.

Published: November 4, 2022

Data and code availability

All data reported in this paper will be shared by the lead contact upon request.

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.