NR2F1 regulates regional progenitor dynamics in the mouse neocortex and cortical gyrification in BBSOAS patients

Abstract

The relationships between impaired cortical development and consequent malformations in neurodevelopmental disorders, as well as the genes implicated in these processes, are not fully elucidated to date. In this study, we report six novel cases of patients affected by BBSOAS (Boonstra-Bosch-Schaff optic atrophy syndrome), a newly emerging rare neurodevelopmental disorder, caused by loss-of-function mutations of the transcriptional regulator NR2F1. Young patients with NR2F1 haploinsufficiency display mild to moderate intellectual disability and show reproducible polymicrogyria-like brain malformations in the parietal and occipital cortex. Using a recently established BBSOAS mouse model, we found that Nr2f1 regionally controls long-term self-renewal of neural progenitor cells via modulation of cell cycle genes and key cortical development master genes, such as Pax6. In the human fetal cortex, distinct NR2F1 expression levels encompass gyri and sulci and correlate with local degrees of neurogenic activity. In addition, reduced NR2F1 levels in cerebral organoids affect neurogenesis and PAX6 expression. We propose NR2F1 as an area-specific regulator of mouse and human brain morphology and a novel causative gene of abnormal gyrification.

Keywords: NR2F1/COUP-TFI; BBSOAS; cell cycle dynamics; cortical folding; neurodevelopmental disease.

Figure 1. Cortical malformations in BBSOAS patients with NR2F1 haploinsufficiency

1. A–F

   Cortical malformations observed in six novel BBSOAS patients (See Fig EV1A–F for additional MRI panel and Table 1 for clinical features). Patient (P)1 shows abnormally elongated occipital convolutions (arrowheads in A), whereas abnormal gyrification of supramarginal and angular gyri is present in five out of six individuals (P2‐P6; B‐F). B′, C′, and C″ highlight the abnormally convoluted regions boxed in (B and C).

2. G

   Schematic representation of the human NR2F1 protein sequence (based on the current annotation: https://www.ncbi.nlm.nih.gov/protein/NP_005645.1), showing novel variants of a new BBSOAS patient cohort. Key amino acids for the functioning of the DNA‐binding domain (DBD) or the ligand‐binding domain (LBD) are listed in boxes.

3. H, H′

   TBR2 (green in H), SOX2 (red in H), and NR2F1 (yellow in H′) immunofluorescence (IF) of a GW11 section of human neocortex (see also Fig EV1G–G‴).

4. I

   NR2F1 expression (quantified by pixel intensity) at different L‐M levels (see also Fig EV1G–G‴) and in different regions along the A‐P extent of the cortex, as indicated. n ≥ 3 sections from n = 1 fetal brain.

5. J–L′

   NR2F1 (red) IF of human GW14 neocortex, showing expression levels in the CP (K, L) and VZ (K′, L′) around primary convolutions in the posterior‐most cortex (see also Fig EV1H). High NR2F1 is detected in the progenitor area of a gyrus (L′).

6. M, M′

   NR2F1 level as measured on single cells in the VZ (M) and CP (M′) of sulci and gyri, as in (J‐L′). n ≥ 4 sections from n = 1 fetal brain.

7. N, O

   NR2F1 (red) and TBR2 (green) immunostaining of the same regions shown in (J‐L′). NR2F1 level as measured in single TBR2⁺ IPs by pixel intensity is shown in graph (O). n ≥ 4 convolutions from n = 1 fetal brain (n = 2 sections).

8. P–R

   NR2F1 (red) and HOPX (green) immunostaining of the same regions shown in (J‐L′) (see also Fig EV1H). Virtually, all HOPX⁺ bRGs are also NR2F1⁺ (magnification in P″) and their number is greatly increased in gyri (P″) compared to sulci (P′), as quantified in (R). At a single cell level, NR2F1 pixel intensity is higher in oSVZ HOPX⁺ cells in gyri compared to sulci (Q). n ≥ 4 sections from n = 1 fetal brain.

Data information: Nuclei (blue) were stained with DAPI. In (I, R), the number of positive cells was quantified in 100 μm‐width boxes, randomly placed across the cortex. In graphs, data are represented as means ± SEM. Two‐way ANOVA (I) and Student′s t‐test (M, M′, O, Q, R) (*P < 0.05, ***P < 0.001). Scale bars: 50 μm. CP: cortical plate; iSVZ: inner subventricular zone; IZ: intermediate zone; oSVZ: outer subventricular zone; SVZ: subventricular zone; VZ: ventricular zone.

Figure EV1. MRI features of six novel BBSOAS patients and NR2F1 expression along the L‐M and A‐P axes

1. A

   Two‐year‐old female patient with thinning of the optic chiasm (left image) and optic nerves (right image).

2. B

   Four‐year‐old male patient showing prominent occipital convolutions (left) together with thinning of the posterior half of corpus callosum (right).

3. C

   Six‐year‐old female patient with thinning of the posterior half of corpus callosum (left) and of the optic chiasm (right).

4. D

   Three‐year‐old female patient with small asymmetry of the occipital cortex (left), thinning of the posterior half of corpus callosum (right), and normal thickness of optic nerves (not shown).

5. E

   Seven‐year‐old female patient with thinning of the optic chiasm (left) and mild thinning of posterior corpus callosum (right).

6. F

   Twelve‐year‐old male patient showing mild thinning of the optic nerves (left) and of the posterior half of corpus callosum (right).

7. G–G‴

   Low‐magnification DAPI staining and schematic representation of a GW11 coronal section (G) displaying the position of high‐magnification views in (G′‐G‴) and in Fig 1H and H′. Colors in (G) display lateral, dorsal, and medial areas taken for NR2F1 pixel intensity measurements at GW11 (see Fig 1I). Representative images of NR2F1 (red) immunofluorescence (IF) at different levels along the latero‐medial (L‐M) extent of the cortex, as shown in (G′‐G″).

8. H

   Low‐magnification DAPI staining of a GW14 sagittal section displaying the position of high‐magnification images showed in (I‐I″, K‐K″) and in Fig 1J–L′, N, N′ and P–P″.

9. I, J

   NR2F1 (red) IF at GW14 depicting increasing expression levels in the posterior‐most cortex (I″), as quantified in (J). n ≥ 4 sections from n = 1 fetal brain.

10. K–L″

    TBR2 (green) and SOX2 (red) IF of a GW14 primary convolution (detail taken from H). The average number of different NP classes in the fissure (K′) and convoluted region (K″) is quantified in (L‐L″). The convoluted region is populated by a higher number of both apical and basal progenitors. n ≥ 4 sections from n = 1 fetal brain.

11. M

    Schematic representation of human NR2F1 expression levels correlated with distinct cytoarchitectures along cortical convolutions. Expression of NR2F1 displays local modules of low versus medium/high expression levels in sulci and gyri, respectively. Medium/high NR2F1 expression in gyri is associated with high basal RG numbers and high neurogenic activity promoting radial and tangential expansion, whereas low expression is correlated with a small progenitor pool and low neurogenic potential.

Data information: Nuclei (blue) were stained with DAPI. In (J, L‐L″), the number of positive cells or the pixel intensity was quantified in 100 μm‐width boxes, randomly placed across the cortex in exemplificative regions, as shown in (H). Data are represented as means ± SEM. Two‐way ANOVA (J) and Student′s t‐test (L‐L″) (**P < 0.01, ***P < 0.001). Scale bars: 50 μm. In (A‐F), arrowheads highlight the morphological features described above the MRI for each patient. CP: cortical plate; IZ: intermediate zone; iSVZ/oSVZ: inner/outer subventricular zone; VZ: ventricular zone.

Figure 2. Nr2f1 regulates neural progenitor self‐renewal in a neurosphere assay

1. A

   Representative images of wild‐type (WT) and mutant (KO) neurospheres obtained from E15.5 neocortices and cultured in vitro. The culture passage (“step”) corresponds to 3 days of culture after dissociation and re‐plating. KO aggregates are more numerous and larger in size.

2. B

   Graph showing the number of neurospheres per P100 cell plate at different steps, as indicated. Nr2f1 KO neurospheres proliferate for a longer time (up to 30 passages tested), whereas WT cells exhaust around step 9. n ≥ 3 culture plates from n = 2 batches.

3. C

   Panel of WT and KO isolated neurospheres grown from day 1 to day 10. Blue and red lines represent the diameter of WT and KO neurospheres, respectively.

4. D, E

   PH3 (green; dividing cells) IF on cross‐sections of day 4 neurospheres. White arrowheads point at mitotic figures, which are quantified in (E) as percentage of dividing cells over DAPI‐stained nuclei (blue). n ≥ 3 neurospheres from n = 2 batches.

Data information: In (B, E), data are represented as means ± SEM. Two‐way ANOVA (B) and Student's t‐test (E) (**P < 0.01, ***P < 0.001). Scale bars: 50 μm. PH3: Phosphohistone H3.

Figure 3. Nr2f1‐mediated control of progenitor amplification in vivo

1. A–D

   Ki67 (green; mitotically active progenitors) and Tuj1 (red; post‐mitotic neurons) IF at E14.5 (A, A′), E16.5 (B, B′), and E18.5 (C, C′) of lateral pallia of wild‐type (WT) versus mutant (KO) brains. Ki67⁺ proliferating cells (quantified in D) accumulate in mutant cortices. n ≥ 3 brains.

2. E–E′

   Tbr2 (green; IPs) and Pax6 (red; radial glia cells) IF of E13.5 (E, E′) and E17.5 (F, F′) lateral pallia of WT and KO brains. See Fig EV3 for Tbr2 Pax6 double staining in E17.5 HET embryos.

3. G–I

   Quantification of different classes of cortical progenitors: apical radial glia cells (aRGs) as Pax6⁺ NPs in VZ (G), basal intermediate progenitors (IPs) as Tbr2⁺ cells in SVZ (H), and basal RGs as Pax6⁺ cells in outer SVZ (I). See Fig EV3 for quantification of distinct progenitor classes in E17.5 HET embryos. n ≥ 3 brains.

4. J–L

   E12.5 dorsal brain views (J, J′) and representative sections of posterior hemispheres (K, K′), showing a vesicle enlargement in mutant brains (K′). The extension of the ventricular surface was quantified along the A‐P axis (L); red color code represents Nr2f1 gradient in WT brains. 12 μm‐thick sections were collected on series of 10 slides; consecutive measurement levels are 120 μm apart from each other. n ≥ 3 brains.

5. M

   Schematic representation of different aRG cell divisions, as visualized in vivo after co‐in utero electroporation (IUE) of Tis21‐RFP and Sox2‐GFP plasmids into E12.5‐old embryos. Green: symmetric proliferative divisions; yellow: asymmetric differentiative divisions; and red: symmetric differentiative divisions.

6. N–P

   GFP (green; expressed under Sox2 promoter) and RFP (red; expressed under Tis21 promoter) IF of E13.5 lateral pallia, 18 h after IUE. The proportion of single‐ or double‐positive NPs is shown in pie charts in (P). n = 3 electroporated brains.

7. Q–R′

   Ki67 (green; progenitors) and EdU (red) IF of E13.5 WT (Q, Q′) and KO (R, R′) embryos, injected with EdU at E12.5. Differentiating cells (EdU⁺Ki67⁻) are located in the IZ/CP; percentages shown in (Q′, R′) were obtained from n ≥ 4 sections of n ≥ 2 brains. See Fig EV3 for neural differentiation index of E13.5 HET embryos.

Data information: Nuclei (blue) were stained with DAPI. In (D, G, H, I), the number of positive cells was quantified in 100 μm‐width boxes, randomly placed across the LP. In graphs, data are represented as means ± SEM. Student's t‐test (Q′, R′; E13.5, EdU injected at E12.5; WT/KO: *P = 0.03967) and two‐way ANOVA (D, G, H, I, L, P) (*P < 0.05, **P < 0.01, ***P < 0.001). Scale bars: 50 μm. CP: cortical plate; IZ: intermediate zone; SVZ: subventricular zone; VZ: ventricular zone.

Figure EV2. Molecular identity of basal (b) RGs in Nr2f1 mutant cortices

1. A–F

   Tbr2 (green), Sox2 (red), and Ki67 (blue) triple IF of WT (A, C, E) and KO (B, D, F) neocortices at E16.5 (A–D) or at E18.5 (E, F) showing abundant NPs in mutant brains. bRGs located out of the Tbr2⁺ SVZ layer are very abundant in the mutant lateral pallium (LP; arrowheads in B).

2. G

   Average number of NPs with distinct molecular profiles in the LPs of E18.5 control and mutant embryos. Cells were evaluated separately in the VZ‐SVZ (left) and in the IZ basal to the Tbr2⁺ SVZ layer (hereinafter oSVZ; right). A high number of Ki67⁺Sox2⁺Tbr2⁻ NPs are found in the KO oSVZ, consistent with bRG identity. n ≥ 4 sections from n = 2 brains.

3. H–J

   Tbr2 (green), Pax6 (red), and Phospho‐Vimentin (blue) triple IF of E17.5 WT (H, I) and KO (J) LPs. The typical morphology of a bRG cell is shown in (H) and magnified in (H′); gray arrowheads point to the apical process, and empty arrowhead highlights the Pax6⁺ bRG soma. Note a high number of P‐Vim⁺ cells in the oSVZ of KO embryos (J; arrowheads point to P‐Vim⁺ cells).

4. K

   Average number of dividing (P‐Vim⁺) cells in the LPs of E18.5 WT and KO embryos. Cells were evaluated separately in the VZ‐SVZ (left) and oSVZ (right). Mutant brains have a higher number of P‐Vim⁺ cells expressing Pax6 or both Pax6 and Tbr2, and located in the VZ‐SVZ or in the oSVZ layer, respectively. n ≥ 4 sections from n = 2 brains.

Data information: In (G, K), the number of positive cells was quantified in 200 μm‐ or 400 μm‐width boxes, as indicated, randomly placed across the LP. In graphs, data are represented as means ± SEM. Two‐way ANOVA (**P < 0.01, ***P < 0.001). Scale bars: 50 μm. CP: cortical plate; IZ: intermediate zone; SVZ: subventricular zone; VZ: ventricular zone.

Figure EV3. Nr2f1 dose‐dependent control of progenitor amplification and differentiation in vivo

1. A–C

   Tbr2 (green; IPs) and Pax6 (red; radial glia cells) IF of E17.5 lateral pallia of WT (A, A′), HET (B) and KO (C) brains.

2. D–F

   Quantification of different classes of cortical progenitors in distinct genotypes as indicated; apical radial glia cells (aRGs) were quantified as Pax6⁺ NPs in VZ (D), basal intermediate progenitors (IPs) as Tbr2⁺ cells in SVZ (E), and basal RGs as Pax6⁺ cells in outer SVZ (F). n ≥ 2 brains.

3. G–I′

   Ki67 (green; progenitors) and EdU (red) IF of E13.5 WT (G‐G″), HET (H, H′), and KO (I, I′) embryos, injected with EdU at E12.5. Differentiating cells (EdU⁺Ki67⁻) are located in the intermediate zone (IZ) or in the cortical plate (CP).

4. J

   Graph showing the average number of EdU⁺Ki67⁻ differentiating cells in WT (blue column), HET (gray column), and KO (orange column) cortices quantified 24 h after EdU injection. Delayed at E13.5 in KO brains, neurogenesis shows a decreasing trend also in HET brains. The number of EdU⁺Ki67⁻ differentiating cells was re‐normalized on total number of Ki67⁺ cells, to factor‐in the increased size of progenitor pool in mutant brains compared to WT. n ≥ 4 sections from n = 2 brains.

Data information: Nuclei (blue) were stained with DAPI. The number of positive cells was quantified in 100 μm‐width boxes, randomly placed across the LP. Data are represented as means ± SEM. Two‐way ANOVA (n.s.=not significant, P > 0.05; *P < 0.05, **P < 0.01, ***P < 0.001). Scale bars: 50 μm. CP: cortical plate; IZ: intermediate zone; SVZ: subventricular zone; VZ: ventricular zone.

Figure EV4. Delayed neurogenesis and posterior cortical thickening in Nr2f1‐deficient mice

1. A

   Graph showing the average number of EdU⁺Ki67⁻ differentiating cells in WT (blue columns) and KO (orange columns) cortices quantified 24 h after EdU injection. Delayed at E13.5, neurogenesis in KO brains increases significantly from E16.5 onwards. n ≥ 4 sections from n = 2 brains.

2. B–C′

   Ki67 (green), EdU (red), and Satb2 (blue; cortical callosal neurons) triple IF of E16.5 WT (B, B′) and mutant (C, C′) embryos, injected 24 h before with EdU. Differentiating cells (EdU⁺Satb2⁺Ki67⁻) can be recognized as they migrate toward the CP.

3. D

   Short‐term EdU injection as in (A), after re‐normalization on the total number of EdU⁺ cells, to take account of the increased size of progenitor pool in mutant brains compared to WT. n ≥ 4 sections from n = 2 brains.

4. E

   Average number of EdU⁺ cells in the LP of control (blue columns) and mutant (orange columns) embryos at P0 upon EdU injection at the indicated stages. Mutant cortices are characterized by a higher density of differentiating neurons at mid‐late corticogenesis. n ≥ 4 sections from n = 2 brains.

5. F–F″

   Ctip2 (green), Tbr1 (blue), and EdU (red; injected at E12.5) triple IF of P0 WT (F) and KO (F′) brains showing a high number of E12.5‐born Tbr1⁺EdU⁺ neurons in lower layers of mutant cortices. Graph in (F″) quantifies the average number of EdU⁺ neurons expressing distinct laminar markers (Tbr1 and Ctip2) in WT and KO embryos injected with EdU at E12.5. n ≥ 4 sections from n = 2 brains.

6. G–G″

   Cux1 (green) and EdU (red; injected at E15.5) double IF of P0 WT (G) and KO (G′) brains showing a high number of E15.5‐born Cux1⁺EdU⁺ neurons in superficial layers of mutant cortices. Quantification of E15.5‐born Cux1⁺ superficial neurons in WT and KO embryos is shown in (G″). n ≥ 4 sections from n = 2 brains.

7. H–J

   NeuN (green; neurons), EdU (red), and GFAP (blue; glia) triple IF of WT (H, H′) and KO (I, I′) animals injected at E17.5 and analyzed at P8. The average number of EdU⁺NeuN⁺ neurons or of EdU⁺GFAP⁺ glial cells is quantified in (J). n ≥ 4 sections from n = 2 brains.

8. K, K′

   Representative images of P0 WT (K) and KO (K′) brains, showing slight elongation of posterior hemispheres (compare red lines with blue ones).

9. L, M

   Tuj1 (red) IF of WT (L) and KO (L′) cortices. Note the increased thickness of KO IZ and CP compared to WT ones. Graph in (L″) indicates CP thickness measured at E17.5 and P0 in the posterior cortical region of control (blue) and mutant (orange) animals. The average number of Tuj1⁺ differentiating neurons in posterior‐most CP is quantified from E12.5 to P5 in (M). L″: n = 2 brains per genotype/age; M: n ≥ 4 sections from n = 2 brains.

10. N–Q

    Ctip2 (green), Tbr1 (blue), and Satb2 (red) triple IF of WT (N,N′,P) and KO (O,O′,P′) posterior cortices at P0 and P8, as indicated, and quantified in (Q). n ≥ 4 sections from n = 2 brains.

Data information: Nuclei (blue) were stained with DAPI. Data are represented as means ± SEM. Two‐way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001). Scale bars: 50 μm. CP: cortical plate; IZ: intermediate zone; SVZ: subventricular zone; VZ: ventricular zone.

Figure EV5. Regional control of cell cycle dynamics

1. A, B

   Tbr2 (green; IPs) and PH3 (red; mitotic figures) double IF of E17.5 WT and KO posterior lateral pallia (LPs). Arrowheads in (A, B) point to PH3⁺ mitotic NPs.

2. C

   Average number of dividing cells in the LP of control (WT) and mutant (KO) animals in different regions along the A‐P axis, as indicated. E17.5 was chosen as a developmental age displaying abundant basal progenitor populations. Values of dividing aRGs, IPs, and bRGs are indicated in orange, yellow, and green, respectively. n ≥ 3 brains.

3. D

   Experimental set‐up for cell cycle time quantification by double EdU/BrdU injection (adapted from Martynoga et al, 2005). Ti corresponds to an interval time between EdU and BrdU injections. For detailed information, see Methods in the Appendix File.

4. E–E″

   BrdU (green) and EdU (red) double IF of WT, HET, and KO E12.5 ventricular zone of LPs. Single EdU⁺ cells (leaving fraction) that exit the S‐Phase during the 2‐h injection period move within the VZ surface (red arrowheads), while double EdU/BrdU label S‐phase cells reside in a more basal VZ position, next to the SVZ.

5. F, F′

   Cell cycle duration along the A‐P axis of E12.5 (F) and E14.5 (F′) LP in mutant (KO) embryos and their control (WT) littermates. Cell cycle is accelerated (i.e., has a shorter duration) in NPs of the posterior‐most region of the mutant cortex, where Nr2f1 is expressed at high levels in control littermates. n ≥ 3 brains per age/genotype. Cell cycle time has been expressed in arbitrary units (A.U.) upon normalization on the WT Ant duration.

6. G

   Cell cycle duration of E14.5 WT, HET, and KO LPs, showing a decreasing trend of cell cycle duration in HET animals and a significant decrease in KO ones compared to WT controls. n ≥ 2 brains.

Data information: Nuclei (blue) were stained with DAPI. In (C, F‐G), the number of positive cells was quantified in 100 μm‐width boxes, randomly placed across the LP. In graphs, data are represented as means ± SEM. Two‐way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001). Scale bars: 50 μm. Ant: anterior; CP: cortical plate; IZ: intermediate zone; Med: medial; Post: posterior; SVZ: subventricular zone; VZ: ventricular zone.

Figure 4. Nr2f1 orchestrates cell cycle dynamics in posterior hemispheres

1. A–D

   Graphs showing the average number of PH3⁺ mitotic figures in the posterior lateral pallium of WT and mutant cortices. The total number of mitoses (A) as well as the number of dividing cells in VZ (Pax6⁺Sox2⁺ aRGs; B), SVZ (Tbr2⁺ IPs; C), and outer (basal‐most) region of SVZ (oSVZ/IZ; Pax6⁺ bRGs; D) is shown. For graphs (B‐D), statistical analysis by two‐way ANOVA is shown in Appendix Table S5. n ≥ 3 brains per age/genotype.

2. E

   Cell cycle duration at different embryonic ages (from E10.5 to E16.5) in WT (blue bars) or KO embryos (orange line), as quantified upon a double EdU/BrdU injection protocol in posterior‐most LP (See Fig EV5D–E″ and Materials and Methods). n ≥ 3 brains per age/genotype.

3. F–G′

   Ki67 (green; NPs) and EdU (red) IF in WT (F,F′) and KO (G,G′) after 6 h consecutive EdU incorporation. Note the higher number of labeled NPs in KO cortices (G′) suggesting faster cell cycle progression and/or longer S‐phase.

4. H

   Best linear fit of EdU⁺Ki67⁺ NP percentage after 8 h EdU cumulative labeling in WT (blue line) and KO (orange line) brains. The x‐ and y‐intercepts are proportional to the S‐phase length and to the number of cycling cells, respectively, while the steepness of the line is proportional to cell cycle duration (see Materials and Methods; WT: y = 11.66x + 24.62; KO: y = 13.54x + 32.96). n ≥ 2 brains per time point/genotype.

5. I–K

   PH3 (green; dividing NPs) and Pax6 (red; aRGs) IF of E13.5 WT (I) and KO (J) lateral pallia. Strong and weak/punctate PH3 patterns show M‐phase and G2‐phase cells, respectively. G2‐phase NP percentage is quantified in (K). As a result of shorter G1‐phase and globally shorter cell cycle time, the percentage of G2‐phase cells is increased in the KO NP pool. n ≥ 2 brains.

6. L

   Cell cycle time of E13.5 NPs in the lateral pallia of WT (upper column) and KO cells (lower column), as calculated by EdU cumulative labeling and PH3 staining (see Materials and Methods). KO NPs have a 16.1% faster cell cycle time, due to a 30.9% shorter G1‐phase compensating a 14.3% longer S‐phase. n ≥ 3 brains.

Data information: Nuclei (blue) were stained with DAPI. In (A‐E, H, K), the number of positive cells was quantified in 100 μm‐width boxes, randomly placed across the lateral pallium. In graphs, data are represented as means ± SEM. Student's t‐test (K; *P < 0.05) and two‐way ANOVA (A‐E, H, L) (*P < 0.05, **P < 0.01, ***P < 0.001). Scale bars: 50 μm. SVZ: subventricular zone; VZ: ventricular zone.

Figure 5. Nr2f1‐mediated molecular control of neural progenitors (NPs)

1. A–B′

   Pax6 (green; aRGs) and Nr2f1 (red) IF of E12.5 WT (A, A′) and KO (B, B′) lateral pallia. See Appendix Fig S4 for Pax6 and Nr2f1 staining and pixel intensity quantification in HET embryos.

2. C

   Pax6 and Nr2f1 pixel intensity quantification at E12.5 and E14.5 indicating increased Pax6 levels upon Nr2f1 removal. n ≥ 3 brains.

3. D

   Real‐time RT–PCR quantification of Nr2f1 and Pax6 expression in E12.5 cortices. n ≥ 3 cortices.

4. E–G″

   GFP (green; electroporated cells), Pax6 (red), and Nr2f1 (blue) IF of pCIG2‐Nr2f1‐IRES‐GFP electroporated brains at E13.5 (24 h after IUE). VZ (F‐F″) and CP (G‐G″) regions are shown at high magnification. Arrowheads point to GFP⁺ electroporated cells overexpressing Nr2f1 (blue in F′, G′), down‐regulating Pax6 (F″), and rapidly migrating out of the VZ (G‐G″).

5. H

   Scatter plot showing Pax6 and Nr2f1 pixel intensity of VZ electroporated progenitors, comparing Nr2f1 overexpressing cells (orange dots) with control pCIG2‐GFP electroporated cells (blue dots). Average Nr2f1 and Pax6 pixel intensities of the 2 populations were compared by two‐way ANOVA and resulted significantly different (***P = < 0.0001). n = 2 electroporated brains.

6. I

   Real‐time RT–PCR of cell cycle genes comparing WT (blue line) and KO (orange line) cortices. CyclinD1 and P21 transcripts are down‐regulated, whereas Dct is up‐regulated in mutants. n = 3 cortices.

7. J–L

   GFP (green; electroporated cells) and P21 (red) IF of E14.5 cortex electroporated 48 h earlier with control PX458 plasmid (J) or CRISPR/Cas9‐expressing plasmid directed against Nr2f1 sequence (PX458‐αNr2f1; K). Percentage of P21/GFP double‐positive cells (L). Arrowheads in (J, K) point to P21⁺ cells. n = 2 electroporated brains.

8. M

   Schematic model of Nr2f1 action on NP cell cycle and neural differentiation. Nr2f1 promotes neurogenesis by repressing Pax6 and Dct expression and thus cell cycle progression, and by activating P21‐mediated cell cycle exit. Both actions modulate the G1‐phase length.

9. N

   Graph showing the number of differentially expressed genes (DEG; see Materials and Methods) up‐regulated (UP) or down‐regulated (DOWN) in KO compared to WT, as detected by RNA‐Seq of E15.5 neocortices.

10. O

    Gene ontology (GO) categories significantly enriched among DEG genes (DAVID Gene Ontology software; see Materials and Methods).

11. P

    Hierarchical clustering and heatmap of the expression level of DEGs belonging to the “neuron differentiation” GO category in (O). Heat map color scale indicates normalized gene expression from high (red) to low (white) level.

Data information: In (A, B, J, K), nuclei (blue) were stained with DAPI. In (C), the pixel intensity was quantified in 100 μm‐width boxes, randomly placed across the lateral pallium. Data are represented as means ± SEM. Student's t‐test (L; *P < 0.05) and two‐way ANOVA (C, D, H, I) (*P < 0.05, **P < 0.01, ***P < 0.001). For statistical analysis of RNA‐Seq data (N‐P), see Materials and Methods. Scale bars: 50 μm. SVZ: subventricular zone; VZ: ventricular zone.

Figure 6. Functional rescue of NP cell cycle dynamics via Pax6 genetic modulation

1. A

   Real‐time RT–PCR analysis of Pax6 expression in neurospheres with different genotypes, as indicated. n = 3 cortices. N KO, P HET: Nr2f1 KO, Pax6 HET.

2. B

   Representative images of Nr2f1 KO Pax6 HET neurospheres obtained from E15.5 neocortices and cultured in vitro.

3. C

   Graph showing the number of neurospheres per P100 cell plate, at different steps, as indicated. While Nr2f1 KO neurospheres (orange line) proliferate for a longer time compared to WT ones (blue line), the loss of one Pax6 allele (gray line) almost restores normal proliferation rate and exhaustion time. Complete Pax6 loss is not compatible with stem cell renewal (yellow line). n ≥ 3 culture wells from n = 2 independent batches.

4. D, E

   Paired‐cell analysis after neurosphere dissociation and 24‐h culture at clonal density. Couples of dividing cells were identified by Map2 (red; N, neurons) and BLBP (green; P, progenitors) IF. Pie charts in (E) show the proportion of P‐P (proliferative; D), P‐N (asymmetric differentiative; D′), or N‐N (symmetric differentiative; D″) couples in WT, Nr2f1KO, and Nr2f1KO Pax6 HET animals. n ≥ 3 samples from n = 2 culture batches.

5. F–I

   Pax6 (green) and Nr2f1 (red) IF in the posterior‐most region of E12.5 cortex, showing Pax6 upregulation in Nr2f1 KO animals (G, G′) compared to WT (F′, F″). Normal Pax6 levels can be restored by loss of one Pax6 allele (Nr2f1 KO, Pax6 HET; H, H′). Pixel intensity quantification of Nr2f1 (blue) and Pax6 (orange) is shown in (I). n ≥ 3 brains. N KO: Nr2f1 KO; P HET: Pax6 HET.

6. J

   Cell cycle duration at E12.5 in different mutants as indicated, quantified in posterior regions with double EdU/BrdU injection protocol (see Materials and Methods). n ≥ 6 sections from n = 2 brains.

7. K

   Insets showing representative γ‐Tubulin (green) and phospho‐Vimentin (red) IF of the cortical VZ surface to evaluate the orientation of the cleavage plane of dividing aRGs. Pie charts show the percentage of vertical (green), oblique (yellow), and horizontal (red) division planes of mitotic figures in the E12.5 lateral pallium of animals with different genotypes, as indicated. n ≥ 2 brains per genotype.

Data information: Nuclei (blue) were stained with DAPI. In (I, J), the pixel intensity or the number of positive cells was quantified in 100 μm‐width boxes, randomly placed across the LP. Data are represented as means ± SEM. Two‐way ANOVA (A, C, E, I, J, K; *P < 0.05, **P < 0.01, ***P < 0.001). For statistical analysis of (E, K), see Appendix Table S5. Scale bars: 50 μm.

Figure 7. Delayed neurogenesis upon NR2F1 down‐regulation in human brain organoids

1. A–C

   NR2F1 expression in cerebral organoids during neural induction and differentiation in vitro (day 40 and day 70, respectively), together with the NP markers PAX6 (green in A), SOX2 (green in B, B′), and DCX as a neural marker (blue in A, B, B″). Quantification of NR2F1 levels in NPs is shown in (C). n ≥ 4 organoids from n = 2 batches.

2. D–G

   NR2F1 (red) and PAX6 (green) IF in day 40 organoids. Neuroepithelia with high NR2F1 levels show weak PAX6 expression (E, E′), and vice versa (F, F′). Quantification by pixel intensity analysis (G). n ≥ 4 organoids from n = 2 batches.

3. H

   Schematic representation of an organoid electroporation. PX458 plasmids were injected in ventricular‐like cavities, and organoids were electroporated and processed after 7 days of in vitro culture.

4. I, J

   NR2F1 (red) and GFP (green) IF in day 40 human brain organoids upon electroporation of the PX458‐αNR2F1 plasmid. In (I, I′), GFP⁺NR2F1⁻ cells (empty arrowheads) and a GFP⁺NR2F1⁺ cell (white arrowhead) are shown; quantification in (J). n ≥ 6 organoids from n = 2 batches.

5. K–M

   TUJ1 (red; differentiating neurons) and GFP (green) IF 7 days after electroporation of control (K) and PX458‐αNr2f1 plasmids (L). Co‐expression of GFP with SOX2 or TUJ1 distinguishes neural progenitors (NPs) from neurons (Ns), as quantified in (M). Arrowheads in (K, L) point to TUJ1⁺ GFP⁺ cells. n ≥ 6 organoids from n = 2 batches.

6. N

   Red and green arrowheads in (N) point to a SOX2⁺ (red) progenitor and a TUJ1⁺ (blue) neuron, respectively.

Data information: Nuclei (blue in D, I, K, L) stained with DAPI. In (C, G, J), the pixel intensity or the number of positive cells was quantified in 100 μm‐width boxes, randomly placed across the cortex/organoid neuroepithelia, while in (M) the number of positive cells was normalized over the total number of GFP⁺ cells. In graphs, data are represented as means ± SEM. Two‐way ANOVA test (*P < 0.05, **P < 0.01, ***P < 0.001). Scale bars: 50 μm.

Figure 8. Schematic representation of cortical cellular and morphological consequences upon Nr2f1 loss

1. Nr2f1 is expressed along an anteromedial‐low to posterolateral‐high gradient (red color code) in the developing mouse cortex, spanning from VZ aRGs to SVZ IPs, bRGs, and CP neurons. Upon Nr2f1 removal (KO), the NP pool of the posterior cortex expands leading to occipital macrocephaly.

2. Early to late effects on progenitor and neuron behavior upon Nr2f1 deficiency. Early in development (E10.5‐E12‐5), Nr2f1 loss causes cell cycle acceleration, increased NP self‐renewal, and sustained Pax6 expression. Neurogenesis (as well as the expression of neurogenetic factors such as Tbr2 and P21) is delayed, thus allowing amplification of the progenitor pool and lateral expansion of the posterior hemispheres. At later stages (E13.5‐E14.5), neurogenesis comes into play and neurons are produced at high rate. Persistent Pax6 expression in the expanded NP population leads to abundant basal RG production at late time points (E16.5‐E18.5). High neuronal output results in radial expansion and generation of a thick posterior cortex.