PHF6-mediated transcriptional control of NSC via Ephrin receptors is impaired in the intellectual disability syndrome BFLS

To begin to examine the function of PHF6 as a transcriptional regulator in the embryonic brain, we performed ChIP-seq analysis of PHF6 in the developing cortex of mouse embryos at embryonic day 17 to 18 (E17-18). We identified 2467 PHF6 binding sites at P-value < 10^–5 (Dataset EV1, Figs. 1 and EV1). These binding sites occurred in various genomic regions, including the proximal region of transcription start sites (TSS’), gene bodies, and intergenic regions (Fig. 1A). Compared to what would be expected from the random distribution of binding sites across the genome, we observed significant enrichment upstream of the TSS as well as in the 5’ untranslated region (UTR) of protein-coding genes (Fig. 1A). Particularly, PHF6 sites are strongly enriched in the 1 kb region around the TSS, with the highest density immediately downstream of the TSS (Fig. 1B). This pattern suggests a role of PHF6 in regulating gene expression.

Follow-up analysis revealed that PHF6-bound regions significantly overlap (CA)_n-microsatellite repeats, as revealed by motif analysis of the top 1000 PHF6 sites (Figs. 1C and EV1B). These microsatellites are specifically located at the centre of PHF6 sites (Figs. 1D and EV1B), suggesting that they are associated with PHF6 binding. Among the top 1000 PHF6 peaks, 609 overlap a (CA)_n repeat on either DNA strand. In comparison, we observed an overlap of only 67 between (CA)_n repeats and shuffled peak coordinates (Fisher’s exact test P < 2.2e–16) (Dataset EV2). An unbiased analysis of the distribution of all genomic (CA)_n repeats revealed that they are largely enriched near genes involved in developmental processes, including central nervous system development, neurogenesis, and neuron differentiation (Fig. 1E, Dataset EV3). Function enrichment analysis of PHF6 sites also revealed the same trend (Dataset EV4), with many Gene Ontology (GO) terms such as forebrain development and regulation of neurogenesis commonly found among the most enriched terms for both PHF6 sites and (CA)_n microsatellite repeats (Fig. 1F, Dataset EV5). These results suggest that (CA)_n repeats are specifically enriched near neural development genes and are bound by PHF6.

Next, we profiled the genome-wide pattern of gene deregulation by analysis of Phf6 knockdown (KD) and control cortical progenitors following their isolation at embryonic day 14 (E14) and expansion for 5 days in culture. RNA-seq analysis (Dataset EV6) revealed that PHF6 functions as a transcriptional activator or repressor (Fig. 2A). In addition, enrichment analysis, performed separately on upregulated and downregulated genes, revealed that a large panel of genes involved in nervous system development are downregulated in the Phf6 KD group (FDR < 0.02) (Fig. 2B). A number of significant PHF6-differentially expressed genes were found to have peaks within the +/−2 kb vicinity of the TSS (Fig. 2C,D).

To further understand the role of PHF6 in the regulation of transcription, we employed Pol II occupancy data (Liu et al, 2017), and examined the association between PHF6 binding and Pol II occupancy in neural progenitor cells. Interestingly, we observed that TSS’ with a PHF6 site within 300 bp tend to be depleted of Pol II, compared to genes with a PHF6 site between 300–1000 bp of the TSS (Fig. 2E). This pattern suggests that the binding of PHF6 within the immediate vicinity of TSS might have a negative effect on the recruitment of Pol II to the TSS. To examine this prediction, we analyzed the association of PHF6 peaks and PHF6 differentially regulated genes. We found that PHF6 inhibition led to an overall increase in the expression of 65% of genes with a PHF6 site at or immediately downstream of the TSS (Fig. 2D,F). These observations suggest a position-dependent role for PHF6 in regulating transcription, which may provide mechanistic insight into the dual role of PHF6 as a transcriptional activator and repressor.

Our data on functional annotation of PHF6 binding sites suggest that PHF6 regulates neurogenesis (Fig. 1F). Interestingly, previous studies in hematopoietic stem cells (HSCs) showed that PHF6 can restrict the self-renewal capacity of HSCs (McRae et al, 2019; Miyagi et al, 2019). These findings led us to investigate whether PHF6 regulates cell proliferation or self-renewal. To begin with, we subjected PHF6-GFP or control GFP-expressing neuroblastoma (N2A) cell lines to KI67 staining and found that PHF6 significantly suppressed the proliferation of these cells (Fig EV2A–C). Next, we induced the KD of Phf6 via a pool of siRNA in primary E14 eNSC cultures followed by limiting dilution assay (LDA). Compared with eNSCs transfected with non-targeting siRNA control, we found a significant increase in eNSC neurosphere numbers upon KD of Phf6 (Fig. 3A,B). Consistent with this observation, immunoblotting analyses of neurospheres following 7 days in culture, revealed upregulation of stem cell markers, SOX2 and NESTIN in Phf6 KD relative to the control cells (Fig. 3C). Our data suggest that PHF6 restricts stem cell self-renewal in primary neurosphere cultures.

We next set out to characterize the role of PHF6 in stemness using a genetic mouse model in which we induced genetic deletion of Phf6 via breeding Phf6^loxP/loxP with Nestin-CreERT2⁺ mice followed by tamoxifen administration at E14 for 24–48 h to delete Phf6 exons 4 and 5 in the Nestin expressing cells (Fig. 3D). Extreme limiting dilution assay (ELDA) (Hu and Smyth, 2009; Rasool et al, 2022) and LDA analyses of eNSCs obtained from Phf6^-/Y / Nestin-CreERT2⁺ (Phf6 KO) and Phf6^loxP/Y / Nestin-CreERT2^- (control) mice revealed a significant increase in self-renewal (Fig. 3E), sphere number (Fig. 3F), and sphere diameter (Fig. 3G,H) in Phf6 KO eNSCs. Importantly, a significant increase in the expression of the stemness markers, Nestin and Sox2, (Fig. 3I) and an increase in EdU incorporation (Fig. 3J) was observed in the eNSCs of Phf6^-/Y / Nestin-CreERT2⁺ mice. Our data shows that the genetic deletion of Phf6 promotes the self-renewal of eNSCs, suggesting that PHF6 loss-of-function may restrict eNSC commitment to differentiated progenies in the developing brain.

In parallel, we employed a second Phf6 KO mouse model wherein Phf6^loxP/loxP were bred with Nestin-Cre⁺ mice to induce deletion of Phf6 from the mouse central and peripheral nervous system at E11.5 (Tronche et al, 1999), the onset of Nestin gene expression, thus producing a highly efficient KO model (Fig. EV3A,B). We subjected the brain sections from Phf6^-/Y / Nestin-Cre⁺ (KO) and Phf6^loxP/Y / Nestin-Cre⁻ (Ctl) mice at post-natal day 0 (P0) to Nissl staining and found a notable decline in neuron density within the forebrain and midbrain sections of Phf6^-/Y / Nestin-Cre⁺ brains compared to Phf6^loxP/Y / Nestin-Cre⁻ controls (Fig. EV3C). Our data suggest that the deletion of Phf6 induces a decline in neuron density. Taken together, these results support a model whereby PHF6 may restrict eNSC self-renewal and promotes eNSC commitment to newly born neurons.

The R324X mutation is the most recurrent BFLS patient-mutation occurring at exon 10 (C.1024 C > T), impairing the ePHD2 domain, whereby PHF6 is proposed to function as a truncated protein (Ahmed et al, 2021; Chao et al, 2010; Crawford et al, 2006; Gecz et al, 2006; Jahani-Asl et al, 2016; Lower et al, 2004, 2002; Todd et al, 2015). Another BFLS patient point mutation (m) in Phf6 is wherein cysteine-99 is replaced with phenylalanine (C99F) at nt.296 G > T impairing the function of the PHD1 domain. To investigate whether impairment in eNSC fate specification may underlie BFLS pathogenesis, we employed both BFLS mouse models, R342X and C99F-m. Analysis of mRNA expression in E14 cerebral cortices revealed a consistent increase in the expression of both Nestin and Sox2 in BFLS relative to wild-type control mice (Fig. 3K,L). We thus conducted additional analysis in eNSCs of R342X mice and found a significant increase in their self-renewal, neurosphere number, and proliferation relative to eNSCs of the wild-type control mice (Fig. 3M–P). Taken together, our findings demonstrate that similar to Phf6^-/Y / Nestin-CreERT2⁺ mice, BFLS patient mouse models exhibit alterations in eNSC expansion.

To identify downstream effectors of PHF6 function in the regulation of neurogenesis, we first analyzed the candidate target genes with their expression significantly deregulated based on the RNA-seq analysis with particular focus on druggable targets (e.g., Receptors, Kinases). These analyses revealed a host of candidate genes that could serve as PHF6 targets to regulate neurogenesis (Dataset EV6). We focused on members of the EphR family (EphA4/7 and EphB1/2) given that EphRs are the largest family of RTKs highly expressed in the developing brain (Barquilla and Pasquale, 2015; Darling and Lamb, 2019; Lisabeth et al, 2013). Importantly, EphRs have been shown to play different roles in regulating neuronal development (Aoki et al, 2004; del Valle et al, 2011; Stuckmann et al, 2001; Wilkinson, 2014). Prior to validation of EphRs as viable targets of PHF6 in the context of BFLS, we conducted additional gene expression analysis using public databases. First, via querying single-cell RNA-seq data [Data ref: (Di Bella et al, 2021)] of the developing mouse brain, we found that Phf6, EphA4/7, and EphB1/2 are expressed in the developing brain of mice ranging from embryonic day 10 (E10) to postnatal mice at day 4 (P4) (Figs. 4A–F and EV4A–E). Furthermore, Pearson correlation analysis between Phf6, EphA4/7, and EphB1/2 revealed a positive correlation between Phf6 and EphR expression in different cell types, in particular progenitors and migrating neurons (Figs. 4G,H and EV4A–E). Second, we analyzed the RNA-seq data of the human ventral frontal cortex (VFC) [Data ref: (BrainSpan Atlas of the Developing Human Brain, 2011)] and found a similar trend in the expression of EPHR genes across development, and their correlation with PHF6 expression (Fig EV4F–J). We, thus, asked if EphR expression levels are altered in Phf6 KO and BFLS mice. Via subjecting eNSCs from Phf6^loxP/Y / Nestin-CreERT2^- and Phf6^-/Y / Nestin-CreERT2⁺ mice to RT-qPCR and immunoblotting analyses, we observed a significant decrease in both mRNA and protein expression of each of the identified EphR upon genetic deletion of Phf6 (Fig. 5A,B). Independently, we also subjected E14 brain tissue from R342X and C99F-m mice to RT-qPCR and immunoblotting analyses. The results revealed downregulation of EphR mRNA and protein expression in both C99F and R342X mice with a more profound impact in R342X mice (Figs. 5C,D and EV5A,B), confirming that the expression of EphRs is altered in BFLS mice harbouring PHF6 patient mutations.

We next set out to investigate if the identified EphRs are direct PHF6 targets. Our ChIP-seq data revealed robust and significant binding of PHF6 to the promoter of EphA4 with a p-value of 1.8E−08 (Dataset EV1). ChIP-seq data also revealed peaks associated with the TSS of EphA7 and EphB1 although the p-values did not reach the cut off values for significance (p-value for EphA7: 7.2E−04; p-value for EphB1, 1.4E−04) (Dataset EV1, Fig. EV1). We designed ChIP-qPCR experiments to specifically investigate the possibility of PHF6 binding to EphA4 but also examined EphA7/EphB1 (Fig. EV1) due to their significant deregulated expression in a PHF6-dependent manner (Dataset EV6). To begin with, ChIP-qPCR experiments were conducted in PHF6-overexpressing N2A cell lines using a ChIP-grade PHF6 antibody. We established PHF6 enrichment on the EphR genes that we examined in N2A cells expressing PHF6-GFP relative to GFP control (Fig. EV5C). We further assessed the functional consequences of PHF6 binding to EphR via loss- and gain-of-function studies. We conducted a firefly luciferase assay in N2A cell lines expressing PHF6-GFP or GFP control (Fig. EV5D). The cells were electroporated with either the control pGL4.23-basic reporter plasmid (pGL4.23), or the luciferase reporter plasmids harbouring the promoters of different EphR genes, including pGL4.23-EphA4, pGL4.23-EphA7, and pGL4.23-EphB1, together with a Renilla expression plasmid, and were subjected to a dual luciferase assay after 48 h. Cells expressing PHF6-GFP showed increased reporter activity for EphR regulatory regions (Fig. EV5D). Second, we induced the KD of Phf6 via a pool of siRNA (Fig. EV5E) and subjected the cells to a firefly luciferase assay. Our data revealed significant downregulation of EphR promoter activity in Phf6 KD cells. Importantly, parallel immunoblotting and RT-qPCR analyses revealed significant deregulation of the EphR protein and mRNA expression levels in a PHF6-dependent manner (Fig. EV5F,G).

To further assess if PHF6 direct regulation of EphR might be perturbed in the patient mouse models, we conducted ChIP assays in either E14 or P0 whole brain tissue of R342X, as well as luciferase assay in primary eNSCs and found that PHF6 regulation of EphR is consistently impaired in R342X mice (Fig. 5E–G). Similarly, the ChIP assay revealed that the binding of PHF6 to EphA4 and EphB1 promoters were significantly attenuated in whole brain tissue of E14 C99F mice relative to the wild-type control (Fig. EV5H). The specificity of the PHF6 antibody used for ChIP was also confirmed in IP and ChIP-PCR experiments using Phf6^loxP/Y / Nestin-CreERT2^- and Phf6^-/Y / Nestin-CreERT2⁺ eNSCs (Fig. EV5I,J).

We have established that PHF6 directly binds to gene regulatory elements of EphR to upregulate their expression. Mice harbouring Phf6 deletion, or BFLS patient mutations exhibit altered NSC self-renewal and deregulated EphR expression, raising the question of whether knockdown of either EphA4/7 or EphB1 can phenocopy the PHF6 mutant induced eNSC phenotype in BFLS.

EphA4 and EphA7 are involved in NSC regulation and neural development. EphA4 has been studied in axon guidance and neural circuit formation, whereas EphA7 plays a key role in apoptosis and cortical patterning (Depaepe et al, 2005; Kania and Klein, 2016; Klein, 2012).

We employed an siRNA approach in primary E14 WT eNSCs to induce the KD of each of these receptors followed by ELDA analysis to assess eNSC self-renewal (Fig. 6A–D). Our results revealed a significant increase in eNSC self-renewal in both EphA4 and EphA7 KD cells with the most profound impact in the EphA4 KD cells (Fig. 6A,B). Although there was a similar trend with EphB KD in eNSCs, no significant changes in self-renewal were induced upon the knockdown of EphB family of receptors (Fig. 6C,D). To investigate the impact of EphR KD on stemness, we also subjected whole protein lysates to immunoblotting using SOX2 and NESTIN antibodies (Figs. 6E,F and EV5K,L). We found that KD of EphA members and EphB1, but not EphB2, induced an increase in the protein expression levels of SOX2 and NESTIN (Figs. 6E,F and EV5K,L). Our studies demonstrate that although PHF6 regulates the gene expression of several EphR family members, KD of EphA4 induces the most significant phenotype on eNSC self-renewal.

In view of our observations that EphA4 KD most closely phenocopies the Phf6 mutant-induced eNSC phenotype, we next assessed if forced expression of EphA4 alters eNSC expansion. We generated an EphA4 plasmid fused with a GFP tag on the C-terminus. E14 WT eNSCs were cultured and electroporated with EphA4-GFP (pLVX.EphA4-GFP), or control GFP plasmid (pLVX.GFP) followed by ELDA and immunoblotting analysis (Fig. 6G–J). Our results showed that the expression of EphA4 induced a significant decline in eNSC self-renewal (Fig. 6G), the protein expression of both SOX2 and NESTIN (Fig. 6I), stem cell frequency (SCF) (Fig. 6H), and eNSC sphere size (Fig. 6J). We thus aimed to examine if the EphA- family of receptors can rescue the PHF6-mutant induced phenotype using the R342X mouse model (Fig. 6K–P). Forced expression of EphA4-GFP and EphA7-GFP was induced in eNSC cultures from the R342X mouse brain, and efficient electroporation of EphA4 and EphA7-GFP plasmids were confirmed by immunoblotting (Fig. 6K,L). LDA and ELDA analysis revealed that both EphA4 (Fig. 6M,N) and EphA7 (Fig. 6O,P) rescue the R342X-induced eNSC phenotype. In particular, EphA4 more profoundly decreased eNSC self-renewal and SCF in R342X eNSC (Fig. 6M,N). These findings assert the potential for the EphA- family of receptors, specifically EphA4, in ameliorating the PHF6-mutant induced eNSC phenotype.

We have established that PHF6 patient mutations alter eNSC fate in BFLS, prompting us to characterize the eNSCs, in their niche, in the developing brain. We analyzed the whole-brain lysates of C99F-m (Fig. 7A) and R342X (Fig. 7B) at E14 by immunoblotting analysis. Similar to results from primary eNSCs, we found a marked increase in the expression of stem cell markers, in BFLS mouse brains (Figs. 7A,B and EV3H), and a decrease in protein expression of mature cell-type markers including oligodendrocytes (OLIG2), astrocytes (GFAP), as well as progenitor cells (ASCL1) (Fig. EV3I). We next analyzed the stem cell marker, SOX2 (Fig. 7C), and the progenitor cell marker, TBR2 (Fig. 7D), via immunohistochemical analysis of R342X E14 coronal brain sections (Fig. 7F). Percent population of both cell types were imaged and quantified in the ventricular zone (VZ) and subventricular zone (SVZ), which are regions of high stem cell density. We observed a reverse correlation between SOX2 positive (SOX2+) and TBR2 positive (TBR2+) cells in their neurogenic niches, whereby BFLS mice exhibited a higher percentage of SOX2+ cells and an attenuated number of TBR2+ cells (Fig. 7C,D,F), with no significant differences observed in the percentage of merged SOX2 + /TBR2+ cells (Fig. 7E,F). A similar trend of increased SOX2+ cells was noted in the Phf6^-/Y / Nestin-Cre⁺ brains (Fig EV3D).

The changes in the proportion of SOX2+ and TBR2+ cells suggest altered cell populations manifesting a disproportionate number of neural stem versus progenitor cells in BFLS, which may contribute to disease pathogenesis. In parallel studies, Nissl staining analyses revealed that similar to Phf6 KO mice, R342X brain sections exhibited a decrease in neuronal density throughout the cortex (Fig. EV3J), suggesting the possibility of impaired neuronal migration. We thus set out to analyze the impact of Phf6 deletion on cortical layer neurons via subjecting Phf6^loxP/Y / Nestin-Cre^- and Phf6^-/Y / Nestin-Cre⁺ brain sections at P0 to immunohistochemical analysis using antibodies to SATB2⁺, CTIP2⁺, and TBR1⁺ to quantify neuronal numbers in cortical layers II-V, layer V, and layer VI, respectively (Fig. EV3E,F). Our results revealed a shift of SATB2+ neurons away from the apical cerebral cortex plate in Phf6^-/Y / Nestin-Cre⁺ mice, with no significant changes in the number of SATB2+, CTIP2+, and TBR1+ neurons. To quantify the migration patterns of SATB2+ neurons in the cerebral cortex influenced by loss of Phf6, a grid consisting of 10 equivalent bins was applied to the image of P0 cerebral cortex to equally divide the cortical wall spanning from the basal of ventricle zone to the pial surface into ten bins. The ten bins were marked sequentially from apical to basal, with bin 1 covering the most superficial (i.e., apical) layer, and bin 10 covering the deepest (i.e., basal) layer. Neurons within each bin were counted and a significant decline in SATB2+ neurons in bin 1 of the cerebral cortex was observed in Phf6^-/Y / Nestin-Cre⁺ mice (Fig. EV3E,F), suggesting impairment in the ability of SATB2+ neurons to migrate to superficial layers of the developing cerebral cortex in Phf6 KO mice. This finding is consistent with the attenuation of neuron density (Fig. EV3C) and suggests that PHF6 is involved in regulating the process of radial neuronal migration during the establishment of cortical lamination.

Together, we report that PHF6 alters the mechanisms that regulate NSC fate in the developing brain, and that loss-of-function of PHF6 in BFLS results in an imbalance in the number of uncommitted stem cells and neural progenitors which may contribute to BFLS pathogenesis.

All animal experiments were approved by the Animal Care Committee (ACC) at the University of Ottawa in Ottawa, Ontario, Canada, and McGill University in Montreal, Quebec, Canada. Mice were maintained in regular housing conditions with standard access to food and drink in a pathogen-free facility. The R342X mouse model was generated using CRISPR/Cas9 and functions as a truncated PHF6 protein (Chao et al, 2010; Crawford et al, 2006; Gecz et al, 2006; Jahani-Asl et al, 2016; Lower et al, 2004; Lower et al, 2002; Todd et al, 2015) This strain was generated through the breeding of R342X female heterozygous (HET) mice with C57BL6/J WT (B6 WT) male mice. Hemizygous (HEMI) males were used as experimental mice, and B6 WT males were used as a control. The C99F-m mouse model was generated using CRISPR/Cas9 where cysteine-99 is replaced with phenylalanine (C99F) at nt.29 G > T (Cheng et al, 2018). This strain was generated through breeding C99F-m female HET mice with B6 WT male mice. HEMI males were used as experimental mice, and B6 WT males were used as control.

The Phf6^-/Y / Nestin-CreERT2⁺ mouse strain (KO) is generated by a brain-specific deletion of Phf6 via breeding Phf6^fl/fl female mice (McRae et al, 2019) with Nestin-CreERT2⁺ male mice and inducing the Cre recombinase via oral gavage of Tamoxifen (Sigma-Aldrich, T5648) in pregnant dams at E14 and embryos collected 24–48 h later. Phf6^-/Y / Nestin-CreERT2^{+ -}were characterized and compared to Phf6^loxP/Y / Nestin-CreERT2^- control mice subjected to tamoxifen administration and used as control in all analyses. Male mice were used throughout.

The Phf6^-/Y / Nestin-Cre⁺ mouse strain was generated by breeding Phf6^fl/fl female mice with Nestin-Cre⁺ male mice to generate Phf6^-/Y/Nestin-Cre⁺ KO males and Phf6^-/Y/ Nestin-Cre^- control littermates. Here, the Phf6 gene was deleted from the mouse central and peripheral nervous system from E11.5 (Tronche et al. 1999), which is the onset of Nestin gene expression.

For genotyping, mouse tissue (tail or ear clipping) was first lysed in alkaline lysis buffer (25 mM NaOH, 0.2 mM EDTA pH 12) and then placed in a heat block at 95 °C for 30 min. The samples were then neutralized using an equal volume of neutralization buffer (40 mM Tris-HCl pH 5.0).

For genotyping of C99F-m and R342x, the PCR reaction mixture was set up as follows using Klentaq Thermostable DNA Polymerase Thermus aquaticus, recombinant, E. coli (Jena Bioscience, #PCR-217L); 2.5 μL 10x PCR buffer, 0.2 μL 25 mM dNTP, 6.5 μL Betaine, 1 μL 10 μM forward primer, 1 μL 10 μM reverse/mutation primer, 0.2 μL Klentaq enzyme, 12.6 μL RNAse-free H₂O, 1 μL DNA for a total mix of 25 μL per PCR tube.

For genotyping of Phf6^fl/fl, the PCR reaction mixture was set up as follows using a 2x Green PCR Master-Mix high performing (ZmTech Scientific, #S2100G); 7.5 μL 2x Green PCR Master-Mix, 0.4 μL 10 μM forward primer, 0.4 μL 10 μM reverse primer, 5.7 μL RNAse-free H₂O, 1 μL DNA for a total mix of 15 μL per PCR tube.

For genotyping of Nestin-CreERT2, the PCR reaction mixture was set up as follows using a 2x Green PCR Master-Mix high performing (ZmTech Scientific, #S2100G); 7.5 μL 2x Green PCR Master-Mix, 1.5 μL 0.5 μM oIMR1084 primer, 1.5 μL 0.5 μM oIMR1085 primer, 1.5 μL 0.5 μM oIMR7338 primer, 1.5 μL 0.5 μM oIMR7339 primer, 0.975 μL 6.5% glycerol, 1 μL DNA for a total mix of 15.5 μL per PCR tube.

The genotyping samples were PCR amplified in a Bio-Rad T100 Thermal Cycler using the following program for C99F-m, R342X, Phf6^Loxp/Loxp: 1. 95 °C for 2 min, 2. 95 °C for 30 s, 3. 60 °C for 30 s, 4. 72 °C for 30 s, 5. repeat steps 2–4 33x, and 6. 72 °C for 4 min.

The Nestin-CreERT2 genotyping samples were PCR amplified using the following program: 1. 94 °C for 2 min, 2. 94 °C for 20 s, 3. 65 °C for 15 s (−0.5 °C per cycle), 4. 68 °C for 10 s, 5. Repeat steps 2–4 10 times, 6. 94 °C for 15 s, 7. 60 °C for 15 s, 8. 72 °C for 10 s, 9. Repeat steps 6–8 28 times, and 10. 72 °C for 2 min.

The Nestin-Cre genotyping samples were PCR amplified using the following program: 1. 94 °C for 2 min, 2. 94 °C for 20 s, 3. 60 °C for 20 s, 4. 72 °C for 25 s, 5. Repeat steps 2–4 35 times, and 6. 72 °C for 2 min.

The PCR-amplified products were run on a 3% agarose gel at 100 V for 40 min for C99F-m, R342X, Nestin-CreERT2, and Nestin-Cre. The Phf6^fl/fl PCR amplified products were run on a 3% agarose gel at 100 V for 60 min. See primers listed in Table EV1.

Pregnant dames (gestation day E14) were given an oral gavage of one 0.1 mL dose of Tamoxifen (Sigma-Aldrich, T5648) at a concentration of 20 mg/mL using a 1 mL syringe and a 22-gauge feeding needle (Instech Solomon, #FTP-22-25-5).

Protein lysates were obtained from whole brain tissue harvested in RIPA lysis buffer containing protease and phosphatase inhibitors (ThermoFisher Scientific, A32959). The concentration of proteins was analyzed by the Bradford Assay (Bio-Rad) with BSA standard. PVDF membranes were activated in Methanol for 5 min and then blocked in 5% BSA in TBST. Membranes were probed with anti-PHF6 (NOVUS, NB100-68262, 1:1000), anti-EphA4 (ThermoFisher, 37-1600, 1:500) or (Santa Cruz, sc-365503, 1:100), anti-EphA7 (ThermoFisher, BS-7034R, 1:500) or (R&D Systems, MAB1495, 1:100), anti-EphB1 (Abcam, ab129103, 1:1000), anti-EphB2 (Abcam, ab252935, 1:500), anti-SOX2 (Abcam, ab97959, 1:250), anti-NESTIN (Santa Cruz, sc-23927, 1:100) or (R&D Systems, MAB2736, 1:500), anti-GFP (Abcam, ab1218, 1:1000), anti–GAPDH (Cell Signaling, 2118S, 1:5000), anti-beta-Actin (Sigma-Aldrich, a5316, 1:2000), alpha-Tubulin (Abcam, 9074, 1:5000), overnight at 4 °C, followed by HRP-conjugated secondary antibody, anti-rabbit IgG HRP (Bio-Rad, 1706515) or anti-mouse IgG HRP (Bio-Rad, 1706516) for 2 h at room temperature. Proteins were visualized with ECL (Bio-Rad), and signals were detected with a Chemidoc imaging system (Bio-Rad).

80 µg of total cell extracts from Phf6^loxP/Y / Nestin-CreERT2^- or Phf6^-/Y / Nestin-CreERT2⁺ eNSCs were employed for immunoprecipitation (IP), using either 1 µg of IgG or PHF6 antibody (NOVUS, NB100-68262, 1:1000). For input, 4 µg of total cell lysates from both Phf6^loxP/Y / Nestin-CreERT2^- and Phf6^-/Y / Nestin-CreERT2⁺ eNSCs were utilized.

RNA was isolated from cells and whole brain tissue with Trizol (Invitrogen) according to the manufacturer’s instructions. Reverse transcription of RNA was performed using 5x All-In-One RT MasterMix cDNA synthesis (Abm, G492). Quantitative real-time PCR was performed using SsoAdvanced™ Universal SYBR®Green Supermix (Bio-Rad, 1725271). Samples were incubated at 25 °C for 10 min, followed by incubation at 42 °C for 15 min, and finally 85 °C for 5 min to inactivate the reaction. See primers listed in Table EV1.

Mouse brains were fixed in 4% paraformaldehyde (PFA) for 24 h, followed by 24 h of 15% sucrose fixation, and another 24 h of 30% sucrose fixation before being snap frozen in OCT on dry ice. 8 μm frozen sections were cut using a cryostat. Antigen retrieval was performed on sections prior to blocking by submerging slides in a slide holder with Dako Target Retrieval Solution (Agilent, S1699) and heating in a beaker of water for 20 min at 95–98 °C. Sections were then cooled for 15 min and blocked in 20% donkey serum, 0.1% Triton-X, 0.1% Tween in PBS, for 20 min at room temperature. We applied the SOX2 (1:250) antibody (Abcam, ab97959) and the TBR2 antibody (1:50) (ThermoFisher, 14-4875-82) overnight at 4 °C in a humid chamber. Secondary antibodies (1:500); Anti-rabbit IgG, Alexa Fluor® 647 Conjugate (Cell Signaling, 4414 S), Anti-rat IgG Alexa Fluor® 488 Conjugate (Cell Signaling, 4416 S), and DAPI (1:1000 of 1 µg/ml) (ThermoFisher, D1306) were applied for 45 min at room temperature in a humid chamber. Slides were mounted with ProLong Gold Antifade Mountant (ThermoFisher, P36934) with a #1.5 coverslip. Images were obtained with a laser scanning confocal microscope (ZEISS LSM 800) at 20× objective. Detection wavelengths were as follows: DAPI detection 400–605, TBR2 (AF488) 400–650, SOX2 (AF647) 645–700, and all with a detector gain of 650 V.

For PHF6 and coronal layer marker immunofluorescent brain section staining, brains were fixed with 4% PFA and equilibrated in 30% sucrose solution at 4 °C until the brains sank to the bottom of the vail. Brains were immersed in 50% OCT (VWR) solution diluted by 30% sucrose for overnight at 4 °C. Brains were transferred to the cryomold (VWR) filled with 50% OCT and flash frozen in liquid nitrogen. Frozen brains were stored at −80 °C. Brain blocks were subjected to cryosection at the thickness of 12 µm and mounted onto SuperFrost slides (Fisher Scientific). Sections were then washed three times with PBS and 0.1% Tween-20 detergent (PBST), then antigen retrieval in Citrate buffer (0.1 M, pH 6.0) by microwave boiling for 10 min and blocked in 10% horse serum/PBST for 30 min at room temperature. After blocking, sections were subjected to the following primary antibody for PHF6 immunofluorescence (overnight at 4 °C): rabbit anti-PHF6 (1:150, Sigma-HPA001023), or for coronal layer markers; mouse anti-SATB2 (1:200, Abcam-ab51502), rat anti-CTIP2 (1:200, Abcam-ab18465), and rabbit anti-TBR1 (1:200, Abcam-ab31940). The next day, after washing three times in PBST, sections were incubated in secondary antibody for 1 h at room temperature: anti-rabbit 555 Alex Fluor (1:500, Invitrogen A21206), anti-mouse 488 Alexa Fluor (1:500, Invitrogen A21202), or anti-rat 647 Alexa Fluor (1:500, Invitrogen A21247). Nuclei were counterstained by incubating sections in Hoechst 33342 dye (ThermoFisher Scientific) for 15 min at room temperature. Finally, slides were mounted onto coverslips (Fisher Scientific) in DAKO Fluorescence Mounting Medium (Agilent Technologies).

For Nissl staining, brain sections were rehydrated by 10 min submersions in 95% ethanol, followed by 1 min submersion in 70% ethanol and 1 min submersion in 50% ethanol. Sections were rinsed in tap water and then in distilled water. After washing, sections were stained in 0.25% Cresyl Violet Stain Solution in distilled water for 5 min, followed by a quick wash in distilled water. Sections were quickly differentiated in 70% ethanol with 1% acetic acid for 10 s to 1 min and checked under the microscope. Sections were then dehydrated via two 5-min submersions in 100% ethanol. Finally, slides were cleared by three 5-min submersions in xylene and mounted onto coverslips (Fisher Scientific) with Permount Mounting Medium (Fisher Chemical). Stained slides were air-dried overnight in the fume hood at room temperature. Immunofluorescent images were acquired by using Zeiss Axiovert Observer Z1 epifluorescent/light microscope equipped with an AxioCam cooled-colour camera (Zeiss) or SP8 confocal microscope (Leica). Nissl-stained slides were scanned by a Zeiss AxioScan Z1.

Embryonic NSCs (eNSCs) were obtained by whole brain culturing of E14 mice (Burban and Jahani-Asl, 2022; Nasser et al, 2018) (excluding cerebellum). Pregnant mice were euthanized, uterine horns were removed, and embryos were placed in cold 1× HBSS. Brain tissue was cut into small pieces and placed in 15 mL falcon tubes containing 1 mL cold 1× HBSS. Tissue was allowed to settle to the bottom, HBSS was replaced with 1 mL fresh HBSS for washing, and then replaced once more with 1 mL stem cell media (SCM) containing 1:1 DMEM-F12 (Wisent, 319-005-CL) (ThermoFisher, 31765035), 50 units/mL penicillin-streptomycin (Wisent, 450-201-EL), 1× B-27 supplement (Invitrogen, 17504044), 2 µg/mL Heparin (Stemcell Technologies, 07980), 20 ng/mL mEGF (Cell Signaling, 5331SC), 12.5 ng/mL bFGF (Abbiotec, 600182). The tissue was mechanically dissociated 15× with P1000 then an additional 15× with P200. The lysate was then plated in 6 mL of SCM and left in the incubator for 6–7 days until spheres grew to 40–200 μm in size, replenishing with 2 mL SCM media at day 4.

For the limiting dilution assay (LDA), NSCs were dissociated to single-cell suspension using Accumax. Single cells were counted and plated in a 96-well plate at different cell doses per well, in triplicates. Spheres were counted 7 days post-plating.

For the extreme limiting dilution assay (ELDA), NSCs were dissociated to single-cell suspension using Accumax. Single cells were counted and plated in a 96-well plate at different cell doses per well with a minimum of 12 wells/cell dose (Rasool et al, 2022). 7 days post-plating, the presence or absence of spheres in each well was recorded and analyzed with http://bioinf.wehi.edu.au/software/elda/33 (Hu and Smyth, 2009).

For cell viability, NSCs were dissociated to single-cell suspension using Accumax. Single cells were counted and seeded at a density of 200 cells/well, in a 96-well plate. Cell viability was evaluated 7 days post-plating using alamarBlue (Thermo Fisher Scientific, #DAL1100) according to the manufacturer’s protocol. 10% resazurin was added to the cells in each well and incubated for 4 h at 37 °C. Fluorescence was read using a fluorescence excitation wavelength of 560 nm and an emission of 590 nm.

Representative images of spheres were taken with the 10× objective lens of an Olympus IX83 microscope with an X-Cite 120 LED from Lumen Dynamics, and an Olympus DP80 camera.

eNSCs were dissociated into single-cell suspension using Accumax, counted and plated at a density of 1 × 10⁶ cells. Cells were incubated with 10 μM EdU upon plating. Following 22 h in culture, eNSCs were fixed, permeabilized, and stained using the Click-iT EdU proliferation kit (Thermo Fisher Scientific, #C10337) according to the manufacturer’s protocol. Fluorescence was analyzed by flow cytometry (BD FACS CantoII & Sony SH800). Data were analyzed using the FlowJo software. The number of cells that had incorporated EdU was defined as the ratio of EdU-positive cells over total number of cells.

PBS containing protease inhibitors (Thermo Fisher Scientific, #A32959) was used as cell washing buffer prior to fixation. Cross-linking was done via 1% formaldehyde in PBS for 10 min and quenched with 0.125 M glycine in PBS for 5 min at room temperature (RT). Washing, fixation, and quenching was done in 15 mL tubes while rotating at RT. Post-quenching, cells were washed twice with PBS containing protease inhibitors. Cells were then pelleted by spinning at 150 × g for 10 min at 4 °C. Cell pellets were dissolved in ChIP lysis buffer (40 mM Tris-HCl, pH 8.0, 1.0% Triton X-100, 4 mM EDTA, 300 mM NaCl) containing protease inhibitors. Chromatin fragmentation was performed through water bath sonication (BioRuptor) at 4 °C, creating an average length of 500 base pairs (bp) of product. Cell lysates were spun down at 12,000 G for 15 min, followed by dilution of supernatant (1:1) in ChIP dilution buffer (40 mM Tris-HCl, pH 8.0, 4 mM EDTA, protease inhibitors).

Immunoprecipitation (IP) was done using a PHF6 antibody (Novus Biological, NB100-68262), rabbit IgG antibody (Cell Signaling, #3900 S). Antibody-protein-DNA complexes were collected, washed, and then eluted. Reverse cross-linking was done as described in Soleimani et al, 2013 (Soleimani et al, 2013). Immunoprecipitated DNA was analyzed by qPCR, and the binding enrichment was expressed as a percentage of the input.

The PHF6 binding regions (based on ChIP-seq peaks) were cloned into the pGL4.23 (Promega) vector to generate the EphA4, EphA7 and EphB1 luciferase reporter genes by digesting the plasmid and the annealed primer pair using EcoRV (NEB, #R0195L) and KpnI (NEB, #R3142) then ligating them with T4 DNA ligase (NEB, #M0202L). The constructs were confirmed by DNA sequencing. Cells were electroporated with the EphA4-pGL4.23, EphA7-pGL4.23, EphB1-pGL4.23 or the empty pGL4.23. Luciferase assays were performed 48 h after transfection with the Dual-Luciferase Reporter Assay system (Promega, #E1910) with a GloMax Luminometer (Promega). In all experiments, cells were electroporated with a Renilla firefly reporter control and the firefly luminescence signal was normalized to the Renilla luminescence signal. See primers listed in Table EV1.

Transient KD of Phf6 and EphA4/A7/B1/B2 using an siRNA approach was performed with ON TARGET-plus SMART pool mouse Phf6 siRNA (Dharmacon, #L-058690-01-0005), mouse EphA4 siRNA (Sino Biological, #MG50575-M), mouse EphA7 siRNA (Sino Biological, #MG50587-M, mouse EphB1 siRNA (Sino Biological, #MG50479-M), mouse EphB2 siRNA (Santa Cruz, #sc-39950), and ON TARGET-plus non-targeting pool (Santa Cruz, #sc-36869). siRNA (100 nM) were nucleofected into eNSCs (10⁶ cells) and cultured in eNSC media at 37 °C in a humidified atmosphere of 5% CO₂.

Single cell RNA-seq data from the mouse cerebral cortex was obtained [Data ref: (Di Bella et al, 2021)]. Log normalized counts, cell type annotation and UMAP coordinates were retrieved from the original publication and used to generate UMAP plots. For the correlation analysis, MAGIC (Van Dijk et al, 2018) was applied to obtain imputed gene expression. Correlation values were obtained on the imputed gene expression after applying MAGIC.

Normalized RPKM (Reads per Kilobase Million) values of RNA-seq data were obtained from the Allen Brain Atlas BrainSpan dataset [Data ref: (BrainSpan Atlas of the Developing Human Brain, 2011)] and data from the ventral frontal cortex (VFC) was taken. The average RPKM values was calculated per developmental time. All plots were generated using R (version 4.0.0).

ChIP-seq was performed by pooling the cortex of three mice (n = 3) prior to sequencing. ChIP-seq data were processed as previously described (Hernandez-Corchado and Najafabadi, 2022). Briefly, raw reads were aligned to the mouse genome assembly version mm10 with bowtie2 (version 2.3.4.1) using the “--very-sensitive-local” mode. Duplicate reads were removed using samtools (version 1.9) (Danecek et al, 2021). ChIP-seq peaks were identified using MACS (version 1.4) (Feng et al, 2012; Zhang et al, 2008) with a permissive p-value threshold of 0.001, using “--nomodel” option. Fragment size was specified using “--shiftsize” argument, with the fragment length obtained by cross-correlation analysis using phantompeakqualtools (Landt et al, 2012). Peak-TSS distances were calculated using bedtools (Quinlan and Hall, 2010) only for peaks that passed p-value threshold of 10^–5, with TSS coordinates obtained from GENCODE (Frankish et al, 2019) (release M9).

Pol II occupancy data were obtained from GEO (accession number GSM2442441) (Liu et al, 2017). The bedGraph file representing Pol II occupancy was directly downloaded from GEO, converted to bigWig, and overlayed on gene TSS coordinates using bwtool (Pohl and Beato, 2014).

Cortical progenitors were established from the cortex of wild-type E14 mice and subjected to electroporation with Phf6 siRNAs (n = 3) and non-targeting control siRNA (n = 3). Cell were subjected to mRNA-Seq analysis following 5 days in culture. mRNA-seq raw reads were mapped to mm10 genome using HISAT2 (Kim et al, 2015), followed by duplicate read removal using samtools. Gene-level read counts were obtained by HTSeq (Anders et al, 2015), using gene annotations from GENCODE (release M9). Genes with a minimum of 150 reads in at least one sample were retained. Gene set analysis was performed using ConsensusPathDB (Kamburov et al, 2011).

Statistical analysis was performed with the aid of GraphPad software 7. Two-tailed unpaired student t-tests were used to compare two conditions (normal distribution). One-way ANOVA was used for analyzing multiple groups (normal distribution). Data are shown as mean with standard error of mean (mean ± SEM). p-values of equal or less than 0.05 were considered significant and were marked with one asterisk (*). p-values of less than 0.01 are denoted by **, and p values of less than 0.001 are denoted by ***. All data presented are from 3 or more independent biological (n) replicates (n ≥ 3), unless otherwise noted in corresponding figure legends, thus no additional statistical methods were used to predetermine sample size. Randomization was used to allocate animals to experimental groups, following genotyping. The researchers were blind to treatment groups for all quantifications as well as imaging analysis. Only male mice were included in this study. Methods of statistical analysis and p-values employed are reported in corresponding figure legends.

Synopsis graphic was created with BioRender.com.