IGF2 interacts with the imprinted gene Cdkn1c to promote terminal differentiation of neural stem cells

Adult neurogenesis in the subventricular zone (SVZ) is an orderly multistep process in which self-renewing astrocyte-like neural stem cells (NSCs or B1 cells) produce mature progeny via transit-amplifying progenitors (TAPs or C cells) (Chaker et al., 2016). These progenitors rapidly divide to give rise to neuroblasts (A cells) that migrate through the rostral migratory stream (RMS) to the olfactory bulb (OB) where they fully differentiate and integrate as interneurons (Calzolari et al., 2015; Chaker et al., 2016; Götz et al., 2016; Ponti et al., 2013). The SVZ also gives rise to astrocytes and oligodendrocytes that integrate into the corpus callosum (CC) (Menn et al., 2006; Sohn et al., 2015) and striatum (Figueres-Oñate et al., 2019). Sustained neurogenesis throughout adult life also occurs in the subgranular zone (SGZ) of the dentate gyrus (DG) in the hippocampus and is hypothesized to be involved in behavioural/cognitive processes, such as memory, and in diseases of the central nervous system (CNS) (Gage et al., 1998; Gonçalves et al., 2016).

Genomic imprinting is an epigenetic process that causes genes to be expressed depending on their parental origin (John and Surani, 2000). A relatively small subset of genes within the mammalian genome (0.4%) is imprinted, showing monoallelic expression in the whole organism or in specific tissues favouring the maternal or the paternal allele (John and Surani, 2000). Imprinted expression is initially determined by differential DNA methylation that is established in the germline (Surani, 1998). The prevalence of genomic imprinting is higher in the brain than in other organs and many genes with a crucial role in neurodevelopment are expressed in a maternal- or paternal-specific manner (Perez et al., 2016). For example, Cdkn1c is an imprinted gene expressed only by the maternal allele (Pateras et al., 2009) and the misregulation of its expression has been associated with human growth disorders, such as Beckwith-Wiedemann and Silver-Russell syndromes, as well as with the onset of several types of cancers (Bastaki et al., 2016; Binder et al., 2020; Nakashima et al., 2015; Stampone et al., 2021). Cdkn1c encodes p57 protein and belongs to the CIP/KIP family of cyclin-dependent kinase inhibitors (CKIs), which also includes p21 and p27. Among them, p57 is the least studied CIP/KIP member, although it has been demonstrated to play a fundamental role in regulating the cell cycle and differentiation during mammalian development (Rossi et al., 2018). Indeed, this cell cycle regulator controls multiple stages of corticogenesis such as cell cycle exit of foetal progenitors, their differentiation and migration (Furutachi et al., 2015; Jadasz et al., 2012; Laukoter et al., 2020; Mairet-Coello et al., 2012; Tury et al., 2011). Importantly, p57 accumulates in proliferating MASH1⁺ telencephalic neural progenitors; it represses neuronal differentiation independently of cell-cycle exit and acts as a direct repressor of transcription (Joseph et al., 2009). p57 has also been shown to be a key factor during embryonic specification of quiescent NSCs and its deletion impairs the emergence of adult NSCs in the SVZ (Furutachi et al., 2015). Moreover, p57 inhibits adult NSC proliferation after focal cerebral ischemia (Guo et al., 2014) and controls NSC quiescence in the adult SGZ (Furutachi et al., 2013). Consistently, a detailed analysis of hematopoietic stem cells (HSCs) lacking p57 has confirmed the role of this cell cycle inhibitor as a key molecule in quiescence and self-renewal (Matsumoto et al., 2011; Tesio and Trumpp, 2011; Zou et al., 2011).

Igf2 (insulin-like growth factor 2) is another imprinted gene expressed only by the paternal allele in most tissues (DeChiara et al., 1991; Ferguson-Smith et al., 1991; Giannoukakis et al., 1993). Remarkably, the silent maternal allele is specifically activated in neurogenic niches, resulting in biallelic expression of the gene (Ferrón et al., 2015; Lehtinen et al., 2011). It has been shown that IGF2 regulates NSCs in the adult mouse SVZ by influencing their cell cycle (Ferrón et al., 2011; Lehtinen et al., 2011; Ziegler et al., 2019). Indeed, IGF2 has been shown to promote expansion of NSCs more potently than either insulin growth factor 1 (IGF1) or standard growth media (Alagappan et al., 2014; Ziegler et al., 2014, 2019). Detailed analysis of the SVZ niche has shown that IGF2 is secreted in a paracrine manner by the choroid plexus (CP) epithelium, and becomes readily accessible to those NSCs contacting the cerebrospinal fluid (CSF) (Bracko et al., 2012; Ferrón et al., 2015; Lehtinen et al., 2011; Ziegler et al., 2014). The brain vasculature and leptomeninges also secrete IGF2 into the neurogenic niches (Ferrón et al., 2015). Accordingly, studies in the hematopoietic system showed that IGF2 acts as a potent growth factor for adult bone marrow cells by stimulating their ex vivo expansion (Barroca et al., 2017; Zhang and Lodish, 2004), and a direct link between IGF2 and the cell cycle regulator p57 has been shown in this system (Thomas et al., 2016). However the mechanisms by which IGF2 governs neurogenesis remain incompletely characterized.

In our study, we report that maternal, but not paternal, deletion of Cdkn1c in adult NSCs abrogates their quiescence and promotes their premature differentiation into terminally differentiated astrocytes. Moreover, we show that IGF2 promotes cell fate commitment by inducing the expression of Cdkn1c mRNA and p57 protein, which elicits cell cycle exit and enhances terminal differentiation of adult NSCs into non-multipotent astrocytes. Consistently, intraventricular infusion of recombinant IGF2 in a conditional mouse model with Cdkn1c-deficient neural progenitors has confirmed that IGF2 and p57 act in a common pathway to regulate quiescence and differentiation of the NSC pool in the adult SVZ. We also identify a methylation-independent mechanism of control of the Cdkn1c imprinted gene after IGF2 treatment, which involves the PI3K-Akt pathway during the differentiation of adult NSCs.

IGF2 plays an important role in NSCs by affecting their proliferation capacity (Bracko et al., 2012; Ferrón et al., 2015; Lehtinen et al., 2011); however, little is known about the effects of IGF2 in the differentiation process of NSCs. In order to study the effects of this factor in neural differentiation, NSCs isolated from the adult SVZ were expanded and differentiated into the three neural lineages of the CNS in the presence or absence of IGF2 (Belenguer et al., 2016). To do this, cells were first expanded as neurospheres in the presence of mitogenic stimulation with epidermal growth factor (EGF) and fibroblast growth factor 2 (FGF2) (Ferrón et al., 2007). Neurospheres were then disaggregated, plated on Matrigel (adherent assay) and cultured for 2 days in vitro (2 DIV) in medium without any insulin stimulation but containing FGF2 to induce neural progenitors differentiation (Belenguer et al., 2016). Afterwards, mitogens were withdrawn and the medium was supplemented with 2% foetal bovine serum (FBS) which is required for terminal differentiation (Fig. 1A). Under these conditions, differentiation is pushed forward and NSCs stop dividing and initiate an orderly program of commitment and differentiation into neurons, oligodendrocytes and astrocytes during the following 5 days (7 DIV) (Fig. 1A).

After 2 DIV under differentiation conditions, IGF2-treated NSCs showed an increase in proliferation, as indicated by an MTS viability assay (Fig. S1A). Consistently, immunodetection of the proliferation marker Ki67 revealed an increase in the proportion of nestin⁺ (Nes) progenitors that were still proliferating after 2 DIV under differentiation conditions in IGF2-treated cultures (Fig. S1B). This increased rate of proliferation was maintained after 3 DIV under differentiation conditions in presence of IGF2, as revealed by a higher proportion of Ki67⁺ and MCM2⁺ cells (Fig. S1C,D). Moreover, although the level of expression of the Nes gene was not altered in presence of IGF2 (Fig. S2A), the expression of the neuronal gene βIII-tubulin (Tubb3), the astrocytic genes S100b (S100β) and Gfap, and the oligodendrocytic gene Olig2 was increased in IGF2-treated cells after 2 DIV under differentiation conditions (Fig. S2A). Accordingly, culturing NSCs in the presence of IGF2 promoted their terminal differentiation into the three neural lineages, increasing the percentage of neurons measured as βIII-tubulin⁺ cells, astrocytes measured as cells expressing high levels of GFAP and S100β, and oligodendrocytes measured as O4⁺ cells formed after 7 DIV under differentiation conditions (Fig. 1B). Moreover, higher proportions of cells that were strongly positive for S100β, a protein largely absent from neurogenic GFAP⁺ cells (Raponi et al., 2007), and lower proportions of Nes⁺ cells, were also observed after 7 DIV under differentiation conditions in the presence of IGF2 (Fig. S2B), suggesting that IGF2 directly promotes terminal differentiation of multipotent NSCs. To confirm whether enhanced differentiation in the presence of IGF2 was accompanied by a reduction in the capacity of differentiated NSCs to form neurospheres, 7 DIV differentiated adult NSCs cultures were detached and replated again in proliferating conditions (Fig. 1A). This led to the re-activation of a small proportion of cells that retained the capacity to form neurospheres in non-adherent conditions after 5 more days (12 DIV) (Fig. S2C). As expected, the higher presence of S100β was consistent with a reduction in the neurospheres formation capacity in IGF2-treated NSCs cultures (Fig. S2C), indicating that the bias toward a more differentiated phenotype in the presence of IGF2 correlated with a reduction in stemness.

A direct link between IGF2 and the cell cycle regulator p57 has been shown in the HSC pool (Thomas et al., 2016). In fact, overexpression of IGF2 in purified adult murine HSCs resulted in the upregulation of Cdkn1c expression, giving rise to an arrest of HSCs in the G₀/G₁ phase of the cell cycle and promoting the formation of multi-lineage colonies in vitro. However, the link between IGF2 and p57 in the NSC population is still unknown. Neurospheres cultures grown in suspension under proliferating conditions exhibited low levels of Cdkn1c mRNA, but mitogen withdrawal induced an increase in Cdkn1c gene expression (Fig. 1C). Moreover, a significant increase in the percentage of cells with detectable levels of the p57 protein in differentiating cultures inversely correlated with the proportion of proliferating Ki67⁺ cells (Fig. S3A), suggesting that increased expression of p57 in the absence of mitogens might correlate with cell cycle exit and differentiation of adult NSCs. Accordingly, p57 protein was detected in Nes⁺ progenitors after 2 and 3 DIV under differentiation conditions, and in GFAP⁺ astrocytes at 7 DIV (Fig. S3B).

To define the potential role of IGF2 in regulating Cdkn1c expression, the levels of expression of this gene was quantified by qPCR in NSCs cultures in absence or presence of IGF2. This study revealed no changes in the levels of Cdkn1c expression in proliferating conditions in the presence of IGF2 (Fig. 1C). However, Cdkn1c levels were significantly increased IGF2-treated cultures in differentiation-promoting conditions showing a maximum difference at the first step (2 DIV) of the differentiation process (Fig. 1C). No changes in the levels of expression of the other two members of the CKIs family of proteins, Cdkn1a and Cdkn1b, were observed at the same time of differentiation (Fig. S3C). Immunofluorescent detection of nuclear p57 also revealed increased levels of the p57 protein in IGF2-treated cultures after 2, 3 and 7 DIV under differentiation conditions (Fig. 1D,E), consistent with the increased levels of the protein observed by immunoblot in differentiation conditions after IGF2 treatment (Fig. 1F). To determine whether mitogen withdrawal could influence the increase in p57 levels in neurospheres cultures, we generated primary cultures from the adult SVZ that were maintained in the absence of mitogens (Costa et al., 2011; Ortega et al., 2011) and determined the levels of p57 in the presence or absence of IGF2 (Fig. S3D). Consistent with the data in neurosphere cultures, the addition of IGF2 to the primary cultures also induced the expression of p57 (Fig. S3D,E), which resulted in a more differentiated state of the cultures, as indicated by the increased percentage of GFAP⁺ astrocytes, βIII-tubulin⁺ neuroblasts and Olig2⁺ oligodendrocytes in IGF2-treated cultures (Fig. S3F,G).

Cdkn1c expression is finely regulated by several epigenetic mechanisms, including genomic imprinting. The mouse Cdkn1c gene belongs to the Kcnq1ot1 imprinted cluster on mouse chromosome 7 and is canonically expressed from the maternally inherited chromosome (Fig. S4A) (Stampone et al., 2018). To explore whether the upregulation of Cdkn1c in IGF2-treated NSCs could be caused by a loss of imprinting of the Cdkn1c gene, we assayed the imprinting state of the gene in NSCs at 2 DIV under differentiation conditions. NSCs from wild-type adult F1 hybrid offspring from reciprocal crosses of Mus musculus domesticus (C57BL6/J) and Mus musculus castaneus (CAST/EiJ) strains, in which a single-nucleotide polymorphism (SNP) was identified at the Cdkn1c gene between the two subspecies, were analysed (Fig. S4B). IGF2-treated and untreated cultures showed the expected maternally inherited imprinted expression of Cdkn1c, and no expression of the paternal allele was observed (Fig. S4C), indicating that genomic imprinting was not altered after IGF2 treatment. The Cdkn1c gene has a somatic differentially methylated region (sDMR) that is located at the promoter of the gene that regulates its expression (Pateras et al., 2009; Stampone et al., 2018). Given that the imprinting state of the gene was not altered, and to further analyse whether DNA methylation of Cdkn1c promoter might be associated with the regulation of the expression of the gene after IGF2 treatment, we next determined the methylation levels of the sDMR by bisulphite sequencing. Consistent with the maintenance of genomic imprinting of Cdkn1c, IGF2-treated NSCs showed the expected levels of methylation (mean methylation percentage: untreated, 61.6±3.5%; IGF2 treated, 61.5±3.6%; P=0.96) (Fig. S4D), supporting a methylation-independent function of IGF2 on the regulation of expression of this gene in the adult NSCs.

IGF2 binds with high affinity to the insulin receptor (IR) and Igf1 receptor (IGF1R), and is also able to interact with IGF2R to target the ligand to lysosomes for degradation (D'Ercole et al., 1996; Stewart and Rotwein, 1996). It has been previously shown that IGF2 activates phosphatidylinositol 3-kinase/protein kinase B (PI3K-Akt) and mitogen-activated protein kinase (MAPK) to promote cell cycle progression and differentiation (Bracko et al., 2012; Chirivella et al., 2017). More precisely, the link between IGF2 and p57 has been shown to be dependent on the activation of PI3K/Akt pathway in HSCs (Thomas et al., 2016). Therefore, to elucidate the intracellular pathways downstream of IGF2 in differentiating NSCs, a study at the mRNA level was performed in NSCs after 2 DIV under differentiation conditions. This study showed that Insr and Igf1r were the most strongly expressed receptors and maintained their level of expression after IGF2 treatment (Fig. S5A). Moreover, treatment with IGF2 induced the phosphorylation of IR and IGF1R, whereas no effect was observed on IGF2R (Fig. 2A). Furthermore, the levels of phosphorylated Akt (pAkt) and ERK1/2 (pMAPK) were determined after IGF2 treatment in differentiation conditions. This study revealed a significant increase of both pAkt and pMAPK in IGF2-treated NSCs after 2 DIV under differentiation conditions (Fig. 2B). Notably, simultaneous treatment of adult NSC cultures with IGF2 and the PI3K pharmacological inhibitor LY294002 blocked the effects of IGF2 on Cdkn1c expression (Fig. 2C and Fig. S5B), whereas treatment with the ERK1/2 inhibitor PD0325901 did not modify the increased levels of Cdkn1c expression induced by IGF2 (Fig. 2C and Fig. S5B). These results demonstrated that IGF2 promoted Cdkn1c expression in NSCs through a mechanism involving PI3K-Akt but not the ERK1/2-MAPK pathway.

To determine whether differentiation effects of IGF2 via p57 are mediated by activation of the PI3K-Akt pathway, wild-type NSCs were differentiated for 7 DIV into astrocytes, oligodendrocytes and neurons in the presence of IGF2 and LY294002 or PD0325901 (Fig. 2D). As previously shown, IGF2 promoted an increase of the percentage of βIII-tubulin⁺ neurons, GFAP⁺/S100β⁺ astrocytes and O4⁺ oligodendrocytes cells (Fig. 2D,E). However, the addition of LY294002 to IGF2-treated cultures reverted the differentiation effects of the factor, resulting in cultures with a smaller proportion of neurons, astrocytes and oligodendrocytes (Fig. 2D,E). Accordingly, the proportion of undifferentiated cells was increased in the presence of LY294002 (Fig. 2D,E). Importantly, the addition of PD0325901 did not have any effect on NSCs differentiation (Fig. 2D,E). These data confirm that the PI3K-Akt pathway mediates regulation of Cdkn1c expression by IGF2.

To evaluate whether p57 could mediate the differentiation effects of IGF2 in adult SVZ NSCs, a murine genetic model was generated by crossing mice carrying loxP sites flanking exons 2 to 4 of Cdkn1c gene (Cdkn1c^loxp/loxp) (Matsumoto et al., 2011) with mice expressing the Cre-recombinase under the control of the mouse Gfap promoter (Gfap-cre^+/0) (Fig. S6A) (Garcia et al., 2004; Montalbán-Loro et al., 2019). Given that Cdkn1c is an imprinted gene expressed only from the maternal allele, crosses were carried out reciprocally to generate heterozygous mice with either the maternal (Cdkn1c-Gfap^mat) or the paternal (Cdkn1c-Gfap^pat) deleted allele (Fig. S6A). To first determine the specificity of the Gfap-cre recombination, we crossed Gfap-cre females with ROSA26R males and performed an X-gal histochemistry in the adult brain of the resulting mice (Gfap-cre/LACZ). This analysis showed positive staining for β-galactosidase in the SVZ and RMS of the adult Gfap-cre/LACZ brains (Fig. S6B), corroborating the deletion of Cdkn1c in the adult GFAP⁺ stem cell population.

A significant reduction of Cdkn1c mRNA and p57 protein was observed in Cdkn1c-Gfap^mat NSCs after 2 DIV under differentiation conditions (Fig. 3A,B), but no downregulation of p57 was detectable in Cdkn1c-Gfap^pat heterozygous NSCs (Fig. 3A,B), coincident with the canonical maternal expression of the gene. No changes in the levels of expression of Cdkn1a and Cdkn1b were observed at the same time of differentiation in Cdkn1c-Gfap^mat and Cdkn1c-Gfap^pat cultures (Fig. S6C). Importantly, body and brain weights were not affected in Cdkn1c-Gfap^mat and Cdkn1c-Gfap^pat compared with controls (Fig. S6D). To investigate the differentiation effects of IGF2 in the presence or in the absence of p57, NSCs from the adult SVZ of heterozygous Cdkn1c-Gfap^mat and Cdkn1c-Gfap^pat and control mice were expanded and differentiated for 7 DIV in the presence or absence of IGF2. We observed that IGF2 promoted the terminal differentiation of NSCs, which increases the percentage of βIII-tubulin⁺ neurons, GFAP⁺/S100β⁺ astrocytes and O4⁺ oligodendrocytes in control cultures, as previously shown (Fig. 3C,D). These effects were also observed in Cdkn1c-Gfap^pat cultures; however, the proportions of these three neural lineages were not equally induced in Cdkn1c-Gfap^mat cultures (Fig. 3C,D). Accordingly, the proportion of cells in the differentiated cultures with the capacity to activate and form new neurospheres in non-adherent cultures was reduced after IGF2 treatment in all cultures; however, Cdkn1c-Gfap^mat cultures were not equally reduced compared with control or Cdkn1c-Gfap^pat cultures (Fig. 3E). All these data confirm that maternal expression of Cdkn1c partly mediates the differentiation effects of IGF2 in adult NSCs in vitro.

Immunohistochemical analysis with antibodies to p57 and to cell-identity antigens in wild-type adult brains revealed nuclear staining for p57 protein in the GFAP population located close to the lateral ventricles (Fig. S7A), in mature neurons in the striatal parenchyma and in DCX⁺ neuroblasts reaching the olfactory bulb (Fig. S7A,B). Moreover, a transcriptomic analysis of the different populations in the neurogenic lineage from the adult SVZ and studied by flow cytometry (Belenguer et al., 2021a) revealed that GLAST⁺ quiescent NSCs (qNSCs) expressed very low levels of Cdkn1c, whereas GLAST⁺ primed NSCs (pNSCs) upregulated the levels of expression of the gene, being even higher in GLAST/EGFR⁺ activated NSCs (aNSCs) (Fig. S7C). Consistent with the in vivo immunostaining analysis, significant levels of expression of Cdkn1c were maintained in more differentiated cells, such as GLAST−/CD24+/PSA-NCAM+ neuroblasts and GLAST−/CD24−/EGFR+ neural progenitors (NPCs) (Fig. S7C). Thus, in order to characterize the function of p57 in the maintenance and differentiation of NSCs in the adult SVZ in vivo, a similar flow cytometry analysis of the different cell fractions from the SVZ was also performed in 3-month-old Cdkn1c-Gfap^mat and control mice (Belenguer et al., 2021a,b). Cdkn1c deletion resulted in reduced percentage of aNSCs (Fig. 4A), correlating with a reduced number of primary neurospheres obtained from the SVZ of Cdkn1c-Gfap^mat mice (Fig. 4B). Consequently, a decrease in the percentage of the neuroblast population was observed in Cdkn1c-deficient mice (Fig. 4C,D). Loss of p57 also resulted in a significant increase in the terminally differentiated astrocytic population in Cdkn1c-Gfap^mat mice compared with wild types (Fig. 4D and Fig. S7D). Importantly, a similar analysis in 8-month-old mice confirmed the loss of NB production and the induction of astrocyte differentiation in Cdkn1c-Gfap^mat mice, which were accelerated with age (Fig. 4D). This correlated with the exhaustion of neurosphere-forming cells isolated from the SVZ of 8- and 24-month-old Cdkn1c-Gfap^mat mice (Fig. 4B). No changes in the EGFR⁺ transit amplifying progenitor population or in the O4⁺ oligodendrocyte progeny was observed in Cdkn1c-deficient mice (Fig. S7E-H).

To further study the role of p57 in NSC differentiation in vivo, we performed brain electroporation in Cdkn1c-Gfap^control and Cdkn1c-Gfap^mat newborn pups at 3 days of age (Fig. 4E). This procedure takes advantage of the brain ventricles to allow the introduction of DNA into the lining NSCs (Fig. 4E). We used a Piggy BAC integrative vector carrying a red fluorescent protein (RFP) together with a GLAST-Transposase to permanently label the NSCs and their progeny with RFP (Fig. 4E). Sixty days after electroporation, we found a considerable number of RFP-labelled cells migrating through the RMS, and reaching and integrating with the OB in control mice (Fig. 4F-H). RFP⁺ cells were far less numerous in the RMS of Cdkn1c-Gfap^mat mice, resulting in fewer NBs in the OB (Fig. 4F-H). Notably, we found terminally differentiated astrocytes in the striatum of Cdkn1c-Gfap^mat mice, as indicated by the increased percentage of RFP⁺ cells that expressed the astrocytic marker S100β (Fig. 4I,J). These data confirmed a role for p57 in the differentiation process of adult NSCs also in vivo.

Finally, in order to functionally test whether p57 could also mediate the differentiation effects of IGF2 in vivo, we infused PBS solution or 100 µg ml⁻¹ of recombinant IGF2 for 7 days into the lateral ventricle (Ferrón et al., 2007; Mellott et al., 2014) of Cdkn1c-Gfap^mat and Cdkn1c-Gfap^control mice brains (Fig. 5A). We first injected mice with BrdU 15 days before pump implantation and euthanized them immediately after pump removal (Fig. 5A). In the SVZ, fast-proliferating transit-amplifying progenitors dilute out the BrdU, which is specifically retained in slowly proliferating NSCs (label-retaining cells, BrdU-LRCs), in newborn neurons and in newborn oligodendrocytes that cease to divide and undergo terminal differentiation soon after the injection in the OB or in the CC, respectively (Ferrón et al., 2007; Menn et al., 2006). The number of BrdU-LRC⁺ cells was significantly increased after IGF2 infusion in both wild-type and Cdkn1c-deficient mice (Fig. 5B-D), and more of them were positive for the proliferation antigen Ki67 (Fig. 5E-G), supporting a role for IGF2 in regulating the number of activated NSCs within the SVZ, although this effect was independent of p57.

Immunohistological analysis of cell populations within the SVZ of Cdkn1c-Gfap^control and Cdkn1c-Gfap^mat mice confirmed the increased astrocytic differentiation of cells in p57-deficient mice, as indicated by the higher proportion of GFAP⁺ cells that were also positive for S100β found in the SVZ of Cdkn1c-Gfap^mat mice compared with control mice (Fig. 6A,B). Notably, IGF2 infusion in the lateral ventricle of Cdkn1c-Gfap^control mice resulted in a higher proportion of GFAP/S100β⁺ cells and this increase was even higher in Cdkn1c-Gfap^mat (Fig. 6A,B). Importantly, more of these GFAP/S100β⁺ cells retained BrdU in Cdkn1c-Gfap^mat mice (Fig. 6A,B and Fig. S8A), suggesting an accumulation of inactive cells with mature astrocytic fate in the absence of p57.

Higher numbers of newly generated BrdU⁺ oligodendrocytes were also found in the CC of wild-type mice after IGF2 infusion (Fig. 6C). In addition, more densely populated DCX⁺ neuroblast chains in the RMS were found in the wild-type SVZ after IGF2 infusion (Fig. 6D,E; Fig. S8B,C), which resulted in higher numbers of newly generated BrdU⁺ neurons in the granular and periglomerular layers (PGL) of the OB (Fig. 6D,F and Fig. S8D). However, no effects of IGF2 infusion were observed in the number of BrdU⁺ newly generated neurons and oligodendrocytes in the absence of p57 (Fig. 6C-F and Fig. S8C,D), confirming that p57 is required for the differentiation effects of IGF2 also in vivo.

This study shows that IGF2 and Cdkn1c interact to regulate the differentiation potential of adult NSCs. Treatment of NSCs with recombinant IGF2 promotes a more differentiated phenotype of NSCs and a reduction of their stemness in vitro. The promotion of terminal differentiation of NSCs correlates with the induction of an increased level of the maternally expressed gene Cdkn1c through activation of the PI3K-Akt pathway. Consistently, intraventricular infusion of IGF2 in wild-type results in a final stimulation of NSCs differentiation. Importantly, infusion of IGF2 in Cdkn1c-deficient mice also induces proliferation; however, the absence of p57 in infused mice results in terminal differentiation of non-multipotent astrocytes at the expense of neurons and oligodendrocytes (Fig. 7). Our data present a model that illustrates the requirement of the correct interaction of IGF2 and the imprinted gene Cdkn1c to ensure long-term neurogenesis in the adult SVZ neurogenic niche.

Genomic imprinting controls the allelic expression of a subset of dose-sensitive genes in a parent of origin-dependent manner. Imprinted genes can be classified into canonical imprinted genes and genes that switch detectable expression of both parental alleles in different tissues and/or during development (Laukoter et al., 2020). Genes belonging to the second category include Igf2, which, although paternally expressed by most tissues, has also been shown to be biallelically expressed by the CP and vasculature in the SVZ niche and has a critical function in postnatal neurogenesis (Ferrón et al., 2015). In contrast, NSCs within the SGZ rely on canonically imprinted autocrine IGF2 expressed from only the paternal allele to regulate hippocampal stem cell function (Ferrón et al., 2015). Along the same lines, we have previously shown that loss of genomic imprinting of the paternally expressed gene Dlk1 is necessary to maintain NSC function in both neurogenic niches: the SVZ and the SGZ (Ferrón et al., 2011; Montalbán-Loro et al., 2021). Notably, many of these imprinted genes work coordinately and even antagonistically to control foetal growth (Daniel et al., 2015; Hoffmann et al., 2014). Thus, dynamic changes in genomic imprinting translate into biologically relevant functions.

The CKI p57 has been described to induce cell cycle exit, differentiation, neurite outgrowth and migration during embryonic cortical development (Imaizumi et al., 2020; Laukoter et al., 2020; Mairet-Coello et al., 2012), and is known to control NSC quiescence in hippocampal neurogenesis (Furutachi et al., 2013). More concretely, p57 has been shown to play a non-canonical role in regulating the activity of transcription factors implicated in neuronal differentiation interacting robustly with a subset of neurogenic bHLH proteins, including Mash1 (Ascl1), Neurod1 and Nex/Math2 (Neurod6), that inhibit neuronal differentiation, independently of the interaction with CDKs and cell-cycle exit (Joseph et al., 2009). Moreover, loss- and gain-of-function approaches suggest that p57 exerts a context-dependent repressive effect on neuronal but not astrocyte differentiation after mitogen withdrawal to allow proper glial differentiation (Joseph et al., 2009). In the SGZ of the hippocampus it has been shown that p57 is expressed at high levels in radial NSCs and that this CKI is required for NSC quiescence. Consistently, p57 diminishes in proliferating active radial NSCs (Furutachi et al., 2013). However, our immunohistochemical analysis in the SVZ shows that p57 is present in GFAP⁺ cells, whereas it is absent in neuroblasts within the SVZ. Interestingly, our data also show that p57 increases during neuronal and astroglial differentiation in vitro and becomes highly expressed in mature astrocytes and in more differentiated neuroblasts that reach the OB in vivo.

Given the absence of decisive markers to identify the neural stem cell population in vivo, part of the analysis of adult NSCs has relied on a neurospheres ex vivo assay obtained from the SVZ neurogenic niche (Bizy and Ferrón, 2014; Ferrón et al., 2007; Reynolds and Weiss, 1992). Moreover, recent FACS studies using different combinations of markers have led to the identification of quiescent and activated NSCs (Basak et al., 2018; Belenguer et al., 2021b; Daynac et al., 2016; Llorens-Bobadilla et al., 2015; Morizur et al., 2018). This strategy, in combination with our primary neurospheres assay in wild-type and p57 deficient cultures, has shown an impairment of self-renewal potential and a consistent lower number of activated NSCs in vivo after p57 deficiency.

In addition, in vitro treatment of NSCs with IGF2 promotes a more differentiated phenotype and a reduction of their sternness capacity. Notably, these IGF2-mediated effects occur along with the induction of an increased level of the maternally expressed gene Cdkn1c, which is not due to an alteration of the imprinting status of the gene but is regulated through activation of the PI3K-Akt pathway. As a result, an IGF2-mediated increase in the levels of p57 in vitro promotes neuronal and glial differentiation of NSCs. This was corroborated with the intraventricular infusion of IGF2 in wild-type mice, which stimulates the differentiation of NSCs within the adult SVZ. These results were consistent with previous studies in a murine model of Alzheimer's disease, in which IGF2 infusion increased the population of DCX⁺ cells in the adult hippocampus through stimulation of NGF, BDNF, NT3 and IGF1 (Mellott et al., 2014). Importantly, our study also revealed that, after infusion of IGF2 in Cdkn1c deficient mice, NSCs accumulate and differentiate more markedly into non-multipotent committed astrocytes, limiting neurogenesis and oligodendrogenesis. Notably, IGF2 infusion in wild-type mice also induces the proliferation of NSC within the neurogenic niche, effects that are also observed after the infusion of other growth factors, such as EGF, FGF, BDNF or GDNF (Kobayashi et al., 2006; Türeyen et al., 2005; Zigova et al., 1998). However, infusion of IGF2 in Cdkn1c-deficient mice resulted in similar rates of proliferation than in wild-type mice, indicating that effects of IGF2 on NSCs activation are not mediated by p57, consistent with its role as a cell cycle inhibitor. Therefore, our data present a model that illustrates the requirement of the correct interaction of the two imprinted genes Igf2 and Cdkn1c to balance neurogenesis, oligodendrogenesis and astrogliogenesis in the adult SVZ neurogenic niche (Fig. 7).

The human imprinted gene CDKN1C is the most frequently silenced or mutated gene in the imprinting genetic disorder Beckwith-Wiedemann syndrome (BWS); aberrant imprinting of IGF2 has also been shown to interact with CDK1C in this imprinting disorder (Caspary et al., 1999; Grandjean et al., 2000). Interestingly, both genes are located in the same imprinted gene cluster on human chromosome II (Grandjean et al., 2000). Moreover, alterations in both imprinted genes can lead to malignant cell transformation in different cancers (Bastaki et al., 2016; Nakashima et al., 2015). Therefore, our study highlights the importance of probing the relative contributions of imprinted genes and their inhibition to the regulation of cell signalling during adult neurogenesis in non-pathological conditions. In future studies, it will be important to investigate the precise roles of these genes in neurodevelopmental disease conditions.

All transgenic mice used in the study were maintained in the C57BL6 background. Gfap-cre [6.Cg-Tg(Gfap-cre)73.12Mvs/J] mice were obtained from the Jackson Laboratory and genotyped as described previously (Garcia et al., 2004). Gfap-Cre mice were generated using a 15 kb mouse Gfap promoter cassette containing all introns, promoter regulatory elements, exons, and 2 kb of 3′ and 2.5 kb of 5′ flanking regions of the mouse Gfap gene (Johnson et al., 1995). To avoid the expression of exogenous Gfap, a small fragment of the first exon of the Gfap gene was removed (Johnson et al., 1995). Cdkn1c^loxp/loxp mice were obtained from Riken BioResource Center (Japan) and contain LoxP sites flanking exons 2 to 4 (which include the entire coding region) of Cdkn1c gene (Furutachi et al., 2013; Matsumoto et al., 2011). To specifically ablate the maternal Cdkn1c allele in GFAP⁺ cells, we crossed female Cdkn1c^loxp/loxp mice with a male harbouring a Cre transgene under the control of the Gfap promoter (Gfap-Cre). Reciprocally, to ablate the paternal allele, female Gfap-Cre mice were crossed with male Cdkn1c^loxp/loxp mice (see Fig. S6A). Gfap-Cre mice were also crossed with a reporter mouse, which carry the β-galactosidase gene under the regulation of the ubiquitous ROSA26 promoter containing a loxP-flanked stop sequence (Gfap-Cre/LACZ). Control mice were generated by crossing mice without Cre-recombinase with ROSA26R mice (Gfap-control/LACZ). Animals were genotyped by PCR analysis of DNA extracted from ear-punch tissue and amplified with the primers for the presence of Cre-recombinase Cre-F and Cre-R (Table S1), and for the presence of LoxP sites p57 CKO-F, p57 CKO-R and p57 3′-Rv (Table S1). For imprinting studies, hybrid F1 offspring from reciprocal crosses between the subspecies Mus musculus domesticus and Mus musculus castaneus were used. Mice were maintained in a 12 h light/dark cycle with free access to food and water and according to the Animal Care and Ethics committee of the University of Valencia.

BrdU administration regimes have been previously detailed (Ferrón et al., 2007, 2011). Briefly, mice at 2-4 months of age were injected intraperitoneally with 50 mg of BrdU per kg of body weight every 2 h for 12 consecutive hours (seven injections in total). Two weeks after injections, animals were deeply anaesthetized with a mixture of ketamine and medetomidine, and fixed them to a stereotactic device (NeuroLab). We infused IGF2 (100 µg ml⁻¹; R&D Sytems) in PBS or PBS alone into the lateral ventricle for 7 days by means of an osmotic mini-pump (Alzet, model 1007D; flow rate 0.5 µl/h) coupled to a cannula for intracerebral delivery (see Fig. 5A). The stereotactic coordinates for targeting the lateral ventricle were anteroposterior −0.1 mm, mediolateral −0.8 mm from bregma and dorsoventral −3.0 mm from the skull surface. Seven days later, animals were transcardially perfused with 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer saline (PBS) pH 7.4 and brains were vibratome sectioned at 40 µm. For in vivo postnatal electroporation, a RFP Piggy BAC plasmid was injected together with a GLAST-Transposase plasmid in mouse pups at postnatal day 2 (P2) using a square wave electroporator (ECM 830 Square Wave Electroporation System, BTX) (Fabra-Beser et al., 2021; Mateos-White et al., 2020). Pups were deeply anesthetized by isoflurane inhalation and once the pedal reflex was lost, ≈1 μl of a 2 µg µl⁻¹ plasmid solution was injected into one of the lateral ventricles (Sonego et al., 2013). Five pulses of 95 ms at 95 V and spaced 950 ms were delivered, using forcep electrodes (Platinum Tweezertrode 5 mm Diameter, BTX) at 45° to target the subpalial SVZ (Mateos-White et al., 2020). Pups were reanimated on a heating pad and returned to mother's cage until culling day. Perfusion and tissue processing was carried out at P60 as previously described. To normalize the electroporation efficiency between animals, the percentage of RFP⁺ cells was calculated relative to RFP⁺ cells in the SVZ. All counting were carried out blinded to avoid the risk of bias.

For immunohistochemistry, sections were washed in PBS and blocked at room temperature for 1 h in PBS with 0.1% Triton X-100 supplemented with 10% Foetal bovine serum (FBS) and then incubated overnight at 4°C with primary antibodies (see Table S2). For BrdU detection, sections were pre-incubated in 2 N HCl for 20 min at 37°C and neutralized in 0.1 M sodium borate (pH 8.5) for 10 min. Detections were performed with fluorescent secondary antibodies incubated for 1 h at room temperature (see Table S3). For SVZ wholemounts, the lateral walls of the lateral ventricles were dissected free and the resulting wholemounts were fixed for 1.5 h in 4% PFA and washed overnight at 4°C in PBS. Wholemounts were washed three times in PBS containing 0.5% Triton X-100 for 15 min each, blocked for 2 h in 10% FBS and 2% Triton X-100 in PBS, then incubated for 48 h at 4°C with primary antibodies in the same blocking solution. After incubation with appropriate secondary antibodies, the stained walls were mounted with Fluorsave (Calbiochem). Nuclei were counterstained with 1 µg ml⁻¹ of DAPI. Primary and secondary antibodies and dilutions used are listed in Tables S2 and S3, respectively. Samples were captured and analysed with an Olympus FV10i confocal microscope (Olympus). For β-galactosidase staining, brain samples were fixed with 4% PFA in 0.1 M PBS (pH 7.4), 2 mM MgSO₄ and 5 mM EGTA for 30 min at 4°C and processed for vibratome sectioning. Sections were incubated in PBS with 2 mM MgCl₂, 5 mM K₃Fe(CN)₆, 5 mM K₄Fe(CN)₆, 0.01% sodium deoxycholate and 0.02% NP-40, and 1 mg ml⁻¹ X-Gal for 24 h at 37°C and washed several times in PBS.

Adult mice were sacrificed by cervical dislocation. To initiate each independent culture, the brains were dissected free and the regions containing the SVZ were isolated from each hemisphere and washed in Earle's balanced salt solution (EBSS; Gibco). Tissues were transferred to EBSS containing 1.0 mg ml⁻¹ papain (Worthington DBA), 0.2 mg ml⁻¹ L-cystein (Sigma), 0.2 mg ml⁻¹ EDTA (Sigma) and incubated for 30 min at 37°C. Tissue was then rinsed in EBSS, transferred to Dulbecco's modified Eagle's medium (DMEM)/F12 medium (1:1 v/v; Life Technologies) and carefully triturated with a fire-polished Pasteur pipette to a single cell suspension. Isolated cells were collected by centrifugation, resuspended in DMEM/F12 medium containing 2 mM L-glutamine, 0.6% glucose, 9.6 g ml⁻¹ putrescine, 6.3 ng ml⁻¹, progesterone, 5.2 ng ml⁻¹ sodium selenite, 0.025 mg ml⁻¹ insulin, 0.1 mg ml⁻¹ transferrin, 2 µg ml⁻¹ heparin (sodium salt, grade II; Sigma) and supplemented with 20 ng ml⁻¹ epidermal growth factor (EGF; Invitrogen) and 10 ng ml⁻¹ fibroblast growth factor (FGF; Sigma) (growth medium) (Belenguer et al., 2016; Ferrón et al., 2007). Neurospheres were allowed to develop for 6 days in a 95% air/5% CO₂ humidified atmosphere at 37°C. To estimate proliferation by the MTS assay, 15,000 cell/cm² were plated after Accutase disaggregation in differentiation medium for 2 DIV and the absorbance at 490 nm was measured using the CellTiter 96 Aqueous One Solution Cell Proliferation Assay following the manufacturer's protocol and the spectrophotometer Victor 3-1420 (PerkinElmer). For culture expansion, cells were plated at a relatively high density (10 cell/μl) and maintained for several passages. For bulk differentiation assays, 80,000 cell/cm² were seeded in Matrigel-coated coverslips and incubated 2 days in neurosphere culture medium without EGF. Medium was then changed with fresh medium without FGF but supplemented with 2% FBS for 5 more days (see Fig. 1A). For reactivation assays, 7 DIV differentiated NSCs were detached and replated in proliferation-promoting conditions (with mitogens) for 5 more days before counting the number of neurospheres formed. For primary cultures in the absence of mitogens, a previously described protocol was used (Ortega et al., 2011). Briefly, SVZ tissue was enzymatically digested with 0.7 mg ml⁻¹ hyaluronic acid (Sigma) and 1.33 mg ml⁻¹ trypsin (Sigma) in Hanks' Balanced Salt Solution (HBSS; Invitrogen) with 2 mM glucose for 30 min at 37°C. Tissue was rinsed in ice-cold medium composed of EBSS, 4% BSA (Sigma) and 20 mM HEPES. Tissue was then carefully triturated with a fire-polished Pasteur pipette to a single cell suspension. Cells were collected by centrifugation and washed first in ice-cold medium containing of 0.9 M sucrose (Sigma) and 0.5×HBSS, and second in ice-cold medium with 4% BSA and 2 mM HEPES. The final pellet was resuspended in culture medium containing DEMEM/F12 Glutamax (Invitrogen) supplemented with B27 (Invitrogen), 2 mM glutamine (Sigma) and 8 mM HEPES. Cells were plated on pretreated p24 wells with Poly-D-lysine (Sigma) at a density of 200-300 cells per mm² and cultured for 5 days. When indicated, cultures were treated with IGF2 (R&D Systems; 100 ng ml⁻¹), the PI3 K/Akt inhibitor LY294002 (Sigma; 50 µM) or the MEK1/2 inhibitor III PD0325901 (Millipore; 1 µM) at the time of plating. IGF2 treatments were performed in insulin-free medium. Cultures were fixed for staining at specific times of differentiation with 4% PFA and 0.1 M PBS for 15 min, and performed immunocytochemistry as described previously (Belenguer et al., 2016). Primary and secondary antibodies and dilutions used are listed in Tables S2 and S3, respectively. DAPI (1 µg ml⁻¹) was used to counterstain DNA. Laser settings were first established on wild-type tissue or untreated samples, and similar regions of interest were acquired in an Olympus FV10i confocal microscope. Maximal projection images were generated and the mean grey intensities of nuclear marker p57 were measured with ImageJ/Fiji software. Intensities are represented as frequency histograms normalized to the maximum count in each comparison.

Cell analysis by flow cytometry were done following previously established protocols (Belenguer et al., 2021b). The SVZ of both brain hemispheres from each mouse were minced and enzymatically digested using the Neural Tissue Dissociation kit (T) in a GentleMACS Octo Dissociator with heaters (Miltenyi). Digested pieces were mechanically dissociated by pipetting up and down 20 times through a plastic Pasteur pipette. Cell suspension was filtered through a 40 mm nylon filter. Cells were pelleted (300 g, 10 min), resuspended in 100 ml blocking buffer (HBSS without calcium and magnesium, 10 mM HEPES, 2 mM EDTA, 0.1% glucose and 0.5% BSA) and incubated with different combinations of specific primary antibodies at 4°C for 30 min (see Table S2). Cells were also incubated with EGF-Alexa488 or EGF-Alexa647 (1:300; Molecular Probes) at 4°C for 30 min. After washing with 1 ml of blocking buffer, labelled samples were centrifuged (300 g for 10 min at 4°C) and resuspended in 0.5 ml of blocking buffer for analysis. All analyses were performed with a LSR-Fortessa cytometer (Becton Dickinson) with 350, 488, 561 and 640 nm lasers. Dead cells were excluded by staining with 0.1 mg ml⁻¹ DAPI before analysis.

RNAs were extracted with RNAeasy mini kit (Qiagen) including DNase treatment, following the manufacturer's guidelines. For quantitative PCR, 1 µg of total RNA was reverse transcribed using random primers and RevertAid H Minus First Strand cDNA Synthesis kit (Thermo Scientific), following standard procedures. Thermocycling was performed in a final volume of 10 µl containing 1 µl of cDNA sample (diluted 1:20), and the reverse transcribed RNA was amplified by PCR with appropriate Taqman probes (see Table S4). Quantitative PCR was used to measure gene expression levels normalized to Gapdh, the expression of which did not differ between the groups. qPCR reactions were performed in a Step One Plus cycler with Taqman Fast Advanced Master Mix (Applied Biosystems). To study specific expression of the paternal and maternal alleles, PCR was carried out for the Cdkn1c gene in NSCs derived from adult F1 mice hybrid offspring from Mus musculus domesticus (C57BL6/J) and Mus musculus castaneus (CAST/EiJ) mice. We have previously identified a single-nucleotide polymorphism (SNP) between the two subspecies at the Cdkn1c gene (Fig. S4B). The SNP was a ‘T’ nucleotide in BL6 and a ‘C’ nucleotide in CAST mice, and was located at position 1382 in the Cdkn1c transcript (Cdkn1c sequence NM_001354981.1). The PCR product was purified and sequenced by Eurofins Genomics and analysed using the Gene Align programme.

DNA methylation level was quantified using bisulfite conversion and pyrosequencing. The DNA was bisulfited-converted using EZ DNA Methylation-Gold kit (Zymo Research) in accordance with the manufacture's protocol. Specifically, for Cdkn1c promoter, bisulfite-converted DNA was amplified by PCR with specific primer pairs: Cdkn1cMETH-F and Cdkn1cMETH-R (Table S1). PCRs were carried out in a 20 µl volume, with 2 U HotStar Taq polymerase (Qiagen), PCR buffer 10× (Qiagen), 0.2 mM dNTPs and 400 mM primers. PCR conditions were: 96°C for 5 min, followed by 39 cycles of 94°C for 30 s, 54°C for 30 s and 72°C for 1 min. For pyrosequencing analysis, a biotin-labelled primer was used to purify the final PCR product using sepharose beads. The PCR product was bound to Streptavidin Sepharose High Performance (GE Healthcare), purified, washed with 70% ethanol, denatured with 0.2 N NaOH and washed again with 10 mM Tris-acetate. Pyrosequencing primer (Cdkn1cMETH-seq; 400 mM) was then annealed to the purified single-stranded PCR product and pyrosequencing was performed using the PyroMark Q96MD pyrosequencing system using PyroMark reactives (Qiagen).

Cells were lysed in cold RIPA buffer. Total protein concentration was determined using the BCA system (Pierce). Equal amounts (30 mg) of protein were loaded on polyacrylamide gels for SDS–polyacrylamide gel electrophoresis. Proteins were transferred to polyvinylidene difluoride (PVDF) membranes and immunoblots were carried out using primary antibodies (Table S2) followed by incubation with appropriate secondary horseradish peroxidase-conjugated antibodies (Table S3) and chemoluminiscent detection (Western Lightning, PerkinElmer). All antibodies were diluted in PBS containing 5% semi-skimmed milk and 0.1% Tween-20. Proteins were revealed using Lightning Plus ECL (Perkin Elmer) and the bands were analysed by densitometry using ImageJ (NIH) software.

All statistical tests were performed using the GraphPad Prism Software, version 7.00 for Windows. Data were first tested for normality with a Shapiro-Wilk test. The significance of the differences between groups were evaluated with adequate statistical tests for each comparison. For data that passed normality tests: when analysing only one variable, a paired t-test was used for comparing two groups and one-way ANOVA followed by Tukey's post-hoc test for three or more groups. When two variables were analyse, two-way ANOVA followed by Tukey post-hoc test was used. For data groups that did not pass normality, Wilcoxon or Mann–Whitney non-parametric tests were performed, depending on whether samples were paired or not, respectively. For variables with more than two categories, Kruskal–Wallis or Friedman tests (for unpaired or paired data, respectively) were used followed by a Benjamini, Krieger and Yekutieli post-hoc test. For temporal measures that passed the normality test, lineal regression analysis was performed. When comparisons were performed with relative values (percentages), data were previously normalized by using arcsin root transformation. Values of P<0.05 were considered statistically significant. Data are presented as the mean±s.e.m.; the number of experiments performed with independent cultures or animals (n) and P-values are indicated in the figures. Box and whisker plots show the mean (+), median (line in box), and maximum and minimum values (whiskers).

We thank Dr Anne Ferguson-Smith and her group for technical support and discussion of the data. We also thank David De Agustín for his technical help with primary cultures and Central Service for Experimental Research (SCSIE) from the University of Valencia for its contributions to the work.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.200563.reviewer-comments.pdf