Summary
The expansion of the neocortex during mammalian evolution has been linked to an enlargement of the subventricular zone during cortical development and an increase in the proliferation of the basal progenitors residing therein. Here, we explored a potential role of YAP, the major downstream effector of the Hippo signaling pathway, in proliferation of basal progenitors. We show that YAP expression and activity are high in ferret and human basal progenitors, which are known to exhibit high proliferative capacity, but low in mouse basal progenitors, which lack such capacity. To induce YAP activity in mouse basal progenitors, we expressed a constitutively active YAP (CA-YAP). This resulted in an increase in proliferation of basal progenitor. In addition, CA-YAP expressing mouse basal progenitors promoted the production of upper-layer neurons. To investigate if YAP is required for the proliferation of basal progenitors, we pharmacologically interfered with the function of YAP in the developing ferret and human neocortex. This resulted in a decrease of cycling basal progenitors. In concert, genetical interference with the function of YAP in ferret developing neocortex resulted in decreased abundance of basal progenitors. Together, our data indicate that YAP promotes the proliferation of basal progenitors and suggest that changes in YAP activity levels contributed to the evolutionary expansion of the neocortex.
Introduction
The neocortex, the seat of higher cognitive functions, undergoes substantial expansion during the evolution of certain mammalian brains such as human. A major factor in neocortical expansion, notably regarding the increase in the number of cortical neurons, is thought to be an increased proliferative capacity of cortical neural progenitor cells (cNPCs) (Fietz and Huttner, 2011; Florio and Huttner, 2014; Geschwind and Rakic, 2013; Lui et al., 2011; Namba and Huttner, 2017; Rakic, 2009; Wilsch-Bräuninger et al., 2016).
Two principal classes of cNPCs exist in the developing neocortex, referred to as apical progenitors (APs) and basal progenitors (BPs) (Florio and Huttner, 2014; Lui et al., 2011; Namba and Huttner, 2017). The defining feature of APs is that they undergo mitosis at the ventricular (apical) surface of the ventricular zone (VZ), the primary germinal zone where the AP cell bodies reside (Florio and Huttner, 2014; Namba and Huttner, 2017). At the onset of neurogenesis, the major AP cell type are apical (or ventricular) radial glia (aRG) (Fietz and Huttner, 2011; Florio and Huttner, 2014; Lui et al., 2011; Namba and Huttner, 2017). The defining feature of BPs is that they undergo mitosis away from the apical surface, typically in a secondary germinal zone called the subventricular zone (SVZ) where the BP cell bodies reside (Haubensak et al., 2004; Miyata et al., 2004; Noctor et al., 2004). BPs originate from APs, delaminate from the apical surface, migrate beyond the VZ, and thus form the SVZ. There are two main types of BPs, basal intermediate progenitor (bIPs) and basal (or outer) radial glia (bRG). In contrast to aRG, which are epithelial cells exhibiting apical-basal polarity with contact to the ventricle and (in the canonical form) to the basal lamina, bIPs are nonepithelial cells that no longer exhibit apical-basal polarity and that have lost contact to both the ventricle and the basal lamina (Haubensak et al., 2004; Miyata et al., 2004; Noctor et al., 2004). bRG, however, though lacking an apical process that reaches the ventricle, retain epithelial features in that they (in the canonical form) possess a basal process that contacts the basal lamina (Fietz et al., 2010; Hansen et al., 2010; Reillo et al., 2011).
BP composition and proliferative capacity may differ greatly between a developing lissencephalic neocortex (e.g. mouse) and a developing gyrencephalic neocortex (e.g. ferret and human). In the developing mouse neocortex BPs mostly comprise bIPs that typically undergo neurogenic consumptive divisions giving rise to two neurons; compared to the aRG they derive from, these neurogenic bIPs characteristically upregulate the transcription factor Tbr2 and downregulate the transcription factor Sox2 (Haubensak et al., 2004; Miyata et al., 2004; Noctor et al., 2004). Only a minor portion of mouse BPs are bRG, and their proliferative potential is limited (Shitamukai et al., 2011; Wang et al., 2011). In contrast, in the developing ferret and human neocortex, the majority of BPs are proliferative bRG (highly proliferative in human) that do not express Tbr2 but rather maintain expression of Sox2 (Fietz et al., 2010; Hansen et al., 2010; Reillo et al., 2011). Moreover, as first shown in a seminal contribution for the developing monkey neocortex (Smart et al., 2002), the SVZ in a gyrencephalic neocortex is characteristically split into two morphologically distinct zones, an inner SVZ (iSVZ) and an outer SVZ (oSVZ). Of note, the evolutionary expansion of the neocortex has been linked to an increase in the proliferative capacity and abundance of BPs in the oSVZ, especially of bRG (Dehay et al., 2015; Fietz and Huttner, 2011; Lui et al., 2011; Namba and Huttner, 2017). However, the molecular players that differentially promote the proliferative capacity of BPs across the various mammalian species remain largely unknown.
To gain insight into this issue, we examined a major molecular mechanism known to regulate organ size, the Hippo/YAP signaling pathway (Barry and Camargo, 2013; Camargo et al., 2007; Lian et al., 2010; Yu et al., 2015). The core of Hippo/YAP signaling is the YAP protein, whose transcriptional activity is regulated by phosphorylation. Phosphorylated YAP (phospho-YAP) is retained in the cytoplasm, whereas dephosphorylated YAP (active YAP) translocates to the nucleus and activates the expression of genes linked to cell proliferation (Zanconato et al., 2015; Zhao et al., 2010; Zhao et al., 2007; Zhao et al., 2008). Recent studies dissecting the roles of the cadherin family members Dchs1 and Fat4 (Cappello et al., 2013) and of the tumor suppressor neurofibromatosis 2 (Lavado et al., 2013; Lavado et al., 2014) and investigating heterotopia formation (Saito et al., 2017) in mouse brain development have reported that YAP promotes the proliferation of mouse APs. These studies, however, have not focused on a potential role of YAP in regulating the proliferation of BPs, nor have they addressed the question whether differences in YAP activity may underlie the differences in the proliferative capacity of BPs across various mammalian species in the context of the evolutionary expansion of the neocortex.
In the present study, we have identified differences in YAP expression and YAP activity between the developing lissencephalic mouse and gyrencephalic ferret and human neocortex that match the differences in the proliferative capacity of BPs across these species. Enhancing YAP activity in mouse BPs induced their proliferation and therefore shifted their fate from neurogenic to proliferative. In contrast, inhibition of endogenous YAP activity by verteporfin or by administration of a dominant-negative YAP reduced BP proliferation in developing ferret and human neocortex. Taken together, these findings suggest that an upregulation of YAP activity contributed to the increased proliferative capacity of BPs in the context of the evolutionary expansion of the neocortex.
Results
BPs with high proliferative capacity, which are abundant in embryonic ferret and fetal human neocortex but are lacking in embryonic mouse neocortex, show high YAP expression
To explore whether YAP is differentially expressed between the principal germinal zones of developing mouse and human neocortex, the VZ vs. the SVZ, and between the major classes of cNPCs therein, the APs vs. BPs, respectively, we first analyzed the levels of Yap/YAP mRNA in these germinal zones and cNPC classes that have been determined in previously published transcriptome analyses (Figure S1) (Fietz et al., 2012; Florio et al., 2015). Yap/YAP mRNA was robustly expressed in the VZ of both embryonic day (E) 14.5 mouse and 13 week-post-conception (wpc) human neocortex (Figure S1A), and accordingly in mouse and human aRG (Figure S1B), the major AP type (Namba and Huttner, 2017). Strikingly, Yap/YAP mRNA was found to be expressed in the human iSVZ and oSVZ, but not the mouse SVZ (Figure S1A), and in human bRG, but not mouse BPs (Figure S1B). Given that both human and mouse APs, and human, but not mouse, BPs are endowed with the ability to expand their population size by cell proliferation (Namba and Huttner, 2017), these data provided a first indication that the proliferative capacity of cNPCs, notably of BPs, may be linked to the expression of YAP. Consistent with this notion, no significant Yap/YAP mRNA expression was detected in the mouse and human cortical plate (CP) (Figure S1A), a zone of the developing neocortex lacking cell proliferation, nor in post-mitotic neurons (Figure S1B; note that the YAP mRNA in the human neuron fraction reflects the presence of bRG in G1 in this fraction (Florio et al., 2015)).
Further evidence in support of a potential role of YAP in cNPC proliferation was provided by the comparison of Yap mRNA levels between Tis21-GFP–negative and Tis21-GFP–positive aRG of E14.5 mouse neocortex, determined previously (Florio et al., 2015) using the Tis21-GFP knock-in (Tis21tm2(Gfp)Wbh) mouse line (Haubensak et al., 2004). This revealed a two-fold higher Yap mRNA level in Tis21-GFP–negative aRG as compared to Tis21-GFP–positive aRG (Figure S1C) (Florio et al., 2015), in line with both cNPC types being capable of self-renewal but only the former cNPC subpopulation being capable of expansion by proliferation (Attardo et al., 2008; Haubensak et al., 2004). Finally, comparison of mRNA levels between a prospective gyrus vs. a prospective sulcus of developing (postnatal day (P) 2) ferret neocortex, available in a previously published transcriptome dataset (De Juan Romero et al., 2015), showed that the Yap mRNA level was higher in the oSVZ of the prospective gyrus than the prospective sulcus (Figure S1D), consistent with the notion that a relative increase in cNPC proliferation in this germinal zone contributes to gyrus formation (Hansen et al., 2010; Reillo et al., 2011; Wang et al., 2011). Taken together, these Yap/YAP mRNA data raised the possibility that YAP may not only have a role in the proliferation of APs, as previously shown for embryonic mouse neocortex (Lavado et al., 2013; Lavado et al., 2014), but that differences in the level of active YAP may underlie the differences in the proliferative capacity of mouse vs. ferret and human BPs.
We therefore examined the expression of the YAP protein in embryonic mouse, embryonic ferret and fetal human neocortex by immunofluorescence (Figure 1A-C, F-H). Consistent with the mRNA expression data (Figure S1A), YAP immunoreactivity was overt in the E14.5 mouse, E36 ferret and 14 wpc human VZ and in the ferret and human SVZ, notably the oSVZ, but was low in the mouse SVZ (Figure 1A-C). In the case of the embryonic ferret oSVZ, the YAP immunostaining revealed cells exhibiting a basal process (Figure 1B’), suggesting that they were bRG.
YAP-positive BPs in embryonic ferret and fetal human neocortex express Sox2 rather than Tbr2, and their YAP is nuclear and active
For YAP to be able to promote the expression of genes linked to proliferation, it needs to be nuclear and unphosphorylated (Zanconato et al., 2015; Zhao et al., 2010; Zhao et al., 2007; Zhao et al., 2008). Interestingly, in line with a potential role of YAP in BP proliferation, the 13-14 wpc human SVZ contained the highest percentage of YAP-positive nuclei (50%), followed by the E36 ferret SVZ (15%), whereas the E14.5 mouse SVZ exhibited only low levels of YAP-positive nuclei (4%) (Figure 1D).
We sought to obtain direct evidence that the higher level of nuclear YAP protein seen in embryonic ferret and fetal human SVZ as compared to mouse SVZ (Figure 1D) is linked to the increased proliferative capacity of ferret and human BPs vs. mouse BPs. To this end, we compared the expression of YAP with that of Sox2 (Figure 1A-C), an indicator of proliferative capacity (Hansen et al., 2010; Reillo et al., 2011; Wang et al., 2011). Specifically, we asked whether SVZ nuclei positive for Sox2 are also positive for YAP. Indeed, the vast majority of the Sox2-positive BP nuclei in the E36 ferret and 13-14 wpc human SVZ were YAP-positive (70% and 86%, respectively), whereas this was the case for only a minor proportion of the Sox2-positive nuclei in the E14.5 mouse SVZ (4%) (Figure 1E).
We next investigated whether the YAP protein in the Sox2-positive BP nuclei was in an active form, i.e. in a dephosphorylated state. A first, qualitative indication that this was the case was obtained by comparing the immunofluorescence signals for total YAP and phospho-YAP in Sox2-positive BP nuclei in the SVZ of human 11 wpc neocortex (Figure S2A). Relative to the cytoplasmic signals for total YAP and phospho-YAP, respectively, this comparison showed a much lower signal for nuclear phospho-YAP than nuclear total YAP, suggesting that most nuclear YAP in fetal human neocortical BPs was in the dephosphorylated state and hence active. We therefore used this approach to compare the levels of active YAP in Sox2-positive BP nuclei in the SVZ of mouse E13.5, ferret E36 and human 11 wpc neocortex. To this end, we subtracted the internal standard-adjusted immunofluorescence signal for nuclear phospho-YAP from that of nuclear total YAP to obtain information about nuclear dephosphoYAP levels (see legend to Figure S2B for details). This revealed substantially higher levels of active YAP in the Sox2-positive BP nuclei of ferret E36 and human 11 wpc neocortex than mouse E13.5 neocortex (Figure S2B).
We complemented these data by determining the effect of lambda protein phosphatase treatment of cryosections on the YAP immunofluorescence signals in Sox2-positive BP nuclei in the SVZ of mouse E14.5, ferret E36 and human 12-13 wpc neocortex. For YAP immunofluorescence, two rabbit monoclonal antibodies were used together, one recognizing YAP irrespective of serine127 phosphorylation (total YAP) and the other recognizing the serine127 phosphorylation site when phosphorylated (phospho-YAP). Hence, in the control, i.e. without lambda protein phosphatase treatment, the YAP immunofluorescence signal reflects the binding of primary antibodies to 1-2 sites, depending on whether serine127 is phosphorylated or not. In the case of mouse, lambda protein phosphatase treatment reduced the YAP immunofluorescence signal by ≈40% compared to control (Figure S2C). As lambda protein phosphatase treatment resulted in complete dephosphorylation of serine127 (see Methods), this ≈40% reduction suggested that about two-thirds (40/60) of the YAP protein in the Sox2-positive BP nuclei of mouse E14.5 neocortex was in phosphorylated form, and thus inactive. In contrast, in the case of ferret and human, lambda protein phosphatase treatment reduced the YAP immunofluorescence signal by only ≈20% and 10%, respectively, compared to the controls (Figure S2C), consistent with only ≈25% (20/80) and ≈10% (10/90) of the YAP protein in the Sox2-positive BP nuclei of ferret E36 neocortex and human 12-13 wpc neocortex, respectively, being in phosphorylated form. These data therefore corroborated our finding that there are markedly higher levels of active YAP in the Sox2-positive BP nuclei of embryonic ferret and fetal human than embryonic mouse neocortex.
Results essentially opposite to the occurrence of YAP in Sox2-positive BP nuclei were obtained when we compared the expression of YAP with that of Tbr2 (Figure 1F-H), which in embryonic mouse neocortex has been established as a marker of differentiating BPs, notably of neurogenic bIPs (Englund et al., 2005; Kowalczyk et al., 2009; Pontious et al., 2008). Thus, only a minor proportion (<16%) of the Tbr2-positive nuclei in the E14.5 mouse, E36 ferret and 11 wpc human SVZ were also positive for YAP (Figure 1K). Similarly, only ≈20% of the Tbr2-positive nuclei in the VZ of developing mouse, ferret and human neocortex, which constitute newborn BPs (Arai et al., 2011), were YAP-positive (Figure 1J). In contrast, the overwhelming majority (>75%) of the Tbr2-negative nuclei in the VZ of developing mouse, ferret and human neocortex, which constitute APs capable of expansion by proliferation, were YAP-positive (Figure 1I).
Taken together, these data suggest that the proliferative capacity of cNPCs, notably of BPs, correlates with the occurrence of nuclear YAP activity, with a high percentage of proliferating APs in developing mouse, ferret and human neocortex and of proliferating BPs in embryonic ferret and fetal human neocortex exhibiting nuclear (and mostly active) YAP, whereas this was the case for only a low percentage of the proliferating BPs in embryonic mouse neocortex. Conversely, only a low percentage of the nuclei of differentiating (rather than proliferating) BPs in the developing mouse, ferret and human neocortex exhibited nuclear YAP (and hence nuclear YAP activity, if any).
Conditional expression of constitutively active YAP in embryonic mouse neocortex decreases Tbr2-positive BPs and increases Sox2-positive BPs
To examine if YAP activity is functionally linked to the proliferative capacity of BPs, we conditionally expressed a constitutively active YAP (CA-YAP) in BPs of the embryonic mouse neocortex, which lack proliferative potential and in which YAP expression normally is very low (Figure 1A,D,E). In the CA-YAP, two serine residues were replaced by alanine (S112A and S382A), mutations that have been shown to stabilize and to increase the nuclear localization of the YAP protein (Camargo et al., 2007; Dong et al., 2007; Zhao et al., 2010). The plasmid used for CA-YAP expression (referred to as CA-YAP–expressing plasmid) contained a strong constitutive promoter (CAGGS) driving a floxed membrane EGFP followed by the CA-YAP and an IRES-linked nuclear RFP reporter (Figure 2A left). Expression of CA-YAP and RFP upon Cre-mediated excision of the floxed EGFP was validated by transfection of HEK293T cells (Figure S3A) The same plasmid but lacking the CA-YAP module was used as control (Figure 2A left). Conditional expression of CA-YAP in mouse BPs was achieved by in utero electroporation (IUE) of the neocortex of E13.5 embryos of the Tis21-CreERT2 knock-in (Btg2tm1.1(cre/ERT2)Wbh) mouse line (Wong et al., 2015) (Figure 2A right). In embryos of this mouse line, the expression of tamoxifen-activated Cre follows that of Tis21, that is, is specific for BP-genic aRG and BPs (Wong et al., 2015). Conditional CA-YAP expression in mouse neocortical BPs by this approach was validated by analysis of E14.5 embryos (Figure S3B,C). Furthermore, conditional CA-YAP expression by this approach drove expression of a previously identified YAP target gene, CTGF (Malik et al., 2015; Zanconato et al., 2015; Zhao et al., 2008), in BPs three days after IUE (Figure S3B,D; for details see supplementary text).
We first examined the effects of conditional CA-YAP expression on BP fate, by analyzing Tbr2 and Sox2 expression 2 days after IUE of Tis21-CreERT2 mouse embryos (Figure 2A-G). This revealed, among the RFP-positive progeny of the targeted cells, a marked reduction in the proportion of Tbr2-positive cells in the VZ and SVZ (Figure 2D) and a striking increase in the proportion of Sox2-positive cells in the SVZ but not VZ (Figure 2G), compared to control. Hence, increasing YAP activity in neocortical BPs of embryonic mouse is sufficient to change their fate, interfering with their differentiation to a neurogenic progenitor type and promoting their neural stem cell properties.
Conditional CA-YAP expression in embryonic mouse neocortex increases proliferative capacity of BPs
Next, we directly examined the potential effects of conditional CA-YAP expression in embryonic mouse neocortex on BP proliferation, using three distinct approaches. First, we performed immunofluorescence for the cycling cell marker Ki67 2 days (Figure 3A-C) and 3 days (Figure S4A,B) after IUE. Conditional CA-YAP expression doubled to tripled the proportion of Ki67-positive cells among the RFP-positive progeny of the targeted cells in the SVZ (Figure 3D, Figure S4C). Concomitant with this effect, conditional CA-YAP expression resulted in an altered distribution of the RFP-positive cells across the various zones of cortical wall, with a greater proportion of these cells in the VZ and a lesser proportion in the IZ and CP (Figure S5). These data raise the possibility of a relationship between the CA-YAP–induced progenitor proliferation and the reduced migration of the RFP-positive progeny of the targeted cells beyond the germinal zones to the basal region of the cortical wall.
Second, we performed immunofluorescence for the mitotic cell marker phosphovimentin (pVIM). Conditional CA-YAP expression increased the proportion of pVIM-positive cells among the RFP-positive progeny of the targeted cells in the SVZ and (albeit not statistically significant) VZ (Figure S4D-F). Consistent with this, conditional CA-YAP expression increased the abundance of mitotic, pVIM-positive BPs and (albeit not statistically significant) APs (sum of RFP-positive and -negative mitoses) (Figure S4D,E,G).
Third, we carried out a cell-cycle re-entry assay. To this end, we performed a 5’-ethynyl-2’deoxyuridine (EdU) pulse labeling 24 h after IUE, followed by Ki67 immunofluorescence 24 h later (Figure 3E). Given the cell-cycle parameters of Tis21-GFP–positive APs and BPs at this stage of embryonic mouse neocortex development (Arai et al., 2011), the latter time interval is sufficient for the incorporated EdU to become inherited by daughter cells. Thus, EdU- and RFP-positive cells that are also positive for Ki67 are daughter cells that re-entered the cell-cycle (Figure 3E). Using this assay, we found that conditional CA-YAP expression doubled the cell-cycle re-entry of BPs in the SVZ (Figure 3, E-H). Taken together, the results of these three lines of investigation demonstrate that conditional CA-YAP expression in embryonic mouse neocortex increases BP proliferation.
Conditional CA-YAP expression in embryonic mouse neocortex results in reduced deep-layer neuron and increased upper-layer neuron generation
Given that BPs generate most cortical neurons (Florio and Huttner, 2014; Lui et al., 2011), we explored the potential consequences of the CA-YAP–induced increase in BP proliferation for neuron generation in the embryonic mouse neocortex. To this end, we performed immunostaining for Tbr1, a deep-layer neuron marker, and Satb2, an upper-layer neuron marker 4 days after IUE (Figure 4A-E). Conditional CA-YAP expression reduced the proportion of Tbr1-positive neurons among the RFP-positive progeny of the targeted cells in the IZ and CP to almost half of control (29% vs. 44%, Figure 4F), and caused a small but statistically significant increase in the proportion of Satb2-positive neurons (from 76% to 87%, Figure 4G).
Considering that the pool size of the RFP-positive progeny in the IZ and CP, four days upon IUE (E13-E17), is not much affected by the conditional CA-YAP expression compared to control (Figure S6), these data are consistent with the notion that the changes in the proportions of specific types of neurons among the RFP-positive progeny reflect the changes in neuron generation in embryonic mouse neocortex. Specifically, on the one hand, the CAYAP–induced increase in BP proliferation initially results in a reduced production of deep-layer neurons because upon conditional CA-YAP expression a certain proportion of the mouse BPs has been induced to undergo symmetric proliferative divisions and hence a lesser proportion of the mouse BPs is available to undergo the symmetric consumptive divisions that generate neurons. On the other hand, the CA-YAP–induced increase in BP proliferation eventually results in an increased mouse BP pool size that at later stages of cortical neurogenesis gives rise to more upper-layer neurons.
Pharmacological inhibition of YAP activity reduces mitotic BP abundance in embryonic ferret and fetal human neocortex
Having established that mimicking a ferret- or human-like expression of active YAP in BPs of embryonic mouse (i.e. lissencephalic) neocortex suffices to increase their proliferation, we next investigated whether the presence of active YAP in BPs of developing ferret and human (i.e. gyrencephalic) neocortex is a necessary requirement for their proliferation. To this end, we applied an inhibitor of YAP, verteporfin (Brodowska et al., 2014; Liu-Chittenden et al., 2012; Song et al., 2014; Wang et al., 2016), in an ex vivo free-floating tissue (FFT) culture system (Schenk et al., 2009), using E33-34 ferret and 11-13 wpc human neocortical tissue (Figure 5). Verteporfin is a small molecule that inhibits the association of YAP with TEAD transcription factors and thereby prevents the expression of genes linked to cell proliferation (Liu-Chittenden et al., 2012).
Treatment of embryonic ferret and fetal human neocortical FFT cultures with 1 µM verteporfin for 48 h decreased the abundance of basal mitoses, identified by phosphohistone H3 (PH3) immunofluorescence (Figure 5A,C), approximately 4-fold and 2-fold, respectively (Figure 5B,D). In ferret, verteporfin treatment also resulted in a decrease, albeit not statistically significant, in apical mitoses (Figure 5B).
Dominant-negative YAP expression reduces mitotic BP abundance in embryonic ferret neocortex
We sought to obtain corroborating in vivo evidence for an essential role of YAP activity in BP proliferation in developing gyrencephalic neocortex. To this end, we used a dominantnegative YAP construct (DN-YAP) to block its transcriptional co-activator function (Nishioka et al., 2009; Sudol et al., 2012) (see Supplemental Experimental Procedures for details). In this DN-YAP construct, the transactivation domain of mouse YAP is replaced with the engrailed domain, a Drosophila repressor of transcription (Nishioka et al., 2009), and nuclear localization of DN-YAP is ensured by replacing YAP serine112 with alanine (Hao et al., 2007; Zhao et al., 2007). Forced expression of DN-YAP creates multiple copies of DN-YAP that outcompete endogenous wtYAP in binding to TEAD transcription factors, which causes downregulation of YAP-driven genes linked to proliferation (Nishioka et al., 2009; Zanconato et al., 2015). We chose embryonic ferret neocortex to examine the effects of DN-YAP expression on BP proliferation.
We performed IUE of ferret embryos to deliver the DN-YAP construct into the developing ferret neocortex (Kawasaki et al., 2012; Kawasaki et al., 2013). Specifically, we coelectroporated the ferret dorsolateral neocortex with either CAGGS-empty vector plus CAGGS-EGFP or with CAGGS-DN-YAP plus CAGGS-EGFP. IUE was performed at E33, the stage which corresponds to mid-neurogenesis in the ferret. Ferret embryos were harvested 2 days later, at E35, to allow for sufficient time for the targeted APs to complete their cell-cycle (Turrero Garcia et al., 2015) and give rise of BPs (Figure 6A). To confirm the expression of DN-YAP, we performed immunostaining using a YAP antibody that recognizes both endogenous YAP and DN-YAP and compared the level of YAP immunoreactivity in GFP-positive cells upon DN-YAP expression to that of control. Upon DN-YAP expression, many of the GFP-positive cells showed a higher level of YAP immunoreactivity, especially in the SVZ, suggesting that the DN-YAP was successfully expressed in BPs (Figure 6B, compare lower and upper rows).
Analysis of the GFP-positive progeny of the targeted APs revealed that expression of DN-YAP decreased slightly the abundance of mitotic APs, and massively that of mitotic BPs, as revealed by PH3 immunofluorescence (Figure 6B,C). Similar results were obtained by quantitation of pVIM-positive mitoses (Figure 6D,E).
There are two main types of BPs, (i) those lacking processes at mitosis, i.e. bIPs, and (ii) those bearing radial processes at mitosis, i.e. bRG (Fietz et al., 2010; Kelava et al., 2012; Reillo et al., 2011). To examine whether any of these two types of BPs is preferentially affected by DN-YAP expression, we analyzed the processes as revealed by pVIM staining in GFP-positive mitotic BPs in the SVZ. We observed a similar decrease in process-lacking and process-bearing BPs, which suggests that the effect of DN-YAP expression is non-selective with regard to the BP population type, affecting equally bIP and bRG (Figure 6F).
Given that conditional CA-YAP expression in embryonic mouse neocortex resulted in a specific increase in Sox2-positive BPs (Figure 2G), we explored whether inhibition of YAP activity in embryonic ferret neocortex would affect the Sox2-positive pool of BPs in the SVZ. Indeed, DN-YAP expression reduced the proportion of Sox2-positive BPs among the GFP-positive progeny of the targeted cells in the SVZ (Figure 6G,H).
Taken together, our data using verteporfin and DN-YAP indicate that YAP activity is necessary for the proliferation of BPs in developing gyrencephalic neocortex in vivo.
Discussion
The present study demonstrates a crucial role of YAP, the central effector of the Hippo signaling pathway, in the development and evolution of the mammalian neocortex. Specifically, our findings advance our understanding of cNPC activity in cortical development, and of its differences across mammals, with regard to two key aspects. First, YAP activity is shown to be necessary and sufficient to maintain the proliferative capacity of BPs at levels typically seen in developing gyrencephalic neocortex. Second, differences in YAP activity in BPs between embryonic mouse, which develops a lissencephalic neocortex, on the one hand, and embryonic ferret and fetal human, which develop a gyrencephalic neocortex, on the other hand, are shown to contribute to, if not underlie, the differences in BP proliferative capacity across these species. Three aspects of our findings deserve particular consideration.
DIFFERENTIAL YAP EXPRESSION BETWEEN VZ AND SVZ ACROSS SPECIES
Previous reports on the role of YAP were confined to mouse brain development and concentrated on APs (Cappello et al., 2013; Lavado et al., 2013; Lavado et al., 2014; Saito et al., 2017). In contrast to these reports, the present study is focused on BPs and in this context has compared YAP expression and activity in the SVZ vs. VZ of developing neocortex of three mammals exhibiting a different extent of neocortex expansion – mouse, ferret and human. Our demonstration that in species in which BPs exhibit significant proliferative capacity, YAP expression and activity are well detectable not only in the VZ but also SVZ, has at least two implications. First, to the best of our knowledge, these data constitute the first report of YAP protein expression in the iSVZ and oSVZ of developing neocortex. Second, the present findings suggest that the role of YAP, and hence of Hippo signaling, in the proliferation and consequently the pool size of cNPCs is far more widespread than previously assumed.
YAP IN BPS – A PUTATIVE DOWNSTREAM TARGET OF SOX2
The co-expression of active YAP with Sox2 in ferret and human BPs suggests an interesting mechanistic scenario regarding the regulation of YAP expression and activity. Recently, it was shown that Sox2 directly drives the expression of YAP in progenitors of the osteo-adipo lineage (Seo et al., 2013). Our results therefore raise the possibility that YAP expression in proliferative BPs is driven by Sox2. Furthermore, while YAP activity is known to be decreased by phosphorylation via upstream Hippo signaling kinases such as Nf2 and Wwc1 (Yu et al., 2015), it was recently reported that in cancer stem cells Sox2 represses transcription of the NF2 and WWC1 genes, which in turn increased YAP activity (Basu-Roy et al., 2015). Our finding of increased YAP activity in proliferative BPs of developing gyrencephalic neocortex could therefore be explained by Sox2 in these cells repressing NF2 and WWC1.
BP FATE SWITCH UPON CONDITIONAL YAP EXPRESSION
However, our data are not only consistent with the notion that in ferret and human BPs Sox2 positively regulates YAP expression levels and activity, but also indicate that forced YAP expression in mouse BPs, directly or indirectly, results in increased Sox2 expression and decreased Tbr2 expression. In other words, the present approach of conditionally expressing CA-YAP in the mouse Tis21-positive BP lineage induced a BP fate switch from neurogenic to proliferative. This in turn led to an expansion of the BP pool, which eventually resulted in an increased generation of upper-layer neurons. Hence, increasing YAP activity in BPs is sufficient to induce features that are hallmarks of an expanded neocortex, as characteristically observed in gyrencephalic mammals (Florio and Huttner, 2014; Lui et al., 2011). This, together with our finding that inhibiting YAP activity in BPs of developing ferret and human neocortex reduced their abundance, in turn leads us to conclude that an increased YAP activity in BPs of developing neocortex is a major contributor to its evolutionary expansion.
Experimental Procedures
Ethics
All animal experiments (mice and ferrets) were performed in accordance with the German Animal Welfare legislation (“Tierschutzgesetz”). All procedures regarding the animal experiments were approved by the Governmental IACUC (“Landesdirektion Sachsen”) and overseen by the Institutional Animal Welfare Officer(s). The license numbers concerning the experiments with mice are: Untersuchungen zur Neurogenese in Mäuseembryonen TVV2015/05 (in utero electroporation, tamoxifen, EdU) and 24–9168.24-9/2012-1 (tissue collection without prior in vivo experimentation). The license number concerning the experiments with ferrets is: Untersuchungen zur Neurogenese in Frettchen" (TVV 2015/02) issued by “Landesdirektion Sachsen”.
Mice
To characterize YAP expression E13-14 mouse embryos (C57BL/6JOlaHsd) were used. For electroporation E13 Tis21-CreERT2+/- heterozygous mouse embryos were used. These were obtained by crossing C57BL/6JOlaHsd females and Tis21-CreERT2+/+ males (Wong et al.,2015). Mice were crossed and kept under strict pathogen-free conditions in the animal facility of the Max Planck Institute of Molecular Cell Biology and Genetics.
Ferrets
Normally pigmented pregnant female sable ferrets (Mustela putorius furo) were purchased from Marshall BioResources (North Rose, NY, USA). They were delivered and housed in the animal facility of the Max Planck Institute of Molecular Cell Biology and Genetics.
Human fetal tissue
Human fetal brain tissue was obtained from two sources. First, from the Klinik und Poliklinik für Frauenheilkunde und Geburtshilfe, Universitätsklinikum Carl Gustav Carus of the Technische Universität Dresden, following elective pregnancy termination and informed written maternal consents, and with approval of the local University Hospital Ethical Review Committees. The age of a 12 wpc fetus (n=1) was assessed by ultrasound measurements of crown-rump length and other standard criteria of developmental stage determination. The second source was the Human Developmental Biology Resource (HDBR). This human fetal brain tissue was provided by the Joint MRC/Wellcome Trust (grant # MR/R006237/1) Human Developmental Biology Resource (www.hdbr.org). The HDBR provided fresh tissue from fetuses aged 11-13 wpc, (11 wpc, n = 4; 12 wpc, n = 2; 13 wpc, n = 2; wpc 14, n=1). Human fetal brain tissue was dissected in 1x PBS and used immediately for culture or fixation (as indicated) when obtained from Dresden. When obtained from HDBR, tissue was dissected and shipped in Hibernate E media (Gibco A1247601). Upon arrival, all tissue was cultured in slice culture medium (SCM, see section on ex vivo FFT cultures) for 2-3 h prior to any further manipulation. All tissue was fixed for at least 24 h at 4°C in 4% paraformaldehyde in 120 mM phosphate buffer (pH 7.4) (referred to in short as 4% PFA).
Tamoxifen preparation and administration
Tamoxifen powder, 200 mg, was dissolved in 10 ml of corn oil (Sigma, T-5648) under constant stirring at ≈40°C. Tamoxifen was administered to trigger activation of Cre recombinase in Tis21::CreERT2 +/- mouse embryos. Pregnant mice received tamoxifen (2 mg, 0.1 ml) orally by gavage at E12.5, i.e. one day before IUE, and once on the day of IUE (E13.5).
In utero electroporation of mice
Tamoxifen-treated pregnant mice carrying E13.5 embryos were anesthetized using initially 5% isoflurane (Baxter, HDG9623), followed by 2-3% isoflurane during the IUE procedure. Endotoxin-free plasmids (Control, pCAGGS–LoxP–Gap43-GFP–LoxP–IRES–nRFP; CA-YAP, pCAGGS–LoxP–Gap43-GFP–LoxP–CA-YAP–IRES–nRFP) were mixed on the day of surgery with Fast Green (Sigma, 0.25% final concentration) to a final plasmid concentration of 2 µg/µl in 1x PBS. Using a borosilicate microcapillary (Sutter instruments, BF120-69-10) the DNA/Fast Green mixture was intraventricularly injected, which was followed by six 50-msec pulses of 30 V at 1 sec intervals (BTX genetronics Inc., 45-0052INT), using a 3-mm diameter electrode (BTX genetronics Inc., 45-0487). After the IUE, the uterus was placed back into the abdominal cavity, and the peritoneum was sutured (VICRYL™ 5-0, V493H). Abdominal skin was closed with clips and animal received 100 µl of painkiller (Rimadyl 1 mg/ml). Pregnant mice were sacrificed by cervical dislocation at the indicated time points (E14.5-E17.5), and embryonic brains were dissected and fixed in 4% PFA, overnight at 4°C.
Verteporfin treatment of ferret and human neocortex in ex vivo free-floating tissue culture
An ex vivo free-floating tissue (FFT) culture system, adapted and modified from (Schenk et al., 2009; Long et al., 2018), was used to perform verteporfin treatment of embryonic ferret and fetal human neocortex tissue. E33 ferret brains were dissected, meninges removed, and the two hemispheres separated. Fetal human neocortex tissue of 11-13 wpc was cut into 2000-2500 µm-thick pieces (tangential dimension). Tissue was cultured in a whole-embryo culture incubator (Ikemoto RKI) in a rotating flask with 1.5 ml of SCM (for composition, see below), and incubated at 37°C in the presence of a humidified atmosphere consisting of 40% O2 / 5% CO2 / 55% N2, with continuous rotation at 26 rpm. Control flasks contained 10 µl of DMSO (Dimethyl sulfoxide, Sigma, 472301) added to the 1.5 ml of SCM. Verteporfin flasks contained 1 µM of verteporfin (Verteporfin, Sigma, SML0534) dissolved in 10 µl of DMSO, added to the 1.5 ml of SCM. FFT cultures were carried out for 48 h, with one change of SCM (containing either DMSO or DMSO plus verteporfin) after 24 h. After 48 h of FFT culture, tissue was fixed in 4% PFA overnight at 4°C.
The SCM medium used for FFT cultures contained: 84 ml of Neurobasal medium (Gibco, 21103049) supplemented with either 10 ml of rat serum (ferret and human cultures) or 10 ml of 5x KnockOUTTM Serum Replacement (for human cultures, Gibco, 10828028), 1 ml GlutaMAXTM (100x), 1 ml penicilin/streptomycin (100x) (Gibco,15140122), 1 ml N-2 (100x) (Gibco, 17502048), 2 ml B-27 (50x) (Gibco, 17504044) and 1 ml of 1 M HEPES-NaOH, pH 7.2, to yield a final volume of 100 ml.
In utero electroporation of ferrets
In utero electroporation of E33 ferret embryos was performed as originally established (Kawasaki et al., 2012; Kawasaki et al., 2013), with the modifications indicated below. Pregnant ferrets (with embryos at E33) were kept fasted for at least 3 h before the surgery and placed in the narcosis box with 4% isoflurane. Subsequently, they were positioned on the operation table and attached to the narcosis mask with 3% isoflurane and injected subcutaneously with analgesic (0.1 ml Metamizol, 50 mg/kg), antibiotic (0.13 ml Synulox, 20 mg/kg or 0.1 ml amoxicilin, 10 mg/kg) and glucose (10 ml 5% glucose solution). The ferret bellies were then shaved, sterilized with iodide and surgically opened. Then, the uterus was exposed. As the ferret uterus is pigmented, a transmitted light source was used for the visualization of embryos. Embryos were injected intraventricularly with a solution containing 0.1% Fast Green (Sigma) in sterile 1x PBS, 2 µg/µl of one of the endotoxin-free plasmids (CTRL-empty or DN-YAP) as indicated. To visualize electroporated cells, CTRL and DN-YAP plasmids were co-electroporated with CAGGS-EGFP (1 µg/µl), using a 5-mm diameter electrode (BTX genetronics Inc., 45-0489). Electroporations were performed with six 50-msec pulses of 100 V at 1 sec intervals. Subsequently, the uterus was placed back in the peritoneal cavity, muscle layer with the peritoneum were sutured (VICRYL™ 4-0), after which the skin was sutured intracutaneously. Animals were carefully monitored until they woke up and then underwent postoperative care for the following 3 days. Pregnant females received subcutaneous injections of 15 ml of 5% glucose and the painkiller Metamizol (50 mg/kg, WDT, 99012) three times per day.
The embryos were obtained by cesarean section at E35, and the cerebral cortex was dissected and fixed in 4% PFA overnight at 4°C. To this end, the mother ferrets underwent a second surgery that followed the same pre-operative care, anesthesia and analgesia as the first surgery. The sutures from the first operation were removed and the uterus exposed, after which the embryos were removed by a caesarian section. Subsequently a complete hysterectomy was performed, after which the muscle layer with peritoneum and skin were sutured and the animal underwent the same post-operative care as after the first surgery. Animals were kept at the BMS of the MPI-CBG for at least two weeks after the second surgery after which they were donated for adoption.
Statistical analysis
Data were tabulated in Excel (Microsoft, Redmond, WA) and analyzed in Prism 5 (GraphPad) software. Statistical analyses were performed using unpaired Student’s t-test, Mann-Whitney U-test, and one-way ANOVA test.
Author contributions
M.K., T.N., J.T.L.M.P., and W.B.H. conceived the project and designed the experiments.
M.K. performed most of the experiments, with technical assistance from J.T.L.M.P., T.N., K.L. and N.K.
B.L. performed the cesarean sections on ferrets and took care of ferrets pre- and post-surgery.
H.K. shared the technical knowledge on IUE of ferrets.
M.K. analyzed the data, with day-to-day supervision by J.T.L.M.P. and T.N.
M.K. and W.B.H. wrote the manuscript, with input from T.N., K.L. and N.K.
W.B.H. supervised the project.
Declaration of interests
The authors declare no competing interests.
Acknowledgements
We are grateful to the Services and Facilities of the Max Planck Institute of Molecular Cell Biology and Genetics for the outstanding support provided, notably J. Helppi and his team of the Biomedical services (BMS), and J. Peychl and his team of the Light Microscopy Facility. We also thank J. Fei (Tanaka Lab, CRTD, Dresden, Germany) for providing YAP antibody and constructive suggestions. We would like to thank all members of the Huttner group for helpful discussions, especially A. Güven for advice, A. Sykes for assistance with molecular cloning and M. Florio for sharing RNA-seq data on YAP expression. We thank D. Gerrelli, S. Lisgo and their teams at the HDBR for the invaluable support from this resource. M.K. was a member of the International Max Planck Research School for Cell, Developmental and Systems Biology, and a doctoral student at the Technische Universität Dresden. W.B.H. was supported by grants from the DFG (SFB 655, A2) and the ERC (250197), by the DFG-funded Center for Regenerative Therapies Dresden and by the Fonds der Chemischen Industrie.
Footnotes
↵4 Lead Contact