Neural Progenitor Cells Undergoing Yap/Tead-Mediated Enhanced Self-Renewal Form Heterotopias More Easily in the Diencephalon than in the Telencephalon

Spatiotemporally ordered production of cells is essential for brain development. Normally, most undifferentiated neural progenitor cells (NPCs) face the apical (ventricular) surface of embryonic brain walls. Pathological detachment of NPCs from the apical surface and their invasion of outer neuronal territories, i.e., formation of NPC heterotopias, can disrupt the overall structure of the brain. Although NPC heterotopias have previously been observed in a variety of experimental contexts, the underlying mechanisms remain largely unknown. Yes-associated protein 1 (Yap1) and the TEA domain (Tead) proteins, which act downstream of Hippo signaling, enhance the stem-like characteristics of NPCs. Elevated expression of Yap1 or Tead in the neural tube (future spinal cord) induces massive NPC heterotopias, but Yap/Tead-induced expansion of NPCs in the developing brain has not been previously reported to produce NPC heterotopias. To determine whether NPC heterotopias occur in a regionally characteristic manner, we introduced the Yap1-S112A or Tead-VP16 into NPCs of the telencephalon and diencephalon, two neighboring but distinct forebrain regions, of embryonic day 10 mice by in utero electroporation, and compared NPC heterotopia formation. Although NPCs in both regions exhibited enhanced stem-like behaviors, heterotopias were larger and more frequent in the diencephalon than in the telencephalon. This result, the first example of Yap/Tead-induced NPC heterotopia in the forebrain, reveals that Yap/Tead-induced NPC heterotopia is not specific to the neural tube, and also suggests that this phenomenon depends on regional factors such as the three-dimensional geometry and assembly of these cells.


Introduction
Undifferentiated neural progenitor cells (NPCs) are central to the production of a variety of cells in developing and mature brains [1][2][3][4][5][6][7][8][9]. During the early embryonic period before neurons emerge, the nuclei/somata of NPCs are diffusely distributed throughout the initial wall of brain primordia [also called the neuroepithelium (NE)]. This is because NPCs, which adopt an elongated (~100 μm) epithelial morphology, move their nuclei/somata along the apicobasal axis in a cell cycle-dependent manner, a process called interkinetic Abstract Spatiotemporally ordered production of cells is essential for brain development. Normally, most undifferentiated neural progenitor cells (NPCs) face the apical (ventricular) surface of embryonic brain walls. Pathological detachment of NPCs from the apical surface and their invasion of outer neuronal territories, i.e., formation of NPC heterotopias, can disrupt the overall structure of the brain. Although NPC heterotopias have previously been observed in a variety of experimental contexts, the underlying mechanisms remain largely unknown. Yes-associated protein 1 (Yap1) and the TEA domain (Tead) proteins, which act downstream of Hippo signaling, enhance the stem-like characteristics of NPCs. Elevated expression of Yap1 or Tead in the neural tube (future spinal cord) induces massive NPC heterotopias, but Yap/Tead-induced expansion of NPCs in the developing brain has not been previously reported to produce NPC heterotopias. To determine whether NPC heterotopias occur in a regionally characteristic manner, we introduced the Yap1-S112A or Tead-VP16 into NPCs of the telencephalon and diencephalon, two neighboring but distinct forebrain regions, of embryonic day 10 mice by in utero electroporation, and compared NPC heterotopia formation. Although NPCs in both regions exhibited enhanced Kanako Saito and Ryotaro Kawasoe have contributed equally to this work. nuclear migration (IKNM) [reviewed in 10,11]. Horizontal assembly of NPCs individually differing in cell-cycle phase and nuclear/somal position makes this NE structure "pseudostratified" (i.e., although there are several layers of nuclei in the NE wall, one layer of cells can host multiple layers of nuclei). As the brain walls thicken (>100 μm), a new zone in which differentiated neurons accumulate is added outside (i.e., basal to) the original NE region, which is subsequently called the ventricular zone (VZ) (~100 μm), and NPCs elongate along the apicobasal axis by adding basal processes [4][5][6][8][9][10][11]. Throughout the neuron-producing period, the elongated NPCs (also referred to as radial glial cells) keep their nuclei/somata in the apical (inner) part of the wall, as in the earlier NE period, maintaining the pseudostratified structure and VZ thickness, which correspond to the range of IKNM along the apicobasal axis. NPCs' nuclei/somata move only in the VZ, and therefore do not intermingle with neurons located basally. Under diverse experimental and pathological conditions, however, NPCs' nuclei/somata assume heterotopic positions; i.e., they invade neuronal territory ~200 μm away from the apical surface. A major reason for the formation of such NPC heterotopias is over-proliferation in the NE/VZ. For example, excessive activation of canonical Wnt-β-catenin signaling [12][13][14], FGF signaling [15], or cyclin D1 [16,17], which can induce NPCs to maintain stem-like proliferative and undifferentiated status rather than undergoing quiescence and differentiation, results in extra-VZ nuclear/somal distribution of NPCs, and these mechanisms may underlie the pathogenesis of congenital brain diseases in humans [18,19]. Furthermore, physiological dislocation of NPCs' nuclei/somata to non-VZ positions seems to have accompanied neocortical evolution [8,9,[20][21][22][23]. Thus, elucidation of the mechanisms that regulate the extra-VZ positioning of NPCs' nuclei/somata is a critical goal in the field of developmental neuroscience.
A previous study that focused on the developing neural tube (designated here as the future spinal cord, not including the regions that form the future brain) showed that excessive activation of the Yes-associated protein 1 (Yap1) and the TEA domain (Tead) proteins, key DNA-binding platforms for Yap, in the NE results in over-proliferation of NPCs [24]. Yap1 and Tead are involved in cell density-dependent regulation of proliferation. At high cell densities, strong Hippo signaling suppresses cell proliferation by inhibiting nuclear Yap accumulation, whereas at low cell densities, weak Hippo signals allow nuclear Yap accumulation, promoting cell proliferation through Tead activation [25][26][27]. The Yap/Tead-induced over-proliferation or expansion of self-renewing stem-like NPCs in the neural tube NE/VZ is almost always accompanied by NPC heterotopias outside the VZ [24]. Interestingly, although Yap/Tead-induced NE/VZ expansion has been consistently reproduced in multiple studies [28][29][30][31][32], NPC heterotopias have not, to our knowledge, been clearly observed in primordial CNS regions other than the neural tube.
Hence, we sought to determine whether Yap/Teadinduced NPC heterotopias occur in a regionally characteristic manner. Although it is possible that NPC heterotopias might arise in a manner dependent upon the molecularly regulated regional identities of NPCs, it is also possible that the occurrence of NPC heterotopias reflects region-specific three-dimensional conditions (e.g., the geometry and assembly of NPCs). Yap1 and Tead play multiple roles in response to different physical/mechanical factors, such as tissue tension or compression [33][34][35][36][37][38], and they are involved in the epithelial-mesenchymal transition [26,27,39].
To address this issue, we took advantage of in utero electroporation (IUE)-mediated gene transfer into the early embryonic mouse dorsal forebrain, which consists of the telencephalon (future cerebrum) and the dorsal diencephalon (future thalamus). The developing telencephalon and diencephalon share some common molecular hallmarks (e.g., Pax6 expression in the NE/VZ of both regions, Fig. 1). Morphologically, however, the extent to which a given wall portion (and NE/VZ contained) expands tangentially/laterally during the embryonic period differs between these regions. The dorso-ventral extension of the thalamic wall is clearly smaller than that of the cerebral wall, which expands to form a large vesicle/hemisphere [40]. These differences in physical limitations (inversely, lateral expandability) might underlie the differential susceptibility to experimental/pathological perturbations [41]. To investigate this possibility, we performed IUE at embryonic day 10 (E10), because (1) the forebrain primordium at this early stage is similar to the neural tube in terms of the composition and structure of the NE, and (2) in previous work, the use of the same protocol to activate Wnt3a induced mild NPC heterotopias in the telencephalon [14], allowing us to comparatively address possible contributions of Yap/Tead in NPC heterotopia formation. We found that artificial activation of Yap/Tead enhanced NPC self-renewal in both telencephalic and diencephalic NE/VZ, whereas NPC heterotopias formed much more frequently in the diencephalon than in the telencephalon.

Animals
Pregnant ICR mice were obtained from SLC (Japan). All protocols for animal experiments were approved by The Animal Care and Use Committee of the Nagoya University.

Yap1 Expression in Single Cell Transcriptome Profiles
Information of Yap1 expression levels in single E11, E14 and E16 telencephalic cells was obtained from the single cell transcriptome profiles available in the GEO database under accession codes: GSE10881 (for E14) and GSE55981 (for E11, E16) with probe set ID: 1448363_at. . Yap1 immunoreactivity was observed throughout the pseudostratified NE of the dorsal diencephalon (c) and the medial telencephalic wall (d), and it was also seen in the similarly pseudostratified VZ of the lateral telencephalic wall (e). Outside the VZ, where Pax6 was not detected (f), radially elongated fibers (i.e., basal processes of NPCs) were also Yap1-positive, whereas neurons dominating in that region were Yap1-negative. NPCs showed nuclear anti-Yap1 staining, with some heterogeneity in the NE/VZ (d'). Simi-lar immunostaining patterns were reproduced at E12 (g, h). Singlecell transcriptome analysis of E11, E14, and E16 telencephalic cells (i) revealed that Yap1 expression was strongest in undifferentiated NPCs [designated as apical progenitors (AP)] and weakest in neurons (N), with intermediate levels of expression in basal progenitors (BP) differentiating while keeping mitotic activity [43,44]. At E11, the NE/VZ was almost fully occupied by AP-type cells (undifferentiated NPCs). Scale bars 100 µm in a and b, 10 µm in c-f, g, and h, 5 µm in d'

In Utero Electroporation
In utero electroporation (IUE) was performed using pregnant ICR mice at E10 as described previously [14]. DNA solution was injected into the lateral ventricle. The head of the embryo in the uterus was placed between the discs of a forceps-type electrode (disc electrodes of 1 mm; CUY560P1, NEPA GENE, Chiba, Japan), and electric pulses (50 V) were charged four times, resulting in gene transfection into the cerebral wall.

Statistical Analysis
Differences were analyzed using Wilcoxon rank-sum test (in Figs. 2c, 3d, 4c) or Permuted Brunner Munzel test (in Fig. 2d) by R, and data were expressed as means ± SD.

Yap1 is Expressed in Early Embryonic Mouse Forebrain NPCs
We first examined the expression of Yap1 in the early (E11 and E12) telencephalic and diencephalic walls. Observation of sections immunostained with anti-Yap1 antibody at low magnification revealed intense Yap1 immunoreactivity in the NE and VZ of both regions, which were also positive for Pax6 (Fig. 1a, b). In the neuronal zone forming outside the VZ, Yap1 immunoreactivity was only faintly detected. In individual E11, E14, and E16 telencephalic cells, single-cell transcriptome analysis, which can discriminate among cells of distinct differentiation status [43,44], revealed prominent Yap1 expression in undifferentiated NPCs but not in neurons (Fig. 1i), confirming our immunohistochemical finding. At a higher magnification (Fig. 1c-h), Yap1 immunoreactivity was most clearly observed in the cytoplasm of the pseudostratified somata of NPCs, as well as in NPCs' radial processes extending basally (towards the pial surface). But, it was also detected in the nucleus (Fig. 1d, d′, e), similar to a pattern previously observed in the neural tube at E8.5 and Fig. 3 Yap/Tead overexpression in the dorsal diencephalon resulted in frequent and massive NPC heterotopias. a Anti-GFP and anti-Pax6 photomicrographs of E13 dorsal diencephalons that had been electroporated at E10 with the control, Yap1-S112A, or Tead2-VP16 vectors. b Basal zone dominated by TUJ1 + neurons in a Yap1-S112A-IUE telencephalic wall is shown at a higher magnification, containing a large cluster of heterotopic NPCs. Distribution of Pax6 + NPCs in a is schematically shown in c (green = GFP + Pax6 + , magenta = GFP − Pax6 + ). Graph d summarizes the distribution of GFP + Pax6 + obtained from multiple sections [n = 4 for control, n = 5 for Yap1-S112A, and n = 4 for Tead2-VP16 (three independent embryos in each group)], showing that Yap1-S112A or Tead2-VP16 IUE induced massive NPCs heterotopias. In addition to GFP + Pax6 + cells predominantly constituting heterotopic clusters, some GFP − Pax6 + cells (arrow in panel b, magenta in panel c) were observed in or near the heterotopic clusters found in the basal-most portion of the telencephalic wall. Scale bars 50 µm in a, 10 µm in b E10.5 [26]. This nuclear Yap1 localization, a sign of Yap1's contribution to proliferation [25][26][27], is consistent with previously reported Yap1's roles in NPCs [24,[28][29][30][31][32]. Nuclear Yap1 immunorectivity was not homogeneous in the NE/VZ (i.e. some nuclei were more strongly stained than others, Fig. 1d′), although the reason for this remains unknown.

Yap1 or Tead Enhances Stem-Like Behaviors of NPCs in Both the Dorsal Telencephalon and the Dorsal Diencephalon
To increase Yap1 expression in the dorsal diencephalic or telencephalic walls, we used the Yap1-S112A mutant, previously shown in NIH3T3 cells to induce proliferation, resulting in denser confluence than that achieved by Yap1 without mutations [26,33]. We electroporated E10 mice with a GFP-tagged Yap1-S112A or a control GFP vector. After 1 day (at E11), we compared anti-Yap1 immunoreactivity between un-electroporated, control GFP-electroporated, and Yap1-S112A-electroporated cells. Yap1 immunoreactivity was more intense in the Yap1-S112A-electroporated NPCs (both cytoplasmic and nuclear) than in un-electroporated or control GFP-electroporated NPCs in both diencephalon (Fig. 2b) and telencephalon (data not shown), confirming that our IUE successfully increased the expression of Yap1 in a region-specific manner.
We next sought to evaluate the stem-like (i.e. self-renewing) status of the Yap1-S112A-electroporated NPCs. Consistent with previous studies revealing the formation of a hypertrophic VZ in the Yap-overexpressed telencephalon Fig. 4 Yap1/Tead overexpression in the dorsal telencephalic VZ resulted only rarely in NPC heterotopias in the ventralmost region. a Anti-GFP and anti-Pax6 photomicrographs of E13 dorsal telencephalons that had been electroporated at E10 with the control, Yap1-S112A, or Tead2-VP16 vectors. Distribution of Pax6 + NPCs in a is schematically shown in b (green = GFP + Pax6 + , magenta = GFP − Pax6 + ). Graph c summarizes the distribution of GFP + Pax6 + obtained from multiple sections [n = 5 for control, Yap1-S112A, and Tead2-VP16 (three independent embryos in each group)], showing that most GFP + Pax6 + NPCs were in the VZ. d An exceptional case in which a relatively large cluster of heterotopic GFP + Pax6 + NPCs was formed by Yap1-S112A IUE of the ventral-most part of the dorsal telencephalic VZ. Note that this region is physically constrained by a hairpin-loop bending of the VZ upon transition to the ventral telencephalon. Scale bars 50 µm in a, 100 µm (left panels) and 10 µm (right panels) in d [30,31], our comparison of the proportion of the total GFP + cells that were also positive for Pax6, a transcription factor expressed in stem-like NPCs [8,43,44] to control the balance between their proliferative and undifferentiated (VZ-remaining) status and differentiation [45,46], between control and Yap1-S112A IUE telencephalons (analyzed at E13) revealed that the descendant cells of the Yap1-S112A-electroporated cells were more frequently Pax6 + (61%) than those of control GFP-electroporated cells (48%) (p = 0.0012) (Fig. 2a, c), suggesting that Yap1-S112A-electroporated NPCs engaged in self-renewal (i.e., they divided to give rise to stem-like daughter cells) to a greater extent than the controls. We also electroporated E10 telencephalons with activator-modified Tead2 (Tead2-VP16), previously shown to mediate Yap-dependent enhancement of proliferation in both NIH3T3 cells [26] and NPCs [24,32]. The frequency of Pax6 expression in GFP + cells (%Pax6 + /GFP + ) in Tead2-VP16 IUE was greater (67%) than in control GFP IUE (48%) (p = 0.0006) (Fig. 2c), as also observed in the Yap1-S112A IUE experiments.

NPC Heterotopias Form More Frequently in the Dorsal Diencephalon than in the Dorsal Telencephalon
We next asked whether telencephalic or diencephalic VZs exhibiting an elevated tendency toward self-renewal might form NPC heterotopias, as previously observed in the neural tube [24]. By E13, the Yap/Tead-overexpressing dorsal diencephalic walls frequently formed massive NPC heterotopias (Fig. 3) [extra-VZ NPCs were present in 61% in Yap1-S112A IUE (p = 0.0022) and 63% in Tead2-VP16 IUE (p = 0.0043), versus 13% in control IUE]. Each heterotopia, 150-200 µm away from the apical/ventricular surface and within the TUJ1 + neuronal territory, consisted of many GFP + TUJ1 − somata (each containing the Pax6 + nucleus) (10-50 per 16-µm-thick section), clustered to become ~50 µm in aggregate diameter. Up to 63% of GFP + Pax6 + cell nuclei were outside the VZ, and formation of heterotopia was very frequent (i.e., in most Yap1-S112A or Tead2-VP16 IUE cases). GFP immunostaining suggested that most (if not all) of these heterotopically aggregated NPCs had lost connections to the apical surface. In the control IUE, by contrast, GFP + cells that were judged to have lost apical attachment were TUJ1 + (Pax6-negative) neurons.
Occasionally in Yap/Tead-overexpressing embryos, Pax6 + but GFP-negative nuclei/somata were observed in or near the heterotopic NPC aggregates (~6% of all heterotopic Pax6 + cells). Although some GFP-negative Pax6 + cells were observed in the control IUE cases, they were distributed within 120-130 µm from the apical surface and close to the basal border of the VZ (Fig. 3c). GFP-negative Pax6 + cells in the Yap/Tead-overexpressing cases were observed in more basal portions (150-200 µm from the apical surface), near the GFP + Pax6 + cells massively occupying such far-basal regions. Therefore, it is possible that not only GFP + electroporated cells or their descendant cells, but also a minor fraction of the surrounding non-electroporated (GFP − ) VZ cells, were detached from the apical surface, presumably due to mechanical effects near the apical surface, such as overcrowding [11,14,47].
Although the apical surface of the dorsal telencephalon is concave, that of the ventral telencephalon is convex, making the dorsal-to-ventral transitional region of the VZ bend abruptly in a hairpin loop. In our previous experiments in which Wnt3a was overexpressed in the dorsal telencephalic VZ, this most ventrolateral part was most extensively thickened along the apicobasal axis, and formed frequent large NPC heterotopias, in contrast to other dorsal telencephalic VZ regions that seemed to have freely expanded tangentially (along the dorso-ventral axis) without apicobasal thickening [14]. Physiological non-surface mitoses in the dorsal telencephalon are observed earliest, and most frequently, in this ventrolateral end region [48].
Although the precise mechanism underlying the differential heterotopia formations between the diencephalic and telencephalic NPCs under enhanced self-renewal due to artificial activation of Yap1/Tead remains unknown, our results suggest that NPC heterotopias may occur as a reflection of region-specific three-dimensional conditions. As schematically (but semi-quantitatively) traced in Fig. 5a, the telencephalic and diencephalic VZs differ in regard to their lateral expandability during early embryonic days. The lateral extension of the dorsal diencephalon is much smaller than that of the dorsal telencephalon, which forms a large vesicle/hemisphere [40,49]. Such differential tangential expandability of VZ (i.e., tissue-level permissiveness for lateral expansion of the NE/VZ: telencephalon > diencephalon) could explain the differential NPC heterotopia formation (telencephalon < diencephalon) that we observed in our Yap1/Tead experiments.
In line with this model, we have observed that E11 or E12 dorsal diencephalic walls to which Yap1/Tead had been introduced at E10 did not form heterotopias when sliced and cultured in collagen gel (Saito et al., unpublished observation). Although collagen gel gently holds embedded slices to assist our time-lapse observations, it allows slices to grow almost freely, without physically restricting slices' expansion and thickening. VZ in slice cultured brain walls may be less laterally compressed than in vivo (i.e. permissiveness to VZ's lateral expansion is greater in slice culture than in vivo), thereby failing to reproduce the in vivo (heterotopia-forming) behaviors of the Yap1/Tead-introduced dorsal diencephalic VZ cells.
Excessive compression at/near the apical surface is involved in epithelial cells' detachment from the apical surface [11,14,47]. Limited lateral expandability of the VZ (Fig. 5b), which is more likely to arise in the diencephalon and the neural tube than in the telencephalon, may thus act as an inducer for NPCs to leave the apical surface. These results provide a basis for future studies aimed at revealing how Yap1/Tead might act in NPCs to sense and/or respond to such hypothetical physical/mechanical limitations on the lateral expansion of VZ, and thereby contribute to heterotopia formation. Fig. 5 Model of differential heterotopia formation by Yap1/ Tead-overexpressing hyper-selfrenewing dorsal diencephalic and telencephalic NPCs, possibly depending on tissue-level permissiveness for lateral expansion of the NE/VZ. A trace of the growing telencephalic and diencephalic walls (a), reproduced based on Altman and Bayer [40], shows that by E13, the telencephalic NE/ VZ laterally expands such that it becomes three times broader than at E10, whereas the diencephalic NE/VZ cannot expand so extensively along the dorsoventral axis. Such differential tangential expandability could explain the differential heterotopia formation by NPCs (b) Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.