*Article* **NXN Gene Epigenetic Changes in an Adult Neurogenesis Model of Alzheimer's Disease**

**Idoia Blanco-Luquin 1,\* , Blanca Acha <sup>1</sup> , Amaya Urdánoz-Casado <sup>1</sup> , Eva Gómez-Orte <sup>2</sup> , Miren Roldan <sup>1</sup> , Diego R. Pérez-Rodríguez <sup>3</sup> , Juan Cabello <sup>2</sup> and Maite Mendioroz 1,4**


**Abstract:** In view of the proven link between adult hippocampal neurogenesis (AHN) and learning and memory impairment, we generated a straightforward adult neurogenesis *in vitro* model to recapitulate DNA methylation marks in the context of Alzheimer's disease (AD). Neural progenitor cells (NPCs) were differentiated for 29 days and Aβ peptide 1–42 was added. mRNA expression of Neuronal Differentiation 1 (*NEUROD1*), Neural Cell Adhesion Molecule 1 (*NCAM1*), Tubulin Beta 3 Class III (*TUBB3*), RNA Binding Fox-1 Homolog 3 (*RBFOX3*), Calbindin 1 (*CALB1*), and Glial Fibrillary Acidic Protein (*GFAP*) was determined by RT-qPCR to characterize the culture and framed within the multistep process of AHN. Hippocampal DNA methylation marks previously identified in Contactin-Associated Protein 1 (*CNTNAP1*), SEPT5-GP1BB Readthrough (*SEPT5-GP1BB*), T-Box Transcription Factor 5 (*TBX5*), and Nucleoredoxin (*NXN*) genes were profiled by bisulfite pyrosequencing or bisulfite cloning sequencing; mRNA expression was also measured. *NXN* outlined a peak of DNA methylation overlapping type 3 neuroblasts. Aβ-treated NPCs showed transient decreases of mRNA expression for *SEPT5-GP1BB* and *NXN* on day 9 or 19 and an increase in DNA methylation on day 29 for *NXN*. *NXN* and *SEPT5-GP1BB* may reflect alterations detected in the brain of AD human patients, broadening our understanding of this disease.

**Keywords:** adult hippocampal neurogenesis; NPCs; Alzheimer's disease; Aβ peptide; DNA methylation; gene expression; *NXN*; *CNTNAP1*; *SEPT5-GP1BB*; *TBX5*

## **1. Introduction**

Adult neurogenesis (AN) is the process of forming functional neurons *de novo*. In the adult mammalian brain, neurogenesis occurs predominantly in specific brain niches: the subgranular zone (SGZ) of the dentate gyrus (DG) of the hippocampus and the subventricular zone (SVZ) lining the lateral ventricles [1,2]. During the process of adult hippocampal neurogenesis (AHN), neural stem cells (NSCs) self-renew and differentiate, giving rise to transient amplifying progenitors (TAPs), neuroblasts, and eventually mature neurons, astrocytes, and oligodendrocytes.

AHN regulators can be divided into intrinsic or extrinsic factors, that is, transcription factors (TFs) synthesized by the developing neural precursors and neurons, and growth factors and neurotrophins secreted from the surrounding niche, respectively [3]. Epigenetic mechanisms tightly regulate extrinsic and intrinsic factors [4], controlling both temporal

**Citation:** Blanco-Luquin, I.; Acha, B.; Urdánoz-Casado, A.; Gómez-Orte, E.; Roldan, M.; Pérez-Rodríguez, D.R.; Cabello, J.; Mendioroz, M. NXN Gene Epigenetic Changes in an Adult Neurogenesis Model of Alzheimer's Disease. *Cells* **2022**, *11*, 1069. https://doi.org/10.3390/ cells11071069

Academic Editors: FengRu Tang and Luisa Alexandra Meireles Pinto

Received: 27 January 2022 Accepted: 20 March 2022 Published: 22 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

and spatial gene expression. Sequential steps of AN are regulated directly or indirectly by *de novo* methylation and maintenance of methylation marks [5]. Each distinct human brain region (cerebral cortex, cerebellum, and pons) has a characteristic DNA methylation signature [6], and even within brain regions such as the hippocampus, global methylation varies between neuronal subtypes [7].

During both physiological and pathological aging in humans, AHN clearly emerges as a robust phenomenon [8]. AHN is involved with the normal functionality of hippocampal circuits, which demonstrates an important link between AN and cognitive processes [9]. Thus, impaired neurogenesis may negatively impact the survival of adult-born neurons and contribute to learning and memory failure, as occurs with aging and neurological disorders, e.g., Alzheimer's disease (AD) [8,10,11].

AD is the most common neurodegenerative disorder, characterized by progressive memory loss and cognitive decline caused by widespread loss of neurons and synaptic connections in the cortex, hippocampus, amygdala, and basal forebrain, and by a gradually significant loss of brain mass. The amyloid precursor protein (APP) plays a key role in normal brain development by influencing NSC proliferation, cell fate specification, and neuronal maturation [10]. However, its derivative, the amyloid β (Aβ) peptide, a cleavage product of the APP enzymatic processing, is the major component of amyloid plaques, one of the hallmark pathologies found in brains of AD patients. Monomeric Aβ can selfaggregate to form oligomers, protofibrils, and amyloid fibrils, which deposit as amyloid plaques. Although the impact of Aβ on neurogenesis is still controversial, it is well known that Aβ plaques can cause severe damage to neurons and astrocytes, which results in the gradual loss of neurons associated with AD symptoms [11].

Remarkable alterations in AHN have been detected at early stages of AD, even before the onset of hallmark lesions or neuronal loss [8,12]. Impairments in epigenetic mechanisms lead to the generation of damaged neurons from NSCs, exacerbating the loss of neurons and deficits in learning and memory that characterize AD pathology [11]. Indeed, we and others have described epigenetic changes in DNA methylation in the hippocampus of AD patients at the genome-wide level [6,13]. In a previous study, we reported altered DNA methylation in the AD hippocampus occurring at specific regulatory regions crucial for neuronal differentiation; moreover, a set of neurogenesis-related genes were identified in the damaged tissue [6]. Hence, a better understanding of AHN impairment observed at the initial and later stages of AD by noninvasive methods may reveal insights into the pathogenesis of AD. What is more, restoration of normal levels of AHN may provide a potential therapeutic strategy to delay or halt AD-linked cognitive decline [8,12].

Here, we propose an intuitive *in vitro* approach to assess a stepwise lineage progression, as occurs during *in vivo* neurogenesis, by using human neural progenitor cells (NPCs) derived from an induced pluripotent stem cell (iPSC) line as the starting source material. In order to infer whether the differentiation of human NPCs into mature neurons is disrupted in the AD microenvironment, we designed an observational descriptive study by generating an *in vitro* model triggered by prolonged exposure to nanomolar concentrations of Aβ peptide 1–42. Next, we evaluated DNA methylation levels and mRNA expression changes of specific neurogenesis-related candidate genes.

#### **2. Materials and Methods**

#### *2.1. NPCs Culture, Neuronal Differentiation and Aβ Peptide Administration*

NPCs Derived from XCL1 DCXpGFP (ACS5005™, American Type Culture Collection, ATCC, Manassas, VA, USA) were cultured following manufacturer recommendations. Briefly, 0.30 <sup>×</sup> <sup>10</sup><sup>6</sup> NPCs were seeded onto a CellMatrix Basement Membrane Gel (ATCC® ACS3035™) coated 12-well plate and incubated in NPC expansion medium: complete growth medium including DMEM/F-12 (Gibco, Fisher Scientific, Waltham, MA, USA), supplemented with the Growth Kit for Neural Progenitor Cell Expansion (ATCC® ACS3003) and then maintained in a humidified incubator (5% CO2, 37 ◦C).

Neuronal differentiation experiments were carried out for 9, 19, and 29 days by plating NPCs at a seeding density of 80,000 viable cells/cm<sup>2</sup> in 6-well coated culture plates. First, NPCs were incubated in an expansion medium (day 0). From day 1 (post-seeding), half of the medium was changed for differentiation medium every 2–3 days throughout the duration of the culture period. Complete Differentiation Medium consisted of serum-free neuronal basal BrainPhys™ Neuronal Medium, formulated to improve the electrophysiological and synaptic properties of the neurons [14], NeuroCult™ SM1 Neuronal Supplement (1:50), N2 Supplement-A (1:100), Recombinant Human Brain-Derived Neurotrophic Factor (BDNF, 20 ng/mL), Recombinant Human Glial-Derived Neurotrophic Factor (GDNF, 20 ng/mL), Dibutyryl-cAMP (1 mM) and ascorbic acid (200 nM) (STEMCELL Technologies, Vancouver, BC, Canada). Half-fresh medium containing Aβ protein fragment 1–42 (50 nM; Sigma-Aldrich, St. Louis, MO, USA) or DMSO (Sigma-Aldrich) as a vehicle was added once a week.

NPCs were harvested on day 0 and 9, 19, and 29 days of differentiation for both conditions by detaching them with Accutase (Innovative Cell Technologies, San Diego, CA, USA), then washed with Dulbecco's phosphate-buffered saline (DPBS, Sigma-Aldrich), centrifuged at 13,000 rpm and frozen at −80 ◦C. All experiments were performed in triplicate.

#### *2.2. Selection of Candidate Epigenetic Marks in AD*

A set of differentially methylated positions (DMPs) in AD was produced from a methylome dataset generated in a previous study described elsewhere [6]. In brief, the Infinium HumanMethylation450 BeadChip array (Illumina, Inc., San Diego, CA, USA) was performed at the Roswell Park Cancer Institute Genomics Shared Resource (Buffalo, NY, USA) to measure DNA methylation levels in CpG sites (also named *positions*) in a cohort of 26 pure AD cases and 12 controls. A total of 118 AD-related DMPs were identified in the hippocampus of AD cases compared to controls. Here, we selected four of the above-identified DMPs in AD patients compared to controls (absolute β-difference ≥ 0.085 and *p*-value ≤ 0.05) and analyzed them due to their relationship with neurogenesis (Table 1 and Supplementary Figure S1).

**Table 1.** Selected differentially methylated positions (DMPs) in AD hippocampus measured by 450 K Illumina BeadChip array. The table shows four DMPs prioritized by beta difference (delta) criteria. Each CpG site was annotated by UCSC hg19 build.


#### *2.3. DNA Methylation Levels Assessed by Bisulfite Pyrosequencing*

Genomic DNA was isolated from frozen cell pellets of basal NPCs and control or Aβ peptide treated NPCs incubated in differentiation media for 9, 19, or 29 days by using the FlexiGene DNA Kit (Qiagen, Redwood City, CA, USA). Next, 500 ng of genomic DNA was bisulfite converted using the EpiTect Bisulfite Kit (Qiagen) according to the manufacturer's protocol. Primer pairs to amplify and sequence the chosen CpG genomic positions were designed with PyroMark Assay Design version 2.0.1.15 (Qiagen) (Supplementary Table S1) and bisulfite PCR reactions were carried out on a VeritiTM Thermal Cycler (Applied Biosystems, Foster City, CA, USA). Next, 20 µL of the biotinylated PCR product was immobilized using streptavidin-coated Sepharose beads (GE Healthcare Life Sciences, Piscataway, NJ, USA) and 0.4 µM of sequencing primer annealed to purified DNA strands. Pyrosequencing was performed using PyroMark Gold Q96 reagents (Qiagen) on a PyroMark™ Q96 ID System (Qiagen). For each particular CpG, DNA methylation levels were expressed as the percentage of methylated cytosines over the sum of total cytosines. Unmethylated and

methylated DNA samples (EpiTect PCR Control DNA Set, Qiagen) were used as controls for the pyrosequencing reaction.

#### *2.4. Extension of NXN Gene Methylation Mapping by Bisulfite Cloning Sequencing*

Previously bisulfite-converted genomic DNA was used to validate pyrosequencing results. Primer pair sequences were designed by MethPrimer [15] (Supplementary Table S1). PCR products were cloned using the TopoTA Cloning System (Invitrogen, Carlsbad, CA, USA); a minimum of 10–12 independent clones were sequenced for each triplicate, cell condition, and region (Sanger sequencing) [16]. Methylation graphs were obtained with the QUMA software [17].

#### *2.5. Neurogenesis Markers mRNA Expression: Analysis by Real-Time Quantitative PCR (RT-qPCR)*

Total RNA was extracted from frozen pellets of basal NPCs and the control or Aβ peptide treated NPCs incubated in differentiation media for 9, 19, or 29 days using the RNeasy Mini kit (QIAGEN, Redwood City, CA, USA) following the manufacturer's instructions. Genomic DNA was digested with DNase I (RNase-Free DNase Set, Qiagen). RNA concentration and purity were determined using a NanoDrop spectrophotometer. Complementary DNA (cDNA) was reversely transcribed from 1000 ng total RNA with SuperScript® III First-Strand Synthesis Reverse Transcriptase (Invitrogen) after priming with oligo-d (T) and random primers. RT-qPCR reactions were performed in duplicate with Power SYBR Green PCR Master Mix (Invitrogen) in a QuantStudio 12 K Flex Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Sequences of primer pairs were designed using a real-time PCR tool (IDT, Coralville, IA, USA) (listed in Supplementary Table S1). Relative mRNA expression levels of lineage-specific genes in a particular sample were calculated as previously described [18] and the geometric mean of the *ACTB* and *GAPDH* genes used as reference to normalize the expression values.

#### *2.6. Immunofluorescence Staining*

NPCs were seeded on Nunc™ Lab-Tek™ II chamber slides (Thermo Fisher Scientific, Waltham, MA, USA), coated with CellMatrix Basement Membrane Gel. Cells were either left untreated or treated with Aβ protein fragment 1–42 (50 nM) in differentiation media, as described above. After 9, 19, or 29 days of incubation, cells were fixed with 4% formalin (OPPAC, Noain, Spain) for 15 min; next, they were permeabilized using 0.5% TWEEN® 20 (Sigma-Aldrich) in DPBS and blocked with 10% fetal bovine serum (Sigma-Aldrich) containing 0.5% Tween in DPBS for 30 min at room temperature. Rabbit monoclonal anti-NeuN [EPR12763] (Cat# ab177487, RRID:AB\_2532109; 1:300), anti-GFAP [EP672Y] (Cat# ab33922, RRID:AB\_732571; 1:300), anti-Synaptophysin [YE269] (Cat# ab32127, RRID:AB\_2286949; 1:200) and anti-Ki67 [SP6] (Cat# ab16667, RRID:AB\_302459; 1:500) primary antibodies (Abcam, Cambridge, UK) diluted in blocking buffer were added and incubated overnight at 4 ◦C. After three washing steps, Alexa Fluor® 647 donkey anti-rabbit secondary antibody (Abcam Cat# ab150075, RRID:AB\_2752244; 1:500) was added and incubated for 30 min at room temperature in the dark. Following three washing steps, the slides were mounted with ProLong™ Gold Antifade Mountant with DAPI (Molecular Probes, OR, USA). Immunofluorescence images were obtained using a Cytation 5 Cell Imaging Multi-Mode Reader and analyzed with the Gen5™ software (BioTek, Winooski, VT, USA).

#### *2.7. Statistical Data Analysis*

Statistical analyses were performed with the SPSS version 21.0 (IBM, Inc., Armonk, NY, USA) and GraphPad Prism version 6.00 for Windows (GraphPad Software, La Jolla, CA, USA). We first checked that all continuous variables had a normal distribution using the one-sample Shapiro–Wilk test. Significance level was set at *p*-value < 0.05. Differences between the various time points for mRNA levels of specific genes and percentages of DNA methylation were assessed by one-way analysis of variance (one-way ANOVA) followed

by post hoc Tukey's honestly significant difference test. In cases where the Levene test did not show homogeneity of variance, Welch's ANOVA followed by Dunnett's T3 were conducted. Non-parametric data were analyzed using the Kruskal–Wallis test. A paired *t*-test was used to analyze differences in methylation or expression levels of the studied genes between Aβ peptide treated and control groups at each time point. GraphPad Prism version 6.00 for Windows was used to draw the graphs.

#### **3. Results**

#### *3.1. Time-Related Changes in Cultured NPCs during Neural Differentiation*

To determine whether neural differentiation was effectively induced, we first examined any morphological modifications of the cells over time. As shown in Figure 1A, NPCs exposure to differentiation medium led to an increase in the number and length of neuritic extensions, which even connected with the extensions of neighboring cells in comparison with basal cells grown in proliferation medium at Time 0. These changes in cell morphology, typical of cells undergoing differentiation [19,20], were noticed from the first time point (day 9), becoming more evident over time in response to directed neurogenesis.

expression. (**A**) Phase-contrast images on days 0, 9, 19, and 29 of basal cells incubated in expansion medium and control cells incubated in differentiation medium (10× magnification with 20× magnification inset lens; the scale bar is 100 µm). (**B**) The graph shows Ki67 proliferation marker expression for control and Aβ-treated NPCs at 9, 19, and 29 days of culture in differentiation medium. Data represent the mean value ± standard error of the mean (SEM).

The total cell number in NPC cultures remained steady because of no proliferation, confirmed by unchanged Ki67 protein marker expression in control or exposed to Aβ peptide cells (Figure 1B), which was associated with a gradual boost of cell differentiation. In fact, immunofluorescence (IF) staining revealed neuronal nuclei (NeuN) and synaptophysin protein expression, which mark neurons and synaptic vesicles in the NPC culture (Figure 2).

**Figure 2.** Immunofluorescence staining of NPC differentiation. (**A**) Representative images show NeuN and synaptophysin protein expression on days 9, 19, and 29 in NPCs incubated in differentiation medium (20× magnification (the scale bar is 100 µm) with 10× magnification 4 × 4 montage inset (the scale bar is 300 µm)). (**B**) The graphs show NeuN and synaptophysin markers expression for control and Aβ treated NPCs at 9, 19, and 29 days of culture in differentiation medium. Data represent the mean value ± SEM.

To confirm the above observations, we explored if gene expression profiles of different TFs and molecular markers had changed in our *in vitro* model across consecutive stages of driven neuronal differentiation. For that, we measured mRNA expression levels of the Neuronal Differentiation 1 (*NEUROD1*), Neural Cell Adhesion Molecule 1 (*NCAM1*), Tubulin Beta 3 Class III (*TUBB3*), RNA Binding Fox-1 Homolog 3 (*RBFOX3*), Calbindin 1 (*CALB1*), and Glial Fibrillary Acidic Protein (*GFAP*) genes by RT-qPCR (Figure 3). Expression levels of all genes but *CALB1* changed over time.

*NEUROD1* mRNA expression levels of NPCs increased in differentiation medium. Statistically significant increases of mRNA expression for this basic helix-loop-helix (bHLH) TF on days 9 (*p*-value < 0.05), 19 (*p*-value < 0.05) and 29 (*p*-value < 0.001) were observed in comparison to basal cells.

In our *in vitro* model, *NCAM1* mRNA expression overlapped that of *NEUROD1*. We found a statistically significant increase from the addition of differentiation medium to the cell culture (F(3,17) = 31.85, *<sup>p</sup>*-value = 3.3634 <sup>×</sup> <sup>10</sup>−<sup>7</sup> ), which was more pronounced on day 19 (*p*-value < 0.001). Significant differences were also seen between days 9 and 19 (*p*-value < 0.001), days 9 and 29 (*p*-value < 0.01) and between basal cells and any of the other time points: from day 0 to day 9 (*p*-value < 0.01) and from day 0 to day 29 (*p*-value < 0.001).

Once the proliferation medium was changed for differentiation medium, NPCs began to express *TUBB3* mRNA, a gene marker with a key role for proper axon guidance and maintenance. This increase remained constant over time in comparison to basal cells (*p*-value < 0.01). However, no changes were observed between the first, second, and third time points.

*RBFOX3* encodes the NeuN antigen, which has been widely used as a marker for postmitotic neurons. In our study, *RBFOX3* mRNA expression progressively rises over time, proving the successful achievement of progenitor-to-neuron differentiation. Statistically significant differences in the rise of mRNA expression between day 0 and day 9 (*p*-value < 0.01), day 9 and day 19 (*p*-value < 0.01) and day 9 and day 29 (*p*-value < 0.05) were seen. Likewise, all other differences between any time point with respect to basal cells were also statistically significant: from day 0 to day 19 (*p*-value < 0.01) and from day 0 to day 29 (*p*-value < 0.05).

124

Regarding *CALB1* mRNA expression, and given that this gene encodes a protein expressed in mature granule cells, no significant changes were detected.

A statistically significant rise in *GFAP* mRNA expression was observed on day 29 in comparison with basal cells (*p*-value < 0.01) and day 9 of differentiation (*p*-value < 0.01). This suggested the presence of NPCs-derived astrocytes in the culture.

None of the neuronal lineage-specific genes showed significant mRNA expression differences between day 19 and day 29.

#### *3.2. Assessment of Epigenetic Markers Involved in Neurogenesis in Differentiating NPCs*

DNA methylation levels of four neurogenesis-related genes previously found to be altered in the AD hippocampus [6] were quantified by bisulfite pyrosequencing. The same genomic loci identified in the human hippocampus were used to assess DNA methylation levels, corresponding to the genes Contactin-Associated Protein 1 (*CNTNAP1*), *SEPT5-GP1BB* Readthrough (*SEPT5-GP1BB*), T-Box Transcription Factor 5 (*TBX5*), and Nucleoredoxin (*NXN*) (Table 1 and Supplementary Figure S1).

No significant differences in DNA methylation levels were observed for *CNTNAP1*, *SEPT5-GP1BB*, and *TBX5* throughout the differentiation process within the time frame of this study (Figure 4A–C).

Nonetheless, changes in *NXN* methylation levels were observed. Two CpG positions were assessed for the *NXN* gene. For the first one, DNA methylation levels increased on day 9 (*p*-value < 0.01) and were maintained over time; statistically significant differences were also seen on day 19 (*p*-value < 0.01) and day 29 (*p*-value < 0.01) with respect to basal cells (Figure 4D). Regarding the CpG following cg19987768, the pyrogram revealed a similar methylation pattern (day 9 vs. day 0: *p*-value < 0.05; day 19 vs. day 0: *p*-value < 0.01; day 29 vs. day 0: *p*-value < 0.05) (Supplementary Figure S2A). The same differences in methylation levels were observed for both CpGs together (day 9 vs. day 0: *p*-value < 0.001; day 19 vs. day 0: *p*-value < 0.001; day 29 vs. day 0: *p*-value < 0.001) (Supplementary Figure S2B). These findings led us to extend the methylation local mapping for the *NXN* gene using bisulfite cloning sequencing. We confirmed that average DNA methylation levels across all CpG sites for the amplicon were statistically significantly higher at every time point in comparison to day 0 (day 9 vs. day 0: *p*-value < 0.001; day 19 vs. day 0: *p*-value < 0.001; day 29 vs. day 0: *p*-value < 0.05) (Figure 5). Additionally, this approach revealed a decrease in *NXN* DNA methylation levels on day 29, which was statistically significant with respect to day 9 (*p*-value < 0.01).

**Figure 5.** *NXN* DNA methylation levels by bisulfite cloning sequencing. (**A**) Percentages of DNA methylation for *NXN* over time. (**B**) *NXN* extended mapping is illustrated by black/white circle-style figures. Black and white circles denote methylated and unmethylated cytosines, respectively. Each column represents a single CpG site in the examined amplicon, and each line represents an individual DNA clone. Average percentages of methylation for each analyzed sample are indicated at the bottom. \* *p*-value < 0.05; \*\* *p*-value < 0.01; \*\*\* *p*-value < 0.001.

We also measured mRNA expression levels of these markers by RT-qPCR (Figure 6).

**Figure 6.** mRNA expression profiles for the *CNTNAP1* (**A**), *SEPT5-GP1BB* (**B**), *TBX5* (**C**), and *NXN* (**D**) genes. Bar graphs represent the percentages of relative mRNA expression for each gene relative to the geometric mean of the *ACTB* and *GAPDH* housekeeping gene expression for NPCs at each time point of culture. Mean values ± SEM. \* *p*-value < 0.05; \*\* *p*-value < 0.01; \*\*\* *p*-value < 0.001.

*CNTNAP1* mRNA expression levels progressively increased over time with statistically significant differences on day 19 (*p*-value < 0.05) and day 29 (*p*-value < 0.001) in comparison to basal cells. Moreover, significant expression differences were noticed between day 9 and day 29 (*p*-value < 0.01) (Figure 6A).

From day 19, a significant increase in mRNA expression for *SEPT5-GP1BB* was detected (*p*-value < 0.01) and maintained on day 29 (*p*-value < 0.01). Furthermore, mRNA expression on day 19 (*p*-value< 0.001) and day 29 (*p*-value< 0.001) was also significantly higher than for day 0 (Figure 6B).

mRNA levels for the *TBX5* gene increased on day 29 with statistically significant differences in comparison to the cells in culture on day 0 (*p*-value < 0.05) and day 9 (*p*-value < 0.05) (Figure 6C).

Finally, significant differences were observed from the addition of the differentiation medium for the *NXN* gene in terms of gene expression (day 9 vs. day 0: *p*-value < 0.01; day 19 vs. day 0: *p*-value < 0.01); day 29 vs. day 0: *p*-value < 0.05) (Figure 6D). The increase in mRNA expression continued to day 19 (0.384 ± 0.117; *p*-value < 0.05).

Overall, similar transcriptional patterns for the *TBX5* and *GFAP* genes and the *NXN*, *NCAM1* and *RBFOX3* genes during the NPCs culture period, were observed.

#### *3.3. Effect of Aβ Peptide Addition on Cultured NPCs during the Stages of Neurogenesis*

To mimic the cell environment in AD, we exposed NPCs to Aβ peptide 1–42 once a week during the differentiation period. First, we assessed whether the expression levels of the genes selected to characterize each stage of neurogenesis in culture were altered due to the addition of the Aβ peptide (Figure 7).

**Figure 7.** Effect of the addition of Aβ peptide 1–42 during the differentiation period. mRNA expression of the *NCAM1* (**A**), *TUBB3* (**B**), *RBFOX3* (**C**), and *SEPT5-GP1BB* (**D**) genes relative to the geometric mean of *ACTB* and *GAPDH* housekeeping genes expression was determined for the controls and Aβ peptide treated NPCs on days 9, 19, and 29. Vertical lines represent the SEM. \* *p*-value < 0.05; \*\* *p*-value < 0.01.

We found transient and mild treatment-specific differences in mRNA expression for some of the studied lineage-specific genes. The Aβ peptide reduced *NCAM1* expression (*p*value < 0.05) on day 19 (Figure 7A), and *TUBB3* (*p*-value < 0.05) and *RBFOX3* (*p*-value < 0.01) expression on day 9 (Figure 7B,C). Interestingly, such decreases occurred at the beginning or in between the studied time window, but these differences were no longer significant at the end time point (day 29).

Next, we assessed how the addition of Aβ peptide affected mRNA expression of neurogenesis-related genes and if the changes had any relationship with their methylation status.

We observed a statistically significant decrease in *SEPT5-GP1BB* mRNA on day 19 (*p*-value < 0.05) (Figure 7D) and an increase in the percentage of DNA methylation with a trend towards statistical significance on day 29 (*p*-value = 0.082) with the addition of the Aβ peptide to the culture.

Finally, the Aβ peptide slightly reduced *NXN* mRNA expression on day 9 (*p*-value < 0.05) which is maintained until day 19 (*p*-value < 0.05) (Figure 8A). *NXN* methylation

seems to decrease on day 9 but does not reach statistical significance (*p*-value = 0.11). One possible explanation for this is that the sample size is insufficient to show statistical significance. Interestingly, a rise in the percentage of *NXN* methylation level of Aβ peptide-treated cells was seen on day 29, measuring all amplicon CpG sites (*p*-value < 0.05) (Figure 8B), when the decrease in *NXN* mRNA expression is no longer observed.

**Figure 8.** Effect of Aβ peptide 1–42 addition on the *NXN* gene during the differentiation period. mRNA expression relative to the geometric mean of *ACTB* and *GAPDH* housekeeping genes expression (**A**). DNA methylation level in the extended mapping amplicon (**B**) were determined for the controls and Aβ peptide-treated neural progenitor cells on days 9, 19, and 29. Vertical lines represent the SEM. \* *p*-value < 0.05.

#### **4. Discussion**

To date, a broad overview of the stages of AHN exists. This complex multistep process can be divided into four phases: a precursor cell phase, an early survival phase, a postmitotic maturation phase, and a late survival phase. Type 1 radial glia-like cells (RGLs) represent the NSC population that can differentiate into TAPs (type 2 cells), which initially have a glial (type 2a) and then a neuronal (type 2b) phenotype. Through a migratory neuroblast-like stage (type 3), lineage-committed cells exit the cell cycle ahead of maturation into dentate granule neurons functionally integrated into the hippocampal circuitry [21,22]. Based on cell morphology TFs expression and a set of marker proteins, distinct milestones have been established [21]. In this study, we examined the expression dynamics of key markers in order to characterize a directed human NPCs differentiation model across distinct differentiation stages (Figure 9) to test new AHN epigenetic and expression markers that might be associated with AD.

During stage 1 (proliferation phase), type 1 RGL cells express GFAP. However, no differences in *GFAP* expression are detected until day 19 after the addition of the differentiation medium. This suggests that our *in vitro* NPCs culture window starts after the proliferative phase, during stage 2, when type-2 cells (differentiation phase) lose the GFAP marker [22]. Thus, in contrast to their *in vivo* counterparts in the SGZ of the brain (some authors describe that the *in vitro* expanded NSCs are less neurogenic and mainly biased

towards an astrocytic fate upon differentiation [20]), *GFAP* expression on day 19 would correspond to a subset of astrocytes present in our NPCs culture [23].

**Figure 9.** Expression pattern of AHN lineage-specific genes, assessed to characterize our NPCs *in vitro* model. The diagram illustrates *NEUROD1*, *NCAM1*, *TUBB3*, *RBFOX3*, *CALB1*, and *GFAP* gene expression profiles during directed neuronal differentiation for our time window NPCs culture model, based on the developmental stages of AHN within the neurogenic niche of the DG.

In stage 3 (migration phase), migrating neuroblasts display the polysialylated form of NCAM (PSA-NCAM), a marker that appears at the late stage of AN and seems to persist in young postmitotic neurons [24]. Accordingly, our results suggest the presence of a plateau between day 19 and day 29 for *NCAM1* mRNA expression. Most PSA-NCAM-positive cells express NeuroD and NeuN, but not GFAP, which supports the abovementioned findings [24]. bHLH TF *NEUROD1* plays an essential role in the differentiation and survival of neuronal precursors in the SGZ. NeuroD1 deletion leads to new granule neurons depletion and their failure to integrate into the DG [25]. In line with findings by Xuan Yu et al. [26], we observed a rise of *NEUROD1* gene expression during our culture time window. Moreover, expression of NeuroD can also be detected in PSA-NCAM-positive cells, precedes it [24], and reaches the highest point in late-stage type 2b and type 3 cells [2]. Once the newly generated neurons become postmitotic, they begin to express the NeuN marker, which is consistent with an earlier *RBFOX3* mRNA expression in our model. We found that *RBFOX3* expression increases until days 19 and 29 of differentiation, showing an expression profile similar to that of *NCAM1*.

Next, cells become postmitotic entering stage 4 (axonal and dendritic targeting). Immature neurons still express PSA-NCAM and, at the same time, can also be marked by NeuN. *TUBB3*, involved in axon guidance and maintenance, is expressed simultaneously; it encodes a class III member of the beta-tubulin protein family, characteristic of early postmitotic and differentiated neurons and some mitotically active neuronal precursors. This is consistent with the increase in *TUBB3* mRNA detected in our model, prior to its translation into protein. *TUBB3* mRNA expression persists in neurons displaying high complexity and electrophysiological properties, such as very low capacitance, high input resistance, depolarized resting membrane potential, and lack of synaptic activity, which show immunoreactivity for NeuN and thus represent postmitotic neurons [24,27].

Finally, mature granule cells establish their synaptic contacts and become functionally integrated into the hippocampus in stage 5 (synaptic integration), expressing calbindin along with NeuN but without co-expressing PSA-NCAM [24]. We do not find variations in *CALB1* mRNA expression within the analyzed culture time window, which may occur later in time. We indeed detect synaptophysin in the IF study on day 29, which suggests that our time window ends early at the synaptic integration phase.

Hence, by culturing NPCs as a monolayer in a medium that accelerates neuronal differentiation by enhancing synaptic activity [14], we achieve a less time-consuming differentiation strategy that resembles the *in vivo* developmental program of human hippocampal DG, which differs from that of the SVZ [28], as we are able to generate developing neurons potentially expressing relevant features of the AHN process.

Once the first objective was accomplished, we evaluated whether a set of AD-related differentially methylated genes targeted specific AHN milestones. These genes had been identified in a previous study of the human hippocampus and annotated as neurogenesis genes following a curated review of the literature [6]. No differences in DNA methylation for the *CNTNAP1*, *SEPT5-GP1BB*, and *TBX5* genes were identified within the period of this study. Only one or two CpGs were analyzed for each gene, those that had been identified as differentially methylated in the hippocampus of AD patients, so changes in DNA methylation may be present in other regions of the gene and may not have been detected with our approach. Still, changes in DNA methylation may occur before or after our time window.

However, it is worth noting that all the above genes undergo mRNA expression changes, suggesting they could be considered potential molecular markers of different AHN stages (Figure 10). Further studies should be carried out to confirm this.

**Figure 10.** Expression patterns of neurogenesis-related genes evaluated in our *in vitro* model on NPCs. The illustration depicts the expression profiles of the *CNTNAP1*, *SEPT5-GP1BB*, *TBX5*, and *NXN* genes during directed neuronal differentiation of our time window culture model on NPCs, according to the developmental stages of AHN within the neurogenic niche of the DG.

*CNTNAP1* and *SEPT5-GP1BB* mRNA expression levels increase on day 19/29, possibly identifying immature neurons, when axonal and dendritic targeting occurs. Indeed, *CNTNAP1* encodes a type I integral membrane protein that regulates the intracellular processing and transport of contactin to the cell surface [29,30], also known as contactinassociated protein (CASPR), which is present in synapses and interacts with AMPA (αamino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) glutamate receptors that mediate fast excitatory synaptic transmission in the central nervous system (CNS) [31]. CASPR is an

adhesion molecule crucial to forming axoglial paranodal junctions surrounding the nodes of Ranvier in myelinated axons [32].

Known to be a negative regulator of neurite outgrowth in CNS neurons [30], CASPR1 plays an essential role in the timing of neuron and astrocyte development in the mouse cerebral cortex by repressing the transcription of the Notch effector Hes1. In radial glial cells, CASPR1 deficiency delays the generation of cortical neurons and induces the early formation of cortical astrocytes without affecting the number of progenitor cells. Thus, during the neurogenic period, CASPR1 is highly expressed, while during the gliogenic period its expression decreases [32,33]. Moreover, CASPR1 has been reported to be under the regulation of the astrocytic methyl-CpG-binding protein 2 (MeCP2) along with key myelin genes and proteins [34].

For its part, *SEPT5-GP1BB* is originated from naturally occurring read-through transcription between the neighboring *SEPT5* (*SEPTIN5*) and *GP1BB* (Glycoprotein Ib Platelet Subunit Beta) genes on chromosome 22. Inefficient use of an imperfect polyA signal in the upstream *SEPT5* gene causes transcription to continue into the *GP1BB* gene. The Genotype Tissue Expression (GTEx) Project established by the National Institutes of Health (NIH) Common Fund shows the highest median expression of this gene in the brain cortex, but to the best of our knowledge, this is the first study describing *SEPT5-GP1BB* as a possible key marker of temporal specification of cell fate in neurogenesis.

The *TBX5* gene displays the highest level of mRNA expression on day 29. It belongs to a phylogenetically conserved family of genes sharing a common DNA-binding domain, the T-box, which encodes TFs involved in the regulation of developmental processes. Accordingly, it is considered pivotal in the establishment of the cardiac lineage [35]. Moreover, *TBX5* regulates the development of the vertebrate eye [36] and limb skeletogenesis [37]. Here, we observe a statistically significant increase in *TBX5* mRNA at the end time point of our culture (day 29), and therefore, we propose it as a transcriptional candidate marker of postmitotic differentiating cells that may exhibit a peak of expression in the transition of immature to mature neurons.

The most relevant findings of our study relate to the *NXN* gene. At CpG site resolution, *NXN* shows differential methylation at every time point in comparison to basal cells. Moreover, when we extend the mapping and further average across all CpG sites of the amplicon, we confirm these findings and show that peak methylation of *NXN* occurs on day 9. Such curve outlined by the percentage of *NXN* DNA methylation would range from type-2a/2b TAPs to immature neurons, peaking at type 3 neuroblasts. This may allow to discriminate the migration stage of neurogenesis.

Interestingly, the increase in *NXN* methylation is associated with higher mRNA expression levels during our culture time window. DNA methylation at gene promoter regions usually represses gene expression through the recruitment of methylated DNA-binding protein family members, such as methyl-CpG-binding protein 1 (MBD1) and MeCP2. Nevertheless, DNA methylation roles in gene regulation appear complex and multi-faceted and genome structure integration becomes of major importance [38]. In the same way that CG-rich and CG-poor regulatory elements may undergo distinct modes of epigenetic regulation [38], DNA methylation has been linked to gene activation within the transcribed regions and the highest levels of gene body methylation may enhance transcription [39]. Indeed, it is precisely in this region where the studied DMPs are located (Supplementary Figure S1). Thus, DNA methylation has been previously correlated with increased expression in human embryonic stem (ES) cells in an *in vitro*-induced differentiation work [40,41]. Furthermore, gene expression is not only regulated by methylation in the same region, but by other epigenetic mechanisms or methylation in other regulatory areas. Several gene regulatory elements seem to communicate on the same or different chromosomes. Enhancers and insulators participate in this higher-order organization of chromatin [42]. In fact, sequential recruitment of lineage-restricted transcription factors leads to enhancers being activated or maintained in a poised state upon stem cell differentiation [43].

*NXN* is a ubiquitously expressed endogenous antioxidant, member of the thioredoxin antioxidant superfamily [44,45]. In brain sections of mice, there is a predominant neuronal expression of *NXN* in septal nuclei and the hippocampus, in which its deletion results are embryonically lethal, mainly due to cranial defects and deformities [46]. Specifically, immunoreactive signals of *NXN* were found in fibers in the cortex, hippocampus, and cerebellum [46].

In proliferating cells, NXN sequesters dishevelled segment polarity protein 2 (DVL2). Upon the increase in ROS, NXN releases DVL2, relaying the WNT signal to downstream effectors. As a result, cytosolic β-catenin accumulates and shuttles to the nucleus where it drives specific expression of target genes relevant to neuronal differentiation [44,45,47]. NXN also retains a pool of inactive Dvl by preventing the possible interaction of Dvl and kelch-like protein 12 (KLHL12) and its subsequent ubiquitination and degradation, ensuring a prompt activation upon Wnt stimulation [46]. In agreement with this, it has been proved that *NXN* knockdown of SH-SY5Y human neuroblastoma cells increases proliferation and cell cycle reentry [48]. Accordingly, in our *in vitro* model, the increased expression of *NXN* mRNA levels is consistent with the absence of cell proliferation.

The literature points to interactions with further partners that include histone deacetylase 6 (HDAC6), heat shock protein 90 kDa (HSP90), and calcium calmodulin kinase 2a (Camk2a), a postsynaptic kinase crucial for neuronal plasticity [46,48]. Moreover, *NXN* may be implicated in transcriptional regulation, promoting the induction of the TFs CREB (cAMP response element-binding protein), NF*κ*B (nuclear factor kappa B), and AP-1 (activator protein-1) [46].

In the context of AD, it is known that Aβ peptides are generated after the cleavage of APP by γ-secretase in the amyloidogenic pathway [10]. In previous models, the physiological concentration of Aβ peptides in the brain revealed a positive effect on neuroplasticity and learning, showing improved hippocampal long-term potentiation (LTP), while high nanomolar Aβ administration resulted in impaired cognition [49,50], suggesting a hormetic nature [51]. Because low picomolar levels of extracellular concentrations of Aβ in the normal brain have been estimated, in our experiments we chose a concentration of Aβ peptide 1–42 in the nanomolar range (50 nM), added once a week during the 29 days of culture, a single dose determined by the average of concentrations used by Gulisano et al. [52] and Malmsten et al. [53].

It has been reported that the synthetic Aβ peptide 1–42 oligomer decreases human NSC proliferative potential and appears to favor glial differentiation; it reduces neuronal cell fates [10] or suppresses the number of functional human ES cells-derived neurons [54]. Nonetheless, Bernabeu-Zornoza et al. showed that 1 µM monomeric Aβ peptide 1–42 promoted human NSCs proliferation by increasing the pool of glial precursors, without affecting neurogenesis [55]. On the other hand, differentiating neurospheres exposed to fibrillar Aβ decreased neuronal differentiation and induced gliogenesis [54]. The existing controversies may be due to Aβ isoforms, peptide concentrations, aggregation state, administration times, or type of NSCs/NPCs from different species or culture systems used in each experiment [55].

In our work, some of the analyzed genes show a mild decrease in mRNA expression after Aβ 1–42 addition. This transient effect is evident on day 9 or day 19, not occurring on day 29, suggesting that, despite affecting genes involved in the fate of neurogenesis, probably before cells maturation and leading to a decrease in differentiation, the addition of nanomolar concentrations of Aβ is somehow counteracted in the long-term. A timedependent reversal of the effects of picomolar Aβ on synaptic plasticity and memory had been already seen by Koppensteiner et al., attributable to the enzyme neprilysin, whose levels are reduced with aging and in the brains of AD patients [56]. In fact, a study in which mutant APP was overexpressed to ensure Aβ release exclusively by mature neurons, found neither a positive nor a negative effect in AHN [57]. Hence, our simplistic model may shed light on early AD neurogenesis events, before Aβ deposition cannot be overcome.

A transcriptomic analysis of several human AD profiles demonstrated upregulation of neural progenitor markers expression and downregulation of later neurogenic markers, implying that neurogenesis is reduced in AD due to compromised maturation [58]. Interestingly, the authors showed downregulation of *NCAM1* expression in the hippocampus of early-stage AD, as well as of *NCAM1*, *TUBB*, and *RBFOX3* in late-stage AD, which is in line with our results after the addition of Aβ to the culture. Moreover, Moreno-Jimenez et al. recently provided evidence for substantial maturation impairment underlying AD progression. They identified a decline in doublecortin-expressing cells that co-expressed PSA-NCAM in the DG starting at Braak stage III, followed by a reduction in the expression of NeuN and βIII-tubulin, among others, at some of the subsequent stages of the disease [12].

Our results also show a decrease in *SEPT5-GP1BB* mRNA expression on day 19 when Aβ 1–42 was added to the culture. Again, this suggests that even low levels of Aβ peptide deposit may already have an effect on neuronal fate. For *NXN*, such a decrease in mRNA expression was also observed on day 9 and day 19 cultures. No changes were seen on day 29 when the percentage of methylation levels in the *NXN* amplicon increased in differentiating cells with Aβ 1–42.

Thus, *NXN* emerges as a candidate gene that needs to be further studied to address its ability to determine not only the temporal sequence of neurogenesis but simultaneously the differences in the AD brain due to Aβ peptide deposition.

AHN confers a unique mode of plasticity to the mature mammalian brain. Research in this field requires non-invasive monitoring to understand the lifelong impact [59]. Easier than manipulating NSCs, in part because of the time saving, our NPCs model facilitates studying gene expression levels in an *in vitro* cell culture platform within a human context [60]. Moreover, this straightforward approach may help further understand the alterations affecting specific lineage cell types in presence of the Aβ peptide, including early pathological changes, possibly associated with prodromal phases. On the other hand, other cell types are involved in pathogenesis, particularly microglia, which play a major role, together with neuroinflammation, in the risk of developing AD and its progression. In consequence, co-cultures with other cell types present at neurogenic niches, such as microglia, may be implemented to overcome the limitations presented by the characteristics of an *in vivo* niche environment.

Finally, the development of AHN monitoring methods as biomarkers for cognitive function in live individuals will be crucial to staging AD progress. Moreover, studying the utility of TF reprogramming to preserve endogenous AHN may contribute to cognitive resilience in AD [58]. However, despite the enthusiasm, the prospect of using adult NSCs therapeutically as a regenerative source needs to address neuronal integration and its impact on host mature neural circuits [59]. It will involve strategies to accomplish the NSC pool maintenance, generation of correct neuronal subtypes, suppression of glial fates, and differentiation and survival of immature neurons [2].

#### **5. Conclusions**

In this work, we present the transcriptional profiles of a number of genes involved in specific stages of the AHN process for a thorough understanding of the lineage-restricted fate during human neuronal differentiation. The addition of Aβ peptide 1–42 to our human NPCs culture model, generates results that are similar to those obtained in human AD samples regarding the expression of the *NCAM1*, *TUBB3*, and *RBFOX3* genes, offering an *in vitro* opportunity to study AHN impairment in the AD context. Considering this approach, the *NXN* gene shows a rise in DNA methylation, the maximum being coincident in time with type 3 neuroblasts and displays differential DNA methylation in immature neurons in presence of the Aβ peptide. Moreover, *CNTNAP1*, *SEPT5-GP1BB*, *TBX5*, as well as *NXN* were revealed as mRNA expression molecular markers for specific stages of AHN. Finally, differentiating NPCs decrease their *SEPT5-GP1BB* or *NXN* mRNA expression at different neurogenesis time points with the addition of the Aβ peptide to the culture.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cells11071069/s1. Supplementary Table S1. Bisulfite pyrosequencing, bisulfite cloning sequencing and RT-qPCR primers. Supplementary Figure S1. Genomic positions of the CpGs analyzed by pyrosequencing. Supplementary Figure S2. *NXN* DNA methylation levels in differentiating NPCs.

**Author Contributions:** I.B.-L. contributed to the conception and design of the work, running the experiments, analysis and interpretation of data, figure design and drawing and drafting/review of the manuscript for content; B.A. contributed to running the experiments, statistical analysis and review of the manuscript for content; A.U.-C. contributed to figure design and drawing and review of the manuscript for content; E.G.-O. contributed to the design and review of the manuscript for content. M.R. contributed to running the experiments; D.R.P.-R. contributed to the interpretation of data and review of the manuscript for content; J.C. contributed to the design and review of the manuscript for content; M.M. contributed to the conception and design of the work, analysis and interpretation of data, statistical analysis, study supervision, drafting/review of the manuscript for content and funding acquisition. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Spanish Government through a grant from the Institute of Health Carlos III (FIS PI17/02218), jointly funded by the European Regional Development Fund (ERDF), European Union, A way of shaping Europe; the Trans-Pyrenean Biomedical Research Network (REFBIO II-MOMENEU project) and the Government of Navarra through two grants from the Department of Industry of the Government of Navarra (PI058 iBEAS-Plus and PI055 iBEAS-Plus). AUC received a grant Doctorandos industriales 2018–2020 and a Predoctoral grant (2019) founded by the Department of Industry and Health of the Government of Navarra. MM received a grant Programa de intensificación- (LCF/PR/PR15/51100006) founded by Fundación Bancaria la Caixa and Fundación Caja-Navarra, and Contrato de intensificación from the Institute of Health Carlos III (INT19/00029).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** We want to kindly thank Valle Coca (Navarrabiomed BrainBank, technical support), Paula Aldaz and Imanol Arozarena (Cancer Signalling Research Unit, Navarrabiomed, technical and scientific support), Natalia Ramirez (Haematological Oncology Research Unit, Navarrabiomed, scientific support), Ibai Tamayo, Arkaitz Galbete, Mónica Enguita, Berta Ibañez, and Julián Librero (Methodology Unit, Navarrabiomed, technical support) for their help.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **Abbreviations**

Aβ: amyloid β; AD: Alzheimer's disease; AHN: adult hippocampal neurogenesis; AMPA: α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; AN: adult neurogenesis; ANOVA: two-way analysis of variance; AP-1: activator protein-1; APP: amyloid precursor protein; BDNF: Brain-Derived Neurotrophic Factor; bHLH: basic helix-loop-helix; *CALB1*: calbindin 1; Camk2a: calcium calmodulin kinase 2a; CASPR: contactin-associated protein; cDNA: complementary DNA; *CNTNAP1*: Contactin-Associated Protein 1; *CREB*: cAMP response element-binding protein; DG: dentate gyrus; CNS: central nervous system; DMPs: differentially methylated positions; DVL2: Dishevelled segment polarity protein 2; ES: embryonic stem; GDNF: Glial-Derived Neurotrophic Factor; *GFAP*: Glial Fibrillary Acidic Protein; *GP1BB*: Glycoprotein Ib Platelet Subunit Beta; HDAC6: histone deacetylase 6; HSP90: heat shock protein 90 kDa; IF: immunofluorescence; iPSCs: induced pluripotent stem cells; KLHL12: kelch-like protein 12; LTP: long-term potentiation; MBD1: methyl-CpG-binding protein 1; MeCP2: methyl-CpG-binding protein 2; *NCAM1*: Neural Cell Adhesion Molecule 1; NeuN: neuronal nuclei; *NEUROD1*: Neuronal Differentiation 1; NF*κ*B: nuclear factor kappa-B; NPCs: neural progenitor cells; NSCs: neural stem cells; *NXN*: Nucleoredoxin; PSA-NCAM: polysialylated form of NCAM; PSCs: pluripotent stem cells; *RBFOX3*: RNA Binding Fox-1 Homolog 3; RGLs: radial glia-like cells; RT-qPCR: real-time quantitative PCR; SE: standard error; SEM: standard error of the mean; *SEPT5*: *SEPTIN5*; *SEPT5-GP1BB*: *SEPT5-GP1BB* Readthrough; SGZ: subgranular zone; SVZ: subventricular zone; TAPs: transient amplifying progenitors; *TBX5*: T-Box Transcription Factor 5; TFs: transcription factors; *TUBB3*: Tubulin Beta 3 Class III.

## **References**


**Gabriela Maria Guerra 1,†, Doreen May <sup>1</sup> , Torsten Kroll <sup>1</sup> , Philipp Koch <sup>1</sup> , Marco Groth <sup>1</sup> , Zhao-Qi Wang 1,2,\*, Tang-Liang Li 1,3 and Paulius Grigaraviˇcius 1,\***


**Abstract:** SMG6 is an endonuclease, which cleaves mRNAs during nonsense-mediated mRNA decay (NMD), thereby regulating gene expression and controling mRNA quality. SMG6 has been shown as a differentiation license factor of totipotent embryonic stem cells. To investigate whether it controls the differentiation of lineage-specific pluripotent progenitor cells, we inactivated *Smg6* in murine embryonic neural stem cells. Nestin-Cre-mediated deletion of *Smg6* in mouse neuroprogenitor cells (NPCs) caused perinatal lethality. Mutant mice brains showed normal structure at E14.5 but great reduction of the cortical NPCs and late-born cortical neurons during later stages of neurogenesis (i.e., E18.5). *Smg6* inactivation led to dramatic cell death in ganglionic eminence (GE) and a reduction of interneurons at E14.5. Interestingly, neurosphere assays showed self-renewal defects specifically in interneuron progenitors but not in cortical NPCs. RT-qPCR analysis revealed that the interneuron differentiation regulators *Dlx1* and *Dlx2* were reduced after *Smg6* deletion. Intriguingly, when *Smg6* was deleted specifically in cortical and hippocampal progenitors, the mutant mice were viable and showed normal size and architecture of the cortex at E18.5. Thus, SMG6 regulates cell fate in a cell type-specific manner and is more important for neuroprogenitors originating from the GE than for progenitors from the cortex.

**Keywords:** SMG6; NMD; neurogenesis; neurodevelopmental syndromes

## **1. Introduction**

Cell fate relies on the correct "read" of the genetic code and its translation into functional proteins. Nonsense-mediated mRNA decay (NMD) is a cellular surveillance mechanism that is involved in controlling the quality of mRNA [1–3]. It degrades transcripts that, after a nonsense mutation or alternative splicing events, harbor a premature termination codon (PTC) before (>50–55 nucleotides) an exon–exon junction complex (EJC). The stable interaction of UPF1 with eRFs at the PTC site recruits the NMD factors UPF2, UPF3 and the kinase SMG1. SMG1 phosphorylates UPF1 and UPF2 thereby promoting the recruitment of the endonuclease SMG6 or the SMG5/SMG7-mediated exonuclease for RNA degradation [1]. The branches of SMG6- and SMG5/7-mediated NMD pathways have been shown to overlap, yet with distinct differences in certain populations of target transcripts [4]. Recently, it was demonstrated that SMG6 (Suppressor with morphogenetic effect on genitalia protein 6) endonuclease activity can depend on the SMG5/SMG7 heterodimer [5]. Moreover, non-mutant transcripts can also be targeted by NMD, for example the transcripts with uORF and long 30UTR, thereby regulating normal gene expression [4,6].

NMD deficiency leads to an accumulation of deleterious mRNA products, which can be translated not only into mutated malfunctional proteins but also into functional

**Citation:** Guerra, G.M.; May, D.; Kroll, T.; Koch, P.; Groth, M.; Wang, Z.-Q.; Li, T.-L.; Grigaraviˇcius, P. Cell Type-Specific Role of RNA Nuclease SMG6 in Neurogenesis. *Cells* **2021**, *10*, 3365. https://doi.org/ 10.3390/cells10123365

Academic Editor: FengRu Tang

Received: 27 October 2021 Accepted: 26 November 2021 Published: 30 November 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

ones, which then may cause tissue dysfunction or pathogenesis [3,7–9]. The complete knockout of NMD genes, such as *Upf1* [10], *Upf2* [11] or *Smg1* [12] resulted in embryonic lethality, highlighting the importance of NMD for early development. Previously, we have shown that the complete deletion of *Smg6* in mouse germline blocks differentiation of embryonic stem (ES) cells into germ layers and thereby results in early embryonic lethality [13]. Furthermore, we demonstrated that SMG6 is required for the production of induced pluripotent stem cells (iPS) from somatic cells [13]. This identified SMG6-mediated NMD as a fundamental regulator of cell fate change, thereby controlling differentiation and development. In that study, the telomeric function of SMG6, originally identified as *Est1a* (Ever Shorter Telomere 1a) in yeast, was proven to be negligible.

In humans, mutations or deletion of NMD factors have been associated with several neurological disorders [14]. Human UPF3B, encoded by the X-linked gene *UPF3B*, was the first NMD factor linked with human neurodevelopmental syndromes such as X-linked intellectual disability (ID) with and without autism, childhood onset schizophrenia (COS) and attention deficit hyperactivity disorder (ADHD) [15,16]. Mutations of *UPF3B* led to the loss or truncation of the protein and caused ID in males of several families [15,16]. Genetic studies have identified mutations or copy number variants in *UPF2*, *UPF3*, *SMG9* and the exon-junction complex component *RBM8A* in human patients with neurological symptoms [17–19]. Furthermore, NMD proteins (UPF3b, UPF1, UPF2, SMG1) have been shown to participate in axon outgrowth, synapse formation and, thus, in various behavioral processes [20–26]. All these observations suggest that, in addition to its general role in the early embryonic development, NMD has specific functions in self-renewal and differentiation of neuroprogenitor cells as well as in neuronal functionality. However, the precise molecular mechanism underlying these processes remains largely unknown.

In order to elucidate the function of SMG6 in neurogenesis during brain development, we applied genetic and cellular studies to inactivate *Smg6* in various neural progenitor cells. We found that SMG6 is more critical for the cell fate determination of interneuron progenitors in the ganglionic eminence (GE) compared to cortical neuroprogenitor cells (NPCs).

#### **2. Material and Methods**

#### *2.1. Mice*

Smg6flox(*Smg6tm1.1Zqw*), Smg6+/∆(*Smg6tm1.2Zqw*) [13], Nes-Cre (*Tg*(*Nes-cre*)*1Kln*) [27], Emx1-Cre (*Emx1tm1*(*cre*)*Krj*) [28], Rosa26-CreERT2 (Gt(ROSA)26Sortm1(cre/ERT2)Tyj), *Smg6*- CNS∆ and *Smg6*-CoHi∆ mice were bred and housed in the mouse facility of Fritz Lipmann Institute (FLI, Jena, Germany). Mice were fed ad libitum with standard laboratory chow and water in ventilated cages under a 12 h light/dark cycle. All animal work was conducted according to the German animal welfare legislation and approved by the Thüringer Landesamt für Verbraucherschutz (TLV). The genotyping of mice was performed by PCR on DNA extracted from tail tissue as previously described [13].

#### *2.2. Immunoblot Analysis*

Proteins were extracted using RIPA buffer supplemented with 1 mM PMSF and 20–50 µg of cell lysates were separated with SDS–PAGE as described [29]. The primary antibodies rabbit anti-*Smg6*/Est1A (1:1500; Abcam, Berlin, Germany), mouse anti-GAPDH (1:20,000; Sigma-Aldrich, Taufkirchen, Germany), mouse anti-Actin (1:20,000; Sigma-Aldrich), and secondary antibodies HRP-conjugated goat anti-rabbit IgG or goat anti-mouse IgG (1:10,000; DAKO, Frankfurt am Main, Germany), were used.

#### *2.3. RT-qPCR*

Total RNA was isolated from neural stem cells using the RNeasy Mini Kit (Qiagen, Hilden, Germany) following the manufacture's recommendations. After genomic DNA removal by DNAse I, the cDNA library was generated using SuperScript™ III Reverse Transcriptase (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). The quantitative

real-time PCR (qPCR) was performed in triplicates for each sample using Platinum™ SYBR™ Green qPCR SuperMix-UDG (Invitrogen) and a LightCycler® 480 Instrument (Roche). The sequences of the used primers are summarized in Table 1.


**Table 1.** Primers used for qPCR analysis.

#### *2.4. Neurosphere Formation Assay and In Vitro Neuronal Stem Cell Differentiation*

Neuroprogenitor cells were isolated and used for neurosphere formation assays from E13.5 cortices and ganglionic eminences as previously described [29]. After isolation, neuroprogenitor cells were plated for neurosphere formation in DMEM/F12 medium (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 2% B-27 supplement (Invitrogen), 1X penicillin and streptomycin (Thermo Fischer Scientific, Waltham, MA, USA), 20 ng/mL EGF and 10 ng/mL bFGF (PeproTech, East Windsor, NJ, USA). Formed neurosphere numbers and cell numbers were counted after 7 days in culture. For the in vitro differentiation, neurospheres were trypsinised and cells were plated on poly-L-lysine (10 µg/mL overnight at room temperature, P5899, Sigma-Aldrich) and then laminin (10 µg/mL, 30 min at 37 ◦C, L2020, Sigma-Aldrich)-coated flat bottom 96-well plates (CellCarrier 96 Ultra, Cat# 6005550). After 2 days, the differentiation was initiated by changing the neural stem cell medium by differentiation medium: DMEM/F12 (Gibco) supplemented with 1% FSC (Thermo Fischer Scientific), 2% B27 (Invitrogen), 1X penicillin and streptomycin (Thermo Fischer Scientific). Cells were fixed, permeabilized and applied for immunofluorescent staining as described previously [34]. β-Tubulin III (TUJ1) antibody (1:200, Covance, MMS-435P, Princeton, NJ, USA) was used to detect neurons, and GFAP antibody (1:400, Dako, Z0334, Frankfurt am Main, Germany) was used to detect astrocytes.

#### *2.5. Histology, TUNEL Reaction and Immunofluorescent Staining*

Brains from E14.5 and E18.5 embryos were fixed overnight with 4% paraformaldehyde (PFA) (pH 7.2) and cryopreserved in 30% sucrose overnight. Neg-50 (Thermo Fischer Scientific) frozen section medium was used to embed the brains followed by cryosectioning (Microm™ HM 550 Cryostat, Thermo Fischer Scientific, Waltham, MA, USA) of 12 µm thick slices. After antigen retrieval in citrate buffer for 40 min at 95 ◦C, immunostaining with the following primary antibodies was performed: rabbit anti-SOX2 (1:200, Abcam, Ab97959); rabbit anti-TBR2/Eomes (1:200, Abcam, Ab23345); rabbit anti-TBR1 (1:200, Abcam, Ab31940); rat anti-CTIP2 (1:200, Abcam, Ab18465); rabbit anti-CUX1/CDP (1:100, Santa Cruz, Sc-13024, Heidelberg, Germany); rabbit anti-A-calbindin D-28k (1:1000, Swant, CB38, Burgdorf, Switzerland); anti-phospho-histone H3 (Ser10) (1:400, Cell Signaling, 9071, Danvers, MA, USA); and anti-Cleaved Caspase-3 (1:200, Cell Signaling, 9661S).

Immunoreactivity was visualized using secondary antibodies anti-rabbit IgG conjugated with Cy3 (1:200, Sigma-Aldrich), Cy2 or Cy5 (1:200, Jackson ImmunoResearch), donkey antimouse IgG conjugated with Cy3 (1:200; Sigma-Aldrich), goat anti-rabbit Biotin conjugated (1:400, Vector Laboratories, Burlingame, CA, USA) and donkey anti-rat IgG conjugated with Alexa-488 (1:200, Sigma-Aldrich), streptavidin-Cy3 (1:800, Sigma-Aldrich). Overall cell death was detected using TUNEL reaction, as described previously [35]. In all cases, the nuclei were counterstained using DAPI (1:10,000, Sigma-Aldrich) and mounted in ProLong Gold Antifade reagent (P36930, Invitrogen).

#### *2.6. Microscopy and Image Analysis*

The images of whole brain sections were acquired using BX61VS Olympus Virtual microscope and processed with Olympus Olyvia 2.9 computer program. The immunofluorescence images of the in vitro differentiation of neural stem cells were acquired using the microscope ImageXpress Micro Confocal (short IXMC) from Molecular Device (MD). In each well, a z-stack of seven images with 1 µm distance was recorded using 10× Plan Apo objective with confocal mode (50 µm slit disc) at four sites. For subsequent image analysis, a custom module in the MetaXpress software from MD (version 6.2.3.) was created and applied on maximum projections of each z-stack. Each cell was defined by its nucleus using a mask derived from the DAPI channel. Depending on the intensity within this mask in the other channels, each nucleus was classified as belonging to a neuron, astrocyte or unclassified cell. This led to a summarized number of each cell type in each well.

#### *2.7. RNA-Seq and Bioinformatic Analysis*

NPCs were isolated from E13.5 embryo brains after crossing *Smg6*flox and Rosa26- CreERT2 mice, and kept in neurosphere cultures for 2 days followed by 4-Hydroxytamoxifen (4-OHT, Sigma) treatment to induce the *Smg6* deletion. Total RNA was isolated 6 days after 4-OHT treatment from cells with genotypes *Smg6*flox/flox; Rosa26-CreERT2 (without 4-OHT) (ctr), *Smg6*flox/flox treated with 4-OHT (ctr + 4-OHT) and *Smg6*flox/flox; Rosa26-CreERT2 treated with 4-OHT (*Smg6*-iKO)) using the RNeasy Mini Kit (Qiagen) and following the manufacturer's manual. The RNA integrity was checked using an Agilent Bioanalyzer 2100 (Agilent Technologies). All samples showed a RIN (RNA integrity number) higher than 9. Approximately 800 ng of total RNA was used for library preparation using a TruSeq Stranded Total RNA (RiboZero Gold) according to the manufacturer's protocol. The libraries were pooled into one and sequenced in three lanes using HiSeq2500 (Illumina) in single-read high-output mode, which created reads with a length of 50 bp. Sequencing reads were extracted using bcl2FastQ v1.8.4. On average, 73 million reads per sample were obtained. For expression analysis, the raw reads were mapped with STAR (version 2.5.4b, parameters: –alignIntronMax 100000–outSJfilterReads –outSAMmultNmax Unique –outFilterMismatchNoverLmax 0.04) [36] to the *Mus musculus* genome (GRCm38) with the Ensembl genome annotation (Release 92). For each Ensembl gene, reads that mapped uniquely to one genomic position were counted with FeatureCounts (version 1.5.0, multi-mapping or multi-overlapping reads were not counted, stranded mode was set to "–s 2", Ensembl release 92 gene annotation) [37]. The table of raw counts per gene per sample was analysed with R (version 3.5.0) using the package DESeq2 (version 1.20.0) [38]. The sample group *Smg6*-iKO (*n* = 4) was contrasted with the sample group Ctr + 4-OHT (*n* = 4), with *Smg6*-iKO being the reference level. The sample group Ctr (*n* = 3) was also contrasted with the sample group *Smg6*-iKO (*n* = 4) and the Ctr + 4-OHT (*n* = 4). For each gene of the comparison, the *p*-value was calculated using the Wald significance test. Resulting *p*-values were adjusted for multiple testing with Benjamini & Hochberg correction. Genes with an adjusted *p*-value < 0.05 were considered differentially expressed (DEGs). The log2 fold changes (LFCs) were shrunk with lfcShrink to control for variance of LFC estimates for genes with low read counts. The changes in molecular pathways as well as their possible upstream regulators were identified by analyzing abovementioned three pairwise comparisons using Ingenuity Pathway Analysis (IPA) program (Qiagen).

#### *2.8. Statistical Analysis*

Depending on the distribution of the data points, unpaired two-tailed Student's *t*-test or Mann–Whitney U (MWU) test were used to calculate significance. Data sets underwent Shapiro–Wilk test for the normal distribution. If the data set passed the Shapiro–Wilk test (*p* value > 0.05), Student's *t*-test was used, if it did not pass (*p* value < 0.05), Mann– Whitney U test was applied. Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA). The type of the test performed is indicated in each figure legend. Indication for significance was used as follows: n.s. > 0.05, \* < 0.05, \*\* < 0.01, \*\*\* < 0.001, \*\*\*\* < 0.0001.

#### **3. Results**

#### *3.1. Smg6 Deletion in CNS Compromises Embryonic Neurogenesis and Newborn Viability*

In order to understand the biological function of SMG6 specifically in the differentiation program of committed lineage NPCs in the central nervous system (CNS), we generated a conditional knockout mouse model in which *Smg6* was deleted in NPCs from embryonic day E10.5 by intercrossing *Smg6F/*<sup>∆</sup> [13] and *Nestin-Cre* transgenic mice [27] to generate CNS-deleted mice (*Smg6*-CNS∆). We obtained an expected number of *Smg6*- CNS∆ embryos at E14.5 and a slightly reduced number at E18.5 (Figure S1A) according to the Mendelian ratio. However, all *Smg6*-CNS∆ newborns died within 1–2 days after birth (Figure S1B).

To investigate the role of SMG6 in neurogenesis, we first analysed brains at E14.5 and found that an efficient *Smg6* deletion in the CNS (Figure S1C) yielded normal embryo body and brain weight as well as cortex (CTX) thickness (Figure S1D–G). Furthermore, cortical cellularity and populations of SOX2+ neural progenitor cells and TBR2+ intermediate progenitors (IPs), as well as populations of early newborn neurons positive for TBR1 and CTIP2, were the same as in controls at this stage of development (Figure 1A–C). Next, we examined the brains just before birth at E18.5 and found that *Smg6*-CNS∆ fetuses had normal body and brain weights (Figure S1H–J). However, we detected a significantly smaller CTX in E18.5 *Smg6*-CNS∆ brains (Figure 1D), indicating mild microcephaly and defects in embryonic neurogenesis. Concomitantly, the *Smg6*-CNS∆ cortices presented a significant reduction of cortical cellularity at this stage (Figure 1E), likely responsible for the reduction of the CTX.

Mouse CTX contain well defined cellular layers composed of neural precursors in the ventricular and subventricular proliferative zones (VZ and SVZ, respectively), early born neurons in the middle part (layers VI and V) and late born neurons in the upper part of the cortical plate (layers II/III). Next, we histologically analysed the structure of the E18.5 *Smg6*-CNS∆ brains. It revealed that all the neuronal layers in *Smg6*-CNS∆ cortices were formed, but with a significant decrease of SOX2+ NPCs in the VZ and TBR2+ IPs in the SVZ, indicating an exhaustion of NPC pools during the late embryonic brain development (Figure 1F,G). The numbers of early born neurons in layers VI and V positive for TBR1 and CTIP2 were normal, in contrast to a significant reduction of the late born neurons in the layers II/III judged by the CUX1+ population (Figure 1F,G). These findings indicate that SMG6 is dispensable for early cortical neurogenesis, but its deletion prematurely depleted NPC pools and compromised cortical neurogenic production, affecting the cellularity of the CTX during later development.

**Figure 1.** SMG6 deficiency in the central nervous system causes perinatal lethality**.** (**A**) Representative immunofluorescent staining of E14.5 cortices, coronal sections. SOX2 and TBR2 stained the progenitors of the VZ and SVZ, respectively. TBR1 and CTIP2 stained the early born neurons of layers VI and V, respectively. DAPI was used to stain the cell nucleus. VZ: ventricular zone, SVZ: subventricular zone. (**B**) Quantification of the neuroprogenitor cells (SOX2 and TBR2) and neurons (TBR1 and CTIP2) in E14.5 cortices in A, Student's *t*-test was applied for all markers. (**C**) Quantification of the total cellularity in the E14.5 cortices, Student's *t*-test. (**D**) Smaller CTX area of the E18.5 *Smg6*-CNS∆ brains, Student's *t*-test. (**E**) Quantification of the total cellularity in the E18.5 cortices, Student's *t*-test. (**F**) Immunofluorescent staining of E18.5 coronal sections showing only cortical region. SOX2 and TBR2 labels NPCs and intermediate progenitor cells in the VZ and SVZ, respectively. TBR1 and CTIP2 stained the early born neurons in layers VI and V. CUX1 stained the later born neurons in layers IV, III and II. DAPI was used to counterstain the cell nucleus. VZ: ventricular zone, SVZ: subventricular zone. (**G**) Quantification of the NPC cells (SOX2 and TBR2) and neurons (TBR1, CTIP2 and CUX1) in E18.5 cortices shown in D **Figure 1.** SMG6 deficiency in the central nervous system causes perinatal lethality. (**A**) Representative immunofluorescent staining of E14.5 cortices, coronal sections. SOX2 and TBR2 stained the progenitors of the VZ and SVZ, respectively. TBR1 and CTIP2 stained the early born neurons of layers VI and V, respectively. DAPI was used to stain the cell nucleus. VZ: ventricular zone, SVZ: subventricular zone. (**B**) Quantification of the neuroprogenitor cells (SOX2 and TBR2) and neurons (TBR1 and CTIP2) in E14.5 cortices in A, Student's *t*-test was applied for all markers. (**C**) Quantification of the total cellularity in the E14.5 cortices, Student's *t*-test. (**D**) Smaller CTX area of the E18.5 *Smg6*-CNS∆ brains, Student's *t*-test. (**E**) Quantification of the total cellularity in the E18.5 cortices, Student's *t*-test. (**F**) Immunofluorescent staining of E18.5 coronal sections showing only cortical region. SOX2 and TBR2 labels NPCs and intermediate progenitor cells in the VZ and SVZ, respectively. TBR1 and CTIP2 stained the early born neurons in layers VI and V. CUX1 stained the later born neurons in layers IV, III and II. DAPI was used to counterstain the cell nucleus. VZ: ventricular zone, SVZ: subventricular zone. (**G**) Quantification of the NPC cells (SOX2 and TBR2) and neurons (TBR1, CTIP2 and CUX1) in E18.5 cortices shown in D (MWU test for SOX2 and TBR2, Student's *t*-test for the rest). For all graphs: *n*—number of embryos analysed. Error bars represent SEM, statistic comparison as indicated in each graph description—n.s. > 0.05, \* *p* < 0.05, \*\* *p*< 0.01.

#### *3.2. SMG6 Is Essential for Survival of Neural Cells*

In order to investigate the cause for the reduction of NPCs and late born neurons of E18.5 *Smg6*-CNS∆ brains, we analysed cell death by TUNEL and Active-Caspase 3 (Act-Cas3) staining. At both E14.5 and E18.5 stages *Smg6*-CNS∆ brains exhibited a significant increase of TUNEL and Act-Cas3 signals in the areas of CTX and GE (Figures 2A–H and S2A–H). The cell death at E14.5 was mainly found in the proliferative VZ and SVZ areas, where double TBR2 + TUNEL+ staining confirmed the death of IPs (Figure 2A,B). Intriguingly, TUNEL staining in GE (Figure 2C,D) detected clearly higher (Figure 2D versus Figure 2Bi) cell death in GE than in CTX at E14.5 brain sections. At E18.5, TUNEL positive cells in CTX were additionally detected in the intermediate zone (IZ) as well as in the cortical plate (CP) (Figure 2E,F), suggesting neuronal death at a late stage of brain development. However, it stayed at comparable levels to E14.5 (Figure 2Bi versus Figure 2Fi) whereas in GE (Figure 2G,H) we detected less TUNEL signals compared to younger embryos (Figure 2H versus Figure 2D). Immunofluorescence staining for Act-Cas3 confirmed elevated Caspase 3-dependent cell death in the CTX as well as in the GE (Figure S2A–H). Counting of phospho- Histone H3 (Ser10)+ cells did not reveal any difference for the number of mitotic cells in the CTX and GE between mutant and control animals at E14.5 (Figure S3A,B). Interestingly, SMG6 seems to be more crucial for the survival of NPCs of GABAergic interneurons (IN) that are generated in the GE. To confirm this, we stained E14.5 cortices with calbindin, an interneuron marker [39–41], and detected significantly less calbindin+ IN in *Smg6*-CNS∆ GE (Figure 2I,J) compared to controls. These observations indicate that at early neurogenesis, SMG6 absence affects the survival of NPCs prominently in the GE, and to a lesser extend in the CTX.

#### *3.3. Smg6 Deletion Compromises the Self-Renewal and Differentiation Capacity of Neuroprogenitors*

To investigate the renewal capacity of SMG6-deficient NPCs, we performed the in vitro neurosphere formation assay using neural stem cells isolated from the CTX or the GE at E13.5 (Figure S4A). The control and *Smg6*-CNS∆ cortical NPCs formed the same number of neurospheres after 7 days (Figure 3A,Ai), containing a comparable number of cells (Figure 3Aii), indicating a dispensable role of SMG6 in the renewal capacity of cortical NPCs. In contrast, mutant GE NPCs gave rise to a similar number of neurospheres (Figure 3B,Bi), but these neurospheres contained much fewer cells (Figure 3Bii). These results indicate that SMG6 is specifically critical for the renewal capacity of the GE NPCs, but less so for those from the CTX. Concomitantly, the mRNA expression levels of the transcription factors *Dlx1*, *Dlx2* and *Mash1,* known to drive interneuron differentiation [42–47], were dramatically reduced in the neurospheres originating from SMG6-deficient GE (Figure 3C). These findings indicate a specific role of SMG6 in interneuron progenitors.

Because SMG6 is essential for ES cell differentiation [13], we next studied the differentiation potential of *Smg6*-deleted neuroprogenitors. To this end, we used neurospheres derived originally from NPCs of the CTX or GE (Figure S4A) and induced their differentiation in vitro. We found significantly less neurons, judged by TUJ1 staining, at 6 and 8 days post differentiation (dpd) of NPCs originated from both the CTX and GE (Figure 3D,G,H,K). The percentage of GFAP-positive cells, representing astrocytes, was modestly but significantly increased in *Smg6* mutant cultures (Figure 3F,J). In addition, we observed overall reduced cellularity at 8 dpd compared to controls. It is of note that the reduction of cell numbers at 8 dpd was more prominent in differentiation cultures from the GE (76%) than in cultures from the CTX (29%) (Figure 3E,I). In summary, SMG6 plays a role in the differentiation program in vitro and is required for the survival of differentiating cells particularly of GE derived NPCs.

*Cells* **2021**, *10*, x FOR PEER REVIEW 8 of 21

**Figure 2.** SMG6 deficiency causing death of neuronal cells. (**A**) Co-staining of dying cells using TUNEL and intermediate progenitors by labelling TBR2 on coronal sections of E14.5 embryo brains counterstained by DAPI for cell nucleus. Yellow head arrow marks TUNEL-positive cells, white arrow—TUNEL- and TBR2-positive cells. Quantification of the total TUNEL-positive cells (**Bi**) their distribution in different cortical layers (**Bii**) and dying TBR2-positive (TUNEL<sup>+</sup>TBR2<sup>+</sup>) cells (**Biii**) at E14.5 indicate neural cell death mainly in area of progenitors. Statistics MWU for IZ and CP in Bii and Biii, Welch **Figure 2.** SMG6 deficiency causing death of neuronal cells. (**A**) Co-staining of dying cells using TUNEL and intermediate progenitors by labelling TBR2 on coronal sections of E14.5 embryo brains counterstained by DAPI for cell nucleus. Yellow head arrow marks TUNEL-positive cells, white arrow—TUNEL- and TBR2-positive cells. Quantification of the total TUNEL-positive cells (**Bi**) their distribution in different cortical layers (**Bii**) and dying TBR2-positive (TUNEL+TBR2<sup>+</sup> ) cells (**Biii**) at E14.5 indicate neural cell death mainly in area of progenitors. Statistics MWU for IZ and CP in Bii and Biii, Welch *t*-test in Bi and Student's *t*-test for the rest. (**C**) TUNEL staining of GE at E14.5 with quantification respectively in (**D**) using Student's *t*-test. (**E**) Co-staining of dying cells using TUNEL and intermediate progenitors by TBR2 on coronal sections of E18.5 embryo brains counterstained by DAPI for cell nucleus. Yellow head arrow marks TUNEL-positive cells, white arrow—TUNEL- and TBR2-positive cells. Quantification performed in the same manner at E14.5 (**Fi**–**Fiii**) shows increased cell death in all cortical layers. Statistics by Student's *t*-test. (**G**) TUNEL staining of GE at E18.5 quantified in (**H**) using Welch *t*-test. (**I**) Immuno-fluorescent staining and quantification (**J**) of calbindin-positive cells in coronal sections of E14.5 embryo brains. DAPI counterstains the cell nucleus. Statistics by Student's *t*-test. For all graphs: VZ—ventricular zone; SVZ—subventricular zone; IZ—Intermediate zone; CP—cortical plate; *n*—number of embryos analysed. Error bars represent SEM. Significance—n.s. > 0.05, \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p*< 0.0001.

**Figure 3.** Neuroprogenitor renewal and differentiation are impaired without SMG6, especially in the NPCs from GE. (**A**) In vitro neurosphere assay on NPCs from CTX shows no significant difference in the number of neurospheres (**Ai**) as well as in the cell number per neurosphere (**Aii**) after 7 days in culture. The numbers of neurospheres formed from GE cells (**B**) also did not differ (**Bi**), but had fewer cells per neurosphere (**Bii**). (**C**) qPCR analysis on GE neurospheres shows relative expression changes of indicated gene mRNAs after *Smg6* deletion. (**D**–**G**) Differentiation capacity of progenitors from CTX and (**H**–**K**) GE at 6 and 8 days after differentiation induction (dpd). (**D**,**H**) Co-stainings of in vitro differentiated cultures **Figure 3.** Neuroprogenitor renewal and differentiation are impaired without SMG6, especially in the NPCs from GE. (**A**) In vitro neurosphere assay on NPCs from CTX shows no significant difference in the number of neurospheres (**Ai**) as well as in the cell number per neurosphere (**Aii**) after 7 days in culture. The numbers of neurospheres formed from GE cells (**B**) also did not differ (**Bi**), but had fewer cells per neurosphere (**Bii**). (**C**) qPCR analysis on GE neurospheres shows relative expression changes of indicated gene mRNAs after *Smg6* deletion. (**D**–**G**) Differentiation capacity of progenitors from CTX and (**H**–**K**) GE at 6 and 8 days after differentiation induction (dpd). (**D**,**H**) Co-stainings of in vitro differentiated cultures from CTX and GE, respectively, at the indicated time points with quantifications of total cellularity in (**E**,**I**), GFAP-positive cells (**E**,**J**) and neurons labelled with TUJ1 staining (**G**,**K**). For all graphs: *n*—number of embryos used for cell isolations. Error bars represent SEM. Statistics by unpaired Student's *t*-test in (**A**–**C**) and Welch *t*-test in (**E**–**G**,**I**–**K**) significance n.s. > 0.05, \* *p*< 0.05, \*\* *p*< 0.01, \*\*\* *p*< 0.001.

#### *3.4. SMG6 Deficiency Exclusively in the Cortex Does Not Inhibit Corticogenesis*

Nestin-Cre drives deletion of the *Smg6* gene in the whole brain; thus, CTX and GE both are affected in *Smg6*-CNS∆ mice. To further dissect the SMG6 function in different populations of NPCs, we deleted *Smg6* only in the cortical/hippocampal progenitor cells at day E9.5 by crossing *Smg6*-floxed mouse with the Emx1-Cre mouse model (designated as *Smg6*-CoHi∆). Intriguingly, in contrast to *Smg6*-CNS∆ mutants, the *Smg6*-CoHi∆ mice were born at expected ratios (Figure S4B) and were viable during the observation period of 20 months. We confirmed specific deletion of the SMG6 protein in the CTX, but not in other parts of the brain such as the GE and the hind/mid brain (Figure S4F). Notably, the SMG6 levels seem to be reduced less efficiently than in *Smg6*-CNS∆ embryos because the CTX of *Smg6*-CoHi∆ embryos contain a high number of cells, i.e., IN that are not affected by the deletion and thus have normal levels of SMG6. The body and brain weight of *Smg6*-CoHi∆ embryos as well as the CTX were normal at E18.5 (Figures S4C–E and 4A,B). Moreover, the cellularity and thickness of *Smg6*-CoHi∆ cortices were the same as controls (Figure 4C–E). Microscopic analysis of immunofluorescent staining revealed a similar number of TBR1+ and CTIP2+ neurons as well as SOX2+ and TBR2+ neuroprogenitor populations between mutant and control littermates (Figure 4D,F). In contrast to *Smg6*-CNS∆, TUNEL assay did not detect obvious cell death in the CTX (Figure 5A,B) nor in the GE (Figure 5C,D). However, we detected a significant increase of cellular death only in the retrosplenial cortex of E18.5 *Smg6*-CoHi∆ embryos (Figure 5E,F), indicating that SMG6 is vital specifically and only for this small population of cortical cells. *Cells* **2021**, *10*, x FOR PEER REVIEW 12 of 21

**Figure 4.** SMG6 deficiency solely in cortical and hippocampal neuroprogenitors does not inhibit corticogenesis**.** (**A**) Comparison of the E18.5 brains and quantification of the CTX area (**B**). (**C**) Quantification of the CTX thickness. (**D**) Immunofluorescent staining of E18.5 coronal sections showing only cortical regions. SOX2 and TBR2 labels NPCs and intermediate progenitors in the VZ and SVZ, respectively. TBR1 and CTIP2 stains early born neurons in layers VI and V. DAPI counterstains the cell nucleus. VZ: ventricular zone, SVZ: subventricular zone. (**E**) Quantification of the total cellularity in the E18.5 cortices. (**F**) Quantification of the neuroprogenitor cells (SOX2 and TBR2) and neurons (TBR1 and CTIP2) in E18.5 cortices shown in D. For all graphs: *n*—number of embryos analysed. Error bars represent SEM. Statistics by unpaired Student's *t*-test, except in (**B**) the MWU was used, significance—n.s. > 0.05. **Figure 4.** SMG6 deficiency solely in cortical and hippocampal neuroprogenitors does not inhibit corticogenesis. (**A**) Comparison of the E18.5 brains and quantification of the CTX area (**B**). (**C**) Quantification of the CTX thickness. (**D**) Immunofluorescent staining of E18.5 coronal sections showing only cortical regions. SOX2 and TBR2 labels NPCs and intermediate progenitors in the VZ and SVZ, respectively. TBR1 and CTIP2 stains early born neurons in layers VI and V. DAPI counterstains the cell nucleus. VZ: ventricular zone, SVZ: subventricular zone. (**E**) Quantification of the total cellularity in the E18.5 cortices. (**F**) Quantification of the neuroprogenitor cells (SOX2 and TBR2) and neurons (TBR1 and CTIP2) in E18.5 cortices shown in D. For all graphs: *n*—number of embryos analysed. Error bars represent SEM. Statistics by unpaired Student's *t*-test, except in (**B**) the MWU was used, significance—n.s. > 0.05.

n.s. > 0.05, \* *p*< 0.05.

**Figure 5.** SMG6 deficiency in cortical and hippocampal neuroprogenitors cause neuronal cell death in the retrosplenial cortex. TUNEL staining comparison of E18.5 brain sections in CTX (**A**), GE (**C**) and retrosplenial cortex (RtCtx) (**D**). (**B**,**E**,**F**) Quantifications of total TUNEL positive cells in the respective brain parts. For all graphs: *n*—number of embryos analysed. Error bars represent SEM. Statistics by unpaired Student's *t*-test, except in (**F**) where the MWU was used, significance— **Figure 5.** SMG6 deficiency in cortical and hippocampal neuroprogenitors cause neuronal cell death in the retrosplenial cortex. TUNEL staining comparison of E18.5 brain sections in CTX (**A**), GE (**C**) and retrosplenial cortex (RtCtx) (**D**). (**B**,**E**,**F**) Quantifications of total TUNEL positive cells in the respective brain parts. For all graphs: *n*—number of embryos analysed. Error bars represent SEM. Statistics by unpaired Student's *t*-test, except in (**F**) where the MWU was used, significance—n.s. > 0.05, \* *p*< 0.05.

The finding that *Smg6* deletion has a stronger effect on IN progenitors prompted us

to investigate the transcriptional programs initiated by the loss of SMG6. We analysed the total transcriptome profile of SMG6-deficient NPCs isolated from E13.5 *Smg6*-CER (Smg6flox crossed with Rosa26-CreERT2) mice brains and cultured for 6 days in the presence of 4-OHT that induces *Smg6* deletion (*Smg6*-iKO). RNA-seq data comparison of controls (Ctr + 4-OHT) with *Smg6*-iKO cells revealed 859 differentially expressed genes (DEGs)

*Dysregulation*

Taken together, using two different neural specific Cre mouse models, we demonstrate that GE NPCs and interneurons are particularly vulnerable to SMG6 loss. We show that cortical NPCs and neurons are affected mainly if *Smg6* is deleted simultaneously in NPCs from both brain parts GE and CTX. We conclude that the defects of cortical neurogenesis in the *Smg6*-CNS∆ model are likely sensitized by *Smg6* deletion in the IN progenitors derived from the GE.

#### *3.5. SMG6 Null Mutation Activates DNA Repair and p53 Pathways Causing Cell Cycle Dysregulation*

The finding that *Smg6* deletion has a stronger effect on IN progenitors prompted us to investigate the transcriptional programs initiated by the loss of SMG6. We analysed the total transcriptome profile of SMG6-deficient NPCs isolated from E13.5 *Smg6*-CER (Smg6flox crossed with Rosa26-CreERT2) mice brains and cultured for 6 days in the presence of 4-OHT that induces *Smg6* deletion (*Smg6*-iKO). RNA-seq data comparison of controls (Ctr + 4-OHT) with *Smg6*-iKO cells revealed 859 differentially expressed genes (DEGs) (cutoff adjusted *p* < 0.05), containing 385 upregulated and 474 downregulated DEGs (Figure 6A, Supplementary Table S1). Confirming its function in NMD, *Smg6* knockout resulted in the presence of the prominent NMD target genes within the upregulated DEGs (underlined in Figure 6A). The analysis using IPA (Ingenuity Pathway Analysis) showed that the majority of the DEGs are involved in DNA repair and cell cycle pathways (Figure 6B). Furthermore, we used the IPA upstream regulator tool to find which possible regulator was upstream, and whether the activation or silencing of it could explain the observed gene expression changes. The upstream analysis of all DEGs predicted the activation of *Trp53*, *Cdkn2a* and *Cdkn1a* genes (positive Z-score) that might have caused the detected changes (Figure 6C), indicating activation of DNA repair pathways that would lead to cell cycle arrest and eventually cell death. Interestingly, we found that *FoxM1* is among the top 30 strongest upstream regulators of DEGs, although with a negative Z-score, indicating that an inhibition of the FOXM1 function could also be a reason for the detected expression changes. FOXM1 is known to promote *Dlx1* gene expression [48] and its silencing could cause the down regulation of *Dlx1*, which we observed in neurospheres after *Smg6* deletion (Figure 3C).

**Figure 6**. RNAseq analysis of neuroprogenitors after 4-OHT induced deletion of *Smg6*. (**A**) Total amount of identified differentially expressed genes (DEGs) with the 25 strongest up- and down-regulated genes. (**B**) Top 30 altered molecular pathways identified using IPA software. Control comparison of control cells treated and not treated with 4-OHT is shown in the left column. (**C**) Top 30 predicted upstream regulators of the detected changes in the transcriptome. Control comparison of control cells treated and not treated with 4-OHT is shown in the left column. For Ctr condition cells from 3 **Figure 6.** RNAseq analysis of neuroprogenitors after 4-OHT induced deletion of *Smg6*. (**A**) Total amount of identified differentially expressed genes (DEGs) with the 25 strongest up- and down-regulated genes. (**B**) Top 30 altered molecular pathways identified using IPA software. Control comparison of control cells treated and not treated with 4-OHT is shown in the left column. (**C**) Top 30 predicted upstream regulators of the detected changes in the transcriptome. Control comparison of control cells treated and not treated with 4-OHT is shown in the left column. For Ctr condition cells from 3 embryos and for each of Ctr + 4OHT and *Smg6*-iKO conditions cells from four embryos were used. DEGs for analysis by IPA were defined by adjusted *p*-value ≤ 0.05 and log2FoldChange ≤ −0.2 and ≥0.2.

#### **4. Discussion**

A complete deletion of key NMD genes such as *Smg1*, *Upf1* or *Upf2* results in early embryonic lethality, demonstrating an essential function of RNA metabolism/quality control during development [10–12]. We previously showed that *Smg6* deletion in ES cells had no impact on their viability but blocked their ability to differentiate into germ layers during mouse development. We concluded that SMG6-mediated NMD is a license factor for the cell fate determination of pluripotent stem cells [13]. Interestingly, in the present study, we show that if *Smg6* is deleted in committed neural stem cells during development at E10.5 (Nestin-Cre) or at E9.5 (Emx1-Cre), these NPCs are viable, can differentiate into various cell types and thus are able to support development of the entire brain. However, later embryonic neurogenesis and thus production of late born neurons is compromised in the Nestin-Cre mediated *Smg6* deletion model. This is in immense difference to the pluripotent ES cells, where the deletion of *Smg6* blocked differentiation through overexpression of the pluripotency gene *c-Myc* [13]. In contrast to ES cells, the overexpression of *c-Myc* during neurogenesis is known rather to support not only the self-renewal but also the neuronal differentiation of the NPCs [49,50].

Nestin-Cre mediated deletion of *Smg6* resulted in defects in both CTX and GE. In the CTX we detected a reduction of the NPC pool in VZ and SVZ and less production of late born neurons at E18.5 (Figure 1F,G), indicating that defects appear rather late in corticogenesis. Interestingly, we found that *Smg6* inactivation affects GE-derived NPCs earlier, i.e., at E14.5, judged by higher cellular death in histology and a decreased selfrenewal of IN progenitors in the neurosphere assay compared to results obtained from CTX. Furthermore, we detected a reduced generation of calbindin-positive GABAergic neurons at E14.5, indicating dysregulation of the interneuron production. The striking finding is that Emx1-Cre mediated *Smg6* inactivation exclusively in cortical and hippocampal progenitors (*Smg6*-CoHi∆ mice), which spared IN progenitors, caused no comparable cortical defects. However, Emx1-Cre was reported to be more efficient and mediating stronger cortical defects than Nestin-Cre in other models [34,51,52]. *Smg6*-CoHi∆ mice were viable during the observation period of 20 months. This cell type specificity of SMG6 is supported by in vitro neurosphere assays, where we found that SMG6 null compromises only progenitors from GE but not from CTX. Thus, we conclude that SMG6 plays a less important role in the cortical progenitor renewal and differentiation, but is specifically required for IN neuroprogenitors and their derived cell lineages during brain development. Given these observations and the fact that GE-derived INs migrate tangentially to the CTX and regulate cortical neurogenesis via the GABA release [53,54], it is plausible that *Smg6* deletion caused malfunction of INs, which affected cortical neurogenesis (Figure 7).

The role of SMG6 on interneuron progenitors is reminiscent of a recent study showing that exon-junction complex factor RBM8A is critical for interneuron development [55]. Consistent with the defects of SMG6 deficient interneurons, key transcription factors that control the cell fate of interneuron progenitors were found dysregulated. *Dlx1* and *Dlx2* expression was dramatically reduced in SMG6-deficient neurospheres. DLX1 and DLX2 transcription factors promote or repress the expression of other transcription factors that are responsible for triggering IN differentiation, migration and maturation [42–47]. The downregulation of *Dlx1* gene correlates well with the IPA upstream regulator analysis of the RNA-seq data (Figure 6C), which predicted the inactivation of FOXM1, a known positive regulator of *Dlx1* [48]. Remarkably, along the same line to *Smg6*-CNS∆ mice, perinatal lethality has also been reported in mice with *Dlx1*/*2* double knock out (KO) [42,45,56] due to impaired differentiation and migration of GABAergic INs in the neocortex [45]. Disruptions of other genes such as *Gad1*, *Gad2*, *Nkx2.1* and *Sox6,* known to regulate GABAergic neuron development, were also often linked to perinatal lethality [57–62]. Unfortunately, the precise reason for the perinatal lethality of the above mentioned and *Smg6*-CNS∆ mice is unclear.

environmental changes caused by defects in interneurons.

CNS∆ mice is unclear.

**Figure 7.** The role of SMG6 in embryonic neurogenesis**.** *Smg6* deletion in cortical and hippocampal NPCs (*Smg6*-CoHi∆) does not inhibit corticogenesis. However, if SMG6 is absent in the whole nervous system (*Smg6*-CNS∆) it causes cell death and also renewal defects of GE neural stem cells leading to defective interneurons tangentially migrating to the CTX. Consequently, the self-renewal of cortical NPCs and the production of late born neurons are impaired possibly because of **Figure 7.** The role of SMG6 in embryonic neurogenesis. *Smg6* deletion in cortical and hippocampal NPCs (*Smg6*-CoHi∆) does not inhibit corticogenesis. However, if SMG6 is absent in the whole nervous system (*Smg6*-CNS∆) it causes cell death and also renewal defects of GE neural stem cells leading to defective interneurons tangentially migrating to the CTX. Consequently, the self-renewal of cortical NPCs and the production of late born neurons are impaired possibly because of environmental changes caused by defects in interneurons.

> The transcriptional changes in the SMG6-deficient NPCs confirmed the defective NMD by changes in known NMD target transcripts. The majority of the top 30 dysregulated pathways belong to DNA repair, cell cycle and p53 related pathways, which can be The transcriptional changes in the SMG6-deficient NPCs confirmed the defective NMD by changes in known NMD target transcripts. The majority of the top 30 dysregulated pathways belong to DNA repair, cell cycle and p53 related pathways, which can be modulated also by the NMD [63–66].

> BAergic neuron development, were also often linked to perinatal lethality [57–62]. Unfortunately, the precise reason for the perinatal lethality of the above mentioned and *Smg6*-

> modulated also by the NMD [63–66]. Since the transcriptome and thus the expression of NMD targets highly depends on the cell type, it is plausible that deletion of *Smg6* may have cell type-dependent consequences in cell fate. In this regard, the conditional knock-out of *Upf2* in the adult hematopoietic system preferentially compromised the viability of hematopoietic stem cells and progenitors, but not that of terminally differentiated T cells [11]. In addition, the *Upf2* null mutation did not affect the proliferation of fetal hepatocytes, but compromised their maturation process [67]. Furthermore, UPF3B was demonstrated to be very important for a Since the transcriptome and thus the expression of NMD targets highly depends on the cell type, it is plausible that deletion of *Smg6* may have cell type-dependent consequences in cell fate. In this regard, the conditional knock-out of *Upf2* in the adult hematopoietic system preferentially compromised the viability of hematopoietic stem cells and progenitors, but not that of terminally differentiated T cells [11]. In addition, the *Upf2* null mutation did not affect the proliferation of fetal hepatocytes, but compromised their maturation process [67]. Furthermore, UPF3B was demonstrated to be very important for a subset of olfactory sensory neurons [68]. Taking these studies together, SMG6, or in general NMD, regulates cell fate programs highly depending on the cell type and developmental stage.

> subset of olfactory sensory neurons [68]. Taking these studies together, SMG6, or in general NMD, regulates cell fate programs highly depending on the cell type and developmental stage. In conclusion, although SMG6 is essential for the differentiation of pluripotent ES cells, it is less important for the differentiation of committed cortical NPCs. Using various In conclusion, although SMG6 is essential for the differentiation of pluripotent ES cells, it is less important for the differentiation of committed cortical NPCs. Using various mouse models, we demonstrate that SMG6, as a general endonuclease of NMD for aberrant RNA, plays distinct roles in different cell types: CTX versus GE neuroprogenitor cells (Figure 7) versus ES cells. These genetic results allow predicting that the importance of NMD varies dramatically, depending on the transcriptional program of a specific cell type.

> mouse models, we demonstrate that SMG6, as a general endonuclease of NMD for aberrant RNA, plays distinct roles in different cell types: CTX versus GE neuroprogenitor cells

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/cells10123365/s1, Supplementary Figure S1: Smg6 deletion in all neuroprogenitors; Figure S2: Smg6 deletion in the central nervous system cause Caspase3 dependent cell death in the cortex and ganglionic eminence; Figure S3: Proliferation of neuroprogenitors in the cortex and ganglionic eminence of E14.5 brains of indicated genotypes; Figure S4: Smg6 deletion only in cortex and hippocampus does not affect embryo development. Supplementary Table S1: The list of identified DEGs.

**Author Contributions:** Conceptualization, G.M.G., Z.-Q.W., T.-L.L. and P.G.; methodology, G.M.G., D.M., T.K., P.K., M.G., T.-L.L. and P.G.; software, T.K.; validation, G.M.G., T.K., P.K., M.G. and D.M.; formal analysis, G.M.G., T.K., P.K., M.G. and P.G.; investigation, G.M.G., T.-L.L., D.M., T.K., P.K., M.G. and P.G.; resources, Z.-Q.W. and P.G.; data curation, G.M.G., T.K., P.K., M.G. and P.G.; writing—draft preparation, review and editing, G.M.G., Z.-Q.W. and P.G.; visualization, G.M.G., Z.-Q.W. and P.G.; supervision, Z.-Q.W. and P.G.; project administration, Z.-Q.W. and P.G.; funding acquisition, Z.-Q.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** G.M.G. was a member of the Leibniz Gradual School on Aging (LGSA) at the FLI. This work was supported by the Deutsche Forschungsgemeinschaft (DFG) (WA2627/8-1) and the DFGfunded RTG1715. TL was supported by grants from National Natural Science Foundation of China (Nos. 31770871 and 81571380) and a grant (No. KF2020005) from NHC Key Laboratory of Birth Defect for Research and Prevention (Hunan Provincial Maternal and Child Health Care Hospital, Changsha, China). The article is published via an IOAP funding of Friedrich-Schiller-University Jena and Thüringer University and State Library Jena (ThULB).

**Institutional Review Board Statement:** All animal work was conducted according to the German animal welfare legislation and approved by the Thüringer Landesamt für Verbraucherschutz (TLV) (License number: 03-042/16).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The RNAseq dataset presented in this study is available on GEO (GSE186964, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE186964, accessed on 26 November 2021) and other datasets upon request from the corresponding author.

**Acknowledgments:** We are grateful to Patrick Elsner for excellent assistance in the maintenance of the mouse colonies as well as to the imaging and histology facilities for their support. We thank all members of the Wang lab for critical discussion of the project.

**Conflicts of Interest:** The authors declare no competing interest.

#### **References**

