EVA1A/TMEM166 Regulates Embryonic Neurogenesis by Autophagy.

Self-renewal and differentiation of neural stem cells is essential for embryonic neurogenesis, which is associated with cell autophagy. However, the mechanism by which autophagy regulates neurogenesis remains undefined. Here, we show that Eva1a/Tmem166, an autophagy-related gene, regulates neural stem cell self-renewal and differentiation. Eva1a depletion impaired the generation of newborn neurons, both in vivo and in vitro. Conversely, overexpression of EVA1A enhanced newborn neuron generation and maturation. Moreover, Eva1a depletion activated the PIK3CA-AKT axis, leading to the activation of the mammalian target of rapamycin and the subsequent inhibition of autophagy. Furthermore, addition of methylpyruvate to the culture during neural stem cell differentiation rescued the defective embryonic neurogenesis induced by Eva1a depletion, suggesting that energy availability is a significant factor in embryonic neurogenesis. Collectively, these data demonstrated that EVA1A regulates embryonic neurogenesis by modulating autophagy. Our results have potential implications for understanding the pathogenesis of neurodevelopmental disorders caused by autophagy dysregulation.

Introduction

Neural stem cells (NSCs) are a class of multipotent cells responsible for generating all types of neural cells in the nervous system and for maintaining the stem cell population by self-renewal (Gage, 2000). The transition of proliferative and multipotent NSCs to fully differentiated neurons is called neurogenesis. Neurons are generated mainly from early embryonic development to early postnatal stages. After this phase, only a few neurogenic zones remain active in the adult brain (Gotz and Huttner, 2005, Ming and Song, 2011, Paridaen and Huttner, 2014). Embryonic neurogenesis is essential for the maintenance of NSCs, as well as for the generation of functional neural cell types and is therefore important for brain development and function.

Autophagy is an evolutionarily conserved process in which cellular proteins and organelles are engulfed by autophagosomes and delivered to lysosomes for degradation. Autophagy is involved in a wide variety of physiological and pathological processes, such as cancer, immune and infectious diseases, and neurodegeneration (Jaeger and Wyss-Coray, 2009, Rosello et al., 2012). Among these diseases, the role of autophagy in neurodegeneration has been studied extensively. Deletion of autophagy-related genes, such as Atg5 and Atg7, caused dysregulation of autophagosome formation and consequent neurodegeneration (Komatsu et al., 2005, Komatsu et al., 2006, Kuma et al., 2004). In fact, many studies have suggested that modulators of autophagy may be used as potential therapeutic strategies to treat diseases caused by autophagy defects (Harris and Rubinsztein, 2012, Nassif and Hetz, 2011, Dalby et al., 2010). Although studies have suggested that autophagy has neuroprotective effects, the role of autophagy in regulating embryonic neurogenesis remains largely unknown.

Previous studies have demonstrated that autophagy occurs in physiological conditions in the brain during the development of the nervous system and that it plays a significant role in cell differentiation (Schweichel and Merker, 1973, Zhao et al., 2010, Zeng and Zhou, 2008). Some autophagy-related genes that function in the inhibition of neurodegeneration are also essential for CNS development. Specifically, AMBRA1, an activating molecule in BECLIN 1-regulated autophagy, is essential for the development of the nervous system (Fimia et al., 2007, Yazdankhah et al., 2014, Vazquez et al., 2012). Another autophagy-related gene, Atg5, plays an important role in neurogenesis and gliogenesis. (Vazquez et al., 2012, Lv et al., 2014, Wang et al., 2014). These findings suggested a critical role of autophagy in the development of the nervous system. However, the precise mechanisms through which neurogenesis is regulated by autophagy is far from being well elucidated. Many molecular targets of this process need to be identified.

EVA1A (Eva-1 homolog A), also known as TMEM166 (Transmembrane protein 166) or FAM176A (Family with sequence similarity 176), is a lysosome and ER-associated protein that can regulate cell autophagy and apoptosis (Wang et al., 2007). It is conserved in humans, chimpanzees, rats, mice, and dogs, indicating that it may have important functions in vertebrate animals. Previous studies revealed that EVA1A is expressed in a cell-type-specific and tissue-type-specific manner and is significantly downregulated in cancer tissues (Xu et al., 2013, Sun et al., 2012, Tao et al., 2015). In vivo and in vitro experiments demonstrated that EVA1A overexpression inhibits tumor cell proliferation by autophagy (Chang et al., 2013, Xie et al., 2014). These findings suggested a close correlation between EVA1A-induced autophagy and cancer suppression. Other studies have suggested that EVA1A-induced autophagy plays a significant role in cell death following focal cerebral ischemic injury (Li et al., 2012). Lu et al. (2015) demonstrated that EVA1A/TMEM166 is a key player in C/EBPα-mediated autophagy induction and protection against starvation. However, the role of EVA1A in neurogenesis has not been reported, and the mechanism by which EVA1A regulates autophagy remains to be determined.

In the present study, we generated an Eva1a knockout (KO) mouse to study the physiological role of EVA1A in vivo. The analysis of Eva1a KO embryos and isolated NSCs revealed the function of EVA1A in self-renewal and neuronal differentiation of NSCs. Our study highlighted a previously unknown role of EVA1A in the CNS.

Results

EVA1A Expression in Brain Development

To determine the role of EVA1A in brain development, we first evaluated the expression of EVA1A in the embryonic cortex from embryonic day 10.5 (E10.5) to E16.5 and postnatal 2 day (P2). Western blotting revealed that EVA1A expression increased substantially from E10.5 to E16.5 and gradually decreased after P2 (Figures S1A and S1B). Consistent with this observation, immunofluorescence and confocal microscopy observations indicated that the fluorescence signals of EVA1A were elevated at E10.5 and peaked at E16.5 (Figure S1C). Interestingly, the level of autophagy, which is indicated by the autophagy marker LC3B-II, also increased from E10.5 to E16.5 (Figures S1A, S1B, and S1D). Simultaneously, β-TUBULIN III, the neuron marker, was markedly upregulated during this period (Figures S1A, S1B, and S1E). These results demonstrate that neuron differentiation occurs at the early embryonic development stage and the initiation of neuron differentiation corresponds to the expression of EVA1A and the autophagy level, suggesting that EVA1A might be involved in brain development including neuron differentiation.

Generation of Eva1a Gene Knockout Mice

To investigate the physiological function of EVA1A in the development of the CNS, we generated Eva1a gene KO mice. To this end, we crossed Eva1a mice (Figure 1A) with transgenic mice expressing Cre recombinase under the control of the Nestin promoter (Nestin-Cre). The Eva1a+/+ and Eva1a−/− mice were identified by PCR analysis of mouse tail DNA (Figure 1B). The results of RT-PCR and western blotting demonstrated that the level of EVA1A was low or absent in the indicated brain areas (cortex, hippocampus, and cerebellum) in Eva1a−/− mice at E12.5 (Figures 1C and 1D). Consistent with this result, an immunofluorescence assay confirmed that EVA1A protein signals were nearly absent in the cortex of the Eva1a−/− mice (Figure 1E versus 1F). We also evaluated the protein level of EVA1A in other tissues, such as liver, lung, heart, and muscle. There was no significant change between the Eva1a−/− and Eva1a+/+ mice (data not shown), indicating that the EVA1A deficiency was restricted to the CNS.

Figure 1

Generation of Eva1a KO Mice

(A) Scheme to generate Eva1a conditional knockout mice.

(B) Genomic DNA was extracted from mouse tails and analyzed by PCR.

(C) Brain homogenates (cortex, hippocampus, and cerebellum) were obtained from Eva1a+/+ mice and Eva1a−/− mice at E12.5. The Eva1a mRNA level was analyzed by RT-PCR, and the protein level was analyzed by western blotting.

(D) Statistical analysis of the level of Eva1a in different regions of the brain. The average value in Eva1a+/+ brain tissues was normalized as 1. Data are means ± SD of the results from three independent experiments. ∗∗∗p < 0.001; t test.

(E and F) Frozen sections were obtained from whole-brain samples of both Eva1a+/+ and Eva1a−/− mice at E12.5, and an immunofluorescence assay was carried out to detect EVA1A expression in the cortex (scale bar, 100 μm). The right panel shows higher magnifications (scale bar, 10 μm). The white arrow indicates the EVA1A protein signal in the whole-brain samples.

Eva1a Deletion Decreases the Ability of NSCs Self-Renewal, but Fails to Induce Cell Apoptosis

To investigate the role of EVA1A in NSC self-renewal, we first detected the expression of the NSCs markers, NESTIN and SOX2, in the whole-brain tissue at E12.5 in both Eva1a−/− and Eva1a+/+ mice. Our results revealed that the levels of both NESTIN and SOX2 were markedly decreased in the Eva1a−/− mice (Figure 2A). We next injected bromodeoxyuridine (BrdU) into both the Eva1a+/+ and Eva1a−/− mice at E12.5 and detected proliferating cells in living tissues. After 4 hr, the frozen sections were obtained from whole-brain samples and subjected to immunofluorescence assay. Confocal microscopy showed that the Eva1a−/− mice displayed a weaker signal for NESTIN staining (Figure 2B versus 2C, left panel) and the number of NESTIN+ BrdU+ positive cells was decreased in the Eva1a−/− cortex compared with those in the Eva1a+/+ mice, indicating the number of newborn NSCs was reduced in the Eva1a−/− mice (Figures 2B–2D). To rule out the possibility that the reduction in NSCs was the result of cell apoptosis, we performed a terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay in brain sections. There was no significant difference in TUNEL-positive cells between Eva1a−/− and Eva1a+/+ mice, indicating that the decrease in NSC self-renewal is independent of apoptosis (Figures 2E–2G).

Figure 2

Eva1a Knockout Decreases NSCs Self-Renewal, but Fails to Induce Cell Apoptosis

(A) Immunoblot analysis of EVA1A, NESTIN, and SOX2 expressions in whole-brain tissue from Eva1a+/+ and Eva1a−/− mice at E12.5.

(B and C) Pregnant female mice at E12.5 were injected intraperitoneally with BrdU 4 hr before euthanasia. Brain frozen sections were obtained and subjected to an immunofluorescence assay. Representative confocal microscopy images of the fluorescence of NESTIN and BrdU were analyzed. Scale bar, 100 μm.

(D) Quantification of BrdU and NESTIN double-positive cells within a 350-μm slab of the brain section in Eva1a+/+ and Eva1a−/− littermate mice. Five randomly selected sections from each mouse were examined. Three independent experiments were performed. Data are means ± SD ∗∗p < 0.01; t test.

(E and F) Frozen sections were obtained from whole-brain samples of both Eva1a+/+ and Eva1a−/− mice at E12.5, and subjected to a TUNEL assay. Representative confocal microscopy images of the fluorescence were analyzed. Scale bar, 100 μm.

(G) The percentage of TUNEL-positive cells was calculated as the mean of ten different fields from each section. Five randomly selected sections from each mouse were examined. Three independent experiments were performed. Data are means ± SD. n.s., not significant.

(H) Eva1a+/+ and Eva1a−/− NSCs were grown from E12.5 embryos under proliferative conditions. Neurosphere formation was viewed under phase contrast microscopy and representative microscopy images are shown. Scale bar, 25 μm.

(I and J) Eva1a+/+ and Eva1a−/− NSCs were cultured in 96-well non-coated plates with 500 or 1,000 cells per well under the same conditions as in (H). The relative neurosphere number and diameter were calculated as shown in the histograms. The average value in the Eva1a+/+ group was normalized as 100%. Data are means ± SD of the results from three independent experiments. ∗∗∗p < 0.001; t test.

(K) Eva1a+/+ and Eva1a−/− NSCs were grown from E12.5 embryos under proliferative conditions. Cell apoptosis was detected using the FITC-Annexin V plus propidium iodide staining kit by flow cytometry.

(L) Apoptotic cells were calculated and each bar represents the mean ± SD from three independent experiments. n.s., not significant.

We next investigated the influence of EVA1A on NSC self-renewal in vitro. As shown in Figures 2H and 2I, the number of secondary neurospheres generated from Eva1a−/− NSCs was markedly lower than that from Eva1a+/+ NSCs. At the same time, deletion of Eva1a in NSCs resulted in smaller-sized neurospheres compared with their Eva1a+/+ counterparts (Figures 2H and 2J), indicating defective self-renewal in Eva1a−/− NSCs. Furthermore, Annexin V staining and flow cytometry analysis indicated there was no significant difference in cell apoptosis between Eva1a−/− NSCs and Eva1a+/+ NSCs (Figures 2K and 2L), excluding the possibility that the defective NSC self-renewal was caused by apoptosis. Collectively, these observations supported a positive role of EVA1A in NSC self-renewal.

NSCs Differentiation Is Impaired in Eva1a−/− Mice

Having demonstrated the correlation between EVA1A and NSC self-renewal, we then investigated the effect of EVA1A on NSC differentiation. We first examined the levels of both EVA1A and β-TUBULIN III (a marker of neurons) in Eva1a−/− and Eva1a+/+ mice at E16.5. As shown in Figures 3A–3D, Eva1a was nearly absent in Eva1a−/−, which is associated with a concomitant decrease in β-TUBULIN III protein. Data from immunofluorescence assay in brain frozen sections confirmed that the β-TUBULIN III showed a stronger fluorescence signal in Eva1a+/+ mice than in Eva1a−/− mice (Figure 3E versus 3F, left panel). Meanwhile, the number of β-TUBULIN III+ BrdU+ double-positive cells within a 350-μm slab of the brain was significantly downregulated in Eva1a−/− mice (from around 31.66% to 14.33%) (Figures 3E–3G), suggesting damage to the newborn neuron population.

Figure 3

NSC Differentiation Is Impaired in Eva1a KO Mice

(A) RT-PCR analysis of Eva1a and β-tubulin III mRNA in whole-brain tissue from Eva1a+/+ and Eva1a−/− mice at E16.5.

(B) Quantifications of the ratio of β-tubulin III/Gapdh. The average value in the Eva1a+/+ group was normalized as 1. Data are means ± SD of the results from three independent experiments. ∗∗∗p < 0.001; t test.

(C) Immunoblot analysis of EVA1A and β-TUBULIN III expression in whole-brain tissue from Eva1a+/+ and Eva1a−/− mice at E16.5.

(D) Histogram shows the ratio of β-TUBULIN III/ACTB. The average value in the Eva1a+/+ group was normalized as 1. Data are means ± SD of the results from three independent experiments. ∗p < 0.05; t test.

(E and F) Pregnant female mice at E16.5 were injected intraperitoneally with BrdU 4 hr before euthanasia. Brain frozen sections were obtained and subjected to an immunofluorescence assay. Representative confocal microscopy images of the fluorescence of β-TUBULIN III and BrdU were analyzed. Scale bar, 50 μm.

(G) Quantification of β-TUBULIN III and BrdU double-positive cells within a 350-μm slab of the brain section in Eva1a+/+ and Eva1a−/− littermate mice. Five randomly selected sections from each mouse were examined. Three independent experiments were performed. Data are means ± SD ∗∗∗p < 0.001; t test.

(H) NSCs from Eva1a+/+ and Eva1a−/− mice at E12.5 were cultured under differentiation conditions for 6 days. The levels of β-TUBULIN III were evaluated by western blotting.

(I) Quantifications of the amount β-TUBULIN III protein relative to ACTB are shown. The average value in the Eva1a+/+ group was normalized as 1. Data are means ± SD of the results from three independent experiments. ∗∗∗p < 0.001; t test.

(J) The treatment of NSCs was the same as in (H), after which, the cells were subjected to immunostaining against β-TUBULIN III and viewed under a confocal microscope. Scale bar, 75 μm.

(K and L) Five randomly selected areas from each slide were examined for the percentage of β-TUBULIN III+-differentiated neurons and β-TUBULIN III+ cells with neurites under the fluorescence microscope. Three independent experiments were performed. Data are means ± SD. ∗∗p < 0.01; t test.

Next, we established NSC cultures, in which the transition from proliferative to differentiated cells can be easily monitored in short-term cell culture, following treatment with all-trans retinoic acid for direct neuronal differentiation. NSCs from Eva1a+/+ and Eva1a−/− mice at E12.5 were cultured under differentiation conditions for 6 days. The levels of β-TUBULIN III were lower in Eva1a−/− NSC cultures (Figures 3H and 3I). Consistent with the results of western blotting, the percentage of β-TUBULIN III+ differentiated neurons was decreased in Eva1a−/− NSC cultures (Figures 3J and 3K). We also found that the neurites of Eva1a−/− neurons were shorter than those from Eva1a+/+ neurons (Figures 3J and 3L), indicating a deficiency in neurite formation and neuron maturation. Collectively, these data suggested that EVA1A may be important for the neuronal differentiation of NSCs.

EVA1A Modulates Autophagy through the PIK3CA/AKT-mTOR Pathway

Previous studies have demonstrated that dysregulation of autophagy can lead to abnormal neuronal hyperproliferation, cell death, and differentiation (Levine and Kroemer, 2008, Cecconi et al., 2007). We observed that autophagy is activated during embryonic neurogenesis, which is accompanied by an increase in EVA1A levels (Figure S1). Therefore, we hypothesized that EVA1A might regulate embryonic neurogenesis via autophagy. Initially, we analyzed the autophagy-related genes in the cortex of both Eva1a+/+ and Eva1a−/− mice at E16.5. We found that the mRNA levels of Atg5 and Lc3b were decreased in Eva1a-depleted cortex tissue (Figure 4A). We next measured the conversion of LC3B using western blotting. The conversion of LC3B-I to LC3B-II occurs during autophagy, and the amount of LC3B-II is correlated with the number of autophagosomes (Klionsky et al., 2012). It was shown that the accumulation of LC3B-II was reduced in the Eva1a−/− cortex compared with the Eva1a+/+ cortex (Figure 4B). Consistent with the results of western blotting, confocal microscopy observation indicated that the number of LC3B puncta was decreased in the Eva1a−/− cortex (Figure 4C versus 4D).

Figure 4

EVA1A Modulates Autophagy through the PIK3CA/AKT-mTOR Pathway

(A) The mRNA levels of autophagy-related genes in whole-brain tissue from Eva1a+/+ and Eva1a−/− mice at E16.5 were detected by RT-PCR.

(B) The protein levels of SQSTM1 and LC3B in whole-brain tissue were detected by western blotting.

(C–F) Brain frozen sections were obtained from both Eva1a+/+ and Eva1a−/− mice at E16.5, which were then subjected to an immunofluorescence assay. Representative confocal microscopy images of the distribution of β-TUBULIN III, LC3B, and SQSTM1 were analyzed.

(G and H) Representative confocal microscopy images of the distribution of ubiquitin are shown. Scale bar, 25 μm.

(I and K) Western blot analysis with the indicated antibodies in whole-brain tissue from both Eva1a+/+ and Eva1a−/− mice at E16.5.

(J and L) Densitometric analysis of the immunoblots. ACTB was assessed as a loading control. The average value in the Eva1a+/+ group was normalized as 1. Data are means ± SD of the results from three independent experiments. ∗p < 0.05, ∗∗p < 0.01; t test.

It has been widely recognized that the SQSTM1 (sequestosome 1)/P62 protein serves as a link between LC3 and ubiquitinated substrates. It incorporates into the completed autophagosome and is degraded in lysosome. Therefore, the level of SQSTM1 is a marker for autophagic flux (Klionsky et al., 2012). As shown in Figure 4B, the basal level of SQSTM1 was upregulated and formed numerous aggregates in the Eva1a−/− cortex compared with the Eva1a+/+ cortex (Figure 4E versus 4F). The presence of strong ubiquitin staining was also observed in the Eva1a−/− cortex (Figure 4G versus 4H), indicating blockage of autophagic flux.

In vitro experiments were next performed to corroborate the above-mentioned autophagic phenotypes. NSCs were generated from both Eva1a−/− and Eva1a+/+ mice at E12.5. RT-PCR showed that loss of Eva1a in NSCs resulted in a reduction of Atg5 and Lc3b mRNA levels (Figure 5A). Simultaneously, decreased LC3B-II and increased SQSTM1 protein levels were also observed in Eva1a−/− NSCs compared with Eva1a+/+ NSCs (Figure 5B). Consistent with the results of western blotting, the number of LC3B dots per cell in Eva1a−/− NSCs was significantly decreased compared with that in Eva1a+/+ NSCs (Figures 5C and 5D). These data were consistent with those from the mouse cortex. Taken together, our data suggested that autophagic activity was impaired in Eva1a−/− cells. Importantly, the decrease of β-TUBULIN III levels in the Eva1a−/− cortex was associated with decreased LC3B puncta, as well as increased SQSTM1 accumulation (Figures 4C–4F), indicating that Eva1a depletion-triggered damage to neurogenesis might result from disrupted autophagy.

Figure 5

Eva1a Knockout in NSCs Increases the PIK3CA/AKT-mTOR Signal Pathway

(A) NSCs from Eva1a+/+ and Eva1a−/− at E12.5 were cultured under proliferation conditions. The mRNA levels of autophagy-related genes were detected by RT-PCR.

(B) The treatment of NSCs was the same as in (A). The levels of LC3B and SQSTM1 were analyzed by western blotting.

(C) The treatment of NSCs was the same as in (A). The cells were then plated on glass coverslips coated with poly-ornithine for 24 hr. NSCs were subjected to immunofluorescence against LC3B and viewed under a confocal fluorescence microscopy. Scale bar, 10 μm.

(D) Histogram showing the quantification of the numbers of LC3B puncta per cell under confocal microscopy. Three independent experiments were performed. Results are means ± SD of at least 100 non-overlapping cells scored. ∗∗∗p < 0.01; t test.

(E and G) The treatment of NSCs was the same as in (A). NSCs were lysed and subjected to immunoblot analysis using the indicated antibodies.

(F and H) Densitometric analysis of the immunoblots. ACTB was assessed as a loading control. The average value in the Eva1a+/+ group was normalized as 1. Data are means ± SD of the results from three independent experiments. ∗p < 0.05, ∗∗∗p < 0.001; t test.

We then investigated the mechanisms of the autophagy defective in Eva1a−/− cells. The mammalian target of rapamycin (mTOR) has been known as a negative regulator of autophagy. Experiments showed that the phosphorylation levels of mTOR at Ser2448 (activation site) and RPS6KB1 (ribosomal protein S6 kinases, 70 kDa, polypeptide 1, a substrate of mTOR) as well as EIF4EBP1 (Eukaryotic translation initiation factor 4E binding protein 1), two downstream effectors of mTOR, were elevated in the Eva1a−/− cortex (Figures 4I and 4J). Furthermore, we examined mTOR signaling in NSCs from both Eva1a−/− and Eva1a+/+ NSCs at E12.5. As shown in Figures 5E and 5F, the levels of p-mTOR, p-RPS6KB1, and p-EIF4EBP1 were elevated in the Eva1a−/− NSCs. Collectively, these results support the hypothesis that an increase of mTOR activity contributes functionally to the autophagy defect in Eva1a−/− cells.

The PIK3CA-AKT signaling pathway plays a key role in various cellular processes, including cell growth, proliferation, survival, and autophagy. Activation of PIK3CA results in phosphorylation of phosphatidylinositol 4,5-bisphosphate (PIP2) to form PIP3. Subsequently, PIP3 recruits downstream effector proteins such as serine/threonine kinase AKT. Once activated, AKT-mediated phosphorylation of TSC2 (tuberous sclerosis complex 2) inhibits the TSC1-TSC2 complex function, and thus activates mTOR signaling. Therefore, further experiments were performed to determine the influence of Eva1a deletion on PIK3CA-AKT signal. As shown in Figures 4K and 4L, there was a significant increase in the phosphorylation of PIK3CA and AKT in the Eva1a−/− cortex at E16.5. Importantly, the basal phosphorylation levels of TSC2 were decreased. Similarly, we analyzed the PIK3CA/AKT signaling pathway in NSCs at E12.5. Consistent results were obtained in the absence of Eva1a in NSCs (Figures 5G and 5H). Taken together, these results suggested that loss of Eva1a leads to the activation of the PIK3CA-AKT axis, which results in the inactivation of TSC2 and the subsequent activation of mTOR signals.

Overexpression of EVA1A Restores Self-Renewal and Neuronal Differentiation in Eva1a−/− NSCs

Eva1a depletion induced defects in NSC differentiation; therefore, we performed reciprocal EVA1A gain-of-function experiments using NSCs at E12.5 and an Ad5-Eva1a construct. The number and the size (diameter) of secondary neurosphere formed from the NSCs were used as a measure of NSC self-renewal. Our results showed that neurosphere formation was decreased in Eva1a-deleted NSCs compared with that of Eva1a+/+ NSCs. However, overexpression of EVA1A in Eva1a−/− NSCs increased the number of neurospheres (Figures 6A and 6B, left panel). Meanwhile, the neurosphere diameter in EVA1A-overexpressing Eva1a−/− NSCs was greater than that of Eva1a−/− NSCs alone (Figures 6A and 6B, right panel). In addition, Eva1a−/− NSCs allowed to differentiate in cultures generated a small number of β-TUBULIN III-positive neurons compared with that in Eva1a+/+ NSCs (Figure 6D, left panel). After EVA1A overexpression in Eva1a−/− NSCs, the percentage of differentiated neurons increased from around 15% to 25% (Figure 6D, left panel). Furthermore, compared with the Eva1a−/− NSCs alone, the Eva1a−/− NSCs overexpressing EVA1A appeared to produce longer neurites (Figures 6C and 6D, right panel), indicating enhanced neuronal differentiation and maturation by EVA1A overexpression. These results suggested that Eva1a depletion-induced deficiency in neurogenesis can be rescued by overexpression of EVA1A, indicating the specificity of EVA1A function.

Figure 6

Overexpression of EVA1A Restores Self-Renewal and Neuronal Differentiation in Eva1a−/− NSCs

(A) NSCs obtained from Eva1a+/+ and Eva1a−/− at E12.5 were infected or not infected with Ad5-Eva1a and further cultured for 12 hr under proliferative conditions. Neurosphere formation in NSCs from Eva1a+/+, Eva1a−/− and Eva1a−/− plus Ad5-Eva1a were viewed by phase contrast optical microscopy. Scale bar, 25 μm.

(B) NSCs from Eva1a+/+ and Eva1a−/− mice were cultured in 96-well non-coated plates with 500 or 1,000 cells per well under the same conditions as in (A). The number and the size (diameter) of secondary neurospheres were evaluated. The average value in the Eva1a+/+ group was normalized as 100%. Data are means ± SD of the results from three independent experiments. ∗∗∗p < 0.001; one-way ANOVA with Tukey's post hoc test.

(C) NSCs obtained from Eva1a+/+ and Eva1a−/− at E12.5 were infected or not infected with Ad5-Eva1a and cultured under differentiation conditions for 6 days. Then the cells were subjected to immunostaining against β-TUBULIN III and viewed under a fluorescence microscope.

(D) Five randomly selected areas from each slide were examined for the percentage of β-TUBULIN III+-differentiated neurons in total cells and β-TUBULIN III+ cells with neurites under the fluorescence microscope. Three independent experiments were performed. Data are means ± SD. ∗∗p < 0.01, ∗∗∗p < 0.001; one-way ANOVA with Tukey's post hoc test.

(E and G) The treatment of NSCs was the same as in (A). NSCs were lysed and subjected to immunoblot analysis using the indicated antibodies.

(F and H) Densitometric analysis of the immunoblots. ACTB was assessed as a loading control. The average value in the Eva1a+/+ group was normalized as 1. Data are means ± SD of the results from three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; one-way ANOVA with Tukey's post hoc test.

Eva1a depletion leads to the activation of the PIK3CA/AKT-mTOR pathway; therefore, we examined whether overexpression of EVA1A could inactivate this pathway and increase autophagy levels. As expected, overexpression of EVA1A in Eva1a−/− NSCs enhanced the conversion of LC3B and SQSTM1 degradation (Figures 6E and 6F, lane 3 versus lane 2), suggesting the activation of autophagy. Simultaneously, overexpression of EVA1A in Eva1a−/− NSCs significantly downregulated the levels of p-RPS6KB1 and p-EIF4EBP1 (Figures 6E and 6F, lane 3 versus lane 2), indicating the inhibition of the mTOR signal. We subsequently detected the PIK3CA-AKT signal in such conditions. As shown in Figures 6G and 6H, the phosphorylation levels of PIK3CA and AKT were decreased and TSC2 phosphorylation was increased in Eva1a−/− NSCs overexpressed by Eva1a. Furthermore, a PIK3CA/AKT inhibitor (perifosine) was used to confirm the relationship between EVA1A-regulated autophagy and the PIK3CA/AKT-mTOR pathway. Data from experiments indicated that perifosine-mediated inhibition of PIK3CA/AKT-mTOR at least partly rescued autophagy deficiency in Eva1a−/− NSCs at E16.5 (Figures S2A–S2D). Importantly, we observed that perifosine treatment could partly reverse the defective self-renewal and defective differentiation caused by Eva1a knockout (Figures S2E–S2H). These data further confirmed that EVA1A regulates autophagy via the suppression of the PIK3CA/AKT-mTOR signaling, and that EVA1A-mediated autophagy is involved in self-renewal and neuronal differentiation.

Eva1a Depletion Induced Impaired NSC Differentiation Can Be Reversed by Methylpyruvate and Rapamycin

Autophagy is a potential energy source during cellular restructuring, which is associated with neuronal differentiation and neurite outgrowth (Hara et al., 2006, Levine and Klionsky, 2004). A recent study has demonstrated that autophagy inhibition with 3-methyladenine (3-MA) reduces ATP levels during retinal neurogenesis, and this inhibition could be reversed by treatment with methylpyruvate (MP) (Mellen et al., 2008). MP is a permeable analog for the citric acid cycle. To clarify the correlation among EVA1A, neurogenesis, and the availability of energy in the course of autophagy, we performed the following test. NSCs from Eva1a+/+ and Eva1a−/− mice at E12.5 were grown as neurospheres and cultured under differentiation conditions for 6 days. In the Eva1a−/− group, the number of β-TUBULIN III+ neurons was decreased (Figures 7A, 7B, and 7E), and the neurite length appeared shorter than that from the Eva1a+/+ NSC group in the normal culture (Figures 7A, 7B, and 7F), which indicated a decrease in neuronal differentiation. However, MP treatment markedly increased the number of β-TUBULIN III+ neurons and the neurite length in the Eva1a−/− group (Figures 7C–7F), while no significant changes were observed in the Eva1a+/+ group with or without MP treatment (Figures 7A and 7C). Simultaneously, we further detected the autophagy levels under such conditions. In normal culture conditions, the expression of β-TUBULIN III protein was downregulated in Eva1a−/− NSCs compared with that of Eva1a+/+ NSCs (Figures 7G and 7H, lane 3 versus lane 1). Meanwhile, decreased β-TUBULIN III was accompanied by weakened LC3B-II accumulation and increased SQSTM1 levels (Figures 7G and 7H, lane 3 versus lane 1), which was consistent with the above observation. After treatment with MP, β-TUBULIN III expression showed weak upregulation. At the same time, accumulation of LC3B-II was significantly elevated and a decrease in SQSTM1 levels was observed in Eva1a−/− NSCs (Figures 7G and 7H, lane 4 versus lane 3). Data from transmission electron microscopy analysis shown that MP treatment increased autophagic structures in the Eva1a−/− NSCs compared with that in Eva1a−/− NSCs alone (Figure S3), suggesting enhanced autophagy. Collectively, these results implied that MP treatment recovers impaired autophagy in Eva1a−/− NSCs.

Figure 7

MP Treatment Rescues NSC Differentiation in Eva1a−/− Mice

(A–D) NSCs from Eva1a+/+ and Eva1a−/− at E12.5 were grown as neurospheres, and then plated under differentiation conditions for 6 days with or without MP (10 mM) treatment. Cells were fixed for immunofluorescence analysis. Representative fields of Eva1a+/+ and Eva1a−/− neurons were stained with β-TUBULIN III antibodies. Nuclei were stained with Hoechst 33342 (scale bar, 75 μm). The right panel shows higher magnifications (scale bar, 25 μm).

(E and F) Five randomly selected areas from each slide were examined for the percentage of β-TUBULIN III+-differentiated neurons in total cells and β-TUBULIN III+ cells with neurites under the fluorescence microscope. Three independent experiments were performed. Data are means ± SD. ∗∗p < 0.01; one-way ANOVA with Tukey's post hoc test.

(G) The treatment of NSCs was the same as in (A–D), after which the NSCs were lysed and subjected to immunoblot analysis using the indicated antibodies.

(H) Histogram showing the quantification of the ratio of the relative proteins/ACTB. The average value in the Eva1a+/+ group (lane 1) was normalized as 1. Data are means ± SD of the results from three independent experiments. ∗∗p < 0.01, ∗∗∗p < 0.001; one-way ANOVA with Tukey's post hoc test.

To further consolidate the hypothesis that defective neurogenesis is caused by autophagy deficiency, we utilized the mTOR-dependent chemical rapamycin and mTOR-independent chemical lithium (LiCl) to induce autophagy for the rescue assay. As shown in Figures S4A and S4B, rapamycin could increase autophagy both in Eva1a+/+ and Eva1a−/− NSCs at E16.5. LiCl treatment could enhance autophagy in Eva1a+/+ NSCs (Figures S4C and S4D, lane 2 versus lane 1), but failed in Eva1a−/− NSCs (Figures S4C and S4D, lane 4 versus lane 3), indicating that the autophagy impairment caused by Eva1a deletion is associated with the mTOR activity.

We further examined the differentiation of NSCs under the same treatment conditions. Data obtained from western blotting indicated that rapamycin treatment could increase the levels of β-TUBULIN III in Eva1a−/− NSCs (Figures S4A and S4B, lane 4 versus lane 3). LiCl, however, had no effect on the level of β-TUBULIN III in Eva1a−/− NSCs (Figures S4C and S4D, lane 4 versus lane3). Simultaneously, we found that rapamycin treatment, but not LiCl, enhanced the formation and maturation of neurons in Eva1a−/− NSCs (Figures S4E–S4H). Together, our data reveal that the defective neurogenesis in Eva1a−/− NSCs could be rescued by mTOR-dependent autophagy modulators, at least partly.

Discussion

In the present study, the role of EVA1A/TMEM166 in the process of neurogenesis was investigated using Eva1a KO mice. We demonstrated that EVA1A regulates NSC self-renewal and neuronal differentiation by modulating autophagy. Genetic disruption of Eva1a caused a deficiency in the autophagic machinery, which consequently impaired the differentiation process, decreasing the number of newborn neurons and impeding their maturation in culture. This effect could be rescued by overexpression of EVA1A or treatment with MP in vitro. Our studies suggest strongly that EVA1A modulates the process of autophagy, which may provide energy for the transition from NSCs to mature neurons.

Several studies have suggested that autophagy is activated during nervous system development and cell differentiation (Nassif and Hetz, 2011, Schweichel and Merker, 1973, Zhao et al., 2010). Initially, we observed that the EVA1A levels increased over the course of neurogenesis, along with a concomitant increase in autophagy in wild-type mice. Based on this, we speculated that increased expression of EVA1A in NSCs may play a role in neurogenesis. Here, we present several lines of evidence that demonstrate that the loss of EVA1A leads to substantial decline in NSC self-renewal, as well as in neuronal differentiation in vivo and in vitro. The phenotype changes observed in Eva1a-deleted NSCs could be rescued by a gain-of-function assay in vitro.

The mechanism by which EVA1A regulates neurogenesis is unclear. Autophagy plays a critical role in neurogenesis and is mainly regulated by the mTOR pathway; therefore, we examined the effect of Eva1a depletion on the mTOR pathway. The phosphorylation levels of mTOR, RPS6KB1, and EIF4EBP1 were increased in NSCs derived from Eva1a−/− mice. Consistent with these findings, the levels of p-PIK3CA and p-AKT (upstream of mTOR) also increased in Eva1a-depleted cells, which was accompanied by a substantial decrease in p-TSC2 levels. Thus, it can be seen that the PIK3CA-AKT signal negatively regulates TSC1/2 and activates mTOR, and consequently, inhibits autophagy. Importantly, transduction of Ad5-Eva1a in Eva1a−/− NSCs attenuated this PIK3CA/AKT-mTOR signaling pathway, and restored the autophagy level. These results were consistent with those of a previous report, which indicated that overexpression of EVA1A/TMEM166 decreased the activity of mTOR in human tumor cells (Chang et al., 2013). Our studies indicated that EVA1A probably regulates neurogenesis through the PIK3CA/AKT-mTOR pathway. However, further studies are necessary to understand how EVA1A regulates the PIK3CA/AKT axis precisely.

Autophagy is a complex catabolic program that functions in the lysosomal degradation of proteins and other subcellular constituents and serves as an effective method of providing metabolic precursors. Previous studies demonstrated that neuroepithelial cells undergo differentiation in an energy-consuming process (Vazquez et al., 2012, Mellen et al., 2008, Qu et al., 2007), and autophagy can provide sufficient energy for the process of neurogenesis. Until recently, only a few studies had demonstrated the effects of autophagy on neurogenesis in vivo. However, how autophagy regulates neurogenesis remains largely unknown. In this study, we observed that NSC self-renewal and differentiation were decreased in Eva1a-deficient mice and autophagy was decreased. Decreased levels of autophagy may lead to a shortage of energy supply and impede the process of neurogenesis; therefore, we treated Eva1a-defective NSCs with MP. Generally, MP treatment restored neuronal differentiation in these cells. Thus, Eva1a deletion-induced defective neuronal differentiation is probably associated with a lack of available energy caused by the inhibition of autophagy.

In summary, our results revealed that loss of Eva1a leads to impaired NSC self-renewal and differentiation, along with a decrease in autophagy, which is inversely associated with the activation of the PIK3CA/AKT-mTOR signaling. Functionally, EVA1A overexpression or MP treatment substantially rescued Eva1a-depletion-induced NSC self-renewal and neuronal differentiation deficiency, demonstrating the crucial function of EVA1A in embryonic neurogenesis. Our investigations provide insights into the activities of EVA1A, as well as the role of basal autophagy in neurogenesis. Moreover, our results provide a basis for the identification of molecular targets for the treatment of neurodevelopmental disorders caused by dysregulation of autophagy.

Experimental Procedures

Generation of Tissue-Specific Eva1a-Deficient Mice

Eva1aflox/flox mice with the C57BL/6J background were constructed by the Model Animal Research Center of Nanjing University (Nanjing, China). Nestin-Cre transgenic mice were provided by Prof. Hong Zhang (Chinese Academy of Sciences, Beijing, China). Progeny containing the Eva1a flox allele were bred with these Cre transgenic mice to generate Eva1aflox/flox; Nestin-Cre mice. All mice used in the study were bred and maintained at the Experimental Animal Center, Peking University Health Sciences Center (Beijing, China). The animal experiment protocol was approved by the Biomedical Research Ethics Committee of Peking University and strictly adhered to the American Physiological Society's Guiding Principles in the Care and Use of Vertebrate Animals in Research and Training.

Determination of the Genotype of Eva1a Mutant Embryos

To genotype embryos, early-stage embryos' tails were cut and then the genome was genotyped by PCR. The primers used in this study are listed in Table S1.

Protein Extraction and Western Blotting

Total proteins from mouse tissues or cells were extracted using lysis buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, and with freshly added proteinase inhibitor cocktail and phosphatase inhibitors). Protein concentrations were determined by the BCA protein assay reagent (Pierce, NCI3227CH). Equal amounts of proteins were separated by SDS-PAGE electrophoresis and transferred to nitrocellulose membranes. After blocking with 5% nonfat milk, the membranes were incubated with primary antibodies, washed, and then incubated with the secondary antibodies. The membranes were then washed and scanned using an Odyssey Infrared Imaging System (LI-COR Biosciences). Scanned bands were quantified using the ImageJ software. The results were representative of at least three experiments.

Immunofluorescence

Brain frozen sections from both Eva1a−/− and Eva1a+/+ mice at E12.5 and E16.5 were fixed with 4% formaldehyde, washed with PBS, and blocked with PBS with 5% BSA and 0.25% Triton X-100 at room temperature for 1 hr. The sections were incubated with the indicated primary antibody at 4°C overnight, washed with PBS three times, and incubated with fluorescein isothiocyanates (FITC) or rhodamine-labeled secondary antibodies at 37°C for 1 hr. Nuclei were stained with Hoechst 33342 for 5 min. Finally, immunofluorescence was detected under a confocal fluorescence microscope (LSM 510 Meta plus Axiovert zoom, Carl Zeiss) with a 63×/1.40 NA oil immersion objective lens (PlanApochromat; Carl Zeiss) and a camera (AxioCam HRm, Carl Zeiss). Images were processed and viewed using the LSM Image Browser software.

RT-PCR

Total RNA was extracted from mice tissues or cells using the Trizol reagent (Invitrogen, 15596-026). RT-PCR was performed using 0.1 μg of cDNA with a 2× Taq mastermix (CWBIO, cw0682) and 10 pmol/μl of primers. Reactions were run on a Bio-Rad MyCycler Thermal Cycler, and PCR products were analyzed by agarose gel electrophoresis. The primers used in this study are listed in Table S2.

BrdU Incorporation Assay

Cell proliferation was detected using bromodeoxyuridine (BrdU) (Sigma-Aldrich, B5002). Briefly, BrdU was dissolved in 0.9% sterile NaCl and filtered through a 0.22-μm filter. BrdU was then applied to pregnant female mice by intraperitoneal injection (50 mg/kg) 4 hr before euthanasia. The embryos were removed, and the tail from each embryo was collected for genotyping. The brain of each embryo was immediately fixed in 4% paraformaldehyde. Two days later, the fixed brains were transferred to a solution of 30% sucrose in PBS for cryopreservation. Cryostat sections of the brains were cut in the coronal plane at a thickness of 10 μm and collected on Superfrost Plus slides (VWR). Brain sections were incubated in 2 N HCl at 37°C for 30 min and rinsed in borate buffer (pH 8.5) for 15 min. The sections were then incubated with an anti-BrdU antibody (Bioworld, MB6004) at 4°C overnight after DNA denaturation.

For double labeling of BrdU and NESTIN or β-TUBULIN III, sections were incubated with an anti-BrdU antibody and the indicated antibodies for each cell marker at 4°C overnight after DNA denaturation. Immunofluorescence analysis was performed as described above.

Cell Culture and Treatments

Primary NSCs generated from E12.5 embryos from Eva1a+/+ and Eva1a−/− mice were cultured in Dulbecco's modified Eagle's medium (DMEM/F12, GIBCO, 42400-028) supplemented with 2% B27, 20 ng/ml basic fibroblast growth factor (bFGF) and epidermal growth factor, 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM glutamine. Cells were resuspended in DMEM/F12 medium, and then plated onto uncoated tissue culture dishes at 3.5 × 104 cells/cm2 and incubated at 37°C in 5% CO2.

To induce cell differentiation, neurospheres from cultures of less than ten passages were plated at a density of 105 cells/cm2 on coverslips coated with 15 mg/ml poly-ornithine (Sigma-Aldrich, P4957), supplemented with N2 (Gibco BRL, 17502048) containing 0.5 μM all-trans retinoic acid (Wako, R117), 0.5% fetal bovine serum without bFGF, and incubated for 6 days. The medium was replaced every 2 days.

To study the correlation between energy availability and neuronal differentiation, 10 mM methylpyruvate was added to the culture medium for 6 days. The culture medium was supplemented with 1.5 mg/ml sodium bicarbonate to maintain the pH balance in the presence of MP, and replaced every 24 hr. Cells were then fixed for immunofluorescence or lysed for western blotting.

To detect LC3B distribution, NSCs from Eva1a+/+ and Eva1a−/− at E12.5 were cultured under proliferative conditions and plated on glass coverslips coated with poly-ornithine for 24 hr. NSCs were subjected to immunofluorescence against LC3B and viewed under a confocal fluorescence microscopy.

The recombinant viral plasmid Ad5-Eva1a was prepared by SinoGenoMax. NSCs were infected by incubation with Ad5-Eva1a for 12–16 hr, and then fixed for immunofluorescence or lysed for western blotting.

Neurosphere Formation Assay

We collected NSCs on culture day 3 and dissociated them using a NeuroCult cell dissociation kit (Stem Cell Technologies, 05707), and cultured them in 96-well non-coated plates with 500 or 1,000 cells per well. Analyses were performed after 5–7 days in culture. Images of neurospheres (NSs) in 8 wells per experimental condition were acquired and analyzed using ImageJ software to quantify both the number of NSs and the diameter of individual NSs. All data presented are the results of three to four separate experiments.

Measurement of Neurite Length

To visualize neuronal morphology, cells were stained with anti-β-TUBULIN III antibodies, followed by incubation with Texas Red secondary antibody (Molecular Probes, T2767) or with Alexa 488 antibody (Molecular Probes, A1101). Cell nuclei were counterstained with Hoechst 33342. Neurites were classified as primary (predominant processes emerging directly from the cell body) or secondary (processes emerging from a primary neurite). Images of the cells were taken at 40× or 63× magnification using a confocal laser scanning microscope (Olympus). Thirty visual fields were chosen randomly and 30 individual images were taken from each group. Neurite length was measured using the NIH ImageJ software.

Apoptosis Detection

Apoptosis in tissue was detected using an In Situ Cell Death Detection Kit (Roche, 1168479591) by TUNEL assays. Brain frozen sections from both Eva1a−/− and Eva1a+/+ mice at E12.5 were treated and stained according to the manufacturer's directions. The nuclei were stained with Hoechst 33342 and imaged under a confocal microscope. The percentage of TUNEL-positive cells was calculated as the mean of ten different fields from each specimen.

Cell apoptosis was detected by flow cytometry. Primary NSCs generated from Eva1a−/− and Eva1a+/+ mice at E12.5 were cultured in Dulbecco's modified Eagle's medium for 3 days, trypsinized, and resuspended in 100 μl of binding buffer containing 5 μl of FITC-conjugated Annexin V, and incubated at room temperature in the dark for 30 min. The cells were stained with propidium iodide and analyzed on a FACS Calibur flow cytometer (Becton Dickinson).

Statistical Analysis

The statistical significance of the mean differences observed between samples was determined by the Student two-tailed t test using SPSS 18.0. One-way ANOVA with Tukey's post hoc test was used to compare means of more than three samples using SPSS 18.0. Data are shown as means ± SD of the results of at least three independent experiments. Values of p < 0.05 were considered significant.

Author Contributions

Conceptualization: Y.B. and Y.C; Methodology: M.L., G.L., J.H., and X.S.; Analysis: M.L., G.L., and X.S. Resources: J.J., Y.G., L.Q., and Y.X. Writing: Y.B., Y.C., and G.L.