METH powder (purity of 99.1%) was obtained from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China) and formulated into a 10 mM stock solution, stored at 2–8 °C, and diluted into 1.5 mM and 3 mM solutions for neurospheres treatment.

Sprague-Dawley rats were provided by the Experiment Animal Center of the Sichuan Academy of Chinese Medical Sciences. The rats were housed in an animal room under a standard condition with a 12 h light/dark cycle (light 7:00–19:00) at room temperature (20–25 °C) and relative humidity (40–70%). All experimental procedures and use of the animals were conducted in accordance with the guidelines established by the Association for Assessment and Accreditation of Laboratory Animal Care and the Institutional Animal Care and Use Committee of Sichuan University.

Neural stem/progenitor cell culture from the embryonic hippocampus of Sprague-Dawley rats was performed according to the procedure from StemCell Technologies with some modifications. Briefly, the hippocampus was dissected from E18 embryos and placed in cold PBS containing 2% glucose on ice. After completing precipitation, the supernatant was discarded and the hippocampal tissue was then resuspended with complete NeuroCult proliferation medium (NeuroCult™ Basal Medium supplemented with 10% NeuroCult™ Proliferation Supplement, 20 ng/mL EGF, 10 ng/mL bFGF, and 2 μg/mL Heparin). The dissected embryonic hippocampus was then mechanically dissociated into individual cells using the glass pipette with a fire-polished tip. After gentle titration 15–20 times, the single-cell suspension was filtered with a 40 μm cell strainer to remove the large cell debris. The single-cell suspension was centrifuged at 200 g for 5 min and cell pellets were resuspended with the complete NeuroCult proliferation medium. Lastly, the cell suspension was seeded on the ultra-low adherent 6-well plate with a cell density of 1 × 10⁵ cells per well and incubated in a humidified incubator with 5% CO₂ at 37°C.

Floating neurospheres were formed and pelleted by centrifuging at 200 g for 5 min after 2 days of culturing. Previous observations suggest a time/dose-dependent effect of METH on the brain (Davidson et al., 2022). Due to the timeline of neurosphere proliferation and differentiation, it was challenging to expose METH at low concentrations for a longer period in order to generate toxicity (Lenzi et al., 2022). Furthermore, according to the previous in vitro study, acute high doses of METH induced dopaminergic neuronal autophagy and apoptosis by increasing DNA damage-inducible transcript 4 (DDIT4) expression leading to neurotoxicity. However, such neurotoxic effects were not observed at low concentrations of METH (Li et al., 2017). Therefore, to explore the neurotoxicity induced by METH and verify the previously reported neurotoxicity mechanisms in our neurospheres model (Li et al., 2017), neurospheres were treated with a high concentration of METH at 1.5 mM and 3 mM (final concentration). After METH exposure for 24 h, neurospheres were collected for the subsequent immunoblotting, immunofluorescence, and differentiation analysis.

After METH treatment for 24 h, neurospheres were pelleted by centrifuging at 200 g for 5 min and resuspended with the complete NeuroCult differentiation medium (NeuroCult™ Basal Medium supplemented with 10% NeuroCult™ Differentiation Supplement). The single-cell suspension was pelleted by centrifuging at 200 g for 5 min and resuspended with the complete NeuroCult differentiation medium. Cells were then plated on a matrigel-coated coverslip in a 6-well plate with a density of 2 × 10⁵ cells per well and incubated in a humidified incubator with 5% CO₂ at 37°C. Cells were differentiated for 7 days, with medium change every 2 days.

Neurospheres from three biological replicates were collected for whole genome transcriptional analysis after METH treatment for 24 h. Total RNA was extracted using a miRNeasy Mini Kit (No. 160040742; QIAGEN) according to the manufacturer’s instructions, and 1 μg of total RNA was used to prepare RNA-seq transcriptome libraries using a TruSeq RNA Library Prep kit (No. RS-122-2001; Illumina). After quantifying the concentration with a TBS380 mini-fluorometer, the constructed cDNA libraries were sequenced as 2 × 150 bp read length with the Illumina HiSeq 4000 sequencer instrument. The sequencing reads were trimmed and quality-controlled using SeqPrep and Sickle with default parameters. Lastly, HISAT2 software was applied to perform directional alignment between clean reads and reference genomes, respectively (Kim et al., 2015).

The gene expression was calculated according to the transcripts per million (TPM) method. The abundance of genes was quantified by adopting RSEM (Li and Dewey, 2011). Genes with fold change (FC) > 2 and Padj ≤ 0.05 were considered to be significantly differentially expressed genes (DEGs) (Robinson et al., 2010; Wang et al., 2010; Love et al., 2014). The principal component analysis (PCA) and correlation, scatter, and hierarchical clustering heatmap analysis were performed on the Majorbio website (http://www.www.majorbio.com). In addition, gene ontology (GO) and KEGG enrichment analysis were performed to determine which DEGs were significantly enriched in GO terms and signaling pathways compared to the whole-transcriptome background at Bonferroni-corrected p-value ≤0.05. GO and KEGG enrichment analysis was carried out by Goatools and KEGG orthology-based annotation system (KOBAS), respectively (Xie et al., 2011). Protein-protein interaction networks were constructed from all DEGs using Cytoscape with 3.8.2 ClueGO/CluePedia plugin (Cox and Mann, 2008).

Neurospheres were harvested for RT-qPCR assay after METH treatment for 24 h. Total RNA was extracted from the neurospheres using the AxyPrepTM Multisource RNA Miniprep kit (Axygen) according to the manufacturer’s instructions. The single-stranded cDNA was reverse transcribed from extracted RNA with a PrimeScript™ RT reagent kit with gDNA Eraser. The qPCR reaction was performed with SYBR Green MasterMix by a QuantStudio Q5 quantitative PCR instrument (ThermoFisher). The expression of each gene was normalized to the expression of GAPDH and analyzed by using the ΔΔCT method. Primers used in this study were listed as follows: Ccna2 F: 5′-GAGGCAGCCA GACATCACTAACAG-3′, Ccna2 R: 5′-TTCACAGCCAAATGC AGGGTCTC-3’; Ccnb1 F: 5′-GTG​CCA​GTG​TGC​GAA​CCA​GAG-3′, Ccnb1 R: 5′- TGG​GCT​TGG​AGA​GGG​AGT​ATC​AAC-3’; Ccne2 F: 5′-TCA​AGA​CGC​AGT​AGC​CGT TT-3′, Ccne2 R: 5′- ATC​TGG​GCT​TCT​TGT​GGA​GAG-3’; Cdca7 F: 5′-GAGCT TCCCCGGCATATTCT-3′, Cdca7 R: 5′-CCG​TTC​TGG​GTT​CCT​TCT​GG-3’; Klf4 F: 5′-CTT​GCA​GCA​GTA​ACA​ACC​CG-3′, Klf4 R: 5′- CGA​TTC​CTG​GTG​GGA​TAG​CG-3’; Ccnf F: 5′-AGC​AAA​CGG​AAG​CGA​GA GAACAG-3′, Ccnf R: 5′- GTCCAG GCT​CCA​GTC​CAG​AAG​G-3’; Nurp1 F: 5′-GCC​ACC​AAC​AGC​CCA​CAC​TTC-3′, Nurp1 R: 5′- TTC​CGA​CCT​CCA​CCG​ACG​AC-3’; Notch1 F: 5′-TGC​CAG​GAC​CGT​GAC​AAC​TAC-3′, Notch1 R: 5′- GCTCGCAC AGTCATCCAGATTG-3’; GAPDH F: 5′-AAC​ATC​ATC​CCT​GCC​TCT​ACT​G-3′, GAPDH R: 5′- AACATCATCCCTGC CTCTACTG-3’; Hes1 F: 5′-TCC​AAG​CTG​GAG​AAG​GCA​GA-3′, Hes1 R: 5′-CTCGTTCATG CACTCGCTGA-3’; Hes5 F: 5′-TCG​CCA​ATC​GCC​TCC​AGA​G-3′, Hes5 R: 5′-GGTCCCGAC GCATCTTCTCC-3’; Hey1 F: 5′-GCT​GAA​GTT​GCC​CGT​TAT​CTG​AG-3′, Hey1 R: 5′-GCT​GGG​ATG​CGT​AGT​TGT​TGA​G-3’; Actin F: 5′-TAC​AAC​CTC​CTT​GCA​GCT​CC-3′, Actin R: 5′- GGA​TCT​TCA​TGA​GGT​AGT​CAG​TC-3’.

Total proteins were extracted by a whole protein extraction kit (No. BC3710; Solarbio) and protein concentration was measured by a BCA assay kit (No. P0010; Beyotime) according to the manufacturer’s instructions. Equal amounts of proteins were resolved on the 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF) membranes. After blocking with non-fat milk, the PVDF membrane was sequentially incubated with the primary and secondary antibodies. Primary antibodies and dilutions used in this study were listed in the following: rabbit anti-CDK2 (1:4000; 22060-1-AP; Proteintech), rabbit anti-CDK4 (1:8000; 11026-1-AP; Proteintech), rabbit anti-CDK6 (1:4000; 19117-1-AP; Proteintech), rabbit anti-Cyclin E1 (1:1000; ab33911; Abcam), mouse anti-Cyclin A (1:500; sc-271645; Santa Cruz Biotechnology), rabbit anti-ɑ-Tubulin (1:2000; #2125; Cell Signaling Technology).

Neurospheres were dissociated into single cells with Accutase™ cell detachment solution (No. 07920; StemCell Technologies). After washing three times with PBS, the dissociated single cells were fixed with 70% ice-cold ethanol overnight at 4 °C. Next, the fixed cells were washed three times with PBS and incubated with 50 μg/mL propidium iodide (PI) in PBS at room temperature for 30 min. The proportion of cells was analyzed by Cytomics FC500 flow cytometer (Beckman Coulter Ltd., United States), and data were analyzed using the CXP Analysis software version 2.2.

Neurospheres were permeabilized and blocked with a blocking solution consisting of 5% normal goat serum and 0.5% TritonX-100 in PBS at room temperature for 1 h. Neurospheres were incubated with the primary antibodies diluted in the blocking solution overnight at 4°C and then incubated with the Alexa-conjugated secondary antibodies at room temperature for 1 h. Neurospheres were further stained with 4′,6-diamidino-2-phenylindole (DAPI) for 5 min and subsequently mounted onto glass slides with ProLong™ Gold Antifade Mountant (No. P36934, ThermoFisher). Images were acquired on a laser scanning confocal microscope (Leica SP8 TCS) with built-in LAS X software (Leica Biosystem). Images were analyzed with ImageJ software (version 1.45). Primary antibodies and dilutions used in this study were listed in the following: rabbit anti-SOX2 (1:400; #3579; Cell Signaling Technology), mouse anti-Nestin (1:400; #4760; Cell Signaling Technology), rabbit anti-Ki67 (1:100; ab15580; BioLegend), mouse anti-β3-Tubulin (1:200; #4466; Cell Signaling Technology), mouse anti-GFAP (1:1000; ab7260; Abcam), rabbit anti-GFAP (1:5000; ab7260; Abcam), mouse anti-Tuj1 (1:1000; ab78078; Abcam).

The apoptosis was determined by using a terminal dexynucleotidyl transferase (TdT)-mediated dUTP nick end labeling (TUNEL) assay kit (No. 12156792910; Roche, Switzerland) according to the manufacturer’s instructions. Briefly, the reaction mixture of enzyme solution and labeling solution was freshly mixed at a ratio of 1:9 (v/v). Then, neurospheres were incubated with the reaction mixture for 1 h at room temperature in the dark. And all cell nuclei in neurospheres were subsequently stained with DAPI for 5 min at room temperature protecting from the light. After washing three times with PBS, neurospheres were mounted onto glass slides with ProLong™ Gold Antifade Mountant (No. P36934, ThermoFisher). Apoptosis was quantified by determining the percentage of TUNEL positive-stained (TUNEL+) cells to all DAPI-stained cells using ImageJ software (version 1.45).

Cellular superoxide anions were detected and quantified using dihydroethidium (DHE, D1168; Thermo Fisher Scientific) as a previously reported procedure (Wang et al., 2019). Briefly, neurospheres were dissociated into single cells with Accutase™ cell detachment solution (No. 07920; StemCell Technologies). After washing three times with PBS, the dissociated single cells were pellet and resuspended with prewarmed PBS containing 10 μM DHE. After incubating for 15 min in the dark, cells were precipitated and washed three times with PBS. Lastly, an equal number of cells (1 × 10⁵ cells/well) was added to the 96-well plate. The fluorescence intensity was measured using a Varioskan LUX multimode microplate reader (Thermo Fisher Scientific) with excitation and emission at 510 and 600 nm, respectively.

`Matrigel was used as a supportive matrix in the neurosphere culture. After neurosphere formation, neurospheres were plated on the 6-well plate pre-coated with Matrigel and cultured in neurobasal plus medium (No. A3582901, ThermoFisher) without growth factor. Neurospheres were then cultured in a humidified incubator with 5% CO₂ at 37°C. To track the migration of NPCs out of neurospheres, images were captured by phase contrast microscopy at 2, 4, and 6 h, respectively. The diameter of migration was measured by ImageJ software (version 1.45). The average outgrowth was quantified by subtracting the inner diameter from the outer diameter as previously reported (Ge et al., 2016).

Data are presented as the mean ± SEM. Statistical analysis was carried out through one-way analysis of variance (ANOVA) followed by the Tukey–Kramer test using GraphPad Prism 8 software (version 8). p < 0.05 was considered statistically significant. The number of neurospheres or biological replicates that were used for the statistical analysis was presented in the corresponding figure legend of the individual figure.

Neural stem and progenitor cells (NSCs/NPCs) were isolated from rat hippocampal tissues of embryonic day 18 (E18), and cultured in the proliferation medium. The floating growing neurospheres appeared after 1–2 days of culturing. On day 2, neurospheres were treated with METH at the final concentration of 1.5 mM and 3 mM, respectively (Li et al., 2017). After METH treatment for 24 h, neurospheres were collected and re-suspended with the differentiation medium without METH, followed by another 7 days for differentiation (Figure 1A). One of the distinguishing features between NSCs and NPCs is the capacity of self-renew and pluripotent differentiation (Hu et al., 2006). To determine these functional properties, SOX2, a well-characterized transcription factor for NSC maintenance and brain development (Zhu et al., 2017; Mercurio et al., 2019), and Nestin, a marker of NSCs/NPCs and radial glial (Onishi et al., 2019), were used to detect the self-renew and pluripotent ability of neurospheres. The immunofluorescent images presented that both SOX2 and Nestin were highly expressed in the neurospheres (Figure 1B), suggesting the high proliferation and differentiation capacity of neurospheres. In contrast to neurospheres in the control group, METH treatment caused a decreased size of neurospheres with increased cell fragments around the neurospheres. The diameter of neurospheres was decreased with the increase in METH concentration (Figure 1C). These results indicate that METH exposure caused the reduced size of the neurosphere.

To explore the molecular mechanisms of reduced neurospheres caused by METH, the whole-genome transcriptional profiles among the control and METH-treated groups were compared using RNA sequencing (RNA-seq). The sample correlation analysis was conducted based on the transcripts per million (TPM) distribution. The Pearson correlation analysis of three biological replicates was almost equal to 1 in each group, indicating the reliability of our RNA-seq results. Furthermore, correlation analysis showed that the differences between the control and METH-treated groups were augmented with the increase in METH concentration (Figure 1D). Similar result was obtained from the unsupervised principal component analysis (PCA) (Figure 1E). Next, the differentially expressed genes (DEGs) between the METH-treated and control groups were separately analyzed. The scatter plots show all the detected genes, with the significantly upregulated and downregulated genes (fold change (FC) > 2 and Padj < 0.05) in blue and red, respectively. Compared to the control group, 868 and 2,247 genes were significantly upregulated in the 1.5 mM and 3 mM METH-treated groups, respectively; nevertheless, 864 and 2036 genes were significantly downregulated, respectively (Figures 1F, G; Supplementary Table S1).

Those DEGs that were affected in both METH-treated groups were analyzed. Venn graph showed that 1570 DEGs were dysregulated in both METH-treated groups (Figure 1H; Supplementary Table S1). Among these common DEGs, 781 genes were downregulated and 789 genes were upregulated following METH treatment (Figure 1I; Supplementary Table S1). The expression level of those common DEGs was visualized by heatmap. The results presented a similar expression profile in the METH-treated groups while a distinct difference with the control group. The expression of most DEGs changed gradually in a dose-dependent manner (Figure 1J; Supplementary Table S1). These results suggest that METH can induce neurodevelopmental toxicity dose-dependently.

To investigate the disturbed cell signaling in response to METH, all DEGs were subjected to gene ontology (GO) enrichment and KEGG analysis, respectively. GO enrichment analysis discovered that several biological processes were significantly enriched in both 1.5 and 3 mM METH groups. As shown in Figures 2A, B, several biological processes involved in DNA replication, cell cycle, cell migration, and cell differentiation were found to be affected by METH. KEGG analysis based on these changed biological processes further revealed that several cell signaling were enriched in both METH-treated groups, such as cellular processes (e.g., cell cycle, p53 signaling pathway, lysosome), genetic information processing (e.g., DNA replication, mismatch repair, homologous recombination), and regulating organismal systems (e.g., axon guidance, glutamatergic synapse, GABAergic synapse) (Figures 2C, D). Notably, cell cycle and DNA replication were the two most affected cell signalings that were enriched. In agreement with previous studies, cell cycle-related signaling pathways displayed significantly high variations following the METH treatment (Jackson et al., 2014). These results indicate the reduced size of neurospheres caused METH may be attributed to the disturbed cell cycle.

How those signaling pathways were disturbed by METH was further addressed. The cell signaling involved in DEGs was visualized via ClueGO/CluePedia plugin from Software Cytoscape (version 3.8.2). Although a similar number of upregulated and downregulated DEGs were identified in the METH-treated groups, the downregulated DEGs dominated the enriched signaling pathways (Figure 2E). Furthermore, almost 80% of downregulated DGEs-enriched pathways were involved in the cell cycle. As shown in the pie chart, the top three downregulated DGEs-enriched pathways were cell cycle, mitotic (40%), and cell cycle (12.75%); moreover, Orc1 is phosphorylated by cyclin A/CDK1 (5.8%). In contrast to the downregulated DGEs, the upregulated DGEs-enriched pathways were diverse, and the top three enriched pathways are involved in the regulation of multicellular organismal processes (21.43%), anatomical structure formation (18.57%), and establishment of localization (10.0%) (Figure 2F). Taken together, these findings suggest that METH exposure may cause cell cycle arrest during the proliferation period of NSCs/NPCs.

The cell cycle is a series of precisely timed and carefully regulated stages of growth, DNA replication, and division into two genetically identical cells including two major phases interphase and mitosis (Lubischer, 2007). We then look into which stage of the cell cycle was arrested by METH. Firstly, those DEGs implicated in the cell cycle were investigated. The expression profile of cell cycle-involved genes was presented in the heatmap. The majority of those DEGs were actually downregulated upon METH exposure (Figure 3A). Then, the altered level of several DEGs that are critical for cell cycle progression was validated through RT-qPCR analysis (Figure 3B). Compared to the control group, genes of cyclin A2 (Ccna2), cyclin B1 (Ccnb1), cyclin E2 (Ccne2), cyclin F (Ccnf), Notch receptor 1 (Notch1), and cell division cycle-associated 7 (Cdca7) were significantly downregulated by METH, while the expression of KLF transcription factor 4 (Klf4) and nuclear protein 1(Nupr1), which are important for cell cycle progression (Choi and Roh, 2018; Li et al., 2020), were significantly upregulated. These results indicated that METH caused an aberrant cell cycle progression of neurospheres.

Moreover, the effect of METH on the cell cycle distribution of neurospheres was analyzed by flow cytometry with propidium iodide (PI) staining. The results showed that compared to the control, METH exposure resulted in a reduction in the percentage of cells in the S and G2/M phases but an increase in the G0/G1 phase (Figures 3C, D), indicating that METH exposure caused a cell cycle arrest at G0/G1 phase. We further measured the expression levels of various cyclin proteins and cyclin-dependent kinases that control the progression of cells by regulating the cell cycle. The results showed that the levels of the proteins involved in the G1/S transition were significantly reduced by METH, such as CDK2, Cyclin E1, CDK4/6, and the G2/M transition, Cyclin A (Figures 3E, F). Similar to a previous finding that METH protracts G1/S phase transition in the T cells (Potula et al., 2018), cell cycle arrest caused by METH may also be associated with the impairment of G1/S transition. Lastly, the cell proliferation capacity following METH exposure was detected by immunostaining Ki67, a marker for cell proliferation protein. By counting the number of Ki67-positive (Ki67+) cells and DAPI-stained cells, we found that the percentage of Ki67+ cells was diminished in the METH-treated neurospheres compared with control neurospheres, suggesting an attenuated proliferation capacity of neurospheres (Figures 3G, H). Collectively, these results suggest that METH causes cell cycle arrest by impairing G1/S transition, resulting in lower self-renew and cell division.

The effect of METH on the cell differentiation of neurospheres was examined. The profile of those DEGs enriched in the GO term was displayed in the heatmap. The change level of the involved genes was presented in a dose-dependent manner (Figure 4A). GO enrichment analysis based on those DEGs showed a number of neural differentiation pathways enriched, such as regulation of dendritic spine morphogenesis, axon guidance, and positive regulation of cell differentiation. Among these GO terms, positive regulation of neural crest cell differentiation, positive regulation of dendritic spine morphogenesis, and regulation of cell morphogenesis involved in differentiation were the top three enriched GO terms (Figure 4B). To explore the relationship of DEGs between cell cycle and cell differentiation, a protein-protein interaction (PPI) analysis was conducted using the STRING tool (https://stringdb.org/) and Cytoscape software (Cytoscape 3.8.2) (Szklarczyk et al., 2017). The interaction network presented that cell cycle-related DEGs, Notch1, klf4, and Cdk2 were co-expressed with cell differentiation-related DEGs, e.g., Notch3, Runx1, Wnt3 (Figure 4C). The decreased Notch1 and Notch3 suggest the inhibition of Notch signaling, a pathway playing a central role in various aspects of development and cellular differentiation, for instance, the maintenance of the undifferentiated state and cell fate decision (Yoon and Gaiano, 2005; Harada et al., 2021). Analysis of the downstream genes of Notch signaling, Hes1, Hes5, and Hey1 also found reduced expression upon METH treatment (Supplementary Figure S1A). Collectively, these results suggest that METH exposure may accelerate cell cycle exit and then cause abnormal cell differentiation.

To investigate the effect of METH on cell differentiation, the morphology of neurons and astrocytes was visualized after differentiation culturing on day 10. For Tuj1-stained differentiating and mature neurons, the dendrites of METH-treated neurons were obviously longer than those of the control group. Additionally, the complexity of neurons was also altered by METH. Compared to the multiple dendrites from the soma in the control group, METH markedly decreased the dendrites number from the soma but increased the branches from the primary dendrites (Figure 4D). To quantify changes in neuronal morphology, we traced the dendrites of individual-stained neurons and performed Sholl analysis to assess dendritic arborization. The line profile displayed that the dendritic branching complexity was indeed changed by METH, exhibiting reduced branches within 60 μm distance area from the soma while increased branches beyond that area (60–180 μm distance area from the soma) (Figures 4E, F). Similarly, the total dendritic length was also extended by METH (Figure 4G). These results indicate that METH promoted abnormal neuronal differentiation. Except for the aberrant differentiation of neurons, we also noticed the changes in the morphology of GFAP-labeled astrocytes. Normally, the astrocytes in the cultured neurospheres present typical star-shaped morphology with large round somata extending thick and short processes that branch out into several small processes. Nevertheless, METH treatment led to an increase in the astrocytes with more elongated and thin processes (Figure 4D). Then, astrocytes were classified into three groups based on their processes: those with no branches, those with long branches below 3, and those with long branches above 3 (Supplementary Figure S1B). The quantified results showed that the percentage of astrocytes with long processes above 3 increased from 0.5% in the control neurospheres to 10% in the 3 mM METH-treated neurospheres (Figure 4H). According to the subpopulation of astrocytes in the CNS (Matyash and Kettenmann, 2010; Köhler et al., 2021; Köhler et al., 2023), the increased ratio of astrocytes with elongated branches upon METH exposure suggests the possibility of decreased protoplasmic astrocytes while with increased fibrous astrocytes in METH-treated neurospheres. Taken together, METH can cause an abnormal development of neurospheres, which may be attributed to the aberrant cell cycle exit and cell differentiation.

NSCs/NPCs migrate to the site of brain development or injury, where they undergo differentiation into specific cell types, such as neuron and glial cells (Csobonyeiova et al., 2019; Akter et al., 2021). To investigate whether the abnormal astrocyte morphology resulted from the altered cell migration in neurospheres, we analyzed those DEGs enriched in the GO term, called regulation of cell migration. The heatmap presented that most of those DEGs involved in cell migration underwent extensive changes following METH treatment (Figure 5A). GO analysis showed a number of migration-related pathways enriched, such as neuron migration, cell adhesion, and cellular component movement (Figure 5B). Moreover, the relationship of DEGs between cell cycle and cell migration was explored using the PPI analysis. The results showed that cell cycle-related DEGs, Notch1, Klf4, and Cdk2 formed an interaction network with cell migration-related DEGs, such as Wnt7a, Pdgfb, Vegfa (Figure 5C), further demonstrating that METH-dysregulated cell cycle exit was accompanied with aberrant cell migration.

To verify whether METH exposure resulted in changes in cell migration of neurospheres, we applied an adherent culture system to investigate the effect of METH on cell migration capacity. After plating neurospheres on Matrigel-coated plates, the distance of cells migrated out of neurospheres was traced at 2 h, 4 h, and 6 h post-plating, respectively. The phase contrast images presented the diminished cell migration distance in the METH group (Figure 5D). We then separately measured the inner diameter and outer diameter, and the average outgrowth was obtained by subtracting the inner diameter from the outer diameter. The decreased distance of outgrowth indicated that METH attenuated the migration ability of NSCs/NPCs out of neurospheres (Figure 5E). Taken together, METH exposure attenuated the migration of NSCs/NPCs to the target site of brain development, which may attribute to the abnormal cell differentiation.

Previous study has found that high doses of METH treatment caused dopaminergic neuronal autophagy and apoptosis (Li et al., 2017). In this study, neurospheres were also subjected to high doses of METH for 24 h. Hence, to systematically assess whether the reported neurotoxicity occurred in the neurospheres, those DEGs involved in cell death, including the P53 signaling pathway, lysosome, and cellular senescence, were further analyzed. The heatmap presented that most of those DEGs involved in cell death were upregulated in a dose-dependent manner following METH treatment (Figure 6A), indicating the progress of cell death. GO enrichment analysis using those DEGs enriched a number of GO terms involving oxidative stress and DNA damage. The enrichment of those GO terms, such as response to hypoxia, lysosome organization, and cellular response to oxygen level, were supposed to be associated with apoptotic cell death (Figure 6B). Moreover, the PPI analysis found the interaction of diverse cell cycle involved DEGs (such as Ccna2, Cdkn1a, Ccnf, and Ccne2) with the Ddit4 (Figure 6C), a gene involved in hypoxia-triggered apoptosis that is significant upregulated in METH-induced neurotoxicity (Li et al., 2017). The occurrence of the previously reported METH-induced neurotoxicity in neurospheres supports the validity of the toxic mechanism of cell cycle exit and abnormal differentiation.

Lastly, to determine whether METH exposure led to oxidative stress and apoptosis in neurospheres, the cellular oxidative stress and cell death of neurospheres were separately evaluated. Analysis of cellular superoxide anions using a fluorescent probe DHE found that METH exposure resulted in increased cellular oxidative stress (Figure 6D). The apoptotic cell death was determined using the TUNEL assay, a commonly used method to detect apoptosis in situ. In contrast to a very low number of TUNEL-positive (TUNEL+) cells in the control neurospheres, the percentage of TUNEL+ cells was evidently increased in a dose-dependent manner upon METH treatment (Figures 6E, F). Collectively, these results indicate that METH causes cell cycle arrest, abnormal cell differentiation, and attenuated cell migration eventually accompanied by increased oxidative stress and apoptotic cell death.