Identification of Novel NRF2-Dependent Genes as Regulators of Lead and Arsenic Toxicity in Neural Progenitor Cells

Abstract

Lead (Pb) and arsenic (As) are prevalent metal contaminants in the environment. Exposures to these metals are associated with impaired neuronal functions and adverse effects on neurodevelopment in children. However, the molecular mechanisms by which Pb and As impair neuronal functions remain poorly understood. Here, we identified F2RL2, TRIM16L, and PANX2 as novel targets of Nuclear factor erythroid 2-related factor 2 (NRF2)—the master transcriptional factor for the oxidative stress response—that are commonly upregulated with both Pb and As in human neural progenitor cells (NPCs). Using a ChIP (Chromatin immunoprecipitation)-qPCR assay, we showed that NRF2 directly binds to the promoter region of F2RL2, TRIM16L, and PANX2 to regulate expression of these genes. We demonstrated that F2RL2, PANX2, and TRIM16L have differential effects on cell death, proliferation, and differentiation of NPCs in both the presence and absence of metal exposures, highlighting their roles in regulating NPC function. Furthermore, the analyses of the transcriptomic data on NPCs derived from autism spectrum disorder (ASD) patients revealed that dysregulation of F2RL2, TRIM16L, and PANX2 was associated with ASD genetic backgrounds and ASD risk genes. Our findings revealed that Pb and As induce a shared NRF2-dependent transcriptional response in NPCs and identified novel genes regulating NPC function. While further in vivo studies are warranted, this study provides a novel mechanism linking metal exposures to NPC function and identifies potential genes of interest in the context of neurodevelopment.

Graphical Abstract

1. Introduction

Exposures to common environmental metal contaminants, such as lead (Pb) and arsenic (As), have been associated with impaired neuronal functioning and neurodevelopment in children. Children in low income countries may be particularly at risk, given their potentially high exposure to various metals (Ericson et al., 2021). Among these metals, Pb has been by far the most actively studied. In 2012, the Centers for Disease Control and Prevention (CDC) started using a blood lead reference value (BLRV) to identify children with elevated blood Pb levels. The current CDC BLRV is 3.5 μg/dL (CDC, 2021). Despite notable declines in population exposures to Pb over time, approximately 632 million children globally (age =<18 yr) are estimated to have blood Pb levels at or above 5 μg/dL (Ericson et al., 2021). The U.S. Environmental Protection Agency (EPA) has determined that Pb exposure in children is associated with impaired cognitive and behavioral development (Bellinger et al., 1987; Bellinger, 2013; Canfield et al., 2003; Dickerson et al., 2015; Froehlich et al., 2009; Goodlad et al., 2013; McLaine et al., 2013; Min et al., 2009; Miranda et al., 2009; Needleman et al., 1996; Opler et al., 2004; Opler et al., 2008; Toscano and Guilarte, 2005). As is also a prevalent and well-recognized environmental toxicant (ATSDR, 2017), however, As-mediated neurotoxicity has been significantly understudied compared to Pb. It is estimated that between 94 and 220 million people are at risk of exposure to elevated As concentrations above the World Health Organization (WHO) provisional guideline value of 10 μg/L in groundwater (Podgorski and Berg, 2020). Some epidemiology studies have reported associations between As exposure and neurocognitive, psychological, and behavioral deficits in children (Calderon et al., 2001; Hamadani et al., 2011; O'Bryant et al., 2011; Rocha-Amador et al., 2007; Rosado et al., 2007; Tsai et al., 2003; von Ehrenstein et al., 2007; Wang et al., 2019; Wang et al., 2007; Wasserman et al., 2007; Wasserman et al., 2011). Decrements in neurobehavioral and cognitive functions following As exposure have also been reported in animal studies (Luo et al., 2009; Rai et al., 2010; Rodriguez et al., 2002; Xi et al., 2010; Xi et al., 2009). Despite evidence from epidemiological as well as animal and cell-culture studies that show Pb and As can alter neural functioning, the molecular mechanisms remain poorly defined.

The current understanding of metal-induced neurotoxicity primarily centers around neuronal effects. Alkondon et al. (Alkondon et al., 1990) and Guilarte et al. (Guilarte and Miceli, 1992) have demonstrated that Pb exerts an inhibitory effect on the N-methyl-D-aspartate (NMDA) receptor, which is involved in in brain development, synaptic plasticity, learning, and memory. Additionally, Pb impedes the release of brain-derived neurotrophic factor from vesicles and hinders the subsequent activation of tropomyosin-related kinase B in presynaptic neurons (Neal et al., 2011; Stansfield et al., 2012). As exposure induced neuronal cell death, DNA degradation and inhibited neurite growth in cultured rat or mouse neuronal cells (Aung et al., 2013; Namgung and Xia, 2001). As also affects levels of neurotransmitters such as dopamine, serotonin, and their metabolites as well as expression of NMDA receptor in vivo (Luo et al., 2009).

Despite previous research, there is still a lack of comprehensive understanding regarding the distinctive vulnerability of early brain development to metal exposure. The potential susceptibility of early brain development to metal exposure may be partially elucidated by the impact of metals on neural progenitor cells (NPCs). NPCs, serving as the precursors of both glial and neuronal cell types in the central nervous system (CNS), play a crucial role in the development of the brain (Martinez-Cerdeno and Noctor, 2018). Consequently, the functions of NPCs, including proliferation and differentiation, are likely associated with neurodevelopmental processes in children (Li et al., 2020; Vaccarino et al., 2001). Indeed, previous studies have shown that Pb and As affect NPC proliferation, neurogenesis, and gene expression both in vitro and in vivo (Bain et al., 2016; Breier et al., 2008; Gilbert et al., 2005; Hong and Bain, 2012; Liu and Bain, 2014; Liu et al., 2012; Ora et al., 2000; Sanchez-Martin et al., 2013; Schneider et al., 2005; Senut et al., 2014; Tyler and Allan, 2013; Tyler et al., 2017; Tyler et al., 2015; Verina et al., 2007; Wang et al., 2010).

Despite the known effects of Pb and As in NPCs, the underlying molecular mechanisms remain poorly understood, and moreover, whether such effects might be linked to impaired neurodevelopment in children is not known. In this study, we performed global transcriptional profiling to assess the impact of Pb and As exposure on human NPCs. We identified and characterized F2RL2, TRIM16L, and PANX2 as novel, direct Nuclear factor erythroid 2-related factor 2 (NRF2) targets and determined their roles in the regulation of NPC proliferation. We further tested whether dysregulation of the genes is associated with neurodevelopmental outcomes such as autism spectrum disorder (ASD), which has been associated with Pb and As exposure (Dickerson et al., 2015; Wang et al., 2019), by analyzing the transcriptomic data on human NPCs derived from ASD patients. By integrating global gene expression profiling with mechanistic studies, our study reveals a potential mechanistic link between metal exposure and neurodevelopmental outcomes.

2. Materials and Methods

2.1. Cells and reagents

ReNcell Cx human neural progenitor cells derived from the cortical region of human fetal brain were purchased from Millipore Sigma (SCC007) and cultured according to the supplier’s protocol. An aqueous solution of 1 mM Pb acetate (cat. no. 316512; Sigma Aldrich) stock or 1 mM sodium arsenite (cat.no. S7400; Sigma Aldrich) was used in all experiments. Transfection of siRNAs was performed with Lipofectamin RNAiMax transfection reagent (ThermoFisher) following the manufacturer’s protocol. siRNAs including nontargeting control (SIC001), si-NRF2-1 (SASI_Hs01_00182393), si-NRF2-2 (SASI_Hs02_00341015), and si-KEAP1 (SASI_Hs01_00080908) were obtained from Sigma. siRNAs including Silencer^™ Select Negative Control No. 1 siRNA (Catalog number:4390843), si-F2RL2 (Catalog number:4390824, siRNA ID:s4929), si-TRIM16L(Catalog number:4392420, siRNA ID:s44970), and si-PANX2 (Catalog number:4392420, siRNA ID:s229443) were purchased from ThermoFisher. Transfection of plasmids was performed with TransfeX (ATCC) following the manufacturer’s protocol. Plasmids include pF2RL2 (Addgene, Plasmid #51877), pPANX2 (Addgene, Plasmid #161697), pTRIM16L (GenScript, OHu31231).

2.2. Cell culture, exposure, siRNA transfection

ReNcell CX cells were exposed to vehicle control (H₂O), 1 μM Pb or 1 μM As for 24 h and total RNA for RNA-seq were isolated using RNeasy kit (Qiagen). For validation of RNA-seq data, cells were exposed to vehicle control (H₂O), Pb (1 or 2 μM) or As (1 or 2 μM) for 24 or 48 h, then total RNA or cell lysates were harvested for qRT-PCR (24 h) or western blotting (48 h). For siRNA transfection experiment, cells were transfected with nontargeting control or siRNA overnight, then total RNA were collected for qRT-PCR.

2.3. RNA-seq Library Preparation and Sequencing

Poly-adenylated RNA species were isolated from 1 μg of total RNA and converted to a cDNA library for RNA sequencing using the TruSeq^® RNA v2 kit (Illumina). Sample preparation involves isolating poly-adenylated RNA, RNA fragmentation, cDNA synthesis, ligation of adapters, PCR amplification using DNA barcodes, and library validation and quantification. Four samples were multiplexed into a single lane of the Illumina HiSeq 2000 for paired-end reads of 100 bp. Sequencing was performed at the Bauer Core Illumina Sequencing Facility (FAS Center for Systems Biology, Cambridge, MA).

2.4. RNA-Seq Data Analysis

Taffeta scripts (https://github.com/HimesGroup/taffeta) were used to analyze the RNA-Seq data, which included using FastQC (Andrews) (v.0.11.7) to obtain overall QC metrics. Reads for each sample were aligned with STAR (v. 2.5.2b) to the reference Homo sapiens build 38 UCSC file (hg38) genome obtained from the Illumina, Inc. iGenomes resource (Dobin et al.). Additional QC parameters were obtained to assess whether reads were appropriately mapped. Bamtools (v.2.3.0) (Barnett et al.) was used to count/summarize the number of mapped reads, including junction spanning reads. The Picard Tools (v.1.96; http://picard.sourceforge.net) RnaSeqMetrics function was used to compute the number of bases assigned to various classes of RNA, according to the hg38 refFlat file available as a UCSC Genome Table. For each sample, HTSeq (v.0.6.1) was used to quantify genes based on reads that mapped to the provided hg38 reference files (Anders et al.). The DESeq2 R package (v. 1.18.1) was used to measure significance of differentially expressed genes between the exposed and control samples and create plots of the results (Love et al.). The reported adjusted p-values are false-discovery rate corrected to 5% according to the procedure in DESeq2 that accounts for the large number of comparisons made. An adjusted p-value <0.05 was considered significant. The NIH Database for Annotation, Visualization and Integrated Discovery (DAVID) was used to perform gene functional annotation clustering using Homo Sapiens as background, and default options and annotation categories (Disease: OMIM_DISEASE; Functional Categories: UP_KW_BIOLOGICAL_PROCESS, UP_KW_CELLULAR_COMPONENT, UP_KW_MOLECULAR_FUNCTION, UP_KW_PTM, UP_SEQ_FEATURE; Gene_Ontology: GOTERM_BP_DIRECT, GOTERM_CC_DIRECT, GOTERM_MF_DIRECT;Interactions: UP_KW_LIGAND; Pathway: BIOCARTA, KEGG_PATHWAY, REACTOME_PATHWAY; Protein_Domains: INTERPRO, PIR_SUPERFAMILY, SMART, UP_KW_DOMAIN).

2.5. qRT-PCR

RNA was extracted and purified from cell cultures using the RNeasy Mini Kit (Qiagen) and reverse transcribed using the iScript cDNA Synthesis Kit (Bio-Rad). The resulting cDNA was amplified using iTaq Universal SYBR Green Supermix (Bio-Rad), 250 nM forward primer, and 250 nM reverse primer in a CFX96 Touch thermocycler (Bio-Rad). Fold changes were calculated using the ΔΔC_t method with β-actin as the housekeeping gene used for normalization. All primer sequences used in this study are listed in Supplementary Table S1.

2.6. Western blot

Whole-cell lysates were prepared in NP40 buffer supplemented with 1x Halt Protease Inhibitor Cocktail (Thermo) and 1 mM PMSF. Cleared lysates were denatured with 1x Laemmli Sample Buffer (Bio-Rad) and beta-mercaptoethanol as a reducing agent, and heated at 95°C for 10 minutes before being run in SDS-PAGE and transferred to a PVDF blotting membrane. Primary antibodies used in this study include anti-F2RL2 (Aviva), anti-TRIM16L (Invitrogen), anti-PANX2 (Aviva), and anti- β-actin (BioLegend).

2.7. Identification of NRF2 binding sites using ConTra V3

To identify putative antioxidant response elements (AREs) in the promoter regions of F2RL2, TRIM16L, and PANX2, we utilized ConTra v3 web server, which analyze any genomic region of interest to detect transcription factor binding sites across species (Kreft et al., 2017). Here are the settings for the analyses in the present study-Type of analysis: visualization; Reference organism: Human; Sequence parts: promoter 1000bp; Transcription factors: NFE2L2 Homo_sapiens,M6360_1.02,M6360_1.02 and Nfe2l2 JASPAR_CORE_2016,MA0150.2,MA0150.2; Stringency: core=0.85, similarity matrix=0.70. The ConTra v3 web server is freely available at http://bioit2.irc.ugent.be/contra/v3.

2.8. Chromatin Immunoprecipitation (ChIP) Assay

The SimpleChIP^® Plus Enzymatic Chromatin IP Kit (cat.no.9005, Cell Signaling Technology) was utilized to detect endogenous levels of NRF2-DNA interactions in ReNcell CX cells. Briefly, 4 million ReNcell CX cells were plated in a 10-cm dish per immunoprecipitation. After 24-h incubation, NRF2 was cross-linked to DNA using 1% formaldehyde. Nuclei were collected and Chromatin DNA was digested with micrococcal nuclease for 20 min into small fragments (150–900 bp). Nuclear pellet was resuspended in ChIP Butter and sonicated for 20 sec three times (Sonics Vibra-Cell). Digested, Crossed linked chromatin was incubated with Normal Rabbit IgG (negative control) or NRF2 antibody (cat. no. 12721; Cell Signaling Technology) overnight at 4°C with rotation. Antibody-bound complexes were captured by ChIP-Grade Protein G Magnetic Beads. Bound DNA was purified and underwent quantitation by PCR using primers for putative F2RL2 ARE, TRIM16L ARE, PANX2 ARE, NQO1 ARE (Positive control) and RPL30-exon 3 (Negative control, Cell Signaling Technology).

2.9. Measurement of Reactive Oxygen Species (ROS)

ReNcell Cx cells were transfected with negative control siRNA, si-F2RL2, si-PANX2, or si-TRIM16L overnight. Alternatively, cells were transfected with blank vector, pF2RL2, pPANX2, or pTRIM16L overnight. pcDNA GFP was cotransfected to confirm the transfection efficiency of the plasmids (Supplementary Figure S1). The cells were then plated in 96-well plates at a density of 20,000 cells per well and exposed to the vehicle control (H₂O), Pb (5 μM), or As (5 μM) for 24 hours. ROS production was measured using 2',7'-dichlorodihydrofluorescein diacetate (H2DCFDA, Invitrogen). DCF fluorescence intensity was measured by Celigo S (Nexcelom) and normalized to the number of nuclei (Hoechst 33342 Fluorescent Stain, ThermoFisher).

2.10. Measurement of Cell Number, Cell Proliferation, and Cell Death

ReNcell CX cells were transfected overnight with either negative control siRNA, si-F2RL2, si-PANX2, or si-TRIM16L. Alternatively, cells were transfected with blank vector, pF2RL2, pPANX2, or pTRIM16L overnight. After transfection, the cells were plated in 96-well plates at a density of 10,000 cells per well and exposed to either the vehicle control (H2O), Pb (2 μM), or As (2 μM) for 48 hours. Cell numbers were determined using the CyQUANT^® Cell Proliferation Assay (Invitrogen, C7026), following the manufacturer's protocol. Cell proliferation was assessed by EdU incorporation, a measure of DNA synthesis, using the Click-iT^™ EdU Proliferation Assay (Invitrogen), following the manufacturer's instructions. Cell death was quantified by Propidium Iodide (PI) staining, which binds to DNA in dead cells, as previously described. (Cohen et al., 2021).

2.11. Differentiation of NPCs

After transfection with siRNA or plasmids, ReNcell CX cells were plated in 12-well plates at a density of 100,000 cells per well and incubated overnight. Then, the culture medium was replaced with medium lacking EGF and FGF to initiate differentiation (Day 0) and exposed to vehicle control (H₂O), Pb (2 μM), or As (2 μM). The culture medium was replaced every two days. On Day 0 and Day 7, cells were collected to measure the expression of NPC, neuronal, or non-neuronal markers by qRT-PCR or immunofluorescence staining

2.12. Immunofluorescence (IF) staining

IF staining of NPCs was performed following the protocols from a previous study (Cohen et al., 2021). Cells were transfected with siRNA and then replated on coverslips coated with laminin and poly-L-lysine the next day. The cells were allowed to attach overnight, and subsets of coverslips were fixed in 4% PFA prior to differentiation (Day 0). The remaining coverslips were transferred to EGF/FGF-free media containing either the vehicle (H2O), Pb (2 μM), or As (2 μM). They were then differentiated and fixed on Day 0 or Day 7. The fixed cells were stained for either the neuronal markers TUBB3 (Mouse, 1:100; Fisher, PIMA1118X) and MAP1B (Rabbit, 1:100, Fisher, NBP304801) or the glial and oligodendrocyte markers GFAP (Mouse, 1:100, Fisher, NBP234365F) and PLP1 (Rabbit, 1:100, Fisher, NBP1600720). Anti-mouse (PIA32723TR, 1:2000 ) or anti-rabbit (PIA32733TR,1:2000) secondary antibodies were purchased from Invitrogen. Additionally, cells were co-stained with wheat germ agglutinin (WGA) and DAPI to label membranes and nuclei. The imaging was conducted using a Zeiss Axio Examiner. The Zeiss Zen Blue software package was employed to determine the staining intensities of all markers. The average staining intensity of each marker was divided by the staining intensity of WGA to normalize for differences in signal strength between regions.

2.13. Gene expression analysis of transcriptomics datasets associated with autism spectrum disorder

To further establish the relevance of our findings to neurodevelopmental outcomes, specifically autism spectrum disorder (ASD), we conducted an analysis utilizing publicly available transcriptomic datasets from Gene Expression Omnibus (accession ID: GSE214323 and GSE221923). Fu et al. (Fu et al., 2023) performed the original study, where they obtained human induced pluripotent stem cells (iPSCs) from both healthy control individuals and ASD patients carrying PTEN (PTEN WT/I135L, ASD line 1) or CTNNB1 (CTNNB1 p.Gln76*, ASD line 2) variants (Figure 6A). They generated panels of iPSC lines, including PTEN wild-type/I135L, PTEN knockout/knockout, PTEN wild-type/wild-type, and CTNNB1 wild-type/wild-type, using CRISPR-Cas9 gene editing. Subsequently, iPSC lines were differentiated into NPCs, and the transcriptomes of each line were profiled using RNA-seq. In our study, we reanalyzed these datasets to investigate the differential regulation of F2RL2, PANX2, and TRIM16L expression in ASD NPCs compared to control NPCs.

2.11. Statistical analysis

Statistical analysis was performed with GraphPad Prism version 10 (La Jolla, CA 92037, USA). Data were analyzed by Two-way ANOVA, one-way ANOVA or t-tests. A p-value < 0.05 was considered statistically different. Data were expressed as means ± SEM.

3. Results

3.1. Transcriptomic profiling of ReNcell CX human neural progenitor cells exposed to Pb or As

To better understand the effects of environmental metal toxicants Pb and As on NPCs, we performed global transcriptomic profiling in a human NPC cell line, ReNcell Cx exposed to Pb or As (Figure 1A). The concentration of Pb (1 μM, or 20.7μg/dL) and As (1 μM or 75 μg/L) used in this study is approximately 6 and 7.5 times higher than the current guideline value and falls within the range observed in exposed human populations (CDC, 2021; Ericson et al., 2021; Podgorski and Berg, 2020). After 24-hour exposure, total RNAs from the cells were extracted and used for RNA-seq-based transcriptional profiling (Figure 1A). Following stringent multiple testing corrections, we identified a total of 160 and 66 differentially expressed (DE) genes after exposure to Pb and As, respectively, as shown in the volcano plots in Figure 1B and 1C (Padj <=0.01, log₂FC <−0.5 or log₂FC >0.5). We identified 26 common DE genes with both Pb and As treatment (Figure 2A and 2B, Supplementary Table S2). The Full list of DE genes are available in Supplementary Excel File Table S1.

Figure 1.

Identification of differential gene expression in human NPCs exposed to Pb or As by RNA-seq. (A) Schematic workflow of the study. Created by BioRender. Volcano plots of RNA-seq results with top genes annotated in NPCs exposed to Pb (B) or As (C). Differentially expressed (DE) genes defined by log2(fold change) <−0.5 or >0.5 , Padj<0.01; Grey dots represent genes that do not meet the significance threshold.

Figure 2.

Common DE genes with both Pb or As treatment. (A) The Venn diagram shows the overlap of DE gene sets (B) Fold expression of common DE genes with Pb and As. (C) qPCR validation of F2RL2, TRIM16L, and PANX2. (D) Western blotting for F2RL2, TRIM16L, and PANX2. *P < 0.05, **P<0.01, ***P<0.001, ****<0.0001. N = 3 experiments.

3.2. Functional Annotation Enrichment using DAVID

We then used the DAVID pathway analysis tool to identify ontological categories that were enriched among the DE genes induced by exposures to Pb or As. The most significantly enriched terms in the DE genes with Pb or As treatment are presented in Table 1 and 2, respectively. Interestingly, with both Pb and As, the terms including “NFE2L2 regulating anti-oxidant/detoxification enzymes”, “Nuclear events mediated by NFE2L2”, “Ferroptosis”, “KEAP1-NFE2L2 pathway”, “Cellular responses to oxidative stress”, and “Cell redox homeostasis” are highly enriched. Together, these data implicate critical roles of oxidative stress and NRF2/KEAP1 pathway in the regulation of Pb and As-mediated neurotoxicity. The full list of enriched terms is available in Supplementary Excel File Table S2.

Table 1.

Enriched terms in the gene list differentially regulated by Pb by DAVID analysis.

Annotation Cluster 1	Enrichment Score: 3.28	Fold Enrichment	Padj*
Category	Term
REACTOME_PATHWAY	R-HSA-9818027~NFE2L2 regulating anti-oxidant/detoxification enzymes	94.23	0.00
REACTOME_PATHWAY	R-HSA-9759194~Nuclear events mediated by NFE2L2	27.04	0.00
REACTOME_PATHWAY	R-HSA-9755511~KEAP1-NFE2L2 pathway	19.96	0.00
KEGG_PATHWAY	hsa04216:Ferroptosis	31.60	0.00
REACTOME_PATHWAY	R-HSA-9711123~Cellular response to chemical stress	10.73	0.00
UP_KW_LIGAND	KW-0521~NADP	8.78	0.00
UP_KW_MOLECULAR_FUNCTION	KW-0560~Oxidoreductase	4.23	0.01
REACTOME_PATHWAY	R-HSA-2262752~Cellular responses to stress	3.72	0.02
REACTOME_PATHWAY	R-HSA-8953897~Cellular responses to stimuli	3.66	0.02
GOTERM_BP_DIRECT	GO:0007568~aging	9.95	0.28
KEGG_PATHWAY	hsa01240:Biosynthesis of cofactors	7.06	0.16
GOTERM_BP_DIRECT	GO:0006979~response to oxidative stress	10.17	0.52
GOTERM_BP_DIRECT	GO:0009725~response to hormone	22.13	0.52
GOTERM_BP_DIRECT	GO:0045454~cell redox homeostasis	21.10	0.52
KEGG_PATHWAY	hsa01100:Metabolic pathways	1.96	0.30
Annotation Cluster 2	Enrichment Score: 2.77	Fold Enrichment	Padj
Category	Term
GOTERM_CC_DIRECT	GO:0005829~cytosol	1.89	0.01
GOTERM_CC_DIRECT	GO:0005737~cytoplasm	1.74	0.05
Annotation Cluster 3	Enrichment Score: 2.76	Fold Enrichment	Padj
Category	Term
REACTOME_PATHWAY	R-HSA-170834~Signaling by TGF-beta Receptor Complex	14.35	0.02
REACTOME_PATHWAY	R-HSA-9006936~Signaling by TGFB family members	10.94	0.04

^*Padj: adjusted p-values for multiple comparisons by the Benjamini Hochberg correction

Table 2.

Enriched terms in the gene list differentially regulated by As by DAVID analysis.

Annotation Cluster 1	Enrichment Score: 3.09	Fold Enrichment	Padj*
Category	Term
REACTOME_PATHWAY	R-HSA-9818027~NFE2L2 regulating anti-oxidant/detoxification enzymes	91.98	0.00
REACTOME_PATHWAY	R-HSA-9759194~Nuclear events mediated by NFE2L2	26.39	0.00
KEGG_PATHWAY	hsa04216:Ferroptosis	35.92	0.00
REACTOME_PATHWAY	R-HSA-9755511~KEAP1-NFE2L2 pathway	19.49	0.00
REACTOME_PATHWAY	R-HSA-9711123~Cellular response to chemical stress	10.48	0.00
GOTERM_BP_DIRECT	GO:0007584~response to nutrient	23.02	0.15
UP_KW_MOLECULAR_FUNCTION	KW-0560~Oxidoreductase	4.22	0.03
UP_KW_LIGAND	KW-0521~NADP	7.24	0.07
GOTERM_BP_DIRECT	GO:0043066~negative regulation of apoptotic process	4.49	0.35
GOTERM_BP_DIRECT	GO:0006979~response to oxidative stress	11.22	0.35
GOTERM_BP_DIRECT	GO:0009410~response to xenobiotic stimulus	6.62	0.35
GOTERM_BP_DIRECT	GO:0009725~response to hormone	24.42	0.35
GOTERM_BP_DIRECT	GO:0045454~cell redox homeostasis	23.28	0.35
REACTOME_PATHWAY	R-HSA-2262752~Cellular responses to stress	2.97	0.22
REACTOME_PATHWAY	R-HSA-8953897~Cellular responses to stimuli	2.92	0.22
GOTERM_BP_DIRECT	GO:0043524~egative regulation of neuron apoptotic process	8.84	0.40
GOTERM_BP_DIRECT	GO:0007568~aging	8.78	0.40
GOTERM_BP_DIRECT	GO:0034599~cellular response to oxidative stress	11.25	0.69
Annotation Cluster 2	Enrichment Score: 2.04	Fold Enrichment	Padj
Category	Term
KEGG_PATHWAY	hsa04216:Ferroptosis	35.92	0.00
KEGG_PATHWAY	hsa04978:Mineral absorption	17.53	0.01
GOTERM_BP_DIRECT	GO:0006879~cellular iron ion homeostasis	23.02	0.15
REACTOME_PATHWAY	R-HSA-917937~Iron uptake and transport	17.97	0.07

^*Padj: adjusted p-values for multiple comparisons by the Benjamini Hochberg correction.

3.3. Validation of differentially regulated genes using qPCR and WB

Among 26 common DE genes with both Pb and As treatment (Figure 2A and 2B, Supplementary Table S2), NQO1 and HMOX1 are well known oxidative stress responsive genes regulated by NRF2, the master transcriptional factor for oxidative stress response (Kensler et al., 2007). NRF2 works to activate transcription by binding to AREs in the promotor region of target genes. In addition to NQO1 and HMOX1, many genes in Figure 2B have AREs in their promotor and identified as direct targets of NRF2. Induction of NQO1, HMOX1, and the other known NRF2 target genes as well as above DAVID pathway analysis data strongly suggest that Pb and As elicit oxidative stress and activate NRF2 in the cells. Because F2RL2, TRIM16L, and PANX2 were most highly upregulated with Pb and As and not previously reported as direct NRF2 targets, they were chosen for the follow-up validation. Up-regulation of F2RL2, TRIM16L, and PANX2 by Pb or As in NPCs was confirmed by qRT-PCR and Western blotting (Figure 2C and 2D). As shown in Figure 2C, mRNA expression of F2RL2, TRIM16L, and PANX2 increased in a concentration-dependent manner with both metals at 24 h. Figure 2D demonstrates the increased protein expression of F2RL2, TRIM16L, and PANX2 following exposure to Pb or As at 48 h.

3.4. The role of NRF2 in F2RL2, TRIM16L, and PANX2 regulation

Based on upregulation of NRF2 target genes and enrichment of NRF2 signaling in NPCs exposed to Pb and As, we hypothesize that F2RL2, TRIM16L, and PANX2 are part of the NRF2-mediated oxidative stress response. To test this hypothesis, we suppressed the expression of NRF2 and KEAP1(an endogenous NRF2 inhibitor) using siRNA, then measured the expression of F2RL2, TRIM16L, and PANX2. As shown in Figure 3A, transfection with si-NRF2-1 and si-NRF2-2 significantly decreased NRF2 expression while transfection with si-KEAP1 decreased the expression of KEAP1 measured by qPCR. In the absence of oxidative stress, KEAP1 interacts with NRF2, resulting in its ubiquitination and subsequent degradation. However, under conditions of oxidative stress, NRF2 dissociates from KEAP1, leading to its accumulation and translocation into the nucleus (Kensler et al., 2007). Thus, the inactivation of KEAP1 serves as an activator for NRF2. Consistent with this, si-KEAP1 transfection significantly upregulated NRF2 expression in the cells (Figure 3A, right panel). Next, we measured expression of F2RL2, TRIM16L, and PANX2 in the absence or presence of si-NRF2 or si-KEAP1. As shown in Figure 3B-3D, knockdown of NRF2 decreased expression of F2RL2, TRIM16L, and PANX2 while KEAP1 knockdown increased their expression as measured by qPCR and western blotting. Together, these data indicate that up-regulation of F2RL2, TRIM16L, and PANX2 is mediated by NRF2.

Figure 3.

NRF2-dependent regulation of F2RL2, TRIM16L, and PANX2 expression (A) Efficiency of siRNA knockdown of NRF2 or KEAP1 (B) Effect of NRF2 knockdown on F2RL2, TRIM16L, and PANX2 expression (C) Effect of KEAP1 knockdown on F2RL2, TRIM16L, and PANX2 expression. qPCR was done 24 hr post siRNA transfection. Negative control: nontargeting siRNA. (D) Western blotting shows NRF2/KEAP1-dependent regulation of protein expression. *P < 0.05, **P<0.01, ***P<0.001, ****<0.0001. N = 3 experiments.

3.5. Direct transcriptional regulation of F2RL2, TRIM16L, and PANX2 by NRF2

NRF2 activates target gene expression by binding to AREs within the promoter of target genes (Kensler et al., 2007). To identify potential NRF2 binding sites in the promoter region of F2RL2, TRIM16L, and PANX2, we utilized ConTra v3 web server, which analyzes any genomic region of interest to detect transcription factor binding sites across species (Kreft et al., 2017). Using ARE position weight matrices (M6360_1.02 and MA0150.2, Figure 4A), we identified multiple putative AREs in the promotor regions of F2RL2, TRIM16L, and PANX2 (Supplementary Figure S2-S4). Among them, we selected the most conserved ARE across species for each gene and designed specific primers for ChIP-qPCR assay (Figure 4B). The chromatin was prepared and digested, and chromatin immunoprecipitations were performed using the NRF2 or IgG (negative control) antibodies. Purified DNA was analyzed by qPCR to detect the presence of the putative F2RL2, TRIM16L, or PANX2 ARE sequences. As shown in Figure 4C, binding of NRF2 to the F2RL2, TRIM16L, or PANX2 ARE was significantly enriched compared to binding of IgG. A significant increase was also observed for a canonical NQO1 ARE, whereas no significant increase was observed for a non-NRF2 target sequence, RPL30-exon 3. Together, these data indicate that the F2RL2, TRIM16L, and PANX2 genes contain a functional ARE and are direct targets of NRF2.

Figure 4.

Regulation of F2RL2, TRIM16L, and PANX2 expression by direct NRF2 binding to the promoter region. (A) Consensus sequence motifs for NRF2 binding. (B) Presence of a putative ARE in the promoter region of F2RL2, TRIM16L, and PANX2. (C) NRF2 ChIP followed by qPCR amplification of the putative F2RL2 ARE, TRIM16L ARE, PANX2 ARE, NQO1 ARE (positive control), and RPL30 Exon 3 (negative control). *P < 0.05, **P<0.01, ***P<0.001. N = 3 experiments.

3.6. Effect of F2RL2, TRIM16L, and PANX2 on ROS formation

Because we found that F2RL2, TRIM16L, and PANX2 are direct targets of NRF2, a master regulator of the oxidative stress response pathway, we next explored the role of F2RL2, TRIM16L, and PANX2 in the regulation of Pb or As-induced ROS formation in NPCs. First, we suppressed the expression of F2RL2, TRIM16L, and PANX2 using siRNA transfection, then measured ROS formation in NPCs in the absence or presence of Pb or As using the H2DCFDA assay. Supplementary Figure S5 shows that siRNA transfection significantly decreased the expression of F2RL2, TRIM16L, and PANX2 in NPCs. As shown in Figure 5, treatment with Pb or As significantly increased DCF fluorescence intensity, indicating increased ROS production. Knockdown of F2RL2 or PANX2 further augmented Pb or As-induced ROS production in NPCs, while knockdown of TRIM16L did not lead to a significant change (Figure 5). Furthermore, overexpression of F2RL2 or PANX2 suppressed Pb or As-induced ROS production in NPCs (Supplementary Figure S6). These data implicate the potential involvement of F2RL2 and PANX2, but not TRIM16L, in the cellular response to Pb or As-induced ROS formation in NPCs.

Figure 5.

Effect of F2RL2, TRIM16L, and PANX2 knockdown on ROS formation in NPCs. After 24-h post transfection, NPCs were treated with Pb or As (5 μM) for 24 h. ROS production was measured by DCF fluorescence intensity normalized to Hoechst nuclei staining. *P < 0.05, **P<0.01, ***P<0.001. ****P<0.0001, significantly different compared to NT/Negative control. ^#P<0.05, ^##P<0.01, ^###P<0.001. ^####P<0.0001, significantly different from each other. N.S. not significant. N = 3 experiments.

3.7. Effect of F2RL2, TRIM16L, and PANX2 on cell number, cell death, and proliferation of NPCs

To test the functional roles of F2RL2, TRIM16L, and PANX2 in NPCs, we suppressed F2RL2, TRIM16L, and PANX2 expression using siRNA transfection, then measured the cell number, cytotoxicity, and proliferation of NPCs in the absence or presence of Pb or As. As shown in Figure 6A-6C, treatment with 2 μM Pb or As significantly decreased the cell number, as measured by the CyQuant assay, which quantifies the amount of DNA. siRNA knockdown of F2RL2 further decreased the cell number both in the presence and absence of metal treatment (Figure 6A), implicating the roles of F2RL2 in the survival or proliferation of NPCs. Conversely, siRNA knockdown of PANX2 significantly increased the cell number (Figure 6B). siRNA knockdown of TRIM16L did not significantly affect cell proliferation (Figure 6C).

Figure 6.

Effect of F2RL2, TRIM16L, and PANX2 knockdown on cell number, cell death, and proliferation of NPCs. After 24-h post transfection, NPCs were treated with Pb or As (2 μM) for 48 h. (A-C) Effect of F2RL2, TRIM16L, and PANX2 knockdown on cell numbers measured by Cyquant assay. (D-F) Effect of F2RL2, TRIM16L, and PANX2 knockdown on cell death measured by PI staining. (G-l) Effect of F2RL2, TRIM16L, and PANX2 knockdown on proliferation measured by EdU assay*P < 0.05, **P<0.01, ***P<0.001. ****P<0.0001, significantly different compared to NT/Negative control. ^#P<0.05, ^##P<0.01, ^###P<0.001. ^####P<0.0001, significantly different from each other. N.S. not significant. N = 3 experiments.

Next, we measured cell death using PI, a fluorescent DNA-binding dye that freely penetrates cell membranes of dead or dying cells. As shown in Figure 6D-6F, Pb or As treatment significantly increased cell death in NPCs, consistent with the decreased cell number in Figure 6A-6C. Knockdown of F2RL2 increased Pb or As-induced cell death; on the other hand, knockdown of PANX2 decreased Pb or As-induced cell death, aligning with the data in Figure 6A and 6B. No significant changes were observed with TRIM16L knockdown (Figure 6F).

Figure 6G-6I shows decreased cell proliferation of NPCs with Pb or As treatment, as measured by the EdU assay. Interestingly, knockdown of F2RL2 increased cell proliferation compared to the negative control with or without metal treatment, implicating potential compensatory responses to decreased cell number (Figure 6A) and increased cell death (Figure 6D). On the contrary, knockdown of PANX2 decreased cell proliferation (Figure 6H). TRIM16L knockdown did not affect cell proliferation either (Figure 6I).

The overexpression of F2RL2 and PANX2 resulted contrasting trends, as demonstrated in Supplementary Figure S6, further strengthening our findings. Interestingly, while TRIM16L knockdown did not impact cell number, cell death, or proliferation (Figure 6C, F, and I), TRIM16L overexpression significantly decreased Pb or As-induced cell death and reversed the Pb or As-induced reduction of cell numbers, suggesting complex regulatory mechanisms governing its role (Supplementary Figure S7:C and F). Together, these data indicate that F2RL2, PANX2, and TRIM16L differentially regulate NPC function and metal toxicity.

3.8. Effect of F2RL2, TRIM16L, and PANX2 on NPC differentiation

Neural Progenitor Cells (NPCs) have the capacity to differentiate into multiple neural cell types and play pivotal roles in early brain development (Li et al., 2020; Martinez-Cerdeno and Noctor, 2018; Vaccarino et al., 2001). In this study, we investigated whether the suppression of F2RL2, TRIM16L, and PANX2 influences NPC differentiation both in the absence and presence of metal treatment. First, we knocked down the expression of F2RL2, TRIM16L, and PANX2 in NPCs through siRNA transfection. Subsequently, we initiated differentiation with or without the treatment of Pb or As. We assessed the expression of various neural cell markers at post-differentiation day 0 and day 7 using qRT-PCR and immunofluorescence (IF) staining. We examined the mRNA expression of neural markers, including NES (Nestin, an NPC marker), TUBB3 (tubulin beta 3 class III, a neuronal marker), MAP2 (microtubule-associated protein 2, another neuronal marker), GFAP (glial fibrillary acidic protein, an astrocyte marker), and OLIG2 (oligodendrocyte transcription factor 2, an oligodendrocyte marker) in NPCs using qRT-PCR. As shown in Figure 7A, knockdown of PANX2 and TRIM16L significantly increased the expression of NES, TUBB3, MAP2, and GFAP at post-differentiation day 0 (Day 0). However, F2RL2 knockdown decreased GFAP expression while TRIM16 knockdown increased OLIG2 expression.

Figure 7.

Effect of F2RL2, TRIM16L, and PANX2 knockdown on NPC differentiation measured by qRT-PCR. After 24-h post transfection, NPCs were differentiated while treated with vehicle (H₂O), Pb or As (2 μM). Expression of neural markers was measured by qRT-PCR at Day 0 or Day 7. NES (Nestin, an NPC marker), TUBB3 (tubulin beta 3 class III, a neuronal marker), MAP2 (microtubule-associated protein 2, another neuronal marker), GFAP (glial fibrillary acidic protein, an astrocyte marker), and OLIG2 (oligodendrocyte transcription factor 2, an oligodendrocyte marker).(A) Expression of neural markers at Day 0. *P < 0.05, **P<0.01, ***P<0.001. ****P<0.0001, significantly different compared to Negative control. ^&P<0.05, ^&&P<0.01, significantly different from each other. (B-F). Expression of neural markers at Day 7. *P < 0.05, **P<0.01, ***P<0.001. ****P<0.0001, significantly different compared to Negative control in each treatment condition (NT, Pb or As). ^#P<0.05, ^##P<0.01, ^###P<0.001. ^####p<0.0001, significantly different compared to NT/Negative control. N = 3 experiments.

At post-differentiation day 7 (Day 7), both Pb or As treatment and siRNA knockdown of each gene increased the expression of most genes compared to the control (NT/Negative control) (Figure 7B-7F). Notably, we observed differential effects of metal treatment with the knockdown of specific genes. For instance, F2RL2 knockdown significantly increased Pb-induced NES upregulation, whereas knockdown of F2RL2, PANX2, and TRIM16L decreased As-induced NES upregulation (Figure 7B). Similar patterns were observed for other neural markers. In Figure 7C, the knockdown of F2RL2 and PANX2 increased Pb-mediated TUBB3 expression, while the knockdown of TRIM16L decreased As-induced TUBB3 expression. In Figure 7D, the knockdown of F2RL2 and PANX2 increased Pb-mediated GFAP expression, whereas the knockdown of all three genes decreased As-induced GFAP expression. PANX2 knockdown also led to increased Pb-induced MAP2 upregulation (Figure 7E), while the knockdown of all three genes decreased Pb-induced OLIG2 upregulation (Figure 7F). Supplementary Figure S8 shows the effect of overexpression of these genes on NPC differentiation. Figure 8 and Figure 9 displays IF staining of NPCs at Day 0 and 7, along with quantification of neural markers, including TUBB3, MAP1B (neuronal markers), GFAP (astrocyte marker), and PLP1 (oligodendrocyte marker), normalized to WGA (Wheat Germ Agglutinin). Comprehensive IF staining images for all neural markers as well as WGA, Hoechst, and overlay can be found in Supplementary Figure S9-S11.

Figure 8.

IF staining of NPC differentiation. After 24-h post transfection, NPCs were differentiated while treated with vehicle (H₂O), Pb or As (2 μM). Expression of neural markers was measured by IF staining at Day 0 or Day 7. Neuronal markers: TUBB3 and MAP1B, Astrocyte marker: GFAP, Oligodendrocyte: PLP1 (A) Day 0. (B) Day 7. Comprehensive IF staining images for all neural markers as well as WGA, DAPI, and overlay can be found in Supplementary Figure S7-S9.

Figure 9.

Quantification of neural markers in IF staining. The Zeiss Zen Blue software package was employed to determine the staining intensities of all markers. The average staining intensity of each marker was divided by the staining intensity of WGA to normalize for differences in signal strength between regions. *P < 0.05, **P<0.01, ***P<0.001. ****P<0.0001, significantly different compared to Negative control in each treatment condition (NT, Pb or As). ^#P<0.05, ^##P<0.01, ^###P<0.001. ^####P<0.0001, significantly different compared to NT/Negative control. ^&P<0.05, ^&&P<0.01, ^&&&P<0.001. ^&&&&P<0.0001 significantly different from each other. N=8-21.

As seen in Figure 9A, TUBB3 expression increased with knockdown of F2RL2, PANX2, or TRIM16L at Day 0. At Day 7, its expression increased under all treatment conditions (NT, Pb, or As/Negative control) compared to Day 0 (NT/Negative control), indicating neuronal maturation. However, treatment with Pb or As decreased TUBB3 expression compared to NT, suggesting an inhibitory effect of Pb and As on neuronal maturation. siRNA knockdown also had a modifying effect on TUBB3 expression in the absence or presence of metal treatment. For instance, the knockdown of TRIM16L resulted in decreased TUBB3 expression in NT or Pb-treated cells compared to the negative control, whereas knockdown of F2RL2 or PANX2 increased TUBB3 expression in As-treated cells.

Similar trends were observed for MAP1B, indicating increased neuronal maturation, and GFAP, indicating increased differentiation into astrocytes (Figure 9B and 9C). Notably, PANX2 or TRIM16L knockdown decreased GFAP expression in Pb-treated cells, while F2RL2 knockdown decreased GFAP expression in As-treated cells.

PLP1 expression did not significantly increase at Day 7, suggesting a later differentiation of oligodendrocytes from NPCs (Figure 9D). However, Pb or As treatment significantly increased PLP1 expression, implying that Pb or As expedites oligodendrocyte differentiation. The effect of siRNA knockdown varied depending on the treatment conditions (NT, Pb, or As), indicating complex interactions.

In summary, our findings suggest that Pb or As treatment influences NPC differentiation, and these effects are differentially regulated by the knockdown or overexpression of F2RL2, PANX2, or TRIM16L.

3.9. Association of F2RL2, PANX2, and TRIM16L with autism spectrum disorder

Previous studies have linked Pb and As exposure to detrimental neurodevelopmental outcomes, including ASD (Bellinger, 2013; Calderon et al., 2001; Dickerson et al., 2015; Froehlich et al., 2009; Goodlad et al., 2013; Needleman et al., 1996; O'Bryant et al., 2011; Opler et al., 2008; Tsai et al., 2003; Wang et al., 2019). To explore the potential involvement of our novel NRF2-regulated genes, F2RL2, PANX2, and TRIM16L, in ASD, we conducted a comparative analysis using publicly available transcriptomic datasets. Specifically, we utilized RNA-seq data on iPSC-derived NPCs from ASD patients and healthy controls (Fu et al., 2023) and examined the expression of F2RL2, PANX2, and TRIM16L in the dataset. As depicted in Figure 10A, Fu et al. (Fu et al., 2023) utilized iPSCs from ASD patients carrying with PTEN or CTNNB1 variants, as these genes are known ASD risk genes, and generated panels of cell lines through genome editing (Butler et al., 2005; Marchetto et al., 2017). We compared the transcriptome of four ASD NPC lines (PTEN WT/I135L, PTEN KO/KO, PTEN WT/WT, and CTNNB1 WT/WT) with that of control NPCs. The expression levels of F2RL2, PANX2, and TRIM16L were summarized in Figure 10B. We observed significant downregulation of F2RL2 and TRIM16L in NPCs with the PTEN variant (PTEN WT/I135L). Conversely, PTEN knockout (PTEN KO/KO) or correction (PTEN WT/WT) using CRISPR-Cas9 gene editing led to significant upregulation of F2RL2 and downregulation of PANX2, suggesting that PTEN variants modify gene expression in ASD NPCs. Furthermore, NPCs with CTNNB1 correction (CTNNB1 WT/WT) displayed significant upregulation of F2RL2 and downregulation of TRIM16L. Notably, these transcriptomic changes correlated with dysregulation of other ASD risk genes and modifications in neurogenesis phenotype (Fu et al., 2023). In summary, these findings suggest potential interactions of F2RL2, PANX2, and TRIM16L with ASD genetic background and ASD risk genes.

Figure 10.

Expression of F2RL2, TRIM16L, and PANX2 in human NPCs derived from ASD patients. (A) The schematic workflow on establishing cell lines and transcriptomic analyses (Fu et al., 2023). (B) The heatmap showing the expression of F2RL2, TRIM16L, and PANX2 in ASD NPCs compared to control NPCs. *Padj<0.05. N=3 experiments

4. Discussion

Studies have reported that exposure to environmental metal toxicants, such as Pb or As, during the early stages of brain development is associated with adverse neurocognitive and neurodevelopmental outcomes in children. However, the molecular mechanisms underlying metal toxicity remain poorly understood. In this study, we identified the NRF2-mediated antioxidant response pathway as a shared cellular response to Pb and As exposure in human NPCs. Specifically, we discovered that F2RL2, TRIM16L, and PANX2 are novel NRF2 target genes upregulated by both metals, and we found that they regulate NPC function including cell death, proliferation, and differentiation. Furthermore, we demonstrated that the dysregulation of F2RL2, TRIM16L, and PANX2 is associated with ASD genetic backgrounds and ASD risk genes, suggesting potential roles of these genes in the pathogenesis of neurodevelopmental disorders.

We identified significant transcriptomic changes related to the NRF2/KEAP1 pathway in human NPCs exposed to both Pb and As. Although this finding is consistent with previous reports of NRF2/KEAP1 pathway activation by Pb or As (Korashy and El-Kadi, 2006; Simmons et al., 2011; Wagner et al., 2016; Yang et al., 2007; Zeller et al., 2010), and the relationship between these genes and NRF2 has been suggested (Beaver et al., 2014; Copple et al., 2019; Liao et al., 2020; Singh et al., 2013), our study is the first to demonstrate that F2RL2, TRIM16L, and PANX2 are regulated by direct NRF2 binding to AREs in the promoter region. Through sequence analysis of the promoter regions of these genes (1000-bp upstream) using ConTra V3, we identified multiple putative AREs. Specifically, we found at least 12, 16, and 3 putative AREs in the promoter region of F2RL2, TRIM16L, and PANX2, respectively (Supplementary Figure S2-S4). From these, we selected a highly conserved ARE for each gene and tested NRF2 binding using ChIP-qPCR. However, we acknowledge that besides the validated AREs in this study, NRF2 may bind to other AREs in the promoter region to regulate the expression of these genes.

Our study provides the first report on the functional roles of F2RL2 in the regulation of NPC function and its potential neuroprotection against environmental stress during early brain development. The F2RL2 gene encodes PAR3, a member of the protease-activated receptor (PAR) family. PAR3 acts as a cofactor in thrombin-mediated cleavage and activation of PAR4, playing a crucial role in hemostasis and thrombosis (Nakanishi-Matsui et al., 2000; Sambrano et al., 2001). However, its functions in the nervous system remain largely unknown. Previous research indicates PAR3's involvement in regulating pain behaviors (Mwirigi et al., 2021) and neuroprotection mediated by activated protein C (APC) and PARs (Cheng et al., 2006; Guo et al., 2004; Guo et al., 2009a; Guo et al., 2009b; Liu et al., 2004; Wang et al., 2009). APC and its mutant form, 3K3A-APC, exhibit neuroprotective effects and promote proliferation and differentiation of NPCs while inhibiting astroglial cell differentiation through PAR-mediated Akt signaling (Goldman et al., 1997; Guo et al., 2011; Nakamura et al., 2011; Petraglia et al., 2010; Thiyagarajan et al., 2008). Our study shows the involvement of F2RL2 in the regulation of Pb or As-induced ROS formation in NPCs, providing a potential mechanism for its neuroprotective effect. Further investigation is needed to understand the interaction between ROS formation, NRF2, PAR3, APC, and Akt signaling in the neuroprotective response to metal exposure.

Furthermore, we found that PANX2 regulates ROS formation and cellular functions in NPCs. PANX2, a member of the pannexin family of ion channels, was initially believed to have selective expression in the CNS but is now known to be widely expressed in various tissues (Baranova et al., 2004; Bruzzone et al., 2003; Le Vasseur et al., 2019; Le Vasseur et al., 2014). PANX2 has been proposed to regulate neuronal development and ischemia-induced brain injury (Bargiotas et al., 2011; Swayne et al., 2010). For example, the suppression of PANX2 in neuro2a cells accelerated neuronal differentiation, neurite extension, and the expression of neuronal cell markers (Swayne et al., 2010), suggesting a role of PANX2 in neuronal commitment. This is consistent with our data showing that PANX2 knockdown increased various markers of neural differentiation. However, its effects varied with the treatment conditions, implicating complex regulations on NPC differentiation. Additionally, loss of PANX1 and PANX2 has been linked to reduced ischemic brain damage and improved neurological outcomes in vivo (Bargiotas et al., 2011). Our data demonstrate that PANX2 knockdown reduces cell death and proliferation of NPCs regardless of metal exposure, contributing to the increased number of NPC whereas its overexpression has an opposite effect. Although Pb and As treatments decrease cell number, si-PANX2 reverses their effects, restoring number to basal levels, suggesting that PANX2's influence on NPCs outweighs the effects of metals at the concentrations used in this study. Decreased cell proliferation with the knockdown may result from compensatory response to increased cell number. Interestingly, although PANX2 knockdown augmented Pb or As-induced ROS formation similar to F2RL2 knockdown, their effects on NPC number, death, proliferation are opposite implicating additional mechanisms. Gap junction/ion channel proteins play important roles in cell-cell communication, which can be perturbed by metal exposure (Hussain et al., 2018). Further research is necessary to fully understand the functions of PANX2 under metal exposure and the complex mechanisms by which PANX2 regulates NPC function.

The TRIM family proteins, characterized as E3 ubiquitin ligases, play diverse roles in cellular processes, including innate immunity, genetic disorders, cancer, and neurodegenerative diseases (Gushchina et al., 2018; Hatakeyama, 2011; Meroni and Diez-Roux, 2005; Ozato et al., 2008). Furthermore, they are involved in sustaining cellular homeostasis under oxidative stress. For example, TRIM72 protects hair follicle stem cells from apoptosis induced by oxidative stress, while TRIM4 and 8 sensitize cells to oxidative stress (Dang et al., 2020; Niu et al., 2022; Tomar et al., 2015). In addition, TRIM22 accelerates NRF2 degradation, while TRIM15 stabilizes NRF2 by promoting KEAP1 ubiquitination and degradation (Liang et al., 2022). In our study, we found that TRIM16L expression is directly regulated by NRF2 binding to the promoter region, although suppression of TRIM16L did not affect ROS production in Pb or As-exposed NPCs. We also found that overexpression of TRIM16L decreased Pb or As-induced cell death and affected NPC differentiation. From our RNA-seq data, another TRIM protein, TRIM16, was also upregulated by Pb and As exposure. Further investigation into the roles of TRIM16L and other TRIM proteins in the NRF2/KEAP1 pathway will elucidate their impact on cellular homeostasis and function under metal exposure.

Furthermore, our findings demonstrate that the dysregulation of F2RL2, PANX2, and TRIM16L is associated with ASD genetic backgrounds, potentially interacting with ASD risk genes. ASDs are complex neurodevelopmental disorders with genetic heterogeneity (Geschwind and Levitt, 2007). Early brain overgrowth is observed in approximately 20% of individuals with ASD (Fombonne et al., 1999; Nordahl et al., 2011). In addition to key ASD risk genes like PTEN and CTNNB1, the genetic background, including variants and mutations throughout the genome and/or epigenome, may contribute to ASD pathology (Pietropaolo et al., 2011; Sokolowski, 2020). Previous research using NPC lines with PTEN variants/ASD backgrounds has shown dysregulation of known ASD risk genes and genes involved in neurogenesis, resulting in an overproduction of NPCs and neurons (Fu et al., 2023). Our analysis of the transcriptomic datasets revealed differential regulation of F2RL2, PANX2, and TRIM16L in these cells. This association with ASD gene expression and phenotype, along with our in vitro evidence of their role in NPC proliferation and differentiation, suggests their potential involvement in NPC proliferation and neurogenesis during early brain development. Performing genome-wide association studies to explore the relationship between genetic variants in these genes and neurodevelopmental outcomes, both in the absence and presence of metal exposure, will provide valuable insights into their roles in early brain development.

This study presents valuable insights into the cellular responses to Pb and As exposure, shedding light on the potential mechanisms underlying their neurotoxic effects. Abnormal NPC proliferation, differentiation, and death have been associated with adverse neurodevelopmental outcomes in vivo (Fukuda et al., 2011; Groszer et al., 2001; Murai et al., 2016; Noguchi et al., 2008). However, it is crucial to acknowledge certain limitations in the research, which have implications for future studies. One notable shortcoming is the absence of animal exposure-related research, which hinders the direct translation of the observed cellular effects to in vivo nerve damage. Therefore, future research should endeavor to address this gap by investigating alterations in the expression of F2RL2, PANX2, and TRIM16L in NPCs within mouse brains following exposure to Pb or As. Additionally, the study should explore the functional significance of these genes in relation to neurodevelopment and behavioral outcomes, employing in vivo models both with and without exposure to these metals.

In conclusion, we shed light on the roles of novel NRF2-regulated genes in regulating NPC function in response to environmental metal toxicants and their implications for neurodevelopmental outcomes. While further research is needed to fully comprehend the implications of these findings, they may identify novel genes of interest in the context of neurodevelopment, providing a foundation for the development of preventative or therapeutic strategies to mitigate environment-related neurological pathologies.

Environmental Implication.

Our study addresses the critical susceptibility of early brain development to ubiquitous metal toxicants lead (Pb) and arsenic (As). Despite efforts to reduce exposure, millions of children globally are still exposed to these toxicants above guideline values. The concentration of Pb and As used in the study falls within the range observed in exposed human populations and has been associated with adverse neurocognitive and developmental outcomes in children. We believe our findings may form a foundation for developing preventative or therapeutic strategies to mitigate environment-related neurological pathologies.

Highlights.

• Pb and As upregulate F2RL2, PANX2, TRIM16L in neural progenitor cells (NPCs).

• F2RL2, PANX2, TRIM16L are novel NRF2-dependent genes.

• NRF2 directly binds their promoter region.

• Knockdown or overexpression of these genes differentially regulate NPC function.

• Dysregulation of these genes is linked to autism spectrum disorder gene signature.