Arsenic impairs stem cell differentiation via the Hippo signaling pathway

Abstract

Arsenic is a ubiquitous toxic metalloid, with over 150 million people exposed to arsenic concentrations above the current 10 ppb drinking water standard through contaminated food and water. Arsenic is a known developmental toxicant as neuronal and muscle development are disrupted following arsenic exposure during embryogenesis. In this study, murine embryonic stem cells were chronically exposed to 0.1 μM (7.5 ppb) arsenic for 32 weeks. RNA sequencing showed that the Hippo signaling pathway, which is involved in embryonic development and pluripotency maintenance, is impaired following arsenic exposure. Thus, temporal changes in the Hippo pathway’s core components and its downstream target genes Ctgf and c-Myc were investigated. Protein expression of the pathway’s main effector YAP in its active form was significantly upregulated by 3.7-fold in arsenic-exposed cells at week 8, while protein expression of inactive phosphorylated YAP was significantly downregulated by 2.5- and 2-fold at weeks 8 and 16. Exposure to arsenic significantly increased the ratio between nuclear and cytoplasmic YAP by 1.9-fold at weeks 16 and 28. The ratio between nuclear and cytoplasmic transcriptional enhancer factor domain was similarly increased in arsenic-treated samples by 3.4- and 1.6-fold at weeks 16 and 28, respectively. Levels of Ctgf and c-Myc were also upregulated following arsenic exposure. These results suggest that chronic exposure to an environmentally relevant arsenic concentration might hinder cellular differentiation and maintain pluripotency through the impairment of the Hippo signaling pathway resulting in increased YAP activation.

Introduction

Arsenic is a ubiquitous toxic metalloid found in soil and groundwater that is well known for its hazardous effects on health. Arsenic can originate from natural sources as well as anthropogenic activities.^1,2 Its main route of exposure is through ingestion of contaminated food and drinking water, threatening more than 150 million people in 70 countries. The current drinking water standard is 10 ppb arsenic, which was established by the U.S. Environmental Protection Agency and the World Health Organization.^3,4 Elevated arsenic concentrations have been documented in various countries including China, Mexico, Chile, Argentina, India, Bangladesh, and the United States.^5–8 Reports indicate that in the United States, approximately 2.1 million people obtain drinking water from domestic wells with arsenic concentrations above 10 μg/L.⁹ In addition, it has been previously estimated that on average, an individual adult in the United States ingests 3.2 μg of arsenic daily, with a total dietary intake ranging from 1.3 to 11.4 μg/day.^2,10 Epidemiological studies have shown that placental and cord blood arsenic concentrations are comparable, indicating that arsenic can readily cross the placental barrier.^11–13 Consequently, exposure during pregnancy and embryonic development can cause both immediate and delayed effects, including increased spontaneous abortion and stillbirths, low birth weight, reduced weight gain, impaired muscle development and locomotor activity, learning and memory deficits, and neurobehavioral alterations.^{1,11,12,14,15}

Previous studies conducted using rodent models have shown that administration of inorganic arsenic to pregnant dams induces fetal malformations, impaired development, disrupted locomotor activity, and behavioral deficits.^16–19 In particular, Chattopadhyay et al.¹⁶ observed that 300 ppb of sodium arsenite administered through drinking water to pregnant rats resulted in offspring with decreased responsiveness to a new environment, along with reduced locomotion and limb movement. Consistently, hippocampal neurogenesis, cell morphology, and gene expression patterns have been impaired following exposure to 50 ppb arsenic during fetal development.²⁰ Moreover, disrupted learning abilities and memory behavior have been reported in adult mice following in utero arsenic exposure to 50 ppb of arsenic as these mice performed poorly in new object identification tests and failed to show improvements in hippocampal-dependent learning tasks compared with control mice.^21,22

While the mechanisms responsible for developmental changes in offspring are not fully understood, one possibility is that arsenic impairs proper cellular development and differentiation. For instance, C2C12 myoblasts exposed to 20 nM sodium arsenite showed impaired differentiation and myotube formation caused by inhibition of myogenic transcription factor expression.²³ Similarly, exposure of P19 mouse embryonic stem cells to 0.5 μM sodium arsenite suppressed skeletal muscle and neuronal differentiation.²⁴ Neuronal differentiation and neurite outgrowth was also impaired in Neuro-2a cells when exposed to <5 μM arsenic trioxide²⁵ or to 5 μM sodium arsenite.²⁶ Moreover, chronic low-level arsenic exposure can increase expression of genes involved in pluripotency maintenance and decrease the expression of genes involved in cellular differentiation.²⁷

Signaling pathways help regulate the balance between pluripotency and cellular differentiation. Arsenic exposure is known to impair several signaling pathways, including TGF-β^27–29 and Wnt/β-catenin.^24,30,31,32 Another pathway involved in regulating cellular differentiation and stem cell maintenance is the Hippo signaling pathway.^33,34 In the Hippo pathway following upstream activation, the signal receiver MST1/2 phosphorylates LATS1/2, which then phosphorylates YAP when it is bound to TAZ.^35,36 Phosphorylation of the YAP/TAZ complex results in cytoplasmic retention, or in ubiquitination and subsequent degradation.³⁷ Conversely, the unphosphorylated YAP/TAZ complex translocates to the nucleus where it interacts with members of the transcriptional enhancer factor domain (TEAD) family to regulate expression of target genes,^33,38,39 including connective tissue growth factor (Ctgf), Axl, and c-Myc.^{33,35,40,41,42} Interestingly, unlike other signaling pathways, its role in arsenic-induced developmental toxicity has been poorly investigated.

The purpose of the current study was to determine whether a long-term arsenic exposure to P19 embryonic stem cells disrupted the Hippo pathway, thereby inhibiting cellular differentiation. Our results showed an increased expression of YAP at week 8 along with a decreased expression of its phosphorylated (inactive) form at weeks 8 and 16. Transcript levels of Lats2 were significantly upregulated at week 24, and protein levels of LATS2 were significantly increased at weeks 8 and 16 of arsenic exposure. Furthermore, at weeks 16 and 28, the ratio of nuclear (active) and cytoplasmic (inactive) YAP and TEAD was increased concomitantly with upregulated expression of YAP’s target genes Ctgf and c-Myc resulting in pluripotency maintenance. These results suggest that the Hippo signaling pathway plays a role in arsenic-induced developmental toxicity and is involved in the disruption of proper cellular differentiation following arsenic exposure.

Methods

Cell culture

P19 mouse embryonic stem cells (Sigma-Aldrich, St. Louis, MO, USA) were cultured in α-MEM media (Hyclone, Logan, UT, USA) with 7.5% bovine calf serum (Hyclone), 2.5% fetal bovine serum (Sigma-Aldrich), and 1% L-glutamine (Thermo Fisher Scientific, Waltham, MA, USA). Cells were cultured in a humidified incubator at 37 °C and 5% CO₂ and passaged every 2 to 3 days to maintain low confluency and avoid premature cellular differentiation. Cells were exposed to 0 or 0.1 μM arsenic as sodium arsenite (Sigma-Aldrich) for up to 32 weeks as previously described.²⁷ The concentration chosen for this study (0.1 μM or 7.5 ppb) is lower than the current drinking water standard of 10 ppb and is a relevant concentration for an embryonic exposure study.^11,13 Control and arsenic-exposed samples were continuously cultured as three independent replicates (n = 3 for each treatment).

Cell differentiation

Every 4 weeks, the control and arsenic-exposed cells were differentiated as previously described.²⁷ In brief, cells were exposed to 1% DMSO and cultured as hanging drops for 2 days, with the newly formed embryoid bodies (EBs) transferred to 96-well ultralow attachment plates and cultured for an additional 3 days. At day 5, half the EBs were harvested, washed in phosphate buffered saline and incubated overnight at 4 °C in 10% neutral buffered formalin (NBF) for subsequent immunohistochemistry. The remaining EBs were transferred to 48-well plates coated with 0.1% gelatin, and cultured for 4 days to promote differentiation into skeletal myoblasts and neural progenitors. Culture medium was refreshed every 48 h. At day 9, cells were collected and stored at −80 °C in Trizol (Thermo Fisher) for subsequent RNA extraction or in radioimmunoprecipitation assay (RIPA) buffer for subsequent protein analysis.

Quantitative PCR

Extracted RNA (2 μg) was converted into complementary DNA (cDNA) using M-MLV Reverse Transcriptase (Promega, Madison, WI, USA). Gene expression was assessed by real-time quantitative PCR on a Bio-Rad iQ5 thermocycler (Hercules, CA, USA) using RT² SYBR Green (Applied Biosystems, Foster City, CA, USA) and gene-specific primers (Supplementary Table S1). Samples were run in triplicate and a standard curve was generated with five concentration points (10⁻³–10⁻⁷ ng cDNA) to determine run efficiency. A melt curve was obtained for each analysis to ensure the absence of nonspecific primer binding. Gapdh and β2-microglobulin were used as housekeeping genes to normalize data. Gene expression for each individual week was analyzed using the delta–delta Ct method,⁴³ by setting the controls to a relative expression of 1.

Immunohistochemistry analysis

Day 5 EBs were fixed overnight in NBF, dehydrated in ethanol, embedded in paraffin, and sectioned at 5 μm. Slides were deparaffinized and Tris–EDTA buffer (pH 9) was used for antigen retrieval. Primary antibodies for YAP (1:100, Cell Signaling, #12395), TEAD4 (1:50, Sigma-Aldrich, #AV38276), and CTGF (1:200, Genetex, #GTX124232) were incubated overnight at 4 °C. After incubation with Alexa Flour 488 secondary antibodies (1:200, anti-rabbit, Invitrogen #A11034) and Alexa Fluor 546 (1:200 anti-mouse, Invitrogen, #A11003), nuclei were counterstained with TO-PRO-3 (1:1,000, Thermo Fisher, #T3605).

Samples double labeled with YAP and TEAD4 were imaged using a Leica SP8X confocal microscope (Leica Microsystems, Buffalo Grove, IL, USA), equipped with a tunable white light laser, HyD detectors, and time gating. Samples were imaged using a 63× objective (numerical aperture (N.A.) = 1.4) and a zoom of 1.5. To image TO-PRO-3, we used an excitation wavelength of 635 nm and collected emission wavelengths of 650 to 700 nm, with a time gate of 0.5 to 6.0 ns and a frame average of 2. To image Alexa Fluor 546, we used an excitation wavelength of 557 nm and collected emission wavelengths of 562 to 620 nm, with a time gate of 0.5 to 6.0 ns and a frame average of 4. To image Alexa Fluor 488, we used an excitation wavelength of 499 nm and collected emission wavelengths of 504 to 550 nm, with a time gate of 0.5 to 6.0 ns and a frame average of 4. Additional overview images were collected using a 20× objective (N.A. = 0.75) and a zoom of 1.0 using the settings outlined above, but with frame averages of 1 (TO-PRO-3), 2 (AF546), and 2 (AF488).

Samples labeled with CTGF were imaged using a Leica SPE confocal microscope (Leica Microsystems, Buffalo Grove, IL), using a 63× objective (N.A. = 1.3) and a zoom of 1.5. To image TO-PRO-3, we used an excitation wavelength of 635 nm and collected emission wavelengths of 650 to 700 nm, with a frame average of 2. To image Alexa Fluor 488, we used an excitation wavelength of 488 nm and collected emission wavelengths of 495 to 550 nm, with a frame average of 4. Additional overview images were collected using a 20× objective (N.A. = 0.6) and a zoom of 1.0 using the settings outlined above, but with frame averages of 2 for both TO-PRO-3 and AF488.

Images were analyzed in ImageJ to determine protein expression and cellular localization. CTGF, YAP, and TEAD expression was determined by assessing the integrated density value (IDV), which was normalized to area. For YAP and TEAD, a section at the outer edge of the EBs containing 30–40 cells was identified, and IDVs were determined for the total section and for the nuclei. Protein expression in the cytoplasm was determined by subtracting the nuclei IDV from the IDV of the whole tissue section. Both were used to calculate a nuclear/cytoplasmic ratio for YAP and TEAD expression.

Immunoblotting analysis

Differentiated cells were harvested at day 9 and proteins were extracted using RIPA buffer supplemented with protease inhibitor cocktail (Thermo Fisher Scientific). Total protein concentration was quantified using the BCA protein assay kit (Thermo Fisher Scientific). Equal amounts of protein (10 μg) were electrophoresed on 5–20% TBX gels (Bio-Rad) and immunoblotting was performed. Primary antibodies included active YAP1 (1:1,000, Abcam, #ab205270), phosphorylated YAP1 (1:1,000, Cell Signaling, #13008S), LATS2 (1:1,000, Proteintech, #20276-1-AP), MST1 (1:1,000, Cell Signaling, #3682S), GAPDH (1:1,000, GeneTex, #GTX627408), and acetylated tubulin (1:1,000, Sigma-Aldrich, #T7451). Horseradish peroxidase-conjugated anti-rabbit (1:5,000, Abcam, #ab6721) and anti-mouse (1:5,000, Invitrogen, #62-6520) secondary antibodies were used. Protein expression was assessed by chemiluminescence (Luminol; Santa Cruz, Santa Cruz, CA, USA) using a Bio-Rad ChemiDoc imaging system.

RNA sequencing

RNA sequencing and subsequent quality control and genome alignment were performed by Novogene (Sacramento, CA, USA) on rRNA-depleted RNA extracted from day 9 differentiated cells at weeks 8, 16, and 24 (n = 3 for each treatment; each replicate is from a separate flask of continuously cultured cells). Paired-end sequencing was performed via the Illumina platform, and reads were checked for quality metrics and trimmed for low-quality bases. Indexes of the reference genome were built using STAR,⁴⁴ and paired-end clean reads were aligned to the Mus musculus (mm10) reference genome using STAR (v2.5). HTSeq v.6.1^45,46 was used to count the mapped reads for each gene. Differential expression analysis was performed using the DESeq2 (v1.29.7) R package,⁴⁷ which can be used to identify genes that react in a condition-specific manner in time-course experiments. Specifically, we performed a Wald test at week 16 and week 24 to compare control and arsenic-treated samples at those specific time points. The Wald test is a hypothesis test commonly used to detect differentially expressed genes when comparing two groups.

Gene counts were normalized to library size using DESeq2. The P-values resulting from the differential expression analysis were subjected to the Benjamini–Hochberg adjustment to control for the false discovery rate (FDR). Genes with a resulting adjusted P-value <0.05 were considered differentially expressed and used for subsequent functional annotation analysis. Differentially expressed genes were annotated using the R package annotables.⁴⁸ The samples were clustered using the hierarchical clustering distance method with the pheatmap package in R.⁴⁹ Gene Ontology (GO) over-representation analysis and gene set enrichment analysis (GSEA) of differentially expressed genes were performed using the R package clusterProfiler^50,51 that adjusted for gene length bias. GO terms with adjusted P-value <0.05 were considered significantly enriched by differentially expressed genes. To assess and investigate the biochemical pathways impaired following chronic low-level arsenic exposure, the clusterProfiler package was also used to test the over-representation and statistical enrichment of differential gene expression in Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways.

Statistical analysis

Results are expressed as mean ± SE. For immunoblotting analysis, the volume of each band was quantified by densitometry in Image Lab software (Bio-Rad) using a local background subtraction method to control for variations in each lane. Intensity values for each individual sample were normalized to Gapdh, or the geometric mean of Gapdh and acetylated tubulin intensity values, which serve as loading controls. The average normalized control intensity was set at 1 to calculate fold changes. Protein expression and cellular localization for immunohistochemistry analysis were determined using ImageJ. Statistical significance for differential gene expression, protein expression, and localization in control and arsenic-exposed cells were determined using Student’s t-test (P < 0.05) using GraphPad Prism software.

Results

Chronic low-level arsenic exposure drives differential expression of several genes over time

P19 mouse embryonic stem cells were chronically cultured as previously described in individually maintained flasks for 32 weeks and exposed to 0 or 0.1 μM (7.5 ppb) arsenic as sodium arsenite.²⁷ Cells were differentiated in clean medium every 4 weeks via hanging drop to form skeletal myoblasts and neural progenitors. While no obvious morphological changes during the long-term exposure were noted, the arsenic-exposed cells were previously shown to have increased expression of pluripotency genes, such as Sox2 and Oct4, and decreased expression of genes involved in differentiation, such as Dcx, Gdf3, and Prom1.²⁷ To investigate the mechanisms responsible for the impaired differentiation, RNA sequencing was performed on day 9 differentiated cells at weeks 8, 16, and 24. Expression analysis shows that differences in RNA expression between the independent control and arsenic-exposed cell lines increases over time. At week 8, the heat map demonstrates that control and arsenic-exposed cells do not separate out into unique clusters, although some divergence is noted (Fig. 1A). However, at week 16 and week 24, sampling clustering is very evident between the control and arsenic-exposed cells, and the clusters become more pronounced over time (Fig. 1B–C).

Fig. 1.

Arsenic exposure drives sample clustering and differential expression of several genes over time. (A–C) Heat map representations of normalized read counts of differentially expressed transcripts from the three independent flasks of day 9 differentiated cells exposed to 0 (pink bars) or 0.1 μM arsenic (blue bars) at weeks 8 (A), 16 (B), and 24 (C). (D) Principal component analysis (PCA) from embryonic stem cells exposed to 0 (blue circles) or 0.1 μM arsenic (pink circles) at weeks 8 (pink dots), 16 (green dots), and 24 (blue dots). The differences in gene expression profile are represented by the distance between the samples (n = 3 independent replicates per exposure group).

In addition, principal component analysis indicates that arsenic-exposed embryonic stem cells express different genes than their control counterparts and that these differences accumulate throughout the exposure (Fig. 1D). Arsenic samples at different time points are mainly distributed along the x-axis (PC1), which accounts for 39% of the variation reported. Conversely, control samples at different time points are mainly distributed along the y-axis (PC2), which account for 20% of the variation. Interestingly, the distance between control and arsenic-treated samples is minimal at week 8 while progressively increasing along the x-axis at weeks 16 and 24 suggesting that the differences in gene expression profile accumulate throughout time (Fig. 1D).

Gene ontology indicates that the most significantly enriched functional annotations in arsenic-treated samples were related to embryonic development, regulation of neuron differentiation, pattern specification processes, and regulation of cell adhesion (Fig. 2A–B). In addition, GSEA was performed to investigate the biological processes in which differentially expressed genes (DEGs) were enriched between control and arsenic samples. This analysis revealed that DEGs with high enrichment scores were associated with cell fate commitment, embryonic skeletal system development, neuron fate commitment, and forebrain neuron differentiation (Fig. 2C).

Fig. 2.

Chronic low-level arsenic exposure impairs biological processes related to embryonic development and cellular differentiation. (A) GO pathway enrichment analysis performed on differentially expressed genes (DEGs) of arsenic-treated samples at week 24. Bar plot showing the 15 most over- or underrepresented GO pathways. (B) Dot plot representing the ratio of over- or underrepresented genes according to the number of DEGs. (C) Significantly enriched GO pathways (FDR ≤ 0.05) are presented. Gene set enrichment analysis (GSEA) performed on DEGs of arsenic-treated samples at week 24. Gene sets for cell fate commitment, embryonic skeletal system development, neuron fate commitment, and forebrain neuron differentiation were significantly enriched in arsenic-treated samples. The genes are represented on the x-axis (vertical black lines) while the enrichment score is identified on the y-axis. Significance threshold was set at FDR ≤ 0.05.

Hippo signaling pathway is disrupted following chronic low-level arsenic exposure

Pathway analysis was performed to assess DEGs that are enriched in KEGG pathways. Our results indicated that signaling pathways regulating pluripotency in stem cells are overall upregulated in arsenic-treated samples (Supplementary Fig. S1A). Interestingly, our results suggest that the Hippo signaling pathway might also be disrupted over time following chronic, low-level arsenic exposure (Fig. 3A).

Fig. 3.

Arsenic impairs expression of key Hippo pathway components. (A) KEGG pathway analysis of the Hippo signaling shows several downregulated (green) and upregulated (red) genes in arsenic-treated samples at week 24. (B–E) RNA sequencing normalized read counts of Lats2 (B), Tead3 (C), Tead4 (D), and E-cadherin (E) at weeks 8, 16, and 24. Data is presented as the average read count ± SE. Statistical differences (*) were determined using Student’s t-test (P ≤ 0.05).

The Hippo signaling pathway regulates embryonic and organ development, tissue regeneration, tumorigenesis, and maintenance of stem cell pluripotency,^52,33 and Yap is highly expressed in embryonic stem cells. In vitro studies have shown that during cellular differentiation, YAP is inactivated and sequestered in the cytoplasm, resulting in downregulation of pluripotency markers such as Nanog, Oct4, and Sox2. Read counts of the Hippo pathway’s key components Lats2, Tead3, and Tead4 indicated that their expression was not different between the control and arsenic-exposed cell lines at week 8 of exposure (Fig. 3B–D). However, at week 16 and at week 24, Lats2 expression was significantly increased in the arsenic-exposed cells by 1.6- and 1.8-fold, respectively (Fig. 3B), Tead2 was increased by 1.3- and 2.2-fold, respectively (Fig. 3C), and Tead4 was increased by 2.1- and 2.5-fold, respectively (Fig. 3D). Read counts of E-cadherin (Cdh1), an upstream regulator of the Hippo pathway, varied between the replicates at week 16, but were significantly downregulated by 2.7-fold in arsenic-treated samples at week 24 (Fig. 3E).

Chronic arsenic exposure impairs protein expression of key components of the Hippo signaling pathway

Immunoblotting was performed to investigate the effects of arsenic exposure on key proteins in the Hippo signaling pathway including the signal receiver MST1, LATS2 and the main effector YAP1. No significant differences were seen in the protein levels of MST1 (Supplementary Fig. S2) throughout the exposure. Immunoblotting of LATS2 revealed multiple bands suggesting the presence of alternative splicing or posttranslational modifications (Fig. 4A). Analysis of the strongest band suggested that arsenic exposure increased the protein expression of LATS2 throughout the exposure, with protein levels of LATS2 significantly upregulated by 6.4-fold and 3.6-fold at weeks 8 and 16, respectively (Fig. 4A–B). Interestingly, significant upregulation of LATS2 expression was reported when analyzing the weakest bands at all time points (Supplementary Fig. S3). Protein levels of active YAP1 were significantly upregulated by 3.8-fold in arsenic-treated samples at week 8. Conversely, protein expression of inactive phosphorylated YAP1 was significantly downregulated by 2.1-fold at week 16 in arsenic-treated cells while it was significantly upregulated at week 24 by 1.7-fold (Fig. 4C–D).

Fig. 4.

Arsenic exposure impairs protein expression of the Hippo pathway’s core components. (A) Protein expression of LATS2 of differentiated cells exposed to 0 or 0.1 μM arsenic at weeks 8, 16, and 24. (B) Protein expression of LATS2 expressed as fold change. (C) Protein expression of YAP1 (left) and phosphorylated YAP1 (right) of differentiated cells exposed to 0 or 0.1 μM arsenic at weeks 8, 16, and 24. (D) Protein expression of YAP1 (left) and phosphorylated YAP1 (right) expressed as fold change. Protein levels were determined by immunoblotting analysis, assessed by densitometry and normalized to GAPDH, or to GAPDH and acetylated tubulin. Statistical differences were determined using Student’s t-test (*; P ≤ 0.05).

We also examined YAP’s nuclear translocation and co-localization with its nuclear effector TEAD through immunohistochemistry in day 5 EBs derived from cells exposed to 0 or 0.1 μM arsenic. Both YAP and TEAD4 were present in the cytoplasm and nucleus (Fig. 5). While no differences in total YAP and TEAD4 expression were observed between control and arsenic-exposed EBs at week 16 (Fig. 5A), the ratio between nuclear and cytoplasmic YAP expression was significantly upregulated in arsenic-treated samples by 1.2- and 1.4-fold at weeks 16 and 28, respectively (Fig. 5A–C). Similarly, the ratio between nuclear and cytoplasmic TEAD was also significantly increased in arsenic-treated samples by 1.2-fold at weeks 16 and 28 (Fig. 5A–B, D). This data implies that the Hippo signaling pathway is dysregulated due to chronic arsenic exposure, leading to YAP activation.

Fig. 5.

Chronic low-level arsenic exposure impairs the cellular localization or Yap and Tead over time. (A–B) Representative images of YAP (red) and TEAD (green) expression and cellular localization in day 5 control and 0.1 μM arsenic-exposed EBs at weeks 16 (A) and 28 (B) of exposure. (C–D) Relative fluorescence and nuclear or cytoplasmic localization of YAP (C) and TEAD (D) were determined in ImageJ and expressed as ratio of integrated density value (IDV) of nuclear protein and IDV of cytoplasmic protein (n = 4–6 per exposure group). Statistical differences were determined using Student’s t-test (*; P ≤ 0.05).

Chronic, low-level arsenic exposure increases the expression of YAP’s target genes

Finally, we investigated levels of two YAP1 target genes, c-Myc and Ctgf. C-Myc mRNA levels were upregulated in arsenic-exposed samples at week 16 and week 24 by 2.5- and 3.6-fold, respectively (Fig. 6A). Similarly, Ctgf transcript levels were upregulated by 6.9-fold in arsenic-treated samples at week 24 (Fig. 6B). While CTGF protein expression was not significantly altered at week 16 in arsenic-treated EBs (Fig. 7A, C), it was significantly upregulated by 1.4-fold at week 28 along with clustering of CTGF-positive cells at the EBs’ periphery (Fig. 7B–C). These results suggest that chronic, low-level arsenic exposure enhances YAP activation and nuclear translocation, thereby promoting interaction with TEAD to increase transcription of target genes.

Fig. 6.

Arsenic impairs transcript levels of the YAP1 target genes c-Myc and Ctgf. (A–B) Transcript levels of c-Myc (A) and Ctgf (B) were assessed in day 9 differentiated cells exposed to 0 or 0.1 μM arsenic at weeks 8, 16, and 24 through qPCR analysis. Fold change was determined using the ddCt method to compare control versus arsenic-exposed cells at each week. Results were normalized to geometric mean of Gapdh and β2-microglobulin. Statistical differences were determined using Student’s t-test (*; P ≤ 0.05).

Fig. 7.

Chronic low-level arsenic exposure impairs CTGF protein levels at week 28. (A–B) Representative images of CTGF (green) in day 5 control and 0.1 μM arsenic-exposed EBs at weeks 16 (A) and 28 (B) of exposure. (C) Relative fluorescence was determined in ImageJ and is presented as integrated density value (IDV) ± SE (n = 3–6 per exposure group). Statistical differences were determined using Student’s t-test (*; P ≤ 0.05).

Discussion

The results of this study suggest that chronic exposure to an environmentally relevant arsenic concentration delays cellular differentiation by disrupting the Hippo signaling pathway. This occurs by enhancing the activation and nuclear translocation of YAP, the main effector in the Hippo signaling pathway, but also of its transcription factor TEAD. This increased nuclear accumulation of YAP and TEAD results in increased expression of YAP’s target genes, which are involved in pluripotency activation and maintenance.

Arsenic maintains stem cell pluripotency through impairment of the Hippo signaling pathway

The Hippo pathway is a conserved signaling pathway that controls tissue homeostasis through regulation of apoptosis, cell proliferation, migration, cellular differentiation, and stem cell renewal. Previous studies have shown that the Hippo pathway plays a role in early embryonic development and cell fate specification.^35,53,33,54 YAP is the main effector protein of the Hippo signaling pathway, and when it is phosphorylated by LATS1/2 kinases upon activation, YAP remains in the cytoplasm and is ultimately degraded. Conversely, when the Hippo pathway is inhibited, active unphosphorylated YAP is able to translocate in the nucleus and regulate transcription of downstream target genes through interaction with TEAD. Yap knockout mice are embryonic lethal, as cell division is arrested at the blastomere stage.^33,55,56 Similarly, double knockout of Yap and Taz results in cell fate specification defects and embryo death at the morula stage.^33,57,58 Yap is highly expressed in cultured stem cells and its activation leads to pluripotency maintenance, progenitor cell expansion, and inhibition of cellular differentiation.^{59–62,63,33,35,64} Conversely, during embryonic stem cell differentiation, YAP is retained in the cytoplasm, resulting in downregulation of genes involved in pluripotency such as Sox2, Oct4, and Nanog.^33,61 Recent in vitro studies have shown that overexpression of YAP in embryonic stem cells not only upregulates pluripotency markers but also maintains stem cell phenotype and inhibits differentiation even under differentiation culture conditions.^61,65 Conversely, YAP knockdown results in decreased stem cell properties and impaired phenotype.⁶¹

In the current study, RNA sequencing and KEGG analysis suggest that the Hippo signaling pathway is impaired following chronic arsenic exposure. Moreover, the sequencing analysis also confirms that arsenic exposure impairs the Wnt/β-catenin signaling pathway (Supplementary Fig. S1B) and TGF-β signaling pathway (Supplementary Fig. S1C), as has been previously described.^{24,28,29,27,32} In contrast, arsenic exposure inhibits neuronal and muscle differentiation, in part by maintaining cells in their pluripotent state (Supplementary Fig. S1A).^{23–27,66–68} For example, gene expression of the pluripotency markers Sox2 and Oct4 were increased in differentiating embryonic stem cells that had been exposed to arsenic for 12 weeks.²⁷ Consistently, our results indicate that chronic arsenic exposure upregulates signaling pathways involved in pluripotency regulation and maintenance.

Our findings also indicate that differential gene expression between independent control and arsenic-exposed samples increases as long-term exposure progresses, as few differences are observed between control and arsenic exposure at week 8. However, clustered, divergent RNA expression between the exposure groups is very apparent at weeks 16 and 24. Similar time frames are seen in other in vitro chronic arsenic exposure studies, although these typically use already differentiated cells and convert them into a transformed, malignant, or cancer stem cell phenotype. For example, HPL-1D human peripheral lung cells become malignant after 32 weeks of continuous 2 μM arsenic exposure.⁶⁹ Similarly, human WPE prostate epithelial stem cells continuously exposed to 5 μM arsenic start acquiring a cancer stem-cell phenotype after 18 weeks of exposure,⁷⁰ while kidney progenitor cells exposed to 0.5 μM arsenic develop loss of contact inhibition starting at 10 weeks of exposure.⁷¹

Arsenic impairs protein levels and cellular localization of Hippo pathway’s core proteins

Yap is crucial in the response to oxidative stress induced by exposure to different xenobiotics and in the resistance to cytotoxic agents and heavy metals, including arsenic.^72,73 Li and coworkers⁶³ reported that arsenic activated Yap independently of the canonical Hippo signaling pathway and induced a Yap-mediated disruption of tight and adherens junctions in the skin. Mice exposed to up to 200 ppm arsenic for 1 month through their drinking water had increased phosphorylation of LATS which, however, did not result in YAP inactivation.⁶³ Conversely, our results did not report significant differences in phosphorylated LATS protein levels following arsenic exposure while we observed downregulation of LATS protein levels at week 24. Importantly, the cadherin/catenin complex is involved in the initiation of the Hippo pathway.^{33,35,74–76} According to our results, transcript expression of E-cadherin was significantly downregulated in arsenic-exposed cells at week 24 and we have previously shown that arsenic treatment significantly increased N-cadherin protein levels.²⁷ These results suggest that the dysregulation of cadherin/catenin complex following chronic, low-level arsenic exposure could be involved in the activation of the Hippo signaling pathway.

Hepatocellular carcinoma cells exposed to extracellular vesicles obtained from THP-1 cells treated with arsenite had increased colony formation, migration, and invasion capacity along with reduced LATS1 and phosphorylated YAP protein levels and elevated protein expression of unphosphorylated YAP and TAZ.⁷⁷ Our findings show that arsenic exposure significantly upregulated YAP protein expression at week 8 and significantly inhibited phosphorylated YAP protein levels at week 16. In addition, the ratio between nuclear (active) and cytoplasmic (inactive) YAP was significantly upregulated at weeks 16 and 28 in arsenic-exposed samples, as determined by immunohistochemistry analysis. We observed upregulation of Lats2 transcript levels at week 24 and increased LATS2 protein expression at weeks 8 and 16. Importantly, previous studies have shown that Lats2 overexpression increased phosphorylation of Yap and induced Yap’s nuclear-to-cytoplasm translocation.^78,79 These results are in line with Li and coworkers’ findings and suggest that arsenic exposure might suppress the activation of the Hippo signaling pathway through the impairment of Lats expression and subsequently enhance YAP activation and nuclear translocation.

Arsenic-induced Yap activation enhances transcription of Yap’s target genes

To confirm our hypothesis that arsenic exposure increases YAP activation and nuclear translocation, we investigated the expression of YAP’s known downstream target genes, c-Myc and Ctgf.^35,40,42C-Myc transcript levels were significantly upregulated at weeks 16 and 24 while Ctgf transcript levels were significantly increased at week 24. Immunohistochemistry analysis confirmed upregulation of CTGF protein levels at week 28 in arsenic-treated EBs. Various studies have previously reported that arsenite exposure increases c-Myc levels in cells,^80–83 and in lung⁸⁴ and intestinal tissues³² of mice exposed to arsenite through drinking water. However, little is known about the changes in Ctgf transcript and protein levels following arsenic exposure. Consistent with our findings, Li and coworkers reported increased Ctgf mRNA levels in the skin of mice exposed to 100 ppm of sodium arsenite for 30 days⁶³; similarly, elevated CTGF protein levels were observed in aortal tissues of rats exposed to 50 ppm of sodium arsenite for 90 days.⁸⁵ These results indicate that arsenic exposure might enhance YAP activation and nuclear translocation resulting in upregulated expression of downstream target genes. Conversely, Zhou et al.⁸⁶ hypothesized that arsenic promotes YAP degradation as they observed increased ubiquitination-mediated degradation of YAP in esophageal squamous cell carcinoma cells exposed to arsenic nano-complexes. While these contrasting results could be explained by the different in vitro models used, further studies are needed to clarify the role of arsenic in the dysregulation of the Hippo signaling pathway.

Arsenic may enhance Yap activation through a non-canonical Hippo pathway

YAP is a transcription coactivator and changes in the expression of its target genes are regulated through interaction with members of the TEAD family of transcription factors in the nucleus. While the role of TEAD in the regulation of the Hippo pathway has been largely overlooked, recent studies have shown that TEAD mediates the expression of YAP’s target genes and is involved in the regulation of YAP’s biological functions.³⁹ In addition, Lin et al.⁸⁷ investigated the effects of TEAD nuclear-cytoplasmic translocation on YAP activation showing that TEAD cellular localization plays an essential role in regulating YAP’s activity and the output of the Hippo signaling pathway. Specifically, Tead knockout cells had normal YAP dephosphorylation upon Hippo pathway inhibition, yet YAP did not accumulate in the nucleus. Similarly, inhibition of TEAD and increased cytoplasmic localization prevented YAP nuclear accumulation.⁸⁷

To our knowledge, the current study is the first to investigate the effects of arsenic exposure on Tead expression and cellular localization. Our results show that Tead transcript levels are significantly upregulated following arsenic exposure at week 24 along with increased ratio between nuclear and cytoplasmic TEAD at weeks 16 and 28. These results, compounded by the findings of Lin et al.,⁸⁷ suggest that chronic low-level arsenic exposure impairs the expression and nuclear localization of YAP not only by impairing the canonical Hippo signaling pathway but also by regulating the subcellular localization of TEAD. Further studies are needed in order to clarify the role of Yap and Tead and their interplay in arsenic-induced impairment of the Hippo signaling pathway.

Conclusion

A chronic 0.1 μM arsenic exposure for 32 weeks impaired the Hippo signaling pathway in P19 mouse embryonic stem cells. Lats downregulation along with an increased ratio between nuclear and cytoplasmic YAP and TEAD suggest that arsenic impairs the Hippo signaling pathway and enhances YAP activation resulting in increased expression of YAP’s target genes Ctgf and c-Myc. Taken together, these results suggest that the impairment of the Hippo signaling pathway and increased YAP activation play a role in pluripotency maintenance and delayed cellular differentiation following arsenic exposure.

Author contributions

MCP was involved in conceptualization, conducted experiments, and drafted and edited the manuscript. BDM conducted experiments and edited the manuscript. LJB was involved in conceptualization, and drafted and edited the manuscript.

Funding

This work was supported by the National Institutes of Health (ES027651).

Conflict of interest statement

There are no conflict of interest to declare.

Data availability

Data can be made available upon reasonable request to the corresponding author.