Diphenylarsinic acid induced astrocyte-preferential cell-type-specific aberrant activation of signal transduction related to oxidative stress, MAP kinase activation, transcription factor regulation, and glutathione metabolism

Abstract

Diphenylarsinic acid (DPAA) was responsible for the 2003 arsenic poisoning incident in Japan, in which DPAA-exposed individuals experienced cerebellum-related neurological symptoms. We previously reported that DPAA targets cerebellar astrocytes rather than neurons in rats in vivo and induced the aberrant activation of particular signal transduction pathways, such as the MAP kinase and transcription factor pathway, as well as the oxidative stress response in cultured normal rat cerebellar astrocytes (NRA). Here, we examined the effects of 10 µM DPAA exposure for 96 hr in a panel of nine cell lines (HepG2, U251MG, T98G, 1321N1, SK-N-SH, SH-SY5Y, MCF7, A549, and C6) as well as NRA, and examined the DPAA-susceptible signal transduction pathways: oxidative-stress responsive factors [heme oxygenase-1 (HO-1), Hsp70, superoxide dismutase-1, and catalase), MAP kinases (ERK1/2, p38MAPK, and SAPK/JNK), transcription factors (CREB, c-Jun, and c-Fos), glutathione (GSH), and GSH-related enzymes (glutamate-cysteine ligase and glutathione synthetase). In NRA, DPAA significantly activated these signal transduction pathways. Although there were cell-type specificities in susceptibility to DPAA, multivariate clustering analyses classified NRA, rat glioma-derived C6, and two human glioma-derived cell lines, U251MG and 1321N1, into an identical group. These results suggest that DPAA might affect cellular signal transduction preferentially in astrocytes among the diverse types of cells.

INTRODUCTION

Diphenylarsinic acid (DPAA) is a pentavalent organic arsenic that was identified as a causative agent in the 2003 arsenic poisoning incident in Kamisu, Ibaraki, Japan (Ishii et al., 2004), in which it leaked from a buried concrete block and contaminated well water. The concentration of arsenic, including DPAA and some metabolites, reached 4.5 mg As/L, which is equivalent to 60 µM arsenicals. DPAA is an artificial arsenic compound that was used as a raw material for chemical weapons such as diphenylcyanoarsine and diphenylchloroarsine, which were developed during World War II. Residents who consumed the DPAA-contaminated well water experienced neurological symptoms such as recent memory and cognitive disturbances, vertigo, somnipathy, uncontrolled or repetitive eye movements, myoclonus, and gait disturbance (Ishii et al., 2004), which suggests neurological symptoms, including cerebellar dysfunction. Although the biological effects of DPAA on human health remain unclear, studies have found abnormal ocular movement (Nakamagoe et al., 2013; Nakamagoe et al., 2006) and decreased regional cerebral blood flow (Ishii et al., 2019) as residual central nervous system damage caused by DPAA intoxication. Even after two decades, some residual symptoms are still observed. It should be noted that DPAA is not a nauseating and vomiting-inducing chemical warfare agent, but a possible raw material for chemical weapons. The characteristics of DPAA revealed in our studies and those of others are unrelated to the pharmacological properties of chemical weapons.

In experimental animal models, developmental subchronic exposure to DPAA induced reversible increases in exploratory behavior and persistent impairment of passive avoidance in rats (Negishi et al., 2013). A previous study showed a decrease in GSH in the blood, liver, and brain and an impairment of spatial learning ability (Ozone et al., 2010). Although subchronically DPAA-exposed mice exhibited increased ambulatory activity, an increased shuttle-type discrete conditioned avoidance rate, and reduced coordination ability on fixed and rotating rods (Umezu et al., 2012), mice receiving a single oral administration of DPAA altered striatal dopamine levels (Umezu and Shibata, 2021). Our previous study revealed that DPAA increased the synthesis and release of neuroactive and vasoactive peptides in rat cerebellar astrocytes in vitro and in vivo (Negishi et al., 2012). Moreover, exposure to DPAA at 10 µM induced a transient increase in cell viability, followed by cell death, and an aberrant activation of cellular signal transduction, including the induction of oxidative stress-responsive proteins (Nrf2, heme oxygenase-1, and Hsp70), phosphorylation of mitogen-activated protein kinases (ERK1/2, p38MAPK, and SAPK/JNK), upregulation/phosphorylation of transcription factors (CREB, c-Jun, and c-Fos), and the release of brain-active cytokines (MCP-1, adrenomedullin, FGF2, CXCL1, and IL-6) in cultured normal rat cerebellar astrocytes (NRA) (Negishi et al., 2016). In addition, ERK1/2, p38MAPK, and SAPK/JNK regulated DPAA-induced aberrant signal transduction in NRA (Negishi et al., 2017). Our studies further demonstrated that DPAA induces NRA-like aberrant activation of cellular signal transduction in cultured normal human cerebellar astrocytes (NHA), although they exhibited greater dose- and time-dependent resistance compared with that in NRA (Sasaki et al., 2022). In particular, NHA required 50 µM DPAA exposure for 96 hr or 10 µM for 288 hr to achieve significant activation, whereas NRA only required a 10 µM exposure for 96 hr. Despite being a commercially available and potentially promising cell model for examining DPAA neurotoxicity in humans, NHA has significant limitations, including its cost-ineffectiveness and medium-specific culture, inconsistent availability, potential ethical issues, and low growth capacity which impedes large-scale culturing. Therefore, alternative human cells that are more convenient, easy-to-handle, and relevant are needed. DPAA-induced aberrant cellular activation has only been observed in astrocytes, although there were several in vitro studies of DPAA using the hepatoblastoma HepG2 cell line, which focused on the cytotoxicity and intracellular metabolism of DPAA and glutathione (GSH) (Kinoshita et al., 2006; Ochi et al., 2006) as well as DPAA-induced glutaminase downregulation (Kita et al., 2009; Kita et al., 2007, 2012).

In the present study, we prepared a human tumor cell line panel consisting of the HepG2 (hepatoblastoma), U251MG and T98G [glioblastoma (WHO Grade IV) and representing highly malignant astrocytoma], 1321N1 [a lower-grade astrocytoma (WHO Grade II)] (Macintyre et al., 1972), SK-N-SH (neuroblastoma), SH-SY5Y (neuroblastoma), MCF7 (breast adenocarcinoma), and A549 (lung adenocarcinoma) cell lines to assess their vulnerability/susceptibility to DPAA exposure, as well as that of rat cells (NRA and C6 glioblastoma). Based on the cytotoxicity of DPAA in these cells and our previous results, the cells were exposed to DPAA under identical conditions (10 µM DPAA for 96 hr in the chemically defined serum-free medium), and the protein expression of oxidative stress-responsive factors [HO-1, Hsp70, superoxide dismutase-1 (SOD-1), and catalase], the phosphorylation of MAP kinases (p38MAPK, SAPK/JNK, and ERK1/2), the phosphorylation and expression of transcription factors (CREB, and c-Jun, and c-Fos), the intracellular/extracellular levels of GSH, and the expression of GSH-related enzymes [glutamate-cysteine ligase catalytic subunit (GCLC), glutamate-cysteine ligase modifier subunit (GCLM), and glutathione synthetase (GSS)] were analyzed. We preliminarily found that 10 µM DPAA exposure for 96 hr increased the protein expression of SOD-1 and decreased that of catalase in NRA, which are important anti-oxidative stress factors. A previous study also demonstrated a DPAA-induced decrease in catalase activity in the liver of mice, although its activity was under the detection limit in the brain (Ozone et al., 2010). Thus, SOD-1 and catalase were added as DPAA-related oxidative stress-response factors in the present study. Previous studies have shown the importance of GSH in DPAA toxicity. These lines of evidence implied that DPAA may be activated intracellularly by glutathione, in which pentavalent arsenic in DPAA is reduced to its trivalent form in rat liver (Kobayashi and Hirano, 2013; Ochi et al., 2006). We measured GSH using an enzyme (glutathione reductase)-based assay in DPAA-exposed NRA and found that DPAA exposure markedly increased the extracellular GSH concentration without a significant change in intracellular GSH concentration. This suggests a marked increase in GSH synthesis and equivalent secretion. This is the reason that GSH-related enzymes were investigated in the present study. The present study aimed to determine whether DPAA-induced activation of cellular signal transduction is astrocyte-preferential or more ubiquitous and to identify more convenient human-derived astrocyte models among the tested human glioma-derived cells, U251MG, T98G, and 1321N1 that respond similarly to DPAA exposure as NRA.

MATERIALS AND METHODS

Chemicals

Diphenylarsinic (DPAA; CAS RN: 4656-80-8; C₁₂H₁₁AsO₂) (> 99% purity) acid was purchased from Tri Chemical Laboratories (Yamanashi, Japan).

Cell culture

Normal rat cerebellar astrocytes (NRA) were derived from Wistar rats at postnatal day 2 as described in our previous studies (Negishi et al., 2017; Sasaki et al., 2022). Briefly, enzymatically dissociated cerebellar cells were cultured in DMEM/F-12 (Thermo Fisher Scientific, MA, USA) containing 10% fetal bovine serum (FBS) for 7–10 days. In these cultures, astrocytes were dominant among cerebellar cells, including neurons and microglia because of their high proliferation rate. To enhance astrocyte purity, the cells were subsequently passaged 5–7 times to obtain a sufficient number of astrocytes in DMEM/F-12/10%FBS and subsequently frozen and preserved until use. HepG2 (ATCC® HB-8065™) (Cellular Engineering Technologies, Iowa, USA), U251MG (DS Pharma Biomedical Co., Ltd., Japan), T98G (DS Pharma Biomedical), 1321N1 (DS Pharma Biomedical), SK-N-SH (ATCC® HTB-11™) (DS Pharma Biomedical), SH-SY5Y (ATCC® CRL-2266™) (DS Pharma Biomedical), MCF7 (Summit Pharmaceuticals International, Co., Ltd., Japan), A549 (Summit Pharmaceuticals International), and C6 (Summit Pharmaceuticals International) cells were subcultured in DMEM/F-12/10% FBS, in which all cell lines grew efficiently.

DPAA exposure

DPAA exposure was performed by replacing the culture medium with DPAA-containing DMEM/F-12/ITS-X. At the final passage, the cells were plated at appropriate cell densities as follows: NRA (500/mm²), HepG2 (200/mm²), U251MG (200/mm²), T98G (50/mm²), 1321N1 (100/mm²), SK-N-SH (500/mm²), SH-SY5Y (500/mm²), MCF7 (300/mm²), A549 (30/mm²), and C6 (25/mm²). These seeding densities were established to avoid over cell confluency, which could result in cell detachment after 192 hr (8 days) when samples were prepared for western blot analysis. They were then cultured in serum-containing medium (DMEM/F-12/10% FBS) for 48 hr and in serum-free medium [DMEM/F-12 with insulin-transferrin-selenite supplement (Invitrogen)] for another 48 hr. The cells were then exposed to DPAA at 0.4–100 µM for 48, 96, or 144 hr for cell viability analyses or at 10 µM for 96 hr for protein expression analyses by western blotting and glutathione assays after exchanging the culture medium for DPAA-containing serum-free medium. In each experiment, the control cells were exposed to the vehicle (dimethyl sulfoxide) at the appropriate concentration (0.01% in serum-free medium).

Cell viability

The cell viability of each DPAA-exposed cell line was quantified using the CellTiter-Blue™ Cell Viability Assay (Promega, Wisconsin, USA). The CellTiter-Blue™ reagent was added directly to the culture medium, and resorufin (pink) reduced from resazurin (dark blue) in viable cells was measured fluorometrically (535_Ex/595_Em) using a plate reader.

Western blotting

Western blotting was performed as described in our earlier study (Negishi et al., 2016) to analyze protein expression and phosphorylation levels. Briefly, cultured astrocytes exposed to DPAA were lysed by CelLytic MT (Sigma-Aldrich) supplemented with protease inhibitors [cOmplete mini ULTRA (Roche Diagnostics, Mannheim, Germany) and protease inhibitor cocktail (Sigma-Aldrich)] and phosphatase inhibitors [PhosStop (Roche)]. Collected cell lysates were thoroughly lysed by ultrasonic disruption and freeze/thaw cycles followed by centrifugation at 8000 × g. Protein samples soluble in the lysis buffer were collected and, after adjusting the protein concentration, denatured with a sample buffer containing sodium dodecyl sulfate (SDS) and 2-mercaptoethanol (2ME). The protein samples were separated by SDS-polyacrylamide gel electrophoresis and blotted onto polyvinylidene fluoride Immobilon P membranes (Merck Millipore, MA, USA). Membranes were incubated in Tris-buffered saline (TBS) containing 0.1% Tween-20 and 5% bovine serum albumin or PVDF Blocking Reagent (Toyobo, Osaka, Japan) for blocking. The membranes were incubated with a primary antibody diluted in BSA-containing blocking buffer or Can Get Signal Solution 1 (Toyobo). In the present study, the following primary antibodies were used: anti-HO-1 rabbit polyclonal antibody (ab13243, Abcam, Cambridge, UK); anti-Hsp70 mouse monoclonal antibody (SPA-810 [clone C92F3A-5], Enzo Life Sciences, Pennsylvania, USA); anti-SOD-1 rabbit monoclonal antibody, (ab51254 [clone EP1727Y], Abcam); anti-catalase rabbit polyclonal antibody (No. 8841, Cell Signaling Technology (CST), Massachusetts, USA); anti-phospho-ERK1/2 rabbit monoclonal antibody (No. 4370 [clone D13.14.4E], CST); anti-ERK1/2 mouse monoclonal antibody (No. 9107 [clone 3A7], CST); anti-phospho-p38MAPK rabbit monoclonal antibody (No. 4511 [clone D3F9], CST); anti-p38MAPK rabbit monoclonal antibody (No. 8690 [clone D13E1], CST); anti-phospho-SAPK/JNK rabbit monoclonal antibody (No. 4668 [clone 81E11], CST); anti-SAPK/JNK rabbit monoclonal antibody (No. 9258 [clone 56G8], CST); anti-phospho-CREB rabbit monoclonal antibody (No. 9198 [clone 87G3], CST); anti- CREB mouse monoclonal antibody (No. 9104 [clone 86B10], CST); anti-c-Jun rabbit monoclonal antibody (No. 9165 [clone 60A8], CST); anti-c-Fos rabbit monoclonal antibody (No. 2250 [clone 9F6], CST); anti-GCLC mouse monoclonal antibody (ab55435 [mouse ascites], Abcam); anti-GCLM rabbit monoclonal antibody (ab126704 [clone EPR6667], Abcam); anti-GSS rabbit monoclonal antibody (ab133592 [clone EPR6563], Abcam); and anti-β-actin mouse monoclonal antibody (A1978 [clone AC-15], Sigma–Aldrich). Depending on the primary antibody, horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (Jackson ImmunoResearch Laboratories, PA, USA) or HRP-conjugated goat anti-rabbit IgG (Jackson ImmunoResearch) was appropriately diluted in the blocking buffer or Can Get Signal Solution 2 (Toyobo). An Immobilon Western Chemiluminescence HRP Substrate (Millipore) was used to visualize immunoreactive molecules and recorded using an EZ-Capture and LuminoGraph III Lite systems (Atto, Japan). Protein expression analyses employed β-actin as an internal loading control. Chemiluminescent images of immunoreactive signals were quantified using ImageJ software. The values for phosphorylated and total protein expression of ERK1/2 and SAPK/JNK with isoforms of p44/p44 and JNK1/2/3, respectively, were summed up without distinction.

Glutathione concentration

Intracellular and released total glutathione (GSH) levels were measured using a glutathione (GSSG/GSH) detection kit (Enzo Life Sciences, New York, USA), which is a glutathione reductase-based assay. Cell lysates at equal protein concentrations or culture medium were deproteinized by adding an appropriate volume of 5% metaphosphoric acid and centrifuged. The supernatant of each sample was subjected to the assay based on the product manual without discrimination of reduced and oxidized forms. 5-Thiobenzoic acid (TNB) produced from 5-5’-dithiobis[2-nitrobenzoic acid] (DTNB) and glutathione and TNB released from glutathione-conjugated TNB by glutathione reductase were measured colorimetrically (absorbance at 405 nm) using a plate reader.

Statistical analyses

The effect of DPAA treatment on cell viability in each cell line was analyzed by a one-way ANOVA at each exposure time. Protein expression levels and glutathione concentrations were analyzed by a two-way ANOVA followed by Dunnett’s test using SigmaPlot 15.0 (Systat Software, California, USA). Graph drawing was performed using R (R 4.3.3)/RStudio (2024.04.2) (R Core Team, 2024). Cluster analyses using Ward’s methods were also performed with R/RStudio based on the logarithm of the ratio of each protein expression level, extracellular or intracellular glutathione level, or cell viability in DPAA-exposed cells to that of the control cells [log2 (DPAA/Cont)] after standardization into the z-score in each cell line.

RESULTS

In the present study, we examined the adverse effects of DPAA on cell viability, the expression and/or phosphorylation of oxidative stress-responsive factors (HO-1, Hsp70, SOD-1, catalase), MAP kinases (ERK1/2, p38MAPK, and SAPK/JNK), and transcription factors (CREB, c-Jun, and c-Fos), as well as intracellular and extracellular glutathione levels in cultured cells derived from established cell lines, including HepG2, U251MG, T98G, 1321N1, SK-N-SH, SH-SY5Y, MCF7, A549, and C6, as well as NRA.

DPAA-induced increase and decrease in cell viability

The cell viability of various types of cultured cells exposed to 0–100 µM DPAA for 48, 96, or 144 hr was determined and compared with that of corresponding control (vehicle only) cells at each exposure time. Following DPAA exposure, NRA exhibited significantly increased (6.25–100 µM for 48 hr; 3.125–25 µM for 96 hr; 3.125–6.25 µM for 144 hr) and decreased (50–100 µM for 96 hr; 25–100 µM for 144 hr) cell viability (Fig. 1A). HepG2 cells showed significantly increased (12.5–25 µM for 144 hr) and decreased (100 µM for 48 hr, 96 hr, and 144 hr) cell viability following DPAA exposure (Fig. 1B). Among the human glioma-derived cell lines, U251MG exhibited significantly increased (12.5–100 µM for 48 hr; 12.5–50 µM for 96 hr) and decreased (100 µM for 96 hr; 50–100 µM for 144 hr) cell viability after DPAA exposure (Fig. 1C). T98G showed significantly increased (12.5–100 µM for 48 hr; 12.5–100 µM for 96 hr; 6.25–50 µM for 144 hr) and decreased (100 µM for 144 hr) cell viability (Fig. 1D). 1321N1 also showed significantly increased (6.25–12.5 µM for 96 hr; 3.125–6.25 µM for 144 hr) and decreased (50–100 µM for 48 hr; 50–100 µM for 96 hr; 25–100 µM for 144 hr) cell viability (Fig. 1E). In contrast, after exposure to DPAA in neuroblastoma-derived cell lines, SK-N-SH showed significantly decreased (100 µM for 48 hr; 12.5–100 µM for 96 hr; 6.25–100 µM for 144 hr) cell viability (Fig. 1F) and SH-SY5Y exhibited slightly but significantly increased (6.25 µM for 144 hr) and decreased (50–100 µM for 48 hr; 25–100 µM for 96 hr; 25–100 µM for 144 hr) cell viability (Fig. 1G). Following DPAA exposure, MCF7 exhibited significantly increased (12.5 µM for 144 hr) and decreased (100 µM for 144 hr) cell viability (Fig. 1H). A549 showed significantly decreased (100 µM for 96 hr; 50–100 µM for 144 hr) cell viability (Fig. 1I). Rat glioma C6 exhibited significantly decreased (12.5–100 µM for 48 hr; 6.25–100 µM for 96 hr; 0.78–100 µM for 144 hr) cell viability following DPAA exposure (Fig. 1J). In some cell lines, DPAA treatment at sub-cytotoxic doses induced morphological changes (Fig. S1). Based on the DPAA cytotoxicity profile of DPAA in NRA and the nine tested cell lines, cultured cells exposed to DPAA at 10 µM for 96 hr and control cells were subjected to western blot analyses to assess cell-type specificity in the DPAA-sensitive protein expression and/or phosphorylation.

Fig. 1

Effect of exposure to diphenylarsinic acid (DPAA) at 0–100 μM for 48 (blue), 96 (red), and 144 (green) hours on the cell viability of NRA (A), HepG2 (B), U251MG (C), T98G (D), 1321N1 (E), SK-N-SH (F), SH-SY5Y (G), MCF7 (H), A549 (I), and C6 (J) cells. Values are indicated as means ± SEM (n = 8 or 12 from 2 or 3 separate experiments, respectively). Asterisk (*): P < 0.05 vs. 0 μM vehicle control in each exposure period.

Protein expression of β-actin as a housekeeping protein in various cultured cells exposed to DPAA

Protein expression analyses by western blotting used β-actin as an internal loading control based on standard protocols. There was little effect of DPAA on the β-actin expression within each cell line, although HepG2 exhibited lower β-actin expression levels compared with the other cell lines, as shown in each panel. Because the origins of the cell lines used were different, variation in the expression levels of β-actin protein appeared reasonable, as well as other functional proteins examined in this study. Consequently, differences among cell types in protein expression will not be described in detail hereafter in the present study, although some significant main effects of cell type were detected in two-way ANOVAs because of cell type-specific diversities.

DPAA-induced increase in the expression of oxidative stress-responsive factors in various cultured cell lines

DPAA exposure at 10 µM for 96 hr induced a significant increase in heme oxygenase-1 (HO-1) expression in NRA as previously reported (Fig. 2A) (Negishi et al., 2016). DPAA exposure induced a significant increase in HO-1 expression in U251MG and C6 cells (Fig. 2A). DPAA exposure also induced a significant increase in Hsp70 expression in NRA, HepG2, T98G, 1321N1, MCF7, and C6 cells (Fig. 2B), whereas it showed little effect on SOD-1 expression (Fig. 2C). Two-way ANOVA revealed a significant main effect of DPAA exposure on catalase protein expression, suggesting a slight DPAA-induced decrease regardless of the cell type (Fig. 2D).

Fig. 2

Effect of DPAA exposure on the protein expression of HO-1 (A), Hsp70 (B), SOD-1 (C), and catalase (D) in NRA, HepG2, U251MG, T98G, 1321N1, SK-N-SH, SH-SY5Y, MCF7, A549, and C6 cells exposed to DPAA at 10 μM for 96 hr. Values were shown as relative values to that in DPAA-exposed NRA (means ± SD, n = 4 from two separate experiments). Asterisk (*): P < 0.05 vs. 0 μM vehicle control for each cell type. The label “DPAA” in each panel indicates the P-value of the main effect of DPAA exposure.

DPAA-induced increase in the phosphorylation of MAP kinases (ERK1/2, p38MAPK, and SAPK/JNK) in various cultured cell lines

DPAA exposure at 10 µM for 96 hr significantly increased ERK1/2 phosphorylation in NRA, U251MG, 1321N1, MCF7, and C6 (Fig. 3A), while it significantly decreased total ERK1/2 protein expression in NRA, HepG2, and C6 (Fig. 3B). DPAA exposure significantly increased p38MAPK phosphorylation in NRA, 1321N1, and C6 (Fig. 3C), while there was no DPAA-related change in the total p38MAPK protein expression (Fig. 3D). DPAA exposure significantly increased SAPK/JNK phosphorylation in NRA, MCF7, and C6 (Fig. 3E) with little change in the total SAPK/JNK protein expression (Fig. 3F).

Fig. 3

Effect of DPAA exposure on the protein phosphorylation (A, C, and E) and expression (B, D, and F) of ERK-1 (A and B), p38MAPK (C and D), and SAPK/JNK (E and F) in NRA, HepG2, U251MG, T98G, 1321N1, SK-N-SH, SH-SY5Y, MCF7, A549, and C6 cells exposed to DPAA at 10 μM for 96 hr. Values were shown as relative values to that in DPAA-exposed NRA (means ± SD, n = 4 from two separate experiments). Asterisk (*): P < 0.05 vs. 0 μM vehicle control for each cell type. The label “DPAA” in each panel indicates the P-value of the main effect of DPAA exposure.

DPAA-induced increase in the phosphorylation of CREB and the expression of c-Fos and c-Jun in various cultured cell lines

DPAA exposure at 10 µM for 96 hr slightly increased CREB phosphorylation in NRA, U251MG, T98G, 1321N1, SH-SY5Y, and C6 cells as a two-way ANOVA suggested a significant main effect of DPAA (Fig. 4A), whereas no effect of DPAA on total CREB protein expression (Fig. 4B). DPAA exposure at 10 µM for 96 hr resulted in a significant increase in c-Jun protein expression in NRA, U251MG, T98G, 1321N1, and C6 cells (Fig. 4C) and significantly increased c-Fos protein expression in NRA and C6 (Fig. 4D).

Fig. 4

Effect of DPAA exposure on the protein phosphorylation (A) and expression (B) of CREB, c-Jun (C), and c-Fos (D) protein expression in NRA, HepG2, U251MG, T98G, 1321N1, SK-N-SH, SH-SY5Y, MCF7, A549, and C6 cells exposed to DPAA at 10 μM for 96 hr. Values were shown as relative values to that in DPAA-exposed NRA (means ± SD, n = 4 from two separate experiments). Asterisk (*): P < 0.05 vs. 0 μM vehicle control for each cell type. The label “DPAA” in each panel indicates the P-value of the main effect of DPAA exposure.

Effect of DPAA exposure on glutathione production and secretion in various cultured cells

Previous studies suggested that GSH might be involved in DPAA toxicity in vitro (Ochi et al., 2006) and DPAA exposure decreased brain glutathione concentration in vivo (Negishi et al., 2013; Ozone et al., 2010). Therefore, the intracellular or extracellular concentration of total GSH in DPAA-exposed cells was quantified using a glutathione reductase-based assay, in which the reduced GSH and its oxidized form, GSSG, were measured without distinction. The intracellular GSH concentration following DPAA exposure at 10 µM for 96 hr (Fig. 5A) was significantly decreased in 1321N1, MCF7, and C6 cells, increased in SH-SY5Y cells, and minimally changed in other cell lines, including NRA. In contrast, the extracellular GSH concentration released from the cells (Fig. 5B) was significantly increased in NRA, 1321N1, SH-SY5Y, and MCF7 cells. When the released GSH concentration was calibrated by the protein concentration (mg protein) of the cells, it significantly increased in NRA, U251MG, 1321N1, and MCF7 cells (Fig. 5C).

Fig. 5

DPAA-induced alterations in intracellular glutathione (GSH) [nmol/mg of cellular protein (mgP)] (A), GSH released in the culture medium (released GSH) (μM) (B), and released GSH normalized to protein concentration (nmol/mgP) (C) in NRA, HepG2, U251MG, T98G, 1321N1, SK-N-SH, SH-SY5Y, MCF7, A549, and C6 cells exposed to DPAA at 10 μM for 96 hr. Values were shown as relative values to that in DPAA-exposed NRA (means ± SD, n = 4 from two separate experiments). Asterisk (*): P < 0.05 vs. 0 μM vehicle control for each cell type. The label “DPAA” in each panel indicates the P-value of the main effect of DPAA exposure.

Effect of DPAA-exposure on the protein expression of glutathione-related enzymes in various cultured cell lines

The disruption of glutathione synthesis/secretion might represent a substantial outcome of DPAA-induced aberrant activation of cellular signal transduction. To identify the underlying mechanism, the expression of three GSH-related enzymes [glutamate-cysteine ligase catalyze subunit (GCLC), glutamate-cysteine ligase modifier subunit (GCLM), and glutathione synthetase (GSS)]. DPAA exposure significantly increased the protein expression of GCLC only in HepG2 cells (Fig. 6A) and GCLM in HepG2, U251MG, T98G, and 1321N1 cells (Fig. 6B). It should be noted that DPAA induced a 2.6-fold increase in GCLM protein in NRA, although statistical significance was not evident because of the low relative values. DPAA exposure showed little effect on GSS protein expression (Fig. 6C).

Fig. 6

Effect of DPAA exposure on the protein expression level of GCLC (A), GCLM (B), and GSS (C) in NRA, HepG2, U251MG, T98G, 1321N1, SK-N-SH, SH-SY5Y, MCF7, A549, and C6 cells exposed to DPAA at 10 μM for 96 hr. Values were shown as relative values to that in DPAA-exposed NRA (means ± SD, n = 4 from two separate experiments). Asterisk (*): P < 0.05 vs. 0 μM vehicle control for each cell type. The label “DPAA” in each panel indicates the P-value of the main effect of DPAA exposure.

Profiling of DPAA-induced aberrant cellular activation in various cultured cells

The expression of each protein normalized to β-actin expression in DPAA-exposed and control cells in the 10 tested cell lines was converted into standardized values (z-score) and expressed as a grayscale heatmap (Fig. 7A) to survey the diversity in protein expression levels among tested cells and to view the profiles affected by DPAA exposure (the combined chemiluminescent images are also shown in Fig. S2). For a more direct view of the effects of DPAA exposure, DPAA-induced changes in 21 experimental values were calculated as the z-score of the logarithm of the ratio [log2 (DPAA/Cont)] and subjected to hierarchical clustering using Ward’s methods. These included protein expression levels of 17 targets; intracellular or released glutathione levels; and relatively low-dose (12.5 μM) or high-dose (50 μM) DPAA-exposed cell viability (Fig. 7B). Cell viability was chosen as an experimental index because DPAA distinctively caused an increase or decrease in cell viability following relatively low-dose (12.5 μM) or high-dose (50 μM) DPAA exposure, respectively, in certain cell lines, including NRA. Ward’s hierarchical cluster analysis of 10 cell types successfully identified two large groups. In the NRA-containing group consisting of NRA, C6, U251MG, 1321N1, and MCF7, 4 of 5 cell types were astrocyte-related cells and 4 among 5 astrocyte-related cells tested were assigned. Rat glioma-derived C6 was classified closest to NRA, followed closely by two human-derived astrocytes (U251MG and 1321N1) and one breast adenocarcinoma (MCF7). Although cluster analyses indicated that DPAA activated signal transduction pathways such as MAP kinase activation in C6 like NRA (upper half of the heatmap), DPAA exposure induced noticeably different effects in certain factors. Specifically, as already mentioned, DPAA exposure at 12.5 μM increased the cell viability of NRA but it decreased that of C6. DPAA exposure also elevated extracellular GSH concentrations in NRA, but had no effect in C6 cells, whereas it decreased intracellular GSH levels in C6 cells without affecting NRA (lower half of the heatmap). On the other hand, in U251MG and 1321N1 cells, DPAA exposure increased MAP kinase phosphorylation (upper half of the heatmap) and other responses (bottom quarter) similar to those observed in NRA, although the intensity of these effects was lower compared with that in NRA. Furthermore, cluster analysis successfully classified 21 experimental values into factors increased by DPAA exposure, i.e. DPAA-responsive factors, such as oxidative stress-responsive factors and MAP kinase activation-related factors, particularly in the astrocyte-rich cell-type group, and the other factors.

Fig. 7

Protein expression profiling of DPAA-induced activation of oxidative stress-responsive proteins, MAP kinases, transcription factors, and GSH synthesis-related enzymes in NRA, HepG2, U251MG, T98G, 1321N1, SK-N-SH, SH-SY5Y, MCF7, A549, and C6 cells as shown by a grayscale heatmap (A), in which the depth of each tile indicates a z-score for the expression of each protein normalized by β-actin expression. Hierarchical clustering analysis using Ward’s methods (B) for visualizing the cell-type specific effects of DPAA exposure in NRA, HepG2, U251MG, T98G, 1321N1, SK-N-SH, SH-SY5Y, MCF7, A549, and C6 cells using the z-score from the logarithm of the ratio [log2 (DPAA/Cont)] of 21 experimental values; each protein expression level, intracellular or released glutathione level (GSHi or GSHr, respectively), and relative low-dose (12.5 μM) or high-dose (50 μM) DPAA-exposed cell viability.

DISCUSSION

This study explored the cell-type-specific effects of DPAA in eight human cell lines (HepG2, U251MG, T98G, 1321N1, SK-N-SH, SH-SY5Y, MCF7, and A549), a rat cell line (C6), and rat cerebellar astrocytes, NRA. We found that DPAA-induced aberrant activation of cellular signal transduction preferentially occurred in astrocytes.

We first identified cerebellar astrocytes as a target of DPAA in cerebellar cell culture, which consisted of neurons, astrocytes, and other cells, and in the in vivo cerebellum (Negishi et al., 2012), in which DPAA exposure increased oxidative stress-responsive proteins, such as HO-1 and Hsp70, preferentially in astrocytes. Selectively cultured rat cerebellar astrocytes, NRA, were then used to determine the effect of DPAA in our previous studies (Negishi et al., 2017, Negishi et al., 2016). DPAA exposure significantly increased the phosphorylation of MAP kinases (ERK1/2, p38MAPK, and SAPK/JNK), and the expression and/or phosphorylation of several transcription factors (CREB, c-Jun, and c-Fos), as well as oxidative stress-responsive proteins. In the present study, these DPAA-responsive factors in NRA were used as endpoints to examine the effect of DPAA in rat or human-derived cells originating from various cell types. In addition, the expression of SOD-1 and catalase was examined as additional anti-oxidative stress factors, and GSH-related enzymes, GCLC, GCLM, and GSS, and intracellular and extracellular GSH levels in DPAA-exposed cells were measured based on the possibility that GSH plays an important role in DPAA-induced aberrant activation of cellular signal transduction.

As expected, NRA showed increased protein expression of HO-1 and Hsp70, increased phosphorylation of ERK1/2, p38MAPK, and SAPK/JNK, increased CREB phosphorylation, and increased protein expression of c-Jun and c-Fos following DPAA exposure. This indicates that these factors are reproducibly DPAA-responsive factors, although for some factors, no statistical significance was observed in NRA because of its extremely low values relative to other cell types, even if there was an apparent alteration. Furthermore, DPAA exposure promoted GSH secretion in NRA, which suggests the importance of GSH metabolism in DPAA toxicity. Indeed, this is the first study to demonstrate DPAA exposure and GSH secretion in NRA. A discussion of the mechanism of DPAA-induced aberrant cellular activation based on the present results alone is difficult. Interestingly, there was considerable diversity in GSH metabolism depending on the type of cell lineage. It appears that GSH release was increased by DPAA exposure in NRA without any changes in the expression of GSH synthesis-related enzymes. This occurred because the absolute values were small compared with those in other cell lines, such as HepG2 and T98G. There was a considerable increase in the protein expression GCLM and GSS following DPAA exposure, although no significant effect of DPAA was evident in NRA, which was implied by the red panels of GCLM and GSS in NRA in the heatmap of the z-scores of log₂(DPAA/Cont) (Fig. 7B). These results suggest that DPAA interferes with GSH metabolism in NRA, which might stimulate GSH synthesis through activation of cystine/glutamate transporter xCT as well as GSH synthesis-related enzymes and the secretion of newly synthesized GSH and possibly intracellularly generated GSH-conjugated DPAA (Kobayashi and Hirano, 2013; Ochi et al., 2006) into the extracellular space through specific transporters such as MRP-1 and P-gp, respectively. This will be an important issue for further study.

In previous studies with the hepatoblastoma-derived cell line HepG2, glutaminase C down-regulation occurred following DPAA exposure at 250–500 µM for 24 hr (Kita et al., 2007; Kita et al., 2009). DPAA promoted glutaminase C degradation mediated by mitochondrial Lon protease (Kita et al., 2012). In the present study, Hsp70 and two GSH-related enzymes (GCLC and GCLM) were determined to be DPAA-responsive (up-regulated) factors in HepG2 in addition to glutaminase C (a down-regulated factor). In the glioblastoma cell line U251MG (Pontén and Macintyre, 1968), DPAA exposure at 10 µM for 96 hr increased cell viability, protein expression of HO-1, c-Jun, and GCLM, phosphorylation of ERK1/2, and GSH secretion. Other factors, such as Hsp70 expression and p38MAPK phosphorylation, also exhibited NRA-like alterations following DPAA exposure; however, these changes were not statistically significant because of their low absolute values, despite appearing to be altered. In T98G (Stein, 1979), another glioblastoma-derived cell line, DPAA exposure resulted in increased Hsp70, c-Jun, and GCLM expression. Although DPAA increased GCLM protein expression in T98G, no change in intracellular of released GSH concentration was observed in DPAA-exposed T98G cells. Unlike other cells, T98G released a large amount of GSH into the culture medium, which was comparable to that observed in DPAA-exposed NRA, regardless of DPAA exposure. A human astrocytoma-derived 1321N1 has also been used in toxicological studies (Kim et al., 2012) as well as in oncological studies of anticancer drugs. Exposure of 1321N1 to DPAA increased Hsp70 and c-Jun protein expression and ERK1/2 and p38MAPK phosphorylation. DPAA exposure increased GCLM protein expression and GSH secretion, while depleting intracellular GSH, unlike NRA. The human neuroblastoma cell line SK-N-SH (Ross et al., 1983) and its subline SH-SY5Y are widely used not only in neuroblastoma-targeting anticancer drug-related research (Stefàno et al., 2022; Li et al., 2022), but also in neurotoxicological fields (Janyou et al., 2015; Humphrey et al., 2005; Wisessaowapak et al., 2021). In SK-N-SH, DPAA exposure resulted in no change in the molecular endpoints described above. However, in the closely related cell line SH-SY5Y, DPAA caused a slight increase in CREB phosphorylation and increased intracellular and extracellular GSH levels without any significant change in GSH-related enzymes. SK-N-SH and SH-SY-5Y are well-known neuroblastoma cell lines that are undifferentiated under normal culture conditions, such as DMEM/F-12 supplemented with 10% FBS or ITS-X (serum-free medium) used in the present study. They can be differentiated by retinoic acid (Li et al., 2022; Påhlman et al., 1984) and/or trophic factors such as NGF (Jensen et al., 1992) and BDNF (Encinas et al., 2000; Targett et al., 2024) into neuron-like cells. Because the primary goal of the present study was to assess cell-type specificity against DPAA exposure, it was necessary to establish an identical culture environment with no additional supplements for all cell types. Another issue to consider is evaluating the differences in the effects of DPAA on undifferentiated and differentiated cells, which may contribute to a better understanding of the effects of DPAA on human neurons. MCF7 is a frequently studied breast carcinoma-derived cell line used for breast cancer research and toxicological studies involving endocrine-disrupting chemicals particularly estrogenic compounds, such as bisphenol A (Dong et al., 2011). Exposure of MCF7 to DPAA induced increased protein expression of Hsp70, increased ERK1/2 and SAPK/JNK phosphorylation, intracellular GSH depletion, and increased GSH secretion. Although there was no statistical significance, DPAA exposure appeared to increase c-Jun and c-Fos protein expression. Interestingly, in hierarchical cluster analysis, MCF7 was included in the astrocyte-rich cluster group containing NRA, U251MG, 1321N1, and C6, even though MCF7 belongs to a lineage distant from astrocytes. A549 is a type II pulmonary epithelial cell model (Foster et al., 1998) used in drug metabolism and toxicant studies in the lungs (Garcia-Canton et al., 2013). DPAA exposure does not affect cellular signal transduction, such as MAP kinase phosphorylation in A549 cells. Therefore, of the nine cell lines examined in the present study, A549 is the most resistant to DPAA-induced aberrant activation of cellular signal transduction pathways. This might be attributed to a higher metabolic activity in A549 cells (Hukkanen et al., 2000). The rat glioma C6 cell line is frequently used in glioblastoma growth and invasion studies (Grobben et al., 2002). It is used to determine whether species (rat or human) or cell attributes (normal or tumor-derived) are important determinants of susceptibility to DPAA exposure. DPAA exposure in C6 triggered the NRA-like activation of aberrant intracellular signaling pathways. It also caused intracellular GSH depletion and did not trigger GSH release unlike NRA. Nevertheless, many similarities between C6 and NRA indicate that C6 is the most closely related cell line to NRA based on the hierarchical cluster analysis. A multivariate analysis revealed that NRA and C6 are the most closely related, followed by U251MG and 1321N1, which indicates an astrocyte preference for aberrant activation by DPAA. Our current and previous results in NHA suggest that, regardless of whether they are normal tissue-derived or tumor-derived, rat cells may be more vulnerable to DPAA than their corresponding human cells, although the results from just two cell types should be interpreted with caution. These results do not compromise the usefulness of rat-derived cells and it is prudent from the standpoint of human safety to consider acceptable daily intake (ADI) by examining the no-observed-adverse-effect levels (NOAELs) for various endpoints using highly susceptible rat-derived cells or rats in vivo.

In the present study, in addition to induction of anti-oxidative stress responses and activation of MAP kinases and transcription factors, which has been reported previously, we added changes in intracellular and/or extracellular glutathione concentrations to the evaluation targets and found a great diversity in the intracellular and extracellular status of glutathione. Although DPAA exposure did not affect intracellular/extracellular GSH level in HepG2, T98G, SK-N-SH, and A549, we found that DPAA exposure increased both intracellular and extracellular GSH in SH-SY5Y, decreased intracellular but increased extracellular GSH in 1321N1 and MCF7, unchanged intracellular but increased extracellular GSH in NRA and U251MG, and decreased intracellular but unchanged extracellular GSH in C6. In addition to NRA, GCLM was increased in HepG2, T98G, and 1321N1 in response to DPAA, suggesting that, at least in HepG2 and T98G, the increase in GCLM expression did not result in elevated intracellular or extracellular GSH and degradation/transport of GSH as well as the production and release should be considered.

Increased extracellular GSH concentration appear to be correlated with aberrant activation of intracellular signaling in NRA, U251MG, 1321N1, and MCF7. However, in C6 cells, GSH synthesis-related enzymes expression is not significantly altered, and it appears that GSH disappears following DPAA exposure as it is no longer measurable in this GSH assay. Although there is no strong evidence, we hypothesize that DPAA is activated intracellularly by glutathione, in which pentavalent arsenic in DPAA is reduced to the trivalent form (Kobayashi and Hirano, 2013; Ochi et al., 2006). This suggests that chemically reactive GSH-conjugated DPAA (Kobayashi and Hirano, 2013; Ochi et al., 2006) binds to cysteine residues in specific intracellular enzymes to inhibit their function. This results in the disruption/activation of intracellular signal transduction, such as protein phosphorylation by organomercurials (Geri et al., 2024).

It is possible that C6 cannot release intracellularly generated DPAA-GS and its cytotoxicity reduces cell viability even at concentrations of 10 µM. On the other hand, NRA exhibits GSH production in response to DPAA with doubled GCLM expression, although the absolute expression of GCLM in NRA was low compared with other cell type. Some of it reacts with DPAA but may be released extracellularly with other free GSH molecules. Although these results do not rule out the possibility that GSH metabolism may play an important role, they do not provide a precise mechanism by which GSH metabolism alone causes MAP kinase activation, etc. Therefore, further studies are needed to determine the chemical form of DPAA inside and outside the cell. In addition, glutathione-related factors [e.g., glutathione peroxidase (GPx) and glutathione S-transferase (GST)] in addition to glutathione synthesis-related enzymes should be evaluated to reveal more precise molecular mechanisms leading to cell-type-specificity in DPAA-induced aberrant signal transduction and glutathione metabolism, because glutathione may play an important and crucial role.

The above represents an overview of the effects of DPAA on each cell type addressed in the present study; however, the following issues should be addressed in future studies. First, the intracellular arsenic concentration in each cell type should be measured using inductively coupled plasma mass spectrometry (ICP-MS), which may explain some of the cell-type-specificity related to DPAA exposure. Second, studies using additional cell types should be conducted to draw more definitive conclusions. Human iPS-derived neurons (Garcia-Leon et al., 2019) and astrocytes (Lundin et al., 2018) are promising candidates for the analysis of DPAA neurotoxicity in humans. Because microglia and vascular endothelial cells, in addition to neurons and astrocytes, are essential for normal brain functioning, the effects of DPAA exposure on these cells should be examined. For example, HMC3 cells (Rawat and Spector, 2017) and human umbilical vein vascular endothelial cells HUVEC should be examined. Commercially available human astrocytes immortalized by viruses carrying hTERT and SV40 genes and other human glioma-derived cell lines, such as U-87MG (Pontén and Westermark, 1978; Allen et al., 2016), are also important candidates. Third, studies on the dose- and time-dependencies following DPAA exposure among DPAA-susceptible human-derived astrocytes, U251MG and 1321N1, should be performed. Although NHA could not be included in the present study, our previous results (Sasaki et al., 2022) indicated that NHA failed to exhibit DPAA-induced aberrant activation of intracellular signaling pathways following DPAA exposure at 10 µM for 96 hr, although NHA responded to DPAA similar to NRA only at high DPAA at a higher dose (50 µM) for 96 hr or 10 µM for a longer period (288 hr). In the present study, DPAA exposure at 10 µM for 96 hr induced some NRA-like cellular signal transduction activation in U251MG and 1321N1 cells, suggesting that these human glioma-derived cells may be more susceptible to DPAA compared with NHA. They may presumably serve as a convenient, easy-to-handle, and promising model for studying the mechanism of DPAA-induced aberrant signal transduction in human astrocytes without consuming precious and difficult-to-use NHA. Fourth, DPAA at relatively low doses showed cell type-specific adverse effects, although a pentavalent organic arsenic is generally recognized as having low general toxicity. It is unclear whether this cell type difference in susceptibility to DPAA applies to other heavy metal compounds, especially inorganic arsenicals, such as arsenite, a trivalent inorganic arsenic having high toxicity, or specific to DPAA, which is interesting from a toxicological perspective. Finally, we were not able to determine the underlying cellular and molecular mechanisms of DPAA toxicity, which should be extensively explored further using these valuable human-derived astrocytes.

In summary, our present findings provided the first evidence that there is a cell-type-specificity in susceptibility to DPAA. Moreover, DPAA may affect signal transduction preferentially in brain astrocytes among the diverse types of cells in the human body. GSH produced by brain astrocytes may be one of key molecules and its metabolism may play a key role in DPAA neurotoxicity. Identifying the underlying mechanisms between DPAA exposure and cerebellar toxicity will lead to the development of preventive and therapeutic strategies against future incidents of DPAA toxicity.

ACKNOWLEDGMENTS

The authors declare no conflicts of interest associated with this study. The authors would like to thank T. Ishida and N. Sumiyoshi (Faculty of Pharmacy, Meijo University) for their technical assistance with the experiments. The authors would like to thank Enago (www.enago.jp) for the English language review. This work was supported by the Ministry of the Environment of Japan (Research on the influence of diphenylarsinic acid and related compounds on human health) and by MEXT/JSPS KAKENHI; Grant-in-Aid for Scientific Research (C) (22K12394, T. N.), Grant-in-Aid for Scientific Research (C) (19K12345, T. N.), Grant-in-Aid for Scientific Research (C) (16K00564, T. N.), Challenging Exploratory Research (26550043, T. N.), Grant-in-Aid for Young Scientists (A) (20681005, T. N.), and Grant-in-Aid for Young Scientists (A) (23681010, T. N.), and by the Research Institute of Meijo University (T.N.). The funding organizations have no control over the resulting publication. Note that the views expressed in this study do not necessarily reflect the positions or policies of the Ministry of the Environment of Japan and MEXT/JSPS.

Conflict of interest

The authors declare that there is no conflict of interest.

REFERENCES

•  Allen,  M.,  Bjerke,  M.,  Edlund,  H.,  Nelander,  S. and  Westermark,  B. (2016): Origin of the U87MG glioma cell line: good news and bad news. Sci. Transl. Med., 8, 354re3.
•  Dong,  S.,  Terasaka,  S. and  Kiyama,  R. (2011): Bisphenol A induces a rapid activation of Erk1/2 through GPR30 in human breast cancer cells. Environ. Pollut., 159, 212-218.
•  Encinas,  M.,  Iglesias,  M.,  Liu,  Y.,  Wang,  H.,  Muhaisen,  A.,  Ceña,  V.,  Gallego,  C. and  Comella,  J.X. (2000): Sequential treatment of SH-SY5Y cells with retinoic acid and brain-derived neurotrophic factor gives rise to fully differentiated, neurotrophic factor-dependent, human neuron-like cells. J. Neurochem., 75, 991-1003.
•  Foster,  K.A.,  Oster,  C.G.,  Mayer,  M.M.,  Avery,  M.L. and  Audus,  K.L. (1998): Characterization of the A549 cell line as a type II pulmonary epithelial cell model for drug metabolism. Exp. Cell Res., 243, 359-366.
• Garcia-Canton, C., Minet, E., Anadon, A. and Meredith, C. (2013): Metabolic characterization of cell systems used in in vitro toxicology testing: lung cell system BEAS-2B as a working example. Toxicology in vitro: an international journal published in association with BIBRA, 27, 1719-1727.
•  Garcia-Leon,  J.A.,  Vitorica,  J. and  Gutierrez,  A. (2019): Use of human pluripotent stem cell-derived cells for neurodegenerative disease modeling and drug screening platform. Future Med. Chem., 11, 1305-1322.
•  Geri,  A.,  Zineddu,  S.,  Massai,  L.,  Ronga,  L.,  Lobinski,  R.,  Gailer,  J. and  Messori,  L. (2024): Mercury binding to proteins disclosed by ESI MS experiments: the case of three organomercurials. J. Inorg. Biochem., 252, 112479.
•  Grobben,  B.,  De Deyn,  P.P. and  Slegers,  H. (2002): Rat C6 glioma as experimental model system for the study of glioblastoma growth and invasion. Cell Tissue Res., 310, 257-270.
•  Hukkanen,  J.,  Lassila,  A.,  Päivärinta,  K.,  Valanne,  S.,  Sarpo,  S.,  Hakkola,  J.,  Pelkonen,  O. and  Raunio,  H. (2000): Induction and regulation of xenobiotic-metabolizing cytochrome P450s in the human A549 lung adenocarcinoma cell line. Am. J. Respir. Cell Mol. Biol., 22, 360-366.
•  Humphrey,  M.L.,  Cole,  M.P.,  Pendergrass,  J.C. and  Kiningham,  K.K. (2005): Mitochondrial mediated thimerosal-induced apoptosis in a human neuroblastoma cell line (SK-N-SH). Neurotoxicology, 26, 407-416.
•  Ishii,  K.,  Nemoto,  K.,  Iwasaki,  N.,  Takeda,  T.,  Masuda,  T.,  Shibata,  Y. and  Tamaoka,  A. (2019): Decreased regional cerebral blood flow in patients with diphenylarsinic acid intoxication. Eur. J. Neurol., 26, 136-141.
•  Ishii,  K.,  Tamaoka,  A.,  Otsuka,  F.,  Iwasaki,  N.,  Shin,  K.,  Matsui,  A.,  Endo,  G.,  Kumagai,  Y.,  Ishii,  T.,  Shoji,  S.,  Ogata,  T.,  Ishizaki,  M.,  Doi,  M. and  Shimojo,  N. (2004): Diphenylarsinic acid poisoning from chemical weapons in Kamisu, Japan. Ann. Neurol., 56, 741-745.
•  Janyou,  A.,  Changtam,  C.,  Suksamrarn,  A.,  Tocharus,  C. and  Tocharus,  J. (2015): Suppression effects of O-demethyldemethoxycurcumin on thapsigargin triggered on endoplasmic reticulum stress in SK-N-SH cells. Neurotoxicology, 50, 92-100.
•  Jensen,  L.M.,  Zhang,  Y. and  Shooter,  E.M. (1992): Steady-state polypeptide modulations associated with nerve growth factor (NGF)-induced terminal differentiation and NGF deprivation-induced apoptosis in human neuroblastoma cells. J. Biol. Chem., 267, 19325-19333.
•  Kim,  M.S.,  Takahashi,  T.,  Lee,  J.Y.,  Hwang,  G.W. and  Naganuma,  A. (2012): Methylmercury induces CCL2 expression through activation of NF-κB in human 1321N1 astrocytes. J. Toxicol. Sci., 37, 1275-1278.
•  Kinoshita,  K.,  Ochi,  T.,  Suzuki,  T.,  Kita,  K. and  Kaise,  T. (2006): Glutathione plays a role in regulating the formation of toxic reactive intermediates from diphenylarsinic acid. Toxicology, 225, 142-149.
•  Kita,  K.,  Sato,  M.,  Suzuki,  T. and  Ochi,  T. (2009): Structure-effect relationship in the down-regulation of glutaminase in cultured human cells by phenylarsenic compounds. Toxicology, 258, 157-163.
•  Kita,  K.,  Suzuki,  T. and  Ochi,  T. (2007): Down-regulation of glutaminase C in human hepatocarcinoma cell by diphenylarsinic acid, a degradation product of chemical warfare agents. Toxicol. Appl. Pharmacol., 220, 262-270.
•  Kita,  K.,  Suzuki,  T. and  Ochi,  T. (2012): Diphenylarsinic acid promotes degradation of glutaminase C by mitochondrial Lon protease. J. Biol. Chem., 287, 18163-18172.
• Kobayashi, Y. and Hirano, S. (2013): The role of glutathione in the metabolism of diphenylarsinic acid in rats. Metallomics: integrated biometal science, 5, 469-478.
•  Li,  C.,  Feng,  C.,  Chen,  Y.,  Wu,  P.,  Li,  P.,  Xiong,  X.,  Peng,  X.,  Wang,  Z. and  Li,  Y. (2022): Arsenic trioxide induces the differentiation of retinoic acid-resistant neuroblastoma cells via upregulation of HoxC9. Adv. Clin. Exp. Med., 31, 903-911.
•  Lundin,  A.,  Delsing,  L.,  Clausen,  M.,  Ricchiuto,  P.,  Sanchez,  J.,  Sabirsh,  A.,  Ding,  M.,  Synnergren,  J.,  Zetterberg,  H.,  Brolén,  G.,  Hicks,  R.,  Herland,  A. and  Falk,  A. (2018): Human iPS-Derived Astroglia from a Stable Neural Precursor State Show Improved Functionality Compared with Conventional Astrocytic Models. Stem Cell Reports, 10, 1030-1045.
•  Macintyre,  E.H.,  Pontén,  J. and  Vatter,  A.E. (1972): The ultrastructure of human and murine astrocytes and of human fibroblasts in culture. Acta Pathol. Microbiol. Scand. [A], 80, 267-283.
•  Nakamagoe,  K.,  Fujizuka,  N.,  Koganezawa,  T.,  Shimizu,  K.,  Takiguchi,  S.,  Horaguchi,  T.,  Ishii,  K. and  Tamaoka,  A. (2013): Residual central nervous system damage due to organoarsenic poisoning. Neurotoxicol. Teratol., 37, 33-38.
•  Nakamagoe,  K.,  Ishii,  K.,  Tamaoka,  A. and  Shoji,  S. (2006): Upward gaze-evoked nystagmus with organoarsenic poisoning. Neurology, 66, 131-132.
•  Negishi,  T.,  Matsumoto,  M.,  Kobayashi,  Y.,  Kojima,  M.,  Sakaguchi,  F.,  Takahata,  K.,  Kanehira,  T.,  Arakaki,  R.,  Aoyama,  Y.,  Yoshida,  H.,  Yamada,  R.,  Sumiyoshi,  N.,  Tashiro,  T.,  Hirano,  S.,  Yoshida,  K. and  Yukawa,  K. (2017): Dysregulation of MAP Kinase Signaling Pathways Including p38MAPK, SAPK/JNK, and ERK1/2 in Cultured Rat Cerebellar Astrocytes Exposed to Diphenylarsinic Acid. Toxicol. Sci., 156, 509-519.
•  Negishi,  T.,  Matsumoto,  M.,  Kojima,  M.,  Asai,  R.,  Kanehira,  T.,  Sakaguchi,  F.,  Takahata,  K.,  Arakaki,  R.,  Aoyama,  Y.,  Yoshida,  H.,  Yoshida,  K.,  Yukawa,  K.,  Tashiro,  T. and  Hirano,  S. (2016): Diphenylarsinic Acid Induced Activation of Cultured Rat Cerebellar Astrocytes: Phosphorylation of Mitogen-Activated Protein Kinases, Upregulation of Transcription Factors, and Release of Brain-Active Cytokines. Toxicol. Sci., 150, 74-83.
•  Negishi,  T.,  Matsunaga,  Y.,  Kobayashi,  Y.,  Hirano,  S. and  Tashiro,  T. (2013): Developmental subchronic exposure to diphenylarsinic acid induced increased exploratory behavior, impaired learning behavior, and decreased cerebellar glutathione concentration in rats. Toxicol. Sci., 136, 478-486.
•  Negishi,  T.,  Takahashi,  M.,  Matsunaga,  Y.,  Hirano,  S. and  Tashiro,  T. (2012): Diphenylarsinic acid increased the synthesis and release of neuroactive and vasoactive peptides in rat cerebellar astrocytes. J. Neuropathol. Exp. Neurol., 71, 468-479.
•  Ochi,  T.,  Kinoshita,  K.,  Suzuki,  T.,  Miyazaki,  K.,  Noguchi,  A. and  Kaise,  T. (2006): The role of glutathione on the cytotoxic effects and cellular uptake of diphenylarsinic acid, a degradation product of chemical warfare agents. Arch. Toxicol., 80, 486-491.
•  Ozone,  K.,  Ueno,  S.,  Ishizaki,  M. and  Hayashi,  O. (2010): Toxicity and Oxidative Stress Induced by Organic Arsenical Diphenylarsinic Acid and Inorganic Arsenicals and Their Effects on Spatial Learning Ability in Mice. J. Health Sci., 56, 517-526.
•  Påhlman,  S.,  Ruusala,  A.I.,  Abrahamsson,  L.,  Mattsson,  M.E. and  Esscher,  T. (1984): Retinoic acid-induced differentiation of cultured human neuroblastoma cells: a comparison with phorbolester-induced differentiation. Cell Differ., 14, 135-144.
•  Pontén,  J. and  Macintyre,  E.H. (1968): Long term culture of normal and neoplastic human glia. Acta Pathol. Microbiol. Scand., 74, 465-486.
•  Pontén,  J. and  Westermark,  B. (1978): Properties of human malignant glioma cells in vitro. Med. Biol., 56, 184-193.
• R Core Team. (2024): R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing.
•  Rawat,  P. and  Spector,  S.A. (2017): Development and characterization of a human microglia cell model of HIV-1 infection. J. Neurovirol., 23, 33-46.
•  Ross,  R.A.,  Spengler,  B.A. and  Biedler,  J.L. (1983): Coordinate morphological and biochemical interconversion of human neuroblastoma cells. J. Natl. Cancer Inst., 71, 741-747.
•  Sasaki,  S.,  Negishi,  T.,  Tsuzuki,  T. and  Yukawa,  K. (2022): Diphenylarsinic acid induced activation of MAP kinases, transcription factors, and oxidative stress-responsive factors and hypersecretion of cytokines in cultured normal human cerebellar astrocytes. Neurotoxicology, 88, 196-207.
•  Stefàno,  E.,  Muscella,  A.,  Benedetti,  M.,  De Castro,  F.,  Fanizzi,  F.P. and  Marsigliante,  S. (2022): Antitumor and antimigration effects of a new Pt compound on neuroblastoma cells. Biochem. Pharmacol., 202, 115124.
•  Stein,  G.H. (1979): T98G: an anchorage-independent human tumor cell line that exhibits stationary phase G1 arrest in vitro. J. Cell. Physiol., 99, 43-54.
•  Targett,  I.L.,  Crompton,  L.A.,  Conway,  M.E. and  Craig,  T.J. (2024): Differentiation of SH-SY5Y neuroblastoma cells using retinoic acid and BDNF: a model for neuronal and synaptic differentiation in neurodegeneration. In Vitro Cell. Dev. Biol. Anim., 60, 1058-1067.
•  Umezu,  T.,  Nakamiya,  K.,  Kita,  K.,  Ochi,  T.,  Shibata,  Y. and  Morita,  M. (2012): Diphenylarsinic acid produces behavioral effects in mice relevant to symptoms observed in citizens who ingested polluted well water. Neurotoxicol. Teratol., 34, 143-151.
•  Umezu,  T. and  Shibata,  Y. (2021): Toxicokinetic characteristics and effects of diphenylarsinic acid on dopamine in the striatum of free-moving mice. Neurotoxicology, 83, 106-115.
•  Wisessaowapak,  C.,  Visitnonthachai,  D.,  Watcharasit,  P. and  Satayavivad,  J. (2021): Prolonged arsenic exposure increases tau phosphorylation in differentiated SH-SY5Y cells: the contribution of GSK3 and ERK1/2. Environ. Toxicol. Pharmacol., 84, 103626.