Background
Arsenic is well-known to be toxic to most of the tissues in the body after direct exposure. Direct exposure may be through inhalation, ingestion, or direct skin contact with arsenic-contaminated air, food and drinks, or products [
1]. Commonly, the exposure is by drinking arsenic-contaminated water. Permissible level for arsenic in drinking water is 10 ppb for most of the countries over the world. There are many evidences of health hazards due to drinking arsenic-contaminated water above this level for a certain period. Severity of health hazards varied with the level of contamination as well as duration of exposure. Types of health hazards also varied from skin lesions to cancerous lesions. Neurological lesions include peripheral neuropathy, memory loss, and behavior changes [
2]. These changes are cumulatively known as neurotoxic effects of arsenic, and its effect is more distinct in the developing brain. From the evidence of animal studies, developmental arsenic exposure leads to definite health hazards in the offspring because arsenic can pass through the placenta [
3]. In experimental design, developmental exposure is also known as maternal exposure because arsenic-contaminated water is only given to pregnant mother and not given directly to the offspring. Exposure occurs from mother to offspring through placenta during gestational period and so it is also called gestational exposure. In some study design, maternal exposure is given during gestation, after delivery, and during lactational period. Arsenic is never given directly to offspring, but it might reach to offspring through placenta circulation and lactation [
4].
Arsenic can pass through not only the placenta barrier but also the blood-brain barrier [
5]. Therefore, toxic effect is possible to adult brain, and it might be more intense for a developing brain. Being a developmental neurotoxicant, neurotoxic effects of arsenic can be seen by maternal arsenic exposure during gestational period [
6,
7]. Our previous study proved that maternal arsenic exposure from gestational days 8–18 impaired the social behavior and related gene expressions in F1 mice. In that study, we concluded that the behavioral changes might be due to impaired serotonergic system [
8]. Effects of maternal arsenic exposure to F1 mice were studied not only in the nervous system but also in the other systems. Nohara et al. studied the tumor-augmenting effect of arsenic on liver cells in a similar study design. They showed that tumor-augmenting effect occurred in both F1 and F2 mice by maternal arsenic exposure [
9].
Although arsenic was not given to the F1 mice, they might get arsenic exposure indirectly from the mother during developing period. For the F2 mice, they were generated by mating among F1 males and females, and there was no direct as well as indirect exposure to arsenic. Etiology of cancer is related to genetics, and so the tumor-augmenting effect in F2 mice might be due to genetics in origin. In the neurotoxic study, it was uncertain to see the similar effects in the next generation. This study aimed to detect the neurotoxic effects in F2 male mice born to gestationally arsenite-exposed F1 mice. In the present study, social behavior, histological feature of the prefrontal cortex, and social behavior-related gene expressions were studied in the 41-week-old and 74-week-old groups of C3H male mice.
Discussion
The major findings of this study were impaired social behavior and decreased 5-HT 5B gene expression in the arsenite-F2 male mice. In our previous study of the F1 male mice, similar findings were observed, and we postulated that impaired social behavior might be due to deranged serotonergic system in the prefrontal cortex [
8]. Prefrontal cortex is concerned with cognitive function and social behavior. Role of serotonergic system in the prefrontal cortex is neuromodulating effect, and it stabilizes the pyramidal cell excitation in the prefrontal cortex. It is important for proper social behavior and normal mood status. Neurotoxic effect of arsenic might involve the serotonergic system, resulting in poor sociability and social novelty preference.
In sociability test, the time spent for stranger 1 was not much different between the control and arsenite-F2 mice. The key difference was the time spent for empty cup; the arsenite-F2 mice tended to spend more time to empty cup. It could be interpreted as a kind of social isolation, and it was more distinct in the 41-week-old group. In social novelty preference test, the arsenite-F2 mice seemed to have no preference to either stranger 1 or stranger 2. These behavioral changes were clearly seen in both age groups. Recently, Valles et al. reported that altered motor activity and increased anxiety-like behaviors in zebrafish were transmitted to the F2 generation after ancestral exposure to arsenic in F0. They pointed out that arsenic was a neurotoxicant and potent epigenetic disruptor [
14].
In the present study, the C3H male mice were used as a good model for social behavior assessment and neurochemical analysis [
8,
15]. We studied the social behavior and the mRNA expressions related to behavior in two age groups of C3H mice; namely the 41-week-old group as a mature adult group and the 74-week-old group as an old adult group. Social behavior might be changed with age, and so we compared the social behavior between the two age groups. However, there was no significant difference between two age groups, regarding social behavior. Social behavior changes in arsenite-F2 mice were also similar to that of gestationally arsenite-exposed F1 mice in our previous study [
8].
In the present study, sodium arsenite (NaAsO
2, 85 ppm (85 mg/L) was given only to F0 pregnant mice from gestational days 8–18 (critical period of neurodevelopment). Since species differences were observed between human and mouse, a high dose is required for detection of neurotoxic effects in C3H mice. This is a standard dose to detect the effects of maternal exposure to arsenic on neurotoxicity, reproductive toxicity, and carcinogenicity in our research group [
9,
10,
16]. No maternal toxicity and teratogenicity were observed at the dose of the present study. We only used the male pups because there is a sex difference in susceptibility to arsenic. Moreover, there is the oestrous cycle in female and possible hormonal effects on social behavior [
17].
Neuronal morphology changes in the prelimbic cortex, a part of the prefrontal cortex, were studied by Aung et al. in 2016. They found that arsenic exposure was associated with a significant increase in the number of pyramidal neurons in layers V and VI of the prelimbic cortex. The prenatal arsenic exposure was associated with a significant decrease in neurite length but not dendrite spine density in all layers of the prelimbic cortex. In the present study, there were no significant histological changes, such as chromatolysis, gliosis, and necrosis, of the prefrontal cortex in the arsenite-F2 mice. It might be due to the absence of direct or developmental exposure to arsenic in F2 mice. Otherwise, immunohistochemical study would be helpful to detect more details than ordinary histological study using H&E stain.
Although there were no gross and histological changes in the prefrontal cortex, we found that the behavioral and neurochemical changes in the prefrontal cortex occurred in the arsenite-F2 mice. We studied the mRNA expressions of 5-HT 5B and BDNF in the prefrontal cortex, and these were found to be decreased in the arsenite-F2 mice compared to the control mice. These changes seemed to be inherited and might be the result of epigenetic disruption. In the present study, we could not prove that it was epigenetic transgenerational inheritance. However, these changes were found consistently in both gestationally arsenite-exposed F1 mice [
6] and arsenite-F2 mice in the present study.
Since serotonin was known as a neuromodulator, we hypothesized that decreased serotonin receptors (5-HT 5B) would be responsible for glutamate-induced hyperexcitation of pyramidal cells and cell destruction in the prefrontal cortex. This led to impaired cognitive function, social memory, and finally social behavior. This effect was augmented by the decreased in BDNF which had a neuroprotective effect on many brain regions. Almeida and co-workers demonstrated the neuroprotective effect of BDNF against glutamate-induced apoptotic cell death in the hippocampus [
18]. The decreased mRNA expression of BDNF in the prefrontal cortex was seen in the F1 of our previous study [
6] and also in the old adult group of arsenite-F2 mice in the present study.
Arsenic-induced BDNF decrease was convinced by both animal and human studies. Valles et al. found that a reduction in BDNF gene expression in the F0 and F2 generation of zebrafish was induced by inorganic arsenic exposure [
14]. In a human study, Karim and co-workers described that serum BDNF level was significantly (
p<0.001) decreased in people living in arsenic-endemic area, and there was a dose-response relationship between arsenic exposure and serum BDNF levels [
19]. Role of BDNF was not only as a neuroprotective agent, but also as a synaptic regulator in the brain. Its synaptogenic and synaptoplastic effects in the prefrontal cortex were also important for cognitive function, and there was an interaction between BDNF and serotonin regarding social behavior.
In addition to 5-HT 5B and BDNF, we also studied the mRNA expression of dopamine receptor in the prefrontal cortex. Dopamine is one of the major markers in social brain, and among the dopamine receptors, dopamine D1 receptor (Drd1) plays a major role in social cognition [
20]. In the present study, decreased Drd1a gene expression was observed in both age groups of arenite-F2 mice, significantly in the old adult group. This finding was also supportive to the occurrence of impaired social behavior in the arsenic group. It was consistent with the findings of Homberg and colleagues. In their study, they stated that sociability and social novelty preference were significantly reduced in the Drd1
I116S mutant rats compared to wild-type rats [
20].
We also investigated the oxidative stress and inflammatory markers, HO-1 and COX-2, because almost all arsenic effects were based on oxidative stress and tissue inflammation [
2]. In both mature adult group and old adult group of arsenite-F2 mice, mRNA expression of HO-1 was significantly increased. Increased HO-1 expression was associated with a variety of conditions, including ultraviolet irradiation, hyperthermia, inflammatory cytokines, heavy metals, apoptosis, and cancers [
21]. In the present study, the oxidative stress indicated by HO-1 expression might not be due to direct effect of sodium arsenite because F2 mice were never exposed to arsenic directly. It might be probably due to glutamate-induced apoptotic cell death in the prefrontal cortex as discussed above.
Otherwise, we could not exclude the other possible conditions of upregulation of HO-1 in this study because HO-1 is a non-specific oxidative stress marker. To explore more information on oxidative stress, we studied plasma 8OHdG which is a sensitive marker of oxidative DNA damage [
22]. In the present study, there was no significant increase in plasma 8OHdG level in arsenite-F2 mice. These findings suggested that increased HO-1 was not likely due to oxidative DNA damage and apoptotic cell death. HO-1 response might be due to other non-specific inflammatory process.
The IL-1β is a potent inflammatory cytokine which is important for regulation of other inflammatory cytokines such as IL-6 and IL-8 [
23]. In the present study, IL-1β mRNA expression was only increased in 41-week-old arsenite-F2 mice. Regarding COX-2, it was a potent inflammatory marker, and we did not detect its upregulation as HO-1 in this study. Similar to HO-1, COX-2 was also associated with inflammation, apoptosis, and cancers [
24]. In contrast, COX-2 mediates inflammation while HO-1 modulates it. The HO-1 has antioxidant, anti-inflammatory, antiapoptotic, and antiproliferative effects [
25]. The overexpression of COX-2 induces HO-1 expression which in turn inhibits COX-2 expression [
26]. Therefore, in the present study, there was an increase in HO-1 expression, but not in COX-2 expression.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.