Comparison of the toxic effects of organic and inorganic arsenic in Caenorhabditis elegans using a multigenerational approach

Abstract

Although arsenic (As) is a persistent contaminant in the environment, few studies have assessed its effects over generations, as it requires an animal model with a short lifespan and rapid development, such as the nematode Caenorhabditis elegans. Furthermore, few studies have evaluated the effects of As metabolites such as dimethylarsinic acid (DMA^V), and several authors have considered DMA as a moderately toxic intermediate of As, although recent studies have shown that this chemical form can be more toxic than inorganic arsenic (iAs) even at low concentrations. In the present study, we compared the toxic effects of arsenate (As^V) and DMA^V in C. elegans over 5 subsequent generations. We evaluated biochemical parameters such as reactive oxygen species (ROS) concentration, the activity of antioxidant defense system (ADS) enzymes such as catalase (CAT) and glutathione-S-transferase (GST), and nonenzymatic components of ADS such as reduced glutathione (GSH) and protein-sulfhydryl groups (P-SH). Exposure to 50 μg L⁻¹ of As^V led to an increase in ROS generation and GSH levels together with a decrease in GST activity, while exposure to DMA^V led to an increase in ROS levels, with an increase in lipid peroxidation, CAT activity, and a decrease in GSH levels. In addition, both treatments reduced animal growth from the third generation onward and caused disturbances in their reproduction throughout all 5 generations. This study shows that the accumulated effects of DMA need to be considered; it highlights the importance of this type of multigenerational approach for evaluating the effects of organic contaminants considered low or nontoxic.

Graphical Abstract

Graphical Abstract.

1. Introduction

The presence of arsenic (As) in the environment has been documented in diverse environmental compartments. Approximately 108 countries are estimated to be affected by As contamination in groundwater (concentration beyond the maximum permitted limit of 10 μg L⁻¹ recommended by the World Health Organization (WHO).^1,2 In Latin America, it is estimated that at least 4.5 million people are exposed to high levels of As through contaminated water bodies (concentrations greater than 50 μg L⁻¹) in highly affected areas.³ In fact, the distribution of As in natural waters can be highly heterogeneous, from very low to very high levels (0.5 μg L⁻¹ to >5,000 μg L⁻¹).⁴ As can be found in different chemical forms, such as inorganic As (iAs), including arsenite (As^III) and arsenate (As^V), which are the main As species in sediments and aquatic environments, as well as organic forms (such as MMA and DMA, mono-and dimethylarsonate, respectively), mainly due to the activity of microorganisms such as aquatic phytoplankton, bacteria, and microalgae.⁵ Although As is released into the environment through natural processes, anthropogenic activities such as mining, production, and use of fertilizers can increase the natural concentration of As, which has harmful effects on living organisms.^6,7

Beyond exposure, the metabolization capacity of animals is a crucial factor in evaluating the effects of As exposure because once it is incorporated by organisms, it can be biotransformed and excreted and/or accumulated in tissues in different chemical forms. This process can be influenced by several factors, such as co-exposure to other contaminants^{8,9,10,11,12,13}, exposure concentration,^8,14,15 and the ability to metabolize As.^5,8,16,17 Following uptake into living organisms, As^V may be reduced to As^III, which can be further methylated to organic species such as MMA, DMA, trimethylarsenoxide (TMAO), arsenocholine (AsC), arsenobetaine (AsB), and various arsenosugars.¹⁸ Fish and crustaceans tend to accumulate As in the form of AsB, while polychaetes accumulate As mainly as methylated forms such as MMA and DMA.^14,16 iAs is considered more toxic, with As^III reacting with thiol-containing molecules, such as reduced glutathione (GSH) and other cysteine residues, and As^V replacing phosphate during the process of oxidative phosphorylation in mitochondria, thus affecting ATP production.¹⁹ On the other hand, organic compounds such as DMA have initially been considered to be only moderately toxic, but the paradigm that methylation is only part of a detoxification process is now confronted by evidence showing that DMA can affect gene transcription and cause reproductive disturbances and genotoxicity^10,19,20.

Bundschuh et al.³ showed a critical link between chronic As exposure and adverse human health effects, including different types of cancer, reproductive impairment, and cognitive problems in children. Furthermore, cancer development is a multistep process; for example, oxidative stress caused by exposure to environmental contaminants can induce DNA damage, which plays an important role in human oncogenesis.²¹ Yamanaka et al.²² proposed that during DMA metabolism, DMA radicals are formed, which interact with the DNA, causing single strand breaks, chromosomal aberrations, and cell cycle arrest, but they did not observe any inhibition of DNA repair enzyme activity. Müller et al.¹⁰ showed a similar result with polychaetes exposed to DMA, where genotoxic damage was observed after 48 h of exposure; however, the enzymatic DNA repair system was not affected.

Most previous studies have focused on acute As toxicity in living organisms, and very little is known about the dynamic effects of chronic exposure.²³ Josende et al.²⁴ showed that exposure to As^III generated different responses over 5 generations in C. elegans, reporting an increase in reactive oxygen species (ROS) generation from the second generation, which was maintained until the fifth generation, and a critical decrease in physiological parameters, such as reproduction and growth, from the third generation onwards. However, as mentioned above, As^III and As^V show different effects; therefore, it is unknown if As^V may affect the organism over generations.

In addition, few studies have addressed the transgenerational or multigenerational effects of As exposure.^24,25,26 This requires an organism with a short life cycle, such as C. elegans. This nematode is a small, free-living, soil-dwelling organism with a short life span of ~2–3 weeks and a rapid generation time of 3–4 days.²⁶  Caenorhabditis elegans was the first multicellular organism for which the genome was fully sequenced, and although it is remarkably smaller than the human genome, both genomes have a similar number of genes (worms 20,000 genes; humans 23,000). There is a substantial overlap between C. elegans and humans in terms of genes and biochemical pathways. Bioinformatics analysis suggest that 60–80% of worm genes are homologous to humans.²⁷ Thus, considering the problem of As contamination in Latin America and the lack of data on the effects of exposure to iAs and organic forms such as DMA in the offspring of animals continuously exposed to these compounds, the aim of this study was to evaluate the biochemical and physiological effects of exposure to As^V and DMA^V over 5 generations in C. elegans.

2. Materials and methods

2.1 Caenorhabditis elegans

Caenorhabditis elegans from wild strain N2 Bristol were cultivated in a nematode growth medium (NGM; 3.0 g L⁻¹ NaCl, 5.0 g L⁻¹ peptone, 0.005 g L⁻¹ cholesterol, dissolved in absolute ethanol, 0.11 g L⁻¹ CaCl₂, 0.12 g L⁻¹ MgSO₄, 5.3 g L⁻¹ KH₂PO₄, 17.0 g L⁻¹ agar, with pH 6.0 with 1 mL of Antibiotic Antimycotic Solution (100×; Sigma-Aldrich) at 20 °C, seeded with Escherichia coli, OP50 strain as a food source. Prior to the experiment, a synchronous population of C. elegans was obtained after a bleaching treatment (50 mL of 0.8 M NaOH, 0.5% NaClO), in order to restrict the animal population to the embryonic development stage (egg) and start with animals at the same stage of larval development (L1). After the synchronization process, the eggs obtained were placed in culture plates containing S-basal liquid medium (5.8 g L⁻¹ NaCl, 1 g L⁻¹ K₂HPO₄, 6 g L⁻¹ KH₂PO₄, 0.005 g L⁻¹ cholesterol, dissolved in absolute ethanol) until hatching. After hatching, E. coli OP50 was added, and the animals were conditioned again at 20 °C for another 96 h, prior to the start of the experiment.

2.2 Experimental design

At the beginning of the exposure, the animals were filtered through a mesh of 25-μm porosity to separate the larvae in the initial stage (L1) for exposure. After filtering, the number of animals was estimated according to Josende et al.,²⁴ and the animals were distributed into different experimental groups. All the experiments were performed using 3 independent assays. Exposures were carried out with food (E. coli OP50 inactivated by freezing) to facilitate the ingestion of contaminants and prevent As metabolism by bacteria.

The experimental designs for the biochemical analysis were carried out in NGM solid medium with a thin layer of S-basal liquid medium on the surface to simulate natural environmental conditions. Biochemical experiments were performed in medium Petri dishes (Kasvi, model k13, 100 × 100 mm in diameter), with ~10,000 animals at the beginning of each generation. The experiments to assess animal growth were carried out in 24-well culture plates (TPP, model 92024), using NGM solid medium with S-basal liquid medium on the surface, with 25 animals per well.

The experiments were carried out for 5 generations after prior exposure of the parental generation and continued exposure of the offspring to the tested contaminants. The aim of this approach was to assess animals that face contamination during embryonic development. Each generation was exposed for 96 h to allow L1 to develop, reach adulthood, and reproduce. After 96 h, the animals of each generation were washed using S-basal liquid medium from the plates and filtered using a 25-μm filter that allowed the passage of L1 and retained adults and individuals in other stages. This filtration procedure was intended to minimize the chemical stress of regular chemical synchronization and did not interfere with the results. Adults from each generation on NGM plates were collected for biochemical analysis, and L1 was transferred to new plates and a new exposure was performed, starting with the subsequent generation. This process was repeated until 5 generations of animals had been exposed at the L1 stage.

The experimental groups and their respective concentrations were set as follows: (i) control group, exposed only to E. coli OP50 and S-basal; (ii) DMA^V group exposed to E. coli OP50, S-basal liquid medium, and DMA^V [50 μg L⁻¹]; and (iii) As^V group exposed to E. coli OP50, S-basal liquid medium, and As^V [50 μg L⁻¹]. The concentrations were determined according to a previous study in C. elegans.²⁴

2.3 Biochemical measurements

2.3.1 Preparation of homogenized samples

Samples containing a pool of animals were homogenized (1:1, v:v) in S-basal liquid medium using a sonicator (QSonica, model Q125, 125-W power). Samples were sonicated for 40 s at 50% amplitude in an ice bath, and the homogenate was centrifuged at 10,000 × g for 20 min at 4 °C. Supernatants were collected, aliquoted, and stored at −80 °C until the respective biochemical analysis was performed, with the exception of ROS quantification, which was performed immediately after sample collection. The samples were normalized to the total protein content present in each sample, which was quantified at 600 nm using a commercial kit (BIOPROT U/LCR, Bioclin).

2.3.2 Reactive oxygen species

ROS content was measured using samples processed as described in Section 2.3.1. Supernatants were used to determine the ROS content using 2,7-dichlorofluorescein-diacetate (H₂DCF-DA, Invitrogen), which generates a fluorochrome detected at 485- and 530-nm wavelengths for excitation and emission, respectively.²⁸ Readings were performed in a fluorescence microplate reader (FilterMax F5, Molecular Devices) at intervals of 5 min for 120 min.

2.3.3 Total antioxidant capacity against peroxyl radicals

The antioxidant capacity against peroxyl radicals was measured according to the method described by Amado et al.²⁹ using a fluorescence microplate reader (FilterMax F5, Molecular Devices) programmed at 37 °C, where peroxyl radicals are generated by the thermal decomposition of 2,2′-azobis (2 methylproprionamidine) dihydrochloride (ABAP, Sigma-Aldrich). Immediately prior to the reading, 2′,7′-dichlorofluorescein diacetate (H₂DCF-DA, Invitrogen) was added to all wells. H₂DCF-DA is cleaved by esterases present in the samples, and the nonfluorescent compound H₂DCF is oxidized by peroxyl radicals generated by the thermal decomposition of ABAP in the fluorescent compound DCF, which is detected at the excitation and emission wavelengths of 485 and 530 nm, respectively. Results are presented as the relative area (the difference between the area with and without ABAP divided by the area without ABAP). For interpretation purposes, a low relative area means a high total antioxidant capacity against peroxyl radicals (ACAP) and vice versa.²⁹

2.3.4 Glutathione-S-transferase activity

Glutathione-S-transferase (GST) activity was determined by monitoring the conjugation of 1-chloro-2,4-dinitrobenzene (CDNB Sigma-Aldrich) with GSH, according to the method suggested by Habig and Jakoby.³⁰ As the reaction media, 0.1-M potassium phosphate buffer (pH 7.0) was used, which was previously heated at 25 °C in a water bath. The absorbance of the formed complex was measured at 340 nm (ELx 800, BioTek).

2.3.5 Reduced glutathione (GSH) and protein-sulfhydryl (P-SH) levels

The concentrations of reduced glutathione (GSH) and protein-sulfhydryl groups (P-SH) were measured using 5,5-dithiobis-2-nitrobenzoic acid (DTNB, Sigma-Aldrich) according to Sedlak and Lindsay.³¹ Samples were precipitated with trichloroacetic acid (50% w/v) and centrifuged at 2,000 × g for 10 min at 4 °C. An aliquot of the supernatant with 0.4 Tris-base at pH 8.9 was transferred to clear microplates. Absorbance (405 nm) was measured using 96-well microplates (ELx 800, BioTek). The protein-containing pellet was resuspended in a buffer for P-SH determination. Then, 100 μL of extract and 160 μL of 0.2 Tris-base were added to the microplate and incubated for 15 min. Absorbance (405 nm) was measured using 96-well microplates (ELx 800, BioTek). Concentrations were expressed in μmol of GSH per mg of protein and in μmol of equivalent P-SH per mg of protein for GSH and P-SH, respectively.

2.3.6 Thiobarbituric acid reactive substances

Lipid peroxides were measured as thiobarbituric acid reactive substances (TBARS), following Oakes and Van Der Kraak.³² The method for determining lipid peroxidation was based on the reaction of malondialdehyde (MDA), a lipid peroxidation product, with thiobarbituric acid (TBA, JT Baker), which under high temperature (95 °C) and acidic conditions (acetic acid 20%) produces a chromogen that can be quantified using fluorometry. The resulting chromogen was measured at the wavelengths of 520 and 580 nm for excitation and emission, respectively (Victor 2, PerkinElmer). The concentration of lipid peroxides was given as nmol/TBARS/mg of protein, using tetramethoxypropane (TMP, Sigma-Aldrich) as a standard.

2.3.7 Catalase activity

CAT activity was measured according to the methodology described by Beutler et al.³³ The decomposition of 50-mM hydrogen peroxide (H₂O₂) was quantified using a spectrophotometer (BioMate3, Thermo Scientific) at 240 nm, and the results were expressed in CAT units (1 unit corresponded to the hydrolysis of 1 μmol of H₂O₂ per min and per mg of protein at 25 °C and pH 8.0).

2.4 Physiological assessments

2.4.1 Growth

To assess the growth of the animals, after 96 h of exposure in each generation, 10 adult animals from each sample in all replicates of the experiment were measured using the free ImageJ software, to obtain the average animal size for each treatment in each generation. The results are expressed as growth length (mm).

2.4.2 Reproduction

Three 250-μL aliquots were collected after washing each plate with 50 mL of S-basal liquid medium at the end of each generation in all plates. Reproduction was evaluated by the number of offspring, counted using the CEM method²⁴ divided by the total number of pregnant hermaphrodites. The results are expressed as the number of offspring per fertile adult.

2.5 Statistical analysis

Statistical data were tested for normality and homoscedasticity using Cochran’s and Levene’s tests, respectively.³⁴ All parameters were analyzed using 2-way analysis of variance (ANOVA), with treatment and replication as factors, followed by the Newman–Keuls method (post-hoc). Linear correlations (r) were performed between generations and each of the analyzed variables. A significance level of 5% was used (α = 0.05). The main objective of this study was to compare the results obtained from the groups exposed to DMA^V against the group exposed to As^V, as well as both treatments with the control group, to evaluate the effect of each treatment. Therefore, the figures related to each analysis are composed of graphs that represent each generation (graphics a, b, c, d, and e). Different letters on the bars indicate statistical differences between treatments (P < 0.05). A table is also presented showing the results of the linear correlation analysis (Table 1), where “*” indicates a significant correlation (P < 0.05).

Table 1.

Summary of linear correlations (r) between the generations and each one of the analyzed variables

Treatment	ROS	ACAP	GSH	P-SH	GST	CAT	TBARS.	Gr.	Rep.
Control	0.33	0.74*	−0.19	0.48	−0.59*	−0.68*	0.52*	0.37	−0.23
DMA	−0.17	0.77*	−0.79*	−0.59*	−0.23	−0.73*	−0.71*	−0.60*	0.77*
As	0.69*	0.22	0.54	−0.02	0.72*	−0.35	−0.31	−0.21	−0.27

ROS: Concentration of reactive oxygen. ACAP: Total antioxidant competence against peroxyl radicals (relative area). GSH: Concentration of reduced glutathione (μmoles of GSH per mg of protein). P-SH: Concentration of protein-sulfhydryl groups (μmoles of GSH equivalents per mg of protein). GST: Glutathione-S-transferase activity (nmoles/mg of protein/min). CAT: Catalase activity (enzymatic units). TBARS: Concentration of thiobarbituric acid reactive substances (nmol/TMP equivalents/mg of protein, using tetramethoxypropane or TMP as a standard). Gr.: Growth (mm). Rep.: Reproduction (# of offspring/total number of pregnant hermaphrodites).

^*Significant correlations (P < 0.05).

3. Results

3.1 Biochemical results

3.1.1 Reactive oxygen species

The group exposed to As^V showed an increase in ROS content in the first, second, and fifth generations (Fig. 1a, b, and e, respectively), while the fourth generation showed a reduced ROS concentration when compared with the DMA^V and control groups (Fig. 1d). Animals exposed to DMA^V showed an increase in ROS content only in the third generation (Fig. 1c) when compared with the control group and As^V group, and a decrease in the last generation compared with the same groups (Fig. 1e). A significant positive correlation (P < 0.05) was observed in animals exposed to As^V across generations (Table 1).

Fig. 1.

ROS levels (relative area) measured over 5 generations: first generation (a); second generation (b); third generation (c); fourth generation (d); and fifth generation (e). Data are expressed as mean + 1 standard error (n = 9). Different letters represent significant statistical differences from treatments (P < 0.05).

3.1.2 Total antioxidant capacity against peroxyl radicals

The group exposed to As^V showed a decrease in total antioxidant capacity in the first generation (Fig. 2a) and an increase in the second to fourth compared with the control group. In the second and fourth generations, the As^V treatment also showed an increase in ACAP when compared with the DMA^V group (Fig. 2b, c, and d). The group exposed to DMA^V showed an increase in total antioxidant capacity compared with the control in the first, third, and fourth generations (Fig. 2a, c, and d). DMA^V and As^V had different effects on the total antioxidant capacity in all generations except for the third, where both DMA and As^V increased the total antioxidant capacity compared with the control group (Fig. 2c). In addition, a positive correlation over generations was observed in the control group and in the animals exposed to DMA^V (Table 1).

Fig. 2.

ACAP—Total antioxidant capacity against peroxyl radicals (relative area): first generation (a); second generation (b); third generation (c); fourth generation (d); and fifth generation (e). Data are expressed as mean + 1 standard error (n = 12). Different letters represent significant statistical differences from treatments (P < 0.05).

3.1.3 GSH and P-SH content

Exposure to As^V increased GSH content in the second and fourth generations (Fig. 3b and d), whereas exposure to DMA^V led to an increase in the GSH content in the first generation (Fig. 3a). But in the third and fourth generations, GSH levels decreased (Fig. 3c and d). The animals exposed to DMA^V showed an increase in P-SH content in the first generation (Fig. 4a), followed by a decrease in the second generation (Fig. 4b) as compared with the control group. In the third generation, both DMA^V and As^V increased the P-SH content in animals as compared with the control group (Fig. 4c). DMA^V treatment showed a negative correlation in both GSH and protein-sulfhydryl group content across generations (Table 1).

Fig. 3.

GSH—Reduced glutathione (μmol of GSH/mg of protein): first generation (a); second generation (b); third generation (c); fourth generation (d); and fifth generation (e). And covariance analysis (f). Data are expressed as mean + 1 standard error (n = 12). Different letters represent significant statistical differences from treatments (P < 0.05).

Fig. 4.

P-SH—Protein-bound sulfhydryl groups (μmol of P-SH/mg of protein): first generation (a); second generation (b); third generation (c); fourth generation (d); and fifth generation (e). Data are expressed as mean + 1 standard error (n = 12). Different letters represent significant statistical differences from treatments (P < 0.05).

3.1.4 GST and CAT activity

Both DMA^V and As^V groups decreased GST activity in the third generation (Fig. 5c). Animals exposed to As^V also showed a decrease in GST activity in the first generation (Fig. 5a) and increased activity in the fourth generation compared with the control (Fig. 5d). While exposure to As^V increased CAT activity in the second generation (Fig. 6b) compared with the control and As^V treatments, animals exposed to DMA^V showed an increase in CAT activity in the first and fourth generations compared with the control group and the As^V treatment (Fig. 6a and d). GST and CAT activity did not show differences over generations in the covariance analysis (Figs 5f and 6f). In the control group, a negative correlation was observed between GST and CAT activity between generations, and DMA^V treatment also showed a negative correlation with CAT activity over generations (Table 1).

Fig. 5.

GST—Glutathione-S-transferase activity (nmol of CDNB conjugated/min/mg of protein): first generation (a); second generation (b); third generation (c); fourth generation (d); and fifth generation (e). And covariance analysis (f). Data are expressed as mean + 1 standard error (=12). Different letters represent significant statistical differences from treatments (P < 0.05).

Fig. 6.

CAT—Catalase activity (μmol of H₂O₂/min/mg of protein): first generation (a); second generation (b); third generation (c); fourth generation (d); and fifth generation (e). And covariance analysis (f). Data are expressed as mean + 1 standard error (n = 12). Different letters represent significant statistical differences from treatments (P < 0.05).

3.1.5 Thiobarbituric acid reactive substances

Although exposure to As^V did not increase lipid peroxidation, the group exposed to DMA^V showed an increase in TBARS content in the first, third, and fourth generations (Fig. 7a, c, and d). Although the TBARS content was not different between As^V and the control group in any generation, a statistical difference was observed between DMA^V and As^V over generations (Fig. 7f). The control group showed a positive correlation over generations, while DMA^V treatment showed a negative correlation with this variable (Table 1).

Fig. 7.

TBARS—Thiobarbituric acid reactive substances (nmol of MDA/mg of protein): first generation (a); second generation (b); third generation (c); fourth generation (d); and fifth generation (e). And covariance analysis (f). Data are expressed as mean + 1 standard error (n = 12). Different letters represent significant statistical differences from treatments (P < 0.05).

3.2 Physiological results

3.2.1 Growth

Exposure to As^V and DMA^V affected the growth of animals from the third to the fifth generation (Fig. 8c, d, and e). In the third and fifth generations, there was no difference between animals exposed to As^V and DMA^V (P > 0.05), while in the fourth generation, the animals exposed to DMA^V showed a decrease in growth when compared with the As^V and control groups (Fig. 8d; P < 0.05). In addition, DMA^V treatment showed a significant (P < 0.05) negative correlation with growth over generations (Table 1).

Fig. 8.

Growth (length in mm): first generation (a); second generation (b); third generation (c); fourth generation (d); and fifth generation (e). Data are expressed as mean + 1 standard error (n = 40). Different letters represent significant statistical differences from treatments (P < 0.05).

3.2.2 Reproduction

Animals exposed to As^V showed a decrease in the number of offspring in the first generation, while in the third and fourth generations, As^V induced an increase in reproduction compared with the control and DMA^V treatments (Fig. 9a, c, and d). The animals exposed to DMA^V showed a decrease in offspring in the second and fourth generations, compared with the control and the group exposed to As^V (Fig. 9b and d), with the greatest difference being observed in the fourth generation, where the animals in the group exposed to DMA^V showed a 35% reduction in the number of offspring compared with the As^V treatment (Fig. 9d). In addition, animals exposed to DMA^V showed a significant (P < 0.05) positive correlation over generations (Table 1).

Fig. 9.

Reproduction (number of offspring per adult): first generation (a); second generation (b); third generation (c); fourth generation (d); and fifth generation (e). Data are expressed as mean + 1 standard error (n = 9). Different letters represent significant statistical differences from treatments (P < 0.05).

4. Discussion

The role of As in oxidative stress has been studied for a long time, and although many studies have evaluated the effects of iAs in living organisms, few studies have evaluated the effects of organic forms of As, such as DMA. Once accumulated in tissues, As can cause damage to the organism, and its main toxicity mechanism is the induction of a pro-oxidative scenario. Park et al. (2009)³⁵ showed that a transcriptional shift from growth and maintenance to the activation of cellular defense mechanisms was caused by oxidative stress; many of these transcriptional changes were SKN-1 (ortholog to nuclear factor erythroid 2-related factor (Nrf2) in mammals)-dependent. In fact, Zhang et al.²⁶ showed that C. elegans exposed to As showed an increase in the expression of SKN-1, which activated the expression of genes of the antioxidant defense system (ADS). Nrf2 is an important transcription factor that can be triggered by oxidative stress, and studies have shown that it is sensitive to As exposure^36,37. Under prooxidant conditions, the cytoplasmic Nrf2/Keap1 (Kelch-like ECH-associated protein 1) senses the chemical changes in the cellular environment, then Nrf2 translocates to the nucleus and upregulates the transcription of multiple antioxidant response element-controlled target genes, and finally initiates the transcription of a variety of antioxidant genes.

In this study, we observed that exposure to both As^V and DMA^V changed the redox state of the exposed animals (Figs 1–3), and this change could be responsible for the observed effects on the animal’s growth from the third to the last generation. Exposure to As^V led to an increase in ROS content in the first, second, and fifth generations (Fig. 1a, b, and e). In the fourth generation, a decrease in ROS content was observed with As^V exposure (Fig. 1d), which can be explained by the increase in total antioxidant capacity and GSH levels (Figs 2d and 3d), an important scavenger of ROS.¹⁷ While the increase in ROS content caused by exposure to As^V did not generate lipid damage in animals in any generation (Fig. 7), the increase in ROS content caused by exposure to DMA^V in the third generation led to an increase in lipid peroxidation (Fig. 7c). These results point to the importance of the multigenerational approach for evaluating the toxic effects of contaminants, as these contaminants can have an accumulated effect over generations. In this study, although both treatments increased the generation of ROS, the antioxidant defenses seem to have been effective in mitigating lipid damage in animals exposed to As^V, while the same was not true for animals exposed to DMA^V. This strongly suggests that the toxicity of DMA has been underestimated and that organic As species can be as toxic as, or even more toxic than iAs.

Among the components of ADS, GSH is an important nonenzymatic antioxidant that protects organisms against damage caused by stressors such as metals and other contaminants.³⁸ The ability of an organism to detoxify xenobiotics can be assessed by the activity of GST, a class of enzymes that participate in the detoxification process through the conjugation of xenobiotics with GSH.³⁹ In the third generation, both treatments negatively modulated GST activity, indicating a lower detoxification capacity of the animals in this generation. Despite the reduction in GST activity, animals exposed to As^V showed an increase in ACAP and in the content of protein-bound sulfhydryl groups (P-SH) (Figs 2c and 4c), which was sufficient to prevent lipid damage, even in a prooxidant setting. The same was not observed for the animals exposed to DMA^V, which, even with an increase in ACAP, showed an increase in the TBARS levels compared with the control group (Figs 2c and 7c), showing that even after increasing the antioxidant defenses, the animals could not mitigate the damage caused by exposure to this methylated form of As. However, a major modulation in GST activity was observed in the fourth generation of the animals exposed to As^V, where the GST activity was 2.65 and 3.13 times the values in the control group and in the DMA^V treatment, respectively. In fact, some studies point to this difference in the modulation of GST activity against exposure to As in aquatic organisms, at the same concentrations tested in this study (50 μg L⁻¹), where different forms of As such as As^III, As^V, and DMA^V induce different enzymatic responses.^10,14,15 Chen et al.⁴⁰ observed an increase in the GST activity of fish Oryzas melastigma following chronic exposure to As^III and As^V (100 μg L ⁻¹), and the GST activity was even higher in animals exposed to As^V. Painefilú et al.⁴¹ showed that fish Oncorhynchus mykiss exposed to As^III presented an increase in GST activity, but only at high concentrations (10,000 μg L ⁻¹). In another study, Josende et al.⁹ observed that C. elegans exposed to As^III positively modulated GST activity over 5 generations, which is different from the results observed in this study, indicating that the modulation of GST activity induced by exposure to As can be dependent not only on the concentration and exposure time but also on the chemical form of As. Changes in GST activity may be directly related to an increase in toxicity, since this enzyme is responsible for the detoxification of xenobiotics, and its malfunction can result in the accumulation of toxic compounds.⁴² In addition, the differences between the responses of animals exposed to As^V and DMA^V were even more notable in the correlation results. Animals exposed to DMA^V showed a negative correlation with CAT activity, GSH and P-SH content, and lipid damage, while As^V exposed animals did not show a significant correlation in these parameters, showing that the animals exposed to DMA^V were more susceptible to changes in the redox status (Table 1).

In a prooxidant scenario, the enzyme catalase (CAT) is another ADS enzyme responsible for decreasing ROS levels. Along with the increase in GSH levels (Fig. 3b) and increase in ACAP (Fig. 2b), the positive modulation of CAT activity induced by exposure to As^V contributed to the absence of lipid damage even when the ROS content in this group was 2-fold that of the control group in the second generation (Fig. 1b). CAT is a peroxisomal enzyme that catalyzes the breakdown of hydrogen peroxide (H₂O₂) into O₂ and H₂O. Exposure to As induces the formation of superoxide radicals that are converted into H₂O₂ by the action of superoxide dismutase (SOD). Accumulation of H₂O₂ is prevented in cells by CAT activity.⁴⁰ However, when CAT activity is not sufficient to contain H₂O₂ levels, this reactive species can interact with transition metals, generating hydroxyl radicals through the Fenton reaction. These hydroxyl radicals, in turn, are highly reactive and can react with macromolecules, causing lipid and DNA damage.⁴³ Exposure to DMA^V, on the other hand, positively modulated CAT activity in the first and fourth generations (Fig. 6a and d). This modulation seems to have been sufficient to maintain ROS levels similar to those of the control treatment, but it was not able to prevent lipid damage (Figs 1a and d, 7a and d). In a previous study, C. elegans exposed to 50 μg L⁻¹ of As^III, using the same multigenerational approach, showed an increase in CAT activity from the third generation onwards.²⁴ The authors suggest that the lack of CAT modulation in the first 2 generations can be explained by the prompt activity of GST in these 2 generations.

Josende et al.²⁴ observed an increase in the TBARS content in C. elegans exposed to As^III, and this increase was maintained over generations. In the present study, As^V did not induce lipid damage (Fig. 7), while DMA^V led to an increase in TBARS levels compared with the control group in the first, third, and fourth generations, even with an increase in ACAP observed in the same generations (Fig. 2a, c, and d). ROS, such as hydrogen peroxide, superoxide anion radicals, singlet oxygen, and hydroxyl radicals, can directly or indirectly damage DNA, proteins, and lipids. For hydroxyl radicals to be involved in lipid oxidation through As exposure, a free transition metal (such as iron) is required for the Haber–Weiss processes to cause lipoperoxidation.⁴⁴ When tested as releasers of iron from ferritin, methylated As forms were more active than As^V or As^III.⁴⁵ Furthermore, Yamanaka et al.²² proposed that during DMA metabolism, hydroxyl radicals can be formed, and these interact with DNA and membrane lipids, causing cell damage. These facts may explain why DMA^V exposure resulted in an increase in TBARS content, while As^V exposure did not.

The physiological cost of metabolizing and detoxifying metals or other contaminants to concentrations within tolerable limits can represent a stress factor that reduces animal fitness and causes changes in energy resource allocation, forcing compensation through other parameters during its life history.⁴⁶ Exposure to anthropogenic pollutants disturbs cellular homeostasis due to direct interactions of contaminants and their metabolites with macromolecules, as well as secondary stress due to the generation of ROS. The recovery of homeostasis involves the activation of elimination mechanisms, detoxification of accumulated compounds, and repair or replacement of macromolecules damaged by contaminant exposure, their metabolites, or ROS. These mechanisms are ATP-dependent, and their activation may result in considerable energy costs.⁴⁷ The cumulative lifetime costs of detoxification can be high, causing a trade-off between toxicant tolerance and fitness traits (such as growth, reproduction, or development) in animals.^48,49 Previous studies have shown that the offspring of animals exposed to metals and metalloids have reduced body size compared with unexposed animals.^24,50,51 In the present study, we observed that both As^V and DMA^V led to reductions in body size in the offspring of exposed animals, and this reduction appeared and was maintained from the third generation onwards. In the first 2 generations, although the animals showed a prooxidant state, with an increase in the ROS concentration, modulations in enzymatic and nonenzymatic components of the ADS—mechanisms used to deal with the oxidative scenario—were sufficient to maintain stable animal growth. Nagashima et al.⁵² proposed that exposure to As can reduce the body size of the offspring of exposed animals through the action of As on dopaminergic neurons, suggesting that dopamine regulates the body size in these animals. In fact, Zhang et al.²⁶ showed that the offspring of C. elegans exposed to As^III showed a decrease in growth compared with the control from the first generation, with an associated decrease in dopaminergic neurons, suggesting an effect of the observed increase in the ROS content. Yu and Liao²⁵ evaluated C. elegans exposed to As^III and observed a decrease in the animal growth from the first generation onwards once the animals presented an increase in H3K4 methylation and a decrease in spr-5 levels. It has been reported in C. elegans that deficiency of spr-5 leads to high levels of sterility for many generations. Greer et al.⁵³ suggested that epigenetic changes that alter the methylation status of histones cause changes in reproduction and transgenerational growth in C. elegans. These results suggest that the toxic effects of As are not only derived from oxidative stress but also from interactions with other biological processes such as epigenetics.

Lipid peroxidation depends on the lipid composition, and lipids that present a higher level of unsaturation, such as polyunsaturated fatty acids (PUFA), are the most attractive for the oxidative action of ROS. The initial consequences of this may include the disruption of cell membrane signaling, fluidity, and integrity.⁵⁴ Furthermore, PUFAs are precursors of several lipid metabolites with potent bioactivities that are produced through enzymatic and/or mediated reactions by free radicals. Caenorhabditis elegans oocytes produce prostaglandins; prostaglandins are an important PUFA-derived signal that recruits sperm for fertilization. Recent studies have shown that PUFAs are synthesized in the gut and transported to oocytes to be converted to F-series prostaglandins to guide sperm under the regulation of insulin signaling in the gut and via TGF-β in sensory neurons, thereby controlling reproduction.⁵⁵ In fact, these data seem to corroborate the results observed in this study, where DMA^V caused greater lipid peroxidation than As^V and simultaneously induced greater negative modulation in the reproduction of exposed animals. The correlation results also showed that C. elegans seemed to be more sensitive to DMA^V treatment than As^V treatment over generations in these physiological parameters. Previously, our group reported the genotoxic effects of DMA in the polychaeta Laeonereis culveri.¹⁰ In the present study, we showed the toxic effects of DMA on another organism, C. elegans, which allows a multigenerational approach showing the accumulated effects of As throughout the life history of these organisms. Our results show that the concept that DMA is moderately toxic needs to be reassessed.

5. Conclusion

In this study, we evaluated the toxicological and physiological effects of multigenerational exposure to 2 forms of As in C. elegans to understand the effects of the persistent presence of As in the environment on these organisms. In addition to using relatively low concentrations compared with concentrations found in contaminated regions in Latin America, in this study, we opted to filtering the animals instead of a standard chemical synchronization with strong bases, to avoid the possibility of animals inheriting resistance due to the chemical stress of the synchronization step. Although both treatments exerted modulations on the ADS and physiological parameters such as growth and reproduction over generations, the most severe effects were seen from the third generation onward. The data indicated that DMA^V may be more toxic, being the main contributor to lipid damage and reduction in the reproduction of the exposed animals. In addition, in the correlation analysis, almost every parameter evaluated showed a significant correlation between DMA^V treatment over generations. Several studies in the literature consider the forms of iAs as the main forms responsible for the toxicity of this metalloid, but it is important to emphasize that the vast majority of these studies did not evaluate these effects in a multigenerational way. Although more studies are needed to understand the mechanisms behind the observed responses, our study sheds some light on the accumulated effects of As, showing that toxicity and the role of DMA in As toxicity deserve more attention.

CRediT authorship contribution statement

Larissa Müller: Conceptualization, Formal analysis, Investigation, Methodology, Project administration, Validation, Writing—original draft. Marcelo Estrella Josende: Conceptualization, Formal analysis, Investigation, Methodology, Validation. Gabriela Corrêa Soares: Formal analysis, Investigation, Methodology. José Marìa Monserrat: Conceptualization, Formal analysis, Investigation; Methodology, Project administration, Validation, Writing—original draft; Funding acquisition. Juliane Ventura-Lima: Conceptualization, Formal analysis, Investigation; Methodology, Project administration, Validation, Writing—original draft; Funding acquisition.

Funding

This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior—Brasil (CAPES)—Finance Code 001. Juliane Ventura Lima and José María Monserrat are research fellows at CNPq (313707/2020-0 and 307888/2020-7, respectively).

Declaration of competing interests

The authors declare that they have no competing interests.