Reproductive and Developmental Toxicity of Arsenic in Rodents: A Review

Abstract

Arsenic is a recognized reproductive toxicant in humans and induces malformations, especially neural tube defects, in laboratory animals. Early studies showed that murine malformations occurred only when a high dose of inorganic arsenic was given by intravenous or intraperitoneal injection in early gestation. Oral gavage of inorganic arsenic at maternally toxic doses caused reduced fetal body weight and increased resorptions. Recently, arsenic reproductive and developmental toxicity has been studied in situations more similar to human exposures and using broader endpoints, such as behavioral changes and gene expression. For the general population, exposure to arsenic is mostly oral, particularly via drinking water, repeated and prolonged over time. In mice and rats, methylated or inorganic arsenic via drinking water or by repeated oral gavage induced male and female reproductive and developmental toxicities. Furthermore, at nonmaternally toxic levels, inorganic arsenic given to pregnant dams via drinking water affected fetal brain development and postnatal behaviors. However, arsenic given by repeated oral gavage to pregnant mice and rats was not morphologically teratogenic. In this review of arsenic reproductive and developmental toxicity in rats and mice, the authors summarize recent in vivo studies and discuss possible underlying mechanisms. The influences of folate, selenium, zinc, and arsenic methylation on arsenic reproductive and developmental toxicity are also discussed.

Arsenic is a major global health concern due to its wide distribution and adverse health effects. Arsenic naturally occurs in the Earth’s crust, and can contaminate drinking water sources through leaching, erosion, and mining. High concentrations of arsenic in water have been reported in Asia (Bangladesh, China, India, Inner Mongolia, and Taiwan), Europe (Hungary), and the Americas (Argentina, Chile, Mexico, and the northeast and western United States of America) (Akter et al. 2005; Ayotte et al. 2003; Karagas et al. 2002; National Research Council 1999, 2001). Grains and vegetables grown in arsenic-contaminated soil or irrigated with arsenic contaminated water incorporate arsenic in their tissue. Arsenic is also used in the manufacture of wood preservatives, glass, pesticides and herbicides, semiconductors, and pharmaceuticals. Therefore, ingestion of arsenic-containing food, inhalation of arsine in industrial settings, and exposure to arsenical herbicides and pesticides can also contribute to arsenic intake (Akter et al. 2005). For the general population, arsenic in drinking water is the main exposure source (National Research Council 2001), and more harmful than arsenic in food, because the bioavailability (actual amount absorbed into the bloodstream) of arsenic from water is greater than that from grains or vegetables (Akter et al. 2005). Furthermore, arsenic in drinking water is mainly inorganic arsenic. Arsenic found in seafood is predominately organic forms, such as arsenobetaine. The organic forms tend to be less potent toxicants than inorganic arsenic (De Gieter et al. 2002; Guillamet et al. 2004; Sakurai et al. 2004).

Developmental effects, cancer, and cardiovascular disease have all been associated with long-term exposure to arsenic in humans. There is a paucity of epidemiological studies of arsenic reproductive toxicity, and there are only a few of arsenic developmental toxicity (DeSesso et al. 1998; Tchounwou et al. 2003). Epidemiological studies have reported that arsenic exposure in utero increased spontaneous abortion and stillbirth and decreased birthweight (Ahmad et al. 2001; DeSesso et al. 1998; Ihrig et al. 1998; Milton et al. 2005). However, these studies lack detailed information on confounders (exposure to other metals, smoking, maternal age, etc.) and accurate maternal arsenic exposure. The use of animal models therefore is essential for investigating arsenic developmental and reproductive toxicities, because it allows greater experimental controls, including the confounders mentioned above, and direct observations of developmental changes before birth. Furthermore, it provides opportunities to investigate the effects of interventions that has not been approved for humans.

In early studies, arsenic induced fetal malformations in rats and mice after an intraperitoneal (i.p.) or intravenous (i.v.) injection in early gestation (DeSesso 2001; DeSesso et al. 1998; Holson et al. 2000a; Hood 1998; Stump et al. 1999). Both inorganic and methylated arsenic injections caused developmental toxicity at maternally toxic or near lethal doses. In contrast, single or repeated oral exposure to inorganic arsenic did not induce fetal gross malformation in mice or rats (Holson et al. 2000b; Stump et al. 1998a, 1999). At maternally toxic doses, repeated oral exposures to dimethylarsinic acid increased resorptions and decreased fetal weight in mice and rats (Rogers, Chernoff, and Kavlock 1981). Although no gross abnormalities were reported in these rats, an increase in incidences of cleft palate was observed in mice by Roger, Chernoff, and Kavlock (1981). Inhalation of inorganic arsenic or arsine did not cause developmental toxicity (Morrissey et al. 1990; Stump et al. 1998b). In summary, the above studies detected arsenic developmental toxicity only at maternally toxic doses.

In the past 5 years, arsenic reproductive and developmental toxicity studies have expanded to include more human-related exposure conditions and diverse endpoints. In particular, following United States of America Food and Drug Administration (US FDA) Guidelines for Developmental Toxicity Studies (US FDA 2000), recent arsenic studies used repeated or prolonged oral exposures (via drinking water or gavage). Reproductive and non–gross developmental endpoints (molecular events, brain development, and behaviors) were also investigated. These studies revealed that prolonged exposure to arsenic could cause developmental toxicity at maternally nontoxic levels. For example, maternal exposures to inorganic arsenic in drinking water throughout gestation affected fetal brain development and newborn behaviors in rats (Chattopadhyay et al. 2002; Rodriguez et al. 2002).

ARSENIC METABOLISM

Inorganic arsenic is the most common form of arsenic in the environment. Inorganic arsenic is readily absorbed through the gastrointestinal tract, and biotransformed in the liver and other tissues. Most mammals, including mice, rats, and humans, biotransform inorganic arsenic by alternating reduction and oxidative methylation (Figure 1) (Carter, Aposhian, and Gandolfi 2003; Vahter 2002). The classical arsenic transformation is as follows: from inorganic pentavalent arsenate (As^V) to inorganic trivalent arsenite (As^III), to monomethylarsonic acid (MMA^V), to monomethylarsonous acid (MMA^III), to dimethylarsinic acid (DMA^V), to dimethylarsinous acid (DMA^III), and in some animals, including rats (Lu et al. 2003), to trimethylarsine oxide (TMA^VO). The reduction is facilitated by reductases and the reduced form of glutathione (GSH) along with possibly other thiols as electron donors. Oxidative methylation is carried out by methyltransferases, and S-adenosylmethionine (SAM) serves as the main methyl donor. Both mice and rats are more effective in arsenic methylation than humans.

Inorganic and methylated arsenicals are excreted mainly in the urine and to a lesser degree in the bile. More importantly, arsenic can pass through the placenta to the developing fetus. When pregnant mice inhaled arsine gas, fetal brain and liver had higher arsenic concentrations than placenta and maternal liver (Miyazaki et al. 2005). Inorganic arsenic in the drinking water of pregnant rats can also result in arsenic accumulation in fetuses, including the fetal brain (Rodriguez et al. 2002). Inorganic, monomethyl and dimethyl arsenicals all have been detected in the fetuses of pregnant mice exposed to As^III or As^V through i.p. injection or oral gavage (Hood et al. 1987, 1988). However, placenta and maternal blood, urine, liver, and kidneys were only analyzed for total arsenic in Hood’s studies (1987 (1988). It is not clear what form(s) of arsenic transfer across the placenta or if fetuses methylate arsenic. Because the exposure to As^III or As^V results in exposure to its methylated metabolites, these studies cannot tell the active form of arsenic that is directly responsible for developmental toxicity. More information is therefore needed to identify the importance of arsenical species and methylation in placental transfer and arsenic developmental toxicity in mice and rats. In humans, DMA^V accounts for nearly 90 % of all detected arsenic in the blood plasma of both the newborns and their mothers (Concha et al. 1998). This form of arsenic accounts for 60% to 70% of the total detected arsenic in urine of the general population, whereas urine from pregnant women contained more than 90% arsenic as DMA^V. This suggested that arsenic methylation may be increased during pregnancy and that DMA^V is the major form of arsenic transferred to the fetus.

Rats are known to retain arsenic, as a consequence of their red blood cells having a high affinity for arsenic, particularly DMA^III (Lu et al. 2004). The whole body retention of inorganic arsenic in rats was 20 times higher than that in similarly exposed mice (Vahter 1981). In spite of the arsenic accumulation in rat erythrocytes, arsenic distributes to other tissues including the fetus as well. Although not an ideal test species for arsenic kinetic studies, rats are an appropriate animal model for arsenic reproductive and developmental toxicity studies.

REPRODUCTIVE EFFECTS OF ARSENIC IN MICE AND RATS

Male Reproductive Toxicity

Arsenite exposure causes male reproductive toxicity when given through drinking water (Chinoy, Tewari, and Jhala 2004; Pant et al. 2001, 2004) or by i.p. injection (Sarkar et al. 2003). As^III interferes with spermatogenesis (Pant et al. 2001, 2004; Sarkar et al. 2003) and alters activities of spermatogenetic enzymes (Chinoy, Tewari, and Jhala 2004; Pant et al. 2001, 2004). Furthermore, As^III lowers levels of testosterone and gonadotrophin (Chinoy, Tewari, and Jhala 2004; Sarkar et al. 2003) (Figure 2). These results suggest that arsenic may act on the brain or pituitary as well as directly on the germ cells (Sarkar et al. 2003).

Male mice exposed to sodium arsenite in drinking water showed reproductive toxicity without clinical effects (Pant et al. 2001). Sodium arsenite was given to mice via drinking water at up to 533.90 μmole/L for 35 days. As^III-treated mice did not show changes in body weight, testes weight, or accessory sex organ weights. However, at 533.90 μmole/L, the activity of 17β-hydroxysteroid dehydrogenase (HSD) was decreased. Conversely, the activities of lactate dehydrogenase (LDH) and γ-glutamyltranspeptidase (γGT) were increased in the testes. LDH was used as a marker of Leydig cell function, and γGT as a marker of Sertoli cell function. In addition to amino acid transport across the plasma membrane, γGT regulates GSH levels and contributes to protection against oxidative stress (Meroni et al. 2000). Because arsenic lowers follicle-stimulating hormone (FSH), which in turn may decrease the activity of γGT (Schteingart et al. 2002), the observed increases in the activity of γGT could result from arsenic-induced oxidative stress. As^III-treated mice also showed decreases in sperm count and motility along with an increase in abnormal sperm. One possible cause of arsenic-decreased sperm motility may be arsenic binding to thiols (Uckun, Liu, and D’cruz 2002). Sperm nuclear chromatins have large amounts of thiol-rich protamines, and the sperm flagellum is rich in thiols. In addition, arsenic might affect sperms through lowering levels of gonadotrophin and testosteronen (see below).

Spermatogenesis and plasma levels of gonadotrophin and testosterone were affected by As^III in rats (Sarkar et al. 2003). Sodium arsenite was given to Wistar rats via i.p. injections at 4, 5, or 6 mg/kg/day for 26 days. At 5 and 6 mg/kg/day, relative testicular weight, accessory sex organ weights and epididymal sperm counts were decreased. The same was true for plasma concentrations of luteinizing hormone (LH), FSH, and testosterone. Because late stages of spermatogenesis were especially sensitive to testosterone, quantitative analysis of spermatogenesis was carried out by counting the relative number of each variety of germ cells at stage VII of the seminiferous epithelium cycle, as defined by Leblold and Clermont (1952). Massive degeneration of all the germ cells at stage VII was observed at 5 and 6 mg/kg/day. These authors suggested the observed arsenic-induced low levels of LH and FSH might be the trigger of suppressed testosterone synthesis. Low testosterone consequently increased spermatid degeneration. Although As^III may act on the brain or pituitary to suppress LH and FSH levels, direct inhibition on germ cells by binding to thiol cannot be ruled out. Another possible cause of the reduction in serum LH, FSH, and testosterone levels could be high serum corticosterone levels. High corticosterone can reduce serum gonadotrophin and testosterone levels (Hardy et al. 2005; Vreeburg et al. 1988), and has been reported in As^III-treated rats (Biswas, Roy Chowdhury, and Sarkar 1994).

In mice, in addition to spermatogenesis, cholesterol metabolism and testicular testosterone level were affected by As^III (Chinoy, Tewari, and Jhala 2004). Male Swiss mice were given arsenic trioxide (As₂O₃) orally at 0.5 mg/kg for 30 days. Treated mice showed increased cholesterol levels and decreased protein levels in the testes. Testicular structural damage observed included degeneration of tubules and denudation of germinal epithelial cells. There was also a lack of sperm in the lumen of seminiferous tubules. In addition, testicular activities of 3β-HSD and 17β-HSD, and testosterone levels in the serum were decreased. In the testis, cholesterol in the interstitial tissue is used for testosterone synthesis (Kabbaj et al. 2003). 17β-HSD converts androstenedione into testosterone. In the seminiferous tubes of the testis, cholesterol in the membrane of developing cells influences the gamete’s fertility (Kabbaj et al. 2003). These data suggest that low serum testosterone after As^III exposure was due to low enzymatic conversion (17β-HSD), rather than a lack of the synthetic precursor (cholesterol). Chinoy, Tewari, and Jhala (2004) also tested the effects of coexposures to arsenic and fluoride (NaF), and found that the recovery from arsenic and fluoride-induced effects can be facilitated by ascorbic acid, calcium, and vitamin E. These results suggest arsenic- and fluoride-induced reproductive toxicity was at least in part mediated by oxidative stress.

Male reproductive effects of long-term exposure to As^III via drinking water were investigated in mice by Pant et al. (2004). Swiss albino mice were given sodium arsenite (NaAsO₂ at 53.39 μmole/L, equivalent to 4 ppm arsenic) via drinking water for 365 days. The mice showed decreases in absolute and relative testicular weights, but no change in epididymal or accessory sex organ weights. Sperm count and sperm motility were decreased, and the percentage of abnormal sperm was increased. Additionally, the activities of marker testicular enzymes were altered. For example, enzymes associated with postmeiotic spermatogenic cells showed changes in both directions. The activities of sorbitol dehydrogenase and acid phosphatase were decreased, and the activity of LDH was increased. The changes in biochemical activities of these testicular enzymes, which are associated with specific types of germ cells, suggested damage to germ cells. The testicular activity of γ-GT, associated with Sertoli cells, was increased. Meanwhile, activity of 17β-HSD, which converts androstenedione to testosterone in Leydig cells, was decreased. These authors suggested the decreased 17β-HSD activity might be due to low levels of plasma gonadotrophins, which has been reported in rats i.p. injected with sodium arsenite (Sarkar et al. 2003).

Female Reproductive Toxicity

In female mice and rats, inorganic arsenic suppresses ovarian steroidogenesis, prolongs diestrus, and degenerates ovarian follicular and uterine cells (Chattopadhyay et al. 2001; Navarro, Liu, and Keefe 2004; Zhang et al. 2000). It also increases meiotic aberrations in oocytes, and decreases cleavage and pre-implantation development (Navarro, Liu, and Keefe 2004).

Arsenic can induce ovarian and uterine toxicity, and influence neuroendocrine regulation of female sex hormones (Chattopadhyay et al. 2001). In female Wistar rats gavaged with 10 ml of 0.4 ppm sodium arsenite daily for 28 days, a consistent diestrous stage was observed. There were also decreases in relative ovarian and uterine weights, activities of Δ⁵-3β-HSD and 17β-HSD in ovary, and the activities of peroxidase in the ovary and uterus. Moreover, levels of LH, FSH, and estradiol in the plasma, and norepinephrine levels in midbrain and diencephalon were decreased, whereas serotonin levels in midbrain and diencephalon were increased. The primary cause of the observed As^III toxicity in the female reproductive system could be arsenic-induced changes in the levels of catecholamines in the brain. The elevation in serotonin and decrease in norepinephrine in the midbrain and diencephalon could lower gonadotrophin synthesis and secretion. Low gonadotrophin levels could in turn decrease activities of ovarian Δ⁵-3β-HSD and 17β-HSD, two important regulatory enzymes for steroidogenesis (Ghersevich et al. 1994a, 1994b; Kaminski et al. 1997; Miro et al. 1995). These observations suggest that low plasma levels of estradiol could be the cause of consistent diestrous. These arsenic-induced ovarian and uterine toxicities and steroidogenic dysfunction were decreased by coadministrations of ascorbic acid orally. Possible mechanisms of ascorbic acid protection included its antioxidant property, facilitating the elimination of arsenic, and influences on hormones. Regarding hormonal influences, ascorbic acid can enhance endogenous norepinephrine secretion and consequently stimulate gonadotrophin releasing hormone release (Miller and Cicero 1987). Ascorbic acid also facilitates the synthesis and secretion of gonadotrophins from the anterior pituitary (Wun et al. 1994), and is a stimulator of gonadal steroidogenesis (Murray et al. 2001).

Using the same strain of rat and As^III treatment, Chattopadhyay et al. later observed ovarian follicular and uterine cell degeneration (Chattopadhyay et al. 2003). This was accompanied by increases in dopamine levels in the midbrain and diencephalon, as well as arsenic levels in the ovary, uterus, and plasma (Chattopadhyay et al. 2003). Similarly to norepinephrine, low dopamine levels could decrease gonadotrophin synthesis and secretion. The observed low FSH level may contribute to the observed decreased number of healthy follicles and increased number of apoptotic follicles. The authors suggested that uterine cell degeneration may be due to low ovarian estradiol and /or increased production of reactive oxygen species (ROS) after arsenic treatment. The As^III toxicity in the female reproductive system was decreased by coadministrations of sodium selenite orally. Conversely, this selenite supplement did not reduce As^III-increased activities of renal and hepatic enzymes. The causes of selenite selective protection against arsenic toxicity are not clear. Other studies showed that arsenic and selenium decrease the tissue concentrations of each other by increasing mutual excretions (Berry and Galle 1994; Csanaky and Gregus 2003; Zeng, Uthus, and Combs 2005). Gastrointestinal and biliary excretion was increased via the formation of seleno-bis(S-glutathionyl) arsinium ion (Csanaky and Gregus 2003; Zeng, Uthus, and Combs 2005). Arsenic and selenium also precipitate each other by forming insoluble selenide (As₂Se), which were seen in lysosomes of renal cells and in dense deposits in the urinary lumen (Berry and Galle 1994; Zeng, Uthus, and Combs 2005). These lysosomes and their precipitate were consequently excreted in the urine. Furthermore, selenite inhibits As^III- and As^V-induced activation of C-Jun N-terminal kinase (JNK), activator protein-1 (AP-1), and nuclear factor-κB (NFκB) signaling. Consequently, selenite inhibits arsenic-induced apoptosis and necrosis.

Oocyte meiotic abnormalities and compromised preimplantation development were observed in As^III-treated mice (Navarro, Liu, and Keefe 2004). Female CD-1 mice were i.p. injected with 0, 8, or 16 mg/kg sodium arsenite every 2 days for a total of 7 injections over 14 days. Superovulation was induced by injections of equine and human chorionic gonadotrophins overlapping the end of As^III treatment. Metaphase II oocytes from these As^III-treated mice had increased meiotic aberrations, characterized by spindle disruption and chromosomal misalignment. Additionally, zygotes from As^III-treated mice showed lower rates of cleavage, decreased morula formation, and decreased development to blastocysts. More apoptotic nuclei were seen in the blastocysts of As^III-treated mice. The authors suggested that arsenic-induced meiotic aberrations could subsequently compromise oocyte fertilization, preimplantation development, and embryo viability. Some of these arsenic effects on oocytes were observed at 8 mg/kg, a previously established maternal non-observed-adverse-effect level (NOAEL).

DEVELOPMENTAL EFFECTS OF ARSENIC IN MICE AND RATS

Developmental toxicity is the adverse effects of an agent on the developing organism. Developmental toxicity can be manifested as death, structural anomaly, altered or retarded growth, and functional deficiency (US FDA 2000). The last includes biological dysfunctions and behavioral deficits that become evident as the animal grows.

Developmental toxicity observed without maternal toxicity is a clear indication of selective toxicity to the embryo/fetus. However, maternal toxicity does not preclude the possibility that an agent is also a developmental toxicant. US FDA 2000 Redbook, Guidelines for Developmental Toxicity Studies (US FDA 2000), stated that “developmental effects that occur in the presence of minimal maternal toxicity are considered to be evidence of developmental toxicity, unless it can be established that the developmental effects are unquestionably secondary to the maternal effects. In situations where developmental effects are observed only at doses where there is a substantial amount of maternal toxicity, then the possible relationship between maternal toxicity and the developmental effects should be evaluated.” Maternal toxicity can be measured as changes in body weight and adjusted body weight, and feed and fluid consumption. Daily clinical observations and necropsy data, such as organ weights, are also maternal toxicity parameters.

In 2001, the US EPA announced a new maximum contaminant level for arsenic in drinking water of 10 μg/L (10 ppb). Several excellent reviews of arsenic developmental toxicity had been published prior to the year 2001 (DeSesso et al. 1998; Holson et al. 2000a; US EPA 1998). In the present review, new findings of arsenic developmental toxicity are emphasized. In vivo studies and those following FDA guidelines (US FDA 2000) provide the most relevant information to humans. Additionally, studies on the influences of folate and methylation on arsenic developmental toxicity may help in identifying susceptible populations. All gestation days (GDs) in this review have been adjusted such that gestation day (GD) 0 was recorded upon demonstration of vaginal sperm plug in mice or sperm in the vaginal smear in rats.

Inorganic arsenicals, As^III and As^V, are more toxic than organic arsenicals to embryos/fetuses (National Research Council 1999). Similar to other systemic toxicity, the teratogenic potential is greater from As^III than As^V (Hunter 2000; Lammon et al. 2003). Few studies have examined developmental effects of arsine gas (Morrissey et al. 1990) or organic arsenic (Chernoff et al. 1990; Rogers, Chernoff, and Kavlock 1981); nor have trivalent organic arsenicals been studied in this respect. Inorganic arsenic given to pregnant dams induces slow development, behavior changes, and malformations in the fetus (Chattopadhyay et al. 2002; DeSesso 2001; DeSesso et al. 1998; Holson et al. 2000b; National Research Council 1999; Stump et al. 1999). Arsenic-induced malformations have been reported in the neural tube, skull and skeleton, eye, and urogenital system.

Arsine Gas

Arsine gas (AsH₃) was not fetotoxic or teratogenic in rats or mice (Morrissey et al. 1990). Pregnant F344 rats and CD-1 mice were exposed to up to 2.5 ppm (8 mg/m³) arsine by inhalation for 6 h/day during GDs 6–15. No developmental or reproductive toxicity was observed, although maternal splenomegaly and evidence of hemolysis occurred in the 2.5 ppm group (Morrissey et al. 1990). When pregnant rats were exposed to up to 5 ppm arsine during GDs 6 to 17, the arsenic concentrations in both maternal blood and fetal liver were increased in a dose-dependent manner. This indicated that the lack of arsine fetotoxicity/teratogenicity was not due to the lack of embryonic arsenic exposure. The arsine NOAEL for maternal toxicity was 0.5 ppm (increased spleen weight in rats and mice). This was also the arsine NOAEL for developmental effects (increase in average fetal body weight per litter in rats).

Methylated Arsenic

Dimethylarsinic acid given to pregnant mice and rats by oral gavage caused developmental toxicity (Chernoff et al. 1990; Rogers, Chernoff, and Kavlock 1981). CD-1 mice were orally gavaged with DMA^V at 200, 400 or 600 mg/kg/day during GDs 7 to 16 (Rogers, Chernoff, and Kavlock 1981). These mice showed lower maternal weight gain and fetal weight at 200 mg/kg/day, and a higher incidence of cleft palate at 400 mg/kg/day. Similarly, during GDs 7 to 16, CD rats were orally gavaged with DMA^V at 7.5 to 60 mg/kg/day (Rogers, Chernoff, and Kavlock 1981). Maternal weight gain and fetal weight were decreased at 40 mg/kg/day and higher. Fetal mortality was increased at 50 or 60 mg/kg/day. However, no fetal gross malformations were seen in these rats. In a later study, DMA^V was given to pregnant Sprague-Dawley (SD) rats at 40 mg/kg/day by oral gavage during GDs 6 to 15. This treatment did not induce maternal weight reduction or maternal lethality, but decreased fetal weight (Chernoff et al. 1990).

Inorganic Arsenic

Fetal malformations were only reported when pregnant rats and mice were i.v or i.p. injected with inorganic arsenic at early gestation (DeSesso 2001; Stump et al. 1999). Maternal inhalation or oral ingestion of inorganic arsenic affected fetal development and behavior, but did not cause malformations (Chattopadhyay et al. 2002; DeSesso 2001; Holson et al. 1999, 2000b; Stump et al. 1999). The importance of administration routes in determining adverse developmental effects of inorganic arsenic can be explained by toxicokinetic differences. The toxicokinetic differences among administration routes include maternal circulation levels of arsenicals as well as rates and pathways of biotransformation and excretion (DeSesso 2001). The maternal circulation levels of arsenicals are influenced by absorption rates. In oral exposure, arsenic is absorbed into the blood from the intestines. It is then transported to the liver and may undergo first-pass metabolism prior to being delivered to the uterus (DeSesso 2001; Stump et al. 1999). Additionally, first-order elimination was observed for maternal arsenic in mice given inorganic arsenic (Hood et al. 1987, 1988). Intraperitoneal injections, on the other hand, allow arsenic to be taken up by blood vessels directly. Some arsenic may bypass first-pass metabolism by going into vessels that line the inner surface of peritoneal cavity (DeSesso 2001; Stump et al. 1999). Oral administration of arsenic at a dose twice that used in i.p. injection resulted in a peak maternal arterial blood arsenic concentration that was roughly only 30% of the i.p. injection (DeSesso et al. 1998; Hood et al. 1987). Furthermore, the uterus may be directly exposed to arsenic in the peritoneal cavity, which may have higher concentrations than blood. As a result, the arsenic concentration differences in embryos from mothers exposed to arsenic from i.p. injection and oral gavage are even greater than those in maternal blood. A maximal embryonic inorganic arsenic concentration after maternal oral exposure to a dose twice that used in maternal i.p. injection was only 22% of that caused by the i.p. injection (Hood et al. 1987). When the same dose of arsenic was given to pregnant rats, the embryonic total arsenic concentration from i.p. injected mothers was more than 10 times higher than that from the orally exposed mother (Holson et al. 2000a). Inhalation is the least effective means of increasing maternal or embryonic arsenic concentrations, compared to i.p. and i.v. injections and oral exposure (Holson et al. 2000a).

Intraperitoneal Injection

Consistent with earlier studies, sodium arsenate given by a single i.p. injection induced fetal malformations in Swiss mice (Fascineli et al. 2002). Without affecting maternal weights, an i.p. injection of 45 mg/kg sodium arsenate on GD 8 decreased placental weight and increased fetal malformations. The increased malformations included external (exencephaly and eye abnormalities), visceral (hydrocephalus and hydronephrosis) and skeletal malformations.

Oral Gavages

In a multiple administration study (Holson et al. 2000b), arsenic trioxide via oral gavage did not cause neural tube defects, even at maternally toxic dose levels. Female Crl:CD(SD)BR rats were gavaged with arsenic trioxide from 14 days prior to mating through GD 19. At the highest dose tested (10 mg/kg/day), fetal weights were decreased. There were no arsenic-induced changes in mating index, fertility index, implantation, or fetal malformation. It is worth noting that maternal toxicity1 was observed at 10 mg/kg/day and the maternal NOAEL was 2.5 mg/kg/day due to transient decreases in food consumption at 5 mg/kg/day.

In Drinking Water

When As^III was given to pregnant rats in the drinking water throughout gestation, fetal behavior and brain development were affected (Chattopadhyay et al. 2002). Sodium arsenite was administered to pregnant rats in drinking water at 0.03, 0.3, and 3 ppm. Although rats exposed to up to 0.3 ppm As^III could complete gestation and parturition on schedule, an exposure of 3 ppm caused 25% neonatal death. At 0.3 and 3 ppm, both post gestational mothers and 1-day-old neonatal pups showed decreased spontaneous behavior, with this effect being more dramatic in neonatal pups. The spontaneous behavior was measured as the changes of weight of an animal when placed on a single pan balance. It included, and was not limited to, movement, shaking, tremors, and grooming. Neonatal rat brain cells derived from pups of 0.3 ppm arsenic–treated mothers had increased cell membrane damage (as measured by Trypan Blue dye exclusion test). They also showed increased intracellular generation of ROS and nitric oxide (NO), and decreased DNA and protein synthesis (measured by ³H-thymidine and ¹⁴C-leucine incorporation, respectively). These data demonstrated that the developing brain can be affected by in utero exposure to non–maternal-lethal levels of As^III in drinking water.

Postnatal developmental changes were observed when As^III was given in the drinking water to pregnant and lactating rats and continually to the newborns (Rodriguez et al. 2002). Sodium arsenite at 36.7 mg/L was administered to SD rats from GD 15 or postnatal day 1, until newborns were approximately 4 months old (Rodriguez et al. 2002). Weaned pups received the same As^III treatments in water as mothers. Whereas both female and male pups were assessed for developmental indices, only male pups were subjected to behavioral tests. The behavioral tests measured spontaneous locomotor activity in a chamber and motor coordination on a rotating cylinder. Two learning tasks, spontaneous and delayed alternation tests, were also included in the behavioral tests. Maternal behaviors (retrieval of pups, cleaning and sniffing pups, nest-building, and self-grooming) and body weights were unaffected by either arsenic treatment. In behavioral tests, the pups in the group exposed from GD 15 showed increased spontaneous locomotor activities, and pups in both exposed groups showed increased numbers of errors in a delayed alternation task in comparison to the pups in the untreated control group. Because performing a delayed alternation task requires sensory information on the body in space, the increased errors in arsenic-exposed pups suggested that the striatum, hippocampus, and prefrontal cortex, along with neurotransmitter metabolism, may be affected by arsenic. Among the developmental indices, the group exposed from GD 15 had more litters showing full pinna detachment on postnatal day 12. Conversely, more litters showed low ratings on eye opening on postnatal day 14, compared to the untreated controls. However, there was no difference in these developmental indices on postnatal day 16. These data showed that arsenic induced an asynchrony of the maturation processes during postnatal development. Furthermore, arsenic caused behavioral changes, including deficits in spontaneous locomotor, activity, and more errors in completing a spatial learning task.

Influence of Selenium

In utero exposures to As^III decreased the activity of thioredoxin reductase (an antioxidant enzyme) in the brain (Miyazaki et al. 2005). Furthermore, combination of As^III and a selenium-deficient diet decreased the activity of type II iodothyronine deiodinase (a selenoenzyme important for brain development) (Miyazaki et al. 2005). Pregnant ICR mice were fed a selenium-sufficient or selenium-deficient diet from GDs 0 to 16. From GDs 7 to 16, half of the rats were orally gavaged with 58 μmol/kg/day sodium arsenite. On GD 17, maternal and fetal tissues were harvested and analyzed. None of the treatments caused changes in maternal weight, litter size, mortality, or fetal body weight. The selenium-deficient diet increased arsenic concentrations in maternal liver and fetal brain in arsenic-treated groups. Arsenic decreased thioredoxin reductase activities in the fetal brain in both diet groups, and in the fetal liver in the selenium-deficient group. This decrease of antioxidant selenoenzyme activity could enhance oxidative stress and damage. Furthermore, with mothers on a selenium-deficient diet (but not selenium-sufficient diet), fetal brains showed arsenic-increased activity of type II iodothyronine deiodinase. Among four types of iodothyronine deiodinases, types II and III are expressed in the brain (Kodding et al. 1986; Polk 1995). Type II iodothyronine deiodinases transform inactive thyroxine (T4) to receptor-active triiodothyronine (T3) by outer ring deiodination. Meanwhile, type III iodothyronine deiodinases transform inactive T4 to inactive reverse T3 by inner ring deiodination. The balance of these activities is important in determining brain T3 level, which affects brain development. Miyazak (2005) therefore suggested that in utero exposure to As^III in conjunction with selenium deficiency might disturb fetal thyroid hormone balance in the brain, and potentially brain development.

Influence of Zinc

Zinc did not lessen arsenic fetotoxicity in mice (Fascineli et al. 2002). Both arsenic and zinc are known to induce expression of metallothionein (Liu et al. 2001), a protective protein that binds heavy metals. Metallothionein isotype 1 may be involved in developmental processes during gestation (Nordberg and Nordberg 2000). Zinc also induced arsenic tolerance in mice in a metallothionein independent manner (Kreppel et al. 1994). Zinc, however, did not ameliorate arsenic teratogenicity, when it was given either prior to arsenic or simultaneously with arsenic (Fascineli et al. 2002). For zinc pretreatment, Swiss mice were gavaged with zinc on GDs 7 and 8, and i.p. injected with As^V on GD 8. For simultaneous treatment, mice were i.p. injected with As^V plus zinc on GD 8. The dose of sodium arsenate (NaHAsO₄ · 7H₂O) was 45 mg/kg. Zinc sulfate (ZnSO₄ · 7H₂O) was given by oral gavage at 40 or 20 mg/kg or i.p. injection at 10 or 5 mg/kg. Controls received no treatment, As^V alone or zinc alone. When dams and fetuses were examined on GD 17, As^V alone decreased fetal and placental weights and increased fetal malformations. Zinc alone caused delayed fetal development, but not malformations, with the exception that an i.p. injection of 10 mg/kg zinc caused exencephaly. Excenphaly is a condition in which the brain is located outside the skull due to defects in neutral tube closure. Exposures to zinc plus As^V (in sequence and simultaneously) decreased maternal weight gain, fetal weight and placental weight, and delayed fetal ossification. Moreover, zinc did not decrease As^V-induced malformations. In fact, vertebrate skeletal anomalies were more frequent in the zinc plus As^V group than As^V alone group. This report demonstrated that neither pretreatment nor simultaneous treatment of zinc prevented arsenic-induced teratogenicity in mice. The fetotoxicity from arsenic and zinc exposure may be a combination of their effects on the fetus, placenta, and mother.

Influence of Folate

Folate can affect arsenic methylation, and both folate deficiency and arsenic can induce malformations, including neural tube defects (NTDs). Using transgenic mice deficient in folate transport, the relationship between folate and arsenic developmental toxicity has been intensively studied in the past 5 years. Arsenic methylation is catalyzed by methyltransferases and requires S-adenosylmethionine (SAM) as the methyl donor. SAM is eventually regenerated through the homocysteine remethylation cycle, a process that requires 5-methyltetrahydrofolate as a cofactor (Figure 3). Thus, inorganic arsenic metabolism is dependent on the folate supply (Spiegelstein et al. 2003). Both folate deficiency and arsenic exposure have been reported to increase congenital malformations, including NTDs, in humans and rodents. Folate enters cells through folate binding proteins (Folbp) in conjunction with reduced folate carriers (RFC). Mouse Folbp1−/− embryos have multiple structural malformations, including NTDs, and die before GD 10.5. Similarly, RFC−/− embryos die immaturely in utero. On the other hand, Folbp1+/−, Folbp2−/−, and RFC+/− embryos develop normally with no apparent congenital abnormalities (Spiegelstein et al. 2005b; Wlodarczyk et al. 2001).

Mice lacking a functional Folbp2 gene (Folbp2−/−) were more sensitive to in utero arsenic exposure than wild-type mice, and a folate-deficient diet further increased arsenic-induced teratogenicity (Spiegelstein et al. 2005b; Wlodarczyk et al. 2001). In a study of embryonic genotype effects on arsenic teratogenicity, female Folbp2−/− and Folbp2 +/+ mice were mated with males of the same genotype, and were i.p. injected with 40 mg/kg sodium arsenate on GDs 7.5 and 8.5, the critical period of neural tube closure (Wlodarczyk et al. 2001). This As^V treatment increased resorption rates and NTDs in the surviving fetuses from both strains. After subtracting spontaneous NTDs in untreated controls of the same genotype, the increase of As^V-induced NTDs was bigger in Folbp2 −/− embryos than in Folbp2 +/+. A folate-deficient diet further increased the NTD frequency in Folbp2−/− embryos, but not in Folbp2+/+ embryos (Wlodarczyk et al. 2001). In addition, maternal genotype affected the sensitivity to in utero As^V exposure of Folbp2+/−embryos. Folbp2+/− embryos were from Folbp2−/− females mated with Folbp+/+ males or from Folbp2 +/+ females mated with Folbp−/− males. The resorption rates of embryos from the Folbp+/+ maternal group were higher than that from the Folbp−/− maternal groups after corresponding As^V 30 and 40 mg/kg single i.p. doses, but there were no differences in excencephaly rates between maternal genotype groups. Conversely, no difference was found in the 24-h urinary arsenical profiles of Folbp2−/− and Folbp+/+ female mice i.p. injected with 30 mg/kg sodium arsenate (Wlodarczyk et al. 2001). In a later study (Spiegelstein et al. 2005b), arsenicals were measured in the 24-h urine of wild-type and Folbp2−/− male mice after a single i.p. injection of 1 mg/kg sodium arsenate. In spite of the decreased plasma folate and SAM levels in Folbp2−/−mice compared to the wild type, there were no differences due to genotype in urinary arsenical profiles of male mice. Overall, impairment of Folbp2-mediated folate transportation due to inactive Folbp2 gene increased developmental defects induced by in utero exposure to As^V without affecting arsenic metabolism/excretion.

In contrast to the elevated sensitivity to As^V teratogenicity seen in Folbp−/− mice, no RFC or Folbp1 genotype-related differences in embryonic susceptibility to As^V exposure were observed (Spiegelstein et al. 2005a). Due to the embryonic lethality in Folbp−/− and RFC−/− mice, Folbp+/− and RFC+/−mice were used by these authors (Spiegelstein et al. 2005a). In a study of Folbp1 effects, Folbp1+/− female mice were mated with Folbp1+/+ males, after which pregnant mice received an i.p. injection of sodium arsenate at a dose of 30, 35, or 40 mg/kg on both GDs 7.5 and 8.5. All three doses increased resorption rates and NTD rates with no embryonic genotype (Folbp+/+ or Folbp1+/−) difference. Similarly, in a study of RFC effects, RFC+/− female mice were mated with RFC+/+ males, and RFC+/+ females were mated RFC+/− males; pregnant mice received an i.p. injection of 40 mg/kg sodium arsenate on GDs 7.5 and 8.5. Arsenate treatment increased NTD rates, but no differences were observed by RFC geneotypes or mating strategy. Regarding RFC effects on arsenic metabolism, RFC+/+ and RFC+/− mice were i.p. injected with 1 mg/kg sodium arsenate after receiving a normal or folate-deficient diet. With a normal diet, RFC+/− mice had the same plasma folate levels, plasma SAM levels, and urinary arsenical profiles as the wild type (RFC+/+). With a folate deficient diet, both RFC+/+ and RFC+/− mice had lower plasma folate and normal plasma SAM levels, and lower total arsenic in the urine. Thus, there were no RFC genotype– or Folbp genotype–related differences in embryonic susceptibility to in utero As^V exposure-induced NTDs, and arsenic metabolism appeared unaltered in RFC+/−mice.

Folate supplements failed to protect mice from arsenic teratogenicity (Gefrides, Bennett, and Finnell 2002). Arsenic acid (Na₂SeO₄) given at 40 mg/kg (i.p.) once each day on GD 7.5 and 8.5 increased NTDs and resorptions in LM/Bc, SWV, and CXL-Splotch mice. LM/Bc embryos were highly susceptible to arsenic developmental toxicity, while SWV embryos were relatively resistant. CXL-Splotch mice carry a mutation in the transcription factor Pax3, and heterozygous litters have high incidences of spontaneous NTDs. Arsenic increased NTDs in both CXL wild-type (+/+ ×+/+) and heterozygous (+/Sp × +/Sp) litters. Neither folic acid nor folate2 given to pregnant mice at 25 mg/kg (i.p.) once daily from GDs 6.5 to 10.5 provided protection against arsenic-induced resorptions or NTDs in CXL wild-type and heterozygous embryos, LM/Bc, or SWV. Folic acid and folate also did not decrease spontaneous NTDs in Splotch heterozygous embryos. Unexpectedly, folate given to arsenic treated pregnant mice caused maternal deaths in all three strains.

Studies of folate effects on arsenic developmental toxicity can be summarized as follows. Folbp2−/− mice, but not Folbp1+/− or RFC+/− mice, had increased sensitivity to in utero arsenate exposure–induced teratogenicity. There was no apparent RFC or Folbp2 genotype–related difference in arsenic metabolism based on urine arsenicals, and therefore the increased sensitivity to arsenate teratogenicity in Folbp2−/− mice was unlikely due to changes in arsenic metabolism. Furthermore, folate and folic acid supplements did not protect mice from arsenic-induced resorptions or malformations.

Influence of Methylation

Inhibition of arsenic methylation, by either a methylation inhibitor or a protein deficient diet, increased As^III and As^V developmental toxicities (Lammon and Hood 2004; Lammon et al. 2003). Periodate-oxidized adenosine (PAD) inhibits SAM-dependent methylation, and therefore inhibits arsenic methylation. In mice treated with PAD and then either As^III or As^V, inhibition of arsenic methylation was evidenced in the urinary arsenic profile (Lammon et al. 2003). Those mice had increased inorganic arsenic and decreased DMA^V in the urine, as compared to mice treated with As^III or As^V alone. This PAD pretreatment enhanced As^III and As^V developmental toxicities, including resorptions and fetal malformations (Lammon et al. 2003). CD-1 mice were given PAD (i.p.) followed by 7.5 mg/kg sodium arsenite or 17.9 mg/kg sodium arsenate (i.p.) on GD 8. Controls received As^III, As^V, PAD alone, or were untreated. Mothers were sacrificed on GD 17, and litters were examined. Compared to untreated mice, PAD alone–treated mice had increased prenatal mortality and percentages of litters with grossly malformed fetuses. Similarly, mice treated with As^III alone had increased prenatal mortality. As^V alone–treated mice had increased incidences of ablepharia (a partial or total absence of the eyelids). Compared to either arsenical alone, PAD pretreatment resulted in higher maternal morbidity and mortality, more complete resorptions and lower fetal weight. Furthermore, PAD pretreatment increased fetal malformations, possibly by PAD-induced maternal toxicity or PAD-enhanced arsenic toxicity. Increased incidences of exencephaly, ablepharia, misshapen vertebral centra, and abnormalities of the ribs, and sternebrae were observed. In summary, both As^III and As^V developmental toxicity can be enhanced by PAD, possibly by inhibition of arsenic methylation.

A low-protein diet also enhanced both As^III and As^V developmental toxicity (Lammon and Hood 2004). Female CD-1 mice were given diets with 20% (protein sufficient), 10% (moderate protein deficiency), or 5 % (severe deficiency) protein. A single i.p. injection of As^III (7.5 mg/kg NaAsO₂) or As^V (17.9 mg/kg Na₂HAsO₄) was given on GD 8. These treatments were expected to induce low levels of malformations; fetuses were examined on GD 17. Although 5% protein diet alone decreased maternal weight, protein deficiency alone did not cause developmental toxicity. With a protein-sufficient diet, increases in malformations were observed in either As^III or As^V treatment. The As^III or As^V alone increased incidences of exencephaly, ablepharia, rudimentary ribs, and sternebrae abnormalities. Arsenic plus protein deficiency decreased maternal weight gain, and increased the incidences of exencephaly, ablepharia, and skeletal defects. The observed skeletal defects included malformed vertebral centra, fused ribs, and abnormal sternebrae (bipartite, rudimentary, or unossified). These data showed that as the dietary protein content decreases, the incidence of fetal malformations of the offspring of arsenic-treated pregnant mice increases. Because the protein deficiency alone did not cause developmental toxicity in this study, the authors suggested that protein deficiency probably enhanced arsenic developmental toxicity by impairing arsenic methylation.

CONCLUSION

The forms of arsenic and administration routes greatly affect the severities and types of arsenic-induced reproductive and developmental toxicity. Inorganic arsenic caused reproductive and developmental toxicity, as demonstrated in in vivo studies using rats and mice. DMA^V caused developmental toxicity, whereas arsine did not. Gross structural malformations were only induced by i.p. or i.v. injections of inorganic arsenic. Other reproductive and developmental toxicity, however, was seen after maternal oral inorganic arsenic exposure. Inorganic arsenic exposure, including via drinking water, affects hormonal regulation and functions of both the male and female reproductive systems. Newborn behaviors and fetal brain development were also affected by As^III in the drinking water.

Selenium, but not zinc or folate, supplement was protective against arsenic-induced reproductive and developmental toxicity. Selenite supplementation decreased arsenic-induced female reproductive toxicity. A selenium-deficient diet increased arsenic-induced changes in selenoenzymes, important for brain development. Folate deficiency is associated with neural tube defects, and may interfere with arsenic methylation (Figure 3). Folbp2 −/− mice, but not Folbp1 +/− or RFC+/− mice, had increased sensitivity to in utero arsenate exposure–induced teratogenicity. However, folate and folic acid supplements did not protect mice from arsenic-induced resorptions or structural malformations. Inhibition of arsenic methylation also increased As^III and As^V developmental toxicity.