Conceived and designed the experiments: CDK-H FZ RIE JWH. Performed the experiments: CDK-H FZ BPJ. Analyzed the data: CDK-H FZ BPJ. Contributed reagents/materials/analysis tools: BPJ. Wrote the paper: CDK-H.
The authors have declared that no competing interests exist.
Arsenic (As) exposure is a significant worldwide environmental health concern. Chronic exposure via contaminated drinking water has been associated with an increased incidence of a number of diseases, including reproductive and developmental effects. The goal of this study was to identify adverse outcomes in a mouse model of early life exposure to low-dose drinking water As (10 ppb, current U.S. EPA Maximum Contaminant Level).
C57B6/J pups were exposed to 10 ppb As, via the dam in her drinking water, either in utero and/or during the postnatal period. Birth outcomes, the growth of the F1 offspring, and health of the dams were assessed by a variety of measurements. Birth outcomes including litter weight, number of pups, and gestational length were unaffected. However, exposure during the in utero and postnatal period resulted in significant growth deficits in the offspring after birth, which was principally a result of decreased nutrients in the dam's breast milk. Cross-fostering of the pups reversed the growth deficit. Arsenic exposed dams displayed altered liver and breast milk triglyceride levels and serum profiles during pregnancy and lactation. The growth deficits in the F1 offspring resolved following separation from the dam and cessation of exposure in male mice, but did not resolve in female mice up to six weeks of age.
Exposure to As at the current U.S. drinking water standard during critical windows of development induces a number of adverse health outcomes for both the dam and offspring. Such effects may contribute to the increased disease risks observed in human populations.
Chronic exposure to arsenic (As), by contamination of drinking water from natural geological sources, is a significant worldwide environmental health concern
Exposure events during critical windows of fetal and postnatal (PN) development pose a serious risk for adverse health outcomes later in life
Epidemiology studies have shown that exposure to As during gestation is associated with a number of adverse effects on the fetus, including low birth weight, survival, and spontaneous abortion
Based on cancer evidence, the U.S. EPA's current Maximum Contaminant Level (MCL) for As in public drinking water supplies is 10 ppb (0.13 µM)
The experimental model of exposure is detailed in
Following the detection of cervical plugs, mated females were exposed to control water or 10 ppb As in drinking water through the gestational period. At the birth of their pups, dams in each exposure group were further divided into sub-groups receiving control water or 10 ppb As in drinking water through 30 days of age (n = 14–17 dams per exposure). Weaning from the dam took place at day 21 PN (or later if a pup did not reach the 7 gram weight cut-off). This resulted in 4 overall exposure groups: 1. Control (no As drinking water exposure; 2. Postnatal (PN, offspring receiving 10 ppb As from PN days 1–30); 3. In utero (IU, offspring receiving 10 ppb As from gestational day 1 through birth); 4. In utero & postnatal (IU&PN, offspring receiving 10 ppb As from gestational day 1 through day 30 PN) At day 30 PN, all offspring were placed on control drinking water and growth was assessed until 6 weeks PN.
Following birth, pups were monitored daily to assess survival and development. Survival and developmental milestones (eye opening, pinna unfolding, appearance of fur) was not differentially affected by the As exposure paradigm. As early as day 10 post natal, offspring exposed to As displayed significant decreases in growth (evidenced by total body weight), regardless of the timing of As exposure (
(A). The weight (grams) of offspring was monitored over the course of development and is shown across all four exposure groups at birth, day 10, 21, and day 28 PN. Male and female mice are included. Birth, n = 21–25; Day 10, n = 17–21; Day 21, n = 65–71; Day 28, n = 44–52. One Way ANOVA, compared to control for each respective time point (B.) Mouse weights separated by gender at day 42 PN across all four exposure groups; n = 49–54. Male mice in all exposure groups are represented with open shapes and female mice with closed shapes. One Way ANOVA for female mice, compared to unexposed female mice. Asterisks indicate statistical significance, * p<0.05, ** p<0.01*** p<0.001, ****p<0.0001. Error bars represent mean ± SEM.
To assess the direct exposure of the F1 offspring to As, total As levels in the placenta, dam's breast milk, pup stomach content (day 10 and 21 PN), and urine were measured by ICP-MS across all exposure groups at various time points pre-, during and post-pregnancy. Arsenic levels in the placentas, dam's milk and pup stomachs were near the limits of detection in all exposure groups. There was no observable increase in the As concentrations in the placenta or milk, but there was a significant increase in the stomach As contents of the pups that received As exposure IU and IU&PN compared to control offspring. This increase in stomach As levels was observed at day 10 PN, during which time the offspring would be primarily consuming breast milk, as opposed to drinking from the water bottle source of As. By day 21 PN, no differences in As levels were detected in the stomach contents of the pups. As expected, urinary excretion of As significantly increased in the As exposure groups. Interestingly, while all mice in the exposure groups received the same level of As (10 ppb), excretion values decreased in the pregnant and lactating As-exposed mice, when compared to As-exposed virgin mice. Mice that were exposed to As only IU no longer excreted significant levels of As during the postpartum period. Urinary As excretion was primarily in the form of dimethylarsenate (DMA). Values are represented in
Control | As (2 week) | |
|
||
|
11.91 (5.0) | 56.01 (13.1) |
Control | IU | |
|
||
|
1.20 (0.4) | 1.00 (0.5) |
|
6.89 (2.5) | 26.75 (6.9) |
Control | IU | PN | IU & PN | |
|
||||
|
0.64 (0.3) | 2.1 (0.5) |
1.17 (0.4) | 2.69 (1.0) |
|
1.56 (0.6) | 0.89 (0.4) | 0.78 (0.5) | 1.73 (1.0) |
|
4.41 (2.5) | 5.94 (3.3) | 19.68 (3.1) |
20.00 (1.3) |
|
||||
|
4.8 (0.4) | 5.0 (0.3) | 4.6 (0.9) | 3.7 (0.8) |
Total As levels were measured by ICP-MS, as described in Methods (n = 3–6). Values represent mean ± SEM. Asterisks indicate statistical significance from respective control group in matching row.
p<0.05,
p<0.01,
p<0.001. Number sign indicates statistical significance from virgin female mouse exposed to 10 ppb As for 2 weeks. One Way ANOVA.
Given that the growth deficit manifested during the postnatal period (as early as day 10) and there was not a significant increase in the exposure of the pups to As via the breast milk, we hypothesized that other alterations in the dam's milk were at least partially responsible. Obvious differences in the rearing of the pups by the dams in different exposure groups were not observed. By day 1 PN, all pups displayed prominent milk spots and by day 10 PN, no significant difference was observed in the weight of the stomach contents of the pups, which confirmed that the offspring in all exposure groups were being fed and general milk production was intact. The milk of the dams in all exposure groups was assessed for total protein and triglyceride (TG) concentrations. Protein concentrations were not significantly different (
Breast milk was collected between day 10–12 PN. Breast milk was analyzed for (A.) total protein concentrations and (B.) total TG concentrations. Asterisks indicate statistical significance, ** p<0.01, *** p<0.001, One Way ANOVA compared to control. Error bars equal mean ± SEM (n = 8–9 mice per exposure).
We were interested to assess how exposure to As might alter lipid metabolism in the dams during the course of pregnancy and lactation. Unexpectedly, we found a significant increase in the incidence of liver steatosis in the As-exposed dams at day 15.5 gestation. None of the control dams displayed gross liver steatosis at day 15.5 gestation, while 55.5% (5/9) of the As exposure dams displayed gross liver steatosis (* p<0.05, compared to control, Fisher's exact probability test). Liver steatosis was confirmed at the histological level (
Dams were sacrificed at gestation day 15.5. (A, B.) Detection of liver steatosis at the gross level was observed in a significant percentage of the As exposed mice (C–F.) Histological hematoxylin and eosin staining in (C.,E.) control and (D., F.) As-exposed dams. Scale bars indicate magnification.
Control | IU | |
|
||
|
119 (8.5) | 74 (0.5) |
|
4.1 (0.1) | 6.85 (0.7) |
Control | IU | PN | IU & PN | |
|
||||
|
110 (10.3) | 70 (4.0) |
71 (9.9) |
65 (5.7) |
|
4.2 (0.6) | 6.1 (1.4) | 5.3 (0.7) | 5.44 (0.4) |
Asterisks indicate statistical significance,
p<0.05 and
p<0.01, compared to respective control. Values represent mean ± SEM. Two tailed student's t-test for gestational exposure; One Way ANOVA for post natal exposures.
Profiling of serum lipoproteins using Fast Protein Liquid Chromatography revealed no clear differences between the control and As-exposed dams for cholesterol levels in the lipoprotein fractions, indicating that 10 ppb As does not substantially affect the number of lipoprotein particles for any of the fractions (
Dams were fasted for 6 hours on PN day 21 and pooled serum of 4 As (IU&PN) treated dams (open triangles) and 3 controls (black circles) was used for lipoprotein profiling by FPLC. Fractions were assayed for TG (top panel) and total cholesterol levels (bottom panel).
To assess the level to which decreased milk nutrients contributed to the growth deficit, we designed a series of cross-fostering experiments. In these experiments, offspring from opposing gestational exposure groups (control or gestational As) were exchanged and fostered. We observed that we could reverse the growth phenotype by simply exchanging the litters and dams from opposing exposure groups (
The average litter weight (grams) of offspring at weaning (day 21 PN) was assessed following fostering. Immediately following birth, all dams were placed on control drinking water through the weaning period. Groups were as follows: (A.) Control offspring remaining with biological mom through weaning/not fostered (n = 11 litters); (B.) 10 ppb IU As exposed mice remaining with biological mom/not fostered (n = 11 litters); (C.) Control offspring fostered by 10 ppb IU As mom (n = 5 litters); (D.) 10 ppb IU As offspring fostered by control mom (n = 5 litters); (E.) Control offspring fostered by a non-biological control mom (n = 3 litters); (F.) 10 ppb IU As offspring fostered by a non-biological 10 ppb IU As mom (n = 3 litters). Asterisks indicate statistical significance, * p<0.05, **** p<0.0001, One Way ANOVA compared to control mice remaining with biological mom (column A). Error bars represent mean ± SEM.
In the conduct of an experiment designed to examine immune effects later in life from in utero exposure to As at the current U.S. EPA drinking water standard, we were surprised to observe significant short-term effects of these exposures on the F1 offspring and the dams. Specific effects included decreased growth of the F1 offspring and altered TG levels and profiles in the dams. Decreases in the nutrient content of the dam's breast milk, specifically TGs, appear to play a role in the growth deficits observed in the F1 offspring. The growth deficit was reversed by cross-fostering of the pups.
Arsenic concentrations were measured in the dam's milk, pup stomachs, placentas, and dam's urine, which confirmed that As was not transferred via the breast milk. In two exposure groups (IU and PN), we observed a trend towards decreased As levels in the milk, which was also recently reported in a similar mouse study with much higher doses of exposure
Interestingly, we also observed significant decreases in As excretion by the dams in all As exposure groups during the gestational and postnatal period, compared to virgin mice ingesting the same level of As. There was a trend towards increased water consumption and urine output in the As-exposed mice during pregnancy (and across all groups during the lactational period), but given the magnitude of these trends, the changes in water consumption and urine output are unlikely to account entirely for the significant decrease in urinary As output. Recent epidemiological data has shown that pregnancy can significantly alter the metabolism of As
The impeded growth of As-exposed offspring observed in this study was unexpected, given the low dose (10 ppb) of As used. Similar growth deficits have been observed in a high dose [10, 50 and 100 mg/L] IU exposure model of rats
Based on our results, it is clear that As effects on the dam play a major role in the growth deficit of the offspring. The results of the fostering experiments suggest that the milk is the major contributor to the effects we have observed. However, we cannot definitively rule out that small behavioral changes in the dams rearing of the pups did not play a role in the growth deficit. Breastfeeding of infants in human populations with As exposure has been shown to be protective against increased As exposure to the infant
We have previously reported that a combined exposure of adult mice to As followed by a sublethal infection with influenza significantly increased the severity of respiratory infection
Overall, this study indicates that low-level As exposure during the IU or PN period results in a number of immediate adverse effects, including altered TG levels and profiles and growth deficits in the F1 offspring. Many of these outcomes were direct effects on the dam, which manifested as adverse outcomes in the F1 offspring. The results suggest that exposure of vulnerable populations to As, perhaps at levels as low as the current MCL of 10 ppb, may induce a significant increase in adverse health outcomes, which has been previously unrecognized. It is well documented that exposure to chemicals and environmental toxicants during the developmental period can have both immediate and long-term health problems. The goal of our future research will be to address the impact of early-life As exposure on the development and relative adverse health risks of these F1 individuals as adults, and on their F2 offspring.
All animal studies were conducted in accordance with Association for Assessment and Accreditation of Laboratory Animal Care (AALAC)-approved guidelines using a protocol approved by the Dartmouth Institutional Animal Care and Use Committee (IACUC). Approval protocol number for Dartmouth IACUC was 10-10-02. All animals were treated humanely and with regard for alleviation of suffering. C57BL/6J mice (Jackson Labs, Bar Harbor, ME) were housed in ventilated cages with AIN-76A chow (Harlan Teklad, Madison, WI, ad lib) and corncob and carefresh bedding (Scott's Distributing, Hudson, NH). Background As concentrations in the diet were less than 20 ppb, which is a level too low to speciate. We have demonstrated in previous studies that the bioavailable fraction of inorganic arsenic in this diet is low and that we can clearly distinguish an experimental signal as compared to control by exposing animals to drinking water As at 10 ppb
Mated mice were acclimated on the AIN-76A diet for 2 weeks prior to mating. In this model, following the detection of cervical plugs, naïve pregnant mice were exposed to control or 10 ppb As in drinking water through the gestational and weaning period. Males were not exposed to As prior to mating. At birth, dams in control and As exposure groups were further split into groups receiving control or 10 ppb As in drinking water, which resulted in 4 overall exposure groups: 1. Control (no As drinking water exposure); 2. Postnatal (PN, offspring receiving 10 ppb As from PN days 1–30); 3. In utero (IU, offspring receiving 10 ppb As from gestational day 1 through birth); 4. In utero & postnatal (IU&PN, offspring receiving 10 ppb As from gestational day 1 through day 30 PN). (
Litters born to a control dam or a gestationally As-exposed dam within 12 hours of each other were eligible for cross-fostering. Fostering of litters always occurred within 12 hours of birth. All biological offspring were removed, the litters were culled to seven pups, and they were transferred, as a group, to the cage of the foster dam. The behavior of the dam was monitored for one hour post-fostering to ensure acceptance of the litter. All dams were maintained on deionized distilled water ddH2O (no As) following birth. Weight (grams) of fostered offspring was assessed at day 21 PN.
Sodium arsenite (Sigma Aldrich, St. Louis, MO) was prepared from stock solution in ddH2O to yield a 10 ppb (µg/L) concentration of drinking water. The arsenic concentration in the final solution was confirmed by induction-coupled plasma mass spectrometry (ICP-MS) metal analysis at Dartmouth's Trace Elements Analysis Core Facility. Drinking water was changed twice weekly.
At birth, the gestational length, number of pups, average litter weight (weight of individual pups/number of pups in litter), number of dead pups, and number of pups with malformations were recorded. Weight (grams) was recorded for offspring at birth, day 10, day 21, day 28 and day 42 PN.
Liver steatosis in the dams was observed at the gross observational level. For histological confirmation, livers were removed and fixed in formaldehyde, paraffin embedded, sliced and stained with hematoxylin and eosin.
Dams (day 10–12 postnatal) were separated from pups for 6 hours to allow for milk accumulation. Dams were lightly anesthetized (i.p.) with 9∶1 ketamine∶xylazine mix at 0.1 ml/30 g body weight and injected (i.p.) with 2 IU (100 uL) of oxytocin (Sigma-Aldrich, St. Louis, MO). Dams were milked by gentle manual stimulation of the teat and collection with a pipette. Milk was stored at −20 degrees C.
Milk samples were diluted in PBS and assayed for protein (1∶400 dilution). Protein concentrations were determined by BCA Protein Assay (Pierce, Rockford, IL), according to manufacturer's instructions.
Dams were fasted for 6 hours. Blood was collected from the vena cava and livers were collected and snap frozen. Liver tissue (∼150 mg) was homogenized in PBS using the Bullet Blender Homogenizer (Next Advance, Cambridge, MA).
Triglyceride concentrations were determined in the dam's breast milk (1∶160 dilution), liver tissue (1∶20 dilution) and serum (1∶2 dilution) samples from pregnant and lactating dams across all exposure groups, using Wako L-type triglyceride (Wako Diagnostics, Richmond, VA), according to manufacturer's instructions.
Lipoproteins were separated using fast protein liquid chromatography (FPLC). 0.2 mL of pooled mouse plasma was injected onto a Superose 6B 10/30 column (GE Healthcare Life Sciences, Piscataway, NJ) and eluted at a constant flow of 0.2 mL/minute with phosphate buffered saline (pH 7.4). The effluent was collected in 0.2 mL fractions and triglyceride and cholesterol levels were determined for each fraction using L-type and Cholesterol E kits from Wako Diagnostics (Richmond, VA).
Total concentrations of trace metals in the breast milk (n = 4–5), pup stomachs (n = 6), maternal urine (n = 3) and placentas (n = 3) were measured by ICP-MS. Placenta, stomach and milk samples were digested with 0.3–0.5 ml optima HNO3 at 70°C followed by addition of 0.05–0.1 ml of H2O2. The digested samples were then diluted to 3 ml (stomach and milk) or 5 ml (placenta) final volume. These samples were then analyzed by ICP-MS (7700x, Agilent, Santa Clara, CA) using He as a collision gas for As determination. Quality control included sample analysis duplicates and spikes. Urine was diluted 10-fold with 1% optima HNO3 and analyzed for As by ICP-MS as above.
Statistical analysis was performed with Graphpad Prism 5.0d for Macintosh (GraphPad Software Inc., La Jolla, CA) using two tailed student t-test, ANOVA (with Bonferroni post test) or Fisher's exact probability test, requiring p<0.05 for statistical significance.
(TIFF)
(TIFF)
(DOCX)
The authors thank Dr. Radu Stan for use of metabolic cages and Roxanna Barnaby for animal breeding and husbandry.