|Home | About | Journals | Submit | Contact Us | Français|
In the present study we characterize the toxic effects of in utero arsenic exposure on the developing lung. We hypothesize that in utero exposure to inorganic arsenic through maternal drinking water causes altered gene and protein expression in the developing lung, indicative of downstream molecular and functional changes. From conception to embryonic day eighteen, we exposed pregnant Sprague-Dawley rats to 500 ppb arsenic (as arsenite) via the drinking water. Subtracted cDNA libraries comparing control to arsenic exposed embryonic lungs were generated. In addition, a broad Western blot analysis was performed to identify altered protein expression. A total of 59 genes and 34 proteins were identified as being altered. Pathway mapping and analysis showed that cell motility was the process most affected. The most likely affected pathway was alteration in integrin signaling through the β-catenin pathway, altering c-myc. The present study shows that arsenic induces alterations in the developing lung. These data may be useful in the elucidation of molecular targets and biomarkers of arsenic exposure during lung development and may aid in understanding the etiology of arsenic induced adult respiratory disease and lung cancers.
Inorganic arsenic is a potent human carcinogen. Chronic environmental arsenic exposure through consumption of geologically contaminated drinking water has been correlated with increased incidence of and mortality due to internal cancers of the lung, skin, kidney, urinary bladder and liver [1–4]. In addition, reports from human studies in Chile, Bangladesh and the West Bengal region of India show that chronic exposure to arsenic via drinking water is correlated with increased incidence of chronic cough, chronic bronchitis, shortness of breath and obstructive or restrictive lung disease [5–8]. Taken together, these studies argue unequivocally that the lung is targeted by arsenic, producing both carcinogenic and non-carcinogenic endpoints.
Growth and development requires the temporal and spatial coordinated expression of genes and gene products. During this critical time, in utero and early postnatal exposure to toxicants has the potential to affect gene expression, altering organ structure and physiological function. However, only limited attention has been paid to the effects of environmentally relevant exposures to toxicants during these critical periods of development. The effects of human exposure to arsenic during these sensitive developmental times have recently been reported by Smith et al. . Exposures to high levels of arsenic, either in utero or during early childhood development led to an increased risk of lung cancers and chronic lung disease in adults. In a mouse model of transplacental carcinogenesis, arsenic exposure (42.5 and 85 ppm) during gestation days 8 through 18 lead to significant increases in tumor incidence and multiplicity in the lung and several other organs in adult offspring . While the doses used in the previous mouse studies are high compared to environmental exposure levels, they do show that tumor formation can occur in an animal model of in utero arsenic exposure. While arsenic has long been recognized as a human carcinogen, the non-cancerous health effects of arsenic ingestion in the drinking water can also lead to significant disease, including cardiovascular disease, arteriosclerosis, diabetes and chronic pulmonary disease. However, the molecular targets for these alterations after exposure to environmentally relevant levels of arsenic are not known.
To date, there are few studies detailing alterations in the developing fetus induced by maternal arsenic exposure. Acute, high dose ip injections of arsenic (30–45 mg/kg) during gestation have been associated with neural tube defects and corresponding aberrant gene expression of developmentally important transcription factors in the neural tube, including Hox 3.1 and Pax 3 . These doses of arsenic also triggered upregulation of bcl-2 and p53 gene expression in the neural tube, indicative of inhibition of cellular proliferation .
In the present study, we examined gene and protein expression changes in developing lungs of arsenic exposed fetal rats. We generated subtracted cDNA libraries from day 18 fetal lungs of rat pups exposed to 500 ppb arsenic (using arsenite) in utero beginning at conception. These doses are environmentally relevant and have been observed in areas of endemic exposure, including Chile  and the West Bengal region of India .
Our results show that arsenic caused aberrant expression of 93 genes and proteins. Analysis revealed that pathways involved in β-catenin signaling were affected. In addition, pathway analysis indicated that c-myc may play a central role in the in utero responses to arsenic. Taken together, these alterations demonstrate that arsenic causes alterations in gene and protein expression in the embryonic rat lung, following in utero exposure through maternal drinking water. Our results may be useful in the elucidation of molecular targets and biomarkers of inorganic arsenic exposure during lung development and may give insight into the etiology of arsenic-induced respiratory disease and lung cancers.
T-PER Tissue Protein Extraction Reagent, Halt Protease Inhibitor Cocktail Kit and Restore Western Blot Stripping Buffer were obtained from Pierce Biotechnology (Rockford, IL). High range and full range Rainbow molecular weight markers were obtained from Amersham Biosciences (Piscataway, NJ). Sodium arsenite was obtained from J.T. Baker (Phillipsburg, NJ). RNA Stat-60 RNA isolation reagent was purchased from Tel-Test Incorporated (Friendswood, TX). All other laboratory chemicals were purchased from Sigma Chemical Company (St. Louis, MO), and were secured in the highest purity available.
Male and female Sprague-Dawley rats were obtained from Harlan Sprague-Dawley (Indianapolis, IN). Animals were housed in a humidity controlled room, maintained at 22°C on a 12 h light-dark cycle and were given standard rat chow and water ad libitum. Rat chow used in our experiments was found to contain 75 ng/g of total arsenic. Based on daily intake of food and water this would increase the intake of arsenic in the exposure group by about 7% over the expected 500 ppb levels. Female rats were weighed nightly and estrus cycles were monitored using an Estrus Cycle Monitor EC40 (Fine Science Tools Inc., Foster City, CA). Upon reaching estrus female rats were mated overnight and pregnancy was confirmed the following morning by the presence of sperm in a vaginal lavage sample. Immediately after mating, pregnant animals were randomly placed in control or 500 ppb arsenic (sodium arsenite) drinking water treatment groups. Arsenic solutions were replaced once daily and arsenic concentration was validated by ICP mass spectrometry (data not shown). Mothers were euthanized with CO2 at embryonic day 18 (day 1 was the second morning after mating). Embryonic lungs were removed from day 18 fetal rats by microdissection and were immediately snap frozen in liquid nitrogen. Pup lung weights and body weights were recorded. Animal protocols were approved by the Institutional Animal Care and Use Committee of the University of Arizona.
Total RNA was extracted from lung tissue of a single control and treated day 18 fetal rat using RNA Stat-60. RNA was ethanol precipitated and mRNA was subsequently isolated using oligo-dT selection with the NucleoTrap mRNA Mini Purification Kit (Clontech, Palo Alto, CA). Synthesis and amplification of first strand cDNA from mRNA was carried out using the SMART PCR cDNA Synthesis Kit (Clontech). Subsequent preparation of subtracted cDNA libraries was performed using the PCR-Select cDNA Subtraction Kit (Clontech). The subtractive hybridizations performed were: 1) Treated (500 ppb arsenate or arsenite) cDNA subtracted with an excess of control cDNA and 2) Control cDNA subtracted with an excess of treated cDNA. Following cDNA subtractions, cDNA fragments were cloned into TOP-10 chemically competent E. coli (Invitrogen, Carlsbad, CA). All cDNA clones were sequenced in both directions by the Arizona Research Laboratories Genetic Analysis and Technology Core Sequencing Facility at the University of Arizona using an Applied Biosystems (Foster City, CA) ABI Prism 3700 DNA Analyzer. Sequences were processed to remove poor quality sequence data and vector sequences using the FAKtory DNA Sequence Fragment Assembly System . Sequences were subsequently blasted with BLAST at NCBI. Genes were classified by function using the BioRag database (bioresource for array genes) found at http://www.biorag.org.
Isolation of total RNA from animal tissues was carried out utilizing the RNA-stat 60 reagent and homogenization with a Tissue-Tearor (BioSpec Products, Inc., Bartlesville, OK). Two ethanol precipitations were performed to ensure RNA purity. Total RNA isolation from cell culture samples was performed as described above. DNase treatment of total RNA for real time PCR was also carried out as described above. Reverse transcription of 3.0 μg total RNA was performed using the Ambion EndoFree RT Kit, according to manufacturer protocol. Primer sets were designed against the first 300 base pairs of the 3' end of the gene of interest using the Primer 3 program from the Massachusetts Institute of Technology: http://www.genome.wi.mit.edu/cgi-bin/primer/primer3_www_slow.cgi. Primers are listed in Table 1 and were purchased from Integrated DNA Technologies (Coralville, IA). Real time PCR reactions were performed using the QuantiTect SYBR Green PCR Kit (Qiagen, Valencia, CA) and were carried out using a Rotor Gene RG-3000 instrument and Rotor Gene Version 4.6 software (Corbett Research, Mortlake, NSW, Australia). Each 10 μL reaction contained: 1 μM each forward and reverse primers, 6.25 mM supplemental MgCl2, 2.0 μL 1:25 diluted cDNA from RT reaction, 1× QuantiTect SYBR Green PCR Master Mix, and 0.25× SYBR Green I (purchased as 10,000× concentrate). Amplification conditions utilized 95°C for 15 min followed by 45 PCR cycles as follows: 95°C for 15 sec, 58°C for 15 sec, 72°C for 20 sec, followed by a final incubation for 45 sec at 72°C and a subsequent melt from 72–99°C at 1°C intervals, 5 sec per interval. Real time PCR analysis of fetal lung cDNA samples was performed in triplicate, using cDNA samples from individual pups, each pup being from a different litter (N=3 control, N=4 treated). All gene expression was normalized to expression of β-actin. Prior to export of real time PCR data, slopes were corrected using the slope correct function of the Rotor Gene software.
Total protein was isolated from snap frozen fetal lung tissue using 5 μL/mg tissue of Tissue Protein Extraction Reagent (T-PER) with 10 μL/mL Halt Protease Inhibitor Cocktail and 1 mM sodium orthovanadate, according to manufacturer protocols. Protein quantitation was performed using the Bicinchonic Acid Kit (Sigma, St. Louis, MO). Protein samples were analyzed by Becton Dickinson using their Powerblot analysis, which stains using 995 well characterized antibodies. (BD Biosciences, San Diego, CA). Of these, 716 of the antibodies are know to react with rat proteins, 16 are known not to react with rat and 263 are untested in rat. A total of 853 individual or modified proteins were detected. Protein samples were run on multiple lanes, blotted and immunostained. Levels of proteins were determined using denistometric analysis. Proteins from three arsenic treated and three control pups were run independently. Significance was assigned only if all three treated samples responded similarly when tested against all three controls.
Analysis of the altered genes and proteins was performed using MetaCore from GeneGo (San Diego, CA). This is a curated analysis that allows for determination of the most likely cellular processes, pathways, transcriptional regulators and receptors that are altered. Statistical analysis of lung weights and lung to body weight ratios were compared using Student's t-test, conducted with Graph Pad Prism 3.0 (Graph Pad Software, San Diego, CA). Statistical analysis of real time PCR data were performed using two-tailed Student's t-tests, requiring p<0.05 for statistical significance.
Treatment with 500 ppb arsenic (as sodium arsenite) resulted in decreased body weight and lung weight in day 18 fetal rats (Table 2). In particular, arsenic caused significant (p<0.0005) reduction of fetal body weight by 6.6 % and lung weights by 13.3 %. Arsenic exposure also caused significant (p<0.05) reduction of lung to body weight ratios by 7.2 %. No changes in maternal weight gain or food and water consumption were observed between control and treated pregnant animals during the course of this study (data not shown). Additionally, there were no alterations in litter size or in the number of fetal resorptions between treatment groups (data not shown).
The cDNA libraries isolated from subtractive hybridization of control versus arsenic treated day 18 fetal rat lungs contained 326 clones. From these clones, 281 sequences were obtained with approximately 160 representing unique genes. Such redundancy in subtracted cDNA libraries is consistent with that seen in the literature . After applying an e value match score maximum of e−50, 129 unique genes were identified with 59 clones representing genes of known function. Genes from arsenic subtracted libraries are listed in Table 3. An expanded Table 3, which includes functional descriptions of the genes, can be found in the Supplemental Material. Major functional areas described by this list of genes include matrix, cytoskeletal and adhesion genes, transcriptional regulators, ribosomal genes and protein modification and turnover genes. Subtractive hybridization selects for low abundance genes that are differentially expressed in two samples. Therefore, the affect of arsenic is listed as either increasing or decreasing expression, based on abundance in controls and treated animals.
Quantitative real time PCR was used to validate differential expression of select genes identified in the subtracted libraries. The following genes were analyzed: collagen type III α1 chain and sprouty-2. Primer sequences are given in Table 1. Data for real-time were consistent with the differential expression data. Sprouty-2 and collagen 3a1 were both down regulated by two-fold with 500 ppb arsenic exposure in vivo.
Proteins from three treated and three control pups were analyzed using BD Powerblot (Table 4). In order for the proteins to be considered as altered, changes in expression had to occur between each treated sample when compared against each individual control. A total of 34 proteins were determined to be altered by the in utero exposure to arsenic (Table 4). Alterations in serine/threonine phosphorylation were also detected in three unidentified proteins. A majority of the proteins that were misregulated are associated with cell motility, adhesion and cytoskeleton. (See expanded Table 4 in the Supplemental Material). The Powerblot analysis contained antibodies against only four of the arsenic altered gene products. (ataxin 2, 14-3-3ε, caspase 2, HSP70). While the protein levels tended to follow the down regulation of the gene expression, none of these four protein levels reached the strict significance criteria used for the Powerblot analysis.
Both genes and proteins that were identified as being differentially expressed by arsenic were entered into a curated data analysis program, MetaCore from GeneGo. This program utilizes published data to determine the most probable pathways, diseases, cellular processes and transcriptional regulation that are defined by the altered expression. For the genes and proteins that were identified, the statistically most likely disease is lung cancer and the most probable cellular function that has been altered is cell motility. Using the receptor pathway analysis, MetaCore identified matrix and extracellular signaling through the β-catenin pathway as a highly likely site of action described by the altered genes and proteins (Figure 1). Finally, the altered genes and proteins interacted mostly with transcription factors c-myc (22 genes/proteins), HNF-4α (16) and p53 (14).
Alterations in the developing rat lung induced by environmentally relevant doses of arsenic (500 ppb as arsenite) were analyzed to assess a potential role for this metalloid in alteration of lung development. The changes observed were chronicled by subtractive hybridization and by Powerblot analysis of proteins.
Environmental arsenic exposure in drinking water has been linked to an increase in respiratory disease. Recently, Smith et al.  have reported in a human population from Chile that in utero and early postnatal exposures to high environmental levels (800 – 900 ppb) greatly increase the incidence on lung disease associated mortalities later in life. Standard mortality ratios (SMR) from lung cancers, bronchiectasis and other chronic lung diseases were increased five to ten fold with early postnatal exposure alone. Combined in utero and postnatal exposures increased the SMR even higher. Our data reported here using 500 ppb arsenic exposures are therefore relevant for providing potential mechanisms that can lead to human diseases.
Analysis of altered genes and proteins in our rat model has indicated that the beta-catenin pathway and c-myc may be important targets of in utero arsenic exposure. The wnt signaling pathway and c-myc have been shown to be important regulators of lung development as well as playing a role in lung cancers and chronic lung diseases. Beta-catenin is required for appropriate proximal/distal lung fate during lung morphogenesis  and c-myc is important for appropriate cell expansion . Both are also over expressed in lung cancer [16–17]. Additionally, beta-catenin and c-myc gene expression were found to be altered in the livers of mice following in utero exposure to 85 ppm arsenite . While the pathway mining suggests beta-catenin pathways and c-myc as important molecular targets, their involvement must be verified in additional experiments.
Additional genes of particular interest to our research group were the arsenic induced modulation of extracellular matrix genes, cell motility genes and those involved with regulation of fetal growth. A number of these genes and proteins were modulated by arsenic in the present study. Collagen type III is an interstitial collagen found within the alveolar walls and pulmonary vasculature . Expression of the gene encoding collagen III peaks at embryonic day 12 in the mouse, with its expression becoming gradually reduced until birth . Decreased immunostaining of this collagen is found in the alveoli in a mouse model of metalloproteinase-induced emphysema . This particular emphysema model is independent of changes in elastin and illustrates that reduced expression of collagen III can contribute to reduced lung function through induction of emphysematous changes. We have also previously reported misregulation of this and other matrix genes following arsenic exposure in adult mice .
Aberrant airway remodeling is a hallmark of many diseases including emphysema, asthma, idiopathic pulmonary fibrosis, tuberculosis and bronchiectasis [22–26]. It consists of an array of persistent tissue structural changes that occur through a process of injury and dysregulated repair that may lead to airway chronic inflammation and altered extracellular matrix (ECM) deposition in the airway wall, leading to airflow obstruction [23, 27–30]. In addition to matrix, cell migration is a critical step in tissue remodeling  and abnormalities in cell migration have the potential to lead to respiratory disease.
Changes that we have seen are consistent with arsenic induced chronic obstructive pulmonary disease (COPD) incidence in exposed humans. Environmental arsenic exposure in drinking water has been linked to an increase in respiratory disease [32–35]. Reduced collagen expression and altered cell motility and evidence of its association with emphysematous changes in the lung  may also be consistent with arsenic induced COPD. Therefore, the effects of arsenic on the developing lungs in the present study and the physiological consequences thereafter should ultimately be assessed by physiological and morphological analyses conducted over a developmental time course.
Sprouty-2 is another gene that is involved in fetal growth that was modulated by arsenic exposure in the present. Sprouty-2 is necessary for lung development, lung maturation and regulation of cellular proliferation in the developing lung and is crucial for embryonic survival. A loss of function mutation of sprouty-2 in Drosophila caused excessive branching and embryonic death . Cultured mouse embryos injected with sprouty-2 antisense oligonucleotides showed a significant increase in tracheal branching and increased expression of the proliferating cell nuclear antigen, along with increased expression of the lung maturation markers, surfactant protein genes SP-A, SP-B and SPC . The main target of sprouty-2 seems to be epithelial cells, where its expression reduces cellular proliferation . Down regulation of sprouty-2 by arsenic in the present study is thus likely to be physiologically significant and merits further study.
In the present study, a general decrease in fetal growth and fetal lung growth were observed in day 18 fetal rats following 500 ppb arsenic (as arsenite) exposure during gestation (Table 2). These findings are consistent with previous studies showing reduced birth weight of pups exposed to arsenic in utero [10, 39] and are also novel, as changes in fetal lung growth have not been documented following gestational arsenic exposure. No changes in maternal weight gain during pregnancy were observed between control and arsenic treated mothers, nor were there any changes in number of pups per litter or number of fetal resorptions (data not shown).
Arsenic is at best a weak mutagen. Therefore several epigentic mechanisms have been proposed as sites of action for arsenic. Alteration of DNA methylation is one of the main epigenetic modifications in humans that controls gene expression . The methylation status of genes is long lived and heritable. Alteration in the methylation of genes will result in long term effects on gene expression. Therefore, in utero alteration of methylation status of genes important for the control or the development of chronic diseases could lead to adult diseases later in life. Exposure to arsenic has been shown to alter DNA methylation. Both global hypomethylation  and specific gene hypo-(cmyc, c-Ha-ras) and hypermethylation (p16, p53) have been reported following arsenic exposure [42–44]. It is believed that arsenic exerts its effect of methylation status due to its metabolism. Inorganic arsenic (III) is methylated to mono- and dimethyl forms using methyltransferases and requires S-adenosyl-methionine as a cofactor. Altering the activity in this pathway could affect DNA methylation status, resulting in altered gene expression. Additionally, arsenic has also been shown to affect DNA repair processes . Our analysis has identified alterations of MSH6, a protein associated with recognition of DNA mismatches and mismatch repair. The levels of this protein were increased in arsenic exposed animals. This may be indicative of increased repair. Alternatively, DNA repair proteins have also been associated with regulation of apoptosis. Therefore, this increase may be a response to additional cellular stress caused by the arsenic exposures.
Arsenic methylation provides a site for interaction between arsenic exposure and dietary methyl donors. Many micronutrients are essential for DNA synthesis and DNA repair. Among these, folate is one of the more extensively studied. Not only is folate important in proper nucleotide synthesis, it also plays an important role in DNA methylation . Thus dietary folate could provide an intervention strategy for reducing the adverse in utero exposure effects of arsenic exposure.
We have used a combination of alterations in mRNA and protein expression to determine altered pathways and sites of action of arsenic. We feel that using information from these two sources strengthens our results. Alterations in levels of expression of genes or proteins can potentially occur independent of each other. For example gene expression could be unchanged, but alterations in protein degradation could affect the levels of the protein. Therefore using both mRNA and protein levels provides a wider range of potential sites of action and more confidence in the pathway analysis.
Finally, it should be pointed out that the levels of exposure in these experiments, while found in the environment, are quite high. It will be important to retest these results using lower, doses of arsenic consumption. In addition, we did not directly measure the levels of arsenic found in the fetal lungs. However, based on work reported by Devesa et al.  we would expect to find that the levels of arsenic in the fetal lung would be about 1% of the administered arsenic, or about 5 ppb, mostly in the form of dimethylarsenic.
The findings presented in this study implicate arsenic as a developmental toxicant in the rat lung. Reduction in fetal weight, lung weight and lung to body weight ratios and altered expression of cellular differentiation markers demonstrate that arsenic is causing changes in fetal growth and development, following in utero exposure from conception to embryonic day 18. These alterations may ultimately result in fetal lungs that are premature, while also being highly differentiated and branched. It is also possible that the observed reduction in lung growth was a compensatory response to increased differentiation. Targets of in utero arsenic exposure in the developing lung appear to be the developing extracellular matrix and the processes of cellular differentiation and branching morphogenesis. The arsenic-induced toxic and growth alterations presented in this study demonstrate effects of arsenic in the developing lung, findings which merit further mechanistic and physiological studies.
This work was supported in part by the Superfund Basic Research Program NIEHS Grant Number ES-04940 and the Southwest Environmental Health Sciences Center P30-ES-06694.