Arsenic-Induced Decreases in the Vascular Matrix

Abstract

Chronic ingestion of arsenic is associated with increased incidence of respiratory and cardiovascular diseases. To investigate the role of arsenic in early events in vascular pathology, C57BL/6 mice ingested drinking water with or without 50 ppb sodium arsenite (AsIII) for four, five, or eight weeks. At five and eight weeks, RNA from the lungs of control and AsIII-exposed animals was processed for microarray. Sixty-five genes were significantly and differentially expressed. Differential expression of extracellular matrix (ECM) gene transcripts was particularly compelling, as 91% of genes in this category, including elastin and collagen, were significantly decreased. In additional experiments, real-time RT-PCR showed an AsIII-induced decrease in many of these ECM gene transcripts in the heart and NIH3T3 fibroblast cells. Histological stains for collagen and elastin show a distinct disruption in the ECM surrounding small arteries in the heart and lung of AsIII-exposed mice. Immunohistochemical detection of α-smooth muscle actin in blood vessel walls was decreased in the AsIII-exposed animals. These data reveal a functional link between AsIII exposure and disruption in the vascular ECM. These AsIII-induced early pathological events may predispose humans to respiratory and cardiovascular diseases linked to chronic low-dose AsIII exposure.

Introduction

It has recently been appreciated that chronic ingestion of arsenic can lead to cardiovascular and respiratory toxicity with a resulting plethora of diseases, including chronic bronchitis, obstructive pulmonary disease, interstitial lung disease, and bronchiectasis (Mazumder 2007); and peripheral vascular disease, atherosclerosis, coronary artery disease, myocardial infarction, ischemic heart disease, and diabetes (Chang et al. 2004; Tseng et al. 2003). Specifically, in diabetes, disregulation of matrix metabolism has been shown to contribute to arsenic-induced vasculopathies (Cooper et al. 2001). Diabetes is commonly associated with both microvascular and macrovascular complications. Expression of growth factors is stimulated in the vasculature in response to this disease, which leads to extracellular matrix accumulation in the diabetic kidney (Cooper et al. 2001). In contrast to this finding, we found a decrease in extracellular matrix transcripts in the lung, heart, and NIH3T3 cells in response to sodium arsenite (AsIII). The mechanism(s) associated with these various arsenic-mediated cardiovascular and respiratory pathologies remains elusive. The majority of biological models of arsenic exposure implicate aberrations in cell proliferation, oxidative damage, and DNA-repair fidelity in arsenic-mediated carcinogenicity and toxicity (Li et al. 2002; Schwerdtle et al. 2003; Yuan et al. 2003). The purpose of this investigation is to determine whether exposure to low concentrations of AsIII can cause early changes in the cardiovasculature, which can lead to the development of cardiovascular diseases.

The large majority of in vivo arsenic studies involve acute time courses and/or high exposure levels. Chronic low-dose arsenic studies in animals are almost nonexistent. The few animal studies that do investigate the chronic, low-dose protocol demonstrate arsenic-associated angiogenic effects (Soucy et al. 2003); increased metastases of tumors to the lung (Kamat et al. 2005); increased oxidative stress (Hughes and Thompson 1996); differential gene expression of cytokines and steroid-, apoptosis-, and cell-cycle–related genes (Chen et al. 2004); and differential gene expression of genes associated with cell growth (Simeonova et al. 2000).

Epidemiology studies from Taiwan link human arsenic exposure to a wide variety of cardiovascular and lung-associated diseases (Chang et al. 2004; Guo et al. 1997). In Taiwan, one of the most studied geographical areas with respect to arsenic in well water, this metal induces blackfoot disease (BFD), a severe form of peripheral vascular disease, in which damage to the blood vessels in the lower limbs results in progressive gangrene. Human studies from BFD-hyperendemic villages are characterized by long-term exposure to well water with high levels of arsenic. In addition to BFD, squamous and small cell lung carcinomas (Guo et al. 2004); ischemic heart disease (Chang et al. 2004; Tseng et al. 2003); carotid atherosclerosis (Wang et al. 2002); cerebrovascular disease and diabetes mellitus (Tsai et al. 1999); and inflammation and oxidative stress (Wang et al. 2003) were significantly associated with arsenic exposure in a dose-response relationship in these villages.

In addition to studies in Taiwan, investigations from around the world demonstrate that arsenic exposure is associated with cardiovascular and respiratory pathologies. Acute and chronic arsenic exposure via drinking water has been reported in Chile, Bangladesh, India, Argentina, Mexico, and Thailand (Ferreccio and Sancha 2006). In these areas, general health effects that are associated with arsenic exposure include lung cancer, cardiovascular and peripheral vascular disease, developmental anomalies, diabetes, lung fibrosis, and hematologic disorders (Nabi et al. 2005; Rahman et al. 2001; Tchounwou et al. 1999). Collectively, these studies demonstrate the vascular and respiratory health consequences observed around the world in people exposed to arsenic.

Findings from epidemiologic studies among populations exposed to lower levels of arsenic (< 100 parts per billion [ppb]) are mixed and generally do not reveal risks of bladder, lung, or other cancers that would be expected from extrapolation of results from the high-exposure studies. When errors in assessment of low exposure are made, about half of the actual risk increase is missed. Thus, bladder cancer studies conducted in populations exposed to low levels of arsenic are limited because of this problem (Bates et al. 1995; Bates et al. 2004; Karagas et al. 2004; Kurttio et al. 1999; Steinmaus et al. 2003). In addition, characterizing past low-dose exposure in detail is challenging, especially in shifting populations in the United States (Bates et al. 1995; Karagas et al. 2004; Steinmaus et al. 2003).

The term reduction–oxidation (redox) state is often used to describe the balance of redox pairs (for example, the glutathione system) in a biological system, which if not balanced, results in oxidative stress. Gene transcripts involved in oxidative stress have been found to be increased in people exposed to arsenic (Wang et al. 2003). Oxidative stress is produced by endothelial, smooth muscle, and fibroblast cells in response to arsenic (Barchowsky et al. 1996; Bau et al. 2002; Bunderson et al. 2002; Lynn et al. 2000; Wang et al. 1997; Yang et al. 2007; Yeh et al. 2002; Yih et al. 2002). Oxidative stress and the resultant disregulation of the extracellular matrix have been linked to both cancer and atherosclerosis through a series of common molecular pathways, which play a significant role in the pathogenesis and progression of these two diseases (Ross et al. 2001).

To investigate the molecular targets involved in long-term, low-dose, arsenic-associated cardiovascular and respiratory toxicity, we conducted a study using Affymetrix mouse 430(A) arrays. Using microarrays, we were able to identify genes with differential expression in the lungs of AsIII-exposed (50 ppb) and control C57BL/6 mice. Hearts were also harvested, and real-time RT-PCR demonstrated an AsIII-associated decrease in the levels of expression of many of the same extracellular matrix genes found to be decreased in the lung. This result was also supported in vitro by using murine embryonic fibroblasts (NIH3T3 cells) exposed to AsIII. The decrease in the extracellular matrix gene transcripts compelled us to investigate the vasculature. In the extravascular matrix of small arteries in the lung and heart, collagen and elastic fiber density and integrity were decreased in the arsenic-exposed groups as compared to controls. The staining for α smooth muscle actin was decreased in the small arteries of the heart. The data from this study indicate that significant disruption in the vascular extracellular matrix and decreases in extracellular matrix gene transcripts occur as a result of exposure to arsenic in drinking water at concentrations routinely found throughout the United States. The specific pattern of gene transcript changes will guide further mechanistic studies of cardiovascular and respiratory diseases induced by arsenic.

Materials and Methods

Animals

Adult C57BL/6 mice were housed in the AAALAC-approved University of Arizona Health Sciences Center animal facility. The animal protocol was approved by the University of Arizona Institutional Animal Care and Use Committee. Animals were housed four mice per cage and fed ad libitum with either normal drinking water or water containing 50 ppb AsIII as sodium arsenite (NaAsO₂) (Sargent-Welch, Tonawanda, NY, USA) for four, five, or eight weeks.

Cell Culture

NIH3T3 murine fibroblasts (ATCC) were grown in DMEM supplemented with 10% fetal bovine serum (Hyclone, Logan, UT, USA) and 1% antimycotic/antibiotic (Cellgro, Herndon, VA, USA). Cells were maintained at 37°C in an atmosphere of 5% CO₂and 95% air and subcultured twice a week. For experimental use, cells were plated into 100 X 20 mm tissue culture dishes (Sarstedt, Newton, NC, USA) at 8 x 10⁶cells per dish for culturing in normal media or media containing 10, 25, or 50 ppb AsIII as sodium arsenite (NaAsO₂) (Sargent-Welch, Tonawanda, NY, USA).

Affymetrix GeneChip System

Using three separate chips for the control and AsIII-treated groups, the Affymetrix GeneChip system was used to screen differential gene expression in C57BL/6 mouse lung tissue from controls or mice exposed to AsIII as sodium arsenite (NaAsO₂) (Sargent-Welch, Tonawanda, NY, USA). Total RNA was extracted from 100 mg of lung tissue homogenized in RNA STAT-60 (Tel-Test, Inc., Friendswood, TX, USA) using a Tissue Tearor (BioSpec products, Bartlesville, OK, USA). Samples were subjected to chloroform extraction and concentrated using a Qiagen RNeasy kit (Qiagen, Valencia, CA, USA). RNA quality was determined using the Agilent 2100 bioanalyzer (Aglient Technologies, Foster City, CA, USA). cRNA was synthesized and hybridized to Affymetrix mouse 430(A) arrays. Probe sets were normalized using Robust MultiChip Analysis (RMA) using GeneTraffic (Iobion) software (Version 3.2, Stratagene, La Jolla, CA; release date June 2004). These normalized data were then analyzed using Statistical Analysis of Microarrays (SAM) ttest or Cyber-T based on Bayesian statistics.

Real-Time RT-PCR Analysis

To validate data from the gene chip analysis, real-time RT-PCR analysis was performed. RNA was collected from whole lung as described above. Reverse transcription was performed using TaqMan reverse transcription reagents (Applied Biosystems, Foster City, CA, USA), and 50 ng of total RNA in a 50 μl reaction. The reverse transcription reaction was primed with random hexamers and incubated at 25°C for ten minutes followed by 48°C for thirty minutes, 95°C for five minutes, and a chill at 4°C. Each real-time PCR reaction consisted of 10 μl of cDNA added to 25 μl of TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA, USA), 2.5 μl of gene-specific primer/probe mix labeled with the 5′ reporter dye 6-FAM, and a 3′ end that contained a nonfluorescent quencher and a minor groove binder (Assays-on-Demand, Applied Biosystems, Foster City, CA, USA), and 12.5 μl of PCR water. The real-time PCR conditions were 95°C for ten minutes, followed by 40 cycles of 95°C for fifteen seconds alternating with 60°C for one minute. Data were collected using the ABI Prism 7000 real-time sequence detection system (Applied Biosystems, Foster City, CA, USA). To determine relative amounts of transcript between samples, the comparative threshold cycle (C_t) method was used. The relative starting amount of cDNA in each sample (control and 50 ppb arsenic) was determined by normalizing the C_tof the extracellular matrix gene-of-interest in the sample to the C_tof the housekeeping gene (β-actin) in the sample. This procedure was also followed for the calibrator cDNA, which was used to compare samples between analyses. Primer sequences are available in Table I of the supplementary data (please refer to http://tpx.sagepub.com/supplemental).

A different protocol was used for real-time RT-PCR analysis of collagen 1a1 (Col1a1), collagen 1a2 (Col1a2), collagen 3a1 (Col3a1), collagen 4a1 (Col4a1), fibronectin (Fn1), fibulin (Fbln1), microfibrillar associated protein 5 (Mfap5), hyaluronan acid synthase-2 (Has2), and elastin (Eln) gene expression in the heart and NIH3T3 murine embryonic fibroblasts. RNA was isolated and a reverse transcription step was performed using 1 μg of total RNA in a 20 μl reaction and the Transcriptor First Strand cDNA Synthesis Kit (Roche Applied Science, Indianapolis, IN, USA). The reverse transcription reaction was primed with oligo dTs and incubated at 65°C for ten minutes followed by 50°C for sixty minutes, 85°C for five minutes, and finally, 4°C until retrieved. Each real-time PCR reaction consisted of 2 μl of cDNA added to 4 μl of Roche Universal PCR Master Mix (Roche Applied Science, Indianapolis, IN, USA), 2 μl of gene-specific primer/probe mix (Universal Probe Library, Roche Applied Science, Indianapolis, IN, USA), and 12 μl of PCR water. The gene-specific probes are from the Universal Probe Library (UPL) (Roche Applied Science, Indianapolis, IN, USA). These probes are labeled with the 5′ reporter dye, 6-FAM, and a 3′ end that contains a nonfluorescent quencher and a minor groove binder to bind to the cDNA. The real-time PCR conditions were 95°C for ten minutes, followed by 40 cycles of 95°C for fifteen seconds alternating with 60°C for one minute. Data were collected using the LightCycler 2.0 detection system (Roche Applied Science, Indianapolis, IN, USA). To determine relative amounts of transcript between samples, the relative quantification method was used. The relative starting amount of each sample was determined by normalizing the crossing point (C_p) of the extracellular matrix gene-of-interest to the C_pof the reference molecule aminolevulinate acid synthase (Alas-1) (Roche Applied Science, Indianapolis, IN, USA). A calibrator was used to arrive at a normalized ratio (LightCycler software 4.0). Primer sequences are available in Table II of the supplementary data (please refer to http://tpx.sagepub.com/supplemental).

Histology

Animals were euthanized by asphyxiation with CO₂. The thoracic and abdominal skin was dissected free, and the pleural cavity was opened along the midline. Hearts were harvested in a 4% paraformaldehyde solution. The trachea was cannulated and in situ instillation of lungs with 10% buffered formalin was performed using a perfusion apparatus at 20 cm of H₂O pressure for one hour. The lungs were then placed in 10% buffered formalin in a scintillation vial and stored overnight at 4° C. Both the heart and lungs were then placed in 70% ethanol (ETOH) and stored overnight at 4°C. The next day, the heart and lungs were dehydrated, embedded in parafilm, and serial-sectioned, and selected samples were stained with Masson’s Trichrome for collagen and Verhoeff-van Gieson’s stain for Eln. Masson’s Trichrome stains the nuclei black, the muscle red, and the collagen blue. Verhoeff-van Gieson stains the nuclei and Eln black and the collagen red.

Analysis of Vascular Wall Complexity

A blinded subjective analysis of vascular wall complexity was performed using animals exposed to 50 ppb AsIII and control drinking water. Animals were euthanized, and hearts were harvested as above. The hearts were dehydrated, embedded in parafilm, and serial-sectioned, and selected samples were stained with Masson’s Trichrome for collagen. Ten random images of blood vessels from each treatment group (control and 50 ppb AsIII) were selected. Four independent researchers did not know which group was which. They described what they observed to be different and/or the same between groups.

Immunohistochemistry

Serial sections of tissue were deparaffinized, and antigen retrieval was performed using Cyto Q tissue recovery solution (Accurate Chemistry and Scientific Corp., Westbury, NY, USA). The slides were microwaved on high for five minutes in the solution and then were left to cool at room temperature for thirty minutes. They were then rinsed with 1X phosphate buffered saline (PBS). Cells were permeabilized with 0.2% triton in 1X PBS for twenty minutes. The slides were incubated in blocking serum for ninety minutes at room temperature to block nonspecific binding. Slides were incubated with monoclonal anti-α smooth muscle actin primary antibody (1:100) (Sigma, Saint Louis, MO, USA) in a humid chamber at room temperature for one hour. Slides were washed and incubated with goat anti-mouse secondary antibody conjugated to AlexaFluor-546 at 1:1000 (Invitrogen, Carlsbad, CA, USA) and Hoechst nuclear stain at 1:500 (Molecular Probes) in a lightproof box in a humid chamber for sixty minutes. Slides were washed and then mounted with Dako mounting medium. The tissues were imaged with a Leica DMLB fluorescence microscope (Leica Microsystems Inc., Bannockburn, IL, USA) and Image-Pro Plus imaging software (Version 5.1, Media Cybernetics Corp., Bethesda, MD; release date 2003).

Computed Morphometric Analysis

Computed morphometric analysis of elastic fibers in lungs from AsIII-treated animals was performed and compared with data obtained from control animals of the same age. To observe the general morphology of the lungs, 4-μm sections of formalin-fixed, paraffin-embedded lungs were stained with Miller’s elastic stain to observe elastic fibers in the alveolar walls. Microscope slides were analyzed using SimplePCI 6.0 (Hamamatsu Corp., Sewickley, PA, USA). The images were captured in bright-field mode on an Olympus IMT-2 inverted microscope with a Hamamatsu OCRA 100 greyscale CCD camera using a previously calibrated 20X objective. Intensity thresholding was performed to select the areas within the lung tissue that were stained for Eln, and these areas were measured by the software. A second intensity threshold was then performed to select the area of the lung tissue in the image. We excluded the airways, pulmonary arterial and venous vessels, interlobular septa, and the visceral pleura from this analysis.

Results

Significant Gene Expression Changes Grouped by Function

Changes in gene expression in the lungs of mice exposed to arsenic for five and eight weeks were grouped by function using information provided in the gene ontology (GO) annotation for each gene. GO categories include: receptors, transcriptional, signal transduction, chaperones, immunological, cell cycle, enzymes, cytokines (Table 1), and structural (Figure 1). Sixty-five genes were observed to be significantly increased or decreased in both the five- and eight-week AsIII-treated groups. These genes were ranked by pvalue, and the ones with pvalues < .0001 are listed in Table I of the supplementary data (please refer to http://tpx.sagepub.com/supplemental). The majority of these genes in the AsIII-exposed group were decreased in comparison to controls.

Decrease in Extracellular Matrix Genes in Response to AsIII

The majority (91%) of the extracellular matrix (ECM) genes, a subcategory of the structural genes, were decreased in the AsIII-treated animals as compared to controls (Figure 1, boxed genes). These include genes encoding for Eln, Mfap5, Lum, Col1a1, Col1a2, Col6a3, Col6a2, Col3a1, Col4a1, and Fn1. Other genes related to the extracellular matrix that showed a decrease in expression include Fbn1, Fbln1, and the enzyme involved in collagen and elastin processing, lysyl oxidase 1 (Lox1) (Table 1). All of these genes play a major role in extra-cellular matrix homeostasis.

Increase in Redox-Sensitive Genes in Response to AsIII

In the present study, AsIII significantly increased redox-sensitive genes, including heat shock protein 40 (hsp40/Dnaja1), heat shock protein 105 (hsp105), heat shock protein 1b (hsp1b), and osmotic stress protein 94 (Osp94), a heat shock protein family member (Table 1). In addition, microarray revealed an increase in several of the glutathione system genes, including glutathione-S-transferase (GST) α 2 and glutathione reductase (both showed a 2.3-fold increase in the eight-week AsIII-treated animals compared to the controls) and GST α 3 (a 2.25-fold increase in the eight-week AsIII-treated animals compared to the controls) (unpublished data). These genes are of interest because arsenic is a known oxidant stressor (Li et al., 2002) and is known to have a high affinity for oxidation-sensitive dithiols, which are targets of arsenic.

AsIII Exposure Decreases Extracellular Matrix Gene Transcript Expression in Fibroblasts

Because stromal fibroblasts are a main source of extracellular matrix molecules in mature tissues, we examined AsIII effects on matrix gene expression in NIH3T3 murine embryonic fibroblasts. Real-time RT-PCR analysis was performed for Col1a1, Col1a2, Col3a1, Col4a1, Has2, and Fn1 gene expression following treatment of fibroblasts with AsIII for one, four, and sixteen hours (Figure 2). Data presented are normalized to percentage of control. Col4a1 at sixteen hours (50 ppb) and Has2 at one hour (10, 25 and 50 ppb) and four hours (10 ppb) do not demonstrate a decrease in transcription levels in these genes. However, in most cases, the normalized values demonstrate significant decreases in transcription levels for the Col1a1, Col1a2, Col3a1, Col4a1, Has2, and Fn1 genes from cells exposed to 10, 25, and 50 ppb AsIII as compared to controls. This finding is in concordance with the gene expression microarray data from lungs of AsIII-treated animals, suggesting a direct effect by AsIII on gene expression in relevant cells.

AsIII Exposure Decreases Extracellular Matrix Gene Transcript Expression in the Heart

To investigate the possibility that AsIII affects the heart in a manner similar to the lung, hearts were harvested from control and AsIII-exposed (50 ppb, eight weeks) mice. Total RNA was isolated, and real-time RT-PCR analysis was performed for Col1a1, Col1a2, Col3a1, Col4a1, Fn1, Eln, Fbn1, Mfap5, and Has2 (Figure 3). In all cases, the values demonstrate significant decreases in transcription levels for these genes similar to that observed in the fibroblasts. This finding is consistent with the decreases in gene transcription found in the lungs of animals exposed to AsIII.

Disrupted Extracellular Matrix Deposition in Blood Vessels Exposed to AsIII

The extracellular matrix is critical for maintaining the integrity of the vasculature, so we used histological stains for collagen in the lung (Figures 4A and 4B) and heart (Figures 5G and 5H) and elastin in the lung (Figures 4C and 4D) and heart (Figures 5C and 5D) to highlight the morphological significance of the decrease in these gene transcripts in response to 50 ppb AsIII as compared to control. This morphological significance is seen dramatically in the vascular matrix. The vascular wall is composed of three cell types: endothelial cells (tunica intima), smooth muscle cells (tunica media), and fibroblasts (tunica adventitia) (Hallmann et al. 2005; Saharinen et al. 1999). Elastin staining shows the internal and external elastic laminae of control arteries in the lung (Figure 4C) and heart (Figure 5C) to be intact, whereas the elastic laminae are considerably less well defined, and in parts of the wall, nonexistent in the 50 ppb vessels in the lung (Figure 4D) and heart (Figure 5D). In the lung, the detection of collagen shows a disrupted and considerably less dense tunica adventitia in small arteries in the 50 ppb (Figure 4B) animals compared to controls (Figure 4A). In the heart, collagen deposition is also disrupted, and there is considerably less dense tunica adventitia in small arteries in the 50 ppb-exposed animals (Figure 5H) compared to controls (Figure 5G). In arteriolar walls, the collagen fibers are frayed, and some fibers may have lost some length as compared to the controls. These results were observed in all three animals exposed to AsIII. These histological stains for elastin and collagen demonstrate the association between the decrease in extracellular matrix gene transcription in organs of arsenic-treated animals and the morphological significance of this decrease on the vascular matrix in these organs.

Analysis of Vascular Wall Complexity

Using heart tissue stained with Masson’s Trichrome for collagen, ten random images of blood vessels from each treatment group (control and 50 ppb AsIII) were selected, and four independent researchers described what they observed to be different and/or the same between groups. In general, they described the control blood vessels in the heart as having a structured collagen network present in the extravascular matrix. In the control tissue, most blood vessel walls contained fairly dense collagen fibers, whereas only a few walls showed scant or loose collagen production. In contrast, they described the AsIII treated blood vessel walls to be surrounded by a dilated extravascular space with little evidence of a structured collagen network, that is, there was scant collagen present. In the AsIII-treated tissue, some blood vessel walls contained collagen fiber networks similar to control tissue, whereas many walls showed scant or loose collagen production.

Decrease in Detection of Anti-α Smooth Muscle Actin in the Tunica Media of Blood Vessels from Animals Exposed to AsIII

The extravascular matrix plays a role in smooth muscle cell maintenance, and the interplay between smooth muscle cells and the surrounding matrix is important in blood vessel wall homeostasis. In animals treated for four weeks, autofluorescence (anti-α smooth muscle actin) demonstrates intact internal and external elastic laminae in control arteries of the heart (Figure 5A), whereas in the 50 ppb vessels (Figure 5B), the elastic laminae are considerably less well defined, and in parts of the wall, nonexistent. In these same animals, smooth muscle cell density in the walls of the arteries of the heart was determined using anti-α smooth muscle actin antibody staining (Figures 5E and 5F). In these animals, exposure to AsIII (50 ppb) resulted in a decrease in detection of anti-α smooth muscle actin staining in the tunica media of arteries compared to controls. This finding may reflect a loss in smooth muscle cell integrity around arteries exposed to AsIII.

Computed Morphometric Analysis

The extracellular matrix component elastin is critical for maintaining the integrity of the lung and vasculature. To highlight the morphological significance of the decrease in elastin gene transcription found in the lung, we used computed morphometric analysis of elastic fibers from AsIII-treated and control animals of the same age. At least 100 fields were examined in each treatment group. The area of stained elastin is expressed as a percentage of the total area of lung tissue that was measured. Computer image analysis by means of thresholding demonstrated that on average, elastic fibers constituted 38.9% of the total tissue in the 50 ppb-treated animal lungs and 56.0% in control lungs.

Discussion

The mechanism(s) of arsenic-induced cardiorespiratory disease is not known. We used gene expression analysis along with complimentary histological and immunohistochemical approaches to correlate differential gene expression and morphological changes in the vascular matrix after exposure to environmentally relevant concentrations of arsenic. Signaling mechanisms affecting gene transcription could be modulated by AsIII. This modulation could be the result of interaction with proteins containing a critical cysteine that is in a vicinal dithiol grouping with a second cysteine. It is known that the stability of these complexes is favored by the exothermic formation of AsIII–thiolate bonds (Spuches et al. 2005). The dithiol complexing ability of AsIII allows it to act as an oxidizing agent, thus creating oxidative stress. For example, activation of NF-κB depends on the integrity of the IκB kinase (IKK) complex, which is inhibited by AsIII binding to Cys-179 in the activation loop of IKK (Kapahi et al. 2000). Arsenite has also been shown to inhibit DNA repair of oxidative damage by interaction with vicinal dithiols of DNA ligase and Poly(ADP-ribose) polymerase (PARP) (Lynn et al. 1997; Yager and Wiencke 1997). Oxidative stress has been implicated in arsenic-associated respiratory and cardiovascular diseases (Bunderson et al. 2002; Han et al. 2005). Because arsenic is a known oxidant stressor, we would expect AsIII to increase redox-sensitive genes. The redox-sensitive gene transcripts that are increased in the lungs of animals exposed for both five and eight weeks include heat shock and stress proteins (Table 1) and glutathione metabolizing enzymes (unpublished data). These enzymes include: glutathione-S-transferase-α 2 (GST-α 2), GST-α3, and glutathione reductase (a 2.3-fold increase compared to controls).

Of particular interest to us in this study are the extracellular matrix molecules collagen and elastin. In the matrix, the collagen isoforms include type I (tensile strength), type III (subendothelium), type IV (basal lamina), and type VI (bridges cells with extracellular matrix). In addition, type I and III collagens are the main extracellular matrix components of the lung (Crystal and West 1997), and they are involved in wound healing (Van Hoozen et al. 2000). Studies of genetic defects in collagen and collagen-associated molecules aid in understanding the phenotypic consequences of decreases in collagen gene transcription, which can be observed in genetic diseases affecting the vasculature. Aortic aneurysms are a consequence of Marfans syndrome, which is associated with defects in fibrillin and type I collagen assembly (Whiteman et al. 2006). Altered blood vessel stability is the result of collagen defects in both osteogenesis imperfecta and Ehlers-Danlos syndrome (Kuznetsova et al. 2004). As demonstrated in the present study, AsIII-associated decreases in collagen and fibrillin may mimic the collagen-related syndromes, resulting in a disrupted blood vessel structure. Because arsenic has a high affinity for vicinal dithiols (Rey et al. 2004), it is important to note that collagen production is regulated by molecules that contain vicinal dithiols. These include: epidermal growth factor (EGF) (Saharinen et al. 1999), the transcription factor c-Krox (Galera et al. 1994; Karsenty and Park 1995), and N-propeptides, which control collagen fibril shape (Wu et al. 1991). It is possible that the decrease in collagen transcription was in part mediated by its interaction with the vicinal dithiols of molecules that modulate collagen production. Further research is needed to confirm the interaction of arsenic with cysteine-rich molecules in this context.

Because exposure to ingested arsenic is associated with increased risk for various vasculopathies (Soucy et al. 2004; Tseng 2002; Yang et al. 2005), we focused on important extravascular components. We used Masson’s Trichrome stain to highlight the destabilized and disorganized collagenous tunica adventitia in arteries of both the lung and heart in the 50 ppb group (Figures 4Aand 5G) as compared to controls (Figures 4Band 5H). In addition, a blinded subjective analysis of vascular wall complexity in the heart using Masson’s Trichrome stain for collagen was performed. This is an excellent method to obtain a large amount of information, as it is easy to make out the extensive collagenous fiber reinforcement of the blood vessel wall, and most aspects of vessel wall pathology can be appreciated. In the control tissue, most blood vessel walls contained fairly dense collagen fibers; in contrast, the AsIII-treated blood vessel walls were surrounded by a dilated extravascular space with little evidence of a structured collagen network.

Taken together, the gene expression and histological studies reveal that collagen and collagen-related gene transcript levels and subsequent collagen disorganization in the arteriolar wall are affected by an environmentally relevant level of arsenic that may predispose exposed individuals to vascular pathologies in general. In addition to collagen, microarray analysis and subsequent real-time analysis (data not shown) revealed that in response to 50 ppb AsIII, elastin gene transcription was decreased in the lung (Figure 1) and heart (Figure 3). Elastin is a major component of the extravascular matrix, and elastin fiber assembly employs elastin microfibril-associated proteins. These proteins include microfibril associated protein-5 and the fibulins and fibrillins, which are a family of secreted adhesive glycoproteins that link cells to the elastic fibers of the extracellular matrix (Freeman et al. 2005; Saharinen et al. 1999). Histological analysis demonstrated that compared to controls, elastin staining is disrupted in the extravascular matrix of arteries in the lung and heart of mice exposed to 50 ppb AsIII (Figures 4C–Dand 5C–D). Because the study of elastin knockout mice reveals elastin to be a major developmental regulator of the vascular smooth muscle cell life cycle and organization (Faury 2001), we looked at α-smooth muscle actin staining in blood vessel walls. The decrease in elastin stain was found concomitant with reduced detection of α-smooth muscle actin staining in blood vessel walls in animals exposed to 50 ppb AsIII as compared to controls (Figures 5E–F).

Immunohistochemical analysis also demonstrated a disruption in the elastic lamellae of the tunica media at 50 ppb AsIII as compared to controls (Figure 5A and 5B). Computed morphometric analysis of elastic fibers in lungs from AsIII-treated animals was performed and compared with data obtained from control animals. Image analysis demonstrated that on average, elastic fibers constituted 38.9% of the total tissue in the 50 ppb-treated group and 56.0% in control lungs. Thus, elastin volume density in the lung in AsIII-exposed animals was less than in control lungs. This morphometric analysis confirms that there are significant differences in elastic fiber density between AsIII-treated and control mice and that 50 ppb AsIII causes abnormal patterns of elastin expression and accumulation. Because elastic fibers give blood vessels important properties, such as elastic recoil, the clinical relevance of AsIII-dependent disruption of elastic fibers would be a progressive impairment of their function with a concomitant production of circulating elastin peptides, which could potentially be measured with a blood test.

Our results provide new insight into the arsenic-associated effects on smooth muscle cells of blood vessel walls. Our observation of altered arteriolar matrix with decreased α-smooth muscle actin and elastin staining suggests smooth muscle cell dysfunction. Altered smooth muscle cell activity has been implicated in arsenic-associated cardiovascular diseases (Engel et al. 1994). To study the in vivo effects of AsIII, we used a fibroblast cell line to substantiate the reduction in the transcription of Col1a1, Col3a1, Col4a1, Col6a2, Eln, Fbln1, Fbn1, and Mfap5 in the lungs and hearts of mice exposed to arsenic. Thus, a substantial number of gene transcripts found to be decreased in AsIII-treated animals were also demonstrated to be decreased in AsIII-treated fibroblasts which are critical cells for extracelluar matrix deposition, structure, and maintenance.

Data presented here are the first to demonstrate the arsenic-associated correlation between genomics, histology, and immunohistochemistry and its disruptive effects in vivo on blood vessels in the lung and heart. These changes are occurring at environmentally relevant, long-term exposure levels. When analyzed in the context of a decrease in critical extracellular matrix gene transcripts (collagen, Eln, Fn1, Lox1, Mfap-5, Has2, Fbln1, and Fbn1), the results suggest a mechanism for arsenic-induced respiratory and cardiovascular pathologies. This mechanism includes both interruption of gene transcription and direct disruption of extracellular matrix components in blood vessels in target organs of arsenic toxicity. Analysis of the molecular mechanisms of altered matrix expression will further our understanding of the sites of action of arsenic.