Inorganic Arsenic–Induced Intramitochondrial Granules in Mouse Urothelium

Abstract

Based on epidemiological data, chronic exposure to high levels of inorganic arsenic in the drinking water is carcinogenic to the urinary bladder of humans. Recently, models have been developed involving transplacental administration of inorganic arsenic and subsequent administration of another substance that produces a low incidence of urogenital neoplasms. Administration of arsenite or arsenate in the diet or drinking water to five-to eight-week-old mice or rats rapidly induces urothelial cytotoxicity and regenerative hyperplasia. In mice administered arsenite, we observed eosinophilic intracytoplasmic granules present in the urothelial cells. These granules were not present in urothelial cells of untreated mice or in treated or untreated rats. By transmission electron microscopy, the granules were located within the mitochondrial matrix, that is, mitochondrial inclusions. Arsenic, primarily as arsenite, was present in partially purified mitochondria containing these granules. Cells containing the granules were not usually associated with degenerative changes. Lack of these granules in rats suggests that they are not necessary for inorganic arsenic–induced urothelial cytotoxicity or hyperplasia. These granules have also been observed with exposures to other metals in other tissues and other species, suggesting that they represent a protective mechanism against metal-induced toxicity.

Introduction

Based on epidemiological data, chronic exposure to high levels of inorganic arsenic in the drinking water is carcinogenic to the urinary bladder, skin, and lungs of humans (National Research Council [NRC] 1999). Inorganic arsenic exists mainly in the environment as trivalent arsenite or pentavalent arsenate (NRC 1999). Although inorganic arsenic has long been known to be carcinogenic to humans, there is still little known regarding its mechanism of action, in part due to the general lack of animal models.

Previous long-term studies in rats and mice with inorganic arsenic were negative (Byron et al. 1967; Soffritti et al. 2006). However, Waalkes et al. (2007) have recently demonstrated that high doses of arsenite (As^III) administered to mouse dams transplacentally induces cancer in the offspring in a variety of tissues. If the pups are subsequently administered diethylstilbestrol or tamoxifen, some eventually develop urogenital neoplasms. Furthermore, Simeonova et al. (2000) demonstrated that As^III administered in the drinking water (0.01%) to C57BL x BDA2 mice for four weeks produced evidence of urinary bladder urothelial simple hyperplasia.

In a previous short-term study involving administration of arsenate (As^V) in the drinking water to rats, we also observed urothelial hyperplasia (Lu et al. 2007). In subsequent experiments involving administration of As^III and As^V to mice, we noticed the presence of intracytoplasmic eosinophilic granules in the mouse urothelium, especially in the superficial cell layer (Suzuki et al. in press). In the present study we examined this phenomenon in greater detail in urothelial cells of mice administered As^III. By transmission electron microscopy (TEM), we demonstrated that the granules are within the mitochondria. The urothelium in these mice also showed evidence of hyperplasia. Interestingly, in contrast to mice, treatment of rats with 200 mg/L sodium arsenite (NaAs^III) in the food did not produce mitochondrial granules (MG) despite detection of simple hyperplasia in the urinary bladder epithelium. MG have been observed in various tissues in different species subjected to exposure to certain heavy metal cations. The granules are mainly associated with the mitochondrial inner membrane and contain the metal within the matrix of the mitochondria (Brown et al. 1985; Jacob et al. 1994). The mitochondrion is an organelle that can remove heavy metals from epithelial cell cytoplasm, sequestering the potentially toxic metal in the granule (Ahearn et al. 2004). In this report, we provide the details of the presence of these MG in mice administered inorganic arsenic and their relationship to arsenic-induced urothelial changes in mice.

Materials and Methods

Chemicals and Diets

NaAs^III (purity, 94%) was purchased from Sigma (St. Louis, MO). Basal diet (Centrified Rodent Diet 5002, PMI Nutrition International, Inc., St. Louis, MO) and the diets containing NaAs^III were pelleted by Dyets, Inc. (Bethlehem, PA) ten weeks or less before use. All diets were stored at −20°C.

Test Animals and Experimental Design

Female F344 rats and female C57BL/6 mice, four weeks old, were purchased from Charles River Breeding Laboratories (Raleigh, NC). On arrival, the animals were placed in a level-4 barrier facility accredited by the American Association for Accreditation of Laboratory Animal Care (AAALAC), in a room with a targeted temperature of 22°C, humidity of 50%, and a 12-h light/dark cycle (0600/1800). The level of care provided to the animals met or exceeded the basic requirements outlined in the Guide for the Care and Use of Laboratory Animals (NIH Publication #86-23, revised 1986). The animals were housed five/cage in polycarbonate cages on dry corncob bedding and fed basal diet. Food and tap water were available ad libitum throughout the study. Fresh diet was supplied to the animals at least once weekly. Food consumption was measured during study week 2. Body weights of all animals were measured the day after arrival, on study day 0, and on the day of sacrifice. Detailed clinical observations of each animal were performed on day 0 and on the last day of the consumption period. All animals were sacrificed by an overdose of Nembutal (150 mg/kg of body weight, i.p.).

Experiment 1

Rats and mice were approximately six weeks of age at the beginning of treatment. Following quarantine for eleven days, animals were randomized using a weight stratification method (Martin et al. 1984) into two groups of ten mice each or five rats each: group 1 was fed basal diet only, group 2 was fed a diet containing 350 ppm NaAs^III (200 ppm As^III). All animals were sacrificed after two weeks of treatment. Prior to removal, with the animal under deep anesthesia, two urinary bladders in each group were inflated in situ with 4% paraformaldehyde fixative, and after removal, the bladders were placed in the same fixative. The bladders were rinsed in 0.1 M Sorensen’s buffer three times, bisected longitudinally, and weighed. One-half of the bladder was placed in 0.1 M Sorensen’s buffer and processed for TEM. The other half of the bladder was placed in 70% ethanol until processing for light microscopic examination. All of the remaining rat and mouse bladders were inflated in situ with and placed in Bouin’s fixative and rinsed in 70% ethanol. At that time, the bladders were embedded in paraffin wax, sectioned at 4 μm, and stained with hematoxylin and eosin (H&E) for light microscopic examination.

Experiment 2

To further investigate the mitochondrial granules in the bladder, a second experiment was performed using mice approximately five weeks of age at the beginning of treatment. Following quarantine for seven days, mice were randomized using a weight stratification method (Martin et al. 1984) into two groups of thirty mice each: group 1 was fed basal diet only; group 2 was fed a diet containing 170 ppm NaAs^III (100 ppm As^III). All animals were sacrificed after two weeks of treatment. Prior to removal, the urinary bladder was inflated in situ with phosphate buffered saline (PBS), and after removal, the bladders were opened and the epithelia scraped gently with the dull edge of a scalpel blade. The epithelia were immediately placed into PBS with protease inhibitor (Sigma) and kept at approximately −80°C until processed for purification of mitochondria.

Purification of Mitochondria

The epithelia collected in experiment 2 were centrifuged, the supernatant was removed, and the pellet was resuspended in isolation buffer (0.25M sucrose, 40mM KCl, 2mM EGTA, 20mM Tris [pH 7.2], 1 mg/ml bovine serum albumin) with 0.01% digitonin and 10% Percoll. After incubation on ice for 10 min, the mitochondrial pellet was recovered from the supernatant after centrifugation at 5,000 ×g for 5 min. The pellet was resuspended in isolation buffer, and centrifuged at 8,000 ×g for 5 min. Pelleted mitochondria were kept at approximately −80°C until processed for speciation and quantitation of arsenicals and examination by TEM.

Determination of Arsenic Speciation in Mitochondria

The quantification of arsenic species in mitochondria samples was performed using ion-pair chromatographic separation (Le et al. 2000) and inductively coupled plasma mass spectrometry (ICPMS) detection (Le et al. 2004; Yuan et al. 2008). An Agilent 1100 series liquid chromatograph (Agilent Technologies, Inc., Santa Clara, CA), equipped with an autosampler and column temperature control, was coupled to an Agilent 7500cs octopole reaction system ICPMS, operated with the helium mode. Chromatographic separation was achieved on a reversed-phase ODS-3 column (Phenomenex, 150 ×4.6 mm, 3-μm particle size) maintained at 50°C. A mobile phase contained 5 mM tetrabutylammonium hydroxide, 3 mM malonic acid, and 5% methanol at pH 5.85. ICPMS was operated with the helium mode (to reduce isobaric interference), and arsenic at m/z 75 was monitored. A certified reference material No.18 Human Urine from National Institute for Environmental Studies (Japan) was used for quality control. Mitochondria samples as received were diluted forty times with deionized water, and a 20-μL diluted aliquot was injected onto the chromatography column for arsenic speciation analysis.

Transmission Electron Microscopy

To investigate bladder tissue by electron microscopy (EM), two bladders in each group in experiment 1 were fixed for at least sixty minutes in 4% paraformaldehyde. Multiple 1 mm × 3 mm strips were dissected and fixed in MPG electron microscopy fixative (pH 7.38) for two hours at 4°C. EM blocks were washed in 0.1 M phosphate buffer (2 × 5 min), processed through a six-hour EM processing schedule, and infiltrated with epon resin (Electron Microscopy Sciences, Fort Washington, PA) on a Leica EM TP automatic tissue processor (Leica Microsystems, Vienna, Austria). Tissues were embedded, oriented longitudinally and polymerized at 60°C for twenty-four hours. Multiple 1μ thick sections stained with 1% toluidine blue in 1% borax were evaluated by light microscopy. Representative areas were selected, ultrasectioned at 70nm (silver sections), mounted on 300 mesh Athene copper grids, and double stained with Reynolds lead citrate and uranyl acetate.

For checking mitochondrial samples in experiment 2 by EM, mitochondrial isolates were fixed in 4% paraformaldehyde in phosphate buffer for two hours and after centrifugation at 1,700 rpm for five minutes, the supernatant was removed. The pellets were resuspended once in 1% osmium tetroxide for forty-five minutes and twice in tripled distilled deionized water for five minutes each wash. The pellets were centrifuged for five minutes and the supernatant was removed after each resuspension. The pellets were then dehydrated in a graded series of ethyl alcohol and infiltrated with propylene oxide (PO) (5 minutes × 2). The pellets were infiltrated with fresh resin according to routine infiltration methods; 2:1 PO/resin for thirty minutes, 1:1 PO/resin for thirty minutes, and finally neat resin overnight at room temperature. Pellets were then embedded and polymerized overnight at 60°C.

All samples were examined with a JOEL 1320 electron microscope (JOEL USA, Inc., Peabody, MA). Representative digital images were taken using a SIS Keenview camera and software.

Results

Animal Parameters

In experiment 1, body weight gain was suppressed in the group fed As^III during the experimental period. The final body weights of mice treated with As^III (12.5 ± 0.7 g) were significantly lower than body weights of mice fed the basal diet (16.7 ± 0.8 g; p < 001). There was a significant difference in food consumption between mice given As^III and their counterparts receiving basal diet (2.0 ± 0.5, 3.0 ± 0.1 g/mice/day, respectively; p < 01). In experiment 2, suppression of body weight gain was also observed in the group of mice fed As^III during the experimental period. The final body weights of mice treated with As^III (15.5 ± 0.8 g) were significantly lower than body weights of mice fed the basal diet (16.8 ± 0.9 g; p < 001). There was a significant difference in food consumption between mice fed As^III and their counterparts receiving basal diet (4.1 ± 0.5, 5.4 ± 0.2 g/mice/day, respectively; p < 001).

In experiment 1, suppression of body weight gain was observed in the group of rats fed As^III during the experimental period. The final body weights of rats treated with As^III (124.2 ± 2.2 g) were significantly lower than with the basal diet (132.8 ± 6.6 g; p < 05). Food consumption in rats treated with As^III tended to be reduced compared with controls.

Histopathological Microscopy

In experiment 1, light microscopic examination of H&E-stained slides showed that most superficial cells of the bladders in all mice treated with As^III for two weeks had mild swelling and small eosinophilic granules in the cytoplasm (Figure 1B–C). Additionally, mild simple hyperplasia of the bladder epithelium was found in two of eight mice treated with As^III for two weeks (Figure 1C). Von Kossa staining (Sheehan et al. 1980) for calcium accumulation in the urothelial cells was negative. In contrast, the bladder epithelium was normal in all control mice (Figure 1A).

In rats, mild simple hyperplasia of the bladder epithelium was observed in all rats treated with As^III for two weeks (Figure 1E). The bladder epithelium was normal in all control rats (Figure 1D). There were no granules in the bladder of rats in either group.

Transmission Electron Microscopy

At low power, toluidine blue-stained sections from all species showed the general structure of the bladder (Pauli et al. 1983). The wall of the bladders consisted of three loosely arranged layers of smooth muscle with intertwining elastic and collagen fibers, at times difficult to distinguish. The transitional epithelium (urothelium) showed the typical distended appearance, that is, loss of folds, and was three to four cell layers thick. A delicate, often incomplete, muscularis mucosa was identified at high power in the bladders from control and As^III-treated mice and rats. The muscularis separated the lamina propria from the submucosa in some but not all sections. The outer adventitia showed normal blood vessels and lymphatics. Several sections from As^III-treated mice showed evidence of edema. Light microscopic examination at high power showed that the superficial and intermediate epithelial cells contained numerous cytoplasmic, dark blue-staining spherical granules of varying sizes scattered throughout the cytoplasm and frequently occurring in clusters. The granules were present in varying amounts in most superficial and intermediate cells and infrequently in basal cells. Granules were not present in nonepithelial cells. No granules were identified in sections of urothelium from the untreated mice or in sections from the bladders of control and As^III-treated rats.

Blocks from the bladders of As^III-treated mice and rats and random blocks from untreated mouse and rat bladders were ultrasectioned at 90 μ thickness. Ultrastructural examination of the urothelium from As^III-treated mice revealed the normal stratified epithelium with three to four cell layers of distended bladder mucosa (Pauli et al. 1983). The basal cells maintained their normal cuboidal appearance with normal cytoplasmic organelles, and they were seen lying on a normal, thin basement membrane. The intermediate cell layer maintained a distended appearance with cells generally cuboidal. The superficial cell layer demonstrated the typical stretched appearance of a distended urothelium. The very large (28–32 μ in diameter, 16–18 μ in thickness), thin superficial (umbrella or dome) cells showed the typical ultrastructural appearance. Their surface plasma membrane in many ultrasections appeared to consist of thickened inflexible plaques, referred to as the asymmetric unit membrane, which were interspersed with narrow zones of normal symmetric membrane. Typical infolding of the cytoplasmic membrane was evident, forming deep clefts and stacks of flattened plasma membrane segments known as fusiform vesicles. The cytoplasmic organelles and nuclei appeared normal except for frequent spherical granules (Figure 2B). Plentiful junctional complexes were present between cells maintaining the cohesion of adjacent cells. There was no evidence of cytoplasmic or nuclear degeneration of intermediate or basal cells. There were some focal degenerative changes of the superficial cells in the control and treated mice. A normal loose lamina propria was demonstrated in all specimens. No calcium deposition was detected in cells of any of the layers.

At high resolution, spherical granules were found to be located within the mitochondrial matrix, that is, mitochondrial inclusions. A majority of the granules were surrounded by a thin margin of mitochondrial matrix (Figure 2D). The membranes and cristae of the mitochondria, even those containing granules, appeared to be intact. Occasional mitochondria contained multiple granules of varying densities (Figure 2E). Most of the granules were semi-electron dense, homogeneous, and lipoid in appearance. They varied in size, ranging from 250–600nm in diameter. Numerous mitochondrial granules were identified in the superficial, intermediate, and basal cells, with the highest concentration found in the superficial cells.

No evidence of mitochondrial granules could be identified in the bladders of untreated mice (Figure 2A) or in the bladders of control or As^III-treated rats (Figure 2C).

Determination of Arsenic Speciation in Mitochondria

The concentrations of arsenic species in the mitochondria sample from treated mouse cells were (mean ± standard deviation) (μM): inorganic As^III 6.53 ± 0.15, dimethylarsinic acid (DMA^V) 0.50 ± 0.01, and inorganic As^V 0.13 ± 0.00. No other arsenic species was detectable (detection limit of 0.003 μM). MG were confirmed by TEM (Figure 2G), and not detected in mitochondria of control epithelium (Figure 2F).

Discussion

To the best of our knowledge, this is the first report of detection of MG in the bladder epithelium of mice treated with As^III. In previous studies, we demonstrated that eosinophilic granules were present in the bladder epithelium of mice treated with DMA^V (Arnold et al. 2006), but not in rats treated with DMA^V or mice or rats treated with monomethylarsonic acid (MMA^V) (Arnold et al. 2003). In other studies in our laboratory, we have discovered MG in urothelial cells in mice treated with As^III or As^V administered in the drinking water, the same as in the present study using administration in the diet (Suzuki et al. in press). Eosinophilic granules in urothelial cells were also observed in the bladders of mice administered NaAs^III in the drinking water in studies conducted by the National Toxicology Program, similar in appearance to what we have observed (Dr. Robert Maranpot, personal communication). We have also observed these granules in mice administered sodium As^V in the drinking water. We have not observed these granules in rats administered inorganic As^III or As^V or organic DMA^V or MMA^V (Arnold et al. 2006, 2003; Lu et al. 2007), nor have we detected them in hamsters administered DMA^V in the diet (Cano et al. 2001).

For heavy metal cations, there are a variety of detoxifying mechanisms involving metallothionein, glutathione, metal transport proteins, and vacuolar sequestration mechanisms in epithelial cells (Ahearn et al. 2004). In a study using human gingival fibroblast cells, the subcellular localization of metal ions differed for each metal examined and, in some cases, for each valence state of the same metal (Messer et al. 2002). The differences in localization correlated with cellular changes in morphology and alterations in toxicity. In mitochondria, some investigators have reported that the toxicity of As^III in vitro induces apoptosis or necrosis through a mitochondrial pathway caused by oxidative stress with release of cytochrome c (Bustamante et al. 2005; Santra et al. 2007). In contrast, however, the mitochondrion can function to sequester heavy metals such as As^III from the cytoplasm (Ahearn et al. 2004). MG in mouse bladder urothelial cells might be a mechanism for cellular protection from As^III toxicity. Regardless, MG in urothelial cells following exposure to arsenicals appear to occur only in mice, even though cytotoxicity and regeneration are similar in rats and mice.

In the present study, organic arsenicals were found in the MG as detected by HPLC and ICPMS. When inorganic arsenic is ingested, it undergoes a series of reductions and oxidative methylations before being excreted in the urine of most species, including humans and rodents (Cohen et al. 2006; Wang et al. 2004; Yoshida et al. 1998). In vitro studies using rat or human urothelial cell lines showed that trivalent organic arsenicals, monomethylarsonous acid (MMA^III) and dimethylarsinous acid (DMA^III), are approximately one thousand times more cytotoxic than the corresponding pentavalent arsenicals (Cohen et al. 2002; Valenzuela et al. 2005). The trivalent arsenicals, especially MMA^III and DMA^III, are believed to be the reactive species producing cytotoxicity in vitro and in vivo. Additionally, cellular uptake of trivalent arsenicals is more extensive than uptake of pentavalent arsenicals (Cohen et al. 2006; Dopp et al. 2004). In this study, organic trivalent arsenicals were not detected because they are unstable. Mitochondria in the bladder epithelium might protect against toxicities of trivalent organic arsenicals, but this appears to be unique to mice. The biologic importance of this phenomenon is unknown, but it does not appear to affect the urothelial cytotoxic and regenerative response to arsenicals in mice since the response is similar in mice and rats.

DMA^V administered in the diet or drinking water produces cytotoxicity, regenerative hyperplasia and tumors of the urinary bladder urothelium in rats but not in mice (Arnold et al. 2006). MMA^V does not produce urothelial lesions in rats or mice (Arnold et al. 2003). As^III and As^V administered in the diet or drinking water induce urothelial cytotoxicity and regenerative proliferation in rats and mice (Lu et al. 2007; Suzuki et al. in press). The MG are detected only in mice. They do not appear to be associated in mice with degenerative changes in the mitochondria of the cells in which they are detected. Based on their absence in rats, even with induction of cytotoxicity, hyperplasia, or tumorigenicity, the MG appear not to be necessary for the urothelial toxicity of arsenicals. Instead, they appear to be acting as a possible mechanism for protecting cells, similar to what has been observed with several other metals in various tissues of other animal species.