An integrated pharmacokinetic and pharmacodynamic study of arsenite action

Toxicology 206 (2005) 389–401

An integrated pharmacokinetic and pharmacodynamic study of arsenite action 2. Heme oxygenase induction in mice Elaina M. Kenyona,∗ , Luz Maria Del Razob , Michael F. Hughesa , Kirk T. Kitchinc a

Experimental Toxicology Division, U.S. Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory/ETD/PKB, Mail Drop B143-01, Research Triangle Park, NC 27711, USA b Environmental Pharmacology and Toxicology Department, CINVESTAV-IPN, Mexico City, Mexico c Environmental Carcinogenesis Division, U.S. Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory, Research Triangle Park, NC, USA Received 9 February 2004; received in revised form 2 July 2004; accepted 2 August 2004 Available online 17 September 2004

Abstract Heme oxygenase (HO) is the rate-limiting enzyme in heme degradation and its activity has a significant impact on intracellular heme pools. Rat studies indicate that HO induction is a sensitive, dose-dependent response to arsenite (As(III)) exposure in both liver and kidney. The objective of this study was to evaluate the relationship of HO induction to administered As(III) dose, and concentrations of inorganic arsenic (iAs) in tissues and urine. Levels of iAs, mono- (MMA) and dimethylated arsenic (DMA) as well as HO activity were determined in liver, lung and kidney over time in female B6C3F1 mice given a single oral dose of 0, 1, 10, 30 or 100 ␮mol/kg As(III). Increased HO activity was a time and dose-dependent response in liver and kidney, but not in lung. Activity peaked in the 4–6 h time range in liver and kidney with the responsiveness in liver being ∼2- to 3-fold greater than kidney. The lowest observed effect levels (LOELs) in this study for HO induction are 30 and 100 ␮mol/kg, respectively, in liver and kidney. The predominant form of arsenic (As) was iAs in liver at all doses, whereas DMA was the predominant form of As in kidney at all doses. Three- to four-fold higher levels of iAs were achieved in liver compared to kidney. MMA was the least abundant form of As in liver and kidney, never exceeding more than 20% of the total As present. The concentration of iAs in tissue or urine demonstrated the strongest correlation with HO activity in both liver and kidney. Results of this study suggest that HO induction is a biomarker of effect that is specific for tissue iAs because a high, but nontoxic, acute dose of DMA (5220 ␮mol/kg) did not induce HO in mice. Thus, HO induction has potential for use as a biomarker of effect for inorganic arsenic exposure and may be used as an indicator response to further the development of a biologically-based dose response model for As. Published by Elsevier Ireland Ltd. Keywords: Arsenic; Arsenite; Mice; Heme oxygenase

∗

Corresponding author. Tel.: +1 919 541 0043; fax: +1 919 541 4284. E-mail address: [email protected] (E.M. Kenyon).

0300-483X/$ – see front matter. Published by Elsevier Ireland Ltd. doi:10.1016/j.tox.2004.08.003

390

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

1. Introduction Inorganic arsenic (iAs) occurs in drinking water in many areas of the world, both as a natural contaminant and as a consequence of industrial activity, at levels sufficient to cause adverse health effects. In epidemiological studies, chronic ingestion of arsenic-contaminated drinking water has been strongly associated with the development of peripheral vascular diseases, and skin and internal cancers (bladder, lung, kidney, liver) in humans (National Research Council (NRC), 1999, 2001). The mechanism(s) by which As induces cancer are not known with certainty. Experimental evidence supports a number of different modes of action including chromosomal abnormalities, effects on DNA methylation, altered cell proliferation, co-carcinogenicity and oxidative stress (Kitchin, 2001; Hughes, 2002; Kitchin and Ahmad, 2003). Heme oxygenase (HO) catalyzes the oxidative degradation of protoheme to biliverdin. HO is the ratelimiting enzyme in heme degradation and thus has a central role in regulating intracellular heme pools (Maines, 1984). Two forms of HO have been identified, an inducible form (HO-1) and a constitutive form (HO-2), which are different gene products and differ markedly in primary structure, regulation and tissue distribution (Maines, 1984, 1997). HO-1 induction is considered a generalized response to oxidative stress and is caused by a variety of stimuli including diverse classes of chemicals (e.g., metals, organics, hormones), radiation and pathophysiological conditions (e.g., heat shock, GSH depletion, ischemia, fever, starvation, etc.) (Maines, 1988). Induction of HO-1 causes modulation of several antioxidant factors – increased bilirubin and biliverdin concentrations, decreased heme and hemoprotein concentration (Maines, 1988) and increased ferritin synthesis (Guzzo et al., 1994). In the rat, acute parenteral or oral administration of arsenite [As(III)] causes a strong dose-dependent induction of both renal and hepatic HO activity (Albores et al., 1989; Cebrian et al., 1988; Sardana et al., 1981; Brown and Kitchin, 1996; Kitchin et al., 1999). Arsenite also induces HO in the liver, kidney and lung of guinea pigs (Falkner et al., 1993). Arsenate induces both hepatic and renal HO in the rat, but is 3- to 6-fold less potent than arsenite (Sardana et al., 1981; Brown and Kitchin, 1996). On the other hand, the pentavalent methylated metabolites of arsenic, monomethylarsonic

acid (MMA) and dimethylarsinic acid (DMA), do not induce HO in either rat liver or kidney (Brown and Kitchin, 1996). Our recent studies indicate HO induction is a sensitive, dose-dependent response to As(III) exposure in both rat liver and kidney and that this response is highly correlated with tissue levels of iAs (Kitchin et al., 1999). Past researchers have preferred the rat to the mouse as an animal model for studying the HO response. In contrast, the mouse is preferred over the rat for pharmacokinetic and toxicity studies of As because the substantial binding of DMA (the primary metabolite of iAs) to hemoglobin in rat erythrocytes results in a much longer biologic half life in the rat compared to the mouse (Vahter et al., 1984). Data on HO induction in mice following oral or parenteral administration of arsenic are sparse and studies to date have not attempted to correlate arsenic dose to target tissue with HO induction. The objective of this study was to evaluate the relationship of HO induction and tissue levels and urinary excretion of iAs in mice.

2. Materials and methods Sodium arsenite and sodium arsenate were obtained from Sigma Chemical Co. (St. Louis, MO). MMA and DMA were obtained from Ventron (Danvers, MA) and Nakarai Chem (Kyoto, Japan), respectively. Sodium borohydride was obtained from EM Science (Gibbstown, NJ). Hydrochloric acid (HCl) was obtained from Fisher Scientific Co (Fair Lawn, NJ). Bovine Liver Standard Reference Material (SRM1577b) for toxic metals was obtained from the US National Institute of Standards and Technology (Gaithersburg, MD). All other chemicals used were of the highest grade commercially available. Ninety-days-old female B6C3F1 mice were obtained from Charles River Laboratories (Raleigh, NC). The animals were maintained according to the guidelines in the NIH, Guide to the Care and Use of Laboratory Animals within an AAALAC-approved animal facility. They were housed in polycarbonate shoebox cages (5–6/cage) with wood chip bedding and were provided with Rodent Chow (Purina, St. Louis, MO) and tap water ad libitum. The manufacturer’s stated level of total arsenic in the rodent chow is less than 0.098 ppm total arsenic. The room was kept on a

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

12/12-h light/dark cycle and at a temperature of 22 ± 1 ◦ C and humidity of 50 ± 10%. Sodium arsenite was dissolved in HPLC grade water and administered to the mice as a single oral dose at 1, 10, 30 or 100 ␮mol/kg (10 ml/kg). To evaluate their capacity to induce HO equimolar doses of arsenate (128 ␮mol/kg) or arsenite (128 ␮mol/kg) and DMA (5220 ␮mol/kg) were also administered orally to mice. Control animals were gavaged with HPLC grade water. To evaluate total flux of arsenite through metabolic pathways, separate groups of mice (n = 4 cages/dose level) received a single oral dose of arsenite (0, 1, 10, 30 or 100 ␮mol/kg) and were placed in Nalgene metabolism cages (Nalge Co., Rochester, NY). Mice were adapted to the metabolism cages for 72 h prior to dosing. Mice were housed three per cage to accumulate enough urine for analysis of metabolites over a 0–24 h period. Urine was collected over ice and stored at −70 ◦ C until analysis. The animals were maintained on the same diets with water ad libitum while in the metabolism cages. Mice were euthanized by cardiac puncture while under CO2 anesthesia at one to 72 h after dosing. Livers, lungs and kidneys to be assayed for HO activity were removed and immediately placed in ice-cold 0.1 M potassium phosphate buffer (pH 7.4). Post-mitochondrial supernatants were prepared and stored frozen at −70 ◦ C until HO activities were assayed. Blood, liver, lung and kidney for speciated arsenic analysis were flash frozen and stored at −70 ◦ C until analyzed. Tissues were analyzed for iAs and metabolites as described in Kitchin et al. (1999). Whole blood, liver, kidney, and urine were digested in 2 M HCl; lung was digested in 2 M phosphoric acid. The digested tissue samples were assayed using a method based on the generation of volatile arsines, followed by chromatographic separation and final detection of these species by hydride generation atomic absorption spectrophotometry (HGAAS) (Crecelius et al., 1986; Del Razo et al., 2001). This particular method generates arsines from both the tri and pentavalent arsenicals, thus it does not distinguish between trivalent and pentavalent forms. Total arsenic was determined by HGAAS, using a Perkin-Elmer 5100 (Norwalk, CT) equipped with a FIAS-200 flow injection atomic spectroscopy system. The biological samples were completely wet digested with sequential addition of nitric, sulfuric, and

391

perchloric acids (Cox, 1980). This procedure converts all arsenicals to iAs for HGAAS determination. The total arsenic analysis was done using an arsenic electrodeless discharge lamp at 197.3 nm in a heated quartz cell. Working stock solutions containing 1 mg arsenic per milliliter were prepared daily. Quality control for total arsenic determinations included the analysis of bovine liver standard (SRM1577b) concurrently with tissues samples; we attained an accuracy of 103–105% and a 3–12% coefficient of variation. The reliability of arsenic species separation procedures was assessed by spiking control liver samples with known amounts of iAs, MMA, and DMA (0.25, 0.25 and 0.50 ␮g/g, respectively). Recoveries ranged from 89 to 111% with coefficients of variation between 3 and 11%. The estimated limit of detection was 0.012, 0.012 and 0.025 ␮g/g of iAs, MMA and DMA, respectively. The experimental animal tissues did not contain any trimethylated arsenicals such as arsenobetaine (a seafood constituent). Therefore, a comparison of tissues concentration estimated by total arsenic analysis and by summation of arsenic species was used to verify the accuracy of the analytical data. HO activity in post-mitochondrial supernatants was determined by the NADPH-dependent oxidation of heme (complexed to horse albumin) to biliverdin and then of biliverdin to bilirubin (via biliverdin reductase) in 0.1 M potassium phosphate buffer (pH 7.4) as described by Tenhunen et al. (1969) and Maines et al. (1977) and modified by Kitchin et al. (1999). Statistical analysis employed analysis of variance using the SAS System for Windows, version 8.02 (SAS Institute, Cary, NC). HO response data was log transformed because of non-homogeneity of variance. Statistically significant differences were evaluated using Dunnett’s test to compare treated groups with the control group (Steel and Torrie, 1980). A value of P < 0.05 was considered a statistically significant effect for two-tailed comparisons. Benchmark dose (BMD) estimates were determined for alterations in HO in liver and kidney of mice and rats (data previously published from Kitchin et al., 1999) using US EPA Benchmark Dose Software (BMDS Version 1.4). The power model was chosen to fit these continuous data according to the following equation: Y [x] = control + slope × xpower

392

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

where Y is the response; x is the dose and power is restricted to be greater than or equal to 1 and variance is modeled as non-homogeneous. The benchmark effect level was set at 100% (i.e. doubling response over control). The BMDL (lower bound confidence limit) was calculated as the 95% lower confidence interval. 3. Results 3.1. Tissue specificity and time course for HO induction The potencies of As(III), arsenate and DMA to induce HO are compared in Table 1. On a molar basis, arsenite is 1.5- and 2-fold more potent than arsenate as an inducer of HO in liver and kidney, respectively. Mice dosed with DMA (5280 ␮mol/kg) do not exhibit HO induction in liver, lung or kidney. HO activity was not induced in lung following administration of either arsenate or arsenite (data not shown). The time course for HO induction in liver and kidney following a single oral dose of 100 ␮mol/kg sodium arsenite is shown in Fig. 1. In the liver HO induction was maximal between 4 and 8 h and returned to control values by 24–72 h post dosing; maximum induction was ∼7.5-fold over controls on average. The time course for HO induction in kidney was more variable than liver and As(III) was also a less potent inducer of HO in kidney with maximal induction being ∼6-fold over control. In both liver and kidney, HO was significantly elevated over control values at 2, 4 and 8 h post dosing. The time course for iAs, MMA and DMA in blood, liver, lung and kidney following a single oral dose of 100 ␮mol/kg sodium arsenite is shown in Fig. 2. The analyses comparing the sum of species (iAs + MMA + DMA) and total arsenic in blood, liver, lung and kidney of arsenite-treated mice indicated that iAs and Table 1 Comparison of heme oxygenase activity in mouse liver and kidney 6 h after oral administration of arsenate, arsenite and DMA Arsenical and dose (␮mol/kg)

Control Arsenate (128 ␮mol/kg) Arsenite (128 ␮mol/kg) DMA (5280 ␮mol/kg) a

HO activity (␮mol/g/h)a Liver

Kidney

0.079 ± 0.027 0.40 ± 0.036 0.59 ± 0.13 0.076 ± 0.043

0.059 ± 0.021 0.19 ± 0.017 0.45 ± 0.13 0.061 ± 0.017

Figures are mean ± S.E. (n = 6).

its methylated metabolites accounted for all of the total arsenic present in these tissues (data not shown). The concentrations of iAs in liver, kidney, lung, and blood peaked within 1 h after a single oral dose of 100 ␮mol/kg sodium arsenite. The peak concentration of iAs was greatest in liver, followed by kidney, lung and blood. Peak concentration of iAs in liver was ∼3fold greater than kidney (Fig. 2A). Liver iAs decreased quickly from 2 to 4 h, and then remained constant from 4 to 24 h at concentrations 3-fold above control values. This is in contrast to lung, kidney and blood in which iAs levels decreased to levels only slightly greater than controls after 4 h. In contrast to iAs, tissue levels of MMA achieved were much lower (on the order of 10fold) and the time course was also more complex, particularly in liver (Fig. 2B). The biological significance, if any, of the apparent rebound of MMA concentration in liver is unclear, but it should be noted that the overall levels of MMA are very low compared to all other arsenicals. MMA peaks later in kidney than in other tissues probably due to its being eliminated primarily in the urine. Kidney achieved the highest concentrations of both MMA (Fig. 2B) and DMA (Fig. 2C) reflecting the rapid metabolism and urinary excretion of these methylated metabolites. The concentrations of DMA achieved and maintained in tissues were also low compared to iAs (Fig. 2C). However, except in liver, DMA levels equaled or exceeded iAs levels at time points longer than 2 h. This again demonstrates the rapidity of iAs methylation. Of all tissues examined, blood tended to have the lowest concentrations of iAs, MMA and DMA at all sampling times, which reflects the rapid tissue uptake of arsenite. Arsenic is excreted in urine of control and arsenitetreated mice (Table 2). The amount of total arsenic excreted increased with increased administered dose of arsenite. The sum of species (iAs + MMA + DMA) excreted in urine accounted for all of the arsenic detected in urine. DMA was the predominant arsenical excreted in control and arsenite-treated animals. However, as the amount of arsenite administered increased, the percentage of DMA excreted decreased. There was a concomitant increase in the percentage of iAs and MMA excreted with increasing dose of arsenite. This alteration in the percentage of the arsenicals excreted with increased dose of administered arsenite suggests that the methylation of arsenite and MMA was either inhibited or saturated.

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

393

Fig. 1. Time course for HO induction in mouse liver (A) and kidney (B) following a single oral dose of sodium arsenite (100 ␮mol/kg). Data are mean ± S.E. (n = 4–8 mice/time point).

3.2. Dose response for HO induction Data examining the correlation of HO induction in liver with administered dose, tissue levels of iAs 6 h after arsenite administration, and cumulative 24-h excretion of inorganic arsenic are shown in Fig. 3(A–C), respectively. Four (data not shown) or 6 h after As(III) administration, HO activity was significantly elevated in liver at doses of 30 and 100 ␮mol/kg. Doses of 1 and 10 ␮mol/kg did not lead to statistically significant increases in HO activity in liver at either 4 (data not shown) or 6 h post As(III) administration. The corre-

lation of hepatic HO activity with either hepatic iAs concentration (Fig. 3B) or cumulative 24 h urinary excretion of iAs (Fig. 3C) followed a similar pattern to that observed for administered dose and HO activity (Fig. 3A). Data examining the correlation of HO induction in kidney with administered dose, tissue levels of iAs 4 h after As(III) administration, and cumulative 24 h excretion of inorganic arsenic are shown in Fig. 4(A–C), respectively. At 4 but not 6 h (data not shown) after arsenite administration, HO activity was significantly elevated in kidney at the highest dose of 100 ␮mol/kg.

394

Table 2 Twenty-four hours cumulative amount in ␮g of arsenicals (and % distribution) excreted in urine of mice orally administered arsenite Dose (␮mol/kg)

iAs (␮g) (% distribution)

MMA (␮g) (% distribution)

Control

0.046 ± 0.013 (2.75± 0.87)

0.011 ± 0.002 (0.66± 0.09)

DMA (␮g) (% distribution) 1.649 ± 0.257 (96.25± 0.94)

Sum of species (␮g) 1.706 ± 0.254

Total As (␮g) 1.914 ± 0.211

1

0.127 ± 0.065 (3.51± 1.14)

0.037 ± 0.009 (1.06± 0.19)

3.292 ± 0.669 (95.43± 1.22)

3.510 ± 0.669

3.510 ± 1.139

10

1.423 ± 0.141 (5.95± 0.36)

0.486 ± 0.155 (2.02± 0.60)

22.00 ± 1.31 (92.03± 0.51)

23.91 ± 1.51

23.89 ± 1.43

30

6.624 ± 1.695 (10.88± 1.86)

2.326 ± 0.298 (3.86± 0.29)

51.51 ± 6.31 (85.37± 2.06)

60.46 ± 7.79

61.19 ± 8.60

100

21.75 ± 6.37 (14.78± 1.75)

9.25 ± 2.21 (6.41± 1.43)

114.09 ± 17.49 (78.81± 1.61)

145.09 ± 24.84

145.22 ± 25.26

Doses of 1, 10 and 30 ␮mol/kg did not lead to statistically significant increases in HO activity at either 4 or 6 h (data not shown) post arsenite administration. The correlation of renal HO activity with either renal iAs concentration (Fig. 4B) or cumulative 24 h urinary

Fig. 2. Tissue concentration time course for iAs (A), MMA (B) and DMA (C) following a single oral dose of sodium arsenite (100 ␮mol/kg). Data are mean ± S.E. (n = 4 mice/time point).

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

Data represents mean ± S.D., n = 4.

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

excretion of iAs (Fig. 4C) followed a similar pattern to that observed for administered dose and HO activity (Fig. 4A). No observed effect levels (NOELs) and model estimates for BMDs and BMDLs are shown in Table 3 for both mice and rats. While the NOELs are the same for liver and kidney for both species, examination of the BMD estimates reveals that As(III) is clearly ∼3-fold more potent at inducing HO in rats compared to mice in both liver and kidney. That the BMD is a better estimator of potency compared to the NOEL is not surprising considering that BMD analysis utilizes data from the entire dose-response curve.

395

4. Discussion Consistent with the results reported by Brown and Kitchin (1996) in rats, we found that arsenite was more potent than arsenate as an inducer of HO in mice. Similarly, orally administered dimethylarsinic acid did not induce HO in mice. Taken together, these results are consistent with our hypothesis that arsenite (among arsenicals tested to date) is a potent, specific inducer of HO and has utility as a biomarker of effect. Under this hypothesis one would also expect arsenate to be a less potent inducer of HO compared to arsenite because it must be metabolized to arsenite; and lower levels

Fig. 3. Correlation of mouse hepatic HO activity at 6 h post exposure with administered arsenite dose (A), hepatic iAs concentration (B), and cumulative urinary inorganic arsenic excretion (C). Data are mean ± S.E. (n = 4–8 mice/group).

396

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

Fig. 3. (Continued).

of arsenic are achieved in target tissues (liver, kidney) following oral administration of equimolar doses of arsenate compared to arsenite in mice (Kenyon et al., 2003). Tissue-specific differences are observed in HO induction among species. In our studies, As(III) induced HO activity in both mouse liver and kidney; induction was greater in liver than in kidney. Similar tissue specificity has been found in rats (Kitchin et al., 1999). However, on an equimolar dose basis, HO induction is higher in rat liver and kidney than in mouse tissues. For example, at a dose of 100 ␮mol/kg, HO induction over controls is over 4-fold higher (400%) in rat liver compared to mouse liver and 75% higher in rat kidney compared to mouse kidney. A possible explanation for differences in tissue responsiveness to HO induction between rats and mice is that peak levels of inorganic arsenic in liver and kidney are maintained 2–8 h longer Table 3 BMD and BMDL values for heme oxygenase induction in liver and kidney of mice and rats Species

Organ

NOEL

BMD

BMDL

Mouse

Liver Kidney

10 30

30.9 62.1

20.8 40.5

Rata

Liver Kidney

10 30

11.9 18.9

4.95 9.34

All values are in ␮mol/kg. The NOEL is the highest dose studied at which statistical significance was not observed. a Rat HO dose response data are from Kitchin et al., 1999.

in rats compared to mice. Notably, higher peak levels of inorganic arsenic are attained in mice liver compared to other tissues. In a previous study with rats administered an equimolar dose of arsenite (Kitchin et al., 1999), both liver and kidney accumulated and retained arsenic over a 24–72 h period, mainly as DMA. Rat red blood cells accumulate DMA (Vahter et al., 1984) and because the livers used in the Kitchin et al. (1999) study were not perfused, accumulation and retention of arsenic loaded erythrocytes in the vascular bed may have produced erroneous estimates of tissue arsenic content. Inorganic arsenic was the predominant arsenical in the liver of mice administered arsenite. In mouse kidney, iAs was initially high, but then decreased to similar to levels of DMA for the remaining time points. MMA was always lower in mouse kidney, unlike that found in rat kidney (Kitchin et al., 1999). This species difference in tissue disposition of arsenicals between the mouse and rat demonstrates the importance of knowing whether or not the animal model has unique biological characteristics. Lung was not responsive to HO induction in mice exposed orally to arsenite in this study or in rats (Brown and Kitchin, 1996). However, Falkner et al. (1993) reported a 2.2-fold induction of guinea pig lung HO activity over controls 24 h after a single s.c. injection of 75 ␮mol/kg arsenite. In this same study, liver and kidney HO activities were also induced 2.5- and 3.4-fold over controls. An interesting interspecies difference in metabolism that may be relevant in this context is that

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

guinea pigs exhibit minimal to no methylation of inorganic arsenic (Healy et al., 1997; Csanaky and Gregus, 2002). Since methylation of iAs usually facilitates excretion (Gebel, 2002), this would be expected to result in higher tissue concentrations of iAs in species deficient in arsenic methylation; however this has not been investigated in the guinea pig. The time course data for HO induction in this paper is unusual in that we used the oral route of administration and measured comparatively early time points. In many published reports of HO induction in animal models, the measured hepatic or renal enzyme activ-

397

ity was 12–24 h after i.p. or s.c. administration of arsenite (Sunderman, 1987). We observed substantial increases in mouse hepatic and renal iAs concentrations that peaked 1 h after oral administration of arsenite, as well as increased HO activity from 2 to 8 h post dosing. These results are qualitatively similar to previous results in rats (Kitchin et al., 1999). However, tissue levels of inorganic arsenic peaked lower, but remained elevated much longer, in rats compared to mice (Kitchin et al., 1999). This difference could be a result of As(III) binding to arsenic-specific binding proteins (Bogdan et al., 1994; Styblo and Thomas, 1997) or

Fig. 4. Correlation of mouse renal HO activity at 4 h post exposure with administered arsenite dose (A), renal iAs concentration (B), and cumulative urinary inorganic arsenic excretion (C). Data are mean ± S.E. (n = 4–8 mice/group).

398

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

Fig. 4. (Continued).

saturation or inhibition of arsenic methylation. The relatively greater responsiveness of rat liver HO induction combined with less variability in response in comparison to mice is probably the reason that the rat has been historically preferred for mechanistic investigations of HO induction. However, the greater similarity in pharmacokinetic behavior of arsenic between mice and humans compared to rats and humans provides a strong incentive for the use of the mouse model. The doses used in this study ranged from those predicted to be below the threshold of biological response to doses as high as mice can tolerate without overt signs of toxicity (a range of 1–100 ␮mol/kg). No signs of toxicity were observed in the mice in this study, e.g. coat condition remained good, and appetite and activity levels remained normal out to 72 h post dosing. To reach statistical significance levels of P < 0.05 for HO induction, a dose of 30 ␮mol/kg of As(III) was required for liver and a dose of 100 ␮mol/kg of As(III) for kidney in the mouse. Thus, the lowest observed response level (LOEL) and no observed response level (NOEL) in this study for hepatic HO induction are 30 and 10 ␮mol/kg, respectively. Whereas for renal HO induction the LOEL is 100 ␮mol/kg and the NOEL is 30 ␮mol/kg. At the LOEL dose in liver (30 ␮mol/kg) and kidney (100 ␮mol/kg), the average concentrations of exogenous (above background) iAs present in the liver and kidney are 0.44 and 0.31 ␮g/g, respectively. The endogenous hepatic and renal concentrations of iAs in our concurrent dose-response experimental an-

imals were 0.026 and 0.027 ␮g/g, respectively. Thus, elevations of 17 (for liver) and 11 (for kidney) times background iAs concentrations were associated with a biological effect–HO enzyme induction – if all measurements are compared at time of peak HO induction (4–6 h) as was done in our earlier rat study (Kitchin et al., 1999). However, because peak concentrations of both hepatic and renal iAs occur several hours before the peak levels of HO activity, it is likely that the signal must be transduced to increase HO activity. The urinary arsenical profile in the mice administered arsenite was unlike the arsenical profile in the tissues. DMA was predominant species of arsenic (>75%) in urine, while except in liver, it had a similar tissue level of iAs. A similar result, high DMA in urine, and a more equal distribution of arsenicals among tissues was observed in mice administered arsenate (Hughes et al., 2003). Given the different toxicological potency of the arsenicals (both inorganic and organic in trivalent and pentavalent forms), assuming that the urinary profile of arsenicals is reflective of tissue arsenical profiles may be misleading and is not recommended. However, in the matter of correlation of different dose metrics (e.g. tissue levels of iAs and cumulative urinary excretion of iAs) with the HO response, urinary and tissue iAs have similar dose response relationships for liver and kidney although the shape of the curves differs for these two tissues. This suggests that urinary iAs may be a reasonable biomarker of exposure for HO induction.

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

In the context of risk assessment and utility of HO induction as a biomarker of effect, it is important to consider two issues: (1) the dose-response relationship in animals relative to expected human exposures and (2) the relationship of HO induction to other toxic endpoints observed in humans and rodent models. The highest doses of arsenite (100 ␮mol/kg = 7.3 mg/kg as As) and DMA (5220 ␮mol/kg = 389 mg/kg as As) used in this study were less than reported LD50s and did not elicit observable toxicity. Note that the oral LD50 of arsenic trioxide, which breaks down to As(III) (or at least trivalent inorganic arsenic) is 26 mg As per kilogram (Kaise et al., 1985) which is ∼3x higher than our dose. According to Kaise et al. (1989), the oral LD 50 for DMA is 648 mg As per kilogram, which is ∼1.7fold higher than the dose used in our study. While these are high exposures when considered in the context of chronic exposures in endemic areas of arsenicosis, rodents and humans differ in their sensitivity to arsenic. Specifically, rodents are considered ∼10-fold less sensitive to the acute toxic effects of inorganic trivalent arsenic compared to humans (Yip and Dart, 2001). In addition, it is well recognized that rodents in general are apparently less sensitive to the carcinogenic effects of arsenic compared to humans (Wang et al., 2002; Rossman, 2003). Thus, while the doses are high, they are still relevant given that they did not induce overt toxicity in rodents and that humans in high exposure areas experience chronic toxicities due to arsenic exposure. HO is the rate-limiting enzyme in heme degradation and has a central role in maintaining intracellular heme pools (Maines, 1984). Arsenic exposure can alter hepatic and renal activities of a number of other enzymes in the heme pathway in mice and this can result in altered urinary excretion of porphyrins in mice exposed to arsenic in drinking water (20 ppm As(III)) for periods of up to 6 weeks (Garcia-Vargas et al., 1995). Altered urinary porphyrin excretion has also been reported in humans chronically exposed to arsenic in drinking water in Mexico (Garcia-Vargas et al., 1994). Induction of HO is considered a generalized response to oxidative stress (Maines, 1988) and oxidative stress may be an important mechanism for arsenic-induced carcinogenicity (Kitchin and Ahmad, 2003). All of the foregoing supports the concept that HO induction is a human-relevant response, suitable for use to develop a biologically-based dose-response

399

model in rodents that can ultimately be extrapolated to humans. Induction of HO in human lymphocytes has also been suggested as a biomarker of arsenic effect in exposed human populations (Menzel et al., 1998) and perhaps adult and neonatal human platelets could also be used since they contain HO activity (Nowell et al., 1998). In conclusion, HO induction is a time- and dosedependent response to arsenite exposure that is positively correlated with iAs concentration in both liver and kidney tissue and cumulative urinary excretion of iAs. The data from this study is consistent with the hypothesis that the cellular concentration of inorganic arsenic is the primary determinant for the induction of HO in liver and kidney. Because the peak concentration of hepatic inorganic arsenic occurs several hours before the peak level of hepatic HO activity, our findings are consistent with other data indicating that the signal must be transduced to increase HO activity. HO induction has potential for use as a biomarker of effect and can be used as a tool to further the development of a biologically-based dose response model for As.

Acknowledgments Dr. Luz Maria Del Razo was supported in part by a CONACYT-Mexico fellowship. We thank Brenda Edwards, Carol Mitchell and Karen Herbin-Davis of the US EPA for their assistance with animal work and Bill Anderson and Janice Brown for their assistance with the heme oxygenase assay. We are also very grateful to Drs. Woody Setzer and Jerry Highfill for advice and assistance with the statistical analysis.

References Albores, A., Cebrian, M.E., Bach, P.H., Connelly, J.C., Hinton, R.H., Bridges, J.W., 1989. Sodium arsenite induced alterations in bilirubin excretion and heme metabolism. J. Biochem. Toxicol. 4, 73–78. Bogdan, G.M., Sampayo-Reyes, A., Aposhian, H.V., 1994. Arsenic binding proteins of mammalian systems: I. Isolation of three arsenite-binding proteins of rabbit liver. Toxicology 93, 175–193. Brown, J.L., Kitchin, K.T., 1996. Arsenite, but not cadmium, induces ornithine decarboxylase activity in rat liver: relevance to arsenic carcinogenesis. Cancer Lett. 98, 227–231.

400

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401

Cebrian, M.J., Albores, A., Connelly, J.C., Bridges, J.W., 1988. Assessment of arsenic effects on cytosolic heme status using tryptophan pyrrolase as an index. J. Biochem. Toxicol. 3, 77– 86. Cox, D., 1980. Arsine evolution-electrothermal atomic absorption method for the determination of nanogram levels of total arsenic in urine and water. J. Anal. Tox. 4, 207–211. Crecelius, E.A., Bloom, N.S., Cowan, C.E., Jenne, E.A., 1986. Determination of arsenic species in limnological samples by hydride generation atomic absorption spectroscopy. In: Electric Power Research Institute Ed., Speciation of selenium and arsenic in natural waters and sediments, volume 2: Arsenic speciation. Palo Alto, California, EA-4641 (Project 2020-2), pp. 1– 28. Csanaky, I., Gregus, Z., 2002. Species variations in the biliary and urinary excretion of arsenate, arsenite and their metabolites. Comp. Biochem. Physiol. C Toxicol. Pharmacol. 131, 355– 365. Del Razo, L.M., Styblo, M., Cullen, W.R., Thomas, D.J., 2001. Determination of trivalent methylated arsenicals in biological matrices. Toxicol. Appl. Pharmacol. 174, 282– 293. Falkner, K.C., McCallum, G.P., Cherian, M.G., Bend, J.R., 1993. Effects of acute sodium arsenite administration on the pulmonary chemical metabolizing enzymes, cytochrome P-450 monooxygenase NAD(P)H:quinone acceptor oxidoreductase and glutathione S-transferase in guinea pig: comparison with effects in liver and kidney. Chem. Biol. Interact. 86, 51–68. Garcia-Vargas, G., Del Razo, L.M., Cebrian, M.E., Albores, A., Ostrosky-Wegman, P., Montero, R., Gonsebatt, M.E., Lim, C.K., DeMatteis, F., 1994. Altered urinary prophyrin excretion in a human population chronically exposed to arsenic in Mexico. Human Expt. Toxicol. 13, 839–847. Garcia-Vargas, G., Cebrian, M.E., Albores, A., Lim, C.K., DeMatteis, F., 1995. Time-dependent porphyric response in mice subchronically exposed to arsenic. Human Expt. Toxicol. 14, 475–483. Gebel, T.W., 2002. Arsenic methylation is a process of detoxification through accelerated excretion. Int. J. Hyg. Environ. Health 205, 505–508. Guzzo, A., Karatzios, C., Diorio, C., Dubow, M.S., 1994. Metallothionein-II and ferritin H mRNA levels are increased in arsenite-exposed HeLA cells. Biochem. Biophys. Res. Commun. 205, 590–595. Healy, S.M., Zakharyan, R.A., Aposhian, H.V., 1997. Enzymatic methylation of arsenic compounds: IV. In vitro and in vivo deficiency of the methylation of arsenite and monomethylarsonic acid in the guinea pig. Mutat. Res. 386, 229– 239. Hughes, M.F., Kenyon, E.M., Edwards, B.C., Mitchell, C.T., Del Razo, L.M., Thomas, D.J., 2003. Accumulation and metabolism of arsenic in mice after repeated oral administration of arsenate. Toxicol. Appl. Pharmacol. 191, 202–210. Hughes, M.F., 2002. Arsenic toxicity and potential mechanisms of action. Toxicol. Lett. 133, 1–16. Kaise, T., Watanabe, S., Itoh, K., 1985. The acute toxicity of arsenobetaine. Chemosphere 14, 1327–1332.

Kaise, T., Yamauchi, H., Horiguchi, Y., Tani, T., Watanabe, S., Hirayama, T., Fukui, S., 1989. A comparative study on acute toxicity of methylarsonic acid, dimethylarsinic acid and trimethylarsine oxide in mice. Appl. Organomet. Chem. 3, 273– 277. Kenyon, E.M., Del Razo, L.M., Hughes, M.F., 2003. Comparative tissue distribution and urinary excretion of inorganic arsenic (iAs) and its methylated metabolites in mice following oral administration of arsenate (AsV) and arsenite (AsIII). Toxicologist Suppl. Tox. Sci. 72 (Suppl. 1), 717. Kitchin, K.T., 2001. Recent advances in arsenic carcinogenesis: modes of action, animal model systems and methylated arsenic metabolites. Toxicol. Appl. Pharmacol. 172, 249– 261. Kitchin, K.T., Del Razo, L.M., Brown, J.L., Anderson, W.L., Kenyon, E.M., 1999. An integrated pharmacokinetic and pharmacodynamic study of arsenite action. 1. Heme oxygenase induction in rats. Teratogen. Carcinogen. Mutagen. 19, 385– 402. Kitchin, K.T., Ahmad, S., 2003. Oxidative stress as a possible mode of action for arsenic carcinogenesis. Toxicol. Lett. 137, 3– 13. Maines, M.D., 1997. The heme oxygenase system: a regulator of second messenger gases. Ann. Rev. Pharmacol. Toxicol. 37, 517–554. Maines, M.D., 1988. Heme oxygenase: function, multiplicity, regulatory mechanisms, and clinical applications. FASEB J. 2, 2557–2568. Maines, M.D., 1984. New developments in the regulation of heme metabolism and their implications. CRC Crit. Rev. Toxicol. 12, 241–314. Maines, M., Ibrahim, N., Kappas, A., 1977. Solubilization and partial purification of heme oxygenase from rat liver. J. Biol. Chem. 252, 5900–5903. Menzel, D.B., Rasmussen, R.E., Lee, E., Meacher, D.M., Said, B., Hamadeh, H., Vargas, M., Green, H., Roth, R.N., 1998. Human lymphocyte heme oxygenase I as a response biomarker to inorganic arsenic. Biochem. Biophys. Res. Comm. 250, 653– 656. National Research Council (NRC), 2001. Arsenic in Drinking Water–2001 Update. National Academy Press, Washington, DC. National Research Council (NRC), 1999. Arsenic in Drinking Water. National Academy Press, Washington, DC, 310. Nowell, S.A., Leakely, J.E.A., Warren, J.F., Lang, N.P., Frame, L.T., 1998. Identification of enzymes responsible for the metabolism of heme in human platelets. J. Biol. Chem. 273, 33342–33346. Rossman, T.G., 2003. Mechanisms of arsenic carcinogenesis: an integrated approach. Mutat. Res. 533, 37–65. Sardana, M.K., Drummond, G.S., Sassa, S., Kappas, A., 1981. The potent heme oxygenase inducing action of arsenic and parasiticidal arsenicals. Pharmacology 23, 247–253. Steel, R.G.D., Torrie, J.H., 1980. Principles and Procedures of Statistics—A Biometrical Approach, second ed. McGraw-Hill, NY. Styblo, M., Thomas, D.J., 1997. Binding of arsenicals to proteins in an in vitro methylation system. Toxicol. Appl. Pharmacol. 147, 1–8.

E.M. Kenyon et al. / Toxicology 206 (2005) 389–401 Sunderman, W., 1987. Metal induction of heme oxygenase. Ann. NY Acad. Sci. 514, 65–80. Tenhunen, R., Marver, H., Schmid, R., 1969. Microsomal heme oxygenase-characterization of the enzyme. J. Biol. Chem. 244, 6388–6394. Vahter, M., Marafante, E., Dencker, L., 1984. Tissue distribution and retention of 74As-dimethylarsinic acid. Arch. Environ. Contam. Toxicol. 13, 259–264.

401

Wang, J.P., Qi, L., Moore, M.R., Ng, J.C., 2002. A review of animal models for the study of arsenic carcinogenesis. Toxicol. Lett. 133, 17–32. Yip, L., Dart, R.C., 2001. Arsenic. In: Sullivan, J.B., Krieger, G.R. (Eds.), Clinical and Environmental Health and Toxic Exposures. Williams and Wilkins, Philadelphia, pp. 858– 866.