|Home | About | Journals | Submit | Contact Us | Français|
In National Toxicology Program 2-year studies, hexavalent chromium [Cr(VI)] administered in drinking water was clearly carcinogenic in male and female rats and mice, resulting in small intestine epithelial neoplasms in mice at a dose equivalent to or within an order of magnitude of human doses that could result from consumption of chromium-contaminated drinking water, assuming that dose scales by body weight3/4 (body weight raised to the 3/4 power). In contrast, exposure to trivalent chromium [Cr(III)] at much higher concentrations may have been carcinogenic in male rats but was not carcinogenic in mice or female rats. As part of these studies, total chromium was measured in tissues and excreta of additional groups of male rats and female mice. These data were used to infer the uptake and distribution of Cr(VI) because Cr(VI) is reduced to Cr(III) in vivo, and no methods are available to speciate tissue chromium. Comparable external doses resulted in much higher tissue chromium concentrations following exposure to Cr(VI) compared with Cr(III), indicating that a portion of the Cr(VI) escaped gastric reduction and was distributed systemically. Linear or supralinear dose responses of total chromium in tissues were observed following exposure to Cr(VI), indicating that these exposures did not saturate gastric reduction capacity. When Cr(VI) exposure was normalized to ingested dose, chromium concentrations in the liver and glandular stomach were higher in mice, whereas kidney concentrations were higher in rats. In vitro studies demonstrated that Cr(VI), but not Cr(III), is a substrate of the sodium/sulfate cotransporter, providing a partial explanation for the greater absorption of Cr(VI).
Chromium (Cr) exists in multiple oxidation states. The hexavalent chromium [Cr(VI)] and trivalent chromium [Cr(III)] states are most important from a biological and an industrial standpoint. Cr(VI) is an industrial contaminant of water and soil and an established human lung carcinogen following inhalation exposure (Cohen et al., 1993; International Agency for Research on Cancer (IARC), 1990; National Toxicology Program [NTP] 1998). Sodium dichromate dihydrate (SDD) is the most water-soluble salt of Cr(VI). In contrast, Cr(III) is an essential element and a natural dietary constituent (Anderson, 1989; National Institutes of Health (NIH), 2007) and is also widely ingested in dietary supplements, most notably chromium picolinate monohydrate (CPM). CPM is an organic complex, with Cr(III) chelated to three molecules of picolinic acid to increase Cr(III) absorption. However, studies have shown that there is significant separation of the Cr(III) from the picolinic acid prior to absorption (Hepburn and Vincent, 2002; NTP, 2008b) and that the absorption of chromium(III) chloride, picolinate, and nicotinate is similar (<1% after 12 h) (Olin et al., 1994). Cr(III) is not classifiable as to its human carcinogenicity (IARC, 1990).
The NTP conducted 2-year toxicity and carcinogenicity studies of Cr(VI), as SDD administered in drinking water (NTP, 2008a; Stout et al., 2009a), and Cr(III), as CPM administered in feed (NTP, 2008b; Stout et al., 2009b). Cr(VI) was carcinogenic in male and female rats and mice, inducing squamous neoplasms of the oral cavity in rats and epithelial neoplasms of the small intestine in mice. In addition, microcytic hypochromic anemia in rats, erythrocyte microcytosis in mice, and increased incidences of histiocytic cellular infiltration in various tissues of rats and mice were observed. In contrast, at much higher doses, Cr(III) was not carcinogenic in female rats or in mice. Increased preputial gland adenomas in male rats may have been related to exposure. Cr(III) exposure did not have any significant effect on survival, body weight, feed consumption, or non-neoplastic lesions. This disparity in toxicity and carcinogenicity is consistent with previous studies (Agency for Toxic Substances and Disease Registry, 2000).
Although the absorption of Cr(VI) and Cr(III) is low following oral exposure, Cr(VI) is absorbed more efficiently than Cr(III) (Donaldson and Barreras, 1966; Fébel et al., 2001; Kerger et al., 1996; Mackenzie et al., 1959). This is thought to occur because Cr(VI) as chromate (CrO42−) structurally resembles sulfate and phosphate and is taken up by all cells and organs throughout the body through sulfate transporters (Costa, 1997), whereas Cr(III) is not a substrate for transport (Proctor et al., 2002) but is thought to enter cells via diffusion or phagocytosis. Consistent with this mechanism of transport, higher chromium tissue concentrations were observed with Cr(VI) compared with Cr(III) following equivalent exposures in drinking water (Costa, 1997; Costa and Klein, 2006; Mackenzie et al., 1958).
Both extracellular and intracellular reduction of Cr(VI) to Cr(III) occur. As a result of the lower bioavailability of Cr(III), extracellular reduction, primarily in the stomach, has been suggested to be protective against the toxic and carcinogenic effects of Cr(VI) following oral exposure (De Flora, 2000; De Flora et al., 1997, 1989; Paustenbach et al., 2003; Proctor et al., 2002). In contrast, intracellular reduction is thought to be a mechanism of carcinogenesis because of DNA damage that occurs when Cr(VI) is reduced through Cr(V) and Cr(IV) to Cr(III) (reviewed in O′Brien et al., 2003). Because of its large redox potential, Cr(III) is not expected to oxidize to Cr(VI) in vivo. In summary, absorption and retention of chromium depend on a number of factors: the rate of reduction of Cr(VI) to Cr(III) outside the cells, the pH of the milieu, the rate of transport of Cr(VI) into the cells, the rate of reduction of Cr(VI) to Cr(III) inside the cells, and the rate of diffusion of Cr(III) from the cells (NTP, 2008a).
As part of the NTP chronic oral studies of Cr(VI) (NTP, 2008a; Stout et al., 2009a) and Cr(III) (NTP, 2008b; Stout et al., 2009b), total Cr was measured in selected tissues and excreta of additional groups of male rats and female mice at selected time points. The objective of these studies was to determine the effect of administered species [Cr(VI) or Cr(III)] on Cr uptake and tissue distribution. Because Cr(VI) is reduced to Cr(III) in vivo, and current analytical procedures do not allow for the speciation of Cr extracted from biological samples, speciation of absorbed Cr(VI) was inferred by comparing tissue uptake of Cr following exposure to Cr(VI) or Cr(III). These are the first studies that provide a comprehensive comparison of chronic toxicity and carcinogenicity with tissue distribution in rats and mice following exposure to Cr(VI) and Cr(III) and may ultimately aid in the extrapolation of the Cr(VI) carcinogenicity data to humans.
SDD (CAS 7789-12-0) was obtained from Aldrich Chemical Company (Milwaukee, WI). The purity was determined using differential scanning calorimetry; titration of the dichromate ion with sodium thiosulfate and potassium ferrocyanide, speciation of the Cr ions using liquid chromatography-inductively coupled plasma-mass spectrometry (ICP-MS), and potentiometric titrimetric analysis with sodium thiosulfate. Based on these analyses, the overall purity was ≥ 99.7%. Dose formulations were prepared approximately every 2 weeks by mixing SDD with tap water. Periodic analysis using ultraviolet/visible/near infrared spectroscopy (350–390 nm) confirmed that all dose formulations varied by less than 10% of the target concentrations. CPM (CAS No. 27882-76-4) used in the 2-year studies was a combination of chemical obtained from TCI America (Portland, OR) and from Sigma-Aldrich (St Louis, MO). Purity was determined by elemental analyses, proton-induced x-ray emission spectroscopy, inductively coupled plasma-atomic emission spectroscopy, high-performance liquid chromatography (HPLC) with ultraviolet-visible (UV) or mass spectrometric detection, or ICP-MS. The overall purity of the chemical was ≥ 95%. The dose formulations were prepared monthly by mixing CPM with feed. Homogeneity and stability of the dose formulations were assessed by HPLC with UV detection. Periodic analysis confirmed that all dose formulations were within 10% of the target concentrations. Concentrations of total chromium in vehicles were 0.604 ± 0.253 ppm in feed [Cr(III) acetate was added because Cr(III) is thought to be an essential element] and below 0.005 mg/l in tap water.
The studies were conducted at Southern Research Institute (Birmingham, AL). Male F344/N rats and female B6C3F1 mice were obtained from Taconic Farms (Germantown, NY). Rats and mice were quarantined for 12 days (CPM) or 14 days (SDD) and were 5–6 (CPM) or 6–7 (SDD) weeks old at the beginning of the studies. Animals were distributed randomly into groups of similar mean body weights and identified by tail tattoo. Rats were housed three to a cage. Mice were housed five to a cage. Feed and tap water were available ad libitum. For the SDD study, feed was irradiated NTP-2000 wafers. For the CPM study, feed was irradiated NTP-2000 open formula meal diet. Both diets were obtained from Zeigler Brothers, Inc. (Gardners, PA). Animals were killed by asphyxiation with CO2.
Animal use was in accordance with the United States Public Health Service policy on humane care and use of laboratory animals and the Guide for the Care and Use of Laboratory Animals. These studies were conducted in compliance with the Food and Drug Administration Good Laboratory Practice Regulations (21CFR, Part 58).
As part of the NTP 2-year bioassays of SDD (NTP, 2008a; Stout et al., 2009a) and CPM (NTP, 2008b; Stout et al., 2009b), additional animals were randomly assigned for measurement of total Cr in selected tissues and excreta; these animals were treated the same as the core study animals used for evaluation of toxicity and carcinogenicity with respect to exposure, housing, and handling. SDD was administered in drinking water to groups of 40 male rats and female mice at concentrations of 0, 14.3, 57.3, 172, and 516 mg/l. CPM was administered in feed to groups of 30 male rats and female mice at concentrations of 0, 2000, 10,000, and 50,000 ppm. On days 4, 11 and 180 (CPM and SDD), and on day 369 (SDD only), up to 10 rats and mice per exposure group were removed from treatment and placed in individual metabolism cages for separate collection of urine and feces. Animals were provided undosed drinking water or feed ad libitum during this period to allow unabsorbed chromium to be excreted. Two collections of urine and feces were made to include the intervals from 0 to 24 and 24 to 48 h; measured values were combined to yield the reported 48-h values. The 48-h washout period was based on an elimination half-life of 8–21 h (Bragt and van Dura, 1983; Vanoirbeek et al., 2003). On days 6, 13, and 182 (SDD and CPM), and on day 371 (SDD only), at the end of the 48-h period, the animals were anesthetized with CO2/O2 and blood was taken from the retro-orbital sinus into heparinized tubes. Erythrocytes and plasma were collected separately. Although the animals were still under anesthesia, the abdominal wall was opened and the aorta was severed. The liver, kidneys, and the stomach (separated into glandular stomach and forestomach) were removed, weighed, and frozen at −20°C until further use. Stainless steel was avoided during the tissue collection; only plastic, Teflon, ceramic, or tungsten carbide instruments were used. Because a washout period occurred prior to tissue collection, increased chromium concentrations in plasma should represent the chromium entering plasma from the tissues. To make comparisons of tissue concentrations between animal species and species of chromium on day 182 (week 26), average time-weighted doses of Cr(VI) and Cr(III) resulting from ingestion of SDD or CPM were calculated using the 1- to 25-week water and feed consumption data and body weight data (NTP, 2008b). These doses are shown in Table 1.
The sample preparation and analysis procedures used to determine the concentration of chromium in tissues (Levine et al., 2010) and excreta (Levine et al., 2009) in support of this investigation are presented in detail elsewhere and are summarized here. Neither of these procedures could differentiate between chromium oxidation states, so data reported are for total chromium, irrespective of oxidation state administered.
Each tissue sample was homogenized using a Polytron tissue homogenizer (Brinkmann Instruments, Westbury, NY) and returned to frozen storage. Prior to analysis, homogenized samples were thawed and an aliquot of homogenate equivalent to 0.500 g of tissue was transferred to a tared 50-ml centrifuge tube and the mass recorded. If insufficient sample was available, the entire sample was used. Samples were digested in 5 ml of concentrated nitric acid using a microwave digestion procedure followed by a second microwave digestion step using hydrofluoric acid. Following digestion, the contents of each tube were transferred to a 25-ml volumetric flask to which was added 0.250 ml of internal standard solution (ISS) and then diluted to volume with distilled water.
A 0.500-g aliquot of plasma was transferred to a tared 50-ml centrifuge tube, and the mass was recorded. Plasma samples were digested as described above for tissue samples. Following digestion, the contents of each tube were transferred to a 25-ml volumetric flask to which 0.250 ml of ISS was added and then diluted to volume with distilled water.
Feces samples were transferred to a tared 50-ml centrifuge tube, and the wet mass was recorded. The tubes were transferred to a freezer prior to lyophilization. All tubes were subjected to freeze-drying for at least 24 h. After lyophilization, the dry mass and percent loss on drying was recorded for each sample. Each sample was manually homogenized with a clean glass rod and stored frozen until analysis. Each freeze-dried feces sample was mixed with a Teflon-coated spatula prior to transferring a target mass of freeze-dried feces homogenate, equivalent to 0.250 g wet mass, to a 50-ml centrifuge tube. Urine samples were vortexed prior to transferring a 0.250-g aliquot to a 50-ml centrifuge tube. Excreta samples were digested as described above for tissue samples. Following digestion, the contents of each tube were transferred to a 25-ml volumetric flask to which was added 0.150 ml of ISS and then diluted to volume with distilled water.
Solvent standards were prepared from two working stock solutions by adding appropriate aliquots of each working stock to 100-ml volumetric flasks along with 10 ml of concentrated nitric acid and diluting to volume with deionized water. Matrix standards were prepared in a similar manner, except that aliquots of the working stocks were spiked into 0.500 g of control tissue or feces homogenate and were then carried through the same digestion procedure as the samples.
Samples and standards were analyzed by ICP-MS using a Thermo X7 instrument (ThermoElectron Corp., Winsford, Cheshire, U.K.) or a Plasma Quad XR Instrument (VG Elemental Ltd., Winsford, Cheshire, U.K.) with a concentric nebulizer and a Peltier impact-bead spray chamber cooled to 5°C.
Brush border membrane vesicles (BBMV) were isolated from renal cortex of six male Sprague-Dawley CD rats (~250 g) using a method previously described by Beck and Sacktor (1978). Ca++ precipitation of a crude membrane homogenate was followed by differential centrifugation to yield a purified vesicle preparation. Marker enzyme analysis showed an enrichment of the brush border membrane marker, alkaline phosphatase, of 10- to 20-fold. The basolateral membrane marker, Na+-K+-ATPase, showed enrichments of less than onefold. The final membrane pellet was suspended in vesicle buffer (100mM mannitol, 100mM KCl, 20mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid [HEPES]/tris (hydroxymethyl)aminoethane [Tris], 1mM MgSO4) at pH 7.4 and stored in liquid nitrogen until use.
For transport experiments, BBMV were thawed, centrifuged, resuspended, and allowed to equilibrate in fresh vesicle buffer for 60 min at room temperature and then placed on ice. Ten microliters of vesicles were incubated at room temperature for 2 min in 90 μl of media (100mM NaCl, 20mM Tris/HEPES, and 1 mM MgSO4, pH 7.4) containing 50μM [35S]-sodium sulfate and 50, 250, and 500μM SDD or CPM. This suspension was transferred to Millipore Glass Fiber Filters (0.7 μm) (Billerica, MA), presoaked in distilled water, washed three times with buffer, and allowed to dry. The filters were counted in a Packard 2900TR Tricarb Liquid Scintillation Analyzer (Waltham, MA). Data were transformed and analyzed using GraphPad Prism software (La Jolla, CA), and values represent analysis from triplicate experiments. Internal controls include sodium-free uptake (KCl substituted for NaCl in transport media) and SITS-sensitive uptake with sodium present but no inhibitor (SITS, 1mM 4-acetamido-4′-isothiocyanostilbene-2,2′-disulfonic acid).
For kinetic experiments, BBMV were thawed, centrifuged, resuspended, and allowed to equilibrate in fresh vesicle buffer for 60 min at room temperature and then placed on ice. Ten microliters of vesicles were incubated for 30 s in 90 μl of media (100mM NaCl, 20mM Tris/HEPES, and 1mM MgSO4) containing 5–250 μM [35S]-sodium sulfate and 250μM SDD in the presence or absence of 1mM SITS. This suspension was transferred to Millipore Glass Fiber Filters (0.7 μm), presoaked in distilled water, washed three times with buffer, and allowed to dry. The filters were counted in a Packard 2900TR Tricarb Liquid Scintillation Analyzer. All transport experiments were carried out at room temperature. Data were transformed and analyzed using Microsoft Excel graphics software. The Michaelis-Menten constant (Km) and the maximum rate of uptake (Vmax) were estimated from x- and y-intercepts of the Lineweaver-Burk plots, and values represent SITS-sensitive sodium sulfate transport from three experiments.
Feces, urine, and tissue concentrations typically have skewed distributions and were analyzed using the nonparametric multiple comparison methods of Shirley (1977) and Dunn (1964). The Jonckheere (1954) test was used to assess the significance of dose-response trends and to determine whether a trend-sensitive test (Shirley’s test) was more appropriate for pairwise comparisons with the controls than a test that does not assume a monotonic dose response (Dunn’s test). Measurements less than the experimental limit of quantitation (ELOQ) were included in the statistical analyses as ½ ELOQ. Trend-sensitive tests were used when Jonckheere’s test was significant at p < 0.01. Departures from a linear dose-response relationship for feces, urine, and tissue concentrations were tested using the lack-of-fit test for simple linear regression (Neter et al., 1996). When a significant departure was found, the shape of the dose-response curve was determined by visual examination. For similar administered doses of Cr that resulted from exposure to 516 mg/l SDD and 2000 ppm CPM, concentrations were compared using Mann-Whitney tests (Hollander and Wolfe, 1973). Concentrations were compared between mice and rats using Mann-Whitney tests at each dose level. For sulfate transport experiments, control and treated groups were compared with the two-tailed Student’s t-test (p < 0.05).
Chromium concentration in tissues and excreta are given in Tables 2 and and33 for male rats and female mice, respectively, following exposure to SDD. The 48-h feces collection contained up to 1.9 mg of chromium for rats and 160 μg for mice. In rats, 49.2%, and in mice, 48.8%, of the ingested chromium was excreted in the feces, indicating that the absorption of chromium was low. The amount of chromium consumed was estimated by calculating the mg SDD consumed from the average daily drinking water consumption data for weeks 1–25 for each dose and converting that amount to mg Cr. The 48-h urine collection contained up to 24 μg of chromium for rats exposed to SDD at 516 mg/l at day 371. Rat urine chromium values ranged from 0.58–2.4% of dose for the lowest two exposure concentrations, to 0.19–0.95% for the highest exposure concentration, with the lower values in each range seen at the earlier time points. Insufficient mouse urine was collected for analysis at most time points.
In rats and mice, the highest mean chromium concentrations were observed in the glandular stomach, kidney, and liver. Mean concentrations of chromium in the kidney were greater than those in the liver at most time points for rats, whereas the reverse was generally true for mice. In general, tissue concentrations at each time point increased with increasing exposure concentration. Statistical analysis of the shapes of the exposure-concentration curves revealed either linear or supralinear responses (Tables 2 and and3;3; Fig. 1A); a supralinear dose response is defined as a decreasing rate of response with increasing exposure concentration.
When tissue concentrations are normalized to external Cr dose (in milligram/kilogram body weight), at 516 mg/l on Day 182 (Table 1) and compared between rats and mice, female mice show statistically higher tissue chromium concentrations in liver (~5×) and glandular stomach, (~4×), whereas male rats had a higher concentration of chromium in the kidney (~1.3×) (Table 4).
In some tissues, such as kidney, liver, forestomach, and glandular stomach, tissue concentrations increased with duration of exposure at each exposure concentration, but the rate of accumulation of total chromium in tissues decreased with longer exposure duration (Fig. 1). This pattern of tissue accumulation was consistent for all tissues when normalized to ingested Cr dose per unit body weight. Concentrations of chromium in erythrocytes and plasma did not increase with increasing duration of exposure at any exposure concentration.
Chromium concentration in tissues and excreta are given in Tables 5 and and66 for male rats and female mice, respectively, following exposure to CPM. The 48-h feces collection contained up to 59.3 mg of chromium for rats and 5.6 mg for mice. In rats, 42.2% of the ingested chromium was excreted in the feces, whereas 20.2% was found in the feces of mice. The total amount of chromium consumed was determined by calculating the mg CPM consumed from the average daily food consumption data for weeks 1–25 for each dose and converting that amount to mg Cr. The 48-h urine collection for CPM-exposed animals contained up to 31 μg of chromium for rats. Insufficient mouse urine was collected for analysis at most time points.
As with Cr(VI), the tissues with the highest mean chromium concentrations after exposure were the glandular stomach, kidney, and liver in rats and mice. Tissue Cr concentrations were higher than would be expected based only on the presence of erythrocytes in the blood of these tissues. Mean chromium concentrations in the kidney were greater than those in the liver at most time points for rats, whereas the reverse was true for mice. However, the observed chromium concentrations were much lower following exposure to CPM than following exposure to SDD, even though exposures to chromium were much higher with CPM. Chromium tissue concentrations generally increased with exposure up to 10,000 ppm; however, levels did not increase above this dose. The exposure-total Cr concentration curves were supralinear or flat (Tables 5 and and6;6; Fig. 1B).
In kidney and liver, chromium concentrations increased with increasing exposure duration, at all exposure concentrations, in both rats and mice; however, the rate of tissue chromium uptake decreased with longer exposure duration. In rats, mean concentrations in the kidney increased dramatically between day 13 and day 182. Tissue chromium concentrations did not increase with exposure duration in erythrocytes, forestomach, or glandular stomach for all doses in both rats and mice. These observations can be interpreted as the tissues approaching equilibrium between Cr(III) uptake and Cr(III) loss via diffusion from the tissues and from cell turnover.
Calculation of Cr dose in milligram/kilogram body weight revealed that exposure to 516 mg SDD/l, which resulted in increased squamous neoplasms of the oral cavity in male and female rats and increased epithelial neoplasms of the small intestine in male and female mice (NTP, 2008a; Stout et al., 2009a), and 2000 ppm CPM resulted in ingestion of comparable doses of Cr per unit body weight. Cr doses from CPM exposure averaged 1.7 and 2.8 times higher, respectively, in rats and mice than those of SDD. All other CPM dose groups resulted in appreciably higher Cr exposures than the 516 mg SDD/l dose group (Table 1).
Exposure to 516 mg/l SDD for 182 days, which was the longest common exposure duration, resulted in significantly higher chromium tissue concentrations compared with those resulting from exposure to 2000 ppm CPM for 182 days. For example, rats exposed to SDD had 13 times more chromium in the liver and 5 times more chromium in the kidney than rats exposed to CPM, and mice exposed to SDD had 39 times more chromium in the liver (Fig. 1C) and 22 times more chromium in the kidney than mice exposed to CPM.
To confirm that interactions with the sodium-sulfate co-transporter might underlie differences in Cr(VI) versus Cr (III) uptake observed in these studies, experiments were conducted in membrane vesicles from kidney, a preparation that has high levels of sodium-sulfate transport. In these experiments, SDD and CPM were incubated with rat kidney BBMV in the presence of 35S to determine if the chromium compounds interacted with the sulfate transporter. Exposure to 50, 200, or 500μM SDD resulted in significant exposure-dependent decreases in sulfate uptake by the vesicles, indicating that SDD inhibited sulfate uptake (Fig. 2). In contrast, exposure of up to 500μM CPM had no effect on sulfate uptake.
Because SDD inhibited the uptake of sulfate, kinetic experiments were conducted to determine the type of interaction with the transporter. The Km for binding of sulfate was significantly increased (p ≤ 0.05) from 12.6 ± 0.69μM in the control experiment to 40.5 ± 3.18μM in the presence of 250μM SDD, whereas the Vmax values were similar between control (65.9 ± 25.3 pmol/30 s/mg protein) and SDD (50.1 ± 15.8 pmol/30 s/mg protein) samples. An increased Km (decreased affinity) without a change in Vmax indicates that the inhibition of sulfate transport was competitive and provides additional evidence that Cr(VI) is taken up by sulfate transporters.
The NTP characterized and compared the carcinogenicity and tissue distribution of SDD and CPM, two chromium compounds with widespread human exposure. The NTP chronic studies of Cr(VI), as SDD, demonstrated that it was carcinogenic in rats and mice after oral exposure (Table 7; NTP, 2008a; Stout et al., 2009a). In contrast, at much higher exposure concentrations, Cr(III), as CPM, may have been carcinogenic in male rats but was not carcinogenic in female rats or mice (NTP, 2008b; Stout et al., 2009b). As part of these studies, total Cr concentrations were measured in tissues and excreta of male rats and female mice at various exposure durations to provide internal dosimetry data to aid in the interpretation of the bioassay results. Because sex differences in chromium tissue accumulation were not expected based on studies conducted by Mackenzie et al. (1958) and Sutherland et al. (2000), measurements were made in only one sex of each species. There were increases in total chromium in multiple tissues in both male rats and female mice, indicating that systemic exposure to chromium occurred following exposure to both Cr(VI) and Cr(III).
It had been hypothesized that oral exposure to Cr(VI) would not produce an increase in cancer, except perhaps in the stomach, because of the efficient capacity of the stomach to reduce Cr(VI) to Cr(III) (De Flora, 2000; De Flora et al., 1997; Proctor et al., 2002). To determine if ingested Cr(VI) was systemically distributed, tissue Cr concentrations resulting from exposure to similar external doses of Cr(VI) and Cr(III) were compared. In all tissues examined, similar external doses of chromium resulted in much higher tissue Cr concentrations following exposure to Cr(VI) relative to Cr(III), indicating that at least a portion of the Cr(VI) was distributed to tissues prior to reduction. These data are consistent with reports in the literature in which the same concentration of Cr(VI) and Cr(III) was administered in drinking water for 11 months (Costa, 1997; Costa and Klein, 2006; Mackenzie et al., 1958). The greater uptake of Cr following exposure to Cr(VI) is consistent with differences observed in toxicity and carcinogenicity, including the absence of small intestine tumors in mice following Cr(III) exposure, and the proposed mechanisms of transport from previous studies (Alexander and Aaseth, 1995); reviewed by Costa (1997) and Costa and Klein (2006) and the present study.
Kerger et al. (1996, 1997) and Paustenbach et al. (1996) have proposed that the pharmacokinetics of Cr(VI) following oral exposure can be explained largely by the reduction of ingested Cr(VI) in the gut, which results in the production of Cr(III) organic complexes. Although Cr(III) is relatively nondiffusible across cellular membranes, Kerger et al. suggest that Cr(III) organic complexes are more easily absorbed into cells than inorganic Cr(III) and once absorbed are more rapidly eliminated than Cr(VI). These investigators point to similar Cr elimination profiles in red blood cells (RBCs) and plasma as evidence for the presence of Cr(III) organic complexes because Cr(VI) absorbed in blood is reduced to unstable intermediates, which form complexes with hemoglobin and other proteins resulting in retention of Cr for the lifetime of the RBC (Gray and Sterling, 1950; Kerger et al., 1996; Ottenwaelder et al., 1988). Our data are not consistent with this hypothesis and suggest that Cr(VI) was taken up by RBCs and tissues following administration of SDD. Several lines of evidence support this conclusion. Cr concentrations were significantly increased in erythrocytes at all exposure durations following exposure to the two highest doses of SDD. Although we utilized a washout period of 48 h, the chromium concentrations in the erythrocytes were approximately sixfold higher than those in the plasma, indicating that the Cr taken up by RBCs was largely retained, rather than diffusing into the plasma from RBCs or other tissues. In contrast, with CPM, an organic complex, chromium concentrations were higher in plasma than in RBCs, indicating limited uptake or more extensive diffusion into the plasma from the RBCs or other tissues. The 15- to 20-fold higher Cr concentrations (on day 182) in the RBC following exposure to Cr(VI), relative to a comparable external dose of Cr(III), and the observed toxicity to RBCs with Cr(VI) but not Cr(III) provides additional evidence that Cr(VI) was preferentially taken up by and was toxic to erythrocytes. These data are consistent with previous reports in the literature demonstrating higher levels of chromium in blood following administration of Cr(VI) compared with Cr (III) (Gray and Sterling, 1950; Mackenzie et al., 1959).
Although chromium was not measured in the small intestine following exposure to Cr(VI) or Cr(III), previous reports suggest that Cr(VI) is also likely to be absorbed in this tissue to a greater extent than Cr(III) (Donaldson and Barreras, 1966; Fébel et al., 2001). The study by Davidson et al. (2004) demonstrating increased susceptibility to skin cancer induction in hairless mice following co-exposure to ultraviolet light and Cr(VI) in the drinking water provides additional evidence that Cr(VI) can have systemic effects that are distant from the site of exposure. The data in the present report do not explain why neoplasms were not observed at sites distant from the alimentary tract.
Because of the observed species differences in sites of induced neoplasms following exposure to Cr(VI), tissue Cr concentrations normalized to external dose were compared between rats and mice. Cr uptake was found to be significantly higher in the kidney of rats and the liver and glandular stomach of mice (Table 4). This is consistent with previous studies in the literature (Coogan et al., 1991; Kargacin et al., 1993; Witmer et al., 1989, 1991). In addition, higher tissue concentrations were achieved in rats than occurred in tissues of mice exposed to the lower concentrations of SDD that also resulted in small intestine neoplasms (57.3 in female mice and 172 mg/l in mice of both sexes). Based on these lines of evidence, the tissue concentration data do not explain the species differences in target sites of carcinogenicity.
It has been previously hypothesized that the small intestine neoplasms observed in the NTP 2-year bioassay of SDD would occur only at doses that exceeded the gastric reduction capacity (De Flora et al., 2008). If the gastric reduction capacity had been exceeded, the dose that resulted in saturation would likely represent an inflection point for a sublinear exposure-response, with doses above this point demonstrating a greater rate of response than lower doses. Following exposure to SDD, the shapes of the exposure-response curves for both tissue concentration data in male rats and female mice (Fig. 1A) and incidences of small intestine neoplasms in male and female mice (Table 7; Stout et al., 2009a) were either linear or supralinear. In addition, Cr(VI) doses from the NTP 2-year mouse study were compared with gastric reductive capacity estimates originally reported for humans (De Flora et al., 1997) and allometrically scaled to mice and compared with average daily doses of Cr(VI) following exposure to SDD (Stout et al., 2009a). This comparison revealed that the calculated dose that might saturate gastric reduction is higher than all the doses in male mice and is nearly equivalent to the highest dose in female mice (Stout et al., 2009a). Collectively, these data indicate that the gastric reduction capacity was not saturated in rats or mice exposed to Cr(VI) in drinking water.
The lowest concentration of Cr(VI) in this study that produced an increase in tumor incidence in the small intestine of female mice was 20 mg/l (57.3 mg SDD/l). Using the time weighted average daily dose of Cr(VI) for the entire 2-year study and assuming mouse and human external exposure concentrations scale by body weight3/4 (body weight raised to the 3/4 power), exposure of mice to 20 mg Cr(VI)/l (1.016 mg/kg) for 2 years would result in a human equivalent daily dose of 0.166 mg/kg. This calculated dose is equivalent or within an order of magnitude to the doses estimated for a 70 kg person drinking 2 l of water per day at the highest concentrations reported in a survey of drinking water sources collected in Texas (5.41 mg/l; Texas Department of State Health Services, 2009) or California (0.603 mg/l; California Department of Public Health, 2007a). The U.S. EPA has set a maximum contaminant level of 100 μg/l total chromium in drinking water (U.S. EPA, 2003), although the limit in several states is 50 μg/l.
In conclusion, the results of these studies support the hypothesis that Cr(VI) is the species of chromium responsible for the induction of carcinogenesis in the NTP chronic toxicity and carcinogenicity studies of SDD. In addition, these results indicate that gastric reduction was not saturated following exposure to Cr(VI) and that differences in tissue uptake cannot account for the species differences in sites of Cr(VI)-induced carcinogenicity. The transport studies confirm previous reports that Cr(VI) is taken up by cells via the sodium/sulfate co-transporter, whereas Cr(III) is not, providing at least a partial explanation for the observed differences in tissue uptake.
This research was supported (in part) by the Intramural Research Program of the National Institutes of Health, National Institute of Environmental Health Sciences under Research Project Number ZO1 ES045004-11 BB and Z01 ES65554.
The authors thank Drs Nigel Walker and Suramya Waidynatha for their critical review of this article.