Document babkEXpe7o5pYL4zQrYm9VEdk

Environmental and Molecular Mutagenesis 32:251 -259 (1998) Studies on the Genotoxicity of Molybdenum Salts in Human Cells In Vitro and in Mice In Vivo Nina Titenko-Holland,* Jianhua Shao, Luoping Zhang, Liqiang Xi, Hailong Ngo, Nong Shang, and Martyn T. Smith Division of Environmental Health Sciences, School of Public Health, University of California, Berkeley, California Molybdenum is an essential element in plants and animals as a cofactor for enzymes. Molybdenum trioxide is used in metallurgical processes, in cos metics as a pigment, and in contact lens solution, yet limited information is available on molybdenum genotoxicity. In the present study the micronucleus (MN) assay in human lymphocytes and mouse bone marrow and the dominant lethal assay in mice were used to assess the genotoxic effects of molybdenum salts in vitro and in vivo. Two salts of molybdenum were tested in whole blood cultures. Ammonium mo lybdate was more potent than sodium molybdate in causing a dose-dependent decrease in viability and replicative index and an increase in MN formation in binucleated lymphocytes (P < 0.001). A doseresponse in both kinetochore-positive MN (caused by chromosome lagging) and kinetochore-negative MN (associated with chromosome breakage) was observed. Based on the results of a toxicity study of sodium molybdate, two doses, 200 and 400 mg/ kg, were assessed in the bone marrow MN assay in mice (two i.p. injections 24 and 48 hr prior to euthanasia). A modest but statistically significant in crease in MN frequency in polychromatic erythro cytes was observed (P < 0.05). The same treatment protocol was used to analyze dominant lethality. A dose-dependent increase in postimplantation loss represented mostly by early resorptions was ob served the first week after treatment (P = 0.003). These preliminary data suggest that sodium molyb date induces dominant lethality at the postmeiotic stage of spermatogenesis. Overall, molybdenum salts produced moderately positive results both in vitro in human cells and in vivo in mice. Environ. Mol. Muta gen. 32:251-259, 1998 1998 Wiley-Liss, Inc. Keywords: molybdenum; human lymphocytes; micronuclei; mouse bone marrow; postimplanta tion loss; dominant lethality INTRODUCTION Molybdenum is an essential microelement which serves as a cofactor in enzymes in humans and other species [US EPA, 1975]. Molybdenum salts and other compounds are used in metallurgical processes as addi tives to steel and corrosion-resistant alloys and occupa tional exposure occurs at roasting plants [Walravens et al., 1979]. Molybdenum is also an environmental pollutant produced from uranium processing and combustion pro cesses, and was recently introduced as a component of contact lens solution and a color additive in cosmetics [ACGIH, 1995]. Human molybdenum intake in the United States has been estimated to be 120-350 mg/day [Tsongas et al., 1980; Goyer, 1996], with a NOEL at 500 mg/day [Chappell, 1974]. The lethal dose for different species has been estimated to be between 100 and 800 mg/day, depending on the route of exposure and chemical structure of the molybde num compound [Maresh et al., 1940; Miller and Engel, 1960; Mills and Davis, 1987]. Pneumoconiosis developed in 3 out of 19 people who had been chronically exposed to metallic molybdenum and molybdenum trioxide [Mog- ilevskaja, 1963]. A gout-like disease has been reported in people living in areas with high levels of molybdenum in the soil [Kovalsky and Vorotnitskaya, 1970; Pitt, 1976]. Teart disease, a form of wasting and copper defi ciency, has been observed in cattle grazing on pastures high in molybdenum [Mills and Davis, 1987]. A recent National Toxicology Program study [NTP, 1997] has shown that exposure of mice and rats to molybdenum trioxide by inhalation caused alveolar inflammation and hyaline degeneration of the respiratory epithelia. Equivo cal evidence of carcinogenic activity of molybdenum tri oxide in rats was obtained in the same 2-year inhalation Contract grant sponsor: National Institute of Environmental Health Sci ences Center at Berkeley; Contract grant number: NIH P30 ESO1896; Contract grant sponsor: NIH; Contract grant number: P42 ES04705 from the National Institute of Environmental Health Sciences, using funds provided by US EPA. Correspondence to: Dr. Nina T. Holland, 217 Warren Hall, SPH, Uni versity of California, Berkeley CA 94720-7360. E-mail: ninah@ uclink4.berkeley.edu Received 29 September 1997; Revised and accepted 19 July 1998 1998 Wiley-Liss, Inc. 252 Titenko-Holland et al. study with exposure to 30 or 100 mg/m3 6 days a week [NTP, 1997]. In a three-generation reproductive study, mice were chronically exposed to 10 ppm molybdenum in drinking water, which resulted in early death of mice in F1, F2, and F3, litters with no live pups, and increased maternal death [Schroeder and Mitchener, 1971]. In rats which received 80 or 140 ppm sodium molybdate in feed for several weeks, fewer litters and impaired growth of pups were observed [Jeter and Davis, 1954]. Only limited and conflicting information is available on the genotoxicity of molybdenum compounds. Ammonium molybdate (10--5 M, 24 hr) induced chromosome aberra tions and sister-chromatid exchanges (SCEs) in human lymphocytes in vitro [Bobyleva et al., 1991]. These indi ces of cytogenetic damage were also increased in lympho cytes of workers with more than 10 years of occupational exposure to molybdenum [Babaian et al., 1980; Bobyleva et al., 1993]. Molybdenum trioxide (8 and 15 mg/kg), but not ammonium molybdate (2.5 and 7.5 mg/kg), induced chromosome damage in mouse bone marrow [Chopikashvili et al., 1991]. The same authors reported dominant lethal mutations were induced and survival in Drosophila decreased when 2-4 g ammonium molybdate were added per kg of feed [Chopikashvili et al., 1991]. On the other hand, compounds of molybdenum were not mutagenic in Escherichia coli [Venitt and Levy, 1974], Bacillus subtilis [Kada et al., 1980], and Salmonella typhimurium, with or without metabolic activation [Zeiger et al., 1992; NTP, 1997]. Negative results were also obtained with 0.5-10 mkg/ml molybdenum trioxide in cytogenetic tests (SCEs and chromosome aberrations) in cultured Chinese hamster ovary cells [NTP, 1997]. Since the information available on the genotoxicity of molybdenum is conflicting, and there were no data on reproductive genotoxicity in mammals, the goal of the present study was to explore further the possible genotoxicity of molybdenum salts in human cells in vitro and laboratory mice in vivo. An additional reason for the study was that applications of molybdenum were recently ex panded from metallurgy to include contact lens solutions, cosmetics, etc., which could immediately affect many people. Three well-established genotoxicity assays were employed. The micronucleus (MN) assay was performed in human lymphocytes in vitro and mouse bone marrow in vivo. Additionally, the dominant lethal assay in mice was applied to assess the reproductive genotoxicity of molybdenum in a mammalian system. Ammonium mo lybdate and sodium molybdate were chosen since they are formed immediately upon addition of molybdenum trioxide to an aqueous buffered solution. They also repre sent the likely products of molybdenum trioxide upon interaction with body fluids. Finally, ammonium molyb date and sodium molybdate are readily soluble and less toxic than other molybdenum compounds. MATERIALS AND METHODS Micronucleus Assay in Human Lymphocytes Blood was collected from two healthy volunteers (donor #1 was a 27-year-old female, and donor #2 a 24-year-old male) in heparinvacutainers and cultured for 72 hr in a humidified incubator with 5% CO2 at 37C. Culture medium consisted of RPMI 1640 supplemented with 10% fetal bovine serum (Hyclone, Logan, UT), 2 mM L-glutamine, 100 units/ml penicillin, 100 mg/ml streptomycin (Gibco, Grand Island, NY), and 1% phytohemagglutinin (PHA, HA15, Burroughs-Wellcome, Greenville, NC). Two salts, ammonium molybdate (NH4)6 Mo7O24 4H2O, and sodium molybdate (Na2MoO4 H2O) (Sigma, St. Louis, MO), were used to as sess the genotoxicity of molybdenum. Both salts were diluted in PBS (Ca2+ and Mg2+ free) immediately before addition to the cultures (24 hr after they were initiated) and were present until the end of cultivation at 72 hr. Four doses of each salt were used to treat duplicate lymphocyte cultures in three experiments. Colchicine (0.05 |jM), a known aneuploidogen, was used as a positive control (incubation for 24 hr). Cytochalasin B (Sigma) (6 mg/ml) was added to the cultures at 44 hr postinitiation as described by Fenech and Morley [1985]. Cytochalasin B prevents the cells from completing cytokinesis, resulting in the formation of multinucleated cells. Prior to harvesting, cell viability was determined by trypan blue (0.16%) exclusion. Easy-Lyse Whole Blood Erythrocyte Lysing Kit (Leinco Technologies, Ballwin, MO) was employed to gently lyse eryth rocytes while maintaining the viability of unfixed cells. Briefly, 0.2 ml lysing solution was added to 0.2 ml whole blood culture medium. After incubation at room temperature for 10 min, mixed solution was centri fuged and cells washed in 1 ml of buffer. After the supernatant was decanted, cells were ready for viability check. Lymphocytes were isolated using Ficoll-Paque (Pharmacia, Piscataway, NJ) density gradient at 2,000 rpm for 35 min at room temperature. The lymphocyte layer was removed and washed in PBS at 1,200 rpm for 10 min, then spun directly (600 rpm, 10 min) onto glass slides using a cytocentrifuge (Shandon, Sewickley, PA). After air drying and before methanol fixation at room temperature for 15 min, slides were stored at --20C in a sealed box and desiccated in an N2 atmosphere. Detailed procedures for performing the MN assay with antikinetochore antibody staining have been described elsewhere [Eastmond and Tucker, 1989; Yager et al., 1990]. Briefly, methanol-fixed slides were incubated for 5 min in PBS containing 0.1% Tween 20. Excess fluid was drained from the slide and 40-50 ml of the antikinetochore antibody (Chemicon, Temecula, CA) diluted 1:1 with PBS containing 0.2% Tween 20 was applied. The slide was then coverslipped and placed in a humidified chamber at 37C for 1 hr. Following two washes in PBS containing 0.1% Tween 20 for 5 min each, excess fluid was again drained and the slides were covered with a 1:50 dilution of fluorescent goat anti-human IgG (Chemicon) and incubated again for 1 hr. Because the fluorescent-labeled antibodies fade upon exposure to light, this and all subsequent steps were conducted in yellow light. The slides were rinsed twice in buffer plus 0.1% Tween 20 and counterstained with DNA-dye 4'6-diamidino-2-phenylindole (DAPI) (2 mg/ml) in an anti fade solution [Johnson and de Nogueira Araujo, 1981]. Slides were stored refrigerated for up to a week prior to analysis by microscopy. Replicative index (RI), a measure of cell division kinetics, was calcu lated by scoring at least 200 cells per dose or sample and counting the percent of cells containing 1, 2, 3, or more nuclei per individual. RI was calculated as follows: RI = [(1x% mononuclear cells) + (2 X% binuclear cells) + (3 X% tri) + (4 X% tetra)]/100. The MN frequency was estimated in 1,000 binucleated cells of sodium and ammonium molybdate/dose/per experiment per donor, except cultures treated with colchicine or high levels of ammonium molybdate, which were cytotoxic and thus yielded smaller numbers of binucleated cells. Genotoxicity of Molybdenum Salts 253 Micronuclei in Mouse Bone Marrow Statistical Analysis Male C57BL/6J and C3H/J female mice (7-12 weeks of age) were purchased from BAK Universal (Fremont, CA). Males were housed singly in plastic cages. The animal room was maintained on a 12-hr light/dark cycle, 20-24C, with 60-70% relative humidity. Mice were fed standard breeding granulated diet and water was supplied ad libitum. Animal husbandry was in accordance with the Guide for Care and Use of Laboratory Animals (NIH Publication no. 78-023, revised, 1985). Fifteen healthy C57BL/6J male mice (10-12 weeks) were randomly assigned to control and treatment groups. Treated males received intraperitoneal injections (0.7 ml) of sodium molybdate (Sigma) in sterile PBS for two consecutive days, while mice in the control group received concurrent injections of PBS. Dose levels (200 and 400 mg/kg) were chosen on the basis of range-finding data which determined the 50% of maximum tolerated dose [MacGregor et al., 1987]. Colchicine was used as a positive control (single i.p. injection of 0.8 mg/kg). Animals were sacrificed by CO2 anesthesia followed by cervical dislocation 48 hr after final injection. The femoral bone marrow cells were flushed out with fetal bovine serum (~1 ml) and the cellular suspension was spun at 1,000 rpm for 5 min. Smears were prepared from the resulting pellet, following the procedure of Schmid [1976], and stained by May-Grunwald-Giemsa. The incidence of MN was as sessed in an average of 2,000 polychromatic erythrocytes (PCE) per dose. The proportion between normochromatic erythrocytes and PCE was estimated in 400 cells from each slide. Slides from each dosepoint were randomized and scored at 630X under oil immersion (Nikon micro scope) by two observers who were unaware of the identity of the slides. No significant difference between scorers was observed. All question able MN were additionally assessed by a third scorer. Dominant Lethal Assay in Mice The dominant lethal assay was performed according to the standard protocol [Green et al., 1985; Shelby et al., 1986] with some modifica tions. Since there was no data on the reproductive genotoxicity of sodium molybdate in mice, we decided to prescreen the effects on the main periods of spermatogenesis in a pilot experiment using a minimal num ber of females. Our goal was to determine whether any reproductive genotoxicity might be observed in either postmeiotic, mitotic, or premeiotic germ cells before committing a significantly larger number of ani mals with daily sampling for 7-8 weeks, as would be required by standard dominant lethal protocol. C57BL/6J male mice were randomly selected from the same group of animals used for the MN assay and assigned to control (10 males) and treated groups (10 males for 200 mg/kg and 12 males for 400 mg/kg). A total of 166 untreated C3H/J female mice (10-12 weeks old) were housed 4-5 per cage and were mated with treated C57BL/6J mice after the end of treatment to sample various postmeiotic cells: spermatozoa (week 1), spermatids (week 3), spermatocytes (week 5), and spermatogonia (week 7) [Bateman and Epstein, 1971]. The mating activity of molybdenum-treated C57BL/6J males was estimated as a percent of females found with vaginal plugs after overnight placement of one male with two females. Females were sacrificed 15-16 days after vaginal plugs were found. The corpus lutea (CL) on both ovaries were counted under a dissecting Nikon stereomi croscope to assess the number of ovulations or potential fertility in each female. The uterine contents were inspected for the number of total implants, early resorptions (moles), and dead embryos. Early postim plantation loss was estimated by the number of moles per female, and late postimplantation loss by the number of dead embryos per female. Total postimplantation loss was calculated as a percent of dead implants, divided by the total number of implants, including moles, live and dead embryos. Dose-response curves for each combination of chemical and MN endpoint were fitted by various polynomial equations using the general ized linear model routine of S language [Chambers and Hastie, 1992]. Best-fit curves were selected by minimizing the Akiki information mea sure. The P-value for each dose-response was generated according to trend of the best-fit curve. Effects of potential confounding factors on the dose-response were also assessed by fitting more complex logistic regression models which considered donor and experiment in addition to dose. However, there was negligible experiment variability, while donors differed significantly (P < 0.001). Thus, the P-values reported for Figures 1 and 2 were calculated after adjustment for effects of experiment and donor variability. Pairwise comparisons were performed by Fisher's exact test for the frequency of bone marrow MN in treated and control animals. When we analyzed the effects of dose and time on postimplantation loss, we pooled the results from all animals of the same dose/week group and used the generalized linear model approach [Chambers and Hastie, 1992]. A value of P < 0.05 was used as the criterion for statistical significance. RESULTS Micronuclei Induced by Sodium and Ammonium Molybdate in Human Lymphocytes Micronucleus (MN) analysis was utilized as a measure of cytogenetic damage in human lymphocytes using antikinetochore-antibody staining to characterize the mecha nism of MN formation (Table I). Two salts of molybde num were used for in vitro treatment of human lympho cytes for 48 hr in three independent experiments for two donors (one female and one male). The range of concen trations chosen for sodium molybdate was 0.1-5 mM, while the more cytotoxic ammonium molybdate was stud ied at 0.1-2 mM. Viability of treated cells varied between 74 and 63% for sodium molybdate and 68 and 61% for ammonium molybdate. A significant inhibition of prolif eration, assessed by RI, was observed only at the highest concentration of 5 mM sodium molybdate (1.5 vs. 2.0 in control, P < 0.01). Ammonium molybdate, on the other hand, caused a significant dose-dependent decrease of RI at three out of four concentrations. The extent of ammo nium molybdate-induced inhibition of proliferation was comparable with the positive control colchicine (1.1 vs. 2.0 in control cultures). This inhibition resulted in a lim ited number of binucleated cells available for scoring, thus a total of only 452 cells were analyzed for MN frequency at the highest concentration of ammonium mo lybdate for both donors. The statistical generalized linear model was used, which allowed comparison between do nors, evaluation of the variability between experiments, as well as interaction between these factors and the dose. In all cases of total MN frequency, as well as kinetochorepositive and kinetochore-negative MN, there was no sig nificant difference between experiments for either donor. However, the female donor had a consistently higher MN frequency than the male donor in all experiments, both in controls (14.7 MN cells per 1,000 vs. 8 MN cells, 254 Titenko-Holland et al. TABLE I. Micronuclei Induced by Molybdenum Salts in Human Lymphocytes Chemical treatment Control Colchicine Sodium molybdate Ammonium molybdate (mM) 0 5 X 10--5 0.1 0.5 1 5 0.1 0.5 1 2 Donor F M F M F M F M F M F M F M F M F M F M Scored BN cells 3,000 3,000 1,562 1,303 3,000 3,000 3,000 3,000 3,000 3,000 3,000 3,000 3,000 3,000 2,577 3,000 2,281 1,624 323 129 MN cells per 1,000 14.7 2.9 8.0 1.7 92.5 14.8 88.0 19.8 19.3 1.5 17.3 4.5 22.6 6.5 19.6 5.7 31.6 4.5 20.6 5.1 32.0 3.0 17.0 1.7 19.3 2.5 13.3 3.1 30.3 12.9 20.6 3.8 40.3 2.1 19.3 10.0 55.5 9.2 31 Total 51 24 167 132 69 54 75 63 100 69 112 57 63 43 91 64 101 39 21 4 MN MN+ 17 9 131 106 41 30 46 40 59 40 68 33 42 29 50 37 61 18 14 2 MN-- 34 15 36 6 28 24 29 23 41 29 44 24 21 14 41 27 40 21 7 2 Replicative index* 2.0 0.0 2.0 0.1 1.1 0.0 1.1 0.1 1.8 0.1 1.9 0.2 1.9 0.1 1.8 0.0 1.7 0.1 1.8 0.1 1.5 0.0 1.6 0.1 1.8 0.1 1.8 0.1 1.7 0.1 1.6 0.0 1.3 0.0 1.2 0.0 1.1 0.0 1.0 0.0 Viability (%) 87.3 2.3 79.6 5.0 78.0 5.8 71.3 2.1 74.0 7.0 69.0 6.2 71.6 7.5 67.3 3.5 69.7 1.1 66.0 4.6 67.6 5.7 62.6 5.5 68.3 4.0 62.7 2.5 67.0 6.6 64.0 5.3 64.3 6.8 67.6 4.6 61.0 4.2 62.5 2.1 Abbreviations: MN, micronuclei; BN, binucleated cells; MN cells, micronucleated cells; +, kinetochore-positive; --, kinetochore negative. *Replicative index = [(1 X % mononucleated cells) + (2 X % binucleated cells) + (3 X % tri- or >trinucleated cells)]/100. respectively) and after molybdenum exposure (P < 0.001). Both molybdenum salts induced a dose-dependent in crease in the frequency of total MN in both donors (P < 0.001) (Fig. 1). The increase in MN per 1,000 cells was modest for sodium molybdate (a little more than twofold over the control level), and more noticeable for ammo nium molybdate (an almost fourfold increase) compared to a nearly tenfold increase caused by colchicine. The level of kinetochore-positive MN+ in control cul tures was 4.3/1,000 cells, and of kinetochore-negative MN-- 8.2/1,000 cells. Similar trends for induction of both types of MN were observed after sodium molybdate treat ment (Fig. 2b), and by ammonium molybdate in the male donor (Fig. 2a). However, in the female donor (#1) notice ably more kinetochore-positive than kinetochore-negative MN were induced by ammonium molybdate, with a five fold increase over control level at the highest dose. A statistical model based on all datapoints estimated that dose-effect for both MN+ and MN-- was significant after treatment with either sodium molybdate or ammonium molybdate (P < 0.001). These data suggest that a plausi ble mechanism of molybdenum genotoxicity includes chromosome lagging, resulting in aneuploidy, and chro mosome breakage. Molybdenum Toxicity in Mice A dose range-finding experiment was performed with sodium molybdate in mice (Table II). This experiment showed that mortality was not affected up to a dose of 200 mg/kg sodium molybdate. No statistically significant increase was observed at 400 mg/kg (85.2% survival), while at 500 mg/kg more than 50% of treated animals died within 24 hr after the second injection (P < 0.01). Based on this mortality data, all experiments with domi nant lethality and bone marrow MN were performed with two doses of 200 and 400 mg/kg of sodium molybdate. Bone Marrow Micronuclei Induced by Sodium Molybdate in Mice The background ratio of poly- and normochromatic erythrocytes was 0.97, and was not significantly affected by sodium molybdate (1.06 and 0.83 at 200 and 400 mg/ kg, respectively) (Table III). Colchicine treatment (0.8 mg/kg), on the other hand, was highly cytotoxic and caused a significant decrease in the PCE/NCE ratio (0.13, P < 0.001). The results of the bone marrow MN assay are shown in Table III. The control level was 2.8 MN cells/1,000 with no multiple MN detected. Sodium mo lybdate induced a statistically significant increase of MN in PCEs. Micronucleated cells with multiple MN were observed mostly in PCEs at 400 mg/kg. As a result, a regression of the MN frequency per 1,000 cells was statis tically significant (P < 0.05), though the dose response for micronucleated cells did not reach statistical signifi cance. The level of molybdenum-induced MN was smaller (twofold increase) than with colchicine treatment used as a positive control (3.5-fold increase, P < 0.01). Genotoxicity of Molybdenum Salts 255 Fig. 1. Dose-response of the micronuclei induced by ammonium mo lybdate (A) and sodium molybdate (B) in lymphocytes of two donors: female, original datapoints shown with circles; male, data shown with triangles. One thousand cells were scored for each of the three experi ments for both salts. Dose-response curves of best-fit present results of statistical modeling, which included assessment variability between donors and experiments and interaction between dose and donor. The difference between experiments was negligible, while donors differed significantly (P < 0.001). Effect of the dose was highly significant for both salts, P < 0.001. This preliminary data suggest that sodium molybdate was weakly genotoxic in mouse bone marrow, though it did not affect cell proliferation at the two concentrations studied. Dominant Lethal Assay in Mice The mating activity of C57BL/6J male mice was not affected by treatment with sodium molybdate (Table IV). No difference in the weight of males (26.0-26.8 g) or females (23.1-23.5 g) between treated and control groups was observed. Slightly fewer corpus lutea were found in females mated with treated males (8.0-8.7 per female) compared with control females (9.6 per female), but this difference was not significant (P > 0.05). Pregnancy rate Fig. 2. Kinetochore-positive (MN+) and kinetochore-negative mi cronuclei (MN-) in human lymphocytes treated with molybdenum salts. MN+ and MN-- were differentiated by the presence of bright yellowgreen signal after antikinetochore-antibody staining. Cells were counterstained with DAPI. Data from two donors (female and male) and three experiments for both sodium molybdate and ammonium molybdate were included in this analysis. Fitted plots present results of statistical model ing of dose effect taking into account variability between experiments and between donors, as well as interaction between donor and dose. MN frequency in two donors differed significantly (P < 0.001), while variability between experiments did not (P > 0.05). The increase in MN+ as well as MN- was dose-dependent and statistically significant for both ammonium molybdate (A) and sodium molybdate (B) (P < 0.001). was estimated as the percent of females with vaginal plugs which had at least one implant at the 15/16th day of pregnancy. There was a 10% decrease in the pregnancy rate in the 400 mg/kg group, but the difference was not statistically significant (P > 0.05) except in the first week. However, an overall dose-dependent increase was ob served in total postimplantation loss (Table IV). From 6.7% in the control group, postimplantation loss was in creased to 10.6% in the 200 mg/kg group and 16.3% in the 400 mg/kg group (P = 0.001). This increase was most noticeable at the first week 256 Titenko-Holland et al. TABLE II. Effect of Sodium Molybdate (Na2MoO4) on Mortality in Mice Dose of Mo (mg/kg) 0 50 100 200 400 500 750 1,500 Number of animals injected 11 3 3 14 27 14 12 12 Number of mice dead* 0 0 0 0 4 8 10 10 Survival (%) 100 100 100 100 85.2 42.8** 16 7*** 16.7*** *All deaths occurred during first 12 hr after i.p. injection of (Na2MoO4) in PBS. Statistically significant by Fisher's exact test, **P< 0.01, ***P < 0.001. (reflecting an effect on postmeiotic male germ cells, P = 0.003). Both early and late postimplantation loss were affected by molybdenum (Table V). A statistically sig nificant and dose-dependent increase in the number of moles (early resorptions) was found at the first week (0.33 per female in control vs. 0.94 in 400 mg/kg group, P = 0.013). The average size of moles, though, in control and treated groups was similar (3.4 1.1 mm in control, 3.8 0.9 mm in the 200 mg/kg group, and 3.3 0.6 mm in the 400 mg/kg group). Late postimplantation loss (dead embryos) were found in only one control female at the third week, and were comparatively rare in the treated group. A significant increase was observed only for the first week after treatment (P = 0.036). Since we observed a gradual increase in postimplantation loss in both control and treated groups at 3-7 weeks after treatment compared to the first week (Table V), we used a generalized linear model to assess which covariates were likely responsible for this trend. This analysis showed that the rise in postim plantation loss was associated with the increasing body weight of females at later dates of mating (P < 0.001) rather than with different stages of spermatogenesis af fected. An appropriate correction for the difference in the weight of the females was made, after which a significant postmeiotic dose-dependent effect of sodium molybdate on postimplantation loss in the treated group of mice was confirmed. These preliminary results indicate a mild but statistically significant postmeiotic effect of sodium mo lybdate on dominant lethality in mice. DISCUSSION All three assays used in the present study yielded posi tive evidence of modest molybdenum salt genotoxicity both in vitro and in vivo. Ammonium molybdate was slightly more cytotoxic and genotoxic in human lympho cytes than sodium molybdate. However, each salt induced both kinetochore-positive MN, which suggests aneu- ploidy as at least one mechanism for their genotoxicity, and kinetochore-negative MN, which was most likely caused by chromosome breakage. Appreciable differ ences between the two donors were observed in both background and molybdenum-induced MN frequency. The female donor had higher MN frequencies, and also had a noticeably higher level of induction of kinetochorepositive MN. This difference may be explained by the known predominant elimination of the inactive X-chromosome in the cells of females [Tucker and Preston, 1996] . Our findings on MN induction by molybdenum salts correlate well with earlier studies published in the Russian literature in which chromosome aberrations and sister-chromatid exchanges in human lymphocytes were analyzed [Babaian et al., 1980; Bobyleva et al., 1991]. In these studies, workers occupationally exposed to molyb denum had a lower RI and increased frequency of single fragments and exchange types of chromosome aberrations in their peripheral blood lymphocytes. Moreover, when lymphocytes from these workers were treated with mo lybdenum salts in vitro an increase in cytogenetic damage was observed, suggesting a sensitization due to previous contact with molybdenum. Together these data show that molybdenum salts may cause chromosome damage in hu man lymphocytes at relatively high concentrations by at least two mechanisms, chromosome breakage and chro mosome lagging which leads to aneuploidy. On the other hand, in a study of molybdenum trioxide performed in Chinese hamster ovary (CHO) cells, no in crease in either SCEs or chromosome aberrations was observed with or without metabolic activation [NTP, 1997] . There are several possible explanations for this discrepancy. Since molybdenum trioxide is more cyto toxic than molybdenum salts, cytogenetic analysis was limited only to concentrations allowing cell growth in the NTP study. These concentrations were 4-10 times lower than equimolar concentrations used in our study with mo lybdenum salts. No significant increase was seen in hu man lymphocytes at the range of concentrations compara ble to those used for molybdenum trioxide. Other possible explanations are interspecies differences in sensitivity to molybdenum and/or differences in sensitivity of short term human lymphocyte culture versus the transformed hamster cell line. There are limited data available on the possible molecu lar mechanism of molybdenum genotoxicity. It is known that it may impair purine metabolism through an enhance ment of xanthine oxidase activity [NTP, 1997]. Molybde num also has a variety of oxidative states, and one can speculate that molybdenum could cause oxidative damage through the formation of free radicals, as was previously shown for the closely related element chromium [Itoh and Shimada, 1996]. Interference with protein metabolism and enzyme impairment described in animals fed a diet with a high level of molybdenum [Mills et al., 1958; NTP, Genotoxicity of Molybdenum Salts TABLE III. Effect of Sodium Molybdate (Na2MoO4) on Bone Marrow Micronuclei in Mice Dose (mg/kg) Number of cells scored MN cells per 1,000 PCE MN per 1,000 PCE PCE/NCEa 0 200 400 Positive control (colchicine, 0.8) 9,000 9,000 8,000 4,600 2.8 4.8* 4.6* 9.6** 2.8 4.9* 5.8* 11.3** 0.97 1.06 0.83 0.13*** a400 cells were scored for the ratio of PCE/NCE, and at least 1,000 PCE were scored for the micronucleus (MN) frequency from each slide. *P < 0.05, **P < 0.01, ***P < 0.001 (one-tailed Fisher's exact test). 257 TABLE IV. Descriptive Values for Male and Female Mice in the Dominant Lethal Assay With Sodium Molybdate Treatment (mg/kg body-weight) 0 200 400 Males Number Body weight1 Mating activity2 Females Number Body weight1 at day of conception Body weight1 at 16th day of pregnancy Number of corpus lutea1 Pregnancy rate (%)3 Postimplantation loss (%)4 10 26.4 2.2 45.2% 57 23.1 2.0 32.8 5.3 9.6 1.5 84.2 6.7 10 26.8 2.4 41.8% 55 23.5 2.5 32.8 5.1 8.7 3.1 87.3 10.6 12 26.0 1.5 38.1% 54 23.4 2.4 31.8 5.8 8.0 2.7 74.1 16.3 'Mean SD. 2Mating activity was calculated as a frequency of successful mating (female/s found with vaginal plugs) for all males in the treatment group at all times they were placed with females. 3Rate of pregnancy was calculated as a portion of females found with implants out of a total number of females with vaginal plugs. 4Postimplantation loss was calculated as follows: 1- (number of dead implants/total number of implants per female). Dose-dependent increase was caused by sodium molybdate (P = 0.001). 1997] may also explain the aneuploidy induction observed as kinetochore-positive micronuclei in our experiments. More studies are warranted in order to understand the mechanism of molybdenum genotoxicity. The comparatively low toxicity of sodium molybdate in vivo allowed concentrations for treatment of mice used for the bone marrow MN and dominant lethality assays to be within the range of concentrations used in vitro for human lymphocytes. The effect of sodium molybdate on bone marrow MN was statistically significant but rela tively mild compared to the effects of the positive control colchicine. In a study of chromosome aberrations in mouse bone marrow, molybdenum trioxide, but not am monium molybdate, induced chromosome damage [Chopikashvili et al., 1991]. These data suggest that mo lybdenum compounds may be genotoxic for mouse bone marrow in vivo. Prior to this study, there were only two reports on the reproductive toxicity of molybdenum in mice and rats [Jeter and Davis, 1954; Schroeder and Mitchener, 1971], but to our knowledge no data were available on germ cell genotoxicity in mammals. The increase in postimplanta tion loss induced by sodium molybdate shown here is the first indication of possible reproductive genotoxicity of molybdenum in mice. The only other available reference reported that ammonium molybdate induced an increase in dominant lethality in Drosophila [Chopikashvili et al., 1991]. A postmeiotic effect of sodium molybdate in mice mostly affected spermatozoa and resulted in an increase in both early and late postimplantation death. A similar slight increase in postimplantation loss was induced in spermatids and spermatocytes, but only the effect on sper matozoa was statistically significant. Thus, like a vast majority of other mutagens studied [Russell et al., 1990], molybdenum appears to have a predominantly postmeiotic effect on male germ cells. However, we cannot reach 258 Titenko-Holland et al. TABLE V. Postimplantation Loss in Mice After Treatment With Sodium Molybdate Treatment (mg/kg body weight) 0 200 400 Spermatozoa (week 1) Females Early loss (moles/female) Dead embryos/female Total loss (%) Spermatids (week 3) Females Early loss (moles/female) Dead embryos/female Total loss (%) Spermatocytes (week 5) Females Early loss (moles/female) Dead embryos/female Total loss (%) Spermatogonia (week 7) Females Early loss (moles/female) Dead embryos/female Total loss (%) 20 0.33 0.62a 0 5.4 29 0.55 0.91 0.03 0.19 8.0 7 1.0 1.0 0 12.5 5 0.8 1.1 0 26.4 18 0.44 0.62 0.11 0.47 7.8 23 0.63 0.83 0.05 0.23 9.3 7 0.5 0.58 0 12.5 8 1.2 0 14.1 21 0.94 1.35b 0.06 0.24c 13.8d 16 0.62 0.77 0.31 0.48 12.1 10 2.0 1.63 0.14 0.38 24.6 8 1.5 0 21.4 aMean SD. bDose-dependent increase, P = 0.013. cDose-dependent increase, P = 0.036. dPostimplantation loss was increased, P = 0.003. a definite conclusion on the genotoxicity of molybdenum on the meiotic and premeiotic stages of spermatogenesis based on the preliminary data available from our pilot experiment. More studies are needed to confirm our find ings, to expand them to different salts of molybdenum, and possibly to use different treatment protocols. Thus, evidence is accumulating that different com pounds of molybdenum are moderately genotoxic in vitro and in vivo, both in somatic and germ cells. Since these compounds are often present in complex mixtures with other metals and chemicals during mining and metal pro cessing, or are found in food and drinking water, one concern about molybdenum is its possible interaction with other mutagens. In a study of chromosome aberrations in the bone marrow of mice treated with ammonium molyb date in combination with the antibacterial drug dioxidine, a synergistic effect between the two agents was observed [Chopikashvili et al., 1991]. On the other hand, in experi ments on survival and dominant lethality in Drosophila treated with ammonium molybdate or molybdenum triox ide in combination with other metals and/or vitamin C, no additional increase was produced by combinations of molybdenum, wolframite, and cadmium [Chopikashvili et al., 1991]. Moreover, vitamin C had a protective effect on the survival of Drosophila treated with ammonium molybdate. In its moderate genotoxicity and possible in teraction with other agents, molybdenum is similar to other metals, such as lead [Kristensen et al., 1993], chro mium and selenium [Itoh and Shimada, 1996], and mer cury and cadmium [Berces et al., 1993]. Further studies are warranted on the interaction of molybdenum com pounds with other mutagens and antimutagens in vivo and in vitro. In summary, molybdenum salts yielded moderately positive results at relatively high doses in three experi mental systems. In light of its significant environmental presence in complex mixtures, particularly related to min ing and metallurgy, and its increasing use in health care products, the possible adverse effects of molybdenum in humans deserve more attention. ACKNOWLDGMENTS Jinhua Shao was supported by a WHO fellowship. The views expressed here are solely those of the authors and not of the funding agencies. The authors are grateful to Barton Bowers for his help with preparation of the manu script. REFERENCES ACGIH (American Conference of Governmental Industrial Hygienists) (1995): 1995-1996 Threshold Limit Values (TLVs) forchemical substances and physical agents and biological exposure indi ces (BEIs). Cincinnati, OH: ACGIH. Babaian EA, Bagramian SB, Pogosian AS (1980): Effect of some chemi cal hazards involved in molybdenum production on the chromo Genotoxicity of Molybdenum Salts 259 some apparatus of experimental animals and of man. Gig Tr Prof Zabol 9:33-36. Bateman AJ, Epstein SS (1971): Dominant lethal mutations in mammals. In Hollaender A (ed): ``Chemical Mutagens: Principles and Methods for Their Detection, Vol. 2.'' New York: Plenum Press, pp 541-568. Berces J, Otos M, Szirmai S, Crane-Uruena C, Koteles GJ (1993): Using the micronucleus assay to detect genotoxic effects of metal ions. Environ Health Perspect Suppl 101(3):11-13. Bobyleva LA, Chopikashvili LV, Alyokhina NI, Zasukhina GD (1991): Chromosome aberrations and sister chromatid exchanges in high risk groups of workers exposed to heavy metals. Cytol Genet (Russia) 25:3:18-23. Bobyleva LA, Chopikashvili LV, Alyokhina NI, Zasukhina GD (1993): Modification of spontaneous and induced levels of chromosome aberrations and sister chromatid exchanges by ascorbic acid in lymphocytes of workers exposed to heavy metals. Cytol Genet (Russia) 29:3:430-434. Chambers JM, Hastie TJ (1992): Statistical models in S. Pacific Grove, California: Wadsworth & Brooks/Cole Advanced Books & Soft ware, 1992. Chappell WR (1974): Introduction and summary. In: ``Transport and the Biological Effects of Molybdenum in the Environment.'' The Molybdenum Project, Progress Report. Boulder and Fort Collins, CO: University of Colorado and Colorado State University, Janu ary 4, 1974, pp 1-29. Chopikashvili LV, Bobyuleva LA, Zolotareva GN (1991): Genotoxicity of molybdenum and its derivatives in Drosophila and mammals. Cytol Genet (Russia) 25:5:45-49. Eastmond DA, Tucker JD (1989): Kinetochore localization in micronucleated cytokinesis-blocked Chinese hamster ovary cells: A new and rapid assay for identifying aneuploidy-inducing agents. Mutat Res 224:517-525. Fenech M, Morley AA (1985): Measurement of micronuclei in lympho cytes. Mutat Res 147:29-36. Goyer RA (1996): Molybdenum. In Klaassen CD, Amdur MO, Doull J (eds): ``Casarett and Doull's Toxicology, 5th Ed.'' New York: McGraw-Hill, p 718. Green S, Auletta A, Fabricant J, Kapp R, Manandhar M, Sheu C, Springer J, Whitfield B (1985): Current status of bioassays in genetic toxicology -- The dominant lethal assay. Mutat Res 154:49-67. Itoh S, Shimada H (1996): Micronucleus induction by chromium and selenium, and suppression by metallothionein inducer. Mutat Res 367:233-236. Jeter MA, Davis GK (1954): The effect of dietary molybdenum upon growth, hemoglobin, reproduction and lactation of rats. J Nutr 54:215-220. Johnson GD, de Norueira Araujo GM (1981): A simple method of reducing the fading of immunofluorescence during microscopy. J Immunol Methods 43:349-350. Kada T, Hirano K, Shirasu Y (1980): Screening of environmental chemi cal mutagens by the rec-assay system with Bacillus subtilis. In de Serres FJ, Hollaender A (eds): ``Chemical Mutagens: Princi ples and Methods for Their Detection, Vol. 6.'' New York: Ple num Press, pp 149-173. Kovalsky VV, Vorotnitskaya IE (1970): Induction of the enzymes of purine metabolism by copper and molybdenum. In Mills CF (ed): ``Trace Element Metabolism in Animals.'' Edinburgh and London: E. & S. Livingstone, pp 176-179. Kristensen P, Eilertsen E, Einarsdottir E, 0vreb0 S, Haugen A (1993): Effect of modification by inorganic lead in the dominant lethal assay. Mutat Res 302:33-38. Maresh F, Lustok MJ, Cohen PP (1940): Physiological studies of rhe nium compounds. Proc Soc Exp Biol Med 45:576-579. Miller RF, Engel RW (1960): Interrelations of copper, molybdenum and sulfate sulfur in nutrition. Fed Proc 19:666-677. Mills CF, Davis GK (1987): Molybdenum. In Mertz W (ed): ``Trace Elements in Human and Animal Nutrition, Vol. 1, 5th Ed.'' San Diego: Academic Press, pp 429-463. Mills CF, Monty KJ, Ichihara A, Pearson PB (1958): Metabolic effects of molybdenum toxicity in the rat. J Nutr 65:129-142. Mogilevskaja OY (1963): Experimental studies on the effect on the organism of rare, dispersed and other metals and their compounds used in industry. In Izraaelson ZI (ed): ``Toxicology of the Rare Metals.'' Jerusalem: Translations Ltd., pp 12-27. NTP (National Toxicology Program) (1997): Molybdenum trioxide. NTP TR 462, NIH Pub. No. 95-3378. Pitt MA (1976): Molybdenum toxicity: Interactions between copper, molybdenum and sulphate. Agents Actions 6:758-769. Russell LB, Russell WL, Rinchik EM, Hunsicker PR (1990): Factors affecting the nature of induced mutations. In Allen J, Bridges M, Lyon M, Moses M, Russell LB (eds): ``Banbury Report 34: Biology of Germ Cell Mutagenesis.'' Cold Spring Harbor, NY: Cold Spring Harbor Laboratories, pp 271 -289. Schmid W (1976): The micronucleus test for cytogenetic analysis. In Hollaender A (ed): ``Chemical Mutagens, Vol. 4.'' New York: Plenum Press, pp 31-53. Schroeder HA, Mitchener M (1971): Toxic effects of trace elements on the reproduction of mice and rats. Arch Environ Health 23:102 106. Shelby MD, Cain LA, Hughes LA, Braden PW, Generoso WM (1986): Dominant lethal effects of acrylamide in male mice. Mutat Res 173:35-40. Tsongas TA, Meglen RR, Walravens PA, Chappell WR (1980): Molyb denum in the diet: An estimate of average daily intake in the United States. Am J Clin Nutr 33:5:1103-1107. Tucker JD, Preston RJ (1996): Chromosome aberrations, micronuclei, aneuploidy, sister chromatid exchanges, and cancer risk assess ment. Mutat Res 365:147-159. US EPA (U.S. Environmental Protection Agency) (1975): Molybde num--A toxicological appraisal: EAP-600/1-75-004. Research Triangle Park, NC: Health Effects Research Laboratory, Office of Research and Development. Venitt S, Levy LS (1974): Mutagenicity of chromates in bacteria and its relevance to chromate carcinogenesis. Nature 250:493-495. Walravens PA, Moure-Eraso R, Solomons CC, Chappell WR, Bently G (1979): Biochemical abnormalities in workers exposed to mo lybdenum dust. Arch Environ Health 34:302-307. Yager JW, Eastmond DA, Robertson ML, Paradisin WM, Smith MT (1990): Characterization of micronuclei induced in human lym phocytes by benzene metabolites. Cancer Res 50:393-399. Zeiger E, Anderson B, Haworth S, Lawlor T, Mortelmans K (1992): Salmonella mutagenicity tests. V. Results from the testing of 311 chemicals. Environ Mol Mutagen 19(Suppl 21):2-141. Accepted by-- J. Yager