In Vivo Effects of ChromiumCharlotte Witmer," Ellen Faria," Hyoung-Sook Park,' Nakissa Sadrieh,lEdward Yurkow," Sean O'Connell,2 Andrea Sirak,3 and Heinz Schleyer41Joint Graduate Program in Toxicology, Rutgers University and Robert Wood Johnson Medical School,Piscataway, New Jersey; 2Department of Surgery, Robert Wood Johnson Hospital, New Brunswick, NewJersey; 3Block Drug Company, Jersey City, New Jersey; 4Department of Surgery, University of Pennsylvania,Philadelphia, Pennsylvania

The production of reactive oxygen species on addition of hexavalent chromium (potassium dichromate, K2Cr207) to lung cells in culture was studiedusing flow cytometer analysis. A Coulter Epics Profile II flow cytometer was used to detect the formation of reactive oxygen species after K2Cr2O7was added to A549 cells grown to confluence. The cells were loaded with the dye, 2',7'- dichlorofluorescein diacetate, after which cellular esterasesremoved the acetate groups and the dye was trapped intracellularly. Reactive oxygen species oxidized the dye, with resultant fluorescence.Increased doses of Cr(VI) caused increasing fluorescence (10-fold higher than background at 200 pM). Addition of Cr(I) compounds, as the picoli-nate or chloride, caused no increased fluorescence. Electron paramagnetic resonance (EPR) spectroscopic studies indicated that three (as yetunidentified) spectral "signals" of the free radical type were formed on addition of 20, 50, 100, and 200 pM Cr(VI) to the A549 cells in suspension.Two other EPR "signals" with the characteristics of Cr(V) entities were seen at field values lower than the standard free radical value. Liver micro-somes from male Sprague-Dawley rats treated intraperitoneally with K2Cr2O7 (130 pmole/kg every 48 hr for six treatments) had decreased activity ofcytochromes P4503A1 and/or 3A2, and 2C1 1. Hepatic microsomes from treated female Sprague-Dawley rats, in contrast, had increased activities ofthese isozymes. Lung microsomes from male Sprague-Dawley rats had increased activity of P4502C11. - Environ Health Perspect 102(Suppl3):169-176 (1994).

Key words: chromium, cytochrome P450 isozymes, electron paramagnetic resonance (EPR), reactive oxygen species (ROS), lung cells

IntroductionPrevious studies in our laboratories (1-3)and those of others (4-6) have shown thatthere are several potentially toxic in vitroeffects of hexavalent chromium [Cr(VI)]which are readily detected. However, some

of these effects are not evident in vivo andothers have not been tested in relevant sys-

tems. For example, it has been shown thatwhen Cr(VI) compounds are reduced byH202 and/or glutathione (GSH), the path-way of reduction is dependent upon theconcentrations of the reductants. Theresultant cellular damage (to DNA) may

depend upon the molar ratios of thereductants and Cr(VI) compound (4).Production of Cr(V) species in unidentified

This paper was presented at the SecondInternational Meeting on Molecular Mechanisms ofMetal Toxicity and Carcinogenicity held 10-17January 1993 in Madonna di Campiglio, Italy.We thank Dr. Paul Thomas, Rutgers University,

for many helpful discussions about isozymes ofcytochrome P450, as well as for the use of his HPLCequipment and the generous donation of the testos-terone and the metabolites for use as standards.This work was partly supported by training grantES07148 from the National Institute of EnvironmentalHealth Sciences.

Address correspondence to Dr. Charlotte Witmer,Joint Graduate Program in Toxicology, RutgersUniversity and Robert Wood Johnson MedicalSchool, Environmental Health Sciences Institute, 681Frelinghuysen Road, Piscataway, NJ 08855-1179.Telephone (908) 932-3751/3720. Fax (908) 932-0119.

forms, as well as the formation of a Cr(V)-tetraperoxo complex, can be detected withaddition of Cr(VI) to several reductants invitro (4,5,7,8). There is evidence thatunder slightly alkaline conditions, in vitro,hydroxyl radicals (-OH) are formed, whichthen react with deoxyguanine to initiateDNA strand breakage (4). The concentra-tions of reactants used to demonstrate theseintermediates in vitro have generally beenhigher than those found in vivo. The car-cinogenicity of chromium compounds alsoappears to result from attack on DNA byCr(III), formed in situ by the reduction ofCr(VI). Paradoxically, there are beneficialeffects of Cr(III) on glucose metabolism (9)and chromium is generally accepted(although controversially) as being arequired trace element (10). Extracellularreduction of Cr(VI) to Cr(III), which can-not readily cross the cell membrane, is cur-rently perceived as a protection reaction.The acute toxicity of Cr(VI), which is alsolittle understood, can result in fatality torats after intraperitoneal administration ofseveral daily doses in the 200 pmole/kgrange (2). This general toxicity must beseparated from the carcinogenicity in stud-ies of mechanisms of action, and is also notextensively studied. Current knowledge ofchromate carcinogenicity and toxicity thusindicates that effects of in vivo exposure to

Cr(VI) compounds depend upon theuptake and metabolic fate of the chromatein target organs.

The lung has been well established asthe principal target organ for chromiumcarcinogenicity (11-13), and as the site ofsome Cr(VI)-induced toxicity (3).However, little is known of the metabolicfate of hexavalent chromium in this organ,except that many reducing agents, such asascorbate and glutathione, are present inthis tissue and can ultimately produceCr(III) (14-16). The liver, on the otherhand, is the major site of xenobiotic metab-olism in mammals. Effects of Cr(VI) onthe liver enzymes of rats have been shownin our laboratory to differ in in vitro and invivo studies (1,2). Addition of potassiumdichromate (K2Cr207) to microsomal frac-tions from livers of both male and femalerats of two strains caused significantdecreases in both cytochrome P450 (P450)content and cytochrome c reductase activ-ity (a measure of cytochrome P450 reduc-tase activity). These decreases were foundin vitro in microsomes from untreated ratsas well as in microsomes from rats that hadbeen treated with inducing agents or withdichromate prior to removal of the livers(2). In in vivo studies, intraperitonealadministration of K2Cr207 (130 pmole/kgbody weight) to rats on alternate days for

Environmental Health Perspectives 169

WITMER ETAL.

six treatments showed that only livers frommale Sprague-Dawley rats exhibited adecrease in P450 content, while the hepaticreductase activities remained unchanged inall rats (4. These results emphasize the factthat in vitro effects are often not mimickedin vivo, and indicate the need for furtherexploration of in vivo effects of Cr(VI) onthe hepatic enzymes.

We have previously demonstrated thatin the rat less than 2% of orally adminis-tered hexavalent chromium is absorbed(17,18). However, lung chromium concen-tration was measurable (12 lig/g tissue) 24hr after 14 days of daily treatment with240 ,imole Cr(VI)/kg (as calcium chro-mate) (12). Intratracheal administration (45pmole/kg, which mimics the exposure ofworkers in smelting plants, results in rela-tively high concentrations of chromium inlung 21 days after the single treatment (8.5pg/g tissue) (E. Faria and C. Witmer,unpublished data). We also compared theeffects of Cr(VI) on A549 cells and L1210cells (19), using the alkaline elution tech-nique (20). The A549 cells are a cell linefrom human lung tissue which has the char-acteristics of Type II cells, and L1210 cellsare a murine cell line of leukemic origin. Inthis study, 5 pM chromate caused single-strand breaks as well as DNA-proteincross-links in A549 cells, but formed onlyDNA-protein cross-links in the L1210 cells.These findings support the premise thatlung tissue and lung cells differ from othertissues in their response to Cr(VI)-initiateddamage.

On the basis of these previous findings,we have used the A549 lung cells and invivo studies in rats to explore the effects ofCr(VI) exposure in in vivo systems. Thepurposes of the present work are to con-tinue our studies of the effects of Cr(VI)treatment on the hepatic and lung mixed-function oxidase enzymes and to determinewhether lung cells in culture respond toexposure to Cr(VI) compounds by produc-ing reactive oxygen species and free radicals(including paramagnetic forms ofchromium), as has been indicated by invitro work (21,22). We have also concen-trated on effects on P450 isozymes inhepatic and lung microsomes followingtreatment of the rats with K2Cr207. Thedose chosen was previously shown todecrease the total P450 in male Sprague-Dawley rats (4. To further study the modeof action of Cr(VI) in these in vivo systems,we used flow cytometry and electron para-magnetic resonance (EPR) spectroscopy.Flow cytometry is a powerful tool fordetecting reactive oxygen species (ROS),

which have been implicated in metal-induced DNA strand breakage in other celltypes (21,22). EPR spectroscopy can beused to detect, under favorable conditions,paramagnetic transition metal ion species,including Cr(V) species, as well as a varietyof free radical entities that may be formedin these A549 lung cells by reactions withchromium compounds.

Materials and MethodsAnimals and Cells

All rats used in these studies were obtainedfrom Taconic Farms (Germantown, NY).They were kept in the animal facility foracclimatization at least 4 days prior to useand conditions in the animal facility werethose recommended by the NationalInstitutes of Health. The animals weighedbetween 120 and 200 g when used. Bothfood (Purina rodent chow) and water weresupplied ad libitum. All water bottles werefitted with plastic tips and the food wasadded directly to the cages to avoidchromium contamination. The rats werekept in plastic cages in groups of three percage.

A549 cells were obtained from theAmerican Type Culture Collection(Rockville, MD) and were maintained in a95% air, 5% CO2 atmosphere, in RPMI1640 medium (Gibco, Grand Island, NY)at 370C, pH 7.2, with 10% fetal calf serum.

ReagentsHexavalent chromium compounds(Na2CrO4, K2Cr2O7) and all other chemi-cals not specifically noted were obtainedfrom Fisher Chemical Company(Somerville, NJ). The 2',7'-dichlorofluores-cein (DCFH) and the diacetate of DCFH(DCFH-DA) were purchased fromMolecular Probes (Eugene, OR). All chemi-cals were of the highest purity available.Chromium picolinate (98% purity) wasgenerously donated by the Nutrition 21Company (San Diego, CA). Testosteroneand testosterone metabolites were obtainedas referenced in Jayyosi et al. (23) and werewere donated to us by Dr. Paul Thomas(Rutgers University).

Treatment ofAnimalsPotassium dichromate was administeredintraperitoneally to the rats in groups of atleast three per group at a dose of 130jimole Cr/kg on alternate days for total ofsix treatments. Animals were sacrificed withether anesthetic 24 hr after the last dose.Organs were removed and weighed, andmicrosomes were immediately prepared.

Prepartion ofMicrosomesMicrosomes were prepared by methodsused routinely in our laboratory (24), andthe microsomal fractions were suspendedin phosphate buffer, pH 7.4 (0.1 M phos-phate, 0.15 M KCl) so that the resultingsuspensions contained the equivalent ofapproximately 15 g of original liver/ml.The microsomes were either used immedi-ately or frozen at -800C for future use.

Testosterone Metabolism (Assay forActities ofCytohrme P450Iszymes)The isozymes of cytochrome P450hydroxylate testosterone in a regio- andstereo-specific manner, and thus themetabolites are indicative of the specificisozyme activities. We used the methoddescribed by Jayyosi et al. (23) for thisassay. Specifically, testosterone (0.25pmole, dissolved in methanol) was incu-bated in a final volume of 1.0 ml of 0.05 Mphosphate buffer (pH 7.4, containing 0.15M KCI and 1 mM EDTA), with livermicrosomes (0.250 mg protein) or lungmicrosomes (0.250 mg protein), and anNADPH-generating system, in a shakingwater bath for 10 and 30 min, respectively,at 37°C. Each assay was carried out induplicate, and the appropriate blanks wererun. The reaction was stopped by the addi-tion of methylene chloride and an aliquotof the organic layer (containing the hydrox-ylated metabolites) was dried under nitro-gen and then dissolved in the mobile phaseand analyzed by using high-pressure liquidchromatography (HPLC). The HPLC col-umn was a reverse-phase C,8 column, usedas described by Jayyosi et al. (23) to sepa-rate testosterone, androstenedione and theeight monohydroxylated testosterone iso-mers at room temperature. The isomerswere identified by comigration withauthentic standards. It should be notedthat, in this assay, the microsomes usedwere from the same animals as those usedto study the in vitro effects of dichromate(4.Generation ofReactive Oxygen SpeciesThe Coulter Epics Profile II flow cytometerwas used to study the Cr(VI)-induced for-mation of reactive oxygen species (ROS),such as H202 and hydroperoxides, by mea-suring the fluorescence of the oxidizedform of 2',7'-dichlorofluorescein (DCF).This dye is fluorescent only in the oxidizedstate. The reduced dye (DCFH) is intro-duced into the cells and is oxidized to itsfluorescent form by any ROS producedafter Cr(VI) addition. The fluorescence

Environmental Health Perspectives170

IN VIVO EFFECTS OFCHROMIUM

intensity is thus directly proportional to theROS produced. To carry out these experi-ments, A549 cells were grown to conflu-ence, harvested by centrifugation at 1000gfor 10 min, and suspended in phosphatebuffer, pH 7.4 (0.1 M, containing 0.15 MKCl) at a concentration of 2 x 106 cells/mi.The reduced nonfluorescent dye (DCFH),as the diacetate (DCFH-DA), which isnonpolar and enters the cells rapidly (25)was added to the cells (5 liM final concen-tration) and the mixture incubated for 30min at 37°C. Intracellular esterases imme-diately hydrolyze the DCFH-DA leaving aresultant polar compound which is trappedin the cells in the reduced, nonfluorescentstate. K2Cr207 (in phosphate-bufferedsaline) at several concentrations (20, 100,200 PM) was then added in 200-jil vol-umes. (Total volume in the tube was 2.2ml). The resulting ROS were determinedat 10-min intervals for 60 min. The excita-tion wavelength of the dye is 488 nm; theemission wavelength, 525 nm. The data(number of cells containing the oxidizedfluorescent dye and the log fluorescence)were automatically recorded on a disc inthe cytometer, and histograms of the datawere generated using Epics Cytologic.Control samples contained either buffer(no chromium compound) or the trivalentcompound, CrCl3, which does not entercells readily, or chromium picolinate (dis-solved in methanol), which enters the cells(both at 200 piM). Positive controls wereH202 or phorbol myristate. To detect anydirect oxidation of the reduced dye by thehexavalent chromium, K2Cr207 was addedto the reduced dye directly in the Perkin-Elmer fluorimeter (Perkin-Elmer, SouthPlainfield, NJ) with a 525-nm bandpass fil-ter, and any fluorescence of the mixture at525 nm was recorded for 60 min.Experiments were repeated three timeswith the same results.

Electron Parmanetic ResonanceThe electron paramagnetic resonance(EPR) data were generated using a VarianE-109 spectrometer (Vairan Instruments,Palo Alto, CA) to detect spectra from freeradicals and/or paramagnetic Cr(V) species,following addition of hexavalent chromium(as K2Cr207) to the A549 lung cells. Thecells were grown to confluence and har-vested immediately prior to the EPR exper-iments. The cells were resuspended inphysiologic saline (by dilution of a stocksolution in phosphate-buffered saline,unless otherwise designated) at a concen-tration of 1 to 2 x 1 07 cells in a total vol-ume of 0.18 ml, and chromate was added

4.0

II.1

3.0

2.0

1.0

0.0

* co

0 C,(I

T

15 7A OS 1 A 16" 2a 2a IISAHP A

T-OH

Figure 1. Hydroxylation of testosterone by cytochrome P450 isozymes in hepatic microsomes from dichromate-treatedmale Sprague-Dawley rats. Microsomes were prepared from dichromate-treated male rats as described in the text. Thetestosterone assay and identification of metabolites were carried out as described by Jayyosi et al. (21). The abscissaindicates the position of hydroxylation of the testosterone.

in a volume of 0.250 ml. The concentrationsof chromate added are given in the legendsfor Figures 7 to 9. Immediately after additionof the chromate and mixing, the suspensionwas transferred to an EPR tube of the "flatcell" type (with minimal dimension in theE-field plane). The cell was inserted in thepretuned microwave cavity, and the instru-ment was retuned as necessary. In mostcases, this entire procedure required lessthan 1.5 min.

The EPR spectra (X-band, 9.1 GHz)were measured with 100 KHz field modu-lation. The Ho magnetic field was centeredat 3230 Oersted. Maximal field scan range

2.0

III

1.01

0.0oI5.

was 200 Oersted; typical scan rates were 50Oe/min. Nominal filter time constants(RC) of 0.1 or 0.3 sec were used.Microwave power levels were set at low,nonlimiting values (typically between 5and 20 mW incident radiation), and nar-row modulation amplitudes were chosen,close to those giving maximal signalresponse. High receiver gain settings wereneeded throughout. All essential experi-mental conditions, induding initial Cr(VI)concentration, presence of other ingredi-ents, and time information, are specified inthe individual figure legends.

U co

ra crVO

-6Ma27A IGA 1I" 2A 26 AD

T-OH

Figure 2 Hydroxylation of testosterone by cytochrome P450 isozymes in hepatic microsomes from dichromate-treatedfemale Sprageu-Dawley rats. Microsomes were prepared from dichromatre-treated female rats as described in the text.The testosterone assay and identification of metabolites were carried out as referenced in Figure 1. See text for identifi-cation of isozymes catalyzing the hydroxylation of testosterone at the positions indicated.

Volume 102, Supplement 3, September 1994 171

WITMER ETAL.

* cm

0 Crv(I

T

Figure 5. Production of reactive oxygen species (ROS)after the addition of Cr(lll) compounds (as CrCI3 and Crpicolinate) and Cr(VI) as potassium dichromate. The his-tograms for the Cr(VI) compounds are the same as thoseshown in Figure 4. The Cr(lIl) compounds (200 pM) wereadded to the A549 cells as described in the text. The his-tograms represent the appearance of ROS; measure-ments were made 20 min after the addition of eachcompound. Conditions were as for Figure 4. Control wasbuffer vehicle.

Figure 3 Hydroxylation of testosterone by cytochrome P450 isozymes in lung microsomes from dichromate-treated maleSprague-Dawley rats. Lung microsomes were prepared from the group of rats referred to in Figure 1. The testosteroneassay and identification of metabolites were carried out as referenced in Figure 1.

Fgure 4. Production of reactive oxygen species (ROS) after addition of Cr(VI) as potassium dichromate to A549 lungcells. The histograms, made using the program, EPICS Cytologic, represent the appearance of ROS after the addition ofbuffer, and 20 and 100 pM Cr(VI) to A549 cells. Measurements were made using the Coulter Epics Profile flow cytometer.Measurements were made 20 min after the addition of the compounds. Control was buffer vehicle.

The A549 cells were carefully examinedbefore and after each of the experimentalruns. In no case was there less than 80%survival as tested by Trypan Blue exclusionand cell count determinations.

ResultsOur studies of the effects of Cr(VI) treat-

ment on testosterone metabolism in hepaticmicrosomal preparations from male andfemale rats are summarized in Figures 1 and2. In microsomes from dichromate-treatedmale rats there was a significant decrease inthe 6p-hydroxylation of testosterone; the

21-hydroxylation was also decreased. Thesespecific hydroxylations are carried out

almost exclusively by both cytochromeP4503A1 (CYP3A1) and cytochrome P4503A2 (CYP3A2), isozymes previously desig-nated as cytochromes P450p and P4501respectively. As there is no constitutiveCYP3A1 enzyme in livers of mature malerats (26), the initial (and perhaps thereduced) activities at these positions appear to

be due to the CYP3A2 isozyme. The hydrox-ylation of testosterone at the 16a position,catalyzed by cytochrome P4502C 1I(CYP2C1 1), previously designated P450h, a

male-specific isozyme, was also significantlydiminished in the male rats following dichro-mate treatment.

The results of the dichromate treatmentof the female rats with Cr(VI) prior to iso-lation of hepatic microsomes were directlyin contrast to the results in hepatic micro-somes from the male rats similarly treated.The hydroxylations catalyzed by isozymesCYP3A1 and CYP3A2 (at both the 21 and613 positions) were significantly increased(Figure 2), as was the testosterone hydroxy-lation by CYP2B1 (at the 16a position,and in the formation of androstenedione).This latter enzyme is the type that is alsoinduced by phenobarbital. The two iso-forms of the 3A family (CYP3A1 andCYP3A2) are both present in the femalerat. These two isoforms are very similar inamino acid content (89% homology) andin substrate specificity, and their activitiescannot be satisfactorily separated by thismethod. For further separation, we arecurrently carrying out SDS gel elec-trophoresis of the microsomal fractionsand making Western blots of the P450proteins, followed by the use of mono-clonal antibodies to the two isozymes asprobes. The monoclonal antibodies wererecently developed by Dr. Paul Thomasand colleagues (personal communication).These results have been discussed else-where (3).

In lung microsomes from the dicro-mate-treated male rats, the specific changewas a significant increase in hydroxylationscarried out by cytochrome P4502B1(CYP2B 1), previously called P450b.Specifically there were increases in thehydroxylation activities at the 16a, and16,B positions and in the production ofadrostenedione (Figure 3).This isozyme

Environmental Health Perspectives

4.0

I

3.0

2.0

1.0

J T

ROS 20 minutesCPc200 UNcontrol

Cr(3) CrC13 200 uM.20 u Cr(6)

100 uml Cr(6)

.5,.,a^"- ;s

~| 5wtttl-v* ;s --'0 %_-g 2+t

I0.0IsU 7A OS UAI" " a 2S IIAKP A

T-OH

*~~~~~~~~~~~~~~~~~p o.! i;' z~! i~:iiii '~' -: ;-~~~~~~~~~~.50 -s: --; : . ; ! i M

*,S~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. .-.... .. x .. , , ii i

%AL

172

IN VIVO EFFECTS OFCHROMIUM

*.'>.~~t e . _s'.'!Ff0 ' ,';0 \0 .iiX*:A: .... : :. :

~#UEUEiv W

lC(6ooma +V.ItE 1lOuX

+ Yltt Ou

Figure 6. Effect of vitamin E succinate on production of reactive oxygen species (ROS) in A549 cells after addition ofCr(VI) as potassium dichromate. Vitamin E succinate (10 and 20 pM) was incubated with the A549 cells for 24 hr prior tomeasurement of the ROS produced by the addition of the Cr(VI) as potassium dichromate to the vitamin-treated cells.Vitamin E succinate in buffer and the control buffer were also added to the A549 cells at zero time. Cr(VI) was added at100 pM concentration, as indicated. Measurements were made as described in legend for Figure 1.

was also induced in the hepatic microsomesfrom the dichromate-treated female rats.

Generation ofROS FollowingAddition ofCr(VI) to A549 CellsFigures 4 to 6 are representative his-tograms generated by the Epics Cytologicprogram, indicating a significant forma-tion of ROS after K2Cr2O7 addition toA549 cells, at 20, 100, and 200 pM con-centrations, with an increase in ROSwith increased dichromate. These figuresall show the response at 20 min. TheROS produced at 30 min were almostidentical to those at 20 min, indicatingthat the reaction was maximal by 20min. No drop in the ROS levels appearsuntil the 40-min period, indicating someleakage of the dye from the cells by thisperiod, since the oxidized dye is stable.No ROS were seen with addition of thechromium compounds (the chloride andpicolinate) to the cells. The fluorescenceof cells containing the picolinate actuallywas less than that of untreated controlcells. We detected no effect of preincuba-tion of the cells with vitamin E succinate(20 pM, 24 hr) on the ROS producedwith addition of Cr(VI) (20 or 200 pM)to the A549 cells (Figure 6). The cellsmaintained their homogeneity for the60-min period of exposure to dichromate(figure not shown), indicating no damagedetectable by this measure. All resultsshown in these figures are from a typicalexperiment.

EPR SpectroscopyWe have used electron paramagnetic reso-nance spectroscopy (EPR) as a probe tostudy the chemical and biochemical eventsthat occur in A549 lung cells with exposureto Cr(VI). A representative example of theapplication ofEPR techniques to these cellsis illustrated in Figure 7. This spectrum is agood representative of results obtained withthe A549 cells under the conditions out-lined in "Materials and Methods."

The spectrum in Figure 7 shows at leastfive major spectral features (signals) whichwe have labeled Peaks I through V, with

field strength increasing from left to right.Actual Ho values and the exact times (aftermixing) when these signals were recordedare given in the legend. The spectrum rep-resents, in approximation, the first deriva-tive of the resonance absorption, sincemagnetic field modulation and phase-sensi-tive detection were employed. The intensi-ties of all observed spectral features (peakamplitude, or integrated areas, when used)exhibit an absolute dependence on the con-centration of the A549 cells. With buffermedium alone, only a flat "baseline" spec-trum is seen. Furthermore, all signals alsodepend on the initial Cr(VI) concentration(see below). A free radical standard(DPPH) has been included in the measure-ment. Under the instrumental conditionsused, this resonance was centered at 3204.9Oe (Figure 7).

Three of the major EPR signals (PeaksI-III) were observed at magnetic fieldslower than that of the marker, and thuswith g-values greater than that of the freeelectron value. Together with other infor-mation, this suggests that the speciesresponsible for this part of the EPR spec-trum have the characteristics of free radicalspecies. On the other hand, the rest ofEPRsignals, which are found at higher fields(Peaks IV and V), and thus to the right ofthe marker position, are consistent withcomplexes of paramagnetic transition ionsand are no doubt Cr(V) species.

Some of the chemical and biochemicalcomplexities caused by the exposure of theA549 cells to Cr(VI) are seen in Figure 8a-d Peak amplitudes of the five major EPRpeaks previously defined are plotted as afunction of time. These data were obtained

3187 3201.0 3218A

323 344

Figure 7. EPR spectrum of A549 cells after exposure to Cr(VI). Signal amplitudes, in relative units (au), are plotted as afunction of the HO field in [Oersted]. The reaction was started by adding Cr(VI) (as K2Cr207 in buffered solution) to har-vested and resuspended A549 cells at "zero time." A total of 2.2 x 107 cells was used, in a volume of 0.43 ml. Final con-centrations at zero time were: Cr(VI) 116 pM, Na+ (as NaCI) 57 mM, Na2-EDTA 4.2 mM, phosphate 4.0 mM. Field valuesof the prominent spectral features (at constant microwave frequency) are indicated in the graph (Peaks l-V). They wererecorded -actual times after mixing-as follows: Peak 1(3184.7 Oe), at 2.98 min; Peak 11(3201.9 Oe), at 3.3 min; PeakIII (3218.4 Oe), at 3.7 min; Peak IV (3233.9 Oe), at 3.9 min; and Peak V (3244.6 Oe), at 4.17 min. All these measurementswere performed at 21 'C. The inserted marker indicates the field position of an organic free radical standard (crystalline1,1-diphenyl-2-picryl-hydrazyl, DPPH, with a g-value, g = 2.0036 ± 0.0003) under the conditions of this experiment. Thisnarrow-line resonance signal of this marker was observed at 3204.9 Oe.

Volume 102, Supplement 3, September 1994

.:..

|-=i-s

173

WITMER ETAL.

25

7

.01

0.r:0 25

20

a

0

time [min]

10 15

time [min]

E

S 10 1S

time [min]20 2

10 15

time [min]

Figure 8 EPR spectra of A549 cells after exposure to Cr(VI): time course. Peak amplitudes (arbitrary units, au),sponding to the Peak to V signals defined in Figure 7, are plotted as a function of time after mixing, in min. The)determined directly from the recorded spectra; no attempts were made here to obtain and use 'refined values'nonlinear curve-fitting procedures. The spectra were recorded as described. Approximately 2 x 1 07freshly harvestewere resuspended in 0.18 ml phosphate-buffered saline. Cr(VI), as K,Cr20,, was added in phosphate buffer (0.17.2) to obtain the final Cr(VI) concentrations specified below. Final phosphate concentration was about 58 mM. /

panels have identical X- and Y- axes. In each panel, the following codes are used: Peak I, open circles, solid line; Ffilled circles, long dashes; Peak Ill, open triangles, medium-length dashes; Peak IV, filled triangles, short dashes; Fopen squares, dotted line. Final Cr(Vl) concentrations for exposure at "zero time" were (a) 23 pM; (b) 58 pM; (c) 1 '

(d 233 pM.

40-p~~~~~~~~~~~400

30

30

2~~~~~~~~~~~~~2, . ,

ED sFi'

time (min) time nmm]

Figure 9. EPR spectra of A549 cells after exposure to Cr(VI): effects of dimethylsulfoxide (DMS0). (a) Meanobserved amplitudes of Peak and Peak II is plotted as a function of time [min] after mixing. DMSO sample: opensolid line; control sample (no DMSO): filled circles, broken line. (b) The amplitude of Peak V is plotted, in analogouion as a function of time after mixing. DMSO sample: open circles, solid line; control sample (no DMSO): filledbroken line. Experimental conditions used: 2.0 x 107 cells in 0.43 ml final volume. Cr(VI) 116 pM; DMSO 320 mM.

from a set of experiments in which the ini-tial Cr(VI) concentration was varied about10-fold, from 23 to 233 FiM final concen-

tration, while the concentration of A549cells was constant. Detailed analysis showsthat the maximal amplitudes of each of themajor signals varied strongly, and in com-

plicated ways, with the Cr(VI) concentra-

tion. Also, these amplitudes were reachedat significantly different times after mixing.

In spectra from the highest concentra-

tion of Cr(VI) used, the signals are

detectable for up to 20 min after mixing,while at the low concentration, these sig-nals last only for about 10 min, indicatingthat the rates of decay are also different.

We have also examined the effects ofthe addition of a several additional com-

pounds, including dimethylsulfoxide andinorganic anions (phosphate and sulfate) to

the A549 cells, by adding them to theextracellular medium. In Figure 9, we showthe effects on the Cr(VI)-induced spectra

(233 pM Cr), of the addition of dimethyl-sulfoxide (DMSO) to the cellular medium.Panel A shows the effects seen on the mean

2 5 of the amplitudes of Peaks I and II; panel Bshows that the opposite effect is seen on

the amplitude of Peak V. Details are out-

corre- lined in the figure legend.y were" using DiScussiondcells The results seen in the studies of in vivoM, pH effects of Cr(VI) were of two kinds: a

Peak II change in critical isozyme activities, whichPeakV, may be detrimental for the animal under16 uM; some circumstances; and the formation of

ROS and paramagnetic species which are

known to be toxic. That these effects on

the isozymes are seen in vivo in both liverand lung after treatment of rats with rela-tively low doses of Cr(VI) indicates that theliver at least responds somewhat differentlythan predicted by the in vitro studies.Relatively high Cr(VI) concentrations were

required in the in vitro system to reducethe total cytochrome P450 content (2).

The findings of significant changes inthe rat hepatic and lung microsomalisozymes of the P450 family of enzymes

following in vivo exposure to dichromateb - indicate that the effects of hexavalent

chromium compounds can indude impor-tant indirect effects, such as changes in the

40 first step of metabolism (the hydroxylation)of other xenobiotics. The liver enzymes,

and specifically the P450 family are themajor site of xenobiotics metabolism. The

cirfclths substrates metabolized by the 3A P450is fash- gene family include many carcinogens, i.e.,circles, CYP3A1 metabolizes aflatoxin B, and

acetylaminofluorene, as well as cortisol,

Environmental Health Perspectives

25

50 uM

20

15

10 _

5 /I ~~~~~~b0 ma

a

174

IN VIVO EFFECTS OFCHROMIUM

progesterone, and testosterone. This genefamily is induced by steroids and has themost widespread substrate activities of allthe P450 families. Aflatoxin B, is a xenobi-otic to which we are frequendy exposed (inpeanut butter), so changes in its rate ofmetabolism could cause concern.

The overall effects of these changes inspecific hepatic P450s on metabolism ofother xenobiotics cannot be effectivelyassessed, however, without information onthe fate of the phase I products (epoxidesand/or hydroxylated metabolites) resultingfrom P450 oxidations. Any such changescould alter the equilibria of the initial reac-tions. For example, chromate has beenshown to inhibit glutathione reductaseactivity in erythrocytes (27) and perhapsthis is also true in other tissues. Our resultsemphasize the differing effects of chromatetreatment on lung. and on one other tissue,liver in male rats. The interesting opposingeffects of Cr(VI) on liver enzymes in maleand female rats, however, were not unex-pected, as steroids are metabolized differ-ently in the two sexes.

CYP2B1, which was increased in activ-ity in lungs of Cr(VI)-treated male rats, isactive in lung Clara cells. The lung is thefirst site of inactivation of inhaled xenobi-otics, and the increased activity of thisform of P450 appears to be a defenseaction. Increases in CYP2B1 can causeincreased metabolism of xenobiotics suchas aniline, barbiturates, and benzo[a]-pyrene and other compounds in the lung.However, if the companion enzyme, epox-ide hydrolase, is not also increased, therecould be increased production of toxic orcarcinogenic intermediates, such as epox-ides. This enzyme is also induced by phe-nobarbital treatment and is relativelynonspecific in its substrate activities. Themechanism of this increased activity wasnot addressed in the current study.

The formation of an abundance ofreactive oxygen species caused by Cr(VI)addition to the A549 lung cells indicatesthat the reduction of hexavalent chromiumis not a simple process in which someCr(V) may be formed which is thenreduced to Cr(III) complex(es) that reactwith DNA. Rather, there is a possibility ofmany interactions of reactive species withintracellular moieties. Several methods ofidentifying these species are now beingtried in our laboratory. The lack of effect ofvitamin E on the ROS formation wasunexpected and suggests that the speciesproduced do not result from lipid peroxi-dation. Vitamin E added to V79 cells inculture at the same concentration by

Sugiyama (28) reduced the number of sin-gle strand breaks in DNA caused by hexa-valent chromate. This suggests that thestrand breaks are not caused directly by thereactive oxygen species.We have also carried out preliminary

studies in which we exposed A549 cells toseveral concentrations of Cr(VI) as dichro-mate for 3 hr. The dichromate exposurewas followed by washout of the chromatewith subsequent growth in a dichromate-free medium for 24 hr, followed by RNAsetreatment and staining of the cells withpropidium iodide. We found an abnormalaccumulation of cells in S-phase in cellsexposed to 5 pM dichromate. This is inagreement with Costa (29) who alsoreported a build-up of S-phase cells inChinese hamster ovary cells after treatmentwith several metals. Our work with the cellcycle analysis is still in progress.

EPR SpectroscopyEPR spectroscopy detects the presence ofparamagnetic entities, allows their quanti-tation, and affords the characterization ofkinetic features of their formation as wellas their decay processes. Compounds thatcan be investigated exhibit paramagnetismthat is derived either from the electronconfiguration of transition metal ions incertain valence states and specific coordi-nation geometries, or from molecularunpaired electrons in other molecules, i.e.,free radical species (S=1/2). Differentspecies can be observed side by side, andtheir individual characteristics can be eval-uated. However, a definitive chemicalcharacterization of the molecules causingthe observed resonance absorption solelyon a spectroscopic basis may be difficult.

EPR spectroscopy has been successfullyapplied to studies of the transition ele-ment, chromium, in its many states andlevels of complexation for many years.Most chromium compounds were foundto have g-values close to the "free electron"value, often at slightly higher fields (i.e.,lower g-values); observed line widths areoften narrow, especially for Cr(V) species.As is characteristic for transition metalions, the actual spectral features are quitesensitive to the nature of the ligands, thegeometry of coordination, and even thematrix in which these complexes are stud-ied. These factors influence the shape ofthe resonance absorption; relaxation timesand temperature play additional importantroles. Finally, for different, purely geomet-ric reasons, the actual EPR spectra for anygiven material obtained in liquid solution(as we have used here) differ substantially

from spectra obtained in the frozen state(with "immobilized" quasi-random orien-tation of the spins). Thus the g-values inimmobilized and randomly oriented sam-ples are inherently very different from theg-values of solution spectra.

Of particular interest in these studieswere paramagnetic entities containingchromium in a formal (V) valence statewith a 3di electron configuration. Thesecoordination complexes play recognized,but not understood, roles as intermediatesin the reduction of Cr(VI) compounds tocomplexes of the thermodynamically stableCr(III). Recent reports include the obser-vation of tetraperoxo-chromate(V) as aproduct of the reduction of Cr(VI) (4,5)and a number of studies of Cr(V) entitiesin a variety of chemical and biochemicalsystems, with reported g-values around1.98 (4-6). Frozen samples were usuallyemployed to facilitate detection and to cir-cumvent problems of chemical stability ofthe Cr(V) complexes.

The aim of our application of EPRspectroscopy was to investigate effects ofCr(VI) in A549 cells, under conditionsthat are as close as possible to physiologicalconditions. These techniques provide themost direct experimental access to themeasurements of the paramagnetic species.The approach requires the operation of theEPR spectrometer close to the fundamen-tal limitations of the method, especiallyregarding detection sensitivity. It necessi-tates compromises in virtually all instru-mental settings (choice of microwavepower levels, modulation amplitudes, scanrates and amplifier gain). Use of cell sus-pensions in solution allowed us to lookclosely at the reactions in which theobserved paramagnetic species [includingCr(V) species] respond kinetically in realtime, under nearly physiological condi-tions, and relatively unperturbed by themeasurements. We have not used spin-trapping agents, so as not to shift equilib-ria or alter kinetic information.

The EPR spectrum seen in Figure 7 isa good example of our results obtainedwith the A549 cells under our defined con-ditions (see Materials and Methods). Fivemajor spectral features (signals) can beidentified (Peaks I to V). Each of theseexhibits its own complex kinetics of for-mation and decay. A representative sum-mary of these kinetic aspects [amplitude vstime, as a function of initial Cr(VI) con-centration] is shown in Figure 8a-d.

Peaks I to III can be classified, spectro-scopically, as generated by free radicalspecies. Like all species we have observed

Volume 102, Supplement 3, September 1994 175

WTMER ETAL.

in the A549 cells, they are transientspecies, varying greatly in rates of forma-tion and decay. It must be stressed, how-ever, that no definite assignments ofchemical structures from the signals ispossible. More information is requiredthan the spectroscopic data provide.Likewise, the kinetic results which canprovide crucial information for under-standing of some of the dynamic intracel-lular processes involving the paramagnetic

species, do not allow definition of thechemical structures. Clearly, much furtherwork is needed to positively identify thespecies responsible for Peaks I to III in theA549 cells. Work toward this aim is inprogress.

The three signals (Peaks I to III)which we have found in these lung cellsfollowing Cr(VI) addition have not beenreported previously in any known system .It is not known whether these signals are

unique to lung cells, because other organshave not been examined by EPR tech-niques. Peaks IV and V, however, indicateby their resonance absorption at highermagnetic fields than the free electronvalue that they are caused by a Cr(V)entity. Thus, our findings are in agree-ment with the reports of others (4,5,7,8)that some form of Cr(V) is found inchemical and biochemical systems on thereduction of Cr(VI).

REFERENCES

1. Faria EC, Witmer CM. Differences in enzyme responses to in vivoand in vitro treatment. FASEB J 13:A292 (1989).

2. Faria EC, Witmer CM. Mixed function oxidase enzyme responsesto in vivo and in vitro chromate treatment. In: Biological ReactiveIntermediataes IV: Molecular and Cellular Effects and TheirIm act on Human Health (Witmer CM, Snyder R, Jollow DJ,KafGF, Kocsis JJ, Sipes IG, eds). New York:Plenum Press,1990;275-280.

3. Faria EC, Sadrieh N, Witmer CM. Effects of hexavalent chromium[Cr(VI)] on cytochrome (P450) isozymes. Toxicologist 13:8(1993).

4. Aiyar J, Berkovits HJ, Floyd RA, Wetterhahn KE. Reaction ofchromium(VI) with glutathione or with hydrogen peroxide: identi-fication of reactive intermediates and their role in chromium(VI)-induced DNA damage. Environ Health Perspect 92:53-62 (1991).

5. Kawanishi S, Inoue S, Sano S. Mechanism of DNA cleavageinduced by sodium chromate (VI) in the presence of hydrogen per-oxide. J Biol Chem 261:5952-5958 (1986).

6. Petrilli FL, DeFlora S. Metabolic deactivation of hexavalentchromium mutagenicity. Mutat Res 54:139-147 (1978).

7. O'Brien P, Barrett J, Swanson F. Chromium(V) can be generatedin the reduction of chromium(VI) by glutathione. Inorg Chim Acta108:L19-L20 (1985).

8. Goodgame DML, Joy AM. Relatively long-lived chromium(V)species are produced by the action of glutathione on carcinogenicchromium(VI). J Inorg Biochem 26:219-224 (1986).

9. Davies DM, Holdsworth ES, Sheriff JL. The isolation of glucosetolerance factors from Brewer's yeast and their relationship tochromium. Biochem Med 3:297-311 (1985).

10. Mertz W, Schwartz K. Chromium as a trace metal and essentialnutrient. Am J Physiol 196:614-618 (1959).

11. IARC. Organic and inorganic chromium compounds. In:Monographs on the Evaluation of Carcinogenic Risk of Chemicalsto Man. Lyon:International Agency for Research on Cancer,1973;100-125.

12. IARC. Chromium and chromium compounds. In: IARCMonographs for the Carcinogenic Risk of Chemicals to Humans:Some Metals and Metallic Compounds Vol 23. Lyon:InternationalAgency for Research on Cancer, 1980;205-323.

13. Bidstrup L, Case RAM. Carcinoma of the lung in workmen in thebichromates-producing industry in Great Britain. Br J Ind Med13:260-264 (1956).

14. Suzuki Y. Reduction of hexavalent chromium by ascorbic acid inrat lung lavage fluid. Arch Toxicol 62:115-122 (1988).

15. De Flora S, Morelli A, Basso C, Romano M, Serra D, DeFlora A.Prominent role ofDT diaphorase as a cellular mechanism reducingchromium(VI) and reverting its mutagenicity. Cancer Res

45:3168-3196 (1985).16. Petrilli FL, Rossi GA, Camoirano A, Romano M, Serra D,

Bennicelli C, DeFlora A, DeFlora S. Metabolic reduction ofchromium by alveolar macrophages and its relationship to cigarettesmoke. J Clin Invest 77:1917-1924 (1986).

17. Witmer CM, Park HS, Shupack SI. Mutagenicity and dispositionof chromium. Sci Total Environ 86:131-148 (1989).

18. Witmer CM, Harris R, Shupack SI. Oral bioavailability ofchromium from a specific site. Environ Health Perspect92:105-110 (1991).

19. Park HS, Witmer CM. Comparison of chromium-induced lesionsin cultured human and mouse cell lines. Toxicologist 8:158 (1988)

20. Kohn KW, Ewig RAG, Erikson RC, Zwelling LA. Measurement ofstrand breaks and crosslinks by alkaline elution. In: DNA Repair:A Laboratory Manual of Research Procedures, Vol 1, Part B.(Freidberg E, Hanawalt PH, eds). New York:Marcel Dekker,1981;379-491.

21. Snyder RD. Role of active oxygen species in metal-induced DNAstrand breakage in human diploid fibroblasts. Mutat Res193:237-244 (1988).

22. Sugiyama M. Role of physiological antioxidants in chromium(VI)-induced cellular injury. Free Radic Biol Med 12:397-407 (1992).

23. Jayyosi Z, Cooper KO, Thomas PE. Brain Cytochrome P450 andtestosterone metabolism by rat brain subcellular fractions: presenceof cytochrome P4503A immunoreactive protein in rat brain mito-chondria. Arch Biochem Biophys 298:265-270 (1991).

24. Thomas P E, Reik LM, Ryan DE, Levin W. Induction of twoimmunochemically related rat liver cytochrome P-450s,cytochrome P-450c and d by structurally diverse xenobiotics. J BiolChem 258:4590-4598 (1983).

25. Bass DA, Parce JW, Dechatelet LR, Szejda P, Seeds MS, ThomasM. 1983. Flow cytometric studies of oxidative product formationby neutrophils: a graded response to membrane stimulation. JImmunol 130:1910-1917 (1983).

26. Cooper KO, Reik LM, Jayyosi Z, Bandiera S, Kelley M, Ryan D,Daniel R, McClusky SA, Levin W, Thomas PE. Regulation of twomembers of the steroid-inducible cytochrome P450 subfamily (3A)in rats. Arch Biochem Biophys 301:345-354 (1993).

27. Srivastava SK, Beutler E. The transport of oxidized glutathionefrom the erythrocytes of various species in the presence of chro-mate. Biochem J 114:833-837 (1969).

28. Sugiyama M. Effects of vitamins on chromium(VI)-induced dam-age. Environ Health Perspect 92:63-70 (1991).

29. Costa M, Cantaonin 0, DeMars M, Swartzenruber DE. Toxicmetals produce an S-specific cell cycle block. Res Commun ChemPathol Pharmacol 38:405-419 (1982).

176 Environmental Health Perspectives