Bioavailable Transition Metals in ParticulateMatter Mediate Cardiopulmonary Injury inHealthy and Compromised Animal ModelsDaniel L. Costa and Kevin L. DreherPulmonary Toxicology Branch, Experimental Toxicology Division,National Health and Environmental Effects Research Laboratory,Research Triangle Park, North Carolina

Many epidemiologic reports associate ambient levels of particulate matter (PM) with humanmortality and morbidity, particularly in people with preexisting cardiopulmonary disease (e.g.,chronic obstructive pulmonary disease, infection, asthma). Because much ambient PM is derivedfrom combustion sources, we tested the hypothesis that the health effects of PM arise fromanthropogenic PM that contains bioavailable transition metals. The PM samples studied derivedfrom three emission sources (two oil and one coal fly ash) and four ambient airsheds (St. Louis,MO; Washington; Dusseldorf, Germany; and Ottawa, Canada). PM was administered to rats byintratracheal instillation in equimass or equimetal doses to address directly the influence of PMmass versus metal content on acute lung injury and inflammation. Our results indicated that thelung dose of bioavailable transition metal, not instilled PM mass, was the primary determinant ofthe acute inflammatory response for both the combustion source and ambient PM samples.Residual oil fly ash, a combustion PM rich in bioavailable metal, was evaluated in a rat model ofcardiopulmonary disease (pulmonary vasculitis/hypertension) to ascertain whether the diseasestate augmented sensitivity to that PM. Significant mortality and enhanced airwayresponsiveness were observed. Analysis of the lavaged lung fluids suggested that the milieu ofthe inflamed lung amplified metal-mediated oxidant chemistry to jeopardize the compromisedcardiopulmonary system. We propose that soluble metals from PM mediate the array ofPM-associated injuries to the cardiopulmonary system of the healthy and at-risk compromisedhost. Environ Health Perspect 1 05(Suppl 5):1053-1060 (1997)

Key words: particulate matter, metals, air pollution, lung injury, animal models, pulmonaryhypertension

IntroductionOver the last several years, a robust estimates across diverse urban airsheds,database has emerged from a series of epi- based primarily on outdoor mass metrics ofdemiology studies demonstrating a statisti- exposure, i.e., total suspended particulatescal association between ambient air (TSP) and PM.O pm in aerodynamicpollution particulate matter (PM) and diameter (PM1o) (3,4). Indeed, as the massmortality/morbidity among the exposed metric for PM is restricted to smaller sizehuman population (1-4). These studies ranges, the strength of the statistical corre-revealed remarkably consistent unit risk lation increases (i.e., PM2 5 > PM1o > TSP)

This paper is based on a presentation at The Sixth International Meeting on the Toxicology of Natural and Man-Made Fibrous and Non-Fibrous Particles held 15-18 September 1996 in Lake Placid, New York. Manuscriptreceived at EHP26 March 1997; accepted 17 July 1997.

The authors acknowledge the contributions of the staff of the Pulmonary Toxicology Branch in this researcheffort and the assistance of D. Doerfler with the statistical analyses. In addition, the authors thank L. Folinsbeeand L. Birnbaum for their constructive critiques of this manuscript. This manuscript has been reviewed andapproved for publication by the NHEERL. Mention of trade names or commercial products does not constituteendorsement or recommendation for use.

Address correspondence to Dr. D.L. Costa, Pulmonary Toxicology Branch, MD-82, National Health andEnvironmental Effects Research Laboratory; U.S. Environmental Protection Agency, ERC Building, ResearchTriangle Park, NC 27711. Telephone: (919) 541-2531. Fax: (919) 541-0026. E-mail: costa.dan@epamail.epa.gov

Abbreviations used: BAL, bronchoalveolar lavage; CFA, coal fly ash; DC, Washington; DOFA, fly ash from adomestic oil-burning furnace; Dus, Dusseldorf, Germany; IT, intratracheal; LDH, lactate dehydrogenase; MCT,monocrotaline; NIST, National Institute of Standards and Technology; Ott, Ottawa, Ontario, Canada; PM, partic-ulate matter; PM10, particulate matter< 10pm in aerodynamic diameter; PMN, polymorphonuclear neutrophil;ROFA, residual oil fly ash; St.L, St. Louis, MO; TSP, total suspended particulates; U.S. EPA, U.S. EnvironmentalProtection Agency.

(3,5,6). Subsequent studies have suggestedthat certain individuals may be at higherrisk for adverse effects of PM (3,7-10).The elderly with chronic cardiopulmonarydiseases, those with pneumonias, andthose with asthma at any age appear to beat higher risk. However, considerableuncertainty remains about specific bio-markers or biological mechanism(s) thatmight underlie the higher risk.

Although the epidemiology of ambientair PM is compelling, these findings sufferfrom the limited ability of epidemiologyto address the issue of causality. However,animal toxicology studies, though encum-bered with other limitations that relate toextrapolation, are capable of directlyaddressing questions regarding the mecha-nisms of response to PM with controlledempirical study designs. We used a toxico-logic approach to address two criticalquestions that have evolved from theepidemiology database:* Are there constituents common to PM

that could account for the consistencyin epidemiologic findings across suchdiverse exposure environments?

* Can animal models of cardiopulmonarydisease elucidate the etiology of appar-ent subpopulation susceptibilities toPM as identified by epidemiology?These two questions transcend the PM

issue and focus attention on the underlyingcausation and toxicologic potency of ambientair PM and how the physicochemical proper-ties ofPM link to the apparent potentiationof effects in predisposed individuals.

Ambient air PM1o is comprised of acomplex mixture of crustal and anthro-pogenic matter that are naturally separatedby size. The crustal material, consistinglargely of insoluble silicates derived fromabrasive processes, is typically larger in size(. 2.5 pm) and often is referred to as coarsemode PM. The PM fraction below 2.5 pmis considered fine mode; it arises from com-bustion processes or atmospheric transfor-mation of combustion emissions. The PMfine mode is of particular concern as apotential health threat. As this fractionremains airborne for long periods of time, itpenetrates the indoor air environment andenters the respiratory tract to deposit in theairways and deep lung.

Fly ash from fossil and waste fuelcombustion contributes > 2.5 x 105 tonsannually to the U.S. ambient PM burden(11). Fly ash is concentrated mostly inurban areas and can be considered a fine

Environmental Health Perspectives * Vol 105, Supplement 5 * September 1997 l1053

COSTAAND DREHER

mode PM -that is relatively rich in metalcontaminants (12). To address the hypoth-esis that PM-related health effects arisefrom anthropogenic combustion-relatedPM with bioavailable transition metals, weelected to use representative emission-source PM. The emission-source fly ashselected for study derives from the combus-tion of domestic and industrial grade fueloil and coal, and represents a range ofbioavailable constitutive metals. An isomet-ric approach was used (equal dose-mass orequal dose of metal) in an attempt to char-acterize the constituents from these sources,which are important determinants of theinflammatory response. In addition, weused several collected TSP samples toascertain the involvement of metals inambient PM-induced acute lung injury.Previous work with various types ofPM suggested that ionizable metals inPM correlate with indicators of inflamma-tion, altered host resistance, macrophagedysfunction, and airway hyperreactivity(13-16), but these studies have notattempted to manipulate mass and metaldosing to directly address this concept.Concomitantly, we established a model ofcardiopulmonary disease in the form ofadvanced pulmonary vasculitis/hypertensionto explore the question of susceptibilityand the possible mechanisms by which PMeffects are potentiated.

MethodsAnimals

Male Sprague-Dawley rats (60 days of age)were obtained from Charles River BreedingLaboratory (Raleigh, NC) and housed inan American Association for Assessment ofLaboratory Animal Care Internationalapproved animal facility for at least 1 weekbefore use. They were maintained at 22 to23°C, 50 to 60% relative humidity, and a

0600- to 1800-hr light cycle. Purina ratchow (Purina Mills, Richmond, IN) andwater were available ad libitum. Studygroup sizes varied.

Model Particulate MatterThree types of fly ash were used as emission-source particulates to model the combustioncomponent of fine mode PM. Residual oil flyash (ROFA) was collected as a fugitive ashdownstream ofan air-cleaning cyclone from apower plant burning industrial grade #6 low-S residual oil (13). Fly ash from a domesticoil-burning furnace (DOFA) that providedheat for a large apartment complex was col-lected as distal flue-deposited ash. It may beconsidered a mix of both fugitive and settledemission PM, and was provided by A. Ghioof the U.S. Environmental ProtectionAgency (U.S. EPA). The coal fly ash ([CFA]#4213.3) was collected by the EnergyAssessment and Control Division, NationalExposure Research Laboratory, U.S. EPA,from an experimental boiler furnace at theResearch Triangle Park, NC research facility.The source coal was Pittsburgh seam coal,burned at 825°C using Grove limestoneabsorbent (Ca/S ratio = 3/1).

St. Louis, MO (St.L) and Washington(DC) PM were purchased from theNational Institute of Standards andTechnology ([NIST] Gaithersburg, MD)as SRM #1648 and SRM #1649, respec-tively. Dusseldorf, Germany (Dus), andOttawa, Ontario, Canada (Ott), PM sam-ples were donated by G. Hatch, U.S. EPA,and R. Vincent, Health Ministry ofCanada, respectively. By their collectionmode, bagged house TSP samples are lessthan ideal. There is little selectivity in thecollection process except for the screeningof relatively large debris and vegetativefragments, and collection occurs usuallyover long periods of time with wide-rangingambient humidity and temperature. Thus,

it is likely that any losses or changes in PMchemistry due to inherent sample instabil-ity would have already taken place, even ifthe sample was subsequently stored undercontrolled conditions designed for preser-vation. In this sense, the samples used heremust also be considered less than ideal.They were not of similar age, nor werethey stored under similar conditions.Their age ranged from about 15 years(St.L, DC, Dus) to 3 years (Ott), and theywere either stored frozen (DC, St.L) or atroom temperature (Dus, Ott) in sealedbottles. However, each sample was sievedthrough a 30-100 mesh to remove largeextraneous debris. As such, though therewere likely to be differences in PM organiccharacter, we felt that the metal contentwould be relatively stable for the purposeof hypothesis testing. Indeed, comparisonsof the metal content of the St.L (SRM#1648) and DC (SRM # 1649) dust inour laboratory with that published by theNIST indicated no change. Similarly, theROFA dust analysis that we conducted formetals (16) was quite similar to that donein 1985 (13). We recently showed thatlow levels of PM-associated endotoxin didnot contribute to the in vivo toxicity ofthese ambient PM samples (17).

The basic physicochemical properties ofeach emission and urban PM studied areprovided in Table 1. The samples were pre-pared similarly for characterization of solu-ble metal and S (presented as SO4) content.To determine the acid-extractable metalcontent, PM samples were suspended in 1 MHCl in polypropylene tubes at a concentra-tion of 6.4 to 7.9 mg/ml, followed by mix-ing end-over-end for 1 to 2 hr at roomtemperature. The acid hydrolysates werethen centrifuged at 17,000 xg for 20 min,then the supernates were transferred to freshpolypropylene tubes for metal analysis.Analogously, double-distilled H20 was used

Table 1. Particulate matter physicochemical characterization.

Transition metal contenta Total Total Sizeb/Sample Fe Cu Ni V Zn metal sulfate geometric mean pHc

Emission PMDOFA 154.51 (85) 2.07 (80) 0.12 (100) 0.00 9.00 (100) 165.70 (86) 358.74 (95) 1.78/2.02 2.59ROFA 23.31 (68) 0.23 (97) 37.51 (92) 41.71 (84) 1.01 (95) 103.77 (83) 560.52 (100) 1.95/2.19 2.99CFA 14.57 (0) 0.03 (0) 0.70 (8.2) 0.37 (0) 0.08 (0) 15.75 (0.4) 49.36 (58) 4.17/5.00 9.72

Ambient air PMStL 8.91 (0.4) 0.37 (30) 0.11 (35) 0.11 (30) 3.56 (73) 13.06 (22) 142.66 (93) 3.72/4.42 5.22Dus 9.65 (0.5) 0.42 (21) 0.24 (49) 0.18 (31) 3.70 (68) 14.19 (20) 137.28 (100) 3.58/4.44 5.28Ott 6.33 (2.3) 0.82 (18) 0.27 (30) 0.21 (0.9) 10.63 (57) 18.26(35) 70.41 (90) 4.09/4.90 6.0DC 6.57(4.7) 0.14(33) 0.02(100) 0.20(76) 1.49(91) 8.42(23) 93.48(91) 3.27/3.88 3.65

"Transition metal content obtained from 1 M HCI hydrolysis of PM and expressed in pg/mg PM. Values in parenthesis represent percent of the metal that is water soluble.bMass median aerodynamic diameter of PM. cpH of a 10- to 11-mg/ml aqueous suspension.

Environmental Health Perspectives * Vol 105, Supplement 5 * September 19971054

METALS IN AMBIENT PM MEDIATE LUNG INJURY

to extract the water-soluble metals, startingwith a concentration of 10 to 11 mg/mi.The recovered supernate from this proce-dure was then acidified to pH < 2.0 with1 M HCI for metal analysis. The concentra-tions of Fe, V, Ni, Cu, Zn, and SO4= ineach extraction were quantified against spe-cific standards (Sigma Chemical, St. Louis,MO) using a PE Model P40 inductivelycoupled plasma emission spectrophotometer(Perkin Elmer, Norwalk, CT). PM aerody-namic size distributions were determinedusing a TSI 3310A aerodynamic particlesizer (TSI, St. Paul, MN). The acid-solubletransition metals (Cu, Fe, V, Fe, Zn) com-mon to all PM were used as the basis fortotal metal comparisons.

Particulate Maner Exposureand Bronchoalveolar lavageEach PM was administered by intratracheal(IT) instillation as previously described byDreher et al. (16). Briefly, a predeterminedmass of PM was diluted with sterile, injec-tion-grade saline (0.9%) to provide suffi-cient dosing mixture to dose all the animalsin a group. The rats were anesthetizedlightly (to lose gagging reflex and responseto pinch) with halothane and instilled ITwith 0.3 ml containing the desired PMsample, via a transoral tube placed underlaryngoscopic view (18). The animals werethen returned to their cages; they regainedconsciousness within approximately 10min. The treatment doses were based oneither a) administration of equal dose bymass (nominal 2.5 mg/rat) of each PM orb) normalization of each PM mass to atotal metal content of 46 pg/dose for theemission PM comparison (see Table 1 forthe appropriate metal content) and 35.5 pgof total metal (as Cu, Fe, V, Fe, Zn) for theambient PM and ROFA comparison.

At 24 and/or 96 hr following IT instil-lation, the rats were anesthetized withsodium pentobarbital (Nembutal, AbbottLaboratories, Chicago, IL) (50 mg/kg bodyweight, ip), then exsanguinated via theabdominal aorta. Rat tracheas were cannu-lated and bronchoalveolar lavage (BAL)performed with phosphate-buffered saline(Mg+2- and Ca+2-free) pH 7, using a vol-ume equivalent to 28 ml/kg body weight,infused and withdrawn three times.

BAL cell counts were determined using aCoulter Counter (Coulter, Miami, FL). Celldifferential analyses were performed on BALcytospin samples prepared using a Shandonmodel cytospin 3 (Shandon Instruments,Pittsburgh, PA) and stained with LeuKoStat(Fisher Chemical, Pittsburgh, PA). BAL

samples were centrifuged at 850 xg for1(0 min at 4°C. The recovered cell-freesupernates were used for the following bio-chemical analyses: total protein was deter-mined using the Coomassie Plus ProteinAssay (Pierce, Rockland, IL); albumin wasdetermined using the MALB SPQ kit(INCSTAR, Stillwater, MN); and lactatedehydrogenase (LDH) activity was deter-mined using kit #228 (Sigma Chemical). Allassays were adapted for automated analysisperformed on a Hoffmann-La Roche CobasFara II dinical analyzer (Roche Diagnostics,Branchburg, NJ).

Animal Model ofPulmonary HypertensionAn animal model was established to assessthe role of preexisting lung disease as apredisposing factor in PM responses.Pulmonary vasculitis/hypertension wasinduced in male Sprague Dawley rats 60days of age, using a single ip injection ofmonocrotaline ([MCT] Aldrich Chemical,Milwaukee, WI, 60 mg/kg), as previouslydescribed (19). After 10 to 12 days vascu-lar remodeling was well under way, andpulmonary arterial pressure was just bebeginning to increase (19). Pulmonaryartery pressure was measured in a limitednumber of rats (n = 3/group) by catheteriz-ing the pulmonary artery via the jugularvein in the anesthetized intact rat (20). Ina similar cohort of rats, BAL was conductedto ascertain the degree of lung inflamma-tion as described above as well as H2O2content (assayed by chemiluminescentreaction with isoluminol and microperoxi-dase). Control and MCT groups (n = 6)were treated IT with ROFA at doses of 0,0.25, 1.0, and 2.5 mg/rat. At 96 hr post-ITROFA, the rats underwent BAL as above.

Statistl AnalysisData were analyzed using one- or two-fac-tor analysis of variance and assessed foroverall differences. Pairwise testing (t test)was used, with adjustment for multiplecomparison. The hypothesis tested wasthat each PM sample would differ fromcontrol and that specifically for metal-based comparison the PM samples wouldnot differ among themselves. Significancewas assigned when the p value was < 0.05.

Results and DiscussionMetal-based ParticulateMatter ToxicityThis study attempted to evaluate furtherthe relative importance of bioavailable

metal to that of the mass dose of PM. Byanalogy, this approach was applied toambient urban PM. To this end, threeemission-source PM, representing a widerange of varied metal composition andbioavailability, were evaluated and com-pared in the context of four urban TSPfrom domestic and international sites. Thefindings were consistent with other studies,which indicated that the toxicity of PMgenerally correlated with total acid-solublemetal content, and that the oxidant poten-tial of the variously solubilized metalextracts might be a contributing mecha-nism (13-15). The variability in correla-tions both here and in the noted studiesmay reflect the varied oxidant potential oftransition metals, their multiplicity ofactions, or potential interactions betweenthe metals (16).

Table 1 displays the physicochemicalcharacterization of the fly ash and urban PMwe examined. Among the fly ash, there wasconsiderable variation in constituent acid-soluble metal; the DOFA contained about50% more total metal than ROFA andnearly 10 times that of the CFA. However,while the metal content ofROFA was largelydistributed among Fe, Ni, and V, DOFA-associated metal was primarily Fe, exceedingthat ofROFA by 7- to 10-fold. The Zn con-tent of DOFA was also relatively high com-pared to ROFA, and like much of theacid-available metal in oil fly ash, themajority was water soluble and thus easilybioavailable. This contrasted with the CFA,which contained a relatively small amount ofacid-extractable Fe and even less V, Ni, Cu,and Zn; thus no metal was readily bioavail-able. The pH values for oil fly ash were lowcompared to the CFA, which probablyreflects the relatively high SO4= content ofthe former. Oil fly ash is known to associatewith the metals during combustion cooling(K Olin, personal communication).

The ambient PM samples representedin Table 1 contained about 10% of theacid-available metal of the oil fly ash, ofwhich only a small amount was water solu-ble. Therefore, fly ash appears to be a rea-sonable, if significantly more potent,surrogate for the study ofPM mechanisms.However, it should be noted that becauseTSP comprises nearly the entire size rangeof ambient PM, the indicated metalbioavailablity likely underrepresents therelatively metal-rich respirable fine modePM that would be of most health concern(19). With regard to bioavailable metal,the TSP may be considered a mass-dilutedform of fine mode PM. In PM S04=, the

Environmental Health Perspectives * Vol 105, Supplement 5 * September 1997 1055

COSTAAND DREHER

*difference was less dramatic (< 50%) thanthat of oil fly ash. This probably reflectsthe contributions of other sources of S04=such as that from photochemical trans-formation of SO2 and perhaps stabilized byNH4+ (21). This would account for thelower acidity of TSP relative to that of oilfly ash.A wide range of metals was represented

in the ambient TSP samples, but weelected to focus on the predominant fivethat also were predominant in fly ash. Theparticle size of the ambient TSP sampleslikely does not represent that which existsin the atmosphere but rather is the productof agglomerations that occurred duringcollection. However, their size and relativemetal insolubility compared to oil fly ashprompted us to include the CFA to makecomparisons on the basis of dose mass ofthe PM challenge.

Figure IA to C illustrates the impactof IT-instilled fly ash on BAL parametersof altered permeability (total protein,albumin) and cellular injury (LDH) at 24and 96 hr after administration. The instil-lation of saline had no significant effecton these parameters when compared tocaged controls (K Dreher, unpublisheddata); similarly, acidic instillates have noeffect on these parameters (16). With aninstillate dose of approximately 2.5 mg flyash, DOFA had the greatest effect, withan acute (24 hr) and dramatic alterationin permeability and clear evidence of cel-lular injury (Figure 1D-F). By 96 hr, per-meability alterations reversed considerably,although the indicators of cell injuryremained high. ROFA also induced con-siderable injury but clearly less than that ofDOFA; their relative patterns of reversal,however, were similar. The degree ofresponse exceeded that predicted solely onthe basis of total metal. As DOFA's metalmix (a ratio of 75:1:17 for Fe:Cu:Zn) dif-fered from that of ROFA (1:1.5:2 forFe:Ni:V) and as there are known interac-tions among some transition metals (16),varied mechanisms may have been operant.Additionally, it appears that Zn may actdifferently (alone or interactively) in highconcentrations, as shown in another flyash that also contained V and Ni, albeit inlower concentrations (22). When com-pared to oil fly ash samples, CFA withmost of its metal tightly bound inducedlittle, if any, injury, and none of its para-meters differed from those of the salinecontrol. Clearly, the differential bioavail-ability of metal content was reflected in thedramatically different responses obtained

with DOFA, ROFA, and CFA whenadministered in an equimass manner.

To address the role of metals as the pri-mary determinant of injury, rats weretreated with each emission fly ash on anequimetal basis (- 46 pg per rat); mass dosevaried. As predicted, the initial injuriesinduced by these emission PMs did notdiffer appreciably from each other at the24-hr time point (Figure ID-F). However,by 96 hr, some significant differencesbecame apparent. In DOFA and CFA,where the predominant metal was Fe, theequimetal comparison shows similarresponses for all BAL inflammatory para-meters at 24 and 96 hr. In contrast, theROFA effect on the BAL parameters at 96hr persisted, which suggested either thatone of the metals not in the other fly ashhad a sustained impact (perhaps Ni) (16)or that metal mixture interactions in theequimetal dose of ROFA prolonged oraltered the longevity of the response, asreported previously (16). Thus, caution isnecessary in interpreting bioavailablemetal effects when multiple metals may beinvolved. However, a primary role of thebioavailable transition metals in themagnitude of the responses is clear.

The impact of fly ash on BAL cellularprofiles (Figure lA-C) when administeredin accordance with the equimass protocolyielded results in general agreement withthe BAL injury parameters. With themetal-rich oil fly ash, the macrophage andpolymorphonuclear neutrophil (PMN)influxes approached an order of magnitudehigher cell number when compared to theCFA, which had only a small amount ofbioavailable metal. When administered onan equimetal basis (Figure ID-F), theinflammatory cell responses are surprisinglysimilar in spite of minor fluctuationsthat are common to HAL cell data. Theeosinophils, normally considered in thecontext of allergic or parasitic inflamma-tion, respond to the instilled PM and simi-larly appear to be metal mediated. Thiseosinophil response to metals is in manyways a curiosity that is characteristic ofmetal exposure in the context of the PMresponse, and suggests a means of indictingmetal-based activity and anthropogeniccombustion sources in the assessment ofambient PM.

Examination of Figure 2A to F revealsthat the inflammatory responses to theurban PM are coherent with the oil fly ashsamples when linked by the hypothesis thatbioavailable metals drive the response.Therefore, ROFA studies could serve as the

positive control and, using the extensivedatabase developed in this laboratory, fur-ther insights could be developed. Theequimass dosing with urban PM samplesand ROFA clearly demonstrated the relativepotency of fly ash compared to urban PMsamples. With some variability in response,all four of the urban PM samples displayedabout 10% of the ROFA activity in terms ofBAL injury markers. On the other hand,when all the PMs including ROFA wereadministered in equimetal doses, the magni-tude of the change in the BAL parameterswas similar. Closer scrutiny of the datareveals that the Ott PM induces a small butconsistently greater response for all inflam-matory markers. Lacking an obvious expla-nation, one might argue that the excessresponse relates to the somewhat highercontent of water-soluble metal in this PM(Table 1) or perhaps a proinflammatory roleof Zn, as suggested by Gavett et al. (22).Alternatively, the Ott was also the freshest(most recently collected) PM of the groupstudied, although a loss of metal potencywith age of the PM seems unlikely. Thepotency of ROFA in this study is virtuallyidentical to that previously tested (13).

As with oil fly ash, there was anappreciable eosinophil response associatedwith the urban PM samples (Figure 2F).What is striking is that even with the dilu-tion effect of coarse mode PM mass onthat of the fine mode within TSP, theresponse persists. If one assumes thateosinophilia in BAL is a marker of metalassociated with fly ash as noted above,then this cell type possibly could serve as apotential marker of fine mode PM expo-sure. Recent studies (K Dreher, unpub-lished data) utilizing selectively collectedfine mode urban PM appear to supportthe contention that eosinophils can serveas such a marker. Likewise, the equimetalstudies demonstrated the remarkable con-sistency of eosinophilia across ambient PMsamples as well as with ROFA itself, whichlends further support to the notion thatcombustion source metals in PM play acontributory role in the biologic responsesto ambient PM.

In conclusion, it appears that theinduction of injury by emission andambient PM samples is determined pri-marily by constituent metals and theirbioavailability. Based on studies with flyash, the earliest phases of the responseappear to be driven by the individual met-als (16), but the persistence of theresponse perhaps reflects the complexityof the bioavailable metal mix or the

Environmental Health Perspectives * Vol 105, Supplement 5 * September 19971 056

METALS IN AMBIENT PM MEDIATE LUNG INJURY

*

SAL DOFA ROFA CFA SAL DOFA ROFA CFA

D1x107 -

*

_f x 106 - *

-a

CSiFxRF C50

C.)

lx 103.SAL DOFA SOFA CFA

B1 x 10

1= I T. -

.0

E *

2 lx1l S _ _SAL DOFA SOFA CFA

E

.

0z

SAL DOFA ROFA CFA

C

SAL DOFA ROFA CFA SAL DOFA ROFA CFA

Flx 106

a)

I

-J

._

0

.Ca0LU

SAL DOFA ROFA CRA SAL DOFA ROFA CFA SAL DOFA ROFA CFA SAL DOFA ROFA CFA

Figure 1. Emission source PM-induced acute pulmonary injury. Results were obtained from analyses performed on BAL samples at 24 and 96 hr postexposure. Rats wereexposed by IT instillation to saline (SAL), ROFA, DOFA, or CFA. Mass refers to responses obtained from animals exposed to an equivalent mass (-2.5 mg/rat) of each PM.Metal refers to responses obtained from animals exposed to an equivalent amount of total transition metal of each PM (46 pg/rat). This dose of metal translates to the fol-lowing mass dose of PM: ROFA, 0.450 mg/rat; DOFA, 0.282 mg/rat; and CFA, 2.97 mg/rat. Values are expressed as means +SEM. Animals per group, n=3 to 6 at 24 and 96 hrpostexposure. *, significant difference from saline control. *, significant difference from the other emission samples, excluding SAL.

unique qualities of one metal compared toanother. How the metals are solubilizedin vivo and interact to induce toxicityrequires further investigation, as redoxpotential is not the same for all the transi-tion metals and may depend in part on itscoordination state (23). The PM challenges

described herein involve complex mixturesof metals and other PM-associated com-pounds (i.e., small molecular weight oracidic water-soluble organics) for whichfurther study is also needed. Clearly, theseconclusions need to be confirmed usingactual inhalation exposures, which involves

both temporal and dosimetric relationshipsin need of examination.

Animal Model ofPulmonary HypertensionMCT is metabolized in the liver to thepyrrole that is specifically toxic to the

Environmental Health Perspectives * Vol 105, Supplement 5 * September 1997

A

E830

0~az

Ix10SAL DOFA ROFA CFA

1 057

COSTAAND DREHER

Mass r* A mtlDI x 104 - 1 x 10o

S~~~~~~~~~~~~~~~~~~~~~~~~~|x105

lxlI$ lxlOt

CD

|D Cj0* ~~~~~~~~~~~~~~~~~~~~~~~1X104-

Sa Duo StL DC Ott ROFA Sd Dus SL DC Ott ROFA Sal Dus StL DC Ott ROFA Sal Dus StL DC Ott ROFA

BE

lx~~~~~~~~~~~~~~~~~~~~~~~~~~~ o

=LI X10 0.

Sal Dus StLDC Ott ROFA Sal Dus St DC Ott ROFA imal Sal Dus StL DC Ott ROFA Sal DusSet DC Ont ROFA

C FIxi103 lIx io5-

pumo ary DuscutlarCenotthelium Sal) DuskSator forOtt Roxicty Sntalstudis StherC OtrastiRoFA SealBs StLe., OttaflstatThigue2inurba hasmbienti epM-indued acut pulmonayel jry Results werte obtainceentfomfnlye perore dotoxcit samplsti atW 24or postexposuenc stateswere expoedby ITistilltiotohe LpahoensistD,Otorf OApumo assrfr stof responstles RObtAie fromanimalstexpossbedoa eqiaethu maableof2mg/rat)lingchPMeletalrefes torrespertnsesotionedfoTypimalsy exposelesion equvaen invouvmnt oftalranstinmentatioofecPmetal-cpgat- Thedospecfirpsoses). ahP raolw:Ds

posresse tosgiicntdifernefrlomnalin contrtesol; alyzednifi dantdifrnce uromthe otherambentonMsaple,th- uinSL

by 10 to 12 days, worsens to cause right way. This is due to the substrate-rich Fe+3 +e-- -oFe+2cardiomegaly, then cor pulmonale and milieu of the lung lining fluid during Fe+2 + H202 <- OH" + Fe+3death over the ensuing 3 to 4 weeks (19). inflammation and the weakened state ofOur choice of the model was to establish a the blood-air barrier. The prototypicpreexistent state of progressive inflamma- Fen'ton reaction is depicted below for Fe In vitro assessment of the feasibility of thistory cardiopulmonary disease as a potential but can be considered applicable to most reaction with ROFA has been positive

Environmental Health Perspectives * Vol 105, Supplement 5 * September 19971 058

METALS IN AMBIENT PM MEDIATE LUNG INJURY

Table 2. Lung inflammation in the monocrotaline cardiopulmonary disease model after 10 to 12 days.

BAL end points Control Monocrotaline

Protein, pg/ml 188.3 ± 30.0 3951.8 ± 423.5*LDH, U/liter 28.3 ± 9.7 423.5 ± 33.5*PMNs/ml x 103 45.9 ± 0.39 462.8 ± 120.0*H202, V/0.4x 106 cells 0.0063 ± 0.0011 0.01 75 ± 0.002*

*p<0.01 compared to control.

(14,25), but the plausibility of this reactionin the in vivo model with PM has not beenverified to date. Theoretically, lung inflam-mation can result in the release of the reduc-tant superoxide and labile H202 from lungcells, which in the presence of lipids canreact in vitro with ROFA to yield more oxi-dants (D Costa, unpublished data). Morework is needed to assess this hypothesis inlight of other pathophysiologic frailties ofthe impaired animal. 0

Over the 10- to 14-day incubationperiod after MCT injection, the rats lostlittle body weight (- 10%); however, pul-monary artery pressure measured in a smallcohort of similarly treated rats indicated adoubling of Ppa from 10 to 20 mm Hg.Baseline BAL values in Table 2 for controland MCT rats reflect the inflammatorystate of the MCT model at the point of ITinstillation ofROFA.

Ninety-six hours after ROFA instillation,there was clear evidence of enhanced neu-trophilic inflammation (Figure 3), with theinflux of PMNs appearing to plateau at 1mg ROFA/rat. However, other BAL endpoints did not show a consistent picture ofenhanced injury because of great variabilityin response (data not shown). Notably, by96 hr, significant mortality had occurred inthe MCT-ROFA groups (050% across alldoses of ROFA), with the remaining animalsin a somewhat more impaired state thaneither MCT or ROFA alone. Thus, interpre-tation of any 96-hr data is representative ofthe survivor response. Table 3 sums ourexperience with the MCT model regardingmortality. To date, there does not appear tobe a dose response over the range of dosesevaluated; time to death also appears not tobe dose dependent. Preliminary studies inour laboratory with Mt. St. Helens volcanic

ash, which has virtually no bioavailable metalor significant toxicity in healthy animals, alsowas innocuous in the MCT animal. Mt. St.Helens volcanic ash was collected for theU.S. EPA by A McFarland, Texas A&MUniversity, from an open field near Ritzville,WA 340 km NE of the volcano. Thus thePM-associated metal component seems to bean essential factor for amplification.

The cause of death in the MCT-ROFAanimals appears related to altered cardiacfunction, although it is unclear whetherdeath was associated with direct cardiacinjury or was secondary to pulmonary fail-ure (26). Electrocardiographic studyrevealed that ROFA alone in the healthyrat clearly induced a mild arrhythmic con-dition over the acute inflammatory timeperiod (96 hr). However, animals pre-treated with MCT exhibited significantlymore and worsened dysrhythmias, amongwhich bundle branch A-V block and S-Tsegment inversion were considered mostsevere. These alterations progressed withtime and correlated with time of death.Though their origin is as yet uncertain,these cardiographic changes appear consis-tent with the formation of ectopic foci inheart muscle and/or hypoxic ischemia, pos-sibly secondary to pulmonary diffusionabnormalities (S Gardner, unpublisheddata). This model has not yet been utilizedwith ambient PM in this laboratory; how-ever, Godleski et al. (27) reported mortal-ity in a similar model with inhalation ofconcentrated ambient Boston, MA air(-350 pg/M3 for 6 hr/day for 3 days).

ConclusionWe propose that soluble metals from PMmediate the array of PM-associated injuriesto the cardiopulmonary system of both the

400 SAL rat model_ MCT rat model b,e

350-

-E b,d3300 b.c

x 250

m

O 100 ba-

so as-' a,d a,s

00mg 0.25mg 1.0mg 2.5mg

ROFA Dose, mg

Figure 3. Neutrophilic response in the BAL fluid in arat model of pulmonary vasculitis/hypertensioninduced by a single ip injection of MCT 96 hr after ITinstillation of various doses of ROFA in 0.3 ml saline.Potentiation of the response occurs at each dose ofROFA; these animals represent the survivor cohort(-50%) at the time of analysis. Like-lettered super-scripts indicate significantly different from the appro-priate control or ROFA alone (- MCT) group.

Table 3. Mortality of monocrotaline rat model treatedwith intratracheal residual oil fly ash.

Treatment Mortality by 96 hr

MCT + saline 4/46MCT + 0.25 mg ROFA 8/16MCT + 1.0 mg ROFA 10/16MCT + 2.5 mg ROFA 25/46

healthy and the at-risk compromised host.When adjusted for bioavailable metal con-tent, combustion source and ambient PM(TSP) show remarkable similarity in theinflammatory response. The data in thecompromised animals appear to parallel therecent epidemiology studies suggesting thatthose at enhanced risk possess chronicinflammatory lung conditions. The currentneed is to ascertain the relevance of thesedata and validate the hypothesis that metalsmediate toxicity when applied to inhalationexposure scenarios that approximate similardoses in humans.

REFERENCES

1. Schwartz J, Dockery DW. Increased mortality in Philadelphiaassociated with daily air pollution concentrations. Am RevRespir Dis 145:600-604 (1992).

2. Pope CA III, Dockery DW. Acute health effects of PM1o pol-lution on symptomatic and asymptomatic children. Am RevRespir Dis 145:1123-1128 (1992).

3. Dockery DW, Pope CA III, Xu X, Spengler JD, Ware JH, FayME, Ferris BG Jr, Speizer FE. An association between air pol-lution and mortafity in six U.S. cities. N Engl J Med329:1753-1759 (1993).

4. Schwartz J. Air pollution and daily mortality: a review andmeta analysis. Environ Res 64:36-52 (1994).

Environmental Health Perspectives * Vol 105, Supplement 5 - September 1997 1059

COSTAAND DREHER

5. Schwartz J. PM1o, ozone, and hospital admissions for theelderly in Minneapolis, MN. Arch Environ Health 49:366-374(1994).

6. Schwartz J, Dockery DW, Neas LM. Is daily mortality associ-ated specifically with fine particles? J Air Waste Manage Assoc46(10):927-939 (1996).

7. Schwartz J. What are people dying of on high air pollutiondays? Environ Res 64:26-35 (1994).

8. Schwartz J. Air pollution and hospital admissions for theelderly in Detroit, Michigan. Am J Respir Crit Care Med150:648-655 (1994).

9. Schwartz J, Slater D, Larson TV, Pierson WE, Koenig JQ.Particulate air pollution and hospital emergency room visits forasthma in Seattle. Am Rev Respir Dis 147:826-831 (1993).

10. Schwartz J, Morris R. Air pollution and hospital admissions forcardiovascular disease in Detroit, Michigan. Am J Epidemiol142:23-35 (1995).

11. EPA Office of Air Quality Planning and Standards. NationalAir Quality Trends Report, 1993. EPA 454/R-94-026.Washington:U.S. Environmental Protection Agency, 1994;52.

12. Schroeder WH, Dobson M, Kane DM, Johnson ND. Toxictrace elements associated with airborne particulate matter: areview. Environ Sci Technol 9:838-845 (1987).

13. Hatch GE, Boykin E, Graham JA, Lewtas J, Pott F, Loud K,Mumford JL. Inhalable particles and pulmonary host defense:in vivo and in vitro effects of ambient air and combustion parti-cles. Environ Res 36:67-80 (1985).

14. Pritchard RJ, Ghio AJ, Lehmann JR, Winsett DW, Tepper JS,Park P, Gilmour MI, Dreher KL, Costa DL. Oxidant genera-tion and lung injury after particulate air pollutant exposureincrease with the concentrations of associated metals. InhalToxicol 8:457-477 (1996).

15. Ghio AJ, Stonehuerner J, Pritchard RJ, Piantadosi CA, QuigleyDR, Dreher KL, Costa DL. Humic-like substances in air pollu-tion particulates correlate with concentrations of transition met-als and oxidant generation. Inhal Toxicol 8:479-494 (1996).

16. Dreher KL, Jaskot RH, Lehmann JR, Richards JH, McGee JK,Ghio AJ, Costa DL. Soluble transition metals mediate residual

oil fly ash induced acute lung injury. J Toxicol Environ Health50:285-305 (1997).

17. Jaskot R, Dreher K. Unpublished data.18. Costa DL, Lehmann JR, Harold WM, Drew RT. Transoral

tracheal intubation of rodents using a fiberoptic laryngoscope.Lab Animal Sci 36(3):256-261 (1986).

19. White SM, Roth RA. Progressive lung injury and pulmonaryhypertension from monocrotaline. In: Handbook of AnimalModels of Pulmonary Disease, Vol II (Cantor JO, ed). BocaRaton, FL: CRC Press, 1989; 75-97.

20. Stinger RB, Iacopino VJ, Alter I, Fitzpatrick TM, Rose JC, KotPA. Catheterization of the pulmonary artery in the closed-chestrat. J App1 Physiol Respir Environ Exercise Physiol51(4):1047-1050 (1981).

21. U.S. EPA. Air Quality Criteria for Particulate Matter, Vol I.EPA/600/P-95/00 1 aF. Washington:U.S. EnvironmentalProtection Agency, 1996.

22. Gavett SH, Madison SL, Dreher KL, Winsett DW, McGee JL,Costa DL. Metal and sulfate composition of residual oil fly ashdetermines airway hyperreactivity and lung injury in rats.Environ Res 72:162-172 (1997).

23. Ghio AJ, Kennedy TP, Whorton AR, Crumbliss AL, HatchGE, Hoidal JR. Role of surface complexed iron in oxidant gen-eration and lung inflammation induced by silicates. Am JPhysiol 263:L51 1-L518 (1992).

24. Wilson DW, Segall HJ, Pan LC, Lame MW, Estep JE, MorinD. Mechanisms and pathology of monocrotaline pulmonarytoxicity. Crit Rev Toxicol 22(5/6):307-325 (1992).

25. Meng ZH, Ghio AJ, Hatch GE, Costa DL. Luminol-enhancedchemiluminescence after in vitro exposures of rat alveolarmacrophages to oil fly ash is metal dependent. Inhal Toxicol (inpress).

26. Watkinson WP, Campen MJ, Costa DL. Arrhythmia induc-tion after exposure to residual oil fly ash particles in the pul-monary hypertensive rat. Fundam Appl Toxicol (in press).

27. Godleski JJ, Sioutas C, Katler M, Koutrakis PO. Death frominhalation of concentrated air particles in animal models of pul-monary disease. Am J Respir Crit Care Med 1 53:A1 5 (1996).

1060 Environmental Health Perspectives * Vol 105, Supplement 5 * September 1997