REVISED DRAFT 1/27/2004, Page 1 Ethyl acrylate risk assessment with a hybrid computational ...

47 Pages·2004·0.39 MB·English

TOXICOLOGICALSCIENCES79,394–403(2004)
DOI:10.1093/toxsci/kfh116
AdvanceAccesspublicationMarch31,2004
Ethyl Acrylate Risk Assessment with a Hybrid Computational Fluid
Dynamics and Physiologically Based Nasal Dosimetry Model
Lisa M. Sweeney,*,1 Melvin E. Andersen,† and Michael L. Gargas*
*TheSapphireGroup,Dayton,Ohio45431;and†CIITCentersforHealthResearch,ResearchTrianglePark,NorthCarolina27709
D
o
ReceivedOctober13,2003;acceptedFebruary15,2004 w
n
lo
(National Toxicology Program, 2002), ethyl acrylate was re- ad
Cytotoxicity in the nasal epithelium is frequently observed in moved from this listing. The decision to delist EA was based ed
rodentsexposedtovolatileorganicacidsandestersbyinhalation. onthelackoftumorigenicitybyotherroutes,evidencelinking from
An interspecies, hybrid computational fluid dynamics and physi- cancer in the gavage studies with site- and concentration- h
ologically based pharmacokinetic (CFD-PBPK) dosimetry model ttp
dependentirritation,andtheunlikelihoodofsignificanthuman s
forinhaledethylacrylate(EA)isavailableforestimatinginternal exposure. Thus, concerns regarding the chronic toxicity of ://a
dose measures for EA, its metabolite acrylic acid (AA), and EA- c
a
mediated reductions in tissue glutathione (GSH). Nasal tissue ethyl acrylate to humans pertain to noncancer effects. This de
m
concentrationsofAAwerepreviouslyusedasthedosemetricfor work identifies an appropriate Reference Concentration (RfC) ic
a chronic Reference Concentration (RfC) calculation with this for chronic noncancer effects of ethyl acrylate for use in risk .o
u
compound.However,EAwasmoretoxicthanexpected,basedon assessments for potentially exposed populations. p.c
o
calculatedtissueAAconcentrations.UnlikeAA,EAcausesdeple- Similar to findings with other esters and volatile organic m
tionoftissueGSH.WehavedevelopedanRfCforEAusingtissue acids(Bogdanffyetal.,1987;Mainwaringetal.,2001;Miller /tox
GofSnHasdaelptliesstiuoendinosteh.eTohlefahcytobrryidepCiFthDel-iPuBmPaKsmthoedperlimwaasryremfienaesdutroe et al., 1981), inhaled ethyl acrylate damaged the rodent nasal sci/a
improve the accuracy of simulations for GSH in rat olfactory olfactoryepitheliuminacuteandchronicstudies(Fredericket rtic
al.,2002;Milleretal.,1985).IntheMilleretal.(1985)study, le
tissues. This refined model was used to determine the concentra- -a
tion for continuous human exposures to EA predicted to reduce mice and rats were exposed to ethyl acrylate for 6 h/day, 5 bs
nasalGSHlevelstothesameextentasseeninratsexposedtoEA days/week for 27 months. This study established a no-ob- trac
attheno-observed-effectlevel(NOEL).Importantly,AAconcen- served-effectlevel(NOEL)of5ppmfornonneoplasticeffects t/7
9
trationsinthehumannasalolfactoryepitheliumattheproposed and a lowest observed effect level (LOEL) of 25 ppm for /2
/3
chronic RfC were predicted to be lower than the AA concentra- olfactory tissue damage in both rats and mice. Other effects, 94
tionsestimatedintheratattheNOEL.Thus,achronicRfCbased suchasirritancyatthebeginningofexposure,lethargylaterin /16
on maintaining GSH in the human nasal olfactory epithelium at exposure,anddecreasedbodyweightgains,occurredathigher 493
levelsequivalenttotheratNOELwouldalsoprovideanadequate 6
ethylacrylateconcentrations(Milleretal.,1985).Usingnasal 3
margin of safety with respect to AA concentrations in nasal tis- b
sues. explants,Fredericketal.(2002)havedemonstratedthatolfac- y g
u
Key Words: ethyl acrylate; acrylic acid; glutathione; olfactory tory lesions similar to those produced by in vivo exposure to es
epithelium;referenceconcentration. ethyl acrylate may be produced by in vitro exposure to ethyl t o
n
acrylate.Theselesionsarenotproducedinthepresenceofthe 3
0
esterase inhibitor, paraoxon. Furthermore, these lesions are M
a
similar to those produced in vivo and in vitro by acrylic acid rc
Ethyl acrylate (EA; CAS No. 140–88–5) is a colorless h
(AA) (Frederick et al., 1998). Similarly, the concentration of 2
liquidusedasanintermediateinproductionofplasticresinsfor 0
1
an acid metabolite has been related to nasal and reproductive 9
industrial and consumer use. Ethyl acrylate induced forestom-
effectsofotherrapidlymetabolizedcompounds(Bogdanffyet
ach tumors in rodents administered high doses by gavage
al., 1987; Sweeney et al., 2001; Welsch et al., 1995). These
(National Toxicology Program, 1986) and was included in a
findings appear to support a mode of action for inhaled ethyl
listofcompoundsknownorreasonablyexpectedtobehuman
acrylate related to the concentration of AA in the olfactory
carcinogens. However, in the Tenth Report on Carcinogens
epithelium.
The chronic inflammation associated with exposure to AA
1To whom correspondence should be addressed at The Sapphire Group,
and its esters is believed to be the result of induction of the
2661CommonsBoulevard,FirstFloor,Dayton,OH45431.Fax:(937)458-
0050.E-mail:[email protected]@thesapphiregroup.com. mitochondrial permeability transition (MPT) by AA. This
chronic inflammation ultimately leads to the nonneoplastic
Toxicological Sciences vol. 79 no. 2 © Society of Toxicology 2004; all rights
reserved. lesions localized to the olfactory epithelium (Custodio et al.,
394
EA CFD-PBPK CHRONIC RISK ASSESSMENT 395
1998).Inthecaseofethylacrylate,esteraseactivity(leadingto
AAproduction)appearstobenecessarytoproducedamageto
the epithelium, since no damage was observed when rat nasal
explants were incubated with high concentrations of ethyl
acrylate(upto1mM)inthepresenceoftheesteraseinhibitor,
paraoxon (Frederick et al., 2002). Depletion of hepatic GSH
and other forms of oxidative stress enhance mitochondrial
permeability(Haouzietal.,2001;Hoeketal.,2002).Asingle
exposure of rats to ethyl acrylate at the LOEL for olfactory
damage, 25 ppm for 6 h, depleted olfactory epithelial GSH to
35–58%ofthelevelincontrols(Fredericketal.,2002).While D
o
AA appears to be required for nasal toxicity of ethyl acrylate, w
n
the depletion of GSH may enhance it. Sustained depletion of lo
a
d
GSHwascriticalforthedevelopmentofforestomachtumorsin e
d
rats administered ethyl acrylate by gavage (Frederick et al., fro
1992). GSH depletion has been implicated in acute lethality m
FIG. 1. Computational fluid dynamics-physiologically based pharmaco- h
and in biochemical effects of ethyl acrylate (Frederick et al., kineticmodelforethylacrylateanditsmetaboliteacrylicacid(AA)inratsand ttp
s
2002; Vodicka et al., 1990). humans.R(cid:1)respiratoryepithelium,O(cid:1)olfactoryepithelium.Solidarrows ://a
Thechronictoxicitystudiesforinhaledethylacrylateinrats denotebloodflow,arrowswithlargedashesdenoteairflow(cyclicbreathing), ca
(Miller et al., 1985) can form the basis for developing a safe anddottedarrowsdenotemetabolism.AdaptedfromFredericketal.(2002). de
m
exposure level for humans. While these interspecies extrapo- ic
.o
lationscanbedoneusingdefaultprocedures,theavailabilityof u
the mode of action information and a computational fluid bsaymthpelersatainodofinrartaetstoisfsmueet(hMylaminewthaarcirnyglaetet aml.e,ta2b0o0l1is)m. Tihnehuadmjuanstendasraaltbtiiospsusye p.com
dynamics–physiologically based pharmacokinetic (CFD- ratesarethehighest(104and63(cid:1)molmetabolized/mloftissue/hforolfactory /to
PBPK) model for ethyl acrylate and its metabolite AA in rats and respiratory epithelium, respectively), and the human tissue rates are the xs
and humans (Frederick et al., 2002) permit use of more infor- lowest (18.1 and 23.0 (cid:1)mol metabolized/ml of tissue/h for olfactory and ci/a
mation in the dosimetry calculation (U.S. EPA, 1994). First, respiratory epithelium). We elected to use the rates determined in human rtic
autopsy tissue, because they are derived from human tissue using the com- le
the interspecies extrapolation of the target tissue dose metrics poundofconcern.Inaddition,thesamplesizes(n)aregreaterthanthesample -ab
can be conducted with this model. Secondly, potential phar- sizesforotherspecies. stra
macokineticvariability/uncertaintyamonghumanscanalsobe Theparametervaluesusedinthecurrentstudyarethesameasthoseused c
estimatedbyexaminingmodeloutputfordistributionsofinput inpreviouslypublishedCFD-PBPKmodelsforethylacrylateandAA(Fred- t/79
ericketal.,1998,2002)withthefollowingexceptions.Forahumanengaged /2
parameters. inlightwork,cardiacoutputisincreasedfrom3.9to6.0ml/sec-kg0.74(Brown /39
4
etal.,1997).Tofacilitatesensitivityanalysis,certainflowcalculationswere /1
6
METHODS modified to ensure mass balance. Specifically, blood flow to the slowly 4
9
perfusedtissuescompartment(muscleandfat)wascalculatedasthedifference 3
6
Literature evaluation and preliminary modeling. Available studies po- betweenthecardiacoutputandthesumofflowstotheliver,kidneys,other 3 b
tentiallyrelevanttothedeterminationofachronicRfCforethylacrylatewere richlyperfusedtissues,nose,andnasopharynx.Likewise,thefractionofthe y g
providedbythesponsorandidentifiedbyliteraturesearchesconductedbythe nasal blood flow perfusing the respiratory epithelium was calculated as the ue
astuutdhyo,rseinndtphoeinatr,eamsoodfeetohfyalcatciroynl,ataenadndanAiAntetornxaiclitdyoasnedmdeistrpiocs(itoilofna.ctAorcyrictiocna-l depififtehreelniucmebaentdwetheenntoatsaallbvleosotidbuflloew. tothenoseandtheflowtotheolfactory st on
centrationofAA)wereselected.Preliminarymodelingwasconductedusing The Frederick et al. (2002) model predicted that the olfactory epithelial 30
themodelofFredericketal.(2002),describedintheAppendix.Evaluationof concentrationofGSHassociatedwiththeethylacrylateNOEL(5ppm)after M
these results in light of the AA literature (described in Results) prompted 6hofexposurewouldbe95.9%oftheinitialvalue.Theexperimentaldata, arc
additional literature evaluation of a potential role for GSH in the mode of however,indicatedgreaterdepletion,toapproximately75%ofcontrolatthe h 2
action.BecauseolfactoryGSHconcentrationwasidentifiedasapotentialbasis NOELand40%attheLOEL(25ppm)(Fredericketal.,2002).Tousethe 01
forchronicRfCderivation,theFredericketal.(2002)modelwasrefinedto modelforinterspeciesextrapolationofolfactoryGSHconcentrations,therat 9
improveitsabilitytodescribeGSHdepletionintheratolfactoryepithelium. model was adjusted to provide a more accurate prediction of the measured
GSH concentrations. This calibration was accomplished by introducing an
adjustmentfactorforGSHconjugationintheolfactoryepithelium(KGSHO)
CalculationofAAandGSHConcentrationsinDorsalOlfactoryEpithelium
toreflectarateincreaseovernonenzymaticratesmeasuredinvitro.Avalueof
CFD-PBPKmodel. PredictionoftheAAconcentrationandGSHconcen- KGSHOequalto10accuratelysimulatedolfactoryGSHconcentrationsfor6h
trationinthedorsalolfactoryepitheliumofaratorhumanexposedtoagiven exposureofratsto5or25ppmethylacrylate(Fig.2)withoutsignificantly
concentration of ethyl acrylate was performed using a refinement of the affectingtissueAApredictions.ItwasassumedthatKGSHOwouldnotdiffer
CFD-PBPKmodeldevelopedbyFredericketal.(2002),depictedinFigure1 between species. While this adjustment factor, KGSHO, was determined
and described in the Appendix. Three possible values for the rate of ethyl empirically,itisconsistentwiththeliterature.PotterandTran(1993)deter-
acrylatemetabolisminhumannasaltissueswereconsidered(Fredericketal., minedthattherateofsecond-orderconjugationofethylacrylateandGSHwas
2002): the rate determined in human autopsy tissue, the rate determined in increased15-foldinlivertissueduetothepresenceofglutathionetransferases
monkeytissue,andtherateofethylacrylatemetabolismrattissuemultiplied (GSTs).Bangeretal.(1993)determined,usingvariousGSTsubstrates,that
396 SWEENEY, ANDERSEN, AND GARGAS
exhibitingdifferingpharmacodynamics.Thedefaultpharmacodynamicadjust-
ment factor for animal-to-human pharmacodynamic differences is 3 (U.S.
EPA,1994).Anadjustmentfactorforintraspecies(human)variabilityof10,to
accountforpharmacokineticandpharmacodynamicdifferencesinresponseto
thetoxiccompound,wasalsoapplied.Potentiallysusceptiblesubpopulations
wereconsidered.
The adjustment factor for intraspecies pharmacokinetic and pharmacody-
namicvariabilitywasevaluatedinlightofsensitivityanalysisofnasalolfac-
toryepitheliumGSHandAAconcentrationsattheNOELforratsandatthe
proposed RfC for humans. Briefly, normalized sensitivity coefficients (SC)
werecalculatedforchangesinpredicteddosemetricsofpotentialinterestfor
changesinmodelparametervalues.Intheanalysespresentedhere,sensitivity
(normalized sensitivity coefficient, SC) was determined by increasing the D
o
parameter values by 1% and dividing the fractional change in the model w
predictionbythe1%fractionalchangeintheinputparameter. nlo
a
Variability analysis. A variability analysis was conducted to assess the d
e
impact of parameter variability/uncertainty on the predicted olfactory tissue d
AAandGSHpredictionsforthehumansexposedtotheethylacrylateatthe fro
m
proposedRfC.Thevariabilityanalysisaccountsforboththemodelsensitivity h
andtheknownorestimatedvariabilityintheinputparameter(Licataetal., ttp
s
2is00p1re;dSiwcteeednuesyinegtaal.,p2ro0p0a3g).atTiohneaopfperrorxoirmfaotremCuVlaf(oVrathrdeemmoadne,l1o9u9t4p,utci(tCedVmin) ://ac
a
Licataetal.,2001,withformulassubstitutedtobeintermsofnormalizedSC) d
e
FIG. 2. Glutathione (GSH) concentration (percent of control) in the ol- shownbelow. m
ic
factory epithelium of the dorsal meatus of rats exposed to ethyl acrylate by (cid:1)(cid:2) .o
i(nbhoattloamtio,nth.iLnilninese:),MreofidneelmpernetdoicftFiornesdefroicrk5etpaplm.(2(0to0p2,)mthoicdkel.liEnex)p,er2i5mepnptmal CVm(cid:2) (cid:3)(cid:3)normalized_SCi(cid:4)2(cid:3)CVi(cid:4) up.co
dataofFredericketal.(2002)(mean(cid:6)SD)(cid:5)(cid:1)5ppm,(cid:2)(cid:1)25ppm. i m
/to
x
tWivhiteyrecoieifnfidciiceanttesretlhaetiviethtpoarcahmanegteers,innoprmaraalmizeetderSiC,ainisdtCheVniosrmthaeliczoeedffisecniesni-t sci/a
twhoeualdctitvhiutysoafntoilcfiapcattoerythGatSTKsGrSanHgOed, fthroemin4c0retaose90in%GofSHthactoinnjuligvaetrio.nOnine ofvariationinparameteri. i rticle
olfactoryepithelium,wouldbe40to90%of15,orbetween6and13.5.Itis CVivalueswerecalculatedfromexperimentaldataorestimatestakenfrom -a
notclearwhythisincreasewasnotdetectedinvitrobyFredericketal.(2002). tthheepsaumbleishdeodselimtereatrtuicrseffoorrpwahraicmhettheersswenitshit(cid:3)iSvCity(cid:3)(cid:5)an0a.l2y.sCesVwmewreasccoanldcuuclatetedd.for bstra
Computations. SimulationswereperformedusingACSLSim11.8(Aegis c
Technologies Group, Hunstville, AL) on a Dell Optiplex GX260 computer t/7
9
withaPentium4processororDellInspiron7500withaPentiumIIprocessor. RESULTS /2
/3
9
4
InterspeciesExtrapolationandApplicationofAdjustmentFactorsforCFD- Selection of Critical Endpoints and Studies for the Chronic /1
PBPKModelingApproach RfC 649
3
Whenusinganinternaldosimetrymodel,adjustmentfactorsfordurationand 6
3
pharmacodynamicdifferencesaremostlogicallyappliedtothetargetinternal Miller et al. (1985) identified nasal lesions of the olfactory b
y
dose prior to calculation of the human-equivalent continuous exposure con- epithelium in mice and rats that were exposed to 25 but not 5 g
u
centration(Andersenetal.,2000).Thepredictedhuman-equivalentexposure ppm ethyl acrylate for 6 h/day, 5 days/week for 27 months. e
s
concentration for a human with average pharmacokinetic and physiological Since effects with NOELs higher than the olfactory toxicity t o
n
parametervaluesisthendeterminedbyiterativelytestingdifferentexposure NOELmayleadtolowerRfCsbecauseofunequaluncertainty 3
concentrations and calculating internal dose. The resulting human exposure factors,allrelevantendpointsshouldbeconsidered.Irritancyat 0 M
concentrationwasthenusedasthebasisconcentration,andahumanpharma- a
cokinetic sensitivity and variability analyses were conducted at this inhaled thebeginningofexposureandlethargyattheendofexposure rch
concentration. were noted in animals exposed to 225 ppm but not 75 ppm 20
1
Time adjustment of internal dose. Area under the curve (AUC), rather ethyl acrylate and the high-dose (225 ppm) animals had sig- 9
thanpeakconcentration,ofAAwasselectedasanappropriateinternaldose nificantlydecreased((cid:5)10%)bodyweightgainsrelativetothe
metric based on the studies of Lomax et al. (1994) that demonstrated that
controls (Miller et al., 1985). Ethyl acrylate does not exhibit
exposuresofmicetoAAwithequivalentconcentration(cid:2)time(C(cid:2)T)(25
selective developmental toxicity (Saillenfait et al., 1999). A
ppmfor4.4h/dayor5ppmfor22h/day)producedsimilarincidenceofnasal
irritation.SincesteadystateconcentrationsofAAintheolfactoryepithelium two-generation reproduction study of ethyl acrylate has not
areexpectedwithin1h(Fredericketal.,2002),timeadjustmentforcontinuous beenconducted,butthelackofreproductivetoxicityofAAin
exposurewasperformedbymultiplyingthetissueconcentrationachievedin rats at doses up to and including 502 mg/kg/day in drinking
the rat after 1 h of exposure to ethyl acrylate at the NOEL (5 ppm) and
water (Hellwig et al., 1997) suggests that ethyl acrylate is
multiplyingby(6h/24h)(cid:2)(5days/7days)toconvertfromtheschedulein
unlikely to be selectively toxic to reproduction. Hence the
Milleretal.(1985)toaccountforcontinuousexposure.
relevantendpointsareolfactorylesions,lethargy,irritancy,and
Adjustment factors. Tissues from individual animals or humans may
responddifferentlytothesameconcentrationofapotentiallytoxiccompound, body weight.
EA CFD-PBPK CHRONIC RISK ASSESSMENT 397
Selection of Internal Dose Metric for Olfactory Toxicity would have to exceed 50% and be sustained in order to
contributetotoxicity.Giventhelimitedexperiencewithroleof
Basedonthemodeofactionfornasaldamage,measuresof
GSH in nasal toxicity, a clear NOEL for GSH depletion (the
internal dose related to AA concentration in the olfactory
experimentally observed value) was used.
epitheliumwereconsidered.Thepreferreddosemetricwasthe
Thus, the dose metrics used for interspecies extrapolation
time-averaged steady state concentration of AA in the nasal
were the predicted concentration of AA in the olfactory epi-
olfactory epithelium, rather than the peak concentration. This
thelium (adjusted for exposure duration) and the predicted
selection was based on the studies of Lomax et al. (1994)
minimumconcentrationofGSHintheolfactoryepithelium(as
indicating that exposures of mice to AA with equivalent C (cid:2)
percent of control).
T (25 ppm for 4.4 h/day or 5ppm for 22 h/day) produced
similar incidence of nasal irritation. Chronic RfC Calculation
D
TheinitialhypothesiswasthattheAAconcentrationwould o
w
be the preferred dose metric for setting a risk reference value InTable2,approachestoidentifyingachronicRfCforethyl n
lo
forethylacrylate.Thepredictednasaltissueconcentrationsof acrylatearesummarized.DefaultapproachestoRfCsbasedon ad
AAattheNOELwerecomparedforinhalationexposuretoAA olfactory tissue damage and lethargy/irritancy/body weight ed
andforexposuretoethylacrylate(6h/day,5days/week)using decreases (Miller et al., 1985) are based on default U.S. EPA fro
m
the model of Frederick et al. (2002). The olfactory tissue AA guidance (U.S. EPA, 1994). A chronic RfC based on CFD- h
concentrations would be similar if the AA concentration were PBPK modeling of AA concentrations in olfactory epithelium ttps
the sole determinant of nasal damage. In fact, the calculated without consideration of GSH depletion is provided for com- ://a
c
concentrations differed by a factor of 60 (Table 1). parisontotheRfCderivedwhenolfactoryGSHconcentrations ad
e
The comparison of predicted AA concentrations associated were also considered. While uncertainty factors were applied m
with inhalation exposure at the NOELs for AA and ethyl topredictedtissueconcentrationsofAA,nouncertaintyfactors ic.o
u
acrylate suggested that AA concentration alone was not suffi- wereappliedtotissueGSHconcentrations.GSHdepletionper p.c
cient to account for the nasal toxicity of ethyl acrylate. Addi- se does not cause the toxicity, but appears to be a modifying om
tional dose measures were considered and tested. These mea- pharmacodynamic factor that enhances the toxicity of AA in /to
x
sures included calculated total esterase metabolism and AA the tissues. sc
i/a
concentrationsinthebloodexchangelayerofthenasalepithe- Sensitivity of Olfactory AA Concentrations to Choice of rtic
lialtissues.Uponreviewingthemodelresults(simulationsnot le
Work Versus Rest and Human Olfactory Metabolism Rate -a
shown)andtheliteratureontheoraltoxicityofethylacrylate, b
s
it appeared that the most reasonable mode of action for ethyl The choice of light work versus rest provided a slightly tra
c
acrylate toxicity was increased susceptibility to mitochondrial lower value for the RfC. The choice of measured human t/7
9
toxicity of AA due to depletion of GSH by conjugation with metabolismratesratherthanmeasuredmonkeyoradjustedrat /2
/3
ethyl acrylate. rates increased the equivalent human ethyl acrylate exposure 9
4
Atargetvaluefora“safe”olfactoryGSHconcentrationwas concentrations (Table 3). These values, however, are still /1
6
4
selected,basedonthedepletionobservedinasingleexposure higher than the proposed RfC that incorporated prediction of 9
3
study(Fredericketal.,2002)usingthesameexposureconcen- nasal olfactory GSH depletion, a contributing pharmacody- 63
tration and duration (5 ppm for 6 h) as the NOEL toxicology namic factor. by
g
study (Miller et al., 1985). A larger body of GSH research u
e
(summarized in Frederick et al., 1992), which did not specif- Sensitivity of Proposed Chronic RfC to Target GSH st o
ically include nasal toxicity, indicated that GSH depletion Concentration n 3
0
The chronic RfC for olfactory toxicity was derived by de- M
a
termining what exposure concentration would be expected to rc
TABLE 1 result in GSH depletion to 75% of control. When 50% GSH h 2
0
AcrylicAcid(AA)ConcentrationsAssociatedwithNoObserved depletion rather than 25% depletion was used as the target 19
Effect Levels and Lowest Observed Effect Levels for Olfactory internal dose metric, the proposed chronic RfC for olfactory
TissueDamageinRatsorRatTissue toxicity increased from 0.25 ppm to 0.8 ppm. The choice of
25%depletionasatargetinsteadof50%depletionproduceda
Endpointforrats/rattissues AssociatedAAconcentration
three-fold (0.8 ppm/0.25 ppm) increase in the RfC for this
endpoint.
AANOEL,LOELinvivo 3mM,6mM(Fredericketal.,1998;
Andersenetal.,2000)
Sensitivity Analysis of Nasal Olfactory Epithelium AA and
EthylacrylateNOEL,LOEL 0.05mM,0.25mM
invivo GSH Concentrations
AANOEL,LOELinvitro* 0.4mM,0.6mM(Fredericketal.,1998)
The nasal toxicity observed in ethyl acrylate-exposed rats
*Ratnasalexplants apparentlyresultsfromenhancedsusceptibilitytoAAtoxicity
398 SWEENEY, ANDERSEN, AND GARGAS
TABLE 2
Default and Computational Fluid Dynamics-Physiologically Based Pharmacokinetic Modeling Approaches to Selection of a
Chronic Reference Concentration for Inhaled Ethyl Acrylate
Endpoint(extrapolationapproach)
Lethargy,irritancy,
decreasedbodyweight Olfactorytissuedamage
CFD-PBPK,GSH, CFD-PBPK,GSH
Default Default CFD-PBPK,AA andAA andAA
D
NOEL(ppm) 75 5 5 5 5 o
w
(Milleretal., n
lo
1985) a
d
LOEL(ppm)(Miller 225 25 25 25 25 e
d
DoesteaMl.,e1as9u8r5e) externalconcentration externalconcentration 0.05mMAAin minimumolfactory minimumolfactorytissuea from
olfactoryepitheliuma tissueaGSH(cid:1) GSH(cid:1)50%ofcontrol h
50%ofcontrol ttps
Durationadjustment (6/24)(cid:2)(5/7) (6/24)(cid:2)(5/7) (6/24)(cid:2)(5/7) NA NA ://a
(hperday,days ca
d
perweek) e
m
Regionalgas 1(Category2,systemic 0.18(Category1,portal NA NA NA ic
dosimetryratio effectathigher ofentryeffectonlyat .o
u
(RGDR) concentrations) lowconcentration) p
.c
UF 3 3 3 NA NA o
A,PD m
UFH 10 10 10 NA NA /to
Targetinternal NA NA 0.05mMAA(cid:2)(6/24) minimumolfactory minimumolfactorytissuea x
s
concentrationfor (cid:2)(5/7)/(3(cid:2)10)(cid:1) tissueaGSH(cid:1) GSH(cid:1)75%ofcontrol ci/a
hPuBmPaKnMCFodDe-la 0o.l0fa0c0t3ormyMepiAthAeliiunma 50%ofcontrol rticle
RfCcalculationb 75ppm(cid:2)(6/24)(cid:2) 5ppm(cid:2)(6/24)(cid:2) humanCFD-PBPK humanCFD-PBPK humanCFD-PBPK -a
b
ChronicRfC 0.4(5p/p7m)(cid:2)1/(3(cid:2)10) 0.0(055/7p)p(cid:2)m0.18/(3(cid:2)10) 1.8mpopdmelc 0.8mpopdmelc 0.2m5opdpemlc (Recommended strac
RfC) t/7
9
/2
aDorsalolfactoryepithelium,firsttissuelayerbelowmucus. /39
bDefaultRfC(cid:1)NOEL(cid:2)TimeAdjustment(cid:2)RGDR/(UFA,PD(cid:2)UFH). 4/1
cCFD-PBPKmodelofFredericketal.(2002),modifiedasdescribedunder“Methods,”assuminglightwork. 64
9
3
6
3
b
y
g
u
e
duetoGSHdepletion.SensitivityanalysisofpredictedAAand st o
GSH concentrations in the olfactory epithelium were con- n
EthylAcrylateExposureCToAnBcLenEtra3tionsCorrespondingtoTar- ducted for the rat at the NOEL (for 6 h of exposure) and the 30 M
human at the proposed RfC (for 1 week of exposure) to assist a
get Internal Concentrations of Acrylic Acid in the Nasal Epithe- rc
lium in the evaluation of the selected Uncertainty (Adjustment) h 20
Factors from Table 2. 1
9
The chemical-specific parameters that had the greatest im-
AtRest Lightworka
pact on olfactory AA predictions in the rat were the exposure
Measuredhumannasalmetabolismratesb 2.2ppm 1.8ppm concentration and mucus:air partition coefficient for ethyl ac-
Measuredmonkeynasalmetabolismratesb 0.9ppm 0.8ppm
rylate (Table 4). The most sensitive physiological parameters
Adjustedratnasalmetabolismratesc 0.6ppm 0.6ppm
werethewidthoftheolfactoryepitheliumandmucusthickness
aApproximately50W. over the olfactory epithelium. Model predictions of olfactory
bEthylacrylatemetabolisminnasalrespiratoryandolfactoryepitheliumin tissueGSHconcentrationafter6hofexposureweregenerally
vitro(Fredericketal.,2002).
lesssensitivetomodelparametersvaluesandweresensitiveto
cEthyl acrylate metabolism rates measured in rat tissue (Frederick et al.,
exposure concentration, kinetic parameters for ethyl acrylate
2002),adjustedbyhuman:ratmetabolismratioformethylmethacrylate(Main-
waringetal.,2001). hydrolysis and GSH conjugation in the olfactory epithelium,
EA CFD-PBPK CHRONIC RISK ASSESSMENT 399
TABLE 4 based, the GSH concentrations, were not sensitive to small
SensitivityofPredictionsofAcrylicAcid(AA)Concentrationin changes in most of the model parameter values.
the First Layer of the Rat Dorsal Nasal Olfactory Epithelium to
ModelParameterValues Consideration of Potentially Sensitive Subpopulations
Normalizedsensitivity TherecommendedchronicRfCwasdeterminedonthebasis
Parametera coefficient of predicted GSH depletion in the olfactory epithelium. As
shown in Table 6, olfactory GSH concentrations are most
Exposureconcentration 1.00
sensitive to olfactory ethyl acrylate-GSH conjugation rate,
Mucus:airpartitioncoefficientforethylacrylate 0.89
Diffusivityinolfactoryepithelium (cid:8)0.42 GSH turnover rate, and the mucus:air partition coefficient for
14ofdistanceacrosstheolfactoryepithelium 0.41 ethyl acrylate. Subpopulations for which these values differ D
Thicknessofmucusoverolfactoryepithelium (cid:8)0.38 from the population means would not be expected for a ther- ow
Mucusdiffusivity 0.35 modynamic parameter, such as the mucus:air partition coeffi- nlo
Mucus:epitheliumpartitioncoefficientforethyl a
cient,andhavenotbeenidentifiedforanyoftheotherparam- d
Naascarlytliastseue:bloodpartitioncoefficientforethyl (cid:8)0.32 eters. Banger et al. (1996) have demonstrated a lack age- or ed fro
acrylate 0.27 sex-related differences in cytosolic olfactory GSTs in rats. No m
Ethylacrylatehydrolysiscapacityinnasal additionaluncertaintyfactorsareanticipatedtobenecessaryto http
olfactoryepithelium 0.27 protectsensitivesubpopulations.Basedonthevariabilityanal- s
Dilutionofenzymaticallyrichlayersofnasal ysis at the proposed RfC, the predicted concentrations of the ://ac
olfactoryepitheliuminvitro 0.27 a
putativetoxicmetaboliteAAinhumanolfactoryepitheliumfor d
Enzymeaffinityforethylacrylatehydrolysisinthe e
m
nasalolfactoryepithelium (cid:8)0.27 the 95th-percentile individual would be only 2.1-fold higher ic
Widthofbloodexchangeregionunderolfactory than in the average individual (1.6 standard deviations higher, .ou
p
epithelium 0.26 with a model CV of 0.71, Table 6), less than the default .c
Nasaltissue:bloodpartitioncoefficientforAA 0.24 intraspeciesUFforhumanpharmacokineticvariability((cid:7)3.2). om
NFroarcmtiaolnizoefdncaasradliabcloooudtpfluotwtoolfactory (cid:8)0.23 For human olfactory GSH, the variability analysis (Table 7) /toxs
epitheliuminanteriorregion (cid:8)0.21 predicted that the 95th-percentile individual would possess ci/a
Fractionofnasalbloodflowtoolfactory olfactoryGSHconcentrationsof58%ofcontrol.Theresultsof rtic
epithelium (cid:8)0.21 the variability analysis thus indicate that uncertainty factors le
Bloooudtpfluto)wtonasalcavity(fractionofcardiac (cid:8)0.21 used for pharmacokinetics (AA concentrations) and pharma- -abstra
Bodyweight (cid:8)0.20 c
t/7
9
aSimulations conducted with the refined CFD-PBPK model for cyclic TABLE 5 /2
/3
breathing of 5 ppm ethyl acrylate by rats for 1 h (steady-state AA). Only Sensitivity of Predictions of Glutathione (GSH) Concentration 94
padardaimtioentearls10wiptahra(cid:3)mnoertmeraslihzaedd(cid:3)nseonrmsitailviziteydcsoeenfsfiitcivieintyt(cid:3)c(cid:5)oef0fi.c2i0enat(cid:3)rebesthwoewenn.0.A05n intheFirstLayeroftheRatDorsalNasalOlfactoryEpitheliumto /164
and 0.20, with the remaining 100 parameters having (cid:3)normalized sensitivity ModelParameterValues 93
6
coefficient(cid:3)(cid:9)0.05. 3 b
Normalizedsensitivity y
Parametera coefficient gu
e
s
tthioenwciodetfhfiocfietnhtefoolrfaecthtoyrlyaecpryitlhaeteliu(Tma,balend5)th.e mucus:air parti- E14xopfodsiustraenccoenaccernotsrsattihoenolfactoryepithelium (cid:8)00..3209 t on 3
0
Tables 6 and 7 summarize the sensitivity of model predic- AugmentationofGSHconjugationinolfactory M
tionsofhumanolfactoryAAandGSHconcentrations,respec- epitheliumbyGSHtransferases (cid:8)0.28 arc
tively, to values of model parameters. The chemical-specific NonenzymaticsecondorderrateconstantforGSH h 2
conjugationwithethylacrylate (cid:8)0.28 0
1
parameters that had the greatest impact on tissue AA predic- Mucus:airpartitioncoefficientforethylacrylate (cid:8)0.26 9
tions were the exposure concentration and the nasal tissue: Ethylacrylatehydrolysiscapacityinnasal
blood partition coefficient for AA. The most sensitive physio- olfactoryepithelium 0.21
Dilutionofenzymaticallyrichlayersofnasal
logical parameters were the respiration rate and tidal volume.
olfactoryepitheliuminvitro 0.21
Predicted olfactory tissue GSH concentration was generally
Enzymeaffinityforethylacrylatehydrolysisin
less sensitive to model parameter values than the AA predic- thenasalolfactoryepithelium (cid:8)0.20
tions, but was sensitive to the exposure concentration (pro-
posed RfC), mucus:air partition coefficient for ethyl acrylate, aSimulations conducted with the refined CFD-PBPK model for cyclic
breathing of 5 ppm ethyl acrylate by rats for 6 h. Only parameters with
and the parameters describing GSH conjugation and turnover
(cid:3)normalizedsensitivitycoefficient(cid:3)(cid:5)0.20areshown.Anadditional4param-
in the olfactory epithelium. The sensitivity analyses indicate etershad(cid:3)normalizedsensitivitycoefficient(cid:3)between0.05and0.20,withthe
that the model predictions upon which the proposed RfC was remaining113parametershaving(cid:3)normalizedsensitivitycoefficient(cid:3)(cid:9)0.05.
400 SWEENEY, ANDERSEN, AND GARGAS
TABLE 6 anistic data support a predominant role for AA in mitochon-
Sensitivity of Predictions of AA Concentration in the First drial toxicity leading to inflammation. However, comparisons
LayeroftheHumanDorsalNasalOlfactoryEpitheliumtoModel to results for inhaled AA were inconsistent—ethyl acrylate is
ParameterValuesandPredictedVariability moretoxicthanAA,basedoncalculationsoftissueAA.Based
on research showing that ethyl acrylate depletes nasal GSH
Normalized Parameter (Frederick et al., 2002) and that GSH depletion can enhance
sensitivity coefficient
mitochondrial toxicity of other compounds (Haouzi et al.,
Parametera coefficient ofvariation
2001; Hoek et al., 2002), we developed a theory that ethyl
Exposureconcentration 1.00 notapplicable acrylate toxicity results from an enhanced sensitivity to AA
Respirationrate 0.57 0.1b due to GSH depletion, and that an appropriate RfC should be
Tidalvolume 0.57 0.33b basedonmaintainingsufficientlevelsofGSHintheolfactory D
Bodyweight (cid:8)0.55 0.15c epitheliumandpreventingelevatedtissuelevelsofAA.There- ow
Nasaltissue:bloodpartitioncoefficient n
fore, we developed one olfactory RfC based solely on pre- lo
forAA 0.54 0.19d a
Ethylacrylatehydrolysiscapacityin dicted AA concentrations (RfC (cid:1) 1.8 ppm) and a second ded
nasalolfactoryepithelium 0.52 0.95e olfactoryRfCbasedonmaintainingaprotectivelevelofnasal fro
Dilutionofenzymaticallyrichlayersof GSH(RfC(cid:1)0.25ppm)andrecommenduseofthelattervalue. m
nasalolfactoryepitheliuminvitro 0.52 0.3f If only predicted AA concentration had been considered for http
Enhzyydmreolayfsfiisniitnytfhoerneathsaylloalcfraycltaotrey olfactory toxicity (olfactory RfC (cid:1) 1.8 ppm), the overall s://a
epithelium (cid:8)0.51 0.61e assessment with uncertainty factors totaling 30 would have ca
d
Widthofolfactoryepithelium 0.46 0.2g resultedinarecommendedRfCof0.4ppmbasedonlethargy, e
m
Mucus:airpartitioncoefficientforethyl irritancy, and decreased body weight. Consideration of nasal ic
acrylate 0.45 0.24e GSH resulted in a slightly lower RfC of 0.25 ppm. .ou
D14oiefffpudistihisvteailtniyucmeinaocrlfoascstothryeeoplfiathcetoliruym (cid:8)00..2475 00..256gh latTehiesraecfoamctmorenodfed(cid:7)c7hr(o1n.8icpRpfmC/o0f.205.2p5ppmp)mlofowreerththyalnactrhye- p.com/to
Nasaltissue:bloodpartitioncoefficient potential RfC derived based on predicted olfactory tissue AA xs
c
forethylacrylate 0.23 0.17e alone (1.8 ppm). The factor of 7 can be considered an addi- i/a
ModelCoefficientofVariation 0.71 tionalpharmacodynamicadjustmentfactortothepredictedAA rtic
le
aSimulations conducted with the refined CFD-PBPK model for cyclic tissue concentration for the contribution of GSH status to the -ab
breathingof0.25ppmethylacrylatebyhumansfor1week(steady-state).Only development of toxicity, in addition to the pharmacodynamic stra
parameters with (cid:3)normalized sensitivity coefficient(cid:3) (cid:5) 0.20 are shown. An c
additional31parametershad(cid:3)normalizedsensitivitycoefficient(cid:3)between0.05 t/79
and 0.20, with the remaining 85 parameters having (cid:3)normalized sensitivity TABLE 7 /2
coefficient(cid:3)(cid:9)0.05. Sensitivity of Predictions of GSH Concentration (Percent of /394
bBrownetal.(1997). Control)intheFirstLayeroftheHumanDorsalNasalOlfactory /16
cAllenetal.(1996). 4
dBlackandFinch(1995). EpitheliumtoModelParameterValuesandPredictedVariability 93
6
eFredericketal.(2002). 3 b
fAllenetal.(1996). Normalized Parameter y
g
gBogdanffyetal.(1998). sensitivity coefficient u
e
Parametera coefficient ofvariation s
t o
n
codynamics (GSH concentrations) (Table 2) should provide Exposureconcentration (cid:8)0.26 notapplicable 30
AugmentationofGSHconjugationin M
ample protection of human health. olfactoryepitheliumbyGSH arc
transferases (cid:8)0.26 0.3b h 2
DISCUSSION Nonenzymaticsecondorderrateconstant 0
1
forGSHconjugationwithethylacrylate (cid:8)0.26 0.3c 9
To determine a chronic risk reference value for ethyl acry- Mucus:airpartitioncoefficientforethyl
late, multiple endpoints and multiple approaches were consid- acrylate (cid:8)0.24 0.24c
First-orderGSHlossinthenasalolfactory
ered.BecausebothNOELsanduncertainty/adjustmentfactors
epithelium 0.24 0.3c
maydifferamongendpoints,itwasnecessarytoconsiderboth
ModelCoefficientofVariation 0.14
olfactorytoxicityandlethargy/irritancy/decreasedbodyweight
aspossiblecriticaleffects.Duetothelackofmechanisticdata aSimulations conducted with the refined CFD-PBPK model for cyclic
to support the use of CFD-PBPK modeling for interspecies breathingof0.25ppmethylacrylatebyhumansfor1week(steady-state).Only
parameters with (cid:3)normalized sensitivity coefficient(cid:3) (cid:5) 0.20 are shown. The
extrapolation of the lethargy/irritancy/decreased body weight
remaining124parametershad(cid:3)normalizedsensitivitycoefficient(cid:3)(cid:9)0.05.
findings, a default approach was used, resulting in an RfC of bAllenetal.(1996).
0.4 ppm for these endpoints. For the olfactory toxicity, mech- cBangeretal.(1993).
EA CFD-PBPK CHRONIC RISK ASSESSMENT 401
uncertainty factors for AA included in the derivation of the tissue concentration of AA could produce a combined dose
“AA only” RfC (1.8 ppm). The derivation of an RfC of 1.8 metric that is mathematically equivalent to the current pro-
ppm based on AA predictions alone (Table 2) included a posed RfC, but would logically seem to pose a greater risk to
pharmacodynamicuncertaintyfactorof3foranimaltohuman human health because adequate GSH protection would be
differences UF , and an uncertainty factor for differences retained, but concentrations of the putative toxicant would be
A,PD
among humans (UF (cid:1) 10) that is generally considered to higher.
H
consist of equal components ((cid:7)3.2) for pharmacokinetic and On initial review, it appears that the olfactory toxicity RfC
pharmacodynamic variability. The additional pharmacody- based on GSH depletion did not incorporate any uncertainty
namicadjustmentfactorforGSH(7)accountedforthegreater factors (Table 2). The selection of 25% depletion of olfactory
susceptibility, compared to rats, of the human olfactory tissue epithelium GSH (equal to the concentration observed at the
to GSH depletion predicted by the CFD-PBPK model, and NOEL in the rat) was somewhat conservative in that other D
o
resulted in a total pharmacodynamic uncertainty/adjustment researchers have identified 50% GSH depletion as a threshold w
n
factor(cid:1)UF (cid:2)UF (cid:2)AF (cid:1)3(cid:2)3.2(cid:2)7 for toxicity (Frederick et al., 1992). Also, this olfactory RfC lo
A,PD(AA) H,PD(AA) A,PD(GSH) a
(cid:1) 67. was not based on GSH alone, but on GSH depletion in com- de
d
ThesensitivityanalysisresultsforGSHpredictionsintherat bination with the presence of AA, which, as noted above, is fro
nose(Table5)aidintheunderstandingofwhytheintroduction predicted to be sven-fold lower than a target concentration m
h
ofKGSHO(enhancementofGSHconjugationintheolfactory developed by applying time and uncertainty factor adjust- ttp
s
tissue by GSTs) improved the predictive capability of the ments. As AA was the putative toxicant with GSH as a mod- ://a
model.TheratolfactoryGSHconcentrationsweresensitiveto ifying factor, customary UFs were most logically applied to ca
d
relatively few model parameters. A reduction in the ethyl setting the target values of AA rather than GSH. Some con- e
m
acrylate hydrolysis rate was considered, but even if it was set servatism in the GSH dose metric was warranted, due to the ic
.o
equal to zero, the Frederick et al. (2002) model would predict lack of validation data for the human GSH model. The selec- u
p
that GSH would only be depleted to 88% of control (vs. 75% tion of 25% depletion rather than 50% depletion of GSH .co
m
observed in the experiment and 97% predicted in Frederick et provided this conservatism and reduced the olfactory toxicity /to
al.,2002).Clearly,someoralloftheadjustmentmustbetothe RfC by a factor of approximately 3 (from 0.8 ppm to 0.25 xs
c
GSH modeling. ppm). i/a
The model sensitivity analyses indicated that the human The proposed mode of action for olfactory toxicity of ethyl rtic
le
nasalolfactorytissueGSHpredictions,forwhichtherewereno acrylate—AAtoxicityenhancedbyGSHdepletion—isconsis- -a
b
validationdata,werenothighlysensitivetothemodelparam- tent with the mode of action for toxicity of ethyl acrylate for stra
eter values. A conservative target concentration of GSH was other endpoints and for another olfactory toxicant, methyl c
t/7
selected,anduncertaintywithrespecttopreciselywhatlevelof iodide.Modesofactionforethylacrylatetumorigenesisinthe 9
/2
GSH was necessary to protect human nasal tissue was of less rodentforestomachandacutelethalityhavebothbeenlinkedto /3
9
concern due to two factors: the observation that GSH predic- GSH depletion (Frederick et al., 1992, 2002). Chamberlain et 4
/1
tions were not highly variable for the population and the al. (1998) have investigated the mode of action for olfactory 64
9
prediction that concentrations of the putative toxic metabolite toxicity of methyl iodide in rats by modulating the status of 3
6
3
wouldbeseven-foldlowerthanthetargetconcentrationsbased GSH and cytochrome P450. They concluded that GSH deple- b
y
on target tissue AA concentration alone. tion (to 40% of control), rather than GSH-metabolite forma- g
u
Model sensitivity analyses of the baseline Frederick et al. tion, initiated nasal toxicity, and that the relatively slow turn- e
s
(2002) model, without the GSH refinement, were also con- over of GSH in the olfactory epithelium (compared to t o
n
ducted(datanotshown).Thesensitivitycoefficientsforolfac- respiratory epithelium and other tissues) prolonged GSH de- 3
0
tory AA predictions were identical to the refined model, and pletion and rendered olfactory epithelium more susceptible to M
a
the results for GSH were very similar, with sensitivity coeffi- oxidative damage than the respiratory epithelium. Ethyl acry- rc
h
cients differing no more that 30% from the values determined late was hydrolyzed to AA at similar rates in nasal olfactory 20
1
for the refined model. and respiratory epithelium, and the tissues have similar GSH 9
TheproposedRfCwasderivedbyconsideringAAandGSH concentrations following single exposures, but toxicity was
separately,thencomparingtheresults.Thatis,apotentialRfC limitedtotheolfactoryregion(Fredericketal.,2002;Milleret
wasderivedbasedonAAonly,andanotherpotentialRfCwas al., 1985), indicating that the slow recovery of GSH concen-
derived on the basis of GSH only, and this latter value was trations after exposure may contribute to the development of
selected as the proposed RfC because it is also believed to be the nasal olfactory lesions.
protective with respect to AA, whereas the “AA-only” RfC Comparisons of ethyl acrylate nasal toxicities between spe-
may not be sufficiently protective of olfactory GSH for ethyl cies,includingconsiderationofacrylicacid,canbeinformative
acrylate-exposedindividuals.Thepossibilityofcombiningthe astopossiblemechanismsofaction.Ourinitialhypothesiswas
two measures was considered but deemed to be unjustifiable. that the nasal lesions produced following ethyl acrylate expo-
For example, a lesser extent of GSH depletion and a higher sure were due to the formation of AA. Comparisons of the
402 SWEENEY, ANDERSEN, AND GARGAS
concentrations of AA predicted by modeling (see Table 1) ModelequationshavepreviouslybeenprovidedbyFredericketal.(1998).
indicated that other factors in addition to AA concentration Model code (in ACSL) will be provided by the corresponding author upon
request.
wereinvolvedinthetoxicity(e.g.,GSHdepletion).Inspection
TheonlysignificantchangefromFredericketal.(2002)wastheaugmen-
oftheNOELsandLOELsforEAandAAinratsandmicealso
tationofGSHconjugationinolfactorytissuesbyafactorof“KGSHO”.The
indicated factors other than AA production might be involved modifiedequationisprovidedbelow
in the formation of nasal lesions following ethyl acrylate ex-
posures(Milleretal.,1985).TheNOELsandLOELsformice RGDOE (cid:2)VDOE (cid:3)(KGSH(cid:3)Kgsho(cid:3)GSDOE (cid:3)CDOE)
ij ij ij ij
and rats exposed chronically to ethyl acrylate are both 5 ppm
and 25 ppm, respectively. The NOEL and LOEL for chronic where
AA exposures in rats are 25 ppm and 75 ppm, values that are RGDOE (cid:1)RateofGSHconjugationindorsalolfactoryepitheliumtissue
ij
higher than those for ethyl acrylate indicating that the ester compartmentj,layeri((cid:1)mol/h). D
may be more potent than the acid alone (Miller et al., 1981). VDOEij (cid:1) Volume of dorsal olfactory epithelium tissue compartment j, ow
The LOELs for mice exposed to AA were 5 ppm for female layeri(ml). nlo
KGSH (cid:1) Second-order rate of nonenzymatic GSH conjugation (ml/h- ad
miceand25ppmformalemice,providingnoclearindication (cid:1)mol). ed
that the ester might be more potent. Kgsho(cid:1)increaseinGSHconjugationraterelativetononenzymaticrate, fro
In summary, a proposed RfC for ethyl acrylate was devel- olfactorytissue(dimensionless). m
opedbasednotonlyontheexpectedconcentrationsofthetoxic parGtmSDenOtEj,ijla(cid:1)yeGrSiH((cid:1)cmoonlc/emnlt)rationindorsalolfactoryepitheliumtissuecom- https
metabolite AA, but also on GSH, which serves both as a CDOE (cid:1)Ethylacrylateconcentrationindorsalolfactoryepitheliumtissue ://a
cofactor for ethyl acrylate metabolism and as an indicator of compartmijentj,layeri((cid:1)mol/ml). ca
d
tissueoxidativestress.TheproposedRfCof0.25ppmislower e
m
than the RfC that would have been calculated based on AA ic
ACKNOWLEDGMENTS .o
prediction alone (1.8 ppm), but 50-fold higher than the RfC u
p
yielded by the default approach, 0.005 ppm. ThisworkwasfundedbytheBasicAcrylicMonomersManufacturers,Inc. .co
m
/to
REFERENCES xs
c
APPENDIX i/a
Allen,B.C.,Covington,T.R.,andClewell,H.J.(1996).Investigationofthe rtic
Theethylacrylatemodel(Fig.1)wasanextensionoftheCFD-PBPKmodel impact of pharmacokinetic variability and uncertainty on risks predicted le-a
forAA(Andersenetal.,2000;Bushetal.,1998;Fredericketal.,1998).The withapharmacokineticmodelforchloroform.Toxicology111,289–303. b
s
Frederick et al. (2002) model incorporated cyclic breathing (i.e., inhalation Andersen, M., Sarangapani, R., Gentry, R., Clewell, H., Covington, T., and tra
c
followedbyexhalation),airflowthroughthenasalpassages,masstransferof Frederick,C.B.(2000).ApplicationofahybridCFD-PBPKnasaldosimetry t/7
ethyl acrylate from the vapor phase to the mucus layer, diffusion from the model in an inhalation risk assessment: An example with acrylic acid. 9
/2
mucuslayertothetissuelayers,anddiffusionthroughtissuelayers(fourfor Toxicol.Sci.57,312–325. /3
olfactoryepithelium,twoforrespiratoryepithelium,oneforremainingnasal Banger,K.K.,Lock,E.A.,andReed,C.J.(1993).Thecharacterizationof 94
surfaces)tothebloodexchangeregion.Themodeldescribedthetransportof glutathione S-transferases from rat olfactory epithelium. Biochem. J. 290 /16
ethylacrylatethroughthebloodstreamtothelungs,liver,kidneys,otherrichly (Pt.1),199–204. 49
3
perfused tissues, and poorly perfused tissues (muscle and fat), which were 6
Banger,K.K.,Lock,E.A.,andReed,C.J.(1996).Regulationofratolfactory 3
modeledassingle,well-mixedcompartments.Thesamemodelstructurewas glutathioneS-transferaseexpression.Investigationofsexdifferences,induc- by
used to describe the distribution of metabolically generated AA, with the tion,andontogenesis.Biochem.Pharmacol.52,801–808. gu
exception that exhalation of the acid was expected to be negligible (rat e
bnleogoledc:ateirdpbayrtiFtiroendecroicekffiectieanlt.(cid:1)(2040320)0.(cid:6)Me1t3a0b;olFisrmedeorficektheytlaal.c,r1y9la9t8e)oacncdurwreads Blapcakrt,itKio.nAc.,oaenffidcFieinntcsh,inL.ra(t19ti9ss5u).esA.cTroyxliiccoalc.idLeotxt.id7a8t,io7n3–a7n8d.tissue-to-blood st on
through two pathways—second-order conjugation with GSH or saturable Bogdanffy, M. S., Randall, H. W., and Morgan, K. T. (1987). Biochemical 30
(Michaelis-Menten)esterase-mediatedhydrolysistoAA.AAmetabolismoc- quantitationandhistochemicallocalizationofcarboxylesteraseinthenasal M
a
curredthroughsaturableoxidation. passagesoftheFischer-344ratandB6C3F1mouse.Toxicol.Appl.Phar- rc
h
macol.88,183–194. 2
0
Bogdanffy, M. S., Sarangapani, R., Kimbell, J. S., Frame, S. R., and 1
9
ScalingofMetabolicParameters Plowchalk,D.R.(1998).Analysisofvinylacetatemetabolisminratand
humannasaltissuesbyaninvitrogasuptaketechnique.Toxicol.Sci.46,
Aspreviouslydescribed(Fredericketal.,1998,2002),ratesofethylacrylate
235–246
andAAmetabolisminlung,liver,kidney,otherrichlyperfusedtissues,and
Brown,R.P.,Delp,M.D.,Lindstedt,S.L.,Rhomberg,L.R.,andBeliles,R.P.
muscle were scaled from in vitro (per ml of tissue, Black and Finch, 1995;
(1997). Physiological parameter values for physiologically based pharma-
Fredericketal.,2002)toinvivo,basedontissuevolume.Metabolicratesfor
cokineticmodels.Toxicol.Ind.Health13,407–484.
nasalolfactoryandrespiratoryepitheliumwerescaled,basedontissuevolume,
andadjustedforinvitrodilutionofmetabolicallyactivetissueswithconnec- Bush, M. L., Frederick, C. B., Kimbell, J. S., and Ultman, J. S. (1998). A
tive tissues, based on microscopic examination of nasal epithelial samples CFD-PBPK hybrid model for simulating gas and vapor uptake in the rat
(Frederick et al., 2002). In vitro rates of AA metabolism determined in rat nose.Toxicol.Appl.Pharmacol.150,33–45.
tissues(BlackandFinch,1995)wereusedforbothratsandhumans.Invitro Chadha,T.S.,Birch,S.,andSackner,M.A.(1987).Oronasaldistributionof
ratesofethylacrylatehydrolysisweredeterminedinrat,monkey,andhuman ventilationduringexerciseinnormalsubjectsandpatientswithasthmaand
tissues(Fredericketal.,2002). rhinitis.Chest92,1037–1041
EA CFD-PBPK CHRONIC RISK ASSESSMENT 403
Chamberlain,M.P.,Lock,E.A.,Gaskell,B.A.,andReed,C.H.(1998).The methacrylatetoxicityinratnasalepithelium:Studiesofthemechanismof
roleofglutathioneS-transferaseandcytochromeP450-dependentmetabo- actionandcomparisonsbetweenspecies.Toxicology158,109–118.
lismintheolfactorytoxicityofmethyliodideintherat.Arch.Toxicol.72, Miller,R.R.,Ayres,J.A.,Jersey,G.C.,McKenna,M.J.(1981).Inhalation
420–428. toxicityofacrylicacid.Fundam.Appl.Toxicol.1,271–727.
Custodio,J.B.,Palmeira,C.M.,Moreno,A.J.,andWallace,K.B.(1998). Miller, R. R., Young, J. T., Kociba, R. J., Keyes, D. G., Bodner, K. M.,
Acrylicacidinducestheglutathione-independentmitochondrialpermeabil- Calhoun,L.L.,andAyres,J.A.(1985).Chronictoxicityandoncogenicity
itytransitioninvitro.Toxicol.Sci.43,19–27. bioassay of inhaled ethyl acrylate in Fischer 344 rats and B6C3F1 mice.
Frederick,C.B.,Bush,M.L.,Lomax,L.G.,Black,K.A.,Finch,L.,Kimbell, DrugChem.Toxicol.8,1–42.
J.S.,Morgan,K.T.,Subramaniam,R.P.,Morris,J.B.,andUltman,J.S. NationalToxicologyProgram(1986).Toxicologyandcarcinogenesisstudies
(1998).Applicationofahybridcomputationalfluiddynamicsandphysio-
ofethylacrylate.ReportNo.259.NIHPublicationNumber87–2515.
logicallybasedinhalationmodelforinterspeciesdosimetryextrapolationof
NationalToxicologyProgram(2002).ReportonCarcinogens,TenthEdition;
acidic vapors in the upper airways. Toxicol. Appl. Pharmacol. 152, 211–
U.S. Department of Health and Human Services, Public Health Service, D
231. o
December2002. w
Frederick, C. B., Lomax, L. G., Black, K. A., Finch, L., Scribner, H. E., n
Potter,D.W.,andTran,T.B.(1993).Apparentratesofglutathioneturnover lo
Kimbell,J.S.,Morgan,K.T.,SubramaniamR.P.,andMorris,J.B.(2002). a
Use of a hybrid computational fluid dynamics and physiologically based inrattissues.Toxicol.Appl.Pharmacol.120,186–192. de
d
inhalation model for interspecies dosimetry comparisons of ester vapors. Saillenfait, E. M., Bonnet, P., Gallissot, F., Protois, J. C., Peltier, A., and fro
Toxicol.Appl.Pharmacol.183,23–40. Fabries,J.F.(1999).Relativedevelopmentaltoxicitiesofacrylatesinrats m
Frederick, C. B., Potter, D. W., Chang-Mateu, M. I., and Andersen, M. E. followinginhalationexposure.Toxicol.Sci.48,240–254. http
(1992). A physiologically based pharmacokinetic and pharmacodynamic Sweeney, L. M., Gargas, M. L., Strothor, D. E., Kedderis, G. L. (2003). s
modeltodescribetheoraldosingofratswithethylacrylateanditsimpli- Physiologically based pharmacokinetic model parameter estimation and ://a
c
cationsforriskassessment.Toxicol.Appl.Pharmacol.114,246–260. sensitivityandvariabilityanalysesforacrylonitriledespositioninhumans. ad
e
Haouzi, D., Lekehal, M., Tinel, M., Vadrot, N., Caussanel, L., Letteron, P., Toxicol.Sci.71,27–40. m
Moreau,A.,Feldmann,G.,Fau,D.,andPessayreD.(2001).Prolonged,but Sweeney, L. M., Tyler, T. R., Kirman, C. R., Corley, R. A., Reitz, R. H., ic.o
notacute,glutathionedepletionpromotesFas-mediatedmitochondrialper- Paustenbach,D.J.,Holson,J.F.,Whorton,M.D.,Thompson,K.M.,and up
meabilitytransitionandapoptosisinmice.Hepatology33,1181–1188. Gargas, M. L. (2001). Proposed occupational exposure limits for select .co
Hegllewniegra,tJio.,nGreepmrboadrudctt,ioCn.,toaxnidcitMyusrtpuhdyy,inS.WRi.st(a1r9r9a7t)s.wAicthrycliocntaicniudo:uTswado-- eTtohxyilceonle.Sgcliy.c6o2l,e1th2e4r–s1u3s9i.ng PBPK models and Monte Carlo simulations. m/tox
s
ministrationinthedrinkingwater.FoodChem.Toxicol.35,859–868. U.S.EnvironmentalProtectionAgency(EPA)(1994).Methodsforderivation ci/a
HoAekd,yJs.fBu.n,cCtiaohnialll,rAel.a,tainodnsPhaipst.oGriansot,rJo.eGnt.e(r2o0lo0g2y).1A2l2c,o2h0o4l9an–d20m6i3to.chondria: oetfryin.hOaflafitcioenorfeRfeerseenacrechcoanncdenDtreavteiolonpsmanendt,apWplaischaitniogntoonf,iDnCha.laEtPioAn/6d0o0s/i8m–- rticle
Licata, A. C., Dekant, W., Smith, C. E., and Borghoff, S. J. (2001). A 90/066F. -ab
s
physiologicallybasedpharmacokineticmodelformethyltert-butyletherin Vardeman,S.B.(1994)StatisticsforEngineeringProblemSolving.PWSPub. tra
humans:Implementingsensitivityandvariabilityanalyses.Toxicol.Sci.62, Co.,Boston.AscitedinLicataetal.(2001). ct/7
191–204. Vodicka, P., Gut, I., and Frantik, E. (1990). Effects of inhaled acrylic acid 9
/2
Lomax,L.G.,Brown,D.W.,andFrederick,C.B.(1994).Regionalhistopa- derivativesinrats.Toxicology65,209–221. /3
9
tahcoryloligcyaocfidthfoermeiotuhseer 6naosarl2c2avhiotyursfoplleorwdinagy.tIwnohawl.eTeokxsicoofl.e6xp(oSsuuprepl.t)o, Weblasscehd, Fph.,aBrmluamcoeknitnheatli,cGm.oMde.,lsaanpdpClicoanbollelyt,oRo.rgBa.no(1g9e9n5e)s.isP:hEyxstiroalpoogliactailolyn 4/16
4
445–449. between species and potential use in prenatal toxicity risk assessments. 93
Mainwaring, G., Foster, J. R., Lund, V., and Green, T. (2001). Methyl Toxicol.Lett.82–83,539–547. 63
b
y
g
u
e
s
t o
n
3
0
M
a
rc
h
2
0
1
9