field methods are not available for several of the suspected endocrine disrupters currently under
review  by the  environmental  scientific community.   An understanding of the  modes  of
lexicological action,  measurements  of the  concentrations of these substances in the various
environmental media, and improved means of modeling their movement and transformation among
environmental media are fundamental to conducting human and ecosystem risk assessments.

       Improved methods and models are needed to measure and to predict the exposure to these
substances. A modicum of understanding of exposure is also needed to design studies related to
endocrine disruption; e.g., studies of biological mechanisms and effects should address chemical
species of endocrine disruptors that are most  prevalent in the environment.   The  consensus
emerging from the scientific debate surrounding endocrine disruptors is that there are insufficient
data to resolve objectively the relative ecological and human health risks associated with these
environmental contaminants.

       The suspected  endocrine disruptors  that have been studied are predominantly organic
compounds or organic forms of heavy  metals that  are persistent, can bioaccumulate,  and  can
biomagnify  in   the  food chain.    Subtle  variations  in chemical  form  and physicochemical
characteristics  (e.g.,  planarity, isomerization, equilibria, and sorption affinities),  may  manifest
themselves in numerous ways that may affect exposure (e.g., differences in transport and routes
of exposure, increased or decreased bioavailability, changes in exposure pathways, potential for
atmospheric and hydrological transformation,  and fate).  Most.polychlorinated biphenyls, for
example, would be expected to have more affinity  for the sediment than for the water, since they
are relatively hydrophobic.  Risk  analysts and exposure researchers must understand  complex
exposure patterns, rather than net annual  exposure estimates. Developmental biology dictates that
certain  exposure windows  of vulnerability  can be  expected to follow  temporal  and  seasonal
patterns of endocrine functions.


       A national endocrine  disrupter research  program should follow EPA's risk  assessment
framework,  and explore methods and  models to  estimate and  to predict exposure  to these
substances. The exposure component of this research plan should  follow  the steps shown in
Figure  1. Calculations of exposure are a function of dose.  The U.S.  Environmental Protection
Agency (1992) defines four types of dose: potential dose (D ); applied  dose (D ); internal dose
(Dp; delivered dose (DD); and biologically effective dose (Dfi ).  The exposure pathways begins
with an organism's first contact with a substance (D ) to  its intake, absorption, and  metabolism
(D  , D  and D ) to its effect on the target organ (D  ).
  AID                          .      BE

      Measurements  of D   can often  provide  a reasonable estimate of  exposure;  i.e.,  the
concentration of a contaminant around an organism. For airborne contaminants, D is a function of
concentration, time, and ventilation.  It is difficult or impossible to measure D  directly, so D  , D
and D  are most often expressed by bioraarkers, i.e., "indicators of changes or events in human
biological systems"  (1991).   Biomarkers may either be the contaminant itself or  metabolites
indicating exposure to the contaminant; e.g., increased concentration of cotinine (a metabolite of
nicotine) in blood resulting from exposure to tobacco  smoke. Similarly, biomarkers in ecosystems
are "biochemical, physiological, or  nistological indicators  of either exposure to  or effects of
xenobiotic chemicals at the suborganismal or organismal level" (Huggett, et al., 1992). Biomarkers
can also apply to ecological exposure, although they are not often classified as measures of dose
(ecologists may apply the terms, "biotic and abiotic accumulation").  For example, Hunsaker et al.
(1990) have suggested measuring chclinesterase levels and porphyrin accumulation to indicate the
level of ecosystem exposure..

            Hazard Identification")
                                       Exposure Assessment

                                     ~~ i** ~- — — — fm,— — •••"""
                                               Effects Assessment
Figure 1   Simplified Exposure Assessment Paradigm.   A substance  is released  to  the environment,  is
           transported, may be transformed chemically and physically, and can move through various pathways;
           e.g., water, air, soil and sediment. After a substance reaches an environmental pathway; i.e., media
           change, this can be tantamount to being a new  source, in essence starting the process again from
           source characterization. The fate is then determined iteratively via mass balance.

       Methods for estimating ecosystem exposure can be similar to those for human exposure
assessment, as indicated in Figure 2,   but the methods may differ in important ways.  Both
ecosystem and human exposure assessments are often concerned with sensitive subpopulations,
many pollutants  are  both human  and ecological stressors,  and ambient  measurements for some
pollutants can be indicators of both human  and ecosystem exposure (e.g., ozone).  Human risk
assessments provide an expression of the likelihood  that an adverse  outcome will result  from a
given hazard; e.g., 10~6 chance of cervical cancer in a  population exposed to  a particular pollutant.
Ecological risk assessments are also expressions of the likelihood of an adverse outcome,  but the
expression depends  upon  the "environmental value" of  concern;  e.g.,  biological diversity,
sustainability, and aesthetics (Environmental Monitoring and Assessment Program, 1993).   A
major difference between human and ecological exposure  paradigms is the level of biological
organization at which contaminants  are  typically studied;  i.e., human  epidemiology considers
population exposure for one species (human) and medical research considers responses at various
doses and exposures for an individual  human being.  Ecological exposure  assessments often
attempt to address substances that affect the  whole ecosystem. These may include exposures to a
community (several  species), as evidenced by contaminant concentrations in certain indicator
organisms  or "sentinel"  species.  Ecosystem exposure also considers population exposures for a
target species  (e.g.,  top predator tissue concentrations of a contaminant suspected of reducing
fecundity). Ecosystem exposures can even be extrapolated from measurements of abiotic  media;
e.g., an estimate of fish community exposure extrapolated from water column concentrations of a


« Source Identification, Characterization
ft Apportionment
• Transport/Transformation/
• Environmental Concentration
• Exposure Measurements (Potential Dose)
Actual Dose
• Applied Dose
• Internal Dose
• Delivered Dose
• Biologically Effective Dose
• Biomarkers
• Source Identification, Characterization
tf Apportionment
• Transport/Transformation/
• Deposition
• Physical/Chemical Measurements of
Srressor In Ecosystem
JL Phys/Chem/Bio
^f Degradation
• Accumulation Into Abiotic and Biotic
Components of Ecosystem
• Biomarkers

Figure 2   Exposure components of risk paradigms are similar for humans and ecosystems (Vallero, 1996a).

Human exposure can be expressed as the lifetime average daily dcse (LADD).  Each route of
exposure must be considered; i.e., ingestion (water,  food, and soil), inhalation of gases  and
particles, and derma)  exposures.  Based .upon  Derelanko's (1995) expressions of LADD, total
LADD may be calculated as the sum of all LADD values via all routes:
               «        1        U

          LADDT = lifetime average daily dose (mg/kg/d) via all routes
          LADDA = lifetime average daily dose (mg/kg/d) via inhalation
          LADDj = lifetime average daily dose (mg/kg/d) via ingestion
          LADDD = lifetime average daily dose (mg/kg/d) via dermal r

Further, each route can be further, subdivided. For example, LADDT = LADDg + LADDp


          LADDg = lifetime average daily dose (mg/kg/d) from inhaling vapors;
          C = concentration in air (mg/m^);
          IR = inhalation rate (m^ /h);
          EL = exposure length (h/d);
          AF = uptake or absorption factor (dimensionless, fraction of inhaled C absorbed);
          ED = duration of exposure (d);
          BW = body wt (kg);
          TL = typical lifetime (d)

where:      LADD  = lifetime average daily  dose (mg/kg/d)  from inhaling particle matter (liquid
            and solid);
            Cp = concentration of contaminant sorbed on or in particle (mg/m^);
            PC = particle concentration in air (mg/m^);
            10"6 = converts kg to mg.

        Equations 2 and 3 also indicate that exposure models should incorporate physicochemical
properties associated with transport, transformation  and fate  in  air, soil,  water, and  sediment
transport capabilities of existing compartmental models.   For example, Figure 3  illustrates  three
idealized bimodal distributions for particles.  Such distributions can provide weight-of-evidence for
                      i             i
       COMPOSITION.-from Hidy (1975).
                                          PRIMARILY NATURAL
                                          OR QUASI-NATURAL

                I       I
               ambient aerosol;
               e.g. East Coast
               such as
               Philadelphia, and
               Washington, DC,
               Also Los Angeles
               when wind is from
LU '
CO 0.01   0.1     1.0     10
<       I       I       I	
                 Soil/dust dominated
                 ambient aerosol.; e.g.
                 desert, farming
                 Industrial grinding
                 andmining. Also, Los
                 Angeles when wind is
                 from tho nearby
                 eastern desert.
                     0.5   ~0.7  1.0-2.5-3.0
                      AERODYNAMIC DIAMETER (urn)
  0.01   0.1     1.0     10
Figure 3    Particles often display a bimodal distribution by mass, originate from multiple sources, show dynamic
            growth and reactivity, and are carriers of other  pollutants (Hidy,  1975).   The upper right mass
            distribution is typical for an area dominated by anthropogenic (combustion) sources, while the bottom
            distribution is typical for areas where particles are generated from noncombustion sources (e.g., re-
            entrained soil and mining activities).

anthropogenic or natural sources of contaminants. Karickhoff and Long (1996) have developed the
SPARC model  which characterizes the  potential environmental  fate  of substances based upon
vapor pressure, lipophih'city (e.g.,  Kow), activity coefficients, water solubility, phase partitioning

(i.e.,  Henry's Constant) and  ionization potential  (pKa).  Therefore,  physical and  chemical
characteristics, such as phase distribution and a substance's affinity to accumulate in various
environmental compartments, can profoundly affect the estimates of exposure to humans and

       Exposure assessments should be  conducted  at appropriate spatial and temporal scales,
depending upon the hypothesis or research question being investigated. Timing is complicated by
triggering and response mechanisms in the endocrine system at certain stages of development in
humans and wildlife.  These windows of exposure, where the organism is particularly vulnerable
to hormonal  dysfunction,  must be  addressed  in  any exposure  calculations.   The author

CWDDeDc = (LADDT + CW)(SF)(MT)                                                  (4)

where:    OV/DD^ = Total critical window endocrine disrupter exposure (mg/kg/d);
          CW = Additive dose during critical windows of vulnerability (mg/kg/d);
          SF =  sensitivity factors;  e.g.,  demographics  for human  populations, species
          sensitivities for wildlife (dimensionless);
          MT = maternal transfer and  transgenerational multiplier (dimensionless).


       The National  Academy of Sciences (1991)  has recommend  approaches for  assessing
human exposure to airborne pollutants (Figure 4),  emphasizing  the need for data from direct
measurements (personal and biomarker monitoring)  and from indirect approaches (especially to
gain knowledge  about activities). At  the  outset, however, exposure  research for  endocrine
disrupters should emphasize the physicochemical characterization of known  or highly suspect
endocrine  disrupters.   As indicated, even slight differences  in  physicochemical properties  can
greatly affect environmental fate; therefore exposure estimates must begin with an understanding of
how these substances  can  be  expected  to behave  in the various media and  their  fate.  The
development and adaptation of compartmental transport and fate models can be a major focus of
this research.

Characterizing   Chemodynamic  Fate

       Chemical substances reside and move through  environmental "compartments;"  i.e.,  air,
soil, water, sediment, and biota, being transported and transformed before reaching the sites of
their ultimate  fate. Such compartments can be simulated in mathematical models to predict the
persistence, bioaccumulation, bioconcentration, and  biomagnification of  these substances within
each medium, according to physicochemical properties such as vapor pressure,  water/lipid
solubility, bioaccumulation factors, and chemical half-life. Modeling can apply to both human  and
ecosystem exposures; such as food chain models, however, compartmental exposure models have
principally been  used to address  outcomes other  than endocrine disruption.   A  number of
researchers, including Cohen and Clay (1994), have  developed models to simulate the movement
and change of chemical substances in  the environment, as a function of their  physicochemical
characteristics. These properties dictate the potential sorption, transformation, transfer, and fate in
soil, sediment, water and air and uptake by biota; the ability of substances to enter the food chain;
and the magnification  of chemical concentrations at higher trophic levels for those substances that

                                     Exposure Analysis
                             Pharmacokinetic &
                          Phartnacodynamic Models
Figure  4   Possible approaches for analysis of air contaminant (From:  National Academy of Sciences,  1991).
           The dashed line between pharmacokinetic and pharmacodynamic models and exposure models have
           been added by author to show that exposure models can be derived from direct measurement data, from
           routines from other models, and from combinations of measured and derived data.

        In sediments, for example,  transformation pathways  and kinetics  are determined by  a
complex interaction of microbial, chemical, and  physical processes.   The interplay of  these
processes should  be an overarching theme  for any study  of chemical fate in the environment.
Many of the suspect endocrine disrupting chemical substances identified to date are low solubility,
neutral  organic compounds that are highly sorbed on the  organic carbon phases of sediments.
Currently available predictive tools are based on hydrophobic solution theory, and are reliable for
estimating the magnitude of sorption of such  compounds  on  sediments.  Comparable tools for
estimating the kinetics of the sorption and desorption processes are lacking.  Work is  also needed
to develop models for predicting the sorption of endocrine disrupters to  particles in sediment, soil,
water, and air, under varying environmental  conditions (e.g.,  pH,  moisture, organic matter types
and concentrations, ionic strength, and concentrations, shapes and sizes of particles).

Providing Improved Exposure Data
       Reliable and standardized data bases are vital in testing effects/exposure hypotheses,  and in
evaluating exposure and effects  models.  A strategic approach for using or modifying existing

monitoring programs to assess  current and historical effects of endocrine disrupters should be
developed. Federal and other data bases need to be reviewed for  reliability (meta-data, quality
assurance, documentation, frequency, and methods), and an assessment be given to the scientific
community regarding the data quality and the means for accessing these data (electronically and
manually).  Several existing monitoring programs that collect data could be used  to help in
problem formulations for risk assessments, or to support exposure or effect characterizations in
retrospective risk assessments.  Examples in the U.S. include the Environmental Monitoring and
Assessment Program (EMAP) of the EPA, the National Status and Trends Program (BEST) of the
Department of the Interior's National Biological Service, the National Water Quality Assessment
Program (NAWQA) administered by the U.S. Geological Survey, and a variety of state and joint
international monitoring programs.

Identifying Major Knowledge  Gaps and  Uncertainties

       The ranges  of  uncertainty  must be identified and  incorporated into exposure  models.
Improved  toxicokinetic and structure  activity models need to be  linked with the physicochemical
characteristics  of suspect endocrine  disrupters, especially at critical and sensitive early  life-stages.
Compartmental models  and laboratory studies must  be linked  to field research by  developing
mechanism-based dose response models.  Exposure levels observed in the field will be used as a
basis for identifying realistic dose ranges in laboratory experiments.

       Exposure  scientists will need specific and sensitive biofnarkers. Effects and exposure
biomarkers must be calibrated to adverse individual- and population-level effects.  Biomarkers of
exposure  are an essential adjunct to environmental measurements  in  developing and verifying
human and ecosystem exposure models.  They are also needed to screen ecosystems for exposures
and  to improve exposure estimates in future epidemiological studies. Field evaluations of these
markers should establish which are most predictive of population-level effects (i.e., which are
most useful for establishing cause and effect relationships). This necessitates the evaluation of
"normal" values and the uncertainty associated with their measurement.

Research  at  Appropriate Spatial, Biologic,  and  Temporal  Scales

       Endocrine disrupter research will take place at spatial  scales ranging from  subcellular
exposure  to regional.  Methods for assessing exposure for an individual organism  (e.g., one
human being)  differ from methods used to assess population exposure.   Likewise, estimating
exposures for  a single  ecosystem component;  e.g., a lake  or wetland, will be different from a
large-scale exposure assessment of  region or biome. Geographic scale also plays a crucial role in
model  selection.   There may  be  a  need  for predictive capability on  the micro-scale (e.g.,
occupational,  residential),  field-scale  (e.g.,  production  plant  emissions impacting  adjacent
ecosystems or  human populations),  regional scale (e.g., farm applications and resulting  human and
ecosystem exposures in an entire  watershed)  and global-scale (e.g., long-range transport and
exposure at remote sites).

       At lower levels of biological organization, a researcher may be able to determine signals of
exposure for a wide array of contaminants, and provide detailed and specific information about a
subject's activity patterns.  Often,  however, scientists are asked to estimate exposure of entire
populations or target groups,  wherein  gathering  detailed  and specific information about the
exposure  of  each  individual  in a  population  is scientifically and  economically  infeasible.
Moreover, in  the case  of ecosystems,  detailed  information  about individuals may have  less
importance than the interrelationships and diversity of a larger ecological community;  true to the
adage, "not seeing the forest for the trees." The hypothesis or study objective determines the scale
of an exposure assessment.

       The temporal scale can range from very short duration, single event to random episodic
events to long-term, discrete exposures (e.g. annual or seasonal) to continuous exposures.  The
temporality of the exposure must be considered in the study design.   For example, an episodic
event may  require monitoring until the chemical and biological indicators regain equilibrium, and
long-term,  continuous exposure studies may call for seasonal and annual time series and status and
trend assessments.   Even episodic exposures may require lengthy follow-up  studies,  however,
since the exposures may be transgenerational, and successive generations become exposed  via
maternal transfer.

Exposure  Hypothesis  Testing

       Comprehensive  investigations of a small  number of experimental  sites/systems with
problems that are known or strongly-suspected to be related to endocrine disrupters could yield
valuable information about human and wildlife exposures. Such an integrated study, conducted at
multiple levels of biological organization, with both laboratory and field components, may provide
insights into the identification of sensitive measurement endpoints and species, and extrapolation
among endpoints, species and chemicals.

       Both biological and exposure measurements need to be collected for areas expected to have
elevated  concentrations  of suspected endocrine disrupters   in  various environmental media.
Biomarker researchers can test screening tools, in situ results can be compared to in vivo and in
vitro findings, and  biologically plausible  hypotheses linking exposure and effects can be tested.
At the outset, using professional judgment, scientists could select pilot study sites based  upon
weight-of-evidence  that populations have been affected by exposure to endocrine disrupters.  Such
evidence could  be  derived from  ecological epidemiology, exposure  screens,  historical data
suggesting a likelihood that endocrine disrupters are present in one  or more environmental media,
or suggestions from source or  fate models of a "hot spot." For ecosystems, pilot studies should
address direct and  indirect effects of endocrine disruption in multiple phylogenetic groups and
trophic levels.

       An emerging area of concern is the impact of endocrine disrupters on mammalian immune
systems. Some agricultural chemicals, such as DDT and its metabolites, can act as both endocrine
disrupters and immunosuppressants.  The initial mode of action is  to suppress adrenal secretions
which, in  turn,  directly and indirectly decreases the  immune response to bacterial infection.  A
possible exposure study may include measurements of mammalian serum antibody tilers.  Some
veterinary Pharmaceuticals and antibiotics are also administered to promote growth in livestock,
which may select out the more resistant strains of enteric bacteria, the so-called "super bugs." The
fate and transport in the various environmental  media of these multi-active compounds should be
examined.  Various species of bacteria  at contaminated sites should  be analyzed  for  increased
resistance to selected antibiotics. An important exposure research question is whether the bacteria
are being transported among environmental compartments, thus spreading resistance to unexposed
bacterial populations. Identifiable bacteria strains may also prove useful as exposure biomarkers
for specific endocrine disrupting agricultural Pharmaceuticals.

       Studies will  be needed  to test and confirm results from predictive, integrated models that
incorporate structure-activity relationships, toxicokinetics, bioenergetics, environmental chemistry,
and population ecology.  They can provide a means for testing effects and exposure screening
tools, and would provide multimedia samples for analytical methods development  Although these
pilots would likely focus upon ecosystem level effects and endocrine disrupter concentrations in
environmental media, they may also present the opportunity to conduct human residential studies
compare  expected exposures from human exposure models to actual exposures under  actual
environmental conditions where weight-of-evidence suggests a human biological response; e.g.
concentrations in carpet, food,  indoor and outdoor ah, and drinking water at a small number of
sites around a former facility where a suspect endocrine disrupting substance was manufactured.

       In addition to measurements of'endocrine disrupters  in highly contaminated  areas, some
estimate of variability of contamination in different regions may be obtained via monitoring sites
where  air,  water, soil, sediment, and vegetation samples can be gathered.  These  samples can
provide valuable information about areas other than the highly contaminated areas  that will be
addressed in  the pilot studies.  The  National Exposure Research Laboratory, for  example, is
establishing such a site near each of its  laboratory facilities in Georgia, Nevada, North Carolina,
and Ohio.  At a minimum, these "near-laboratory" sites will provide a means for developing and
testing measurement protocols  for a wide range of pollutants, including suspected endocrine


       Agricultural  operations have long used many neurologically active  pesticides that have
subsequently been shown to effect the endocrine systems of animals.  Several suspect endocrine
disrupters presently under review by the U.S. EPA (1996) are agricultural chemicals or their decay
products; e.g., DDT and DDE, that are relatively lipophilic  and  tend to bioaccumulate  in the
environment.  Recently, in  response to  concerns  about groundwater contamination, pesticide
manufacturers have reformulated  pesticides to dissipate upwardly to the atmosphere to prevent
downward migration.  These changes in chemodynamics accentuate the importance of advancing
the understanding of soil-to-air fluxes.

       Several fungicides are form dated with active organochlorine and organometaUic functional
groups (Meister, 1996), often making them persistent and semi-volatile substances (vapor pressure
=. 10'2 to 10'8 kilopascals). Semi volatile compounds do not readily dissipate and can remain active
for longer periods of time  than  volatile  compounds  (Lewis  and Gordon,  1996).   Unlike
nonvolatiles,  after being incorporated into the soil, semivolatiles may later be transported from soil
to air.   This flux  rate is diffusion  controlled,  and is  proportional to vapor  pressure. In the
atmosphere, they may remain as gaseous pollutants or may be sprbed to particles that can travel
long distances and later concentrate in various environmental media and biotic tissue.

       Within soil and sediment, sorption and degradation processes exert the largest controls over
the fate  of agricultural chemicals in the environment. Adsorption  to soil particles decreases the
vapor pressures of pesticides, and is dependent upon soil conditions. The rate of evaporation from
the soil  column also affects xenobiotic  flux.  Plant uptake  rates of dissolved chemicals and
transport to the ground and surface waters are controlled by the chemicals' physical and chemical
properties and the conditions of the soil (Rao, et al. 1993).  Spencer and Cliath (1990) identified
soil water  content,  physical and chemical properties, concentration of the compound, and soil
properties, especially sou organic matter, as the most important factors controlling adsorption rates
(Nash  and Hill,  1990).   Fungicide half-life in the soil  and sediment, degradation  product
formation, and kinetics in the soil matrix and air column and in the sediment and. water column
must   be characterized properly  before reliable  endocrine  risk assessments are possible.
Characteristics of the compounds affect its potential for transport and transformation, such as water
solubility, lipophilicity, dissociation,  and molecular weight.  Characteristics of the environmental
media  also determine  transport and  transformation,  such  as  soil and solution pH and  redox
conditions, soil moisture, soil texture and structure, and type and amount of soil organic matter.

       The modes  of toxic actions of these substances are determined by the physicochemical
characteristics. These same characteristics may influence a compound's ability to persist and to
bioaccumulate, and to elicit longer term effects in humans and  wildlife.  Slight variations  in
physicochemical characteristics, such as planarity and isomerization, may drastically change soil
fungicide exposure and effects.

       Movement and  transformation in the environment can  be illustrated by two  fungicide
classes  associated with endocrine  effects in humans and  wildlife;   i.e.,  dicarboximides  and


       Vinclozolin is a dicarboximide fungicide whose structure is shown in Figure 5. Since it has
a dichlorobenzene group, Vinclozolin may be classified as an organochlorine compound, the group
presently most often associated with endocrine disruption. It may also be classified as a carbamate
pesticide, since  the  right  side of the structure is derived from  carbamic acid.   Two principal
degradation  products  result from  opening  the  carbamate  ring:    a  buteonic  acid  (2-[[(3,5-
Dichlorophenyl)-carbamoyl]oxy]-2-methyl-3-butenoic acid), referred to as  "Ml"; and an enanilide
(3',5'-Dichloro-2-hydroxy-2-methylbut-3-enanilide),  called "M2".  Both degradation products
have been isolated in plants and soils (Kelce,  et al., 1994). Ml is a reversible reaction, where the
carbamate ring closes and returns  to the Vinclozolin structure, whereas M2 is  a non-reversible
degradation product.

       Vinclozolin is  registered  in the  United States and  Europe  as  a fungicide for grapes,
strawberries, sunflowers, rape seed, soft fruits, hops, ornamental plants. It has been shown to
alter mammalian sex differentiation by inhibiting androgen receptor activity (Kelce et al, 1994).
Developing fetuses are extremely sensitive to vinclozohn exposure; exposures to rat fetuses has
been associated with infertility, deformed genitalia, and reduced sperm counts (Wong et al., 1995;
Gray, et al., 1994).  Kelce, et al.  (1996) have found that it induces antiandrogenic developmental
effects in vivo and that it inhibits androgen receptor (AR)  binding and AR gene expression in vitro.
           Vln clozolln
Figure 5   Structural formulae and degradation pathways of Vinclozolin and its principal degradation products
           (from Szeto, et al., 1989a).

       Physical and chemical processes determine the amount and form of Vinclozolin that may
reach the soil.  Application rates and methods affect Vinclozolin degradation and persistence,
since it may applied be to foliage and above-ground plant parts and migrate downward to soils,
or it may be incorporated directly into soil; e.g., it is used to prevent onion white rot and other
fungi in bulb crops (Meister, 1996).

       Szeto,  et al. (1989b) found that the pH of the soil, sediment, and  water column  is a
principal factor in vinclozolin degradation rates.  The degradation is quite rapid at higher pH, and
much slower at low pH. At 35 ° C, the half-life at pH 8.3 is less than one hour, and at pH 4.5 is
530  hours.  This difference can be explained  in part to vinclozolin's increased resistance  to
hydrolysis  at  lower  pH.    The pH also  determines the  principal degradation pathway that
vinclozolin will take, with higher pH yielding more Ml and lower pH yielding more M2. A third
degradation product, 3,5-dichloroaniline, has been detected after considerable time  (672.3 h at pH
6.5,  1537  h at pH 5.5,  and 505.8  h at pH 4.5).  This points to  important considerations for
estimating  the fate of vinclozolin.  Not only does increased  soil and solution acidity increase
vinclozolin's persistence, but acidity also influences the degradation  pathways and the appearance
of secondary degradation products.

       The type of application solvent also influences  vinclozolin degradation and bioavailability
of vinclozolin.  Szeto, et al (1989c) compared the persistence of vinclozolin by applying the
fungicide in water and acetone solutions to young pea plants and analyzing the concentration of
vinclozolin in leaflets.  The acetone-vinclozolin solution was  significantly more persistent than the
water solution. This is likely the result of the dichorophenyl  group's influence on  the lipophilicity
of vinclozolin. The persistence observed in the field  was lower than in  the laboratory studies,
likely to do increased photodegradation and  greater  moisture gradients. However, these findings
indicate the importance of the original application solutions and suspensions when estimating the
ultimate fate of vinclozolin.  They also  indicate the importance of available  moisture in the soil
column, which could serve to dilute the lipophilic application solvents over time.

       The amount and degradation pathways in the atmosphere and fluxes  from soil to  air are
potentially  important research areas.  Since vinclozolin's vapor pressure is 1.6 xlO'5 kilopascals
(kPa) at 20° C, which is considered semivolatile, the atmospheric transport could be an important
exposure pathway. The  volatility of its degradation products should also  be  considered in these

Organotin Fungicides

       Organometals are an  important group of endocrine disrupters.  These are compounds with
covalent bonds between  a carbon  (C) and metal atom (Pelletier, 1995).  Organotins have  been
associated with endocrine dysfunction (Briischweiler, etai, 1996; Ochlmann, et al., 1993). Their
structures are shown in Figure 6.
     Triphenyltin Acetate                  Triphenyltin Hydroxide

Figure  6   Structural formulae of triphenyltin acetate and triphenyltin hydroxide.

       Agricultural fungicide and scab biocide applications comprise the principal source of the
triphenyltins. Triphenyltin hydroxide is directly applied to soil and plant tissue as a fungicide and is
used to treat seeds, root, and tubers.  Triphenyltin hydroxide is also the degradation product of
triphenyltin acetate, a fungicide for potatoes, rice, and sugar beets, and scab biocides for peanuts
(Tnayer, 1974). Other tin (Sn) compounds are used as fungicides and  antifouling  agents, most
notably tributyltin; but their usage has been greatly restricted in the past decade.

       The mechanisms of toxic action for organotins are more  diverse than those of vinclozolin.
They include cytotoxicity in  the liver, disturbance of calcium homeostasis and  induction of
apoptosis   in   thympcytes,   inhibition   of  ATP-synthesis   and   mitochondrial  oxidative
phosphorylation, inhibition  and uncoupling of  chloroplasts, ion  pump inhibition and   cell
membrane damage, Cytochrome P450 inhibition, and intracellular enzyme inhibition (Pent, 1996).
The principal endocrine endpoint observed to date is increased imposexual response in gastropods.
Mud snails and other  gastropods exposed to tributyltin compounds exhibit an increased incidence
of pseudohermaphroditism (Bryan, et al., 1989 and Smith, 1981). Like mercury, Sn is neurotpxic.
The endocrine response may be the indirect result of Sn  activity in  neurological system, which
induces a chain of endogenous responses that ultimately elicit an endocrine response.

       Most Sn-related research has focused on the aquatic environment, but fate and transport in
soils and the atmosphere must also be better explained. Figure 7 shows  the principal degradation
pathway from the less stable acetate compounds  to alkylated tin species mediated by microbes,
which theoretically will ultimately degrade the fungicides and metabolites to elemental Sn  (Keijzer
and Loch,  1995).   The process would likely move in  the opposite direction under  aerobic,
microbially mediated conditions (from oxides and elemental tins to alkylated species).

       The amount of Sn species fluxing  to the atmosphere is an area in need of research.  The
volatility varies  by species.  The alkyltins are more  volatile than  the aryltins.  Methyltins, for
example, are significantly more volatile than the phenyltins.  Within the phenyltins, triphenyltin
acetate's vapor  pressure  is  1.9 xlO"6 kPa  at 20° C, which is considered semivolatile,  but
triphenyltin hydroxide is a salt that is nonvolatile. The volatility of each degradation product should
also be studied.
MkrobM    2
Figure 7   Degradation pathway expected for tin fungicides under aerobic soil conditions (from Keijzer and Loch,
           1995).                                          '                             '   •

       Even  the less stable organotin species can remain undegraded in soil under certain redox
and acidity conditions, and if soil organic matter is plentiful. For example, triphenyltin acetate has
been shown to accumulate in organic-rich upper soil horizons as a result of heavy application rates
in the Netherlands (Keijzer and Loch,  1995). This phenomenon can have profound effects on the
ability of certain tin compounds to migrate, after deposition under certain atmospheric and aquatic
conditions, but  to remain in  place  for  much longer  time  periods under  other  conditions.
Environmental acidity and redox conditions can also be affected by type of organic matter present
in soils and sediment.  Kuballa, et al. (1995) found that humic substances act as methylating agents
for Sn in sediment in reduced soil profiles. Therefore, anaerobic soil conditions; e.g.,  in wetlands
and rice paddies, can play a major role in Sn methylation. Little is known about the persistence of

the triphenyliin fungicides, but the persistence appears to be enhanced in reduced environments
(Pent, 1996).   Speciation of organotins is also pH dependent.  Within normal  natural ranges,
higher pH values, where the hydroxides dominate, generally increase acute toxicity and uptake by
organisms.  However, this is mediated by the presence  of  organic material.  Soil moisture will
affect speciation of the tin compounds.  For example, triphenyltin acetate is  relatively hydrophilic
(9 mg/1 at 20° C), and is not lipophilic. Application rates will also affect soil-air fluxes by kinetic
and cheraicd processes; e.g., mass action.

       In aquatic systems, increasing dissolved  organic carbon  reduces bioavaUability by the
creation of Sn-organic complexes.  Therefore, organic matter in soils likely will play similar roles,
especially under saturated conditions.

       Accurate means of extracting, separating, and detecting the  various organotin species is
necessary  to  generate reliable estimates  of exposures and risks.   Analytical techniques for
speciating  tins are  improving. Barshick,  et alr  (1996) recently  found new  chromatographic
methods that show promise in improving the ability to speciate inorganic and organic forms of Sn
from a single soil.  Future research applying th•• procedures to different soil types and specific Sn-
related research should advance analytical capability even further.


       The challenge of addressing endocrine disrupter risks and exposures is daunting; however,
is has been shown that EPA's risk paradigm relates well to both human and ecosystem exposure
assessment.  Presently, some strong weight-of-evidence in isolated studies of invertebrates, fish,
reptiles, birds and mammals has provided compelling reasons for linking exposures to a number of
chemical compounds to  endocrine disruption mechanisms in populations.   These  ecosystem
observations suggest that similar human  weight-of-evidence data bases, could  be valuable  in
directing future human endocrine disrupter exposure research.

       The chemical classes represented by  agricultural fungicides are sufficiently different in
chemical structure and mechanisms of endocrine action to provide insights into the chemodynamic
factors that are likely to influence  human and ecosystem  exposures. The two fungicides reviewed
in the present study represent chemical groups that have been associated with  in vitro, in vivo, and
in situ endocrine effects.   .

       In soils and sediments, transformation pathways and  kinetics are determined by a complex
interaction of processes that affect the amount and the degree of speciation  of substances in the
soil, and the potential for their release and uptake by the atmosphere, surface water, groundwater,
and biota. Transport and transformation are  function of characteristics of the compounds and
characteristics of the environmental media. Slight variations in physicochemical characteristics may
drastically change the potential for exposure and effects.

       The cheraodynamic behavior of fungicides in soil is basic to predicting future exposures
and the efficacy of agricultural endocrine disruptor exposure prevention strategies. Improved flux
measurement methods to screen and to model exposures  to endocrine disrupting pesticides in
various media nee1.' to  be developed, validated, and incorporated into test  guidelines, especially
those required unc~: Toxic Substances and  Control Act and Federal  Insecticide, Fungicide and
Rodenticide Act.

ACKNOWLEDGMENTS              ,

       I thank Dr. Ellen Cooler, Dr. Nancy Wilson, and Joe Bumgarner of  the National Exposure
Research  Laboratory for critical  reading  of the  manuscript  .  Dr.  Cooler's insights regarding
•modeling and deposition  were very helpful. Dr. Wilson's advice regarding chemical structures and

Mr. Bumgamer's insights about nontraditonal pollutants improved the quality of the paper.  I am
grateful to Dr. William Kelce of the .National Health and Environmental Effects  Laboratory for
recommendations concerning toxicity  and fate of vinclozolin. Finally, I acknowledge the valuable
advice of Dr. J. Jeffrey Peirce of Duke University and Dr. Viney P. Aneja of North Carolina State
University regarding contaminant fluxes from soil to the atmosphere.


Barshick, C.M., Barshick, S-A., Mohill, M.L., Britt, P.P. and Smith, D.H. (1996)  Elemental
      and  organometallic analyses  of soil, using  glow discharge mass spectrometry  and gas
      chromatography/mass spectrometry. Rapid Comm.  Mass Spec.  10:341-346.
Briischweiler, B.J.,  WUrgler, F.E., and Pent, K.  (1996) Inhibition  of Cytpchrome P4501A by
      organotins in fish hematoraa  cells PLHC-1.  Environ. Toxicol  Chem.  157: 728-735.
Bryan, G.W., Gibbs, P.E., Huggett, R.J., Curtis,  L.A., Bailey, D.S., and Dauer, D.M. (1989)
      Effects of tributyltin pollution on the mud snail. Ilyanassa obsoleta, from the York River and
      Sarah Creek, Chesapeake Bay. Mar. Pollut. Bull., 20: 458.
Cohen, Y. and Clay, E.G.  (1994).  Multimedia Partioning of Particle-Bound Organics. J: Haz.
    Materials. 37: 507-526.
Colborn, T., von Saal, F.S., and Soto, A.M. (1993) Environ. Health Perspect., 101(5), 378-384.
Derelanko, M.J.  (1995). Risk Assessment Calculations.  In:   CRC Handbook  of Toxicology.
    Derelanko, M.J. and Hollinger, M.A. editors. CRC Press, Boca Raton, 653-656.
Environmental Monitoring  and Assessment  Program (1993).   Environmental Monitoring and .
    Assessment Program:  Master  Glossary. EPA/620/R-93/013. Research Triangle Park,  NC:
    U.S. Environmental Protection  Agency, Office of Research and Development.
Feni,  K. (1996) Ecotoxicology of organotin compounds.  Cm. Reviews Toxicol. 26(1):  1-117.
Gray, L.E., Monosson, £., and Kelce, W.R.  (1996)  Emerging issues: the effects of endocrine
      disrupters on  reproductive  development.   In:   Interconnections Between Human and
      Ecosystem Health. Ed.: R.T. iJiGiulio and E. Monosson, Chapman and Hall: London, 45-
Gray, L.E.,  Ostby, J.S., and  Kelce, W.R.  (1994)  Developmental  effects of an environmental
      antiandrogen:  the fungicide vinclozolin alters sex differentiation of the male rats.   Toxicol
     Appl. Pharmacol. 129: 46-52.
Hidy, G.M. (1975). Summary of the California Aerosol Characterization Experiment. J.  Air Poll
    Com. Assoc.  25:  1106-1114.
Huggett, R.J., Kimerle, R.A.,  Mehrle,  Jr., P.M., and Bergman, K.L. eds.  1992.  Biomarkers-
    Biochemical, Physiological, and Histological Markers of Anthropogenic Stress.  Boca Raton,
    LA:  Lewis Publishers.
Hunsaker, C.T., MacCarthy, J.F., Shugart, L.R., and O'Neill, R.V. (1990).  Indicators  Relevant
    to Multiple Resource Categories.  In: Ecological Indicators for the  Environmental Monitoring
    and Assessment Program.       Hunsaker, C.T and Carpenter, D.E., editors , 2-1 - 2-18.
    EPA 600/3-90/060.  Research  Triangle Park, NC:  U.S. Environmental Protection  Agency,
    Office of Research and Development.
Karickhoff, S.W.  and  Long, J.M.   (1996).  Research Program: SPARC.  National Exposure
    Research Laboratory Briefing.
Kavlock, R.J., Daston, G.P.,  DeRosa  C., Fenner-Crisp P.,  Gray L.E., Kaattari  S., Lucier G.,
     Luster M., Mac M.J., Marzka C.,  Miller R., Moore J.,  Rolland R.,  Scott G.,   Sheehan
     D.M., Sinks T.,  and Tilson,  H.A. (1996).  Research Needs for the Risk  Assessment of
     Health  and Environmental Effects of Endocrine Disrupters: A Report of  the US  EPA
     Sponsored Workshop. Environmental Health Perspectives (in press).
Keijzer, T.J.S. and Loch, J.P.G. (1995)  Accumulation of HN03-extractable tin in agricultural and
     non-agricultural soils by use of triphenyltin acetate. Water, Air and Soil Pollution. 83: 287-

Kelce, W.R.,  Lambright, C.R., Gray, L.E.,  and Roberts, K.P. (1996)  Vinclozolin and p,p'-
     DDE alier androgen-dependent gene expression:  In vivo confirmation of an androgen
     receptor-mediated mechanism.  Toxicol. Appl. Pharmacol. 141: 275-285.
K(":e,  W.R.,  Monosson, E., Gamcsik, M.P.  (1994)  Environmental  hormone  disrupters:
     evidence that vinclozolin developmental toxicity is mediated by anti-androgenic metabolites.
     Toxicol. Appl. Pharmacol. 126: (in press).
Kuballa, J., Desai, M.V.M., and Wilken, R-D. (1995)  Influence of humic substances on sorption
     and methylation processes in inorganic- and organo-tin species. Applied Organomet.  Chem.
     9: 629-638.
Lewis, R.G. and Gordon, S.M.  (1996)  Sampling for organic chemicals in air. In: Principles of
     Environmental Sampling. Ed.:  American Chemical Society: Washington, DC, 401-470.
McCurdy, T.  (In press).  Modeling the  "Dose Profile" in Human Exposure Assessments:  Ozone
    as an Example. In:  Reviews in Toxicology. Series D - In Vivo and Risk Assessment.  Ed.:
    E. Hodgson, IOS Press, Inc.: Amsterdam.
Meister, R.T., Ed. (1996)   Farm Chemicals Handbook '96.  Meister Publishing  Company:
     Salem, MA.
Nash,  R.G.  and Hill,  B.D.  (1990) Modeling pesticide volatilization and  soil decline .under
     controlled conditions.  In:  Long  Range Transport of Pesticides. Ed.: D.A. Kurtz, Lewis
     Publishers, Inc.: Chelsea, MI,  17-28.
National Academy of Sciences, National Research Council (1991).  Human Exposure Assessment
    for Airbone Pollutants: Advances and Opportunities.  Washington, DC: National Academy
Ochlmann, J. Stroben,  E., Schulte-Ochlmann, U.. Bauer, B., Fioroni,  P., and  Markert, B.
     (1993)  Tributyltin biomonitoring using prosbranchs as sentinel organisms. Fresenius J.
     Anal, Chem. 7:  1003-1008. .
Pelletier, E. (1995) Environmental organometallic chemistry of mercury, tin, and lead:  Present
     status and perspectives.  In: Metal Speciation and Bio availability in Aquatic Systems. Ed:
     A. Tessier and D.R. Turner, John  Wiley and Sons Ltd., West Sussex: 103-148.
Rao, P.S.C.,  Bellin, C.A...& Brusseau, M.L. (1993)  Coupling  biodegradation  of organic
     chemicals to sorption and transport in soils and  aquifers. In:  Sorption  and degradation  of
     pesticides and organic  chemicals in soil.  SSSA Special Publication no.  32.    SSSA,
     Madison, WI:  1-26.
Science Advisory Board, Environmental  Futures Committee,  U.S.   Environmental Protection
     Agency, January 1995; Beyond The  Horizon: Using Foresight to Protect the Environmental
Smith, B.S.  (1981) Tributyltin compounds induce male characteristics on female  mud snails
   Nassarius obsoletus ~ Ilynanassa obsoleta. J. Appl. Toxicol. 1: 141.
Spencer, W.F. and Cliath, M.M. (1990) Movement of pesticides from soil to the atmosphere. In:
   Long Range Transport of Pesticides. Ed.: D.A. Kurtz, Lewis Publishers, Inc:  Chelsea,  MI, 1-
Szeto, S.Y., Burlinson,  N.E., Rahe, J.E.,  and Oloffs,  P.C. (1989a) Kinetics of hydrolysis of the
   dicarboximide fungicide vinclozolin. J. Agric. Food Chem.  37:  523-529.
Szeto, S.Y., Burlinson, N.E., Rettig, S.J., and Trotter, J. (1989b) Identification of hydrolysis
   products of the fungicide vinclozolin by spectroscopic and x-ray crystallographic methods.  J.
   Agric. Food Chem. 371103-1108.
Szeto, S.Y., Burlinson, N.E., Rahe,  J.E.,  and Oloffs, P.C. (1989c) Persistence of the fungicide
   vinclozolin on p~a leaves under laboratory conditions. J. Agric. Food Chem.  37:  529-534.
Thayer, J.S. (1974) Organometallic compounds in living organisms. J. Organomet. Chem.  76:
U.S. Environmental Protection Agency  (1996) Working list of endocrine disrupters for  exposure
   research (unpublished). National Exposure Research Laboratory.
U.S.  Environmental  Protection Agency.  (1992)   Federal Register.   Guidelines  for Exposure
    Assessment. Vol. 57. No. 104. FRL-4129-5.

Vallero, D.A. (1996a) Transport, transformation and fate of endocrine disrupters:  potential areas
   of exposure research. In:  Measurement of Toxic and Related Air Pollutants: VIP-64.  Air and
   Waste Management Association:  Pittsburgh, pp. 541-552.
Vallero, D.A. (1996b).  Exposure  Assessment Methodologies  for Humans and Ecosystems. IN:
    NATO AS1  Series, Partnership Sub-Series, 2.  Environment:  Urban Air Pollution. AUegrini, I.
    and De Sands, F. editors. Springer-Verlag: Berlin Heidelberg, 8: 449-466.
Wong, C-I,  Kelce, W.R., Madhabananda, S., and Wilson, E.M.  (1995)  Androgen  receptor
   antagonist versus agonist activities of the fungicide vinclozolin relative to hydroxyflutamide. J.
   Biol. Chem.  270 (34): 19998-20003.


       This  paper has been reviewed  in  accordance with  the United  States  Environmental
Protection Agency's peer and administrative review policies and approved for presentation and
publication.  Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.