  SEQ CHAPTER \h \r 1 UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

WASHINGTON, D.C. 20460

OFFICE OF                  

PREVENTION, PESTICIDES AND 

TOXIC SUBSTANCES        

MEMORANDUM

DATE:		April 5, 2010

SUBJECT:	Charge questions for the FIFRA Scientific Advisory Panel
meeting on “Re-Evaluation of Human Health Effects of Atrazine:  Review
of Experimental Animal and In Vitro Studies and Drinking Water
Monitoring Frequency” to be held April 26-29, 2010

FROM:	Tina Levine, Ph.D. 

		Director

		Health Effects Division (7509P)

		

TO:		Joseph Bailey

		Designated Federal Official

		FIFRA Scientific Advisory Panel

		Office of Science Coordination and Policy (7201M)

On March 30, 2010, EPA transmitted its draft issue paper entitled
“Re-Evaluation of Human Health Effects of Atrazine:  Review of
Experimental Animal and In vitro Studies and Drinking Water Monitoring
Frequency” to the FIFRA Scientific Advisory Panel (SAP) in preparation
for the April 26-29, 2010 meeting.  This memo contains the charge
questions for the April SAP meeting.

SAP CHARGE

As part of the re-evaluation of the health effects of atrazine, three
meetings of the FIFRA Scientific Advisory Panel (SAP) are scheduled in
2010.  The first of these was held February 2-4, 2010.  In preparation
for the April meeting, the Agency developed a draft issue paper focused
on 1) a review of the recent scientific literature on the mode of action
and toxicological profile of atrazine and 2) approaches to determine
appropriate drinking water monitoring frequencies.   The purpose of the
April meeting is to solicit feedback from the SAP on the status and
overall scientific direction of the Agency’s re-evaluation in these
two areas including preliminary conclusions about some aspects of the
experimental toxicology data.  The Agency will use the Panel’s
feedback to inform the weight of the evidence (WOE) analysis that will
be developed in the coming months and reviewed at the September 2010
SAP.  The experimental toxicology portion of the draft April issue paper
(Sections 3 & 4, Appendices A & B) represents the state of the science
up to January 30, 2010.  Since that time, new data have become available
and more will become available in the spring and summer of 2010 that
will provide further characterization of some key areas, such as
describing the causal linkage between key events in the mode of action,
the dose-response concordance of these key events, and the relationship
between the precursor events with functional outcomes.  These new
studies will be incorporated into the September SAP.

Question 1.0:  In Vivo & In Vitro Experimental Toxicology in Mammals

In 2003, the human health risk assessment was based on a mode of action
in which exposure to atrazine leads to a reduced release of gonadotropin
releasing hormone (GnRH) from the hypothalamus thereby lessening the
afternoon pituitary luteinizing hormone (LH) surge in female Sprague
Dawley rats.  As a result, the estrus cycle lengthens.  This, in turn,
leads to increased estrogen levels and an increased incidence of mammary
tumors in female Sprague Dawley rats.  EPA determined that atrazine’s
cancer mode of action (i.e., premature reproductive aging) in the
Sprague-Dawley rat is not likely to be operative in humans (a conclusion
consistent with the SAP recommendation in 2000).  The Agency will be
evaluating the cancer classification further as the experimental
toxicology data are integrated with the epidemiology literature; new
experimental toxicology studies (Section 3.6 of the draft issue paper)
do not suggest a change to the Agency’s previous conclusion that
atrazine is “not likely to be carcinogenic to humans.”   

Although the cancer mode of action may not be operative in humans, it is
not unreasonable to assume that atrazine might cause adverse effects on
hypothalamic-pituitary function in humans.  Thus, the same endocrine
perturbations that induce tumors in rats may play a role in at least
some reproductive/developmental effects (not associated with
reproductive aging) that may be relevant to humans.  Accordingly, the
Agency identified disruption of estrous cyclicity and delays in puberty
onset (males and females) occurring as a consequence of disruptions to
the hypothalamic-pituitary-gonadal (HPG) axis as the critical endpoints
of concern (Sections 2.0 & 3.2 of the draft issue paper).  

The Agency continues to believe that this disruption of the HPG axis is
critical to affecting estrous cyclicity and delayed puberty onset. 
However, as discussed in Questions 1.2-1.6 (See Figure 3 of the draft
issue paper), new information has been published that indicates a more
expanded understanding of how atrazine may perturb the pathway leading
to reproductive effects.   These new data indicate that atrazine
disrupts the hypothalamic-pituitary-adrenal (HPA) axis altering the
central nervous system’s control of the pituitary and adrenal which,
in turn, disrupts the HPG axis. Furthermore, a hypothesis for a direct
effect of atrazine on the adrenal cortex and the gonads which may impact
steroidogenesis directly is also supported by the available data.   

Question 1.1

With the caveat that the review of the epidemiological literature is
still on-going, please comment on the Agency’s preliminary conclusion
that new experimental data from in vitro and in vivo laboratory animal
studies do not support a change in the conclusions from the 2003 risk
assessment that atrazine is unlikely to be a human carcinogen.

Question 1.2

Based on an evaluation of the studies examining the mode of action of
atrazine on neuroendocrine function (Section 3.2 of the draft issue
paper), the Agency has preliminarily concluded that atrazine affects
both the HPG axis and the HPA axis.  With respect to the temporal
concordance, recent studies show that atrazine induces a rapid (within
minutes) increase in ACTH and adrenal cortical hormones (corticosterone
and progesterone) in both male and female rats (Fraites et al., 2009;
Laws et al., 2009; Pruett et al., 2009), while changes in the HPG axis,
such as the suppression of the LH surge, may take up to 2-3 days (Cooper
et al., 2009).  This sequence of HPA alterations followed by HPG changes
indicates that the suppression of the surge may be, in part, mediated by
activation of the HPA axis (i.e., a corticosterone suppression of the
GnRH and LH release).  In addition, with respect to dose-response
concordance, atrazine-induced increases in ACTH, corticosterone and
adrenal progesterone are seen following a single dose of atrazine (50
mg/kg).   In contrast, the HPG is not altered following a single dose up
to 200 mg/kg. However, when treated for multiple days, the dose
necessary to alter the HPA axis is lower than or equal to the one needed
to affect the HPG axis.  

Please comment on the Agency’s preliminary hypothesis for the mode of
action involving atrazine’s alteration of both the HPA and HPG axes. 
Does the document adequately and clearly describe the hypothesis in the
context of a “toxicity pathway” (i.e. cellular response pathways
that, when sufficiently perturbed, are expected to result in adverse
health effects)?  Does the document clearly describe the data used to
test the proposed hypothesis? To what extent do the available data
establish key events in the proposed mode of action hypothesis?  What
are the strengths and limitations of the data available on this
hypothesis?  

 

Please include in your comments a discussion of the Agency’s
interpretations of the data linking the initial perturbations in HPA
axis to changes in the HPG axis.  

Please comment on the evidence that the initial perturbations in the HPA
axis may lead to impairment in reproductive function and/or
developmental consequences.  Which event(s) is/are viewed as critical in
leading to health consequences? Are there data on other substances that
would inform this question?

Question 1.3

The Agency has preliminarily concluded that atrazine directly targets
cells within the HPA axis (Sections 3.2 & 3.3.2.3 of the draft issue
paper) and that the increased activity of the adrenal axis does not
reflect nonspecific stress.  Evidence for this conclusion is based on
the following observations.  

Several studies report increased adrenocorticotropic hormone (ACTH) and
corticosterone levels immediately following a single exposure to
atrazine and its two metabolites (deisopropyl-atrazine, DIA;
deethyl-atrazine, DEA) but not diamino-s-chlorotriazine (DACT; Laws et
al., 2009; Fraites et al., 2009; Pruett et al., 2009). 

A single intravenous administration of DEA in vivo (Fraites et al.,
2009) also causes an immediate increase in ACTH and adrenal hormone
release, indicating that a chlorotriazine-induced gastrointestinal
distress is not driving the hormonal response.

The adrenal corticosterone response to continued oral exposure to
atrazine did not habituate in rats (Fraites et al., 2009; Laws et al.,
2009) or mice (Pruett et al., 2009).  

Although, atrazine induced a rapid increase in ACTH, prolactin secretion
was not affected in the same animals (Laws et al., 2009).  This is
unusual in that an increase in both ACTH and prolactin are typically
observed in response to a number of physiological and psychological
stressors.

Please comment on the extent to which the available evidence supports
the preliminary conclusion that atrazine and its intermediate
metabolites (DIA and DEA) induce changes in the HPA axis as a result of
a direct action on the HPA tissue and such changes are not due to a
generalized or non-specific stress response.  Are there data on other
substances that would support this conclusion?

Question 1.4

Based on a review of in vitro studies evaluating the effects of atrazine
on estrogen production, the Agency has preliminarily concluded that
atrazine does not have a direct effect on the catalytic activity of
aromatase.  However, with continued exposure (> 24 hrs) atrazine can
cause increased estrone and estradiol production in the H295R and JEG-3
cell lines (Sanderson et al., 2000; Laville 2006; Higley et al., 2010). 
 These changes in estrogen production have been associated with
increased cAMP and CYP19 mRNA (Sanderson et al., 2000, 2001) that are
part of a complex mode of action through which atrazine up-regulates the
gene expression of aromatase and possibly other enzymes within the
steroidogenic pathway (Section 3.3.2.3 & Figure 3, Arrow 7of the draft
issue paper).  The hypothesis that atrazine alters multiple steroids is
supported by  increased testosterone concentrations in H295R cells
following exposure to atrazine (Higley et al., 2010), and in vitro
studies demonstrating that atrazine affects a number of signal
transduction pathways and/or transcription factors in a variety of cell
lines (Suzawa and Ingraham, 2008; Albanito et al., 2008).  Up-
regulation of the expression of a group of major steroidogenic genes has
been observed in JEG-3 cells (Suzawa and Ingraham, 2008).  In vivo,
there is little evidence that atrazine alters aromatase per se (Modic,
2004) and the evidence associating atrazine exposures to increases in
serum estrogens appears to be related to a general increase in gonadal
and adrenal progesterone and androstanedione (Modic 2004; Laws et al.,
2009).  

Please comment on the extent that the available data do or do not
support the Agency’s preliminary conclusion that atrazine has a
general stimulatory effect on steroidogenesis, as opposed to a direct
effect on aromatase.  

Question 1.5

A series of studies evaluating the effect of either peripubertal or
gestational exposure of atrazine in the male rat indicate that atrazine
decreases testosterone concentrations and may lower androgen dependent
tissue weights but that it has little effect on basal LH secretion in
the male.  This response generally occurs at doses of 50 mg/kg and
above.  In contrast, atrazine in vitro has been found to consistently
increase steroid hormone production.  Thus, there is an apparent
discrepancy between the in vivo and in vitro data.  However, our
understanding of the interplay between the adrenal and the gonadal axes
may provide an explanation.   At the doses used in the in vivo studies,
atrazine induces an activation of the adrenal axis (increased ACTH and
corticosterone) and the increase in corticosterone does not habituate
(i.e., it does not lessen with repeated dosing). There are a number of
studies in the literature demonstrating that prolonged corticosterone
stimulation will impair testosterone production by a direct action on
steroidogenesis in the Leydig cells. Therefore, the in vivo effects on
testosterone synthesis are likely the result of an adrenal
hormone-mediated down regulation of HPG axis (through corticotropin
releasing hormone [CRH] modulation of GnRH pulsatility) or directly on
the testis (through a glucocorticoid receptor mediated change in
steroidogenesis).  

Please comment on the degree to which the proposed mode of action
(Figure 3 of the draft issue paper) provides a biologically plausible
explanation for the decrease in serum and testicular testosterone
identified in the in vivo studies.  Please comment on the extent that
the available data do or do not support this hypothesis.  

Question 1.6

Based on a review of the neurotoxicity studies (Section 3.4 of the draft
issue paper), the Agency has preliminarily concluded that several recent
studies provided further support for the concern that dopaminergic
neurotransmission may be affected by atrazine, a concern that was raised
by studies prior to 2003.  However, the Agency has also concluded that
several aspects of these studies of the dopaminergic neuronal pathways,
in particular the changes noted in stereological evaluations and the
observed changes in behavior, should be considered as preliminary
findings.  This conclusion is based on the notable limitations
identified in the data including:  (1) lack of clear dose-response
relationships, (2) lack of inclusion of suitable positive controls to
confirm the competency and reliability of the procedures utilized in
examining dopaminergic systems in the brain, (3) limited data to
corroborate stereological findings, (4) limited or no supporting
histological and behavioral assessments, and (5) no consideration of the
potential role of the HPA axis (e.g. alterations in corticosterone).  
EPA has further determined that two non-dopaminergic neurotoxicity
studies--one on brain somatostatinergic systems and the other on
neurobehavior in mice -- also have significant limitations (e.g., lack
of details on source and purity of atrazine, and age and body weights of
the mice; poor quality of amino cupric silver staining photomicrographs;
limited presentation of reverse-transcriptase polymerase chain results;
lack of data to corroborate conclusion of neuronal degeneration;
reference citations did not support statements made in the text; no
explanation of biological plausibility of alterations in
somatostatinergic receptor subtypes; inappropriate statistical
evaluation of data (litter vs. pup); discrepancies in number of pups
examined for behavioral endpoints, and lack of objective and validated
behavioral tests).  Please comment on these preliminary conclusions
regarding the neurotoxicity findings.  

The Agency has preliminarily concluded that the available studies
indicate that the neurotoxicity endpoints examined are not more
sensitive than those evaluated for neuroendocrine function following
atrazine exposure (e.g., attenuation of LH surge and estrous cycle
disruptions in female Sprague-Dawley rats which form the basis for the
current chronic RfD).  In addition, the Agency has concluded that there
is no association between atrazine exposure and development of
Parkinson’s Disease on the basis of non-specificity of effects on
brain dopaminergic systems, lack of histological and behavioral features
characteristic of Parkinson’s Disease, and results of epidemiological
studies. Please comment on this preliminary conclusion. 

Question 1.7

Based on a review of two studies of the potential effects of atrazine on
the prostate (Section 3.3.2.2 of the draft issue paper), the Agency has
concluded that the results of the Rayner et al. (2007) study of pregnant
rats treated during gestation support previous observations (Stoker et
al., 1999) that atrazine treatment to the rat dam either perinatally or
early postnatally can increase prostate weights (due to an increase in
inflammatory infiltrate) in the male offspring.  This effect on the
offspring was shown to be due to a suppression of prolactin in the
atrazine-exposed dams during lactation, and is consistent with the mode
of action of atrazine on neuroendocrine function.   The Agency also
preliminarily concluded that further research is needed to provide more
convincing evidence that atrazine-mediated suppression of prostate
cancer in the probasin/SV40T antigen transgenic rat (androgen-dependent
prostate cancer rodent model)  may possibly be due to caloric
restriction rather than endocrine-related as postulated by Kandori et
al. (2005).  Please comment on these preliminary conclusions regarding
the recent prostate findings.

Please comment on the extent to which the available data support the
proposed mode of action for prostatitis and the appropriateness of the
rodent model in the context of human health. 

Question 1.8

Based on data obtained in studies to assess the effects of atrazine on
the immune system following developmental and adult exposure, as well as
in vitro mechanistic studies, EPA concluded that atrazine has the
potential to affect the immune system (Section 3.5 of the draft issue
paper).  However, the Agency has also concluded that underlying
mechanisms of atrazine-mediated immunotoxicity and its relevance to
potential adverse health effects in humans are still not thoroughly
understood.  It is well established that products of the endocrine
system modulate immune function.  However, adult rodent studies have not
determined whether immune system effects are caused by direct effects of
atrazine and/or its metabolites or whether they are the result of
modulated endocrine hormone production, which in turn affects immune
cells.  Two published studies indicate that exposure to atrazine during
immune system development may result in altered immune function in
offspring, although no conclusions were drawn on the potential adversity
of the effects, because immunosuppression was observed in one study and
immunoenhancement was observed in the other.   Please comment on
potential explanations for the disparate findings reported by Rooney et
al. and Rowe et al.  Many immunotoxicologists consider immunomodulation,
i.e., suppression or enhancement of immune function, as a potentially
adverse alteration of homeostasis, because both have been associated
with disease states.   Thus, while the results Rooney et al. and Rowe et
al. may appear to be contradictory, together the data indicate
unintended immunomodulation at approximately the same dose in two
species.  Please comment on characterizing the effects of gestational
atrazine exposure as immunomodulation, when describing the
immunotoxicological outcome of developmental immunotoxicity studies.

Appendix B of the draft issue paper describes experiments conducted by
EPA scientists on the potential developmental immunotoxicity of
atrazine.  The results of the experiments provide conflicting results. 
Additional studies did not provide a suitable explanation of the
differences.  The Agency believes both sets of data are of high quality.
 However, in the context of hazard assessment, such differences are
difficult to interpret.  Please comment on the information contained in
Appendix B and provide suggestions for interpreting such data as part of
the atrazine re-evaluation.

The available data do not indicate that atrazine-induced immunotoxicity
is a more sensitive endpoint than the atrazine-induced effects on
neuroendocrine function, (e.g., attenuation of LH surge and estrous
cycle disruptions which form the basis for the current chronic RfD) in
female Sprague-Dawley rats.  Please comment on the degree to which these
preliminary conclusions are supported by the available data

Question 1.9

After the April 2010 SAP, the Agency will evaluate the weight of
evidence (WOE) for atrazine by integrating the experimental toxicology
data with the epidemiological studies.  As part of this work, the Agency
will consider the available data on hormonal changes and functional
outcomes that may be used as endpoints for deriving PoDs across
different durations of exposure and for different populations (including
potentially sensitive ones).  Important studies evaluating the
dose-response relationships for hormones involved in the HPA axis are
still on-going.  These studies are expected to provide high quality data
that can be used to characterize the entire dose response curve.  With
the caveat that these data are not available now, given the current
understanding of atrazine’s mode of action, please comment on what
would be appropriate endpoints to consider for use in deriving points of
departure in a possible future risk assessment.

Given that the duration of the toxicological concern is a key factor in
evaluating the frequency of drinking water monitoring, please include in
your response a consideration of the magnitude and duration of changes
in key events in the toxicity pathway that are sufficient perturb normal
function and comprise human health.   In defining the exposure window of
interest, please comment on the toxicokinetic/dynamic considerations
with respect to atrazine's effects on the HPA/HPG axis.  

Question 2.0:  Approaches to Evaluating Water Sampling Strategies &
Frequency of Monitoring

In conjunction with the toxicological review presented in the issue
paper, the Agency has also discussed methods for re-evaluating the
sampling frequency that is necessary for determining, with confidence,
concentrations of the pesticide in water that sources drinking water. 
These have included different methods for estimating pesticide
concentrations between known sampling events and examining the
performance of different sampling strategies for averaging periods of
different durations.  The Agency seeks feedback from the Panel with
regard to how the uncertainty and variability in both the monitoring
data and in the toxicity data (i.e., point of departure) can be
integrated to characterize and to interpret the potential significance
of atrazine concentrations in drinking water.  

Question 2.1

  

Given the nature of the temporal patterns of pesticide occurrence in
surface waters described in Section 5.2, including serial correlations
from day to day, periodicity in elevated concentrations within seasons
and from year to year, below quantitation data, and uncertainty in the
shape of the pesticide distributions in surface waters--what statistical
approaches should the Agency consider in determining confidence bounds
on exposure estimates from monitoring data? Please comment on how the
approach may vary depending on the duration of concern.

Question 2.2

The first two simulation methods presented in Section 5.5 are applicable
to the specific data sets they describe, although some generalities
regarding shape patterns appear to exist.  Given this information,
please comment on the strengths and weaknesses of the approaches and on
the practical merits of pursuing them or some other numerical approach
with a larger set of higher concentration systems.  Please comment on
how the methods for determining confidence bounds might apply given
these considerations.     

Question 2.3

As described in Section 5.4.2, the Agency is considering the use of a
confidence interval or prediction interval approach to characterize the
uncertainty of exposure estimates derived from monitoring data of
varying sampling frequencies.  Please comment on the strengths and
weaknesses of either placing confidence bounds on the rolling average
estimates and comparing the upper limit from monitoring against the
level of concern (LOC) or, conversely, placing confidence bounds on the
LOC.  

Question 2.4

Please comment on the relative merits of the various modeling approaches
the Agency described in Section 5.4.1 and 5.6 for interpolating
pesticide concentrations between sampling points and, in particular, on
the strengths and weaknesses of these methods as the frequency of
samples decreases. Considering the health endpoint(s) being considered
for atrazine, particularly data for the HPA axis, and the exposure time
frame needed to induce the health effect(s) which is shorter than that
used in the 2003 risk assessment, please comment on the advantages and
disadvantages of each model for evaluating the likely occurrence and
exposure via drinking water of short, moderate, and long duration
concentrations.  Please comment on the Agency’s proposed approach for
evaluating these methods, as described in Section 5.7.1.  To what extent
should the Agency consider other factors, such as the shape of the
chemograph (Section 5.5.3), weather patterns, stream flow, and/or
pesticide use patterns in evaluating the modeling approaches?

Page   PAGE  10  of   NUMPAGES  10 

