Slide 1 of 38

1
Drs. Vicki Dellarco Anna LowitHealth Effects
DivisionOffice of Pesticide Programs

• Mode of Action/ Human Relevance Analysis For
  Incorporating Mechanistic Data in Human Health
  Risk

May 2006
2
Uses of Mechanistic Data in Risk Assessment

• Identify
• Key biological (precursor) events leading to
  adverse toxicities (Mode of Action)
• Inform
• Human relevance of animal findings
• Dose response extrapolation
• Life stage susceptibilities
• Understand
• Common pathways of toxicity (cumulative risk
  assessment)
• Promote
• Consistent harmonized approach to risk assessment
  for all health endpoints

3
Mechanism of action (more detailed
understanding at biochemical molecular level)
versus Mode of action (identification of
key obligatory steps)
Exposure
Key event
Key event
Key event
4
Mode of Action Framework

• EPAs Guidelines for Carcinogen Risk Assessment
• 1996 Proposed Revisions
• put forth the notion of understanding mode of
  action versus mechanism of action
• 1999 Interim Guidance
• introduced mode of action framework
• 2005 Final Guidance
• minor rewording of MOA framework

5
Mode of Action Framework

• Postulated mode of action
• Identify sequence of key events on the path to
  cancer
• Experimental support
• Concordance of dose-response for key events with
  that for tumors
• Temporal relationships for key events tumors
• Biological plausibility Coherence
• Strength, consistency specificity
• Other modes of action
• Identify uncertainties
• Conclusion

6
Human Relevance Framework

• Risk Sciences Institute-ILSI
• Comparability or concordance analysis of the key
  events relevant biology between the laboratory
  species humans
• Tumor Responses Meek et al., 2003, Critical
  Reviews in Toxicology Vol 33/Issue 6, 581-653
• Reproductive, Developmental, Neurtoxocity
  Responses Seed et al., 2005 Critical Reviews in
  Toxicology Vol 35/Issue 8-9, 63-781
• extended human relevance analysis to include
  mutagenic carcinogens noncancer end points
• WHO/IPCS Human Relevance Framework (in
  preparation)

7
Human Relevance Framework

• Based on three analyses
• Is the Weight of Evidence sufficient to establish
  a MOA in animals (MOA Framework)?
• Are the key events in the animal MOA plausible in
  humans?
• Taking into account kinetic/dynamic factors, is
  the animal MOA plausible in humans?

8
Continue with the Dose Response Exposure
Assessment
No,
9
Assessing an Animal Mode of Action

• General Points
• Applicable to all chemicals, to all endpoints,
  and to all modes of action
• Evaluation of MOA for tumors or (other adverse
  effects) in different organs
• MOA in different organs may or may not be the
  same
• Site concordance between animals humans

10
Assessing an Animal Mode of Action

• General Points
• When a substance operates via a novel MOA, the
  analysis is focused on the chemical entails a
  detailed evaluation via the MOA Framework
• When a substance produces an adverse effect
  consistent with an already established peer
  reviewed MOA through which other chemicals have
  been shown to operate, the analysis is focused on
  the established MOA a determination of whether
  the substance operates via the same key events
  established for the pathway

11
Assessing an Animal Mode of Action for Human
Relevance

• General Points
• Concordance Analysis of key events is for the MOA
  is not necessarily a chemical specific
  evaluation.
• Chemical specific generic information relevant
  to the toxicity process can be valuable

12
MOA Inform Human Relevance
EPAs Cancer Assessment Review Committee (CARC)
classified atrazine as not likely to be
carcinogenic to humans.
Vinclozolin--Since the androgen receptor is
widely conserved across species lines,
antiandrogenic effects would be expected in
humans.
13
MOA Inform Dose Response Extrapolation
Alachlor - . . . a margin of exposure (MOE)
approach (indicative of a non-linear dose
response) should be used for the risk
assessment.
Chloroform . . . a nonlinear approach is more
appropriate for low-dose extrapolation.
Cacodylic Acid . . . nonlinear default
approach (i.e., derivation of a reference dose or
margin of exposure) is regarded as the more
appropriate dose response extrapolation approach.
. .
14
Case StudyCacodylic Acid (Dimethylarsinic acid)
15
DMAV Mode of Action

• Science Issue Paper Mode of Carcinogenic Action
  for Cacodylic Acid (Dimethylarsinic Acid, DMAV)
  and Recommendations for Dose Response
  Extrapolation (July 26, 2005)
• http//www.epa.gov/oppsrrd1/reregistration/cacodyl
  ic_acid/
• Revised issue paper will be publicly available
  this spring.
• EPAs Science Advisory Board (SAB) reviewed the
  special issue paper in September, 2005
• Draft SAB report December 27, 2005

16
Metabolism of Arsenic
Pesticide Chemical
Alternate steps of oxidative methylation
reduction
Methylation Reduction
TMAsV
TMAsIII
17
DMAV MODE OF ACTION ANALYSIS
Weight of Evidence

• Extensive experimental cellular and laboratory
  animal data

18
DMAV Available Cancer Data

• No epidemiology data
• Standard rodent bioassay
• Bladder carcinogen in rats
• via feed -100 ppm (9.4 mg/kg bw per day)
• via drinking water- 50 200 ppm
• females more sensitive than males
• Not carcinogenic in mice
• Up to 500 ppm in B6C3F (Gurr et al., 1989)
• 121 ppm in C57 XC3H/Anf or AKR (NCI 1969)

19
Mode of Action Measurable Key Events in Target
Tissue
DMAIII Metabolite
Urothelial Toxicity
Urinary bladder from a female F344 rat treated
with 100 ppm DMAV
Sustained
Regenerative Proliferation
BrdU Labeling
Hyperplasia
Urinary Bladder Tumors
20
Compensatory regeneration in rat bladder at weeks
8 10 following ingestion of DMAV
21
Association of Key Precursor Events Bladder
Tumors in F344 Rats
Temporal
Dose (mg/kg bw/day) Metabolism DMAV?DMAIII Urothelial Toxicity Regenerative Proliferation Urothelial Hyperplasia Transitional Cell Carcinoma
0.2 (2 ppm) (wk 3-0.03 0.01 uM) (wk 10-6/10, grade 3 or 4) - - -
1 (10 ppm) (wk 3-0.12 0.02 uM) (wk 3-2/7, grade 3) (wk- 10 8/10, grade 3 or 4) slight (wk 10-1.5X inc) - -
4 (40 ppm) (wk 3-0.28 0.09 uM) (wk 3-7/7, grade 3) (wk 10-5/10, grade 3 or 4) (wk 10-4.3X inc) (wk 10- 4/10) -
9.4 (100 ppm) (wk 3-0.55 0.15 uM) (6 hrs-6/7, grade 3) (24 hrs-4/7, grade 3 or 4) (wk 2 6/10, grade 5)(wk 10-0/10, grade 4 or 5) (wk 1- 2.2X inc) (wk 2-3.9X inc) (wk 10-4.2X inc) (wk 8-7/10) (wk 10-9/10) (papilloma first obs at wk 107 carcinoma first obs at wk 87)
Dose Response Concordance
22
Cacodylic Acid Key Events Temporal Relationship
DMAIII ? DMAV
Urinary bladder from a female F344 treated with
100 ppm DMAV
Urothelial Cytotoxicity Regenerative
Proliferation Hyperplasia Tumors
6 hours 1 Week 8-10 weeks 104 weeks
BrdU labeling
Urinary bladder tumors
23
Cacodylic Acid Key EventsCytotoxicity/Regenerati
ve Proliferation

• Strength, Consistency Specificity
• Consistency of association found in repeated
  experiments within a lab among different labs
• Inhibition of DMAV ?DMAIII reduced cytotoxicity
• Cessation of exposure to DMAV results in recovery
  of tissue (i.e., hyperplasia)
• Biological Plausibility Coherence
• Regenerative proliferation associated with
  persistent toxicity appears to be a risk factor
  for bladder cancer in humans

24
Characterization of Cacodylic Acids Genotoxicity

• Neither DMAV or DMAIII are direct acting
  point/gene mutagens
• Both are clastogenic but DMAIII is the more
  potent
• In vitro data only
• DNA damage appears to result from an indirect
  mechanism (ROS/oxidative damage)
• DMAIII ? DMAV

25
Chromosomal Aberrations

• For the oxidative DNA damage to be relevant to
  the carcinogenic process (i.e., clonally
  expanded), stable chromosomal mutations must be
  formed
• Formation of chromosomal mutations requires DNA
  replication because chromosomal alterations are
  produced by errors of replication on a damaged
  DNA template.
• frequency of chromosomal mutations will be a
  function of the regenerative proliferative
  response.
• All these events--genetic errors, cytotoxicity,
  stimulation in cell proliferaiton -- must occur
  to result in bladder tumors.

26
Other Modes of Action or Key Events

• No other MOA with sufficient scientific support
• Direct DNA reactivity
• Formation of solids
• Changes in urinary chemistry physiology

27
Mode of Action Conclusions

• Sequence of key events leading to bladder tumors
  measurable supported by robust data
• Biologically plausible
• Uncertainties do not discount scientific support
• cellular target for cytotoxicity not understood
• unknown cytotoxic metabolites found in urine
  (after drinking water exposure)

28
Human Relevance of DMAVs Mode of Action
Metabolism to DMAIII
Urothelial Cytotoxicity
?
Regenerative Proliferation
Hyperplasia Bladder tumors
29
Concordance Analysis of Key Events in Rats
Humans Qualitative Quantitative Plausibility
Key Event Rats Humans
Presence of DMAIII in urine Yes Yes (based on Asi)
Persistent cytototoxicity Yes Possible
Persistent regenerative prolif/hyperplasia Yes Possible
Bladder Tumors Yes Possible
30
DMAV Mode of Action (MOA)

• The SAB concurred with EPAs conclusions
• Rat data developed for DMAV most appropriate data
  for quantifying cancer risk
• MOA for the development of bladder tumors in rats
  established
• The rat MOA is expected to be plausible in humans
• The MOA supports nonlinear extrapolation of
  cancer risk to DMAV

31
Dose response extrapolation approach

• Dose response extrapolation should be based on
  considerations of MOA which supports nonlinearity
• Must be sufficient DMAIII to produce cell
  killing sufficient cell killing to lead to
  regenerative proliferation
• Cytotoxicity enhanced proliferation need to be
  sustained
• Frequency of chromosomal mutations dependent on
  enhanced proliferation on generation of ROS
  (DMAIII ?DMAV)
• Point of Departure based on cell proliferation
  should be protective of DMAs carcinogenic
  promoting effects

32
Dose Response Considerations

• Cancer Guidelines describe a two-step
  dose-response process which separates
• Modeling the observable range of data
• Extrapolation to lower doses
• Nonlinear extrapolation
• Preferred approaches
• PBPK Model--internal dosimetry at the target
  tissue
• e.g. DMAIII
• BBDR Modelpredict biological effect
• e.g., two stage clonal growth
• Interim approach
• Identify a point of departure (POD) based on
  benchmark dose modeling
• Apply uncertainty and safety factors

Key event
POD
Response
Dose
33
Quantitative Dose-response Assessment
MOA Established?
No
Yes

• 1. Fit data in observable range
• 2. Linear extrapolation from POD

BBDR model?
Yes
Use model
No
Yes, nonlinear
MOA informs low-dose extrapolation?
No
RfD/RfC or MOE
Yes, linear (including mutagenic MOA)
34
Benchmark Dose Modeling Regenerative
Proliferation
Hill Model with 0.95 Confidence Level
Mean Response

BMD10
BMDL10
1
2
3
4
5
6
7
dose
35
Cacodylic Acid Summary of benchmark dose
estimates and lower 95 confidence limits for
cytotoxicity, BrdU labeling index, hyperplasia
and tumor data. (Doses in mg/kg/day)
Biological Event Duration Feeding Feeding Feeding Feeding Duration Drinking water Drinking water Drinking water Drinking water
Biological Event Duration 10 10 1 1 Duration 10 10 1 1
Biological Event Duration BMD (mg/kg/day) BMDL (mg/kg/day) BMD (mg/kg/day) BMDL (mg/kg/day) Duration BMD (mg/kg/day) BMDL (mg/kg/day) BMD (mg/kg/day) BMDL (mg/kg/day)
Tumor 104 weeks 7.74 5.96 6.80 2.22 104 weeks 1.92 1.21 0.88 0.14
Hyperplasia 10 weeks 1.36 1.04 0.42 0.32 104 weeks 1.63 1.04 0.74 0.14
Hyperplasia 104 weeks 1.97 1.61 0.93 0.66 104 weeks 1.63 1.04 0.74 0.14
BrdU labeling (proliferation) 10 weeks 0.65 0.29 0.54 0.07 Not determined. Available data not suitable for modeling. Not determined. Available data not suitable for modeling. Not determined. Available data not suitable for modeling. Not determined. Available data not suitable for modeling. Not determined. Available data not suitable for modeling.
Cytotoxicity 3 weeks 0.68 0.18 0.31 0.02 No reliable dose-response data available No reliable dose-response data available No reliable dose-response data available No reliable dose-response data available No reliable dose-response data available
Cytotoxicity 10 weeks 0.02 0.008 0.002 0.0007 No reliable dose-response data available No reliable dose-response data available No reliable dose-response data available No reliable dose-response data available No reliable dose-response data available
36
PBPK Model Application to DMA Risk Assessment
Q What human exposure to DMAV is required to
produce the same target tissue dose of DMA to
bladder that results in tumors in rats exposed to
DMAV?

• Use PBPK model to estimate the environmental
  exposure to DMAV required to achieve the same
  target tissue dose to bladder.

oral exposure
DMAV
metabolism
elimination

• Estimate target tissue dose using various dose
  metrics (e.g., DMAV, DMAIII or TMAO concentration
  in urine or bladder tissue) associated with
  bladder tumor development using DMAV PBPK model.
  
• Current status Agency developing mouse model
  first then scale to rats and humans

37
Summary

• Human Relevance Framework
• Identify key events
• Assist in dose response assessment
• Assist in rodent to human extrapolation
• Promote harmonization of risk assessment for all
  endpoints