Toxicology data on susceptible populations and its implications for risk assessment

Lauren ZeiseOffice of Environmental Health Hazard Assessment California Environmental Protection Agency

Presentation to The Food Advisory CommitteeUS Food and Drug AdministrationCenter for Food Safety and Applied NutritionDecember 16, 2014 Silver Springs, Maryland

The views in this presentation are those of the author and do not necessarily reflect those of the California Environmental Protection Agency or the its Office of Environmental Health Hazard Assessment

2

Outline

• Human variability, susceptibility and population risk

• Toxicology data-based approaches to address susceptibility generically

• Case examples of approach to susceptible populations in regulatory advice

• Emerging data stream example

Human variability and individual and population risk

3

Sources of differences in response among people

4

Food/nutrition

Psycho-social

stressors

Coexposu

re

Existing Health

Conditions

Heredity

Gender, Lifestage

Source-to-outcome continuum and variability

5

Source-to-Outcome Continuum

Source/media concentrations

Internal concentrations

Biological response measurements

Physiological/health status

External doses

Exposure

Toxicokinetics

Toxicodynamics

Systems dynamics

Types of Biological Variability

Co-exposures

Food/Nutrition

Gender, Lifestage

Heredity (genetic & epigenetic)

Existing health

conditions

Psychosocial stressors

Modifying source-to-outcome parameters

Modifying baseline conditions.

Outcome latency, likelihood, and

severity

Baseline biomarker values

Background and co-exposure doses

Endogenous concentrations

Exposure parameters

Pharmacokinetic parameters

Pharmacodynamicparameters

Systemsparameters

Susceptibility Indicators

Zeise et al. 2014

Context dependent, low dose effects

6Kim Boekelheide slide, NRC Emerging Sciences Workshop, June 2012

Decreased AR activity at target tissue

Interference with androgen mediated development

Reproductive tract malformationsAGD

Nipple Retention

Hypospadias

Sperm quality

Leydig cell tumors Cryptorchism Other reproductive

tract malformations

Other Decreased Decreased Blockade of Androgen Mutated Stressors Testosterone Dihydrotestosterone Receptor (AR) Receptor

Varied related outcomes – upstream endpoint

“Phthalate Syndrome”

8

Variability: Implications for Risk

• Population

• Subpopulations

9

Distinct Susceptible Groups

10

Assay limitations to addressvariabilty and susceptibility

Data, models, theories, concepts

Source-to-Outcome Continuum

Source/media concentrations

Internal concentrations

Biological response measurements

Physiological/health status

External doses

Exposure

Toxicokinetics

Toxicodynamics

Systems dynamics

Epidemiology

Animalbioassays

PBPKmodels

In vitroassays,Toxicitypathways

AdverseOutcome Networks

In vitro systems are typically genetically homogeneous.

Traditional epidemiology:• Limited power to examine

susceptibility factors.• Generalizing from

occupational cohorts to the overall population.

Use genetically homogeneous experimental animals in uniform environments.

Few examples and data to support population PBPK

modeling

Currently do not focus on variability and susceptibility

Addressing susceptibility generically

11

Examples of data and analyses

Early-life susceptibility to carcinogens

• Human examples• Diethylstilbestrol (DES)

in utero • Radioactive iodine early

childhood• Immunosuppressive

agents during childhood• X-irradiation during

adolescence

• Animals• Many examples • Systematic study of

literature (e.g., EPA,

Age windows evaluated

13

Prenatal

Postnatal

Adult

Juvenile

conception birth day 22 day 49//

23 Chemicals to Evaluate Cancer Age-Susceptibility • Benzidine

• Benzo[a]pyrene• Butylnitrosourea• DDT• Dibutylnitrosamine• Diethylnitrosamine• Diethylstilbesterol (DES)• 7,12-

Dimethylbenz[a]anthracene (DMBA)

• 1,2-Dimethylhydrazine• Dimethylnitrosamine• Di-n-propylnitrosamine • 1-Ethylnitrosobiuret• Ethylnitrosourea• 2-Hydroxypropylnitrosamine• 3-Hydroxyxanthine• 3-Methylcholanthrene • 4-(Methylnitrosamino)-1-(3-

pyridyl)-1-butanone (NNK)• Methylnitrosourea• -Propiolactone• Safrole• Tetrachlorodibenzodioxin

(TCDD)• Urethane• Vinyl chloride

Carcinogenic activity ratio – Early Postnatal : Adult

Early Postnatal

16

Early-life increased sensitivity to cancer

LifestageFold Increase more than adult

MeanPercentile

50th 70th 95th

Prenatal 21 3 10 115

Postnatal 79 13 28 350

Juvenile 7 5 7 20

Source: OEHHA, 2009

17

Variable physiology and pharmacokinetics

Data sources • Phase I and II metabolism

and renal excretion with age(Dorne 2010; Ginsberg et al. 2002, 2004; Hattis et al. 2003)

• Bodymass, surface area, BMI, health status in older adults(Thompson et al. 2009)

• Metabolic enzyme polymorphisms(Ginsberg et al. 2009a,b; Bois et al. 1995, 1996; Walker et al. 2009; Neafsey et al. 2009a,b)

Respiratory Tract Tissue

Gas Exchange

Respiratory Tract Lumen (Inhalation)

Respiratory Tract Lumen (Exhalation)

Venous Blood

Rapidly Perfused

Slowly Perfused

Fat

Gut

Liver

Kidney

Oxidation & Conjugation

Oxidation

(Dead space)

Stomach

Duodenum

Oral

IV

IA

PV

Inhaled air Exhaled air

18

In vitro: Exploring genetic variability in human cell lines• Originally studied to:

• Quantify variability in drug response

• Drug class signaturesof cytotoxicity

• Prioritize drugs for investigation

• Utility in toxicology:• Obtain measures of

genetically derived variability

• Explore mode of action hypotheses• Identify determinants of genetic susceptibility

19

Variable activation in 240 lymphoblastoid cell lines

O’Shea et al. (2011), Toxicol Sci 119:398-407Individuals

6

2

4

Activ

ity

20TD Variability Factor

Nu

mb

er

of C

om

po

und

s

1 101

102

103

0

10

20

30

40

50

Example: Toxicodynamic variability estimated using population of human cells (Abdo et al. 2014)

Distribution of inter-individual TD

variability after correction for

technical variability

Repeat with 179

compounds

Cytotoxicity across 1086 human cell

lines

Slide source: Weihsueh Chiu

Differential Susceptibility with Hazard Trait

Genotoxicity

Traits named in NRC 2014, p. 35

Oncogenesis

Cardiovascular

Developmental

Neurologic and sensory

Hepatic

Renal

Gastrointestinal

Endocrine

Metabolic Disease

Respiratory

Reproductive

Hematopoietic

Immunologic

Musculoskeletal

Dermal

Other

California’s consideration of sensitivityEvidence based adoption of defaults

• Infant and child specific exposure intakes- food, water, breast feeding, home grown food consumption, …

• Revised default intraspecies adjustment, absent data:

- Pharmacokinetics 10, if no model and no specific data

- Pharmacodynamics 3, if no additional child susceptibility 10, otherwise: (e.g., asthma exacerbation or neurotoxicity)

22

Respiratory Tract Tissue

Gas Exchange

Respiratory Tract Lumen (Inhalation)

Respiratory Tract Lumen (Exhalation)

Venous Blood

Rapidly Perfused

Slowly Perfused

Fat

Gut

Liver

Kidney

Oxidation & Conjugation

Oxidation

(Dead space)

Stomach

Duodenum

Oral

IV

IA

PV

Inhaled air Exhaled air

Addressing Specific Susceptible Populations

23

Case examples

Case examples for populations with identified susceptibility to toxicant

• Age• Perchlorate interference of I- uptake in bottle fed infants (CalEPA

OEHHA)• Cadmium and renal toxicity in older adults (CalEPA OEHHA 2006)• Developmental and reproductive toxicity – many examples

• Health status• Effects of ozone on asthma sufferers (EPA 2006)• Effects of copper on Wilson’s disease heterozygotes (NRC 2006)

• Co-exposure and health status• Cancer risks to smokers for radon (EPA 2006) and arsenic (CA

2009)

24

25

Genetic susceptibility example: Wilson’s heterozygotes and copper toxicity

• Wilson’s disease• Autosomal recessive disorder• Defective biliary excretion of copper• Copper ccumulation in brain and liver leads to chronic

cirrhosis and neuropsychiatric disorder• Frequency: at least 1 in 40,000 births

• Wilson’s heterozygotes• Abnormal biliary excretion observed in 50% of

heterozygotes• Cases of liver toxicity also observed• Frequency: at least 1% of general population

26

Urinary copper in 206 Wilson’s disease siblings

24-Hour Urine Copper (μg/day)

No.

of S

ubje

cts

Distribution in normal adults

Wils

on’s

Dis

ease

Adapted from NRC (2000)

27

NAS Committee on Copper in Drinking Water

Concern for infants with altered copper metabolism• High drinking water consumption compared to adults

Bottle fed infant with several fold higher consumption• Factor of 3 higher levels of liver copper compared to

adults• Liver toxicity from in Wilson’s heterozygotes and others

genetic polymorphisms that affect copper elimination• Concern supported by experimental animal strains

sensitive to copper due to genetic alterations Models provide insights on effects and mechanisms

Perchlorate Mode of Action

28

Pituitary

THYROIDGLAND

T4 and T3

TSHPerchlorate

Iodide

Perchlorate

Considerations

• Thyroid hormone-dependent brain and neurodevelopment (Haddow et al., 1999; Klein et al., 2001; Kooistra et al., 2006; Pop et al., 2003; Pop et al., 1999; Vermiglio et al., 2004)

• Low stores of thyroid hormone (van den Hove, 1999)

• Low iodine intakes (Pearce et al., 2007: 47% women with breast milk iodine levels below Institute of Medicine recommendations for infant intake)

30

Maternal influences on infant susceptibility

• About 2.5% of pregnant women in U.S. suffering from subclinical hypothyroidism

• About 15% women had low iodide excretion in NHANES III

• Mother is infant’s major source of iodine if she is breast-feeding

• Perchlorate excreted in milk if mother is consuming it in food and water

• If mother is a smoker, she delivers less iodine in milk

California Advisory Level for Perchlorate

• Upstream adverse event• Interference of I- uptake

• Identifiable subgroup at risk• Infants • Pregnant women and fetus via mom’s exposure

• Exposure estimate to ensure adequate margin of safety

• 95th %tile water intake : body weight ratio Addresses bottle fed infant

• Other exposures to perchlorate• Food• Infant formula

Emerging data stream example

32

33

Experimental in vivo data: genetically diverse animal models

• Large panel of new inbred mouse strains

• Combines genomes of 8 genetically diverse founder strains

• Population structure randomizes existing genetic variation

• Captures ≈ 90% known variation in lab mice

Example: Collaborative Cross

Image by D. Threadgill

34

Proof of concept studies: Genetically diverse “Mouse model of human population” • Acetaminophen liver

injury (Harrill 2009)o panel of 36 inbred mouse

strainso whole-genome

association analysis o polymorphisms in Ly86,

Cd44, Cd59a, and Capn8 correlated strongly with liver injury

o orthologous human gene, CD44, associated with susceptibility in two independent cohorts.

• Trichloroethylene liver toxicity (Bradford 2011)o interindividual

differences in 14 inbred stains - TCE metabolism and molecular signaling in the liver.