Revision of the Cramer
Decision Tree
T. Cachet
EFSA / WHO Stakeholder Meeting
2nd December, 2014
The Cramer/Ford/Hall Decision Tree
Published in 1978
 Food Cosmet. Toxicol. (1978) 16, 255-276
Screening for toxicity testing
Classified substances :




Structure (key)
Metabolism
Toxicology
Natural occurrence (in body and/or food)
 Validated against chemicals with data on biological and
toxicological properties (pesticides, drugs, food
additives, industrial chemicals, flavorings, fragrances)
2
The Original Cramer/Ford/Hall Decision Tree
Heterocyclics
Open chain
Alicyclics
Heteroaromatics
Aromatics
3
Class Structure
Class I
Structures and related data suggest a low order of
oral toxicity. If combined with low human exposure,
require a low priority for investigation
Class II
Less clearly innocuous than Class I, but no firm
indication of toxicity or the lack thereof
Class III
Structure and related data permit no initial
presumptions of safety, or may suggest significant
toxicity. These substances deserve the highest
priority for investigation
4
Relation of DT Class to Threshold of Toxicological Concern
(TTC) (Munro, 1996)
5
Structural Class
(examples)
TTC
(µg/person/day)
I
(ethyl butyrate, cinnamaldehyde)
1800
II
(3,6-dimethylpyrazine, pulegone)
544
III
(estragole, anethole)
90
Role of the Decision Tree within Flavoring Safety Assessments
• Close to 3000 flavoring substances have been evaluated in
procedures that rely on the Cramer/Ford/Hall DT as critical
component to guide the safety assessment
– JECFA
– EFSA
• US-based Flavor and Extract Manufacturers Association (FEMA)
Expert Panel has utilized the Cramer/Ford/Hall DT as critical part of
safety assessment process for the constituent-based assessment of
natural complex substances used as flavoring ingredients since
publication of a paradigm in 2005
• This has given us the opportunity to better understand the strength
and also some of the weaknesses of the Cramer decision tree
6
DT: Reconciling New Knowledge
Pulegone
DT Class II
Carvone
DT Class II
Anethole
DT Class III
7
Coumarin
DT Class III
Estragole
DT Class III
Dihydrocoumarin
DT Class III
Flavor Industry Work on Revising Cramer/Ford/Hall DT
• In a co-funded project, IOFI and FEMA have developed
suggestions for strengthening the Cramer/Ford/Hall DT
“Package”
• Some proposed reworking of the DT questions/routing
• Updating of the database that supports the application
of the TTC concept to the DT
– Expanded Munro database
– Opportunity to further strengthen and even expand
current TTC classes
8
Proposed Revisions



9
Revise “trunk” of the tree steps
Steps lacking biochemical basis eliminated
 Biological normality Step 1
 Common component of food Step 22
 Remove need for ‘lookup’ lists
Increase elements
 Other than C, H, O, N, divalent S, to include
higher oxidation state S, Cl, F, and P in a
biologically stable oxidation state
Factors taken into consideration during revision
•
•
•
•
•
•
10
Skeletal structure
Functional group
Presence or absence of other functional groups
Extent of conjugation
Impact of electron donating groups
Positional & geometric isomers
Consideration of Biochemistry
• Low Concern Substances: predominantly undergo detoxication or
complete metabolism via high capacity enzyme-catalyzed pathways
– oxidation in the TCA cycle, fatty acid pathway, cytosolic carbonyl
reduction, etc.
• Medium Concern Substances: excellent substrates for low capacity
detoxication pathways leading to stable, excretable metabolites
myrcene
CYP450 oxidation/hydroxylation
or
(FEMA 2762)
epoxidation/hydrolysis
• High Concern Substances: weak substrates for low capacity
detoxication enzyme pathways leading to stable, excretable
metabolites or metabolites of unknown stability
11
Even Higher Biochemical Concern
Could Argue for Additional Structural Classes
• “weak” substrates for low capacity intoxication pathways
leading to reactive metabolites
O
O
• “excellent” substrates for low capacity intoxication
pathways leading to reactive metabolites
MeO
MeO
OH
MeO
12
MeO
Consideration of Chemistry
• Low Concern Substances/metabolites w/functional
moieties
– Substance/metabolites unlikely to react with biomolecules in vivo
leading to adverse response
O
O
OH
N
H
OH
O
• Medium Concern/metabolites
– Interact with biomolecules → products unlikely to result in
adverse biological responseO
HO
13
Proposed Revisions would lead to Revised Structure
1
As of 3/2014
Note
that Structure Class
assignments to question
outcome are based on
current tree at this stage
3
Alicyclics
17
Class 2
Class 3 18
(A-I)
Cl.3
18
(J-M)
14
Cl.2
Class 3
8
16
19
Cl.3
Cl.1
26
24
27
Class 3
Cl.1 Class 2 Class 1
Class 3
28
25
18 Class 3
(A-I)
18
(J-M)
Class 1
Heterocyclic
29
Cl.1
Cl.2
29
Cl.3
Cl.3
Cl.1
10
17
29
18 Cl.3
(A-I)
Cl.1 Cl.3
23
Class 3
22
5
9A
20
29
Class 3
6
Open chain
15
21
Class 1
4
7
Aromatics
Treat hydrolysis
products separately
2
18
(J-M)
Lactone12
11 Class 3
29
13 Class 2
Class 3
14 Class 3
29
Class 2 Class 1
Class 1
Class 3
Class 3
Heteroaromatic
2
9B
Class 3
9C
15
9D
Proposed Revisions: Work Continues!
• Relation of new question outcomes to Structure
Class Assignments?
15
Improving and Refining the Munro Database
Chemical Space:
Foundation: Munro Database (~ 600 chemically defined
compounds)
Additions (~1400 chemically defined compounds):
1. Flavoring Substances
2. Compounds with published studies
National Toxicology Program (NTP) Studies (USA)
Environmental Protection Agency (EPA) Studies (USA)
Scientific Literature
NOEL/NOAEL Values determined for each study are extracted.
Munro, I.C. et al. (1996) “Correlation of Structural Class with No-Observed-Effect Levels: A
Proposal for Establishing a Threshold of Concern”, Food and Chemical Toxicology.”, 34, 829 867.
16
Improving and Refining the Database
Studies:
Foundation: Munro Database
• Exclusively oral (dietary, gavage) studies
• Mainly chronic studies (short term and acute studies not
included)
Additions:
• Chronic and short term oral studies added
(Some studies suggest the validity of using a conversion factor to
equate NOAEL values between studies of differing lengths.)
17
Conclusions
• Overall, we trust that both the review of
the Cramer Decision Tree and
strengthening the Munro Database will
increase the robustness of the TTC
concept and we are
working/planning towards sharing our
data with the scientific community
including the WHO/EFSA review project
on TTC
18