Transcript Carcinogenic Effects of Low Doses of Ionizing Radiations

Carcinogenic Effects of Low
Doses of Ionizing Radiations
R.Julian Preston
Environmental Carcinogenesis Division
U.S Environmental Protection Agency
Research Triangle Park, NC
AAAS Meeting, 2004
Types of Dose-response Curves
THE CHALLENGE
How to estimate human tumor
responses at low doses.
THE REQUIREMENT
The best fit tumor dose response
curve has to be best for the total range
of tumor types and for different
radiation scenarios.
THE APPROACHES
•
•
•
Extrapolate from human data at
higher doses.
Use animal tumor data
Use surrogates for cancer responsescellular and molecular indicators
What is a Low Dose?
The background radiation levels are about
1mSv (0.001Sv) not including radon.
Cancer risk levels can be computed for
this level of radiation that for the present
discussion is considered to be a low dose.
Examples of Low-Dose Exposures
•
•
•
•
•
•
•
Round-trip flight, NY to London 0.1 mSv
Background radiation (no radon) 1.0 mSv
Background radiation with radon 3.0 mSv
Single screening mammogram
3.0 mSv
Pediatric abdominal CT scan
25 mSv
International Space Station (per yr)170 mSv
Radiation Worker yearly limit
20 mSv
Cancer Risk for Ionizing Radiation
Based on human tumor data –
• A-bomb survivors
• Radiation Workers
• Medical exposures
Assumes linear-no threshold (LNT) response
for extrapolation to low, environmental exposures
Radiation Cancer Risk Estimates
Fatal cancer
Nonfatal cancer
4% per Sv
0.8% per Sv
Background frequency 25%
CANCER EPIDEMIOLOGY
(1)
A-bomb survivors
•
•
•
•
Doses less than 0.5 Sv.
7000 tumors; 50,000 survivors.
Reliable risk estimates to 0.05-0.1 Sv.
Not overestimated by linear
extrapolation from estimates computed
over higher dose ranges.
CANCER EPIDEMIOLOGY
(2)
Medically exposed groups.
• Fractionated and protracted
exposures can increase total
cancer risk.
• Clear evidence of risk not
available for low exposures –
limited precision, large errors.
CANCER EPIDEMIOLOGY
Conclusion
• Finite risk at low dose for solid tumors.
• Plausible for leukemias.
• Low-dose linearity is the best estimate
of dose responsiveness.
ANIMAL CARCINOGENICITY
• Few generalizations can be drawn.
• Variability of responses for different species,
strains, tumor types, radiation types.
• All types of dose-response curve shape observed.
• Radiation-induced life-shortening generally
linear with dose.
• Low dose linearity cannot be ruled out for the
great majority of studies; similar generalization
cannot be made for other curve shapes.
ANIMAL CARCINOGENICITY
Conclusion
Overall LNT is the best estimate of
dose-response for radiation- induced
tumors in animal models.
CELLULAR AND MOLECULAR BIOLOGY
STUDIES
• Cancer is a multi-step process.
• Cancer results from mutations or chromosomal
alterations in oncogenes and tumor suppressor
genes.
• Dose-response information for radiationinduced mutations and chromosome
alterations is informative to tumor dose
response.
GENE MUTATIONS
•
Radiation-induced mutations increase
as a linear, quadratic or linear-quadratic
function of dose.
CHROMOSOMAL ALTERATIONS
• 99.9% of all tumors contain chromosomal
alterations
• Radiation-induced chromosomal alterations
can predict tumor response at low doses, in a
qualitative sense.
• Two DNA lesions required.
• Produced by one or two tracks (low LET).
Produced by one track (high LET).
• Significant increases over background at
100 mGy.
• For low LET, dose-response curves are
linear or linear quadratic.
Cellular Responses to Ionizing Radiation
– Potential Modifiers of Low Dose
Cancer Risk
Bystander effects
Genomic instability
Adaptive responses
Note: Current cancer risk estimates for radiation are
based on human tumor frequencies and so
incorporate any influence of these responses
Conclusions
• Bystander responses and induced genomic
instability could lead to enhanced tumor
responses at low doses.
• Adaptive responses could lead to a
reduction in tumor responses at low doses
• Underlying mechanisms need to be better
understood.
SUMMARY STATEMENT
Based upon information for radiation-induced
human tumors, animal tumors and surrogates
for tumors, LNT is frequently the best fit to the
low dose data and generally cannot be ruled
out.
LNT is the best solution – To-day.
However, more needs to be done and is being
done. The arguments presented are for To-day.
Bystander Effects
Mutation
Mutation
Mutation
Bystander Effects
Mediated by cell signalling events either cell
cell communication or possibly through the
tissue culture medium.
Very little evidence for in vivo responses.
Difficult to demonstrate for chemicals- except
by single cell injection or using medium
transfer experiments
Bystander Effects
Impact on risk assessment ?
Increases target cell population for a given
exposure – how can dose be defined in
biological terms?
Genomic Instability
• Tumor cells generally show extensive genomic instability
(structural and numerical).
• Early stage event – selection for specific phenotypes drives
the cancer process.
• Following radiation exposures (in vivo and in vitro)
chromosomal alterations and mutations can be observed
many cell divisions later – delayed response. Described as
genomic instability.
• Need to distinguish limited response in experimental
systems with extensive response in tumor cells. Are these
the same event?
Genomic Instability
• Induced instability results in higher mutagenic
response than predicted on the basis of initial
cellular response.
• Important for dose-response considerations- what
is the response per unit dose?
• Could lead to supralinear response at low doses
for tumor outcomes.
• Very limited evidence for genomic instability
following chemical exposures.
Adaptive Response
• Small priming dose can reduce response to
larger challenge dose.
• Demonstrated for chemical and radiation
exposures for a range of endpoints, e.g.
mutation, chromosome alterations.
• Reduces responses in most cases by a factor
of about 2.
ADAPTIVE RESPONSE
• Does an adaptive response lead to a
threshold for tumor induction?
• Adaption only reduces the response –
does not eliminate it.
• No evidence for adaptive response at
very low dose rates.
ADAPTIVE RESPONSE
Conclusion
No version of an adaptive response
necessitates departure from LNT (based
on cellular studies); a change in slope can
be predicted.
Interactions of Responses
Adaptive responses and genomic instability
have been reported to be induced in bystander
cells, further complicating dose-response
assessment.