|
Ray Sachs, UCB. Computational Radiation Biology
|
Skip to bystander effect DOE grant
halfway down this page.
Modeling the interrelations among radiation-induced bystander effects,
genomic instability, and cancer risk.
DOE grant. Sachs PI. 2003-2005
Abstract and specific aims
Bystander effects are considered important to radiation risk, especially
for low doses of high LET radiation, where few cells are hit directly.
The objective of the proposed research is to develop quantitative
estimates of radiation bystander effects applicable to low dose
radiation risk estimation. Mathematical formalisms and software will be
designed for determining a key bystander effect parameter, the influence
number. In addition, qualitative scenarios in the literature for
interrelations between the bystander effect and genomic instability will
be adapted to risk estimation by making them quantitative. These two
specifically focused projects will be integrated into the broader,
multi-group, cooperative modeling effort on radiation risk using an
archetype called for in the Program Announcement.
The three specific aims are the following.
Aim 1. Computational modeling of bystander signal influence number for
tissues and for multicellular in vitro systems. A key parameter for
analyses of bystander effects, especially for risk estimation, is the
average number, N, of cells close enough to a cell sending out bystander
signal that they can be strongly influenced by the signal. This
"influence number" N is related to signal range in each particular
situation. Typically, the larger N is, the smaller is the high LET dose
sufficient to elicit most of the bystander response in the multicellular
system. Our pioneering quantitative model of radiation bystander effects
has recently been generalized by several groups. By estimating influence
number N using dose-response relations, the model was applied to radon
risk estimation. At present, estimates of N from epidemiological data on
human lung cancer vary widely. In recent experiments at Columbia using
human 3-D artificial human tissue systems the range found for bystander
signals is ~1 mm, corresponding to large values of N; large values of N,
up to ~1000 in some cases, have also been found for some other
multicellular in vitro systems.
We will develop mathematical bystander effect models for systematic and
quantitative estimates of influence number N. Various in vivo and 2-D or
3-D in vitro multicellular systems will be considered. We shall refine
previous methods that were used to estimate N from localized irradiation
experiments such as microbeam experiments and from dose response
relations. In addition, N will be estimated using models for
intercellular signaling. Classic reaction-diffusion equations for the
signaling molecule(s) may be an adequate approximation in many cases.
More complex scenarios involving specifics on gap junctions or
extracellular signaling (e.g. by cytokines/growth factors) are also
needed. Various groups have recently developed, for a variety of
biological situations, a number of deterministic and stochastic
computational models for intercellular signaling. Appropriate
quantitative approaches of this kind in the literature will be adapted
to the available experimental evidence on radiation bystander effects,
which is now rapidly accumulating, and used to estimate N. We will start
with models having a minimum of adjustable parameters, then gradually
generalize.
Aim 2. Modeling bystander/instability interrelations. Qualitative
mechanistic models have been suggested by several different groups for
bystander effects inducing genomic instability and for bystander
signaling arising a number of cell generations after irradiation. These
qualitative models will be quantified and made more specific,
emphasizing chromosomal instability. In addition to being central to low
dose risk estimation, establishing bystander signal influence number N
(aim 1) can give estimates of effective target size for radiation,
relevant to the large target size sometimes found experimentally for
chromosomal instability. Our quantifications of bystander/instability
interrelations will incorporate recent results on instability in
non-irradiated cells, and on bystander signals from progeny of
irradiated cells. Emphasis will be on stochastic models of
bystander/instability interrelations. Stochastic models are
computationally more intensive than corresponding deterministic ones,
but are superior at analyzing cell-to-cell fluctuations, are usually
conceptually clearer, are sometimes equally parsimonious as regards the
number of adjustable parameters, and are particularly convenient for
estimating statistical significance during model validation.
Aim 3.
Cooperative radiation risk modeling using an archetype. The two
aims above will be quite focused: we plan to start with the simplest
applicable quantitative models, use particular data sets to fix
parameters for each particular model and test its hypotheses, initially
keep the number of adjustable parameters to a minimum, and gradually
move from highly specific models to somewhat more general ones. The
Low-Dose Radiation Risk Program Announcement also calls for more
comprehensive, coordinated risk modeling jointly by the different funded
groups, using a biological archetype, i.e. a computational model of a
biological system. The archetype choice is to be determined by
discussions among the different groups. Results from specific aims 1 and
2 will be incorporated into this multi-group, cooperative risk modeling
effort, especially as regards extrapolating results from higher to lower
doses. We anticipate that this multi-group, cooperative project will be
more integrative, i.e. deal with a broader range of effects, have more
adjustable parameters, and involve more modular subunits compared to
narrower projects such as those of specific aims 1 and 2.
Back to top of page, for signaling
grant and links to other pages on this site.
Significance of the Bystander Effect
Sachs Subcontract. DOE grant. Columbia University, DJ Brenner PI,
2001-2004
Abstract
There has been much recent interest in
responses in cells which are not themselves struck by a
radiation track, but are "bystanders" of a directly
hit cell. Bystander-mediated effects could have profound
implications for low-dose risk estimation, where many cells are
not directly hit. Work from many laboratories has covered a
broad spectrum of endpoints showing bystander effects; however,
in parallel to this experimental work, there has been little
quantitative modeling of the bystander effect which would
facilitate systematic experimental design, or permit
consideration of its implications (if any) for radiation risk
estimation, at low doses, at low LET, or at low dose rate. Our
goal is to develop tractable, testable, biophysical models
describing bystander effects to use as a bridge between A)
experimental data on bystander effects (from our own and other
labs), and B) low-dose risk estimation. These biophysical model
We have developed a preliminary
biologically-based model formalism for bystander effects. This
formalism considers the overall radiobiological response as a
sum of both "direct" damage and bystander-mediated
damage. It encompasses several more specific models and further
experimentation is needed to test and differentiate between
them. At present the formalism applies to acute doses of
high-LET radiation, and a key goal is the extension to lower
doses per hit cell, to lower LETs, and to more
protracted irradiation. Specific goals relate to
- DOSE: Use our current models to
define additional key low-dose high-LET experiments, which will
be undertaken at the Columbia microbeam facility. These
high-statistics chromosome aberration and transformation
experiments will involve microbeam and broad-beam irradiations
with 0, 1 and 2 alpha particles, the results of which are
central to refining biophysical bystander models. In addition,
we shall compare yields of complex chromosome aberrations (a
marker of high-LET type damage) in those situations where more
or less bystander-induced damage is expected, with the goal of
understanding the nature of the bystander-induced damage. The
results of these experiments will be used to refine our models
of high-LET bystander responses.
- RADIATION TYPE: Both from a mechanistic
and a public-health perspective, it is important to extend these
bystander studies to sparsely-ionizing radiation.
Mechanistically, we can then study the effect of smaller amounts
of energy in hit cells, while from a public health perspective,
in any given month about 90% of all our cells are actually
non-hit bystanders to low-LET damaged cells. We will use 9
keV/m protons in these low-LET microbeam studies. We
shall modify the current formalism so that it includes low-LET
radiation, and design and perform experiments, similarly
motivated to those in Goal 1, to specifically test mechanisms of
low-LET bystander effects. The results will in turn be used to
improve our initial models of low-LET bystander-mediated
response.
- DOSE
PROTRACTION: Our preliminary models of the bystander effect
strongly suggest the possibility of inverse dose-rate effects,
where increased protraction at a given dose could increase the
bystander effect. With both mechanistic and pragmatic
motivations, we will design and perform experiments in which a
given proportion of cells are microbeam irradiated with exact
numbers of particles, and the procedure repeated several hours
later, either on the same or a different group of cells on the
dish. The results will be compared to the corresponding acute
irradiations, and the conclusions from these studies will be fed
back into the biophysical models.
- SIGNIFICANCE: The improved biophysical
models, which we will refine using experimental results both
from our own experiments as well from other laboratories, will
be used to make an overall best assessment of the public-health
significance of bystander-mediated responses at low doses, both
at low and high LET, and at low and high dose
rates.
DOE