Mechanistic and quantitative studies of bystander responses in 3-D tissues for
low-dose radiation risk estimation
Sally A. Amundson, David J. Brenner and Alexandre Mezentsev
Center for Radiological Research, Columbia University Medical Center, New York, NY.
We have previously used the MatTek EPI-200 3D human tissue model in studies of
alpha-particle bystander effect. We have now demonstrated a similar bystander effect
using low LET proton irradiation of EPI-200 tissue. In our current studies, we have
developed a masking approach to shield all of the tissue with the exception of a narrow
(25 micron) strip of directly irradiated tissue in the center of the dish. This approach has
improved the precision of alignment of the tissues over what was previously achievable
with the original microbeam technique.
We have also developed and optimized a tissue-disassociation technique to maximize
cell division in this system. This in turn has allowed us to employ the well-characterized
cytochalasin B binucleate micronucleus assay in order to improve our scoring of
micronuclei in the bystander tissue. In this approach, which is the standard for
micronucleus scoring, cytochalasin B is used to block cytokinesis after nuclear division,
resulting in binucleate cells. Scoring micronuclei only in these binucleate cells ensures
that the scored cells have divided since the genotoxic insult was applied. This has been
shown to be especially important in cases where the genotoxic insult may be only mildly
damaging to chromosomes, or where fewer than 90% of the cells are actively dividing.
In these cases, scoring in mononuclear cells (without knowing if they have divided or
not) tends to bias towards false negative results. Since the bystander effect in general
may not be highly clastogenic, and since the cells in our 3D tissue model typically have
a fairly low division rate in situ, the application of the binucleate micronucleus assay in
this model will strengthen conclusions based on this endpoint.
Using masked irradiation, we have determined the most informative time after irradiation
to score cellular responses in the bystander tissue. In this model, elevated levels of
apoptosis and binucleate micronuclei were detected in bystander tissue by 24 hours
after irradiation, and reached maximum expression at 72 hours after irradiation. In
subsequent studies, therefore, these cellular endpoints were assayed 72 hours after
irradiation. Bystander effects were observed up to at least 1 mm from the site of low
LET irradiation, consistent with results from tissue that was bystander to alpha particle
irradiation. For instance, 8% (+/- 1.4%) of control binucleate cells compared to 15% (+/-
2.4%) of bystander binucleate cells located between 250 and 1000 microns from the
tissue center (site of irradiation or mock-irradiation) had micronuclei at 72 hours post
exposure. We are also developing other endpoints that may allow us to detect earlier
manifestations of bystander effect, such as staining for phosphorylation of gamma-
H2AX.
We have optimized recovery of extremely high quality RNA from the MatTek 3D tissues.
Initial quantitative real-time PCR studies monitoring expression of genes, such as
CDKN1A and PTGS2, that respond to direct and bystander irradiation in other model
systems have indicated a rapid gene expression response in EPI-200 following both
high- and low- dose proton exposures. Analysis of global gene expression changes in
the 3D tissues following high and low doses of low LET proton irradiation is ongoing.
During the fall NSRL run at Brookhaven National Laboratories we performed our first
bystander experiments using high LET iron ions. These experiments used specialized
holders designed to allow registration of the 3D tissues directly on top of customized
pieces of track-etch plastic. The tissue-and-track-etch stacks were exposed together to
very low fluence iron ions, and the location of tracks and their predicted penumbra were
determined from the track-etch data. The number of tracks observed in the in-beam
samples agreed well with the predicted number, whereas tracks were not seen in the
out-of-beam controls. Superimposition of the tracks and predicted penumbra onto
exposed or control tissue sections will allow scoring of damage endpoints in
unirradiated bystander tissue. The out-of-beam control tissues can be used to
distinguish between true bystander effects and possible contributions of radiation
resulting from activation of air or objects in the beam line, although such "stray"
radiation was not detected during the experiment.
This research was supported by the Office of science (BER), U.S. Department of Energy,
Grant No. DE-FG02-07ER64336
» Return to the top
