Mapping the radiobiological effectiveness of a pristine carbon beam peak.
Abstract.
Following a successful application for beam time at the INFN-LNS, we have planned an
experimental session aimed to map how radiobiological effectiveness (RBE) varies along a
monochromatic beam of carbon ions. Carbon beams of 62 MeV/u from the INFN-LNS
cyclotron (Catania) and 30 MeV from a Tandem accelerator (Naples) were used to exposed
human glioblastoma (U87 MG) and normal human fibroblasts (AG01522) cells. Samples
were irradiated in 10 different positions across a monochromatic carbon beam in a 0-5 Gy
dose range. Cell survival and chromomal aberrations were investigated. Dosimetry was
performed using a Marcus ionization chamber, Gaffchromic films and CR39 plastic detectors.
The experiment was successful with over 100 cell plates collected for the survival assay
(waiting to be scored for colony formation) and more than 10 cell culture flasks currently
incubated and followed up for chromosomal aberration. Data collected will be analyzed with
a view to formulate a RBE dependency as a function of dose and position along the ion path.
Visit to the INFN provided also the opportunity to strengthen links between our two
institutions and setting up the basis for future projects (i.e. radiobiology using laser driven
proton sources and Geant4 training schools).
Introduction.
Although energetic photons are widely used for cancer radiotherapy, ion beams represent, in
theory, the most effective type of radiation due to their physical dose deposition pattern.
Whilst X-rays are characterised by a maximum dose deposition just after entry into a
medium, the depth-dose distribution of heavy ions is initially flat (plateau region) exhibiting a
sharp maximum (Bragg peak) near the end of their range. This range can be precisely
controlled by adjusting the ion beam energy, allowing for precise targeting of deep-seated
tumours and those close to sensitive organs while at the same time sparing normal tissues. Ion
beams have been used for radiotherapy for a couple of decades with protons used in many
facilities worldwide while heavier ions (i.e. carbon) have been exploited in Japan and
Germany with impressive results. Tumour control probabilities are comparable to, or exceed,
those achieved with conventional radiotherapy and the side effects are much lower.
Supported by a progressively more established and cheaper technology, particle therapy is
currently the fastest growing cancer treatment approach.
Despite these results, there are still key uncertainties of the biological effects caused by ion
beams especially related to late effects including secondary cancer. These uncertainties will
impact on further optimization of cancer particle therapy. The main issue is related to the
change in the pattern of ionization produced as the ion slows down (i.e. loss of energy during
penetration) which causes different biological effectiveness along the ion path (Bragg curve).
Therefore, in contrast to X-rays, not only the dose but also the severity of the effect changes
with depth. For virtually all types of beam and end points of primary interest to radiation
oncologists, the relative biological effectiveness (RBE) of charged particles (relatively to Xrays) varies non-linearly with the energy deposition and is ≥1. While lethal damage leads to
cell death through necrosis or apoptosis and is related to acute tissue effects, sub-lethal
damage doesn’t prevent cell from going through the cell cycle and may ultimately result in
genetic instability, transformation, mutation and carcinogenesis. Importantly, while lethal
effects increase continuously with dose, sub-lethal effects are expected to exhibit a peak as
increasing damage (i.e. dose) prevents cells from progressing through the cell cycle. Nonlethal cellular effects are therefore most likely to occur in the plateau region where, by
definition, healthy tissues are exposed to low but not negligible doses and lethality is low.
Ultimately, normal tissue effects, including risks of secondary cancers, will determine the
overall treatment outcome.
To date, only a few studies have attempted to determine cellular responses other than cell
killing along the ion path and they have been limited to very few positions (usually centre of
the plateau and the peak) making it difficult to evaluate how the corresponding RBE varies
across the ion trajectory. This lack of knowledge results in averaged RBE values being used
in clinical treatment: 1.1 for protons and ~3 to 5 for carbon ions at the Bragg peak. These
values are generally estimated by averaging the limited in vitro and in vivo cell survival data.
Correction factors of up to 15-20% could be expected for protons and much higher values for
heavier ions. When considering the required dose accuracy in radiotherapy (3.5%) and the
fact that any uncertainty on the RBE will translate to the same uncertainty in biological
effective dose, the need for improvement is evident.
Using a multidisciplinary (physics and biology) approach, this experimental session is part of
a project aimed to investigate in detail the damage and resulting lethal and sub-lethal cellular
response caused by therapeutically relevant ion beams across and around the Bragg curve.
Combined assessment of early (cell survival) and late (chromosome aberrations) cellular
response and DNA damage (γH2AX, 53BP1 assay) in a variety of relevant cell lines (normal
fibroblasts, cancerous glioma and prostate) will provide detailed systematic information to
help developing a rigorous theory of ion radiation action at the cellular and molecular level.
Successful validation of the hypothesis will lead to the establishment of a model for
biological dose curves that could be used to design much improved radiotherapy treatments
plans.
Method.
Human glioblastoma (U87 MG) and normal human fibroblasts (AG01522) cells were shipped
to the INFN prior to the experimental session and cultured in conventional T175 flasks in a
5% CO2, 37o C incubator. 24 hrs prior to the irradiation, cells were trypsinised and reseeded
in 9 cm2 polystyrene slide flask (http://www.nuncbrand.com/en/page.aspx?ID=231) for the
high energy ion beam or custom made wells with 6 μm Mylar base for the low energy
exposures. Samples were then incubated to let cells attach to the Slide Flask or the Mylar
substrate. Slide Flasks were irradiated using an automated stage connected to an in-line beam
monitor and the ion beam shutter. This assured that the desired dose was precisely delivered
to our samples. Dose uncertainty due to shutter response time and dose rate (~1-2 Gy/min)
was of the order of ~1%. For each position along the ion path, 10 slide flasks were irradiated
at different doses (0-5 Gy) in duplicate. On the other hand, Mylar wells were mounted
directly on the beam line with the Mylar substrate functioning also as a vacuum window to
minimise energy loss. In this case, the dose delivered was monitored by two solid state
detectors placed either side of the Mylar well. Following the irradiation, cells were
trypsinised, counted and reseeded at pre-established densities in 6 multiwell plates. Cells
from each Slide Flask/Mylar well were reseeded in 6 wells at different densities. The
multiwell plates were then incubated for 10 days for colony formation. After 10 days, cells
were fixed and stained using Crystal Violet (500 mg diluted in 1 L of 70% methanol).
Colonies are currently been scored and survival fractions will be calculated based on the
original number of cells seeded.
For the chromosomal aberrations, cells were left to grow for 24, 48 or 72 hrs and harvested in
calyculin A to induce chromosome condensation. Cells will then be hybridized and mounted
on glass slides for metaphases spread which will then be analysed and scored using the
MetaSystem automated facility.
Carbon ion beams were produced by either a 62 MeV/u cyclotron or by a 3 MeV tandem
accelerator. The beam was collimated to a ~2x2 cm2 to assure >90% uniformity across its
profile. This was confirmed by measurement using both Gaffchromic films (HD-810) and
CR39 plastics. The dose rate for the cell irradiation was set between 1 and 2 Gy/min.
Accurate dosimetry was performed using a Markus ionization chamber mounted on a
micropositioning stage and submerged in a water phantom. Where necessary, PMMA
degraders of precise thickness (± 10 μm) were placed in front of the slide flasks to simulate
appropriate thickness of water and reduce the energy of the carbon beam entering the cells.
Graph showing dosimetry of the 62 MeV/u carbon beam from the INFN-LNS cyclotron.
Dose, fluence and LET are reported as a function of depth in PMMA. Blue arrows indicate
positions along the ion path at which survival experiments have been performed. For the
chromosomal aberrations, cells have been exposed only in the 1st and 8th position
corresponding to entrance and peak.
Results & Discussion.
Dosimetry measurements from the Markus chamber, the Gaffchromic films and the CR39 are
currently being analysed and compared in order to determine precisely the dose absorbed by
the cells. This is a critical step due to the rapidly increasing dose around the Bragg peak area.
Monte Carlo based calculations using Geant4 code will also be implemented with details
about the physical setup used to validate the dosimetry measurements.
All irradiation were successful with colonies present in all the reseeded multiwells (over
100). Colonies for the cell survival experiments are currently being scored. This will result in
10 full survival curves, one for each position along the ion path (i.e. at a different depth) from
which we will calculate the linear quadratic parameters (α and β). Such parameters will be
used to formulate the RBE dependency as a function of dose absorbed and position along the
ion path. Next, we are planning to estimate the RBE across a modulated spread out carbon
beam (as used for clinical applications) and compare it with experimental observations (next
beam time planned for February 2012).
Samples irradiated for the chromosomal aberrations have also been fixed and treated with
calyculin A for the premature chromosome condensations. Metaphases spread have been
prepared for a limited number of samples (remaining still to be processed) and are currently
being analysed. Frequency and type (transmissible Vs lethal) of chromosomal aberrations
will be analysed as a function of dose absorbed and survival fraction for the 2 positions along
the ion path (entrance and peak). Data are expected to provide information on the long term
risks associated with cells exposed to the entrance channel of carbon ion beams.
Full set of data/analysis is expected by mid February 2012, before the next beam time session
at the INFN-LNS.