Radiobiology experiments with a laser driven x-ray source: exploring the UHDR regime

. Laser Driven Plasma Accelerators (LDPA) can generate both characteristic and Bremsstrahlung x-rays in ultrashort bunches. Because of the high intensity of the laser pulses, these laser-induced x-ray sources can achieve peak dose rates compatible with the FLASH regime. At the Laser Laboratory for Acceleration and Applications (L2A2 / USC), a setup for the irradiation of in vitro cell cultures with a laser-driven x-ray source has been developed. The suitability of this source for studying the influence of Ultra High Dose Rates (UHDR) on DNA damage has been proven through the irradiation of in vitro cell cultures.


Laser-driven acceleration and FLASH radiotherapy
Since the development of the Chirped Pulse Amplification (CPA) technique, the generation of ultra-short and ultra-intense laser pulses and its applications in medical physics have become a very active field of research.These laser pulses have a temporal component previous to the main pulse which is produced due to the amplification of the spontaneous emission, generated during the process of the stimulated emission in the laser's gain medium.
If the laser pulse impinges focused on a solid target [1], this component can be intense enough to strongly ionize the surface of the target, creating a plasma.When the electromagnetic fields of the main pulse interact with the free electrons of the plasma, these will experience different acceleration mechanisms [2][3][4] depending on several parameters such as the laser intensity on focus, its angle of incidence or the target thickness.Consequently, the accelerated electrons will interact with the non-ionized atoms of the target producing a Bremsstrahlung continuum as they decelerate, and characteristic x-rays.
These specific particle acceleration mechanisms result in the capability of generating pulsed x-ray sources for preclinical research.In addition, since the acceleration occurs due to the generation of electric fields of the order of TV/m within a few microns, laser-driven acceleration gives rise to more compact sources if compared to conventional radiofrequency accelerators.Laser-induced radiation sources can produce relevant doses in short periods of time, which makes them suitable for radio- * email: alicia.reija@rai.usc.esbiological research related to Ultra High Dose Rate (UHDR/FLASH) therapy techniques.
The interest of delivering doses at very high rates relies on the so-called FLASH effect: while in conventional radiotherapy doses are fractionated and delivered at rates < 1 Gy/s, FLASH radiotherapy (FLASH-RT) proposes to shorten irradiation times while increasing the dose per fraction, so that the average dose rate becomes greater than 40 Gy/s; and often achieving peak dose rates of > 10 4 Gy/s.The main observed benefit of this new radiotherapy technique is a sparing of normal tissue without lowering the antitumoral efficacy achieved in a conventional treatment.This protection of healthy tissues has been experimentally observed in numerous pre-clinical in vivo studies in the last decade [5,6] and the safety and feasibility of the procedure has been tested in human studies [7].
The underlying biochemical mechanism that results in the FLASH effect remains unclear, although there are three main hypothesis that could explain such reduced toxicity of FLASH-RT to normal tissue.The first one proposes the possibility of a greater oxygen depletion in healthy cells during the irradiation time if high dose rates are employed: a transient state of hypoxia (low oxygen concentration) in a cell makes it significantly radiorresistant [8].The second hypothesis relies on the assumption that if the dose is delivered at a very short irradiation time, the number of lymphocytes damaged by the effects of ionizing radiation is significantly reduced, so the immune response that repares healthy cells within a FLASH irradiation may be higher if compared to conventional treatments.Finally, the last hypothesis states that the irradiation at very high instantaneous dose rates enhances high concentrations of radicals in the process of water radiolysis, favoring its recombination instead of its reaction with molecular oxygen, which would increase the probability of producing DNA permanent damage [9] [10].
Biological effects of laser-based x-ray sources have been investigated previously, addressing the dose and dose rate-dependent survival of certain mammalian cell lines [11,12].The setup presented here is characterized by dose concentration in ultra-short pulses.In addition, it allows for studying DNA damage induced by low-energy xrays (<30 keV) which can differ significantly from those at 100-200 keV and beyond [13,14].

Experimental setup of a laser-driven x-ray source at L2A2
In the Laser Laboratory for Acceleration and Applications (L2A2) at the University of Santiago de Compostela we have set up a table-top, laser-driven x-ray source for radiobiology studies in the FLASH regime.It is induced by Ti:Sa laser pulses of 1 mJ of energy and 35 fs of temporal FWHM operating at a 1 kHz repetition rate.These pulses are tightly focused with a microscope objective (Mitutoyo 20×) onto the surface of a copper plate.These laser pulses are p-polarized, and its angle of incidence with respect to the normal direction of the target is 45 • .The FWHM of the laser focal spot in the horizontal direction is 1.5 µm, so at the focus the reached intensities are > 10 18 W/cm 2 .At every laser shot, the copper surface becomes damaged, so the target is mounted on a rotatory motor, which combined with a linear motor that moves within the focal plane of the laser, refreshes the target surface at every laser shot, thus x-ray spectrum stabilisation is ensured.In addition, these two motors are mounted on another linear motor that corrects the wobble movement of the rotatory motor, so that during the entire irradiation the copper plate is placed within the focal plane of the laser with a precision comparable to the Rayleigh length.This entire motorized system was previously described in [15], and it is schemed in Figure 1.The induced x-rays are emitted in front of the copper target, with the direction of the maximum flux centered around the target's normal axis.

Energy distribution and dose evaluation
The energy spectrum of the x-ray source has been measured with a CdTe detector (Amptek XR-100T-CdTe) at 25 cm from the target-normal direction.A 1 mm diameter lead collimator was placed in front of its berylium window.In order to reduce pile-up effects, the spectrum was acquired with an aluminum filter of 165 µm width, and the de-piled spectrum was retrieved applying an algorithm described in [16].In Figure 2, the energy distribution shows a maximum x-ray flux around the superimposed K α and K β characteristic Cu peaks (8.03-8.91 keV).However, the de-pile algorithm was not capable of correcting the characteristic Cu peak repetition, and recent measurements in our laboratory have confirmed that the maximum cutoff energy is still overestimated.On the other hand, the delivered dose was measured with unlaminated EBT3 radiochromic films (RCF).Since this type of films have a low energy dependency when irradiated in air [17], the batch was calibrated with a 6 MV Linac.The RCF were scanned with an Epson Perfection V750 Pro at 300 dpp and in tiff format, and the dose was evaluated in the red channel.
Several measurements of the average dose ratedefined as the relation between the dose read at the RCF and the total irradiation time-are displayed in Figure 3.The average value at 10 cm is 28.14 mGy/s and the peak dose rate within a pulse is 8.04 • 10 8 Gy/s, which was es-timated assuming that the x-ray pulses are emitted at the same repetition rate and duration as the laser pulses, The relative error of the average dose rate shown in Figure 3 states that the source is not totally stabilized.This instability comes from the difficulty in the online correction of the wobble movement of the rotatory motor: since the maximum laser intensity on target is only achieved within a micrometric spatial interval (Rayleigh length), possible imprecisions during this correction significantly lower the x-ray generation efficiency.This issue has currently been addressed through the addition of an ionization chamber, which coupled to the laser shutter ensures a better precision in the total delivered dose.

Irradiation and analysis of in vitro cell cultures
To demonstrate the suitability of the source to perform radiobiology experiments at ultra-high dose rates, we tested the antitumoral response of the laser-induced x-rays through the irradiation of in vitro cell cultures.We chose the A549 line (human basal alveolar epithelial adenocarcinoma cells), and the cultures were prepared, standardized, handled and analysed in the Fundación Pública Galega de Medicina Xenómica (FPGMX).A monolayer of cells was grown on a 70 µm thick polyethylene (PE) foil sealing a sterile culture flask.Cell cultures were safely transported between FPGMX and L2A2.The samples were placed at 10 cm from the source, with irradiation times that ranged from 60 s to 150 s.
The total radiation dose was measured on radiochromic films placed outside of the cell culture flask.This films were covered with 11 µm of Al to filter the possible reflected laser energy that could saturate the film.
To determine the dose on the cells we performed separate measurements with radiochromic films placed inside one of the flasks, after the PE window and without the Al filter.The following correlation between the inside and the outside film was revealed: The correction factor k D = 0.702 was applied to the dose values recorded in the RCF placed after the Al foil.The effects of radiation toxicity were evaluated 24 h after the irradiation with the quantification of DNA repair foci assessed by immunofluorescence analysis of γ-H2AX and 53BP1 foci (see Figure 4).For each of the samples, 60 cell nuclei have been identified and the corresponding damage foci have been counted using confocal microscopy.In Figure 5, the number of colocalized γ-H2AX and 53BP1 foci versus the dose absorbed by the cultures is represented: while the effects of radiation were comparable to those shown by control samples for doses up to 3 Gy, from 3 to 5 Gy the number of foci quantified per cell showed a linear response with the absorbed dose (represented in red in Figure 5).
In addition, as a proof-of-principle, a series of identical cell samples were irradiated with a conventional radiotherapy source (6 MV Varian Linac) at FPGMX, in a similar range of total dose in order to establish the optimum culture time and analysis protocols.

Conclusions and future prospects
We have developed a laser-induced x-ray source for the irradiation of in vitro cell cultures.In a commissioning experiment, DNA damage has been quantified through γ-H2AX and 53BP1 immunofluorescence, and the average foci per cells versus the absorbed dose showed a linear response for doses greater than 3 Gy.Details of the total dose applied to the cells and the manipulation of biological samples have been addressed.
Further stabilisation of the dose rate is still ongoing.For real-time control of the total radiation dose an ionization chamber has recently been implemented.Our aim is to perform systematic studies of DNA damage by lowenergy x-rays and to compare them to other radiation sources such as a clinical Linac or laser-accelerated protons with an experimental setup described previously [18].

Figure 1 .
Figure 1.Scheme of the x-ray source setup.

Figure 2 .
Figure 2. Raw vs depiled x-ray energy spectrum, measured at 25 cm and at 0 • with respect to the target normal with a 165 µm Al filter.

Figure 3 .
Figure 3. Dose rate values and relative error.

Figure 4 .
Figure 4. Analysis of the immunofluorescence staining of foci γ-H2AX and 53BP1 in A549 cells in response to irradiation.

Figure 5 .
Figure 5. Combined γ-H2AX and 53BP1 foci per cell versus absorbed dose.Absorption on PE foil was taken into account in the dose evaluation.The red line is a linear fit to the mean values for doses > 3 Gy.