High throughput targeted cell irradiation with single ions for low dose radiobiological experiments

Dipl.-Phys. Torsten Koal

For low dose radiobiological effects (Bystander effect - deviation from the linear no threshold dose response model) targeted single cell irradiation with counted single ions is required. For this purpose, a reliable recognition software of cell nuclei/compartments, based on dark-field microscopic images to avoid UV-light, and a platform for high precision ion irradiation of the detected targets were developed. It delivers single ions in air with a lateral precision of 0.5 µm. This allows precise irradiation of cell nuclei and other cell structures. A new in-house program Cellcognition was developed using the Matlab platform. The software recognizes the cell nuclei positions in dark field microscope pictures (no staining and no UV-illumination is needed) within a few seconds on a standard PC with a high correctness of >\percentval{96}. The missing 4% were caused by bi-nucleated cells, grains in the Mylar foil or blurred cells. These false positive detections can be manually discarded during the workflow of Cellcognition. For transforming the cell coordinates into beam coordinates two microscopes are necessary (offline at cell laboratory and online at the irradiation platform) as well as fiducial markers on the irradiation Petri dishes (a commercial 35 mm Petri dish with the bottom replaced by a 2.5µm Mylar foil or a Si₃N₄ membrane window).

In the high current mode ~1 nA, as an initial calibration, the program also recognizes reference beam positions on a glass substrate positioned in the focus plane of the cells for the determination of the beam spot positions. In the low current mode (<1 fA) a STIM (Scannig Transmission Ion Microscopy) measurement of a reference structure (a grid with a dimension of 2x2 mm² will be performed with the maximum scan size 2x2 mm² using 4095x4095 pixels. The structure can be recognized in the STIM image as well as in the online microscope image. This allows a projection of the cell coordinate system onto the beam coordinate system.

For the hit verification crosses with the dimension of 3x3 µm and a fluence of 10E8 protons/µm² were written on an artificial test structure (10x10 circles with 3 µm in diameter) in Mylar foil at predetermined coordinates. The crosses on the test structure showed a hit accuracy of 1.5 µm at the recognized coordinates

Figure 1: Top: Fluorescent crosses 10 µm x 10 µm in Mylar at the located nuclei positions. Bottom: DAPI stained nuclei show reduced fluorescence due to DNA destruction after irradiation.

These tests were extended to irradiation experiments with living cells (EaHy, human Fibroblasts). The hit accuracy tests with cells on Mylar foil were also carried out with a high fluence instead of single ions. Fluorescent crosses were written at recognized positions with the ion beam. Additionally, the (post irradiation) DAPI stained nuclei showed a reduced fluorescence at the irradiated areas which confirms the accuracy. The crosses for the nucleus hit verification are shown in Figure 1. The target hit accuracy is in total better than 2 µm.

The Petri dishes have to be in a vertical position due to the horizontal beam at LIPSION. Therefore, the online experimental procedure comprises medium removal and medium refilling. The fast irradiation ~5000 ions/s allows the irradiation setup to irradiate up to 1.500 cells in less than 1.5 minutes including Petri dish handling time.

We gratefully acknowledge the financial support by the NOTE IP 036465 (FI6R), Euratom specific programme for research and training on nuclear energy, 6th FP of the EC.