IntroductionMicronuclei (MN) formation is routinely used as a bio-dosimeter for acute radiation exposures1 and has historically been used as a measure of DNA damage in cells2,3. MN can be formed from whole centromeres lost during cell division due to damage to the mitotic apparatus or can form from acentric fragments caused by DNA damage4. Ionising radiation induces cellular DNA damage, with the most lethal being the creation of double strand breaks (DSBs). These breaks can repair correctly; incorrectly; or can remain as residual damage, where incorrect or residual repair may lead to chromosomal aberrations. Following cell division, fragments of DNA not attached to the centromere cannot be taken into daughter cells and remain as exposed DNA in the cytoplasm. As the new daughter nuclei form, the lost fragments of DNA may gather leftover nuclear envelope proteins and form a smaller (micro) nucleus within the new cell. These MN are fragile in comparison to the main nucleus and are prone to rupture3,5. Not all DNA fragments are drawn into MN and not all MN are prone to rupture6. Several recent studies have explored the properties of MN with particular emphasis on differences in the DNA damage and micro-nuclear DNA sequence and structure (3, 5).Interest in MN has recently grown due to their association with the immune response to radiation through the cyclic GMP–AMP synthase/ stimulator of interferon genes (cGAS /STING) pathway7. The cGAS/STING pathway responds to cytosolic DNA and triggers innate and adaptive immune responses8. It has been suggested to be a major contributor to abscopal effects from ionising radiation and its activation is the focus of many studies aiming to hijack the pathway for use in cancer treatment9. MN are closely associated with cGAS/STING and studies have suggested that MN rupture is the main mechanism by which DNA becomes cytosolic following ionising radiation exposure and triggering the pathway3. Understanding the exact relationship between MN production and ionising radiation could potentially be vital to the development of new treatment planning strategies to exploit natural immunity.Proton therapy is preferentially used in some cancer types due to its favourable physical dose distribution which delivers very little dose following the range of the protons resulting in a lower integral dose for normal tissue compared to photons. This is therefore suited to treating tumours close to radiation-sensitive structures, reducing dose to healthy tissue distal to the tumour site10. In addition, protons have higher linear energy transfer (LET) in the centre and towards the distal end of the Bragg peak, which is associated with more clustered damage around the particle tracks meaning there are more breaks within the localised region which may interact11,12,13. Generally protons have been demonstrated to produce slightly higher levels of damage than traditional photon irradiation, however there are also suggestions that the type of damage may be different. Proton damage may be more complex due to a more focused energy deposition, involving base damage or multiple strand breaks, which are thought to be harder to repair14. The combination of these factors may result in a greater cell kill from protons than photons, quantified by the relative biological effectiveness (RBE). RBE is clinically described as 1.1 for protons to represent their slightly higher killing power, however multiple studies demonstrate a variable RBE in vitro and in vivo, specifically changing with LET15,16. If irradiation at higher LET results in an increase in DNA damage quantity and complexity, an increase in MN induction may be expected due to greater fragmentation of the DNA.Previous in vitro studies demonstrate that MN formation is cell line dependent. Guo et al., Slavotinek et al. and Akudugu et al. analysed the MN yield following photon radiation in a combination of cancer cell lines17,18,19. All three studies found that, although there was a strong linear relationship in the number of MN per cell with dose, there was a large difference in the generation of MN across cell lines. This was independent of cell type, species or cancer type. Guo et al. noted an inverse relationship between extent of micronucleation and apoptosis in cell lines18. The link between radio-sensitivity and MN yield also remains controversial; Slavotinek et al. demonstrated an increased radio-sensitivity was linked with a higher proportion of MN19, however other studies have contrasting results18,20.Several previous papers reported the effect of radiation type and LET on MN induction, with most showing a positive correlation between number of cells containing MN and LET. Sun et al. demonstrated a larger increase in MN production following 18 keV/µm carbon ions and Staaf et al. illustrated a greater MN yield following 97 keV/µm alpha particles21,22. However, these MN formation as a function of radiation modality studies only used a single cell type, and the particles used have a much higher LET range than protons. More subtle differences between particle and photon data and lower LET ranges have not been specifically studied. A meta-analysis of studies published containing MN data revealed a large difference in the number of MN generated per dose between and within cell types. Thus, although there is a large amount of published data on MN, it should not be used to study the RBE of different radiation types due to the biological variability in the data. The major contributor to this variation was mainly attributed differences in experimental methods23.To better quantify the relationship between ionising radiation quality, cell line and MN, we studied MN formation in seven cell lines of different cell types following irradiation with photons and protons. To determine any relationship with LET, two proton LETs were used, 0.6 and 6.5 keV/µm (track averaged LET). We demonstrated a linear relationship between dose and the percentage of micronucleated cells for all cell lines and radiation types. Responses between cell lines varied greatly and did not cluster for cell type or p53 status. Trends in response to radiation type were seen for some cell lines, but these were not statistically significant, and the correlation between MN production per dose and LET was not consistent or quantifiable.Table 1 Cell line Characteristics.Full size tableMethodCell cultureRPE1, H358, HT1080, MG63, RD, and U2OS cell lines, Table 1, were sourced from research groups at the Manchester Cancer Research Centre. All cell lines were authenticated by the Cancer Research UK Manchester Institute Molecular Biology Core Facility. The facility uses the Verogen ForenSeq MainstAY kit for short tandem read (STR) profiling, looking specifically at 27 autosomal STR and 25 Y chromosome STR markers. They compare test results to ATCC references and an internal database of previously tested samples from within the institute. RPE1 p53(−/−) knock out (KO) cells were gifted from Prof Karen Vousden at the Francis Crick Institute, cells were edited as described by Muller et al.25. RPE1 and RPE1 p53(−/−) cells were cultured in DMEM-F12 HEPES buffered, supplemented with 2 mM v/v glutamine (Gibco, 11330032). All other cells were cultured in RPMI 1640 media with 2mM v/v glutamine (Gibco, 21875034). All media was supplemented with 10% v/v fetal bovine serum (Gibco, 10270106) and 1% v/v penicillin/ streptomycin (Sigma Aldrich, P0781). Cells were expanded to at least two passages before each experiment and were not passaged more than 25 times accounting for the previous passage count when the samples were acquired.IrradiationCells were seeded at 750–2000 cells/well into black optical bottomed 96 well plates (Thermo Scientific, 165305) in 200 µl complete media. Plates were irradiated with 2–4 Gy of photon or proton irradiation 32–40 h after seeding. Photon irradiation was performed using a CIX3 irradiator (Xstrahl Inc.) at 300 kV, 10 mA, source-to-surface distance of 400 mm and 2.3 mm thick Cu filter. The time required to irradiate samples was manually inputted based on a measured dose rate of 2.07 Gy/min.For proton irradiation, two different set ups were created to mimic entry protons and protons at the distal edge of a spread-out Bragg peak (SOBP). The two setups are henceforth referred to as low LET, describing the entry dose protons, and high LET, referring to the distal edge protons. Low LET protons used a single energy 245 MeV beam with a nozzle current of 1.29 nA to produce a track averaged LET of 0.6 keV/µm. High LET protons used a single energy 75 MeV beam with a nozzle current of 0.06 nA and 4 cm of solid water to further degrade the beam to 11.3 MeV on target (calculated through Geant4 Monte Carlo simulation), producing a track averaged LET of 6.5 keV/µm. Figure 1 demonstrates both proton setups with the pristine Bragg peaks from the two energies and their corresponding LET in a and the physical placement of the solid water in front of the irradiation cabinet for the high LET setup shown in b. Overall irradiation times for low LET protons and photons were similar (around one minute to irradiate with 2 Gy), however irradiation was slower for high LET protons (around 2–3 min to irradiate with 2 Gy).Monte Carlo simulations were carried out using Geant4 (version 10.7)23 for both the experimental set-up with a cell monolayer in a 96-well plate and the dosimetry set-up with a PTW microDiamond (PTW, Freidburg) in place of the cell layer. Dose and LET was scored and a calibration factor was determined to convert dose measured by the microDiamond to that seen by the cells. Proton irradiations were delivered by spot scanning using a pre-defined spot map with spots spacing of 2.5 mm scanning in a snake pattern. Doses were verified prior to sample irradiation with a PTW microDiamond held in situ. Gafchromic EBT3 film was used to ensure homogenous dose across the whole sample at both setups. Radiation was gated using an ionisation chamber. The dose error was