Increased elasticity of melanoma cells after low-LET proton beam due to actin cytoskeleton rearrangements

Cellular response to non-lethal radiation stress include perturbations in DNA repair, angiogenesis, migration, and adhesion, among others. Low-LET proton beam radiation has been shown to induce somewhat different biological response than photon radiation. For example, we have shown that non-lethal doses of proton beam radiation inhibited migration of cells and that this effect persisted long-term. Here, we have examined cellular elasticity and actin cytoskeleton organization in BLM cutaneous melanoma and Mel270 uveal melanoma cells. Proton beam radiation increased cellular elasticity to a greater extent than X-rays and both types of radiation induced changes in actin cytoskeleton organization. Vimentin level increased in BLM cells after both types of radiation. Our data show that cell elasticity increased substantially after low-LET proton beam and persisted long after radiation. This may have significant consequences for the migratory properties of melanoma cells, as well as for the cell susceptibility to therapy.

Cells are much more elastic after low-Let proton beam. Cellular elasticity was determined at 20 and 40 days post-irradiation. Untreated Mel270 and BLM cells exhibited a normal distribution of the Young's modulus values, although the mean value of Mel270 cells was almost two times higher than that of BLM cells (1.5 vs 0.8 kPa, respectively, Fig. 3A,B). This shows that Mel270 are much stiffer than BLM cells and confirms the widely accepted paradigm that metastatic cells are more elastic than cells from the primary tumor. In general, a substantial decrease of the Young's modulus values was observed in both cell lines after low-LET proton beam. For the highest dose of the low-LET proton beam, Mel270 were approximately 1.5-times more elastic and BLM 5 times more elastic than control (at 20 days post-radiation). In the case of X-rays and proton beam at 40 days the elasticity of Mel270 was not changed significantly. On the other hand, BLM cells displayed a noticeable decrease in the Young's modulus values at both 20 and 40 days after X-ray radiation. The histograms (Suppl. Fig. A1) show more random distribution at 40 days post-treatment than at 20, especially for Mel270 cells. BLM cells at 20 days after low-LET proton beam exhibit very narrow distributions, especially for the dose of 5 Gy. Moreover, morphological changes could be observed by AFM imaging in the examined cells (Fig. 4). As evident from the images,  proton beam radiation or X rays. Individual cell movements were evaluated at 20 and 40 days after irradiation with either proton beam or X-rays and three parameters were calculated: 'Mean rate of displacement' , i.e. the distance from the starting point direct to the cell's final position/time of recording; 'Distance' , i.e. the total cell trajectory (μm) and CME (coefficient of movement efficiency), i.e. the ratio of cell displacement to the cell trajectory length. Mean values presented as percent of control; *p < 0.05, **p < 0.01, ***p < 0.001. Data partially presented in Jasinska et al. 18 . www.nature.com/scientificreports www.nature.com/scientificreports/ Mel270 after both types of radiation became more constrict with no cytoskeleton features being pronounced. On the other hand, in BLM cells after proton irradiation less actin fibers were present in the center region of the cell, whereas in BLM cells after X-rays thick marginal actin fibers were visible. Furthermore, BLM cells displayed more fibroblastoid-like phenotype with higher number of internal actin fibers, known as stress fibers, combined with numerous protrusions being visible (Fig. 4B).
Alterations in marginal and internal actin fiber thickness following radiation. Analysis of actin cytoskeleton revealed disintegration of internal actin fibers in both types of cells (Fig. 5A,B). The thickness of internal actin fibers decreased in all groups in both cell lines at 20 and 40 days, with values in the range of 36-69% of the control level. The change in thickness of marginal actin fibers following radiation treatment varied based on cell line treated. In Mel270 at 20 days only proton beam leads to increase in thickness (with maximum value of 171% of control for 1 Gy), and at 40 days the effect is similar for both types of radiation (Fig. 5A, left). In BLM cells there was an increase in thickness of marginal fibers after proton beam, and a decrease after X rays (Fig. 5A, right). In summary, the effect on internal fibers was similar for both types of radiation and the effect of proton beam on marginal fibers was more pronounced. Actin aggregates were also seen in Mel270 cells (Fig. 4B). Taken together, these changes mark substantial alterations in the cytoskeleton structure.

Increased vimentin level in cutaneous melanoma after radiation. Vimentin is an intermediate fila-
ment that maintains cell and tissue integrity and confers resistance to mechanical stress 20 , but also have complex regulatory functions 21 . In order to determine the level of vimentin in the cells, the corrected total cell fluorescence (CTCF) was calculated from confocal microscopy images of cells stained for vimentin (Fig. 6). In Mel270 cells, a decrease of vimentin was seen at 20 days post-treatment with the lowest values of 49% and 38% after 5 Gy of the low-LET proton beam and X-rays, respectively. At 40 days there was no change, except for an increase of up to 134% for 1 Gy of the proton beam. However, in BLM cells an increased level of vimentin was seen, especially at 40 days after proton beam and 20 days after X-rays. BLM cells treated with 3 Gy of X-rays displayed an increase of the level of this protein to 206% of control. At 40 days we observed values above the control only in the groups treated with 1 or 3 Gy of the low-LET proton beam. These results correspond well with our previous analysis of vimentin levels performed by western blot analysis 18 . www.nature.com/scientificreports www.nature.com/scientificreports/ Adhesion and migration proteins. Radiation did not change the level of three adhesion and migration proteins studied: TNC, LAMB3 and ICAM1, except for an increase seen in the level of LAMB3 in BLM cells 20 days after X-rays. Decreasing trends in LABM3 and ICAM1 in Mel270 cells were not statistically significant (Fig. 7).

Discussion
In this work we have examined two types of melanoma cells: uveal melanoma Mel270 and cutaneous melanoma BLM cells, differing in their genetic phenotype. Cutaneous melanoma is generally not managed with radiation, except for lentigo malignant melanoma or as a palliative treatment 22 . Uveal melanoma, on the other hand, is routinely treated with both photon brachytherapy and low-LET proton beam radiation 5,8,23,24 . Since both types of melanoma metastasize, during which the cells undergo significant deformation of their body; the influence of radiation on their mechanical properties and thus their metastatic capabilities had to be addressed.
In our study, untreated BLM cells, originating from a metastasis of cutaneous melanoma, were twice more elastic than uveal Mel270 cells, originating from a primary uveal tumor. Irradiation led to an increased elasticity www.nature.com/scientificreports www.nature.com/scientificreports/ -in BLM cells after both types of radiation, whereas in Mel270 cells only after low-LET proton beam radiation. It has been reported that modifications of the actin cytoskeleton affect the mechanical properties of the cells to the highest extent [25][26][27][28][29][30] , therefore, in this study we have focused primarily on this cytoskeleton component.
Even though the total levels of actin were not changed after irradiation, actin cytoskeleton rearrangements were detectable long-term after the treatment. Significant decrease in thickness of inner actin stress fibers in all treated groups and an increase in the thickness of marginal fibers in all proton-beam treated groups was observed. The internal actin fibers seem to be more relevant to cellular elasticity 9,31 . Both Young modulus decrease with dose as well as migration parameters decrease with dose was accompanied with a decrease in thickness of internal fibers. These parallel changes were seen in Mel270 cells after proton beam radiation, and in BLM cells for both types of radiation (see Suppl. Fig. A2). Therefore, we concluded that the observed actin rearrangements may be partially responsible for higher elasticity in the treated cells. Of course, it should be noted that other cell constituents that build up the cytoskeleton of a cell such as intermediate filaments and microtubules could also be affected by radiation. However, their contribution to cell mechanics is less important than that of actin cytoskeleton 32 . Cellular elasticity is an end-result of many factors, including the length and thickness of the stress fibers, which are anchored to the adhesion focal points, as well as the number of internal fibers/arcs in the subcortical part of the cell 33 . Moreover, higher vimentin level in BLM cells might also play a role, as vimentin has been shown to enhance cellular elasticity 20 . Further elucidation of the mechanism of cellular softening after low-LET proton beam radiation might be provided by studying the changes in the expression of vinculin, myosin and microtubules organization of irradiated cells.
The increase in cell deformability directly correlates with progression from normal to transformed malignant cell 34 , as well as towards more metastatic phenotype 25,35,36 . Higher elasticity of cells is linked to weaker actin filament polarization and therefore leads to lower migration abilities 37 . Furthermore, migrating cells generally have thicker stress fibers than non-motile cells 37 and thicker stress fibers correspond with higher values of the Young's modulus of cells 30 . Mechanical softening of cancer cells and modification of their adhesion to extracellular matrix increased their capacity to escape the primary tumor 38 . It is worth pointing out that a substantial increase in elasticity might be connected to loss of membrane viability and cellular well-being, however, in our experiments cells were viable and proliferating.
Higher elasticity is thought to ease invasion and migration 11,28 . It should be noted that elasticity is a general parameter describing cell mechanical properties, which results from different factors such as: cytoskeleton organization, in particular actin organization, osmotic pressure, adhesion, etc 31 . Indeed, higher elasticity of a cell is related with increased invasiveness, because it is thought that the softer and more deformable a cancer cell is, the more likely it will invade different tissue barriers 11,25,39 . Accordingly, number of studies have shown that cells with similar migration capabilities but differing in their elasticity had different invasion capabilities 35 . However, in our study, we show that changes in the elasticity of the cells were accompanied by the disruption of actin cytoskeleton. Taking into consideration that actin cytoskeleton is one of the most important components of a cell for migration, it is very unlikely that cells could become more invasive when they have disrupted cytoskeleton. On the contrary, one may speculate that such substantial decrease in cellular stiffness as seen in BLM cells (up to 5 times) may completely unable cellular movement and result in decreased migratory properties. www.nature.com/scientificreports www.nature.com/scientificreports/ The effect of X-rays on elasticity and migratory properties was very much cell line dependent, i.e. the elasticity and migration of Mel270 cells were only slightly affected. In contrast, we observed a decrease in the Young's modulus values in BLM cells after X-rays (Fig. 3), and slight changes in migration potential of BLM cells after X-rays (Fig. 2). This points out that the same range of doses leads to slightly different effects in the tested cell lines. In general, the changes in cellular elasticity in Mel270 choroidal melanoma were much less pronounced. This may stem out from the fact that Mel270 cells origin from the primary tumor, and BLM from a metastasis of a cutaneous melanoma. Along with higher cell elasticity, BLM cells also exhibited a higher level of vimentin (after X-rays at 20 days and low-LET proton beam at 40 days), which may suggest the possibility of EMT transition switching phenotype in melanoma cells post-radiation. However, we have not detected any N-cadherin in the cells, which would be a confirmation of the switching phenotype (data not shown). Moreover, the effect of irradiation on the tumor microenvironment and interactions between cancer and other types of cells, should be further studied in this context, as it affects cellular elasticity and migratory properties as well 40 .
Radiation is known to enrich cellular population with cancer stem-like cells. The changes that we have observed at 20 and 40 days after radiation might represent an average over several cell subpopulations. We have shown previously the existence of three distinct cell populations in the untreated BLM and Mel270 cells, with a shift towards less-proliferative populations after irradiation. Therefore, the observed effects might be even higher if isolated subpopulations were studied. Other authors showed that the Young's modulus of tongue carcinoma cells was decreased in a dose-dependent manner in 24 hrs after X-rays (1-4 Gy) with over 3-fold decrease for the highest dose, with accompanying decrease in actin filaments and increased migration 14 . The increased thickness of peripheral actin fibers was also shown in endothelial cells in response to photon radiation 12 . Other authors suggest that the effects of radiation undergo intensive time changes up to 7 days post-radiation 41 . However, long term effects that are more suitable in terms of a life span of a patient have not been described in literature yet, and thus were examined in this study.
We have checked the level of three proteins involved in adhesion and migration, TNC, LAMB3 and ICAM1, but did not found any significant changes in the protein level, except for an increase in LAMB3 20 days after X-radiation. Our previous proteomic study revealed increase in several proteins involved in migration and metastasis, i.e. moesin (actin remodelling, motility), actinin 4 (migration and metastasis), FAB-2 (migration, microtubule destabilizer), lamine A/C (PI3K/AKT/PTEN, adhesion, motility), lamine B (motility) 42 . In the same study we also observed a decrease in annexin 7 (motility) and vimentin. The latter might result from post-translational modifications, as here we have seen a clear increase in the vimentin level (Fig. 6).
Although the exact mechanisms of the observed elasticity and cytoskeleton changes are not fully clear, there are some indicators in the literature, which prompt further research. Several actin-associated proteins have been implicated in cytoskeleton changes after radiation. For example, ankyrin-1 (ANK1) alters the structure of the actin cytoskeleton and sustains limited cell migration during DNA damage 43 . Similar alterations in the arrangement of actin microfilaments and detachments from junctions were seen in HeLa cells exposed to IR and to ROS 44 . Such actin filaments rearrangements might be induced by RhoA, as rapid formation of actin filaments accompanied by redistribution of VE-cadherin adherent junctions in microvascular endothelial cells after radiation was mediated by RhoA/ROCK 12 . It was shown that GNAQ mutation, one of the driver mutation in uveal melanoma, is present in Mel270 cells 45 . GNAQ mutation induces viability and migration of uveal melanoma cells via Notch signaling activation, which is mediated by YAP dephosphorylation and nuclear translocation 46 . Gαq activates YAP by acting on a www.nature.com/scientificreports www.nature.com/scientificreports/ Hippo-independent signaling network initiated by actin polymerization. Disruption of the actin cytoskeleton, e.g. due to ROS generated by radiation, diminishes both the basal activity of YAP and YAP hyperactivation 47 .
Interestingly, clinical data gathered from patients with uveal melanoma, do not indicate any significant difference in the mortality rate between radiation-treated patients and patients that undergone surgery 48,49 . This may suggest that irradiation of the tumor in the eye does not affect the process of metastasis. However, most authors would argue that metastases are seeded before the treatment or even before the diagnosis and remain dormant for many years 50 . Still, radiation may influence the cells seeding from the tumor after the treatment. Low, non-lethal doses tested may be relevant in the margin of the tumor treated with radiation.
A low-LET proton beam and photon radiation led to considerable rearrangements in the cytoskeleton of cells, affecting strongly their mechanical properties. Cellular elasticity increased more after proton beam irradiation. These changes were persistent long-term after the treatment and may be connected to an inhibition in cellular migratory properties. Taken together, these results suggest that low-LET proton beam radiation might have inhibitory effects on the migratory properties of melanoma cells. The exact molecular mechanism and signaling pathways leading to these effects require further studies.  53 . During the irradiation, the doses 1, 3, or 5 Gy have been delivered with a dose rate of 1 Gy/min, 2 Gy/min, and 6.6 Gy/min, respectively. At the center of cell container position i.e. at the depths 15.8 mm of the SOBP, the calculated Continues Slowing Down Approximation (CSDA) dose averaged LET d was 2.8 keV/µm. A detailed description of the irradiation and dosimetry can be found elsewhere 18 . Doses between 1 and 5 Gy were chosen as they allowed long term post-treatment studies.

Methods
Cell counting. The cells were seeded into 24-well plates (10 4 cells per well), and cell numbers were determined at each consecutive day during the five days duration at 20 days (4th passage) and 40 days (7th passage) following radiation. Cells were trypsinized and counted using Bürker hemocytometer. The experiment was conducted three times and for each repetition, cells were counted from 4 wells. Immunofluorescence staining and images analysis. Cells were plated on to microscope slides (Menzel-Glaser, Germany) in 3 cm dish (TPP, Switzerland) at a density of 15 × 10 4 cells. After 24 hours, the cells were fixed with 4% paraformaldehyde (Sigma-Aldrich, USA) at room temperature. Permeabilization and blocking steps were performed using 0.1% Triton X-100 (Avantor Performance Materials, Poland) and 3% bovine serum albumin (Sigma-Aldrich, USA), respectively. The primary antibodies used for staining were Vimentin (D21H3) Rabbit mAbs (Cell Signaling Technology, USA) in 0.1% BSA at a concentration 1:300. In the second step, a secondary solution in 0.1% BSA was used which contained Alexa Fluor ® 546 phalloidin (Thermo Fisher Scientific, USA), Alexa Fluor ® 488 goat anti-rabbit IgG (H + L) (Thermo Fisher Scientific, USA) and Hoechst 33258 (Thermo Fisher Scientific, USA). The cells were incubated at room temperature in the dark. Glass cover slides were placed on basic microscope slides (Elektromed, Poland) using fluorescent mounting medium (Dako, USA). Samples were kept in dark at 4 °C until analyzed with a scanning laser confocal microscope (Zeiss LSM 880 with Airyscan). The experiment was done at two time-points, 20 and 40 days after irradiation.
The corrected total cell fluorescence (CTCF) was measured using ImageJ v.1.43U (Wayne Rasband, National Institute of Health, USA). Fluorescence intensity was measured from maximum intensity projection of the stack images according to the formula 56 : CTCF = Integrated Density − (Area of selected cell × Mean fluorescence of background readings). For each group, 15 cells were analyzed. Thickness of actin fibers in the cells was determined based on fluorescent images of the cells with stained F-actin using ImageJ v.1.43U (Wayne Rasband, National Institute of Health, USA). Two types of fibers were determined: marginal fibers and internal fibers also www.nature.com/scientificreports www.nature.com/scientificreports/ called stress fibers. From each experimental group, including the untreated control, the fibers of 15 cells were quantified with 15 measurements for each type of fibers per cell.

Atomic force microscopy. Atomic force microscopy (AFM) analysis of the cells was conducted using a
Bruker BioScope Catalyst AFM coupled with an inverted optical microscope (Axio Observer Z1 from Zeiss). Measurements were performed on cells maintained in culture medium at 37 °C. Mechanical analysis of the cells was made in force spectroscopy mode. Before the measurements, the AFM probe was positioned on top of an individual cell and aligned at the center of the cell body using bright field optical microscopy live view at ×400 magnification. Once aligned, force curves from a grid of 5 × 5 points were collected at a rate of 1 Hz. 15 cells per each condition were analyzed. Mechanical measurements were conducted using soft cantilevers with a nominal tip radius of 20 nm and spring constant of 0.01 N/m. For precise mechanical characterization, spring constants of the used cantilevers were routinely determined based on the thermal tune procedure 57 . Analysis of force curves was made using AtomicJ software 58 . In brief, the collected force-displacement curves were first converted into force-indentation curves and fitted with the Sneddon model. A detailed description of the mechanical analysis used in this work can be found elsewhere 59 . Western blot. Cell monolayers were lysed in RIPA buffer (Thermo Fisher Scientific) protease inhibitor cocktail (Roche, Switzerland), PMSF and sodium orthovanadate. The amount of protein was measured using the BCA kit (Sigma-Aldrich, St. Louis, MO) and stored at −80 °C until used. Equal amounts of protein (20 μg) were run on Bolt1Bis-Tris Plus gels (Thermo Fisher Scientific) and transferred to a nitrocellulose membrane. The membranes were blocked with 5% skim milk or in 5% BSA in a TBS buffer with 1% of Tween 20 for 1 h and incubated with primary antibodies against ICAM-1 and Tenascin C (D16C4) (Cell Signaling Technology, MA, USA), LAMB3 (CL3363) (Thermo Fisher Scientific) at 4 °C overnight. Membranes were washed 3 times in TBS and incubated with suitable secondary antibodies and then washed 3 times in TBS. Signals were detected using SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Fisher Scientific). statistical analysis. Statistical analysis was carried out using Statistica v12 (StatSoft. Inc.). Due to the comparison of more than three experimental groups, statistical significance was estimated using evaluation of homogeneity of variances with Levene's Test followed by one-way analysis of variance (ANOVA). The differences were considered to be statistically significant at probability levels of p < 0.05, p < 0.01 and p < 0.001.

Data Availability
The datasets generated and analysed during the current study are available from the corresponding author on reasonable request.