The role of irradiation induced changes in tumor vasculature on the infiltration of immune cells in colon cancer : research data underlying the doctoral dissertation

Šantek, Iva (avtor), Serša, Gregor (mentor), Markelc, Boštjan (komentor)

Colon cancer is the second most prevalent cancer in women and the third in men. The conventional therapies for treating colon cancer include surgery, chemotherapy, and radiotherapy (RT); however, they are often ineffective. The main mode of action of RT is the killing of proliferating cancer cells by causing double-strand DNA breaks. Besides, it leads to a release of reactive oxygen species (ROS) and reactive nitrogen species (RNS), as well as damage-associated molecular patterns (DAMPs) that activate dendritic cells to present tumor antigens to naive T lymphocytes. Consequently, naive T lymphocytes differentiate into effector T lymphocytes, primarily cytotoxic CD8+ T cells and helper CD4+ T cells, which begin to secret chemokines and cytokines, triggering an immune response. Tumors consist of cancer cells and tumor microenvironment (TME). An important part of TME are tumor blood vessels, which are required for tumor progression due to supplying them with nutrients and oxygen. In contrast to normal vasculature, tumor vasculature is not fully functional, as it is immature, disorganized, and inconsistent, which results in inadequate blood flow and the formation of hypoxic microregions, resistant to RT. Importantly, irradiation (IR) does not only affect cancer cells but also TME, including tumor vasculature. However, the response of the TME, including tumor vasculature is poorly understood. The effect of IR on tumor endothelium is time- and dose-dependent. IR can cause apoptosis of endothelial cells (ECs), with proliferating ECs being more susceptible. However, there is little evidence of vascular destruction with single doses of radiation below ~15 Gy or with daily fractionated doses. In addition to anti-angiogenic agents, IR alone can also act as a vascular normalization therapy. Besides the increased sensitivity of tumor cells to IR due to reduced levels of hypoxia, vascular normalization can also improve access of immune cells to previously non-perfused regions. Additionally, doses higher than 2 Gy may lead to endothelial activation, during which ECs switch from anti-inflammatory into pro-inflammatory phenotype. The underlying mechanism for the endothelial activation after IR remains unknown. The activated endothelium is characterized by the overexpression of adhesion molecules and pro-inflammatory chemokines and cytokines, contributing to the infiltration of immune cells. Recently, a subtype of blood vessels termed tumor associated high endothelial venules (TA-HEVs) was shown to represent the entry site for lymphocytes into tumors. However, the exact crosstalk between IR-activated tumor ECs, whether they are connected to TA-HEVs, and different populations of immune cells is yet to be elucidated. Therefore, the aim of the doctoral dissertation was to improve the understanding of the response of tumor blood vessels to RT by determining how the IR-induced changes to the vascular structure and endothelial activation are connected to the homing and extravasation of immune cells into the TME. In the first Work package, we investigated the effects of IR on the vascular network and endothelial activation in vitro. We assessed the response of murine EC lines bEnd.3, 2H11, and SVEC4-10 and human endothelial cell lines HUVEC, EA.hy926, Hulec-5a grown as monolayers to different IR doses and time after the exposure to IR. In the cell proliferation and survival experiments, cells were subjected to 2-10 Gy of IR one day after seeding. Immediately before IR treatment (0 h) and at different timepoints (24, 48, 72, and 96 h) after IR, cells were incubated in phosphate-buffered saline (PBS) solution containing Hoechst 33342 and Propidium Iodide (PI). The number of live cells (Hoechst 33342 positive nuclei) and the number of dead cells (PI positive nuclei) were quantified to calculate relative change of proliferation, number of normalized dead cells and doubling time. To evaluate IR-induced changes in the HUVEC transcriptome, RNA sequencing analysis was performed on HUVECs irradiated with either 2 or 5 Gy and compared to non-irradiated samples. Total RNA of non-irradiated and irradiated samples was extracted at 24 h or 72 h after IR. Transcriptome sequencing was conducted by Novogene Company LTD, using the Illumina platform. RNA sequencing analysis was performed using R software, focusing on significantly changed genes and pathways related to endothelial activation and immune response. In the next step, we established a vasculature-on-a-chip model, which enabled us to examine IR-induced changes in a physiologically relevant 3D microvascular environment. First, we attempted to reproduce a microfluidic chip protocol based on published literature the Chen et al. model. To establish a model capable of forming perfusable vascular networks, we tested the murine EC lines bEND.3, 2H11, and SVEC4-10, and the human EC lines EA.hy926, Hulec-5a and HUVEC. In addition, we optimized the EC seeding density, the EC:fibroblast ratio, the concentrations of thrombin and fibrinogen, and the introduction of interstitial flow. However, due to reproducibility issues, we adopted the OrganoPlate® microfluidic chips (Mimetas). Once established, the microvascular networks were irradiated with 2 or 5 Gy, based on prior IR dose-response results. To assess vascular permeability and integrity, Dextran dye was perfused into the microchips at 24 and 72 h after IR. Endothelial activation was evaluated by immunofluorescence staining for von Willebrand factor (VWF), Intercellular Adhesion Molecule 1 (ICAM-1), Interferon Regulatory Factor 9 (IRF9), and Interferon α/β (IFNα/β). Imaging was conducted with ZEISS ELYRA 7 Lattice SIM system. To assess interaction between irradiated endothelium and immune cells, fluorescently labeled CD8+ T cells, isolated from human buffy coat, were perfused through the microchips at 24 h and 72 h after IR. Imaging was conducted using Axio Observer 7 equipped with Colibri 7 LED light source UV and a Hamamatsu Orca Flash camera. The acquired images were processed and visualized using Imaris software (version 9.9.1). In the second Work package, we determined the level of vascular normalization and endothelial activation in vivo (Permit No. U34401-36/2020/7 and U34401-19/2024/9). We first irradiated MC38 and CT26 colon carcinoma tumors subcutaneously grown in C57Bl/6 or Balb/c mice, respectively, with either a single dose of 15 Gy or fractionated dose of 5 x 5 Gy and determined the response to RT. IR was conducted with a Gulmay CP225 X-RAY generator with 0.55 mm copper and 1.8 mm aluminum filtering, operated at 200 kV, 9.2 mA, with a dose rate of 1.92 Gy/min. To assess the effect of IR on the induction of endothelial activation and the formation of TA-HEVs, frozen sections of CT26 colon carcinomas excised at 24 h, 48 h, and 7 days after IR were immunofluorescently stained for the endothelial activation markers VWF and ICAM-1, the TA-HEV marker Peripheral Node Addressin Antibody (MECA-79), and markers associated with immune response IRF9, Cluster of differentiation 8 (CD8), and Cluster of differentiation 4 (CD4). The Cluster of differentiation 31 (CD31) was used to identify vasculature. Cell proliferation and hypoxia were evaluated by incorporation of thymidine analogue EdU and nitroimidazole derivative EF5 staining, respectively. Immunofluorescently stained tumor sections were imaged using Axio Observer 7 equipped with Colibri 7 LED light source UV and a Hamamatsu Orca Flash camera and analyzed using Imaris software (version 9.9.1). To investigate the transcriptional changes following IR, we used the 10x Genomics Visium Spatial Transcriptomics technology. Formalin-Fixed Paraffin-Embedded (FFPE) control and irradiated CT26 tumors excised at 48 h and 7 days after completion of fractionated 5 x 5 Gy RT, as well as tumor biopsies from three patients (Permit No. 0120-394/2023/3) obtained before neoadjuvant 5 x 5 Gy RT and samples from tumor resection after RT were analyzed. FFPE sections were mounted onto Visium Spatial Transcriptomics slides with capture probes, H&E stained, imaged, and processed for library preparation. Sequencing was conducted by Novogene, and the transcriptomic data were analyzed using using 10X Genomics SpaceRanger analysis pipeline (version 4.0) and a visualization software LoupeBrowser (version 8.0). We focused on IR-induced vascular changes associated with endothelial activation, immune cell infiltration, and TA-HEV formation. Immunohistochemistry on FFPE sections was performed for validation of the Spatial transcriptomics results using antibodies against CD31, CD4, CD8, IRF9, and Granzyme B. In the last Work package, the homing of immune cells following IR was investigated using the dorsal skinfold window chamber model, which enables real-time visualization of the tumor vasculature over time. Two symmetrical titanium frames, were surgically implanted into the dorsal skin of VE-TOM transgenic mice expressing fluorescent protein tdTomato in ECs. A circular window (~10 mm in diameter) was created by removing one skin layer, exposing the subcutaneous tissue and vasculature. MC38 colon carcinoma cells were injected into the remaining subcutaneous tissue of VE-TOM mice. Once the tumors reached a visible size of ~4 mm in diameter, they were irradiated with a single dose of 15 Gy using a specially designed collimator that allows IR of only the tissue within a 6 × 6 mm square inside the window chamber. Tumors were irradiated using the Gulmay 225 X-ray system with 0.55 mm Cu and 1.8 mm Al filtering with IR dose rate if 0.97 Gy/min. Vascular and immune cells' dynamics were monitored using intravital microscopy. A fluorescently labeled anti-MECA-79 antibody was injected intravenously (i.v.) to visualize the TA-HEVs. Further, fluorescently labelled CD8+ T cells isolated from mouse spleen were injected i.v. before imaging to determine their interaction with ECs. Imaging was conducted under isoflurane inhalation anesthesia using a Zeiss LSM 800 confocal microscope. The acquired 3D images were processed and visualized using Imaris software (version 9.9.1). For data analysis, visualization, and statistics, GraphPad, R software, and 10x Genomics software were used. Statistical methods were selected individually for each type of data. Statistical significance was determined using One-way, Two-way and mixed-design analysis of variance (ANOVA), linear mixed model (LMM), Wald test, and t-test. Not normally distributed data were analyzed with non-parametric tests (Wilcoxon test, Mann-Whitney test, and Kruskal-Wallis test).IR of murine bEnd.3, 2H11, and SVEC4-10 and human HUVEC, EA.hy926, and Hulec5a ECs with single doses of 2-10 Gy resulted in a dose-dependent reduction of proliferation and increase in EC death. Among the murine EC lines, bEnd.3 was the most radiosensitive, followed by 2H11 and SVEC4-10. In the bEnd.3 EC line, a reduced proliferation was already determined after the 2 Gy single dose IR; whereas, in the 2H11 and SVCE4-10 EC lines reduced proliferation was determined after 4 Gy. The percentage of dead cells confirmed that the most radiosensitive EC line was bEnd.3, as the percentage of dead cells was statistically significantly increased in all timepoints after IR with 4–10 Gy, which was not the case in 2H11 and SVEC4-10. In all murine EC lines tested, the doubling time was significantly prolonged after IR with 10 Gy; however, in bEnd.3 cells, a significantly longer doubling time was also observed after IR with 6 and 8 Gy. Among all tested human EC lines, HUVECs were the most radiosensitive, while the radiosensitivity of EA.hy926 and Hulec5a was lower but comparable. A statistically significant decrease in proliferation of HUVECs and Hulec-5a, compared to control, was observed regardless of the dose and timepoint, whereas EA.hy926 did not show significance at 24 h after IR with 2 Gy. In case of HUVECs, the most pronounced effect on cell death was determined 24 h after IR, while the percentage of death cells in EAhy.926 and Hulec5a was the highest at 72 h and 96 h after IR, respectively. IR with 10 Gy caused significantly prolonged doubling times in all tested human EC lines, while a significantly prolonged doubling time after IR with lower doses was determined in EAhy.926 (4–8 Gy), and HUVEC (6 and 8 Gy) cell lines. To understand the dependence of EC gene expression on IR dose and time after IR, we compared RNA transcriptomic profiles of HUVECs at 24 and 72 h after 2 or 5 Gy IR with corresponding control (non-irradiated) samples. In addition, we investigated the temporal dynamics of HUVEC gene expression by comparing their transcriptomic profiles at 72 h and 24 h after IR with the same doses. Differential gene expression (DGE) analysis revealed a more pronounced effect of IR on HUVEC global gene expression at the dose of 5 Gy compared to 2 Gy. Furthermore, regardless of the dose, a greater effect of IR was observed at 72 h compared to 24 h after IR. A Functional Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis confirmed our results on the effect of IR on EC proliferation, showing, irrespective of the dose or time after IR, down-regulation of cell cycle progression-related signaling pathways and up-regulation of signaling pathways related to cell cycle arrest. To assess whether IR alone could lead to endothelial activation, the genes and pathways associated with it, including pathways related to changes in the extracellular matrix (ECM) were assessed. Compared to the control group, a significant up-regulation of Cell adhesion molecules, ECM-receptor interaction, Focal adhesion pathways, and Regulation of actin cytoskeleton were observed 24 h after IR with 2 Gy. Contrary, IR with 5 Gy did not induce similarly pronounced effect on ECM-related pathways in HUVECs, with only C-type lectin receptor signaling pathway being up-regulated at 24 h. Finally, to assess whether IR alone could lead to the changes in ECs favorable for the activation of the immune system, the genes and pathways specific for the activation of the immune response were examined. IR of HUVECs activated pathways related to the immune response at the transcriptional level. Regardless of the dose, significant up-regulation of the innate immune response, including NF-κβ signaling, TNFα signaling, and Cytokine-cytokine receptor interaction pathways, was observed at 24 h after IR. This suggests that immune response and pro-inflammatory pathways are activated in irradiated ECs as early as 24 h after IR. Additionally, pathways associated with the immune response remained significantly up-regulated at 72 h after IR. In order to establish an in vitro model that ensures the formation of perfusable 3D vasculature, thus provides a more physiological relevant model compared to monolayer-based assays, we have established a vasculature-on-a-chip model as a microfluidic assay that allows real-time monitoring of changes in the vasculature after IR. We evaluated two different vasculature-on-a-chip models.In the first model (Chen’s model), we established the optimal protocol, which resulted in perfusable microvasculature after 7 days of incubation as follows: 1.3 × 10⁴ HUVECs/μl mixed with NHLF in a 6:1 ratio in 5.5 mg/ml fibrinogen and 100 U/ml thrombin in EBM-2 medium. In addition, the microchips were incubated on a rocker with a 7° tilt changing every 8 min. To visualize the formed vasculature, we performed immunofluorescence staining for Vascular Endothelial cadherin (VE-Cad), VCAM-1, and VWF. However, the reproducibility of this model was low. Therefore, we tested an alternative model, OrganoPlate® microchips (Mimetas). Using this system, we successfully established reproducible vasculature-on-a-chip networks. The seeding and culturing procedure followed the manufacturer’s protocol, with the following changes: 1.0 × 10⁴ HUVECs were seeded and angiogenic factors were added daily from day 4 after seeding. On day 7 after seeding, we perfused the microchips with 70 kDa FITC labelled Dextran, to confirm that the formed vasculature networks were perfusable and not leaky. We then performed IR with either 2 or 5 Gy and followed changes in vascular permeability, morphology, and expression of endothelial activation markers VWF, ICAM-1, IRF9, and IFNα/β within the course of 72 h. The most pronounced effect on EC activation was 72 h after IR with 5 Gy, when we observed increased expression of VWF and ICAM-1, within the cytoplasm and membrane of ECs. This correlated with nuclear translocation of IRF9, characterizing its transcriptional activity; and with an increased expression of IFNα/β, altogether indicating that activated endothelium exhibit a pro-inflammatory phenotype and contribute to the activation of immune response. To determine the level of vascular normalization, we analyzed the percentage of microchips growing perfusable, branched vascular networks. Results showed that IR did not lead to increased vascular normalization 24 h after IR, regardless of the dose. However, 72 h after IR, there was higher percentage of normal microchips after IR with 5 Gy, than in the control. This indicates that the vascular normalization induced by 5 Gy IR occurs at later timepoints. Interestingly, the percentage of perfusable microchips after 2 Gy was lower than the control suggesting that 2 Gy of irradiation does not have a vasculature normalizing potential. In contrast to vascular normalization, the number of CD8+ T cells interacting with ECs was significantly higher 24 h after IR with 2 or 5 Gy, in comparison to control. On the other hand, the number of CD8+ T cells interacting with ECs at 72 h after IR was comparable in irradiated versus non-irradiated microchips. In the second Work package, the effect of a single 15 Gy dose or a fractionated 5 × 5 Gy regimen of IR on tumor growth in CT26 and MC38 murine colon cancer models was evaluated in vivo. In both tumor models, a significant reduction in tumor volume was observed in irradiated tumors compared to controls, regardless of the IR regimen used. To assess whether and at which timepoint IR induces endothelial activation, fresh-frozen sections of irradiated and control colon carcinomas were stained for tumor endothelial cell (TEC) activation markers Vwf and Icam-1, as well as the TA-HEV marker Meca-79. IR induced a significant up-regulation of Vwf expression at all timepoints, regardless of the IR regimen, with the most pronounced effect observed 24 h after fractionated 5 × 5 Gy IR. Furthermore, fractionated IR dose caused significant up-regulation of Icam-1 and Meca-79, which were most elevated 24 h after IR. IR also caused up-regulation of Irf9, and increased infiltration of CD8+ and CD4+ T cells. The changes in hypoxia and the number of proliferative cells after IR were assessed using EdU and EF5 staining, respectively, showing IR-induced decrease of cell proliferation and hypoxia. In order to investigate dose- and time-dependent effect of IR on the transcriptome of irradiated murine CT26 colon tumors, within the spatial context of the TME, we then used the Visium Spatial Gene Expression platform (10X Genomics, US) to analyze FFPE sections of murine CT26 colon tumors. IR with fractionated 5x5 Gy dose led to up-regulation of TA-HEV associated genes in TECs at both 48 h and 7 days after IR. Additionally, we observed up-regulation of genes associated with immune response, which spatially colocalized with TA-HEV markers, suggesting an involvement of TA-HEVs and immune infiltration. This was then further confirmed at the protein level by immunohistochemical staining, with focus on regions with high TA-HEV transcript signatures. Despite different optimization parameters, we were not able to detect a Meca-79 positive TECs in any of the samples. However, immunohistochemical analysis confirmed increased expression of Irf9 and number of CD4+ and CD8+ T cells in irradiated tumors at both timepoints, supporting IR-induced immune activation. We next performed a spatial transcriptomic analysis on matched tumor biopsies from colorectal cancer patients taken before neoadjuvant radiotherapy (5×5 Gy) and samples from tumor resection taken after RT. The results showed that IR led to increased expression of TA-HEV-associated genes in TEC-positive regions. Consistent with in vivo experiments, IR led to up-regulation of immune-related genes that were spatially colocalized with TA-HEV markers. This was also confirmed on a protein level using immunohistochemical staining. In the third Work package, we determined how IR alters the homing and infiltration of immune cells, focusing on the interaction between TECs and CD8+ T cells, and the presence of TA-HEVs. In irradiated tumors, CD8+ T cells transmigrating across the vasculature and infiltrating into the TME were observed to a greater extent compared to control tumors. Next, we examined whether IR promotes the formation of TA-HEVs, by intravenously injecting anti-Meca-79 antibody into irradiated and control tumors growing in VE-TOM mice bearing dorsal skinfold window chambers. Although the Meca-79 signal was detected in one irradiated tumor, it was not observed in others. To validate the antibody delivery and functionality, lymph nodes from the same mice were harvested and imaged, showing Meca-79+ CD31+ vessels. This finding suggested that the low signal of Meca-79 observed in the tumors may be due to the limited abundance of Meca-79+ TA-HEVs or low expression levels of Meca-79 antigen on TECs, which made it impossible to perform intravital imaging of the interaction between CD8+ T cells and TA-HEVs. In this doctoral dissertation, we elucidated how IR modulates tumor vasculature to support anti-tumor immunity. Specifically, using a state-of-the-art 3D vasculature-on-a-chip model and in vivo analyses, we demonstrated that IR, dose- and time-dependently, induces activation of ECs and TECs, leading to increased expression of chemokines, cytokines, and over-expression of ECM markers, thus enhancing immune response. Furthermore, we showed that IR supports the induction of TA-HEVs, and, therefore, lymphocyte infiltration into tumors. The results of our research could be translated into the clinical setting to further improve the efficacy of RT and its scheduling in combination with immunotherapy. 

• Jezik: Angleški jezik
• Ključne besede: colon cancer, radiotherapy, tumor microenvironment, endothelial cells, endothelial activation, tumor vasculature
• Tipologija: 2.20 - Zaključena znanstvena zbirka raziskovalnih podatkov
• Časovno pokritje: 2021-2025
• Organizacija: MF - Medicinska fakulteta
• Leto izida: 2025
• PID: 20.500.12556/RUL-177242
• Metode zbir. podat.: Meritve in testi
• Datum objave v RUL: 30.12.2025

Licence

• Licenca: CC BY 4.0, Creative Commons Priznanje avtorstva 4.0 Mednarodna
• Povezava: http://creativecommons.org/licenses/by/4.0/deed.sl
• Opis: To je standardna licenca Creative Commons, ki daje uporabnikom največ možnosti za nadaljnjo uporabo dela, pri čemer morajo navesti avtorja.

Sekundarni jezik

• Jezik: Slovenski jezik
• Naslov: Vloga z obsevanjem povzročenih sprememb tumorskega žilja na infiltracijo imunskih celic pri raku debelega črevesa : raziskovalni podatki, obravnavani v doktorskem delu

Izvleček

Rak debelega črevesa je drugi najpogostejši rak pri ženskah in tretji pri moških. Običajne metode zdravljenja raka debelega črevesa vključujejo kirurški poseg, kemoterapijo in radioterapijo (RT), vendar so pogosto neučinkovite. Glavni mehanizem delovanja RT je uničenje proliferirajočih rakavih celic s povzročanjem dvojnih prelomov DNA. Poleg tega RT povzroča sproščanje reaktivnih kisikovih in dušikovih spojin ter s poškodbo povezanih molekulskih vzorcev (angl. damage-associated molecular patterns, DAMPs), ki aktivirajo dendritične celice za predstavitev tumorskih antigenov naivnim limfocitom T. Posledično limfociti T začnejo izločati kemokine in citokine ter sprožijo imunski odziv. Tumorji so sestavljeni iz rakavih celic in tumorskega mikrookolja. Pomemben del tumorskega mikrookolja predstavljajo tumorske krvne žile, ki so ključne za napredovanje tumorja, saj mu dovajajo hranila in kisik. Za razliko od normalnega žilja je tumorsko žilje nezrelo, neorganizirano in nepravilno, kar povzroča slab pretok krvi in nastanek hipoksičnih območij, odpornih na RT. Pomembno je, da obsevanje ne vpliva le na rakave celice, temveč tudi na tumorsko mikrookolje, vključno z žiljem. Odziv tumorskega mikrookolja, zlasti žilja, pa je še vedno slabo raziskan. Učinek obsevanja na tumorski endotelij je odvisen od časa in doze. Obsevanje lahko povzroči apoptozo endotelijskih celic, pri čemer so proliferirajoče celice bolj občutljive. Kljub temu je malo dokazov o uničenju tumorskega žilja pri enkratnih dozah, nižjih od 15 Gy, ali pri frakcioniranem obsevanju. Pokazano pa je bilo tudi, da lahko, podobno kot antiangiogene terapije, tudi obsevanje povzroči normalizacijo tumorskega žilja. Tako normalizirano žilje pa lahko vodi do zmanjšanja hipoksije, kar poveča občutljivost tumorskih celic na RT in izboljša dostopnost imunskih celic do tumorskih celic. Poleg tega lahko doze nad 2 Gy povzročijo aktivacijo endotelija, pri čemer endotelijske celice preidejo iz proti-vnetnega v pro-vnetni fenotip. Mehanizem te aktivacije po obsevanju pa je še vedno neznan. Za aktiviran endotelij je značilno povečano izražanje adhezijskih molekul ter vnetnih kemokinov in citokinov, kar spodbuja infiltracijo imunskih celic. Nedavno je bil opisan podtip krvnih žil, imenovan s tumorjem povezane visoke endotelijske venule (angl. tumor-associated high-endothelial venules, TA-HEV), ki predstavljajo mesto vstopa limfocitov v tumor. Vendar pa natančna povezava med z obsevanjem aktiviranim endotelijem, prisotnostjo TA-HEV in različnimi populacijami imunskih celic še ni razjasnjena. Cilj doktorske disertacije je bil izboljšati razumevanje odziva tumorskega žilja na radioterapijo z določitvijo, kako so spremembe, ki jih povzroča obsevanje v strukturi žilja in aktivaciji endotelija, povezane z infiltracijo in ekstravazacijo imunskih celic v tumorsko mikrookolje. V prvem sklopu raziskave smo preučevali učinke obsevanja na žilje in aktivacijo endotelijskih celic in vitro. Ocenili smo odziv mišjih endotelijskih celičnih linij bEnd.3, 2H11 in SVEC4-10 ter človeških endotelijskih linij HUVEC, EA.hy926 in HULEC-5a, gojenih v enoslojih, na različne doze obsevanja in različne časovne točke po izpostavitvi. V poskusih proliferacije in preživetja smo celice obsevali z odmerki 2–10 Gy en dan po nasaditvi. Neposredno pred obsevanjem (0 h) in po 24, 48, 72 in 96 urah smo celice inkubirali v fosfatnem pufru (PBS), ki je vseboval barvili Hoechst 33342 in propidijev jodid (PI). Število živih celic (pozitivnih na Hoechst 33342, negativnih na PI) in mrtvih celic (pozitivnih na Hoechst 33342 in pozitivnih na PI) smo kvantificirali za izračun relativne spremembe proliferacije, števila mrtvih celic in časa podvojitve populacije. Za določitev sprememb v transkriptomu endotelijskih celic HUVEC, ki jih povzroči obsevanje, smo izvedli analizo sekvenciranja RNA. Celice HUVEC smo obsevali z dozama 2 ali 5 Gy in jih primerjali z neobsevanimi vzorci. Skupno RNA smo izolirali 24 in 72 ur po obsevanju. Sekvenciranje transkriptoma je izvedlo podjetje Novogene Company Ltd. z uporabo platforme Illumina, analiza diferecialnega izražanja genov pa je bila izvedena z uporabo programske opreme R, s poudarkom na genih in poteh, povezanih z aktivacijo endotelija in imunskim odzivom. V naslednjem koraku smo vzpostavili model žilja–na–čipu, ki omogoča preučevanje sprememb po obsevanju v fiziološko relevantnem tridimenzionalnem mikrožilnem okolju. Najprej smo poskusili reproducirati model žilja po protokolu Chen in sod., s katerim smo optimizirali gostoto nasaditve endotelijskih celic, razmerje med endotelijem in fibroblasti, koncentracije trombina in fibrinogena ter uvedbo intersticijskega pretoka. Zaradi težav s ponovljivostjo vzpostavitve žilja v tem modelu smo za nadaljnje poskuse uporabili komercialni sistem OrganoPlate® (Mimetas). V tem modelu smo celice HUVEC nasadili v zunajcelični matriks (angl. extracellular matrix, ECM) in jih gojili v gojišču EBM-2, da so oblikovale pretočno mikrožilno mrežo. Ko je bila pretočna mikrožilna mreža vzpostavljena, smo jo obsevali z dozo 2 ali 5 Gy. Prepustnost in integriteto žil smo ocenjevali s perfuzijo barvila (70 kDa FITC označen dekstran) 24 in 72 ur po obsevanju. Aktivacijo endotelija smo določali z imunofluorescenčnim barvanjem naslednjih proteinov: von Willebrandov faktor (VWF), medcelična adhezijska molekula 1 (ICAM-1), interferonski regulatorni faktor 9 (IRF9) in interferon α/β (IFNα/β). Slike smo zajeli z mikroskopom ZEISS ELYRA 7 Lattice SIM. Interakcijo med obsevanim endotelijem in imunskimi celicami smo ocenili tako, da smo skozi nastali žilni sistem spustili fluorescenčno označene CD8⁺ T-limfocite, izolirane iz človeške krvi. Slike smo zajeli z mikroskopom Axio Observer 7 in analizirali s programom Imaris (različica 9.9.1). V drugem sklopu smo in vivo določili stopnjo normalizacije žilja in aktivacije endotelija. Tumorje MC38 in CT26 kolorektalnega karcinoma smo nasadili podkožno v miši sevov C57Bl/6 oziroma Balb/c ter jih obsevali z enkratno dozo 15 Gy ali s frakcionirano dozo 5 × 5 Gy. Učinek obsevanja smo ocenili z vidika rasti tumorjev, aktivacije endotelija in nastanka TA-HEV. Na zamrznjenih rezinah tumorjev, odvzetih po 24 in 48 urah in po 7 dneh po obsevanju, smo izvedli imunofluorescenčno barvanje za proteine, pomembne pri aktivaciji endotelija (VWF, ICAM-1), protein, značilen za TA-HEV (MECA-79), ter proteine, pomembne pri aktivaciji imunskega odziva (IRF9, CD8, CD4). Žilje smo določili z imunofluorescenčim barvanjem proteina CD31, proliferacijo celic z določevanjem vključitve molekule EdU v DNA ter hipoksijo z barvanjem EF5. Slike smo zajeli z mikroskopom Axio Observer 7 in analizirali s programom Imaris. Za določitev transkripcijskih sprememb po obsevanju smo uporabili tehnologijo prostorske transkriptomike Visium (10x Genomics), ki združuje kvantifikacijo izražanja RNA z mikroskopskimi slikami tkiva. Analizirali smo vzorce tumorjev CT26, odvzete 48 ur in 7 dni po končanem frakcioniranem obsevanju (5 × 5 Gy), fiksirane v formalinu in položene v parafin (FFPE), bioptične vzorce tumorja pred RT ter vzorec tumorja iz resekcije rektuma po RT (5 × 5 Gy). Rezine smo obarvali s hematoksilin-eozinom (H&E), zajeli slike vzorcev in jih nato uporabili za pripravo knjižnic. Sekvenciranje je izvedlo podjetje Novogene, analiza podatkov pa je bila opravljena s programoma SpaceRanger (različica 4.0) in LoupeBrowser (različica 8.0). Osredotočili smo se na spremembe v žilju, povezane z aktivacijo endotelija, infiltracijo imunskih celic in nastankom TA-HEV. Rezultate prostorske transkriptomike smo potrdili z imunohistokemijo na zaporednih rezinah z uporabo protiteles proti proteinom CD31, CD4, CD8, IRF9 in grancimu B. V zadnjem sklopu smo preučevali infiltracijo imunskih celic v tumor po obsevanju z uporabo modela dorzalnega okna, ki omogoča spremljanje žilja tumorja v realnem času. Dva simetrična titanova okvirja sta bila kirurško vstavljena v kožo na hrbtu transgenih miši VE-TOM, ki v endotelijskih celicah izražajo fluorescenčni protein tdTomato. Po odstranitvi ene plasti kože smo v preostalo tkivo injicirali tumorske celice MC38. Ko so tumorji dosegli premer približno 4 mm, smo jih obsevali z enkratnim odmerkom 15 Gy z uporabo kolimatorja, ki omogoča natančno obsevanje območja 6 × 6 mm znotraj okenca. Vpliv obsevanja na tumorsko žilje in imunske celice smo spremljali z intravitalno mikroskopijo s konfokalnim mikroskopom Zeiss LSM 800. Za vizualizacijo TA-HEV smo intravensko injicirali fluorescenčno označeno protitelo MECA-79, za sledenje interakcij med endotelijem in imunskimi celicami pa fluorescenčno označene CD8⁺ limfocite T, izolirane iz vranic miši. Za analizo, vizualizacijo in statistično obdelavo podatkov smo uporabili programe GraphPad, R in 10x Genomics. Statistične metode so bile izbrane glede na vrsto podatkov. Statistično značilnost smo določali z enosmerno, dvosmerno ali mešano analizo variance (ANOVA), linearnim mešanim modelom, Waldovim testom ali t-testom. Podatke, ki niso sledili normalni porazdelitvi, smo analizirali z neparametričnimi testi (Wilcoxonov, Mann-Whitneyjev in Kruskal-Wallisov test). Obsevanje mišjih celičnih linij bEnd.3, 2H11 in SVEC4-10 ter človeških endotelijskih celičnih linij HUVEC, EA.hy926 in HULEC-5a z enkratno dozo 2–10 Gy je povzročilo od doze odvisno zmanjšanje proliferacije in povečanje celične smrti. Med mišjimi celičnimi linijami je bila najbolj radiosenzitivna linija bEnd.3, sledili sta ji 2H11 in SVEC4-10. Pri liniji bEnd.3 je bila zmanjšana proliferacija opažena že po enkratni dozi 2 Gy, medtem ko je bila pri 2H11 in SVEC4-10 zaznana po 4 Gy. Odstotek mrtvih celic je potrdil, da je linija bEnd.3 najbolj občutljiva, saj se je delež mrtvih celic statistično značilno povečal pri vseh časovnih točkah po obsevanju z 4–10 Gy, kar ni bilo opaženo pri drugih linijah. V vseh mišjih celičnih linijah se je čas podvojitve populacije po 10 Gy statistično značilno podaljšal, pri bEnd.3 pa se je statistično značilno podaljšal že po 6 in 8 Gy. Med človeškimi celičnimi linijami so bile HUVEC celice najbolj radiosenzitivne, medtem ko sta celični liniji EA.hy926 in HULEC-5a izkazovali nižjo, a med seboj podobno radiosenzitivnost. Statistično značilno zmanjšanje proliferacije HUVEC in HULEC-5a celic v primerjavi s kontrolami je bilo opaženo ne glede na dozo in časovno točko, medtem ko pri celični liniji EA.hy926 nismo opazili sprememb 24 ur po 2 Gy. Pri HUVEC celicah je bil največji delež mrtvih celic opažen po 24 urah, pri EA.hy926 in HULEC-5a pa po 72 oziroma 96 urah. Obsevanje z 10 Gy je povzročilo značilno podaljšanje podvojitvenega časa pri vseh človeških celičnih linijah, medtem ko je bil ta učinek po nižjih dozah opažen le pri celičnih linijah EA.hy926 (4–8 Gy) in HUVEC (6 in 8 Gy). Za določitev, ali so spremembe izražanja genov po obsevanju odvisne od doze in časa po obsevanju, smo primerjali transkriptomske profile celic HUVEC po 24 in 72 urah po obsevanju z 2 ali 5 Gy v primerjavi z neobsevanimi vzorci. Analiza diferecialnega izražanja genov je pokazala bolj izrazit vpliv obsevanja s 5 Gy kot z 2 Gy, ne glede na časovno točko. Večji vpliv obsevanja je bil opažen 72 ur po obsevanju kot po 24 urah. Analiza KEGG je potrdila zmanjšano aktivacijo signalnih poti, povezanih s celičnim ciklom in aktivacijo signalnih poti, povezanih z zaustavitvijo celičnega cikla in aktivacijo p53 signalne poti. Za določitev, ali obsevanje samo lahko sproži aktivacijo endotelija, smo analizirali gene in signalne poti, povezane s spremembami zunajceličnega matriksa. V primerjavi s kontrolno skupino je bilo 24 ur po obsevanju z 2 Gy zaznano značilno povečano izražanje molekul celične adhezije, ter aktivacijo poti interakcije ECM-receptor, fokalne adhezije in regulacije aktinskega skeleta. Obsevanje s 5 Gy ni povzročilo enako izrazitih sprememb v teh poteh, vendar pa je bila opazna aktivacija poti lektinskih receptorjev po 24 urah. Za določitev, ali obsevanje sproži spremembe, ki podpirajo aktivacijo imunskega sistema, smo analizirali signalne poti, specifične za imunski odziv. Obsevanje HUVEC celic je povzročilo povečano aktivacijo signalnih poti, povezanih z imunskim odzivom. Ne glede na dozo sevanja je bila po 24 urah opažena povečana aktivacija signalne poti prirojenega imunskega odziva, vključno s signalizacijo NF-κB, TNFα in interakcijo citokin–receptor. To kaže, da so vnetne poti aktivirane že zgodaj po obsevanju. Povečana aktivacija teh poti je vztrajala tudi po 72 urah. Za vzpostavitev in vitro modela, ki omogoča razvoj pretočne 3D mikrožilne mreže, smo razvili model žilja–na–čipu. Preizkusili smo dva pristopa – prvi pristop je temeljil na protokolih, objavljenih v literaturi (»model Chen«), drugi pa na komercialno dostopnih modelih OrganoPlate® (Mimetas). Model OrganoPlate® se je izkazal kot bolj ponovljiv in je omogočil oblikovanje stabilnih pretočnih mikrožilnih mrež, ki smo jih potrdili s perfuzijo dekstrana (70 kDa FITC). Žilne mreže smo obsevali z 2 ali 5 Gy in spremljali spremembe v prepustnosti, morfologiji in izražanju označevalcev aktivacije endotelija (VWF, ICAM-1, IRF9, IFNα/β). Najizrazitejša aktivacija je bila opažena 72 ur po obsevanju s 5 Gy, ko se je povečalo izražanje VWF in ICAM-1, translokacija IRF9 v jedro pa je potrdila transkripcijsko aktivacijo. Skupaj z večjim izražanjem IFNα/β ti rezultati kažejo, da obsevani endotelij pridobi vnetni fenotip in prispeva k aktivaciji imunskega odziva. Za oceno stopnje normalizacije žilja smo analizirali delež čipov z organizirano, razvejano in pretočno žilno mikromrežo. Rezultati so pokazali, da obsevanje po 24 urah ni povzročilo izboljšanja normalizacije, ne glede na odmerek. Po 72 urah pa je bil delež čipov s pretočno žilno mikromrežo po 5 Gy večji kot pri kontrolah, kar kaže, da normalizacija žilja po obsevanju z 5 Gy nastopi pozneje. Po 2 Gy je bil delež čipov s pretočno žilno mikromrežo nižji od kontrol, kar nakazuje, da ta odmerek nima normalizacijskega učinka. Za določitev, če z obsevanjem inducirana aktivacija žilja vpliva na interakcijo CD8⁺ T-limfocitov, smo le-te vnesli v pretočne žilne mikromreže. Število CD8⁺ T-limfocitov, ki so bili v stiku z endotelijem, je bilo 24 ur po obsevanju z 2 ali 5 Gy statistično značilno večje kot pri kontrolnih čipih, medtem ko so bile po 72 urah razlike med skupinami manjše. V drugem sklopu smo in vivo analizirali učinke enkratnega (15 Gy) in frakcioniranega (5 × 5 Gy) obsevanja na rast tumorjev CT26 in MC38. V obeh modelih je bilo po obsevanju opaženo značilno zmanjšanje volumna tumorjev v primerjavi s kontrolami. Obsevanje je povzročilo povečano izražanje označevalcev aktivacije endotelija (VWF, ICAM-1) in označevalca TA-HEV (MECA-79), pri čemer je bil učinek najbolj izrazit 24 ur po frakcioniranem obsevanju (5 × 5 Gy). Istočasno je bilo opaženo povečano izražanje IRF9 ter večja infiltracija CD8⁺ in CD4⁺ limfocitov T. Imunofluorescenčno barvanje z EdU in EF5 je pokazalo zmanjšano proliferacijo celic in hipoksijo po obsevanju. Za razumevanje transkripcijskih sprememb po obsevanju v prostorskem kontekstu tumorskega mikrookolja smo uporabili tehnologijo prostorske transkriptomike Visium. Frakcionirano obsevanje (5 × 5 Gy) je v endotelijskih celicah po 48 urah in 7 dneh povzročilo povečano izražanje genov, povezanih s TA-HEV in sočasno povečano izražanje genov povezanih z imunskim odzivom, ki so se prostorsko prekrivali z označevalci TA-HEV, kar kaže na povezavo med aktivacijo TA-HEV in infiltracijo imunskih celic. Ti rezultati so bili potrjeni tudi na proteinski ravni z imunohistokemičnim barvanjem, ki je pokazalo večje izražanje IRF9 ter večje število CD4⁺ in CD8⁺ T limfocitov v obsevanih tumorjih. Podobne spremembe smo zaznali v ujemajočih bioptičnih vzorcih tumorja pred RT in vzorcu tumorja iz resekcije rektuma po RT (5 × 5 Gy). Po obsevanju je bilo v regijah z endotelijskimi celicami opaženo povečano izražanje s TA-HEV-povezanih genov in genov, povezanih z imunskim odzivom, kar je bilo potrjeno tudi z imunohistokemičnim barvanjem. V tretjem sklopu smo z intravitalno mikroskopijo v mišjem modelu dorzalnega okna spremljali vpliv obsevanja na privabljanje in infiltracijo imunskih celic v tumor. V obsevanih tumorjih je bilo opaženo večje število CD8⁺ T limfocitov, ki so prehajali skozi žilje in vstopali v tumorsko mikrookolje, kot v kontrolnih neobsevanih tumorjih. Obsevanje je torej povečalo ekstravazacijo CD8+ T limfocitov in s tem funkcionalno potrdilo aktivacijo endotelija po obsevanju. Po drugi strani pa so bile MECA-79+ tumorske endotelijske celice zaznane le v enem obsevanem vzorcu, kar kaže, da je nizka prisotnost TA-HEV ali nižji nivo izražanja antigena omejila neposredno vizualizacijo teh struktur v tumorjih v dorzalnem oknu. V tej doktorski disertaciji smo pojasnili, kako obsevanje spreminja tumorsko žilje in spodbuja protitumorski imunski odziv. Ugotovili smo, da obsevanje povzroča aktivacijo endotelijskih celic, izražanje kemokinov in citokinov ter povečano izražanje proteinov zunajceličnega matriksa, kar skupaj prispeva k vzpostavitvi vnetnega imunskega mikrookolja. Poleg tega obsevanje spodbuja normalizacijo žilja, kar izboljša pretočnost tumorskih žil in izboljša dostopnost imunskih celic do regij, kjer predhodno ni bilo krvnega pretoka. Pokazali smo tudi, da obsevanje podpira nastanek s tumorjem povezanih visokih endotelijskih venul (TA-HEV) in s tem omogoča večjo infiltracijo limfocitov v tumorje. Ti rezultati dokazujejo, da obsevanje tumorski endotelij preoblikuje v imunološko podporno stanje, ki spodbuja vdor T-celic in aktivacijo protitumorskega imunskega odziva. Ugotovitve naše raziskave ponujajo mehanistično osnovo za optimizacijo radioterapije in njeno kombinacijo z imunoterapijo, kar bi lahko izboljšalo terapevtsko učinkovitost pri raku debelega črevesa.

• Ključne besede: rak debelega črevesa, radioterapija, tumorsko mikrookolje, endotelijske celice, endotelijska aktivacija, tumorsko žilje

Projekti

• Financer: ARIS - Javna agencija za znanstvenoraziskovalno in inovacijsko dejavnost Republike Slovenije
• Številka projekta: P3-0003
• Naslov: Razvoj in ovrednotenje novih terapij za zdravljenje malignih tumorjev

• Financer: ARIS - Javna agencija za znanstvenoraziskovalno in inovacijsko dejavnost Republike Slovenije
• Številka projekta: J3-2529
• Naslov: Vloga endotelija pri odgovoru tumorja na radioterapijo