Can Cancer Cell Line In Vitro Radioembolization Model Help to Understand How Beta Radiation Affects Liver Tumor and Improve Treatment Results?

TL;DR

This study uses an in vitro model to explore how beta radiation from 90Y-microspheres affects liver tumor and healthy cells under different oxygen conditions.

Contribution

The study introduces an in vitro model to investigate the effects of beta radiation on cancer and healthy liver cells under normoxic and hypoxic conditions.

Findings

Hypoxia reduces sensitivity to 90Y radiation and promotes a pro-angiogenic tumor environment.

HCT 116 and THLE-2 cells show radiosensitivity, while HepG2 cells are more resistant to radiation.

Radiation effects vary by cell type and oxygen levels, influencing apoptosis and angiogenic signaling.

Abstract

Radioembolization with 90Y-labeled microspheres is an established locoregional therapy for primary and metastatic liver tumors, yet the molecular mechanisms underlying tumor response and resistance, particularly within hypoxic tumor microenvironments, remain poorly understood. Here, we developed an in vitro model of 90Y-microsphere irradiation to investigate the effects of beta radiation under normoxic and hypoxic conditions in human colon cancer (HCT 116), hepatocellular carcinoma (HepG2), and non-malignant liver (THLE2) cell lines. Cells were exposed to two microsphere dilutions, and absorbed doses were estimated using FLUKA-based Monte Carlo simulations. Cellular viability, proliferation, apoptosis-related gene expression, and secretion of angiogenic and inflammatory mediators were assessed. 90Y-microspheres exerted both cytotoxic and cytostatic effects in a cell type- and oxygen-dependent manner. In HCT 116 cells, radiation reduced metabolic activity and proliferation and induced Bax expression in normoxia, whereas hypoxia conferred radioresistance associated with Bcl-2 upregulation. HepG2 cells displayed pronounced resistance, with preserved viability but marked inhibition of proliferation, indicating a predominantly cytostatic response that was further attenuated by hypoxia. In contrast, THLE-2 cells were radiosensitive, showing reduced proliferation and increased Bax expression, while hypoxia partially protected against radiation-induced damage. Both hypoxia and microsphere exposure promoted a pro-angiogenic phenotype, characterized by increased VEGF secretion in radiosensitive tumor and healthy cells, whereas resistant HepG2 cells exhibited angiogenic signaling linked to TIMP2 suppression and IL-8 induction. Collectively, our findings demonstrate that hypoxia diminishes cellular sensitivity to 90Y beta radiation while fostering a microenvironment conducive to tumor progression. This in vitro model provides mechanistic insight into radioembolization responses and may support the development of more personalized therapeutic strategies.

Figures

Full text

1. Introduction

Radioembolization of primary and metastatic liver tumors is proven to be a safe and effective method of intra-arterial brachytherapy that leads to better control of liver cancer disease. High-rate treatment response and prolonged time to progression is observed after one-session treatment. None from a few controlled multicenter trials have shown a positive impact on overall survival rate [1,2,3]. In radioembolization Yttrium-90 serves as a beta-emitting radionuclide and is incorporated into resin or glass microspheres that are administered via the hepatic artery. Owing to the preferential arterial blood supply of liver tumors, these ^90^Y-loaded microspheres accumulate more readily within malignant tissue than in normal liver parenchyma, with an approximate tumor-to-healthy liver uptake ratio of 3:1. However, the intratumoral distribution of microspheres is highly heterogeneous and differs among lesions, depending largely on the architecture and density of the tumor vasculature. Yttrium-90 is a pure, high-energy beta emitter with a physical half-life of approximately 64 h, allowing the microspheres to deliver a clinically relevant radiation dose over a period of about two weeks following administration. Current evidence suggests that an absorbed radiation dose exceeding 70 Gy is necessary to achieve effective eradication in liver tumors [4,5]. In order to achieve satisfactory radioembolization treatment response several issues need to be addressed. On one hand, there is a trend for better patient selection for radioembolization utilizing a dedicated software, which allows more personalized dose calculation and safer ^90^Y-micropshere deposition within liver tumors with lower risk of excessive irradiation of healthy liver tissue. Moreover, a focus on the investigation of the mechanisms behind beta radiation effects on liver cancer is necessary [4,6,7,8,9]. Accordingly, we further developed an in vitro model employing ^90^Y-microspheres to better elucidate how beta radiation induces cancer cell death and to identify the mechanisms underlying radiation resistance. The model was expanded to include additional cancer cell lines as well as normal liver cell lines. Cellular viability and proliferation, apoptosis-related gene expression, and the secretion of angiogenic and inflammatory mediators were systematically evaluated.

2. Results

2.1. Calculated Absorbed Doses for Cell-Lines Cultures

Table 1 summarizes the absorbed dose values calculated for each experimental configuration, including hypoxic and normoxic conditions and both microsphere dilutions. For each case, the nominal absorbed dose in the cell-line culture is reported together with the corresponding upper and lower uncertainties expressed in Gy. These uncertainties were derived from the fluctuations in the specific activity of the microspheres, assuming +13.7% and −6.7% deviations from the nominal value and thus represent the maximal range of variation expected for the delivered dose in the experimental setup.

2.2. Effects of Oxygen Level on Colon Cancer Cells Response to 90Y-Microsphere Treatment

We compared the effect of ^90^Y-microspheres on colon cancer cells in normoxia and hypoxia in vitro. Microscopic observation showed that the addition of microspheres changes cells morphology—they become visibly enlarged (Figure 1A). Moreover, in wells treated with spheres we observed fewer cells than in control wells and this effect is more visible in standard culture conditions (21% pO2).

These observations correspond to the results of the assessment of cell viability indirectly determined by measuring the metabolic activity of cells, using Alamar Blue assay (Figure 1B). In normoxia, there is a decrease in the metabolic activity of cells after adding microspheres; a smaller number of spheres (Spheres 1) caused a decrease by about 24% compared to the control (p-value = 0.021), while a larger number of spheres (Spheres 2) by 50% (p-value < 0.001). We can also observe statistically significant differences in normoxia between the two doses of microspheres, which suggests that in standard in vitro conditions the effect of microspheres is dose-dependent. However, in hypoxia, which better reflects tumor microenvironment (especially after oxygen depletion during radioembolization), a statistically significant reduction in metabolic activity was achieved only with a larger number of spheres (Spheres 2)—about 21% decrease compared to hypoxic control cells (p-value = 0.0064). The decreased metabolic activity after the addition of microspheres indicates both cytotoxic and cytostatic effects of the microspheres.

Additionally, we assessed the effect of microspheres on cell proliferation using BrdU incorporation assay (Figure 1C). In normoxia, a strong reduction in proliferation was observed after adding more spheres (Spheres 2) (it decreased by about 50%, p-value = 0.001). However, a lower dose of spheres (Spheres 1) did not cause statistically significant changes in proliferation. This indicates a strong cytostatic effect at a high radiation dose in normoxia. In hypoxia, a statistically significant effect was also observed after adding higher dose, but the reduction was much weaker than in normoxia (it decreased by about 20%, p-value = 0.0169, compared to hypoxic control).

HCT 116 cells not treated with microspheres (control cells) show lower metabolic activity and reduced proliferation under hypoxia compared to standard culture conditions (normoxia).

As cell viability was affected by treatment, we checked the levels of pro- and anti-apoptotic genes. Transcript analysis showed that exposure to a high dose of radiation (Spheres 2) in normoxia results in a strong upregulation expression of the Bax gene encoding the pro-apoptotic protein Bax (Figure 1D), additionally supporting the observed cytotoxic effect. The increase in Bax expression in hypoxia did not reach statistical significance; however, in hypoxia cells after exposure to microspheres the anti-apoptotic gene Bcl2 was strongly expressed (Figure 1E).

2.3. Effects of 90Y-Microspheres Treatment on Liver Cancer Cell Line Proliferation and Metabolic Activity

The effect of the microspheres on liver cancer cells was similarly assessed. Morphologically, significant changes were observed; after exposure to radiation cells become visibly enlarged, do not form tight connections and become more dispersed (Figure 2A).

After exposure to microspheres cells remain metabolically active—there are no statistically significant differences in Alamar Blue reduction between the control cells and those treated with microspheres, both in normoxia and hypoxia (Figure 2B).

At the same time, proliferation (measure using BrdU) is strongly reduced in normoxia, about 47% (p-value = 0.02) after adding a lower number of spheres (Spheres 1) and about 65% (p-value = 0.0011) with a higher number of spheres (Spheres 2) (Figure 2C). In hypoxia, the statistically significant effect is visible only at a higher dose of microspheres (Spheres 2), where the decrease in proliferation is by about 47%.

There were no differences in both proliferation and metabolic activity between normoxic and hypoxic control cells. No evidence of sphere cytotoxicity in used concentration range was observed; it had cytostatic effect when in higher dose.

Although in Alamar blue assay, we could not observe a reduction in cell viability, we observed a tendency towards an increase in the expression of the pro-apoptotic Bax gene in response to microspheres treatment in normoxia, suggesting some cytotoxic activity of the spheres (Figure 2D). However, the expression of the anti-apoptotic Bcl2 was not shown (it was below the limit of detection in both normoxia and hypoxia).

2.4. Effects of 90Y-Microspheres on Proliferation and Viability of Healthy Liver Cells

The effect of microspheres on healthy tissue was assessed using the immortalized cells of the left lobe of the liver—THLE-2. Microscopic observations showed that the addition of microspheres reduces the number of cells in both normoxic and hypoxic conditions, but morphologically cells remain similar.

In normoxia, metabolic activity, measured by Alamar Blue, decreases dose-dependently (Figure 3B), but the reduction in cell proliferation (BrdU assay) becomes statistically significant only when using a larger number of spheres (Spheres 2) (Figure 3C). The decreased metabolic activity and proliferation of THLE-2 cells after the addition of microspheres indicates that such therapy may have both cytotoxic and cytostatic effects on healthy tissue. At the same time, the increased expression of the pro-apoptotic Bax gene was observed (Figure 3D). In low oxygen tension, only a high dose of radiation causes a decrease in metabolic activity and cell proliferation, but it was not enough to induce Bax gene expression.

2.5. Effect of Various Oxygen Conditions on Secretory Potential of Tested Models

Firstly, we checked how the oxygen condition affects the secretory potential in all three tested models.

All cell lines showed increased secretion of the main pro-angiogenic factor—VEGF-A in hypoxic cells compared to normoxic cells, which confirms the effectiveness of the in vitro hypoxia model used (Figure 4A). In cancer models (HCT 116 and HepG2), increased secretion of VEGF-A in hypoxia was statistically significant, but no changes were observed at the transcript level (Figure 4B). In turn, healthy liver cells showed tendencies to increase VEGF-A secretion in hypoxia, but at the same time there was a strong increase in the expression of the Vegfa gene (Figure 4B). In tumor models a significant decrease in TIMP2 secretion in hypoxia compared to normoxia was observed (Figure 4C). The decrease in TIMP2 secretion was not observed in healthy liver cells in hypoxia. Low oxygen caused a tendency towards decreased secretion of Il-8 (CXCL8) in cancer models, while healthy liver cells did not produce Il-8 at detectable levels (Figure 4D). Strong upregulation of PAI-1 secretion in hypoxia was observed in all tested models (Figure 4E). The tendency towards Osteopontin deregulation in low oxygen tension was observed only in healthy liver cells (Figure 4F). Hypoxia did not affect CD44 secretion and Il-6 (Supplementary Figure S1).

2.6. Effect of Radiation on Secretory Potential of Tested Models

Additionally, we investigated how the radiation may influence the secretory potential of tested models.

As an effect of microsphere therapy, a dose-dependent increase in VEGF-A secretion was observed in HCT 116 cells in normoxic conditions (Figure 5A). Due to the different dynamic OD changes in VEGF-A production in healthy liver cells, we observed a strong tendency to increase Vegfa expression in response to treatment in normoxia (Figure 5B). Liver cancer cells (HepG2) in response to microspheres in normoxia did not produce more VEGF-A (Figure 5A,B), but we observed a tendency to upregulate Il-8 and TIMP2 production after microspheres treatment in normoxia (Figure 5C,D). In the remaining models, we observed no effects of microspheres on the secretion of TIMP2 and Il-8. Higher doses of microspheres upregulate PAI-1 secretion in both cancer models, but not in healthy liver cells (Figure 5E). The secretion of CD44, OPN and Il-6 is not dependent on radiation treatment in normoxic conditions (Supplementary Materials, Figure S2A–C).

In hypoxic conditions, cells were not as sensitive to microsphere treatment in terms of changes in the secretory potential. VEGF-A (in both transcript and protein level) and TIMP2 were stable between treatment conditions (Figure 6A–C). In hypoxia we observed tendency to increase Il-8 secretion after exposure to higher doses of radiation in liver cancer cells (Figure 6D). In turn, in colon cancer models in hypoxia, microsphere treatment caused upregulation of PAI-1 secretion (Figure 6E). The secretion of CD44, OPN and Il-6 was not dependent on the radiation treatment in hypoxic conditions (Supplementary Materials, Figure S3A–C).

3. Discussion

Despite its widespread clinical use, ^90^Y radioembolization has been implemented with limited research into the effects of beta radiation on human liver and cancer cells. Findings from external beam radiation therapy studies cannot fully illustrate the processes involved in ^90^Y radioembolization, which functions as a form of brachytherapy employing beta radiation. Our current study elaborated on the effects of ^90^Y-derived beta radiation on human colorectal and human liver cancer cells, that we initiated previously [10]. This study aimed to investigate the influence of microspheres emitting low doses of beta radiation on colorectal, liver cancer and normal liver cell growth under diverse oxygenation conditions, reflecting the tumor microenvironment and the normal liver in the tumor’s vicinity. Although our primary focus was the mechanistic evaluation of low-dose beta radiation effects, we aimed to achieve clinically relevant radioembolization dose ranges (1X vs. 3X concentration of microspheres) in order to best align dose equivalence with patient treatment. The selection of two microsphere dilutions (1X and 3X) enables a dose-dependent evaluation of beta radiation effects. Rather than relying on nominal microsphere concentration alone, varying the dilution allows assessment of how incremental increases in delivered radiation dose influence cancer cell responses, including viability, proliferation, apoptosis, and stress signaling. This approach better reflects the heterogeneous microsphere distribution observed in vivo, where tumors are exposed to regions of lower and higher microsphere densities. Furthermore, comparing 1X and 3X dilutions facilitate the identification of threshold or saturation effects, distinguishing sublethal from cytotoxic radiation doses and providing insight into mechanisms of radiation sensitivity and resistance. Overall, this design supports a more biologically and clinically relevant characterization of dose–response relationships for beta radiation in cancer cell lines. We found that colon cancer hypoxia alone yielded considerable cell viability and proliferation reduction compared to control culture. A similar trend was observed in the other cell lines tested, but these changes were not statistically significant. Following irradiation, metabolic activity and cell proliferation in normoxic cells decreased significantly in a dose-dependent manner. In contrast, the effect in hypoxic conditions was less pronounced, with substantial changes observed only at higher doses (a 50% reduction compared to normoxic conditions, with a 21% reduction in low oxygen tension). It is worth emphasizing that the significantly lower effect of microspheres in hypoxia may be due to reduced metabolic activity and cell proliferation in hypoxia compared to normoxia (untreated, control cells), but also potentially the activation of survival mechanisms (like bcl2 pathway). This demonstrates the validity of conducting cytotoxicity studies in hypoxic conditions, which better reflect the tumor microenvironment. Moreover, these observations confirm the usefulness of the proposed in vitro microsphere research model. Low oxygen tension may lead to cell phenotype change and selection of radiation-resistant population, which provides additional evidence to earlier report [10]. Protective effect of hypoxia to irradiation was found to be reflected by stable expression of Bax—gene, encoding a pro-apoptotic protein Bax, which was strongly upregulated in normoxia after exposure to radiation. Similar responses were observed in other tested cell lines, including healthy hepatocytes, showing general cellular reaction to stress posed by low oxygen and radiation.

Additionally anti-apoptotic gene Bcl2 could be detected in colon cancer cells, and its expression was upregulated by hypoxia. The activation and expression of Bcl-2 in colon cancer are critical factors influencing tumor behavior and patient prognosis. Overexpression of Bcl2 has been linked to resistance in HCC treatments [11]. For instance, a study found that Bcl-2 confers protection to human hepatoma cells against Fas-mediated apoptosis, suggesting its role in therapy resistance, yet not all hepatoma cells rely on Bcl2 expression (HepG2 vs. HCC-T) as we confirm in our observations [12]. While high Bcl-2 expression has been associated with favorable outcomes in certain populations, its role in inhibiting apoptosis makes it a potential target for therapeutic intervention [13,14]. Our data suggests that hypoxia preconditions cancer cells to an anti-apoptotic bax/bcl2 phenotype that strongly affects their response to radiation.

Therapeutic targeting is utilized in the treatment of cancer by influencing the tumor microenvironment (TME), which is regulated by a complex network of molecules that interact to influence tumor growth, progression, and response to therapy [15]. One of these for both HCC and colon cancer liver metastases is VEGF. VEGF plays a significant role in both colon cancer and primary liver cancer (hepatocellular carcinoma, HCC) by promoting angiogenesis, tumor growth, and metastasis [16,17,18,19]. In our cancer models (HCT 116 and HepG2), hypoxia significantly increased the concentration of secreted VEGF-A, while transcript levels remained unchanged. In contrast, control cells from a healthy liver exhibited a trend toward increased VEGF-A secretion under hypoxia), accompanied by a strong upregulation of Vegfa gene expression. This discrepancy suggests distinct molecular response dynamics between cancerous and healthy cells. Beta radiation induced a dose-dependent increase in VEGF-A secretion under normoxic conditions in HCT 116 cells. However, in hypoxia, VEGF-A levels remained consistently high, regardless of the absorbed dose, in both colon cancer and healthy liver cells. This shows that radiation is a weaker angiogenesis stimulator than hypoxia. In contrast, HCC cells showed no response at either the transcript or protein secretion level. Performed analyses showed alternative angiogenesis pathway could be employed, namely inhibition of TIMP-2 and/or induction of IL-8.

TIMP2 is a soluble inhibitor of matrix metalloproteinases, which inhibits MMP-independent angiogenesis [20]. Ka-Lun Kai et al. found reduced TIMP2 expression was observed in 41.8% of HCC samples and was significantly linked to enhanced liver invasion and poorer patient survival outcomes [21]. In HCC cell lines, stable TIMP2 silencing led to increased invasiveness and extracellular matrix degradation, along with the emergence of invadopodia-like structures, indicating that TIMP2 serves as a negative regulator of HCC metastasis [21]. Under hypoxic conditions, TIMP2 suppression is driven by a regulatory feedback loop involving hypoxia-inducible factor 1 alpha (HIF-1α), microRNA-210 (miR-210), and hypoxia-inducible factor 3 alpha (HIF-3α). In our tumor models, a significant decrease in TIMP2 secretion in hypoxia compared to normoxia was observed. However, in the liver cancer model, the action of microspheres in normoxia tended to increase its level.

We found that HepG2 cells tended to secrete IL-8 in response to microsphere treatment. This upregulation is part of the cellular stress and inflammatory response to radiation damage, and the interplay between IL-8 and irradiation highlights a complex response where radiation not only damages tumor cells but also activates pathways that may help tumors recover and resist further treatment. IL-8 plays a pivotal role in promoting tumor angiogenesis and reshaping the immune microenvironment by attracting a greater number of immunosuppressive cells. In preclinical models using IL-8 transgenic mice for both murine and human colorectal cancer elevated IL-8 levels were linked to accelerated tumor growth through a significant increase in CD31-positive blood vessels around the tumor [22].

PAI-1 (plasminogen activator inhibitor-1) regulates proteolysis by inhibiting enzymes that break down the extracellular matrix, thereby influencing tumor invasion and metastasis. The PAI-1 signaling network is notably intricate, involving a variety of signaling molecules and complex interaction points and recently, research has uncovered a novel aspect of PAI-1’s function, highlighting its unexpected role in promoting cellular senescence [23,24]. As we found previously, cells exposed to radiation showed typical signs of senescence. These were especially conspicuous in hypoxia, namely cells were larger and polyploidy was more prevalent than in those grown in more oxygen-rich conditions [10]. In the current study, PAI-1 secretion increased in all cell lines in response to hypoxia, while tumor lines showed a rise after irradiation. Premature senescence can be triggered by a range of stress factors, such as oxidative stress, DNA damage, hypoxia, or treatments like chemotherapy and radiotherapy [25,26,27]. When senescence is induced during chemotherapy, it can result in cancer cells developing resistance, potentially leading to tumor regrowth [28,29]. Although these senescent cells cease to proliferate and become resistant to apoptosis, they remain metabolically active. Moreover, the senescence-associated secretory phenotype (SASP) comprises an assortment of growth factors, cytokines, and enzymes—such as IL-1, IL-6, IL-8, and metalloproteinase-3—that can influence inflammation, tumor progression, and the development of metastases [29].

In cases where tumors fail to receive an adequate absorbed dose during radioembolization, cellular senescence may be triggered. These senescent cells could potentially serve as the origin for new, therapy-resistant cancer cell populations—a phenomenon that has also been observed in cancer cells rendered senescent by chemotherapy [25,27,29,30].

To ensure that liver tumors receive an adequate cumulative absorbed dose, it is essential to minimize blood flow stasis caused by the embolic material. This can be achieved by delivering a higher radiation dose while reducing the number of spheres used. Additionally, a personalized strategy could combine radioembolization with targeted inhibition of VEGF, Bcl-2, or TIMP-2 signaling pathways to mitigate the hypoxic effects induced by embolization and thereby enhance treatment effectiveness.

Interleukin-6 (IL-6) is a central cytokine in hepatic inflammation and the acute-phase response. Following radioembolization, beta radiation can induce cellular stress and damage in both tumor and surrounding liver tissue, leading to IL-6 release from hepatocytes, Kupffer cells, and hepatic stellate cells. Elevated IL-6 signaling may promote inflammatory liver injury, contributing to radiation-induced liver disease (RILD) through activation of STAT3 and downstream pro-inflammatory pathways, enhance fibrogenic responses by activating hepatic stellate cells, potentially worsening long-term liver dysfunction, and disrupt normal liver regeneration, as excessive or prolonged IL-6 signaling may shift from a protective, regenerative role to a pathogenic one under high radiation stress [31,32,33,34,35,36]. No detectable IL-6 secretion was observed in tumor cells under hypoxia or after radiation exposure. In contrast, healthy liver cells secreted measurable levels of IL-6. While irradiation was associated with increased IL-6 levels, this effect did not reach statistical significance in our experiments (Supplemental Figures S2 and S3).

While IL-6 can play a protective role in controlled regeneration, dysregulated or sustained IL-6 production after radioembolization may amplify hepatic toxicity, particularly in patients with limited liver reserve or pre-existing liver disease. Consequently, understanding IL-6 dynamics may help refine dose individualization strategies and identify patients at higher risk of treatment-related liver injury.

We recognize that our study is not without its limitations. We could not replicate the 3-dimensional multi-cellular nature of tumor with 2-dimentional cell cultures, yet we simulated tumor microenvironment with various oxygenation conditions. TME secretory and gene encoding potential were evaluated with a tailored immunoassay and qRT-PCR. Although a wide spectrum of secretions modulating TME were evaluated, it did not exhaust its potential. Due to small amounts of collected material and limited numbers of tests possible, we focused on the pathways that we measured directly, to avoid speculation. Further, mechanistic studies should be performed, to better understand the background of our observations. Lastly, the intrinsic physical properties of the radiated microspheres prevented their uniform distribution among the wells. Nevertheless, this limitation partly mirrors the in vivo conditions encountered during radioembolization.

4. Materials and Methods

4.1. Cell Lines

The human colon cancer HCT 116 cell line, was cultured in McCoy’s 5A (Modified) Medium with 10% Fetal Bovine Serum (FBS) (both form Thermo Fisher Scientific, Waltham, MA, USA). The human liver cancer cell line HepG2 was cultured in RPMI (Thermo Fisher Scientific) with 10% FBS. Non-malignant liver cell line THLE-2 was cultured in BEGM™ Bronchial Epithelial Cell Growth Medium BulletKit™ (Lonza, Basel, Switzerland) containing BEBM Basal Medium and supplements: 10% FBS, hydrocortisone, epidermal growth factor (EGF), insulin, triiodothyronine, transferrin, retinoic acid, 6 ng/mL recombinant EGF and 80 ng ml/1 o-phosphorylethanolamine (Sigma Aldrich, St. Louis, MO, USA). All cell lines were obtained from the American Type Culture Collection (ATCC). Cells were passaged at 80% confluence by detaching with Trypsin (0.25%) EDTA solution (VWR International, Radnor, PA, USA). Cells were mycoplasma free, as assayed by PCR Mycoplasma Test (PromoCell, Heidelberg, Germany). All experiments were performed in standard culture condition—normoxia (~19% O_2_, 5% CO_2_, 37 °C) or hypoxia (1% O_2_ 5% CO_2_, 37 °C) using XVivo X3 workstation (Biospherix, Parish, NY, USA). Hypoxic conditions reflect the tumor microenvironment, where due to pathological angiogenesis, oxygen supply is insufficient. Because of the heterogeneity of the tumor microenvironment in terms of oxygen delivery, hypoxic conditions in experiments were set at 1% pO2. This reflects data presented in the review by Muz et al. [30]. in various types of cancer, where the range varies from 0.3% for pancreatic cancer to 2.2% in non-small cell lung cancer.

4.2. Experiment Protocol

Cells were seeded in 96-well plates in appropriate density (HCT116—2000 cells/cm^2^, HepG2 9500 cells/cm^2^, THLE-2 6250 cells/cm^2^) for 24 h in normoxic conditions. For each biological replicate, 6 wells were technical replicates. Next, according to the protocol (Figure 7), the medium was changed to normoxic or hypoxic. Normoxic medium was preincubated in standard incubator (~19% O_2_) for 24 h before use. Appropriately hypoxic medium was preincubated in incubator of XVivo X3 workstation (1% O_2_) for 24 h before use. The medium change was performed under hypoxic conditions (in the working chamber of XVivo X3 workstation) to avoid reoxygenation. Then cells were cultured for 48 h in normoxic (~19% O_2_) or hypoxic conditions (1% O_2_). Next the medium was changed again (again, the medium for the change was preincubated for 24 h under the appropriate conditions, as described above) and ^90^Y-micropsheres in 10% volume of medium were added for another 72 h. Two dilutions (1X and 3X) of microspheres were used to investigate the relationship between radiation dose and biological effects. A FLUKA-based (FLUKA Version 4-5.1) numerical model was constructed separately for each cell type (HCT 116, HepG2, and THLE-2) using the experimentally determined microsphere distributions and microscopic images [37,38]. The activity of an individual microsphere was determined from its diameter and a nominal specific activity of 9.15 × 10^6^ kBq/g, as in the earlier works [10,39].

The simulations yielded the normalized absorbed dose in the medium (culture medium plus cell culture). For each case, the number of decays occurring over 72 h and the corresponding absolute absorbed dose were determined. After incubation with radiating microspheres, microscopic observations were made, Alamar Blue and BrdU assays were performed and biological material (conditioned medium and RNA) was collected and frozen for further analysis.

4.3. Alamar Blue

Alamar Blue assay was performed to check cell viability based on mitochondrial activity. A solution of 10% Alamar Blue dye was added to the wells of a 96-well plate in cell line-appropriate culture medium according to the manufacturer’s instructions. After 3 h of incubation fluorescence (excitation 540 nm, emission 590 nm) was measured using VarioScan Lux, ThermoFisher Scientific, Warsaw, Poland) plate reader. Results are presented as a fold change of % reduction in Alamar Blue compared to normoxic control cells.

4.4. BrdU

BrdU incorporation assay (Sigma-Aldrich, Saint Louis, MO, USA) was used to assess cell proliferation based on DNA synthesis. Approximately 2 h before the end of culture, BrdU was added to the wells and its incorporation was assessed according to the manufacturer’s protocol. Absorbance was measured at a wavelength of 450 nm using spectrophotometric reader (VarioScan Lux, ThermoFisher Scientific, Warsaw, Poland). Results are presented as a fold change in cell proliferation compared to normoxic control cells.

4.5. Secreted Factors

Secretory potential was determined by bead-based immunoassay LEGENDPlex which allows the detection of multiple factors in one sample (Custom Panel, Biolegend, San Diego, CA, USA). For samples requiring greater optimization of preparation, results were performed using an enzyme-linked immunosorbent assay (ELISA).

TIMP2, CXCL8, CD44, OPN, PAI-1, were measured by Legenplex. Briefly, conditioned medium samples (pooled form 6 wells of technical replicates) were incubated with antibody-coated beads for 2 in RT with shaking. Then, washed twice and incubated with the detection antibodies for 1 h in RT with shaking. Detection was performed by SA-PE incubation for 30 min in RT with shaking, then analyzed by flow cytometry using CYTOFLEX software v.2.3.0.84 (Beckman Coulter, Brea, CA, USA) according to manufacturer’s instructions. Factors concentration was calculated against standard curve using recombinant proteins provided in the kit. Results are shown as fold change compared to the normoxic control with recalculation to the appropriate number of cells.

VEGF secretion was checked ELISA (DuoSet ELISA, R&D System, Minneapolis, MN, USA). Additionally, due to exceeding the range of the standard curve in bead-based immunoassay LEGENDPlex PAI-1 (for all cell lines) and OPN (for HepG2) and Il-6 were re-assayed by ELISA (DuoSet ELISA, R&D System, Minneapolis, MN, USA). HepG2 conditioned medium samples were diluted 5x to determine the concentration of secreted OPN. All the ELISA enzyme immunoassays were performed according to the manufacturer’s protocol. A 100 µL sample was tested on primary antibody coated and blocked 96-well plates. Detection was performed with the use of HRP substrate, TMB. Optical density was detected at plate spectrophotometric reader (VarioScan Lux, ThermoFisher Scientific, Warsaw, Poland) at 450 nm. Standard curve for recombinant PAI-1, OPN and VEGF and blanked absorbance measurements were performed.

4.6. Gene Expression

TaqMan™ Fast Advanced Cells-to-CT™ Kit was used for RNA isolation. RNA was obtained from cells pooled from 6 technical replicates from 96-well plates for each condition. Cells were detached with Trypsin (0.25%) EDTA solution (VWR International, Radnor, PA, USA), washed with PBS, centrifuged 300× g 5 min and 50 µL of Lysis Solution, containing Dnase I, was added. After 5 min, Stop Solution was added to stop the reaction of removing genomic DNA contamination and the cells were incubated for another 2 min at RT. After two hours keeping on ice, the lysates were centrifuged in order to remove the radiating microspheres and protect the collected material. Lysates were stored in −80 deg. Celsius.

A reverse transcription reaction was performed using 22.5 µL of RNA lysate, 2X Fast Advanced RT Buffer and 20X Fast Advanced RT Enzyme Mix according to the manufacturer’s instructions.

qRT-PCR reactions were performed using TaqMan^®^ Fast Advanced Master Mix or with TaqMan^®^ or probes (Bcl2 Hs00608023; Bax Hs00180269; Vegfa HS0090055; Gapdh HS02786624). Reactions were run on Bio-Rad CFX384 qPCR System (BioRad, Hercules, CA, USA), according to the manufacturer’s protocol. The relative mRNA levels were calculated with 2(-Delta C(T)) method, with normalization to the expression of Gapdh reference gene.

4.7. Statistical Analysis

Each in vitro experiment was performed at least 3 times in independent biological replicates. The results are shown as a mean +/− SEM from all biological replicates (the result of one biological replicate is the mean of 6 technical replicates). Where appropriate results are presented as fold change as compared to normoxia. All statistical analyses were performed based on all biological replicates, using GraphPad Prism 9.0 software. The statistical test used is described in the figure’s captions.

5. Conclusions

In conclusion, hypoxia stands as a hallmark of solid tumors, profoundly influencing their aggressive and malignant behavior. Radiating microspheres exhibit both cytostatic and cytotoxic effects on tumor and healthy cells, with cellular sensitivity largely determined by the tissue of origin; for example, in colon cancer, hypoxia induced radioresistance through Bcl-2 activation, leading to a diminished cytotoxic effect. In liver cancer cells, resistance to radiotherapy was even more pronounced, with only a cytostatic effect observed; hypoxia further reduced their responsiveness to treatment. Conversely, in healthy liver cells, radiation inhibited proliferation and induced apoptosis, although hypoxia appeared to confer a protective effect.

Both low oxygen levels and microsphere exposure triggered a pro-angiogenic response. In radio-sensitive colon cancer cells and healthy liver cells, this response was mediated by the induction of VEGF. In contrast, in the more resistant liver cancer cells, angiogenesis was likely promoted through TIMP2 inhibition and/or IL-8 induction.

Overall, hypoxia and radiation treatments produced similar stress responses, including the induction of apoptosis, reduced proliferation, and a pro-angiogenic shift. Notably, hypoxia primarily decreases cellular sensitivity to radiation while enhancing a microenvironment that promotes cancer progression.