Repurposing Sulfasalazine as a Radiosensitizer in Hypoxic Human Colorectal Cancer

Simple Summary Radiotherapy is a cornerstone for the treatment of colorectal cancer. Tumor cells present in an environment lacking oxygen (hypoxia) are resistant, leading to patient relapse. Altering redox homeostasis and inducing cell death of hypoxic cancer cells is a promising strategy to overcome radioresistance. In this study, redox homeostasis was targeted, and cell death (ferroptosis) was induced in colorectal cancer cells by treating them with the FDA-approved drug sulfasalazine. Overall, sulfasalazine treatment improved the response to radiotherapy in a model system of human colorectal cancer cells. Abstract xCT overexpression in cancer cells has been linked to tumor growth, metastasis and treatment resistance. Sulfasalazine (SSZ), an FDA-approved drug for the treatment of rheumatoid sarthritis, and inflammatory bowel diseases, has anticancer properties via inhibition of xCT, leading to the disruption of redox homeostasis. Since reactive oxygen species (ROS) are pivotal for the efficacy of radiotherapy (RT), elevated levels of ROS are associated with improved RT outcomes. In this study, the influence of SSZ treatment on the radiosensitivity of human colorectal cancer (CRC) cells was investigated. Our principal finding in human HCT116 and DLD-1 cells was that SSZ enhances the radiosensitivity of hypoxic CRC cells but does not alter the intrinsic radiosensitivity. The radiosensitizing effect was attributed to the depletion of glutathione and thioredoxin reductase levels. In turn, the reduction leads to excessive levels of ROS, increased DNA damage, and ferroptosis induction. Confirmation of these findings was performed in 3D models and in DLD-1 xenografts. Taken together, this study is a stepping stone for applying SSZ as a radiosensitizer in the clinic and confirms that xCT in cancer cells is a valid radiobiological target.


Introduction
Sulfasalazine (SSZ), an FDA-approved drug for the treatment of rheumatoid arthritis and inflammatory bowel diseases (IBD), is a potent xCT inhibitor and has attracted a lot of interest to be repurposed as an anticancer drug. SSZ already showed anticancer effects in a range of different tumor types, including colorectal cancer (CRC) [1].
There are two main mechanisms that are widely accepted to drive the observed anticancer effects of SSZ. The most reported mechanism is a disturbance of redox homeostasis. High levels of reactive oxygen species (ROS) are responsible for oxidative damage leading to cell death. Thus, it is to the cancer cell's advantage to retain the intracellular ROS levels under a non-toxic threshold. This results in a tight balance between ROS and antioxidant

Superoxide Dismutase Assay
The activity of superoxide dismutase (SOD) was determined by a commercial kit (Sigma-Aldrich, Antwerp, Belgium, 19160). In short, cells were lysed with a Triton-X lysis buffer supplemented with β-mercaptoethanol (5 mM) and protease inhibitor. Cells were centrifuged at 14,000× g, and the supernatant was collected. Next, 20 µL of the supernatant was mixed with a 200 µL assay cocktail and incubated at 37 • C for 20 min. Afterward, the absorbance was measured at a wavelength of 450 nm.

NAD(P)H Levels
The intracellular levels of NAD(P)H were determined by using a commercial kit (AAT Bioquest, Pleasanton, CA, USA, 15291) following the manufacturer's instructions, as reported elsewhere [21]. Shortly, cells were treated overnight with SSZ. Cells were stained in 250 µL serum-free medium containing 0.5 µL JZL1707 NAD(P)H sensor dye at 37 • C for 1 h. Afterward, cells were analyzed by flow cytometry (FACSCanto, BD Bioscience, Erembodegem, Belgium).

ROS Levels
The quantity of intracellular ROS was determined after SSZ treatment with the ROS assay (Abcam, Cambridge, UK, ab113851), according to the manufacturer's instructions. Briefly, staining with 5 µM CM-H2DCFDA for 30 min at 37 • C was executed. Stained samples were analyzed by flow cytometry. Incubation of spheroids with 10 µM CM-H2DCFDA for 30 min at 37 • C was carried out, and pictures were taken with the Incucyte ZOOM at a magnitude of 10X (Essen Bioscience, Royston, UK).

DNA Damage Analysis
The extent of DNA damage was determined after irradiation with 8 Gy by staining for γ-H2AX, as described elsewhere [20]. Briefly, cells were treated with SSZ and irradiated. Next, 1 h after irradiation, cells were fixed and permeabilized. Afterward, cells were incubated for 40 min at 4 • C with 0.1 µg γ-H2AX antibody (Abcam, Cambridge, UK, ab228655) and analyzed by flow cytometry.

Ferroptosis Levels
The levels of ferroptosis were determined by using C11BODIPY (ThermoFisher Scientific, Merelbeke, Belgium, D3861). Briefly, cells were incubated with 2 µM C11BODIPY for 15 min at 37 • C and measured by flow cytometry. Spheroids were stained with 4 µM C11BODIPY for 15 min at 37 • C, and pictures were taken with the Incucyte ZOOM at a magnitude of 10X (Essen Bioscience, Royston, UK).

Mitochondrial Membrane Potential
A commercial kit was used to discover the influence of SSZ on the mitochondrial membrane potential (Abcam, Cambridge, UK, ab113852) following the manufacturer's instructions. Shortly after overnight treatment of the cells with SSZ, staining with 400 nM TMRE dye for 30 min at 37 • C was conducted. Stained samples were analyzed by the use of flow cytometry.

Seahorse Metabolic Profiling
The Seahorse XF96 analyzer (Agilent Technologies, Santa Clara, CA, USA) was consulted for determining the extracellular acidification rate (ECAR) and oxygen consumption rate (OCR). The assay started by seeding 2.5 × 10 4 cells in 96-well plates. Treatment with SSZ happened overnight. Afterward, cells got equilibrated with unbuffered Dulbecco's Modified Eagle Medium (DMEM) supplemented with 2 mM glutamine and 10 mM glucose at pH 7.4. Cells were put in a 37 • C and CO 2 -free incubator. After this incubation period, cells were put in the Seahorse analyzer, and specific inhibitors were sequentially added: Oligomycin (1 µM), FCCP (1 µM), and rotenone/Antimycin A (0.5 µM). OCR levels were normalized to the protein content of each condition.

Three-Dimensional (3D) Cell Cultures
Spheroids were generated by seeding 5000 cells of DLD-1 and 4000 cells of HCT116 in 96-well ultra-low attachment plates. Overnight treatment with SSZ was initiated when the spheroids reached a diameter of ±500 µm. The next day, spheroids were irradiated with 8 Gy, and the treatment was removed and replaced with fresh medium. The medium was changed every 3 days, and the growth was followed up by EVOS for 11 days.

In Vivo Tumor Xenograft Model
The Ethical Committee of the Vrije Universiteit Brussel granted approval for the in vivo experiments (VUB-22_552_5). The animal experiment was conducted in accordance with National care regulations. DLD-1 cells (6 × 10 6 ) were subcutaneously inoculated into female 6-week-old athymic nude mice (Charles River Laboratories, Miserey, France). After inoculation, DLD-1 xenografts were irradiated as described elsewhere [22]. In short, SSZ was diluted in 0.1 M NaOH and PBS, with an adjusted pH of 7. SSZ (250 mg/kg) was administered intraperitoneally when the tumors reached a size of 200 ± 50 mm 3 for 5 consecutive days. On day 3 of SSZ treatment, mice were irradiated with a first fraction. Eventually, the mice were irradiated 3 times with 4 Gy using a Linac with an energy of 6 × FFF with a dose rate of 625 cGy/min. A unique CT scan of each mouse was made for situating the lesions and dosimetric determination. The entire duration of the experiment, tumor size and weight was checked with an electrical caliper. Tumor growth was calculated by measuring the tumor volume with the ensuing formula: (length × width 2 ) × 0.5.

Data Analysis and Statistics
Analysis of flow cytometric data was executed with FlowJo 7.6.5 Software. Incucyte pictures were analyzed by ImageJ. GraphPad Prism 9.4.0 was used for all the statistical analyses. Except otherwise stated, at least three independent experiments were performed. Data are presented as mean ± SEM. For statistical analyses, One-way ANOVA followed by a Dunnett's multiple comparison test and Two-way ANOVA followed by a Dunnett's or Bonferroni multiple comparison test was used. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

SLC7A11/xCT Expression Levels Are Linked to Mutational Status of CRC Patients and Correlated with Hypoxic Conditions
Firstly, the expression levels of SLC7A11 mRNA were investigated (gene encoding xCT) in patients suffering from CRC, extracted out of the TCGA database, and analyzed with the cBioPortal tool for Cancer Genomics. Upregulated levels of SLC7A11 were present in  Figure 1A). Additionally, the expression level of SLC7A11 in tumor tissue was compared to the expression in normal tissue. Tumors were divided into two separate groups, mRNA high and mRNA normal, and were linked to three different hypoxia scores. A significant correlation between high SLC7A11 mRNA levels and two hypoxia scores could be observed; both the Buffa ( Figure 1B) and Winter hypoxia score ( Figure 1C) correlated with upregulated SLC7A11. The third investigated score, namely the Ragnum hypoxia score, did not correlate with SLC7A11 mRNA levels ( Figure 1D). A possible explanation for this discrepancy could be that the Ragnum hypoxia score was developed specifically for prostate cancer and is not validated in other cancer types. Additionally, the Ragnum hypoxia score was limited to 32 genes, while the Buffa and Winter hypoxia scores included 51 and 99 genes, respectively, indicating more comprehensive hypoxia scores.   Data are shown as mean ± SEM, n > 3. An unpaired t-test was used in the case of a normal distribution (panel (A)), while a Mann Whitney test was used if normality was not present (panels (B-D)), or a two-way ANOVA followed by Bonferroni's test (panel (F)) was used for statistical analyses. ** p < 0.01, *** p < 0.001, **** p < 0.0001. The uncropped blots are shown in Figure S3.
Subsequently, we evaluated the expression of xCT in human MSI CRC cell lines, namely DLD-1 and HCT116 cells, under normoxic (21% O 2 ) and hypoxic (1% O 2 ) conditions. Our results demonstrated a clear xCT expression in both cell lines ( Figure 1E). Consistent with the data obtained from the cBioportal website, xCT expression was upregulated under hypoxic conditions compared to normoxic conditions (1.3-fold in DLD-1 cells and 1.2-fold in HCT116 cells) ( Figure 1F). This could indicate a higher dependence on system xC− within the hypoxic fraction of CRC cells, marking this cystine-glutamate antiporter as an interesting target. Additionally, preliminary data suggest that radiation downregulates the SLC7A11/ xCT expression levels. Similar results were obtained by Lang et al., in fibrosarcoma cells [13]. Altogether, these data support the use of CRC cell lines as a good model to investigate SSZ in combination with RT to alleviate hypoxic radioresistance.

SSZ Radiosensitized Hypoxic CRC Cells While Exerting No Effect on Normoxic CRC Cells
The toxicity profile of SSZ in DLD-1 and HCT116 cells was determined. SSZ decreased cell viability in a dose-dependent manner in both cell lines, independently of the oxygen levels. IC50 values of 3.2 mM and 3 mM were determined for DLD-1 and HCT116, respectively ( Figure S1A). Additionally, cell confluency was followed in real-time. Doses of SSZ up to 1 mM did not significantly decrease the growth rate of both cell lines ( Figure S1B). Since the primary aim of this study was to investigate radiosensitizing properties, doses lower than 1 mM (namely, 100 µM, 300 µM, and 500 µM) were used in subsequent experiments to avoid toxicity originating from SSZ alone.
The effect of SSZ on the radiosensitivity of CRC cell lines was examined under both normoxic and hypoxic conditions. Under normoxic conditions, SSZ exhibited no effect on the radiosensitivity of DLD-1 and HCT116 cells (Figure 2A,C). In contrast, under hypoxic conditions, a clear dose-dependent radiosensitizing effect could be observed in both cell lines, with enhancement ratios (ER) of 1.9 in DLD-1 cells and 1.6 in HCT116 cells ( Figure 2B,D) [23]. SSZ overcame hypoxic radioresistance but did not alter the intrinsic radioresistance of both cell lines. . Data are shown as mean ± SEM, n ≥ 3. A two-way ANO followed by Dunnett's multiple comparison test was used to calculate statistics. * p < 0.05, *** 0.001, **** p < 0.0001.

SSZ Inhibits Antioxidants GSH and TrxR, Inducing Disturbed Redox Homeostasis, Lead to Increased DNA Double-Strand Breaks under Hypoxic Conditions
SSZ-induced toxicity has been commonly linked to decreased levels of GSH in can cells, leading to aberrant levels of ROS [5]. Consequently, the influence of non-toxic do of SSZ on GSH levels was examined. SSZ significantly decreased GSH levels in a do dependent manner in both cell lines, with complete depletion of GSH in DLD-1 cells lowing treatment with 500 µM SSZ ( Figure 3A). Of note, the baseline GSH levels HCT116 cells (13.2 µM) were markedly higher compared to DLD-1 cells (6.1 µM), indi ing an intrinsic difference in balancing redox homeostasis between both cell lines. To ther investigate this, the influence of SSZ on other AOs was determined. It has been est lished that compensation mechanisms can occur to avoid lethal toxicity when one system is compromised [24]. Moreover, a clear difference in ER was present between b cell lines, and differences in AO activity are possibly accountable. A second AO enzy of interest was thioredoxin reductase (TrxR) since TrxR collaborates in functional red . Data are shown as mean ± SEM, n ≥ 3. A two-way ANOVA followed by Dunnett's multiple comparison test was used to calculate statistics. * p < 0.05, *** p < 0.001, **** p < 0.0001.

SSZ Inhibits Antioxidants GSH and TrxR, Inducing Disturbed Redox Homeostasis, Leading to Increased DNA Double-Strand Breaks under Hypoxic Conditions
SSZ-induced toxicity has been commonly linked to decreased levels of GSH in cancer cells, leading to aberrant levels of ROS [5]. Consequently, the influence of non-toxic doses of SSZ on GSH levels was examined. SSZ significantly decreased GSH levels in a dosedependent manner in both cell lines, with complete depletion of GSH in DLD-1 cells following treatment with 500 µM SSZ ( Figure 3A). Of note, the baseline GSH levels in HCT116 cells (13.2 µM) were markedly higher compared to DLD-1 cells (6.1 µM), indicating an intrinsic difference in balancing redox homeostasis between both cell lines. To further investigate this, the influence of SSZ on other AOs was determined. It has been established that compensation mechanisms can occur to avoid lethal toxicity when one AO system is compromised [24]. Moreover, a clear difference in ER was present between both cell lines, and differences in AO activity are possibly accountable. A second AO enzyme of interest was thioredoxin reductase (TrxR) since TrxR collaborates in functional redundancy with GSH [24]. A significant dose-dependent decrease in TrxR activity was observed in both cell lines after SSZ treatment, with again an increased depletion in DLD-1 cells compared to HCT116 cells ( Figure 3B). In continuation, a third AO enzyme was investigated, namely superoxide dismutase (SOD). No difference in SOD activity was observed after SSZ treatment ( Figure 3C). Lastly, we investigated the influence of SSZ treatment on NAD(P)H, the co-factor of GSH and TrxR. A dose-dependent increase in the amount of NAD(P)H was observed in DLD-1 cells, with a significant increase after treatment with 500 µM SSZ (1.5-fold) ( Figure 3D). However, this increase was not observed in HCT116 cells ( Figure 3D). Theoretically, increased levels of NADPH are present as a compensation mechanism to create more reduced Trx and GSH. Additionally, increased levels of NADH are associated with a dysregulated mitochondrial metabolism [25]. s 2023, 15, x FOR PEER REVIEW 10 dancy with GSH [24]. A significant dose-dependent decrease in TrxR activity wa served in both cell lines after SSZ treatment, with again an increased depletion in D cells compared to HCT116 cells ( Figure 3B). In continuation, a third AO enzyme w vestigated, namely superoxide dismutase (SOD). No difference in SOD activity wa served after SSZ treatment ( Figure 3C). Lastly, we investigated the influence of SSZ ment on NAD(P)H, the co-factor of GSH and TrxR. A dose-dependent increase i amount of NAD(P)H was observed in DLD-1 cells, with a significant increase after ment with 500 µM SSZ (1.5-fold) ( Figure 3D). However, this increase was not observ HCT116 cells ( Figure 3D). Theoretically, increased levels of NADPH are present as a pensation mechanism to create more reduced Trx and GSH. Additionally, increased l of NADH are associated with a dysregulated mitochondrial metabolism [25]. A one-way ANOVA performed, followed by Dunnett's multiple comparison test for statistical analyses. * p < 0.05, 0.01, *** p < 0.001, **** p < 0.0001.
The AO system tightly regulates the levels of ROS present within the cells; h altered levels of AO could disrupt redox homeostasis. As a logical next step, the lev ROS were determined. A dose-dependent increase in the amount of ROS was obse after SSZ treatment in both cell lines (4.1-fold in DLD-1 cells and 6.8-fold in HCT116 c reaching statistical significance from 300 µM ( Figure 4A). In order to evaluate wh these increased levels of ROS were responsible for the previously observed radiosen and HCT116 (bottom) cell lines. Data are shown as mean ± SEM, n = 3. A one-way ANOVA was performed, followed by Dunnett's multiple comparison test for statistical analyses. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.
The AO system tightly regulates the levels of ROS present within the cells; hence, altered levels of AO could disrupt redox homeostasis. As a logical next step, the levels of ROS were determined. A dose-dependent increase in the amount of ROS was observed after SSZ treatment in both cell lines (4.1-fold in DLD-1 cells and 6.8-fold in HCT116 cells), reaching statistical significance from 300 µM ( Figure 4A). In order to evaluate whether these increased levels of ROS were responsible for the previously observed radiosensitizing effect under hypoxic conditions ( Figure 2B,D), cells were treated with SSZ in combination with ROS scavenger NAC and irradiated with 8 Gy. NAC partially reversed the radiosensitizing effect after IR in both DLD-1 and HCT116 cells and induced significant changes after treatment with 300 µM and 500 µM SSZ in DLD-1 cells and 500 µM SSZ in HCT116 cells ( Figure 4B). A one-way ANOVA followed by Dunnett's ple comparison test or a two-way ANOVA followed by Dunnett's, or Bonferroni's multiple co ison test was performed for statistical analyses. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.00 Lastly, ROS are crucial for the generation of DNA damage following IR [2 Therefore, the amount of DNA double-strand breaks after SSZ treatment was determ SSZ dose-dependently increased the amount of DNA double-strand breaks (up to fold) in both DLD-1 and HCT116 cells, reaching statistical significance with the hi dose of SSZ (500 µM) ( Figure 4C). Since NAC only partly reverses the radiosensit effect of SSZ, induction of DNA double-strand breaks by increased ROS levels is no only mechanism responsible for the observed radiomodulatory properties of SSZ. Data are represented as mean ± SEM, n ≥ 3. A one-way ANOVA followed by Dunnett's multiple comparison test or a two-way ANOVA followed by Dunnett's, or Bonferroni's multiple comparison test was performed for statistical analyses. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

SSZ Causes Lipid Peroxidation Resulting in Ferroptosis in the DLD-1 Cell Line under Hypoxic Conditions
Lastly, ROS are crucial for the generation of DNA damage following IR [23,26]. Therefore, the amount of DNA double-strand breaks after SSZ treatment was determined. SSZ dose-dependently increased the amount of DNA double-strand breaks (up to 1.5-fold) in both DLD-1 and HCT116 cells, reaching statistical significance with the highest dose of SSZ (500 µM) ( Figure 4C). Since NAC only partly reverses the radiosensitizing effect of SSZ, induction of DNA double-strand breaks by increased ROS levels is not the only mechanism responsible for the observed radiomodulatory properties of SSZ.

SSZ Causes Lipid Peroxidation Resulting in Ferroptosis in the DLD-1 Cell Line under Hypoxic Conditions
Since SSZ has been described as a ferroptosis inducer [13], we evaluated the contribution of ferroptosis induction to the radiosensitizing effect. First, the amount of lipid peroxidation, a hallmark of ferroptosis, was determined. It became evident that SSZ increased the levels of lipid peroxidation in a dose-dependent manner, reaching significance starting from 300 µM in DLD-1 cells (up to 2.7-fold), while 500 µM was necessary for HCT116 cells (1.5-fold) ( Figure 5A). Recently, IR has been reported to be capable of inducing ferroptosis in cancer cells [9,12,13]. Hence, we investigated the amount of lipid peroxidation after a single high dose of IR (16Gy) in combination with SSZ treatment. IR alone was effective at increasing the amount of lipid peroxidation in both cell lines (4-fold in DLD-1 cells and 3-fold in HCT116 cells). However, the combination with SSZ further elevated the levels of lipid peroxidation in DLD-1 cells (7.6-fold increase at 500 µM), while no additional increase was observed in HCT116 cells ( Figure 5B). A two-way ANOVA followed by a Dunnett's or Bonferroni multiple comparison test, or a one-way ANOVA followed by a Dunnett's multiple comparison test, was performed for statistical analyses. * p < 0.05, ** p < 0.01.
Although the measurement of lipid peroxidation is mostly used to investigate ferroptosis, this type of cell death is also characterized by additional features, such as dysregulated mitochondria. To confirm the difference in ferroptosis induction between DLD-1 and HCT116 cells, we aimed to map the mitochondrial activity after SSZ treatment. The mitochondrial membrane potential (ΔΨm) was increased after SSZ treatment in DLD-1 cells (up to 1.7-fold), reaching significance with 500 µM ( Figure 5C). In line with the obtained results by measuring lipid peroxidation, no difference in ΔΨm was observed in HCT116 cells ( Figure 5C).
As a validation step, cells were treated with a ferroptosis inhibitor, Ferrostatin-1, and the impact on the radiosensitizing effect of SSZ was assessed. Ferrostatin-1 clearly re- represented as mean ± SEM, n ≥ 3. A two-way ANOVA followed by a Dunnett's or Bonferroni multiple comparison test, or a one-way ANOVA followed by a Dunnett's multiple comparison test, was performed for statistical analyses. * p < 0.05, ** p < 0.01.
Although the measurement of lipid peroxidation is mostly used to investigate ferroptosis, this type of cell death is also characterized by additional features, such as dysregulated mitochondria. To confirm the difference in ferroptosis induction between DLD-1 and HCT116 cells, we aimed to map the mitochondrial activity after SSZ treatment. The mitochondrial membrane potential (∆Ψm) was increased after SSZ treatment in DLD-1 cells (up to 1.7-fold), reaching significance with 500 µM ( Figure 5C). In line with the obtained results by measuring lipid peroxidation, no difference in ∆Ψm was observed in HCT116 cells ( Figure 5C).
As a validation step, cells were treated with a ferroptosis inhibitor, Ferrostatin-1, and the impact on the radiosensitizing effect of SSZ was assessed. Ferrostatin-1 clearly reversed the radiosensitizing effect of SSZ under hypoxic conditions in DLD-1 cells ( Figure 5D). No counteracting effect was observed in HCT116 cells, confirming the previously obtained results ( Figure 5D).
Considering the observed alterations in NADH levels and ∆Ψm after SSZ treatment in DLD-1 cells, we hypothesized that adaptations in mitochondrial metabolism could also increase radiosensitizing properties. Mitochondria are important consumers of oxygen, enhancing the hypoxic state of cancer cells and, therefore, an appealing target for radiosensitizing properties [27]. As a measure of metabolic activity, the oxygen consumption rate (OCR) of SSZ-treated cells was measured by a Seahorse analyzer. No changes in OCR or any other parameters were observed ( Figure S2A,B), indicating that the observed changes in mitochondrial membrane potential are likely due to the induction of ferroptosis and do not originate from an altered mitochondrial metabolism.

SSZ Radiosensitizes 3D-Colorectal Cancer Models
To confirm the obtained results, 3D-cancer models (spheroids) were generated. Spheroid growth was followed-up after treatment with SSZ alone (300 µM, 500 µM, and 1 mM) and in combination with 8Gy. Higher concentrations of SSZ were used in order to substantiate drug penetration into the core of the spheroid. No effects could be observed with non-toxic doses of SSZ, while a small growth delay was present after treatment with 8Gy in both DLD-1 and HCT116 spheroids. In combination with SSZ, this growth delay was more pronounced, and dose dependency could be observed ( Figure 6A,B).
We validated the mechanisms uncovered in the 2D models and could confirm a clear increase in the amount of ROS after SSZ treatment in DLD-1 and HCT116 spheroids ( Figure 6C,D) and an increase in lipid peroxidation in DLD-1 spheroids but not in HCT116 spheroids ( Figure 6E,F). These results are fully in line with the 2D results.

SSZ Radiosensitizes 3D-Colorectal Cancer Models
To confirm the obtained results, 3D-cancer models (spheroids) were generated. S roid growth was followed-up after treatment with SSZ alone (300 µM, 500 µM, and 1 and in combination with 8Gy. Higher concentrations of SSZ were used in order to stantiate drug penetration into the core of the spheroid. No effects could be observed non-toxic doses of SSZ, while a small growth delay was present after treatment with in both DLD-1 and HCT116 spheroids. In combination with SSZ, this growth delay more pronounced, and dose dependency could be observed ( Figure 6A,B).  . Data represented as mean ± SEM, n ≥ 3. A one-way ANOVA followed by Dunnett's multiple comparison test was performed for statistical analysis. * p < 0.05, ** p < 0.01, *** p < 0.001.

SSZ Radiosensitizes DLD-1 Tumor Xenografts
Additionally, the in vitro data were further validated in DLD-1 xenografts. Briefly, nude mice received five consecutive days of i.p. treatment with SSZ (250 mg/kg). On day three of the treatment, the first fraction of radiation was given to the tumor. In total, the DLD-1 xenografts were irradiated 3 times with 4Gy/fraction. Afterward, tumor growth was followed-up until the tumor volume exceeded 1500 mm 2 ( Figure 7A). Radiation alone induced a tumor growth delay of 13 days, while a single treatment of SSZ did not delay tumor growth. The combination therapy induced a significant tumor growth delay of 36 days ( Figure 7B). Consequently, the median survival of single-treated mice was extended from 49 days in the control group to 65 days in the irradiated group. In conclusion, adding SSZ to the irradiation led to a synergistic effect, with a prolongation in median survival of 89 days ( Figure 7C).
was followed-up until the tumor volume exceeded 1500 mm ( Figure 7A). Radiation alone induced a tumor growth delay of 13 days, while a single treatment of SSZ did not delay tumor growth. The combination therapy induced a significant tumor growth delay of 36 days ( Figure 7B). Consequently, the median survival of single-treated mice was extended from 49 days in the control group to 65 days in the irradiated group. In conclusion, adding SSZ to the irradiation led to a synergistic effect, with a prolongation in median survival of 89 days ( Figure 7C).

Discussion
CRC still accounts for 10% of cancer-related mortality worldwide, and the standard of care for locally advanced rectal cancer consists of RT followed by surgical resection. Despite continuous advances in radiotherapeutic treatment options, CRC patients often develop radioresistant tumors resulting in relapse and poor prognosis [28], urging the need to develop novel strategies.
In CRC, xCT upregulation has been associated with disease recurrence, venous invasion, and lymph node metastasis [29]. Therefore, the purpose of this study was to potentiate the efficacy of RT by targeting xCT with SSZ. SSZ, an xCT inhibitor and FDA-approved drug, was used to facilitate the translation toward a clinical setting in CRC patients. Since both intrarectal and systemic administration in these patients is feasible and well tolerated.
Our results demonstrated that SSZ enhances the radioresponse in human CRC cells under hypoxic conditions but not under normoxic conditions, as opposed to results published in glioblastoma and melanoma [30,31]. Tumor hypoxia is recognized as one of the principal causes of clinical radioresistance [23]. As a result, modification by hyperbaric oxygen and nitroimidazoles have been exploited to overcome hypoxic resistance. Nevertheless, every effort remained largely ineffective for clinical application [32]. Alternating pharmacokinetics are reported for SSZ with maximal achievable plasma levels between 16 and 300 µM [33,34], suggesting the validity of the acquired concentrations in this study. Recent studies showed elevated levels of ROS present under hypoxic conditions, leading to increased AO in order to cope with this phenomenon. The generation of these elevated AO levels is, among others, dependent on the presence and activity of xCT [35,36]. Pursuant to these findings, we identified a correlation between upregulated expression levels of SLC7A11 and higher hypoxia scores. Next to this, xCT was upregulated at the protein level when hypoxia was induced. xCT overexpression has been proven to be implicated in chemoresistance against both 5-fluorouracil and cisplatin in CRC. These chemotherapeu-tic agents are used as first-line treatment in CRC patients in combination with RT [1,37]. Therefore, SSZ treatment can potentially enhance more than solely the radiosensitivity of CRC cells.
It has previously been published that xCT inhibition decreases the levels of GSH and increases the levels of ROS in multiple cancer types, inducing anticancer effects [38]. In this study, we presented similar findings, with increased levels of ROS and decreased GSH levels in both cell lines. However, SSZ did not only decrease the levels of GSH but also the levels of the second most prominent AO, TrxR. Dual targeting of the redundant GSH and TrxR AO systems is perhaps necessary to fully disturb redox homeostasis and, consequently, enhance radioresponses [39,40].
It is well-known that under hypoxic conditions, hypoxia-inducible factors (HIF-1α and HIF-2α) become active. The induction of ferroptosis under hypoxic conditions sparked a lot of controversy in the literature. HIF-1α is described as a negative regulator of ferroptosis through the inhibition of PUFA synthesis and upregulation of SLC7A11. On the contrary, HIF-2α is described as a positive regulator of ferroptosis by stimulating the synthesis of PUFA, increasing the intracellular levels of ROS, and elevating the levels of iron transport within the cells [41]. In accordance with the findings of Singhal et al., our results confirmed that ferroptosis is present in CRC cells under hypoxic conditions. This demonstrates the vulnerability of hypoxic CRC cells to ferroptosis [42]. Further investigation into the expression of genes within the HIF family was beyond the scope of this project. Additionally, induction of ferroptosis after IR has been described in a variety of cell lines and patient samples [9,12,13], indicating the significance of ferroptosis within IR. This further corroborates the potential of combination therapy with SSZ, which could synergize at the level of cell death.
Important to mention is that cystine can be derived intracellularly via the transsulfuration pathway, independently of system xC-. However, this contribution is limited and expected to be insufficient to overcome xCT inhibition [43]. Next to the SLC7A11-GSH-GPX4 ferroptosis defense system, two other ferroptosis defense systems have been described, namely the ubiquinol pathway and the tetrahydrobiopterin pathway. Targeting these pathways is worth investigating alone or in combination with SSZ. Unfortunately, the exact contribution to ferroptosis resistance of these mechanisms is still unknown [44].
Our findings suggest that SSZ was more potent to increase responses to RT in the DLD-1 cell line compared to the HCT116 cell line. A clear difference in the reduction of both GSH and TrxR after SSZ treatment could be observed between DLD-1 and HCT116 cells, with a more substantiated decrease in DLD-1. Secondly, the observed differences in radioresponses could be linked to the individual activity of xCT on the different cell lines. xCT seems to be more active on the HCT116 cell line compared to the DLD-1 cell line, while both cell lines are MSI. We performed similar radiation experiments with SSZ in MSS CRC cell lines which resulted in an even greater reduction of the observed radiosensitizing effects. To prove this hypothesis, higher concentrations of SSZ are required to reach the same amount of potency in the HCT116 cell line. However, this could not be further examined in an in vitro setting due to the observed antiproliferative effects at higher SSZ doses.
Moreover, SSZ was potent at inducing ferroptosis in DLD-1 cells but not in HCT116 cells. Literature suggests that the sensitivity of CRC cells to ferroptosis can be associated with their p53 status. TP53 is described as negatively regulating ferroptosis in CRC cells, which is in sharp contrast with the function of p53 in other cancer types, where it promotes ferroptosis [45]. Our initial results are in line with this observation since DLD-1 is characterized by mutated p53, while HCT116 is characterized as wild-type P53 [46]. Although this finding should be further validated, the p53 status could be an interesting biomarker for selecting patients who would benefit from ferroptosis-stimulating treatments.
Lastly, our in vitro results were confirmed in human DLD-1 xenografts. Since this model deprived the natural tumor environment, possible immunogenic effects induced by SSZ were lacking. Further research should be performed since activated T-cells are shown to suppress xCT, hence adding to the observed effects of SSZ [13,47].
In this study, we have solely focused on the inhibition of xCT as the mechanism of action of SSZ in CRC. By extension, we consider the present study's finding highly relevant for patients with low and mid-rectal cancers. SSZ may be administered as an enema to these patients in an attempt to improve the complete clinical response during (total) neoadjuvant chemoradiation. Besides, it is important to keep in mind that SSZ also has anti-inflammatory effects, considering its daily use as a drug with anti-arthritic activity and its use in IBD. The exact anti-inflammatory working mechanism still needs to be elucidated; however, SSZ is assumed to inhibit prostaglandin synthesis [48]. Importantly, prostaglandin inhibition was already linked to radiosensitization three decades ago [49]. Within the clinic, a prime safety concern with the use of SSZ could be the induction of ROS to lethal levels within not only tumor cells but healthy cells as well. Multiple studies have suggested that healthy cells remain unaffected by xCT inhibition since the expression levels of this antiporter are significantly lower compared to tumor tissue and the presence of healthy and robust AO systems [30,31]. In contrast, preliminary experiments showed upregulation of ROS levels in human dermal fibroblasts upon SSZ treatment, albeit not to lethal levels with the concentration range used in this study. In order to bypass these possible toxicity issues in the clinic, local administration of SSZ in the rectum could be beneficial [50].

Conclusions
Altogether, we have demonstrated that SSZ radiosensitizes hypoxic CRC cells by decreasing the levels of GSH and TrxR, leading to elevated levels of ROS. These elevated ROS levels induced a higher amount of DNA double-strand breaks. Induction of ferroptosis could be beneficial for radiosensitization purposes; however, its contribution could vary between tumors of different genetic backgrounds. Consequently, since the pharmacokinetics and pharmacodynamics of SSZ are well established, the translation of these results into a clinical setting is optimistic.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cancers15082363/s1, Figure S1: SSZ induces a dose-dependent toxicity in human colorectal cancer cells; Figure S2: xCT inhibition does not influence the mitochondrial metabolism of human colorectal cancer cells.; Figure S3: Uncropped Western blots.

Institutional Review Board Statement:
The animal experiment was approved by the Ethical Committee of the Vrije Universiteit Brussel (VUB-22_552_5, approval date 24 June 2022) and was conducted in accordance with National care regulations. Informed Consent Statement: No patients were included or identiable data used in this study.
Data Availability Statement: Publicly available data generated by others were used by the authors (TCGA database).