Dihydroartemisinin-Transferrin Adducts Enhance TRAIL-Induced Apoptosis in Triple-Negative Breast Cancer in a P53-Independent and ROS-Dependent Manner

Triple-negative breast cancer (TNBC) is a highly aggressive breast cancer subtype independent of estrogen receptor, progesterone receptor, or human epidermal growth factor receptor 2. It has a poor prognosis and high recurrence. Due to its limited treatment options in the clinic, novel therapies are urgently needed. Single treatment with the death receptor ligand TRAIL was shown to be poorly effective. Recently, we have shown that artemisinin derivatives enhance TRAIL-induced apoptosis in colon cancer cells. Here, we utilized transferrin (TF) to enhance the effectiveness of dihydroartemisinin (DHA) in inducing cell death in TNBC cell lines (MDA-MB-231, MDA-MB-436, MDA-MB-468 and BT549). We found that the combination of DHA-TF and the death receptor 5-specific TRAIL variant DHER leads to an increase in DR5 expression in all four TNBC cell lines, while higher cytotoxicity was observed in MDA-MB-231, and MDA-MB-436. All the data point to the finding that DHA-TF stimulates cell death in TNBC cells, while the combination of DHA-TF with TRAIL variants will trigger more cell death in TRAIL-sensitive cells. Overall, DHA-TF in combination with TRAIL variants represents a potential novel combination therapy for triple-negative breast cancer.


INTRODUCTION
Breast cancer is the most prevalent malignant carcinoma in women who were diagnosed with cancer worldwide. Endocrine therapy targeting estrogen receptor or progesterone receptor is a welltolerated treatment for some breast cancer patients (1). However, about 15-20% of breast cancers are triple-negative breast cancer (TNBC), which is known as a subtype lacking estrogen receptor, progesterone receptor, or human epidermal growth factor receptor 2 (2,3). As a result, TNBC is resistant to endocrine therapy and shows high metastasis and poor prognosis with a low median survival time (4). Thus, there is an urgent demand for an effective, low-cost, and easy to apply therapy for TNBC.
It has been documented that artemisinin (ART) and its semisynthetic derivatives are effective in treating cancers (5)(6)(7), including breast carcinomas (8,9). They are sesquiterpene lactone compounds used as standard drugs in treating malaria. Compared with artemisinin and some derivatives, dihydroartemisinin (DHA) exhibits higher efficacy, solubility, and bioavailability as a therapeutic agent (10,11). In malaria, parasites digest hemoglobin resulting in an iron-rich internal environment essential for the activation of artemisinins (12). The low-valent transition irons cleave the endoperoxide bridge of artemisinins generating intraparasitic carbon-centered free radicals or reactive oxygen species (ROS) and ultimately kill the Plasmodium falciparum (13,14). Intriguingly, artemisinins also shows a specific cancerkilling nature due to the iron-rich environment present in proliferating cancer cells. However, compared to other breast cancers or other tumors in TNBC, artemisinins provide a relatively weak cell killing activity. However, its anti-cancer activity can be enhanced by forming drug-protein adducts with transferrin (TF) as was demonstrated in lung and liver cancer cells (15). Hence, it is meaningful to utilize both DHA and TF in TNBC cells and investigate the efficacy.
Human transferrin is a blood-plasma glycoprotein, which binds and transports iron in the human body to maintain iron homeostasis and metabolism. Iron-saturated holo-transferrin (holo-TF) transports irons into cells and releases irons in the acidic endosome to accomplish intracellular iron delivery (16). Using fluorescence spectra and UV-vis absorption assays, it was established that DHA binds to transferrin via hydrophobic and van der Waals interactions (15,17). The DHA-TF adducts are formed with high binding affinity in neutral environments, allowing transferrin to transport DHA into cells by transferrin receptor (TfR)-mediated endocytosis. Subsequently, the decreased binding capacity at acidic conditions in endosomes leads to the dissociation of DHA and iron from transferrin, which activates DHA to form ROS in cancer cells (15,18). Thus, it is of interest to evaluate the effect of DHA-TF adducts in TNBC cells for stimulating ROS production and inducing cell death.
It has been established that artemisinin's anti-cancer mechanism acts mainly through the cleavage of the endoperoxide bridge, which eventually induces cell ferroptosis, apoptosis, autophagy, cell cycle arrest, or anti-angiogenesis (19). ROS, which are highly reactive chemical molecules formed by losing one oxygen electron from the cleaved endoperoxide bridge, are essential to the above-mentioned physiological processes (20,21). There are multiple ROS species, including superoxide (O − 2 ), hydrogen peroxide (H 2 O 2 ), hydroxyl radicals (·OH), peroxide (O 2− 2 ), and hydroxyl ion (OH -). Many studies have shown that artemisinins produce ROS in parasites and cancer cells, while Egwu et al. reported specifically that superoxide is generated when parasites are treated with artemisinins (22). We, therefore, analyzed the production of both ROS and superoxide from DHA or DHA-TF within TNBC.
To further establish the efficacy of DHA-TF on TNBC, we became interested in combining it with tumor necrosis factorrelated apoptosis-inducing ligand (TRAIL). TRAIL-induced apoptosis is an irreversible programmed cell death initiating intrinsic pathway and/or extrinsic pathway (23). Trimerized TRAIL binds to the extracellular domain of death receptor 4 (DR4) or death receptor 5 (DR5) to stimulate caspase cascade activation, eventually leading to cell shrinkage, DNA fragmentation, and cell apoptosis (24). TRAIL is harmless to non-transformed cells and our previous research on colon cancer cells showed that DHA enhances TRAIL efficacy, especially the DR5-specific TRAIL variant DHER-induced apoptosis (25).
In this study we found that the treatment of DHA-TF adducts perform a better efficacy than individual DHA in TNBC. It also results in DR5 upregulation and thereby enhances DHER-induced apoptosis. Above all, we confirm that the DHA-TF combination with DHER shows a promising medication effect in TNBC cell lines making it a potential therapeutic for TNBC in the future.

The Formation of DHA-TF Adducts
Fluorescence emission spectra were recorded by a BioTek ™ Synergy ™ H1 Hybrid Multi-Mode Fluorescence Microplate Reader (Biotek, Winooski, VT, USA) using UV-STAR microplates (Greiner bio-one, Alphen a/d Rijn, Netherlands). Different concentrations of transferrin or DHA were diluted in PBS (pH 7.4). Fluorescence spectra were recorded from 300 nm to 400 nm at room temperature (RT) (lex=270 nm) (15,17). The DHA was added into TF solution and incubated for 30 min. to perform the measurements. For the continued analysis on cells, the DHA and TF were diluted in medium (without FBS) with a 1:1 molar ratio and incubate at RT for 30 min. to form adducts.

Caspase 3/7 Activity Assay
Cells were seeded into white-walled 96-well plates (Greiner bioone, Alphen a/d Rijn, The Netherlands) and cultured overnight. MDA-MB-231 and MDA-MB-436 were pretreated with DHA, DHA-TF for 30 min., followed with rhTRAIL WT, 4C7, or DHER for 12 h. The Caspase-Glo 3/7 Reagent (Promega Corporation, Madison, WI, USA) was equilibrated to RT, and 50 µL of the reagent was added to each well. The luminescence was collected after the plate was gently mixed and incubated at RT for 1 h.

Clonogenic Assay
The cells were seeded in 6-well plates at a density of 300,000 cells/ well overnight. They were treated with DHA, TF or DHA-TF for 24 h. Then, the treated cells were harvested and seeded in 6-well plates with 400 living cells/well and cultured for 12 days. Finally, the cells were washed with PBS and fixed with methanol for 5 min., then the air-dried plates were stained with Crystal Violet (Sigma-aldrich, St. Louis, USA) and the colonies were counted.

Apoptotic Assay
The cell apoptosis detection was performed using eBioscience ™ Annexin V-FITC/PI Apoptosis Detection Kit (Thermo Fisher Scientific, Carlsbad, CA, USA). After 20 h. treatment, cells were harvested and washed with the binding buffer from the kit. Following the manufacturer's instructions, cells were incubated with Annexin V-FITC and PI sequentially. Finally, all samples were analyzed by NovoCyte Quanteon Flow Cytometer (Agilent Technologies, California, USA), and 10,000 events were recorded for each analysis.

ROS Production Assay
The cells were treated with DHA, TF, or DHA-TF adducts for 20 h. Then, all washed and collected samples were incubated with Oxidative Stress Detection Reagent and Superoxide Detection Reagent from ROS-ID ® ROS/RNS detection kit (Enzo Life Sciences, Lausen, Switzerland) for 2 h. in the incubator at 37°C and 5% CO 2 . Then the cells were washed and suspended with Wash Buffer. Oxidative stress and Superoxide were detected using a BD FACSCanto ™ Flow Cytometry (BD Biosciences, Aalst, Belgium), and 10,000 cells were analyzed for each sample.

Western Blot Analysis
Cells were harvested after treated with/without DHA-TF for 24h. and western blot was carried out as previously described (25). GPX4 (R&D Systems, Minneapolis, MN) primary antibody was used. Anti-vinculin (Sigma Aldrich, St Louis, USA) was used for confirmation of equal loading of proteins. IRDye ® 680RD Goat anti-Mouse IgG secondary antibody was purchased from LI-COR Biosciences (Miami, FL).

Flow Cytometry of Death Receptors and Transferrin Receptor Expression
Cells were seeded in 6-well plates at a density of 300,000 cells/ well overnight. After 24 h. treatment, cells were harvested and washed with FACS buffer (PBS with 2% FBS). Anti-DR4 antibody (Abcam, Cambridge, UK), anti-DR5 antibody (Exbio, Praha, Czech Re-public) or anti-Transferrin antibody was immobilized with cells for 1 h. on ice. The Goat anti-Mouse IgG Alexa Fluor 633 (Thermo Fisher Scientific, Carlsbad, CA, USA) was used as secondary antibody and the mouse IgG Isotype (Dako, Glostrup, Denmark) performed as control.

Transferrin Enhances DHA Induced Cell Death in TNBC
The cell viability of selected TNBC cell lines was recorded after treatment with increasing concentrations of TF, DHA, or DHA-TF adducts by MTS assay ( Figure 1A). Individual treatment with TF shows no toxicity to all the cell lines (up to 50 µM), while DHA leads to dose-dependent cell death in the range tested (0 µM to 50 µM). Remarkably, the DHA-TF adducts significantly reduced metabolic activity in all cell lines compared with the single TF or DHA treatment. The formation of DHA-TF adducts was checked using fluorescence spectra and the addition of DHA successfully quenches the intrinsic fluorescence of TF at 310 nm in a dose-dependent manner ( Figures S1A, B). Both DHA and DHA-TF show limited impact at lower concentration (lower than 10 µM) in MDA-MB-231 and MDA-MB-436. These two cell lines are relatively resistant to the individual application of DHA (IC 50 >50 µM), but sensitive to DHA-TF adducts treatment (IC 50 = 31.68 and 23.46 µM, respectively). TF especially triggers MDA-MB-468 and BT549 to become more susceptible to DHA induced cell viability reduction (IC 50 of DHA-TF=4.49 and 5.86 µM, respectively). These cell lines also have a more pronounced response to individual DHA treatment ( Table 1).
This enhancement of cell death under DHA-TF treatment was also observed in clonogenic assays. Figure 1B shows that DHA-TF results in a significant reduction of colony numbers across all cell lines, which matches the MTS results. Moreover, colonies with DHA-TF challenges are generally smaller in size than untreated or other treated cells ( Figure 1B, left panel). The statistical analysis from three independent experiments confirmed that the effects of DHA-TF adducts in reducing colony numbers is reproducible ( Figure 1B

DHA-TF Adducts Generates ROS and Superoxide in TNBC
Artemisinins have been reported to generate ROS production in cancer cells and parasites. We performed the ROS detection assay in TF, DHA, or DHA-TF treated TNBC cells to understand the potency of these treatments in generating ROS. As seen in all the samples, TF treatment shows almost no impact on total ROS production compared with untreated ones (Figure 2A). In both DHA and DHA-TF treated cells, more total ROS were generated  and the most substantial increase was evident from the DHA-TF treatment in all cell lines. Furthermore, we found increased superoxide production in DHA or DHA-TF treated cells, which is the essential ROS species responsible for lipid peroxidation and membrane destabilization (28). In summary, TF has negligible effects on both total ROS and superoxide generation, whereas DHA and DHA-TF significantly increase their production in all tested cell lines ( Figure 2B).

DHA-TF Adducts Induce Cell Apoptosis and Ferroptosis in TNBC
Above, we proved that DHA-TF adducts significantly reduce cell viability and clonogenic capacity in TNBC cells. It has been established that artemisinins induce cell apoptosis in colon cancer cells (25). Thus, cell apoptosis of the DHA-TF treated TNBC cells was analyzed by Annexin V/PI assay and a significant increase was observed in all four cell lines compared with the untreated cells ( Figure 3A). The DHA-TF adducts induce 4% and 7% cell apoptosis in MDA-MB-231 and MDA-MB-436 respectively. Whereas more cell apoptosis was observed in MDA-MB-468 (15%) and BT549 (13%), indicating a higher sensitivity to DHA-TF treatment ( Figure 3B).
In the follow-up experiment, the cells were pretreated with the ferroptosis inhibitor ferrostatin-1 (Fer-1), followed by increasing concentrations of DHA or DHA-TF ( Figure 4A). Similar as described above, both DHA and DHA-TF reduce cell viability, and DHA-TF is more effective than DHA alone. Interestingly, the ferroptosis inhibitor Fer-1 partially rescues all the cell lines from the DHA or DHA-TF induced cell death. MDA-MB-468 and BT549 are most susceptible to DHA-TF treatment, and Fer-1 significantly reduces their cell death. This indicates that the cell death triggered by both DHA and DHA-TF in all TNBC cell lines is partially caused by ferroptosis. We also analyzed the expression of glutathione peroxidase 4 (GPX4)

DHA-TF Enhances DHER-Induced Apoptosis in MDA-MB-231 and MDA-MB-436
To enlarge the therapeutic window of DHA-TF in TNBC, we combined it with TRAIL variants to evaluate their efficacy in inducing cell apoptosis. We treated all the cells with wildtype (WT) recombinant human TRAIL (rhTRAIL) and the cell viability was determined. After 24 h incubation with increasing concentration of rhTRAIL, MDA-MB-231 is highly sensitive, MDA-MB-436 is slightly sensitive, MDA-MB-468 and BT549 are resistant to TRAIL-induced apoptosis ( Figure 5A). Then, we combined DHA or DHA-TF with rhTRAIL WT, DR4-specific TRAIL (4C7) or DR5-specific TRAIL (DHER) to understand their influence on TRAIL pathway. In Figure 5B In order to confirm whether the cell viability reduction is from the TRAIL-induced apoptosis pathway, caspase 3/7 activity was measured in MDA-MB-231 and MDA-MB-436 ( Figure 6A). Individual DHA or DHA-TF treatments show limited effects on caspase 3/7 activation in both cell lines, while a minor effect is observed with rhTRAIL WT or 4C7. Most significantly, both DHA and DHA-TF in combination with DHER strongly improve the caspase activation in both cell lines tested, and DHA-TF establishes the best response. Because of the increases in caspase 3/7 activity, we analyzed the effects in early and late stages apoptosis in the DHA-TF and DHER treated cells (Figures 6B, C). Whereas DHA-TF alone shows limited effects on the cells, the individual DHER significantly induces cell apoptosis in MDA-MB-231. Importantly, DHA-TF remarkably enlarges DHER-induced apoptosis compared to untreated cells (p ≤ 0.0001) or to DHER treated cells (p ≤ 0.0001) in both cell lines.

DHA-TF Induces DR5 Expression on TNBC Cell Surface
Judging from the different performance of rhTRAIL WT, 4C7, and DHER, it might be speculated that DHA-TF affects the expression of different death receptors in TNBC cells. Therefore Apart from P53 and caspase inhibitors, several ROS inhibitors [N-Acetyl-L-cysteine (NAC), ferrostatin-1 (Fer-1), and liproxstatin-1(Lip-1)] were used to eliminate the effects from the combination treatment. All the tested ROS inhibitors significantly increase the cell viability in the DHA-TF and DHER treated cells ( Figure 8B). We also found that these ROS inhibitors rescued MDA-MB-436 from DHA-TF-induced cell death. This did not occur in MDA-MB-231, which can be explained by its lower sensitivity to DHA-TF. Afterwards, the DR5 expression was analyzed in the cells which were pretreated with NAC followed with DHA-TF ( Figures 8C, D). As we presented above, DHA-TF increases cell surface DR5 expression, which can be significantly down-regulated by NAC in MDA-MB-231. The same pattern was observed in MDA-MB-436. In summary, ROS production triggers DR5 expression in the DHA-TF treated cells, making the cells more susceptible for DHER treatment.

DISCUSSION
In 2020, approximately 531,068 new cases were diagnosed with breast cancer in Europe, and 141,765 people of both sexes died, according to Globocan 2020. As the most aggressive breast cancer subtype with high mortality and poor prognosis, TNBC's metastasis leads to less than two years median overall survival time, which is substantially shorter than the other subtypes (29). For treating TNBC, the apoptosis-inducing agent TRAIL has been extensively investigated due to its tumor-specific properties (30). To enhance TRAIL-induced apoptosis in TNBC, various natural products have been proved effective, such as pterostilbene and ursolic acid (31,32). However, most of them are still under clinical trials (31,33). In order to accelerate clinical trials, dihydroartemisinin, World Health Organization (WHO)-approved drug, was used as a combination therapy with TRAIL in cancer cells (34). At the same time, transferrin was also used to further improve the efficacy of DHA by the formation of the protein-drug adducts. It has been proven that DHA binds to TF mainly through van der Waals forces and hydrogen bonds, and the molar binding ratio is 1:1. Furthermore, it is important to form the DHA-TF adducts in serum-free medium to avoid the formation of DHA-BSA (15). The quenching of intrinsic fluorescence indicates that DHA changes the microenvironment of chromophores and binds to TF ( Figure S1B), which confirms that DHA interacts with TF at neutral pH (17). Normally, artemisinins enter the cells through passive diffusion, which is uncontrolled, slow and inefficient (35). By forming the DHA-TF adducts, these will be delivered into cells via TfR-mediated endocytosis. On one hand, higher TfR expression on the cancer cell surface makes it a promising method of delivering DHA to tumors (36). On the other hand, more iron transported into cells through transferrin increased DHA-derived ROS production and induced more cell death (37).
In the present paper, we demonstrated that DHA-TF decreases cell viability and produces more ROS/superoxide compared with DHA in TNBC (Figures 1 and 2). Interestingly, DHA-TF adducts exhibited different effects in the tested cell lines, which does not correspond to their TfR cell surface expression level. Similar findings were reported previously in the human prostate carcinoma cell line DU 145 (38). Compared between cell lines, both MDA-MB-468 and BT-549 are more susceptible to DHA treatment, making them more sensitive to DHA-TF ( Table 1).
It has been reported that DHA or DHA-TF activates caspase-9 and induces cell apoptosis via the intrinsic pathway (38,39). Here, we found that DHA-TF also stimulates DR5 expression on TNBC cell surface (Figure 7), suggesting the possibility to induce cell apoptosis through the extrinsic pathway with the combination with TRAIL. What's more, with the DR5-specific TRAIL variant DHER, the cell viability in MDA-MB-231 and MDA-MB-436 is significantly decreased (Figures 5 and 6). Even though DHA-TF successfully increases DR5 expression in all It is reported that the downregulation of c-FLIP or XIAP expression enhances TRAIL sensitivity in MDA-MB-468, which could be implemented in the near future (40,41). Besides, we only analyzed the effectiveness of the combination treatment in cultured cells, the specificity of the combination to tumor cells in vivo still needs further investigation. DHA-TF or TRAIL loaded nanoparticles have been proved to be effective and specific to tumor cells in vivo (42)(43)(44), Which will be a promising delivery method for the combination treatment for further analysis. The P53 status is important for artemisinin induced DR5 upregulation (25). However, most of the TNBC cell lines contain mutated P53 (45), for example, MDA-MB-231 (R280K) (46), MDA-MB-436 (E204 frameshift to stop codon) (47), MDA-MB-468 (R273H) (47) and BT-549 (R249S) (48). All the mentioned mutations are located on the P53 DNA binding site, interfering the regulation of P53 to its target genes, such as DR5 (49). Here, we found that DHA induces DR5 expression in a P53independent and ROS-dependent manner ( Figure 8A). It has been reported that the ART-type drugs are able to induce endoplasmic reticulum (ER) stress response and upregulate the CCAAT/-enhancer-binding protein homologous protein (CHOP) expression in various tumor cells (7,50,51). What's more, we and other groups have demonstrated that CHOP is responsible for the DR5 upregulation under certain stresses (52)(53)(54). Therefore, it is of interest to investigate the influence of DHA-TF on the expression of CHOP, and the linkage between CHOP and DR5 expression under DHA-TF stress. Based on our presented data, we visualized the combination treatment of DHA-TF with DHER in the TNBC cells in a schematic illustration (Figure 9). DHA was incubated with TF in serumfree medium to form DHA-TF adducts. Then, the mixture was added to TNBC cells. Holo-transferrin transports iron and DHA into cells through transferrin receptor-mediated endocytosis. With the decreased pH in the endosome, DHA is released from TF. At the same time, ferrireductase reduces Fe 3+ to Fe 2+ (55), which converts DHA into ROS by cleaving the endoperoxide bridge. The accumulated ROS induces apoptosis, ferroptosis and upregulates DR5 expression. Then, TRAIL or DHER was added into the DHA-TF pretreated TNBC cells and significantly increases cell apoptosis.
In summary, DHA-TF adducts induce more cell death in TNBC cells by triggering both the apoptosis and ferroptosis signaling pathways. In addition, this treatment increases DR5 expression on the cell surface, making the cells more susceptible to the DR5-specific TRAIL variant. Furthermore, DR5 The geometric mean of DR5 expression relative to isotype is analyzed using FlowJo v10 software and shown in bar graphs. Data shown are mean ± SEM from three independent experiments. *0.01 < p ≤ 0.05, **0.001 < p ≤ 0.01, ***0.0001 < p ≤ 0.001, ****p ≤ 0.0001. upregulation is P53-independent and ROS-dependent in the TNBC cells. Our data showed that the combination of DHA-TF with DHER can be a potent strategy for TNBC patients in clinical trials.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
XZ, AS-G, and FN designed and performed the experiments. XZ and AS-G analysed data. XZ wrote the original draft. XZ, AS-G, RS, and WQ reviewed the manuscript. WQ supervised the research and acquired funding. All authors contributed to the article and approved the submitted version.

ACKNOWLEDGMENTS
Thanks to our colleague Fengzhi Suo for giving comments to the manuscript.