Synergistic ferroptosis‐starvation therapy for bladder cancer based on hyaluronic acid modified metal–organic frameworks

Abstract Bladder cancer (BCa) is one of the most common malignancies of the urinary tract. Metastasis and recurrence of BCa are the leading causes of poor prognosis, and only a few patients can benefit from current first‐line treatments such as chemotherapy and immunotherapy. It is urgent to develop more effective therapeutic method with low side effects. Here, a cascade nanoreactor, ZIF‐8/PdCuAu/GOx@HA (ZPG@H), is proposed for starvation therapy and ferroptosis of BCa. The ZPG@H nanoreactor was constructed by co‐encapsulation of PdCuAu nanoparticles and glucose oxidase into zeolitic imidazolate framework‐8 (ZIF‐8) modified by hyaluronic acid. The vitro results indicated that ZPG@H enhanced intracellular reactive oxygen species levels and reduced mitochondrial depolarization in the tumor microenvironment. Therefore, the integrated advantages of starvation therapy and chemodynamic therapy endow ZPG@H with a perfect ferroptosis inducing ability. This effectiveness, combined with its excellent biocompatibility and biosafety, means that ZPG@H could make a critical contribution to the development of novel BCa treatments.


| INTRODUCTION
Bladder cancer (BCa) is one of the most common urinary tract malignancies. According to the National Comprehensive Cancer Network Clinical Practice Guidelines, metastasis and recurrence of BCa are the main causes of poor prognosis, especially in patients who are diagnosed at late stages. 1,2 In recent years, cisplatin (CDDP)-based chemotherapy and immune checkpoint inhibitors have been used as first-line treatments for muscle-invasive and metastatic BCa; however, it has been estimated that only 50% of BCa patients benefit from chemotherapy and 20% benefit from immunotherapy. 3 The overall survival of patients with BCa remains unsatisfactory because of adverse effects, resistance, and intolerance. 4,5 Therefore, novel antitumor treatments that are less toxic and resistant are urgently needed.
Ferroptosis is an iron-dependent form of programmed cell death characterized by lipid peroxidation, which is usually caused by an imbalance in the generation and degradation of intracellular reactive oxygen species (ROS). 6,7 The increase in intracellular ROS and decrease in the mitochondrial volume are considered biomarkers of ferroptosis. [8][9][10] Ferroptosis is reported to be a natural barrier to cancer development because certain tumor suppressors, such as p53 and BRCA 1-associated protein 1 (BAP1), engage in the ferroptosis pathway. 11,12 Moreover, the high load of ROS and specific mutations in tumor cells make them intrinsically more susceptible to ferroptosis. 10 Therefore, ferroptosis is an important target for BCa treatment. Nanomaterials are essential ferroptosis inducers because of their high ROS production efficiency, potential for glutathione (GSH) depletion, and other features. 13 On the one hand, the specific modifications on the surface of nanomaterials endow them with long blood circulation time and reduced renal clearance. Conversely, nanomaterials with passive and active target abilities can enhance accumulation at the tumor site and reduce adverse effects. 14 For example, Liao et al. reported a nanodrug that specifically targets BCa cells to deliver iron ions, produce excess ROS, and ultimately lead to ferroptosis. 15 Moreover, AuNRs, IONs@Gel, 16 and gold clusters (PAA4 and PAA5) 17 have also been reported have potential for BCa treatment targeting the ferroptosis pathway. Therefore, it is significant to develop nanomedicine that can actively target BCa and induce subsequent ferroptosis pathway.
Chemodynamic therapy (CDT) has been designed to convert endogenous chemical energy into ROS via Fenton and Fenton-like reactions, 18,19 ultimately resulting in cell apoptosis or even ferroptosis. 20 However, a single CDT treatment usually does not completely eradicate tumors because of low ROS generation efficiency. [21][22][23][24] Therefore, a CDT combined with a synergistic treatment strategy should be developed. Owing to the Warburg effect, the proliferation of cancer cells consumes more glucose than normal tissues. 25 Once glucose supply is interrupted, the growth of tumor cells is inhibited. 25,26 Thus, starvation therapy, which is related to glucose metabolism, is increasingly being regarded as a potential clinical therapy approach. 27 As glucose oxidase (GOx) can catalyze the oxidation of glucose into gluconic acid and hydrogen peroxide, 28 it has been reported to be involved in the regulation of tumor metabolism and can play a role in starvation therapy. More importantly, GOx-based starvation therapy can synergistically enhance the efficiency of CDT by increasing H 2 O 2 levels in the tumor microenvironment. [29][30][31] Moreover, Zhou and his colleagues indicated that GOx/BSO@CS PVs effectively inhibit the growth of 4T1 tumors and provide the basis of a promising strategy to prepare pH-sensitive nanomedicines for synergistic starvation-ferroptosis tumor therapy. 32 Here, a cascade nanoreactor intended for starvation therapy and ferroptosis of BCa was proposed. The nanoreactor ZPG@H was constructed by co-encapsulation of PdCuAu nanoparticles and GOx into zeolitic imidazolate framework-8 (ZIF-8) modified with hyaluronic acid (HA). After entering the tumor environment, HA-modified ZPG@H can specifically target the cluster determinant 44 (CD44) receptor, which is expressed on the surface of BCa cell. [33][34][35] After HA has been decomposed by intracellular hyaluronidase (Hyase), 36

| Ethics statement
All animal studies were carried out in strict accordance with the Guide for the Care and Use of Laboratory Animals as adopted by the Institutional Animal Care and Use Committee of Southern Medical University.

| Cell culture
Human BC cell lines T24 were used in this study. The cells were cultured in DMEM medium (Gibco, USA) supplemented with 10% FBS (Gibco, USA) and incubated at 37 C in 5% CO 2 . The cell lines were authenticated by short tandem repeat (STR) profiling before receipt and were propagated for 6 months after resuscitation.

| Synthesis of ZIF-8/PdCuAu/GOx@HA
The preparation of PdCuAu nanoparticles were synthesized as follows. K 2 PdCl 4 , CuCl 2 , and HAuCl 4 aqueous solution with the molar ratio of 1:1:1 was added to 9.1 mL of the mixed solution containing 0.256 g CTAB and 0.21 g citric acid. The mixed solution was stirred for several minutes, then 0.03 g ascorbic acid was added and stirred for 20 min. It was then transferred to stainless steel autoclave lined with polytetrafluoroethylene. The sealed vessel was heated at 190 C for 4 h and then cooled to room temperature. The product was separated by 9000 rpm for 10 min and washed with deionized ethanol treated three times.
2.5 | Cellular uptake ability of ZIF-8/PdCuAu/ GOx@HA T24 cells were seeded in confocal dish at a density of 5 Â 10 3 cells and cultured for 24 h. After the T24 cells incubated with RhBconjugated ZPG@H for 0, 2, 4, 6, 8, and 10 h, the cells were fixed with 4% paraformaldehyde, sequentially stained with DAPI, and finally observed with a CLSM. Additionally, the treated cells were collected and analyzed with flow cytometry assay (BD FASCVerse, USA).
Additionally, in order to detect PTT of ZPG@H for T24 cells, the cells were planted in 96-well plates, and the non-NIR groups were treated with ZPG@H NPs as described above. As for NIR irradiation groups, the cells were pre-incubated with these NPs for 6 h, and then irradiated with an NIR laser (1064 nm, 1 W/cm 2 , 10 min), and finally incubated for another 18 h. Afterwards, cell viability was quantified with CCK8 assay through a microplate reader.
For the live/dead staining assay, the T24 cells were seeded in 6-well plates and cultured overnight. Subsequently, different preparations of ZG@H, ZPG@H, or ZP@H were added and incubated for 12 h. After that, the treated cells were stained with Calcein-AM/PI double stained kit, and finally observed via fluorescence microscopy and evaluated through flow cytometry.

| Detection of mitochondrial function
The mitochondrial transmembrane potential was detected with JC-1 mitochondrial membrane potential (MMP) Assay Kit. Briefly, after treated with ZG@H, ZPG@H or ZP@H, with or without NIR laser, the T24 cells were collected and incubated with JC-1 staining. Then, the cell samples were observed and imaged with fluorescence microscope, as well as analyzed through flow cytometry assay.

| Detection of ROS
The reactive oxygen species was detected with DHE (ROS probe).
T24 cells treated with ZG@H, ZPG@H, or ZP@H, with or without NIR laser, were harvested and incubated with DHE. After that, the cells were observed via fluorescence microscope and assessed by flow cytometry assay.

| Detection of the level of lipid peroxidation
In order to further evaluate the level of lipid peroxidation, the treated cells were collected and detected with via liperfluo probe, and finally observed with fluorescence microscope and measured by flow cytometry.

| In vivo therapeutic effect of ZPG@H
Three groups of the BCa-bearing mice mentioned above were also Additionally, the hearts, livers, spleens, lungs, and kidneys of all groups BCa-bearing mice mentioned above were collected and fixed in 4% formalin for H&E to assess the biocompatibility.

| Statistical analysis
All data are expressed as the mean ± standard deviation (n ≥ 3), and the significance of differences among groups was evaluated using one-way ANOVA and a Student's t-test (*p < 0.05, **p < 0.01, ***p < 0.001).

| Characterization and ROS production of ZIF-8/PdCuAu/GOx@HA (ZPG@H)
The morphology and structure of ZPG@H were characterized using TEM, FT-IR, and X-ray diffraction (XRD). The TEM images showed that the obtained ZPG@H had an average size of approximately 85 nm (Figure 1a), and PdCuAu nanoparticles with a diameter of 5 nm were uniformly distributed inside the ZIF-8 framework. FT-IR spectroscopy was used to analyze the functional groups of the ZPG@H nanocomposites ( Figure 1b); 1567 cm À1 showed the vibration absorption of the imidazole framework, 3450 cm À1 showed the vibration absorption peak of OH in water molecules, and 1549 cm À1 , the vibration absorption peak of the C C double bond on the benzene ring, 37 indicating that the ligand was completely deprotonated at this position and providing the necessary conditions for the preparation of the ZIF-8 crystal. 38 In addition, the peak at 1648 cm À1 can only be observed in the IR spectrum of ZPG@H, and this was corresponding to the N H and C H stretching of GOx, 39 Figure S1 shows the double inverse relationships between the glucose concentration and initial reaction rate for two formations. Then, the K m value can be calculated as 8.23 mM for free GOx and 9.24 mM for encapsulated GOx, based on the equation K m = Slope/Intercept. 40 The above result implied that the catalytic activity of GOx slightly decreased after the formation of ZPG@H, but there was not a significant difference between the two formations.
The ratio of GOx and PdCuAu nanoparticles was optimized; 3 mg PdCuAu nanoparticles and GOx of different masses (3, 4, 5, or 6 mg) were used for the synthesis of ZPG@H. Figure S2 reveals that absorbance intensity increased with the percentage of GOx increase and reached to a platform at 5 mg. Therefore, it was thought that 3 mg GOx and 5 mg PdCuAu nanoparticles were the optimal ratio to synthesize ZPG@H.

DLS measurements of ZPG@H in different mediums or in PBS
with different pH values were conducted. As shown in Figure S3A,

| In vitro therapeutic efficacy and cellular internalization of ZIF-8(PdCuAuGOx)@HA (ZPG@H)
The CD44 receptor is typically highly expressed on the surface of BCa cells (e.g., T24); however, it is barely expressed in normal cell lines (e.g., SV-HUC-1 cells). Moreover, HA-conjugated NPs preferentially targeted CD44 receptor-overexpressing cells. We explored the cellular uptake of ZPG@H labeled with rhodamine B (RhB) to investigate its active targeting ability through CD44 receptor-mediated endocytosis. As shown in Figure 2a fluorescence microscopy (Figures 2f and S4E), T24 cells in the ZPG@H group exhibited the highest level of mortality. Moreover, flow cytometry revealed that T24 cells treated with ZPH@H showed a higher percentage of apoptosis than the other groups (Figures 2g and   F I G U R E 3 Legend on next page. S4F). Together, these results demonstrate the excellent killing effect of ZPG@H on T24 cells in vitro. In addition, the flow cytometry results, which were consistent with fluorescence microscopy, revealed that ZPG@H had the best therapeutic efficacy among the three drug formulations, and its cytotoxicity was enhanced with increasing concentration (Figures 2h and S4B-D).

| In vitro ROS production and ZPG@H starvation therapy
Previous studies have reported that ROS can cause irreversible damage to mitochondria. 41

| Ferroptosis of BCa cells based on ZPG@H
To provide further evidence on the death mechanism of T24 cells induced by ZPG@H, a CCK8 assay was conducted to evaluate T24 cell viability in the presence of various inhibitors, including autophagy inhibitors chloroquine (CQ), ferroptosis inhibitors (Ferrostatin-1, Fer-1), apoptosis inhibitors (ZVAD-FMK, ZVAD), and necrosis inhibitor (Necrostatin1, Nec1). As shown in Figure 4a, ZPG@H-induced cell death was markedly reduced by Fer-1, implying that ZPG@H induced ferroptosis in T24 cells. To confirm the role of the ferroptosis pathway in ZPG@H-induced cell death, ROS and JC-1 probes were used to assess the process of ferroptosis via fluorescence microscopy and flow cytometry. The highest red fluorescence was observed in the group treated with ZPG@H without inhibitor, and the lowest red fluorescence in the group treated with Fer-1 (Figures 4c,d, and S6A).
Meanwhile, the aggregate/monomer fluorescence ratio of the group with inhibitor was higher than that of the group without inhibitor in the fluorescence microscopy and flow cytometry assay results (Figures 4b,e,f, and S6B).
Additionally, LiperFluo was used to evaluate intracellular lipid hydroperoxides. As shown in Figure S6C,E, the green fluorescence of T24 cells incubated with 40 μg/mL ZPG@H was significantly stronger that the groups with 0 and 20 μg/mL ZPG@H. In addition, the increase in lipid hydroperoxide following ZPG@H treatment showed a concentration-dependence in cytometry assay ( Figure S6G,I). In the rescue experiment, ZPG@H-induced increase of lipid peroxidation was rescued by the iron death inhibitor Fer-1 both in fluorescence microscopy and cytometry assay ( Figure S6D,F,H,J).
Western blotting was performed to measure the protein levels of the ferroptosis-related genes. Interestingly, the protein expression  In addition, when T24 cells were co-cultivated with different concentrations of ZPG@H, the expression of SLC7A11 and GPX4 was reduced as the ZPG@H concentration increased, while SLC3A2 expression remained stable. Moreover, the reduction in the expression levels of these proteins was enhanced by Fer-1. (Figure 4g). It made certain that the ZPG@H can induce ferroptosis of T24 cells by inhibiting the expression of SLC7A11 protein.  (Figure 5e). In addition, our results showed that the ZPG@H group had the slowest growth in the three drug formulations ( Figure S7B,C). Additionally, compared with control group, the average tumor weigh of ZPG@H group was only 23.06% ( Figure S7D). Notably, body weight of mice did not differ significantly between groups ( Figure S7E). All results mentioned above revealed that ZPG@H had the best therapeutic effect in those three drug formulations.
After verifying the therapeutic effect of ZPG@H in vivo, the expression of GPX4 and SLC7A11 in tumor tissues suffering from ferroptosis was examined using IHC staining. As shown in Figure 5e,g, the IHC scores of SLC7A11 decreased from 4.4 to 2.4 and 0 for the control, 10 and 20 mg/kg groups, respectively. This revealed that the import of cystine, used in glutathione biosynthesis, decreased in T24 cells treated with ZPG@H. For GPX4, the expression decreased by approximately 58.5% and 31.7% after injection of 10 and 20 mg/kg ZPG@H, respectively (Figure 5f,h). The inhibition of GPX4 can result in lipid peroxidation and ferroptosis. Simultaneously, malondialdehyde levels were elevated in the treated groups, suggesting an important role for the ferroptosis pathway in ZPG@H-mediated antitumor therapy ( Figure S7F). The results showed a marked decrease in SLC7A11 and GPX4, triggered by starvation and ferroptosis in vivo.
Fluorescence imaging was used to detect the biodistribution of Cy5.5 modified ZPG@H in vivo, as shown in Figure S8A,B; ZPG@H can access the tumor region in 3 h. In Figure S8C, the fluorescence can mainly be observed in the tumor, liver, and kidney, revealing that ZPG@H was metabolized through liver and kidney. The metabolism of ZPG@H in the blood of mice was examined by ICP, the Zn 2+ in blood decreased gradually with time prolongation (Figure S8D), and the decreased trend was consistent with the fluorescence results. t 1/2 was calculated as 12.1 h.
It is vital to evaluate the biosafety and biocompatibility of ZPG@H; therefore, these two qualities must be assessed. Figure 2e shows that the effect of ZPG@H on tumor cells (T24) was more pronounced than that on normal urothelial epithelial cells (SV-HUC-1), indicating that the cytotoxicity induced by ZPG@H was selective. In addition, blood compatibility was evaluated using red blood cell (RBCs) hemolytic analysis with mouse RBC. According to our data, hemolysis could hardly be observed with different concentrations of ZPG@H, even at 120 μg/mL. Phosphate-buffered saline and purified water were used as negative and positive controls, respectively, and the maximum hemolysis rate was less than 5% (Figure 5i,j). In addition, H&E staining of the heart, liver, spleen, lung, and kidney of all groups after the intervention showed little off-target damage and inflammatory reaction ( Figure S9A). Moreover, as shown in Figure S8A, intravenously administered ZPG@H was gradually cleared from the blood over 3 days. Therefore, it is safely eliminated from the body and does not interfere with metal metabolism in the host animals.
As shown in Figure 5b, the weight of the mice increased smoothly during the entire experimental cycle, and there was no significant difference between the group injected with PBS and the groups injected with ZPG@H. In addition, routine blood examination and biochemical analysis demonstrated that the injection had no apparent side effects in the experimental groups ( Figure S9B,C). Therefore, based on the above results, it can be deduced that ZPG@H has no obvious side effects in vivo. Taken together, these results confirm the biocompatibility and biosafety of ZPG@H.

| CONCLUSION
In conclusion, the ZPG@H nanocomposite was devised, and its cytotoxicity was produced by promoting starvation therapy and According to the inhibitor treatment results, cell death caused by ZPG@H occurred mainly via the ferroptosis pathway. An intracellular cystine transporter, SLC7A11, was found to be a potential target of ZPG@H. In addition, when T24 cells were incubated with a certain concentration of ZPG@H, the protein level of SLC7A11 decreased, leading to cytotoxicity through ferroptosis. Altogether, this study provides a synergistic strategy to enhance intracellular ROS levels and reduce mitochondrial depolarization, which has significant value for updating traditional BCa treatments.