Tumor Microenvironment-responsive Nanodrug for Clear-cell Renal Cell Carcinoma Therapy via Triggering Waterfall-like Cascade Ferroptosis

Abstract


Introduction
Renal cell carcinoma (RCC), one of the most common genitourinary malignant tumors, causes 170,000 deaths each year worldwide. Clear cell renal cell carcinoma (ccRCC) accounts for more than 75% of RCC cases, which is characterized by high invasiveness and is inclined to metastasize. [1,2] However, the treatment of ccRCC remains a challenge in clinical setting, owing to the facts that it is easy to develop metachronous distant metastases after surgical resection, and insensitive to chemotherapy or radiotherapy. [3,4] Therefore, it is still an urgent clinical need for advanced ccRCC to develop the new drugs and treatment strategies.
As far as we know, metabolic reprogramming is an important feature of ccRCC, which is usually associated with mutations in VHL tumor suppressor genes, and about 90% of ccRCC have VHL mutations. [5] Studies showed that mutations in the VHL gene can lead to the stable expression of hypoxia-inducible factors HIF-1α and HIF-2α. [6] In addition to the aberrant polyunsaturated fatty acids (PUFAs) elevation mediated by HIF-2α, the stable expression of HIF-1α and HIF-2α also reduce the catabolism of fatty acids through inhibiting β oxidation, [7] endowing the signi cantly higher level of PUFAs in ccRCC than that of normal kidney tissue. [8] It is worth noting that PUFAs containing diallyl groups are more likely to be oxidized to form lipid peroxides (L-OOH) compared with saturated fatty acids (SFA) and monounsaturated fatty acids (MUFA). [9] Therefore, the changes of metabolic feature in ccRCC makes it inherently prone to a new form of regulated cell death, ferroptosis, which is characterized by the accumulation of iron-dependent reactive oxygen species (ROS) and LPO to lethal levels. [7,8,10,11] Brie y, PUFAs in cells are more likely to be oxidized by ROS (e.g., hydroxyl radical ( • OH)) to form L-OOH. Then, the L-OOH react with redox-active transition metals such as ferrous iron (Fe 2+ ) to produce reactive accumulation of alkoxyl radicals (L-O • ) that can deplete nucleic acids and proteins. Eventually, the extensive peroxidation of lipids including L-OOH and L-O • results in bio-membrane destabilization and micelle formation, driving cells to undergo ferroptosis. [12][13][14][15] In recent years, numerous studies have focused on the development of iron-based nanomaterials that catalyze • OH generation by the Fenton reaction between H 2 O 2 and Fe 3+ /Fe 2+ to attack PUFAs and initiate ferroptosis. [16][17][18] However, the intratumoral H 2 O 2 concentration (50-100 × 10 −6 M) is limited and insu cient to produce desirable and enough • OH for inducing satisfactory nanocatalytic-therapeutic e cacy via Fenton or Fenton-like reactions. [19] In addition, tumor cells have evolved powerful antioxidant systems to prevent lipid peroxidation and ferroptosis, resulting in limited antitumor e ciency when relying solely on Fenton reaction. Reduced glutathione (GSH), as a vital antioxidant molecule, can cycle between oxidized glutathione (GSSG) by participating in redox biochemical reactions, enabling GSH/GSSG to be an essential antioxidant defense conserved in eukaryotes. [20,21] Cysteine is a building block for the biosynthesis of GSH, and system x c -, as a transmembrane cystine-glutamate antiporter, imports cystine into cells, which plays a key role in GSH synthesis and cellular redox regulation. [22,23] On the other hand, glutathione peroxidase 4 (GPX4), as a major antioxidant enzyme can directly reduce lipid hydroperoxides to form lipid alcohols by using two GSH molecules as electron donors, which prevents the transition metal (e.g., Fe 2+ )-dependent formation of toxic lipid reactive oxygen species.
[9] Therefore, to induce ferroptosis using iron-based nanomaterials, the Fe dose needs to be very high, even up to 75 mg kg −1 body weight, via intravenous injection, [17] and most of the reported iron-based nanomaterials need to either include an additional component that can contribute to ferroptosis, or be combined with other therapeutic approaches (e.g., chemotherapy). [19,24] It is worth noting that erastin, sulfasalazine, ras-selective lethal small molecules (RSL3) etc. have been found to trigger accumulation of lipid hydroperoxides and ferroptosis by blocking the synthesis of GSH or directly inhibiting GPX4 activity.
These small molecules have emerged as the most potent and selective compounds for killing ccRCC cells. [8,25] However, their low speci city for tumors, short half-life in the blood, poor water solubility and de cient accumulation at the tumor site prevents their reliable application in vivo. Nanotechnology has been widely investigated to develop advanced drug delivery system for cancer therapy, with the intrinsic physicochemical properties of nanomaterials. Metal-organic frameworks (MOFs) with a high loading capacity attributed to their large surface area and ultra-high porosity enable them to be promising platforms for hydrophobic small molecule drugs delivery. [26] Particularly, Iron based MOFs nanoparticles (MIL-101(Fe) NPs) with great biocompatibility and degradability provides the possibility of achieving stimuli-responsive controlled release of various pharmacological molecules. [27,28] In

Instruments
The morphology and elemental mapping images were recorded with a transmission electron microscope (FEI Tecnai F20, acceleration voltage = 200 kV). The hydrodynamic sizes and zeta potentials of the NPs were detected with a Malvern Zetasizer Nano ZS (Malvern, UK). Confocal microscopy images were obtained with a confocal laser scanning microscope (Leica TCS SP8 STED, Germany), and the ow cytometry assay was performed with a ow cytometer (BD FACSAria™ III, USA). In vivo imaging was performed with an IVIS imaging system (PerkinElmer, USA).

Synthesis of MIL-101(Fe) NPs
The synthetic procedure of MIL-101(Fe) NPs was consistent with a previous report. [  Then, GPX4 activity was detected according to the manufacturer's protocol. The absorbance at 340 nm was recorded every 1 min with a SpectraMax microplate reader with the temperature maintained at 25 °C.
We also employed a Western blotting approach to assess GPX4 levels. In brief, the lysate of formulationtreated 786-O cells was collected for analysis. The level of protein in these samples was quanti ed. Then, samples containing equal amounts of protein (20 μg) were loaded and subjected to standard SDS polyacrylamide gel electrophoresis. The proteins were separated at a constant voltage, followed by electrical transfer to a 0.45 μm poly (vinylidene uoride) (PVDF) membrane. Upon membrane blocking with 5% milk phosphate buffered saline with Tween 20 (PBST) solution for 1 hour at room temperature, the target proteins were incubated with GPX4-speci c antibody and β-actin antibody overnight at dilutions of 1:2500 and 1:5000, respectively, at 4 °C. Thereafter, HRP-conjugated anti-rabbit IgG H&L at a dilution of 1:5000 was applied as the secondary antibody, and the speci c bands were developed using ECLTM Western blotting detection reagents.

Cytotoxic and anti-tumor effect in vivo
The cytotoxicity measurement was quanti ed according to the standard CCK-8 approach. 786-O cells were seeded in 96-well plates at a density of 5×10 3 cells/well and incubated for 24 hours. Then, MIL-101(Fe), free RSL3 or MIL-101(Fe)@RSL3 was added to each well at different concentrations. After 24 hours of incubation, CCK-8 solution was added to each well, and the cells were incubated for an additional 1 hour. Afterwards, the absorbance was measured with a microplate reader at 450 nm. Cell viability was calculated by the ratio of the absorbance of the sample to the absorbance of the control, and each test was repeated at least three times. Then, calcein-AM and PI were used for live/dead cell staining after different treatments.

In vivo uorescence imaging and biodistribution of nanodrugs
To track the biodistribution of MIL-101(Fe)@RSL3, ICG-labelled MIL-101(Fe)@RSL3 was intravenously injected into mice. Fluorescence images of mice were recorded at different time points with an IVIS imaging system, and then the tumor and main organs were harvested for uorescence imaging. Blood circulation was measured by drawing 10 µL of blood from the tail vein of the tumor-bearing mice at different times after intravenous injection of ICG-labelled PEG-MIL-101(Fe)@RSL3. Each blood sample was dissolved in 140 µL of lysis buffer (1% sodium dodecyl sulfonate, 1% Triton X-100, 40×10 -3 M Tris acetate), and the uorescence intensities were measured using a microplate reader (excitation wavelength: 789 nm, emission wavelength: 814 nm, n=3). The blank blood sample was also measured as the control to subtract blood auto uorescence.

Side effects evaluation
Human normal hepatocyte LO2 and human renal tubular epithelial cell HK-2 cells were seeded into 96well plates and treated with a series of MIL-101(Fe)@RSL3 concentrations (RSL3 dose 0, 0.35, 0.70, 1.40, 2.80, 5.60 μM). Cell viability was evaluated by CCK-8 assay, which was conducted according to the manufacturer's protocol.
Whole blood was collected from the mice and used for haemolytic analysis. After centrifugation at 800 rpm, the upper serum was discarded, and the lower blood cells were washed with cold PBS three times and mixed with a certain amount of MIL-101(Fe)@RSL3. Twenty-four hours later, the mixture was centrifuged at 800 rpm, and the absorbance of the supernatant was measured at 350 nm to calculate the haemolysis rate.
To evaluate side effects in vivo, the mice were sacri ced seven days after treatment. Approximately 0.8 mL of blood was collected from each mouse to conduct complete blood panel analysis and serum biochemistry assays at the Laboratory of Comparative Medicine, Guangdong Medical Laboratory Animal Center. The major organs of the mice from each group were collected and stained with H&E according to standard techniques to evaluate toxicity.
Anti-tumor effect in vivo BALB/c nude mice were randomly divided into four groups (n=6). The nanodrugs were intravenously injected into mice three times a week for a total of 5 injections. The tumor volume was measured every three days according to the formula: V=(L×W 2 )/2, where L is the tumor length and W is the tumor width.
The relative tumor volume was calculated according to the equation: V/V 0 , where V 0 is the initial tumor volume before the start of treatment. Then, 21 days later, the tumor tissues were excised for weighing and histological examination.

Statistical analysis
The data are presented as the mean±standard deviation (SD). The difference of each treatment group was statistically compared via either Student's t test or analysis of variance integrated with Tukey's post hoc analysis. The threshold P value was set at 0.05.  Figure 1C), and the zeta potential reversed from 34.5±0.6 to -23.7±0.5 after modi cation ( Figure 1D), which was measured by dynamic light scattering (DLS). The increased size and reversed surface charge suggested that MIL-101(Fe) had been successfully modi ed with PEG. It is well known that MOF NPs possess tunable porous structures and high speci c surface areas, which facilitate drug loading and release. Herein, a hydrophobic small molecular (1S,3R-RSL3) was loaded into MIL-101(Fe) to form a drug delivery composite (MIL-101(Fe)@RSL3). Energy-dispersive X-ray spectroscopy (EDX) elemental mapping of MIL-101(Fe)@RSL3 showed C, O, Fe and Cl, which con rmed that RSL3 was loaded into MIL-101(Fe) ( Figure   1E). Fourier transform infrared spectroscopy (FTIR) was used to analyze the PEGylated MIL-101(Fe)@RSL3 ( Figure S1)  Figure S2A). In contrast, when exposed to acidic conditions (pH 5.0), the structure of MIL-101(Fe) NPs showed a tendency to degrade and collapse ( Figure S2B), which was consistent with previous reports. [27,31] Moreover, whether the MIL-101(Fe) NPs could release iron ions were colourimetrically determined with the products of o-phenanthroline and ferrous ions. After incubation with MIL-101(Fe) NPs in PBS pH 7.4 ( Figure S3C) and pH 5.0 ( Figure S3D) for 1, 2, 4, 8, 12, and 24 hours, the supernatants became reddish-orange in color, which was attributed to the iron ion release into solution. It could be clearly observed that the number of iron ions released into the acidic solution increased signi cantly. According to calculations from the standard curve ( Figure S3A, B), the iron ion concentrations in the supernatants were 0.21 mM at pH 7.4 and 0.50 mM at pH 5.0 ( Figure 1G). Hydroxyl radicals ( • OH) were generally considered to be the most toxic ROS, which could be generated by the  Figure 1H). Particularly, the stronger • OH signals were observed at acidic solution, indicating the acidic microenvironment promoted the release of iron ions and thus increased the production of • OH. Moreover, the production of • OH was further detected by methylene blue (MB) decolorization experiments. More obvious color fading of MB was observed in acidic solution after mixed with MIL-101(Fe) and H 2 O 2 ( Figure S4A) than neutral solution ( Figure S4B), which further con rmed the acidic environment accelerated the • OH-production. Overall, with excellent biocompatibility and biodegradable properties, MIL-101(Fe) NPs serve as an ideal drug delivery platform and continuously provide iron ions for • OH-production through the Fenton reaction.

Intracellular iron ion release and ferroptosis of ccRCC cells.
To further investigate the intracellular iron ion released from MIL-101(Fe)@RSL3, the NPs were pre- M). The • OH would cause the peroxidation of polyunsaturated fatty acid-containing phospholipids (PUFAs) to form lipid peroxides (L-OOH) and propagate the radical process to adjacent lipid molecules. [22,37] Although there had been reported that an increase in the labile iron pool caused by intracellular iron overload would trigger ferroptosis, cellular iron overload was not su cient to produce enough intracellular • OH by solely depending on the Fenton reaction. [17,38] The strong antioxidant system in cells would block the propagation process of L-OOH, and prevented the generation of highly reactive lipid alkoxyl radicals (L-O•) via Fenton chemistry between Fe 2+ and L-OOH. [9,13] Therefore, it was expected to be an effective way to induce ferroptosis by producing the • OH and inhibiting the antioxidant system to disrupt the redox balance of cells.
As an antioxidant enzyme, GPX4 is essential to maintain lipid homeostasis in cells, preventing the accumulation of lipid peroxidation and blocking ferroptosis. RSL3, a direct inhibitor of GPX4, was expected to be a small molecule drug that induces L-OOH in tumors, but its insolubility and poor pharmacokinetic properties have limited its application. Fortunately, in addition to providing iron ions to cancer cells, MIL-101(Fe) NPs can also serve as an RSL3 delivery platform. To evaluate the inhibition e ciency of RSL3, nicotinamide adenine dinucleotide phosphate (NADPH) levels, which indirectly re ect GPX4 activity, were detected. MIL-101(Fe) did not cause a loss of activity of GPX4, showing activity analogous to that of the control group. In contrast, intracellular NADPH remained almost constant in the free RSL3 and MIL-101(Fe)@RSL3 treatment groups, indicating a decrease in GPX4 activity ( Figure 3A). Moreover, GPX4 expression showed a remarkable decrease after RSL3 and MIL-101(Fe)@RSL3 treatment ( Figure 3B), which can be easily explained by the mechanism of action of RSL3 as a powerful GPX4 inhibitor. This result implied that RSL3 could be released from MIL-101(Fe) NPs and exert an inhibitory effect. Then, BODIPY-C11, a membrane-targeted uorescence probe of L-OOH, was used for staining the various treatment groups. [39] MIL-101(Fe) NPs only slightly increased the oxidized BODIPY-C11 signal, whereas RSL3 and MIL-101(Fe)@RSL3 strongly enhanced the uorescence of oxidized BODIPY-C11 ( Figure 3C), indicating a breakdown of the redox equilibrium. In addition, the intracellular oxidative stress of 786-O cells after various treatments was carefully studied by utilizing DCFH-DA as the intracellular total ROS probe. Both the uorescence images ( Figure S6A) and quantitatively analyst results ( Figure  S6B) showed that MIL-101(Fe)@RSL3 triggered ROS burst and accumulation, which was owing to the iron overload and the GPX4 inhibition. Therefore, the accumulation of lipid ROS induced by MIL-101(Fe)@RSL3 is expected to trigger more violent ferroptosis.
Cell viability assays indicated that MIL-101(Fe)@RSL3 showed signi cant cytotoxicity against ccRCC cells compared to RSL3 treatment. Even at a low concentration of loaded RSL3 (1.4 μM), the cell survival rate in the MIL-101(Fe)@RSL3 group was only ~28% ( Figure 3D). In contrast, the high concentration of MIL-101(Fe) NPs showed negligible cytotoxicity. Furthermore, live/dead cell uorescence staining showed that MIL-101(Fe)@RSL3 induced evident cytotoxicity compared with the other treatment groups ( Figure  3E). However, normal cells, such as human hepatocyte cells (LO2) and human renal proximal tubular cells (HK-2), maintained a high cellular survival rate even at 11.2 μM MIL-101(Fe)@RSL3 NPs, which was 8 times higher than the concentration for ccRCC treatment ( Figure S7). Interestingly, MIL-101(Fe)@RSL3 did not cause severe cytotoxicity in normal human hepatocytes or renal cells, and it is well known that the liver and kidney are vital organs that metabolize NPs.
In vivo antitumor e ciency of MIL-101(Fe)@RSL3 against ccRCC Good compatibility is a prerequisite for the clinical application of nanodrugs. The negligible cytotoxicity of MIL-101(Fe)@RSL3 in normal cells inspired us to further investigate the antitumor effects of MIL-101(Fe)@RSL3 in ccRCC. As we expected, after incubation with a high concentration of MIL-101(Fe)@RSL3 (2 mg/mL) for 24 hours, the haemolysis rate was only 0.5%, which can almost be ignored ( Figure S8). Then, the nanodrugs were labelled with ICG to analyze their tissue distribution and tumor accumulation in vivo. After intravenous injection of ICG-MIL-101(Fe)@RSL3, the whole bodies of tumorbearing mice showed bright uorescence, and the tumor region showed an obvious uorescence enhancement ( Figure 4A). This result suggested that ICG-MIL-101(Fe)@RSL3 accumulated to a great extent in tumors, especially one-hour post-injection. Subsequently, the uorescence gradually faded away; however, the tumor region remained bright up to 12 hours later ( Figure 4B). The blood circulation behavior of ICG-MIL-101(Fe)@RSL3 was determined by evaluating the ICG uorescence from the blood samples collected via the tail vein at various time points, and the half-life of MIL-101(Fe)@RSL3 was calculated to be ~2.5 hours ( Figure 4C). Then, the main organs and tumor were isolated for uorescence imaging. Similar to the uorescence imaging in vivo, the uorescence intensity of the tumor was signi cantly higher than that of the heart, lung, etc. (Figure 4D), while the high-intensity uorescence in the liver, intestine, etc. suggested that the MIL-101(Fe)@RSL3 NPs would be cleared by the liver and nally excreted in bile and faeces ( Figure 4E).
Antitumor effects from the MIL-101(Fe)@RSL3 NPs were detected after intravenous administration.
Tumor growth was constantly recorded over 21 days ( Figure S9), and the changes in the tumor volumes indicated that MIL-101(Fe)@RSL3 remarkably inhibited tumor growth ( Figure 5A). It is worth noting that the RSL3 treatment group showed only limited tumor suppression, which can probably be attributed to the insolubility of RSL3 in water and its de cient accumulation at the tumor site ( Figure  Compared with the control group, all parameters in the treatment groups appeared to be normal and did not show a signi cant difference (P value>0.05). This result indicated that neither RSL3 nor MIL-101(Fe) NPs caused obvious in ammation in mice. In addition, blood biochemical analyses were carried out, and various parameters, including alanine transaminase (ALT), aspartate transaminase (AST), creatinine (CRE) and blood urea nitrogen (BUN), were examined ( Figure 6B). There was no meaningful difference between the groups, suggesting that MIL-101(Fe)@RSL3 did not cause obvious hepatic or kidney toxicity to mice. During the treatment process, the body weights of the mice stably increased ( Figure 6C). Then, histological analysis of the major organs, including the heart, liver, spleen, lung, and kidney, was conducting by slicing for H&E staining after 21 days ( Figure 6D). No apparent histological abnormalities or lesions could be observed in any of the treatment groups throughout the whole treatment period. Especially, there was no noticeable damage in the liver or kidneys, which con rmed the biocompatibility of MIL-101(Fe)@RSL3.

Discussion
It was well known that cancer cells have a higher demand for iron than the normal cells to support their rapid proliferation. [40,41] In addition, most of cancer cells showed upregulated expression of transferrin receptor 1 and hepcidin along with downregulated expression of ferroportin, enabling cancer cells to accumulate iron highly e ciently via transferrin receptors. [42,43] However, the intracellular concentration of iron ions still lies at a low level, making it insu cient to induce ferroptosis by catalyzing the Fenton reaction due to the massive demand of rapid growth cancer cells. [15,31] In this study, iron-based MOFs nanoparticles with pH-responsive degradation properties could be uptake via endocytosis and release Fe 3+ ions in lysosomes for remarkably increasing the intracellular content of Fe 3+ ions. Subsequently, the intracellular Fe 2+ referred to as the labile iron pool (LIP) also increase accordingly, because Fe 3+ ions reduced to Fe 2+ by the intracellular iron reductase (e.g., STEAP3) and superoxide. [44,45]  Therefore, it is reasonable to believe that further elevating the intracellular oxidative stress by combining with chemotherapy, PDT and radiotherapy was expected to aggravate ferroptosis and achieve better curative effect. [46][47][48]

Conclusion
In this study, MIL-101(Fe) nanoparticles rich in iron were used to load a ferroptosis inducer (RSL3) for targeted delivery and responsive release. MIL-101(Fe) showed high encapsulation e ciency and the tumor-targeted delivery of RSL3. Under the acidic tumor microenvironment, MIL-101(Fe)@RSL3 with gradual degradation released iron ions and RSL3, thereby aggravating ferroptosis in ccRCC cells by triggering iron overload and inhibiting GPX4. Therefore, compared with free RSL3, MIL-101(Fe)@RSL3 showed more remarkable antitumor e cacy against ccRCC, which displays an abnormal PUFAs metabolic state, while these nanodrugs did not cause obvious cytotoxicity in normal cells even at very high concentrations. In addition, MIL-101(Fe)@RSL3 caused negligible side effects, enabling it to be a promising nanodrug for systemic ccRCC therapy.