(1S,3R)-RSL3 (cat. HY-100218A), erastin (cat. HY-15763), and Mito-TEMPO (cat. HY-112879) were purchased from MedChemExpress (MCE). Liposomal 2000 Transfection Reagent (cat. 40802ES03), propidium iodide (cat. 40710ES03), and polybrene (cat. 40804ES76) were purchased from Yeasen. Puromycin (cat. ant-pr-1) was purchased from InvivoGen. SYTOX™ Green nucleic acid stain (cat. C1181S), GSH/GSSG Assay Kit (cat. S0053), JC-1 Mitochondrial Membrane Potential Assay Kit (cat. C2005), Blasticidin S HCl (cat. ST018), MPTP Assay Kit (cat. C2009S), and Lipid Peroxidation (MDA) Assay Kit (cat. S0131S) were purchased from Beyotime Biotechnology. RIPA Lysis Buffer (cat. AR0101 and AR0103) were purchased from Boster Biological Technology. Protease and Phosphatase Inhibitor Cocktail (cat. P002) was purchased from NCM Biotech. Bradford 1× Dye Reagent (cat. 5000205) was purchased from Bio-Rad. Cell Counting Kit-8 (cat. K1018) was purchased from APExBIO. BODIPY™ 581/591 C11 (cat. D3861) and MitoSOX™ Red mitochondrial superoxide indicator (cat. M36007) were purchased from Thermo Fisher Scientific. IKE (cat. HF3347) was purchased from Hi-Future, while ZK53 (cat. T80725) and Liproxstatin-1 (Lip-1; cat. T2376) were purchased from TargetMol.

HT-1080, HeLa, HEK-293T, and HCT-116 cell lines were obtained from the Cell Bank of the Chinese Academy of Sciences. HT-1080, HeLa, and HEK-293T cells were cultured in DMEM Medium, while HCT-116 cells were cultured in McCoy’s 5A Medium (Meilunbio). All culture media were supplemented with 10% fetal bovine serum (ExCell Bio) and 1% penicillin-streptomycin (Gibco), and the cells were maintained at 37°C with 5% CO₂ in a humidified incubator. All cell lines were regularly tested and confirmed to be free from mycoplasma contamination.

A mitochondrial-targeted library containing ~900 compounds was purchased from MedChemExpress (MCE) (Catalog number: HY-L089). HT-1080 cells were seeded into 96-well plates at a density of 6 × 10³ cells per well. On the second day, the cells were treated with 5 µM of compounds, which were co-incubated with either DMSO or RSL3 (100 nM). After 12 hours of incubation, cell viability was assessed using the CCK8 assay. Cell viability (%) was calculated using the following formula:

where OD represents the optical density at 450 nm for the treated group (OD_treated), the control group (OD_control), and the blank group (OD_blank).

The HA-tagged wild-type ClpP construct was generated by cloning the full-length ClpP cDNA into the pCDH vector. Short hairpin RNAs (shRNAs) targeting ClpP were cloned into the pLKO lentiviral vector. The following shRNA sequences were used: sh-ClpP:5’-GCCCATCCACATGTACATCAA-3’ and 5’-GCTCAAGAAGCAGCTCTATAA-3’.

All primers and shRNA sequences used in the experiments were synthesized by Tsingke Biotechnology. All plasmid constructs were verified by sequencing. Stable cell lines were selected and analyzed by Western blotting to confirm knockdown or overexpression efficiency.

HEK-293T cells were seeded in 60 mm dishes and transfected with the plasmid of interest, psPAX2 packaging plasmid, and pMD2.G-VSV-G envelope plasmid at a 4:3:1 ratio using Liposomal Transfection Reagent (Yeasen) when cell confluence reached 90%. Fresh medium was replaced 6 hours post-transfection. The culture medium was collected 48 hours later, filtered through a 0.22 µm sterile filter, and supplemented with 10 μg/mL polybrene (Yeasen) for infection of HT-1080 cells. Lentivirus-transduced cells were selected for 48 hours with 2 μg/mL puromycin (Invitrogen) or 10 μg/mL blasticidin S HCl (BSD, Beyotime). Stable cell lines were confirmed by Western blotting.

Cells were trypsinized, washed twice with ice-cold PBS, and lysed in lysis buffer (BOSTER) with a protease inhibitor cocktail (NCM Biotech). The supernatant was collected after centrifugation at 13,000 × g for 20 minutes. Protein concentration was measured using Bradford reagent (Bio-Rad). Protein were resolved by SDS-PAGE and transferred to PVDF membranes (Millipore). Membranes were blocked with 5% nonfat milk for 1 hour, then incubated with primary antibodies followed by HRP-conjugated goat secondary antibodies (Abclonal). Immunodetection was performed with an ECL kit (epizyme) and captured using a chemiluminescence imager (Bio-Rad).

The following blotting reagents and antibodies were used: Rabbit anti-β-Actin (cat. AC038, 1:20,000 for IB), anti-GAPDH (cat. A19056, 1:50,000 for IB), mouse anti-β-Tubulin (cat. AC021, 1:5,000 for IB), HRP-conjugated Goat anti-Mouse IgG (H+L) (cat. AS003, 1:5,000 for IB), HRP-conjugated Goat anti-Rabbit IgG (H+L) (cat. AS014, 1:5,000 for IB) antibodies were purchased from ABclonal. Rabbit anti-CLPP (cat. 15698-1-AP, 1:1,000 for IB), anti-HA tag (cat. 51064-2-AP, 1:5,000 for IB) antibodies were purchased from Proteintech. Mouse anti-GPX4 (cat.sc-166570, 1:500 for IB) antibodies were purchased from Santa Cruz.

Cells were seeded in 96-well plates and cultured overnight at 37°C in a humidified atmosphere containing 5% CO₂. The next day, cells were treated with the indicated drugs for the specified duration. The medium was replaced with PBS containing 10% CCK8 reagent (APEx BIO). After 1 hour of incubation at 37°C, absorbance at 450 nm was measured using a microplate reader (Thermo Scientific). The combination index (CI) for drug interactions was calculated according to the Chou-Talalay method (41). Specifically, a CI value of 1 indicates an additive interaction, CI values greater than 1 indicate antagonism, and CI values less than 1 reflect a synergistic interaction between the drugs.

Cells were seeded in 96-well plates and cultured overnight at 37°C in a humidified atmosphere containing 5% CO₂. The next day, cells were treated with the indicated drugs for the specified duration. Subsequently, cells were incubated with HBSS containing SYTOX™ Green nucleic acid stain (Beyotime) at 37°C for 30 min in the dark. Following incubation, cells were washed once with PBS and then replaced with fresh HBSS. Fluorescence images were acquired directly using an inverted fluorescence microscope equipped with a GFP channel (Ex/Em ≈ 488/500–550 nm).

Cells were seeded in 96-well plates and cultured overnight at 37°C in a humidified incubator with 5% CO₂. The following day, cells were treated with the indicated compounds for the specified durations. After treatment, cells were directly observed without washing to preserve their native morphology and treatment-associated features.

Cell morphology images were acquired using an inverted microscope equipped with a phase-contrast objective lens. Representative fields were selected and captured under identical imaging settings across all conditions to ensure consistency. The acquired images were used to visualize and document morphological changes induced by drug treatment.

Cells were seeded in 6-well plates and cultured overnight at 37°C in a humidified atmosphere containing 5% CO₂. The next day, cells were treated with the indicated drugs for the specified duration. After treatment, cells were stained with propidium iodide (PI) (Yeasen) at 37°C for 30 minutes. Cells were then trypsinized, washed, and resuspended in PBS. The fraction of dead cells was analyzed using a flow cytometer (Advanteon V6B5R3, Agilent), and data were processed with FlowJo software.

Cells were seeded in 6-well plates and cultured overnight at 37°C in a humidified atmosphere containing 5% CO₂. The next day, cells were treated with the indicated drugs for the specified duration. The cells were stained with 5 µM BODIPY™ 581/591 C11 (Thermo Scientific) at 37°C for 30 minutes. After trypsinization, the cells were washed and resuspended in PBS. Lipid peroxidation levels were analyzed using a flow cytometer (Advanteon V6B5R3, Agilent) and data were analyzed using FlowJo software.

Cells were seeded onto 100 mm dishes and cultured overnight at 37°C in a humidified atmosphere containing 5% CO₂. The next day, cells were treated with the indicated drugs for the specified duration. Cells were collected for GSH and GSSG measurement using GSH and GSSG Assay Kit (Beyotime) according to the manufacturer’s protocol. The GSH and GSSG concentrations were calculated using a standard curve and normalized to the total protein level.

Cells were seeded in 6-well plates and cultured overnight at 37°C in a humidified atmosphere containing 5% CO₂. The next day, cells were treated with the indicated drugs for the specified duration. The cells were stained with JC-1 (Beyotime) at 37°C for 30 minutes. After trypsinization, the cells were washed and resuspended in PBS. Mitochondrial membrane potential was analyzed using a flow cytometer (Advanteon V6B5R3, Agilent) and data were processed with FlowJo software.

Cells were seeded in 6-well plates and cultured overnight at 37°C in a humidified atmosphere containing 5% CO₂. The next day, cells were treated with the indicated drugs for the specified duration. The cells were stained with 5 µM MitoSOX (Invitrogen) at 37°C for 30 minutes. After trypsinization, the cells were washed and resuspended in PBS. Mitochondrial ROS levels were analyzed using a flow cytometer (Advanteon V6B5R3, Agilent) and data were processed with FlowJo software.

Cells were seeded in 6-well plates and cultured overnight at 37°C in a humidified atmosphere containing 5% CO₂. The next day, cells were treated with the indicated drugs for the specified duration. The cells were stained using the MPTP Assay Kit (Beyotime) according to the manufacturer’s instructions. After trypsinization, the cells were washed and resuspended in PBS. Mitochondrial permeability transition pore levels were analyzed using a flow cytometer (Advanteon V6B5R3, Agilent) and data were processed with FlowJo software.

This study was conducted in compliance with all relevant ethical guidelines. All animal procedures were approved by the Institutional Animal Care and Use Committee of Zhejiang University and were carried out in accordance with its established protocols, under the ethical approval number ZJU20250193. Four-week-old female, specific pathogen-free (SPF) BALB/c nude mice were obtained from Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China). The mice were provided with standard rodent chow and water ad libitum and were housed in a specific pathogen-free environment with a 12-hour light/12-hour dark cycle.

For the xenograft tumor assay, 6 × 10⁶ HCT-116 cells were subcutaneously injected into the flanks of six-week-old female BALB/c nude mice. When tumors reached a size of 50–100 mm³, the mice were randomly assigned to different treatment groups and administered either vehicle, IKE (25 mg/kg; HI-FUTURE), ZK53 (20 mg/kg; TargetMol), or Lip-1 (10 mg/kg; TargetMol) via intraperitoneal injection every other day. The vehicle consisted of 10% dimethyl sulfoxide (DMSO), 40% polyethylene glycol 300 (PEG300), 5% Tween-80, and 45% saline (v/v/v/v), which served as a common solvent system for all compounds. The maximum allowable tumor burden, as per the ethics committee, was 1,500 mm³, and this threshold was not surpassed during the study.

MDA levels were measured using a lipid peroxidation MDA assay kit (Beyotime). Briefly, fresh tumor tissues were lysed with RIPA buffer (BOSTER). The supernatant was incubated with thiobarbituric acid (TBA) at 100°C for 15 minutes to form MDA-TBA adducts. After centrifugation, the absorbance of the supernatant at 532 nm was measured using a microplate reader (Thermo Scientific). MDA levels were normalized to protein concentration.

Fresh tissues were fixed in 4% paraformaldehyde for 48 hours, dehydrated, and embedded in paraffin. Five-micrometer sections were cut for immunohistochemistry (IHC). Sections were incubated overnight at 4°C with primary antibodies against Ki-67 (1: 300, ab16667, Abcam) and 4-HNE (1:400, ab46545, Abcam). Images were captured from five random regions per tumor at 200× magnification using an EVOS M7000 microscope (Invitrogen). The expression of Ki-67 and 4-HNE was quantified using ImageJ software by measuring the integrated optical density (IOD) of positive staining in each section.

All experiments were performed with at least three independent replicates. Statistical analyses were conducted using GraphPad Prism V8. Data are presented as mean ± SD. Statistical significance was assessed using an unpaired Student’s t-test, one-way ANOVA, or two-way ANOVA, as appropriate. A p-value of < 0.05 was considered statistically significant.

To identify potential ferroptosis sensitizers, we conducted a high-throughput screening of a mitochondrial-targeted drug library using the HT-1080 cell line, a well-established model for ferroptosis research. Our goal was to discover compounds that could enhance ferroptosis induced by RSL3, while minimizing general cytotoxicity (Figure 1A). This was accomplished by comparing the effects of individual compounds (with cell viability >80%) to the combined treatments with RSL3, where the single RSL3 treatment yielded ~60% cell viability, resulting in cell viability <20% when combined.

Among the identified compounds, we confirmed several inhibitors targeting known ferroptosis-related regulators, including the DHODH-targeting inhibitor hDHODH-IN-4 and the FSP1-targeting inhibitor iFSP1, the targets of which, DHODH (17) and FSP1 (42, 43), have been previously implicated in ferroptosis regulation (Figures 1B-D). These findings validate the robustness and reliability of our screening strategy for identifying ferroptosis sensitizers. Notably, in addition to these compounds targeting previously characterized modulators, we identified ZK53 as a novel hit that significantly increased cellular susceptibility to RSL3 (Figures 1B-D).

To further validate our screening results, we recharacterized the identified compound ZK53 in different cell lines to ensure that its ferroptosis-sensitizing effect was reproducible and not cell line–specific. First, we assessed its cytotoxicity across a range of concentrations. Cell viability assays indicated that low concentrations of ZK53 (below 15 µM for HT-1080 cells, below 1 µM for HeLa cells, and below 10 µM for HCT-116 cells) did not significantly affect cell viability (Figures 1E, I, M). However, at higher concentrations, ZK53 partially inhibited cell viability, which is consistent with previous reports suggesting its potential impact on cell proliferation (40). To minimize cytotoxic effects and ensure the consistency of subsequent experiments, we selected non-toxic concentration of 10 µM for HT-1080 cells, 1 µM for HeLa cells, and 5 µM for HCT-116 cells for all subsequent studies.

Next, we validated the sensitizing effect of ZK53 on RSL3. Cell viability assays, using both concentration gradients of the drug combination and individual agents, demonstrated that ZK53 significantly enhanced RSL3-induced cell death across different cell lines (Figures 1F, G, J, K, N, O). The drug interaction between ZK53 and RSL3 was assessed by calculating the combination index values (Figures 1Q-S), which revealed a significant synergistic effect of ZK53 when combined with RSL3. Similar results were obtained through Sytox staining, further confirming the enhanced cell death induced by the combination treatment (Figures 1H, L, P).

In sum, these results suggest that ZK53 may serve as a promising candidate for therapeutic strategies targeting ferroptosis.

Given that RSL3 targets not only GPX4 but also other potential targets (44), we conducted further investigations to confirm that the sensitizing effect of ZK53 on RSL3-induced cell death is mediated via ferroptosis.

In HT-1080 cells, the classical ferroptosis inhibitor Lip-1 completely abolished the sensitizing effect of ZK53 on RSL3-induced cell death (Figure 2A). Lipid peroxidation, a key hallmark of ferroptosis, was assessed using the lipid peroxidation sensor BODIPY staining. The results showed that ZK53 significantly enhanced RSL3-induced lipid peroxidation, an effect that was fully reversed by Lip-1 (Figure 2B). Consistently, the intracellular GSH/GSSG ratio, serving as a quantitative indicator of the redox equilibrium (45, 46), and GPX4 protein expression, representing the cellular antioxidant defense against lipid peroxide accumulation (2, 47), were both significantly reduced following combined treatment with ZK53 and RSL3 (Figures 2C, D). In parallel experiments, we employed another ferroptosis inducer, erastin, and obtained similar results (Figures 2E–H). These findings were further validated in HeLa (Figures 2I–L) and HCT-116 cells (Figures 2M–P).

Together, our data confirm that ZK53 synergizes with ferroptosis inducers such as RSL3 and erastin to promote ferroptosis, supporting its potential as a therapeutic sensitizer for ferroptosis-based treatments.

Having established that ZK53 sensitizes cells to ferroptosis, and given that ZK53 is a selective inhibitor of the human Clp protease subunit (HsClpP) (40), we next explored the potential involvement of ClpP in the regulation of ferroptosis. ClpP plays a critical role in mitochondrial homeostasis and stress responses by degrading misfolded or damaged proteins, especially under oxidative stress conditions (36, 48–50).

We hypothesized that ClpP might contribute to the cellular mechanisms underlying ferroptosis. To investigate this, we performed functional studies in HT-1080 cells with ClpP knockdown using specific shRNA (Figure 3A). Cell morphology, cell death, lipid ROS levels, and GPX4 protein expression were examined in both control and ClpP-knockdown HT-1080 cells.

ClpP depletion significantly attenuated cell death induced by both RSL3 and erastin (Figures 3B, C, F, G). Consistently, knockdown of ClpP markedly inhibited lipid peroxidation and preserved GPX4 protein levels in response to RSL3 and erastin treatment (Figures 3D, E, H, I).

To determine whether ferroptosis inducers affect ClpP expression itself, we assessed ClpP protein levels in cells treated with RSL3 or erastin. As shown in Figures 3J, K, ClpP expression remained largely unchanged after treatment with either compound. These findings indicate that RSL3 and erastin do not regulate ClpP expression directly. Instead, the observed synergistic effect of ZK53 with RSL3 is likely mediated through ClpP’s functional involvement in ferroptosis rather than changes in its abundance.

To further confirm the role of ClpP in ferroptosis, we overexpressed ClpP in HT-1080 cells (Figure 4A). ClpP overexpression potentiated RSL3- and erastin-induced ferroptosis, as evidenced by increased cell death and lipid peroxidation, along with a concomitant decrease in GPX4 protein levels (Figures 4B–G). These effects were completely abrogated by the ferroptosis inhibitor Lip-1 (Figures 4C, D, F, G).

Building on these findings, we next examined whether hyperactivation of ClpP could also enhance ferroptosis. The Y118A mutation in ClpP, a gain-of-function mutation, increases its protease activity and leads to hyperactivation of ClpP’s proteolytic function (51, 52). Overexpressing ClpP Y118A in HT-1080 cells resulted in significantly increased sensitivity to RSL3-induced cell death and lipid peroxidation compared to control cells (Figures 4H-J).

Together, these results further confirm that ClpP and its proteolytic activity play an important role in promoting ferroptosis.

While previous studies have established that ZK53 acts as a ClpP activator—identified through screening and later optimized for its potency (40)—it remains possible that ZK53 could promote ferroptosis through pathways independent of ClpP.

To explore whether the ferroptosis-sensitizing effect of ZK53 is reliant on ClpP, we treated both wild-type and ClpP-deficient HT-1080 cells with ZK53, in combination with ferroptosis inducers. Our analysis of cell death and lipid peroxidation showed that ZK53’s ability to enhance RSL3-induced ferroptosis was significantly diminished in ClpP-deficient cells compared to wild-type controls (Figures 5A, B). Similarly, when we used erastin as an alternative ferroptosis inducer, the sensitizing effects of ZK53 on both cell death and lipid peroxidation were also notably reduced in the absence of ClpP (Figures 5C, D).

Although the ferroptosis-sensitizing effect of ZK53 was substantially attenuated by ClpP knockdown, a modest but significant residual enhancement persisted compared with treatment with RSL3 or erastin alone. This remaining activity may result from incomplete suppression of ClpP or potential ClpP-independent effects of ZK53.

Together, these findings indicate that the sensitizing effect of ZK53 on ferroptosis is primarily mediated through ClpP, and this mechanism is not specific to the choice of ferroptosis inducer.

Activation of ClpP has been shown to disrupt mitochondrial protein homeostasis, affecting critical mitochondrial components, including electron transport chain (ETC) subunits and TCA cycle enzymes like aconitase (ACO2) (51, 53). Given the pivotal role of mitochondrial dysfunction in ferroptosis, we hypothesized that ZK53 might amplify ferroptosis through its effects on mitochondrial integrity.

To test this hypothesis, we first assessed key mitochondrial function indicators in HT-1080 cells following ZK53 treatment. Although no significant changes were observed in mitochondrial permeability transition pore (MPTP) opening (Figure 6A), ZK53 treatment caused a concentration-dependent decrease in mitochondrial membrane potential (ΔΨm) (Figure 6B), as indicated by JC-1 staining. Additionally, MitoSOX staining showed a marked, concentration-dependent increase in mitochondrial ROS production, which was effectively mitigated by the mitochondrial ROS scavenger MitoTEMPO (Figure 6C). These findings suggest that ZK53 treatment disrupts mitochondrial normal function.

We then examined whether co-treatment with ZK53 and the ferroptosis inducer RSL3 would further exacerbate mitochondrial dysfunction. Indeed, combined treatment led to a more pronounced reduction in mitochondrial membrane potential, and a further increase in mitochondrial ROS levels compared to treatment with either agent alone (Figures 6D, E). This synergistic effect suggests that ZK53 and RSL3 collaborate to amplify mitochondrial damage.

To further validate the role of mitochondrial dysfunction in ZK53-induced ferroptosis, we examined whether mitochondrial ROS scavengers could attenuate the ferroptosis-sensitizing effect of ZK53. Treatment with MitoTEMPO markedly mitigated the enhancement of RSL3-induced ferroptotic cell death and lipid peroxidation caused by ZK53 (Figures 6F–H), as evidenced by cell morphology imaging, PI staining, and lipid peroxidation assays. These results demonstrate that mitochondrial ROS are essential mediators of the ZK53/ClpP-dependent ferroptosis amplification, supporting the critical role of mitochondrial dysfunction in ZK53-induced ferroptotic signaling.

In conclusion, our data demonstrate that ZK53 promotes ferroptosis through mitochondrial dysfunction, with mitochondrial ROS generation serving as a key driver of this process.

Targeting ferroptosis has emerged as a promising strategy in cancer therapy, but the effectiveness of small-molecule inducers is often limited by resistance mechanisms. Consequently, combination therapies have become a key focus of research to enhance ferroptosis-based treatments (8).

Given the significant ferroptosis-sensitizing effect of ZK53 observed in vitro, we further assessed its therapeutic potential in vivo. To this end, we established a xenograft tumor model by subcutaneously implanting HCT-116 cells (a ferroptosis-resistant cell line) into BALB/c nude mice, followed by treatment with ZK53, IKE, or their combination, along with a ferroptosis inhibitor treatment group (Figure 7A).

The results demonstrated that while treatment with either low-dose ZK53 or IKE alone did not significantly inhibit tumor growth, ZK53 markedly enhanced tumor sensitivity to SLC7A11 inhibition (Figure 7B). Specifically, combination therapy led to a substantial reduction in tumor volume and weight, with the tumor size being less than half of that observed with single-agent treatments (Figures 7C, D). Further analysis of ferroptosis markers, including MDA and 4-HNE, showed significantly increased levels in the combination treatment group compared to the single-agent groups, indicating enhanced ferroptosis. Meanwhile, Ki-67 levels were notably decreased in the combination group, reflecting reduced tumor cell proliferation (Figures 7E, G–I). This synergistic effect was effectively reversed by the ferroptosis inhibitor Lip-1, further confirming the role of ZK53 in promoting ferroptosis. Notably, no significant changes in body weight were observed during the treatment (Figures 7F), indicating that the combination therapy did not cause overt toxicity or adverse effects in the animals.

In conclusion, ZK53 enhances tumor regression by amplifying ferroptosis, highlighting its potential as a powerful ferroptosis-sensitizing agent for cancer therapy.