Ferroptosis Induction in Glioma by Calceolarioside A via Modulation of the PI3K/Akt/Nrf2 Pathway

Gliomas, particularly high-grade glioblastomas, represent one of the most lethal forms of brain tumors, characterized by rapid proliferation, treatment resistance, and poor prognosis¹. In recent years, various anti-glioma agents have been explored, including traditional chemotherapeutics like temozolomide and a range of natural compounds such as curcumin², resveratrol³, and epigallocatechin gallate⁴. These agents exert their effects through multiple mechanisms, including induction of apoptosis, inhibition of angiogenesis, and modulation of oxidative stress pathways. However, many of these compounds face limitations such as poor bioavailability or incomplete mechanistic characterization. This highlights the critical need for comprehensive methodologies that not only confirm anti-tumor activity but also systematically dissect the underlying pharmacological mechanisms. Establishing such mechanistic workflows is especially important for accelerating the development of novel anti-glioma compounds from discovery to preclinical validation.

To support this goal, an increasing number of studies now integrate advanced computational and experimental tools to enhance mechanistic resolution^5,6. Structure-based drug design and artificial intelligence (AI) models -- such as AlphaFold3 and molecular docking -- are facilitating the identification of promising drug-target interactions with unprecedented speed and accuracy^7,8. These tools complement experimental methods by predicting binding affinities and identifying potential molecular targets before laboratory validation. At the same time, the emergence of pathway-based functional assays, live-cell imaging, and high-resolution proteomic techniques allow researchers to evaluate the cellular impact of new compounds across multiple regulatory axes, including oxidative stress responses, signal transduction cascades, and cell death modalities such as ferroptosis^9,10,11.

Among the signaling pathways involved in glioma biology, the Phosphatidylinositol 3-kinase (PI3K)/Protein Kinase B (Akt) and Nuclear factor erythroid 2-related factor 2 (Nrf2) pathways are of particular interest due to their roles in tumor growth, redox homeostasis, and therapy resistance^12,13. PI3K/Akt signaling promotes survival, proliferation, and metabolic adaptation in glioma cells, while Nrf2 functions as a master regulator of the antioxidant response, suppressing ferroptotic cell death¹⁴. Persistent activation of these pathways contributes to treatment evasion, making them attractive targets for pharmacological intervention¹⁵. Conventional methods for pathway interrogation include western blotting for total and phosphorylated proteins, as well as qRT-PCR for gene expression analysis. However, in-depth analysis of Nrf2 signaling requires additional assays that assess protein stability, nuclear translocation, and post-translational modifications such as ubiquitination. These multifaceted detection strategies are essential for understanding how specific compounds disrupt or reinforce redox regulation and ferroptotic control.

Calceolarioside A (CaA), a hydroxycinnamic acid derivative, offers a dual mechanism by simultaneously inhibiting PI3K/Akt signaling and promoting proteasomal degradation of Nrf2, thus enhancing ferroptotic sensitivity. This dual-targeting strategy may improve reproducibility and mechanistic depth over existing agents like erastin or RSL3, which act downstream of antioxidant defenses. This protocol presents an integrated experimental approach to investigate the pharmacological mechanism of CaA, which includes molecular docking to identify PI3K binding sites, immunoblotting and immunoprecipitation to detect Nrf2 expression and ubiquitination, and ferroptosis markers such as reactive oxygen species (ROS), glutathione (GSH), and malondialdehyde (MDA). Furthermore, a xenograft mouse model is used to assess CaA's therapeutic efficacy and systemic safety. This comprehensive workflow provides a reproducible and informative platform for studying the mechanism of action of anti-cancer agents, with emphasis on integrating signal pathway modulation and cell death regulation.

Cell culture and treatment
Murine HT22 neuronal cells and human glioma U87 and U251 cell lines were cultured in Dulbecco's Modified Eagle Medium supplemented with 10% fetal bovine serum and 4 mM L-glutamine, under humidified conditions at 37 °C with 5% CO₂. To streamline experimental consistency, all cell treatments were standardized and described here. This section outlines the specific conditions used in each treatment group and the corresponding downstream assays.

CaA treatment:
HT22 cells were treated with increasing concentrations of CaA (0, 0.1, 0.5, 1, 10, 50, 100, 250, and 500 µM) for 72 h. U87 and U251 glioma cells were treated with CaA at concentrations of 5, 20, or 50 µM for 24 h for most experiments, or with 50 µM for 0, 6, 12, or 18 h for time-course studies. These treatments were used in the following analyses: western blotting for cysteine/glutamate antiporter subunit (xCT), glutathione peroxidase 4 (GPX4), ferritin (FTH1), glutaminase, and PI3K/Akt/mammalian target of Rapamycin (mTOR)/ eukaryotic Initiation Factor 4B (EIF4B) (total and phosphorylated levels); real-time quantitative polymerase chain reaction (RT-qPCR) for Nrf2 and Kelch-like ECH-associated protein 1 (Keap1); ROS measurement using 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA); GSH and MDA quantification; FerroOrange-based Fe²⁺ detection; 5-ethynyl-2′-deoxyuridine (EdU) proliferation assay; Colony formation assay; 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) cell viability assay.

Liproxstatin-1 (Lip1) co-treatment:
Cells were pretreated with 50 nM Lip1 for 2 h before the addition of 50 µM CaA. These conditions were used in MTT assay (0, 24, 48, 72 h); colony formation assay; EdU assay; Annexin V/PI apoptosis assay (with/without 20 nM cisplatin).

MG132 co-treatment:
Cells were treated with 10 µM MG132 alone or combined with 50 µM CaA for 24 h. This was applied to western blotting for Nrf2 protein stabilization (total and nuclear); Immunoprecipitation (IP) assays for ubiquitinated Nrf2 and Keap1.

Cisplatin co-treatment:
Cells were treated with 20 nM cisplatin alone or in combination with 20 µM CaA for 24 h. These treatments were used in Annexin V/propidium iodide (PI) apoptosis assay.

Nrf2 overexpression:
Cells were transfected with 2.5 µg Nrf2 overexpression plasmid or empty vector using transfection agent for 18 h, then treated with 50 µM CaA for 24 h. This condition was used in western blotting for xCT, GPX4, FTH1, and glutaminase; ROS, MDA, GSH assays; FerroOrange staining for intracellular iron.

A workflow diagram (Supplemental Figure S1) summarizes the above experimental conditions and corresponding analyses. This figure visually links each treatment protocol to its specific downstream assays for improved clarity and reproducibility.

Assessing cell viability
Cell viability was assessed using the conventional MTT colorimetric method as previously described¹⁶. HT22, U251 and U87 cells were plated into 96-well culture plates and incubated overnight to allow adherence. After completing the treatment period, 15 µL of MTT solution was introduced into each well and the plates were incubated for 4 h at 37 °C. The medium was then discarded, and 150 µL of dimethyl sulfoxide (DMSO) was added to solubilize the formed formazan crystals. Absorbance was subsequently recorded at 490 nm using a microplate reader. Higher absorbance values indicate greater cell viability, whereas a reduction in absorbance suggests cytotoxic effects.

EdU incorporation assay
For assessing DNA synthesis and cell proliferation, EdU incorporation assays were performed using an EdU Imaging Kit as previously described¹⁷. Briefly, U251 and U87 glioma cells were seeded in 24-well plates. After treatment, cells were incubated with EdU (10 µM final concentration) and cultured for an additional 24 h. Following the incubation step, the cells were fixed using 4% paraformaldehyde and then permeabilized with 0.5% Triton X-100., and stained according to the kit instructions. The percentage of EdU-positive cells (proliferating cells) was quantified by fluorescence microscopy and normalized to total 4′,6-diamidino-2-phenylindole (DAPI)-stained nuclei.

Colony formation assay
For the colony formation assay, U251 and U87 glioma cells were plated in 3.5 cm dishes at low density (500 cells per dish). Following treatment, the culture medium was exchanged for fresh, drug-free medium, and the cells were maintained for another 14 days to permit colony formation, with medium changes performed every three days. The resulting colonies were subsequently fixed using 4% paraformaldehyde, stained with 0.1% crystal violet., and counted manually under a microscope. Clusters exceeding 75 µm in diameter were scored as colonies. Each experiment was performed in triplicate and repeated three times. This assay followed commonly accepted protocols for evaluating long-term proliferative capacity¹⁸. Colonies should appear as dense, circular clusters; fewer and smaller colonies suggest reduced proliferative capacity.

Assessing cell apoptosis
Apoptotic cell death was assessed using the FITC-Annexin V apoptosis detection kit as previously described¹⁶. U87 and U251 glioma cells were treated with 20 nM cisplatin in the presence or absence of 20 µM CaA for 24 h prior to staining. For experiments involving ferroptosis inhibition, cells were pretreated with 50 nM Lip1 for 2 h, followed by co-treatment with 20 µM CaA and 20 nM cisplatin for an additional 24 h. After the indicated treatments, cells were harvested (300 × g, 5 min) and stained with Annexin V-FITC and PI according to the manufacturer's instructions. After labeling, samples were examined on a flow cytometer. Data acquisition was performed using FL1 (FITC) and FL2 (PI) channels, with compensation settings applied to minimize spectral overlap. The proportion of Annexin V-positive cells was subsequently calculated to determine the degree of apoptosis.

Western blot analysis
Western blotting was conducted according to established procedures^19,20. Briefly, after the indicated treatments, cells were lysed in RIPA buffer containing protease and phosphatase inhibitors, and the protein levels were determined using the Bradford assay. For experiments assessing Nrf2 subcellular distribution, nuclear and cytoplasmic proteins were extracted using a commercial nuclear extraction kit according to the manufacturer's instructions.

A total of 40 µg of protein from each sample was resolved on a 10% SDS-PAGE gel and subsequently transferred onto PVDF membranes. The membranes were blocked with 5% non-fat milk prepared in PBST (PBS with 0.1% Tween-20) for 1 h at room temperature, followed by overnight incubation with primary antibodies at 4 °C. After washing, HRP-conjugated secondary antibodies (1:5,000) were applied for 1 h at 37 °C, and the signals were detected using an enhanced chemiluminescence (ECL) system.

The expression levels of xCT, glutaminase, GPX4, and FTH1 were measured, using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the loading control. Phosphorylated proteins (p-PI3K, p-Akt, p-mTOR, and p-EIF4B) were quantified relative to their respective total protein levels, while the total proteins were normalized against GAPDH. Nuclear Nrf2 was normalized to Lamin B. Expected outcomes include a dose- and time-dependent reduction in Nrf2 and p-PI3K/Akt/mTOR/EIF4B, and downregulation of ferroptosis-related markers in CaA-treated cells.

For ubiquitination assays, cells were treated with MG132 (10 µM) with or without CaA (50 µM) for 24 h. After lysis, protein complexes were immunoprecipitated using anti-Nrf2 or anti-Keap1 antibodies; normal rabbit IgG was used as a negative control. The immunoprecipitates were then probed with anti-ubiquitin antibodies to detect ubiquitinated forms, while total lysates were also probed for Nrf2 and Kelch-like ECH-associated protein 1 (Keap1) levels to confirm input.

RNA extraction and quantitative real-time PCR (qRT-PCR)
RNA extraction and qRT-PCR were performed following standard protocols²¹. Briefly, U87 and U251 glioma cells were treated with or without 50 µM CaA for different durations (6 and 18 h). Total RNA was isolated and converted into cDNA through reverse transcription. Quantitative real-time PCR (qRT-PCR) was performed using a SYBR Green qPCR Master Mix on a real-time PCR platform. Relative gene expression was determined using the 2-ΔΔCt calculation. The primer sequences used for Nrf2, Keap1, and the housekeeping gene GAPDH were as follows:

Nrf2 forward, 5′-CACATCCAGTCAGAAACCAGTGG-3′;
Nrf2 reverse, 5′-GGAATGTCTGCGCCAAAAGCTG-3′

Keap1 forward, 5′-CAACTTCGCTGAGCAGATTGGC-3′;
Keap1 reverse, 5′-TGATGAGGGTCACCAGTTGGCA-3′

GAPDH forward, 5′-CGGAGTCAACGGATTTGGTCG-3′;
GAPDH reverse, 5′-AGCCTTCTACATGGTGGTGAAGAC-3′

Nrf2 transduction
U251 and U87 glioma cells were transiently transfected with either an Nrf2 overexpression plasmid or a corresponding empty control vector using the lipid-based transfection reagent; the procedure was based on conventional transient transfection methods²². Briefly, cells were seeded in 6-well plates at a density of 3 × 10⁵ cells per well and cultured overnight to achieve approximately 70-80% confluence at the time of transfection. For each well, 2.5 µg of plasmid DNA was diluted in 125 µL of Reduced Serum Medium and mixed with 5 µL of the transfection reagent. Separately, 3.75 µL of the transfection reagent was diluted in another 125 µL of the reduced serum medium. The diluted DNA and transfection reagent mixtures were then combined, gently mixed, and incubated for 10-15 min at room temperature to allow complex formation. The resulting transfection complexes (total volume 250 µL) were added dropwise to each well containing cells in fresh complete medium.

After 18 h of incubation at 37 °C in a CO₂ incubator, the transfection medium was replaced with fresh culture medium. Cells were then subsequent assays, including western blotting, ROS analysis, MDA/GSH quantification, and iron accumulation studies. This protocol follows widely accepted transient gene expression procedures and allows for robust Nrf2 expression within 24-48 h post transfection.

Measurement of intracellular iron
Intracellular Fe²⁺ levels were measured using the fluorescent indicator FerroOrange. Brief, U251 and U87 glioma cells were seeded in 24-well plates and treated with varying concentrations of CaA (5, 20, and 50 µM). To assess the effect of Nrf2 overexpression on ferroptosis, U87 and U251 cells were transfected with an Nrf2 expression plasmid or empty vector. After 18 h, medium was replaced, and cells were treated with 50 µM CaA for 24 h. Subsequently, cells were exposed to 1 µM FerroOrange in serum-free medium and incubated at 37 °C for 30 min before analysis. After incubation, cells were rinsed and observed under a fluorescence microscope, using a TRITC filter set (excitation ~543 nm, emission 560-600 nm), where increased orange fluorescence intensity indicates higher levels of labile Fe²⁺, typically localized in the cytoplasm. This method was adapted from established fluorescent probe-based iron assays²³.

Measurement of GSH and MDA
Intracellular GSH and MDA levels were determined following standardized evaluation protocols to evaluate oxidative stress²⁴. Brief, U251 and U87 glioma cells were seeded in 24-well plates and treated with varying concentrations of CaA (5, 20, and 50 µM). To assess the effect of Nrf2 overexpression on ferroptosis, U87 and U251 cells were transfected with an Nrf2 expression plasmid or empty vector. After 18 h, medium was replaced and cells were treated with 50 µM CaA for 24 h. Subsequently, cells were lysed in RIPA buffer, and total protein concentrations were measured using a BCA protein quantification kit. GSH and MDA contents were subsequently analyzed using commercial detection kits in accordance with the manufacturers' protocols.

ROS assay
Intracellular ROS generation was assessed using the fluorescent probe DCFH-DA in accordance with established protocols²⁵. Brief, U251 and U87 glioma cells were seeded in 24-well plates and treated with varying concentrations of CaA (5, 20, and 50 µM). To assess the effect of Nrf2 overexpression on ferroptosis, U87 and U251 cells were transfected with an Nrf2 expression plasmid or empty vector. After 18 h, medium was replaced and cells were treated with 50 µM CaA for 24 h. Subsequently, cells were treated with 5 µM DCFH-DA at 37 °C for 1 h, followed by PBS washes to remove excess dye. Fluorescence signals were visualized using an inverted fluorescence microscope, equipped with a FITC filter (excitation 488 nm, emission 525 nm), and images were captured for analysis. Increased green fluorescence intensity corresponds to elevated intracellular ROS levels.

Animal experiments
All animal experiments were performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Institutional Animal Care and Use Committee of Weifang Peoples Hospital under approval number [2025SDL579]. To generate the glioma xenograft model, 2 × 10⁶ U87 cells were resuspended in 0.2 mL of PBS and subcutaneously inoculated into the left flank of 4-week-old male BALB/c nude mice. After allowing tumors to develop for 7 days, the mice were randomly assigned to three groups (n = 8 per group): untreated control, CaA-treated (2.5 mg/kg), and CaA (2.5 mg/kg) combined with Lip1 (10 mg/kg). Treatments were administered via daily intraperitoneal injection for 14 consecutive days. Tumor dimensions were measured every 3 days using calipers, and volumes were calculated using the formula: volume = (length × width²) / 2. On day 21, all mice were euthanized, and tumors were excised and photographed for further analysis.

RNA sequencing and KEGG pathway analysis
RNA sequencing was performed to profile transcriptional changes following CaA exposure. RNA integrity was confirmed by using a microvolume UV-VIS spectrophotometer and capillary-based electrophoresis system prior to sequencing on the Illumina platform. Clean reads were mapped to the human reference genome (GRCh38) with HISAT2, and gene expression was quantified as FPKM. Differentially expressed genes were identified using DESeq2 (FDR < 0.05, |log₂FC| ≥ 1), and functional enrichment was assessed by KEGG pathway analysis (adjusted P < 0.05).

Molecular docking analysis
Molecular docking was employed to predict the interaction between CaA and the PI3K/Akt pathway. The crystal structure of the human PI3K catalytic subunit p110α (PIK3CA) was obtained from the Protein Data Bank (PDB ID: 8BFU). The structure of CaA was retrieved from PubChem and energy-minimized using molecular modeling software. Using molecular docking preparation software, all water molecules and non-essential ligands were removed from the PIK3CA structure, and hydrogen atoms were added. A docking grid was defined around the active site of PIK3CA (based on known binding site coordinates), and a docking algorithm was used to perform the docking. The top-scoring binding poses of CaA were selected according to binding affinity (kcal/mol) and examined for key interactions. Protein-ligand interactions, such as hydrogen bonds and hydrophobic contacts, were visualized with molecular visualization software and a 2D interaction mapping tool.

Safety and waste disposal
All experiments involving chemical reagents such as DMSO, fluorescent dyes (DCFH-DA, FerroOrange), and fixatives were conducted in a fume hood with appropriate personal protective equipment (PPE), including gloves and safety goggles. DMSO and fluorescent probes were handled with caution due to their ability to penetrate skin and potential cytotoxicity. Contaminated materials, including pipette tips, culture plates, and gloves, were disposed of in designated hazardous waste containers. Animal tissues and carcasses were treated as biohazardous waste and were incinerated or autoclaved according to institutional biosafety regulations and local environmental guidelines.

Statistical analysis
All results are presented as the mean ± standard deviation (SD). Differences between two groups were evaluated using a two-tailed Student's t-test. For analyses involving multiple groups, one-way ANOVA followed by suitable post hoc comparisons (such as Tukey's test) was applied. A p-value below 0.05 was regarded as indicative of statistical significance.

Proliferation suppression and enhanced chemosensitivity of U251 and U87 cells in vitro by CaA
Figure 1A illustrates the structure of CaA. We first evaluated the cytotoxic effects of CaA using a MTT assay on normal murine HT22 neuronal cells and human glioma U251 and U87 cells. Compared with untreated cells (0 µM CaA), CaA had minimal impact on HT22 cell viability at concentrations below 100 µM over 72 h (Figure 1B). However, treatment with 250 µM and 500 µM CaA for 72 h reduced HT22 cell viability by approximately 40% and 60%, respectively (Figure 1B). In contrast to untreated control cells (0 µM CaA), treatment with CaA for 72 h significantly inhibited the proliferation of U251 and U87 glioma cells in a dose-dependent manner. At 20 µM and 50 µM, CaA reduced cell viability by 41.10% and 59.99% in U87 cells, and by 33.98% and 46.34% in U251 cells, respectively (Figure 1C). Similarly, EdU incorporation assays showed that, compared to 0 µM CaA, treatment with 20 µM and 50 µM CaA decreased the proportion of EdU-positive cells by 52.8% and 56.8% in U87 cells, and by 48.6% and 57.9% in U251 cells, respectively (Figure 1D).

CaA also significantly suppressed clonogenic survival: relative to untreated cells, 5 µM, 20 µM, and 50 µM CaA reduced colony formation by 23.6%, 47.2%, and 67.2% in U87 cells, and by 27.9%, 44.7%, and 49.9% in U251 cells (Figure 1E). Notably, CaA enhanced the cytotoxic efficacy of cisplatin: co-treatment with 20 µM CaA and 20 nM cisplatin increased the percentage of apoptotic cells by 19.45% in U87 and 15.58% in U251 cells compared to cisplatin alone (Figure 1F). Statistical analyses were performed using one-way ANOVA or Student's t-test as appropriate. Each group included three technical replicates, and all experiments were independently repeated three times. Collectively, these data indicate CaA not only effectively suppresses glioma cell growth but also increases their sensitivity to chemotherapy in vitro

Ferroptosis induction in glioma cells
CaA treatment significantly increased the generation of ROS in U251 and U87 glioma cells, as measured by DCF fluorescence intensity (Figure 2A). Compared to untreated controls (0 µM), 20 µM and 50 µM CaA increased ROS fluorescence intensity by 6,840-fold and 7,233-fold in U87 cells, and by 10,333-fold and 40,152-fold in U251 cells, respectively. Concurrently, CaA caused a marked depletion of intracellular GSH: 20 µM and 50 µM CaA decreased GSH levels by 37.4% and 46.7% in U87 cells, and by 40.3% and 43.9% in U251 cells (Figure 2B). CaA-treated cells also exhibited increased lipid peroxidation, reflected by elevated MDA content: MDA levels rose by 242% and 268% in U87 cells, and by 80% and 165% in U251 cells following treatment with 20 µM and 50 µM CaA, respectively (Figure 2C).

Western blot analysis revealed dose-dependent downregulation of several ferroptosis-associated proteins following CaA treatment (Figure 2D). In U87 cells, 20 µM and 50 µM CaA reduced FTH1 by 40% and 60%, GPX4 by 50% and 70%, glutaminase by 40% and 60%, and xCT by 60% and 70%, respectively. Similar reductions were observed in U251 cells, with FTH1, GPX4, glutaminase, and xCT decreased by 40% and 60%, 50% and 70%, 55% and 70%, and 55% and 75%, respectively. Lastly, assessment of labile iron accumulation via FerroOrange staining revealed that 20 µM and 50 µM CaA increased fluorescence intensity by 236-fold and 418-fold in U87 cells, and by 30-fold and 37-fold in U251 cells, respectively (Figure 2E).

All data are presented as mean ± SD; comparisons were made using one-way ANOVA with Tukey's post hoc test. P < 0.05 was considered statistically significant. Each experiment was performed with three replicates per group and repeated independently three times.

Reversing inhibitory effects of CaA on glioma cells by ferroptosis inhibition
To confirm the involvement of ferroptosis in CaA's anti-glioma effects, we employed the ferroptosis inhibitor Lip1 in a series of rescue experiments. As shown in Figure 3A, co-treatment with Lip1 (50 nM) significantly attenuated CaA-induced cytotoxicity: compared to CaA treatment alone, the addition of Lip1 increased cell viability by 44% in U87 cells and 43% in U251 cells after 72 h of treatment. Similarly, in the EdU incorporation assay (Figure 3B), Lip1 co-treatment effectively reversed the proliferation suppression caused by CaA. The percentage of EdU-positive proliferating cells rose by 88.6% in U87 cells and 186% in U251 cells, compared to the CaA-only group, after 24 h of exposure.

In the colony formation assay (Figure 3C), Lip1 markedly rescued the clonogenic capacity suppressed by CaA. The number of colonies increased by 74% in U87 cells and 124.6% in U251 cells when Lip1 was combined with CaA, compared to CaA alone. Furthermore, Annexin V/PI staining (Figure 3D) revealed that Lip1 reduced CaA-induced apoptosis, with apoptotic cell percentages decreased by 35% in U87 cells and 40% in U251 cells relative to CaA-only treatment. These results collectively indicate that inhibition of ferroptosis by Lip1 substantially reverses the anti-proliferative, anti-clonogenic, and pro-apoptotic effects of CaA in glioma cells.

All experiments included n = 3 replicates per group, repeated independently three times. Data are shown as mean ± SD. Statistical comparisons were made using t-tests or one-way ANOVA with post hoc tests, as appropriate. P < 0.05 was considered statistically significant.

Nrf2 protein level reduction by promoting Nrf2 degradation
Given Nrf2's critical role in regulating oxidative stress and ferroptosis, we next investigated how CaA affects its protein expression and stability in glioma cells. As shown in Figure 4A, treatment with increasing concentrations of CaA (5, 20, and 50 µM) for 24 h led to a dose-dependent reduction in both total and nuclear Nrf2 protein levels. In U87 cells, total Nrf2 levels decreased by 20%, 55%, and 60%, respectively, while nuclear Nrf2 levels declined by 30%, 70%, and 70%. Similarly, in U251 cells, total Nrf2 protein dropped by 25%, 50%, and 55%, with nuclear levels decreasing by 25% and 55% at 20 and 50 µM, respectively.

We also examined the time course of Nrf2 suppression following CaA treatment at 50 µM. As illustrated in Figure 4B, total Nrf2 protein in U87 cells was reduced by 20%, 55%, and 60% after 6, 12, and 24 h, respectively, while nuclear levels dropped by 15%, 45%, and 60%. In U251 cells, the corresponding decreases in total protein were 30%, 50%, and 60%, and in nuclear Nrf2 were 13%, 55%, and 60%, demonstrating a consistent time-dependent trend. In contrast, CaA treatment did not significantly alter mRNA levels of either Nrf2 or its repressor Keap1, as measured by qRT-PCR (Figure 4C,D), suggesting that the regulation occurs at the post-transcriptional level.

To explore the mechanism underlying this post-translational regulation, we co-treated cells with the proteasome inhibitor MG132 (10 µM). As shown in Figure 4E, MG132 rescued the CaA-induced reduction in Nrf2 protein, increasing total Nrf2 levels by 3.5-fold in U87 cells and 3-fold in U251 cells, indicating that CaA promotes Nrf2 degradation via the proteasomal pathway. Further supporting this conclusion, Figure 4F shows that CaA significantly enhanced the ubiquitination of Nrf2 without affecting Keap1 ubiquitination, suggesting that CaA accelerates Nrf2 turnover through a Keap1-independent, ubiquitin-proteasome-mediated mechanism.

All experiments were performed with n = 3 replicates per group, and repeated in three independent experiments. Results are expressed as mean ± SD, and comparisons were analyzed by one-way ANOVA or Student's t-test where appropriate. A P-value < 0.05 was considered statistically significant.

CaA-induced ferroptosis attenuation by Nrf2 overexpression
To further elucidate the role of Nrf2 in CaA-induced ferroptosis, we overexpressed Nrf2 in U87 and U251 glioma cells prior to CaA treatment (50 µM, 24 h). Nrf2 overexpression markedly attenuated the oxidative and ferroptotic responses triggered by CaA. Specifically, as shown in Figure 5A, enforced Nrf2 expression reduced DCF fluorescence intensity by 32-fold in U87 cells and 16-fold in U251 cells, indicating a significant suppression of intracellular ROS accumulation. Consistently, Nrf2 overexpression restored intracellular antioxidant capacity. As shown in Figure 5B, GSH levels increased by 50% in U87 cells and 60% in U251 cells compared to cells transfected with an empty vector under CaA treatment. In parallel, MDA levels -- indicative of lipid peroxidation -- were reduced by 45% in U87 cells and 55% in U251 cells with Nrf2 overexpression, relative to CaA-treated controls.

Western blot analyses in Figure 5C further confirmed that Nrf2 overexpression restored the expression of key ferroptosis-related proteins. Compared to vector controls, Nrf2-overexpressing U87 cells showed increased levels of FTH1 (+140%), GPX4 (+100%), glutaminase (+63%), and xCT (+137%). In U251 cells, the corresponding increases were FTH1 (+69%), GPX4 (+72%), glutaminase (+33%), and xCT (+50%). Moreover, as shown in Figure 5D, Nrf2 overexpression dramatically reduced labile iron accumulation, with FerroOrange fluorescence intensity decreased by 95% in U87 cells and 98% in U251 cells, compared to cells with basal Nrf2 levels under CaA treatment. Collectively, these data support the conclusion that Nrf2 downregulation is a key mediator of CaA's ferroptotic activity in glioma cells.

All results were derived from n = 3 replicates per group, confirmed across three independent experiments, and are expressed as mean ± SD. Statistical significance was assessed using Student's t-test or one-way ANOVA, with P < 0.05 considered significant.

PI3K/Akt signaling pathway modulation in glioma cells
Global gene expression analysis suggested that CaA impacts the PI3K/Akt pathway. RNA sequencing of CaA-treated U87 cells identified 310 upregulated and 496 downregulated genes relative to control (|log2 fold change| ≥ 1, p ≤ 0.05; Figure 6A). KEGG pathway enrichment analysis of these differentially expressed genes highlighted the PI3K/Akt signaling pathway as significantly enriched (Figure 6B).

To validate these findings at the protein level, we assessed the expression of total and phosphorylated PI3K, Akt, mTOR, and EIF4B in U87 and U251 cells treated with increasing concentrations of CaA (5, 20, and 50 µM). Western blot analysis revealed that CaA did not alter total protein levels, but dose-dependently suppressed phosphorylation across all targets (Figure 6C). In U87 cells, CaA at 5, 20, and 50 µM reduced p-PI3K by 23%, 53%, and 82%; p-Akt by 44%, 62%, and 74%; p-mTOR by 43%, 57%, and 72%; and p-EIF4B by 11%, 29%, and 67%, respectively. Similarly, in U251 cells, the same concentrations of CaA led to reductions in p-PI3K by 26%, 69%, and 85%; p-Akt by 18%, 58%, and 72%; p-mTOR by 36%, 56%, and 72%; and p-EIF4B by 24%, 59%, and 75%, respectively. These results confirm a concentration-dependent inhibition of PI3K/Akt pathway activation by CaA. All quantitative data are presented as mean ± SD, derived from three independent experiments. Statistical significance was determined using one-way ANOVA, with P < 0.05 considered significant.

Molecular docking simulations further indicated that CaA can bind with high affinity to the active site (Ser671, Lys700, Tyr734, Val749, Ser752, and Asp831) of the PI3K p110α (PIK3CA) protein (Figure 6D), suggesting a potential direct interaction. Collectively, these findings indicate that CaA attenuates PI3K/Akt signaling in glioma cells, potentially by directly targeting PIK3CA.

Glioma tumor growth inhibition in vivo by inducing ferroptosis
Finally, we assessed the therapeutic efficacy of CaA in vivo using a subcutaneous U251 glioma xenograft model in athymic nude mice. After 21 days of treatment, CaA markedly suppressed tumor growth, reducing tumor volume by 77.6% compared to the untreated control group (Figure 7A). In contrast, co-administration of the ferroptosis inhibitor Liproxstatin-1 (Lip1) attenuated this effect, with tumor volume reduced by only 21% relative to control, indicating a partial reversal of CaA-induced tumor suppression. Similarly, final tumor weights were reduced by 75.3% in the CaA-treated group but only by 11.7% in the CaA + Lip1 group (Figure 7B). These findings reinforce the notion that ferroptosis contributes substantially to the anti-glioma effect of CaA in vivo.

Importantly, body weight monitoring revealed no significant weight loss across all treatment groups throughout the experimental period, suggesting minimal systemic toxicity (Figure 7C). Histological examination of major organs, including heart, liver, lungs, and kidneys, revealed no overt pathological changes (Figure 7D). Together, these results confirm that CaA exerts potent anti-tumor effects in vivo, primarily through the induction of ferroptosis, without evident adverse effects at the tested dose.

All quantitative data are presented as mean ± SD. Each experimental group consisted of at least eight mice. Statistical significance was determined using one-way ANOVA, with P < 0.05 considered significant.

Data Availability:
All supplemental materials supporting the findings of this study have been uploaded to the public repository Zenodo and can be accessed via DOI 10.5281/zenodo.1777630.

Figure 1: Proliferation inhibition and enhanced chemosensitivity of glioma cells in vitro. (A) Chemical structure of CaA. (B) MTT assay assessing the viability of normal murine HT22 neuronal cells treated with CaA at concentrations of 0, 0.1, 0.5, 1, 10, 50, 100, 250, and 500 µM for 72 h. (C) MTT assay detecting the viability of U251 and U87 glioma cells treated with 0, 5, 20, and 50 µM CaA for 72 h. (D) EdU incorporation assay (10 µM EdU, 24 h incubation post treatment) showing the proportion of proliferating U251 and U87 cells after treatment with 0, 20, and 50 µM CaA for 24 h. (E) Colony formation assay: U251 and U87 cells were seeded at 500 cells per 3.5 cm dish, treated with 0, 5, 20, and 50 µM CaA, then cultured in drug-free medium for 14 days; colonies exceeding 75 µm in diameter were counted. (F) Annexin V/PI apoptosis assay of U251 and U87 cells treated with 20 nM cisplatin alone or combined with 20 µM CaA for 24 h. Error bars represent the standard deviation of the mean. Data are presented as mean ± SD, with n = 3 technical replicates per group and all experiments independently repeated three times. Statistical analyses were performed using (B-E) one-way ANOVA for multi-group comparisons and (F) two-tailed Student's t-test for comparisons between cisplatin alone and cisplatin + CaA groups. *P < 0.05; **P < 0.01. Scale bars = 50 µm (D). Please click here to view a larger version of this figure.

Figure 2: Ferroptosis inhibition in glioma cells. (A) DCFH-DA assay (5 µM DCFH-DA, 1 h incubation at 37 °C) showing intracellular ROS levels in U251 and U87 cells treated with 0, 20, and 50 µM CaA for 24 h; higher green fluorescence intensity indicates higher ROS levels. (B) Quantification of intracellular GSH levels in U251 and U87 cells treated with 0, 5, 20, and 50 µM CaA for 24 h. (C) Measurement of intracellular MDA content in U251 and U87 cells treated with 0, 5, 20, and 50 µM CaA for 24 h, reflecting lipid peroxidation levels. (D) FerroOrange assay (1 µM FerroOrange, 30 min incubation in serum-free medium at 37 °C) showing labile Fe2+ accumulation in U251 and U87 cells after treatment with 0, 20, and 50 µM CaA for 24 h; increased orange fluorescence indicates higher Fe2+ levels. (E) Western blot analysis of ferroptosis-related proteins (xCT, GPX4, FTH1, glutaminase) in U251 and U87 cells treated with 0, 5, 20, and 50 µM CaA for 24 h, with GAPDH as the internal control. Error bars represent SD of the mean. Data are presented as mean ± SD, with n = 3 technical replicates per group and experiments independently repeated three times. Statistical comparisons were performed using one-way ANOVA followed by Tukey's post hoc test. *P < 0.05; **P < 0.01. Scale bars = 50 µm (A,E). Please click here to view a larger version of this figure.

Figure 3: Antiglioma effects of CaA attenuated by ferroptosis inhibition. (A) MTT assay of U251 and U87 cells pretreated with 50 nM Lip1 for 2 h, then co-treated with 50 µM CaA for 0, 24, 48, and 72 h; cell viability was measured to evaluate the rescue effect of Lip1. (B) Colony formation assay: U251 and U87 cells were pretreated with 50 nM Lip1 for 2 h, co-treated with 50 µM CaA, and then cultured in drug-free medium for 14 days to assess clonogenic capacity. (C) EdU assay showing the proliferation of U251 and U87 cells after Lip1 + CaA co-treatment as described above. (D) Annexin V/PI apoptosis assay of cells pretreated with 50 nM Lip1 for 2 h, then co-treated with 20 µM CaA and 20 nM cisplatin for 24 h. Error bars represent SD of the mean. Data are presented as mean ± SD, with n = 3 technical replicates per group and all experiments independently repeated three times. Statistical comparisons were performed using two-tailed Student's t-test (between CaA alone and CaA + Lip1 groups) and one-way ANOVA with Tukey's post hoc test for multi-group comparisons. *P < 0.05; **P < 0.01. Scale bars = 50 µm (B). Please click here to view a larger version of this figure.

Figure 4: Nrf2 protein levels downregulation by promotion of ubiquitin-proteasome-mediated degradation in U251 and U87 cells. (A) Western blot analysis of total and nuclear Nrf2 protein levels in U251 and U87 cells treated with 0, 5, 20, and 50 µM CaA for 24 h; GAPDH was used as the cytoplasmic internal control and Lamin B as the nuclear internal control. (B) Western blot detecting total and nuclear Nrf2 protein levels in U251 and U87 cells treated with 50 µM CaA for 0, 6, 12, and 24 h. (C,D) qRT-PCR analysis of Nrf2 and Keap1 mRNA levels in U251 and U87 cells treated with 50 µM CaA for 6 and 18 h; GAPDH served as the internal reference, and relative expression was calculated using the 2-ΔΔCt method. (E) Western blot of Nrf2 protein in cells treated with 50 µM CaA alone or combined with 10 nM MG132 for 24 h. (F) Immunoprecipitation assay: cells were treated with 10 nM MG132 with or without 50 µM CaA for 24 h; anti-Nrf2 or anti-Keap1 antibodies were used for immunoprecipitation, and anti-ubiquitin antibodies were used to detect ubiquitinated proteins; normal rabbit IgG was the negative control. Error bars represent SD of the mean. Data are presented as mean ± SD, with n = 3 technical replicates per group and experiments independently repeated three times. Statistical analyses were performed using one-way ANOVA with Tukey's post hoc test for multi-time point or multi-concentration groups, and two-tailed Student's t-test for pairwise comparisons. *P < 0.05; **P < 0.01. Please click here to view a larger version of this figure.

Figure 5: Nrf2-mediated suppression of CaA-induced ferroptosis. U251 and U87 cells were transfected with 2.5 µg of Nrf2 overexpression plasmid or empty vector using lipid-based transfection reagent for 18 h, then treated with 50 µM CaA for 24 h. (A) DCFH-DA assay (5 µM, 1 h incubation) showing intracellular ROS levels. (B) Quantification of intracellular GSH and MDA contents. (C) Western blot analysis of ferroptosis-related proteins (FTH1, GPX4, glutaminase, xCT) with GAPDH as the internal control. (D) FerroOrange assay (1 µM, 30 min incubation) detecting labile Fe2+ accumulation. Error bars represent SD of the mean. Data are presented as mean ± SD, with n = 3 technical replicates per group and all experiments independently repeated three times. Statistical comparisons between the Nrf2 overexpression group and the empty vector group were performed using two-tailed Student's t-test. *P < 0.05; **P < 0.01. Scale bars = 50 µm (A,D). Please click here to view a larger version of this figure.

Figure 6: PI3K/Akt signaling pathway inhibition by CaA in glioma cells. (A) Volcano plot of differentially expressed genes in U87 cells treated with 50 µM CaA for 24 h versus untreated controls; Red points indicate genes that are upregulated, while green points denote downregulated genes (FDR < 0.05, |log₂FC| ≥ 1). (B) KEGG pathway enrichment analysis of the differentially expressed genes reveals significant enrichment of the PI3K/Akt signaling pathway (adjusted P < 0.05). (C) Western blot analysis of total and phosphorylated PI3K, Akt, mTOR, and EIF4B in U251 and U87 cells treated with 0, 5, 20, and 50 µM CaA for 24 h; phosphorylated protein levels were normalized to corresponding total proteins, and total proteins were normalized to GAPDH. (D) Molecular docking model showing the binding pose of CaA in the active site of PI3K p110α (PIK3CA, PDB ID: 8BFU), with key interacting amino acid residues (Ser671, Lys700, Tyr734, Val749, Ser752, Asp831) labeled. Error bars represent SD of the mean. Data are presented as mean ± SD, with n = 3 technical replicates per group and experiments independently repeated three times. Statistical significance was determined using one-way ANOVA with Tukey's post hoc test. *P < 0.05; **P < 0.01. Please click here to view a larger version of this figure.

Figure 7: Glioma growth suppression in vivo by ferroptosis induction. Subcutaneous U251 glioma xenografts were established in 4-week-old male BALB/c nude mice (n = 8 per group). After 7 days of tumor establishment, mice were treated with daily intraperitoneal injection of vehicle, 2.5 mg/kg CaA alone, or 2.5 mg/kg CaA combined with 10 mg/kg Lip1 for 14 consecutive days. (A) Tumor growth curves measured every 3 days; tumor volume was calculated using the formula: volume = (length × width²)/2. (B) Final tumor weights after euthanasia on day 21. (C) Body weight changes of mice during the entire treatment period. (D) Representative H&E-stained sections of heart, liver, lungs, and kidneys from each group to evaluate organ toxicity. Error bars represent SD of the mean. Data are presented as mean ± SD, with n=8 mice per group. Statistical analysis was performed using one-way ANOVA with Tukey's post hoc test. *P < 0.05; **P < 0.01. Scale bars = 100 µm (D). Please click here to view a larger version of this figure.

Figure 8: Mechanism of CaA in glioma treatment. CaA effectively treats glioma by promoting Nrf2 degradation and inhibiting the PI3K/Akt pathway, leading to ferroptosis and suppressed tumor growth in vitro and in vivo. Please click here to view a larger version of this figure.

Supplemental Figure S1: A workflow diagram summarizing the experimental conditions and corresponding analyses. Please click here to download this File.

CaA exhibits a compelling antitumor effect in glioblastoma by uniquely modulating two critical survival axes: the PI3K/Akt/mTOR pathway and the Nrf2-ferroptosis defense mechanism. The PI3K/Akt pathway is frequently hyperactivated in glioblastoma multiforme (GBM, altered in up to ~88-90% of cases) due to PTEN loss or RTK activation, driving proliferation and therapy resistance²⁶. Yet, targeting this pathway alone has yielded limited apoptosis in GBM, as tumors often escape PI3K inhibition via cytostatic or compensatory mechanisms²⁶. Here, CaA's dual action is novel: by inhibiting PI3K/Akt signaling (through direct binding to PIK3CA) it not only suppresses a key pro-growth pathway but also promotes the degradation of Nrf2, a master regulator of antioxidant and ferroptosis-protective genes²⁷. This multi-pronged mechanism allows CaA to collapse two of GBM's major defenses -- oncogenic signaling and redox homeostasis -- leading to potent ferroptotic cell death. To our knowledge, CaA is the first natural compound reported to coordinately downregulate the PI3K/Akt/mTOR axis and the Nrf2-driven cytoprotective network in glioblastoma, highlighting a promising therapeutic strategy that leverages the crosstalk between metabolic vulnerability and signal transduction.

Mechanistically, CaA-induced PI3K/Akt inhibition appears to facilitate Nrf2 destabilization and ferroptosis sensitivity in GBM cells. In untreated cancer cells, active Akt can phosphorylate and inactivate GSK3β or other components of the Nrf2 degradation machinery, thereby stabilizing Nrf2 and upregulating its target genes (including xCT and GPX4) that collectively mitigate lipid peroxidation^14,28. This Akt-Nrf2 axis is well documented to impede ferroptosis; for instance, AKT hyperactivation in IDH-mutant gliomas sustains Nrf2 levels and confers resistance to ferroptotic cell death¹⁴. One possible intermediary is the glycogen synthase kinase 3 beta (GSK3β) axis. PI3K inhibition can lead to GSK3β activation, which in turn facilitates β-TrCP-mediated ubiquitination of Nrf2, promoting its degradation independently of Keap1^14,29. In our study, although CaA inhibited PI3K/Akt signaling, further validation is needed to confirm whether this downregulation of Nrf2 occurs through the GSK3β pathway or other PI3K-independent mechanisms. Future work will explore these possibilities using specific pathway inhibitors and rescue experiments. Our findings align with this paradigm. CaA-treated U87/U251 cells showed reduced Nrf2 protein, suggesting enhanced turnover, which would, downstream, diminish the expression of SLC7A11 and GPX4, key executors of the antioxidant defense. Prior studies support that downregulation of Nrf2 lowers SLC7A11/GPX4 levels and makes tumor cells more susceptible to ferroptosis^29,30. Indeed, the tumor suppressor phosphatase and tensin homolog (PTEN) was recently shown to promote ferroptosis via this mechanism: PTEN loss (common in GBM) elevates PI3K/Akt activity and Nrf2-dependent xCT expression, whereas restoring PTEN or inhibiting Akt reactivates GSK3β, reduces Nrf2, and sensitizes cells to ferroptosis²⁸. In our model, CaA effectively mimics the PTEN-restored state by inhibiting PI3K/Akt, thereby accelerating Nrf2 turnover. The consequence is a tipping of the redox balance: with Nrf2-driven antioxidants impaired, CaA triggers rampant lipid peroxidation and ferroptotic cell death as evidenced by our observed increases in ROS and malondialdehyde in treated cells. This mechanistic link between PI3K/Akt inhibition and ferroptosis induction underscores how CaA's two targets synergize: Akt inhibition removes a brake on ferroptosis, and ferroptosis activation in turn ensures cell lethality rather than just growth arrest.

Natural compounds are emerging as ferroptosis triggers in gliomas: juglone (a naphthoquinone from walnuts) was reported to potently induce ferroptosis in GBM via activation of p38 MAPK and suppression of the Nrf2/GPX4 pathway³¹, and even the mineral borax can drive ferroptotic death in GBM cells by reducing HSPA5 and NRF2 levels³². Compared to these agents, CaA targets the well-validated PI3K p110α oncogenic driver directly, which may offer greater specificity in tumors reliant on that kinase. By suppressing PI3K/Akt, CaA effectively pushes GBM cells off their proliferative "track", and by promoting Nrf2 degradation, it removes the guardrails, ensuring the cells cannot escape oxidative damage. This mechanism contrasts with conventional ferroptosis inducers, which act downstream (on cystine transport or GPX4), and with PI3K inhibitors, which do not necessarily inflict lethal damage. CaA's ability to connect these pathways is a distinctive advantage, potentially overcoming the limitations of either approach alone. It is also noteworthy that many GBMs harbor hyperactive PI3K signaling and antioxidant adaptations; thus, a single agent like CaA that tackles both could be especially efficacious in such contexts.

Critical experimental steps that determine assay success include transfection efficiency for Nrf2 overexpression or knockdown, optimal Liproxstatin-1 (Lip1) pre-treatment (1 µM for 2 h prior to CaA exposure), and accurate timing of ferroptosis assays (e.g., ROS and iron detection within 24 h of treatment). These checkpoints are essential to ensure meaningful and reproducible results. Troubleshooting guidance includes confirming cell density and viability prior to transfection and ensuring complete removal of unincorporated fluorescent probes in ROS and iron assays to minimize background. Xenograft variability may be reduced by consistent injection volumes and monitoring of tumor growth to standardize CaA treatment onset. Compared with existing methods, our experimental workflow integrates molecular docking, multi-parametric cell-based assays, and in vivo evaluation to link pathway inhibition directly to functional ferroptosis induction. This offers a higher mechanistic resolution than most single-endpoint ferroptosis studies.

The advantages of this protocol include high reproducibility, scalability across cell lines, and a lack of observed systemic toxicity in animal models. These features make it particularly suitable for both mechanistic studies and preclinical therapeutic exploration. A limitation of our approach is that it does not assess CaA activity in orthotopic or BBB-penetrant models. Further, our xenograft work uses subcutaneous tumors, which may not fully recapitulate brain microenvironment conditions. Additionally, transfection and inhibitor dosing may vary across GBM subtypes and require optimization for reproducibility.

Finally, CaA, a natural phenylpropanoid glycoside, represents a promising lead for GBM therapy by combining targeted inhibition of PI3K/Akt signaling with ferroptosis induction. Acting upstream on PI3K, CaA not only blocks proliferative signaling but also disables the antioxidant defenses that protect against ferroptosis, potentially overcoming limitations of single-target agents (Figure 8). This dual mechanism may reduce adaptive resistance and target multiple tumor subpopulations, including glioma stem-like cells. Future work should focus on medicinal chemistry optimization, confirming molecular interactions with PI3K, and mapping the Nrf2 degradation pathway in detail. Overall, our findings bridge mechanistic insights and experimental workflow, demonstrating how the integration of molecular, cellular, and in vivo assays can uncover a potent dual-action antiglioma strategy.