Therapeutic optimization of LIPA targeting to induce endoplasmic reticulum stress and cell death in ovarian cancer

Abstract

Ovarian cancer (OCa) remains the most lethal gynecologic malignancy in the United States, with a five-year survival rate below 20%. Elevated basal levels of endoplasmic reticulum stress (ERS) have recently emerged as a therapeutic vulnerability in OCa. We have previously shown that the tris-benzamide ERX-41 can induce ERS and cancer cell death in OCa by targeting LIPA. In this study, using iterative structure–activity relationship–guided studies to enhance activity in OCa, we identified a more potent ERX-41-derived analog, ERX-208. Importantly, ERX-208 consistently and significantly reduced cell viability in 23 OCa cell lines spanning five major histological OCa subtypes, with IC₅₀ values ranging from 50–100 nM, compared to ∼500 nM for ERX-41. Notably, ERX-208 showed minimal cytotoxicity toward normal ovarian surface epithelial cells, indicating cancer cell selectivity. ERX-208 induced apoptosis and suppressed colony formation in vitro in OCa cells. Mechanistic studies using RNA sequencing, Western blotting, RT-qPCR, transmission electron microscopy, and immunohistochemistry validated robust activation of ERS pathways upon ERX-208 treatment. Through in silico molecular docking simulation and confirmatory detailed site-directed mutagenesis, we identified that ERX-208 binds to LIPA over a broader interaction surface than ERX-41. At the 10 mg/kg dose, ERX-208 demonstrated favorable biodistribution, no observable toxicity, and potent antitumor efficacy in vivo against established cell line-derived xenograft (CDX), patient-derived xenograft (PDX), and patient-derived explant (PDE) models. Immunohistochemical analysis of treated tumors demonstrated changes in expression of proliferative marker (ki67, decreased) and the ERS marker (GRP78, increased). These findings support the clinical advancement of ERX-208 for the treatment of patients with OCa.

Introduction

Epithelial ovarian cancer (OCa) is the most lethal gynecological malignancy in the United States. There are five major histological and molecular subtypes of OCa: high-grade serous (HGSOC, ~70%); low-grade serous (LGSOC, <5%), mucinous (MOC, ~3%), endometrioid (ENOC, ~10%), and clear cell (CCOC, ~10%) ovarian carcinoma [1, 2]. HGSOC is the most common and aggressive subtype of OCa, and the deadliest gynecologic malignancy in the United States [3]. Nearly 90% of HGSOC patients relapse with metastatic, treatment-resistant disease, despite improvements in cytoreductive surgery, platinum-based chemotherapy, and PARP inhibitors [4]. The five-year survival rate for metastatic OCa remains below 20% and has changed little in decades [5]. This poor prognosis is largely driven by extensive intra- and intertumoral heterogeneity, which undermines the efficacy of precision therapies and underscores the urgent need for treatments that exploit shared vulnerabilities across OCa subtypes.

Among cellular organelles, the endoplasmic reticulum (ER) plays a critical role in protein folding and maturation [6]. Genome-scale genetic screens have identified components of the ER stress (ERS) pathway as selective vulnerabilities in several malignancies, including OCa [7–9]. To overcome the deleterious effects of ERS, cells activate adaptive strategies collectively known as the unfolded protein response (UPR) [10, 11]. UPR includes inhibition of de novo protein synthesis and induction of UPR genes at the transcriptional and translational level [12]. The cellular protein, Protein kinase R-like ER kinase (PERK) functions as a master sensor of ERS and ameliorates ERS through activation of the translation initiation factor Eukaryotic translation initiation factor 2α (eIF2α), decreasing de novo protein translation and reduces stress on ER protein processing [13, 14]. UPR can resolve ERS and restore homeostasis: however unresolved ERS can be lethal to cells via ERS-induced apoptosis [14].

OCa exhibits elevated basal ERS due to its high proliferative index and increased demand for de novo protein synthesis. Additionally, the accumulation of misfolded proteins driven by somatic mutations commonly found in OCa further amplifies ERS, resulting in a compensatory upregulation of the UPR [15]. High expression of key UPR mediators such as GRP78, eIF2α, PERK, and ATF6 has been correlated with disease progression and poor survival outcomes in HGSOC patients [16]. These findings suggest that OCa elevated basal ERS represents a potential therapeutic vulnerability that may be exploited by pharmacologic agents that further activate this stress response, thereby overwhelming protective UPR mechanisms and triggering apoptotic cell death.

Our group previously identified that a small molecule ERX-41 binds stereospecifically to lysosomal acid lipase A (LIPA) [17] and induces ERS and cell death in OCa. Mechanistically, ERX-41 binding to LIPA reduces the expression of multiple ER-resident protein-folding chaperones, leading to the buildup of unfolded proteins in the ER and consequently ERS [17]. Clinical translation of ERX-41 is limited by modest potency (IC₅₀ ~500 nM) against OCa.

In the present study, we screened a curated tris-benzamide library of ERX-41 analogs for increased potency in OCa. We identified a potent ERX-41 analog, ERX-208, which induces ERS and apoptosis in a panel of OCa cell lines in vitro (IC₅₀ ~100 nM). ERX-208 activity is dependent on LIPA expression, as LIPA knockdown abrogates ERX-208 activity, while LIPA reconstitution restores drug sensitivity. ERX-208 reduces OCa proliferation in patient-derived explants (PDEs) and demonstrates robust anti-tumor activity in both cell line-derived xenografts (CDXs) and patient-derived xenografts (PDXs) in vivo. Collectively, our findings establish ERX-208 as a viable therapeutic agent that exploits critical vulnerability in OCa, its elevated basal ERS, and support its further development toward clinical translation.

Materials and Methods

Cell lines and reagents

Established OCa cell lines ES2, OVCAR3, OV90, SKOV3, TOV21G, TOV112D, A2780 (Supplementary Table 2), were obtained from the American Type Culture Collection (ATCC, Manassas, VA) and were maintained using ATCC recommended media. Normal human immortalized primary ovarian surface epithelial cells (IOSE-80) were described in earlier publication [18]. OVSAHO cells were purchased from AcceGen™ and OVCAR4 cells were purchased from Sigma. OVCAR8 cells were obtained from the NCI DCTD repository. SKOV3 Cells that were LIPA-KO were generated and characterized in our previous publication [17]. All cancer cell lines were authenticated using short tandem repeat profiling and Mycoplasma screened by a PCR-based approach. Cancer cell lines were used within 10-15 passages. OCa patient-derived primary cells, and ascites derived cells were obtained from the Ob/Gyn tissue core and were described in earlier publication [19]. Recent articles were utilized to categorize existing cells into subgroups [20, 21]. The antibodies for p-eIF2α (#3398), eIF2α (#5324), p-PERK (#3179), PERK (#5683), CHOP (#2895), GRP78 (#3177), ATF4 (#11815), and ATF6 (#65880) were acquired from Cell Signaling Technology (Danvers, MA, USA). The Ki67 antibody (#ab1667) was acquired from Abcam (Cambridge, MA, USA). The LAL (LIPA) antibody (#sc-58374) was obtained from Santa Cruz (Dallas, TX, USA). The anti-FLAG antibody (#F3165) was purchased from Sigma (Rockville, MD, USA). The histone H3 antibody was purchased from Upstate (#06-755).

Synthesis of ERX-208 and tris-benzamides as ERX-41 analogs

The synthesis of ERX-208 and tris-benzamides is summarized in Scheme S1 (Supplementary Data). To compound 1 that was prepared as previously reported [22], were coupled structurally diverse amines comprising trans-4-aminocyclohexanecarboxylic acid and a variety of amines with HATU. The resulting product was treated with TFA, and the final product was precipitated with diethyl ether and dried in vacuo. The detailed synthetic procedures and characterizations are described in the Supporting Information.

Generation of LIPA mutant cell lines

HEK-293T cells were transfected with the packaging mix of MDL-RRE, MD-2.G, and RSV-REV, along with overexpression plasmids. Lentivirus was collected, and SKOV3 LIPA-KO OCa cells were transduced for various LIPA mutants and selected with 7.5 µg/ml of blasticidin for two weeks. Clones were pooled and tested for the over-expression of LIPA by anti-Flag or anti-LIPA antibodies.

Cell viability assays

Cells were seeded in 96-well plates (1 × 10³ cells per well) 1 day before treatment. Cells were treated with varying concentrations of ERX-41 analogs for 6 days. The effects of ERX-41 analogs along with ERX-208 on cell viability were then measured in triplicate using MTT assay as previously described.

Colony formation, Cell invasion, and Apoptosis assays

The effect of ERX-208 on colony formation and apoptosis were assessed using established methodologies, as previously described [17, 19]. Caspase activity was assessed using the Caspase-Glo® 3/7 Assay Kit (G8090, Promega) to detect apoptosis. The effect of ERX-208 on cell invasion was determined by using the Corning BioCoat Growth Factor Reduced Matrigel Invasion Chamber assay. Briefly, Asc25 and Asc34 cell lines were treated with vehicle or ERX-208 for 22 h, and invaded cells were determined according to the manufacturer’s protocols.

Western Blotting and Real Time-quantitative PCR (RT-qPCR)

The OCa cell lines were treated with ERX-208 for indicated time points and lysed using RIPA lysis buffer with protease and phosphatase inhibitors. Nuclear fractionation was performed using NE-PER™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific, #78833) following the manufacturer’s protocol. Subsequently, we performed Western blotting analysis using ERS-specific antibodies. The activation of known ERS genes was validated using RT-qPCR, with total RNA isolated from OCa cells using the Qiagen RNA extraction kit (Valencia, CA, USA). The primer sequences were listed in Supplementary Table S3. The RT-qPCR analysis was conducted utilizing the CFX96 real-time PCR system from Bio-Rad (Hercules, CA, USA), with the SYBR Green Master Mix provided by Applied Biosystems™ (Waltham, MA, USA).

XBP1 mRNA splicing assay

OCa cells were treated with ERX-208 (1 μM) for 0, 8, and 16 h, and total RNA was extracted using the RNeasymini kit. First-strand cDNA was synthesized using the SuperScript III First-Strand kit (Invitrogen, Waltham, MA, USA). To evaluate relative expression levels of unspliced and spliced XBP1, semi-quantitative RT-PCR analysis was performed using PCR SuperMix (Invitrogen). XBP1u/XBP1s cDNA fragments were amplified by PCR using the following primers: XBP1u/XBP1s-F- 5′CCTGGTTGCTGAAGAGGAGG3′ and XBP1u/XBP1s-R-5′CCATGGGGAGATGTTCTGGAG3′. GAPDH was used as a loading control, with primers as follows: 5′-GGGTCAGAAGGATTCCTATG-3′ and 5′-GGTCTCAAACATGATCTGGG-3′. PCR products were analyzed on a 3% agarose gel.

In silico molecular docking simulations

AutoDock Tools 1.5.6 (Scripps Research Institute, La Jolla, CA) was used to create input PDBQT files of protein and ligand. The input file of human lysosomal acid lipase (LAL) was prepared using published coordinates (PDB code: 6V7N). Water molecules were removed from the protein structure and hydrogens added. All other atom values were generated automatically by ADT. The grid box was centered on the helical motif (²³⁸NLCFLLC²⁴⁴). To facilitate free movement of the ligand, the grid box was set to 35 ×35 x 35 Å, with the x, y, and z coordinates of the center of the grid box set to 123, 29, and 140, respectively. The input files of ERX-41 and ERX-208 were created from their energy-minimized conformation using ADT. Docking calculation was performed with AutoDock Vina 1.2.7. A search exhaustiveness of 64 was used with a model number of 16 while all other parameters were left to their default values. Predicted binding poses were visualized using Maestro (Schrödinger, New York, NY).

Transmission Electron Microscopy (TEM) studies

OVCAR8 cells treated with control and ERX-208 (1 μM for 16 h) were fixed using a solution of 4% formaldehyde and 1% glutaraldehyde in PBS. The cells were subsequently treated with a 2% solution of OsO4, dehydrated using ethanol, and then embedded in Poly/Bed® 812 (Polysciences, Warrington, PA, USA). The ultrathin slices were prepared using a Leica ultramicrotome (Wetzlar, Germany) and subsequently stained with uranyl acetate and lead citrate. The imaging process was carried out using a JEOL JEM-1400 transmission electron microscope (Tokyo, Japan).

Immunohistochemistry (IHC)

IHC using Ki67, p-eIf2α, p-PERK, CHOP, and GRP78 antibodies was done as described previously [17, 19]. Briefly, formalin-fixed, paraffin-embedded primary and xenograft tumor sections were used for IHC. Sections were dewaxed with xylene and rehydrated with a descending ethanol series to water. Antigen retrieval was performed using Antigen Unmasking Solution, Citrate-Based (H-3300-250) for 15 minutes under high pressure. Endogenous peroxidase was eliminated with 3% H₂O₂ for 30 minutes. Slides were blocked with 2.5% normal horse serum for 1 h at room temperature, followed by primary antibodies overnight at 4°C. After washing, slides were incubated with HRP rabbit polymer for 60 minutes before visualization with DAB substrate. All slides were counterstained with hematoxylin. The images were captured under 20X magnification. GRP78, p-eIf2α, p-PERK, and CHOP H-score and number of Ki67 positive cells were quantified using Image J software. The proliferative index was calculated using Ki67 positive cells in five randomly selected microscopic fields at 20X per slide.

RNA sequencing and analysis

Following the manufacturer’s instructions, total RNA was extracted from SKOV3 cells treated with ERX-208 (1 µM) for 48 h (Qiagen, Valencia, CA). RNA-sequencing and analysis were conducted by UTHSA sequencing core using established protocols [23]. RNA-seq data that has been deposited with GEO accession number (GSE311167). The raw reads were aligned to the reference human genome (UCSC hg19) with TopHat2. Genes were annotated (using NCBI Ref Seq) and quantified by HTSeq, and DESeq was used to identify differentially expressed genes and significant genes with fold change > 1 and multiple-test adjusted p value < 0.01 were used for interpreting the biological pathways using Gene Ontology analyses [24].

Human tissue samples

Tissue specimens belonging to various subtypes of OCa (13 HGSOC, 2 ENOC, 1 neuroendocrine, and 1 granulosa cell) were obtained from the UTHSA OB/Gyn tissue core (Supplementary Table 1). Sample collection was conducted in accordance with the Declaration of Helsinki ethical guidelines and approved by the Institutional Review Board of Human Subjects Research Ethics Committee of UTHSA (HSC20190695N and 9 March 2019). Written informed consent was obtained from all patients, and specimens were encoded to protect patients. Tumors were histologically classified according to the World Health Organization criteria.

ELISpot assay and Flow cytometry

Single-cell suspensions were flushed from tibiae and fibulae with sterile DPBS using a 10-ml syringe and 30G needle. Red blood cells were removed by incubation with ACK lysis buffer (Lonzo) for 2 min. Bone marrow cells (2 × 10⁶) were first stained in Hank’s buffered salt solution plus 0.1% BSA (HBSS-BSA) for 20 min with fixable viability dye (FVD, eFluor 506; eBioscience. no. 50-246-097) and fluorophore-labeled mAbs to surface markers, including CD3 (APC-Cy7, Clone 17A2; BioLegend, no. 100221), IgD (FITC, Clone 11-26 c.2a; BioLegend, no. 405704), CD138 (PE-Cy7, Clone 281-2; BioLegend, no. 142514), SCA-1 (PerCP, Clone D7; BioLegend, no. 108121) and B220 (BV421, Clone RA3-6B2; BioLegend, no. 103240), in the presence of mAb Clone 2.4G2, which blocks FcγII and FcγIII receptors. After washing, cells were then fixed and permeabilized by incubation for 30 min in 250 μl of BD Cytofix/Cytoperm buffer (BD Fixation/Permeablization Kit, no. 554714) at 4 °C. After washing twice with BD Perm/Wash buffer, cells were counted again and 10⁶ cells resuspended in 100 μl of BD Cytofix/Cytoperm buffer for intracellular staining with anti-Igκ mAb (PE, eBioscience, Clone 1871, no. MKAPPA04) for 30 min. After washing with BD Perm/Wash buffer, cells were analyzed by Cytek Aurora (Cytek Biosciences). FACS data were analyzed by FlowJo software (BD).

For ELISpot analysis of total IgM⁺ and IgG⁺ ASCs, Multi-Screen filter plates (Millipore) were activated with 35% ethanol, washed with PBS and coated with either anti-IgM (SouthernBiotech, no. 1020-01) or anti-IgG (SouthernBiotech, no. 1030-01) in PBS. Single bone marrow cell suspensions were prepared as above and cultured at 250,000 cells ml^–1 in RPMI 1640 medium (Invitrogen) supplemented with FBS (10% v/v, Invitrogen), penicillin/streptomycin/amphotericin B (1% v/v) and 50 μM β-mercaptoethanol (RPMI-FBS) at 37 °C for 16 h. Following removal of supernatants, plates were incubated with either biotinylated goat anti-mouse IgM (SouthernBiotech, no. 1020-08) or goat anti-mouse IgG1 antibiotic (SouthernBiotech, no. 1070-08) for 2 h and, after washing, incubated with horseradish peroxidase-conjugated streptavidin. Plates were developed using the Vectastain AEC peroxidase substrate kit (Vector Laboratories). The stained area in each well was quantified using CTL ImmunoSpot software (Cellular Technology) and is depicted as the number of spots for quantification.

Pharmacokinetic studies

Pharmacokinetic (PK) studies were done using a single intraperitoneal injection of ERX-208 (10 mg/kg). These studies were conducted utilizing the established protocols by the UTSA Preclinical Pharmacology Core. PK evaluation of ERX-208 was performed in 8-week-old female C57BL/6 mice (RRID:IMSR_JAX:000664). The compound was administered intraperitoneally (i.p.) at a dose of 10 mg/kg body weight. For i.p. administration, ERX-208 was formulated in 0.3% hydroxypropyl cellulose at a final volume of 0.1 mL per mouse. A total of three mice were included in the study. Blood samples were collected in K2EDTA tubes from the tail vein at 10 min, 30 min, 1 h, 2 h, 6 h, and 24 h post-dose. Plasma was separated by centrifugation and stored at -80 °C until analysis. Plasma concentrations of ERX-208 were quantified by LC-MS/MS using a Thermo TSQ Altis Plus triple quadrupole mass spectrometer equipped with a Vanquish UHPLC. Samples were chromatographed on a Thermo Hypersil GOLD aQ column (2.1 ×50 mm, 1.9 µm) using a gradient elution with mobile phase A (water, 0.1% formic acid) and mobile phase B (acetonitrile, 0.1% formic acid). MS data were acquired in single reaction monitoring (SRM) mode, monitoring the MS transition m/z = 861.52→401.17. Pharmacokinetic parameters, including the area under the plasma concentration–time curve (AUClast and AUC ∞ ), maximum concentration (Cmax), time to reach maximum concentration (Tmax), biological half-life (T½), apparent clearance (CL), and apparent volume of distribution (Vd) were determined. Non-compartmental analysis was performed using Phoenix WinNonlin (version 8.0; Pharsight Corporation, Mountain View, CA, USA). Tissue distribution of the compound was done using UT Southwestern Core. PK studies were conducted in female BALB/cAnNHsd mice. Animals received i.p. administration of ERX-208 at a dose of 10 mg/kg once daily for three consecutive days in a formulation volume of 0.1 mL. At predetermined time points (0, 1, 8, and 24 hours following the final dose), groups of three mice per time point were euthanized. Blood was collected via cardiac puncture for plasma isolation. Tissues including brain, liver, kidney, and ovary were harvested without perfusion. Plasma and tissue samples were subsequently processed for drug concentration analysis using core standard protocol. The compound average in tissue is calculated by subtracting the amount of compound in the residual blood/plasma within that tissue [25].

Maximum tolerated dose (MTD) studies

All animal studies were carried out using IACUC approval from UT Health San Antonio. To identify MTD, and target organs toxicity profile of ERX-208, we conducted the toxicity of ERX-208 by repeated intra peritoneal (ip) administration for 5 consecutive days to 8-week-old C57BL/6 female mice. The study included four treatment groups: (1) vehicle control, (2) 10 mg/kg, (3) 20 mg/kg, and (4) 50 mg/kg of ERX-208, administered ip. Each group consisted of three mice (n = 3) and was monitored daily for adverse toxic effects. On the sixth day, the mice were euthanized, and vital organs, including the heart, lung, liver, kidneys, pancreas, spleen, ovaries, and uterus, were harvested. The collected organs were fixed in 10% neutral-buffered formalin for 24 hours, processed, and embedded in paraffin blocks. For histological analysis, 5 µm sections were prepared from the paraffin-embedded tissues and stained with hematoxylin and eosin (H&E). The stained sections were then examined under a microscope for pathology screening to assess any histological changes or abnormalities.

Patient-derived explant (PDE) assay

Excised OCa tissues were processed and grown ex vivo for PDE experiments as previously described [17, 19]. Tumor samples were removed and chopped into small pieces before being incubated for 24 h on gelatin sponges in culture media containing 10% FBS. Tissues were treated with vehicle or ERX-208 in culture media for 72 h before being fixed in 10% buffered formalin overnight at 4 °C and processed into paraffin blocks. Following that, sections were processed for Ki67 immunohistochemical examination.

In vivo orthotopic OCa model

All animal studies were conducted using protocol approved by UTHSA IACUC. For xenograft investigations, ES2 cells (1 × 10⁵ cells) stably expressing GFP-Luciferase were intraperitoneally injected into 8-week-old female SCID mice. Following the development of tumors, treatment groups for either the vehicle or ERX-208 were assigned at random. The number of mice required to show treatment impact was determined based on our pre-existing data as well as published findings. The calculations are based on the unpaired data model, power = 0.8, p value = 0.05. The study was not blinded, and no data was excluded. The body weight, tumor weight, and number of metastatic tumor nodules were measured at the completion of the treatment period. Tumor burden was serially monitored using bioluminescence imaging. At the study endpoint, mice were euthanized, and metastatic nodules were counted throughout the peritoneal cavity. Tumor weight was determined by excising and weighing the tumors. At the end of the experiment, mice were anesthetized with isoflurane inhalation followed by cervical dislocation.

In vivo PDX tumor models

PDX tumors were procured from the Ob/Gyn tissue core. All animal studies were conducted using protocol approved by UTHSA IACUC. 8-week-old female SCID mice were used as the recipients of 2 mm³ implants containing PDX (OCa10, OCa14, OCa30) tumor tissue for ectopic tumor model experiments. When tumors were detectable in size, mice were split between control and treatment groups. Following the development of tumors, treatment groups for either the vehicle or ERX-208 were assigned at random. The number of mice required to show treatment impact was determined based on our pre-existing data as well as published findings. The calculations are based on the unpaired data model, power = 0.8, p value = 0.05. The ERX-208 was administered to the treatment groups while the vehicle (0.3% hydroxypropyl cellulose) was given to the control group. Toxicological consequences on the mice were observed every day. Using a digital caliper, tumor development was monitored every 3-4 days. To prepare the tumors for IHC staining, they were removed, weighed, and processed after the mice were euthanized at the end of each experiment. The study was not blinded, and no data were excluded.

Statistical analyses

GraphPad Prism 10 software was used to analyze statistical differences between groups, utilizing an unpaired Student’s t-test and one-way ANOVA. All the data represented in plots are shown as means ± SE. A p-value of p < 0.05 was considered statistically significant.

Results

Development and characterization of ERX-208

A comparison of the structures of two tris-benzamides, ERX-11 [22] (which does not bind LIPA) and ERX-41 [17] (which binds LIPA) suggests that the 4-methylcyclohexylamine moiety at the C-terminus of ERX-41 may be crucial for the interaction with LIPA. To improve the potency of ERX-41, we designed and synthesized a small library of ERX-41 analogs that preserved the core tris-benzamide [22, 26–29] and the cyclohexane parts, but introduced amines, carboxylic acids, esters, and N-alkyl and N-arylamides in lieu of the 4-methyl group, as illustrated for representative analogs in Fig. 1A.

Fig. 1. Development and characterization of ERX-208.

A Structures of selected ERX-41 analogs. B Heatmap depicting the antiproliferative effects of selected ERX analogs on OCa cells in vitro. C MTT assay results demonstrating the potency of ERX-208 over ERX-41 on the viability of OCa cells in vitro. D–F MTT assay results showing the impact of ERX-208 on the viability of OCa cells in vitro. G, H, Colony formation assay results illustrating the effect of ERX-208 on the clonogenic potential of OCa cells in vitro. The left panel (G) presents representative images of colonies formed, while the right panel (H) quantifies the number of colonies in control versus ERX-208-treated cells. I ERX-208 induces caspase-dependent apoptosis in treated ES2 OCa cells. Caspase 3/7 activity (fold change) was measured after treatment with ERX-208 (500 nM) alone or in combination with Q-VD-Oph (caspase inhibitor), ferrostatin-1 (ferroptosis inhibitor), or necrostatin-1 (necroptosis inhibitor). Data are represented as mean ± SEM. ns not significant; ***p < 0.001; ****p < 0.0001.

To introduce positive charge and hydrogen bonding capability, we replaced the 4-methyl group of ERX-41 with an ammonium group (ERX-188). The amine of ERX-188 was then coupled with various carboxylic acids bearing a branched aliphatic chain (ERX-191), aromatic groups, such as phenyl (ERX-192), benzyl (ERX-193), and naphthyl moieties (ERX-194 and ERX-195), and charged functional groups like a positively charged ammonium group (ERX-196) and negatively charged carboxylate group (ERX-197). In addition, the 4-methyl substituent was replaced with a carboxylate group (ERX-209) that places a negative charge and hydrogen bond acceptor capability. Further modifications of ERX-209 included esterification to remove the negative charge (ERX-198) and coupling of the carboxylic acid of ERX-209 with a variety of amines to introduce additional C-terminal substituents, such as cyclohexane (ERX-271) and 4-methoxyaniline (ERX-272). Heterocycles like thiazole (ERX-205), benzothiazole (ERX-206), and quinoline (ERX-208) were introduced. In addition, the cyclohexane ring was replaced with a linear n-hexyl chain (ERX-203) to evaluate the role of the conformational rigidity imparted by the cyclohexane structure.

Functional screening was performed using MTT-based cell viability assays in OVCAR3 and OVCAR8 OCa cell lines (Fig. 1B) and identified several key structure-activity relationships. First, compounds with the trans-1,4-diaminocyclohexane moiety (ERX-191 to ERX-197) contain an amide bond in the reverse direction compared to the ones in the core tris-benzamide backbone and found to be inactive against OCa cells (Fig. 1B). In contrast, compounds with the trans-4-aminocyclohexanecarboxylic acid moiety (ERX-205 to ERX-272), which preserved the original amide bond orientation relative to the tris-benzamide core, showed partial or strong OCa cell growth inhibition. Together, these data underscore the importance of the orientation of the amide bond for compound activity, presumably involved in forming hydrogen bonds with amino acid residues of LIPA at the binding interface.

The size of substituents on the cyclohexane ring also impacted the anti-proliferative activity. Analogs with monocyclic substituents on the cyclohexane ring (ERX-205, ERX-271, and ERX-272) were found to be significantly less effective compared to their counterparts with bicyclic substituents (ERX-206 and ERX-208). Overall, ERX-208 was noted to have the most potent anti-proliferative activity on OCa cells. The conformational rigidity in the cyclohexane moiety seems to be critical for ERX-208’s activity, as an analog with an opened cyclohexane ring (ERX-203) had no biological activity.

Dose-response studies further demonstrated that ERX-208 significantly reduced viability in SKOV3 and OVCAR4 cells, with an IC₅₀ of ~100 nM, representing a five-fold increase in potency compared to ERX-41 (IC₅₀ ~500 nM) (Fig. 1C). Broader evaluation of ERX-208 across multiple models representing all five subtypes of OCa revealed consistent cytotoxicity in nine established OCa cell lines (ES2, OV90, SKOV3, A2780, OVCAR3, OVCAR8, TOV21G, OVSAHO, TOV112D), five primary tumor-derived OCa cultures (OCa27, OCa30, OCa35, OCa39, OCa67), and seven ascites-patient-derived primary HGSOC cells (AS21, AS23, AS25, AS28, AS29, AS34, AS39) (Fig. 1D–F). In contrast, ERX-208 exhibited minimal cytotoxicity in immortalized normal ovarian surface epithelial (IOSE) cells (Supplementary Fig. 1A). ERX-208 also significantly reduced long-term proliferative potential, as shown by colony formation assays (Fig. 1G-H). ERX-208 induced apoptosis in OCa cells, as demonstrated by increased Annexin V staining (Supplementary Fig. 1,B). Treatment with ERX-208 significantly increased caspase 3/7 activity compared to the control group, indicating enhanced apoptotic signaling. Co-treatment with Q-VD-Oph, a pan-caspase inhibitor, completely abolished ERX-208-induced caspase activation, confirming that cell death was caspase-dependent. In contrast, co-treatment with ferrostatin-1 (a ferroptosis inhibitor) or necrostatin-1 (a necroptosis inhibitor) did not reduce caspase 3/7 activity, suggesting that ERX-208 primarily induces apoptosis rather than ferroptosis or necroptosis (Fig. 1I, Supplementary Fig 1, C). These data suggest that ERX-208 is a more potent ERX-41 analog, with a broad range of activity against multiple OCa subtypes.

ERX-208 induces ERS pathways in OCa cells

In our prior quest to improve potency from ERX-11 to ERX-41, we had noted both a change in the primary mechanism of action and the primary molecular target. Using RNA-sequencing data from SKOV3 cells treated with vehicle or ERX-208 for 48 hours, we noted robust alterations of gene expression by ERX-208. Gene ontology (GO) analysis of upregulated genes highlighted significant activation of stress-related genes, including UPR, ERS, and general cellular response to stress (Fig. 2A, red bars), while cell cycle-associated genes were significantly enriched in downregulated transcripts (Fig. 2B, blue bars). These findings are represented in the heatmap of transcriptomic signatures between control and ERX-208-treated groups, with consistent upregulation of ERS markers and downregulation of cell cycle regulators (Fig. 2C). Gene set enrichment analysis (GSEA) also confirmed significant enrichment of the UPR, p53, and ERS-associated pathways in ERX-208-upregulated genes (Supplementary Fig. 2,A) and cell cycle pathways including DNA replication, and mitotic processes in ERX-208-downregulated genes (Supplementary Fig. 2,B). RT-qPCR studies validate that ERX-208 treatment significantly upregulated the expression of multiple ERS-related genes including XBP1, ERN1, ATF3, HERPUD1 and HSPA5 (BiP) (Fig. 2D) and suppressed the expression of critical cell cycle genes such as CDK2, CCNA2, E2F1, PCNA, PLK1, and SKP2 (Fig. 2E). Collectively, these findings confirm that, like its parent compound ERX-41, ERX-208 induces a robust ERS response and disrupts key cell cycle regulators in OCa cells.

Fig. 2. ERX-208 induces endoplasmic reticulum stress in OCa.

Gene ontology ananlyses of RNA sequencing data highlighting treatment-enriched gene sets, with pathways upregulated (A) and downregulated (B) in response to ERX-208. C Heatmap depicting the enrichment of genes associated with cell cycle regulation and UPR following ERX-208 treatment. D RT-qPCR analysis showing upregulation of UPR target genes. E RT-qPCR analysis demonstrating the downregulation of cell cycle target genes. Data are represented as mean ± SEM. ns not significant; **p < 0.01; ***p < 0.001; ****p < 0.0001.

ERX-208 induces canonical ERS pathways across diverse OCa models

We then examined the ERX-208 activation of ERS across heterogeneous OCa models. RT-qPCR revealed time-dependent upregulation of spliced XBP1 (sXBP1) and CHOP transcripts across multiple OCa cell lines (ES2, SKOV3, OVCAR3, and OVCAR4) following ERX-208 treatment (Fig. 3A). sXBP1 showed robust induction within 8–16 hours, while CHOP increase peaked at 16–20 hours. RT-PCR analysis confirmed splicing of XBP1 mRNA in ERX-208-treated cells (Fig. 3B). Western blot analysis revealed ATF6 activation through cleavage and increased protein levels of key ERS markers, including CHOP, and ATF4 following ERX-208 exposure in OCa39, and OVCAR8 cell lines. In addition, ERX-208 increased phosphorylation of eIF2α (Fig. 3C). Further validation in TOV112D, A2780, TOV21G cells also confirmed consistent upregulation of UPR pathway proteins, supporting that ERX-208 robustly activates ERS in all five subtypes of OCa (Supplementary Fig. 3A). To confirm that ERX-208-induced ER stress marker upregulation occurs in the nucleus, subcellular fractionation was performed. ERX-208 treatment increased nuclear ATF4 and modestly enhanced CHOP in OVCAR8 and OCa39 cells, with negligible cytoplasmic levels. Fractionation integrity was verified using H3 (nuclear) and β-actin (cytoplasmic) markers (Supplementary Fig. 3B). Transmission electron microscopy (TEM) revealed marked ultrastructural changes in the endoplasmic reticulum of ERX-208-treated cells. Compared to control cells, which display tightly stacked, smooth ER membranes, ERX-208-treated cells exhibited swollen, dilated, and fragmented ER, hallmarks of ERS and protein overload (Fig. 3D). Collectively, these data demonstrate like ERX-41, ERX-208 robustly activates the ERS response in diverse OCa models at transcriptional, post-transcriptional, and ultrastructural levels.

Fig. 3. ERX-208 induces ER stress in OCa.

A Time-course analysis of ERX-208 (1 μM) treatment on mRNA expression of ER stress-related genes, sXBP1 and CHOP, in SKOV3, OVCAR3, ES2, and OVCAR4 cells. B RT-PCR analysis showing the temporal effects of ERX-208 (1 μM) on the expression of XBP1 (unspliced XBP1 (XBP1u) and spliced XBP1 (XBP1s)) in OCa39, OVCAR8, OVCAR3, and ES2 cells. C Western blot analysis of UPR component activation in OCa39 and OVCAR8 OCa cell lines treated with ERX-208 for the indicated time points. D Transmission electron microscopy of OVCAR8 cells illustrating the effects of vehicle and ERX-208 treatments on subcellular structures after 16 h.; yellow arrows indicate the ER. Scale bar represents 100 nm.

ERX-208-mediated ERS response is dependent on LIPA expression

To validate the central importance of LIPA in ERX-208–induced signaling, we used LIPA knockout (LIPA-KO) SKOV3 cells [17] and compared their response with parental controls (Fig. 4A). Cell viability assays demonstrated that parental SKOV3 cells were highly sensitive to ERX-208, whereas LIPA-KO cells displayed markedly reduced sensitivity (Fig. 4B). RT-PCR analysis further validated the absence of LIPA expression prevented splicing of XBP1 upon addition of ERX-208 (Fig. 4C). Time-course quantitative RT-PCR analyses revealed a robust induction of UPR-associated transcripts, including spliced XBP1 and CHOP, in parental SKOV3 cells following ERX-208 exposure, whereas LIPA-KO cells exhibited a significantly blunted response (Fig. 4D, E). Together, these data indicate that LIPA is required for ERX-208–induced ERS signaling.

Fig. 4. Specificity and interaction sites of ERX-208 to LIPA.

A Western blot shows knockout of LIPA in SKOV3 cells. B Cell viability assay of SKOV3 parental and LIPA-KO cells treated with increasing concentrations of ERX-208. Results show reduced sensitivity in LIPA-KO cells. C RT-PCR analysis of sXBP1 expression in SKOV3 parental and LIPA-KO cells treated with ERX-208 (1 µM) over various time points. Results show reduced sXBP1 induction in LIPA-KO cells compared to parental cells. RT-qPCR analysis of mRNA levels for sXBP1 (D) and CHOP (E) in SKOV3 parental and LIPA-KO cells treated with ERX-208 (1 µM) over a time course, showing reduced ER stress marker expression in LIPA-KO cells. F Predicted binding poses of ERX-41 (magenta) and ERX-208 (teal) to LIPA (PDB code: 6V7N) with mutated LIPA residues highlighted (red: critical for activity; orange: moderately important; green: no impact). Heatmap shows quantification of colony formation assay in SKOV3 LIPA-KO cells stably expressing LIPA mutants treated with or without ERX-208 (G) and ERX-41 (H). Data are represented as mean ± SEM. ns not significant; ****p < 0.0001.

Identification of differences in ERX-41 and ERX-208 interaction with LIPA

Importantly, ERX-208 appeared to be more dependent on LIPA expression than ERX-41, indicating a potentially more targeted therapy. To study this finding, we performed in silico molecular modeling and docking simulation studies to identify how ERX-41 and ERX-208 bind to LIPA (Fig. 4F). To validate the modeling, we created point mutations with alanine substitutions at multiple modeled contact residues within the binding pocket in LIPA expression constructs. Using lentiviral constructs, we then expressed either wild-type (WT) or mutant (MT) LIPA in SKOV3 LIPA-KO OCa cells and created stable clones. Stable expression of WT and mutant LIPA proteins was confirmed by immunoblotting with both anti-Flag and anti-LIPA antibodies (Supplementary Figure 4A). Functional assessment using colony formation assays revealed that five mutations (T103A, E234A, T261A, K220A, and T224A) conferred marked resistance to ERX-208–mediated growth inhibition, whereas mutations at other sites (M14A, N15A, W104A, H225A, T228, K233A, N238A, and H262A) did not significantly affect the sensitivity to ERX compounds (Fig. 4G; Supplementary Figure 4B). In contrast, while mutations in T103A, T261A and K220A blocked ERX-41 activity, the T224A and E234A mutations did not (Fig. 4H; Supplementary Figure 4B). This differential responsiveness supports our modeling showing that the T224 and E234 residues in LIPA are involved in interaction with ERX-208, but not with ERX-41 (Fig. 4H). Docking models indicate that ERX-208 has a large bicyclic aromatic quinoline moiety that may interact with T224, as shown in Fig. 4F (ii and iv): in contrast, the small methyl group at the 4-position of the cyclohexane structure of ERX-41 does not interact with T224 (Fig. 4Fi). Similarly, a mutation in LIPA residue E234 may block interaction with the amide bond in the core structure of ERX-208. Our docking models also confirm that residues T103 and T261 interact with the nitro group and 2-hydroxyethyl group of the tris-benzamide (common to both ERX-41 and ERX-208) and thus mutations at these sites may affect both ERX-41 and ERX-208 activity (Fig. 4Fiii).

To validate the functional significance of this modeling, we performed rescue experiments by re-expressing either WT or MT LIPA in LIPA-KO cells. Reintroduction of WT LIPA restored ERX-208–induced ER stress, as evidenced by increased CHOP and ATF4 expression. In contrast, cells expressing the E234A LIPA MT failed to activate ER stress markers (Supplementary Fig. 4C). Together, these studies corroborate the modeling studies that ERX-208 interacts over a broader surface on LIPA than ERX-41 and may explain the improved potency of ERX-208.

Pharmacokinetics and tissue distribution of ERX-208

To study the biodistribution of ERX-208, we evaluated drug levels in plasma and tissues following a single intraperitoneal (IP) dose of 10 mg/kg. We noted peak levels (~40–56 ng/mL) in plasma at ~2–3 hours, consistent with a Tmax of ~2 hours. The overall systemic exposure was modest, with an AUCinf of 497.4 h·ng/mL (≈0.58 nM·h) (Supplementary Figure 5A). Approximately 18% of the total AUC was extrapolated from the terminal phase, which is within acceptable pharmacokinetic criteria. ERX-208 demonstrated rapid absorption, a moderate half-life (~9 hours), and sub-nanomolar systemic exposure at this dose. Plasma concentrations declined steadily and dropped below quantifiable limits by 24 hours, consistent with rapid clearance and a large apparent volume of distribution. Tissue distribution studies revealed preferential accumulation of ERX-208 in the liver (~70,000–80,000 ng/g) and ovary (~20,000 ng/g), with sustained levels over 24 hours (Supplementary Fig. 5B). Kidney concentrations were intermediate and declined more rapidly, whereas brain penetration was minimal, remaining an order of magnitude lower than in other tissues. These findings indicate that ERX-208 exhibits limited plasma exposure but extensive tissue distribution, particularly to the liver and ovary.

Toxicity assessment of ERX-208

To evaluate tolerability, we performed a 7-day toxicity study in C57BL/6 mice using IP administration of ERX-208 at 0, 10, 20, and 50 mg/kg. Body weights remained unchanged across treatment groups compared with vehicle controls (Supplementary Fig. 5C). Histological examination of major organs, including pancreas, uterus, ovary, lung, liver, kidney, spleen, and heart, revealed no overt morphological abnormalities or tissue damage following ERX-208 treatment, indicating the absence of gross systemic toxicity (Supplementary Fig. 5D). We next examined the effects of ERX-208 on B-cell populations and antibody secretion. Flow cytometric analysis of bone marrow cells demonstrated comparable frequencies of lymphocytes, plasma cells (IgM⁺ and IgG⁺ antibody-secreting cells), and total live cell fractions between vehicle- and ERX-208–treated groups. ELISPOT assays confirmed no significant differences in IgM- or IgG-secreting antibody-producing cells (Supplementary Figure 5E-F). Collectively, these data suggest that ERX-208 does not impair normal B-cell differentiation or antibody production.

Dose ranging studies of ERX-208

To evaluate the therapeutic efficacy of ERX-208, we tested its antitumor activity in an OCa-30 PDX model using three doses of ERX-208 (2.5, 5, and 10 mg/kg/ip). ERX-208 treatment significantly and dose-dependently inhibited tumor growth compared with vehicle controls, with suppression observed across all treatment groups, and most significant at the 10 mg/kg dose (Fig. 5A). Correspondingly, final tumor weights were significantly lower in ERX-208 treated group, with the greatest reduction at 10 mg/kg (Fig. 5B). Importantly, ERX-208 was well tolerated, as body weights remained unchanged throughout the treatment period across all dosing groups, indicating the absence of overt systemic toxicity (Fig. 5C). Immunohistochemical analysis revealed a marked decrease in the proliferation marker Ki67 in tumors from ERX-208–treated animals compared with vehicle controls, consistent with reduced tumor cell proliferation. In addition, ERX-208 treatment led to increased GRP78 staining, indicating enhanced ERS in ERX-208-treated tumors (Fig. 5D–F). These findings demonstrate that ERX-208 exerts potent antitumor activity in vivo, mediated in part through activation of ERS, without eliciting measurable toxicity in treated animals.

Fig. 5. ERX-208 potently reduces tumor growth in a dose-dependent manner in OCa30-PDX models.

A–C Effects of ERX-208 treatment on OCa-30 PDX tumor models. A Tumor volumes measured over the treatment period in mice treated with vehicle control or ERX-208 at doses of 2.5 mg/kg, 5 mg/kg, and 10 mg/kg administered intraperitoneally (i.p.). B Tumor weights at the endpoint of the study in each treatment group. C Average body weights of mice during the treatment period, demonstrating the tolerability of ERX-208 at all doses tested. D–F IHC analysis of proliferation marker Ki67 and ER stress marker GRP78 in tumor tissue sections from OCa-30 PDX tumors treated with vehicle or ERX-208 at the indicated doses. Significance was determined by an unpaired two-tailed Student’s t-test. Data are presented as mean ± SEM, **p < 0.01; ***p < 0.001; ****p < 0.0001.

ERX-208 suppresses growth of OCa tumor explants

To further assess the therapeutic potential of ERX-208 in clinically relevant models, we evaluated its activity in PDX explants. Tumor tissue pieces established from multiple independent tumors were treated with ERX-208 for 3 days, and proliferation was assessed by Ki67 immunohistochemistry (IHC). Representative IHC images from explants are shown (Fig. 6A). Quantification of Ki67 staining revealed a significant reduction in proliferative activity following ERX-208 treatment (Fig. 6B). Together, these results demonstrate that ERX-208 effectively suppresses the growth of OCa explants, highlighting its translational potential as a therapeutic strategy in OCa.

Fig. 6. ERX-208 potently reduces the growth of PDE and orthotopic ES2-CDX tumors.

A Representative IHC images of tissue sections from PDE models treated with ERX-208 (1 µM), stained for Ki67, a proliferation marker. B Quantification of Ki67 staining in PDE models. Quantitative data are presented as mean ± SEM. C, D Invasion assay results illustrating the impact of ERX-208 on OCa ascites model cells. The left panel (C) presents representative images of invaded cells, while the right panel (D) quantifies the number of invaded cells following treatment with ERX-208. E–G Effects of ERX-208 on ES2 xenograft tumor models treated with a single dose of 10 mg/kg i.p. E Tumor volume measurements of mice treated with vehicle or ERX-208. F Tumor weights at the endpoint of the study. G Quantification of the number of tumor nodules in treated and vehicle groups. Data are represented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

ERX-208 suppresses metastasis

Since OCa is a highly metastatic tumor, we sought to evaluate if ERX-208 could have activity against metastatic tumors. We first noted that ERX-208 markedly suppressed the invasive capacity of primary ascites-derived cells, as measured by BD Biocoat Matrigel invasion assays (Fig. 6C, D). Next, we utilized an OCa model for metastases with ES2 cells stably expressing GFP/Luc were injected intraperitoneally into female SCID mice. Following tumor establishment, mice were randomized into treatment groups (n = 5/group) and treated with either vehicle or ERX-208 (10 mg/kg, ip, 5 days/week). Bioluminescence imaging revealed that ERX-208 significantly suppressed tumor progression, resulting in approximately a 60% reduction in tumor burden compared with vehicle controls (Fig. 6E). Consistent with imaging results, ERX-208 treatment produced reduction in both tumor weight (Fig. 6F) and the number of tumor nodules (Fig. 6G). Collectively, these findings demonstrate that ERX-208 exerts potent antitumor activity in orthotopic OCa xenografts, leading to reduced tumor growth and metastatic dissemination.

ERX-208 suppresses growth of OCa PDXs

To evaluate the therapeutic efficacy of ERX-208 in PDX models of OCa, we tested its effects in OCa-14 and OCa-10 tumors established in SCID mice. In OCa-14 PDX tumors, treatment with ERX-208 (10 mg/kg, ip) significantly inhibited tumor growth compared with vehicle controls, as reflected by reduced tumor volumes throughout the study (Fig. 7A) and a marked decrease in final tumor weights at the endpoint (Fig. 7B). Importantly, ERX-208 was well tolerated, with no significant changes in body weight observed (Fig. 7C). Immunohistochemical analysis demonstrated that ERX-208 treatment decreased expression of the proliferation marker Ki67 while increasing GRP78 levels, consistent with induction of ERS (Fig. 7D, E). Similar results were obtained in the OCa-10 PDX model. ERX-208 treatment significantly reduced tumor growth kinetics (Fig. 7F) and final tumor weights (Fig. 7G), without affecting body weight (Fig. 7H). Consistent with the OCa-14 findings, IHC analysis revealed reduced Ki67 staining and elevated GRP78 expression along with activation of other ER stress markers such as p-eIf2α, p-PERK and CHOP in tumors from treated animals (Fig. 7I, J and Supplementary Fig. 6). Collectively, these results demonstrate that ERX-208 exerts robust antitumor activity in multiple OCa PDX models, mediated by inhibition of tumor proliferation and induction of ER stress, without evidence of systemic toxicity.

Fig. 7. Potency of ERX-208 in reducing the growth of PDX tumors.

A–E Evaluation of OCa-14 PDX tumors. A Tumor volume measurements over time in SCID mice treated with a single dose of ERX-208 (10 mg/kg, ip). B Final tumor weights at the end of the treatment period. C Body weight measurements of mice throughout the study to assess systemic toxicity. D IHC staining of tumor sections for Ki67 and GRP78. E Quantification of Ki67 and GRP78 IHC staining in tumor sections, expressed as percentage positive cells. Evaluation of OCa-10 PDX tumors. F Tumor volume measurements over time in SCID mice treated with ERX-208 (10 mg/kg, ip). G Final tumor weights at the end of the treatment period. H Body weight measurements of mice throughout the study. I IHC staining of tumor sections for Ki67 and GRP78. J Quantification of Ki67 and GRP78 IHC staining in tumor sections, expressed as percentage positive cells. Data are presented as mean ± SEM. Statistical significance was determined by 2-way ANOVA and student’s t-test, with p < 0.05 considered significant. **p < 0.01; ***p < 0.001; ***p < 0.0001.

Discussion

OCa exhibits marked genomic instability, rapid proliferation, and elevated protein synthesis, collectively leading to increased basal ERS and dependence on UPR signaling for survival [30, 31]. These emerging findings indicate that OCa-heightened basal ERS may be a therapeutic vulnerability that can be exploited by a drug that stimulates this stress response to overcome protective UPR mechanisms and cause apoptosis. Genome-wide loss-of-function shRNA screens revealed several ERS pathway components as essential vulnerabilities in OCa [7–9]. The ERS protein GRP78 is essential for OCa cell survival, and its expression is elevated in OCa and correlates with OCa malignancy, metastasis development, and drug resistance [16]. Our data suggest ERX-208 as a potent inducer of ERS and showed anti-tumor activity across diverse preclinical models, including PDEs, PDXs, and orthotopic xenografts. Importantly, ERX-208 did not elicit overt systemic toxicity in vivo.

To mitigate the deleterious effects of ERS, cells engage adaptive mechanisms collectively referred to as the UPR [10, 11]. The UPR acts by transiently suppressing de novo protein synthesis and inducing a set of genes involved in restoring ER homeostasis at both the transcriptional and translational levels [12]. Transcriptomic profiling revealed that ERX-208 treatment upregulates canonical ERS and apoptotic pathways while concomitantly suppressing cell cycle pathways. The induction of XBP1 splicing, ATF4/CHOP activation, and eIF2α phosphorylation across multiple histological subtypes including HGSOC, ENOC, and CCOC supports the broad applicability of this approach. Further, our results suggest that ERX-208 is distinct in that it activates all three branches of the UPR pathway: (1) PERK–eIF2α–ATF4, (2) IRE1–XBP1s, and (3) ATF6. TEM analyses confirmed ultrastructural hallmarks of ERS, further corroborating the molecular findings. Importantly, ERX-208 consistently suppressed proliferation in PDEs, CDXs, and PDXs, supporting its translational relevance across genetically diverse and clinically representative OCa models. These data support the clinical translation of ERX-208 as it is likely to overcome the tumor heterogeneity of OCa.

The molecular target of ERX-208 is the LIPA gene encoding lysosomal acid lipase (LAL) protein. LAL is essential in modulating the capacity of mesenchymal stem cells to promote tumor development and metastasis. LAL is significantly upregulated in OCa tissues relative to normal tissues [19]. Using CRISPR-mediated LIPA knockout, mutational mapping, and in silico molecular docking analyses, we identified critical residues (T103, K220A, T224, E234, and T261) within LIPA necessary for ERX-208–mediated ERS induction and cell death. These findings extend prior work with ERX-41 [17] and suggest that the enhanced potency of ERX-208 derives from structural optimization that improves its binding affinity.

Pharmacokinetic and toxicological studies demonstrated that ERX-208 achieves nanomolar plasma concentrations consistent with its pharmacologically active range, while preferentially accumulating in the liver and ovary, the latter representing a major site of therapeutic relevance. Repeated dosing at efficacious levels did not affect body weight, immune cell composition, or histology of major organs, suggesting a favorable safety margin. This contrasts with standard OCa therapies such as platinum agents and PARP inhibitors, which, despite efficacy, are limited by hematologic and systemic toxicities. The selective accumulation of ERX-208 in ovarian tissue raises the possibility of achieving therapeutic concentrations in tumors while minimizing off-target effects.

While our study establishes proof-of-concept for LIPA-targeting ERS in OCa, several questions remain. First, the precise molecular mechanism by which the broader engagement of ERX-208 with LIPA more potently perturbs ER homeostasis requires further resolution. Second, although our models encompassed major OCa subtypes, future studies should address evolved intratumoral heterogeneity and specific resistance mechanisms that may emerge under chronic treatment. Finally, expanded pharmacologic assessments in higher-order species will be necessary to fully define the safety and therapeutic window of ERX-208.

In summary, our findings identify ERX-208 as a potent, selective, and clinically tractable small-molecule inducer of ERS in OCa. By targeting LIPA, ERX-208 exploits a conserved vulnerability in OCa biology, leading to robust anti-tumor efficacy in multiple preclinical models with minimal toxicity. Together, these data provide a strong rationale for advancing ERX-208 into clinical evaluation and for further exploration of LIPA-dependent ERS modulation as a therapeutic strategy in ovarian and potentially other cancers.

Author contributions

SV, J-MA, MR, GVR, and RKV designed the experiments and interpreted the results. SV, GS, RG, DMP, KP, TR, MH, UPP, XY, XL, AB, and PS conducted the experiments; PS, RG, and BE contributed to TEM studies; TKL, SE, HN, CYC, KK, and J-MA designed and conducted chemistry studies; LZ and YC conducted bioinformatic studies; JWB, GRS, PML, NAC, and UPP designed and conducted PK studies; CC and ZX conducted toxicity studies; SV, GRS, PR, YAL, and EK contributed to the collection of de-identified ovarian cancer tumors; RKV and SV conducted mice studies; GVR, J-MA, RKV, and SV wrote the manuscript.

Data availability

The data generated in this study are available within the article and its Supplementary Data. RNA-seq data analyzed in this study were deposited in the GEO database under the GEO accession number GSE311167.

Competing interests

J.-M.A., G.V.R., X.L., and R.K.V. are joint holders of issued and pending patents on ERX-208, which have been licensed to EtiraRx. The remaining authors declare no conflict of interest.

Supplementary information

The online version contains supplementary material available at 10.1038/s41388-026-03689-w.