RHOV is a Detachment-Responsive Rho GTPase Necessary for Ovarian Cancer Peritoneal Metastasis

Abstract

A defining feature of epithelial ovarian cancer, irrespective of histologic subtype, is its predominant spread through transcoelomic metastasis, where tumor cells disseminate into the peritoneal fluid, resist anoikis, and form multicellular aggregates that invade the peritoneum. This tumor progression represents the main driver of mortality for ovarian cancer patients. Identification of the earliest adaptations necessary for metastasizing ovarian cancer cells to survive matrix detachment could help develop strategies to prevent the initiation of transcoelomic metastasis. In this study, we identified a conserved detachment-sensitive gene signature activated shortly after matrix-detachment across multiple ascites-derived ovarian cancer cell lines. Within this signature, RHOV, an atypical and fast-cycling Rho GTPase, emerged as a top transcript that was confirmed to be highly induced in patient-ascites derived cells. Loss of RHOV impaired anoikis resistance, multicellular aggregate compaction, migration, and invasion in vitro, and it completely abolished metastasis in vivo. Mechanistically, RHOV enhanced c-Jun signaling and cytoskeletal remodeling to support pro-metastatic signaling. Rescue experiments showed that both GTP-binding and membrane localization were required for the pro-metastatic function of RHOV. Together, these findings define RHOV as a unique detachment-sensitive Rho GTPase and establish RHOV as a critical and necessary mediator of early adaptations that prime ovarian cancer cells for peritoneal metastatic progression. This work provides key insights into the molecular vulnerabilities of disseminating tumor cells, establishes the targeting of early molecular adaptations following matrix detachment as a potential therapeutic strategy for metastatic disease, and uncovers functions of an understudied member of the Rho GTPase family.

Introduction

Epithelial ovarian cancer is the second most lethal gynecologic malignancy and the sixth leading cause of cancer-related deaths among women in the United States (1). This high mortality is largely attributed to the fact that over 70% of patients are diagnosed at an advanced stage, where the five-year survival rate falls to approximately 30%. Late-stage disease is marked by widespread peritoneal dissemination and the presence of malignant ascites, features that reduce the effectiveness of standard cytoreductive surgery and chemotherapy (2–4). Given that metastasis accounts for most ovarian cancer-related deaths, there is an urgent need to further elucidate the mechanisms driving dissemination and to identify unique vulnerabilities that can be leveraged for the development of novel therapeutic strategies targeting metastatic disease.

Regardless of histologic subtype, epithelial ovarian cancer uniquely spreads through transcoelomic metastasis, in which tumor cells directly disseminate from the ovary or fallopian tube into the peritoneal cavity (2). At advanced stage, dissemination also occurs from peritoneal and omental tumors and is a major source of anchorage-independent (a-i) cells in malignant ascites. Critical steps in ovarian cancer metastasis include detachment of cells, resistance to anoikis (detachment-induced cell death) within the peritoneal fluid, and subsequent assembly into multicellular aggregates (MCAs) (5–7). These anoikis-resistant cells and MCAs function as metastasizing units, colonizing both neighboring and distant peritoneal sites and perpetuate recurrence via continuous shedding from both primary and secondary lesions. The detection of anoikis-resistant cells or MCAs in patient ascites correlates with poor prognosis and chemoresistance (5,8–11). Research on metastasizing ovarian cancer cells has largely focused on MCAs as metastasizing units. However, the immediate molecular adaptations triggered by matrix detachment, facilitating epithelial cell transition to an a-i state and subsequent anoikis resistance and MCA formation, remain largely unexplored. Our study goal was to uncover novel early events that prime ovarian cancer cells for metastatic success.

Employing both ascites-derived cell lines and patient ascites-derived ovarian cancer cells to identify genes upregulated immediately following matrix detachment, the atypical Rho GTPase, RHOV, emerged as a consistent top hit across all models, and was notably the sole Rho GTPase represented in this signature. Rho GTPases translate extracellular cues into intracellular biochemical signals, an essential process during cellular state transitions (12–15). While the role of Rho GTPases has been well described in the context of migrating and metastasizing cells with an emphasis on the most well characterized members RHOA, RAC1, and CDC42 (14); little is known about the involvement of Rho GTPases in the initial steps following matrix detachment in general and specifically in the context of OVCA metastasis.

RHOV is an atypical Rho GTPase characterized by its rapid GDP/GTP cycling that is hypothesized to render it almost constitutively active once expressed (16–18) RHOV has been studied primarily in developmental processes, such as Xenopus neural crest development, a phenomenon analogous to epithelial to mesenchymal transition (EMT), and zebrafish epiboly where it regulates migration, adhesion, and differentiation (19–21). More recently, RHOV was found to be implicated in promoting EMT and metastatic progression in lung and breast cancer models (22–25). Despite these emerging findings, RHOV remains one of the least characterized Rho GTPases in tumor cells, and the functional relevance of RHOV in ovarian cancer progression and metastasis remains entirely unexplored. This work defines RHOV as a novel detachment-sensitive Rho GTPase and establishes RHOV as an essential regulator of ovarian cancer peritoneal metastasis for the first time, offering new insights into the molecular underpinnings of transcoelomic spread.

Materials & Methods:

Patient Ascites-Derived Cells

Primary ascites-derived cells were isolated from malignant ascites obtained from patients with high-grade serous ovarian cancer (HGSOC) treated at Magee-Womens Hospital of UPMC. Specimen acquisition was approved by the University of Pittsburgh Institutional Review Board (IRB) (Protocol #STUDY19060197) and facilitated through the ProMark Biospecimen Bank at Magee-Womens Research Institute. Written informed consent was obtained from patients by the clinical care teams prior to specimen collection, and all samples were de-identified before laboratory processing. Ascites samples were processed immediately upon receipt for ascites-derived epithelial cell isolation and cultured as previously described (26). Cells were maintained at 37 °C in a humidified incubator with 5% CO₂ in MCDB/M199 medium supplemented with 10% fetal bovine serum (FBS) and penicillin-streptomycin.

Patient-Matched Tumor Specimens

Archived, patient-matched tissue specimens, including normal fallopian tube or ovary, primary ovarian tumor, and omental metastases, from individuals diagnosed with high-grade serous ovarian cancer were obtained via an honest broker from the ProMark Biospecimen Bank at the University of Pittsburgh Magee-Womens Research Institute. The study protocol was approved by the University of Pittsburgh Institutional Review Board (IRB), (Protocol # STUDY19060197). All specimen collection and use adhered to ethical guidelines established by the World Medical Association’s Declaration of Helsinki and the U.S. Department of Health and Human Services Belmont Report.

Patient Survival Analysis

Kaplan-Meier survival analysis was performed on Pan-cancer ovarian cancer specimen (n=374) survival data using KMplotter (RRID:SCR_018753) (27), based on mean gene expression of the identified 32 upregulated genes of the detachment response signature. High and low expression cut-offs were based on upper and lower quartile. Differences were analyzed on all specimens in the patient cohort, independent of other prognostic signatures, mutational status or HRD score. Overall survival was compared between groups using the log-rank test. Plot generation and analysis were performed using GraphPad Prism (RRID:SCR_002798).

Transcription Factor Enrichment and Promoter Binding Analyses

Transcription factor enrichment analysis was performed on the detachment-sensitive gene signature using ChEA3. The input gene list consisted of genes that were differentially expressed at 2 hours under anchorage-independent conditions compared with attached conditions, as identified by RNA sequencing across three ovarian cancer cell lines. The ENCODE ChIP-seq library was used to rank transcription factors based on predicted regulatory association with the input gene set. To evaluate predicted transcription factor binding at the RHOV promoter, the Signaling Pathway Project Ominer tool was utilized. Promoter binding scores were derived using integrated ChIP-Atlas MACS2 peak-calling datasets. at the RHOV promoter region based on curated ChIP sequencing experiments.

Cell Lines and Culture Conditions

OVCA433 cells (RRID:CVCL_0475) were kindly provided by Dr. Susan K. Murphy (Duke University) and cultured in RPMI 1640 medium (Corning, 10–040-CV) supplemented with 10% fetal bovine serum (FBS). SKOV3 cells (RRID:CVCL_0C84) stably expressing luciferase were a gift from Dr. Mythreye Karthikeyan (University of Alabama at Birmingham) and maintained in RPMI 1640 medium (Corning, 10–040-CV) with 10% FBS. OVCAR3 cells (RRID:CVCL_0465) expressing luciferase were a gift from Dr. Lan Coffman (University of Pittsburgh) and cultured in RPMI 1640 (Gibco, 11875) supplemented with 10% FBS and 0.01 mg/mL bovine insulin (MilliporeSigma, I0516). FT282 cells (RRID:CVCL_A4AX) were provided by Dr. Ronny Drapkin (University of Pennsylvania) and grown in a 1:1 mixture of DMEM and Ham’s F-12 (Corning, 10–090-CV) with 2% FBS. ZTGFP mesothelial cells, stably expressing GFP, were provided by Dr. Ioannis Zervantonakis (University of Pittsburgh) and maintained in a 1:1 mixture of Medium 199 (Corning, 10–060-CV) and MCDB105 (Cell Applications, 117–500), supplemented with 10% FBS and 1% Penicillin-Streptomycin (Gibco). HEK293FT cells (Invitrogen, R70007, RRID:CVCL_6911) were cultured in DMEM (Corning, 10–017-CV) with 10% FBS. TOV21G (RRID:CVCL_3613), ES-2 (RRID:CVCL_3509), and OV90 (RRID:CVCL_3768) ovarian cancer cells were purchased from the American Type Culture Collection (ATCC). TOV21G was maintained in a 1:1 mixture of MCDB 105 (Cell Applications, 117–500) and Medium 199 (Corning, 10–060-CV) and supplemented with 15% FBS. ES-2 was cultured in McCoy’s 5A Medium (Corning, 10–050-CV) supplemented with 10% FBS. OV90 cells were maintained in a 1:1 mixture of MCDB 105 (Cell Applications, 117–500) with 1.5 g/L sodium bicarbonate (Gibco, 25080094) and Medium 199 (Corning, 10–060-CV) with 2.2 g/L sodium bicarbonate, supplemented with 15% FBS. Pt412 patient-derived cells were a kind gift from Dr. Ronald Buckanovich (University of Pittsburgh) and maintained in RPMI 1640 medium (Corning, 10–040-CV) with 10% FBS. All cell lines were maintained at 37 °C in a humidified incubator with 5% CO₂, were regularly tested for mycoplasma contamination using the EZ-PCR Mycoplasma Detection Kit (Captivate Bio, 20–700-20) and authenticated by short tandem repeat (STR) profiling (Labcorp).

Cell Culture in Ultra-Low Attachment (ULA) Conditions

Cells were trypsinized, counted, and seeded at equal viable densities into Ultra-Low Attachment (ULA) plates (Corning) to induce anchorage-independent (a-i) growth. For RNA and protein extraction, 2 × 10⁵ cells were seeded in 2 mL of medium per well in 6-well ULA plates (Corning, 3471) and incubated for either 2 hrs (early a-i, single-cell state) or 24 hrs (later a-i, multicellular aggregate formation). Parallel control cultures representing the attached state were seeded at the same density in standard tissue culture-treated 6-well plates. For functional assays, cells were seeded at 1,000 cells per well in 96-well round-bottom ULA plates (Corning, 7007) to promote multicellular aggregate formation, or at 10,000 cells per well in 96-well flat-bottom ULA plates (Corning, 3474) in serum-free medium to maintain cells in a single-cell state. To assess the impact of dissociation method on RHOV expression, cells were enzymatically or non-enzymatically dissociated using one of the following: 0.25% Trypsin-EDTA (Fisher Scientific, 25–200-056), TrypLE Express Enzyme (Thermo Scientific, 12605028), Accutase (BD Biosciences, 561527), or 10 mM EDTA in PBS (Fisher Scientific, AM9260G). Following dissociation, cells were counted and seeded at a density of 2 × 10⁵ viable cells per well in 6-well Ultra-Low Attachment (ULA) plates (Corning, 3471) in complete medium. Cells were incubated under anchorage-independent conditions for 2 hrs prior to RNA extraction for analysis of RHOV expression.

CRISPR-Cas9 Mediated Knockout of RHOV

Isogenic RHOV-knockout (KO) derivatives were generated from the three ascites-derived ovarian cancer cell lines used in our initial screen: OVCAR3, OVCA433, and SKOV3. A multiplex single-guide RNA (sgRNA) CRISPR–Cas9 strategy was employed (28) to target exon 1 of RHOV, including the translation start site. Two sgRNAs were used: gRNA1 (sequence: 5’-ACAAACCCGCTGGTAGCGCT-3’) and gRNA2 (sequence: 5’-ATGGGTACCCCGCGCGCTAC-3’), each expressed from U6 promoters on separate plasmids, each co-expressing wild-type SpCas9. To facilitate sorting, one plasmid contained a GFP marker (pSpCas9(BB)-2A-GFP (PX458), RRID:Addgene_48138) and the other an RFP marker (pU6-(BbsI)_CBh-Cas9-T2A-mCherry, RRID:Addgene_64324). Cells were transiently co-transfected with both constructs, and dual-positive cells were identified by fluorescence microscopy and sorted as single cells into 96-well plates 48 hrs post-transfection. Co-expression of Cas9 and the two sgRNAs resulted in excision of the intervening genomic region via non-homologous end joining (NHEJ) repair. Clonal populations were screened by genomic PCR using primers flanking the targeted locus. The outer primers used were: OUT-forward (5′-GGCCGCCTGAAACATGTAGA-3′) and OUT-reverse (5′-GCTGTGTCCCAGAGCTCAAT-3′). Successful deletion of the 296 bp region spanning exon 1 resulted in a 329 bp product, while wild-type alleles yielded a 625 bp amplicon. An additional screening PCR was performed using an internal primer (IN-forward: 5′-CAAGAGCAGCCTCATCGTCA-3′) paired with OUT-reverse; this yielded a 321 bp product in WT cells, while KO clones lacked amplification, confirming complete excision.

siRNA-Mediated Knockdown of RHOV

For transient RHOV knockdown, cells were transfected with either a non-targeting siRNA SMARTpool control (Dharmacon, D-001810–10-05) or a SMARTpool of siRNAs targeting RHOV (Dharmacon, L-006374–00-0005). Transfections were performed using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer’s protocol.

Cloning and Lentiviral Transduction

Wild-type (WT) human RHOV (RefSeq: NM_133639) was subcloned from the pCMV6-Entry vector (Origene, RC211121) into the pLX304 lentiviral expression vector (RRID:Addgene_25890) using Gateway recombination cloning (Thermo Scientific). Site-directed mutagenesis was used to introduce specific point mutations (G40V, S45N, and C234S) into the pLX304-RHOV WT plasmid using the Phusion Site-Directed Mutagenesis Kit (Thermo Scientific, F541). Separately, RHOV was also cloned into the pcDNA3-APEX2 vector (RRID:Addgene_49386, a kind gift from Dr. Mohamed Trebak, University of Pittsburgh) for proximity labeling experiments. Control constructs expressing Flag-tagged APEX2, with or without a nuclear export signal (NES), were used in parallel. For lentivirus production, 293FT cells were co-transfected with the pLX304-based constructs and packaging plasmids psPAX2 (RRID:Addgene_12260) and pMD2.G (RRID:Addgene_12259) using lipid-based transfection method. Viral supernatants were collected and used to transduce target cells seeded in 6-well plates the day prior. For transduction, 200 µL of viral supernatant was added per well in the presence of 0.8 µg/mL polybrene (MilliporeSigma, TR-1003) and incubated for 24 hrs. Three days post-transduction, cells were subjected to antibiotic selection using 15 µg/mL blasticidin (Thermo Scientific, A1113903).

RNA Sequencing and Transcriptomic Analysis

RNA sequencing was performed to identify both detachment-sensitive gene signatures and RHOV-regulated transcriptional pathways. For the detachment-response dataset, OVCAR3, OVCA433, and SKOV3 cells were cultured under standard adherent conditions or in ultra-low attachment (ULA) plates for 2 or 24 hrs to represent early and late anchorage-independent (a-i) states, respectively, with four biological replicates per condition. In a separate experiment, OVCA433 wild-type (WT) and RHOV-knockout (KO) cells were cultured under attached, early a-i (2 hrs), or late a-i (24 hrs) conditions, with three biological replicates per group. Total RNA was extracted using the Direct-zol RNA Miniprep Kit (Zymo Research, R2052). cDNA library preparation, sequencing, and bioinformatics analysis were performed by Novogene. Libraries were sequenced on the Illumina NovaSeq 6000 platform (RRID:SCR_016387), and reads were aligned to the human genome using HISAT2 (RRID:SCR_015530). Differential expression analysis was conducted using the DESeq2 R package (RRID:SCR_015687), with genes considered differentially expressed at an adjusted p-value ≤ 0.05. Functional enrichment analyses, including Gene Set Enrichment Analysis (GSEA), Gene Ontology (GO), and transcription factor analysis (ChEA), were performed using the BioJupies platform (RRID:SCR_016346). Heat maps and volcano plots were generated in GraphPad Prism (RRID:SCR_002798).

RNA Isolation and Semi-Quantitative Real Time RT-PCR

Total RNA was isolated using the Direct-zol RNA Miniprep Kit (Zymo Research, R2052) according to the manufacturer’s instructions, and cDNA was synthesized using the qScript cDNA Synthesis Kit (Quantabio, 95047). Semi-quantitative real-time RT-PCR was performed on a CFX Opus 96 Real-Time PCR System (Bio-Rad) using iTaq Universal SYBR Green Supermix (Bio-Rad). Gene expression was normalized to the geometric mean of housekeeping genes and analyzed using the ΔΔCt method. Primer sequences for housekeeping genes included GAPDH (forward 5′-GAGTCAACGGATTTGGTCGT-3′, reverse 5′-TTGATTTTGGAGGGATCTCG-3′), ACTB (forward 5′-AGAGCTACGAGCTGCCTGAC-3′, reverse 5′-AGCACTGTGTTGGCGTACAG-3′), TBP (forward 5′-TTGGGTTTTCCAGCTAAGTTCT-3′, reverse 5′-CCAGGAAATAACTCTGGCTCA-3′), and HPRT1 (forward 5′-TGACCTTGATTTATTTTGCATACC-3′, reverse 5′-CGAGCAAGACGTTCAGTCCT-3′). Two RHOV primer sets were used: one for general RHOV detection (forward 5′-GCATTGAGCTCTGGGACACA-3′, reverse 5′-TGGTCCAGCTGAATTAGTACG-3′), and a second set optimized for detecting RHOV loss in knockout samples (forward 5′-CCTCATCGTCAGCTACACCTG-3′, reverse 5′-GAACGAAGTCGGTCAAAATCCT-3′). Primers for early anchorage-independent (a-i) response genes included FOSB (forward 5′-CAGTCCTGTGTGAGGATTAAGG-3′, reverse 5′-CACTTCTGCCCAGAGTAAAGAG-3′), EGR3 (forward 5′-GCGACCTCTACTCAGAGCC-3′, reverse 5′-CTTGGCCGATTGGTAATCCTG-3′), IER5 (forward 5′-TTTCTCGGGACTCCTACGGAA-3′, reverse 5′-GCTCCAGGGGTTCATGTCTC-3′), and NRARP (forward 5′-TCAACGTGAACTCGTTCGGG-3′, reverse 5′-ACTTCGCCTTGGTGATGAGAT-3′).

Polysome Profiling

Polysome profiling was performed using cell lysates prepared as previously described in our prior publication (29). Briefly, cells cultured under adherent or anchorage-independent (ULA) conditions were treated with cycloheximide (100 µg/mL) for 10 minutes at 37 °C to arrest translation. After washing with ice-cold PBS containing cycloheximide, cells were lysed in polysome extraction buffer and clarified by centrifugation. The resulting supernatants were layered onto linear 20–47% sucrose gradients and subjected to ultracentrifugation at 34,000 rpm for 4 hrs and 15 minutes at 4 °C using an SW41 rotor (Beckman Coulter). Gradients were fractionated, and RNA from each fraction was extracted using TRIzol and analyzed by qRT-PCR following cDNA synthesis. For this study, previously generated and validated polysome fractionation samples were reused for downstream analysis.

Tumor Xenografts

All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Pittsburgh and conducted in compliance with institutional guidelines (approved protocol IS00023174). Mice were housed in pathogen-free barrier facilities maintained for immunodeficient strains. For intraperitoneal (IP) xenograft experiments, SKOV3-luciferase wild-type (WT) and RHOV-knockout (RHOV-KO) cells were harvested by trypsinization, washed, and counted. A total of 1 × 10⁶ viable cells suspended in 150 µL PBS were injected intraperitoneally into female NOD scid gamma (NSG) mice (Jackson Laboratory, RRID:IMSR_JAX:005557), within 2 hours of detachment form tissue culture flasks. Mice were randomized into groups (n = 10 per group) based on pilot studies. Tumor progression was monitored every 2–3 days via bioluminescent imaging using an IVIS imaging system. Mice were injected intraperitoneally with D-Luciferin (15 mg/mL, PerkinElmer, 122799) at 10 µL/g body weight, and images were acquired 10 minutes post-injection. On day 35, mice were euthanized via CO₂ asphyxiation followed by cervical dislocation. At necropsy, ascites fluid was collected and quantified, and omentum and other visible peritoneal tumor sites were excised and weighed. Omentum was imaged ex vivo following additional luciferin administration. Tissues were fixed in 10% neutral-buffered formalin (Fisher Scientific, 5735), processed for paraffin embedding, sectioned, and stained with hematoxylin and eosin (H&E). For orthotopic intrabursal xenografts, OVCAR3-luciferase WT and RHOV-KO cells (1 × 10⁶) were resuspended in a mixture of 5 µL Matrigel (Corning, 354248) and 15 µL serum-free media, and injected into the ovarian bursa of NSG female mice (n = 5 per group), (RRID:IMSR_JAX:005557). Mice were imaged every 3–4 days using IVIS, following the same luciferin injection and imaging protocol as described above. The study was terminated on day 57, at which point mice were euthanized and necropsied. Bioluminescence in the ovary and omentum was recorded ex vivo, and the number of peritoneal tumor nodules was documented.

Subcellular Fractionation

Cytoplasmic, membrane and nuclear protein fractions were isolated using the Subcellular Protein Fractionation Kit for Cultured Cells (Thermo Scientific, Cat. No. 78840), following the manufacturer’s instructions. Briefly, cells were collected and sequentially extracted using the supplied compartment-specific buffers to selectively enrich proteins from distinct cellular compartments. At each step, lysates were incubated and centrifuged as directed to isolate the soluble fraction while retaining the remaining pellet for the next extraction. The resulting fractions were stored at −80 °C until further analysis by SDS-PAGE and immunoblotting.

Western Blotting

Cells were collected either by scraping (for attached conditions) or centrifugation (for anchorage-independent conditions) and lysed in RIPA buffer (Thermo Scientific, 89901) supplemented with protease and phosphatase inhibitor cocktails (Thermo Scientific, 78443). Lysates were rotated at 4 °C for 30 minutes, followed by centrifugation at 21,000 × g for 30 minutes at 4 °C to clear debris. Protein concentrations were determined using the Pierce BCA Protein Assay Kit (Thermo Scientific, 23225). Equal amounts of protein were resolved on SDS-PAGE gels (Bio-Rad) and transferred to nitrocellulose membranes (Fisher Scientific, 45–004-001). Membranes were blocked for 1 hour at room temperature in 5% BSA (Fisher Scientific, BP1600–100) prepared in TBS containing 0.1% Tween-20 (MilliporeSigma, 9005–64-5), then incubated overnight at 4 °C with primary antibodies diluted in blocking buffer. Primary antibodies included: RHOV (Proteintech, 26620–1-AP, RRID:AB_2880576, 1:300), phospho-c-Jun (S63) (Cell Signaling Technology, 2361, RRID:AB_490908; 1:1000), total c-Jun (Cell Signaling Technology, 9165, RRID:AB_2130165; 1:1000), phospho-PAK1+PAK2+PAK3 (phospho S144 + S154 + S144 + S141) (Abcam, ab40795, RRID:AB_777032; 1:1000), total PAK1 (Invitrogen, 71–9300, RRID:AB_2534005; 1:1000), total PAK2 (Cell Signaling Technology, 2608, RRID:AB_2283388; 1:1000), phospho-PAK4 (Ser474)/PAK5 (Ser602)/PAK6 (Ser560) (Cell Signaling Technology, 3241, RRID:AB_2158623; 1:1000), total PAK4 (Cell Signaling Technology, 62690, RRID:AB_2827508; 1:1000), total PAK6 (Thermo Scientific, PA5–34938, RRID:AB_2552287; 1:1000), EEA1 (Cell Signaling Technology, 2411, RRID:AB_2096814; 1:1000), phospho-Smad2 (Ser465/467)/Smad3 (Ser423/425) (Cell Signaling Technology, 8828, RRID:AB_2631089; 1:1000), total SMAD2/3 (Cell Signaling Technology, 8685, RRID:AB_10889933; 1:1000), MMP2 (Cell Signaling Technology, 4022, RRID:AB_2266622; 1:1000), TGFBI (Proteintech, 10188–1-AP, RRID:AB_2202311; 1:1000), INHBA (Santa Cruz Biotechnology, sc-166503, RRID:AB_2125854; 1:1000), ATP5A (Abcam, ab14748, RRID:AB_301447; 1:1000), FAK (Cell Signaling, 3285, RRID:AB_2269034; 1:1000), pFAK (Cell Signaling Technology, 3283, RRID:AB_2173659; 1:1000), SRC (Cell Signaling Technology, 2109, RRID:AB_2106059; 1:1000), pSRC (Cell Signaling Technology, 2101, RRID:AB_331697; 1:1000), GAPDH (Santa Cruz Biotecnology, sc-47724, RRID:AB_627678, 1:1000), Vinculin (Sigma-Aldrich, V9131, RRID:AB_477629, 1:1000). After washing with TBST (0.1% Tween-20 in TBS), membranes were incubated with either HRP-conjugated secondary antibodies (Cytiva, NA931, RRID:AB_772210 or NA934, RRID:AB_772206; 1:10,000) for chemiluminescent detection or fluorescently labeled secondary antibodies (LI-COR Biosciences, 926–32211, RRID:AB_621843 or 926–68072 RRID:AB_10953628; 1:10,000) for near-infrared imaging. Chemiluminescent signals were developed using SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific, 34096) and visualized with the ChemiDoc XRS+ system (Bio-Rad, RRID:SCR_019690). For fluorescence detection, membranes were imaged using the Odyssey CLx imaging system (LI-COR, RRID:SCR_014579). For analysis of extracellular proteins including MMP2, TGFBI, and INHBA, cells were seeded and the next day media changed to serum-free, phenol red-free medium (Gibco, 11835). After 72 hrs, conditioned media were collected and cleared of cellular debris by centrifugation. Supernatants were concentrated using Amicon Ultra Centrifugal Filters (MilliporeSigma, UFC901024) by centrifuging at 4,000 rpm for 2 hrs at 4 °C. Protein concentrations of the concentrated media were determined using the Pierce BCA Protein Assay Kit (Thermo Scientific, 23225). SDS-PAGE and transfer conditions matched those described for whole-cell lysates, and equal protein loading was confirmed by Ponceau S staining (MilliporeSigma, P7170–1L). Band intensity was quantified using ImageJ (NIH, RRID:SCR_003070).

APEX2-Based Proximity Labeling: Affinity Pulldown

Expression of Flag-APEX2-RHOV or the control construct Flag-APEX2 was achieved via lentiviral transduction as previously described. For affinity pulldown experiments, cells expressing either Flag-APEX2-RHOV or Flag-APEX2 were incubated with 2.5 mM biotin-phenol (Sigma-Aldrich, SML2135) for 30 minutes at 37 °C, followed by H₂O₂ (Sigma-Aldrich, H1009; 30% w/w in H₂O) treatment for 30 seconds. at final concentrations of 500 µM Flag-APEX2-RHOV and 100 µM for Flag-APEX2 control to achieve equivalent biotinylation levels. The reaction was immediately quenched using PBS containing 10 mM sodium ascorbate (Sigma-Aldrich), 10 mM sodium azide, and 5 mM Trolox (Cayman Chemical, 10011659; pH 8.0), followed by three washes in the same solution (30). Biotinylated proteins were captured using Streptavidin Magnetic Beads (New England Biolabs, S1420S) and eluted by boiling in 4X Laemmli sample buffer for subsequent immunoblot analysis.

Actinomycin D Treatment

For mRNA stability analysis, transcription was inhibited using actinomycin D (Sigma-Aldrich) at a final concentration of 10 μg/mL. Cells were pre-treated with actinomycin D for 30 minutes under attached conditions before being transferred to ultra-low attachment plates, where they were maintained in a-i in the continued presence of the inhibitor for subsequent time-point collection.

Proteasomal Inhibition

Cells were treated with 10 µM MG132 (Sigma-Aldrich, 474790; Calbiochem) or an equivalent volume of DMSO as a vehicle control for 24 hrs under standard culture conditions. After treatment, cells were harvested and lysed, and protein extracts were analyzed by Western blotting to evaluate the effects of proteasomal inhibition on target protein levels.

TGF-β1 Stimulation and SMAD Activation

To assess SMAD2/3 phosphorylation in response to TGF-β1, cells were seeded and allowed to adhere overnight under standard culture conditions. The following day, culture medium was replaced with serum-free medium to minimize basal pathway activation. After 24 hrs of serum starvation, cells were stimulated with 10 ng/mL recombinant human TGF-β1 (R&D Systems, 7754-BH-005) for 0, 15, 30, 60, or 120 minutes. At each time point, cells were lysed and processed for protein extraction and subsequent Western blot analysis.

Cellular Proliferation

Cell proliferation assays were performed using the FluoReporter Blue Fluorometric dsDNA Quantitation Kit (Invitrogen, F2962) according to the manufacturer’s instructions. Briefly, OVCAR3, OVCA433, and SKOV3 cells were seeded in 96-well plates at densities of 2000, 1000, and 500 cells per well, respectively, in a volume of 200 μL of standard culture medium. Plates were harvested at 24-hour intervals over a period of three days. At each time point, culture medium was removed, and plates were stored at −80 °C. Subsequently, plates were processed with Hoechst 33258 staining for dsDNA quantification. Fluorescence intensity was measured using a Victor X fluorescence microplate reader (PerkinElmer) with excitation and emission wavelengths set at 360 nm and 460 nm, respectively. Proliferation rates were calculated by normalizing fluorescence intensity values to the initial measurement taken at day 1.

Cell Cycle Analysis

OVCA433 and SKOV3 cells were plated in six-well dishes at a seeding density of 100,000 cells/well. 48Hrs after seeding, cells were trypsanized and stained with propidium iodide (Abcam, ab139418), cell cycle status was assessed using flow cytometry (Beckman Coulter, CytoFLEX, RRID: SCR_019627) and analyzed using FlowJo software (BD Life Sciences, RRID: SCR_008520).

Live/Dead Staining

Cell viability was assessed using Calcein AM and ethidium homodimer staining. For viability assays in flat-bottom conditions, cells were seeded at a density of 10,000 cells per well in 96-well flat-bottom Ultra-Low Attachment (ULA) plates (Corning, 3474) in serum-free medium to maintain single cells or small clusters. Cells were incubated for 2, 24, 48, and 72 hrs prior to staining. For round-bottom conditions, cells were seeded at a density of 1,000 cells per well in 96-well round-bottom ULA plates (Corning, 7007) in normal culture medium to form single aggregates per well and incubated for 72 hrs prior to staining. Cells were stained with 2 µM Calcein AM and 4 µM ethidium homodimer (Sigma) diluted in PBS and incubated at 37°C for 30 minutes. Images were captured immediately afterward using a Leica Thunder Imager (RRID:SCR_023794) to determine the proportion of live (Calcein-positive, green) and dead (ethidium homodimer-positive, red) cells.

Caspase 3/7 Activity

Caspase 3/7 analysis in a-i: Caspase activity was evaluated using the Caspase-Glo 3/7 3D reagent (Promega, G8981) following the manufacturer’s protocol. Cells were seeded in 96-well flat-bottom ultra-low attachment (ULA) plates as described above, and caspase activity was measured after 48 hrs under anchorage-independent (a-i) conditions. An equal volume of Caspase-Glo 3/7 reagent was added directly to each well, and plates were incubated for 30 minutes at room temperature in the dark. Luminescence was recorded using GloMax Explorer plate reader (Promega), and background signal from medium-only wells was subtracted from all measurements.

Caspase 3/7 analysis in attached conditions: Caspase activity was evaluated using the Caspase-Glo 3/7 reagent (Promega, G8090). Cells were seeded in 96-well tissue culture treated dishes at 1000 cells/well for SKOV3 cells and 2000 cells/well for OVCA433 cells. Caspase activity was measured at 24, 48 and 72hrs post seeding. equal volume of Caspase-Glo 3/7 reagent was added directly to each well, and plates were incubated for 30 minutes at room temperature in the dark. Luminescence was recorded using GloMax Explorer plate reader (Promega), and background signal from medium-only wells was subtracted from all measurements.

Clonogenic Survival

Clonogenic assays were performed to assess single-cell survival following detachment. Briefly, 100 cells per well were seeded into standard 6-well tissue culture plates and maintained under appropriate growth conditions. After colony formation (10–14 days), cells were fixed and stained with 0.05% crystal violet solution. Colonies were imaged and quantified manually.

Spheroid Compaction

Cells were transiently labeled with CellTrace CFSE (Invitrogen, C34554) and seeded at a density of 1,000 cells per well in 200 µL of culture medium in 96-well round-bottom Ultra-Low Attachment (ULA) plates (Corning, 7007). Spheroid formation was monitored by time-lapse imaging using the IncuCyte S3 live-cell analysis system (Sartorius, RRID:SCR_003935), with images acquired every hour over a 24-hour period. Spheroid area was measured at each time point and normalized to the 1-hour time point, used as the baseline to account for initial cell settling within the imaging field.

Mesothelial Clearance

Mesothelial clearance assays were performed by initially preparing mesothelial monolayers (31). Flat-bottom 96-well plates were coated with calf skin collagen type I (Sigma-Aldrich, 8919) at a concentration of 0.05 mg/mL, incubated at 37°C, and washed three times with PBS prior to cell seeding. Mesothelial cells were then trypsinized, counted, and plated at a density of 50,000 cells per well in culture medium. Monolayer formation was confirmed 24 hrs post-seeding. For ovarian cancer spheroid formation, SKOV3 wild-type (WT) and RHOV-knockout (RHOV-KO) cells were trypsinized, counted, transiently labeled with CellTrace Far Red (Invitrogen, C34572), and seeded at a density of 1000 cells per 200 µL medium in round-bottom 96-well Ultra-Low Attachment (ULA) plates (Corning, 7007). The plates were incubated overnight to allow spheroid formation. The following day, a single spheroid was carefully transferred onto each pre-established mesothelial monolayer. Images were captured at multiple time points (0, 4, 8, 12, and 24 hrs for initial clearance, and at 120 hrs for long-term displacement) using a Leica Thunder Imager (RRID:SCR_023794). Areas cleared by the spheroids were quantified using ImageJ software (NIH, RRID:SCR_003070).

Mesothelial Adhesion

Mesothelial cells (ZT-GFP) were seeded at a density of 50,000 cells/well in 96 well dishes coated with calf skin collagen type I (Sigma-Aldrich, 8919) at a concentration of 0.05 mg/mL as described above. The following day, SKOV3 cells (WT and RHOV-KO) were trypsanized, stained with CellTrace Far Red (Invitrogen, C34572) and added to the mesothelial cells at a density of 15,000cell/well. Cells were left to incubate for 30 mins, after which each well wash washed three times using 1XPBS, 5 mins each to wash away non adherent cells. The adherent cells were then examined under a microscope (Leica Thunder Imager) and analyzed using Fiji software (RRID: SCR_002285). The number of adherent cells were then normalized to the number of cells in a seeding control dish that was not subjected to washing steps.

Wound healing

Wound healing assays were performed using the ibidi µ-Dish with Culture-Insert 3 Well (ibidi, Cat. No. 80366), which features two defined 500 µm cell-free gaps per insert to facilitate standardized migration analysis. SKOV3 cells were seeded at a density of 30,000 cells per chamber in 70 µL of complete medium and incubated for 24 hrs to allow monolayer formation. Following insert removal, wells were gently washed with PBS to eliminate residual serum, and replaced with serum-free medium. Cells were imaged every 24 hrs for 5 days to monitor wound closure, with media changed every 48 hrs.

Transwell Invasion

Invasion assays were conducted using Boyden chambers with 8 µm pore-size Transwell inserts coated with Matrigel at 0.4 mg/mL. The coating was allowed to solidify for 3 hrs at 37 °C. For baseline invasion assays, 5 × 10⁴ were suspended in serum-free medium and seeded into the upper chamber, while the lower chamber contained complete growth medium (RPMI with 10% FBS) to act as a chemoattractant. Cells were incubated overnight to allow invasion through the Matrigel-coated membrane. To evaluate the effect of TGF-β1 on cell invasion, 2.5 × 10⁴ cells were seeded in the upper chamber and treated with or without 10 ng/mL recombinant human TGF-β1 (R&D Systems, 7754-BH-005; human cell-expressed) for 24 hrs under serum-free conditions. After the incubation period, non-invading cells were removed from the upper side of the membrane using a cotton swab. Invaded cells on the lower surface were stained with 0.5% crystal violet solution, washed, and air-dried. Stained cells were visualized under a microscope, and invasion was quantified by counting cells in multiple representative fields.

Immunofluorescence Staining

For staining of adherent cells, cells were seeded into 8-well chambered slides (Falcon, 08–774-26) and allowed to attach overnight. The next day, cells were fixed with 4% paraformaldehyde (BTC BeanTown Chemical, 30525–89-4) for 10 minutes at room temperature, permeabilized with 0.25% Triton X-100 (Fisher Scientific, BP151500) in PBS (Corning, 21–040-CV) for 10 minutes, and blocked with Pierce SuperBlock T20 (TBS) Blocking Buffer (Thermo Scientific, 37535) for 30 minutes. Cells were then incubated overnight at 4 °C with anti-EEA1 primary antibody (Cell Signaling Technology, 2411, RRID:AB_2096814; 1:250) diluted in blocking buffer. The following day, cells were washed three times with PBS and incubated with Alexa Fluor™ 488-conjugated goat anti-rabbit secondary antibody (Thermo Fisher Scientific, A-11008, RRID:AB_143165; 1:1000) and Alexa Fluor 647-conjugated phalloidin (Thermo Fisher Scientific, A22287) for 1 hour at room temperature in the dark. After washing, coverslips were mounted using ProLong Gold Antifade Mountant with DAPI (Cell Signaling Technology, 8961) and cured overnight in the dark at room temperature. For wound healing experiments, cells were fixed directly in their culture wells with 4% paraformaldehyde for 10 minutes, permeabilized with 0.25% Triton X-100 in PBS, and blocked with SuperBlock for 30 minutes at room temperature. Cells were stained with Alexa Fluor 647-conjugated phalloidin and DAPI for 1 hour at room temperature in the dark, followed by PBS washes. For immunofluorescence staining of early anchorage-independent (a-i) cells, single-cell suspensions grown in ultra-low attachment (ULA) conditions were collected, fixed with 4% paraformaldehyde, and permeabilized in 0.5% Triton X-100 in TBS for 30 minutes with rotation at room temperature. Cells were blocked in SuperBlock for 1 hour at room temperature, incubated with primary antibody overnight at 4 °C under rotation, washed three times with 0.1% Triton X-100 in TBS, and then incubated with secondary antibody for 3 hrs at room temperature. After a 15-minute incubation with DAPI and Alexa Fluor™ 647-conjugated phalloidin, cells were washed again and mounted using fluorescence-compatible mounting medium. Images for all conditions were acquired using either a Leica Thunder Imager or Leica DMI8 confocal microscope.

Statistical Analysis

Unless specified otherwise, data represent the mean ± standard error of the mean (SEM) from a minimum of three independent experiments. Statistical analyses were conducted using GraphPad Prism version 10 (GraphPad Software, RRID:SCR_002798), with methods selected according to the experimental design described. A p-value of less than 0.05 was considered statistically significant.

Data Availability

The data generated in this study are publicly available in Gene Expression Omnibus (GEO) at GSE305375 and GSE305420. All other raw data generated in this study are available upon request from the corresponding author.

Results:

A Conserved Detachment-Sensitive Transcriptome Highlights RHOV as a Unique Immediate Early Response Gene in Disseminating Ovarian Cancer Cells

Building on our previous findings that ovarian cancer multicellular aggregates (MCAs) cultured under anchorage-independent (a-i) conditions regulate critical survival genes (32,33), we sought to define the immediate early transcriptomic adaptations acquired by metastasizing cells following matrix detachment. RNA-sequencing was performed 2 hours post-detachment, an interval selected to capture the nascent detachment-sensitive transcriptome before secondary stress responses emerge. Based on the common nature of transcoelomic spread across epithelial ovarian cancer subtypes, we employed ascites-derived NIH:OVCAR3, OVCA433, and SKOV3 cell lines. Differential gene expression changes revealed a conserved, detachment-sensitive gene signature shared across all three cell lines in 2 hrs a-i (Fig. 1A, B). The 32 upregulated genes in this detachment signature correlated with significantly poorer overall survival of ovarian cancer patients (Pan-Cancer; HR=1.54, P=0.003; Fig. S1A).

Figure 1. Detection of a detachment-sensitive transcriptomic signature in ovarian cancer cells led by the atypical Rho GTPase, RHOV, that is enriched in metastatic ovarian cancer tumors.

A, Schematic overview of transcoelomic metastasis. Ascites-derived ovarian cancer cell lines (OVCAR3, OVCA433 and SKOV3 cells) were grown in adherent (A) cultures, and in anchorage-independent conditions (a-i) using ultra-low attachment plates for 2 hrs (a-i 2hrs), to mimic early disseminating single suspended cells. In addition, cells were cultured for 24 hrs in a-i, where cells form MCAs. Venn diagram shows overlap of differentially expressed genes (n=4, log₂FC > 1, adjusted P < 0.05) across all three cell lines at 2hrs a-I (Created in BioRender. Abdelnaby, A. (2026) https://BioRender.com/fb39l2r). B, Heat map representing individual genes comprising the detachment-induced signature identified in A, ranked based on significance in OVCAR3 cell line. C, Functional annotation of genes in the detachment-induced signature, highlighting RHOV as the sole Rho GTPase. D, Volcano plots of differentially expressed genes at 2 hrs a-i. E, RHOV is upregulated in HGSOC patient ascites-derived tumor cells (Ascites-1, Ascites-2, and Ascites-3) cultured under a-i conditions (left schematic created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/4bwo6wb). F, RHOV mRNA is enriched in omental metastatic lesions (Tumor-OM) in 4/5 patients compared to matched primary tumors at the ovary (Tumor-OV) and normal tissue (N-FT=Normal Fallopian Tube; upper schematic created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/xrh7el0).

Functional annotation of the detachment-responsive gene signature revealed enrichment across diverse biological processes, including transcriptional regulation, metabolic pathways, chemokine, and receptor-mediated signaling, and included NRARP, FOSB, IER5, and EGR3, well-characterized immediate early response genes that have been implicated in early breast cancer hematogenous metastasis (34) (Fig. 1C; Fig. S1B). Within this signature, RHOV stood out as the sole Rho GTPase, ranking as the most significantly upregulated gene in the high-grade serous ovarian cancer (HGSOC) OVCAR3 cell line (Fig. 1D). Given the central role of Rho GTPases in metastasis and translating extracellular biophysical cues into intracellular signaling, we focused on RHOV, which had not been previously explored in ovarian cancer. To test whether RHOV induction merely reflected a global activation of the Rho GTPase family, the expression of all 20 canonical Rho GTPases was examined. Strikingly, compared to RHOV, none displayed a consistent and robust transcriptional increase during early a-i, including RHOU, the closest sequence homolog to RHOV (Fig. S2). This demonstrates that detachment specifically induces RHOV. Time-course analysis of RHOV transcript levels demonstrated that RHOV transcripts were robustly elevated as early as 30 minutes post-detachment and reached a peak at around 2 hrs (Fig. S3A-D). The magnitude of induction was inversely correlated with basal RHOV expression (Fig. S3E), and RHOV transcription was induced regardless of the method used for cellular detachment, including enzymatic and non-enzymatic dissociation (Fig. S3F). Notably, the detachment-responsive induction of RHOV expression was absent in non-mullerian human embryonic kidney HEK293 cells (Fig. S3G), while it was preserved in the TP53 mutant fallopian tube epithelial cell line FT282 (Fig. S3D). Although RHOV transcription peaked at 2hrs post-detachment, higher levels of RHOV are maintained in longer-term a-i culture (Fig. S3 A-D) but revert to baseline when reattached to plastic (Fig.S3H-J). Actinomycin-D completely abrogated the observed RHOV induction (Fig. S3K), indicating that the detachment-induced surge in RHOV reflects de novo transcription rather than mRNA stabilization. Transcription factor enrichment analysis (TFEA) on the early detachment-sensitive gene signature using ChEA3-ENCODE ranked TCF7L2 as the second most highly ranked transcription factor predicted to regulate the expression of genes within this signature (Fig. S3L). Similarly, TCF7L2 was identified as the highest-scoring transcription factor binding the RHOV promoter region (Fig. S3M). While the regulatory relationship between TCF7L2 and RHOV requires further exploration, this is notable given that RHOV was originally described as a Wnt-responsive Cdc42 homolog 2 (WRCH-2) (35), and implicates the Wnt-responsive transcription factor TCF7L2 as a potential regulator of RHOV following detachment.

Due to the understudied nature of RHOV, only a limited number of antibodies targeting RHOV protein are commercially available. Although western blotting indicated that endogenous RHOV is induced in response to detachment and enriched in the membrane fraction (Fig. S4A), we found that all commercial antibodies tested exhibited poor signal-to-noise ratios. To further validate that RHOV is actively translated, polysome profiling was carried out, which revealed that RHOV transcripts rapidly redistributed into the heavy polysome fraction shortly after detachment, indicating that RHOV transcripts are primed to be translated (Fig. S4B&C). Notably, RHOV is an atypical Rho GTPase distinguished by an accelerated GDP/GTP exchange cycle that is hypothesized to render RHOV constitutively active upon expression, suggesting that RHOV is regulated primarily at the level of transcription (16–18). The above findings show that RHOV de novo transcription is rapidly deployed upon detachment and may represent an important regulatory mechanism of this atypical GTPase in cancer cells.

To validate the clinical relevance of the above findings, we isolated ascites-derived tumor cells from patients diagnosed with metastatic HGSOC. Closely recapitulating the rapid upregulation observed in cell line models, these primary tumor cells exhibited a marked increase in RHOV mRNA within 2 hours following detachment (Fig. 1E). In addition, in four of five matched clinical specimens from patients diagnosed with metastatic disease, RHOV was significantly enriched in metastatic nodules compared to both the primary tumor and non-malignant control fallopian tube tissue (Fig. 1F), suggesting that cells with elevated RHOV during initial detachment gain a selective advantage in successful metastatic colonization to the omentum.

Genetic Ablation of RHOV Impairs Transcoelomic Dissemination and Peritoneal Colonization of Ovarian Cancer Cells

To elucidate the functional role of RHOV, we established isogenic RHOV-knockout (KO) derivatives from the three ascites-derived ovarian cancer cell lines used in our initial screen. Using a multiplex single-guide RNA (sgRNA) CRISPR-Cas9 strategy to target RHOV exon 1, inclusive of its translation start site (Fig. S5A), RHOV was successfully knocked out in OVCA433 and SKOV3 cells, whereas OVCAR3 cells, which harbor RHOV gene amplification (DepMap; RRID:SCR_017655), displayed a partial deletion (Fig. S5B, C). RT-PCR confirmed the loss of detectable RHOV transcripts in both attached and a-i conditions in OVCA433-KO and SKOV3-KO clones and partial KO in OVCAR3 cells (Fig. S5D).

We first investigated whether RHOV is critical for transcoelomic ovarian cancer metastasis using two complementary in vivo models. To assess the role of RHOV in metastatic dissemination from the ovary, an orthotopic intra-bursal xenograft model was employed (Fig. 2A). Following ovarian tumor establishment, parental wild-type OVCAR3-WT cells efficiently disseminated throughout the peritoneum, while tumors from partial RHOV-KO OVCAR3 cells remained confined to the primary ovarian site (Fig. 2B, C), with complete absence of peritoneal lesions in the OVCAR3 RHOV-KO group (Fig. 2D), indicating that RHOV is critical for metastatic progression from the primary tumor site. Ex vivo analysis showed that ovarian bioluminescence was reduced in RHOV-KO versus WT tumors, and omental metastasis was completely abrogated by partial RHOV-KO (Fig. 2E, F).

Figure 2. Genetic deletion of RHOV impairs transcoelomic dissemination and peritoneal colonization of ovarian cancer cells.

A, Schematic of orthotopic intra-bursal xenograft model. Luciferase-labeled OVCAR3 cells were injected into the left ovary-bursa of NSG mice (Created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/onry5aw). B, Partial RHOV-KO OVCAR3 tumors remain confined to the ovary, while OVCAR3 WT tumors disseminate widely as demonstrated by bioluminescence imaging. C, Quantification of total peritoneal bioluminescence (n=5 mice/group, two-way ANOVA, P < 0.0001). D, Quantification of peritoneal tumors in OVCAR3-WT vs RHOV-KO mice (n=5; Unpaired t-test). E, Mice injected with OVCAR3 RHOV-KO cells do not form metastatic omental tumors. Representative gross (left) and bioluminescence (right) images of post-necropsy ovary and omental tumors. F, Quantification of ex vivo bioluminescence, demonstrating reduced primary tumor burden and complete absence of metastasis to the omentum by RHOV-KO OVCAR3 cells (n=5; Ovary: Unpaired t-test, Omentum: Mann Whitney). G, Schematic of the intraperitoneal tumor model where a-i suspended SKOV3 cells are directly injected into the peritoneal cavity (Created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/r9ie7ly). H, I, Representative in vivo bioluminescence images and quantification of total photon flux show significantly reduced metastatic burden in SKOV3 RHOV-KO tumor-bearing NSG mice (n=10; two way ANOVA, P < 0.0001). J, Volume of ascites collected at necropsy (n=10, Unpaired t-test). K, L, Representative images and quantification of peritoneal tumor nodules show reduced size and incidence of peritoneal metastases in the SKOV3 RHOV-KO group (n=10, Unpaired t-test). M, N, Representative ex vivo bioluminescence images of excised omentum and quantification (n=10, Unpaired t-test). O, P, Representative gross images and H&E staining (scale bar=2000 μm) of omentum from SKOV3 WT and RHOV-KO injected mice, and quantification of omental weight (n=10, Unpaired t-test).

In addition, an intraperitoneal (IP) xenograft metastasis model was employed, where cells are directly introduced into the peritoneal cavity in an anchorage-independent state within two hours of their detachment (Fig. 2G). This model specifically assesses the competence of disseminated cells forced into suspension to resist anoikis and colonize the peritoneal surfaces following injection. NSG mice injected with SKOV3 RHOV-KO cells exhibited significantly reduced bioluminescent signal intensity and spatial metastatic distribution compared to their WT counterparts (Fig. 2H-I). Consistent with diminished peritoneal colonization, RHOV-KO tumor-bearing mice also accumulated substantially less ascites (Fig. 2J). Moreover, both the size (Fig. 2K) and the frequency (Fig. 2L) of peritoneal tumor nodules were significantly lower in RHOV-KO xenografts, and these findings were validated using an independent SKOV3 RHOV-KO clone (Fig. S6A-D). Ex vivo imaging (Fig. 2M, N) and histological evaluation (Fig. 2O, P) of excised omentum further confirmed negligible tumor cell colonization by RHOV-KO tumor cells. These data underscore that RHOV is essential for the metastatic colonization of suspended ovarian cancer cells within the peritoneal cavity, and collectively, the above in vivo findings support a role for RHOV in facilitating the metastatic potential of ovarian cancer cells.

Detachment-Induced RHOV Expression Mediates Anoikis Resistance

Having established a critical role for RHOV in peritoneal metastasis in vivo, we next delineated the functional role of RHOV throughout distinct stages of the ovarian cancer metastatic cascade. We first tested the effect of RHOV deletion under attached culture conditions. Comparison of growth rates between WT and RHOV-KO cells across multiple ovarian cancer cell lines demonstrated only modest, cell line-dependent reductions in proliferation (Fig. S7A-C) with no detectable differences in cell cycle dynamics (Fig. S7D-E), or apoptosis (Fig. S7F-G). These findings suggest that intrinsic defects in proliferation or survival under adherent conditions are unlikely to account for the marked suppression of metastatic tumor burden observed in vivo following RHOV-KO.

Given the robust induction of RHOV following detachment, we next defined the role of RHOV during post-detachment stages of the metastatic cascade. Single-cell survival following detachment is a key prerequisite for successful transcoelomic dissemination (5). To model this, OVCA433 cells were cultured in a-i as dispersed single entities or small clusters, using flat bottom ULA plates (Fig. 3A). While RHOV-KO did not influence cell viability in adherent conditions (Fig. S7F-G), RHOV-KO significantly increased susceptibility to anoikis, with a continuous rise in cell death up to 72 hrs in a-i conditions (Fig. 3B-D). Clonogenic survival assays further showed that RHOV-KO cells have markedly reduced colony-forming capacity when seeded as dispersed single cells (Fig. S8A). Subsequent to detachment ovarian cancer cells typically aggregate into MCAs, which further promotes collective anoikis resistance and metastatic efficiency (10,36,37). MCAs were generated by culturing cells in round-bottom ULA plates to promote aggregation and live/dead staining revealed significantly increased cell death in RHOV-KO MCAs across all tested cell lines (Fig. 3E-I). These findings were further validated in a second RHOV-KO clone in SKOV3 cells (Fig. S8B,C). Additionally, siRNA-mediated RHOV depletion in OVCA433 and SKOV3 similarly increased cell death and significantly elevated caspase activity under a-i conditions (Fig. S8D,E). Together, these findings illustrate that RHOV induction during early detachment is critical for both immediate single-cell anoikis resistance and continued viability of MCAs in a-i (Fig. 3J; i).

Figure 3. Detachment-induced RHOV expression promotes anoikis resistance in single ovarian cancer cells and multicellular aggregates.

A, Cells were cultured in flat-bottom ULA plates in the absence of serum to maintain cells dispersed as single cells and small clusters in suspension in a-i. Live and dead cells were stained using Calcein-AM (green) and ethidium homodimer (red), respectively (Created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/bni1l7k). B, RHOV-KO in OVCA433 cells significantly increases single-cell anoikis susceptibility. Representative fluorescent images of Live/Dead cell fractions in flat-bottom ULA a-i culture (scale bar=500μm). C, Representative 10x magnification from 72Hrs time point (scale bar=500μm). D, Quantification of cell death in a-i (n=3, two way ANOVA P <0.0001, Sidak’s multiple comparisons post hoc test P values shown). E, Representative images of MCAs formed by OVCA433, SKOV3, and OVCAR3 WT cells and following RHOV-KO after 72 hrs in round-bottom ULA plates (scale bar=500μm). F, Cells cultured under a-i conditions in round-bottom ULA plates in the presence of serum promote MCAs (schematic created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/uflz3sn). G-I, Quantification of live/dead cells (Calcein-AM/ethidium homodimer) at 72 hrs in a-i reveals significantly increased cell death in RHOV-KO MCAs (n = 3, Unpaired t-test). J, RHOV promotes anoikis resistance and is necessary for single-cell survival and sustained MCA viability following matrix detachment (Created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/v3gfk7b).

RHOV Is Required for Structural Integrity and Invasive Behavior of Ovarian Cancer Cell MCAs

MCAs play a pivotal role in invading the peritoneum and colonizing peritoneal organs, processes shown to be influenced by the physical properties of the aggregates, including their compaction (38,39). We noted that RHOV-KO MCAs often appeared morphologically loose and poorly compacted (Fig. 3E). Using a spheroid compaction assay RHOV KO significantly impaired compaction of MCAs compared to WT controls in both SKOV3 and OVCAR3 lines, which form compact spheroids in a-i (Fig. 4A, B). This was further confirmed in a second SKOV3 RHOV-KO clone (Fig. S9A). OVCA433 cells, which inherently form loosely aggregated MCAs and exhibit low invasive capacity, showed no measurable difference in MCA compaction upon RHOV deletion (Fig. S9B). These findings suggest that RHOV’s role in early a-i facilitates subsequent MCA compaction in cells that inherently have the capacity to form tight MCA clusters. Since MCA compaction reflects, in part, the efficiency of homotypic tumor cell-cell adhesion, we also examined whether this adhesion defect in RHOV-KO cells extended to heterotypic interactions with the mesothelium, the cellular monolayer lining the peritoneal organs and omentum. Using a mesothelial adhesion assay, we found that RHOV knockout significantly reduced the ability of SKOV3 cells to adhere to the mesothelial monolayer, indicating impaired tumor-mesothelium attachment, a necessary step for peritoneal colonization (Fig. S9C).

Figure 4. RHOV promotes MCA compaction, mesothelial clearance, and invasive behavior in ovarian cancer cells.

A, B, MCA spheroid compaction assay shows reduced compaction in SKOV3 RHOV-KO (A) and OVCAR3 partial RHOV-KO cells (B; scale bar = 500 μm; n=3, two way ANOVA, P < 0.0001). C, Schematic of the mesothelial clearance assay. RFP-labeled ovarian cancer spheroids were co-cultured atop a GFP-expressing mesothelial monolayer on collagen to mimic peritoneal colonization (Created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/p77an1v). D, Representative images of mesothelial clearance at indicated timepoints demonstrate that SKOV3 RHOV-KO spheroids fail to displace the mesothelium and maintain cleared zones (scale bar = 500 μm). E, Quantification of mesothelial clearance area over time (n=3; two way ANOVA P value shown). F, Representative images and quantification of wound healing assay demonstrate loss of migratory capacity in SKOV3 RHOV-KO cells (scale bar=100 μm, n=3; two way ANOVA). G, Transwell invasion assay demonstrates severely reduced Matrigel invasion in SKOV3 RHOV-KO cells, scale bar = 275 μm (n=3, unpaired t-test). H, Summary schematic illustrating RHOV’s role across sequential steps of metastatic progression (Created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/76fw27w).

We next tested whether impaired MCA compaction and mesothelial adhesion following RHOV loss translated into functional defects in peritoneal colonization using a mesothelial clearance assay, in which pre-formed, RFP-labeled ovarian cancer MCAs were co-cultured atop a confluent monolayer of GFP-expressing mesothelial cells on a collagen matrix (Fig. 4C). This assay models mesothelial cell displacement by cancer cells and their subsequent engagement with the submesothelial extracellular matrix required for implantation (31,40–42). Using SKOV3 cells, the most invasive line in our panel, we found that RHOV-depleted MCAs were markedly impaired in their ability to clear the mesothelium (Fig. 4D, E). Notably, after prolonged co-culture, RHOV-KO MCAs failed to maintain mesothelial clearance, appearing to grow atop the intact mesothelial monolayer rather than successfully displacing it or competing for binding to the collagen substrate below. This highlights a key deficiency in the metastatic competence of RHOV-deficient MCAs. Moreover, RHOV-KO in SKOV3 cells, which exhibit the highest RHOV expression in attached conditions (Fig. S3E), completely abolished migration (Fig. 4F). This was similarly observed by siRNA-mediated RHOV knockdown (Fig. S9D). Complementary transwell invasion assays revealed that RHOV-KO cells also exhibited severely reduced invasion through Matrigel (Fig. 4G). Taken together, we show that RHOV is a pivotal regulator of sequential metastatic events that follow the detachment of ovarian cancer cells, including anchorage-independent survival (Fig.4H; i), MCA compaction (Fig. 4H; ii), mesothelial adhesion and clearance (Fig. 4H; iii), as well as migration and invasion (Fig. 4H; iv).

RHOV Regulates Pro-Metastatic Transcriptional Programs in Anchorage-Independence

To dissect the mechanistic underpinnings of RHOV-mediated regulation of ovarian cancer metastasis, we performed bulk RNA sequencing following RHOV-KO in OVCA433 (Fig. 5 Fig. S10), as these cells display the lowest basal RHOV expression in adherent conditions and demonstrate robust induction of RHOV expression upon detachment that is maintained at relatively high levels in later a-i time points (Fig. S3B). In adherent cells, Gene Ontology (GO) enrichment analysis revealed that RHOV loss was associated with downregulation of cytokine-mediated signaling, including interleukin and interferon responses (Fig. 5A, Fig. S10A). Upon transition to early anchorage-independent conditions (2 hrs a-i) where RHOV expression peaks, RHOV-KO cells displayed robust suppression of pro-metastatic pathways, including cell-cell junction organization, positive regulation of cell adhesion, cell migration, negative regulation of apoptosis, and extracellular matrix organization (Fig. 5B). Downregulation of these pathways correlates with the observed phenotypes in RHOV-KO cells, such as diminished survival in a-i conditions, impaired MCA compaction, mesothelial adhesion, and reduced migratory capacity. Genes driving the topmost downregulated pathway included MYLK, a key regulator of actomyosin contractility, as well as SERPINE1, F2RL1, and F2R, all of which are known to govern migratory signaling in cancer cells (Fig. S10B) (43–46). Notably, 23% of the signaling pathways induced in WT OVCA433 cells in response to detachment are also RHOV-dependent, including regulation of cell migration, motility, adhesion, and ECM organization (Fig. S10C, D).

Figure 5. RHOV regulates c-Jun signaling and cytoskeletal remodeling in ovarian cancer cells.

A-C, Bulk RNA-sequencing of OVCA433 WT and RHOV-KO cells cultured under three conditions: adherent (A), early a-i (2 hrs; B), and late a-i (24 hrs; C). Left panels: volcano plots showing distribution of differentially expressed genes between WT and KO cells. Right panels: Gene Ontology (GO) Biological Pathways enrichment analysis of significantly downregulated pathways in RHOV-KO cells in each condition (n=3) (cell images created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/yo1h84e). D, Transcription factor enrichment analysis (TFEA) of differentially expressed genes at the 2 hrs a-i timepoint identifies c-Jun as the top predicted downregulated transcriptional regulator in RHOV-KO cells. E, GO enrichment of RHOV-dependent c-Jun–regulated genes shows their convergence on pathways related to actin cytoskeleton organization, anchoring junctions, and cell motility. F-G, Western blots representative image (F) and quantification (G) showing reduced phospho-c-Jun (Ser63) in RHOV-KO OVCA433 cells (n=3, two-way ANOVA, Sidak post hoc P values shown). H, Phalloidin staining shows reduced polymerized F-actin in RHOV-KO cells under early a-i (2hrs) conditions, indicating impaired actin cytoskeletal remodeling in both OVCA433 and SKOV3 RHOV-KO cells (scale bar = 100 μm). I, Quantification of phalloidin integrated density (normalized to area of DAPI) in SKOV3 and OVCA433 2 hrs ai cells in (H), (8–10 spheroids/condition, unpaired t-test).

The transcriptional suppression of pro-metastatic pathways observed during early a-i persisted into the 24-hour anchorage-independent condition, following RHOV-KO. At this later stage, downregulated pathways included cell migration, extracellular matrix (ECM) organization, cytoskeletal organization, and cell-matrix adhesion (Fig. 5C). The leading-edge genes within the top pathway included several ECM structural components (COL4A1, COL4A2) and a broad repertoire of integrins (ITGA1, ITGA3, ITGA5, and ITGA7) all of which are critical for mediating MCA compaction, binding to collagen underlying mesothelium, and subsequent invasion (Fig. S10E) (47–49), processes that we show to be impaired in RHOV-KO cells. Given the coordinated downregulation of multiple integrins in RHOV-KO MCAs, we examined whether this transcriptional remodeling translated into altered canonical integrin signaling. We assessed phosphorylation of the integrin-associated kinases FAK and SRC under both attached conditions and in MCAs. No differences in either basal or anchorage-independent FAK or SRC phosphorylation were observed between WT and RHOV-KO cells at either stage (Fig. S11), indicating that, despite broad integrin downregulation, the functional consequences of RHOV loss appear to occur independently of canonical FAK/SRC signaling.

Together, these results demonstrate that RHOV drives pro-metastatic transcriptional programs during anchorage independence, and that this transcriptional rewiring is initiated as early as 2 hours following detachment. These early adaptations may subsequently shape the pro-metastatic capacity of MCAs, suggesting that RHOV-driven programs activated during the first hours following detachment contribute to the establishment of long-term pro-metastatic traits.

RHOV Orchestrates c-Jun Signaling and Actin Cytoskeletal Remodeling in Disseminating Ovarian Cancer Cells

To uncover key effectors downstream of RHOV that mediate RHOV-dependent transcription and pro-metastatic function during early dissemination, we focused on the 2 hrs a-i timepoint, where RHOV expression peaks and RHOV-driven pro-metastatic transcriptional programs are first induced (Fig. 5B). Transcription factor enrichment analysis (TFEA) revealed c-Jun as the top-ranked transcription factor predicted to be functionally suppressed in RHOV-KO cells in early a-i (Fig. 5D). Gene ontology analysis of RHOV-dependent c-Jun-associated transcripts again revealed a strong enrichment in pathways regulating cell motility, anchoring junction assembly, and actin cytoskeleton organization (Fig. 5E). These findings position c-Jun as a potential effector of the RHOV pro-metastatic transcriptional response. This was confirmed by immunoblotting in both OVCA433 and SKOV3 cells, where WT cells exhibited a significant induction of c-Jun (S63) phosphorylation upon transition from attached to early a-i conditions, and this response was markedly attenuated in RHOV-KO cells (Fig. 5F-G, Fig. S10F). Total c-Jun levels were also reduced, aligning with reports that phospho-c-Jun can enhance its own transcription through positive feedback mechanisms (50). The above data support an established role of RHOV in activating c-Jun signaling (23,51), but demonstrate for the first time that this RHOV-dependent signaling axis is functionally engaged in metastatic ovarian cancer cells following detachment. Given the convergence of RHOV and c-Jun on pathways governing cytoskeletal dynamics, we assessed polymerized F-actin assembly by phalloidin staining at the 2-hour a-i time point in both OVCA433 and SKOV3 cells. RHOV-KO cells exhibited markedly reduced polymerized F-actin intensity relative to their WT counterparts (Fig. 5H, I). These data show that RHOV facilitates actin remodeling following cancer cell detachment, a process well established as essential for metastatic success and likely critical during early stages of dissemination (52,53).

RHOV has been reported to localize to both the plasma membrane and early endosomes, two hubs of actin-mediated activity (54). We similarly observed enrichment of RHOV in membrane fractions in detached conditions (Fig. S4A) and association with the early endosome antigen, EEA1 using a proximity labeling approach (Fig. S12A). Moreover, RHOV-KO cells displayed accumulation of EEA1-labeled vesicles, both during early a-i in OVCA433, where RHOV is highly induced, and under attached conditions in SKOV3, which express high basal RHOV levels (Fig. S12 B, C), suggesting that loss of RHOV may impair endosomal maturation or trafficking, frequently linked to actin disruption (55).

Given that SMAD2, SMAD3, and SMAD4, canonical mediators of TGFβ receptor (TGFBR) signaling, were among the top predicted downregulated transcriptional regulators in RHOV-KO cells (Fig. 5D), and that TGFBR signaling requires trafficking through EEA1-positive early endosomes (56–58), we further tested if RHOV is also an upstream regulator of TGF-β signaling in ovarian cancer cells. While WT cells robustly increased SMAD2/3 phosphorylation and enhanced their invasive capacity following TGF-β treatment, RHOV-KO cells completely failed to respond (Fig. 6A, B, Fig. S13A). RHOV-KO cells also showed impaired induction of canonical SMAD targets, including MMP2, TGFBI, and INHBA (Fig. 6C-E, Fig. S13B, C). Although we did not directly test the link between the effects on the actin cytoskeleton and EEA1 vesicles in RHOV KO cells to the defects in TGF-β signaling, these findings demonstrate that RHOV is an important regulator of multiple pro-metastatic signaling pathways that have previously been linked to regulation by alterations in cytoskeletal dynamics.

Figure 6. RHOV is required for TGFBR signaling and downstream transcriptional activation in ovarian cancer cells.

A, Western blot image and quantification of SMAD2/3 phosphorylation over a 0–120 min time course following treatment with TGFβ1 (10 ng/mL) in SKOV3 WT and RHOV-KO cells. WT cells show robust p-SMAD2/3 induction, while KO cells fail to respond (n=3, two way ANOVA p-value shown). B, Matrigel invasion assay with exogenous TGFβ1 treatment (10ng/ml), or vehicle control (NT), demonstrates that SKOV3 RHOV-KO cells fail to invade, scale bar = 275 μm (one way ANOVA P<0.0001, Sidak’s multiple comparisons test P value shown). C, Volcano plot of RNA-sequencing data comparing RHOV-KO to WT OVCA433 cells in 2 hr a-i conditions, demonstrates significant decrease in transcript expression of canonical SMAD2/3 target genes MMP2, TGFBI, INHBA following RHOV-KO. D, E, Western blot (D) analysis and densitometric quantification (E) of secreted MMP2, TGFBI, and INHBA in media from SKOV3 WT and RHOV-KO cells, shows reduced secretion in RHOV-KO cells (Ponceau staining shown as loading control; normalized to total protein; n = 3, mean ± SEM, Unpaired t-test).

RHOV’s GTP-binding and Membrane Localization Are Required for Signaling and Pro-metastatic Function

To define the molecular features of RHOV essential for its function, we rescued RHOV-KO cells with wild-type (WT) RHOV and three targeted mutants: constitutively active (CA; GTP-bound; G40V); dominant-negative (DN; GDP-bound; S45N), and a membrane localization-deficient (ΔMEM) mutant (C234S), which harbors a cysteine-to-serine substitution in the c-terminal CFV motif that blocks palmitoylation and membrane targeting, while preserving GDP/GTP cycling (Fig. 7A) (54). These constructs were stably expressed in RHOV-KO SKOV3 cells to evaluate their capacity to rescue key RHOV-dependent functions. Both wild-type RHOV (WT) and the constitutively active (CA; G40V) mutants restored resistance to detachment-induced cell death, indicating that RHOV GTP-binding is required for anchorage-independent survival. In contrast, neither the dominant-negative (DN; S45N) mutant nor the membrane localization-deficient (ΔMEM; C234S) mutant conferred any survival advantage under a-i conditions (Fig. 7B, C), suggesting that both GTP-binding and membrane targeting are essential for this function. A similar pattern was observed in wound healing assays: only RHOV-WT and CA restored migratory capacity in RHOV-KO cells, while DN and ΔMEM remained ineffective (Fig. 7D, E). Moreover, phalloidin staining of cells at the leading edge of migrating cells in the wound healing assay showed that WT-RHOV re-expression in RHOV-KO cells reinstated actin-rich protrusive structures, underscoring the restoration of RHOV-dependent cytoskeletal remodeling (Fig. 7F). This was further accompanied by reactivation of c-Jun, a key driver of cytoskeletal transcriptional programs downstream of RHOV (Fig. 7G, H). As with earlier assays, RHOV-DN and ΔMEM did not rescue c-Jun phosphorylation, unlike RHOV-WT and CA constructs. Notably, the indistinguishable behavior of RHOV-WT and the GTP-locked CA mutant across both assays further supports the hypothesis that WT RHOV behaves as a constitutively active Rho GTPase under physiological conditions.

Figure 7. RHOV’s GTPase activity and membrane localization are required to restore pro-metastatic functions in RHOV-deficient ovarian cancer cells.

A, Schematic of RHOV constructs used in rescue experiments: wild-type RHOV (WT), GTP-locked constitutively active G40V (CA), dominant-negative GDP-bound S45N (DN), and palmitoylation-deficient membrane localization C234S (ΔMEM; created in BioRender. Abdelnaby, A.E. (2026) https://BioRender.com/7xr48mg). B, C, WT and CA RHOV but not DN or ΔMEM RHOV-mutants restore anoikis resistance in RHOV-KO SKOV3 aggregates after cells were cultured in a-i conditions for 72Hrs. Representative images of live/dead cell fractions (B) (Calcein-AM/Ethidium Homodimer, scale bar = 500 μm), and quantification (C; n=3; One-way ANOVA P <0.0001, Sidak’s post hoc test P value shown). D, E, RHOV-WT and CA restore migratory capacity of SKOV3 RHOV-KO cells. Representative wound healing assay images, scale bar = 100 μm (D) and quantification of % wound closure at day 7 (E; n=3; One-way ANOVA P=0.0009, Sidak’s post hoc test P value shown). F, Phalloidin staining reveals rescue of actin architecture at the leading edge of migrating SKOV3-RHOV-KO cells by RHOV-WT expression. Images acquired in a wound healing assay (scale bar = 500 μm). G, H, Characterization of c-Jun/Pak pathway components in SKOV3 RHOV-KO+EV vs RHOV-rescue mutants. Representative Western blots (G) and densitometry quantification (H; n=3; One-way ANOVA, Sidak’s post hoc test P value shown).

Given that RHOV-induced c-Jun activation likely involves upstream kinase signaling, and prior studies implicating RHOV in regulating p21-activated kinases (PAKs) (16,25,59–61), we investigated how RHOV-rescue affects PAK phosphorylation. In RHOV-WT and CA expressing cells, we observed a striking reduction in phospho-PAK1/2/3 (Ser144), accompanied by a loss of total PAK1 and PAK2 protein levels (PAK3 was undetectable) (Fig. 7G, H). This aligns with previous reports suggesting that following PAK activation, RHOV induces PAK1 autophosphorylation-dependent degradation (61) and further extends this RHOV-dependent regulatory mechanism to additional group I PAKs in ovarian cancer cells. To confirm that the observed reductions in PAK levels were due to rapid proteasomal degradation, we treated cells with the proteasome inhibitor MG132 (Fig. S14). This restored both phospho- and total PAK1/2/3 levels in WT and CA expressing cells, further supporting the activation-dependent degradation model (61). We also observed a reduction in phospho-PAK4/5/6 in RHOV-WT and CA cells, while total PAK4 and PAK6 protein levels remained unchanged (PAK5 was undetectable). This is intriguing as group II PAK activity is regulated by GTPases, but their phosphorylation is typically GTPase-independent (62), and RHOV was previously shown to interact with PAK6, but not alter its phosphorylation (60).

The above data collectively demonstrate that RHOV’s ability to drive metastatic behavior requires both its GTP-bound state and proper membrane localization. When these features are intact, RHOV promotes anoikis resistance, migration, actin cytoskeletal remodeling, and signaling via PAK/c-Jun. Together, these results support a model in which RHOV operates as a constitutively active, membrane-associated Rho GTPase that is highly induced in response to detachment during peritoneal metastasis.

Discussion:

The majority of ovarian cancer patients present with advanced disease, which is characterized by widespread peritoneal dissemination. The ongoing cycle of tumor cell shedding, dissemination, and reseeding underlies the poor survival associated with metastatic disease. Consequently, considerable attention has focused on the adaptations that enable multicellular aggregates (MCAs) to persist in ascites and colonize secondary sites (32,33,47,63–65). Yet, the earliest cellular responses triggered immediately upon detachment from the extracellular matrix remain less well defined. Our findings underscore the importance of investigating these acute detachment-induced adaptations, which we show to play a foundational role in priming disseminating cells for subsequent anoikis resistance, MCA formation, and peritoneal colonization.

By profiling transcriptional changes of ascites-derived ovarian cancer cells immediately after detachment, we identified a conserved, detachment-sensitive gene signature induced within 2 hours of a-i culture. We hypothesize that this signature reflects an acute stress response adopted by ovarian cancer cells upon detachment, as it is enriched in immediate early genes, including the transcription factors FOSB, IER5, and EGR3, which are known stress-induced genes largely studied in neuronal plasticity and memory formation (66). Although less is known about their function in tumor cells, several of these immediate early genes were recently independently identified in an in vivo model of early hematogenous metastatic spread in breast cancer (34), underscoring their potential role in peritoneal metastatic adaptation. Within this detachment-sensitive gene signature, we focused on RHOV, the sole Rho GTPase and most significantly upregulated gene in OVCAR3 cells. RHOV induction was validated in primary tumor cells isolated from patient ascites, supporting the pathophysiological relevance of this response. The functional importance of rapid RHOV transcriptional upregulation was immediately apparent by the observation that genetic ablation of RHOV profoundly impaired metastatic progression in vivo. These findings establish RHOV not merely as a marker of detachment-induced stress but as a critical effector of early dissemination and peritoneal colonization, whose induction is essential for successful metastatic progression.

Unlike classical Rho GTPases, whose activity is dynamically modulated by Guanine nucleotide exchange factors (GEFs), GTPase-activating proteins (GAPs), and GDP dissociation inhibitors (GDIs), RHOV (also known as Chp, Cdc42 homologous protein or Wrch-2 protein, Wnt responsive Cdc42 homolog-2) is an atypical Rho GTPase, predicted to exist predominantly in a constitutively active GTP-bound state (16–18). Additionally, distinct from prenylated Rho GTPases, RHOV undergoes reversible C-terminal palmitoylation, a modification that may limit GDI-mediated cytosolic sequestration while enabling dynamic regulation of membrane engagement (54,67). Thus, RHOV is predicted to largely bypass canonical regulation through GEFs and GAPs; instead, its activity is thought to be governed through tightly regulated expression. Here, we show for the first time that loss of cell attachment can rapidly drive RHOV de novo transcription in tumor cells.

This mirrors findings in Xenopus neural crest development, where RHOV is transiently induced in a narrow window during specification, just prior to detachment, migration, and differentiation (19,20). This tight upregulation was found to be essential for proper development. Such transient, tightly controlled expression may explain why RHOV has evaded prior detection in ovarian cancer studies. Indeed, RHOV has only surfaced as a top hit in CRISPR-based functional screens, including in models of breast cancer metastasis and Zika virus entry (25,68). Aside from those studies, RHOV’s role in cancer has been explored only through targeted approaches, primarily in lung cancer, where it was shown to promote migration and invasion (22,23). Our work now extends these observations, placing RHOV as a novel mediator of detachment-induced stress adaptations, cytoskeletal regulation, and metastatic fitness. Our findings also reveal the context-dependent nature of RHOV induction. While RHOV expression peaks immediately following detachment at 2hrs a-i, it remains elevated when cells continue to be cultured in a-i long term, albeit at lower levels (Fig 1E, Fig. S3A-D). This contrasts with cells placed back into attached culture conditions, which revert to baseline levels (Fig. S3H-J). In addition, RHOV expression is elevated in omental metastatic tumors compared to matched ovarian tumors (Fig.1F). These findings suggest that different tumor microenvironments may contribute to the regulation of RHOV and may be related to differences in the biophysical cues exerted by these tumor environments. It also suggests that RHOV may have sustained function in later stages of metastasis, although this requires further exploration. We also found that RHOV is upregulated in TP53 mutant FT282 cells upon detachment. This is particularly interesting given that transformed fallopian tube secretory epithelial cells must detach from the fallopian tube STIC to spread to the ovary during early stages of HGSOC development. These findings suggest a potential role for RHOV not only in peritoneal metastasis, but also in the earlier phases of spread from the fallopian tube to the ovary. Whether RHOV also has specific roles during early tumor dissemination from the fallopian tube to the ovary requires further investigation.

RHOV-deficient cells exhibit failure to resist anoikis, form structurally loose and disorganized aggregates in suspension, and are markedly impaired in their ability to clear mesothelial layers, migrate, and invade through extracellular matrices, all key steps of the transcoelomic metastatic cascade. This dependence on RHOV likely explains the dramatic loss of metastatic burden observed in vivo. From the above findings, we anticipate that a failure to metastasize in the absence of RHOV is largely due to increased susceptibility to anoikis and the inability to form MCAs. In turn, this limits the cells’ capacity to engage with and clear the mesothelium lining the peritoneal cavity. RHOV is required for clonogenic survival, which further supports a role for RHOV in single-cell survival. Moreover, the RHOV-induced transcriptional re-programming during early stages of detachment, such as the initiation of downstream c-Jun-dependent pro-metastatic signaling pathways, is likely key for the initiation of metastasis and colonization of the peritoneal space during later stages of metastasis. Interestingly, transcriptomics profiling following RHOV KO identified several immune-related signaling pathways, including interferon response and cytokine signaling downstream of RHOV. Future studies investigating the role of RHOV in immune-competent models will help to elucidate if RHOV affects the tumor immune environment and how this influences tumor metastasis.

A caveat to our study is that we did not utilize an inducible knock-down approach to delineate the temporal order of RHOV action during different stages of metastasis. For example, it is currently unclear if the lack of MCA formation in response to RHOV KO drives anoikis or vice versa. In the orthotopic intra-bursal model, RHOV KO resulted in smaller primary tumors and a complete absence of omental and peritoneal metastases. While the reduced primary tumor size may partially reflect impaired proliferation (even though this was not observed in attached cultures) or reduced local invasion into adjacent stromal tissue, the absence of secondary lesions strongly supports a failure of RHOV KO cells to disseminate. However, to better test this, a future approach would be to induce RHOV KO after primary tumor establishment.

Mechanistically, RHOV coordinates cytoskeletal remodeling and downstream signaling during metastatic adaptation. RHOV engages the PAK-c-Jun signaling axis, driving transcriptional programs that support motility and cytoskeletal machinery. While prior studies have implicated RHOV in cytoskeletal regulation and filopodia formation in breast cancer cells (25), our findings represent the first demonstration of this pathway being functionally engaged in ovarian cancer and in response to detachment. RHOV-mediated actin cytoskeletal remodeling likely exerts broader functional consequences on cellular structural integrity, locomotion, and contractility, all of which are essential for cell survival following detachment, MCA formation, and invasion. For instance, alterations in actin architecture can directly influence the localization, clustering, and signaling efficacy of adhesion molecules, thereby modulating their capacity to transduce essential survival signals and reinstate cell-cell adhesion in a-i. RNA-sequencing analysis revealed significant disruption of cell-cell adhesion and junction-related transcriptional pathways in RHOV-deficient cells specifically under anchorage-independent conditions. This finding correlates strongly with the observed phenotypes, including looser spheroid formation, impaired aggregation, reduced tumor-mesothelial cell adhesion, and reduced tumor cell binding efficiency to collagen-rich mesothelial substrates. These observations resonate with previous studies implicating RHOV in the regulation of E-cadherin dynamics during zebrafish epiboly (21), underscoring a conserved role for RHOV in mediating adhesion-related signaling events. RNA-seq additionally identified multiple integrin subunits as downregulated in RHOV-KO MCAs. While these changes didn’t translate into altered FAK or SRC signaling, whether integrin-mediated engagement with the extracellular matrix is functionally altered in this context remains an important question for future investigation. Our data further suggest that RHOV contributes to endosomal homeostasis, as it localizes to early endosomes and its loss leads to an accumulation of EEA1-positive vesicles. These trafficking defects, likely driven by altered actin dynamics, appear to impair receptor-mediated signaling, exemplified by the strong repression of TGF-β-mediated SMAD2/3 phosphorylation in RHOV-KO cells as well as repression of downstream target genes. Together, these findings identify RHOV as a key modulator of TGFBR signaling and raise the possibility that it more broadly regulates endosomal trafficking and receptor responsiveness. Follow-up studies are needed to define how RHOV governs endosomal dynamics and to determine which additional receptors may be affected.

We find that WT-RHOV can rescue signaling and metastatic functions in RHOV-KO cells similar to the constitutively active G40V mutant, while the ΔMEM mutant acts similar to the dominant negative rescue construct. These findings indicate that RHOV’s GTP-binding and membrane localization are essential for its function. Unlike other Rho GTPases, RHOV possesses a distinct, 16-residue carboxyl-terminal extension ending with a palmitoylation motif (CFV). Targeting this unique palmitoylation site or the adjacent C-terminal sequence could therefore disrupt RHOV membrane localization and selectively impair its oncogenic function. Future studies exploring this mode of RHOV inhibition may yield valuable insights into potential future therapeutic targeting of RHOV in ovarian cancer.

Collectively, this work sheds new light on the function of RHOV, an understudied and atypical member of the Rho GTPase family, and positions it as a critical mediator of metastatic fitness in ovarian cancer. While multiple GTPases have been implicated at different stages of cancer progression in both adherent and anchorage-independent states (14), our study establishes RHOV as the first identified detachment-sensitive Rho GTPase. Our data place RHOV within the conceptual framework of immediate early response genes, which are classically described as the first transcriptional wave activated by external stimuli to initiate downstream programs that drive long-term cellular outcomes (66,69). Here, we show that a similar hierarchical response architecture operates in the context of cell detachment, in which RHOV functions as an immediate early resp onder that subsequently initiates pro-metastatic transcriptional programs that persist beyond the window of peak RHOV expression. In this model, rapid transcription of RHOV, which is constitutively active once expressed, primes ovarian cancer cells for sustained survival, invasion, and peritoneal colonization. Consistent with this, inhibition of this early response, exemplified by RHOV loss, leads to a profound suppression of long-term metastatic success. These findings carry important therapeutic implications, as they suggest that targeting the earliest immediate responses to detachment may represent novel anti-metastatic strategies.

Statement of Significance:

Identification of a detachment-sensitive transcriptional signature in ovarian cancer uncovers RHOV as an essential mediator of transcoelomic metastasis, establishing these early adaptations as crucial regulators of peritoneal metastasis and potential therapeutic targets.