DUSP4 inhibited tumor cell proliferation by downregulating glycolysis via p-ERK/p-PGK1 signaling in ovarian cancer

Abstract

Ovarian cancer (OC) remains a leading cause of gynecological cancer-related mortality, with poor prognosis and limited therapeutic options, underscoring the urgent need for a deeper understanding of OC biology. In this study, we identified a marked reduction in dual-specificity phosphatase 4 (DUSP4) expression in OC tissues compared to benign ovarian masses, with even further decreases observed in metastatic lesions. Moreover, DUSP4 expression varied among OC subtypes, with the lowest levels observed in serous ovarian cancer, and was associated with P53 and KI67 protein levels, altered TP53 mutation rates, advanced tumor stages, and poorer prognosis. Functional experiments demonstrated that DUSP4 overexpression suppressed OC cell proliferation, migration, and invasion in vitro. Phosphoproteomic profiling via LC-MS/MS analysis identified the MAPK pathway and cellular metabolism as key downstream targets of DUSP4. Notably, DUSP4 overexpression reduced phosphorylation of PGK1 at Ser203, a critical regulator of anaerobic glycolysis, and decreased its mitochondrial localization, leading to reduced lactate production and increased ROS levels. Mechanistically, DUSP4 dephosphorylated p-ERK, disrupting its interaction with PGK1 and subsequently reducing PGK1 S203 phosphorylation. In vivo, DUSP4 overexpression significantly inhibited tumor growth in mouse models, accompanied by decreased p-ERK and PGK1 S203 levels. These findings highlight a regulatory axis involving DUSP4, p-ERK, and PGK1, through which DUSP4 modulates glycolysis and tumor progression. This study establishes DUSP4 as a prognostic biomarker and a potential therapeutic target for OC, offering new insights into its role in tumor metabolism and growth.

Introduction

Ovarian cancer (OC) is one of the most common malignancies among women and ranks second in mortality among gynecological cancers [1,2,3]. The disease is often diagnosed at advanced stages due to the lack of effective screening methods and subtle or absent early symptoms, resulting in approximately 70% of patients presenting with metastasis and ascites at the time of diagnosis [4, 5]. While advancements such as poly-(ADP-ribose) polymerase (PARP) inhibitors have shown efficacy in extending survival among the small subset of OC patients with homologous recombination repair gene mutations, only 14-18% of patients significantly benefit from these treatments [6]. For most patients, relapse remains common following either conventional therapies or targeted approaches like bevacizumab, leaving the five-year survival rate for advanced-stage OC persistently below 40% [4]. These challenges underscore the urgent need for further research into the molecular mechanisms driving OC progression and for the identification of novel therapeutic targets to improve patient outcomes.

The evolutionarily conserved mitogen-activated protein kinase (MAPK) pathways are central regulators of cellular processes, including proliferation, differentiation, metabolism, motility, and apoptosis. Dysregulation of these pathways contributes to the development of various pathologies, including malignancies [7,8,9,10]. Dual-specificity phosphatases (DUSPs) are a group of enzymes that act as major negative regulators of MAPKs by specifically dephosphorylating threonine or tyrosine residues. The DUSP family comprises 11 typical members, categorized into three subfamilies based on subcellular localization and substrate specificity. Nuclear DUSPs, including DUSP1, DUSP2, DUSP4, and DUSP5, are primarily induced by growth factors or stress-related signals, while cytoplasmic DUSPs, such as DUSP6, DUSP7, and DUSP9, preferentially modulate the extracellular signal-regulated kinase (ERK1/ERK2) pathway. The third group, including DUSP8, DUSP10, and DUSP16, primarily targets the c-Jun N-terminal kinase (JNK) and p38 MAPK pathways [11]. Increasing evidence indicates that DUSPs are actively involved in tumor development and progression [12,13,14,15], with their expression often linked to treatment responses [16,17,18,19]. For instance, DUSP6 has been identified as a key driver of myeloproliferative neoplasms, with its aberrant overexpression conferring resistance to JAK2 inhibitors in secondary acute myeloid leukemia [18].

Dual-specificity phosphatase 4 (DUSP4), a member of the nuclear DUSP subfamily, plays a critical role in regulating cell behavior through the inactivation of the ERK1/2, JNK, and p38 pathways [11, 20,21,22,23]. However, the role of DUSP4 in cancer progression remains highly context-dependent and varies across tumor types. In some cases, DUSP4 has been associated with tumor-promoting properties, such as its involvement in colorectal cancer progression and resistance to trastuzumab in HER2-positive breast cancer [24]. Conversely, in certain cancers, including pancreatic cancer and basal-like breast cancer, DUSP4 is downregulated, leading to enhanced tumor invasiveness, resistance to chemotherapy or anoikis, and poorer patient outcomes due to unchecked MAPK activation [25,26,27,28]. In the context of OC, previous studies have reported significantly lower DUSP4 expression in serous ovarian cancer tissues compared to borderline ovarian tumors. This suggests that DUSP4 may act as a key regulator of less aggressive tumor behavior in OC [29]. Moreover, DUSP4 knockdown has been shown to increase spheroid formation in OC cells, further supporting its potential role as a tumor suppressor [30]. Despite these observations, the precise mechanisms through which DUSP4 regulates OC progression remain incompletely understood.

Herein, this study reveals a strong association between reduced DUSP4 expression and poor prognosis in OC patients. Phosphoproteomic profiling showed that DUSP4 regulates cellular metabolism and signaling by dephosphorylating p-ERK, which in turn reduces PGK1 Ser203 phosphorylation. This action suppresses anaerobic glycolysis while increasing ROS levels, leading to redox imbalance and oxidative stress in tumor cells. These findings underscore a critical connection between DUSP4, mitochondrial metabolism, and oxidative stress, providing new insights into OC progression and positioning DUSP4 as a promising therapeutic target and prognostic biomarker.

Materials and methods

Patient samples

This study included a total of 116 cases of epithelial OC in situ tumor tissues, 71 cases of metastatic OC tissues, and 62 cases of benign ovarian masses (mainly benign ovarian cysts). The pathological types of all tumor tissues were checked by experienced pathologists. All samples were collected when patients underwent the first surgery without neoadjuvant chemotherapy. All participants were informed of the purpose of this study and signed an informed consent form. This study was approved by the Ethics Committee of Xinhua Hospital, affiliated with Shanghai Jiao Tong University School of Medicine.

Gene expression and survival analysis of TCGA and GEO datasets

The extra datasets with gene expression and prognostic information were obtained from the TCGA data portal and the GEO website (including GSE14764, GSE15622, GSE18520, GSE19829, GSE23554, GSE26193, GSE26712, GSE27651, GSE30161, GSE3149, GSE51373, GSE63885, GSE65986, and GSE9891). The expression of DUSP4 in OC tissues were analyzed with proper standardization and filtering. For survival analysis, the most significant cut-off value of DUSP4 expression was selected as the best cutoff to group patients with high expression and low expression. The Survival package implemented with the Cox regression was used to analyze the progression-free survival rate and the overall survival rate, and to generate the final Kaplan-Meier curves. The p values were calculated using log-rank test. In addition, using TCGA data, a total of 288 OC samples with mutation data were included for mutation analysis. To evaluate the differences in mutation frequency between sample groups, a chi-square test (χ² test) was performed for all detected genes.

Cell culture and reagents

All OC cell lines used in this study were obtained from previous members of our research team with genetic authentication by short tandem repeat analysis (STR). All cell lines were ensured to be free from mycoplasma contamination when used for experiments. SKOV3 and its DUSP4-OE cell line were cultured in RPMI-1640 medium (Gibco, 11875119) supplemented with 10% fetal bovine serum (FBS; Gibco, 10091148) and 1% penicillin-streptomycin-neomycin (PSN) antibiotic mixture (Gibco, 15640055). Hey A8 and its DUSP4-OE cell line were cultured in DMEM medium (Gibco, 11965092) supplemented with 10% FBS and 1% PSN.

Immunohistochemistry assay (IHC)

Paraffin-embedded tissues were cut into 4-µm-thick sections, and the slides were first heated, deparaffinized and rehydrated. Then, the slides were incubated in 10 mM citrate buffer (pH 6.0) at boiling temperature for 20 min for antigen retrieval, and the slides were incubated in a 3% hydrogen peroxide solution for 10 min to block the activity of endogenous peroxidases. Next, after the antigens on the slides were blocked with 5% BSA in PBS for 30 min, the slides were incubated with the following primary antibody at 4 °C overnight: anti-DUSP4 (Abcam, ab216576), anti-PGK1 (Proteintech, 17811-1-AP), anti-PGK1 (Phospho-Ser203) (Signalway Antibody, SAB487P), anti-ERK1/2 (Abcam, ab184699), and anti-ERK1/2 (Phospho-Thr202/Tyr204) (Proteintech, 28733-1-AP). Subsequently, the slides were incubated in HRP-labeled streptavidin-conjugated secondary antibody (Servicebio, G3431) at room temperature for 45 min and the color was developed using freshly prepared DAB solution (Servicebio, G1212). Finally, the slides were counterstained with hematoxylin (Servicebio, G1004), dehydrated, made transparent and sealed with resin.

The images of stained sections were scanned using the Pannoramic MIDI slide scanner (3D HISTECH, Japan), and the Quant Center was used to automatically read all the measured areas, and calculate the total scores of each image by analyzing the percentage of weak (light yellow, scoring 1), moderate (brown-yellow, scoring 2), and strong (brown, scoring 3) positive areas within the measured regions. In this study, the immunohistochemical score (H-Score) was used for quantitative analysis of the staining intensity. The formula is H-Score = ∑(pi × i) = (percentage of weakly stained area × 1) + (percentage of moderately stained area × 2) + (percentage of strongly stained area × 3), where pi represents the proportion of positive staining area, and i represents the staining intensity. The higher H-Score scores indicating stronger overall positivity.

Lentiviral transduction

To establish stable DUSP4 overexpression cell lines, the OC cells were transfected with the overexpression lentivirus of DUSP4 and an empty vector purchased from Shanghai Jiman Biotechnology (China). The day prior to transfection, cells were seeded in a 6-well plate to ensure that the cell density is around 50-60% the next day. Then, on the second day, the old culture mediums were replaced by fresh complete medium containing 6 µg/ml polybrene (Jiman Biotechnology, GM-040901). Virus solution was added to each well based on the multiplicity of infection (MOI), ranging from 6 to 20 µl per well. After 24 h of infection, the mediums were replaced with fresh complete medium. Then, after an additional 48 h of culture, 4–12 µg/ml puromycin was added to select cells effectively transfected by the virus for 2 weeks.

Real-time quantitative PCR assay

Trizol was used to extract total RNA from the collected cells or tissue samples and a reverse transcription reagent kit (Takara; RR036A) was used to reverse transcription RNA into cDNA. Then, a SYBR Green kit (Yeasen, 11202ES08) was used for real-time quantitative PCR amplification. The mRNA levels of target genes were normalized to the mRNA levels of the TUBLIN gene and were calculated using the 2^(-ΔΔCt) method. The paired primers used in this study are listed as follows:

DUSP4-Forward Primer GGCGGCTATGAGAGGTTTTCC.

DUSP4-Reverse Primer TGGTCGTGTAGTGGGGTCC.

TUBLIN-Forward Primer TCGATATTGAGCGTCCAACCT.

TUBLIN-Reverse Primer CAAAGGCACGTTTGGCATACA.

Western blot (WB) and co-immunoprecipitation assay (Co-IP)

200 µl of RIPA lysis buffer containing PMSF (Solarbio, R0020), protease inhibitors (Beyotime, P1006) and phosphatase inhibitors (Merk- Roche, 4906845001) were used to lyse cells. The total cell protein concentrations were detected using the BCA kit (Thermo Fisher Scientific, 23227) according to the instructions. Then, reduced SDS-PAGE sample loading buffers (Beyotime, P0286) were added to the protein solution, and the proteins were boiled at 99 °C for 5 min for fully denaturation.

For western blot, 10% SDS-PAGE gels (Epizyme Biotech, PG112) with 1.5 mm 10-hole combs were prepared using the quick preparation kit according to the instructions. The electrophoresis buffer (25 mM Tris, 250 mM glycine, 0.1% SDS, pH 8.0-8.6) and transfer buffer (20 mM Tris, 192 mM glycine, pH 8.2–8.6, pre-chilled) were also prepared pre-experiment. Then, the denatured protein samples were loaded into the sample wells, 20 µg protein in 15 µl per well. Pre-stained protein markers were loaded at the appropriate well. The voltage was maintained at 80 volts during electrophoresis until the protein runs out of the upper layer. The voltage was then adjusted to 120 volts for electrophoresis until the bottom protein is dispersed close to the bottom of the lower gel. When the electrophoresis was finished, the proteins inside the PAGE gel were transferred into the PVDF membranes (Milipore, IPVH00010) pre-activated in methanol using an ice bath at 300 mA for 70 min. After the transfer is completed, the PVDF membranes were incubated in protein blocking solution at room temperature for 15 min to remove non-specific binding. Then, the PVDF membranes were incubated using the primary antibody solution on a shaker at 4 °C overnight. The next day, the primary antibody was removed, and the PVDF membranes were washed three times with TBST, and then incubated in the HRP-conjugated secondary antibody (Beyotime, A0208 and A0216) solution at room temperature for 1 h. Finally, the PVDF membranes were washed with TBST, and the protein strips were exposed using ultra-sensitive ECL chemiluminescence solution (Beyotime, P0018FM) with the chemiluminescence instrument.

For co-immunoprecipitation assay, the methods for obtaining the total protein lysates and detecting the protein concentrations were the same as above. Firstly, the protein concentration was adjusted to 1 mg/ml. For 1 mg protein sample, 1 µg rabbit normal IgG (Cell Signaling Technology, 2729 S) and 20 µl fully resuspended Protein A + G Agarose (Beyotime, P2012) were added, and the samples were incubated at 4 °C for 2 h to remove non-specific binding proteins. After centrifugation, the supernatant samples without non-specific binding proteins were collected for subsequent co-immunoprecipitation. These samples were divided into three parts according to the ratio of 1:2:2. SDS-PAGE sample loading buffer was added to the first portion (input), and then it was denatured at 99 °C for 5 min. 2 µg normal IgG antibody or target primary antibody were added to the remaining two portions separately, and the samples were incubated at 4 °C overnight. On the next day, 40 µl fully resuspended Protein A + G Agarose were added to each protein sample, which were then incubated at 4 °C for 3 h. After incubation, the supernatant was removed, and the agarose beads were washed with PBS for 5 times. Then, 30 µl 1X SDS-PAGE sample loading buffer was added to resuspend the pellet, and the protein samples were denatured at 99 °C for 5 min for western blot.

Cell proliferation assay using cell count kit (CCK-8)

Cells in fresh complete culture medium were seeded to the labeled 96-well plate, 2000–5000 cells in each well, two duplicate wells were set. The cells were cultured for 5 consecutive days with detection of the number of live cells. CCK8 kit (TargetMol, C0005) was used for cell viability detection according to the instructions. The absorbance value at 450 nm was detected using a microplate reader. Meantime, blank wells without cells were set during detection. After 5 days of continuous detection, a curve was draw based on the corrected absorbance values (OD value of the detection hole - OD value of the blank hole), and statistical analysis was performed. Doubling time was calculated using the formula DT = t⋅ln(2)/ln(N_t​/N_0​)​, where N₀​ and N_t​ represent the cell number at the initial (day1) and final time points (day3), respectively.

Migration and invasion assay

First, 750 ml of complete culture medium were added to each well of the lower chamber of the 8-micron well transwell plate (Corning, 3422), so that the liquid level exceeded the microporous membrane at the bottom of the upper chamber but did not enter the upper chamber. Then, cells resuspended in culture medium without FBS were added into the upper chamber (2 × 10⁴ cells in 200 µl of FBS-free culture medium). Two duplicate wells were set for each group. After cultured in the well plate for 8–12 h, cells in the outer layer of the upper chamber were washed gently with PBS for twice, fixed with 4% paraformaldehyde, and stained with crystal violet. The cells in the upper layer of the chamber were gently wiped with a cotton swab to remove the interference of unpenetrated cells, and they were left to dry naturally. Finally, the transwell chambers were observed and photographed under a microscope, and the cells that have migrated into the lower layer were counted and statistically analyzed.

For invasion assay, the matrigel diluted with FBS-free culture medium were added evenly on the bottom of the upper chamber. 50 µl diluted matrigel (Corning, 356237) for each chamber to cover the bottom, and then, the gels were solidified in a 37 °C incubator for 4–6 h. Be careful to prevent the formation of bubbles. The remaining steps are the same as the migration assay shown above.

Colony formation assay

Cells were seeded in a 6-well plate with 2 ml of complete culture medium, 300–500 cells per well. The cells were cultured continuously for 7–14 days. During this period, the cell growth status and cell confluence should be observed, and the culture was terminated at the appropriate time. Then, the cells in the 6-well plate were washed twice with PBS, fixed with 4% paraformaldehyde, and stained with crystal violet. After staining, they were left to dry naturally for 30 min. Finally, the well plate was photographed, and the cell clones were counted under a microscope.

Phosphoproteomic detection by mass spectrometry and data analysis

The cells of the DUSP4-OE group and the vector group were collected, washed with PBS, and then fully lysed with SDT buffer (4% SDS, 100mM Tris-HCl, 1mM DTT, pH7.6). The protein was quantified using BCA protein assay kit. Protein digestion by trypsin was performed according to filter-aided sample preparation (FASP) procedure described by Matthias Mann. Then, the peptides were desalted using a C18 column (Empore™ SPE Column C18 (standard density), bed I.D. 7 mm, volume 3 ml, Sigma), concentrated by vacuum centrifugation and reconstituted in 40 µl of 0.1% (v/v) formic acid. Next, 100 µg peptide mixture were taken from each sample for labeling using TMT reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. Then, High-SelectTM Fe-NTA Phosphopeptides Enrichment Kit (Thermo Scientific) was used to enrich the phosphorylated peptides according to the manufacturer’s instructions.

Liquid chromatography-tandem mass spectrometry (LC-MS/MS) was performed on a Q Exactive HF-X mass spectrometer (Thermo Scientific) that was coupled to Easy nLC (Thermo Fisher Scientific) for 120 min. The peptides were loaded onto a reverse phase trap column (Thermo Scientific Acclaim PepMap100, 100 μm*2 cm, nanoViper C18) connected to the C18-reversed phase analytical column (Thermo Scientific Easy Column, 10 cm long, 75 μm inner diameter, 3 μm resin) in buffer A (0.1% Formic acid) and separated with a linear gradient of buffer B (84% acetonitrile and 0.1% Formic acid) at a flow rate of 300 nl/min controlled by IntelliFlow technology. The mass spectrometer was operated in positive ion mode. MS data was acquired using a data-dependent top10 method dynamically choosing the most abundant precursor ions from the survey scan (300–1800 m/z) for HCD fragmentation. Automatic gain control (AGC) target was set to 3e6, and maximum inject time to 10 ms. Dynamic exclusion duration was 40 s. Survey scans were acquired at a resolution of 70,000 at m/z 200 and resolution for HCD spectra was set to 17,500 at m/z 200, and isolation width was 2 m/z. Normalized collision energy was 30 eV and the underfill ratio was defined as 0.1%. The instrument was run with peptide recognition mode enabled. Finally, the MS/MS spectra were searched using MASCOT engine (Matrix Science, London, UK; version 2.2) embedded into Proteome Discoverer 2.4. The specific parameters involved are as follows: trypsin was used for peptide digestion, the maximum number of missed cleavage sites allowed was 2, the primary ion mass tolerance was + 20ppm, the secondary ion mass tolerance was 0.1Da, the fixed modification was Carbamidomethyl(C)TMT6/10/16plex(N-term), TMT6/10/16plex (K), the variable modification was Oxidation(M), Phospho(ST), Phospho(Y), the protein sequence database used for searching was Swissprot_Homo_sapiens_20395_20210106.fasta, the database mode used to calculate FDR was Decoy, and the screening criteria for credible proteins was ≤ 0.01.

Bioinformatic analysis were performed later, including hierarchical clustering of the phosphorylated peptides using Cluster 3.0 and Java Treeview software, motif analysis by MeMe (http://meme-suite.org/index.htm), protein subcellular localization predicted by CELLO (http://cello.life.nctu.edu.tw/), domain annotation using the InterProScan software, and pathway enrichment by gene ontology (GO) terms and the online Kyoto Encyclopedia of Genes and Genomes (KEGG) database. For enrichment analysis, the Fisher’ exact test was applied, with the whole quantified proteins as background dataset. Benjamini- Hochberg correction for multiple testing was further used to adjust derived p-values. Functional categories and pathways with p-values < 0.05 were considered as significant.

Immunofluorescence staining assay

The cells were seeded in glass slide put inside the 6-well plate one day before experiment, to ensure that the cells were attached to the slide and the cell density the next day would be around 50%. On the second day, the old culture medium was removed and 100nM Mito Tracker^® Green FM mitochondrial probe solution (Yeasen, 40742ES) diluted in the complete culture medium was added into the cell plate. After incubated at 37 °C for 45 min, the cells were washed with PBS, fixed in 4% paraformaldehyde at room temperature for 30 min in the dark. After washed with PBS, the cells were incubated in quick immunostaining blocking solution containing Triton X-100 at room temperature for 15 min in the dark. Next, the cells were incubated with the primary antibody dilution (PGK1 s203) at 4 °C overnight in the dark. On the next day, after washed with PBS, the cells were incubated with Cy3-labeled donkey anti-rabbit IgG antibody (Servicebio, GB21403) at room temperature for 1 h in the dark. Then, the cells were washed with PBS, and DAPI-containing antifade agent (Beyotime, P0131) was used for nuclear staining and slides mounting.

Reactive oxygen species (ROS) detection

The cells of experimental and control group were seeded on glass slides placed inside a 6-well plate one day prior to the experiment to ensure cell attachment and a confluency of approximately 50–70% on the following day. Then, the cells were washed with PBS, and the DCFH-DA probe solution (Yeasen, 50101ES01) was prepared by diluting the probe 1:1000 in serum-free medium to a final concentration of 10 µM. The prepared DCFH-DA solution was added to the cells, ensuring the slides were fully covered. The cells were incubated at 37 °C in the dark for 15 min. Following incubation, the cells were washed 1–2 times with PBS to remove excess probe that had not entered the cells. Cells were harvested after probe loading and ROS levels were measured using flow cytometry (Beckman) with the FITC channel. The fluorescence intensities of ROS were analyzed using GeoMean (geometric mean) and Mode (peak fluorescence intensity) statistics in FlowJo (v10.6.2).

Lactate detection

Cells were seeded in the 12-well plate the day before experiment, and were cultured with the same amount of fresh medium overnight. The next day, the cell supernatants and cell pellets were collected respectively. The cell supernatants were diluted to a certain proportion and directly used for detection. The cells were lysed using 200 µl RIPA lysis buffer containing protease inhibitors on ice for 30 min. After centrifugation, the supernatant protein samples were diluted and used directly for detection. The cellular protein concentrations were detected using BCA kit.

Next, the enzyme working solution and chromogenic reagent (Nanjing Jiancheng Bioengineering Institute, A019-2-1) were prepared according to the instructions. The enzyme working solution needs to be used immediately after prepared. The chromogenic reagent is valid within 2 weeks when stored at 2–8℃ in the dark. 20 µl distilled water, 3 mmol/L standard sample and the sample to be tested were added to the blank tubes, standard tubes, and measurement tubes respectively, with 2 duplicate holes for each group. Then, 1 ml enzyme working solution and 0.2 ml chromogenic reagent were added to each tube. After reacted in a 37 °C water bath for 10 min, 2 ml stop reagent was added to each tube, and the absorbance value (A) of each tube was detected at a wavelength of 530 nm. The calculation formula for lactic acid in cell supernatant is: lactic acid content (mmol/L) = (A measurement - A blank) / (A standard - A blank) × C standard × N; and the calculation formula for lactic acid in cells is: lactic acid content (mmol/g protein) = (A measurement - A blank) / (A standard - A blank) × C standard ÷ Cpr, where C standard is the standard concentration of 3 mmol/L, and N is the dilution factor when the sample is tested, Cpr is the cellular protein concentration.

In situ tumor cell line transplantation mouse model

The BALB/c nude mice used in this study were provided by Shanghai Jihui Company and kept in Xinhua Hospital Affiliated to Shanghai Jiaotong University School of Medicine. All the mice used were female, 4–5 weeks old, and weighed between 15 and 20 g. The animal experiment protocol involved in this study was reviewed and approved by the Ethics Committee of the Experimental Center of Xinhua Hospital.

After anesthetized, the mouse was put in left lateral position, and was cut a small opening in the skin and peritoneum. Then, the ovary under the peritoneum was exposed, and 5 × 10⁵ tumor cells of the DUSP4-OE group or the vector group were directly injected into the left ovarian tissue of the mouse. After injection, the ovary was gently placed into the abdominal cavity, and the peritoneum and skin were sutured in turn. The growth of tumor was observed every week with in vivo imaging of the mouse. The tumor tissues were harvested until 7–8 weeks after transplantation, and the weigh and volume of the tumors in situ were measured and analyzed.

Statistical analysis

All experiments were repeated three times. Data were analyzed and visualized using RStudio (R version 4.3.0), SPSS (version 20.0) and GraphPad Prism (version 9.0), and were shown as the mean ± s.d. in the plots. Statistical significance was assessed by two-sided t-test, log-rank test, Fisher’s exact test, Kruskal-Wallis test, and chi-square test as described in the figure legends. The p value < 0.05 was considered significant. In figures, ns means no significant difference, * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001.

Results

Decreased expression of DUSP4 in OC tissues is related to poor prognosis

Utilizing customized ovarian tumor/mass tissue microassay for immunohistochemical detection, we examined the protein expression of DUSP4 in benign ovarian masses and OC tissues. As shown through H-score, DUSP4 was significantly downregulated in primary OC tissues (116 cases) compared to benign ovarian mass (62 cases) (Fig. 1A-B). In addition, we found that the expression of DUSP4 in metastatic tumor tissues (71 cases) was lower than that in primary tumor tissues, although it was not statistically significant (Fig. 1A-B). To further compare the expression of DUSP4 in benign ovarian masses and OC tissues, we took the median expression level of DUSP4 in all tissues as the cutoff value, and divided the benign and OC tissue samples into two subgroups (DUSP4^high and DUSP4^low). The results showed that in patients with benign tissues, the patient number of DUSP4^high group accounted for 90%, while this proportion was 67% in OC patients (Fig. 1C), suggesting that OC patients tend to express lower levels of DUSP4. Then, we analyzed the correlation between DUSP4 expression and clinical parameters, including age, tumor stage, histological type, and other relevant factors, using our own dataset. The results showed that DUSP4 expression levels vary across different pathological types of ovarian cancer, with significantly lower expression observed in serous ovarian cancer (Table 1; Fig. 1D). In addition, DUSP4 expression levels were correlated with the intensity of P53 and KI67 protein expression, as demonstrated by immunohistochemical staining (Table 1). Analysis of TCGA data also revealed significant differences in the mutation status of TP53 between OC samples stratified by high and low DUSP4 expression levels (Fig. S1). Furthermore, our results indicated that lower DUSP4 expression is significantly associated with advanced tumor stages (Fig. 1E), which was further confirmed using publicly available data from the TCGA OV datasets (Fig. 1F).

Fig. 1

Low DUSP4 expression in OC tissues is linked to poor prognosis. (A) Representative images of DUSP4 expression levels in human tumor tissue samples detected by immunohistochemical assay. The scale bar is 50 μm. (B) Statistical analysis of DUSP4 expression levels in benign cyst (n = 62), primary tumor (n = 116), and metastatic tumor samples (n = 71). P-value is determined by two-sided t test; the data is shown as mean ± SD; ns means no significant difference, and ***p indicates statistical difference. (C) Statistical analysis of the proportion of high and low DUSP4 expression levels in benign cyst (n = 62), and primary tumor samples (n = 116). (D) Expression levels of DUSP4 in different pathological types of ovarian cancer. Statistical analysis was performed using the Kruskal-Wallis test. *p indicates statistical difference. (E) Expression levels of DUSP4 in OC across different stages. Statistical analysis was performed using the Kruskal-Wallis test based on our dataset. ***p indicates statistical difference. (F) DUSP4 mRNA expression across different stages of OC using data from TCGA OV database. (G) Overall survival analysis of OC patients with high and low DUSP4 expression levels in tumor tissues using Kaplan-Meier curves. The level of significance and p-value is determined by log-rank test. (H-I) Survival analysis of OC patients with high and low DUSP4 expression levels in tumor tissues using data from TCGA and GEO database, including overall survival (H) and progression free survival rate (I) shown by Kaplan-Meier curves

Table 1 Correlations between DUSP4 expression and clinical parameters in 116 OC patients

Moreover, we collected the follow-up information of all patients in the tissue microassay, and drew a Kaplan-Meier survival curve based on the overall survival period. Patients with high DUSP4 expression levels had a significantly higher overall survival (OS) than those with low DUSP4 expression levels, with a p value of 0.0241 (Fig. 1G). Due to the limited number of microassay samples, we also analyzed the correlation between mRNA expression of DUSP4 in OC tissues and patient prognosis using the TCGA and GEO data. Similar to the analysis from our microassay data, OS was significantly higher in patients with high DUSP4 expressions (OS 48.37 vs. 38.93 months) (Fig. 1H). In addition, progression-free survival (PFS) was also significantly higher in patients with high DUSP4 expressions (PFS 21 vs. 18 months) (Fig. 1I). These results showed that OC patients with higher DUSP4 expressions tend to possess a better prognosis.

DUSP4 suppresses proliferation, and motility of OC cells

To select appropriate OC cell lines to clarify the role of DUSP4, we detected the basic expression level of DUSP4 protein in five cell lines. The basic expression of DUSP4 in HEY was the highest, and that in HEY A8 and SKOV3 were the lowest (Fig. 2A-B). In addition to protein expressions, we also used RT-qPCR to detect the mRNA levels of DUSP4 in the five cell lines. Consistent with the protein expression levels, HEY A8 and SKOV3 harbor the lowest basic expression of DUSP4 mRNA (Fig. 2C). Here, we chose HEY A8 and SKOV3 cell lines for further experiments to explore the regulatory function of DUSP4 on malignant behaviors of OC cells, such as proliferation and migration ability.

Fig. 2

DUSP4 suppresses proliferation, migration, and invasion of OC cells. (A-B) Background DUSP4 expression levels in five OC cell lines detected by WB. And quantitative analysis of DUSP4 expressions by calculating the relative gray values of protein bands. (C) RT-qPCR detection of relative DUSP4 mRNA levels in OC cell lines. (D-E) DUSP4 expression levels in DUSP4-overexpression and vector-loaded OC cell lines detected by WB. And the quantitative analysis using the relative gray values of protein bands. (F) RT-qPCR detection of relative DUSP4 mRNA levels in DUSP4-overexpression and vector-loaded OC cell lines. (G-H) Growth curves of DUSP4-overexpression and vector-loaded OC cell lines detected by CCK8 assay. The doubling time for each cell line are as follows: HEY A8-Control 23.67 h, HEY A8-DUSP4-OE 29.41 h; SKOV3-Control 27.27 h, SKOV3-DUSP4-OE 30.85 h. (I-J) Representative images of migration and invasion assay using DUSP4-overexpression and vector-loaded OC cell lines. The scale bar is 200 μm. And statistical analysis of migration and invasion abilities of DUSP4-overexpression and vector-loaded OC cell lines. Experiments above in this figure were all replicated for three times. Significance is determined by two-sided t test; *p indicates statistical difference; the data are presented as mean ± SD

Firstly, for HEY A8 and SKOV3, we successfully constructed DUSP4-OE cell lines and the control ones using DUSP4 plasmid and an empty vector coated with virus. As shown, the DUSP4 protein levels of the two DUSP4-OE cell lines (HEY A8-DUSP4-OE and SKOV3-DUSP4-OE) were significantly higher than those of the control group (Fig. 2D-E). At the same time, the DUSP4 mRNA level of the DUSP4-OE cells were also higher (Fig. 2F). Next, in HEY A8 and SKOV3 cell line, the CCK8 proliferation assay both showed that overexpression of DUSP4 significantly decreased the cell viability of OC cells (Fig. 2G-H). In addition, compared with the control group, clones formed by the DUSP4-OE cells were smaller determined by the cell counting result (Fig. S2A-B), which further indicated that overexpression of DUSP4 inhibited the proliferation ability of OC cells in vitro.

Previously, the immunohistochemistry result of the ovarian tissue microassay revealed that DUSP4 expression levels were obviously lower in metastatic OC tissues compared to primary OC tissues. This finding indicated that DUSP4 may play a regulatory role in invasive behaviors of tumor cells. Then, transwell assays were conducted and the number of cells in the DUSP4-OE group penetrating the chamber membrane and the matrix gel membrane was significantly lower than in the control group (Fig. 2I-J). This result suggested the overexpression of DUSP4 could inhibit the migration and invasion abilities of OC cells.

Protein phosphorylation network in OC cells regulated by DUSP4

Given that DUSP4 is a dual-specific phosphatase, it mainly functions in cells by regulating the phosphorylation level of downstream proteins [7, 31]. Therefore, to comprehensively investigate the impact of DUSP4 on downstream proteins and signal pathways in OC cells, phosphoproteomic detection via LC-MS/MS was conducted on SKOV3-DUSP4-OE and SKOV3-DUSP4-Control cells. A total of 14,103 quantifiable phosphorylated peptides and 15,124 quantifiable phosphorylation sites on 4,069 proteins were identified and included in the analysis. Next, we conducted a preliminary screening of differentially expressed phosphorylated peptides between the two groups. The hierarchical clustering heatmap demonstrated that the differentially phosphorylated modified peptides could effectively distinguish the SKOV3-DUSP4-OE and SKOV3-DUSP4-Control samples (Fig. 3A). Detailly, the SKOV3-DUSP4-OE group exhibited a down-regulation of an average of 695 phosphorylation-modified peptides and an up-regulation of 285 modified peptides compared to the SKOV3-DUSP4-Control group (Fig. 3B).

Fig. 3

Protein phosphorylation profiling in OC cells regulated by DUSP4. (A) Clustering of differentially expressed phosphorylated peptides in DUSP4-overexpression and vector-loaded OC cell lines. The screening criteria were as follows: fold change (FC) up-regulation > 1.2 or down-regulation < 0.83, and a P-value < 0.05. Significance is determined by t test. (B) Histogram of quantitative differential expression results of phosphorylated peptides. (C) Pie chart of subcellular localization statistics of proteins belonging to differentially expressed phosphorylated modified peptides. (D) GO functional enrichment analysis of phosphorylated differentially expressed proteins. Significance is determined by Fisher’s exact test, p-value < 0.05. (E-F) KEGG enrichment pathway annotation of phosphorylation-modified differentially expressed proteins. Significance is determined by Fisher’s exact test, p-value < 0.05

Based on the protein assignments of the peptides, we conducted further investigation into the subcellular localizations of differentially modified proteins, including those present in the nucleus, cytoplasm, cell membrane, and other compartments. The results revealed that DUSP4 overexpression primarily regulates protein phosphorylation modification in the nucleus, followed by the cytoplasm (Fig. 3C). Notably, among the differentially modified peptides, 43 of them belong to proteins localized in the mitochondria (Fig. 3C). This finding suggested a potential regulatory effect of DUSP4 on cellular energy metabolism.

Considering the large number of proteins associated with the differentially modified peptides, GO functional annotation was conducted on all the differentially modified proteins to gain preliminary insights into the regulatory role of DUSP4 in OC cells. The analysis revealed significant changes in various biological processes, including cell metabolism, signaling, cell proliferation, and cell adhesion (Fig. 3D). Moreover, molecular functions such as molecular binding, catalytic activity, and transcriptional regulation were also affected (Fig. 3D). In addition, KEGG pathway enrichment was also performed, which helps to gain a more comprehensive understanding of the changes in various pathways. The results demonstrated that the differentially modified proteins are mainly enriched in MAPK signaling pathway, glycoproteins, focal adhesion and tight junction (Fig. 3E-F), partially consistent with previous reports [7].

DUSP4 alleviates the phosphorylation of PGK1 at Ser203 and reduces the lactate production in OC cells

Previous studies on DUSP4 mainly focused on the regulation of intracellular MAPK pathways [23, 32]. In our above phosphoproteomic data analysis, we observed that a significant proportion of the altered phosphorylated proteins were located in mitochondria (Fig. 3C). This finding prompted us to further explore the potential regulatory role of DUSP4 in cellular metabolic pathways, particularly in mitochondrial functions. Previous studies have highlighted the critical role of PGK1 phosphorylation in cellular metabolism. Notably, under hypoxic conditions or in the presence of specific genetic mutations, phosphorylated PGK1 at the S203 site (PGK1 S203) is known to translocate into mitochondria, where it phosphorylates PDHK1. This phosphorylation inhibits mitochondrial pyruvate metabolism and shifts cellular energy production toward anaerobic glycolysis, contributing to tumor adaptation and progression [33, 34]. Interestingly, among the altered phosphorylated proteins identified in our phosphoproteomic profiling, PGK1 S203 phosphorylation was significantly downregulated in DUSP4-OE cells (Fig. 4A), which was further verified via WB detection (Fig. 4B-C). Then, we also detected the expression level of PGK1 in DUSP4-OE and control cells. The results showed that overexpression of DUSP4 could cause a decrease in the level of PGK1 S203 protein without changes of PGK1 in OC cells (Fig. 4B-C). Further, we investigated the levels of mitochondrial PGK1 S203 in DUSP4-OE and control cells using an immunofluorescence assay with the MitoTraker staining. As depicted, compared to the control cells, the co-localization of PGK1 S203 and MitoTraker was less prominent in the DUSP4-OE cells (Fig. 4D). This observation suggested that DUSP4 overexpression not only reduced the abundance of PGK1 S203 in OC cells but also decreased its presence within mitochondria. DUSP4 overexpression may have a certain impact on cell metabolism by modulating the levels of PGK1 S203 in both the cytoplasm and mitochondria.

Fig. 4

DUSP4 alleviates the phosphorylation of PGK1 at Ser203 and reduces the lactate production in OC cells. (A) Differential expression of PGK1 S203 in DUSP4-overexpression and vector-loaded OC cell lines detected by LC-MS/MS. (B-C) Expression levels of PGK1 and PGK1 S203 in DUSP4-overexpression and vector-loaded OC cell lines detected by WB. And the quantitative analysis using the relative gray values of protein bands. (D) Co-localization expression of PGK1 S203 and MitoTraker in DUSP4-overexpression and vector-loaded OC cells using immunofluorescence. The scale bar is 50 μm. (E) Intracellular lactic acid content in DUSP4-overexpression and vector-loaded OC cells. (F) Lactic acid content in cell supernatant of DUSP4-overexpression and vector-loaded OC cells. (G) Histogram analysis of intracellular ROS levels measured by flow cytometry. (H) Geometric mean fluorescence intensity (GeoMean) of ROS levels. GeoMean values represent the overall ROS levels across the cell population for each experimental group. (I) Peak fluorescence intensity (Mode) of ROS levels. Mode values represent the most frequent fluorescence intensity within the population for each experimental group. For experiments above, the significance of difference is determined by two-sided t test; *p indicates statistical difference; the data are presented as mean ± SD

Next, to explore whether DUSP4 overexpression affects the anaerobic glycolysis process due to downregulation of PGK1 S203 in mitochondria, we conducted further analysis by measuring the lactate content in cells and the cell culture supernatant from both the DUSP4-OE and control groups. The results indicated that both the intracellular and extracellular lactate levels in the DUSP4-OE group were lower compared to the control group (Fig. 4E-F). Further, we investigated the effects of DUSP4 overexpression on mitochondrial metabolism and oxidative stress by analyzing ROS levels in both DUSP4-OE and control cells. The results showed that DUSP4-OE cells produced significantly higher levels of ROS compared to the control group (Fig. 4G-I), suggesting that DUSP4 may enhance mitochondrial-related aerobic metabolism, leading to disrupted redox balance and increased oxidative stress in tumor cells.

Phosphorylation of PGK1 at Ser203 is regulated by DUSP4/p-ERK signaling

The above results demonstrated that DUSP4 overexpression reduced the level of anaerobic glycolysis in cells by downregulating the content of PGK1 S203 in mitochondria. However, the molecular mechanism involved is still unclear. Previous studies have reported that phosphorylated ERK (p-ERK) binds to PGK1 in the cytoplasm, leading to an increase in the level of PGK1 S203 through phosphorylation modification. This subsequently promotes the translocation of PGK1 into mitochondria, inhibiting the TCA cycle and promoting anaerobic glycolysis [33]. Therefore, we hypothesized that a similar molecular mechanism may exist in OC cells. Then, we examined the levels of p-ERK and total ERK in both the DUSP4-OE and control cells. The results revealed that the p-ERK levels in the DUSP4-OE cells were significantly lower compared to the control group, while the level of total ERK remained relatively unchanged (Fig. 5A-B). These findings indicated that DUSP4 overexpression might modulate the PGK1-mediated anaerobic glycolysis pathway by reducing the level of p-ERK in OC cells.

Fig. 5

DUSP4 alleviates the phosphorylation of PGK1 at Ser203 in OC cells via p-ERK signaling. (A-B) Expression levels of ERK and p-ERK in DUSP4-overexpression and vector-loaded OC cells detected by WB. And the quantitative analysis using the relative gray values of protein bands. The significance of difference is determined by two-sided t test; *p indicates statistical difference; the data are presented as mean ± SD. (C-D) Co-immunoprecipitation detection of ERK and PGK1. (E-F) Co-immunoprecipitation WB detection of ERK and DUSP4

Although DUSP4 could simultaneously downregulate the levels of p-ERK and PGK1 S203, but whether there is a relatively close binding relationship between the three remains unknown. Thus, protein immunoprecipitation assays were conducted to investigate the potential binding relationship between ERK, PGK1, and p-ERK. The results revealed a direct binding effect between ERK and PGK1 in OC cells (SKOV3) (Fig. 5C-D). Moreover, PGK1 S203 was successfully detected in immunoprecipitation assay with anti-ERK antibody (Fig. 5C), while p-ERK was also detected in immunoprecipitation assay with anti-PGK1 antibody (Fig. 5D). These findings suggested that p-ERK could directly bind to PGK1 and phosphorylate it at the S203 site in OC cells. Furthermore, we also demonstrated that DUSP4 could bind to ERK (Fig. 5E-F). Combined with previous results in the study, it can be inferred that DUSP4 overexpression exerted a dephosphorylation effect on p-ERK. As a result, the reduced levels of p-ERK may lead to a decrease in its binding affinity towards PGK1, thereby resulting in less phosphorylation of PGK1 at the S203 site. This suggested that one potential mechanism by which DUSP4 overexpression influences the anaerobic glycolysis pathway is through modulating the phosphorylation status of ERK and its subsequent interaction with PGK1.

Decreased phosphorylation of PGK1 at Ser203 by DUSP4 retards OC tumor growth in vivo

In addition to in vitro experiments, we also conducted in vivo studies using a nude mouse ovarian orthotopic transplant tumor model with two cell lines, HEY A8 and SKOV3. Tumors derived from the DUSP4-OE group exhibited significantly smaller sizes compared to those from the control group. Quantitative analysis showed that the average ROI value of tumors in the control group was 2.3 × 10⁹ p/s for HEY A8 and 5 × 10⁸ p/s for SKOV3, whereas the average ROI value of tumors in the DUSP4-OE group was 1.99 × 10⁸ p/s for HEY A8 and 4 × 10⁷ p/s for SKOV3 (Fig. 6A-B). The tumor weight measurements displayed the same trend (Fig. 6C-D). These results indicated that DUSP4 overexpression reduced the tumor-forming ability of OC cells in vivo.

Fig. 6

DUSP4 suppresses tumor growth in vivo through p-ERK/ PGK1 Ser203. (A-B) In vivo imaging of the orthotopic ovarian tumor in nude mice injected with DUSP4-overexpression or vector-loaded OC cells. And the quantitative analysis using total luciferase flux value. (C-D) Size of orthotopic tumors in nude mice transplanted with DUSP4-overexpression or vector-loaded OC cells. And the quantitative analysis using the mass of tumors. (E-F) Protein expression levels in mouse tumor tissues detected using immunohistochemistry. The scale bar is 50 μm. And the relative quantitative analysis of protein levels in mouse tumor tissues. (G-H) Protein expression levels in mouse tumor tissues detected by WB. And the relative quantitative analysis of protein levels in mouse tumor tissues. (I-J) Linear regression analysis between DUSP4 and PGK1 S203 expression levels in mouse tumor tissues detected by immunohistochemistry. Pearson correlation coefficient and two-tailed p-value are calculated. For experiments above, the significance of difference is determined by two-sided t test; *p indicates statistical difference; the data are presented as mean ± SD

Furthermore, immunohistochemical staining analysis of the tumor samples revealed significantly lower levels of p-ERK and PGK1 S203 in the DUSP4-OE tumors compared to the control tumors (Fig. 6E-F). However, there was no significant difference in the expression levels of ERK and PGK1 between the two groups of tumors (Fig. 6E-F). This finding was further confirmed by WB analysis of the tumor samples (Fig. 6G-H). Based on the H-Score quantitative values obtained from immunohistochemical staining, we performed a correlation analysis between the expression levels of DUSP4 and PGK1 S203. The analysis revealed a significant negative linear correlation between the two, with R² values of 89.14% and 75.79%, respectively (Fig. 6I-J). Additionally, we observed a significant positive correlation between the levels of p-ERK and PGK1 S203 (Fig. S3A-B), while a negative correlation was found between the levels of DUSP4 and p-ERK (Fig. S3C-D). These findings provided further support to the in vitro experimental results mentioned earlier. Collectively, these data indicated that the decreased levels of p-ERK and PGK1 S203 induced by DUSP4 overexpression were important factors for suppressing OC progression (Fig. 7).

Fig. 7

The mechanisms of how elevated DUSP4 suppresses tumor growth through p-ERK/ PGK1 Ser203 signaling. Schematic illustration depicts the regulatory role of DUSP4 in tumor growth through the dephosphorylation of p-ERK and p-PGK1 at the Ser203 site. In OC, increased expression of DUSP4 interacts with p-ERK, resulting in the indirect dephosphorylation of p-PGK1 at Ser203. This dephosphorylation event facilitates the entry of p-PGK1 into the mitochondria, subsequently leading to the downregulation of anaerobic glycolysis and the proliferation of tumor cells

Discussion

Despite advancements in OC treatments, including PARP inhibitors and bevacizumab, survival outcomes remain poor, with five-year survival rates below 40% for advanced cases [4, 6]. This highlights the need for better understanding of OC mechanisms and new therapeutic strategies. Our study showed significant reductions in DUSP4 expression in OC tissues compared to benign masses, with further decreases in metastatic lesions. These results align with previous findings linking low DUSP4 levels to more aggressive OC behavior [29]. In addition, DUSP4 levels varied among OC subtypes, with significantly lower expression in serous ovarian cancer. Reduced DUSP4 expression was also linked to higher P53 and KI67 protein levels and differences in TP53 mutation rates, highlighting its role in tumor proliferation, mutation status, and key oncogenic pathways. Furthermore, low DUSP4 expression showed a strong correlation with advanced tumor stages and poorer prognosis, as confirmed by both our dataset and TCGA OV data. Collectively, this establishes DUSP4 as a valuable prognostic biomarker and a potential therapeutic target in OC.

DUSP4, a member of the DUSP family, acts as a phosphatase regulating key cellular processes like growth, apoptosis, and cell state transformation by inactivating ERK1/2, P38, and JNK [11]. It has been implicated in tumor progression across various cancers [24, 26, 30, 35, 36]. For example, DUSP4 overexpression promotes drug resistance in gastric cancer by enhancing EMT [37], while its downregulation activates the Ras-ERK pathway in breast cancer, leading to poor prognosis [27]. Similarly, in pancreatic cancer, loss of DUSP4 enhances tumor cell invasiveness and anoikis resistance by inactivating the ERK pathway [28]. In our study, we demonstrated that DUSP4 overexpression inhibits proliferation, migration, and invasion of OC cells in vitro. Phosphoproteomic profiling using LC-MS/MS further revealed the involvement of DUSP4 in regulating the MAPK pathway, cellular metabolism, proliferation, adhesion, and other critical processes in OC cells. Importantly, we observed that DUSP4 overexpression led to the downregulation of intracellular PGK1 S203 levels, a discovery that warrants further investigation. These findings emphasize the multifaceted role of DUSP4 in OC and its potential as a therapeutic target.

PGK1, a key enzyme in glycolysis, catalyzes the production of ATP by converting 1,3-bisphosphoglycerate to 3-phosphoglycerate, providing energy for cells [38]. Previous studies have shown that Its upregulation in tumor tissues is linked to tumor progression and poor prognosis [39,40,41]. Recent research has highlighted that phosphorylation of PGK1 at the S203 site enhances its catalytic activity, promotes glycolysis, and helps tumor cells adapt to hypoxic environments by interacting with HIF-1α [33, 34, 42]. Additionally, mitochondria-localized S203-phosphorylated PGK1 activates PDHK1, inhibiting mitochondrial pyruvate metabolism and promoting glycolysis [33]. Our study adds to this understanding by exploring the effects of DUSP4 overexpression on the localization and activity of PGK1 S203. We observed that DUSP4 overexpression reduces the mitochondrial localization of PGK1 S203, leading to decreased anaerobic glycolysis and lactate production in OC cells. These findings suggest that DUSP4 plays a role in modulating mitochondrial energy metabolism and disrupting tumor glycolytic pathways. Additionally, DUSP4 overexpression significantly increased ROS levels. The elevated ROS levels may disrupt tumor cell redox balance, contributing to oxidative stress and influencing metabolic dynamics. These findings highlight a potential link between DUSP4 expression, mitochondrial function, and oxidative stress regulation in cancer cells.

We further observed that DUSP4 overexpression led to a significant downregulation of p-ERK levels in OC cells. However, whether DUSP4 regulates PGK1 S203 via components of the MAPK pathway remains unclear. To explore this, we investigated potential interactions among them and confirmed interactions between ERK and PGK1, as well as between DUSP4 and ERK. Taken together, these findings suggest that DUSP4 overexpression likely directly dephosphorylates p-ERK, which in turn reduces PGK1 S203 phosphorylation, ultimately impacting glycolysis in OC cells. To extend these findings, we further demonstrated that DUSP4 overexpression inhibits the growth of OC cells in vivo. This was accompanied by reduced phosphorylation levels of key downstream molecules, including ERK and PGK1 S203, in tumor tissues from mouse models. Previous study has found the deficiency of ARID1A as a cause for DUSP4 downregulation, and demonstrated that ectopic DUSP4 expression inhibited cell proliferation via the MAPK pathway [23]. Our study aligns with these observations but expands on them by uncovering the mechanistic link between DUSP4, p-ERK, and PGK1, through which DUSP4 modulates tumor glycolysis and growth, providing new insights into its role in OC progression.

Through our current exploration, we have elucidated that DUSP4 overexpression primarily dephosphorylates p-ERK, leading to a reduction in mitochondrial PGK1 S203 phosphorylation. This decrease in PGK1 S203 levels suppresses anaerobic glycolysis, disrupts tumor metabolism, and inhibits the proliferation of OC cells. However, whether DUSP4 directly interacts with PGK1 in the nucleus to regulate its S203 phosphorylation, or whether additional proteins are involved in this regulatory process, remains unclear and needs further investigation. Also, the influence of TP53 mutation status on DUSP4 expression or function was not fully explored in this study, and the mechanisms behind their association still need further investigation. Although these issues remain, our findings provide new insights into the role of DUSP4 in OC progression by uncovering its comprehensive impact on protein phosphorylation and tumor metabolism. These results position DUSP4 as a potential therapeutic target and prognostic biomarker in OC, providing insights that could contribute to exploring improved treatment strategies.