Hepatoma-derived growth factor and non-coding RNA network in ovarian cancer patients

The study was performed in line with the principles of Helsinki declaration. Iran university of medical sciences ethics committee has approved and supervised the study. Ethical code #IR.IUMS.FMD.REC.1401.701 written informed consent was obtained from all participants.

Not applicable.

Iran university of medical sciences deputy of research and innovation has funded this project. Grant no #25760 .

Fifty ovarian cancer tissue samples were obtained from patients undergoing surgery at Rasool-e-Akram Hospital between May 2023 and February 2024. Additionally, fifty normal ovarian tissue samples were collected from patients who underwent total abdominal hysterectomy with bilateral salpingo-oophorectomy (TAH-BSO) for non-ovarian pathologies, such as early-stage endometrial carcinoma or acute uterine bleeding. All tissue samples were histopathologically verified prior to inclusion. None of the participants had received chemotherapy or radiotherapy before surgery. Preoperative blood samples were collected, and serum was isolated by centrifugation at 2000 g for 10 min. The sera were stored at −80 °C until further analysis. Control samples were obtained from healthy female volunteers attending routine gynecological check-ups. All participants, both cases and controls, were confirmed to not be pregnant at the time of sampling. This study was approved by the Institutional Review Board of Iran University of Medical Sciences (ethical code: IR.IUMS.FMD.REC.1401.701). To predict miRNAs that target the 3′ UTR of HDGF, we utilized a multi-database approach. First, we employed miRDB ( http://www.mirdb.org/ ) to identify potential miRNA candidates based on sequence complementarity and binding affinity. The HDGF mRNA sequence ( NM_004494 ) was analyzed, and candidates with binding scores above 60 were selected. Subsequently, we cross-referenced predictions with miRWalk ( http://mirwalk.umm.uni-heidelberg.de/ ), which assesses miRNA binding sites along the entire mRNA sequence. This dual approach allowed for a robust selection of potential miRNAs. For the reconstruction of the ceRNA network, human lncRNAs sequences were obtained from LNCipedia ( https://lncipedia.org/ ), focusing on lncRNAs expressed in ovarian cancer tissues identified through GEPIA2 ( http://gepia2.cancer-pku.cn/ ). To explore interactions between miR-345-5p and lncRNAs, we utilized LncBase v.2, which allowed us to identify LINC00839 as a potential ceRNA based on predicted binding sites. Lastly, the HDGF, miR-345-5p and, LINC00839 interaction was depicted using IntaRNA as shown in Fig. 4 via BioRender. Total RNA was isolated from tissues using an RNA extraction solution (BioBasic, Canada) following the manufacturer's protocol. RNA quality and quantity were assessed by gel electrophoresis and spectrophotometry using NanoDrop One instrument (Thermo Scientific, USA). Reverse transcription into cDNA was performed using a first-strand cDNA synthesis kit (Yektatajhiz, Iran), with specific RT stem-loop primers for miR-345-5p, RNU44, and miR-384 ( Table 1 ). Quantitative real-time PCR (qRT-PCR) was conducted using Real Q Plus 2x Master Mix Green High ROX (Ampliqon, Denmark) on a Step-One machine (Applied Biosystems, USA). Reactions were carried out in duplicate in a total volume of 10 μL. Primer sequences used for qPCR are detailed in Table 1 . GUSB was used as the reference gene to normalize HDGF and LINC00839 expression, while RNU44 was used to normalize miR-345-5p and miR-384 expression. The relative expression levels were determined using the 2 -ΔCt method. Table 1 Primers' sequence (5’>3’). Table 1 Gene Forward Reverse GUSB GTGGTGCGTAGGGACAAGAA AGGGGTCCAAGGATTTGGTG HDGF AGAAGGAGTACAAATGCGGGG GGGAAGAGGTCTTTGGGGC LINC00839 TGTGGCGTCCAGCCCTGAAA CCCCAACCCAGTAGTGCCTCG miR-384 CACCCAGTACACACACATTCCTAGAAAT CCAGTGCAGGGTCCGAGGTA U44 ACTCCAGCAGTTCAGGGATGATGATAA miR-345-5p GTCCAAGGCTGACTCCTAGTCCAG miR-384 RT GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACTATGAACA miR-345-5p RT GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACGAGCCCTG U44 RT GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACAACTGACT Primers' sequence (5’>3’). Enzyme-linked immunosorbent assays (ELISAs) were performed to measure serum levels of HDGF (BT Laboratories, China), HE4 (Xema, Russia), and CA125 (Pishtaz-Teb, Iran) using commercially available kits, following the manufacturers’ protocols. Data analysis and visualization were conducted using IBM SPSS version 27 and GraphPad Prism 8 software. The Mann Whitney U test was employed to compare expression levels across different groups, while the Kruskal-Wallis test was utilized to assess gene expression variations among various subtypes. Spearman correlation coefficient was calculated to determine the correlation between gene expression levels. Receiver operating curve (ROC) analysis was used to test biomarker performance. A significance level of p < 0.05 was considered statistically significant. The data is presented as median values with range.

Of our fifty patients, 27 were post-menopausal. Patients’ mean age was 49.8 years with 14.1 standard deviation. Most of our cases were diagnosed with serous carcinoma (36 %). Also, most of our cases were at stage III according to FIGO classification. Mean tumor size was 12.5 cm with 6.37 SD. Mean BMI was 24.26 kg.m −2 with 3.4 SD. Twenty percent of studied tumors were metastatic. The data is shown in Table 2 . Table 2 Patients’ characteristics. Table 2 Variable Range Age 49.8 ± 14.1 Menopause status Pre 23 (46 %) Post 27 (54 %) Subtype Serous 18 (36 %), Mucinous 13 (26 %), Clear cell 7 (14 %), Sero-mucinous 6 (12 %), Endometrioid 6 (12 %) FIGO stage I 10 (20 %) II 11 (22 %) III 20 (40 %) IV 9 (18 %) BMI 24.26 ± 3.4 Tumor size 12.5 ± 6.37 Lymph node metastasis Yes 10 (20 %) No 40 (80 %) Patients’ characteristics. In our candidate selection process, miR-345-5p and miR-384 were identified as significant miRNAs targeting HDGF through bioinformatic predictions, demonstrating a high binding affinity and stable interaction. The ceRNA network reconstruction revealed LINC00839 as a key lncRNA capable of sequestering miR-345-5p, thereby derepressing HDGF expression. HDGF revealed higher expression levels in ovarian cancer tissues compared to normal ovaries (p < 0.001, Fig. 1 A). miR-345-5p revealed lower expression levels in ovarian cancer tissues compared to normal ovaries (p < 0.001, Fig. 1 B). miR-384 revealed slightly lower expression in ovarian cancer tissue; however, it's on the brink of significance (p < 0.056, Fig. 1 C), while LINC00839 was significantly upregulated in OC (p < 0.001, Fig. 1 D). Also, RNA expression levels were compared in different variable groups. Age, BMI, tumor size, FIGO stage, lymph node metastasis, menopausal status, tumor subtype and familial history were the variables that were investigated. The comparison between these groups is shown in Table 3 . miR-345-5p and miR-384 showed no statistically significant variation among groups. LINC00839 showed higher expression levels in large tumors than small ones (p = 0.008). Also, HDGF is highly up-regulated in large tumors than small ones (p < 0.001). HDGF tend to have higher expression levels in advanced stages rather than early stages (p < 0.001, Fig. 1 E), likewise metastatic tumors showed higher expression levels than non-metastatic ones (p = 0.004). Fig. 1 A, B, C, D Comparison of gene expression levels between tumor and control. E. comparison of HDGF expression level between early and advanced ovarian cancer. Fig. 1 Table 3 The association between relative gene expression levels and clinicopathologic features of patients. Table 3 Variable HDGF LINC00839 miR-345-5p miR-384 Age <50 10.48 0.930 1.50 0.712 34.05 0.548 0.52 0.720 ≥50 14.32 1.77 47.83 0.85 BMI <25 13.97 0.533 1.71 0.494 39.32 0.566 0.58 0.259 ≥25 6.93 1.70 55.01 1.07 Tumor size <10 0.09 <0.001 0.07 0.008 60.12 0.215 0.56 0.385 ≥10 16.56 2.01 41.64 0.85 FIGO stage I, II 0.42 <0.001 0.38 0.057 47.17 0.867 0.56 0.630 III, IV 18.12 2.01 41.64 0.82 Lymph node metastasis Yes 21.08 0.004 2.56 0.113 44.73 0.480 0.84 0.971 No 4.78 1.45 47.17 0.71 Menopausal status Pre 7.41 0.239 1.40 0.572 52.34 0.697 0.56 0.930 Post 17.50 2.23 44.32 0.85 Tumor subtype Serous 11.50 0.483 1.07 0.322 44.73 0.263 1.04 0.646 Clear cell 4.11 0.74 186.10 0.95 Mucinous 14.32 2.23 30.48 0.67 Sero-mucinous 11.27 1.44 58.90 0.38 Endometrioid 19.66 2.23 28.34 0.64 Familial history Yes 13.18 0.693 1.69 0.542 44.73 0.990 1.14 0.319 No 10.72 1.70 47.17 0.65 A, B, C, D Comparison of gene expression levels between tumor and control. E. comparison of HDGF expression level between early and advanced ovarian cancer. The association between relative gene expression levels and clinicopathologic features of patients. miR-345-5p expression was inversely correlated with HDGF expression (p < 0.001, r = −0.58, shown in Fig. 2 A), while LINC00839 was positively correlated with HDGF expression (p < 0.001, r = 0.70, shown in Fig. 2 C) and, miR-345-5p and LINC00839 had inverse correlation (p < 0.001, r = −.0.52, shown in Fig. 2 D); Supporting the hypothesis that LINC00839 might act as a molecular sponge for miR-345-5p to augment expression of HDGF. miR-384 and HDGF showed no significant correlation in their expression (p = 0.31, shown in Fig. 2 B). Fig. 2 Correlation of expression levels between genes A. HDGF and miR-345-5p B. HDGF and miR-384 C. HDGF and LINC00839 D. LINC00839 and, miR-345-5p. Fig. 2 Correlation of expression levels between genes A. HDGF and miR-345-5p B. HDGF and miR-384 C. HDGF and LINC00839 D. LINC00839 and, miR-345-5p. HDGF protein levels in patients’ sera were measured. In addition, two available biomarkers for OC diagnosis, CA-125 and HE4, were measured. ROC analysis showed AUC of 0.88 (95 % CI: 0.82–0.95), 0.98 (95 % CI: 0.95–1.00) and, 0.95 (95 % CI: 0.91–0.99) for HDGF, CA-125 and, HE4 respectively (p < 0.001). The ROC curves are depicted in Fig. 3 . Fig. 3 ROC of CA125, HE4 and, HDGF. Fig. 3 Fig. 4 RNA:RNA interactions and binding sites of miR-345-5p, miR-384, LINC00839 and, HDGF. Fig. 4 ROC of CA125, HE4 and, HDGF. RNA:RNA interactions and binding sites of miR-345-5p, miR-384, LINC00839 and, HDGF.

Not applicable.

R.G conceptualized the study, performed experiments and, wrote the manuscript. S.N, N.Z and, Z.S collected clinical data. N.H and M.E analyzed the data. M.N acquired the grant and supervised the study.

Ovarian cancer (OC) is the most lethal of all gynecologic malignancies, with a 5-year survival rate of only 30 %, primarily due to the asymptomatic nature of early disease and the lack of effective screening tools, which result in more than 80 % of patients being diagnosed at advanced stages [ 1 , 2 ]. These limitations underscore the urgent need to identify novel biomarkers and molecular pathways that can enable early detection, predict prognosis, and inform targeted therapies. Hepatoma-derived growth factor (HDGF) has emerged as a multifunctional protein involved in diverse tumor-promoting processes including proliferation, angiogenesis, and metastasis [ 3 ]. Initially discovered in hepatoma cells, HDGF has since been implicated in numerous cancers such as breast, lung, glioma, and hepatocellular carcinoma where it is often associated with poor prognosis and aggressive tumor behavior [ [4] , [5] , [6] , [7] , [8] ]. In ovarian cancer, preliminary evidence suggests that HDGF might contribute to tumor progression by promoting cell proliferation, resistance to apoptosis, and metastasis [ [9] , [10] , [11] ]. Understanding the dysregulation of HDGF and mechanisms influencing its expression in ovarian cancer could lead to novel diagnostic or therapeutic targets. Recent advances in non-coding RNA biology have revealed that microRNAs (miRNAs) and long non-coding RNAs (lncRNAs) are critical regulators of gene expression in cancer [ 12 ]. An emerging concept, the competing endogenous RNA (ceRNA) hypothesis, suggests that lncRNAs can act as molecular sponges for miRNAs, thereby modulating the availability of miRNAs to target mRNAs [ 13 , 14 ]. In this context, computational pharmacology and network-based approaches have become invaluable tools in cancer research. These methodologies integrate transcriptomic data, RNA-RNA interaction predictions, and pathway enrichment to identify critical regulatory hubs and ceRNA interactions. Such systems biology frameworks enhance our understanding of complex molecular interplay and aid in the discovery of actionable targets. By leveraging bioinformatics tools to reconstruct gene–miRNA–lncRNA networks, researchers can prioritize candidate biomarkers with higher accuracy and functional relevance, especially in heterogeneous diseases like OC [ 15 ]. This study aims to investigate the expression levels of HDGF in ovarian cancer and to construct a ceRNA network that elucidates the regulatory mechanisms influencing HDGF expression. Using a combination of experimental validation and in silico network reconstruction, we propose a novel ceRNA axis involving LINC00839 sponging miR-345-5p to regulate HDGF expression. By identifying HDGF as a potential biomarker, we seek to advance its clinical applicability, ultimately contributing to enhanced diagnostic and therapeutic approaches for ovarian cancer.

Hepatoma-derived growth factor (HDGF) has emerged as a multifaceted player in tumor biology, acting not only as a mitogen but also as a nuclear protein with transcriptional regulatory activity. In the context of ovarian cancer (OC), our findings corroborate growing evidence that HDGF is markedly upregulated in tumor tissues, suggesting it plays an oncogenic role. Mechanistically, HDGF engages with multiple intracellular pathways that are critically implicated in carcinogenesis across various cancers, several of which are also relevant to ovarian cancer [ 3 , 16 ]. HDGF has been shown to interact with nucleosome assembly protein 1-like 1 (NAP1L1) and c-JUN to activate downstream targets such as CCND1, promoting cell cycle progression in OC cells [ 11 ]. This interaction not only enhances proliferation but may also contribute to chemoresistance by stabilizing oncogenic signaling circuits. Additionally, HDGF has been implicated in regulating β-catenin activity [ 17 ]—a pathway known to be aberrantly activated in many ovarian tumors [ 18 ]. Through its interactions with DDX5 and Wnt/β-catenin components, HDGF may reinforce transcriptional programs that sustain tumor growth and survival [ 17 ]. Beyond its role in tumor biology, HDGF holds potential as a clinical biomarker. Liu et al. have demonstrated that HDGF expression is significantly associated with poor prognosis and lymphatic metastasis in OC, supporting its role as an independent prognostic indicator [ 19 ]. Giri et al. also reported that HDGF is secreted upon apoptosis or necrosis of OC cells [ 9 ], suggesting its possible detectability in patient serum during active disease. This notion is supported by Zhang et al.'s findings in lung cancer, where serum HDGF levels correlated with prognosis [ 20 ], implying a similar utility in ovarian cancer. While we observed that current diagnostic markers such as CA-125 and HE4 outperform HDGF in sensitivity and specificity for early detection, HDGF may still serve a complementary role, especially for monitoring disease progression or recurrence. Furthermore, the involvement of HDGF in key oncogenic pathways like PI3K/Akt, ERK, and Wnt/β-catenin underscores its potential as a therapeutic target. Inhibition of HDGF or its interacting partners (e.g., NAP1L1, c-JUN, DDX5) could disrupt these critical signaling axes and sensitize tumors to existing treatments [ 21 ]. Long non-coding RNAs (lncRNAs) have increasingly been recognized as important regulators in cancer, exerting their effects through diverse mechanisms including the competing endogenous RNA (ceRNA) network. Among these, LINC00839 has emerged as a potent oncogenic lncRNA [ 22 ]. Our findings reveal that LINC00839 is significantly upregulated in ovarian cancer (OC) tissues, with a strong positive correlation with HDGF expression and a negative correlation with miR-345-5p levels. This suggests a ceRNA mechanism whereby LINC00839 may act as a molecular sponge for miR-345-5p, thereby derepressing HDGF—a known promoter of proliferation and metastasis in OC. In line with our data, LINC00839 has been implicated in tumor progression across multiple cancer types [ 22 ]. In breast cancer, high LINC00839 expression is associated with poor prognosis and chemoresistance through activation of the PI3K/AKT signaling pathway. Knockdown of LINC00839 sensitized cells to paclitaxel and suppressed tumor growth both in vitro and in vivo [ 23 ]. Similarly, in glioblastoma, LINC00839 promotes radiation resistance by activating the Wnt/β-catenin pathway through METTL3-mediated m6A modifications [ 24 ]. In colorectal cancer, its nuclear localization facilitates recruitment of RUVBL1/Tip60 complexes to activate NRF1, thereby enhancing tumor proliferation and invasion [ 25 ]. From a clinical perspective, elevated LINC00839 expression correlates with poor overall survival in several cancers, including neuroblastoma, liver cancer, and nasopharyngeal carcinoma [ 22 ]. Its expression is also associated with aggressive clinicopathologic features such as increased metastasis, resistance to gemcitabine in bladder cancer [ 26 ], and enhanced epithelial-mesenchymal transition (EMT) in nasopharyngeal carcinoma [ 27 ]. These studies suggest that LINC00839 may serve as a biomarker not only for prognosis but also for predicting therapeutic responses. In the context of ovarian cancer, where resistance to chemotherapy and disease recurrence remain significant clinical challenges, LINC00839's role in modulating HDGF expression and potentially activating PI3K/AKT and Wnt/β-catenin signaling cascades underscores its therapeutic relevance. Targeting LINC00839 could offer a dual benefit: attenuating oncogenic signaling and sensitizing tumors to standard therapies. Moreover, the correlation between its expression and clinical outcomes in other malignancies highlights its promise as a biomarker for disease progression, prognosis, and treatment stratification in OC. MicroRNAs (miRNAs) are well-established post-transcriptional regulators of gene expression, with a pivotal role in cancer progression. miR-345-5p, which is downregulated in our OC samples, has been shown to act as a tumor suppressor in various malignancies [ [28] , [29] , [30] , [31] ]. Our results indicate an inverse correlation between miR-345-5p and HDGF expression, supporting the hypothesis that miR-345-5p directly targets the 3′ untranslated region (3′UTR) of HDGF, as has been demonstrated in glioma cells [ 32 ]. The reduced levels of miR-345-5p in OC might relieve its suppression on HDGF, leading to heightened HDGF expression, thereby accelerating tumor growth. In contrast, while miR-384 has also been implicated as a tumor suppressor in multiple cancers by regulating proliferation, migration, and metastasis [ [33] , [34] , [35] , [36] , [37] ] and it is perceived that it suppresses cancer cell proliferation and invasion in glioma and breast cancer cells by directly targeting HDGF [ 38 , 39 ], our data did not show a significant change in miR-384 expression in OC tissues. This could suggest a context-dependent role for miR-384, or that its regulation of HDGF in OC is less prominent compared to miR-345-5p. The discrepancy in miR-384's activity across different cancer types underscores the complexity of miRNA-mediated regulation and the need for further investigation into its specific role in OC.

While this study provides valuable insights into the potential role of miR-345-5p and HDGF in ovarian cancer, it is important to acknowledge several limitations that must be considered when interpreting the results. A key limitation is the relatively small sample size of the patient cohort. Working with a larger, more diverse cohort would be essential to strengthen the validity of the findings and allow for more definitive conclusions. Future studies involving larger cohorts could provide the statistical power necessary to confirm these results, potentially leading to the validation and development of HDGF as biomarker for ovarian cancer prognosis and treatment. In addition to expanding patient cohorts, further research is needed to explore the therapeutic potential of targeting HDGF in ovarian cancer. Given the biological role of HDGF in tumor progression, the use of antibodies against HDGF could be investigated in animal models. Studies employing animal models or even early-phase clinical trials might reveal whether targeting HDGF could induce tumor shrinkage or regression [ 40 ]. This approach holds promise as a novel therapeutic strategy, particularly for patients with advanced or treatment-resistant ovarian cancer. Moreover, miR-345-5p offers potential as a therapeutic agent [ 41 ]. Its role in modulating tumor-related pathways suggests that miR-345-5p could be incorporated into drug delivery systems to selectively target ovarian cancer cells, improving treatment outcomes. Future research should focus on the development and optimization of miRNA-based delivery systems, such as nanoparticles, to enhance the efficacy and specificity of these approaches.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

The data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request.