Cell line–dependent effects of β‑estradiol and testosterone on aurora kinases, GATA3, and associated miRNAs in hepatocellular carcinoma models HepG2 and Huh‑7 Authors: Taraneh Rezaei, Mobina Abidi, Yaser Mohammadi, Farnoosh Farzam, Elham Bahreini Published: 2026 ABSTRACT AbstractBackgroundHigher prevalence of hepatocellular carcinoma (HCC) in men is largely attributed to the sex hormones and behavioral risk factors. This study investigated the effects of male (testosterone) and female (estradiol) hormones on Aurora-A/B kinases, GATA-3, and microRNAs (miR-3941, miR-4500, miR-4742) in HepG2 and Huh-7 cell lines.MethodsIn this study, two human hepatocellular carcinoma cell lines, HepG2 (with a differentiated phenotype) and Huh-7 (with more aggressive tumorigenic behavior), were treated with the sex hormones β-estradiol and testosterone. Hormonal effects were evaluated using the MTT assay (to assess cytotoxicity), flow cytometry (for cell cycle analysis), RT-qPCR (to quantify the expression of target genes and miRNAs, including Aurora-A, Aurora-B, GATA-3, miR-3941, miR-4500, and miR-4742), and western blotting (to assess protein levels of Aurora-A/B kinases and GATA-3).ResultsIn HepG2 cells, estradiol and testosterone decreased Aurora-A/B kinase expression and increased GATA3, causing G0/G1 and G2/M cell cycle arrest and reduced DNA synthesis. Conversely, in the more aggressive Huh7 cell line, these hormones accelerated G1–S transition and led to late cell cycle accumulation. Both cell lines exhibited inverse miRNA regulation (miR-3941 and miR-4500 upregulated; miR-4742 downregulated), suggesting a post-transcriptional mechanism affecting cell division genes. Multivariate regression indicated Aurora-A/B independently regulated G0/G1–S transition in HepG2, while temporal factors predominated in Huh7 cell cycle control.ConclusionA simultaneous comparison of female and male sex hormones in two HCC cell models—one less aggressive (HepG2) and one highly aggressive (Huh-7)—demonstrated that the direction of hormonal effects (inhibitory or stimulatory) depends not on hormone type but on cellular phenotype, and downstream regulatory networks. This duality of response was similar for both hormones within each cell line. This paradox may be attributed to the predominance of distinct signaling pathways—protective in HepG2 versus proliferative in Huh-7 cells.Keywords: Hepatocellular carcinoma, Sex hormones, Aurora kinases, MicroRNAs Introduction Hepatocellular carcinoma (HCC) is the predominant form of primary liver cancer and a major cause of cancer-related deaths globally. A striking feature of HCC epidemiology is its pronounced gender disparity, with males showing significantly higher incidence and mortality rates than females [1]. Some studies attribute this disparity to biological factors, particularly the role of sex hormones in liver physiology and tumorigenesis, while others point to behavioral risk factors more common in men, such as alcohol consumption, smoking, and higher rates of chronic hepatitis infections [1, 2]. These findings have spurred interest in how sex hormones influence molecular pathways in liver cancer, including gene expression, oncogenic signaling, and epigenetic mechanisms. Dysregulation of the cell cycle and uncontrolled cell proliferation are key drivers of HCC development [3]. Aurora kinases, a conserved family of serine/threonine kinases, are critical for regulating cell division. Aurora-A and Aurora-B, key family members, ensure proper sister chromatid segregation and genomic stability [4]. Aberrant expression or activity of these kinases promotes uncontrolled cell division and tumor growth. Aurora-A accumulates at centrosomes from late G1 to early S phase, becoming activated through phosphorylation events [5]. Its activity peaks during the G2/M transition, driving centrosome maturation, mitotic spindle assembly, and chromosome segregation. It remains associated with the spindle until telophase, then relocates to the midbody during cytokinesis to complete cytoplasmic division [4]. The Aurora-A gene, located on chromosome 20q13.2, is overexpressed in cancers such as breast, prostate, and HCC, leading to aneuploidy, abnormal differentiation, and poor prognosis [4, 5]. The tumor suppressor p53 regulates Aurora-A activity, and disruptions in this pathway cause aberrant kinase activation. Studies show that inhibiting Aurora-A with agents like VE-465 has significant anticancer effects, making it a promising therapeutic target [6]. Aurora-B, a core component of the Chromosomal Passenger Complex (CPC), is essential for chromosome segregation and cytokinesis [7]. Despite sharing high catalytic domain similarity with Aurora-A, it differs in genomic location (chromosome 17p13.1), subcellular distribution, and substrate specificity [8]. Aurora-B concentratanaphase anderes during mitotic entry, relocates to the spindle midzone in anaphase, and moves to the midbody in telophase [7]. It phosphorylates histone H3 at Ser10 to promote chromosome condensation, correct biorientation, and activate the spindle assembly checkpoint [9]. Its expression peaks during the G2/M transition and remains elevated throughout mitosis [8, 9]. Immunofluorescence reveals dynamic Aurora-B binding to centromeres with continuous cytoplasmic exchange [4]. Aurora-B is often overexpressed in solid tumors and hematologic malignancies, contributing to chromosomal instability, aneuploidy, and poor prognosis [10]. Genetic or pharmacologic inhibition of Aurora-B disrupts chromosome alignment and cytokinesis, leading to polyploidy and cell death [11]. Notably, some early Aurora kinase inhibitors, initially designed for Aurora-A, exert antitumor effects primarily through Aurora-B inhibition [4]. Thus, distinguishing the roles of Aurora kinases and developing isoform-specific inhibitors are critical for advancing targeted cancer therapies. In this context, GATA-3, a transcription factor encoded by the GATA3 gene on chromosome 10p14, is increasingly recognized as a key regulator of cell cycle and division in HCC [12]. Its zinc finger domains enable binding to specific DNA sequences in gene promoters, influencing cellular differentiation, tissue development, inflammation, and malignancies. Recent studies show GATA-3 directly binds to Aurora-A and Aurora-B gene promoters, modulating their expression and controlling cell cycle progression and mitosis [13]. In cancers like breast cancer, elevated GATA-3 expression correlates with Aurora kinase activation and enhanced estrogen receptor (ER-α) signaling [14]. In HCC, emerging evidence links GATA-3 dysregulation to increased Aurora-A/B expression, potentially driving tumor aggressiveness and poor prognosis [15]. These findings suggest GATA-3, through its interaction with Aurora kinases, plays a pivotal role in HCC cell division and tumor progression. Understanding their shared regulation, particularly via microRNAs (miRNAs), could yield new targeted therapies for liver cancer. Aberrant miRNA expression is observed across various cancers, including HCC, with distinct profiles between normal hepatocytes and HCC cells, highlighting miRNA profiling’s potential for identifying biomarkers and therapeutic targets. Bioinformatic analysis of Aurora-A, Aurora-B, and GATA-3 identified three strongly associated miRNAs: miR-3941, miR-4500, and miR-4742, respectively. Studies by Vasuri et al., Sun et al., and Cai et al. report their downregulation in HCC tissues [16–18]. Given the critical roles of these regulatory factors in cell proliferation and tumorigenesis, and evidence that sex hormones and miRNAs modulate their expression, this study investigates the effects of female (17-β-estradiol) and male (testosterone) hormones on Aurora-A, Aurora-B, GATA-3, and their associated miRNAs in HCC. The findings may offer new insights into hormonal and post-transcriptional regulation of oncogenic pathways in liver cancer and identify potential therapeutic targets. Materials and methods Chemicals and reagents The materials utilized in this study included β-estradiol (Sigma-Aldrich, Cat. No. E8875) and testosterone (Sigma-Aldrich, Cat. No. T1500). Cell culture was performed using DMEM/F12-Glutamax medium (Bio Idea, Cat. No. BI-1011) supplemented with Fetal Bovine Serum (FBS; Biosera, Cat. No. F-1001/100) and Penicillin-Streptomycin (Pen-Strep; Bio Idea, Cat. No. BI-1203). For cell detachment, Trypsin-EDTA 10x (Bio Idea, Cat. No. BI-1602) was employed. Cell viability assays were conducted using MTT reagent (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; Sigma-Aldrich, Cat. No. M5655), and formazan crystals were solubilized in DMSO (Dimethyl Sulfoxide; Sigma-Aldrich, Cat. No. 102952). Lipid accumulation was assessed using Oil Red O staining (Sigma-Aldrich, Cat. No. 1320-06-5). For gene expression analysis, total RNA was extracted using a kit from Yekta Tajhiz (Cat. No. FAPDE050), followed by cDNA synthesis using the same company’s cDNA Synthesis Kit (Cat. No. YT4500). Quantitative PCR was performed using Super SYBR Green qPCR Master Mix (Yekta Tajhiz, Cat. No. YT2552). Cell culture Cell line characteristics The HepG2 human hepatocellular carcinoma cell line was obtained from the Iranian Biological Resource Center (IBRC), and the Huh7 cell line was sourced from the Pasteur Institute of Iran. HepG2 and Huh7 are widely used HCC cell lines with distinct molecular and metabolic features representing liver cancer heterogeneity. HepG2 cells, derived in 1975 from a liver tumor in a 15-year-old Caucasian male, exhibit a well-differentiated, hepatoblastoma-like phenotype. These cells exhibit epithelial morphology, grow as adherent clusters, and retain hepatic functions, including albumin secretion, lipoprotein metabolism, and glucose transport. However, their limited cytochrome P450 activity restricts their use in drug metabolism studies. HepG2 cells serve as a model of differentiated hepatocytes, commonly used in lipid metabolism, toxicology, and hormonal regulation research. In contrast, Huh7 cells, derived from a Japanese male with well-differentiated HCC in 1982, are more tumorigenic and less differentiated than HepG2. They exhibit higher proliferation rates, greater invasiveness, and more active oncogenic signaling. Huh7 cells express higher levels of drug-metabolizing enzymes, such as cytochrome P450 isoforms (e.g., CYP3A4), making them suitable for drug metabolism and hepatotoxicity studies. They are often used to model viral hepatitis, metabolic reprogramming, and advanced HCC behavior. Metabolically, HepG2 cells primarily rely on oxidative phosphorylation, resembling a more “normal” energy profile, whereas Huh7 cells show a shift toward glycolysis (Warburg effect), characteristic of aggressive tumor cells. This metabolic divergence designates HepG2 as a model for early-stage and Huh7 for advanced-stage HCC. Their combined use offers a complementary system to study HCC heterogeneity, allowing thorough analysis of molecular pathways, metabolism, and therapeutic targets across tumor stages. Cell culture conditions HepG2 and Huh7 cells were cultured in DMEM/F12 (GlutaMAX™) medium (Gibco) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin, and maintained at 37 °C in a humidified incubator with 5% CO₂. Cells were monitored daily using an inverted microscope to assess morphology, confluency, and potential contamination. Due to the slower proliferation rate of HepG2 cells, their culture medium was replaced every 24 h. For Huh7 cells, medium replacement was adjusted based on their faster proliferation rate. Cytotoxicity and effective concentration assessment of β-estradiol and testosterone To assess the effects of testosterone, and 17-β-estradiol on HepG2 and Huh7 cell viability, the MTT colorimetric assay was used. In this assay, viable cells reduce yellow tetrazolium salt (MTT) via mitochondrial dehydrogenase enzymes (e.g., succinate dehydrogenase) to form purple formazan crystals, which are dissolved in DMSO. Absorbance is measured at 570 nm using an ELISA plate reader, with color intensity proportional to viable cell count. Briefly, 25 × 10³ HepG2 cells were seeded per well in a 96-well plate and stabilized for 24 h. Cells were treated with various concentrations of the compounds. After 24 and 48 h, MTT solution was added and incubated for 4 h. The medium was removed, DMSO was added to dissolve formazan crystals, and after 15 min of incubation on a shaker, absorbance was measured at 570 nm. The percentage of viable cells relative to the control was calculated, and a survival curve was plotted to determine the effective concentration for 50% viability (V50 or IC50). The MTT assay was performed using four biological replicates (n = 4) for each experimental condition. To investigate dose-dependent effects, we utilized specific hormone concentrations (E2 or T) shown to yield approximately 70% (or V70 defined as a lower/‘less toxic’ dose), 50% (or V50 defined as an intermediate toxic’ dose), or 40% (or V40 defined as the highest / more toxic’ dose) cell viability in HepG2 and Huh7 cell lines, as detailed in Table 1.: Table 1Hormone concentrations for dose-dependent studiesConc.CellE2-V70 (µM)E2-V50 (µM)E2-V40 (µM)T-V70 (µM)T-V50 (µM)T-V40 (µM) HepG2 250320370220320360 Huh7 400600720350620700Specific concentrations (µM) of Estradiol (E2) and Testosterone (T) were used to achieve approximately 70%, 50%, and 40% cell viability in HepG2 and Huh7 cells. These concentrations were determined via MTT and represent a range from less toxic (V70) to more toxic (V40) doses relative to untreated controls Hormone concentrations for dose-dependent studies Specific concentrations (µM) of Estradiol (E2) and Testosterone (T) were used to achieve approximately 70%, 50%, and 40% cell viability in HepG2 and Huh7 cells. These concentrations were determined via MTT and represent a range from less toxic (V70) to more toxic (V40) doses relative to untreated controls Cell cycle analysis by flow cytometry The effects of sex hormones (testosterone and 17-β-estradiol) on the cell cycle were investigated using flow cytometry. This technique quantifies the distribution of cells in the G0/G1, S, and G2/M phases by measuring DNA content. Cells were stained with propidium iodide (PI), a DNA-binding dye that intercalates with DNA and emits fluorescence upon laser excitation in the flow cytometer. Fluorescence intensity reflects cellular DNA content, with distinct patterns for each cell cycle phase: G0/G1 cells exhibit diploid DNA content (2n), S-phase cells show intermediate DNA content due to active DNA synthesis, and G2/M cells have tetraploid DNA content (4n). HepG2 and Huh7 cells (1 × 10⁶ per well) were seeded in 6-well plates with 2 mL DMEM/F12 medium (10% FBS, 1% penicillin-streptomycin) and incubated at 37 °C with 5% CO₂ for 24 h to allow attachment. Cells were treated with testosterone and 17-β-estradiol for 24 h, trypsinized, and centrifuged at 1500 rpm for 5 min at 4 °C. After washing with cold PBS, cells were fixed in cold 70% ethanol (added dropwise while vortexing) and stored at − 20 °C for at least 2 h. Fixed cells were washed with PBS, resuspended in 500 µL staining solution (50 µg/mL PI, 100 µg/mL RNase A), and incubated for 30 min at 37 °C in the dark. Finally, stained cells were analyzed using a flow cytometer (BD FACSCalibur), and cell cycle distribution (G0/G1, S, and G2/M phases) was determined based on DNA content using FlowJo software. The experiments included three biological replicates (n = 3). RT-qPCR analysis of mRNA and miRNA expression Expression levels of Aurora A, Aurora B, and GATA3 mRNAs, as well as miR-3941, miR-4500, and miR-4742, were quantified using quantitative real-time PCR (RT-qPCR). Total RNA was extracted from treated HepG2 and Huh7 cells in 6-well plates using an RNA extraction kit (FABRK001, Favorgen) per the manufacturer’s instructions, with lysis buffer supplemented with 2-mercaptoethanol. RNA integrity was confirmed by agarose gel electrophoresis, and concentration and purity (A260/A280 ≈ 1.8–2.0) were measured using a NanoDrop spectrophotometer (BOE 8637000, Model N-1). For mRNA analysis, cDNA was synthesized from 1 µg total RNA using a cDNA synthesis kit (YT4500, Yekta Tajhiz) with oligo(dT) and random hexamer primers. For miRNAs, cDNA was generated using stem-loop-specific reverse transcription primers for miR-3941, miR-4500, and miR-4742, alongside a universal RT primer system, with U6 small nuclear RNA as the endogenous control. RT-qPCR was performed using cDNA derived from each experimental condition, and Super SYBR Green qPCR Master Mix (YT2552, Yekta Tajhiz) with gene-specific forward primers and either gene-specific (for mRNAs) or universal (for miRNAs) reverse primers. For each cDNA sample, technical triplicates (n = 3) were run to assess amplification. Amplification was conducted under optimized cycling conditions, with primer specificity verified by melt curve analysis. Ct values were normalized to TBP for mRNAs and U6 for miRNAs, and relative expression was calculated using the 2–ΔΔCt method. Table 2 shows the sequences of the primers used. Table 2Sequences of primers used for RT-qPCRGeneSequenceTm-annealing °CAurora-AFAGGCTCAGCGGGTCTTGTGTC64 °CRTGGAGACAGGATGAGGTACACTGGAurora-BFTCTGCCCCAGCGAGTCCTCC64 °CRAGATGTCGGGTGTCCCACTGCGATA3FTCCGCACCTCTTCACCTTCCC64 °CRTGAGGCACTCTTTCTCGTCCTGCTBPFGTTCAGCAGTCAACGTCCCAG61 °CRGTCATGGGGGAGGGATACAGTG Sequences of primers used for RT-qPCR Western blot analysis of protein expression To assess Aurora A, Aurora B, and GATA3 protein expression, Western blotting was performed on HepG2 and Huh7 cells following treatment. Cells in 6-well plates were lysed in buffer containing 1% SDS, 10% glycerol, 50 mM Tris-HCl (pH 6.8), and protease inhibitor cocktail, incubated on ice for 15 min, and centrifuged at 13,000 rpm for 10 min at 4 °C to remove debris. Supernatants were collected for protein analysis. Protein concentrations were determined using the Bradford assay. Proteins (30–40 µg per sample) were separated by 10% SDS-PAGE and transferred to PVDF membranes via wet or semi-dry transfer. Membranes were blocked with 5% non-fat dry milk in TBS-T (0.1% Tween-20) for 1 h at room temperature, then incubated overnight at 4 °C with primary antibodies against Aurora A, Aurora B, GATA3 (1:1000; Cell Signaling, Abcam), and β-actin (loading control). After TBS-T washes, membranes were incubated with HRP-conjugated secondary antibodies (1:5000) for 1 h at room temperature. Protein bands were detected using enhanced chemiluminescence (ECL) reagents and imaged with a gel documentation system. Band intensities were quantified using ImageJ software, normalized to β-actin, and expressed as a percentage of the control group. This analysis was based on three biological replicates (n = 3). Statistical analysis Data was analyzed using GraphPad Prism 9. The Shapiro-Wilk test was applied to verify the normal distribution of the datasets. Differences between multiple groups were analyzed by one-way analysis of variance (ANOVA), followed by Tukey’s post-hoc test for pairwise comparisons. All values are expressed as mean ± standard deviation (SD), and differences were considered statistically significant at P < 0.05. Results Cytotoxicity assessment and determination of optimal working concentrations of β-estradiol and testosterone in HepG2 and Huh7 cells The MTT assay was employed to characterize the cytotoxic effects of β-estradiol (E2) and testosterone (T) and to define appropriate concentration ranges for subsequent experiments. Cell viability was quantified following 24- and 48-h treatments, and concentration–response relationships were generated by plotting normalized viability values against hormone concentrations. Half-maximal inhibitory concentrations (IC₅₀) were derived by nonlinear regression using a four-parameter logistic model (see Methods; Fig. 1A–D). Dose–response profiles illustrating the effects of β-estradiol and testosterone on HepG2 cell viability after 24 h of exposure are shown in panels A and B, respectively. Corresponding viability responses of Huh7 cells under the same treatment conditions are presented in panels C and D Although IC₅₀ values were comparable between the two exposure durations, prolonged treatment resulted in a consistent decline in cell viability, indicative of cumulative cytotoxic effects. To limit time-dependent variability and enhance reproducibility across downstream mechanistic assays, a 24-h exposure duration was selected for all subsequent analyses. Based on 24-h treatments, the estimated IC₅₀ values for E2 were approximately 320 µM in HepG2 cells and 600 µM in Huh7 cells (Fig. 1A, C), whereas corresponding values for T were approximately 320 µM and 620 µM, respectively (Fig. 1B, D). The in‑vitro system used here was specifically designed to identify concentrations at which differential hormone effects become detectable; these effects were not observed at physiological (nanomolar) levels. Consequently, we employed the MTT assay, and the resulting dose–response curves are shown in Fig. 1. The use of micromolar hormone concentrations to achieve detectable biological effects in hepatic cell lines is consistent with previous in vitro studies reporting similar exposure ranges [19, 20]. This requirement is commonly attributed to extensive binding of steroid hormones to FBS proteins in the culture medium, which substantially reduces the freely bioavailable fraction, as well as to the absence of metabolic activation mechanisms under in vitro conditions. Consequently, higher concentrations of active hormones are required to elicit detectable biological effects. Nevertheless, the selection of a 24-h exposure window for subsequent experiments minimized cumulative cytotoxic effects and enhanced experimental reproducibility. Effects of E2, T, and combined treatment on cell cycle distribution To evaluate the effects of β-estradiol (E2), testosterone (T), and their combination (Mix) on cell cycle progression, cell cycle phase distributions were analyzed using a two-way analysis of variance (Two-way ANOVA). Treatment type (E2, T, and Mix) and incubation duration (24, 36, and 48 h) were defined as independent variables. This analytical approach allowed assessment of the main effects of each factor as well as their interaction on cell cycle phase distribution in HepG2 and Huh7 cells. Cell cycle responses to hormonal treatments in HepG2 cells In HepG2 cells, negative control cultures (untreated) exhibited stable cell cycle phase distributions across G0/G1 (Fig. 2. A), S (Fig. 2. B), and G2/M (Fig. 2. C) throughout the incubation period from 24 to 48 h, with no significant temporal changes observed (all P > 0.05). In contrast, treatment with E2, T, or their combination (Mix) induced pronounced time-dependent alterations in cell cycle progression. These effects emerged as early as 24 h and reached their peak at 36 h, whereas extending the incubation to 48 h did not result in further significant modifications (P > 0.05). Specifically, the proportion of cells in G0/G1 increased significantly by 36 h (P < 0.05), accompanied by a concomitant and substantial decrease in S-phase cells, indicating delayed progression through G1 and reduced DNA synthesis. Concurrently, the percentage of cells in G2/M was significantly elevated (P < 0.05), suggesting impeded or delayed mitotic progression. This pattern was consistent across all three hormonal treatments, with no statistically significant differences between E2, T, and Mix (P > 0.05). While no significant interaction between treatment type and incubation time was observed for the G0/G1 phase, this interaction reached statistical significance for the S and G2/M phases (P < 0.05), highlighting that the observed changes in these phases are jointly dependent on both hormone type and exposure duration. Time-dependent effects of β-estradiol (E2), testosterone (T), and their combination [Mix (E2/T)] on cell cycle phase distribution in HepG2 cells. Panels A–C illustrate the percentages of cells in the G0/G1, S, and G2/M phases, respectively, following 24, 36, and 48 h of incubation after hormonal treatments, compared with the negative control (Neg-C). Tukey’s multiple comparisons tests indicate (I) time-dependent differences and (II) treatment-dependent differences for each cell cycle phase of G0/G1, S, and G2/M. (NS, not significant; **P < 0.01; ***P < 0.001) Cell cycle responses to hormonal treatments in Huh7 cells The simultaneous decrease in the G0/G1-phase (Fig. 3. A) and S-phase (Fig. 3. B) cells from 24 to 36 h of incubation (P < 0.05), reflecting the natural slowing of the cell cycle and gradual exit of cells from proliferation, were likely due to increased cell density or resource limitation. These changes plateaued between 36 and 48 h, with no significant alteration in G2/M-phase (Fig. 3. C) distribution (P > 0.05). In contrast, treatment with E2, T, or their combination (Mix) induced pronounced time-dependent effects, promoting cell cycle progression and facilitating cellular proliferation. Initially (24 h), hormonal treatments modestly increased the G0/G1 population compared with the negative control (P < 0.05); however, with extended incubation, this percentage decreased, indicating accelerated transition from G1 to S and faster exit from the resting phase. E2 triggered this reduction earlier (36 h), whereas T exerted a similar effect at a later time point (48 h), and Mix followed a comparable temporal pattern, ultimately producing a similar overall decrease relative to control (interaction P < 0.05). In the S phase, all three treatments significantly increased the proportion of cells, reflecting enhanced DNA synthesis and accelerated entry into the proliferative phase, with Mix exerting a stronger effect up to 36 h and T showing a more pronounced effect at 48 h (interaction P < 0.05). In the G2/M phase, while the negative control showed no significant changes (P > 0.05), all hormonal treatments significantly increased the fraction of cells at 36 and 48 h (P < 0.05) without significant differences among E2, T, and Mix (P > 0.05), indicating accumulation in late cell cycle stages in preparation for mitosis. Significant interactions between treatment type and incubation time were observed for G0/G1, S, and G2/M phases (P < 0.05), highlighting the combined influence of hormone type and exposure duration on cell cycle dynamics (Fig. 3). Time-dependent effects of β-estradiol (E2), testosterone (T), and their combination [Mix (E2/T)] on cell cycle phase distribution in Huh7 cells. Panels A–C illustrate the percentages of cells in the G0/G1, S, and G2/M phases, respectively, following 24, 36, and 48 h of incubation after hormonal treatments, compared with the negative control (Neg-C). Tukey’s multiple comparisons tests indicate (I) time-dependent differences and (II) treatment-dependent differences for each cell cycle phase of G0/G1, S, and G2/M. (NS, not significant; *P < 0.05; **P < 0.01; ***P < 0.001) Dose-dependent effects of sex hormones on the expression of cell cycle–related genes and miRNAs in HepG2 and Huh7 cells To investigate the dose-dependent effects of sex hormones on molecular regulators of the cell cycle, HepG2 and Huh7 cells were treated for 24 h with increasing concentrations of E2, T, or their combination (Mix). Changes in the expression of Aurora-A, Aurora-B, and GATA3 genes, along with their associated miRNAs (miR-3941, miR-4500, and miR-4742), were quantified relative to the negative control. Hormone-dependent modulation of Aurora-A, Aurora-B and GATA 3 gene expression in HepG2 cells Aurora-A: Treatment with both E2 and T significantly altered Aurora-A mRNA expression across all tested doses compared with the negative control (P < 0.001) (Fig. 4. A). E₂ treatment resulted in a consistent downregulation, with similar average reductions observed across the lowest to the highest doses about ~ 0.41 (~ 59% reduction), indicating a dose-independent inhibitory effect. T, in contrast, exhibited a dose-dependent downregulation, with fold changes of 0.70, 0.59, and 0.39 (approximately 30%, 41%, and 60% reductions) from the lowest to the highest dose, respectively. Comparative analysis at equivalent doses revealed that E₂ exerted a stronger suppressive effect than T only at the lowest dose (fold change ratio E₂/T ≈ 0.58, P < 0.001), while no significant differences were observed at the intermediate and high doses (fold change ratios ≈ 0.7 and 1.05, respectively, P > 0.0001). Combined hormone treatment at the V50 dose also significantly decreased Aurora-A expression relative to the control (fold change ≈ 0.29, representing a 71% reduction; P < 0.001), without showing a significant difference compared with either hormone alone (fold change ratios Mix/E₂ ≈ 0.7 and Mix/T ≈ 0.5, P > 0.05). These transcriptional changes were corroborated at the protein level, as all treatments significantly reduced Aurora-A protein expression (E₂: fold change ≈ 0.49 (51% reduction), T: fold change ≈ 0.53 (47% reduction), Mix: fold change ≈ 0.37 (63% reduction); P < 0.05). Analysis of regulatory miRNAs indicated that compared to negative control, E2 did not significantly alter miR-3941 expression (fold change:1.18, P = 0.53), whereas T induced a significant upregulation of this miRNA (fold change:2.1, P = 0.0001). The combined treatment also significantly increased miR-3941 levels (fold change:1.64, P = 0.0008), suggesting that this effect is largely driven by the T component. Aurora-A: Treatment with both E2 and T significantly altered Aurora-A mRNA expression across all tested doses compared with the negative control (P < 0.001) (Fig. 4. A). E₂ treatment resulted in a consistent downregulation, with similar average reductions observed across the lowest to the highest doses about ~ 0.41 (~ 59% reduction), indicating a dose-independent inhibitory effect. T, in contrast, exhibited a dose-dependent downregulation, with fold changes of 0.70, 0.59, and 0.39 (approximately 30%, 41%, and 60% reductions) from the lowest to the highest dose, respectively. Comparative analysis at equivalent doses revealed that E₂ exerted a stronger suppressive effect than T only at the lowest dose (fold change ratio E₂/T ≈ 0.58, P < 0.001), while no significant differences were observed at the intermediate and high doses (fold change ratios ≈ 0.7 and 1.05, respectively, P > 0.0001). Combined hormone treatment at the V50 dose also significantly decreased Aurora-A expression relative to the control (fold change ≈ 0.29, representing a 71% reduction; P < 0.001), without showing a significant difference compared with either hormone alone (fold change ratios Mix/E₂ ≈ 0.7 and Mix/T ≈ 0.5, P > 0.05). These transcriptional changes were corroborated at the protein level, as all treatments significantly reduced Aurora-A protein expression (E₂: fold change ≈ 0.49 (51% reduction), T: fold change ≈ 0.53 (47% reduction), Mix: fold change ≈ 0.37 (63% reduction); P < 0.05). Analysis of regulatory miRNAs indicated that compared to negative control, E2 did not significantly alter miR-3941 expression (fold change:1.18, P = 0.53), whereas T induced a significant upregulation of this miRNA (fold change:2.1, P = 0.0001). The combined treatment also significantly increased miR-3941 levels (fold change:1.64, P = 0.0008), suggesting that this effect is largely driven by the T component. Aurora-B: The suppressive effect of E2 on Aurora-A mRNA expression was largely dose-independent, with fold changes of 0.55, 0.44 and 0.45 observed across the lowest to the highest doses, respectively (approximately, 45%, 56%, and 55%, respectively; P > 0.05). Whereas T again displayed a dose-dependent response with fold changes of 0.84, 0.45, and 0.463 (approximately 16%, 55.4%, and 53.7% reductions P < 0.0001) from the lowest to the highest dose, respectively. Comparative analysis demonstrated that E₂ exerted a stronger inhibitory effect than T only at the lowest dose (fold change ratio E₂/T ≈ 0.65, P < 0.0001), with no significant differences between the hormones at higher doses (fold change ratios ≈ 0.98 and 0.97, respectively, P > 0.05). Combined treatment at the V50 dose also significantly reduced Aurora-B expression compared with the control (fold change ≈ 0.50, representing a ~ 50% reduction; P < 0.0001), without differing from the individual hormone treatments (fold change ratios Mix/E₂ ≈ 1.14 and Mix/T ≈ 1.11, P > 0.05). Consistent with the mRNA data, Aurora-B protein levels were significantly reduced following all hormonal treatments (E₂: fold change ≈ 0.62 (38% reduction), T: fold change ≈ 0.52 (48% reduction), Mix: fold change ≈ 0.47 (52.8% reduction); P < 0.05). In line with these findings, expression of miR-4500, a predicted regulator of Aurora-B, was significantly increased in response to E2, T, and Mix treatments compared with the negative control (fold change: 1.45, 1.7, and 1.3, respectively; P < 0.01), with no significant differences observed among the treatment groups (fold change ratios all close to 1, P > 0.05). Aurora-B: The suppressive effect of E2 on Aurora-A mRNA expression was largely dose-independent, with fold changes of 0.55, 0.44 and 0.45 observed across the lowest to the highest doses, respectively (approximately, 45%, 56%, and 55%, respectively; P > 0.05). Whereas T again displayed a dose-dependent response with fold changes of 0.84, 0.45, and 0.463 (approximately 16%, 55.4%, and 53.7% reductions P < 0.0001) from the lowest to the highest dose, respectively. Comparative analysis demonstrated that E₂ exerted a stronger inhibitory effect than T only at the lowest dose (fold change ratio E₂/T ≈ 0.65, P < 0.0001), with no significant differences between the hormones at higher doses (fold change ratios ≈ 0.98 and 0.97, respectively, P > 0.05). Combined treatment at the V50 dose also significantly reduced Aurora-B expression compared with the control (fold change ≈ 0.50, representing a ~ 50% reduction; P < 0.0001), without differing from the individual hormone treatments (fold change ratios Mix/E₂ ≈ 1.14 and Mix/T ≈ 1.11, P > 0.05). Consistent with the mRNA data, Aurora-B protein levels were significantly reduced following all hormonal treatments (E₂: fold change ≈ 0.62 (38% reduction), T: fold change ≈ 0.52 (48% reduction), Mix: fold change ≈ 0.47 (52.8% reduction); P < 0.05). In line with these findings, expression of miR-4500, a predicted regulator of Aurora-B, was significantly increased in response to E2, T, and Mix treatments compared with the negative control (fold change: 1.45, 1.7, and 1.3, respectively; P < 0.01), with no significant differences observed among the treatment groups (fold change ratios all close to 1, P > 0.05). GATA3: In contrast, hormonal treatments significantly upregulated GATA3 mRNA expression (Fig. 4. C). E₂ induced a marked increase in GATA3 expression at all tested doses relative to the control (fold changes from the lowest to the highest dose: 19.03, 22.9, and 21.6, respectively; P < 0.001), without evidence of dose dependency. T also significantly increased GATA3 expression in a dose-dependent manner (fold changes from the lowest to the highest dose: 4.8, 9.6, and 14.4, respectively; P < 0.0001); however, the magnitude of induction by E₂ was significantly greater than that observed with T (P < 0.001). Combined treatment at the V50 dose significantly increased GATA3 expression compared with the negative control (fold change: 22.5, P < 0.001) and exhibited an effect comparable to E₂ (fold change ratio Mix/E₂ ≈ 0.98, P > 0.05) but stronger than T alone (fold change ratio Mix/T ≈ 2.3, P < 0.0001). These transcriptional changes were further confirmed at the protein level, where all treatments significantly elevated GATA3 protein expression (E₂: fold change ≈ 2.4, T: fold change ≈ 1.9, Mix: fold change ≈ 2.07; P < 0.05). Finally, analysis of miR-4742, a predicted post-transcriptional regulator of GATA3, revealed a significant reduction in its expression following E2 and combined treatments (E₂: fold change ≈ 0.585 (41.5% reduction), Mix: fold change ≈ 0.64 (36% reduction); P < 0.001), whereas the decrease induced by T alone was less pronounced (fold change ≈ 0.82, representing an 18% reduction; P = 0.12). Notably, the suppressive effect of E₂ on miR-4742 expression was more pronounced than that of T (fold change ratio E₂/T ≈ 0.71), consistent with the observed upregulation of GATA3. GATA3: In contrast, hormonal treatments significantly upregulated GATA3 mRNA expression (Fig. 4. C). E₂ induced a marked increase in GATA3 expression at all tested doses relative to the control (fold changes from the lowest to the highest dose: 19.03, 22.9, and 21.6, respectively; P < 0.001), without evidence of dose dependency. T also significantly increased GATA3 expression in a dose-dependent manner (fold changes from the lowest to the highest dose: 4.8, 9.6, and 14.4, respectively; P < 0.0001); however, the magnitude of induction by E₂ was significantly greater than that observed with T (P < 0.001). Combined treatment at the V50 dose significantly increased GATA3 expression compared with the negative control (fold change: 22.5, P < 0.001) and exhibited an effect comparable to E₂ (fold change ratio Mix/E₂ ≈ 0.98, P > 0.05) but stronger than T alone (fold change ratio Mix/T ≈ 2.3, P < 0.0001). These transcriptional changes were further confirmed at the protein level, where all treatments significantly elevated GATA3 protein expression (E₂: fold change ≈ 2.4, T: fold change ≈ 1.9, Mix: fold change ≈ 2.07; P < 0.05). Finally, analysis of miR-4742, a predicted post-transcriptional regulator of GATA3, revealed a significant reduction in its expression following E2 and combined treatments (E₂: fold change ≈ 0.585 (41.5% reduction), Mix: fold change ≈ 0.64 (36% reduction); P < 0.001), whereas the decrease induced by T alone was less pronounced (fold change ≈ 0.82, representing an 18% reduction; P = 0.12). Notably, the suppressive effect of E₂ on miR-4742 expression was more pronounced than that of T (fold change ratio E₂/T ≈ 0.71), consistent with the observed upregulation of GATA3. Impact of β-estradiol (E2), testosterone (T), and combined (Mix) treatment at varying doses and V50-equivalent concentration on mRNA and protein expression of (A) Aurora-A, (B) Aurora-B, and (C) GATA3, as well as miR-3941, miR-4500, and miR-4742 levels in HepG2 cells relative to negative control, analyzed using one-way ANOVA. (NS, not significant; *P < 0.05; **P < 0.01; ***P < 0.001) Hormone-dependent modulation of Aurora-A, Aurora-B and GATA 3 gene expression in Huh7 cells Aurora-A: Treatment of Huh7 cells with sex hormones induced a significant dose-dependent modulation of Aurora-A expression (Fig. 5A). Both E2 and T significantly reduced Aurora-A mRNA levels at all tested doses compared with the negative control (P < 0.001). Notably, lower doses exerted a stronger inhibitory effect than intermediate and high doses for both hormones. Specifically, E2 showed fold changes of 0.44, 0.82, and 0.80 (approximately 56%, 18%, and 20% reductions, respectively), while T exhibited fold changes of 0.24, 0.35, and 0.45 (approximately 76%, 65%, and 55% reductions, respectively) from the lowest to the highest dose. At equivalent doses, T consistently suppressed Aurora-A expression more effectively than E2 (fold change ratios E2/T ≈ 1.8, 2.3, and 1.8 at low, mid, and high doses, respectively; P < 0.001). Combined hormone treatment at the V50 dose also significantly decreased Aurora-A expression relative to the control (fold change ≈ 0.44, representing a 56% reduction; P < 0.001). This effect appeared predominantly driven by T (Mix/T ratio ≈ 1.26, P = 0.35), as the contribution of E2 was not significant (Mix/E₂ ratio ≈ 0.54, P < 0.0001). Consistent with the mRNA data, Aurora-A protein levels were significantly reduced following T and combined treatments (T: fold change ≈ 0.35 (65% reduction); Mix: fold change ≈ 0.44 (56% reduction); P < 0.01). In contrast, E2 alone had no significant effect on protein levels (fold change ≈ 0.82, representing an 18% reduction; P = 0.67). The reduction induced by T was significantly greater than that observed with E2 (fold change ratio of E2/T ≈ 2.3, P = 0.041), and the effect of the combined treatment remained largely attributable to T (fold change ratio of Mix/T ≈ 1.25, P > 0.05). Regarding miR-3941 expression, while all hormonal treatments showed an increasing trend relative to the negative control, this increase reached statistical significance only following combined treatment (fold changes: E2 ≈ 1.06, T ≈ 1.19 (both P > 0.05), Mix ≈ 1.34 (P = 0.01)). No significant differences were observed among the treatment groups in their effect on miR-3941 (P > 0.05). Aurora-A: Treatment of Huh7 cells with sex hormones induced a significant dose-dependent modulation of Aurora-A expression (Fig. 5A). Both E2 and T significantly reduced Aurora-A mRNA levels at all tested doses compared with the negative control (P < 0.001). Notably, lower doses exerted a stronger inhibitory effect than intermediate and high doses for both hormones. Specifically, E2 showed fold changes of 0.44, 0.82, and 0.80 (approximately 56%, 18%, and 20% reductions, respectively), while T exhibited fold changes of 0.24, 0.35, and 0.45 (approximately 76%, 65%, and 55% reductions, respectively) from the lowest to the highest dose. At equivalent doses, T consistently suppressed Aurora-A expression more effectively than E2 (fold change ratios E2/T ≈ 1.8, 2.3, and 1.8 at low, mid, and high doses, respectively; P < 0.001). Combined hormone treatment at the V50 dose also significantly decreased Aurora-A expression relative to the control (fold change ≈ 0.44, representing a 56% reduction; P < 0.001). This effect appeared predominantly driven by T (Mix/T ratio ≈ 1.26, P = 0.35), as the contribution of E2 was not significant (Mix/E₂ ratio ≈ 0.54, P < 0.0001). Consistent with the mRNA data, Aurora-A protein levels were significantly reduced following T and combined treatments (T: fold change ≈ 0.35 (65% reduction); Mix: fold change ≈ 0.44 (56% reduction); P < 0.01). In contrast, E2 alone had no significant effect on protein levels (fold change ≈ 0.82, representing an 18% reduction; P = 0.67). The reduction induced by T was significantly greater than that observed with E2 (fold change ratio of E2/T ≈ 2.3, P = 0.041), and the effect of the combined treatment remained largely attributable to T (fold change ratio of Mix/T ≈ 1.25, P > 0.05). Regarding miR-3941 expression, while all hormonal treatments showed an increasing trend relative to the negative control, this increase reached statistical significance only following combined treatment (fold changes: E2 ≈ 1.06, T ≈ 1.19 (both P > 0.05), Mix ≈ 1.34 (P = 0.01)). No significant differences were observed among the treatment groups in their effect on miR-3941 (P > 0.05). Aurora-B: Analysis of Aurora-B expression showed that compared with the negative control, both E2 and T significantly reduced mRNA levels at all tested doses (Fig. 5B). E2 exhibited fold changes of approximately 0.49, 0.87, and 0.89 (P < 0.001), while T showed fold changes of approximately 0.34, 0.49, and 0.74 (P < 0.05 for all doses). Lower concentrations produced a more pronounced inhibitory effect for both hormones. At equivalent doses, T induced a significantly greater suppression of Aurora-B expression than E2 (P < 0.001). Combined treatment at the V50 dose also markedly decreased Aurora-B mRNA expression (fold change: 0.57, P = 0.002). This effect was largely attributable to the T component, as indicated by a non-significant ratio of Mix/T (≈ 1.16, P > 0.05), while the contribution of E2 was significant (Mix/E₂ ratio ≈ 0.65, P < 0.05). At the protein level, both E2 and T significantly reduced Aurora-B expression relative to the negative control (fold changes: E2 ≈ 0.64 (P < 0.05); T ≈ 0.7 (P < 0.05)). However, no significant difference was observed between the effects of E2 and T (fold change ratio of E2/T ≈ 0.91, P = 0.89). Notably, the combined treatment produced a significantly greater reduction than either hormone alone, indicating a synergistic effect (Mix fold change ≈ 0.3, P = 0.002). This synergy was confirmed by comparing the combined treatment to individual hormones: Mix/T ratio ≈ 0.42 (P = 0.02) and Mix/E₂ ratio ≈ 0.47 (P = 0.03), signifying a significantly stronger effect of the combination compared to T or E2 individually. Assessment of miR-4500 revealed that E2 and the combined treatment significantly increased miRNA expression compared with the negative control (E2: fold change ≈ 1.5, P < 0.05; Mix: fold change ≈ 1.97, P < 0.05). T alone had a minimal, non-significant effect (fold change: T ≈ 1.13, P > 0.05). The upregulation observed following combined treatment was significantly greater than that induced by either hormone individually. Specifically, the Mix/E₂ ratio was approximately 1.3 (P < 0.05), and the Mix/T ratio was approximately 1.7 (P < 0.05), suggesting that the combination’s effect on miR-4500 was significantly amplified compared to E2 alone and also compared to the non-significant effect of T alone. Aurora-B: Analysis of Aurora-B expression showed that compared with the negative control, both E2 and T significantly reduced mRNA levels at all tested doses (Fig. 5B). E2 exhibited fold changes of approximately 0.49, 0.87, and 0.89 (P < 0.001), while T showed fold changes of approximately 0.34, 0.49, and 0.74 (P < 0.05 for all doses). Lower concentrations produced a more pronounced inhibitory effect for both hormones. At equivalent doses, T induced a significantly greater suppression of Aurora-B expression than E2 (P < 0.001). Combined treatment at the V50 dose also markedly decreased Aurora-B mRNA expression (fold change: 0.57, P = 0.002). This effect was largely attributable to the T component, as indicated by a non-significant ratio of Mix/T (≈ 1.16, P > 0.05), while the contribution of E2 was significant (Mix/E₂ ratio ≈ 0.65, P < 0.05). At the protein level, both E2 and T significantly reduced Aurora-B expression relative to the negative control (fold changes: E2 ≈ 0.64 (P < 0.05); T ≈ 0.7 (P < 0.05)). However, no significant difference was observed between the effects of E2 and T (fold change ratio of E2/T ≈ 0.91, P = 0.89). Notably, the combined treatment produced a significantly greater reduction than either hormone alone, indicating a synergistic effect (Mix fold change ≈ 0.3, P = 0.002). This synergy was confirmed by comparing the combined treatment to individual hormones: Mix/T ratio ≈ 0.42 (P = 0.02) and Mix/E₂ ratio ≈ 0.47 (P = 0.03), signifying a significantly stronger effect of the combination compared to T or E2 individually. Assessment of miR-4500 revealed that E2 and the combined treatment significantly increased miRNA expression compared with the negative control (E2: fold change ≈ 1.5, P < 0.05; Mix: fold change ≈ 1.97, P < 0.05). T alone had a minimal, non-significant effect (fold change: T ≈ 1.13, P > 0.05). The upregulation observed following combined treatment was significantly greater than that induced by either hormone individually. Specifically, the Mix/E₂ ratio was approximately 1.3 (P < 0.05), and the Mix/T ratio was approximately 1.7 (P < 0.05), suggesting that the combination’s effect on miR-4500 was significantly amplified compared to E2 alone and also compared to the non-significant effect of T alone. Impact of β-estradiol (E2), testosterone (T), and combined (Mix) treatment at varying doses and V50-equivalent concentration on mRNA and protein expression of (A) Aurora-A, (B) Aurora-B, and (C) GATA3, as well as miR-3941, miR-4500, and miR-4742 levels in Huh7 cells relative to negative control, analyzed using one-way ANOVA. (NS, not significant; *P < 0.05; **P < 0.01; ***P < 0.001) GATA3: Both hormonal treatments significantly upregulated GATA3 mRNA expression at all tested doses compared with the negative control (P < 0.001) (Fig. 5C). Although not significant, the lowest dose of E2 induced a slightly stronger upregulation than the highest dose (fold changes: ≈ 5.8 versus ≈ 4.1, respectively, P = 0.2; ~ 26% difference). In contrast, testosterone treatment showed a significant elevation in GATA3 mRNA levels at the highest dose compared to the lowest dose (fold changes: ≈ 5.7 versus ≈ 8.3, P = 0.02; ~ 0.46% increase). Combined treatment at the V50 dose also significantly increased GATA3 mRNA expression relative to the negative control (fold change: ≈ 5.5, P < 0.0001). However, this effect did not exceed that of either hormone alone, as indicated by the ratios of combined treatment to individual hormone effects (Mix/E2 ratio ≈ 1.08, P = 0.99; Mix/T ratio ≈ 1.06, P = 0.99). At the protein level, both hormonal treatments significantly increased GATA3 expression compared with the negative control (fold changes: E2 ≈ 2.1 (P = 0.17), T ≈ 2.3 (P = 0.008)). While E2 and T did not differ significantly from each other (fold change ratio: E2/T ≈ 0.9, P = 0.64), the combined treatment induced a significantly greater increase in GATA3 protein levels (fold change: ≈ 3.3, P = 0.0009) than either hormone alone (Mix/E2 ratio ≈ 1.6, P = 0.009 ~ 38%; Mix/T ratio ≈ 1.43, P = 0.018, ~ 31%). Evaluation of miR-4742 revealed that all three treatments significantly reduced miRNA expression relative to the negative control (for both E2 and T ≈ 0.46, P = 0.001, ~ 54% reduction; Mix ≈ 0.44, P = 0.0008, ~ 58% reduction). No significant differences were observed among the treatment groups (P > 0.05). Time- and hormone-dependent modulation of gene expression in HepG2 and Huh7 cells To evaluate the effects of treatment type and incubation time on the expression of cell cycle-related genes, HepG2 and Huh7 cells were exposed to E2, T, and their combination (Mix) for 24, 36, and 48 h. Gene expression data were analyzed using two-way ANOVA, with the 24-hour negative control (NC) serving as the reference. The analysis assessed the main effects of treatment, incubation time, and their interaction on gene expression. Time-dependent effects of sex hormones on Aurora-A, Aurora-B, and GATA3 expression in HepG2 cells In HepG2 cells, all treatments significantly reduced Aurora-A (Fig. 6A) and Aurora-B (Fig. 6B) expression after 24 h compared to the negative control (P < 0.001). For Aurora-A, E2 and T exhibited similar inhibitory effects (~ 65% and ~ 55% inhibition, respectively, P > 0.05), and their combination (Mix) did not significantly differ from either hormone alone (~ 68% inhibition, P > 0.05). With incubation extended to 36 h, Aurora-A expression in the negative control showed a notable decrease (~ 50% reduction relative to NC, P = 0.004), which was further amplified by hormonal treatments (~ 75% and ~ 77% inhibition, respectively, relative to NC, both P < 0.001). T had a stronger inhibitory effect than E2 (P = 0.002), and Mix demonstrated a cumulative inhibitory effect (~ 82% inhibition relative to NC, P = 0.001). By 48 h, no significant change was observed in the negative control (P > 0.05), while all treatments continued to reduce Aurora-A expression (E2: ~86%, T: ~85%, and Mix: 86% inhibition relative to NC; all P < 0.001), with no significant differences among the treatments (P > 0.05). Aurora-B expression also decreased following 24-hour treatment, with the Mix treatment showing a weaker effect than individual components, particularly E2 (E2: ~56%, T: ~55%, Mix: 48% inhibition relative to NC; all P < 0.0001). Over time to 36 and then 48 h, the negative control continued to show a downward trend (approximately 30–40% reduction relative to baseline), and hormonal treatments amplified this decrease (E2: 65–85%, and T: 62–86% relative to NC; all P < 0.001), although the combined treatment remained weaker than individual components (Mix: 54–60% inhibition, P = 0.1–0.05). Regarding GATA3 (Fig. 6C), all treatments induced a significant increase after 24 h (fold changes: E2 ≈ 23, T ≈ 10, Mix ≈ 22.5, relative to NC; all P < 0.001), with Mix effects largely attributed to E2. Over 36 and 48 h, negative controls showed no significant changes (P > 0.05), whereas all treatments continued to increase GATA3 expression (fold changes: E2 ≈ 60–68 (P = 0.0001-0.1), T ≈ 46–49 (P = 0.0001-0.5), and Mix ≈ 60–80 (P = 0.0001–0.004), relative to NC). Time-dependent effects (24, 36, and 48 h) of β-estradiol (E2), testosterone (T), and combined (Mix) on (A) Aurora-A, (B) Aurora-B, and (C) GATA3 mRNA expression in HepG2 cells, analyzed using two-way ANOVA. (NS, not significant; *P < 0.05; **P < 0.01; ***P < 0.001) Over 36 and 48 h, the negative control showed no significant changes (P > 0.05), but all treatments continued to increase GATA3 expression (fold changes: E2 ≈ 60–68 (P = 0.0001-0.1), T ≈ 46–49 (P = 0.0001-0.5), and Mix ≈ 60–80 (P = 0.0001–0.004), relative to NC). The effect of the combined treatment (Mix) appeared to be primarily driven by the E2 component. Time-dependent effects of sex hormones on Aurora-A, Aurora-B, and GATA3 expression in Huh7 cells In Huh7 cells, similar patterns to those observed in HepG2 cells were noted, with some notable differences. After 24 h, treatments decreased Aurora-A expression (Fig. 7. A). E2 exhibited a stronger inhibitory effect than T (65% vs. 55% inhibition, respectively; P = 0.009), while the Mix treatment (68% inhibition; P < 0.0001) did not significantly alter E2’s effect (only a 3% increase; P = 0.085). As incubation time progressed to 36 h and then 48 h, the negative control exhibited a marked decrease in Aurora-A expression (45–47% reduction relative to untreated cells, P = 0.004–0.08). This decrease was further amplified by all hormonal treatments (E2: ~75–86% inhibition; P = 0.01-<0.0001; T: ~78–85% inhibition; P < 0.0001; Mix: 83–84% inhibition; P = 0.005-0.1), all relative to the negative control (for all P < 0.0001). For Aurora-B (Fig. 7. B), all treatments reduced expression at 24 h, with T showing a stronger inhibitory effect than E2 (fold changes relative to NC: E2 ≈ 0.85 (P = 0.01), T ≈ 0.45 (P < 0.0001)). The effect of the Mix treatment (fold change ≈ 0.57 relative to NC, P < 0.0001) was primarily attributable to its T component. With an increase in incubation time to 36 h, the negative control showed significant upregulation (fold change: 1.3, P = 0.003). E2, T, and Mix treatments significantly attenuated this increase (fold changes relative to NC: E2 ≈ 0.92 (P = 0.52), T ≈ 0.87 (P < 0.0001)). While E2 and T showed similar effects in attenuating this increase (P > 0.05), no significant differences were observed among the treatments overall in their attenuation of the negative control’s increase (P > 0.05). As incubation time extended to 48 h, gene expression in the negative control continued to increase significantly (fold change: 1.6, P = 0.001), and hormonal treatments maintained their inhibitory effect on this rise (fold changes relative to NC: E2 ≈ 0.95 (P = 0.7), T ≈ 0.7 (P = 0.18)). In this regard, the effects of E2 and T were similar (P > 0.05), but the Mix treatment demonstrated a synergistic effect compared to each component alone (P < 0.05). For GATA3 (Fig. 7. C), all treatments increased mRNA expression at 24 h (P < 0.001), with the effect of Mix largely driven by T (fold changes relative to NC: E2 ≈ 2.1, T ≈ 5.2, Mix ≈ 5.4; all P < 0.001). At 36 and 48 h, negative controls exhibited a decreasing trend (80–86% reduction relative to initial levels, P < 0.0001–0.005). However, all hormonal treatments maintained significant upregulation of GATA3 (fold changes relative to NC: E2 ≈ 2.6–4.8 (P = 0.03 − 0.005), T ≈ 5.3–4.8 (P = 0.9 − 0.01), and Mix ≈ 5.7–5.6 (P = 0.5–0.8)). The effect of the combination therapy (Mix) appeared to be primarily driven by the T component. Time-dependent effects (24, 36, and 48 h) of β-estradiol (E2), testosterone (T), and combined (Mix) on (A) Aurora-A, (B) Aurora-B, and (C) GATA3 mRNA expression in Huh7 cells, analyzed using two-way ANOVA. (NS, not significant; *P < 0.05; **P < 0.01; ***P < 0.001) Multivariate regression analysis of cell cycle–related gene expression in HepG2 and Huh7 cells Multivariate linear regression was performed to assess the independent effects of Aurora-A, Aurora-B, and GATA3 on cell cycle phase distribution, adjusting for hormonal treatment (E2, testosterone, Mix) and incubation time (36 and 48 h). As shown in Fig. 8, in HepG2 cells, both Aurora-A and Aurora-B displayed significant phase-specific effects: Aurora-A was positively associated with G0/G1 (β = 36.7, P = 0.024) and negatively with S phase (β = −38.1, P = 0.036), while its effect on G2/M was not significant (P = 0.11). Aurora-B similarly promoted G0/G1 (β = 51.2, P = 0.025) and reduced S phase (β = −53.6, P = 0.034) without affecting G2/M (P = 0.15). GATA3 showed no independent effect on any phase (G0/G1: P = 0.11; S: P = 0.18; G2/M: P = 0.47), indicating that Aurora kinases are the primary regulators of phase-specific progression, particularly the G0/G1–S transition, whereas GATA3 does not independently influence cell cycle distribution. Multivariate regression plots showing the associations between Aurora-A (A–C), Aurora-B (D–F), and GATA3 (G–I) expression levels and the percentages of cells in the G0/G1, S, and G2/M phases, respectively, in HepG2 cells In contrast, Huh7 cells (Fig. 9) showed no significant independent effects of Aurora-A, Aurora-B, or GATA3 on any phase (P > 0.1), although the models explained a considerable portion of variability (R² = 0.36–0.90). Here, incubation time was the dominant factor, with 36-hour exposure significantly affecting G2/M (P = 0.0047), while hormonal treatments had limited and inconsistent effects. These results indicate that in Huh7 cells, cell cycle regulation is largely time-dependent, and Aurora kinases or GATA3 do not exert independent phase-specific control, highlighting the cell line–specific context of molecular regulation. Multivariate regression plots showing the associations between Aurora-A (A–C), Aurora-B (D–F), and GATA3 (G–I) expression levels and the percentages of cells in the G0/G1, S, and G2/M phases, respectively, in Huh7 cells. Discussion HCC exhibits marked sexual dimorphism in incidence and progression, suggesting a pivotal role for sex hormones in modulating tumor behavior [21]. It is pertinent to highlight that this study differentiates itself from prior epidemiological investigations of HCC’s sexual dimorphism. While many studies compare incidence and hormonal effects between sexes, often reporting opposing roles for sex hormones (e.g., inhibitory for estradiol, stimulatory for testosterone) [22–24], our research focused specifically on male-derived HCC cell lines (HepG2 and Huh7). This allowed us to examine hormonal impacts within a male context naturally enriched with androgen receptors and potentially lower estrogen receptor expression, characteristics of male HCC. Consequently, we observed largely similar directional effects for both β-estradiol and testosterone within each cell line, a finding that contrasts with some previous reports of opposing hormonal actions [21, 25]. The results collectively suggest that the biological outcome of hormonal signaling may depend not solely on the hormone itself, but rather on the intrinsic molecular landscape of each tumor cell, including its reported p53 status, metabolic phenotype, and dominance of downstream signaling pathways. Perhaps if a female HCC cell line, typically with high estrogen receptor expression levels, had been used in this study, we would have seen stronger effects of β-estradiol. Our in vitro approach further aimed to dissect these hormonal influences. Specifically, we treated cells with testosterone alone to mimic male hormonal influence, with a combination of testosterone and β-estradiol to investigate potential synergistic or counteracting effects against a backdrop of known anti-proliferative roles for estradiol, and with β-estradiol alone as a control for its individual impact. A key innovation of this work lies in the comparative analysis of two phenotypically and functionally distinct cell models: HepG2, a more differentiated, less aggressive line with wild-type p53, and Huh7, an aggressive line harboring a p53 mutation (Y220C). This parallel examination revealed paradoxical responses, with hormonal treatment inducing G0/G1 arrest in HepG2 cells but accelerating G1/S phase progression in Huh7 cells. These divergent outcomes align with the known influence of p53 status on cellular sensitivity to various treatments, thereby offering valuable insights into potential context-dependent clinical applications. A key observation of this study is the strikingly opposite cell cycle responses elicited by identical hormonal treatments in HepG2 and Huh7 cells. In HepG2, hormone exposure consistently induced a dual cell cycle arrest characterized by accumulation of cells in G0/G1 and G2/M phases, accompanied by a reduction in S phase population. Such a pattern is compatible with preserved checkpoint control associated with functional p53, as documented for HepG2 in prior literature [26]. In contrast, Huh7 cells displayed a pro-proliferative response, with reduced G0/G1 and increased S and G2/M fractions, reflecting accelerated DNA synthesis and mitotic entry. Given the known p53 deficiency of Huh7 cells [27], one plausible explanation is that impaired checkpoint regulation facilitates the observed pro-growth response [28]. Importantly, within each cell line, β-estradiol and testosterone exerted similar directional effects, suggesting—rather than establishing—that the dominant signaling network may shape the cellular outcome more strongly than the hormonal class. Aurora kinases, as central regulators of G2/M transition, mitotic spindle assembly, and chromosomal segregation, were expected to parallel the observed cell cycle phenotypes [29]. In HepG2 cells, downregulation of Aurora-A and Aurora-B at both mRNA and protein levels aligned with cell cycle arrest, which may reflect functional mediation of hormone dependent growth restraint. However, in Huh7 cells, despite a clear reduction in Aurora-A/B expression following hormonal treatment, cell cycle progression remained enhanced. This observation is consistent with the notion—supported by prior reports—that loss of p53 function can lead to rewiring of cell cycle control in aggressive HCC lines [30–32]. One layer of this discrepancy could be explained by post-transcriptional regulation mediated by microRNAs. The significant upregulation of miR 3941 and miR 4500 in HepG2 cells, concomitant with reduced Aurora protein levels, may suggest an active miRNA dependent silencing mechanism that reinforces mitotic suppression [33, 34]. In contrast, Huh7 cells exhibited limited miRNA responsiveness, with detectable changes only under combined hormonal treatment, which could imply a less effective miRNA regulatory network. This diminished post-transcriptional control might attenuate the functional impact of Aurora downregulation in Huh7 cells [35]. Beyond post-transcriptional mechanisms, activation of alternative and rapid signaling pathways appears to play a dominant role in sustaining proliferation in Huh7 cells. Pathways such as PI3K/AKT and MAPK/ERK, which have been reported in several previous studies [36, 37], are frequently hyperactivated in HCC and particularly enriched in glycolytic, poorly differentiated tumors. These pathways may bypass canonical mitotic checkpoints [38] and drive S-phase entry and mitotic progression independently of Aurora kinase abundance. The persistence of cell cycle advancement despite reduced Aurora expression in Huh7 cells strongly supports the notion that these cells rely on compensatory proliferative circuits rather than classical mitotic regulators [39]. Crosstalk between estrogen and androgen receptor signaling is also likely to contribute to the context specific hormonal responses observed [40, 41]. In HepG2 cells, higher ERα expression and a more differentiated phenotype- as described in previous characterization studies- may favor protective estrogenic signaling, enabling even testosterone to converge toward growth inhibitory pathways through receptor interaction, metabolic conversion, or non-genomic signaling convergence [2, 42]. Such plausible interactions illustrate how hormone receptor balance, rather than absolute hormone levels, shapes downstream cellular behavior. Conversely, in Huh7 cells, reported elevation of androgen receptor expression combined with diminished or altered ER signaling could bias hormonal inputs toward pro-proliferative outcomes [43, 44]. Such plausible interactions illustrate how hormone receptor balance, rather than absolute hormone levels, shapes downstream cellular behavior. The role of GATA3 provides additional insight into the hypothetical dependency of tumor-suppressive mechanisms on cellular context. Although hormonal treatments increased GATA3 expression in both cell lines, this upregulation translated into growth restraint predominantly in HepG2 cells. In Huh7 cells, where p53 function is compromised [45], GATA3 may become functionally uncoupled from its classical anti-proliferative role [46]. This interpretation should be considered provisional, indicating that effective engagement of differentiation-associated transcription factors probably requires an intact checkpoint and tumor suppressor network. This suggests that effective engagement of differentiation-associated transcription factors requires an intact checkpoint and tumor suppressor network, reinforcing p53’s central role as a molecular gatekeeper [47, 48]. Overall, while direct assessment of p53 functionality and receptor signaling was beyond the current study’s scope, our data—combined with existing molecular characterizations of these cell lines—support a conceptual model in which the tumor’s molecular identity profoundly influences hormone driven regulation of cell cycle progression in HCC [48, 49]. In differentiated, p53 competent cells, hormonal signals appear to act cohesively with Aurora kinases and miRNAs to restrain proliferation [50]. In contrast, in p53-deficient, metabolically reprogrammed cells, hormone signaling might be rerouted through alternative pathways, rendering classical regulators functionally dispensable [46]. However, these interpretations should be viewed as hypothesis-generating rather than causal. This context dependency provides a framework for understanding the paradoxical responses of sex hormones in HCC and underscores the necessity of stratifying tumors based on molecular features when considering hormone-related therapeutic strategies. Conclusions The results of this study indicate that the effect of sex hormones β-estradiol and testosterone on cell cycle regulation and biological behavior of HCC is highly dependent on the molecular context and the differentiation status of tumor cells. Comparison between the HepG2 and Huh7 cell lines suggests that differences in the functional status of p53, metabolic patterns, and dominant signaling pathways likely determine the cells’ response to hormonal treatments. In the HepG2 cell line, characterized by a more differentiated phenotype and active p53, hormonal treatments were predominantly associated with the inhibition of cell cycle progression. This response correlated with decreased Aurora kinases expression, increased GATA3, and activation of miRNA-dependent regulatory mechanisms. This suggests that in this cellular context, sex hormones can exert antiproliferative and genome-stabilizing roles through efficient regulatory networks. In contrast, the Huh7 cell line, with a more aggressive phenotype and impaired p53 function, exhibited differential effects of hormones on the cell cycle. Despite a reduction in Aurora kinases expression, cell cycle progression and entry into proliferative phases continued. This finding implies that in this context, cell cycle control may be less dependent on canonical mitotic regulators and more influenced by alternative proliferative signaling pathways, complex interactions of hormone receptors, and the high plasticity of tumor regulatory networks. However, further comprehensive studies in more complex models are necessary to validate these hypotheses and fully elucidate the underlying mechanisms. Acknowledgements The authors gratefully acknowledge the Iran University of Medical Sciences for their financial support and valuable guidance throughout this project (Grant No. 28805). Author contributions T.R. and E.B. designed and supervised the project. T.R., F.F. and M.A. performed all the experiments. E.B. analyzed the data. E.B. and Y.M. wrote the main manuscript. All authors reviewed and approved the manuscript. Y.M. submitted the manuscript. Funding This work was supported by a research grant from the Deputy of Research and Technology, Iran University of Medical Sciences (Grant No. 28805). Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval This study was approved by Ethics Committee of Iran University of Medical Sciences (IR.IUMS.FMD.REC.1403.119). All cell culture practices adhered to international biosafety guidelines for in vitro studies. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.