Differential Modulation of Galectin-3 by Tamoxifen and Progesterone Receptor Modulators: An In Vitro and In Silico Study

Abstract

Background/Aim

Tamoxifen remains one of the most widely used endocrine treatments for hormone-receptor-positive breast cancer. However, its partial estrogen-agonist activity in the endometrium is associated with endometrial proliferation, hyperplasia, polyps, and, rarely, carcinoma. Antiprogestins and selective progesterone receptor modulators, including mifepristone and ulipristal acetate, have been proposed as alternative or adjunctive endocrine strategies, but their effects on endometrial molecular remodeling remain insufficiently characterized. This study investigates whether Galectin-3, a multifunctional lectin that is involved in epithelial integrity, angiogenesis, receptor signaling, and tumor-associated remodeling, could serve as a candidate marker of the endometrial epithelial response to endocrine modulation.

Materials and Methods

Ishikawa endometrial adenocarcinoma cells were exposed to different concentrations of tamoxifen, mifepristone, and ulipristal acetate. Galectin-3 protein expression was evaluated using Western blot analysis, and densitometric measurements were normalized to β-actin. Complementary STRING and STITCH bioinformatics analyses were performed to examine the molecular interactions linking Galectin-3 with steroid receptor and growth-factor-associated signaling pathways.

Results

Tamoxifen produced a biphasic response, characterized by modest induction of Galectin-3 expression at low concentration and marked suppression at intermediate and high concentrations. Mifepristone preserved Galectin-3 expression at low and intermediate concentrations but reduced its expression at the highest concentration. Ulipristal acetate increased Galectin-3 expression at low-to-intermediate concentrations, followed by a return toward baseline at high concentration. STRING and STITCH analyses indicated that Galectin-3 is indirectly linked to steroid receptor signaling through growth-factor-associated pathways, particularly EGFR/ERBB- and VEGF-related networks.

Conclusion

Galectin-3 expression differs according to the type and concentration of endocrine modulator used, suggesting that it may reflect distinct endometrial epithelial responses to hormonal treatment. Galectin-3 therefore warrants further investigation as a potential biomarker for assessing endometrial safety during endocrine therapy.

Introduction

Hormonal therapy is a complementary approach to hormone-dependent breast cancer and is administered continuously for up to 10 years after the completion of surgical and adjuvant therapy (1). To date, tamoxifen is the most widely used hormonal therapy with the best clinical results in terms of breast cancer recurrence (1). However, tamoxifen, as a selective estrogen receptor modulator (SERM), effectively antagonizes estrogen receptor signaling in the breast but exerts partial agonist activity in the endometrium. This dual action is associated with endometrial proliferation, polyps, hyperplasia, and an increased risk of endometrial malignancy, with recent systematic-review evidence also highlighting gynecological risks in premenopausal and perimenopausal women (2-4). The endometrial effects of tamoxifen have prompted investigation of alternative endocrine strategies that attempt to normalize these phenomena without losing the effectiveness of their action.

Antiprogestins such as mifepristone (RU-486) (5, 6) and ulipristal acetate (CDB-2914) (7-10) have been proposed as alternatives or adjuncts, targeting progesterone receptor-positive tumors. Progesterone became a major point of interest in breast tumorigenesis because breast cancer biology is not driven only by estrogen signaling (11). Progesterone receptor (PGR) signaling regulates mammary epithelial proliferation, lobuloalveolar development, stem/progenitor cell expansion (12, 13), and paracrine pathways such as receptor activator of nuclear factor kappa-B ligand (RANKL) and wingless/integrated signaling pathway (WNT) signaling (14). Although progesterone receptor (PR) expression is often estrogen-regulated and is used clinically together with estrogen receptor (ER) and human epidermal growth factor receptor 2 (HER2) status to classify breast cancers, PR is not merely a passive marker of ER activity. PR can actively modify estrogen receptor chromatin binding and transcriptional programs, thereby influencing tumor growth, endocrine responsive-ness, and potentially treatment resistance. For this reason, progesterone signaling is increasingly viewed as a biologically active pathway in hormone-receptor-positive breast cancer and as a potential direct therapeutic target (15).

Although antiprogestins are being considered as direct hormonal treatments for PR-positive breast cancer, their long-term endometrial safety remains unclear. This is important because tamoxifen, despite its anti-estrogenic action in breast tissue, can act as a partial estrogen agonist in the endometrium and is associated with endometrial proliferation, polyps, hyperplasia, and rarely carcinoma. Progesterone normally opposes estro-gen-driven endometrial proliferation; therefore, blocking progesterone signaling could theoretically disturb endometrial homeostasis. However, antiprogestin-related endometrial effects are not necessarily identical to classical hyperplasia, as selective progesterone receptor modulators can induce distinct progesterone-receptor-modulator-associated endometrial changes (16). Thus, the endometrial impact of antiprogestins depends on dose, duration, estrogenic background, and compound-specific activity, and requires dedicated safety evaluation.

Accordingly, biomarkers capable of detecting endometrial molecular changes warrant investigation that will diagnose endometrial transformation in a timely manner, as well as to propose the mechanisms that link the proposed alternative antiprogestin therapies with the potential biomarkers. Galectin-3 (LGALS3) is a lectin that regulates cell adhesion, migration, angio-genesis, and tumor progression (17). It forms oligomers on the cell surface that crosslink receptors such as vascular endothelial growth factor receptor (VEGFR) and epidermal growth factor receptor (EGFR) (18), modulating proliferative and migratory responses. Its expression in the endometrium varies with hormonal milieu, and altered galectin-3 levels have been linked to hyperplasia and neoplastic progression (17). Based on the above, we hypothesized that galectin-3 is a sensitive molecular biomarker of endometrial remodeling under endocrine modulators.

Here we present an integrated analysis combining (i) cell-based studies in Ishikawa cells exposed to tamoxifen, mifepristone, and ulipristal acetate, and (ii) bioinformatic modeling of galectin-3 interactions with hormone and growth factor pathways.

Materials and Methods

Ishikawa cellular cultures. The experimental setting was carried out using Ishikawa cell lines. Ishikawa cells were selected as a reproducible hormone-responsive endometrial epithelial model appropriate for preliminary screening of direct antiprogestin effects on epithelial molecular markers, despite the fact that they are carcinoma-derived and not comparable to normal endometrial epithelium (19). Ishikawa human endometrial epithelial cells were obtained from European Collection of Authenticated Cell Cultures (ECACC)/Sigma-Aldrich/Merck, Darmstadt, Germany (Cat. No. 99040201-1VL) and used between passages 5 and 25. Although they are useful for researching progesterone-dependent stromal decidualization, non-malignant stromal models like telomerase-immortalized human endometrial stromal cells (T-HESC) do not accurately reflect the epithelial compartment that the current study’s goals are aimed at (20). As a result, the current results should be considered hypothesis-generating and need to be verified in endometrial models based on organoids or non-malignant epithelial cells.

Cell culture and drug treatments were performed according to previously published international protocols (21), with minor modifications adapted to the experimental aims of the present study (22). Ishikawa human endometrial epithelial cells were cultured at 37°C in a humidified 5% CO₂ atmosphere in phenol-red-free Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F-12) (Sigma-Aldrich/Merck, cat. No, D6434) supplemented with 10% charcoal/dextran-stripped fetal bovine serum (Sigma-Aldrich/Merck, cat. No, F6765), 1% L-glutamine (Sigma-Aldrich/Merck, cat. No, G7513), and 1% penicillin/streptomycin (Sigma-Aldrich/Merck, cat. No P0781). Experiments were performed using cells between passages 5 and 25, and cells were subcultured at 70-80% confluence using 0.05% trypsin-Ethylenediaminetetraacetic Acid (EDTA) or TrypLE (Sigma-Aldrich/Merck, cat. No, T2601).

For Western blot experiments, cells were seeded in 6-well plates at 1.5-2.0 × 10⁵ cells/well in 2 ml medium to reach 60-70% confluence at treatment, providing sufficient protein for loading 20-30 μg per lane. Twenty-four hours before treatment, medium was replaced with 2% charcoal-stripped FBS medium to reduce background hormonal stimulation. Cultures exceeding 90% confluence were excluded.

Drug exposures were performed at the same time of day. Each experiment included vehicle controls, hormone-stimulated controls, and endocrine modulator-treated wells. Dimethyl sulfoxide (DMSO) (Sigma-Aldrich/Merck, cat. No D2650) was used as vehicle at a final concentration ≤0.1% in all drug-treated conditions. Tamoxifen, mifepristone, and ulipristal acetate were tested at the concentrations specified below.

Tamoxifen (Sigma-Aldrich/Merck, cat. No T5648), mifepristone (Sigma-Aldrich/Merck, cat. M8046), and ulipristal acetate (Sigma-Aldrich/Merck, cat. No SML 1469) were tested over compound-specific concentration ranges selected to evaluate concentration-dependent changes in Galectin-3 expression. The ranges were informed by previously published in-vitro studies and available human pharmacokinetic data. Mifepristone was tested at 0.1, 1, and 10 μM, and ulipristal acetate at 0.05, 0.4, and 4 μM. These ranges were based in part on previous experiments in Ishikawa cells using mifepristone at 0.1, 1, and 10 μM and ulipristal acetate at 0.04, 0.4, and 4 μM (23). In that study, approximately 1.96 μM was used as a pharmacokinetic reference for mifepristone following a 25-mg oral dose (23, 24), whereas approximately 0.4 μM was used as a reference for ulipristal acetate following a 30-mg oral dose (23, 25). Tamoxifen was tested at 0.01, 0.1, and 1 μM. Reported steady-state plasma concentrations of parent tamoxifen in patients receiving 20 mg/day are generally within the submicromolar range, with one clinical study reporting a mean concentration of 378±176 nM (26). Tamoxifen and its metabolites may also accumulate in endometrial tissue at concentrations higher than those measured in serum (27). However, plasma concentrations cannot be directly extrapolated to drug concentrations within the endometrial epithelium. Accordingly, the concentrations used in the present study should be regarded as pharmacologically informed experimental exposures intended to characterize concentration-response relationships rather than as direct equivalents of clinically achieved endometrial concentrations. Cells were treated for 24 h. Treatment positions were randomized between experiments, and drug stocks were prepared at 10 mM in DMSO or ethanol, stored at −20°C protected from light, and freshly diluted before use.

Western blot for Galectin-3 expression. Following treatment, cells were washed twice with ice-cold phosphate-buffered saline (PBS) (Sigma-Aldrich/Merck, cat. No D8537) and lysed on ice using radioimmunoprecipitation assay (RIPA) buffer (Sigma-Aldrich/Merck, cat. No R0278) supplemented with protease (Sigma-Aldrich/Merck, cat. No P8340) and phosphatase inhibitor cocktails (Sigma-Aldrich/Merck, cat. No P5726 and P0044). Lysates were clarified by centrifugation at 12,000×g for 10 min at 4°C, and total protein concentration was determined using a bicinchoninic acid (BCA) assay. Equal amounts of protein, 20-30 μg per lane, were mixed with Laemmli sample buffer (Sigma-Aldrich/Merck, cat. No S3401) containing DTT (Dithiothreitol) (Sigma-Aldrich/Merck, cat. No D9779), denatured at 95°C for 5 min, resolved by SDS-PAGE on 10-12% gels, and transferred to polyvinylidene fluoride (PVDF) membranes (Sigma-Aldrich/Merck, cat. No IPVH08130) (28).

Membranes were blocked with 5% non-fat dry milk in TBS-T for 1 h at room temperature and incubated overnight at 4°C with mouse monoclonal anti-Galectin-3 antibody, clone M3/38, unconjugated (Sigma-Aldrich/Merck, Cat. No. MABT51; dilution 1:1,000). β-actin was used as the loading control and detected using mouse monoclonal anti-β-actin antibody, clone AC-15, unconjugated (Sigma-Aldrich/Merck, Cat. No. A1978; dilution 1:10,000).

After washing three times with Tris-Buffered Saline with Tween 20 (TBS-T), membranes were incubated for 1 h at room temperature with horseradish peroxidase (HRP)-conjugated secondary antibody, clone 1H11, peroxidase conjugate, suitable for western blot (WB), immunofluorescence (IF) and fluorescence-activated cell sorting (FACS) applications (Sigma-Aldrich/Merck, Cat. No. SAB5201891; dilution 1:10,000). Signals were detected using enhanced chemiluminescence (ECL), and densitometric analysis was performed using ImageJ/Fiji and ImageLab software. Galectin-3 expression was normalized to β-actin and expressed as fold-change relative to vehicle-treated control cells (29).

Bioinformatic analyses. To place galectin-3 within larger molecular networks, bioinformatic studies were conducted concurrently with the wet-lab investigations. Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) version 12, which combines computational predictions, curated databases, and experimental data, was used to analyze protein-protein interactions. Because they are important nodes in endometrial signaling and angiogenesis, the focus was on galectin-3, PGR, VEGF, and EGF. In order to reduce the number of spurious relationships, high-confidence thresholds (≥0.7) was applied. Visualization of the resulting networks identified indirect connections between Galectin-3 and progesterone receptor signaling mediated through angiogenic and growth factor pathways.

Tamoxifen, mifepristone, and ulipristal acetate were used to query the Search Tool for Interactions of Chemicals (STITCH) database in order to investigate drug-protein connections. This program provides insights into how small compounds may alter protein networks by combining chemical-protein interactions from prediction and experimental sources. It was feasible to compare predicted drug-protein interactions with noted changes in galectin-3 expression by lining up STITCH outputs with the experimental results.

This work used a dual strategy by integrating computational interaction mapping with controlled cell culture studies. While the bioinformatic analyses presented molecular explanations for these effects within the broader signaling context of the endometrium, the cellular experiments demonstrated directly how endocrine modulators affect galectin-3 expression.

Statistical analysis. Statistical analysis was performed using R software (R Foundation for Statistical Computing, Vienna, Austria). Western blot densitometric data were obtained from three independent experimental runs. For each lane, Galectin-3 band intensity was normalized to the corresponding β-actin band intensity from the same lane, generating a normalized Galectin-3/β-actin ratio. Normalized values were then expressed as fold change relative to untreated control cells, which were set as 1.0 in each independent experiment.

Data are presented as mean±standard deviation (SD). For each endocrine modulator, low, intermediate, and high concentration groups were analyzed separately. Dose-dependent differences in Galectin-3 expression were assessed using one-way ANOVA. When appropriate, post hoc pairwise comparisons were per-formed to identify differences between concentration levels. A p-value <0.05 was considered statistically significant.

The statistical analysis demonstrated significant dose-dependent modulation of Galectin-3 expression for all three compounds: tamoxifen, F(2, 6)=333.91, p<0.001; mifepristone, F(2, 6)=139.01, p<0.001; and ulipristal acetate, F(2, 6)=120.10, p<0.001. These p-values were reported in the Results section and in the corresponding figure legend. The in-silico STRING and STITCH network analyses were interpreted descriptively and were not subjected to inferential statistical testing.

Results

In vitro cellular analysis. Western blot and densitometric analysis were performed to evaluate Galectin-3 protein expression in Ishikawa human endometrial epithelial cells following exposure to tamoxifen, mifepristone, and ulipristal acetate. β-actin was used as the internal loading control, and Galectin-3 band intensity was normalized to the corresponding β-actin signal from the same lane (Figure 1). Normalized values were expressed as fold change relative to untreated control cells, which were set as 1.0. The experiment was performed in three in-dependent Western blot runs, and values are presented as mean±standard deviation (SD).

Figure 1.

Representative western blot analysis of Galectin-3 expression in Ishikawa cells following treatment with tamoxifen, mifepristone, and ulipristal acetate. Galectin-3 (~26 kDa) and β-actin (~42 kDa) protein bands are shown, with β-actin used as the loading control. MW markers are indicated in kDa. Lanes 1-3 represent untreated control cells. Lanes 4-6 represent cells treated with tamoxifen at 0.01, 0.1, and 1 μM, respectively; lanes 7-9 represent cells treated with mifepristone at 0.1, 1, and 10 μM, respectively; and lanes 10-12 represent cells treated with ulipristal acetate at 0.05, 0.4, and 4 μM, respectively. Galectin-3 band intensity was normalized to the corresponding β-actin signal for densitometric analysis. MW: Molecular-weight.

Basal Galectin-3 expression was consistently detectable in untreated Ishikawa cells. The normalized Galectin-3/β-actin ratio in control cells was set as 1.00 in each independent experiment and was used as the reference baseline for comparison with endocrine modulator-treated conditions.

Following tamoxifen exposure, Galectin-3 expression showed a biphasic dose-response pattern. At the lowest concentration tested, tamoxifen produced a modest increase in Galectin-3 expression, with a mean fold change of 1.15±0.05 relative to control. In contrast, intermediate and high tamoxifen concentrations were associated with a clear reduction in Galectin-3 expression. At the intermediate concentration, Galectin-3 decreased to 0.54±0.04-fold of control, while at the high concentration it decreased further to 0.32±0.03-fold of control. Thus, high-dose tamoxifen reduced Galectin-3 expression to approximately one-third of basal levels.

Mifepristone produced a more gradual expression profile. At low and intermediate concentrations, Galectin-3 expression remained close to control levels, with mean fold changes of 0.99±0.04 and 0.93±0.03, respectively. This indicates relative preservation of Galectin-3 expression under low and intermediate mifepristone exposure. At the highest concentration, however, Galectin-3 expression decreased to 0.51±0.04-fold of control, corresponding to an approximately 50% reduction. This pattern suggests that mifepristone affected Galectin-3 expression primarily at higher exposure levels.

Ulipristal acetate showed a distinct non-linear response. At low concentration, Galectin-3 expression increased to 1.36±0.04-fold relative to control, corresponding to an approximately 36% increase above baseline. At intermediate concentration, Galectin-3 expression increased further to 1.46±0.04-fold, corresponding to an approximately 46% increase above baseline. At the highest concentration, this increase was no longer observed, and Galectin-3 expression returned close to control levels, with a mean fold change of 1.02±0.04. This pattern is compatible with a bell-shaped response, with increased Galectin-3 expression at low-to-intermediate ulipristal acetate concentrations and normalization at higher exposure.

Direct comparison of the three agents showed distinct Galectin-3 modulation profiles. Statistical analysis of normalized Galectin-3/β-actin fold-change values demonstrated significant dose-dependent modulation for all three compounds: tamoxifen, F(2, 6)=333.91, p<0.001; mifepristone, F(2, 6)=139.01, p<0.001; and ulipristal acetate, F(2, 6)=120.10, p<0.001 (Figure 2). Tamoxifen produced the strongest suppressive effect at intermediate and high concentrations, with the high-dose condition reducing Galectin-3 to approximately one-third of control levels. Mifepristone preserved Galectin-3 expression at low and intermediate concentrations but reduced expression at the highest concentration. Ulipristal acetate uniquely increased Galectin-3 expression at low and intermediate concentrations, followed by return toward baseline at high concentration (Table I).

Figure 2.

Dose-response modulation of Galectin-3 expression in Ishikawa cells following treatment with tamoxifen, mifepristone, and ulipristal acetate. Values are expressed as fold change relative to untreated control cells and represent mean±SD from three independent Western blot experiments. Error bars indicate SD. Galectin-3 expression was quantified by densitometric analysis and normalized to β-actin for each lane before calculation of fold change. SD, Standard deviation.

Table I. Values represent normalized Galectin-3/β-actin ratios expressed as fold change relative to untreated control cells. Data are derived from three independent western blot experiments and are presented as mean±standard deviation (SD).

In silico analysis. Protein-protein and drug-protein interaction networks were constructed using STRING and STITCH databases to identify signaling pathways associated with galectin-3 (LGALS3) in the context of endocrine modulation. High-confidence interaction thresholds (confidence score ≥0.7) were applied to minimize spurious associations. Network topology analysis revealed a modular organization with clearly defined signaling clusters inter-connected through steroid hormone receptors and growth factor intermediates.

Within the network, estrogen receptor α (ESR1) and progesterone receptor (PGR) emerged as high-degree nodes, exhibiting extensive connectivity with transcriptional regulators, signaling intermediates, and cell cycle-associated proteins. Degree and between-ness centrality measures indicated that ESR1, in particular, functioned as a major network hub linking multiple signaling modules. PGR displayed similar, though slightly less ex-tensive, connectivity patterns. Galectin-3 itself showed lower direct connectivity but was positioned downstream of multiple highly connected nodes, suggesting an indirect regulatory role (Figure 3).

Figure 3.

Integrated protein-protein interaction network illustrating endocrine, growth factor, developmental, and cell-cycle signaling modules associated with galectin-3. Nodes represent proteins and edges represent known or predicted functional interactions. The network highlights interconnected endocrine, growth-factor, developmental, and cell-cycle modules. The green cluster mainly includes steroid receptor and cell-cycle-associated proteins, the blue cluster includes growth-factor and receptor tyrosine kinase pathways, and the red cluster includes WNT/β-catenin and NOTCH-related developmental signaling. Galectin-3 appears indirectly connected to endocrine signaling through growth-factor and developmental intermediates, supporting its potential role as an integrative marker of endometrial remodeling. WNT, Wingless/Integrated signaling pathway; NOTCH, Notch signaling pathway.

Direct physical interactions between galectin-3 and ESR1 or PGR were not detected in the high-confidence network. Instead, galectin-3 was consistently connected to steroid receptor signaling through intermediary growth factor pathways, most prominently epidermal growth factor (EGF/EGFR) and vascular endothelial growth factor (VEGF/VEGFR) signaling. These intermediary nodes formed dense interaction clusters with both hormone receptor-associated proteins and downstream signaling cascades, indicating that galec-tin-3 may be influenced by changes in signaling flux through these pathways rather than by direct receptor-mediated transcriptional control.

The network revealed a tightly interconnected growth factor signaling module centered on EGFR, Erb-B2 receptor tyrosine kinase 2 (ERBB2), and VEGF receptors. This module exhibited strong connectivity to intracellular signaling cascades, including the phosphoinositide 3-kinase-protein kinase B (PI3K-AKT) and mitogen-activated protein kinase signaling pathway (MAPK) pathways. Multiple nodes within these cascades showed shared connections to ESR1- and PGR-associated proteins, indicating convergence of endocrine and growth factor signaling at the level of intracellular signal transduction. Galectin-3 was linked to this module through interactions associated with receptor trafficking, signal amplification, and pathway stabilization (Figure 4).

Figure 4.

The network highlights the close connectivity between EGFR, ERBB2, ERBB3, and ERBB4 with adaptor and signaling proteins including GRB2, SHC1, SRC, and HSP90AA1. EGF and NRG1 are shown as upstream growth-factor ligands, while CD44 is included as a cell-surface adhesion/signaling component linked to receptor activation and downstream pathway modulation. The dense interaction pattern supports the role of the EGFR/ERBB module as a major growth-factor signaling hub potentially connecting Galectin-3-associated pathways with epithelial proliferation, survival, and endometrial remodeling. EGFR, Epidermal growth factor receptor; ERBB, erb-b receptor tyrosine kinase family; GRB2, growth factor receptor-bound protein 2; SHC1, SHC adaptor protein 1; SRC, SRC proto-oncogene, non-receptor tyrosine kinase; HSP90AA1, heat shock protein 90 alpha family class A member 1; EGF, epidermal growth factor; NRG1, neuregulin 1; CD44, cluster of differentiation 44.

A distinct signaling module corresponding to the WNT/β-catenin pathway was identified, comprising WNT ligands, Frizzled receptors, and downstream transcriptional mediators. Although topologically separate from the growth factor module, this cluster maintained indirect connections to ESR1- and PGR-associated nodes through shared regulatory proteins. Additionally, a NOTCH signaling cluster, involving NOTCH receptors and their ligands, was detected and displayed connectivity patterns consistent with roles in epithelial differentiation and tissue organization. Both pathways were linked indirectly to galectin-3 through shared intermediates rather than direct interactions.

Proteins involved in cell cycle regulation and genome stability, including cyclin D1 (CCND1), myelocytomatosis gene (MYC), tumor protein p53 (TP53), and breast cancer gene 1 (BRCA1), formed an interconnected module that bridged hormone receptor signaling with growth factor and developmental pathways. These nodes showed multiple interactions with ESR1- and PGR-associated proteins as well as with downstream signaling cascades, indicating integration of proliferative control mechanisms within the broader endocrine-responsive network.

Drug-protein interaction analysis revealed that tamoxifen showed strong and extensive connectivity with ESR1 and ESR1-associated transcriptional regulators and signaling intermediates. Downstream of ESR1, tamoxifen-associated nodes were linked to growth factor signaling pathways, including EGFR-related and VEGF-related components, as well as to intracellular signaling cascades such as PI3K-AKT and MAPK. Galectin-3 was indirectly connected to these tamoxifen-associated pathways through nodes involved in receptor trafficking, signal propagation, and transcriptional regulation. The network structure indicated that tamoxifen engages multiple signaling layers capable of influencing galectin-3-associated pathways.

In contrast, mifepristone displayed a more restricted interaction pattern. Its primary connectivity was centered on PGR and a limited subset of PGR-associated regulatory proteins. Compared to tamoxifen, fewer downstream signaling intermediates were engaged, and connectivity with growth factor signaling pathways was reduced. Indirect links between mifepristone and galectin-3 were present but involved fewer intermediary nodes and lower network density. This interaction profile suggests a more confined signaling footprint relative to tamoxifen.

Ulipristal acetate exhibited an intermediate interaction pattern between tamoxifen and mifepristone. The drug demonstrated connectivity with PGR as well as with selected downstream transcriptional regulators and signaling intermediates. Unlike mifepristone, ulipristal acetate showed broader indirect associations with growth factor pathways, including EGFR- and VEGF-related nodes, though less extensive than those observed for tamoxifen. Galectin-3 was indirectly connected to ulipristal-associated pathways through shared growth factor and intracellular signaling components, indicating engagement of signaling modules relevant to proliferation and tissue remodeling (Figure 5).

Figure 5.

Drug-protein and protein-protein interaction network illustrating endocrine modulation of steroid receptor signaling. ESR1 and PGR appear as central endocrine signaling nodes, connected with transcriptional regulators and coactivators including NCOA1, NCOA2, NCOA3, EP300, CREBBP, and SP1. Additional signaling and regulatory proteins, including SRC, BRCA1, and NR3C1, link steroid receptor activity with broader transcriptional and cellular response pathways. The network illustrates that tamoxifen, mifepristone, and ulipristal acetate have distinct interaction profiles, supporting the hypothesis that these agents may differentially influence downstream endometrial signaling responses. ESR1, Estrogen receptor 1; PGR, progesterone receptor; NCOA, nuclear receptor coactivator; EP300, E1A binding protein p300; CREBBP, CREB-binding protein; SP1, Sp1 transcription factor; SRC, SRC protooncogene, non-receptor tyrosine kinase; BRCA1, BRCA1 DNA repair associated; NR3C1, nuclear receptor subfamily 3 group C member 1.

Discussion

Galectin-3 is biologically well suited as a candidate marker of endometrial remodeling because it participates in epithelial cohesion, cell-matrix adhesion, receptor trafficking, and angiogenic signaling. Previous research has shown that galectin-3 regulates VEGFR2 cell-surface expression and activation, while also interacting with integrin mediated adhesion pathways (30). The functions galectin-3 exhibits are important for the endometrium, due to its role in maintaining epithelial integrity and regulating both stromal and epithelial interactions, as well as vascular remodeling which are all hormone de-pendent (31). Thus, altered galectin-3 levels would likely indicate some level of alteration in the organization of the epithelium and growth factor signaling.

Our findings suggest that Galectin-3 may serve as a candidate molecular marker of endometrial responses to endocrine modulators (32), according to the current study’s accumulating experimental and computational data. We show that galectin-3 is located at the intersection of progesterone receptor signaling and angiogenic growth factor pathways (33), and that tamoxifen, mifepristone, and ulipristal acetate differentially regulate its expression. We accomplish this by combining wet-lab experiments in Ishikawa cells with bioinformatic analyses of protein and drug interaction networks.

The concentration-dependent suppression of Galectin-3 by tamoxifen represents one of the main findings of the present study. Within the tested concentration range, the lowest concentration (0.01 μM) produced a modest increase in Galectin-3 expression, whereas 0.1 and 1 μM resulted in progressive suppression, with the highest tested concentration reducing Galectin-3 expression to approximately one-third of baseline. Although reported steady-state plasma concentrations of parent tamoxifen during standard 20-mg/day therapy are generally submicromolar (26), tamoxifen and its metabolites may accumulate in endometrial tissue (27). Nevertheless, direct correspondence between the concentrations used in vitro and those experienced by endometrial epithelial cells in vivo cannot be assumed.

Mifepristone displayed a different concentration-response profile. Galectin-3 expression remained close to control levels at 0.1 and 1 μM, whereas a marked reduction was observed at the highest tested concentration of 10 μM. Previous pharmacokinetic data indicate that systemic mifepristone exposure can reach the low-micromolar range, and approximately 1.96 μM has been reported following a 25-mg oral dose (23, 24). Thus, the lower and intermediate concentrations used in the present study overlap with pharmacologically plausible systemic exposures, whereas the 10-μM condition was included primarily to characterize the higher end of the experimental concentration-response range (34-39, Table II). Because plasma exposure does not necessarily reflect endometrial tissue concentration, these conditions are not classified here as ‘therapeutic’ or ‘supratherapeutic’.

Table II. Participating studies.

Ulipristal acetate demonstrated a non-linear concentration-response pattern, with increased Galectin-3 expression at 0.05 and 0.4 μM and a return toward baseline at 4 μM. The intermediate concentration of approximately 0.4 μM is comparable to reported peak systemic exposure following a 30-mg oral dose of ulipristal acetate (23, 25). However, the 4-μM concentration represents a higher experimental exposure included to characterize the concentration-response relationship and should not be interpreted as equivalent to routine clinical tissue exposure. Accordingly, the present findings are described in terms of the tested experimental concentrations rather than as effects occurring at ‘therapeutically relevant’ concentrations.

Combining the in vitro and in silico results provides evidence supporting a model where galectin-3 acts as a molecular reporter of the balance of signals present in the endometrium. The down-regulation of galectin-3 by both tamoxifen and high doses of mifepristone is consistent with an environment where there is disrupted communication between the stroma and epithelium that is similar to what has been observed during the transformation to hyperplasia. The up-regulation of galectin-3 by ulipristal at low to moderate concentrations and by the physiological levels of estradiol-progesterone found in normal reference conditions are indicative of maintaining normal tissue structure. Galec-tin-3 is thus potentially a highly sensitive biomarker for the effect of endocrine-modulating agents on the endometrium and could be used as a marker for the assessment of safe treatment effects (Figure 6).

Figure 6.

In vitro vs. in silico findings: Galectin-3 as biomarker of endometrial response to endocrine modulators.

The bioinformatics analyses support and extend these cellular observations. STRING interaction mapping showed that galectin-3 is indirectly linked to progesterone receptor signaling through VEGF and EGF pathways. Galectin-3’s capacity to stabilize VEGFR2 at the cell surface and regulate EGFR trafficking underscores its role as an integrator of angiogenesis and epithelial migration. Endometrial remodeling is critically dependent on these processes, and disruption of galectin-3 could therefore have profound consequences for tissue homeostasis. STITCH analysis further suggested that tamoxifen, mifepristone, and ulipristal exert distinct influences on these networks, with tamoxifen strongly associated with suppression of VEGF and EGFR pathways, mifepristone showing weaker predicted effects, and ulipristal displaying mixed interactions consistent with its biphasic cellular response.

Combining the in vitro and in silico results provides evidence supporting a model where galectin-3 acts as a molecular reporter of the balance of signals present in the endometrium. The down-regulation of galectin-3 by both tamoxifen and high doses of mifepristone is consistent with an environment where there is disrupted communication between the stroma and epithelium that is similar to what has been observed during the transformation to hyperplasia. The up-regulation of galectin-3 by ulipristal at low to moderate concentrations and by the physiological levels of estradiol-progesterone found in normal reference conditions are indicative of maintaining normal tissue structure. Galectin-3 is thus potentially a highly sensitive biomarker for the effect of endocrine-modulating agents on the endometrium and could be used as a marker for the assessment of safe treatment effects (Figure 6).

Taken together, these findings position galectin-3 as more than a passive marker of endometrial state. Instead, it appears to be an active participant in the signaling balance that determines whether the endometrium remains in a physiological state or progresses toward pathological remodeling. Given the established roles of galectin-3 in cell adhesion and angiogenic signaling, its reduced expression may reflect alterations in pathways associated with pathological endometrial remodeling, whereas maintained or increased expression may reflect a molecular profile more consistent with preserved epithelial homeostasis.

Clinically, this work suggests that monitoring galectin-3 could provide valuable in-formation for assessing endometrial safety in patients receiving long-term endocrine therapy for breast cancer. In women treated with tamoxifen, declining galectin-3 expression may serve as an early molecular indicator of endometrial risk, prompting closer surveil-lance or consideration of alternative therapies. In contrast, the relative preservation of galectin-3 under mifepristone and its induction under ulipristal acetate suggest that these agents may offer safer profiles for the endometrium, though this requires validation in clinical tissue studies.

Study limitations. This study used a single carcinoma-derived Ishikawa cell model, which does not fully reproduce the biology of normal endometrial epithelium or stromal-epithelial interactions. The experiments were based on three independent western-blot runs and a single molecular readout, and western-blot densitometry provides only semi-quantitative assessment of Galectin-3 expression. In addition, cell viability was not formally evaluated across the tested drug concentrations; therefore, particularly at higher concentrations, nonspecific cytotoxic effects cannot be excluded. The clinical relevance of the tested concentrations and the use of Galectin-3 as a marker of endometrial safety require validation in non-malignant epithelial models, organoids, co-cultures and clinical endometrial tissue.

The STRING and STITCH analyses identify known or predicted molecular associations but do not establish direct signaling effects or causal relationships. Consequently, integration of the in vitro and in silico findings should be considered hypothesis-generating, and the present results cannot establish that Galectin-3 predicts clinical endometrial safety or that one endocrine modulator is safer than another. Additional functional and translational studies are required before Galectin-3 can be considered a clinically useful biomarker.

The tested concentrations were selected as pharmacologically informed experimental ranges and were not based on direct measurements of drug concentrations within human endometrial tissue. Plasma pharmacokinetic values cannot be directly translated into concentrations experienced by cultured endometrial cells because tissue distribution, protein binding, metabolism, active metabolites, and duration of exposure differ between in vitro and in vivo conditions. Therefore, the clinical relevance of the observed concentration-response relationships requires validation in non-malignant endometrial models and clinical endometrial tissue.

Future studies should extend these findings beyond the carcinoma-derived Ishikawa model by incorporating non-malignant epithelial cells, stromal-epithelial co-cultures, and human endometrial organoids. Recent advances in endometrial organoid technology have generated hormonally responsive models that more closely reproduce key epithelial and stromal characteristics of the native endometrium (40), while newer organoid systems can model dynamic epithelial transitions across an in vitro menstrual cycle and can be benchmarked against human endometrial tissue (41).

Conclusion

Galectin-3 expression was differentially modulated by tamoxifen, mifepristone, and ulipristal acetate in Ishikawa endometrial adenocarcinoma cells following short-term exposure. These findings suggest that Galectin-3 may serve as a candidate molecular marker of endometrial responses to endocrine modulation. However, the present in vitro and in silico data do not establish endometrial safety, comparative clinical safety between endocrine modulators, or the clinical utility of Galectin-3 as a biomarker. Further validation in non-malignant endometrial models, organoids, stromal-epithelial co-cultures, in vivo systems, and clinical endometrial tissue is required.

Conflicts of Interest

The Authors declare that there are no conflicts of interest in relation to this study.

Authors’ Contributions

Conceptualization: AK, SNK, AM. Data Curation: AK, DK, GA, EZ, IB. Formal Analysis: AK, TSD, GA, IB. Funding Acquisition: IB, SNK. Investigation: AK, TSD, GA, EZ, IB. Methodology: GA, EZ, TSD. Project Administration: IB, SNK, TSD. Resources: IB, EZ, AK. Software: AK, DK, IB. Supervision: SNK, AM. Validation: TSD, GA, EZ, SNK, IB. Visualization: AK, GA. Writing – Original Draft: AK, GA, EZ, IB. Writing – Review & Editing: TSD, SNK, AM. All Authors reviewed/approved the final manuscript and accept accountability for the work.

Artificial Intelligence (AI) Disclosure

During the preparation of this manuscript, a large language model (ChatGPT 5.6, 2026, OpenAI) was used solely for language editing and stylistic improvements in select paragraphs. No sections involving the generation, analysis, or interpretation of research data were produced by generative AI. All scientific content was created and verified by the authors. Furthermore, no figures or visual data were generated or modified using generative AI or machine learning–based image enhancement tools.