Distinctive GABA A receptor subunit expression modulates cell specific EMT and functional responses in glioblastoma breast and ovarian cancer

Abstract

Given the emerging understanding of neurotransmitter involvement in cancer biology, GABA receptors have garnered attention for their diverse and sometimes controversial roles across cancer types. Hence, this study aimed to investigate the expression patterns of GABA-A receptors in glioblastoma, breast, and ovarian cancer cells and their impact on cellular responses via an agonist‒antagonist approach. Cell proliferation and cytotoxicity were assessed using MTT and AO/PI assays, respectively. GABA treatment significantly influenced proliferation, stimulating it in ovarian A2780CP cells overexpressing GABRG3 and inhibiting it in U87 glioblastoma cells, which showed increased expression of GABRR3. Furthermore, GABA reduced CD133+ and CD44+ stem-like populations in A2780CP cells and decreased the CD44+ fraction in MDA-MB231 cells, correlating with specific GABA-A receptor subunit expression. Migration assays revealed that GABA significantly reduced the motility of MDA-MB231 and MCF-7 cells, possibly through modulation of GABRR2 expression. EMT-related transcription factors and markers were evaluated using qPCR, flow cytometry, and Western blot analysis. Protein-level changes in EMT markers confirmed the transcriptional data, with GABA modulating E-cadherin and Vimentin expression in a cell-specific manner. These findings underscore the critical role of GABA-A receptor subtypes in promoting or suppressing cancer progression through context-dependent regulation of proliferation, stemness, migration, and EMT.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-18894-y.

Introduction

Throughout the centuries, from ancient Egyptian descriptions of tumors to groundbreaking discoveries in modern oncology, our understanding of cancer has undergone significant advancements¹. Traditional approaches such as surgery, chemotherapy, and radiation therapy have significantly improved many patients’ survival rates and quality of life². Although promising, these advancements are not universally effective; however, cancer is regarded as one of the leading causes of death in both developed and developing countries³, and the persistent challenges of drug resistance, metastasis, and tumor recurrence remind us of the challenging nature of cancer cells, highlighting the need for further research and personalized treatment strategies⁴.

Despite ongoing debates on cancer stem cells (CSCs), tumor-initiating cells (TICs), phenotypic plasticity⁵, and other factors, such as the heterogeneity of cancer cells⁶, their ability to adapt and the intricate nature of the tumor microenvironment contribute to the elusive nature of this formidable process. While our understanding of CSCs has increased, the quest for effective treatments continues. Given that more than 90% of cancer-related deaths result from metastasis of the primary tumor to other parts of the body, it has been suggested that CSCs are the primary source of metastasis⁷.

Epithelial‒mesenchymal transition (EMT) is vital in embryonic development, tissue repair, and cancer progression. During metastasis, cancer stem cells acquire characteristics that permit them to migrate and infiltrate other tissues. Invaders and metastases can occur due to cellular and microenvironmental changes caused by EMT, which is one of several hallmarks of cancer EMT-related transcription factors, including Snail, Slug, Twist, ZEB1, and ZEB2, that play crucial roles in orchestrating this transition. By controlling the expression of genes associated with cell adhesion, migration, and invasion, these transcription factors contribute to profound changes in cancer cells. EMT-related transcription factors influence various signaling pathways, such as TGF-β, Wnt/β-catenin, Notch, and receptor tyrosine kinase (RTK) signaling. Activation of these pathways triggers the expression of EMT-inducing transcription factors, initiating downstream alterations associated with EMT. The consequences of EMT in cancer cells are significantly associated with poor prognosis⁸. EMT and cancer cell metabolism are interconnected⁹.

The discovery of γ-aminobutyric acid (GABA) and its receptors outside the nervous system in the 1990s opened avenues for investigating their roles beyond traditional neurotransmitter function¹⁰. The emergence of neurotransmitters in cancer research has highlighted their intricate relationship with tumor biology, which impacts tumor growth, immune responses, angiogenesis, and metastasis¹¹. Despite their traditional association with nervous system functions, neurotransmitters such as GABA play pivotal roles in shaping the tumor microenvironment and influencing patient outcomes¹². GABA exerts its effects through two main types of receptors, i.e., ionotropic GABA-A and metabotropic GABA-B. GABA-A receptors are a family of ligand-gated ion channel heteropentamers formed from a selection of 19 subunits with significant physiological and therapeutic implications, including six α (alpha1–6), three β (beta1–3), three γ (gamma1–3), three ρ (rho1–3), and one each of the δ (delta), ε (epsilon), π (pi), and θ (theta) subunits. Nevertheless, the most abundant form has a stoichiometry of 2α, 2β, and 1γ subunit, typically arranged as β–α–β–α–γ to form a functional chloride channel. All isoforms of GABA-A receptors channel chloride movement, but the distribution of the expression of the subunits in different regions and the relationships between the expression pattern and several different diseases and malfunctions have been reported^13–15. A potential paracrine/autocrine mode of action has been described for GABA and its receptors in nonneuronal tissues¹⁶. In cancer biology, the role of GABA and its receptors has garnered significant attention, including as potential targets for novel treatments involving different agonist/antagonist agents¹⁷, prognostic factors¹⁸, and indicators of therapeutic response¹⁹. However, findings concerning the impacts of GABA on cancer cells have been controversial, ranging from stimulatory to inhibitory, which can be attributed to the roles of different subtypes of receptors^20,21, which are linked with different signaling pathways^22–24 and interplay between tumors and the tumor microenvironment²⁵. Thus, a significant scientific gap in the understanding of the involvement of GABAergic signaling in cancer biology, specifically its relation to cancer stem cells and epithelial‒mesenchymal transition (EMT) signaling, remains, necessitating further investigation.

In this study, our aim was to elucidate the expression profiles of GABA-A receptor subtypes in selected breast and ovarian cancer cell lines, as well as glioblastoma, in comparison with those in human brain tissue. Additionally, we investigated the effects of activating/inhibiting GABAergic signaling on various cellular processes, including proliferation, migration, anchorage-independent growth, the expression of cell surface cancer stem cell markers, and EMT-related transcription factors. The findings of this study are expected to provide valuable insights into the role of GABAergic signaling in cancer progression and may aid in identifying potential candidates for targeted therapies.

Materials and methods

Cell lines and chemicals

Human ovarian A2780CP, glioblastoma (U87, U251), and breast (MCF-7, and MDA-MB231) cancers were obtained from the National Cell Bank of Iran (NCBI) with STD profiling authentication certificates. Cell culture grade chemicals, including Dulbecco’s modified Eagle’s medium (DMEM), GABA, and picrotoxin (PTX), were obtained from Sigma‒Aldrich (Merck KGaA, Darmstadt, Germany). In addition, other chemicals and supplements procured from Biosera Europe (Nuaille, France), cell culture vessels and consumables were procured from SPL Life Sciences (Korea).

Study design

For in vitro cell-based studies, the cell lines were cultured and maintained in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin in a humidified incubator at 37 °C and 5% CO₂. The cell culture vessels were divided into three groups: GABA at 100 µmol/L in the presence or absence of 100 µmol/L PTX and control cells receiving DMEM (CTRL). After written consent was obtained, three surgical human brain tissue samples from anatomically identical areas were obtained from patients without malignancy. All patients underwent emergency surgery for intracranial hemorrhage (IHC) within six hours of injury, during which marginal cortical brain tissues were collected from the site of the injury. The surgeries were done with informed consent from patients or legal representatives and adhered strictly to established medical therapeutic protocols with no modifications made for the current study, thereby upholding ethical and medical principles in accordance to the Declaration of Helsinki. The tissue samples were immediately cut into small pieces and snap-frozen in liquid nitrogen. These brain tissues were used as the baseline for normalizing the GABA-A receptor expression profiles of the studied cancer cells. The research project was evaluated and approved by ethics committee of the Shahid Beheshti University of Medical Sciences (IR.SBMU.MSP.REC1399.603). No animal study was included.

Cell proliferation assay

The proliferation of cells was measured in vitro via colorimetric 3-(4,5-dimethyl thiazol-2-yl)-2 and 5-diphenyl tetrazolium bromide (MTT) assays. Briefly, cells were seeded into 96-well plates (2 × 10³ cells per well), grown overnight, washed in PBS, and incubated for 48 h with 100 µM GABA with or without 100 µM PTX; MTT was then added (10 µl/well) for a period of 4 h. The formazan products were solubilized with DMSO, and the optical density was measured at 570 nm. For poorly adherent A2780CP cells, we adapted the protocol by gently centrifuging the plates before removing the supernatant to avoid cell loss.

Acridine orange/propidium iodide (AO/PI) cytotoxicity assay

To complement the MTT assay and provide a membrane integrity-based measure of cytotoxicity, an Acridine Orange/Propidium Iodide (AO/PI) dual staining assay was performed. Briefly, A2780CP, MCF-7, MDA-MB231, U251, and U87 cells were seeded in black 96-well clear-bottom plates at a density of 8 × 10³ cells/well and allowed to attach overnight. Cells were then treated for 48 h under the same conditions as the MTT assay. After treatment, cells were washed gently with PBS and stained with a freshly prepared AO/PI solution (final concentrations: AO 5 µg/mL, PI 5 µg/mL in PBS). Plates were incubated for 15 min at room temperature in the dark. Excess dye was removed by gentle PBS wash, and fluorescence was immediately measured using the BioTek Cytation 5 Cell Imaging Multi-Mode Reader. Fluorescence signals were recorded using the following settings: AO: Ex 485 nm / Em 528 nm (green channel; total nucleated cells, PI: Ex 535 nm / Em 617 nm (red channel; dead cells).

Image-based quantification was conducted using Absorbance values were collected at 450 nm using a BioTek microplate reader, and data were analyzed with Gen5™ software (version 3.17, BioTek Instruments, Inc., USA, https://www.agilent.com/cs/library/software/public/5320200-FW-REV-AQ-V3_17_17.zip). Live and dead cells were counted based on green/red fluorescence intensity and nuclear morphology. The percentage of dead cells was calculated as: % Cytotoxicity = (PI-positive cells / total AO-stained cells) × 100. All experiments were performed in triplicate wells and repeated in at least three independent experiments.

Gene expression assay

All primers (Table 1) were designed and checked for specificity and thermodynamic properties via Primer Blast²⁶ and Oligo Analysis Software (Molecular Biology Insights, Inc., USA).

Table 1.

Sequence of primers used for qPCR.

Symbol	Forward primer (5–3)	Reverse primer (5–3)
GABRA1	ACCACTGTCTTCACCAGGATTT	TCGGTTACACGCTCTCCCAA
GABRA2	CAAACCTATCTGCCTTGCATCA	GGACAGTTGTTACTCCAAACACA
GABRA3	GCCCGTACAGTCTTTGGTGT	AGACGGCTATGAACCAGTCC
GABRA4	GTGCCTGGCGGTTTGTTTA	CGCAGCCTGTTGTCATAACC
GABRA5	GTCCGACACGGAAATGGAGT	GAGGTTGTTGAGAGGGAGGC
GABRA6	ACTCAGAAAACGTCAGTCGGA	TCAGTGACAGCACCTCCAAA
GABRB1	GTACAAAATCGAGAGAGTCTGGG	CTGGGTTCATTGGTGCTGTGT
GABRB2	GAGTCCACGCCATCTTCAAAAAT	TCAGGATACACAGGTCGATCAGC
GABRB3	CTTGCGGGAGGAAGGCTTTT	CGGGATCGTTCACACTCTGG
GABRD	CATCTCAGAGGCCAACATGGAGTA	GGGTCTCGTTGGTGTGGTTG
GABRE	CTCCAGTCGAGGGTCGAG	TGTTTCCTCAGAGAGGAGCTG
GABRG1	CCTTTTCTTCTGCGGAGTCA	TCTGCCTTATCAACACAGTTTCC
GABRG2	TGCTCTACACCCTAAGGTTGAC	CAAGGGGCAGGAGTGTTCAT
GABRG3	CCTTTACACTTTGAGGCTCACC	AGCTGGAGAAAATCAGCGGG
GABRP	GAGGTCGGCAGAAGTGACAA	TCTGTACGGGTTCTCCACCA
GABRQ	GTAGGAAACGTGCAGGATGG	GGTGGACGTCAAAGAACCGA
GABRR1	GGATGTGCAGGTGGAGAGTT	CGTCCTTCCAGTAGTGCCTC
GABRR2	ATGCCCAAGCCAAGTCACTTAT	GTCTCATGCTGAAGTCGTGC
GABRR3	TGGACCAGACTTCCCTCTCT	ATCCGAGATGAATCGGTGCT
GABBR1	GACAGGGTCATCGACCAACA	ACACGCTCCTCTTTCTCAGC
GABBR2	CCATCGAGCAGATCCGCAA	CTTTTGCGTTGTCGCACTCC
TWIST1	TCGGTCTGGAGGATGGAGG	TCTCTGGAAACAATGACATCTAGG
TWIST2	GCCTCAGCTACGCCTTCTC	ACGGACAGCCCTGGCG
SNAI1	CCAGTGCCTCGACCACTATG	CTGCTGGAAGGTAAACTCTGGA
SNAI2	CATTAGAACTCACACGGGGGAGAA	ATGAGCCCTCAGATTTGACCTG
VIM	GGACCAGCTAACCAACGACA	TCCTCCTGCAATTTCTCCCG
CDH	GGGGTTAAGCACAACAGCAAC	CAAAATCCAAGCCCGTGGTG
B2M	TGTCTTTCAGCAAGGACTGGT	TGCTTACATGTCTCGATCCA

RNA extraction was carried out via TRIzol™ LS reagent (Thermo Fisher Scientific, USA) according to the manufacturer’s protocol. All first-strand cDNA amplification (100 ng) was performed via BioFACT™ 2X RT PreMix (BIOFACT Co., Ltd. North Korea). RNA quality was assessed using spectrophotometric absorbance ratios (A260/280 and A260/230), and agarose gel electrophoresis. Only samples with acceptable purity profiles were used for downstream analysis. All the qPCRs were performed via a StepOnePlus™ Real-Time PCR System (Thermo Fisher Scientific Inc., USA) with high-Rox SYBR-Green master mix (RealQ Plus 2x Master Mix Green, Amplicon AS, Denmark). The cycling protocol was 95 °C for 15 s, followed by 60 °C for 30 s, which was repeated 40 times. For normalization of qPCR data, several candidate housekeeping genes were initially evaluated based on exon-exon junction primer design, amplification efficiency, and consistent expression across treatments. Beta-2-microglobulin (B2M) was selected as the most stable gene across all tested conditions and cell lines, and thus used for normalization in the study. The expression of genes of interest (GOIs) was calculated via the 2 ^-∆∆Ct equation and was converted to the fold change via the following formula:

The amplification efficiency was studied via linear regression of amplification curves via LinReg²⁷.

The calculated expressions, i.e., fold changes, were normalized by dividing by the main expression of each subunit in human brain tissue and were prescribed as the relative or normalized expression.

Western blotting

Cells from the control and GABA-treated experimental groups were collected by scraping and lysed using ice-cold RIPA buffer supplemented with a protease inhibitor cocktail (Sigma-Aldrich, P8340, USA). The lysates were incubated on ice for 15 min, briefly sonicated (5 s), and then returned to ice for an additional 15 min. Cell debris was removed by centrifugation at 15,000 × g for 10 min at 4 °C. Protein concentrations were determined using the Bradford assay (Bio-Rad Laboratories GmbH, Germany).

Equal amounts of total protein (20–30 µg) from each sample were separated by SDS-PAGE and transferred onto nitrocellulose membranes (Macherey-Nagel GmbH, Germany). Membranes were blocked in 3% bovine serum albumin (BSA) in TBS-T (Tris-buffered saline with 0.1% Tween-20) for 1 h at room temperature, followed by overnight incubation at 4 °C with the following primary antibodies:

1. Anti-Vimentin (Mouse monoclonal, 1:500; R&D Systems, MAB2105),

2. Anti-E-Cadherin (Mouse monoclonal, 1:500; R&D Systems, MAB1838), and

3. Anti-GAPDH (Mouse monoclonal, 1:10,000; R&D Systems, MAB5718).

After primary antibody incubation, membranes were washed three times with TBS-T and incubated for 3 h at room temperature with horseradish peroxidase-conjugated secondary antibody (Goat Anti-Mouse IgG-HRP, 1:1000; R&D Systems, HAF007). Detection was carried out using Clarity™ Western ECL Substrate (Bio-Rad Laboratories GmbH, Germany), and protein bands were visualized using a chemiluminescence imaging system.

Densitometric analysis was performed using GelAnalyzer software (version 2010a), and band intensities of E-Cadherin and Vimentin were normalized to GAPDH. All experimental conditions were evaluated using two independent biological replicates, each comprising two technical replicates (n = 4).

Cell surface stem cell marker assay

The statuses of CD133 and CD44, which are ovarian cancer and glioblastoma stem cell markers, and CD24 and CD44, which are breast cancer stem cell markers, were studied via flow cytometry (BD FACS Aria; BD Biosciences, USA). Briefly, the cells were washed once with phosphate-buffered saline (PBS) and then harvested with 0.05% trypsin/0.025% EDTA. The detached cells were resuspended in 200 µl of PBS/2% FBS containing CD133-PE-, CD44-FITC-, and CD24-FITC-conjugated antibodies and incubated for 120 min at 4 °C in the dark. Afterward, the cells were precipitated at 500×g for 10 min and suspended in 200 µl of PBS supplemented with 2% FBS. Unstained and immunoglobulin isotypes were used as controls. Flowing software (version 2.5; Turku Center for Biotechnology, Turku University, Turku, Finland) was used to analyze the results.

Wound-healing assay

The wound healing assay, the scratch test, was conducted by growing a fused cell monolayer in a 6-well plate to create an acellular wound zone in which the monolayer was scratched slightly with a pipette tip. For migration (wound healing) assays, serum levels were reduced to 1% FBS to minimize proliferation and emphasize motility-related effects. An inverted microscope was used to obtain images at 0 h after wound formation to assess the wound healing rate at 24 h and 48 h after scratching the treated or untreated cultured cells. To calculate the healing percentage, the wound areas were measured via Optika Vision Pro software (Optika Srl, Ponteranica BG, Italy). The wound ratio was calculated via the following formula:

Anchorage-independent cell growth assay

The effects of manipulating GABA-A signaling on anchorage-independent growth were investigated via soft agar colony formation assays (SOCFAs). In brief, A2780CP cells were not suitably subjected to a scratch assay and were studied by SOCFA and treated with 100 µM GABA or with 100 µM GABA and 100 µM PTX for 48 h. The cells were subsequently trypsinized and seeded (5000 cells/well) in 0.5% noble agar (Sigma) mixed with 2× DMEM (containing 20% FBS and 2% penicillin/streptomycin solution) at a 1:1 ratio. The upper layer of 5% noble agar containing cells was added to each 6-well plate, which formed the bottom layer of agar (containing a mixture of 0.8% noble agar and 2X DMEM). The plate was transferred to an incubator at 37 °C and 5% CO₂ for 21 days and then stained by adding 200 µl of nitro blue tetrazolium chloride solution (Sigma, Germany) per well for 24 h. The images of the plates were captured via a cell imaging system (BioTek Cytation 5 Cell Imaging Multimode Reader, Agilent, USA). The relative number of colonies in treated cells was adjusted to that in control cells and analyzed via ImageJ software; each experiment was performed in triplicate, and statistical analysis was performed via GraphPad Prism software.

Statistical analyses

The data were analyzed via GraphPad Prism (version 8.00, https://www.graphpad.com). At least three biological replicates were analyzed in duplicate or triplicate via appropriate statistical tests. A significant difference of less than 0.05 (p value < 0.05) was considered significant.

Results

GABA receptor expression profiles in ovarian, breast cancer, and glioblastoma cells

The GABA receptor mRNA expression profiles of the A2780CP, MDA-MB231, MCF-7, U251, and U87 cell lines were investigated and compared with the mRNA expression profiles of noncancerous human brain tissue as a reference. As shown in Fig. 1, GABRG2 and GABRG3 were highly expressed genes in the two breast cancer cell lines. Moreover, MCF-7 cells also expressed relatively high levels of GABA-A3, but MDA-MB-231 cells expressed relatively high levels of GABRR2. The expression of GABRG2 and GABRG3 was upregulated in both glioblastoma cell lines, whereas U251 cells presented increased expression of GABA-A4, and U87 cells presented increased expression of GABRR3. Furthermore, A2780CP cells presented GABRG3 and GABRB2 overexpression. A heatmap revealed that GABA receptor expression shows different expression profile patterns despite the common tissue source of cancer.

Fig. 1.

Heatmap of relative GABA-A receptor gene expression in selected cell types. The expression of each subtype has been divided by expression in the human brain. 100 indicates that the expression is equal to that in the brain, and the blue, purple, green, and red colors indicate downregulated, mildly upregulated, moderately upregulated, and highly upregulated genes, respectively.

Effect of GABA on cell proliferation

A2780CP, MCF-7, MDA-MB231, U251, and U87 cells were treated with GABA with or without PTX for 48 h and subjected to the MTT assay. The results revealed that GABA slightly affected the proliferation of different cancer cells. GABA stimulated the proliferation of A2780CP cells (Fig. 2a, p = 0.007) but inhibited the proliferation of U87 cells (Fig. 2e, p < 0.0001). PTX blocked these effects as a GABA-A receptor antagonist (Fig. 2a,e). Compared with nontreated cells, PTX-treated cells did not significantly respond to the treatments (p > 0.05). GABA neither significantly stimulated nor inhibited the proliferation of MDA-MB231, MCF-7 and U251 cells (Fig. 2b–d; p > 0.05). These findings suggest that the response of cancer cells to GABA is a complex process that varies among specific cancer cells, from promoting cell growth to ceasing it. The antagonizing effect of PTX on the stimulatory or inhibitory effects of GABA on A2780CP and U87 cells proves that GABA elicits these effects by binding with GABA-A receptors. Hence, the possible causes for the observed differences mainly stem from the different GABA-A receptor expression profiles in different cells or downstream processes after GABA-A receptor activation.

Fig. 2.

Effects of GABA and a GABA-A antagonist (PTX) on the proliferation of A2780CP (a), MDA-MB231 (b), MCF-7 (c), U251 (d), and U87 (e) cells were measured via MTT assays. One-way ANOVA with Tukey’s multiple comparisons test was used. The data are expressed as the mean value ± SD. (N = 6. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Effect of GABA on cell toxicity

To validate the findings of the MTT assay and distinguish between metabolic changes and true cytotoxicity, AO/PI dual fluorescence viability staining was performed in the same treatment conditions. As shown in Fig. 3, consistent with MTT results, GABA treatment slightly increased the proportion of live cells in A2780CP, while it significantly increased PI-positive (non-viable) cells in U87 cultures. GABA had no significant cytotoxic effect in MCF-7, MDA-MB231, or U251 cells. Importantly, PTX co-treatment reversed both the GABA-induced proliferation in A2780CP and the cytotoxicity in U87, with live/dead ratios comparable to untreated controls, reinforcing the role of GABA-A receptor interaction. AO/PI staining revealed that GABA reduced PI-positive (dead) cells in A2780CP (p < 0.01), suggesting a mild cytoprotective effect, which was reversed by PTX cotreatment. In U87 cells, GABA significantly increased PI-positive cells (p < 0.0001), indicating enhanced cytotoxicity, also reversed by PTX. No significant changes were observed in MCF-7, MDA-MB-231, or U251 cells (p > 0.05), aligning with MTT results.

Fig. 3.

AO/PI staining analysis of cytotoxicity in A2780CP, MCF-7, MDA-MB-231, U251, and U87 cell lines following treatment with GABA or GABA + PTX. Cytotoxicity is expressed as the ratio of PI-positive cells to total cells. Data represent mean ± SD (n = 5; biological duplicate × methodological triplicate). Statistical analysis was performed using two-way ANOVA followed by Tukey’s multiple comparisons test. **p < 0.01, ***p < 0.0001.

Cell surface stemness markers

Flow cytometry analysis was performed to investigate the effects of GABA, a surrogate indicator of CSC populations in different cell lines, on cell surface stemness markers. As shown in Fig. 4a, the percentages of CD133+ and CD44+ cancer stem-like cells among the GABA-treated A2780CP cells were significantly lower than those among the nontreated controls (p = 0.0004 and p = 0.001, respectively). PTX significantly antagonized the GABA-mediated changes in A2780CP cells (p < 0.0001 and p = 0.0002, respectively), suggesting that GABA-GABA-A receptors mediate the observed effects on A2780CP cells. Conversely, GABA significantly decreased the CD44+ population in MDA-MB231 cells (p = 0.007), which was significantly antagonized by PTX (p = 0.0120) (Fig. 4b). No significant effects of GABA on the expression of cell surface stemness markers were detected in the MCF-7, U251, or U87 cell lines (Fig. 4c–e; p > 0.05). Consequently, the potential reasons behind the observed differences are primarily the distinct GABA-A receptor expression patterns in various cells or subsequent processes following GABA-A receptor activation. Flow cytometry images of the selected cells are shown in the Supplementary File (S1 File).

Fig. 4.

Comparison of cancer stem cell marker expression after GABA treatment with or without PTX in A2780CP (a), MDA-MB231 (b), MCF-7 (c), U251 (d), and U87 (e) cells via flow cytometry. Two-way ANOVA with Tukey’s multiple comparisons test. All the data are presented as the means ± SD. (N = 6, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Wound healing assay

A wound healing assay was performed to investigate the effect of GABA on cancer cell migration. The results revealed that treatment with GABA significantly decreased the migration of MDA-MB231 and MCF-7 cells, which demonstrated significantly decreased healing of the wounded area after 48 h (p = 0.0006 and p = 0.015, respectively). PTX neutralized the effects of GABA on migration, so no significant difference was found between the GABA/PTX-cotreated cells and the untreated cells (p > 0.05) (Fig. 5a,b). No significant effects on the migration of U251 or U87 glioblastoma cells were observed (Fig. 5c,d, p > 0.05). Therefore, GABA appears to affect cancer cell migration differently in different cell lines. The omission of the wound healing assay for A2780CP cells was due to their loose attachment to the cell culture surface, making it challenging to create a well-defined scratch and obtain highly reproducible results. As an alternative, soft agar colony formation was performed for these cells.

Fig. 5.

Effects of GABA treatment with or without PTX on MDA-MB231 (a), MCF-7 (b), U251 (c), and U87 (d) cell migration were evaluated via a wound scratch assay. These cells were incubated for 24 h and 48 h. Two-way ANOVA with Tukey’s multiple comparisons test. All the data are presented as the means ± SDs. (*p < 0.05, **p < 0.01, ***p < 0.001).

Anchorage-independent cell growth

Soft agar colony formation assays were conducted to assess the impact of GABA on the anchorage-independent growth of A2780CP cells (Fig. 6a). The results demonstrated that the number of colonies formed by the GABA-treated cells was notably greater than that formed by the control cells (p = 0.004), whereas the population of colonies formed by the GABA\PTX-treated cells did not significantly differ from that formed by the control cells (p = 0.17) (Fig. 6b). In brief, these findings suggest that GABA potentially enhances cancer cell survival in the absence of anchorage, a characteristic of cancer progression.

Fig. 6.

Effects of GABA treatment with or without PTX on soft agar colony formation in A2780CP cells. Soft agar assays of GABA- and PTX + GABA-treated and control cells are shown in (a), and the statistical analysis is shown in (b). Colonies were counted via ImageJ software. The results are presented as the means ± SDs of duplicate samples from representative data from two independent experiments. (N = 6, **p < 0.01). One-way ANOVA with Tukey’s multiple comparisons test was used.

Expression of transcription factors involved in epithelial‒mesenchymal transition

In A2780CP cells, the expression of selected EMT transcription factors was examined following GABA treatment. The results showed that GABA treatment led to an increase in SNAI2, TWIST1, TWIST2, and VIM (Vimentin) but decreased the expression of CDH1. Interestingly, when A2780CP cells were cotreated with GABA and PTX, there was a significant decrease in the expression of SNAI2, TWIST1, TWIST2, and VIM and a significant increase in CDH1 (E-cadherin) expression compared with those in cells treated with GABA alone (Fig. 7a, p < 0.0001). However, SNAI1 expression remained unaffected by GABA-PTX treatment.

Fig. 7.

Alterations in selected EMT-related gene expression profiles in A2780CP cells (a), MDA-MB231 cells (b), MCF-7 cells (c), U251 cells (d), and U87 cells (e). Significant changes were defined on the basis of a criterion of “more than ± twofold changes” plus “p value less than 0.05”. (N = 6. **p < 0.01, ****p < 0.0001). Two-way ANOVA with Tukey’s multiple comparisons test.

GABA treatment had diverse effects on EMT transcription factors in different breast cancer cell lines. Compared with the control, GABA increased CDH1 expression and decreased vimentin levels in MDA-MB231 cells. Compared with GABA treatment alone, GABA-PTX cotreatment significantly increased vimentin gene expression and decreased E-cadherin gene expression (Fig. 7b, p < 0.0001). On the other hand, the GABA-treated MCF-7 cell line presented increased TWIST2 expression, whereas SNAI1, SNAI2, TWIST1, VIM, and CDH1 gene expression decreased. Interestingly, compared with GABA treatment alone, GABA-PTX cotreatment led to a decrease in TWIST2 gene expression and an increase in SNAI1, SNAI2, TWIST1, VIM, and CDH1 gene expression (Fig. 7c, p < 0.0001). These findings highlight the complex and context-dependent regulatory effects of GABA and GABA-PTX cotreatment on EMT transcription factors in breast cancer cell lines.

There were distinct responses to GABA treatment in the U87 and U251 glioblastoma cell lines. In U251 cells, GABA administration increased vimentin expression, which was antagonized by GABA-PTX cotreatment, along with decreased TWIST1 gene expression and a noteworthy increase in E-cadherin gene expression compared with those in the GABA-treated group (Fig. 7d, p < 0.0001). On the other hand, in the U87 cell line, GABA treatment specifically led to an increase in CDH1 gene expression, with no significant effect on other EMT-related transcription factors. Notably, compared with GABA treatment, GABA-PTX cotreatment neutralized the overexpression of E-cadherin and significantly reduced its expression (Fig. 7e, p < 0.0001). Furthermore, GABA-PTX cotreatment induced SNAI2, TWIST2 and VIM gene expression in the U87 cell line.

Indeed, the observed differences in the effects of GABA and GABA-PTX cotreatment on EMT transcription factors in various cancer cell lines and among different cancer types underscore the complexity of EMT regulation. These findings offer valuable insights into the context-dependent nature of GABA’s actions in cancer cells and emphasize the need for tailored therapeutic strategies targeting EMT in specific glioblastoma contexts.

Expression of protein markers of epithelial‒mesenchymal transition

To evaluate the impact of GABA on EMT-related markers, we assessed the protein expression levels of E-cadherin and Vimentin across five human cancer cell lines (MCF-7, MDA-MB-231, U251, U87, and A2780) following GABA treatment. Western blot analysis was performed using samples from the same experiment, with all gels and blots processed in parallel under identical conditions. Band intensities were quantified and statistically analyzed using two-way ANOVA followed by Tukey’s post hoc test to determine significance(Fig. 8). E-cadherin expression was significantly altered by GABA in a cell-specific manner. In MCF-7 cells, GABA treatment led to a marked downregulation of E-cadherin levels compared to untreated controls (mean difference = 0.6609, p < 0.0001), suggestive of an EMT-promoting effect. In contrast, U251 and A2780 cells exhibited significant upregulation of E-cadherin following GABA exposure (mean differences = − 0.6492 and − 0.8280, respectively; both p < 0.0001), indicative of a shift toward an epithelial phenotype. No significant changes were observed in MDA-MB-231 cells (p = 0.6422), while U87 cells showed a modest, non-significant reduction in E-cadherin levels (p = 0.0587). Vimentin expression was also significantly affected by GABA treatment. Two-way ANOVA revealed that the main effects of treatment (p < 0.0001), cell line identity (p < 0.0001), and their interaction (p < 0.0001) were all statistically significant. The GABA-treated group exhibited a significant overall reduction in Vimentin expression compared to controls (mean difference = 0.1053; 95% CI: 0.0723 to 0.1382), suggesting a global attenuation of mesenchymal features. The cell line (column factor) accounted for the majority of the variance (77.05%), while treatment and interaction effects contributed 2.83% and 18.13%, respectively. Together, these findings suggest that GABA differentially regulates EMT markers in a cell line–dependent manner, promoting either an epithelial or mesenchymal shift, depending on the cellular context. Notably, A2780 and U251 cells demonstrated a MET-like response, whereas MCF-7 cells exhibited EMT-associated changes, and MDA-MB-231 cells remained largely unresponsive, consistent with their intrinsic mesenchymal phenotype.

Fig. 8.

Western blot analysis of E-Cadherin and Vimentin expression in MCF-7, MDA-MB-231, U251, U87, and A2780CP cell lines following GABA treatment. GAPDH was used as a loading control. Bar graphs represent densitometric quantification of normalized protein levels. Data are shown as mean ± SD (n = 4). Statistical analysis was performed using two-way ANOVA followed by Tukey’s post hoc test. The original blots are presented in Supplementary Figs. 2, 3 and 4.

Discussion

Research into breast, ovarian, and glioblastoma cancers is highly important, as recent global statistics underscore the significant impact of these diseases on public health. Breast cancer, the most frequently diagnosed cancer in women worldwide, continues to pose a substantial threat, with millions of new cases emerging annually. It remains a leading cause of cancer-related fatalities among women, particularly in developing nations²⁸. Ovarian cancer also presents tough challenges, with a high mortality rate, primarily due to late-stage diagnoses and limited therapeutic options. It ranks as the eighth most prevalent cancer in females²⁹.

Similarly, glioblastoma is associated with a poor prognosis, with a five-year survival rate of less than 10%³⁰. Its aggressive nature and resistance to conventional treatments necessitate urgent research to develop more efficacious therapies. A comprehensive understanding of the molecular and cellular intricacies of these cancers is imperative for devising targeted interventions, early detection methodologies, and personalized approaches, all of which could enhance patient outcomes and alleviate the global burden these devastating diseases impose. Consequently, our study incorporated five distinct established cancer cell lines, each with unique characteristics. In our study, we deliberately selected specific breast cancer and glioblastoma cell lines known for their distinct molecular profiles and phenotypic characteristics, aiming to capture the diverse spectrum of responses to GABAergic modulation. The inclusion of the MCF7 and MDA-MB 231 breast cancer cell lines allowed us to investigate the impact of GABA-A receptor modulation across luminal and triple-negative breast cancer subtypes, which exhibit differential hormone receptor expression and metastatic potential. Similarly, the choice of U251 and U87 glioblastoma cell lines provided an opportunity to explore GABAergic effects in both classical and mesenchymal molecular subtypes of glioblastoma, reflecting the heterogeneity observed in this aggressive brain tumor. While we acknowledge the importance of including multiple ovarian cancer cell lines to comprehensively assess GABAergic modulation in this context, our selection was constrained by the availability of suitable cell lines that could provide a meaningful contrast analogous to the diversity observed in breast cancer and glioblastoma.

Utilization of the agonist-antagonist approach is not only a well-known and scientifically proven method to elucidate the role of a neurotransmitter but also a novel therapeutic³¹. Picrotoxin (PTX), a prototypic antagonist of GABA, is a polycyclic compound lacking a nitrogen atom and was isolated from loco weed plants and shown to be a stimulant causing tonic‒clonic convulsions at quite low doses; however, it was also indexed as an antidote for barbiturate overdose³². It acts as a noncompetitive antagonist and an excellent example of allosteric modulation, not via the GABA recognition site, possibly somewhere within the ion channel, and has been used in GABAergic signaling studies in cancer research^17,33,34. We employed GABA itself, which acts as an orthosteric full agonist, or PTX, which acts as an antagonist, to activate or inhibit GABAergic signaling, respectively.

To the best of our knowledge, these cell lines have not been previously explored in terms of elucidating the expression patterns of GABA-A receptors and the impact of GABA-A receptor stimulation or inhibition on cancer stem cell surface markers, pivotal transcription factors associated with epithelial‒mesenchymal transition (EMT), and the aggressive behaviors of these cells, including migration and anchorage independence.

The selection of experimental assays in this study was guided by the biological characteristics of each cancer cell line and the specific research objectives. Cell viability and proliferation were primarily assessed using the MTT assay, which reflects mitochondrial metabolic activity and provides a reliable index of net cell viability. Although MTT does not distinguish between reduced proliferation and increased cell death, it remains an effective tool for detecting overall phenotypic responses to GABAergic modulation. To complement and validate MTT findings, AO/PI staining was employed to differentiate viable cells from dead cells based on membrane integrity. This dual-staining approach offered mechanistic clarity by distinguishing between suppressed metabolism and actual cytotoxicity, thereby reinforcing the context-dependent effects of GABA observed across cell lines.

Rather than applying a uniform methodology across all experiments, functional assays were selected based on the inherent growth characteristics of each cell type. For the A2780CP ovarian cancer cell line, which exhibits semiattached growth, the soft agar colony formation assay was used to simulate anchorage-independent growth—a condition that closely mimics in vivo tumor behavior and is not accurately captured in monolayer cultures. In contrast, a scratch assay was applied to the remaining cell lines to evaluate migratory behavior under adherent conditions. Notably, our previous studies demonstrated a strong correlation between scratch and soft agar assays across different cancer models, supporting the use of both methods as complementary tools for assessing clonogenicity and migration³⁵.

The findings of this study demonstrate that GABA modulates EMT in a context-dependent manner across diverse cancer cell types, as evidenced by both transcriptional and protein-level analyses. In A2780CP ovarian cancer cells, GABA treatment significantly upregulated the expression of EMT-promoting transcription factors, including SNAI2, TWIST1, TWIST2, and VIM, while concurrently downregulating CDH1 expression, consistent with a mesenchymal shift. These gene-level changes were substantiated by Western blot analysis, which revealed decreased E-cadherin and increased Vimentin protein levels. Importantly, co-treatment with paclitaxel (PTX) reversed these transcriptional effects, restoring epithelial marker expression and suppressing mesenchymal factors, supporting PTX’s role in mitigating GABA-induced EMT. These molecular alterations were further aligned with flow cytometry data, which indicated that GABA promoted a CD133⁺/CD44⁺ stem-like phenotype, an effect neutralized by PTX. Together, these results underscore GABA’s role in promoting EMT and stemness in ovarian cancer cells, hallmarks linked to tumor progression, invasiveness, and chemoresistance^17,36. In breast cancer cell lines, the regulatory impact of GABA and GABA-PTX cotreatment displayed greater complexity. In MDA-MB-231 cells, GABA upregulated CDH1 and downregulated VIM transcripts, a noncanonical EMT profile, which was only partially consistent at the protein level. PTX cotreatment led to increased Vimentin and reduced E-cadherin gene expression, suggesting that in mesenchymal-like MDA-MB-231 cells, PTX may not fully counteract GABA’s effects. In contrast, MCF-7 cells exhibited a unique response pattern: GABA treatment elevated TWIST2 expression while suppressing multiple EMT-related genes including CDH1 and VIM. These transcriptional patterns were reflected in Western blot analysis, which demonstrated downregulation of both E-cadherin and Vimentin protein levels following GABA exposure. Notably, PTX cotreatment restored the expression of these markers at the transcript level, implying that GABA induces a dynamic or partial EMT phenotype in MCF-7 cells, rather than a complete transition. In glioblastoma models, GABA’s effects were again cell-type specific. In U251 cells, GABA induced EMT-like changes characterized by increased VIM and reduced CDH1 gene expression, with corresponding alterations at the protein level. These effects were reversed by PTX cotreatment. Conversely, U87 cells exhibited a more restricted response, with GABA treatment resulting in isolated upregulation of CDH1 and minimal changes in other EMT transcription factors. Although PTX neutralized this upregulation and enhanced mesenchymal gene expression (SNAI2, TWIST2, and VIM), these effects did not reach statistical significance at the protein level, suggesting limited EMT plasticity in U87 cells. A notable limitation of this study is that Western blot analysis was not performed for the GABA/PTX co-treatment group due to budget constraints. As a result, protein-level validation of PTX’s regulatory effects on E-cadherin and Vimentin remains unavailable. However, the strong consistency observed between transcriptional changes and protein expression in the GABA-treated groups provides confidence in the gene expression findings and supports the overall interpretation. Collectively, this study highlights the heterogeneous EMT responses elicited by GABA, and the modulating effects of PTX, across a spectrum of cancer cell types. The observed concordance between mRNA and protein expression, particularly in A2780CP, MCF-7, and U251 cells, strengthens the case for GABA as a functional regulator of EMT and cancer cell plasticity. These results emphasize the importance of cellular context—including lineage, baseline EMT status, and drug responsiveness—in determining GABA’s impact. Given previous reports of elevated systemic GABA levels in ovarian cancer patients³⁷, targeting GABAergic signaling pathways may represent a promising therapeutic strategy for limiting EMT-driven progression and enhancing chemosensitivity in a tumor-specific manner.

Colony soft agar formation assays revealed that GABA treatment stimulated colony formation in A2780CP cells, another indicator of carcinogenesis, whereas PTX inhibited this effect. These findings align with observations in metastatic prostate cancer patients, where elevated GABA levels promoted metastasis through MMP production³⁸. These findings collectively suggest that GABA regulates EMT and stemness in ovarian cancer cells, suggesting that GABA is a potential therapeutic target for improved treatments^39–41.

In breast cancer cells, GABA treatment had different effects, with increased CDH1 expression and decreased VIM expression in MDA-MB231 cells. Conversely, GABA increased TWIST2 expression while reducing SNAI1, SNAI2, TWIST1, and VIM expression in MCF-7 cells. These changes influence cell invasion and metastasis, as evidenced by wound healing assays. Cotreatment with GABA-PTX negatively impacted the antimigratory effect of GABA, possibly by altering the expression of EMT transcription factors. Flow cytometry analysis indicated that GABA decreased the CD44 + cell population in MDA-MB231 cells, suggesting a dual role of GABA in reducing the number of CD44 + cells and suppressing their migration^42–45. The impact of GABA on cell viability was cell type specific, with no significant difference observed in MDA-MB231 cells, which is consistent with previous findings^42,43. Clinical studies have linked low GABA levels and E-cadherin expression with breast cancer progression and increased mortality rates, possibly due to neoangiogenesis and increased oxygen levels in the tumor microenvironment^46,47. GABA treatment reduced migration rates by targeting EMT transcription factors in MCF-7 cells. The impact of GABA on breast cancer cells appears to be context dependent^48–52.

In glioblastoma cell lines (U87 and U251), GABA had contrasting effects, with an increase in E-cadherin expression in U87 cells and an increase in vimentin expression in U251 cells. Wound healing assays revealed no change in cell migration despite differences in EMT gene expression patterns. The flow cytometry results indicated no significant alterations in the CD133 + or CD44 + cell populations. The effect of GABA on proliferation varied; it was inhibited by PTX in U87 cells but did not affect U251 cells. The underlying mechanisms of these differences may be related to GABA oxidation.

While Boyden chamber and Matrigel invasion assays are widely used for evaluating chemotaxis, invasiveness, and extracellular matrix (ECM) degradation, we did not employ these methods in the current study. Our primary objective was to assess cell migration rather than invasion or chemotactic behavior. Accordingly, the wound healing assay performed under low-serum (1% FBS) conditions was selected to minimize proliferation and isolate migration as the variable of interest. Although Matrigel and Boyden chamber systems offer valuable insights into ECM remodeling and aggressive cellular behavior, they are more suited for studying invasion and chemotactic responses—which were beyond the scope of this study. Therefore, the scratch assay provided a more direct and controlled method to quantify lateral migratory capacity without confounding effects from matrix digestion or serum-driven chemotaxis.

Alterations in GABA-A receptor subunit expression were also noted, with increased expression of specific subunits in various cell lines. These changes could influence cancer progression, emphasizing the potential of GABA-A receptors as therapeutic targets⁵³. The variability in GABA-A receptor subunit expression raises questions about its potential impact on cancer progression, underscoring the importance of targeting GABA receptors for therapeutic purposes. In our examination of GABA-A receptor subunit expression, we observed a notable increase in the β2 and γ2 subunits in A2780CP cells compared with those in the control group. The upregulation of GABRB2 in human papillary thyroid cancer tissues and cell lines has also been reported to be positively correlated with lymphatic metastases⁵³. Inhibition of GABRB2 expression led to reductions in colony formation, migration, invasion, and proliferation in papillary thyroid cancer cell lines, suggesting a pivotal role of GABRB2 in the development and metastasis of thyroid cancer. While we hypothesize that the increased expression of the β2 subunit may contribute to the increased aggressiveness of A2780CP cells, further research is warranted to elucidate the exact role of GABRB2 in cancer development.

Furthermore, our investigations revealed elevated expression of the γ2 and γ3 subunits in the MDA-MB231, MCF-7, U251, and U87 cell lines. Additionally, an increase in the ρ2 and ρ3 subunits was observed in the MDA-MB231 and U87 cell lines, whereas the MCF-7 and U251 cell lines presented increased expression of the α3 and α4 subunits. These observations suggest that the alterations in γ2 and γ3 subunit expression across these cell lines may be attributed to a common underlying mechanism. Previous studies have demonstrated the essential roles of γ2 and γ3 in forming and maintaining functional synaptic clusters of GABA-A receptors, which are critical for postnatal brain development⁵⁴. The shifts in γ subunit expression suggest a potential role in the neurogenesis of cancerous cells, providing valuable insights for further research in this area.

Additionally, both the MDA-MB231 and U87 cell lines presented increased expression of the ρ subunit, which aligns with our findings in gliomas. In contrast, the MCF-7 cell line presented increased α3 subunit expression, and the overexpression of α3 did not impact the cell migration response induced by GABA, contrary to our initial observations. Given that GABRA3 expression is associated with increased migration and invasiveness in breast cancer cells⁵⁵, further research is needed to elucidate the underlying mechanisms by which GABA inhibits the migration of α3 subunit-expressing cells. On the other hand, U251 cells presented increased expression of the GABRA4 and GABRG2 subunits, in contrast to the findings of Belotti et al., which suggested that decreased expression of these subunits may be involved in tumor inhibition in high-grade gliomas. The role of subunit GABA receptors appears to be multifaceted and warrants further investigation to enhance our understanding of the impact of GABA on activating mechanisms related to EMT and to provide valuable insights into novel therapeutic approaches in this area.

Conclusion

Our study sheds light on the complex and context-dependent role of GABA-A receptors in modulating cancer progression in breast, ovarian, and glioblastoma cell lines. By employing a receptor-specific agonist–antagonist approach using GABA and picrotoxin (PTX), we were able to demonstrate that GABA-A receptor activation leads to diverse phenotypic responses, including changes in proliferation, migration, cancer stem cell marker expression, and EMT-related gene expression. These effects were shown to be reversible by PTX, underscoring the functional relevance of GABA-A receptor signaling in these models.

The use of the MTT assay was intentionally chosen to provide a rapid and integrative measure of cell viability, allowing detection of both growth-promoting and cytotoxic effects. While additional assays to delineate precise mechanisms of cell death or proliferation are valuable, the observed responses were sufficiently distinct to establish cell-specific effects of GABAergic modulation. Similarly, although protein-level validation of receptor subunits and EMT markers would provide further depth, our transcript-level analysis—normalized against brain tissue and validated using the most stable housekeeping gene (B2M) identified through rigorous testing—offered a reliable and biologically informative platform for comparison.

Our findings support the hypothesis that the expression pattern of GABA-A receptor subunits plays a critical role in defining cancer cell behavior, and that this pattern varies significantly between cell types. These insights suggest that profiling GABA-A receptor subunit expression could aid in identifying tumor-specific vulnerabilities and tailoring personalized therapeutic strategies.

Furthermore, we recognize the importance of extending our findings through additional functional assays. Future investigations should include protein-level validation of GABA-A receptor subunits and EMT markers, transwell migration/invasion assays, and dedicated analyses to distinguish between cell death and proliferation. Although these were beyond the scope of the current study due to time and funding constraints, they represent logical next steps to deepen the mechanistic understanding of GABAergic modulation in cancer. Collectively, our results establish a strong foundation for such follow-up studies and highlight the promise of GABA-A receptor profiling as a tool for advancing personalized oncology.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: MK, NH, MS, and SS. Investigation: MK, NH, MS, SHS, MRS and SS.Writing—original draft; review and editing and final manuscript: MK, NH, MS, SHS, MRS and SS.

Data availability

The data generated in this study are available upon request from the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.