Loss of ABAT-Mediated GABAergic System Promotes Basal-Like Breast Cancer Progression by Activating Ca2+-NFAT1 Axis.

Basal-like breast cancer (BLBC) is the most aggressive subtype with a poor clinical outcome; however, the molecular mechanisms underlying aggressiveness in BLBC remain poorly understood.
Methods: The effects of gamma-aminobutyrate aminotransferase (ABAT) on GABA receptors, Ca2+-NFAT1 axis, and cancer cell behavior were assessed by Ca2+ imaging, Western blotting, immunostaining, colony formation, and migration and invasion assays. We elucidated the relationship between ABAT and Snail by luciferase reporter and ChIP assays. The effect of ABAT expression on BLBC cells was determined by in vitro and in vivo tumorigenesis and a lung metastasis mouse model.
Results: We showed that, compared to other subtypes, ABAT was considerably decreased in BLBC. Mechanistically, ABAT expression was downregulated due to Snail-mediated repression leading to increased GABA production. GABA then elevated intracellular Ca2+ concentration by activating GABA-A receptor (GABAA), which contributed to the efficient activation of NFAT1 in BLBC cells. ABAT expression resulted in inhibition of tumorigenicity, both in vitro and in vivo, and metastasis of BLBC cells. Thus, loss of ABAT contributed to BLBC aggressiveness by activating the Ca2+-NFAT1 axis. In breast cancer patients, loss of ABAT expression was strongly correlated with large tumor size, high grade and metastatic tendency, poor survival, and chemotherapy resistance. Conclusions: Our findings have provided underlying molecular details for the aggressive behavior of BLBC. The Snail-mediated downregulation of ABAT expression in BLBC provides tumorigenic and metastatic advantages by activating GABA-mediated Ca2+-NFAT1 axis. Thus, our results have identified potential prognostic indicators and therapeutic targets for this challenging disease.

Introduction

Basal-like breast cancer (BLBC) subtype accounts for approximately 10-20% of breast cancers and frequently occurs in younger patients. BLBC tumors are usually of larger size and higher grade with a tendency for recurrence and metastasis and have a poor response to chemotherapy 1, 2. This subtype is often triple negative—lacks the expression of estrogen receptor (ER), progesterone receptor (PR) and epidermal growth factor receptor 2 (HER2)—which limits the use of targeted treatments such as endocrine and anti-HER2 therapies, often leading to a fatal clinical outcome. Furthermore, BLBC has an increased propensity to metastasize to the brain and lungs, sites that are associated with poor prognosis and short survival 3-7. Because of its aggressiveness and lack of effective therapeutics, there is a critical need to elucidate the underlying molecular mechanisms and identify molecular targets in BLBC.

Gamma-aminobutyric acid (GABA) is a major inhibitory neurotransmitter in the mammalian central nervous system 8. The biological effects of GABA are mediated by ionotropic GABA-A receptor (GABAA), a family of ion channels, and by the metabotropic GABA-B receptor (GABAB), which is a G protein-coupled receptor. In addition to functioning as an inhibitory neurotransmitter, it can also operate as a trophic factor during neural development to regulate the proliferation, migration, differentiation, and death of neuronal cells 9. Additionally, it has become clear that GABA and GABA receptors are also present in non-neuronal tissues 10. Studies have revealed the involvement of GABA in cancer growth and metastasis. GABA has inhibitory effects on colon cancer 11, gastric cancer 12, and hepatocellular carcinoma 13. GABA has also been shown to promote pancreatic cancer growth through upregulation of the pi-subunit expression of the GABAA receptor 14, and elevated levels of GABA and glutamate decarboxylase (GAD) were observed in metastatic prostate cancer 15. These findings suggest a critical role for the GABAergic system in cancer and enforce the need for further evaluation of this pathway in distinct types of cancers.

GABA levels are maintained by its biosynthetic and catabolic pathways. Much attention has been focused on modulation of the biosynthetic pathway of GABA, whereas its catabolic pathway is less studied but may play an equally critical role in the GABAergic system in tumor cells. Gamma-aminobutyrate aminotransferase (ABAT), a key enzyme responsible for the catabolism of GABA, catalyzes the transfer of the amino group of GABA to α-ketoglutarate, producing succinic semialdehyde and L-glutamate. Patients with ABAT deficiency display elevated GABA levels along with a severe phenotype including psychomotor retardation, lethargy, refractory seizures, hypotonia, and hyperreflexia 16, 17 attributed to the loss of ABAT-mediated disruption of the GABAergic system. In this study, we report that loss of ABAT expression occurs specifically in BLBC and predicts poor prognosis. Loss of ABAT expression promotes tumorigenic and metastatic potential of BLBC cells by activating GABA-mediated Ca2+-NFAT1 axis. Our study provides a molecular understanding of how loss of ABAT contributes to tumor growth and metastasis in BLBC.

Methods

Plasmids and antibodies

Human ABAT and Snail genes were amplified from MCF7 and MDA-MB-231 cDNA libraries, respectively, and sub-cloned in pBABE-Puro.

Antibodies against ABAT, G9a, H3K9me2, H3K9Ac, and NFAT1 were purchased from Abcam. Antibodies against Vimentin and Snail were acquired from Neomarkers and Cell Signaling Technology, respectively. Antibody for E-cadherin was purchased from BD Transduction Laboratories. Antibodies for mCherry and β-actin were obtained from Sigma-Aldrich.

Cell culture

MDA-MB-231 and SUM159 cells were grown in DMEM/F12 supplemented with 10% fetal bovine serum (FBS). BT-549, BT-483, MCF7, and HCC1428 cells were grown in RPMI1640 plus 10% FBS. For establishing stable transfectants with ABAT expression, BLBC cells were transfected with pBABE-ABAT; stable clones were selected with puromycin (300 ng/mL) for 4 w.

Quantitative real-time PCR

Total RNA was isolated using RNeasy Mini kit (Qiagen) according to the manufacturer's instructions. Specific quantitative real-time PCR experiments were performed using SYBR Green Power Master Mix following the manufacturer's protocol (Applied Biosystems).

Luciferase reporter assay

The assay was performed according to the procedure described previously 18, 19. All experiments were performed in triplicate.

Chromatin immunoprecipitation (ChIP)

ChIP assays were performed as described previously 18, 19. The following primers were used for ChIP assays: 5'- CAAATACCTCTAGAAAGCTGT -3′ and 5'- GAAGGTGCCTTTCTACCGTTG -3′ for the ABAT promoter. The cells were prepared to perform ChIP assay with the Imprint ChIP Kit (Sigma) according to the manufacturer's instructions and as described recently 18.

Ca2+ imaging

MDA-MB-231 cells plated on glass bottom cell culture dishes were loaded with the calcium-sensitive fluorescent dye Fluo-4/AM (4 mM; Invitrogen) in Hank's Balanced Salt Solution containing 0.02% pluronic acid (Sigma) for 45 min at 37 °C. GABA (Sigma), CGP (Santa Cruz), picrotoxin (Santa Cruz) and cyclopiazonic acid (CPA; Sigma) were applied at concentrations of 2 mM, 10 μM, 10 μM, and 10 μM, respectively. The Ca2+-free buffer contained 1 mM EDTA. Fluorescence was measured by an Olympus Confocal Laser Scanning Microscope (DU-897D-CS0) using MetaMorph software. Serial scanning was performed at 488/530 nm excitation/emission wavelengths at 1 s intervals. Fluorescence intensity changes (F%) were shown as the percentage of baseline fluorescence.

HPLC

The cells were washed twice with PBS and suspended in 0.4 M perchloric acid for 10 min, and then supernatants were collected by centrifugation. Supernatants were reacted with 1-dimethyl aminonaphthalenesulfonyl chloride (10 g/L in acetone) at 80 ℃ and terminated by adding 100 μL acetic acid (1 M in acetone). After filtering, supernatants were subjected to HPLC system using C18 column (5 μm, 4.6 × 250 mm). Chromatography was carried out with 42% methanol and 58% 0.1 M sodium acetate (v/v) containing 1% tetrahydrofuran and 14 mM 1-heptanesulfonic acid sodium salt. The flow rate was 1 mL/min and the detection was performed at 360/500 nm.

Immunostaining

Cells were grown on glass bottom cell culture dishes and incubated with NucBlue™ Live ReadyProbes™ Reagent (Invitrogen) for 25 min at room temperature. After three washes with HBSS, fluorescence was measured using an Olympus Confocal Laser Scanning Microscope (OLYMPUS IX83-FV3000-OSR).

Colony formation assay

Colony formation assay was performed using double-layer soft agar in 24-well plates with a top layer of 0.35% agar and a bottom layer of 0.7% agar. Cells were seeded in 24-well plates in desired medium and cultured at 37 °C for 15 to 20 days, and the colonies were stained and counted.

Migration and invasion assays

Migration and invasion assays were performed as described previously 18, 19. All experiments were repeated at least twice in triplicate. Statistical analysis was performed using the Student's t-test; a p-value of <0.05 was considered significant.

Tumorigenesis assay and lung metastasis model

Animal experiments were performed according to procedures approved by the Institutional Animal Care and Use Committee at Zhejiang University. To examine the effect of ABAT on tumorigenesis, female SCID mice (5-8 w old) were injected with 1×106 exogenous ABAT-overexpressing cells on the left flank and vector control cells on the right flank. Tumor formation was monitored every 2 to 4 days for 30 days. Tumor size and weight were measured. To test the effect of ABAT on tumor metastasis, SCID mice were injected via tail vein with MDA-MB-231 cells (1×106 cells/mouse) with stable empty vector or ABAT overexpression (6 mice/group). After 4 w, lung metastasis was analyzed by an IVIS-100 imagining system (Xenogen). Mice were sacrificed and lung metastatic nodules were detected in paraffin-embedded sections by staining with hematoxylin and eosin. Data were analyzed using the Student's t-test; a p-value <0.05 was considered significant.

Statistical analysis

Results are expressed as mean ± SD or SEM as indicated. Comparisons were made by the two-tailed Student's t-test or one-way ANOVA. Correlations between ABAT and Snail were analyzed by Pearson's correlation method and Spearman's rank correlation test. Survival curves were plotted using the Kaplan-Meier method, and differences were analyzed by the log-rank test. In all statistical tests, p < 0.05 was considered statistically significant.

Results

ABAT expression is downregulated in BLBC subtype

We recently reported that several metabolic genes—fructose-1,6-biphosphatase (FBP1), aldo-keto reductase 1 member B1 (AKR1B1), and urine diphosphate-galactose ceramide galactosyltransferase (UGT8)—were associated with BLBC aggressiveness 20-22. To explore other clinically relevant determinants for BLBC, we analyzed gene expression profiles of breast cancer in multiple publicly available cDNA microarray datasets, GSE1456, GSE25066, NKI295, TCGA, and MEBTABRIC, which contain 159, 508, 295, 1215, and 1904 breast cancer patients, respectively 23-25. Several known genes previously shown to have critical roles in BLBC exhibited remarkable differences between BLBC and other subtypes, such as lactate dehydrogenase B (LDHB), AKR1B1, UGT8, and FBP1. Notably, in contrast to other subtypes, ABAT expression was markedly downregulated in BLBC (Figure and Figure ). Consistent with this observation, ABAT protein expression was also significantly decreased in BLBC by proteogenomic analysis of a TCGA dataset that contains 105 breast tumor samples 26 (Figure ). To confirm this observation, we collected freshly frozen breast tumor tissues from 21 cases of luminal subtype and 9 cases of the triple-negative subtype that have a significant overlap with BLBC. ABAT expression was elevated in the luminal subtype of breast cancers but significantly downregulated in triple-negative breast cancer (Figure and Figure ). To better characterize the link between ABAT and basal subtype, we evaluated ABAT expression in four other gene expression datasets, GSE12777, GSE10890, E-TABM-157, and E-MTAB-181, containing 51, 52, 51, and 56 breast cancer cell lines, respectively 27-29. Consistently, loss of ABAT was correlated with the basal subtype of breast cancer cell lines (Figure ). Subsequently, we confirmed these findings by semi-quantitative RT-PCR or quantitative real-time PCR in a panel of breast cancer cell lines containing 5 luminal and 5 BLBC cell lines. ABAT mRNA expression was consistently much lower in BLBC cells than in luminal cells (Figure and Figure ). We also detected loss of expression of the epithelial marker E-cadherin but elevated levels of mesenchymal markers vimentin and Snail in BLBC cells. Like E-cadherin, ABAT protein was also lost in BLBC cell lines, whereas it was elevated in luminal cell lines (Figure and Figure ). Thus, our data confirmed that loss of ABAT expression was primarily restricted to BLBC, underscoring its underlying functions in this subtype of breast cancer.

ABAT expression is downregulated by Snail-mediated repression

While investigating the expression of ABAT in BLBC, we noticed inverse expression patterns of ABAT and Snail in breast cancer cell lines (Figure ). To confirm this observation, we examined the expression of ABAT and Snail in the TCGA dataset. As expected, ABAT expression negatively correlated with Snail expression (Figure ). We also analyzed Snail expression in different subtypes of breast cancer and found that, contrary to ABAT, Snail was significantly upregulated in BLBC in the TCGA dataset (Figure ). To elucidate the causal relationship between ABAT and Snail, we analyzed ABAT expression in MCF7 cells with ectopic Snail expression in two previous datasets (GSE29672 and GSE58252) 30, 31 and observed dramatic downregulation of ABAT expression by Snail (Figure ). Next, we expressed Snail in three luminal breast cancer cell lines, BT483, HCC1428, and MCF7. As expected, Snail expression significantly downregulated ABAT expression in these cell lines (Figure ). These results indicated that Snail, as a transcriptional repressor, may suppress ABAT expression through direct transcriptional regulation.

We next investigated whether ABAT expression is regulated directly by Snail. Analysis of the DNA sequence in the ABAT promoter revealed that it contains four putative Snail-binding E-boxes (CAGGTG) from -3300 bp to the transcription start site (TSS) (Figure ). To determine whether these E-boxes are crucial for Snail-mediated transcriptional repression, we cloned the human ABAT promoter and created several deletion mutants of promoter-luciferase constructs based on the location of the E-boxes, including FL1 (-3300 bp), FL2 (-2901 bp), FL3 (-2489 bp), and FL4 (-1293 bp) (Figure ). Upon expression of FL1 in BT483, HCC1428, and MCF7 cells, Snail significantly repressed ABAT promoter luciferase activity (Figure ). Notably, the constructs without the regions between -3300 bp and -1293 bp still maintained low reporter activity induced by Snail, indicating that the E-box between -1293 bp and TSS might be critical for Snail-mediated ABAT repression (Figure ). To test this, we introduced a point mutation in the E-box at -728 bp. As anticipated, a mutation in this E-box (mut) remarkably abolished Snail-mediated ABAT repression (mut vs. FL1) (Figure ), suggesting that Snail represses the ABAT promoter in an E-box-dependent fashion and that the E-box at -728 bp is required for Snail-induced transcriptional repression.

To determine whether Snail targets ABAT directly, we performed chromatin immunoprecipitation (ChIP) assays in BT483, HCC1428, and MCF7 cells with Snail overexpression. The results showed that Snail directly bound to the ABAT promoter (Figure ), indicating that ABAT is a direct target of Snail. Our previous study showed that Snail-G9a complex binds to the E-cadherin promoter for epigenetic silencing of its expression 20. We speculated that this complex might repress ABAT expression by binding to its promoter. Indeed, the downregulation of ABAT was associated with increased H3K9me2 and decreased H3K9 acetylation in the ABAT promoter in breast cancer cell lines (Figure ). We also detected a dramatic enrichment of G9a, a major methyltransferase responsible for H3K9me2, in the ABAT promoter (Figure ), suggesting involvement of the Snail-G9a complex in the up-regulation of H3K9me2 in the ABAT promoter. The increased Snail, G9a, and H3K9me2 at the ABAT promoter correlated well with the downregulation of ABAT expression (Figure ). Together, these data indicated that Snail-mediated epigenetic modification is critical for silencing ABAT expression.

ABAT expression reduces GABA level and enhances breast cancer cell migration and invasion

GABA is a substrate of ABAT in the GABA catabolic pathway (Figure ). We first analyzed the association of GABA with aggressive breast cancer using the previous metabolomics data 32 and found that the GABA level was significantly elevated in BLBC and ER-negative breast cancer (Figure ). To explore the association between GABA and ABAT expression levels, we generated stable clones with the ABAT expression vector or the empty vector in MDA-MB231, SUM159 and BT549 cells (Figure ). Subsequent analysis showed that ABAT expression caused a remarkable decrease in GABA level (Figure ), indicating that ABAT is required for decreasing its expression in breast cancer cells. We also examined the effect of ABAT expression, vigabatrin (an inhibitor of ABAT), and GABA on breast cancer cell proliferation, migration, and invasion. There was no significant effect on the growth of MDA-MB231, SUM159, and BT549 cells by all three treatments (Figure and Figure ). However, ABAT expression markedly repressed the migration and invasion of all three breast cancer cell lines, whereas vigabatrin and GABA significantly restored the decreased migration and invasion of these cells with stable ABAT expression (Figure and Figure ). These data suggest a key role of GABA in the loss of ABAT-mediated migratory and invasive ability in breast cancer cells.

ABAT expression downregulates intracellular Ca2+ concentration and represses NFAT1 activation

In immature neurons, GABA increased the intracellular Ca2+ concentration by activating GABAA receptor 33-35. We examined the intracellular Ca2+ concentration in MDA-MB231 cells by monitoring Ca2+ changes. GABA caused a sharp increase of fluorescence in these cells, indicating a quick induction of transient Ca2+ followed by long-lasting oscillations (Figure ). To investigate whether activation of GABA receptors is critical for Ca2+ changes in tumor cells, we determined the effect of picrotoxin and CGP, inhibitors of GABAA and GABAB, respectively, on Ca2+ entry. Treatment with picrotoxin (10 μM) almost completely blocked the GABA-mediated Ca2+ rise, whereas application of CGP (10 μM) only caused a partial decrease of intracellular Ca2+ concentration in MDA-MB231 cells (Figure ), indicating that GABA functions largely via GABAA-mediated signaling. To determine whether the Ca2+ rise was caused by Ca2+ influx through voltage-gated Ca2 + channels (VGCCs) or intracellular Ca2+ store release, we used the specific inhibitor CPA, a sarco-endoplasmic reticulum Ca2+-ATPase (SERCA) blocker. In the presence of CPA (10 μM), GABA evoked partially reduced transient Ca2+ (Figure ). When Ca2+ was omitted from the buffer, GABA-evoked transient Ca2+ was completely blocked by CPA or picrotoxin but slightly reduced by CGP (Figure ). These data suggest that both Ca2+ influx and intracellular Ca2+ store release trigger GABA-mediated Ca2+ rise, and intracellular Ca2+ store release is controlled mainly by GABAA-mediated signaling. We examined the basal Ca2+ concentration in MDA-MB231 and BT549 cells transfected with stable empty vector or ABAT expression vector. Remarkably, ABAT expression reduced the intracellular Ca2+ concentration, providing evidence for the loss of ABAT-mediated Ca2+ rise (Figure ).

To explore the association of ABAT with basal Ca2+ concentration, we first analyzed Ca2+ concentration in four luminal and BLBC cell lines. The Ca2+ concentration was much higher in BLBC cells without ABAT expression than in luminal cells with endogenous ABAT-expression (Figure and Figure ), confirming that loss of ABAT expression contributes to Ca2+ rise. Nuclear factors 1-4 (NFAT1-4) of activated T cells are important Ca2+ sensors and NFAT1 and NFAT2 are activated in aggressive breast cancer 36-41. To explore the association of ABAT-mediated Ca2+ changes with NFAT1 and NFAT2, we determined the effect of ABAT on nuclear translocation of both factors. Our results showed that ABAT expression led to a significant decrease in nuclear translocation of NFAT1 in MDA-MB231 and SUM159 cells as detected by immunostaining-confocal analysis (Figure ). Similar results were observed by Western blotting analysis (Figure ). However, ABAT expression did not significantly change the nuclear entry of NFAT2 (Figure ), suggesting that ABAT functions mainly via NFAT1-mediated signaling.

ABAT suppresses tumorigenicity of breast cancer

Given the strong association of ABAT loss with GABAergic signaling in BLBC, we assessed the functional role of ABAT in tumor formation using the soft-agar assay. Although ABAT expression did not significantly affect the proliferation of MDA-MB231, SUM159, and BT549 cells (Figure ), it resulted in a remarkable decrease in colony formation of these cells, which was significantly restored by vigabatrin in cells with stable ABAT expression (Figure and Figure ). To test the in vivo tumorigenicity, we performed tumor xenograft experiments in which female SCID mice were injected with MDA-MB231 and SUM159 cells with the stable empty vector or ABAT expression vector. As shown in Figure , MDA-MB231 and SUM159 cells with stable ABAT expression led to significantly reduced tumor growth compared with their corresponding vector control cells. We extended our observations to a clinicopathologically relevant context by exploring a possible association between ABAT expression and clinical specimens. We first analyzed ABAT expression and its correlation with tumor size of breast cancer patients in NKI295 and MEBTABRIC datasets. Patients were divided into two groups according to the primary tumor size. There was a significant association between small tumor size and ABAT expression (Figure ). We then evaluated the relationship between ABAT expression and histological grades of the tumors in MEBTABRIC, GSE25066, NKI295, GSE7390, and GSE1456 datasets in which tumors had been scored for tumor grade. We segregated patients into three groups according to histological grades of tumors. ABAT expression was present predominantly in grade 1 and 2 tumors but to a much lesser extent in grade 3 tumors (Figure and Figure ). These data reinforced the notion that loss of ABAT expression is critical for tumorigenicity and functions as an important mediator of tumor aggressiveness.

ABAT suppresses breast cancer metastasis

Because NFAT1 can promote cell migration, invasion, and metastasis, and because loss of ABAT is associated with NFAT1 activation, we speculated that loss of ABAT expression might be critical for breast cancer metastasis. To test this notion, we first tested whether ABAT expression affected tumor metastasis in a xenograft model in which MDA-MB231 cells were injected via the tail vein to generate pulmonary metastases. Remarkably, ABAT expression suppressed lung metastasis (Figure ). Next, we sought to elucidate the clinical relevance of this observation. We first assessed whether ABAT expression was correlated with metastasis in the GSE25066 dataset. Patients were divided into two groups according to their metastatic status. Tumors with low ABAT expression had a higher probability of developing metastasis than those with high ABAT expression (Figure ). We then evaluated if there was a correlation between ABAT expression and metastatic sites in the GSE12276 dataset with 204 breast cancer patients 42. Consistent with the metastatic tendency of BLBC, primary tumors with low ABAT expression preferentially metastasized to the brain and lungs (Figure ).

Given the critical function of ABAT expression in breast cancer, we performed Kaplan-Meier analyses to determine whether ABAT is a prognostic marker for clinical outcomes by analyzing NKI295 and GSE25066 datasets 23, 24. Patients were divided into two groups based upon ABAT expression levels, with low ABAT expression having shorter overall (OS), relapse-free (RFS), and distant metastasis-free survival (DMFS) (Figure and Figure ). We also used an aggregate breast cancer dataset to determine its clinical relevance 43, showing that tumors with low ABAT expression exhibited shorter OS, RFS, and DMFS (Figure ). A similar result was observed by analyzing RFS in BLBC patient samples of this dataset (Figure ). We then determined whether ABAT expression was associated with chemotherapy sensitivity in the GSE25066 dataset in which patients were treated with chemotherapy containing sequential taxane and anthracycline-based regimens. A significant trend was observed between reduced ABAT expression and chemotherapy resistance (Figure ). These data suggest that ABAT expression is potentially useful in prognostic stratification of patients with breast cancer.

Discussion

ABAT expression is specifically down-regulated in BLBC, the most aggressive subtype of breast cancer. Our study provides several mechanistic and clinical insights into the essential role of ABAT loss in BLBC aggressiveness by elucidating the upstream and downstream molecular events leading to decreased levels of ABAT and consequently activating the Ca2+-NFAT axis.

Snail-mediated repression leads to loss of ABAT

Snail is a key transcriptional repressor that binds to the E-box motif (CAGGTG) controlling cell proliferation, migration and metastasis, and therapeutic response 44, 45. Furthermore, Snail is highly expressed in BLBC 18, 20, 46, 47, suggesting it is a master regulator of the BLBC phenotype. Correlation analysis in a large breast cancer gene expression dataset demonstrated a negative correlation between ABAT and Snail expression. Interestingly, ectopic expression of Snail in breast cancer cells significantly repressed ABAT expression. Our data identified Snail as a direct transcriptional repressor of ABAT. We recently showed that Snail formed a complex with H3K9 methyltransferase G9a, which is required for Snail-mediated H3K9me2 on the E-cadherin promoter in BLBC cells 18. Consistent with this finding, Snail recruited G9a to ABAT promoter for H3K9m2 to directly inhibit the transcription of ABAT. Our data indicated that Snail-mediated epigenetic modification is critical for downregulation of ABAT expression in BLBC.

Loss of ABAT expression activates Ca2+-NFAT1 axis

It has been reported that ABAT expression is downregulated in ER-negative breast cancer 48, 49. In this study, we extended this observation and found low ABAT expression and high GABA levels in BLBC. Aberrant GABA levels have been described in many tumor tissues such as neuroblastoma, colorectal, ovarian, and pancreatic carcinomas 14, 15, 50, but the underlying mechanisms remain unclear. We showed that BLBC, compared with other subtypes of breast cancers, had significantly elevated GABA content due to loss of ABAT expression, which was consistent with the previous observation that patients with ABAT deficiency had increased GABA production 16, 17. When the tricarboxylic acid cycle was inhibited, GABA could function as a trophic source to confer a survival advantage for cells through GABA shunt 51. However, even if tumor cells had relatively high GABA production, it was not sufficient for cancer cells proliferation 52. These studies suggest that loss of ABAT expression regulates BLBC aggressiveness by activating GABAergic signaling.

The altered Ca2+ signaling was associated with critical events during tumor progression, such as proliferation, migration, invasion, and metastasis 53. In immature neurons, GABA activated GABAA to produce sufficient depolarization to elevate the intracellular Ca2+ concentration by activating voltage-dependent Ca2+ channels 33-35. Consistent with this notion, loss of ABAT expression mediated intracellular Ca2+ rise through activation of GABAA because picrotoxin, an inhibitor of GABAA, almost completely blocked GABA-mediated Ca2+ rise. Additionally, the intracellular Ca2+ rise could be partially inhibited in the presence of SERCA blocker but could be almost entirely blocked after Ca2+ was omitted from the buffer, indicating that loss of ABAT-mediated intracellular Ca2+ rise was due to Ca2+ influx through VGCCs and intracellular Ca2+ store release.

Ca2+ signaling functions in cancer cells through upregulating oncogenes and/or downregulating tumor suppressors. NFAT, as a Ca2+ sensor, has a remarkable ability to sense dynamic changes of intracellular Ca2+ and frequency of Ca2+ oscillations in cells 54. NFAT proteins are phosphorylated and reside in the cytoplasm in resting cells; upon stimulation by Ca2+, they are dephosphorylated and translocate to the nucleus where they are transcriptionally active, thus providing a direct and important link between intracellular Ca2+ signaling and downstream gene expression. The oncogenic potential of NFAT1 proteins has been documented by their involvement in controlling migration and invasion of tumor cells 38-41. Importantly, NFAT1 is constitutively activated in triple-negative breast cancer and promotes tumorigenesis and metastasis 36, 37. Indeed, in our study, basal Ca2+ rise and nuclear translocation of NFAT1 were observed in BLBC cells. Consistently, ectopic ABAT expression in BLBC cells suppressed tumorigenicity and metastasis in vitro and in vivo by keeping NFAT1 in a heavily inactive state. These data support the crucial role of ABAT loss-mediated Ca2+-NFAT1 axis in the aggressive behavior of BLBC.

Taken together, ABAT expression was downregulated due to Snail-mediated repression. Loss of ABAT expression then resulted in Ca2+ rise by activating the GABA receptor, which contributed to the activation of NFAT1 in BLBC cells. These findings provide a link between loss of ABAT expression and remodeling of Ca2+-NFAT signaling that facilitates BLBC progression.

Our study provides potential prognostic indicators and therapeutic targets for BLBC

We have shown that loss of ABAT expression is associated with several factors that identify patients who are at risk of cancer progression and predict patient prognosis. These include: 1. Breast cancer subtypes (low ABAT expression occurs specifically in BLBC); 2. Grade (a significantly high frequency of low ABAT expression exists in patients with higher grade tumors); 3. Tumor size (low ABAT expression is significantly correlated with larger tumor size); 4. Tumor metastasis (low ABAT expression has a significantly higher risk of metastasis and is highly associated with metastatic dissemination to the brain and lungs, consistent with the metastatic tendency of BLBC); 5. Survival rate (low ABAT expression predicts poor overall, relapse-free, and distant metastasis-free survival); and, 6. Chemotherapy (low ABAT expression is correlated with poor treatment outcome in breast cancer patients). These findings strongly suggest that ABAT is a useful and independent prognostic factor and its expression needs to be evaluated in breast cancer patients. This may be especially critical for determining which breast cancer patients may benefit from chemotherapy and which may not and should therefore avoid the unnecessary side effects of chemotherapy 55.

Treatment of BLBC represents a significant clinical challenge due to the lack of effective targeted agents and poor response to standard chemotherapy. Therefore, identification of novel molecular targets in BLBC is urgently needed. GABAergic signaling molecules may provide potential targets for controlling BLBC progression. Blockade of GABA or its receptors by small molecules or antibodies may provide a promising new approach to molecular therapy for BLBC. Notably, several inhibitors of GABA and GABA receptors are available for treatment of epilepsy. Whether these inhibitors are effective and safe in patients with BLBC needs to be investigated. Furthermore, CsA and FK506, as potent NFAT inhibitors that prevent its nuclear translocation, have been widely used as immunosuppressive agents in organ transplant to prevent rejection. Both agents might be promising targeted drugs for treating BLBC. Further elucidating GABAergic signaling and corresponding antagonistic drugs may help the GABAergic system become a valuable target for treating BLBC.

Conclusions

To summarize, we demonstrated that loss of ABAT drives BLBC progression by activating Ca2+-NFAT1 axis. Our data suggested that loss of ABAT-mediated GABAergic system is associated with the aggressive behavior of BLBC, providing potential prognostic indicators and therapeutic targets for BLBC.

Supplementary figures and tables.

Click here for additional data file.