microRNA regulation of the serine synthesis pathway in endocrine-resistant breast cancer cells

Abstract

Despite the successful combination of therapies improving survival of estrogen receptor α (ER+) breast cancer patients with metastatic disease, mechanisms for acquired endocrine resistance remain to be fully elucidated. The RNA binding protein HNRNPA2B1 (A2B1), a reader of N(6)-methyladenosine (m6A) in transcribed RNA, is upregulated in endocrine-resistant, ER+ LCC9 and LY2 cells compared to parental MCF-7 endocrine-sensitive luminal A breast cancer cells. The miRNA-seq transcriptome of MCF-7 cells overexpressing A2B1 identified the serine metabolic processes pathway. Increased expression of two key enzymes in the serine synthesis pathway (SSP), phosphoserine aminotransferase 1 (PSAT1) and phosphoglycerate dehydrogenase (PHGDH), correlate with poor outcomes in ER+ breast patients who received tamoxifen (TAM). We reported that PSAT1 and PHGDH were higher in LCC9 and LY2 cells compared to MCF-7 cells and their knockdown enhanced TAM-sensitivity in these-resistant cells. Here we demonstrate that stable, modest overexpression of A2B1 in MCF-7 cells increased PSAT1 and PHGDH and endocrine-resistance. We identified four miRNAs downregulated in MCF-7-A2B1 cells that directly target the PSAT1 3’UTR (miR-145–5p and miR-424–5p), and the PHGDH 3’UTR (miR-34b-5p and miR-876–5p) in dual luciferase assays. Lower expression of miR-145–5p and miR-424–5p in LCC9 and ZR-75–1-4-OHT cells correlated with increased PSAT1 and lower expression of miR-34b-5p and miR-876–5p in LCC9 and ZR-75–1-4-OHT cells correlated with increased PHGDH. Transient transfection of these miRNAs restored endocrine-therapy sensitivity in LCC9 and ZR-75–1-4-OHT cells. Overall, our data suggest a role for decreased A2B1-regulated miRNAs in endocrine-resistance and upregulation of the SSP to promote tumor progression in ER+ breast cancer.

Introduction

Breast cancer (BCa) is the most common malignant disease among women worldwide (Ciriello, et al. 2015) and the second most common cause of cancer mortality in women in the United States (Siegel, et al. 2023). Approximately 70% of primary breast tumors express estrogen receptor alpha (ERα) (Ring, et al. 2004). The current adjuvant treatments for ERα+ BCa target ERα either by competitively inhibiting the binding of estrogens to ERα using selective ER modulators (SERMs), e.g., tamoxifen (TAM) for premenopausal patients (Rani, et al. 2019) and reducing the level of estrogens available to bind ERα by inhibiting aromatase activity with aromatase inhibitors for postmenopausal women (Visvanathan, et al. 2019). A major limitation of current adjuvant endocrine therapies (ET) for ERα+ BCa patients is the development of acquired resistance in ~30–40% of patients- even up to 30 years after initial therapy (Dowsett, et al. 2005). Although combination targeted therapies with fulvestrant, e.g., everolimus, alpelisib, palbociclib, ribociclib, and abemaciclib, have demonstrated a significantly longer progression free survival advantage of ~ 11 months (Samuel Eziokwu, et al. 2020) in ~ 39.9% of patients (Lin, et al. 2020), these data indicate the number of patients who still die of metastatic disease is high. A variety of mechanisms have been implicated in the progression of endocrine-resistance (Clarke, et al. 2015; Fan and Jordan 2019), including PI3K mutations, epigenetic and metabolism changes, and altered expression of non-coding (nc) RNAs, e.g., miRNAs (Egeland, et al. 2015; Muluhngwi and Klinge 2015), but the complete understanding of endocrine-resistant metastatic spread remains to be determined (Petri and Klinge 2020; Rani, et al. 2019).

miRNAs are critical for post-transcriptional regulation of gene expression, but miRNA regulation and targets in oncogenesis and metastasis remain to be fully elucidated. miRNA biogenesis has been reviewed (Klinge 2015). The processing of primary-miRNAs (pri-miRs) to precursor-miRNAs (pre-miRs) is facilitated by the RNA binding protein HNRNPA2B1 (A2B1) (Alarcon, et al. 2015b). A2B1 acts as a reader of N(6)-methyladenosine (m6A), a common modification of mRNAs and pri-miRNAs (Alarcon, et al. 2015a). Previously, we reported higher A2B1 in tamoxifen (TAM) and fulvestrant -resistant, ERα+ LCC9 and LY2 cells derived from TAM- and fulvestrant- sensitive, ERα+ MCF-7 BCa cells (Petri, et al. 2021). Transient overexpression of A2B1 in MCF-7 cells reduced cell sensitivity to 4-hydroxytamoxifen (4-OHT) and fulvestrant, mimicking the endocrine-resistance of LCC9 and LY2 cells (Klinge, et al. 2019). Using non-targeted miRNA sequencing (miRNA seq), we identified miRNAs regulated by A2B1 in MCF-7 cells and used MetaCore pathway analysis to identify A2B1-regulated Gene Ontology (GO) pathways. One of the GO pathways identified as A2B1-regulated in MCF-7 cells was “serine family amino acid metabolic processes.”

Serine is a non-essential amino acid that fuels glycine synthesis and nucleotide metabolism (De Marchi, et al. 2017). The serine biosynthetic pathway (SSP) contributes to tumorigenesis and progression of BCa, although the mechanisms by which serine supports tumor progression and metastasis remain to be fully elucidated (Geck and Toker 2016). This pathway converts 3-phosphoglyerate into serine in a series of three reactions starting with rate-limiting step catalyzed by D-3-phosphoglycerate dehydrogenase (PHGDH) (Supplementary Figure 1), an enzyme upregulated by MYC, which is overexpressed in ER+ BCa (Fallah, et al. 2017). MYC is suppressed by p53 (Ou, et al. 2015), which is mutated in triple negative BCa (TNBC) (Ji, et al. 2022). Inducing expression of PHGDH in mice generated with a PHGDH tetO allele allowing for tissue-specific, doxycycline-inducible PHGDH expression resulted in accelerated breast tumor growth (Sullivan, et al. 2019). The second rate-limiting step in the SSP is a transamination reaction that converts 3-phospho-hydroxypyruvate to 3-phospho-serine and is catalyzed by phosphoserine aminotransferase (PSAT1). Dephosphorylation of 3-phospho-serine to serine is catalyzed by phosphoserine phosphatase (PSPH). The expression of PSAT1 is higher in TAM-resistant breast tumors and was reported to enhance cancer cell proliferation, metastasis, and TAM-resistance (De Marchi, et al. 2017; Gao, et al. 2017; Martens, et al. 2005a). Although PHGDH and PSAT1 are increased in breast tumors, the mechanism for this upregulation and roles for these proteins and serine synthesis in endocrine-resistance and/or metastasis is not yet understood. We reported increased PSAT1 expression was positively correlated with higher tumor grade in TNBC tumor samples (Metcalf, et al. 2020a). In addition, we found that both PSAT1 and PHGDH transcript expression correlated with reduced relapse-free survival in ER+ BCa patients treated with TAM (Metcalf, et al. 2020b). We recently reported higher levels of PHGDH and PSAT1in LCC9 and LY2 TAM-resistant cells compared to the parental MCF-7 cell line (Metcalf, et al. 2020b). PSAT1 overexpression in MCF-7 TAM-sensitive cells and transient PSAT1 knockdown in LCC9 TAM-resistant cells altered TAM-responses, i.e., inhibiting and enhancing TAM-induced cell growth inhibition (Metcalf, et al. 2020b). These data suggest a role for elevated PSAT1 in TAM-resistance (Metcalf, et al. 2020b). Downregulation or pharmacological inhibition of PHGDH increased TAM-sensitivity in LCC9 cells. We also demonstrated that shPSAT1 reduced MDA-MB-231 TNBC cell migration and lung tumor metastasis in vivo after tail vein injection in mice (Metcalf, et al. 2020a). However, knockdown of PHGDH had no impact on these parameters, suggesting that PSAT1 increases migratory potential in TNBC cells, independent of its role in serine synthesis (Metcalf, et al. 2020a). It is important to note that a number of metabolic enzymes have cellular roles in addition to their metabolic activities (Xu, et al. 2021) (Pan, et al. 2021).

The aim of this study was to determine if miRNAs decreased with A2B1 overexpression in MCF-7 cells, computationally associated with mRNA targets in the gene ontology/biological process (GO:BP) “serine family amino acid metabolic processes” (Klinge, et al. 2019), play a role in the increased expression of PSAT and PHGDH detected in TAM-resistant cell lines. We tested the hypothesis that A2B1 downregulates miRNAs that directly target PHGDH and PSAT1, resulting in increased protein levels. We identified A2B1-regulated miRNAs, miR-145–5p and miR-424–5p, that directly target PSAT1, and miR-34b-5p and miR-876–5p, that directly target PHGDH. Transfection of miR-145–5p and miR-424–5p reduced endogenous PSAT1 protein expression in some BCa cell lines; however, the effects of transfected miR-34–5p and miR-875–5p on PHGDH protein expression were cell line-selective. Functionally, we demonstrated that the A2B1-downregulated miRs increased endocrine-sensitivity on endocrine-resistant LCC9 cells.

Materials and Methods

Chemicals

4-hydroxytamoxifen (4-OHT) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Fulvestrant was purchased from Tocris Bioscience (Ellisville, MO, USA). Both were dissolved in DMSO, which was used as the vehicle control in experiments.

Cell culture and treatments

MCF-7, T47D, and ZR-75–1 BCa cells were purchased from ATCC (Manassas, VA, USA) and were used within 9 passages. MCF-7 cells were maintained in Modified IMEM (Gibco, ThermoFisher Scientific, Waltham, MA, USA) with 10% fetal bovine serum (FBS) and 1% Pen/Strep (P/S). T47D cells were maintained in RPMI 1640 (GenClone Cat# 25–506, Genesee Scientific, El Cajon, CA, USA) supplemented with 5% FBS, 6 μg/ml insulin (Sigma-Aldrich), and 1% P/S. ZR-75–1 cells were maintained with RPMI 1640 (GenClone Cat# 25–506, Genesee Scientific) with 10% FBS and 1% P/S. ZR-75–1 cells were exposed to 100 nM 4-OHT for ~20 weeks to create ZR-75–1-4-OHT cells which were maintained in half ZR-75–1 media and half Improved MEM (Richter’s Media) (Corning, Corning, NY, USA) with 5% dextran-coated charcoal-stripped FBS and 1% P/S. LCC9 and LY2 tamoxifen/fulvestrant endocrine-resistant BCa cells were generously provided by Dr. Robert Clarke, Georgetown University Medical Center (Brunner, et al. 1997; Crawford, et al. 2010; Davidson, et al. 1986) and maintained with Modified IMEM (Gibco) with 5% FBS and 1% P/S. MCF-7-A2B1 cells were stably transfected and G418-selected with pcDNA3.1+C-DYK into which A2B1 was cloned (GenScript, Piscataway, NJ, USA) as described (Klinge, et al. 2019). All cell lines were verified by short tandem repeat (STR) genotyping (Genetica, LabCorp, Burlington, NC, USA). STR profiles were compared with publically available profiles using Cellosaurus STR (ExPASy).

Transient transfection

Silencer^™Select siHNRNPA2B1 (#4390824) and Negative Control No. 1 (#4390843) were purchased from Thermo Fisher Scientific. Prior to transfection, cells were grown in phenol red-free Opti-MEM (cat. # 11058021, Thermo Fisher Scientific) for ~ 18–24 h. Lipofectamine RNAiMAX reagent (cat # 13778075 Thermo Fisher Scientific) was used for transfection per the manufacturer’s protocol. LCC9, LY2, T47D, ZR-75–1 and ZR-75–1-4-OHT cells were transiently transfected for 48h or 72h with pre-miR-145–5p, pre-miR-424–5p, pre-miR-34b-5p, or pre-miR-876–5p (Pre-miR^™s, Ambion, Austin, TX, USA), and anti-miR-145–5p, anti-miR-424–5p, anti-miR-34b-5p, or anti-miR-876–5p (Anti-miR^™s, Ambion) using Lipofectamine3000 (Ref # L3000–015, ThermoFisher) in Opti-MEM^® Reduced Serum Medium (Invitrogen, Carlsbad, CA, USA). Negative controls were Anti-miR ^™ negative control #1 (Ambion) and Pre-miR^™ negative control #1 (Ambion).

RNA extraction and quantitative real-time PCR (qPCR)

RNA and miRNA were extracted from cells using the RNeasy and miRNeasy Mini Kits, respectively (Qiagen, Gaithersburg, MD, USA). RNA concentration and quality and reverse transcription for cDNA used reagents and followed previously published (Klinge, et al. 2019). Quantitative real-time PCR (qPCR) for PHGDH and PSAT1 was performed using TaqMan assays (ThermoFisher). 18S rRNA and GAPDH were used as normalizers. qPCR for miR-145–5p and miR424–5p used TaqMan assays and were normalized to RNU6B (Thermo Fisher Scientific). qPCR was performed using an ABI Viia 7 Real-Time PCR system (LifeTechnologies) with each reaction run in triplicate. The comparative threshold cycle (C_T) method was used to determine ΔC_T, ΔΔC_T, and fold-change, log 2, relative to control (Schmittgen and Livak 2008).

MTT assays

Cells were “serum starved” for 48 hours in IMEM supplemented with 5% DCC-FBS (dextran-coated charcoal stripped fetal bovine serum). Cells were treated every 48 hours with vehicle control (DMSO), E₂ (estradiol), 4-OHT, or fulvestrant (ICI 182,780) at the concentrations and number of days indicated in Figure legends. MTT assays used CellTiter (Promega). Each treatment was in quadruplicate within each experiment.

Western blotting

Whole cell extracts (WCE) were prepared as described (Radde, et al. 2016) and separated on 10% SDS-PAGE, transferred to PVDF membranes (Bio-Rad Laboratories, Hercules, CA, USA) and immunoblotted with antibodies: HNRNPA2B1 (B1 epitope-specific: IBL # 18941, IBL America, Minneapolis, MN, USA), PHGDH (cat # HPA021241, Sigma-Aldrich), PSAT1 (cat # 10501–1-AP, Proteintech, Rosemont, IL, USA), α-tubulin (Thermo Fisher Scientific # MS-81-P1), and GAPDH (Santa Cruz Biotechnology, Dallas, TX, USA cat # sc-365062). Membranes were washed with TBS-Tween followed by incubation with anti-mouse (#7076S) or anti-rabbit (#7074S) (Cell Signaling Technology, Danvers, MA, USA) secondary antibodies. Membranes were incubated with Clarity Western ECL (Bio-Rad) and imaged on a Bio-Rad ChemiDoc^™ XRS+ System with Image Lab^™ Software (Bio-Rad). Blots were stained with Ponceau S (Moritz 2017; Romero-Calvo, et al. 2010) or Amido Black (Goldman, et al. 2016) for additional quantification. We observed that the protein levels of GAPDH and α-tubulin varied between some of the breast cancer cell lines and thus we used Ponceau S or Amido Black for quantification where indicated in the Figure legends.

Immunohistochemistry

Tissue microarray (TMA) including samples from 48 patients with histologically confirmed estrogen receptor-positive BCa was purchased from Reveal Biosciences (San Diego, CA). The tissue sections were dried at 57°C for 30 min and subsequently dewaxed with Xylene and decreasing concentrations of ethanol. Antigen retrieval was performed in 10mM sodium citrate buffer (pH 6.0) at 100 °C for 2.5 h. Slides were washed with 1x PBS and blocked with 10% goat serum for 1 h at room temperature followed by an additional wash with 1x PBS. Each section was incubated with the polyclonal primary rabbit antibody against PSAT1 (cat # 10501–1-AP, Proteintech) or HNRNPA2B1 (IBL # 18941, IBL America) at a 1:200 dilutions overnight at 4 °C. After washing with PBS (3 × 5 min). Slides were subsequently incubated with 1:500 dilution of horseradish peroxidase (HRP)-conjugated anti-rabbit secondary antibody (Invitrogen, Cat No.32260) at room temperature for 30 min. After another wash in PBS (3 × 5 min), each section was immersed in 500 μl of diaminobenzidine (DAB) working solution at room temperature for 1.5 min. Finally, the slides were counterstained with hematoxylin and mounted in crystal mount medium. Images were captured using a Nikon Eclipse Ci microscope with dedicated Nikon DS-Fi3 microscope camera (Tokyo, Japan) for higher magnification. Semi-quantitative measures were performed by two independent observers. Cytoplasmic expression of PSAT1 was evaluated based on intensity of staining.

Statistical analysis

Prism^™ software (version 9, GraphPad Inc.) was used to perform data transformations and plotting, linear and nonlinear regression, and statistical analyses. Where two data sets are compared, Student’s two-tailed t-test was used. Where more than two data sets were compared, one way ANOVA followed by Tukey’s multiple comparison post-hoc test or two-way ANOVA followed by Bonferroni post-hoc test were performed. P values are provided in Figures with p < 0.05 considered significant.

Results

PSAT1 is increased in breast tumors from patients with ERα+ BCa

Higher PSAT1 mRNA levels and protein expression was reported in primary breast tumors compared to normal breast tissue (De Marchi, et al. 2017). Lower PSAT1 and PHGDH transcript expression is associated with better prognosis in BCa patients (Mascia, et al. 2022) (Supplementary Figure 2A and 2B). IHC was performed to examine PSAT1 staining in a TMA containing primary breast tumors paired with normal breast tissue or lymph node metastases (LNM) from the same patient (Figure 1A–D). Strong cytoplasmic PSAT1 staining was identified in the epithelial cells of normal breast tissue (Nl. Br.), invasive ductal carcinomas (IDC), and in LNM. PSAT1 expression intensity was higher in the ER+ IDC tumors than the paired normal ER+ breast tissue (8.3% of stained cells with an intensity score of 3) (Figure 1A and 1B) and was lower in LNM compared to normal ER+ breast tissue (60% of stained cells with an intensity score of 1) (Figure 1B and 1D). Similarly, PSAT1 staining intensity was lower in ER+, PR+, HER2- LNM compared to IDC from the same molecular subtype (Figure 1E). These results show that PSAT1 expression is higher in ER+ invasive breast tumors compared to normal breast tissue, although the increased PSAT1 expression does not extend to associated LNM.

Figure 1. IHC staining for PSAT1 in ER+ invasive breast carcinoma (IDC) tumors paired with normal breast or lymph node (LN) metastasis from the same patient.

Shown are representative images from a stained TMA. The tumor in A is paired with normal breast (nl. Br.) tissue (B) from the same patient. The tumor in C is paired with its LN metastasis in D. Bar in 200 μm. Images were taken under 20X and 40X magnification. E) Quantification based on 100% staining and intensity of cells stained are indicated as the percent total number of that sample in the TMA.

Stable HNRNPA2B1-overexpressing MCF-7 cells have increased PSAT1 and PHGDH protein expression.

We recently reported higher expression of PSAT1 and PHGDH in LCC9 and LY2 TAM-resistant cell lines derived from parental luminal A, endocrine-sensitive MCF-7 human BCa cells (Metcalf, et al. 2020b). We also demonstrated that stable overexpression of A2B1 in MCF-7 cells (MCF-7-A2B1 cells) results in TAM-resistance (Petri, et al. 2021). We examined the protein expression of PSAT1 and PHGDH in TAM-resistant MCF-7-A2B1, LCC9, and LY2 cells (Figure 2A). MCF-7-A2B1 cells showed higher expression of PSAT1 and PHGDH protein compared to MCF-7 cells with values similar to those observed in LCC9 and LY2 cells (Figure 2B and 2C). We investigated the effect of A2B1 knockdown (Supplementary Figure 3A) (Petri, et al. 2021) on the expression of PSAT1 and PHGDH transcript and protein in A2B1 cells and observed that siA2B1 reduced PSAT1 and PHGDH transcripts (Supplementary Figure 3B and 3C) and protein (Supplementary 3D and 3E), suggesting a regulatory role for A2B1 in PSAT1 and PHGDH expression.

Figure 2. PSAT1 and PHGDH expression in breast cancer cell lines.

A) WCE were prepared from MCF-7, MCF-7-A2B1, LCC9, and LY2 cells and probed for PSAT1, then stripped and re-probed for PHGDH and for α-tubulin. B and C) Individual values and the mean ± SEM for PSAT1/α-tubulin and PHGDH/α-tubulin. *p ≤ 0.01, one-way ANOVA, Tukey’s post hoc multiple comparisons test. D) WCE were prepared from the indicated breast cancer cell lines and probed for PSAT1, then stripped and re-probed for PHGDH. The blot was stained with Ponceau S and with Amido Black for protein normalization. Shown is one representative western blot of four separate blots. E and F) Values of individual cell lysates from separate cell lysates as biological replicates. Data were analyzed by one-way ANOVA followed by Tukey’s multiple comparisons post-hoc test. * p < 0.05; **p < 0.01, **p < 0.001; ****p < 0.0001.

To determine the expression of SSP enzyme proteins PSAT1 and PHGDH in luminal A BCa cell lines unrelated to MCF-7 cells, we selected T47D (a p53 mutant cell line) and ZR-75–1 (no p53 or other established BCa-associated mutations) (Kenny, et al. 2007). PSAT1 and PHGDH protein levels were higher in T47D and ZR-75–1 than MCF-7 cells, but lower than in LCC9 and LY2 TAM-resistant BCa cell lines (Figure 2D, E and F). To develop a TAM-resistant ZR-75–1 cell line (ZR-75–1-4OHT), we grew ZR-75–1 cells in medium containing 100 nM 4-OHT for > 8 mos. These cells have higher basal proliferation and are TAM- and fulvestrant-resistant compared to ZR-75–1 parental cells (Supplementary Figure 4A and B). ZR-75–1-4-OHT cells have higher PSAT1 and PHGDH protein expression compared to the parental ZR-75–1 cell line (Figure 2D–2F). Compared to MCF-7, all the BCa cell lines showed higher PSAT1 transcript abundance (Supplementary Figure 5A); however, only MCF-7A2B1 cells showed higher PHGDH transcript abundance (Supplementary Figure 5B). The difference between PHGDH protein and mRNA transcript abundance in these BCa cell lines suggest post-transcriptional modulation of PHGDH protein expression.

A2B1 expression decreases miRNAs predicted to target PSAT1 and PHGDH

Gene ontology identified “serine family amino acid metabolic processes” as a pathway regulated by A2B1 overexpression in MCF-7 cells (Klinge, et al. 2019). To determine if A2B1 regulates PSAT1 and PHGDH expression by downregulating miRNAs that target the 3’UTR of PSAT1 and PHGDH, we searched miRTarBase for miRNAs that target each transcript. There were no verified miRNAs targeting PSAT1 or PHGDH in that database, despite the reports that miR-340 (Yan, et al. 2015), miR-424 (Fang, et al. 2018), miR-195–5p (Wang, et al. 2023), and miR-145–5p (Ding, et al. 2022) directly target PSAT1. For PHGDH, miR-107 and miR-103a-3p were reported to directly target PHGDH (Zhang, et al. 2021). When we compared miRNAs predicted to target PSAT1 and PHDGH with the A2B1-regulated miRNAs in MCF-7 cells (Klinge, et al. 2019) using mirDIP 4.1 (Tokar, et al. 2018), we identified 21 miRNAs predicted to target the 3’UTR of PSAT1 that were downregulated after transfection of MCF-7 cells with A2B1 (Table 1). Four miRNAs had either ‘very high’ (miR-4790–3p) or ‘high’ (miR-424–5p, miR-145–5p, and miR-6794–3p) confidence according to mirDIP prediction algorithms (Tokar, et al. 2018). We selected miR-145–5p and miR-424–5p for further analysis since they had ‘high’ confidence and a high integration score (0.68 for miR-145–5p and 0.75 for miR-424–5p) (Tokar, et al. 2018). Further, miR-145–5p and miR-424–5p have been reported to regulate other targets in BCa (Petri and Klinge 2020). We found 19 A2B1-downregulated miRs predicted to target PHGDH mRNA with mirDIP (Table 2). Three miRNAs (miR-34b-5p, miR-6801–5p, and miR-937–3p) were rated with ‘high’ confidence (Tokar, et al. 2018). We selected miR-34b-5p and miR-876–5p for their relevance in BCa. miR-34b-5p was reported to inhibit BCa progression by targeting ARHGAP1 (Dong, et al. 2020) and miR-876–5p inhibited BCa cell proliferation and invasion by targeting TFAP2A (Xu, et al. 2019). Figure 3A shows the alignment of miR-145–5p, miR-424–5p, miR-34b-5p, and miR-876–5p with their miRNA response elements (MREs) in the 3’UTR of PSAT1 or PHGDH. We examined whether the expression of the selected miRNAs associate with overall survival in BCa using Kaplan Meier (KM) plotter (Gyorffy, et al. 2010). These data suggest that low expression of miR-145–5p and miR-424–5p is associated with reduced overall survival (OS) of all BCa patients (Fig. 3B and 3C). The association of reduced miR-34b-5p with OS was not significant and high expression of miR-876–5p was associated with lower OS (Figure 3D – 3E).

Table 1. A2B1-downregulated miRNA predicted to target the 3’UTR of PSAT1 mRNA.

Using mirDIP 4.1 prediction tools we identified human miRNAs downregulated in MCF-7 cells overexpressing A2B1 (miRNA-seq. FC > 0.50, p-value <0.05). The miRNA selected are predicted to bind to the MRE in the 3’UTR of PSAT1mRNA with very high, high, or medium confidence. Also shown are the integration score, number of sources that identify PSAT1 as a target of corresponding miRNA, and the sources that have identified the relationship.

miRNA	hr post TF	logFC	P value	Integration score	# of Sources	Score Class	Sources
miR-4790–3p	48	−1.66	4.79E-02	0.33	7	Very High	bitargeting_May_2021|DIANA|MBStar|MirAncesTar|miranda_May_2021|miRDB_v6|mirzag
miR-424–5p	48	−1.18	8.96E-03	0.75	14	High	BCmicrO|DIANA|MBStar|MirAncesTar|miranda_May_2021|miRcode|miRDB_v6|mirmap_May_2021|MiRNATIP|miRTar2GO|MirTar2|mirzag|MultiMiTar|TargetScan_v7_2
miR-145–5p	48	−1.85	1.17E-02	0.68	10	High	BCmicrO|CoMeTa|MirAncesTar|miranda_May_2021|miRcode|miRDB_v6|MiRNATIP|mirzag|RNA22|TargetScan_v7_2
miR-6794–3p	48	−2.57	4.35E-02	0.13	2	High	MirAncesTar|mirzag
miR-17–5p	48	−0.77	1.06E-02	0.38	8	Medium	BCmicrO|bitargeting_May_2021|MBStar|MirAncesTar|miranda_May_2021|mirbase|MiRNATIP|miRTar2GO
miR-20a-5p	48	−0.64	4.35E-02	0.37	9	Medium	bitargeting_May_2021|DIANA|MBStar|MirAncesTar|miranda_May_2021|mirbase|MiRNATIP|miRTar2GO|RNA22
miR-3125	48	−3.38	3.27E-03	0.32	10	Medium	bitargeting_May_2021|DIANA|MBStar|MirAncesTar|miranda_May_2021|mirmap_May_2021|MiRNATIP|mirzag|PITA_May_2021|RNA22
miR-224–5p	48 & 72	−2.86	1.63E-02	0.31	8	Medium	BCmicrO|DIANA|MBStar|MirAncesTar|miranda_May_2021|miRDB_v6|MiRNATIP|miRTar2GO
miR-193a-3p	48	−0.82	1.72E-02	0.27	4	Medium	BCmicrO|MirAncesTar|miRcode|miRTar2GO
miR-5008–3p	48 & 72	−2.63	2.93E-03	0.20	6	Medium	MBStar|MirAncesTar|miranda_May_2021|miRDB_v6|MiRNATIP|mirzag
miR-520g-3p	48	−2.63	2.83E-02	0.17	5	Medium	BCmicrO|MBStar|MirAncesTar|miranda_May_2021|MiRNATIP
miR-934	48	−1.46	3.52E-03	0.17	4	Medium	BCmicrO|DIANA|MirAncesTar|mirzag
miR-2053	72	−3.82	1.46E-03	0.15	5	Medium	DIANA|MBStar|MirAncesTar|miranda_May_2021|mirCoX
miR-497–3p	48	−0.66	4.08E-02	0.13	5	Medium	DIANA|MirAncesTar|miranda_May_2021|mirmap_May_2021|mirzag
miR-518d-5p	48 & 72	−2.37	4.98E-02	0.13	5	Medium	BCmicrO|bitargeting_May_2021|MBStar|MirAncesTar|miranda_May_2021
miR-520c-5p	48 & 72	−2.37	4.98E-02	0.13	5	Medium	BCmicrO|bitargeting_May_2021|MBStar|MirAncesTar|miranda_May_2021
miR-1283	48 & 72	−2.27	2.51E-02	0.10	5	Medium	BCmicrO|MBStar|MirAncesTar|miranda_May_2021|MiRNATIP
miR-486–5p	48 & 72	−1.24	2.52E-02	0.10	4	Medium	DIANA|MirAncesTar|MiRNATIP|mirzag
miR-4720–5p	72	−3.05	7.52E-03	0.08	4	Medium	bitargeting_May_2021|MirAncesTar|miranda_May_2021|RNA22
miR-3118	72	−3.99	2.77E-04	0.08	3	Medium	MBStar|MirAncesTar|miranda_May_2021
let-7i-3p	48	−0.76	1.18E-02	0.04	2	Medium	MBStar|MirAncesTar

Table 2. A2B1-downregulated miRNA predicted to target the 3’UTR of PHGDH mRNA.

Using mirDIP 4.1 prediction tools we identified human miRNAs downregulated in MCF-7 cells overexpressing A2B1 (miRNA-seq. FC > 0.50, p-value <0.05). The miRNA selected are predicted to bind to the MRE in the 3’UTR of PHGDH mRNA with very high, high, or medium confidence. Also shown are the integration score, number of sources that identify PHGDH as a target of corresponding miRNA, and the sources that have identified the relationship.

miRNA	hr post TF	logFC	P value	Integration score	# of Sources	Score Class	Sources
miR-34b-5p	48	−2.20	2.90E-02	0.38	11	High	bitargeting_May_2021|DIANA|MirAncesTar|miranda_May_2021|mirbase|mirmap_May_2021|MiRNATIP|miRTar2GO|PITA_May_2021|RNA22|rnahybrid_May_2021
miR-6801–5p	72	−3.01	1.42E-06	0.20	8	High	bitargeting_May_2021|MirAncesTar|miranda_May_2021|mirmap_May_2021|mirzag|PITA_May_2021|RNA22|rnahybrid_May_2021
miR-937–3p	48	−2.17	3.69E-02	0.21	6	High	BCmicrO|bitargeting_May_2021|MirAncesTar|miranda_May_2021|miRTar2GO|PITA_May_2021
miR-150–5p	72	−2.83	2.39E-02	0.22	9	Medium	BCmicrO|bitargeting_May_2021|DIANA|MirAncesTar|miranda_May_2021|miRcode|mirCoX|mirmap_May_2021|RNA22
miR-34c-5p	48	−2.36	8.62E-03	0.27	6	Medium	BCmicrO|MirAncesTar|miranda_May_2021|mirbase|miRTar2GO|RNA22
miR-876–5p	72	−3.05	1.83E-02	0.19	6	Medium	BCmicrO|DIANA|MirAncesTar|mirbase|mirzag|RNA22
miR-193a-3p	48	−0.82	1.72E-02	0.14	4	Medium	BCmicrO|bitargeting_May_2021|miranda_May_2021|miRcode
miR-5088–5p	72	−2.95	1.80E-02	0.13	6	Medium	bitargeting_May_2021|MirAncesTar|miranda_May_2021|mirmap_May_2021|MiRNATIP|RNA22
miR-3663–5p	48	−2.96	2.38E-02	0.09	5	Medium	bitargeting_May_2021|MirAncesTar|miranda_May_2021|MiRNATIP|RNA22
miR-518d-5p	48 & 72	−2.37	4.98E-02	0.09	4	Medium	BCmicrO|mirbase|mirmap_May_2021|RNA22
miR-6789–5p	72	−3.05	5.93E-03	0.09	5	Medium	bitargeting_May_2021|MirAncesTar|mirmap_May_2021|PITA_May_2021|RNA22
miR-6794–3p	48	−2.57	4.35E-02	0.09	5	Medium	bitargeting_May_2021|MirAncesTar|MiRNATIP|PITA_May_2021|RNA22
miR-1271–3p	72	−3.37	2.15E-03	0.08	4	Medium	bitargeting_May_2021|MirAncesTar|miranda_May_2021|RNA22
miR-5682	48	−2.37	4.98E-02	0.07	4	Medium	bitargeting_May_2021|MirAncesTar|miranda_May_2021|RNA22
miR-6733–3p	48	−1.74	3.49E-02	0.06	3	Medium	bitargeting_May_2021|MirAncesTar|RNA22
miR-6856–3p	72	−3.38	1.74E-03	0.06	3	Medium	MirAncesTar|miranda_May_2021|mirmap_May_2021
miR-4767	48	−1.63	3.97E-02	0.06	4	Medium	MirAncesTar|mirmap_May_2021|PITA_May_2021|RNA22
miR-7975	48	−2.00	1.86E-02	0.06	3	Medium	bitargeting_May_2021|miranda_May_2021|RNA22
miR-6872–5p	72	−2.86	1.28E-03	0.04	2	Medium	bitargeting_May_2021|RNA22

Figure 3. A2B1-downregulated miRs are predicted to target PSAT1 and PHGDH mRNA.

miR-424–5p, miR-145–5p, miR-34b-5p, and miR-876–5p are downregulated in MCF-7-A2B1 cells (Klinge, et al. 2019) (Tables 1 and 2). A) The seed element and MRE for selected miRNA in the 3’UTR of PSAT1 and PHGDH. B-E) Kaplan-Meier Plotter was used to examine the role of miR-145–5p, miR-424–5p, miR-34b-5p, and miR876–5p expression in overall survival of breast cancer patients. Low expression of miR-145 and miR-424 is significantly associated with lower overall survival (OS).

MCF-7-A2B1, LCC9, and LY2 cells have lower miR-145–5p, miR-424–5p, miR-34b-5p, and miR-876–5p expression compared to MCF-7 cells

To confirm the downregulation of miR-145–5p, miR-424–5p, miR-34b-5p, and miR-876–5p by A2B1 in MCF-7 cells, RT-qPCR was performed. The expression of miR-145–5p and miR-424–5p was significantly reduced in MCF-7-A2B1 cells compared to parental MCF-7 cells (Figure 4A–4B), verifying that modest, stable A2B1 overexpression reduces their expression as seen initially in miRNA-seq data from A2B1-transiently transfected MCF-7 cells (Klinge, et al. 2019). In contrast, miR-34b-5p expression was higher in MCF-7-A2B1 cells while miR-876–5p expression was unaltered in MCF-7-A2B1 cells compared to MCF-7 cells, suggesting differences between A2B1 transient and stable A2B1 regulation of these miRNAs (Figure 4C and 4D). TAM-resistant LCC9 and LY2 cells have higher endogenous A2B1 compared to MCF-7 cells (Klinge, et al. 2019). If A2B1 downregulates miR-145–5p, miR-424–5p, miR-34b-5p, and miR-876–5p, we expect lower expression of these miRNAs in LCC9 and LY2 cells relative to MCF-7 cells. As predicted, compared to MCF-7 cells, LCC9 and LY2 cells have significantly lower abundance of miR-145–5p, miR-424–5p, and miR-876–5p (Figure 4E, 4F, and 4H). However, miR-34b-5p transcript abundance was lower only in the LY2 cells (Figure 4G), suggesting factors in addition to A2B1 regulate miR-34–5p abundance in LCC9 cells. Taken together, these data show reduced miR-145–5p, miR-424–5p, and miR-876–5p in both LCC9 and LY2 TAM-resistant cells and that miR-34b-5p is lower in LY2 cells compared to ET-sensitive MCF-7 cells.

Figure 4. Examination of miR-424–5p, miR-145–5p, miR-34b-5p, and miR-876–5p expression in MCF-7-A2B1, LCC9, and LY2 cells.

qPCR was performed on independent samples of miRNA isolated from MCF-7, MCF-7-A2B1, LCC9, and LY2 breast cancer cells. Values were normalized by RNU48. Data in A and B were evaluated by Student’s unpaired t test, ** p < 0.01, *** p < 0.001, **** p < 0.0001. Data in E, F, G, and H were evaluated by one-way ANOVA followed by Tukey’s multiple comparison test.

Validation of PSAT1 and PHDGH as bona fide targets of A2B1-regulated miRNAs

To determine whether PSAT1 and PHGDH are bona fide targets of miR-145–5p and miR-424–5p and miR-35b-5p and miR-876–5p, respectively, we tested the ability of each miRNA to repress luciferase activity from reporter vectors containing the 3’UTR of PSAT1 or PHGDH. Co-transfection of pre-miR-145–5p and pre-miR-424–5p repressed luciferase activity from the reporter containing the PSAT1 3’UTR (Figure 5A). Similarly, co-transfection of pre-miR-34b-5p and pre-miR-876–5p inhibited luciferase reporter activity from the construct containing the PHGDH 3’UTR (Figure 5B). Co-transfection of pre-miR-145–5p with a luciferase reporter containing the miR-145–5p MRE from the 3’UTR of PSAT1 also repressed luciferase activity; however, miR-145–5p did not inhibit luciferase activity from the reporter containing the MRE for miR-424–5p (Figure 5C). Conversely, miR-424–5p inhibited luciferase activity from the miR-424–5p MRE in PSAT1, but not the miR-145–5p MRE in PSAT1 (Figure 5D). While miR-34b-5p and miR-876–5p inhibited luciferase activity when co-transfected with the reporter containing the respective MRE from the PHGDH 3’UTR, miRNA inhibition was not significantly blocked when co-transfected with a non-target MRE reporter (MRE 424) (Figure 5E and 5F). These data demonstrate the specificity of the miR-MRE recognition and direct miR-MRE interaction for miR-145–5p and miR-424–5p for targeting PSAT1; however, the ability of miR-34b-5p and miR-876–5p to directly target their putative MREs in the 3’UTR of PHGDH is not fully supported. Together, these data validate that PSAT1 is a bona fide direct target of miR-145–5p and miR-424–5p.

Figure 5. A2B1-downregulated miRNAs regulate luciferase report activity.

A and B) HEK-293T cells were transiently transfected with pEZX-MT06 containing the full length 3’UTR of PSAT1 or PHGDH, pre-miR-Control (Cont), pre-miR-145–5p, pre-miR-424–5p, pre-mR-846–5p, or pre-miR-34b-5p, as indicated. C-F) HEK-293T cells were transiently transfected with pmirGLO containing the indicated MRE sequence from PSAT1 or PHGDH and the indicated pre-miR-Control or miR as in A-B. After 48 h, dual luciferase assays were performed with the FF/Renilla ratios normalized to the control transfection. n = 5 separate experiments. Statistical analysis used One-way ANOVA followed by Tukey’s multiple comparison test: *p ≤ 0.05, **p<0.01, ****p ≤ 0.001

miR-145–5p and miR-424–5p regulate endogenous expression of PSAT1 in ER+ BCa cells.

To investigate if miR-145–5p and miR-424–5p regulate endogenous PSAT1 expression in TAM-resistant BCa cells with high PSAT1, we transfected pre-miR-145–5p or pre-miR-424–5p into LCC9 cells. Increased miR-145–5p and miR-424–5p was detected in the transfected LCC9 cells (Supplementary Figure 6A and 6B). Only the transfection of pre-miR-424–5p significantly reduced PSAT1 mRNA abundance (Figure 6A); however, PSAT1 protein abundance was decreased 48 h post-transfection of either pre-miR-145–5p or pre-miR-424–5p (Figure 6B–D). Further, this decrease was inhibited by co-transfection of anti-miR-145–5p or anti-miR-424–5p (Figure 6B–D). The effect on PSAT1 protein abundance was lost at 72h post-transfection with pre-miR-145–5p and pre-miR-424–5p, suggesting a transient miRNA activity (Supplementary Figure 6C–CE). We transfected pre-miR-34b or pre-miR-876b, putatively targeting PHGDH (Figure 4) into LCC9 cells and increased miR-34b and miR-876–5p was detected (Supplementary Figure 7A and 7B). Transfection of LCC9 cells with pre-miR-34b-5p or pre-miR-876–5p did not reduce PSAT1 protein 48 h or 72 h post-transfection (Supplementary Figure 7C–7F), demonstrating specificity. Thus, both miR-145–5p and miR-424–5p specifically regulate endogenous PSAT1 protein abundance in TAM-resistant LCC9 cells. Transfection of pre-miR-145–5p and pre-miR-424–5p (Supplementary Figure 8A and 8B) did not alter PSAT1 protein expression in LY2 cells at 48 h or 72h (Supplementary Figure 8C–8H), suggesting cell-specific miRNA-mediated regulation of PSAT1 abundance.

Figure 6. miR-145–5p and miR-424–5p target PSAT1 in LCC9, ZR-751, and ZR-75–1-4-OHT cells.

A) LCC9 cells were transfected with pre-miR-control pre-miR-145–5p, or pre-miR-424–5p and RT-qPCR was performed on RNA isolated 48 h after transfection. Values were normalized by 18S. Values are the mean +/−SEM of 3 independent, biological replicate experiments. B) WCE were prepared from LCC9 cells 48 h post transfection with as pre-miR-control pre-miR-145–5p, pre-miR-424–5p or co-transfection with anti-miR-145–5p or anti-miR-424–5p, and probed with a PSAT1 antibody. The blots were stained with Ponceau S as a protein loading control. C and D) Analysis of western blots from three independent samples. Values were normalized to LCC9 miR control. Values are normalized to GAPDH or Ponceau S. E) WCE were prepared from ZR-75–1 cells 72 h post transfection with pre-miR-145–5p, pre-miR-424–5p, or co-transfection with anti-miR-145–5p or anti-miR-424–5p and probed with a PSAT1 antibody. F and G) Analysis of western blot performed on 3 independent samples. Values were normalized to ZR-75–1 miRNA control H) WCE were prepared from ZR-75–1 4-OHT cells 72 h post transfection with pre-miR-145–5p, pre-miR-424–5p, or co-transfection with anti-miR-145–5p or anti-miR-424–5p and probed with a PSAT1 antibody. I and J) Analysis of western blots performed on 3 independent samples. Values are normalized to GAPDH or Ponceau S. Statistical analysis used one-way ANOVA followed by Tukey’s multiple comparison test: *p ≤ 0.05, **p<0.01,***p ≤ 0.001

Since T47D cells showed higher PSAT1 protein and mRNA abundance relative to MCF-7 (Figure 2D), we also investigated the effects of miR-145–5p and miR-424–5p on endogenous PSAT1 expression. We transfected pre-miR-145–5p or pre-miR-424–5p into T47D cells. Increased miR-145–5p and miR-424–5p was detected in the transfected cells (Supplementary Figure 9A–9B). miR-424–5p decreased PSAT1 protein and the reduction in PSAT1 by miR-424–5p was abrogated by co-transfection anti-miR-424–5p 48 h, but not 72 h, post-transfection (Supplementary Figure 9C,9E,9F and 9H). In contrast, transfection with pre-miR-145–5p did not significantly reduce PSAT1 protein in T47D cells 48 h or 72h post-transfection (Supplementary Figure 9C,9D,9F and 9G). These observations support the previous data showing that miR-424–5p targets PSAT1 in LCC9 cells 48 h post-transfection and suggest that cell-line-specific factors modulate the ability of miR-145–5p to target PSAT1.

Similarly, ZR-75–1-4-OHT cells showed higher PSAT1 protein expression compared to parental ZR-75–1 or MCF-7 cells (Figure 2D). We transfected ZR-75–1 and ZR-75–1-4-OHT cells with pre-miR-145–5p or pre-miR-424–5p (Supplementary Figure 10A – 10D). At 48 h post-transfection with either pre-miR-145–5p or miR-424–5p, PSAT1 protein was unchanged (Supplementary Figure 11). However, 72 h transfection of ZR-75–1 or ZR-75–1-4-OHT cells with either pre-miR-145–5p or miR-424–5p reduced PSAT1 protein abundance and this reduction was blocked by co-transfection with the respective anti-miR (Figure 6E–6J). These data demonstrate that miR-145–5p and miR-424–5p specifically reduce PSAT1 protein in these cells.

miR-34b-5p and miR-876–5p regulate endogenous expression of PHGDH in T47D cells.

To investigate the effects of miR-34b-5p and miR-876–5p overexpression on endogenous PHGDH, LCC9, T47D, ZR-75–1, and ZR-75–1-4-OHT cells were transfected with either pre-miR-34b-5p or pre-miR-876–5p and increased miRNA abundance was detected with RT-qPCR (Supplementary Figures 7A and 7B, 12A–H). Both miR-34b-5p and miR-876–5p reduced PHGDH mRNA in LCC9 cells (Figure 7A). We detected a significant decrease in PHGDH protein in LCC9 cells 72 h post-transfection with pre-miR-876–5p and although co-transfection of anti-miR-876–5p did not affect this inhibition, the level of PHGDH protein was similar to the control with co-transfection (Figure 7B – 7D). In T47D cells, which are less 4-OHT-sensitive compared to MCF-7 cells (Petri, et al. 2021), PHGDH was reduced by pre-miR-34–5p in T47D cells and this reduction was blocked by anti-miR-34b-5p co-transfection demonstrating specificity (Figure 7E–7F). We did not detect any change in PHGDH protein abundance in ZR-75–1 or ZR-75–1-4-OHT cells transfected with pre-miR-34b-5p or pre-miR-876–5p at 48 h or 72 h post-transfection (Supplementary Figure 13).

Figure 7. miR-34b-5p and miR-876–5p target PHGDH in T47D cells.

A) qPCR was performed on independent samples of RNA isolated from LCC9 cells 48h post transfection (TF) with pre-miR-34b-5p and pre-miR-876–5p. Values were normalized by 18S. Data are the average of 3 individual experiments. B) WCE were prepared from LCC9 cells 72h post transfection with pre-miR-34b-5p, pre-miR-876–5p, or co-transfection with pre-miRs and anti-miRs and probed with a PHGDH antibody. C and D) Analysis of western blot performed on 3 independent samples. Values were normalized to LCC9 media control E) WCE were prepared from T47D cells 72h post transfection with pre-miR-34b-5p, pre-miR-876–5p, or co-transfection with pre-miRs and anti-miRs and probed with a PHGDH antibody. F and G) Analysis of western blot performed on 3 independent samples. Values were normalized to T47D media control. Individual values are the mean ± SEM for PHGDH/Ponceau S. ***p ≤ 0.001, **p ≤ 0.01 one-way ANOVA, Tukey’s multiple comparisons test. Values were normalized to GAPDH. Statistical analysis used One-way ANOVA followed by Tukey’s multiple comparison test: *p ≤ 0.05, **p<0.01, ***p ≤ 0.001

Taken together, these results demonstrate inhibition of PHGDH protein expression by miR-876–5p in LCC9 cells, and by miR-34b-5p in T47D cells. The inability of miR-34b-5p and miR-876–5p to suppress PHGDH expression in ZR-75–1 and ZR-75–1-4-OHT cells suggest that cell-specific factors mediate the effect of miRNA inhibition of PHGDH protein synthesis.

Transfection of miRNAs targeting PSAT1 and PHGDH reduce LCC9 and ZR-75–1-4-OHT cell viability, increases ET sensitivity and results in E2 growth inhibition.

Since inhibition of the SSP by inhibiting PHGDH activity increased TAM-sensitivity of LCC9 (Metcalf, et al. 2020b), we examined the effects of the miRs targeting PSAT1 and PHGDH on the cell viability of LCC9 and ET-resistant ZR-75–1-4-OHT cells (Figure 8). Cells were transfected with pre-miRs for miR-145–5p, miR-424–5p, miR-34b-5p, and miR-876–5p and then treated with 100 nM or 1 μM 4-OHT, 100 nM fulvestrant, or 10 nM E₂. The concentration of 4-OHT models the concentrations detected in BCa tissues when patients received 5 or 20 mg tamoxifen citrate/day (Kisanga, et al. 2004). Co-transfection of the four miRs reduced the viability of ZR-75–1-4-OHT cells and increased sensitivity to 4-OHT and fulvestrant in LCC9 and ZR-75–1-OHT cells (Figure 8). The miR-transfected cells were also growth-inhibited by 10 nM E₂, suggesting again in sensitivity to E₂ as an antagonist, as seen in MCF-7 cells that adapted under “long-term estrogen deprivation (LTED) (Lewis-Wambi and Jordan 2009). These data suggest that the reduction of miR-145–5p, miR-424–5p, miR-34b-5p, and miR-876–5p detected in TAM-resistant LCC9 BCa cells promotes resistance to antiestrogens and rescuing the miRNA expression restores antiestrogen-sensitivity in these ER+ BCa cells.

Figure 8. Transfection with miRs targeting PSAT1 and PHGDH restores growth inhibition by 4-OHT and fulvestrant in endocrine-resistant LCC9 cells.

(A) and ZR-75–1-4-OHT (B) cells were transfected with miR Control (−) or pre-mir-145–5p, pre-miR-424–5p, pre-miR-34b-5p, and pre-miR-876–5p for 48 h prior to treatment: DMSO control, 100 nM 4-OHT, 1 μM 4-OHT, 100 nM fulvestrant, or 10 nM E₂. Treatment was for 5 days. Values are the mean ± SEM from 3 and 4 replicate experiments in A and B respectively where data from B include data from two plates/experiment. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 two-way ANOVA, Sidak’s multiple comparisons test with single pooled variance.

Discussion

Endocrine therapies are the first line of treatment for patients with ER+ BCa resulting in significant reduction of BCa-related mortality (Robertson, et al. 2021). However, acquired resistance to endocrine therapies continues to present major challenges in the treatment of ER+ BCa (Clarke, et al. 2015) and the mechanisms of ER+ endocrine resistance are not completely understood. In this study, we tested the hypothesis that four miRNAs downregulated by the m6A reader RNA binding protein A2B1, miR-145–5p, miR-424–5p, miR-34b-5p, and miR-876–5p, are regulators of key enzymes in the SSP and that overexpression of these A2B1-regulated miRNAs increase the sensitivity ER+ endocrine-resistant LCC9 cells to TAM and fulvestrant.

Using computational analysis, we identified four miRNA that were downregulated by A2B1 and predicted to target PSAT1 (miR-145–5p and miR-424–5p) or PHGDH (miR-34b-5p and miR-876–5p). We demonstrated for the first time that PSAT1 is a bona fide direct target of miR-145–5p and miR-424–5p, miRs that are reduced in TAM-R BCa cell lines LCC9, LY2, and MCF-7-A2B1 cells. In addition, miR-145–5p and miR-424–5p reduced PSAT1 protein in T47D and ZR-75–1 and TAM-resistant ZR-75–1-4-OHT BCa cells, demonstrating that this targeting is not specific for MCF-7 and its TAM-resistant derivative cell lines LCC9 and LY2. We found that the ability of miR-145–5p and miR-424–5p to reduce endogenous PSAT1 protein varied between cell lines.

Serine is central to the biosynthesis of many molecules (Locasale 2013) and increased serine biosynthesis is a metabolic change that has been reported in cancer cells (Davis, et al. 1970; Locasale 2013). We reported that overexpression of two key enzymes of the serine synthetic pathway, PSAT1 and PHGDH, contributed to TAM-resistance in ER+ BCa (Metcalf, et al. 2020b). Serine is also an allosteric activator of PKM2 (Chaneton, et al. 2012), an isoform of pyruvate kinase that is highly expressed in breast tumors and negatively associates with reduced overall and disease-free survival (Zhu, et al. 2017). PHGDH was demonstrated to be a gene essential for breast tumorigenesis (Possemato, et al. 2011) and levels of PHGDH, PSPH, SHMT, and PSAT1 (Supplementary Figure 1) are elevated in breast tumors and correlate with reduced OS (Geck and Toker 2016). Conversely, low PSAT1 was correlated with higher progression-free survival in TAM-treated BCa patients with ERα + primary breast tumors (Martens, et al. 2005b). The higher intensity of PSAT1 expression in the intraductal carcinoma of patients with ER+ BCa compared to normal breast tissue, suggests that serine biosynthesis is more active in ER+ tumors; however, due to the lack of treatment and disease progression data from the TMA, it is unclear if the increased activity results from acquired therapy resistance. Examination of RNA seq data from MCF-7 versus LCC9 cells revealed that PHGDH, PSAT1, and PSPH transcript expression was significantly higher in LCC9 endocrine-resistant cells compared to MCF-7 cells (Muluhngwi, et al. 2017).

The expression of PSAT1 and PHGDH is regulated by many transcription factors including Sp1 (Jun, et al. 2008), NFYB (Jun, et al. 2008), p53 (Simabuco, et al. 2018), ATF3 (Li, et al. 2021; Luo, et al. 2022), ATF4 (Gao, et al. 2017; Ye, et al. 2012), NRF2 (DeNicola, et al. 2015), and YAP (Li, et al. 2021)). In addition, PSAT1 expression is downregulated by miRNAs in cancers: miR-195–5p (ovarian), miR-145–5p (colon), and miR-195–5p (TNBC), respectively (Ding, et al. 2022; Wang, et al. 2023). The processing of primary miRNAs (pri-miRNAs) to precursor miRNAs (pre-miRNAs) is facilitated by the DROSHA/DGCR8 microprocessor complex in the nucleus. For some pri-miRNAs, this process is facilitated by A2B1, an RNA binding protein that recognizes the RNA modification N6-methyladenosine (m6A) and associates with DGCR8 in the DROSHA complex to regulate pri-miRNA processing (Alarcon, et al. 2015a) or recruit negative miRNA regulators. We reported that overexpression of A2B1 in ER+ MCF-7 cells dysregulates the expression of miRNAs (Klinge, et al. 2019). We have also reported that stable overexpression of A2B1 confers an endocrine-resistant phenotype in MCF-7 cells (Petri, et al. 2021). Here we found that, similar to the LCC9 and LY2 TAM-resistant cell lines, MCF-7-A2B1 cells had higher PSAT1 and PHGDH expression compared to MCF-7 cells. These results suggest a role for A2B1 in the processing of miRNAs that target enzymes in the serine biosynthetic pathway contributing to the progression from endocrine sensitivity to endocrine resistance.

miR-145–5p has been suggested to have anticancer effects in various cancers, e.g., gastric cancer (Zhou, et al. 2020), cervical cancer (He, et al. 2020), and ovarian cancer (Pan, et al. 2022). Overexpression of miR-145–5p in ovarian cancer cells reduced PSAT1 mRNA (Pan, et al. 2022). Co-transfection of miR-145–5p with a dual-luciferase reporter vector containing the PSAT1 3’UTR decreased luciferase intensity implicating PSAT1 as a direct target of miR-145–5p. In addition, overexpression of miR-145–5p in ER+ BCa cells decreased PSAT1 mRNA and PSAT1 protein. Taken together with previous reports, our new data suggest a common inverse relationship between miR-145–5p and PSAT1 in some cancers. Although miR-145–5p was negatively correlated with CCAT2 and AKT3 which have higher expression in TAM-resistant cells (Moradi, et al. 2022), this is the first evidence of a negative correlation between miR-145–5p and PSAT1 in endocrine-resistant BCa cell lines.

The activity of miR-424–5p varies between cancer types. For example, miR-424–5p levels were higher in laryngeal squamous cell carcinoma and promoted cancer progression by targeting the tumor suppressor CADM1, suggesting onco-miRNA activity (Li, et al. 2019). However, miR-424–5p plays a tumor-suppressive role in other cancers by targeting oncogenes including ACSL4 in ovarian cancer (Ma, et al. 2021) and SOCS6 in pancreatic cancer (Wu, et al. 2013). More specifically, miR-424–5p was downregulated in colorectal cancer and played a tumor-suppressive role by targeting AKT3 and PSAT1 (Fang, et al. 2018). miR-424–5p reduced BCa cell viability in MDA-MB-231 TNBC cells by targeting PD-L1 (Dastmalchi, et al. 2020). No previous studies have investigated the effects of miR-424–5p on endocrine resistance in BCa. Importantly, we demonstrated that miR-424–5p directly targets PSAT1 mRNA and decreases PSAT1 protein in endocrine resistant BCa cells. Additionally, we demonstrated that co-transfection of miR-145–5p, miR-424–5p, miR-34b-5p, and miR-876–5p increases sensitivity of endocrine-resistant LCC9 cells to 4-OHT and fulvestrant.

There are no reports of miR-34b-5p activity in BCa, however mir-34b-5p was downregulated in papillary thyroid cancer (Rogucki, et al. 2022), colon cancer (Wang, et al. 2022), cervical cancer (Ye, et al. 2022), and lung cancer (Mizuno, et al. 2017). In colon cancer cells, HuR was a target of miR-34b-5p (Wang, et al. 2022). HuR is a key post-transcriptional regulator in BCa progression and higher expression of HuR correlated to aggressive BCa phenotypes (Liao, et al. 2023). HuR stimulated the expression of BCa oncogenes including CCL2 (Lee, et al. 2017), Cyclin D1 (CCND1) (Yuan, et al. 2011), MMP-9 (Heinonen, et al. 2007), and VEGF (Ortega, et al. 2008). These studies demonstrate an anti-oncogenic role for miR-34b-5p. Our study is the first to identify PHGDH as a direct miR-34b-5p target. There is some evidence that miR-876–5p plays a role in decreasing cell proliferation in BCa (Dong, et al. 2019; Xiu, et al. 2022). While we confirmed PHGDH as a bona fide target of miR-876–5p, the ability of miR-34b-5p to reduce endogenous PHGDH protein expression was less conclusive. We have shown that miRNA activity is dependent on cell type and length of time post-transfection, but there may be other regulatory factors influencing miRNA modulation of PSAT1 and PHGDH protein expression, including regulation of miRNA by other non-coding RNA.

There are a number of non-coding RNAs, including circular RNAs (circRNAs), long non-coding RNAs (lncRNAs) that have critical regulatory roles in tumorigenesis and drug-resistance in cancer (Hua, et al. 2019; Wu, et al. 2019; Yu, et al. 2020). Mechanistically, lncRNAs and circRNAs bind directly to miRNA seed regions and act as competing endogenous RNAs (ceRNA) to “sponge” miRNAs, resulting in downregulation of miRNAs and upregulation of miRNA targets (Li, et al. 2017; Zheng, et al. 2016). Several of these ceRNA-miRNA relationships have been identified with miR-145–5p, miR-424–5p, miR-34b-5p, and miR-876–5p in cancer (Table 3). In MDA-MB-231 cells, lncRNA ST8SIA6-AS1 sponged miR-145–5p which increased CDCA3 expression and inactivated p53/p21 signaling (Qiao, et al. 2022). Upregulation of TGFBR2 in MCF-7 and MDA-MB-231 cells due to LINC00052 suppression of miR-145–5p resulted in increased cell proliferation, migration, and invasion (Dong, et al. 2021). LINC00473 is elevated in BCa tissue and cell lines and LINC00473 acted as ceRNA of miR-424–5p in BCa cells, blocking its inhibition of CCNE1 and increasing EMT, cell migration and cell invasion (Zhang and Yang 2023). Inhibition of GRK5 by miR-876–5p in TNBC cells was abrogated by LINC01315-miR-876–5p sponging, promoting TNBC progression (Xiu, et al. 2022). The cross-regulatory network between ceRNAs and miRNAs offers a possible explanation for the incomplete decrease in PSAT1 and PHGDH protein expression observed in this study. Future research is necessary to understand the ceRNA-miRNA-mRNA mechanisms involved in ER+ BCa progression and endocrine-resistance.

Table 3. Summary of competing endogenous RNA (ceRNA) interactions with miR-145–5p, miR-424–5p, miR-34b-5p, or miR-876–5p in various types of cancer.

The table summarizes evidence of ceRNAs and their respective interactions with miR-145–5p, miR-424–5p, miR-34b-5p, or miR-876–5p in cancer. Two databases were searched for ceRNA expression in breast cancer cell lines (LncExpDB - https://ngdc.cncb.ac.cn/lncexpdb/ and circBase - http://www.circbase.org/cgi-bin/simplesearch.cgi).

miRNA	Evidence of ceRNA activity by lncRNA or circRNA	ceRNA Expressed in Breast Cancer cell lines (database)
miR-145–5p	CircZNF609 acted as ceRNA of miR-145–5p in MCF-7 and MDA-MB-231 cells, upregulating the expression of oncogene p70S6K1 expression and increasing cell proliferation (Wang, et al. 2018).
	CircZNF236 absorbed miR-145–5p in SCC-15 and CAL-27 and increased proliferation, invasion, and migration of oral squamous cell carcinoma (OSCC) cells (Lu, et al. 2023).	(circBASE)
	circGOLPH3 was upregulated in OSCC cells and bound miR-145–5p, decreased its abundance and prevented miR-145–5p suppression of KDM2A which promotes tumorigenesis and metastasis (Cheng, et al. 2023).
	Circ_0015756 indirectly regulated PSAT1 by binding miR-145–5p resulting in increased migration and invasion of ovarian cancer (OC) cells (Pan, et al. 2022).
	Circ0001955 was increased in a genome-wide study of colorectal cancer (CRC) patients and was negatively correlated with miR-145–5p (Kadkhoda, et al. 2022).
	Circ 0058063 is increased in OSCC cells and bound to miR-145–5p, upregulating the downstream target SERPIN1, contributing to cancer cell proliferation (Yu, et al. 2022)
	circMYOF bound to miR-145–5p, and prevented the inhibition by miR-145–5p of OTX1 in the exosomes of laryngeal squamous cell carcinoma (LSCC) patients, increasing cell migration and invasion ability (Li, et al. 2022).	(circBase)
	LncRNA PVT1 acted as ceRNA of miR-145–5p in MDA-MB-231 cells and enhanced glycolysis and cell proliferation (Qu, et al. 2023)	(LncExpDB)
miR-424–5p	LncRNA HCG28 inhibited miR-424–5p suppression of SOX9 expression in cholangiocarcinoma tissues and cell lines which increased cell proliferation (Ni, et al. 2023).	(LncExpDB)
	Circ-HSP90A upregulated PD-L1 by binding to miR-424–5p and reduced the effects of immunotherapy in non-small cell lung cancer (NSCLC) cells (Lei, et al. 2023).
	Depletion of circRBMS3, which regulated eIF4B and YRDC by acting as ceRNA of miR-424–5p, inhibited malignancy in an osteosarcoma xenograft mouse model (Gong, et al. 2023).
	The upregulation of lncRNA AGAP2-AS1 targeted and reduced miR-424–5p in keratinocytes, leading to an upregulation of miR-424–5p target AKT3 and activation of the AKT/mTOR pathway in keratinocyte proliferation (Xian, et al. 2022).	(LncExpDB)
	circACTN4 (circ_0050898) acted as ceRNA of miR-424–5p in an intrahepatic cholangiocarcinoma (ICC) xenograft mouse model, increasing YAP1, a miR-424–5p target and activated Wnt/β-catenin signaling (Chen, et al. 2022).
	LINC00355 induced EMT in bladder cancer cells by inhibiting miR-424–5p which increased HMGA2 expression (Li, et al. 2021).	(LncExpDB)
miR-34b-5p	lncRNA SNHG3, upregulated in CRC patients, sponged miR-34b-5p in CRC cells which increased miR-34b-5p target HuR. Depleting lncRNA SNHG3 abundance inhibited cell proliferation (Zhao, et al. 2022).	(LncExpDB)
	lncRNA OIP5-AS1, which is elevated in colon cancer (CC), targeted and inhibited miR-34b-5p in CC cells, preventing miR inhibition of HuR expression and driving CC malignancy (Wang, et al. 2022).	(LncExpDB)
	High expression of lncRNA PVT1 in diffuse large B-cell lymphoma (DLBCL) patients correlated to poor prognosis. lncPVT1 acted as ceRNA of miR-34b-5p in BLBCL cells, increasing Foxp. Compared to non-transfected tumor cells, cells with shRNA knockdown of PVT1 injected into mice resulted in slower tumor growth (Tao, et al. 2022)	(LncExpDB)
	Upregulation of LINC00355 increased ABCB1 by inhibiting miR-34b-5p in bladder cancer cells which promoted bladder cancer cell resistance to cisplatin (Luo, et al. 2021).	(LncExpDB)
	LINC02418 inhibited miR-34–5p suppression of BCL2 in CRC cells, promoting cell proliferation (Tian, et al. 2020).	(LncExpDB)
	circBFAR acted as a ceRNA of mir-34b-5p, inhibiting its suppression of MET expression in pancreatic ductal carcinoma (PDAC) cells and promoted proliferation, migration, and invasion (Guo, et al. 2020b).	(circBASE)
	LncNEAT1 inhibited miR-34b-5p suppression of GL1 in DLBCL cells, and increased cell proliferation (Qian, et al. 2020).	(LncExpDB)
miR-875–5p	circCDR1 acted as a ceRNA of miR-876–5p, inhibiting miR suppression of SLC7A11 in OSCC cells which induced cell proliferation, migration, and invasion (Cui, et al. 2023).
	circ_0001686 inhibited miR-876–5p suppression of SPIN1 in esophagus cancer cells, inhibiting apoptosis and radiosensitivity (Yu, et al. 2023).
	LncRNA HOXC-AS2 decreased miR-876–5p and inhibited miR suppression of HKDC1 in endometrial cancer cells, which increased cell proliferation (Guo, et al. 2022).	(LncExpDB)
	ciRS-7 inhibited miR-876–5p suppression of tumor antigen MAGE-A family members in esophageal squamous cell carcinoma cells and induced cell proliferation, migration, and invasion (Sang, et al. 2018).
	circRPS16 decreased miR-876–5p and inhibited miR suppression of SPINK1 in hepatocellular carcinoma (HCC) cells, which increased cell proliferation, invasion, and cell cycle (Lin, et al. 2021).	(circBASE)
	CircVAPA acted as ceRNA of miR-876–5p, which inhibited miR suppression of WNT5A in NSCLC cells and increased cell viability and proliferation (Zhao, et al. 2021).	(circBASE)
	lncRNA SNGHG14 inhibited miR-876–5p suppression of SSR2 in HCC cells which induced cell proliferation, migration, and invasion (Liao, et al. 2021).
	circHIPK3 inhibited miR-875–5p suppression of PIK3R1 in gastric cancer cells and increased cell proliferation, migration, invasion, and glutaminolysis (Li, et al. 2020a).	(circBASE)
	LncRNA PITPNA-AS1 acted as a ceRNA of miR-875–5p, reducing miR suppression of c-MET in cervical cancer cells and increased cell proliferation and cell cycle (Guo, et al. 2020a).	(LncExpDB)
	MCM3AP-AS1 acted as a ceRNA of miR-876–5p, reducing miR suppression of WNT5A in prostate cancer cells, which increased cell proliferation (Wu, et al. 2020).	(LncExpDB)
	lncRNA MINCR inhibited miR-876–5p suppression of GSPT1 in glioma cells and increased cell migration and invasion (Li, et al. 2020b).	(LncExpDB)
	LncRNA PITPNA-AS1 acted as a ceRNA of miR-875–5p, reducing miR suppression of WNT5a in HCC cells and promoted EMT (Sun, et al. 2019).	(LncExpDB)
	LncRNA HOXC-AS2 decreased miR-876–5p and inhibited miR suppression of ZEB in glioma cells, which increased EMT (Dong, et al. 2019).	(LncExpDB)

Chen Q, Wang H, Li Z, Li F, Liang L, Zou Y, Shen H, Li J, Xia Y, Cheng Z, et al. 2022 Circular RNA ACTN4 promotes intrahepatic cholangiocarcinoma progression by recruiting YBX1 to initiate FZD7 transcription. J Hepatol 76 135–147.

Cheng T, Huang F, Zhang Y & Zhou Z 2023 Knockdown of circGOLPH3 inhibits cell progression and glycolysis by targeting miR-145–5p/lysine demethylase 2A (KDM2A) axis in oral squamous cell carcinoma. Head & Neck 45 225–236.

Cui L, Huang C & Zhou D 2023 Overexpression of circCDR1as drives oral squamous cell carcinoma progression. Oral Diseases 29 957–967.

Dong N, Guo J, Han S, Bao L, Diao Y & Lin Z 2019 Positive feedback loop of lncRNA HOXC-AS2/miR-876–5p/ZEB1 to regulate EMT in glioma. Onco Targets Ther 12 7601–7609.

Gong Z, Shen P, Wang H, Zhu J, Liang K, Wang K, Mi Y, Shen S, Fang X & Liu G 2023 A novel circular RNA circRBMS3 regulates proliferation and metastasis of osteosarcoma by targeting miR-424-eIF4B/YRDC axis. Aging (Albany NY) 15 1564–1590.

Guo J, Ye F, Xie W, Zhang X, Zeng R, Sheng W, Mi Y & Sheng X 2022 The HOXC-AS2/miR-876–5p/HKDC1 axis regulates endometrial cancer progression in a high glucose-related tumor microenvironment. Cancer Science 113 2297–2310.

Guo Q, Li L, Bo Q, Chen L, Sun L & Shi H 2020a Long noncoding RNA PITPNA-AS1 promotes cervical cancer progression through regulating the cell cycle and apoptosis by targeting the miR-876–5p/c-MET axis. Biomedicine & Pharmacotherapy 128 110072.

Guo X, Zhou Q, Su D, Luo Y, Fu Z, Huang L, Li Z, Jiang D, Kong Y, Li Z, et al. 2020b Circular RNA circBFAR promotes the progression of pancreatic ductal adenocarcinoma via the miR-34b-5p/MET/Akt axis. Mol Cancer 19 83.

Kadkhoda S, Ghafouri-Fard S, Noorbakhsh F, Ravaei S, Darbeheshti F, Amoli MM, Taslimi R & Shakoori A 2022 The importance of regulatory pathway mediated by Circ0001955 in colorectal cancer. Experimental and Molecular Pathology 128 104819.

Lei J, Zhu J, Hui B, Jia C, Yan X, Jiang T & Wang X 2023 Circ-HSP90A expedites cell growth, stemness, and immune evasion in non-small cell lung cancer by regulating STAT3 signaling and PD-1/PD-L1 checkpoint. Cancer Immunology, Immunotherapy 72 101–124.

Li Q, Tian Y, Liang Y & Li C 2020a CircHIPK3/miR-876–5p/PIK3R1 axis regulates regulation proliferation, migration, invasion, and glutaminolysis in gastric cancer cells. Cancer Cell Int 20 391.

Li S, Zhang Y, He Z, Xu Q, Li C & Xu B 2022 Knockdown of circMYOF inhibits cell growth, metastasis, and glycolysis through miR-145–5p/OTX1 regulatory axis in laryngeal squamous cell carcinoma. Functional & Integrative Genomics 22 1–13.

Li WJ, Li G, Liu ZW, Chen ZY & Pu R 2021 LncRNA LINC00355 promotes EMT and metastasis of bladder cancer cells through the miR-424–5p/HMGA2 axis. Neoplasma 68 1225–1235.

Li Z, Xie X, Fan X & Li X 2020b Long Non-coding RNA MINCR Regulates miR-876–5p/GSPT1 Axis to Aggravate Glioma Progression. Neurochemical Research 45 1690–1699.

Liao Z, Zhang H, Su C, Liu F, Liu Y, Song J, Zhu H, Fan Y, Zhang X, Dong W, et al. 2021 Long noncoding RNA SNHG14 promotes hepatocellular carcinoma progression by regulating miR-876–5p/SSR2 axis. J Exp Clin Cancer Res 40 36.

Lin S, Lin Y, Wu Z, Xia W, Miao C, Peng T, Zhao Z, Ji C, Mo Z, Liu X, et al. 2021 circRPS16 Promotes Proliferation and Invasion of Hepatocellular Carcinoma by Sponging miR-876–5p to Upregulate SPINK1. Front Oncol 11 724415.

Lu Q, Che H, Che Y & Hu M 2023 CircZNF236 facilitates malignant progression in oral squamous cell carcinoma by sequestering miR-145–5p. Clin Transl Oncol.

Luo G, Zhang Y, Wu Z, Zhang L, Liang C & Chen X 2021 Exosomal LINC00355 derived from cancer-associated fibroblasts promotes bladder cancer cell resistance to cisplatin by regulating miR-34b-5p/ABCB1 axis. Acta Biochim Biophys Sin (Shanghai) 53 558–566.

Ni Q, Zhang H, Shi X & Li X 2023 Exosomal lncRNA HCG18 contributes to cholangiocarcinoma growth and metastasis through mediating miR-424–5p/SOX9 axis through PI3K/AKT pathway. Cancer Gene Ther 30 582–595.

Pan Y, Huang Q, Peng X, Yu S & Liu N 2022 Circ_0015756 promotes ovarian cancer progression via the miR-145–5p/PSAT1 axis. Reprod Biol 22 100702.

Qian CS, Li LJ, Huang HW, Yang HF & Wu DP 2020 MYC-regulated lncRNA NEAT1 promotes B cell proliferation and lymphomagenesis via the miR-34b-5p-GLI1 pathway in diffuse large B-cell lymphoma. Cancer Cell Int 20 87.

Qu H, Li X, Chen F, Zhang M, Lu X, Gu Y, Lv M & Lu C 2023 LncRNA PVT1 influences breast cancer cells glycolysis through sponging miR-145–5p. Genes Genomics 45 581–592.

Sang M, Meng L, Sang Y, Liu S, Ding P, Ju Y, Liu F, Gu L, Lian Y, Li J, et al. 2018 Circular RNA ciRS-7 accelerates ESCC progression through acting as a miR-876–5p sponge to enhance MAGE-A family expression. Cancer Letters 426 37–46.

Sun J, Zhang Y, Li B, Dong Y, Sun C, Zhang F, Jin L, Chen D & Wang W 2019 PITPNA-AS1 abrogates the inhibition of miR-876–5p on WNT5A to facilitate hepatocellular carcinoma progression. Cell Death Dis 10 844.

Tao S, Chen Y, Hu M, Xu L, Fu C-B & Hao X-B 2022 LncRNA PVT1 facilitates DLBCL development via miR-34b-5p/Foxp1 pathway. Molecular and Cellular Biochemistry 477 951–963.

Tian J, Cui P, Li Y, Yao X, Wu X, Wang Z & Li C 2020 LINC02418 promotes colon cancer progression by suppressing apoptosis via interaction with miR-34b-5p/BCL2 axis. Cancer Cell Int 20 460.

Wang S, Xue X, Wang R, Li X, Li Q, Wang Y, Xie P, Kang Y, Meng R & Feng X 2018 CircZNF609 promotes breast cancer cell growth, migration, and invasion by elevating p70S6K1 via sponging miR-145–5p. Cancer Manag Res 10 3881–3890.

Wang Y, Lin C & Liu Y 2022 Molecular mechanism of miR-34b-5p and RNA binding protein HuR binding to lncRNA OIP5-AS1 in colon cancer cells. Cancer Gene Therapy 29 612–624.

Wu J, Lv Y, Li Y, Jiang Y, Wang L, Zhang X, Sun M, Zou Y, Xu J & Zhang L 2020 MCM3AP-AS1/miR-876–5p/WNT5A axis regulates the proliferation of prostate cancer cells. Cancer Cell Int 20 307.

Xian J, Shang M, Dai Y, Wang Q, Long X, Li J, Cai Y, Xia C & Peng X 2022 N(6)-methyladenosine-modified long non-coding RNA AGAP2-AS1 promotes psoriasis pathogenesis via miR-424–5p/AKT3 axis. J Dermatol Sci 105 27–36.

Yu J, Lou Y, Hou M, Ma X & Wang L 2022 Circ_0058063 contributes to oral squamous cell carcinoma development by sponging miR-145–5p and upregulating SERPINE1. J Oral Pathol Med 51 630–637.

Yu W, Ning K, Bai Q & Xiao J 2023 Circ_0001686 knockdown suppresses tumorigenesis and enhances radiosensitivity in esophagus cancer through regulating miR-876–5p/SPIN1 axis. Pathol Res Pract 241 154216.

Zhao J, Lin H, Huang K & Li S 2022 Cancer-associated fibroblasts-derived extracellular vesicles carrying lncRNA SNHG3 facilitate colorectal cancer cell proliferation via the miR-34b-5p/HuR/HOXC6 axis. Cell Death Discov 8 346.

Zhao Y, Dai Q, Fu X, Chen Q, Tang Y, Gao X & Zhou Q 2021 CircVAPA exerts oncogenic property in non-small cell lung cancer by the miR-876–5p/WNT5A axis. The Journal of Gene Medicine 23 e3325.

Conclusion

Mechanisms for endocrine resistance in BCa are diverse and depend on a number of influences (reviewed in (Clarke, et al. 2015; Hanker, et al. 2020). Here we identified the downregulation of miR-145–5p, miR-424–5p, miR-34b-5p, and miR-876–5p by A2B1 in ER+ endocrine-resistant BCa cells. We show that miR-145–5p and miR-424–5p directly target PSAT1, and that miR-34b-5p and miR-876–5p directly target PHGDH, key enzymes in the serine biosynthetic pathway. Overexpression of miR-145–5p and miR-424–5p decreased PSAT1 expression in LCC9, T47D, ZR-75–1, and ZR-75–1 4-OHT cells, demonstrating a role for these miRNAs in endogenous PSAT1 regulation. We also observed a cell-line specific role for miR-34b-5p and miR-876–5p regulating PHGDH. Further, overexpression of all four miRs restored tamoxifen- and fulvestrant- sensitivity in endocrine-resistant LCC9 and ZR-75–1-4-OHT cells and sensitized them to inhibition by E₂, a property reported in LTED MCF-7 cells(Lewis-Wambi and Jordan 2009). In addition, conjugated equine estrogens (CEE) protected postmenopausal women in the Women’s Health Initiative (WHI) trial were noted to have a prolonged decrease in breast cancer incidence due to estrogen-induced apoptosis (Abderrahman and Jordan 2022; Jordan 2015). Future studies are needed to identify ceRNA that influence A2B1-regulated miRNAs that target PSAT1 and PHGDH and their cell line expression. Examining the role of miRNAs in regulating serine synthesis in ER+ endocrine-resistance further elucidates the function of the non-coding RNA cross-regulatory network in the deregulation of gene expression in BCa toward the goal of improving therapeutics for endocrine-resistant disease in BCa patients.