Introduction
Breast cancer is the most frequently diagnosed cancer in women and the second highest cause of cancer-related death in United States women in 2016 [
1]. Selective estrogen receptor modulators (SERMs) and aromatase inhibitors are currently used to treat ER-positive breast cancers [
2‐
6], while anti-HER2 drugs are used to treat ER-negative HER2-positive breast cancers [
7‐
9]. However, there are few treatment options available for breast cancers that do not express ER, progesterone receptor (PR), or the HER2 protein, otherwise known as “triple-negative” breast cancers (TNBCs). As no targeted therapy is available for most TNBC tumors, cytotoxic chemotherapy is the current treatment strategy.
We previously conducted cDNA microarray studies to identify potential therapeutic targets of ER-negative breast cancers [
10]. These studies identified several kinases critical for the growth of TNBC. In this study, we sought to identify phosphatases differentially expressed in ER-negative tumors that also regulate breast cancer growth. We identified 11 underexpressed and 31 overexpressed phosphatases in ER-negative compared to those in ER-positive breast cancers. We initially focused on those 11 phosphatases underexpressed exhibiting decreased expression in ER-negative tumors and found that of these 11 underexpressed phosphatases,
DUSP4 is the most frequently deleted phosphatase in ER-negative breast cancer.
DUSP4 specifically regulates extracellular regulated kinase (ERK), and the phosphatase activity of DUSP4 is enhanced upon interaction with ERK and p38 [
11‐
13]. DUSP4 expression has been shown to play an important role in senescence [
14] and emerging evidence suggests DUSP4 is involved in the growth and progression of cancer [
15,
16]. Our results demonstrate that DUSP4 is frequently deleted in breast cancer and underexpressed in TNBCs, and that DUSP4 overexpression halts TNBC tumor growth and invasion. Our results also demonstrate that all three MAPK proteins (ERK1/2, p38, and JNK1) are negatively regulated by DUSP4. Moreover, re-expression of DUSP4 also inhibits other growth inducing pathways, including NFκB and Rb. Overall, this study demonstrates that DUSP4 is frequently deleted in breast cancers and is a critical regulator of growth and invasion of TNBCs, suggesting DUSP4 is an important tumor suppressor gene in these aggressive breast cancers.
Materials and methods
Breast tumor and microarray datasets
All tumors were collected by Dr. Jenny Chang and approved by the institutional review board of Baylor College of Medicine. This tumor set, including tumor stage, size, and patient race and menopausal status, has been previously described [
10]. In this study, four samples were removed from the analysis. These four samples had an unconfirmed ER status or appeared as outliers on the principal component analysis (PCA) plot and thus removed from the dataset. Therefore, 98 samples, 56 ER-positive and 42 ER-negative invasive breast cancers, were used for this analysis. Gene expression was estimated using the Robust Multi-array Average (RMA) procedure. We limited our data and clustering analysis to 262 genes (454 probesets on Affymetrix U133A chip), which encode known phosphatases and proteins with phosphatase in their name. We used the most variable probeset for each of the 262 genes. Statistical analysis was done using Partek software (
http://partek.com). The following criteria were used to find genes differentially expressed between ER-negative and ER-positive tumors: Benjamini–Hochberg false discovery rate (FDR)
p value ≤ 0.1, fold change ≥1.2 or ≤0.8.
Cell lines and cell culture
The MCF-7 (HTB-22), MDA-MB-231 (HTB-26), and MDA-MB-468 (HTB-132) cell lines (American Type Culture Collection (ATCC), Manassas, VA, USA) were cultured in DMEM (Cellgro by Mediatech, Inc., Manassas, VA, USA) supplemented with 10 % fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomycin. These were grown and maintained as described in Supplementary Methods. STR profiles were compared to (1) known ATCC fingerprints (ATCC.org); (2) the Cell Line Integrated Molecular Authentication database (CLIMA) version 0.1.200808 (
http://bioinformatics.istge.it/clima/) [
17]; and (3) the MD Anderson fingerprint database.
Reagents and antibodies
The DUSP4 antibody was purchased from BD Transduction Laboratories (Lexington, KY, USA). Phospho-ERK1/2 (#4370) and total ERK1/2 (#4372) were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). Antiactin antibody was purchased from Sigma-Aldrich Corp. (St. Louis, MO, USA). Antimouse (#NA931 V) and antirabbit (NA934 V) secondary antibodies were obtained from GE Healthcare Bio-sciences Corp. (Piscataway, NJ, USA).
Viral vectors and modification of cell lines
The DUSP4 ORF clone was obtained from Open Biosystems (Huntsville, AL, USA) and cloned into tetracycline (Tet)-inducible vector (pTIPZ) through a gateway LR clonase reaction. Final constructs were verified through restriction digestion and sequencing. Lentiviral vectors were prepared as described previously [
18]. Stable cell lines expressing inducible cDNAs were generated by lentiviral infection using a pTIPZ lentiviral expression system in the presence of 4 μg/ml polybrene, followed by puromycin selection at 48 h post infection. All pTIPZ stable cell lines were maintained in media with Tet-safe serum (Clontech Laboratories Inc., Mountain View, CA, USA).
Quantitative RT-PCR (qRT-PCR) assessments
Quantitative PCR assays of reverse-transcribed transcripts (4 replicates per transcript) were carried out using an ABI PRISM 7900 Sequence Detection System (Life Technologies (formerly Applied Biosystems Inc., Foster City, CA, USA), as previously described [
10].
Cell growth assays
Cellular proliferation and anchorage-independent growth were measured as described previously [
19].
Cell cycle assays
To measure cell cycle distribution, Tet-inducible cells were treated for 4 days with or without doxycycline (2 μg/ml) to induce DUSP4 gene expression. Cells were then harvested and fixed overnight in 70 % ethanol at −20 °C. Cells were then stained with propidium iodide (PI) (1 μg/ml) in 0.1 % Triton X-100 and RNase in PBS and analyzed using a FACSCalibur Flow Cytometer (BD Biosciences, Franklin Lakes, NJ, USA).
Western blot analysis
Western blot analyses were performed as described previously [
19]. Antibodies used include anti-DUSP4 antibody (BD Biosciences, San Jose, CA, 1:1000), anti-phospho-ERK1/2 antibody (1:1000), and anti-ERK1/2 (1:1000) (both from Cell Signaling Technology Inc., Danvers, MA, USA). Anti-Vinculin antibody (Sigma-Aldrich Corp., St. Louis, MO, USA) was used at a 1:2000 dilution.
Boyden chamber invasion assay
Matrigel-coated transwell cell culture chambers (BD Transduction Laboratory, Lexington, KY, USA) and inserts of 8 μm pore size (Corning Inc. Life Sciences, Lowell, MA, USA) were used to measure invasion of TIPZ-DUSP4 in MDA-MB-231 and SUM 159 cells. 3 × 105 cells/well were placed in the upper chamber of the transwell inserts in serum-free media. Media containing 10 % FBS was placed in the lower chambers. All samples were incubated for 20 h at 37 °C in a humidified atmosphere with 5 % CO2. Noninvasive cells in the upper chambers were removed by cotton swab. Invasive cells were then fixed and stained with HEMA3 (Fisher Scientific Company, LLC, Kalamazoo, MI, USA). Invasive cells on the lower surface of the filters which penetrated through the Matrigel were then mounted on glass slides, counted, and photographed using a light microscope at ×40 magnification. Migration assays were performed using transwell inserts without Matrigel. Percent invasion was calculated based on number of cells invaded divided by number of cells migrated, and the results were multiplied by 100. All assays were performed in triplicate and the results shown as average ± standard deviation.
Reverse phase protein array (RPPA)
RPPA assays were performed in the MD Anderson Core facility. Experimental details are provided in Supplementary Materials.
Mouse experiments
Experiments using nude mice (The Jackson Laboratory, Bar Harbor, ME, USA) were performed in accordance with M.D. Anderson Institutional Animal Care and Use Committee (IACUC)-approved protocols. Experimental details of our mouse experiments are provided in Supplementary Materials.
Discussion
In this study, we identified phosphatases differentially expressed in ER-negative compared to those in ER-positive breast cancers. We found 31 phosphatases significantly overexpressed and 11 phosphatases significantly underexpressed in ER-negative versus ER-positive breast cancers. Included in the set of underexpressed phosphatases are phosphatases regulating growth factor signaling pathways (INPP4B, PTPRT, and DUSP4), as well as phosphatases involved in gluconeogenesis (FPB1), nucleotide cleavage (ENPP1), and metabolism (CANT1). Of these 11 underexpressed phosphatases, DUSP4 is the most commonly deleted phosphatase, deleted in approximately 50 % of human breast cancers. Our results show that DUSP4 overexpression in TNBCs suppresses breast cancer cell growth by suppression of MAPKs (ERK1/2, JNK1, and p38), NFkB, and Rb signaling pathways, ultimately causing a cell cycle block.
DUSP4 is an early response gene synthesized after growth factor stimulation [
11,
32,
33] and has also been mapped to a gene locus that is frequently lost in breast and prostate cancer [
34].
DUSP4 is localized in chromosome 8p, part or all of which is commonly lost in multiple cancers including breast cancer [
35‐
38]. Armes et al. [
34] have shown loss of the DUSP4 gene and protein in early-onset and high-grade breast cancer. DUSP4 loss or epigenetic silencing has been described in lung cancer and glioblastomas [
15,
39,
40]. Re-expression of DUSP4 in lung cancer cells with 8p loss and low endogenous DUSP4 reduces growth, while knockdown of DUSP4 in cell lines with high DUSP4 expression enhances cell growth [
15]. Our results are in agreement with these previously published studies [
34,
40,
41]; together, our results and those of others suggest that DUSP4 is an important tumor suppressor gene in breast cancer.
Balko et al. have demonstrated that
DUSP4 mRNA expression levels correlate inversely with MEK inhibitor sensitivity, suggesting that DUSP4 expression is a biomarker for MEK inhibitor sensitivity in PTEN-positive tumors [
16,
42]. Our results also show that DUSP4 inhibits ERK1/2 phosphorylation in TNBC cells, as well as p38 and JNK1/2 phosphorylation. Since termination of MAPK signaling is maintained by MAP phosphatases, lower expression of DUSP phosphatases results in increased MAPK activity [
32,
43]. p38 has been shown to be involved in cell proliferation and tumorigenesis [
44], and high levels of p38 in breast cancer patients correlates with invasiveness and poor prognosis [
45]. Previously, we have shown that inhibition of the p38 kinase suppresses proliferation of ER-negative breast cancer cells [
19]. Our present study suggests that inhibition of all three MAPKs by DUSP expression has a stronger growth inhibitory effect than inhibition of only ERK1/2 alone. Creighton et al. [
46] previously demonstrated that increased MAPK activity causes loss of ERα expression and plays a role in the generation of the ERα phenotype. Our results suggest reduced expression of DUSP4 activates phosphorylation of all three MAPKs. Therefore, regulation of DUSP4 expression may influence the generation of ERα-negative breast cancer via MAPK activation. Expression of DUSP4 also inhibits NFkB and Rb signaling, in addition to MAPK signaling, ultimately resulting in profound suppression of growth.
In this study, we demonstrated that several other phosphatases (PPP1R3C, INPP4B, FBP1, PTPRT, MTMR9, and CANT1) are underexpressed in ER-negative compared to those in ER-positive breast cancers. The PPP1R3C gene is hypermethylated in colorectal cancer (CRC) [
47], and is a candidate tumor suppressor gene in melanoma and is inactivated through promoter methylation [
48]. Gewinner et al. have shown that INPP4B regulates the PI3 K pathway and that its gene is located in a region frequently deleted in both breast cancer cell lines and high-grade breast tumors [
49‐
51]. Deletion of INPP4B has been shown to increase growth of breast cancer cells in vitro, and overexpression of INPP4B reduces growth in vivo [
49]. Another phosphatase underexpressed in ER-negative breast cancer is FBP1. The FBP1 protein regulates glycolysis and epithelial-to-mesenchymal transition in breast cells, and overexpression of FBP1 reduces growth of breast cancer cells [
52]. The other phosphatases we identified as underexpressed in ER-negative breast cancer (PTPRT, MTMR9, CANT1, ENPP1, and CILP) may play important roles in signal transduction and tumorigenesis in breast as well as other cancers. Thus, this approach of examining the differential expression of phosphatase genes in ER-positive and ER-negative breast cancer has identified many important phosphatases that regulate tumor growth and tumorigenesis.
Through this genomic study of RNA from human breast cancers, we identified specific phosphatases differentially expressed in ER-negative breast cancers compared to those in ER-positive breast cancers. For this study, we focused on those phosphatases underexpressed in ER-negative versus ER-positive breast cancer. Such phosphatases may be important tumor suppressor genes in ER-negative breast cancer. We also demonstrated that DUSP4 controls the growth and invasiveness of ER-negative breast cancer, and alters the phosphorylation of several growth-promoting signaling proteins, including three MAPKs (ERK, p38, and JNK) and NFκB. These results suggest that targeting this pathway by targeting the downstream genes ERK, p38, JNK, and NFκB, or by reactivating DUSP4 (possibly by using demethylation agents), provides a novel approach for the treatment of ER-negative, and particularly triple-negative, breast cancer.
Compliance with ethical standards