Introduction
Triple-negative breast cancer (TNBC) is an aggressive and lethal form of cancer that lacks estrogen receptor alpha (ERα), progesterone, and human epidermal growth factor receptors with no approved targeted therapeutic options. Despite numerous advances, treatment resistance and metastasis are the main causes of death in patients with TNBC. Resistance to conventional treatment and onset of metastases may arise from a subpopulation of cancer stem cells (CSCs) with self-renewal and tumor-initiating capacities [
1,
2]. Thus, combinatorial treatments with conventional chemotherapy and anti-CSC therapies would be necessary to reduce tumor burden, recurrence, and metastasis to distant organs [
3,
4].
Nitric oxide (NO) is a bioactive molecule that exhibits pleotropic effects within cancer cells and tumors, with concentration-dependent pro- and anti-tumor effects. NO is produced by three different nitric oxide synthase (NOS) isoforms: neuronal (nNOS/
NOS1), inducible (iNOS/
NOS2), and endothelial (eNOS/
NOS3) [
5]. Increased iNOS expression has been found in breast cancer [
6-
9] and other different cancers such as lung [
10], colon [
11], melanoma [
12], and glioblastoma [
13]. Previous reports have demonstrated a correlation between high iNOS expression, aggressiveness, and poor prognosis in patients with breast cancer [
6-
9]. Increased iNOS expression has recently been postulated as a prognostic factor for reduced survival in patients with basal-like ERα-negative breast cancer through the induction of interleukin-8 (
IL-8),
CD44,
c-Myc [
7] and partially due to the activation of the transcription factor Ets-1 [
14]. Here, we hypothesize that enhanced endogenous iNOS expression drives poor patient survival by promoting tumor relapse and metastases through modulation of CSC self-renewal properties and tumor cell migration. We further hypothesize that, in combination with conventional chemotherapy, the inhibition of endogenous iNOS would reduce the aggressiveness of residual TNBC cells and mesenchymal features and the number of metastases to distant organs, thus improving survival of patients with TNBC. We studied the inhibition of iNOS with different small-molecule inhibitors: the selective iNOS inhibitor 1400 W and two pan-NOS inhibitors: L-NMMA and L-NAME. L-NMMA has been extensively studied in hundreds of patients for cardiogenic shock [
15] and, if efficacious, would enable immediate translation into clinical trials without the need of extensive preclinical testing.
Methods
Reagents
N-[[3-(aminomethyl)phenyl]methyl]-ethanimidamide (1400 W) and N5-[imino(nitroamino)methyl]-L-ornithine methyl ester (L-NAME) were purchased from Cayman Chemical (Ann Arbor, MI, USA). Tilarginine (NG-Monomethyl-L-arginine) (L-NMMA) was from Santa Cruz Biotechnology (Dallas, TX, USA) and kindly supplied by (Arginox Pharmaceuticals, Redwood City, CA, USA). Tunicamycin and recombinant human TGF-β1 were obtained from Abcam (Cambridge, UK) and PeproTech (Rocky Hill, NJ, USA), respectively. iNOS (N-20), eNOS (C-20), nNOS (R-20), Twist1 (L-21), Twist1 (2C1a), ATF3 (C-19), and CREB-2 (C-20) antibodies were from Santa Cruz Biotechnology. Antibodies Snail (C15D3), Slug (C19G7), TCF8/Zeb1 (D80D3), PERK (C33E10), TGFβ, phospho-Smad2/3 (D6G10), Smad2/3, IRE1α (14C10), phospho-PERK (16 F8), PERK (C33E10), phospho-eIF2α (119A11), eIF2α, β-Actin, anti-rabbit, and anti-mouse IgG were obtained from Cell Signaling Technology (Danvers, MA, USA). Hypoxia-inducible factor 1α (HIF1α) (EP1215Y) was from Abcam.
Oncomine gene expression data analysis
Relative levels of
NOS2 mRNA expression in human TNBC were investigated by Oncomine Cancer Microarray database analysis [
16] of The Cancer Genome Atlas (TCGA) database (n = 593). Patient survival analysis of two different gene expression data sets was obtained [
17,
18].
Cell culture
Mesenchymal-like TNBC cell lines MDA-MB-231 and SUM159 were purchased from American Type Culture Collection (Manassas, VA, USA) and Asterand Bioscience (Detroit, MI, USA), respectively. These cell lines were chosen on the basis of their high expression of epithelial-mesenchymal transition (EMT) markers, metastatic properties, percentage of CD44+/CD24− cells, iNOS protein levels, similar protein levels of iNOS downstream targets, and similar production of total NO (data not shown). Cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Life Technologies, Grand Island, NY 14072 USA) supplemented with 10% fetal bovine serum and 1% antibiotic-antimycotic. Unless otherwise specified, cells were treated daily with 1400 W (0.1, 1, 10, 100 μM; 1, 2, 4 mM), L-NMMA (0.1, 1, 10, 100 μM; 1, 2, 4 mM), or L-NAME (0.1, 1, 10, 100 μM; 1, 2, 5 mM) for 96 hours.
For mammosphere (MS) formation (MSFE), cells were cultured for 96 hours under treatment in 0.5% methylcellulose and MammoCult basal medium (StemCell Technologies, Vancouver, BC, Canada) supplemented with 10% proliferation supplements, 4 μg/mL heparin, and 0.48 μg/mL hydrocortisone. Primary MSs were scanned and counted with GelCount (Oxford Optronix, Abingdon, UK). Secondary MSs were grown in the absence of treatment. For the mouse model of lung metastasis, MDA-MB-231 cells were transfected with a luciferase/GFP-based dual-reporter plasmid and stable clones (MDA-MB-231 L/G) selected with blasticidin (InvivoGen, San Diego, CA, USA).
Cell proliferation assay
Proliferation of SUM159 and MDA-MB-231 was determined by adding premixed WST-1 reagent (Clontech, Mountain View, CA, USA). For transient knockdown in SUM159 and MDA-MB-231 cells (500 cells per well), proliferation was determined after 72 hours of transfection.
Wound healing assay
Confluent cells were treated in starvation conditions (1% serum) for 72 hours. Medium was changed by regular growth medium in the presence of inhibitors for 24 hours more. For transient knockdown, cells were transfected for 72 hours in growth media. A ‘wound’ was then created in the cell monolayer with a 100-μL pipette tip. Images were taken at 0 and 14 hours. Data were replicated in three independent experiments.
RNA interference experiments
SUM159 and MDA-MB-231 cells were transiently transfected with Scrambled siRNA, siRNA1, or siRNA2 (100 nM) (Silencer Select; Ambion, Life Technologies, Grand Island, NY 14072 USA) for 96 hours using Lipofectamine RNAiMAX (Invitrogen, Life Technologies, Grand Island, NY 14072 USA) in accordance with the instructions of the manufacturer. GIPZ lentiviral NOS2 (shRNA1 - V3LHS_360691) and empty vector shRNAs were purchased from Thermo Fisher Scientific. MDA-MB-231 cells were transduced with lentiviral particles and selected with puromycin. Cells were then harvested and plated for immunocytochemistry of iNOS.
Nitric oxide production
Cells were treated with L-NMMA or 1400 W for 24 hours in phenol red- and serum-free DMEM. Aliquots of cell culture supernatant were taken at 0, 0.5, 2, 6, and 24 hours for total NO production with the nitrate/nitrite fluorometric assay kit (Cayman Chemical) in accordance with the instructions of the manufacturer.
Western blot
Whole cell lysates were made in 1X lysis buffer (Cell Signaling Technology) and 1X protease/phosphatase inhibitor cocktail (Thermo Scientific). Samples (30 μg protein) were boiled in sample buffer (Thermo Scientific) containing β-mercaptoethanol (Sigma-Aldrich, St. Louis, MO, USA) and subjected to SDS-PAGE electrophoresis in 4% to 20% polyacrylamide gels (Bio-Rad, Hercules, CA, USA). Proteins were transferred onto nitrocellulose membranes (Bio-Rad). Membranes were incubated overnight at 4°C with primary antibodies (1:1,000) and the appropriate secondary antibodies for 1 hour (1:2,000). Protein bands were developed in autoradiography films (Denville Scientific Inc., South Plainfield, NJ, USA).
RT-PCR analysis of spliced XBP1
cDNA was synthetized from total RNA and subsequently amplified by polymerase chain reaction (PCR). The primers were XBP1 (5′-GGGTCCAAGTTGTCCAGAATGC-3′ and 5′-TTACGAGAGAAAACTCATGGC-3′) and β-Actin (5′-CTGGAACGGTGAAGGTGACA-3′ and 5′-AAGGGACTTCCTGTAACAATGCA-3′). PCR conditions were 1 cycle at 95°C for 5 minutes, 25 cycles of 30 seconds at 95°C, 1 minute at 50°C, and 1 minute at 68°C, followed by 1 cycle at 68°C for 5 minutes. cDNA amplicons were resolved in 2% agarose.
Immunohistochemistry
After antigen retrieval (Tris-Cl, pH 9.0), paraffin-embedded sections of human patient samples and xenograft tumors were incubated for 1 hour at room temperature with iNOS (1:50) (Novus Biologicals, Littleton, CO, USA), Ki67 (1:100) (Abcam), and cleaved caspase-3 (1:50) (Cell Signaling Technology) antibodies. The iNOS score method was as follows: intensity (0 to 3): negative, weak, moderate, strong; distribution (0 to 4): <10%, 10% to 30%, >30% to 50%, >50% to 80%, >80%. Total score can be divided into four groups: negative (0 or 1), weak (2 or 3), moderate (4 or 5), and strong (6 or 7) as previously reported [
7]. MDA-MB-231 cells transfected either with
NOS2-directed shRNA (shRNA1) or empty vector were used as negative and positive control of iNOS staining, respectively.
Animal studies
Either MDA-MB-231 or SUM159 cells (3 × 106) were injected in the right mammary fat pad of female severe combined immunodeficiency (SCID) Beige mice. Once the tumors reached 150 to 200 mm3, the mice were randomly assigned as follows (n = 10 per group): (1) vehicle (saline, intraperitoneal, or i.p.), (2) L-NMMA (either 80 mg/kg or 200 mg/kg, i.p., daily), (3) docetaxel (20 mg/kg), and (4) combo (L-NMMA and docetaxel). For the lung metastases-preventing study, MDA-MB-231 L/G cells were implanted as described above. The mice were randomly assigned, and treatments started 48 hours after cell injection (n = 5 per group): (1) vehicle (saline, i.p.) and (2) L-NAME (80 mg/kg, i.p., daily for 35 days). Lungs were removed and incubated in cold complete DMEM containing 50 μM luciferin for 10 minutes. Luminescent cancer cells were detected with an IVIS-200 in vivo imaging system (PerkinElmer, Waltham, MA, USA). The clinically relevant dose regimen consisted on two cycles of docetaxel (20 mg/kg, i.p., on day 0), L-NMMA (400 mg/kg on day 1, and 200 mg/kg for 4 additional days by oral gavage), and amlodipine on day 0 (10 mg/kg, i.p., daily, for 6 days). Docetaxel alone as well as saline (i.p.) + sterile water (oral gavage) were used as controls.
MS formation and limiting dilution assay (LDA) were assayed as previously described [
4]. All animal procedures and experimental protocols were approved by the Houston Methodist Research Institute Animal Care and Use Review Office that ensured adherence to the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
Xenograft tissue as well as plasma samples were prepared as previously described [
19]. L-NMMA (200 mg/kg) was orally administered by gavage to female SCID Beige mice (n = 5). Blood was drawn before (baseline, 0 hours) and after (0.5, 2, 12, and 24 hours) L-NMMA administration. Ratiometric quantification of methylarginine (L-NMMA) and citrulline was determined as ion abundance levels in plasma and tumor tissue [
19].
Blood pressure
Blood pressure (BP) was measured in 15 female SCID Beige mice for 3 days (basal BP) and subsequently treated with one cycle of the clinically relevant dose regimen (n = 5 per group). The average daily BP was determined by averaging the last 10 of 20 BP measurements for the last three consecutive days of the cycle treatment using a computerized tail cuff monitor (BP-2000 Series II; Visitech Systems, Napa Place, Apex, NC, USA).
Statistical analysis
All data were analyzed by using GraphPad Prism (GraphPad Software, La Jolla, CA, USA). Data are presented as mean ± standard error of the mean. Statistical significance between two groups was analyzed by two-tailed Student’s t test. Experiments with more than three groups were analyzed with one-way analysis of variance (ANOVA) and Bonferroni’s post hoc test. Statistical analysis of tumor volume was assessed by two-way ANOVA and Bonferroni’s post hoc test. Fisher’s exact test was used to determine significant differences in LDA. Survival proportions were assessed by using a Kaplan-Meier method and further analyzed with either Wilcoxon or log-rank test. Proliferation, MSFE, migration index, and Ki67 staining are normalized to the vehicle group (100%). A P value of less than 0.05 was considered significant.
Discussion
TNBC is an extremely aggressive and lethal form of cancer lacking effective targeted therapies. Patients with TNBC show higher risk of metastasis and tumor relapse [
1,
2]. We have recently described two novel cancer genes (
RPL39 and
MLF2) that are important for tumor initiation and metastasis and are regulated by the NO signaling pathway [
32]. iNOS predicts for worse survival in patients with basal-like ER-negative breast cancer and has been suggested to increase tumor aggressiveness by modulating CSCs as well as the metastatic propensity of cells [
7]. This is the first report demonstrating that iNOS inhibition decreases tumorigenicity of TNBC cells by affecting cell proliferation, CSC self-renewal, and migration. Our
in vivo experiments demonstrate the efficacy of a small-molecule iNOS inhibitor L-NMMA as a potential novel targeted therapy in patients with TNBC with immediate translation into human clinical trials.
We report that
NOS2 is commonly increased in invasive TNBC and is associated with poor survival of patients with invasive breast carcinoma. We provide additional data that high iNOS protein levels by immunohistochemistry in samples of patients with TNBC also correlate with worse outcome, consistent with earlier reports in ERα-negative [
8] and invasive breast carcinoma [
33]. Kaplan-Meier analysis of the Van de Vijver and Curtis databases as well as of the human TNBC patient samples strongly indicate that high iNOS expression is associated with poor overall survival in patients with TNBC. These observations establish that increased iNOS expression may be a marker of poor prognosis in patients with breast cancer [
6-
9].
iNOS expression has been correlated with increased tumor grade and aggressiveness of breast cancer cells [
7,
9]. Here, we describe the effect of iNOS on CSC self-renewal, tumor initiation, and the migrating capacity of TNBC cells. The anti-tumor activity of iNOS inhibitors has been previously reported in oral [
34], glioblastoma [
13], and breast cancer [
35-
38] and is consistent with our findings. Increased iNOS expression has been described to contribute to resistance to conventional treatment by promoting tumor initiation in glioblastoma cells [
13,
38]. Additionally, iNOS may influence CSC self-renewal by modulating
CD44 and
c-Myc in ERα-negative breast cancer [
7]. We demonstrate for the first time that iNOS inhibition decreased CSC self-renewal and tumor initiation in TNBC.
NO may either promote or inhibit metastatic events depending on endogenous levels [
5]. The role of NOS inhibitors on metastasis has been previously studied, but the underlying mechanisms remain unclear. An early study demonstrated that the pan-NOS inhibitor L-NAME may decrease tumor growth and lung metastasis in a murine breast cancer model (EMT-6 cells) [
39]. Similarly, L-NAME inhibited the invasive and migrating potential of two metastatic mammary cell lines (C3L5 and C10) [
40]. In a study with the metastatic human adenocarcinoma HRT-18 cells, the invasiveness was substantially decreased by daily treatment with 500 μM of the selective iNOS inhibitor 1400 W [
41]. More recently, 1400 W was shown to markedly inhibit spontaneous lung metastasis in a mouse model of adenoid cystic carcinoma of the oral cavity [
34]. Our present results demonstrate that iNOS inhibition decreases cell migration and lung metastases in TNBC. It has been suggested that NO and iNOS may lead to early metastasis by inducing
IL-8 [
7] and the CXC chemokine receptor 4 [
42]. CSCs display mesenchymal features [
2], resulting in increased cell migration and metastasis. iNOS inhibition decreased CSC self-renewal and tumor initiation, thus indicating that inhibitors against this pathway could reverse the transition of tumor cells to a more mesenchymal-like phenotype. Consistent with the effect on cell migration, selective iNOS inhibition and
NOS2 knockdown decreased transcription factors driving EMT in all of the TNBC cell lines tested.
EMT may be promoted by different signal transduction pathways like TGFβ, Wnt/β-catenin, Notch, and Hedgehog and multiple growth factors. EMT transcription factors (Snail, Slug, Twist1, or Zeb1) are activated by diverse intermediate effectors like c-Myc, Ets, HIF1α, or NFκB [
43]. Additionally, ER stress has been linked to EMT in thyroid, alveolar epithelial, and human renal proximal tubule cells through activation of PERK, XBP1, or Grp78 [
27,
44-
46]. Interestingly, among these disparate signaling networks, iNOS is the common denominator between HIF1α and ER stress [
25-
27]. Inhibition of endogenous iNOS-derived NO production was able to reduce HIF1α stabilization and protein levels in colon carcinoma cells [
47]. Transcription factors Twist1, Snail, Slug, and Zeb1, among others, are directly or indirectly influenced by HIF1α [
26]. Additionally, hypoxia induces ER stress and unfolded protein response and was recently linked to migration and sphere formation in breast cancer cells by activation of the PERK/ATF4/LAMP3 arm under hypoxic conditions [
48]. Our results suggest that iNOS inhibition correlates with an impairment of TGFβ signaling via the ER stress ATF4/ATF3 axis. It is known that TGFβ stimulates ATF4 protein levels to suppress differentiation in calvarial osteoblasts [
29]. Certain conditions such as ER stress through the PERK/eIF2α axis may activate ATF4, which in turn induces ATF3 transcription [
49], whereas ATF3 itself is an activating transcription factor that enhances TGFβ, MS formation, and EMT in cooperation with Twist1 [
28].
To determine a safe and effective regimen with clinical applicability was the main challenge of these preclinical studies. L-NMMA is a pan-NOS inhibitor that has been extensively studied in several clinical trials of circulatory shock [
30]. In the cardiogenic shock trial, L-NMMA was safe and had few adverse events other than transient reversible hypertension [
31]. In normotensive patients, L-NMMA was administered to patients with metastatic renal cell carcinoma prior to infusion of IL-2 [
31]. Doses of 3 and 6 mg/kg did not induce clinically apparent side effects, and BP remained unchanged. At a dose level of 12 mg/kg, patients experienced an increase in systolic BP up to 25 mm Hg, without any clinical symptoms, which normalized rapidly on stopping the L-NMMA infusion. The dose rate in the present study was chosen, with modifications, on the basis of a previous clinical trial in patients with septic shock [
31]. Our results demonstrate that tumor growth can be restrained by an attenuated regimen of 5 days of L-NMMA after chemotherapy, given together with amlodipine. López
et al. [
50] carried out a randomized, placebo-controlled, double-blind study of L-NMMA in patients with septic shock up to a maximum of 14 days. The regimen followed consisted of an initial dose of 2.5 mg/kg per hour and then was adjusted at different rates (0.5, 1, 2.5, 5, 7.5, 10, 15, and 20 mg/kg per hour). The current proposed dose regimen to be clinically tested as a novel anti-cancer therapeutic is at much lower total doses, which have been described [
50].
Competing interests
SSG is a partner in Arginox Pharmaceuticals. The other authors declare that they have no competing interests.
Authors’ contributions
SG-P performed and designed the majority of the experiments, analyzed the data, and drafted the manuscript. YL assisted in the generation of the metastatic model of MDA-MB-231 L/G xenografts and the acquisition of bioluminescence data. MLG assisted with the in vitro data acquisition and Western blot experiments with L-NMMA. EB and DSC assisted in the treatments and measurement of tumor volume in the animal experiments. WQ performed the immunohistochemistry in TMA of xenograft and patient tissues. TP and AAR managed patients and collected the clinical outcome and patient samples. JC assisted with evaluating and scoring the levels of iNOS in TMAs from xenograft and patient tissues. HLW performed the biostatistics and computational analysis to correlate the iNOS levels in the patients’ TMA and the clinical outcome. HZ performed the analysis of NOS2 mRNA expression in human TNBC with the Oncomine Cancer Microarray database of TCGA. MDL and BD assisted in the statistical analysis, interpretation of the clinical data, and review of the manuscript. SSG gave financial support, performed the metabolite profiling by liquid chromatography-tandem mass spectrometry, provided L-NMMA, and assisted in the interpretation of the animal data and revision of the manuscript. JCC conceived the project, supervised the study, provided financial support, and contributed to the writing and critical reviewing of the final manuscript. All authors have made substantial contributions to the design of experiments, data analysis, and drafting of the manuscript. All authors have reviewed and approved the final version of this manuscript and agree to be accountable for all aspects of the work presented herein.