After IRB approval, female breast cancer patients diagnosed with stage II or III primary breast cancer who underwent mastectomy or segmentectomy at the City of Hope between the years 2000 and 2010 were identified through the City of Hope's Cancer Registry. A total of 1,029 subjects were included in the study. To determine the subjects taking perioperative β-blockers, a large-scale string-parsing query across electronic records of physician dictations and pharmacy data was conducted. Anesthesia records were neither computerized nor included in this study. All brand and generic names of β-blockers not specifically intended for glaucoma treatment and available in the United States were included in the study. Patients were considered to have been on perioperative β-blockers if the computational query of medical records, limited to 45 days from the patient's date of surgery, indicated β-blocker use. The time to recurrence was the primary endpoint during the data collection. Recurrence-free survival was measured from the date of surgery to the first recurrence (local or metastasis), death from oncological cause, or the date of the last follow-up, whichever occurred first. Kaplan-Meier estimates and Cox regression were performed to determine the effect that the perioperative administration of β-blockers had on cancer recurrence in our cohort.

The MDA-MB-231 (231) cell line, its brain-trophic derivative MDA-MB-231Br (231Br), a low passage TN brain metastasis cell line COH-BBM3 (BBM3), primary Her2⁺ breast cancer SkBr3 cell line, and low passage Her2-amplified brain metastasis cell lines COH-BBM1 (BBM1), COH-BBM2 (BBM2) were cultured in Dulbecco's modified Eagle's medium (DMEM)-F12 media (Life Technologies) supplemented with 10% fetal bovine serum (FBS; Hyclone), 1% glutamax and 1% antibiotic-antimycotic (both from Life Technologies) (7).

Patient breast-to-brain metastasis (BBM) tissue specimens were obtained with IRB approval, formalin-fixed, and embedded. Paraffin blocks were sectioned onto slides (10 µm) and baked for 3 h. Wax was removed from the patient BBM tissue specimens and primary breast cancer array slides (US Biomax) by xylene, and the tissue was hydrated with a gradient of ethanol washes and PBS wash. For antigen retrieval, slides were boiled in 10 mM sodium-citrate buffer [Tris sodium citrate (J.T. Baker), pH 6.0] for 1 h. Slides were permeabilized in 0.3% Tween-20 (Sigma-Aldrich) at 37°C for 45 min and blocked with 1% bovine serum albumin (BSA; Sigma-Aldrich), 10% normal goat serum (NGS; Invitrogen), and 0.3 M glycine (Fisher Scientific) at room temperature for 1 h. Primary antibodies in a primary carrier of 1.5% NGS and 1% BSA in PBS were applied overnight at 4°C. Fluorescent secondary antibodies (Jackson ImmunoResearch Laboratories) were applied the following day for 1 h at room temperature protected from light. Coverslips (Warner Instruments) were placed on slides mounted with Immunogold with DAPI (Life Technologies). Slides were imaged on a Carl Zeiss LSM confocal microscope. Antibodies used were the following: ADRB1 (Cell Signaling Technology) and ADRB2 (Thermo Scientific).

mRNA was extracted from 1.0×10⁶ 231, 231Br, BBM1, BBM2, BBM3 or SkBr3 cells using an mRNA extraction kit (Qiagen). qPCR was performed on a Neurotransmitters and Receptors qPCR Array plate (SaBiosciences). Data for ADRB2 in the metastasis cell lines were quantified relative to their primary breast cancer subtype counterparts after normalization to actin, i.e., BBM3 and 231Br were compared to 231, and BBM1 and BBM2 were compared to SkBr3.

Cells were trypsinized and then centrifuged at 1,500 rpm for 5 min. Collected cells were homogenized by lysis buffer [glycerol, 3 M KCl and 10% NP-40 (all from Sigma-Aldrich), 1 M Tris (Bio-Rad) and 1 M DTT (Sigma-Aldrich)] in the presence of phosphatase inhibitor, EDTA, and protease inhibitor (all from Thermo Scientific) for 15 min. The tubes were spun down at 12,000 rpm for 10 min and the protein in the supernatant was collected. Forty micrograms of protein boiled for 5 min with loading buffer [95% laemmli, and 5% β-mercaptoethanol (both from Bio-Rad)] were loaded onto Mini-PROTEAN TGX gels (Bio-Rad). The gel was run at 200 V and transferred onto a membrane at 100 V for 1 h. After confirming protein transfer on the membrane with Ponceau staining, the membrane was blocked for 1 h with blocking buffer (Thermo Scientific). Primary antibody diluted in blocking buffer was applied overnight at 4°C. The following day, horseradish peroxidase (HRP)-conjugated secondary antibody (Thermo Scientific) was applied for 1 h at room temperature. Membranes were exposed to SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific) for 5 min. Membranes were scanned by Li-COR c-DiGit membrane scanner (Li-COR). Antibodies used were as follows: ADRB1 (Cell Signaling Technology), ADRB2 (Thermo Scientific) and β-actin (Cell Signaling Technology).

231 or 231Br cells (3.0×10³/well) were seeded on 24-well plates. After 24 h, they were treated with one of the following treatments (day 0): propranolol (33.3 µM), terbutaline sulfate (16.67 µM), or a combination of propranolol and terbutaline sulfate. Cells were counted daily for 5 days.

231 or 231Br cells (7.0×10⁴/well) were plated into a migration assay plate (Ibidi). Twenty-four hours after plating, a silicon separator was removed, mimicking scratch conditions, and one of the following treatments was applied: propranolol (100 µM), terbutaline sulfate (50 µM), or a combination of propranolol and terbutaline sulfate. Tiled images (2×5) were captured at 0 and 9 h under bright field microscopy and phase contrast microscopy at a ×5 magnification. Quantification at 9 h of migration was carried out through ImageJ relative to time 0 under bright field microscopic images. Representative qualitative images of the quantified areas were displayed by phase contrast images.

Matrigel (3 mg/ml) was coated onto 0.8-µm invasion inserts (Millipore), which were placed each in a well of a 24-well plate in a 37°C incubator and allowed to solidify for 1 h. Cells (1.5×10⁵) were placed in the insert in serum-free media. Serum-free media with or without drugs were placed in the well in the following concentrations: propranolol (100 µM), terbutaline sulfate (50 µM), or a combination of propranolol and terbutaline sulfate. After 24 h, the inserts were washed in PBS, cells were fixed with 4% paraformaldehyde for 10 min at room temperature and stained with crystal violet (0.2% in 25% methanol) for 15 min at room temperature. Remaining cells and Matrigel were swabbed out with Q-tips. Five sections on each insert were imaged by bright-field at a ×5 magnification under a Zeiss Observer Z1 Live Cell microscope and quantified. Calculations were carried out relative to the invasion of the control cells.

231Br cells (5.0×10⁴) labeled with firefly-luciferase remained untreated or were pre-treated with propranolol (33.3 µM) for 12 h. Cells were injected into NOD-SCID mice in the intracardial space under IACUC approval (City of Hope IACUC #10044). Tumors were monitored weekly by bioluminescence imaging (BLI). Upon euthanasia, the brains were dissected, fixed in formalin, and sectioned onto slides in 10-µm sections for hematoxylin and eosin staining to confirm tumor location. Images were acquired and tiled at ×5 magnification.

The median age (range) of the patients at diagnosis was 53 years (21–92). Sixty-eight percent of our subjects were stage II patients and 91 of these subjects received perioperative β-blockers. We found that for stage II patients who received perioperative β-blockers the hazard ratio for recurrence and metastases calculated with the Cox regression was 0.51. (95% CI: 0.23–0.97; p=0.041). The log-rank test for this stage II subgroup did not achieve statistical significance at the α=0.05 level (p=0.0588). The Kaplan-Meier curves appear in Fig. 1. There was no statistical difference in post-operative recurrence associated with β-blocker usage for stage III breast cancer patients. In addition, there was no difference in overall survival with β-blocker use for either group.

We then investigated the effects of β-adrenergic receptor activation and blockade on both primary breast cancer and brain metastatic cells. Since β1- and β2-adrenergic receptor subtypes are frequently targeted by β-blockers, we initially analyzed expression of these subtypes in formalin-fixed patient brain metastatic specimens. ADRB1 expression was low in the Her2⁺ and TN brain metastatic tissues (Fig. 2). There was low β2-adrenergic receptor (ADRB2) expression in the Her2⁺ brain metastatic tumors compared to the primary Her2⁺ breast tumors and consistently high β2-adrenergic receptor expression in TN primary and brain metastatic tumors (Fig. 3A). For this reason, we decided to focus our experiments on the β2-adrenergic receptor in the TN tumor subtype. For the purposes of our study, we utilized the TN primary breast cancer cell line 231 and its brain-derivative 231Br. The high expression of β2-adrenergic receptor protein (ADRB2) in TN primary and metastatic cells was observed by western blot analyses (Fig. 3B). Finally, to confirm decreased ADRB2 in Her2 brain metastases and increased ADRB2 in TN brain metastases compared to their primary breast cancer subtypes, we analyzed ADRB2 expression at the mRNA level with qRT-PCR. We analyzed 2 separate Her2⁺ brain metastasis cell lines (COH-BBM1, and COH-BBM2) with an Her2⁺ primary breast cancer cell line (SkBr3) and 2 separate TN brain metastatic cell lines (COH-BBM3 and 231Br) with a TN primary breast cancer cell line (231). Her2⁺ brain metastatic cells expressed a 16-fold decrease in ADRB2 compared to primary Her2⁺ cells and TN brain metastatic cells with a 2-fold increase compared to primary TN cells (Fig. 3C).

We investigated the effects of β2-adrenergic receptor activation on characteristics of metastasis, namely proliferation, migration and invasion (20,21). Studies have shown that terbutaline sulfate, a β2-adrenergic receptor agonist, increases tumor weights similar to those under stress (22,23). Hence we used terbutaline sulfate to mimic stress conditions in vitro. Clinical data suggest a β1 to β2 ratio activity of propranolol to be 1:2, and its clinical effects may be duration-dependent (5). Propranolol also antagonized noradrenaline-induced β2-adrenergic receptor activation 2 to 3 times more than β1-adrenergic receptors (24), thus we utilized this commonly used antagonist to inhibit β2-adrenergic receptors.

We performed a proliferation assay over a 5-day period. Terbutaline sulfate significantly increased proliferation in brain metastatic cells (p<0.0001). Propranolol decreased the rate of cell proliferation compared to the control cells and also eliminated the effects of terbutaline sulfate when co-administered to cells in vitro (p<0.0001) (Fig. 4A). 231 cells were not sensitive to terbutaline sulfate. They were, however, sensitive to propranolol and displayed decreased cell proliferation when compared to the control cells (p<0.0001), including when combined with terbutaline sulfate (Fig. 4B).

We proceeded to observe another characteristic of metastasis, migration. After creating a scratch in the plate, we imaged the scratch after 9 h. At the concentrations utilized for this migration assay, we observed significant migration (p<0.0001) in 231Br cells treated with terbutaline sulfate. Approximately 50% of the scratch was filled after 9 h with terbutaline sulfate, and propranolol lessened these effects to ~80% compared to the control migration (p<0.01) (Fig. 5A). At 9 h, there was no significant difference in migration between the control and the propranolol-treated 231Br cells, as expected from data in a previous study (25). In 231 cells, no drug treatments at the utilized concentration made any significant difference in the migration into the scratch (Fig. 5B).

The effects of the β2-adrenergic receptor agonists and antagonists on the third characteristic of metastasis, invasion, was analyzed through an invasion assay. Cells were placed on an invasion insert coated with Matrigel and allowed to invade for 24 h. There was a significant difference (p<0.05) in invasion between 231Br cells treated with propranolol and 231Br cells treated with terbutaline sulfate (Fig. 6A). There was also a significant decrease in 231Br cell invasion following treatment with propranolol + terbutaline sulfate (p<0.05) compared to cells treated with terbutaline sulfate. In the primary 231 breast cancer cells, there was no significant difference in cells treated with propranolol and/or terbutaline sulfate compared to the control cells at the concentrations used for 24 h (Fig. 6B).

Finally, we investigated the effects of propranolol in vivo. 231Br cells labeled with firefly luciferase (231Br-FF) with or without a 12-h non-cytotoxic pre-treatment of propranolol were injected into female NOD/SCID mice and monitored for brain metastases by bioluminescence imaging (BLI). 231Br cells pretreated with propranolol established brain metastasis at significantly decreased rates (p<0.001) compared to the control cells (Fig. 7A and B). By day 27, the BLI showed increased brain metastases in the control cells when compared with the cells pre-treated with propranolol. In fact, the cells pre-treated with propranolol had higher qualitative BLI intensity in the lung, indicating increased lung metastasis. Tumor formation in the brain for mice injected with control 231Br-FF cells was confirmed by staining with hematoxylin and eosin (H&E) (Fig. 7C, left). Lack of tumor formation in the brain for mice injected with 231Br-FF cells pre-treated with propranolol was also confirmed by H&E staining (Fig. 7C, right).