Breast cancer remains the second most common cancer in women worldwide and the fourth leading cause of cancer death [1
]. Breast cancer has been classified traditionally according to characteristics such as tumour size, lymph node involvement, histological grade and the receptors expressed: estrogen receptor (ER+) and/or progesterone receptor (PR+), human epidermal growth factor receptor 2 (HER2) amplified (referred to as HER2+), and triple negative breast cancer (TNBC), which lacks the expression of all three markers (ER and PR, and HER2) [2
]. Endocrine therapy is the main initial therapy for ER+/PR+ breast cancer; however, inherent and acquired resistance leads to disease progression and incurable disease (reviewed in [3
]). Similarly, whilst anti-HER2 therapies, such as trastuzumab in combination with chemotherapy, have significantly increased overall survival for patients with early and advanced HER2-amplified breast cancer, resistance mechanisms lead to recurrent and metastatic disease [4
]. TNBC is unresponsive to endocrine or anti-HER2 therapies, since it lacks the expression of the necessary receptors. TNBC accounts for 10–15% of all diagnosed breast cancer and has the poorest prognosis due to its biologically aggressive nature. Treatment relies on surgery, chemotherapy and radiotherapy; however, the disease-free interval with neoadjuvant or adjuvant therapy is shorter and metastatic disease more aggressive than other types of breast cancer [5
]. Despite a number of clinical trials in TNBC, drug treatment regimens are yet to be optimised, highlighting a need for additional novel drug therapies.
Terpenoids are a diverse family of natural products found in a variety of fruits, vegetables and medicinal plants [6
]. Classified according to the number of cyclic structures that they contain, terpenoids are a major class of anticancer compounds. Perhaps the most well-known terpenoid in clinical use is paclitaxel, a diterpenoid used for the last 30 years to treat early and metastatic breast cancer as a first-line or second-line treatment, either alone or in combination with other chemotherapeutics [7
]. Disadvantages of its use, however, include adverse effects and multi-drug resistance [8
]. Another subgroup, triterpene saponins, are noted to have cytotoxic, anti-mutagenic, anti-inflammatory and anti-cancer activities (reviewed in [10
]). However, limited studies have investigated the anti-cancer potential of the bacopasides, which are members of this subgroup, derived from the medicinal herb Bacopa monnieri
], and no studies have been performed using the purified compounds bacopaside (bac) I or II. Use of the whole plant extract has been used in traditional Indian medicine for 3000 years, primarily to enhance memory and cognitive function [11
], with no evidence of adverse reactions or side effects in humans (reviewed in [16
]). We therefore aimed to explore the anti-cancer potential of bac I and bac II against breast cancer.
To further support the use of bac I and bac II in cancer, we previously showed that these drugs block the functional activity of the transmembrane protein aquaporin (AQP) 1 [12
]. AQP1 is involved in ion and water transport across the cell membrane and is expressed in human tissues, including in renal tubules, intestinal epithelia, breast epithelia and vascular endothelial cells [17
]. Bac I specifically blocks both the ion and water channel function of AQP1, while bac II blocks the water channel function only [12
]. Several studies support a role for AQP1 in tumour progression: AQP1-null mice showed reduced tumour growth, metastasis and angiogenesis compared to AQP1-expressing mice [18
]; mammary gland tumour cells transfected with AQP1 showed increased migration in vitro and increased tumour cell extravasation and lung metastases in vivo [20
]; and in vitro inhibition of AQP1 reduced endothelial cell tube formation and the migration, invasion, and growth of colorectal cancer cell lines [12
]. Clinically, AQP1 expression was increased in breast cancer compared to adjacent normal tissue [23
], and high AQP1 expression was associated with TNBC and poorer progression-free and overall survival [24
Together, these studies provide a strong rationale to investigate the effects of bac I and II in breast cancer. We therefore tested the anti-cancer activity of the drugs individually and combined on breast cancer cell lines from the major molecular subtypes, MDA-MB-231 (TNBC), T47D (ER+/PR+), MCF7 (ER+/PR−), and BT-474 (HER2+), and looked for potential synergy between the two drugs.
The cell membrane permeabilisation effects of high concentrations of saponins lead to hemolysis and cytotoxicity. At lower doses, however, anti-cancer properties such as the inhibition of proliferation, induction of apoptosis and attenuation of invasiveness have been demonstrated (reviewed in [27
]). Here, we have shown for the first time that the triterpene saponins bac I and bac II in combination act in synergy to reduce the viability, proliferation, migration, and invasion of breast cancer cell lines in vitro. This anti-cancer activity of bac I and bac II occurred at high dose combinations that induced G2/M arrest and apoptosis but also at lower dose combinations that did not induce apoptosis. These lower non-cytotoxic combinations of bac I and II significantly reduced the migration of all four cell lines, with MDA-MB-231 and BT-474 being the most sensitive, and significantly reduced the spheroid invasion of MDA-MB-231.
There are no studies investigating the anti-cancer effects of combining bac I and II. Previous studies have focused on the anti-cancer effects of either the whole plant extract, bacoside A (a mixture of saponins from B. monnieri
containing approximately 20% bac II w/w [28
]), or bac I and bac II alone. In vitro, B. monnieri
extract was cytotoxic to human cancer cell lines of the colon, lung, cervix, and breast [29
]; bacoside A to human kidney carcinoma [32
] and glioblastoma [33
] cell lines; and bac I and bac VII to prostate, glioma, ileocecal, lung and breast cancer cell lines [11
]. Similarly, we have shown bac II to induce apoptosis in vitro in colorectal cancer (CRC) cell lines [14
] and in endothelial cells, suggesting anti-angiogenic potential [13
]. We have also shown reduced migration of CRC cell lines expressing high AQP1 with bac I and bac II alone [12
]. In vivo studies in mice have shown the chemopreventive activity of bacoside A in chemically-induced hepatocellular carcinoma (HCC) [34
] and of B. monnieri
extract in skin carcinogenesis [35
], as well as anti-cancer activity of the extracts bac I and bacoside A in mouse models of melanoma [36
], sarcoma [11
], and Ehrlich ascites carcinoma (EAC) [32
These studies provide strong evidence to investigate the anti-cancer efficacy of bac I and bac II combined in breast cancer cell lines from each of three main molecular subtypes, as performed here for the first time. We found that all cell lines were sensitive to the synergistic effects of bac I and bac II combined, but particularly MDA-MB-231 and BT474, representative of the more aggressive subtypes of breast cancer of TNBC and HER2+, respectively. We also found that certain dose levels and combinations were cytotoxic, consistent with previous studies using the whole extract and its derivatives.
A novel aspect of our study, however, was the finding that lower doses of bac I and II combined, at doses that did not induce apoptosis, also had potent anti-cancer activity in vitro. Previously, we showed that non-cytotoxic concentrations of bac II alone inhibited endothelial cell tube formation—an in vitro measure of angiogenesis [13
]—but otherwise, only one study has shown the reduced motility of prostate cancer cell line DU145 with non-cytotoxic concentrations of B. monnieri
]. This promising finding suggests that these drugs may have the potential to slow cancer progression and prevent metastasis, without the cytotoxicity to normal cells associated with chemotherapy [37
In terms of safety, the highest doses used in our in vitro study to induce apoptosis were 10 μM bac I and 5 μM bac II, equivalent in mice to approximately 7 mg/kg and 4 mg/kg, respectively, while the human equivalent dose [38
] would be approximately 0.6 mg/kg and 0.3 mg/kg, respectively. Extracts of the dried herb typically contain bac I and bac II in a 1:1 ratio, depending on the extraction method [39
]. Rat toxicity studies showed no toxicity of B. monnieri
extract containing bac I and bac II at levels higher than these. A large single oral dose of 5000 mg/kg of whole extract, containing 51.5 mg/kg bac I and 91 mg/kg bac II, caused no acute toxicity in Sprague-Dawley rats; in addition, rats receiving up to 1500 mg/kg of whole extract daily for 270 days, containing 15.5 mg/kg bac I and 27.3 mg/kg bac II, showed no evidence of chronic toxicity [41
]. Others have studied the pharmacokinetics of bac I [42
] and clinical effects of bac I, but not bac II, with no adverse effects in rats receiving 30 mg/kg orally for 6 days [43
] and mice 50 mg/kg orally for 7 days [43
]—5 to 8 times more bac I than the highest dose of bac I used in the present study. However, other animal studies and human clinical trials have used B. monnieri
whole extract at too low a dose to be comparable to the amounts of bac I and bac II used here [44
]. In terms of translation to human trials, based on the guidelines of the U.S. Department of Health and Human Services, Food and Drug Administration (FDA) [48
], the estimation of the maximum safe starting dose in initial clinical trials is 10% of the predicted therapeutic dose, so for bac I and bac II in adults, this would be doses of 0.06 and 0.03 mg/kg, respectively, which is equivalent to approximately 4.2 mg or 2 mg, respectively, in an average 70 kg human. Even at the predicted therapeutic doses of 42 mg Bac I or 20 mg Bac II, these doses are much lower in comparison with other commonly used agents for metastatic breast cancer.
The rationale for combining the two drugs was based on their ability to block AQP1 ion and water channel function. Both bac I and bac II block the water channel function of AQP1, but only bac I blocks the ion channel function [12
]. We therefore hypothesised their anti-cancer potential when combined, given that separately we have shown they reduced the migration of CRC cell lines that expressed high levels of AQP1 [12
]. We found that the MDA-MB-231 cell line expressed the highest level of AQP1 transcript of the four breast cancer cell lines tested. This is consistent with a study showing relatively high AQP1 expression in TNBC [26
]. Additionally, we found that, for the MDA-MB-231 cell line, AQP1 transcript expression could be reduced by bac I and II combined. This suggests the effects of the combined drugs might, in part, be mediated through AQP1. A limitation of our study, however, is the lack of AQP1 silencing or knockout experiments to confirm the role of AQP1 in the observations, which will be investigated in future studies.
There have been reports of other suggested mechanisms of anti-cancer activity for extracts of B. monnieri
and bacoside A, the latter of which contains bac II but not bac I. Both induced the death of human glioblastoma cell lines in vitro through macropinocytosis, a process whereby cell membrane buckling leads to the accumulation of extracellular fluid inside large vacuoles within the cell [33
]. This build-up of fluid within the cells might be as a result of the inhibition of AQP1 channel activity and hence inhibition of water flux. The morphological changes described in glioblastoma cells are consistent with the vacuolisation we observed in breast cancer cell lines with combined bac I and bac II treatment. In addition, the bacoside A co-treatment of male Wistar albino rats treated with N-nitrosodiethylamine (DEN) to induce hepatocellular carcinoma (HCC) reduced serum marker enzymes that are markedly elevated in HCC and increased the level of enzymatic and non-enzymatic antioxidants to near normal, thereby reducing disease-associated hepatic damage [34
]. They also reported that bacoside A attenuated the increased expression of matrix metalloproteinase (MMP)-2 and -9 in DEN-induced HCC. Since MMPs are involved in tumour cell invasion and metastasis, an anti-metastatic effect of bacoside A was suggested [49
]. The potential of combined doses of bac I and II to act via similar mechanisms in breast cancer should therefore be explored.
4. Materials and Methods
4.1. Reagents and Cell Lines
The analytical standards bacopaside I (CAS No. 382148-47-2, 89.6% purity by HPLC, Lot no. 00002002-T17H) and bacopaside II (CAS No. 382146-66-9, 98% purity HPLC, Lot Number: 00002002-T17H), derived from the medicinal herb bacopa monnieri, were obtained from ChromaDex (Irvine, CA, USA), solubilised in methanol at 10 mM and 1.5 mM stock solutions, respectively, and stored at −20 °C. Cell lines MDA-MB-231, T47D, MCF7 and BT-474 were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA). Cells were maintained in complete medium, either in Dulbecco’s modified Eagle’s medium (DMEM; Life Technologies, Carlsbad, CA, USA) for MDA-MB-231, MCF7 and BT-474 or Roswell Park Memorial Institute (RPMI) 1640 medium (Life Technologies, Carlsbad, CA, USA) for T47D, containing 10% foetal bovine serum (FBS) (Corning, NY, USA), 200 U/mL of penicillin, 200 μg/mL of streptomycin (pen strep; Life Technologies, Carlsbad, CA, USA) and 2 mM L-alanyl-L-glutamine dipeptide (GlutaMAX Supplement; Life Technologies, Carlsbad, CA, USA). Cells were grown under standard culture conditions at 37° C with 5% CO2 in air and used within 4 passages.
4.2. Cell Viability Assay for Determining IC50 and Drug Synergy
The effect of bac I and bac II, alone and in combination, on cell viability was determined by MTS assay as described previously [13
]. Cells were seeded in complete medium at 1 × 104
cells per well of a 96 well plate and incubated under standard culture conditions overnight. Next, cells were treated with either vehicle (1% methanol), bac I, bac II or combinations of bac I and bac II, for 24 h. MTS assay was performed using the CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay (Promega, Madison, WI, USA) according to the manufacturer’s instructions. Absorbance was read at 492 nm and results were calculated as the mean absorbance normalized to the vehicle control. The half maximal inhibitory concentration (IC50
) for each drug alone and in combination was determined by non-linear regression analysis.
To determine the IC50
of bac I and bac II combined, bac II was used at a constant concentration of 2.5 μM whilst the concentration of bac I was varied. The isobologram method [50
] was used to produce a theoretical line by plotting the IC50
s for bac I and bac II alone on the y
- and x
-axis respectively. The IC50
combination of bac I and bac II was also plotted on the same graph. Its position would indicate whether the IC50
combination was antagonistic, additive, or synergistic, depending on whether the point was above, on or below the theoretical line respectively. This was also reflected by the combination index, calculated by the formula, CI = d1/Dx
1 + d2/Dx
2, where Dx
1 was the IC50
of bac I, Dx
2 was the IC50
of bac II, and d1 and d2 were the doses of bac I and bac II respectively that combined gave an IC50
. The combination was antagonistic, additive or synergistic if the combination index was greater than, equal to or less than one [50
4.3. Cell Proliferation
The crystal violet assay was used to measure cell proliferation and was performed as described previously [52
]. Briefly, cells were seeded in complete medium at 2 × 103
(MDA-MB-231 and MCF7), 3 × 103
(T47D) or 4 × 103
(BT-474) cells per well of a 96 well plate followed by overnight incubation. Cells were then treated with the vehicle or bac I and/or bac II, and crystal violet absorbance at 595 nm was measured on days 0, 1 and 3 of treatment.
4.4. Cell Cycle Analysis
Cell cycle analysis was performed as previously described [13
]. Briefly, 1 × 105
cells were seeded per well of a six-well plate, incubated overnight, then treated with either the vehicle or combinations of bac I and bac II for 24 h. Cells were resuspended in ice-cold Dulbecco’s phosphate buffered saline (DPBS) with drop-wise addition of 100% ice cold ethanol to a final concentration of 70% ethanol. Cells were centrifuged at 300× g
for 5 min and cell pellets resuspended in 0.25% Triton X-100 in DPBS. Cells were stained for 2 h in darkness with 25 μg/mL propidium iodide (Sigma-Aldrich, St Louis, MO, USA) and 40 μg/mL bovine pancreas ribonuclease A (Sigma-Aldrich, St Louis, MO, USA) in DPBS and analysed on the BD FACSCanto II cell analyser (BD Biosciences, San Jose, CA, USA). Doublet populations were excluded and 50,000 single cell events were captured per sample. Data was analysed using FlowJo software v10.4.0 (FlowJo, LLC, Ashland, OR, USA), specifically by applying the Watson model to determine the percentage of cells in each stage of the cell cycle.
4.5. Apoptosis Assay
Apoptosis was determined by annexin-V and propidium iodide (PI) staining, using the Annexin-V-FLUOS staining kit (Roche Diagnostics, Mannheim, Germany), as previously described [14
]. Briefly, cells were seeded at 1 × 105
per well of a six well plate, cultured overnight then treated with either vehicle or bac I and bac II for 24 h. Cells were harvested and stained with annexin-V and propidium iodide for 15 min in darkness at room temperature then run on a BD FACSCanto II cell analyser (BD Biosciences, San Jose, CA, USA). Compensation was performed to account for the spectral overlap between fluorochromes. Cells were gated to exclude debris and doublets and at least 10,000 single cell events were acquired per sample. Data was analysed using FlowJo software v10.4.0 (FlowJo, LLC, Ashland, OR, USA).
4.6. Scratch Wound (Wound Closure) Migration Assay
The scratch wound assay was performed as described previously [13
]. Cells were seeded in complete medium at 3 × 104
(T47D and MCF7), 4 × 104
(MDA-MB-231), or 5 × 104
(BT-474) cells per well of a 96 well plate, incubated under standard conditions until 80–90% confluent, then serum starved overnight. A circular wound was made with a p10 pipette tip. Cells were resuspended in complete medium supplemented with 1 μg/mL mitomycin C (Sigma-Aldrich, St Louis, MO, USA) to inhibit cell proliferation, and either vehicle or bac I and/or bac II. The area of wound closure was measured using NIS Elements software (Nikon, Tokyo, Japan).
4.7. Spheroid Invasion Assay
The Cultrex® 3D culture spheroid cell invasion assay (Trevigen, Gaithersburg, MD, USA) was set up as per the manufacturer’s instructions using the kit contents, including spheroid formation extracellular matrix (ECM) and invasion matrix. MDA-MB-231 cells were seeded in the spheroid formation ECM at 3000 cells per well of a Costar® 96-well ultra-low attachment plate (Corning, NY, USA) followed by incubation for 72 h. Spheroids were subsequently embedded in invasion matrix and vehicle or drug treatments were added. The area of invasion on day 3 was measured using NIS elements software (Nikon, Tokyo, Japan) and was expressed normalised to the mean invasion of the vehicle control.
4.8. Analysis of Aquaporin-1 Expression by Quantitative PCR
Breast cancer cell lines were seeded at 3 × 105
cells per well in six-well plates and incubated until 70 to 80% confluent. RNA was isolated with the PureLink RNA mini kit (Life Technologies, Carlsbad, CA, USA) at time 0 and following 24 h treatment with vehicle or combinations of bac I and bac II. The iScript cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, CA, USA) was used to reverse transcribe 200 ng of RNA. Transcript expression was determined as previously described [14
], using multiplex TaqMan Gene Expression Assays for aquaporin-1 (AQP1; Hs01028916_m1; Applied Biosystems, Foster City, CA, USA) as the target gene, and coiled-coil serine rich protein 2 (CCSER2; Hs00982799_mH; Applied Biosystems, Foster City, CA, USA) as the reference gene. Reactions were performed using the Applied Biosystems ViiA 7 Real Time PCR System (Life Technologies, Carlsbad, CA, USA) with activation for 30 s at 95 °C, followed by 40 cycles of 15 s at 95 °C, and 30 s at 60 °C. Results were calculated using the 2−ΔCt
relative quantification method, normalizing to CCSER2 reference gene.
4.9. Statistical Analysis
Data were analysed using GraphPad Prism version 7.0c for Mac OS X (GraphPad Software, La Jolla, CA, USA) and is presented as means ± standard deviation (SD). Statistical analysis was done by one-way analysis of variance (ANOVA) with Bonferroni post-hoc test unless otherwise stated.