Bile acids destabilise HIF-1α and promote anti-tumour phenotypes in cancer cell models

Cell lines were acquired from the American Tissue Culture Collection (ATCC) (LGC Standards, Middlesex, UK). Human prostate carcinoma cells, DU-145 (ATCC-HTB-81), human lung adenocarcinoma cells, A-549 (ATCC-CCL-185) and human mammary epithelial adenocarcinoma cells, MCF-7 (ATCC-HTB-22) were maintained in basic Dulbecco’s modified Eagle’s medium (Gibco-BRL, Paisley, UK) supplemented with 10 % foetal calf serum (FCS) and 1 % (vol/vol) Penicillin/Streptomycin (both from Sigma, Ireland) (complete medium). Depending on cell type, supplements were added according to ATCC instructions. Cells were maintained in 5 % CO₂ in a humidified atmosphere at 37 °C and were routinely seeded into T75 tissue culture flasks (Sarstedt, Ireland) and grown to a maximum of 80–90 % confluence. At seeding and prior to experiments, cells were washed in Phosphate Buffered Saline pH 7.4 (PBS) to remove cellular debris and dead cells. Cells were detached using 0.5 % Trypsin-EDTA (Sigma-Aldrich) for 5 min, counted on the Countess Cell Counter (Life Technologies, Paisley, UK). Passage numbers did not exceed 25. For all experiments (except untreated cells), cells were supplemented with 200 μm dimethyloxaloglycine (DMOG).

Dihydroxylated bile acids, DCA and CDCA (Sigma) were solubilised in sterile distilled water to a concentration of 50 mM. For all procedures, a final concentration of 100 μM in complete medium was used.

The intracellular HIF-1α staining of cells grown on coverslips in 6 well plates was carried out using the mouse monoclonal antibody for the HIF-1α subunit (Abcam ab16066). Cells were fixed with 4 % formaldehyde in PBS pH 7.4 and incubated with a 1:20 dilution of the monoclonal antibody overnight at 4 °C. The next day, cells were washed with PBS and incubated with a DyLight-488 conjugated secondary antibody. All images were captured on a Zeiss LSM5 using the Zeiss HBO-100 microscope illuminating system. Images were processed using the Zen AIM application imaging program and converted to JPGs using Axiovision 40 Ver. 4.6.3.0. At least three biological repetitions were carried out.

Cells were seeded onto 6 well plates in complete medium and incubated in 5 % CO₂ in a humidified atmosphere at 37 °C for 1–2 days. BAs were added in fresh medium and cells re-cultured for a further 24 h. After this period, cells were scraped into 500 μl chilled PBS. Intracellular HIF-1α activity was assayed using the HIF-1α subunit Human SimpleStep ELISA kit (Abcam) according to manufacturer’s instructions with plates assayed on a 96-well plate reader at 450 nm.

Lactate dehydrogenase (LDH) release was assayed as a measure of cytotoxicity using an LDH colorimetric kit (Roche) according to manufacturers’ instructions. Briefly, cells were seeded onto 96 well plates and treated with 0.1 % Triton X-100 (control) or 100 μM BAs. Following a 16 h incubation at 37 °C and 5 % CO₂, supernatants were removed and added to catalyst reaction mixture in a fresh plate and further incubated at 37 °C and 5 % CO₂ for 30 min to allow for colour development which was quantified on a spectrophotometer plate reader at 490 nm. Cytotoxicity was expressed as a percentage of cells treated with 0.1 % Triton (100 % cytotoxicity). All assays were carried out in triplicate.

Cells were seeded into 6 well plates on sterile coverslips (Sarstedt) and treated as above for 16 h at 37 °C and 5 % CO₂. Coverslips were removed and fixed with 4 % paraformaldehyde for 15 min, washed and stained with 0.1 % crystal violet (Sigma) for 15 min followed by rinsing in water several times and allowing the coverslips to dry. Once dry, coverslips were viewed on a BX21 Olympus microscope and images collected using the Cell’B program and processed using Adobe Photoshop.

Assays were carried out according to manufacturers’ (Invitrogen, UK) instructions. Briefly, cells were grown on coverslips in 6 well plates for 1–2 days under experimental conditions. Cinnabinaric acid (positive control) was added to cells at a final concentration of 150 μm. Cold PBS pH 7.4 was added to wells to wash away debris, after which Annexin V binding buffer was added to wash cells (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl₂ pH 7.4). 15 μl Annexin V conjugate was added per 100 μl Annexin binding buffer along with 0.1 μM 4′,6-Diamidine-2′-phenylindole dihydrochloride (DAPI) for nucleus visualisation. Cells were incubated for 15 min at room temperature after which coverslips were mounted on glass slides and viewed. All images were captured on a Zeiss LSM5 using the Zeiss HBO-100 microscope illuminating system. Images were processed using the Zen AIM application imaging program and converted to JPGs using Axiovision 40 Ver. 4.6.3.0. At least three biological repetitions were carried out.

96 well plates (Sarstedt) were coated with either 15 μg/ml collagen (Sigma) or Matrigel 120 μg/ml (Corning) and stored at 4 °C until required. On the day of assay, plates were equilibrated in at 5 % CO₂ in a humidified atmosphere at 37 °C for 2 h prior to use. Newly passage cells in 100 μl aliquots were seeded onto plates in the presence or absence of BAs and left for 24 h at 5 % CO₂ in a humidified atmosphere at 37 °C hours to adhere. Matrix free wells acted as controls. Following adhesion, cells were washed with PBS × 3 and fixed with 100 μl methanol at −20 °C for 5 min. Once removed, cells were stained for 15 min with 0.1 % crystal violet after which they were rinsed with tap water. Following 10 min air-drying, 50 μl of 0.5 % Triton X-100 was added to wells and left overnight at room temperature with gentle agitation. Plates were assayed on a plate reader at 590 nm. All assays were carried out in triplicate.

Twenty-four-well inserts with an 8 μm pore size polyethylene terephthalate membranes (Corning) were used to assess cell migration of cells using the extracellular matrix, Matrigel (120 μg/ml) which was pipetted onto the membrane. Once set, transwells were either stored at 4 °C until required or equilibrated for 2 h with media in 5 % CO₂ in a humidified atmosphere at 37 °C prior to use. The lower chambers were filled with complete media acting as a chemoattractant. Cells were seeded in each transwell and allowed to migrate over a 24 h period in 5 % CO₂ in a humidified atmosphere at 37 °C. Cells were fixed by replacing culture medium at the top and bottom of the transwell with 4 % formaldehyde in PBS pH 7.4 for 15 min. Once fixed, transwells were washed in PBS, stained with 0.2 % crystal violet for 10 min and washed several times in distilled water. Non-migrated cells were wiped from the transwell top using cotton buds leaving only migrated cells at the bottom of the membrane. These cells were counted using a stereomicroscope and plotted as the percentage of invading cells of the total number of plated cells. All assays were carried out in triplicate.

Twenty-four-well inserts with an 8 μm pore size polyethylene terephthalate membranes (Corning) were used to assess cell migration of cells. The lower chambers were filled with complete media acting as a chemoattractant. Cells were seeded into each transwell insert and allowed to migrate over a 24 h period in 5 % CO₂ in a humidified atmosphere at 37 °C. Cells were fixed by replacing culture media at the top and bottom of the transwell with 4 % formaldehyde in PBS pH 7.4 for 15 min. Once fixed, transwells were washed in PBS, stained with 0.2 % crystal violet for 10 min and washed several times in distilled water. Non migrated cells were carefully wiped away from the top of the transwell using cotton buds leaving only migrated cells at the bottom of the membrane. These cells were counted using a stereomicroscope and plotted as the percentage of invading cells of the total number of plated cells. All assays were carried out in triplicate.

Cells, seeded in 6 well plates at low density in complete medium, were incubated with BAs for a period of 1–3 weeks in 5 % CO₂ in a humidified atmosphere at 37 °C. Plates were left undisturbed for this period. After 3 weeks, medium was removed and cells rinsed with PBS. Next, cells were incubated in 4 % formaldehyde in PBS pH 7.4 for 20 min after which they were rinsed in PBS and incubated with 0.5 % crystal violet (Sigma, Ireland) for 2 h at room temperature. Wells were rinsed in tap water to remove crystal violet and left to air dry overnight. Once dry, colonies were counted using a stereomicroscope to determine plating efficiencies (PE) and survival fractions (SF). PE = no. of colonies formed/no. of cells seeded × 100 %. SF = no. of colonies formed after treatment/no. of colonies seeded × PE. All experiments were carried out in triplicate.

Cells were plated into individual round dishes (Corning) and grown to 80–90 % confluency for 2–4 days in 5 % CO₂ in a humidified atmosphere at 37 °C in serum free media. Once confluent, a “wound” was generated across the cell lawn using a sterile blue tip after which media was removed and replaced with media containing BAs. Time zero dishes were prepared along with 24 h dishes to assess closing of the wound relative to time. After this period, dishes were fixed with 4 % formaldehyde in PBS pH 7.4 for 15 min. Once fixed, dishes were washed in PBS, stained with 0.2 % crystal violet for 10 min and washed several times in distilled water. All dishes were scored for migration distances on a stereomicroscope and distances migrated between “wound” viewed and captured on an Olympus BX-51.

DU-145 cells (untreated, DMOG treated and BAs + DMOG treated cells) were assessed for metabolic activity using the XTT (2,3-Bis-(2-Methoxy-4-Nitro-5-Sulfophenyl)-2H-Tetrazolium-5-Carboxanilide) assay at 24 h intervals (over 4 days) in 96 well plates in 5 % CO₂ in a humidified atmosphere at 37 °C. Briefly, 100 μl freshly prepared XTT (Sigma) and menandione was added to cells in 96 well plates. After a 2 h incubation, cell metabolic activity was determined by measuring the absorbance at 490 nm on a microplate spectrophotometer. All assays were performed in triplicate.

DU-145 cells (untreated, DMOG treated and BAs + DMOG treated cells) were seeded into 24 well plates and assessed for proliferation over a 15 day period. For staining, culture media was aspirated and cells fixed with 4 % paraformaldehyde at room temperature for 10 min. Cells were then stained with 0.05 % crystal violet for 30 min after which cells were washed in water and air dried on filter paper. Crystal violet dye was dissolved in 500 μl methanol and emission spectra measured at a wavelength of 595 nm on a microplate spectrophotometer. All assays were performed in triplicate.

DU-145 cells (untreated, DMOG treated and BAs + DMOG treated cells) were assessed for cell viability using trypan blue (live/dead) staining. Cells were seeded into T25 flasks and assessed for cell viability every 24 h over a 5 day period. On the day of assay, cells were gently scraped from the flask surface and 2 ml of PBS added. The volume was mixed thoroughly (to remove clumps) and centrifuged at 1200 rpm for 2 min to pellet cells. The supernatant was decanted and the pellet suspension mixed. Cells were removed and added to trypan blue and mixed (to remove clumps) thoroughly. An aliquot was loaded onto a disposable chamber slide and placed into the Countess™ Automated Cell Counter (Invitrogen) to count live/dead cells (viability), represented as percentage cell viability. All assays were performed in triplicate. Random counting of cells was also carried out using a haemocytometer and inverted microscope to assess the validity of Automated Cell Counter data. All visual counts were within +/- 5 % of Cell Counter data (data not shown).

DU-145 cells (untreated, DMOG treated and BA + DMOG treated) were collected in RNA Protect Cell Reagent (Qiagen, West Sussex, UK) and stored at −80 °C until RNA extraction, using the RNeasy Minikit as per the manufacturer’s instructions (Qiagen). DNase treatment and reverse transcription were both performed using the Quantitect Reverse Transcription kit (Qiagen). Approximately 1 ug of RNA was reverse transcribed into cDNA as per protocol. Quantitative RT-PCR was performed using the FastStart Taqman probe master mix (Roche) on a PTC-200 thermocycler (Bio-Rad) according to manufacturer’s instructions. Primers were designed using the Universal Probe Library Assay Design Centre (Roche). HPRT-1 (Forward); gaccagtcaacaggggacat, HPRT-1 (Reverse); gtgtcaattatatcttccacaatcaag (Probe 22). HK II (Forward); tcccctgccaccagacta, HK II (Reverse); tggacttgaatcccttggtc (Probe 54).

Three independent biological replicates (N = 3) were performed for each experiment unless otherwise stated. Statistical significance was measured using a paired two-tailed Student’s T-test. Differences were considered to be statistically significant if the p-value were; * P < 0.05; ** P < 0.01.

We previously demonstrated that BAs induced destabilisation of HIF-1α not only in IB3-1 and S-9 epithelial cells but in A-549 lung adenocarcinoma cells [17]. Here, we assessed HIF-1α destabilisation in three cancer cell lines using two independent methods; immunofluorescence with an anti-HIF-1α antibody in A-549, MCF-7 and DU-145 cells (Fig. 1a-c) and an anti-HIF-1α subunit ELISA in DU-145 cells (Fig. 1d). Untreated cells (normoxic conditions) were devoid of green fluorescent punctae suggesting low HIF-1α levels that would indicate rapid turnover of HIF-1α under these conditions. DMOG treated (hypoxia mimetic conditions) cells exhibited widespread HIF-1α punctate staining indicative of stabilised HIF-1α expression as expected. In cells treated with DMOG plus either DCA or CDCA, HIF-1α detection was faint and dispersed in DU-145 cells (Fig. 1c), while in A-549 and MCF-7 cells (Fig. 1a and b), no HIF-1α staining was observed. This suggests a reduction in HIF-1α expression and implicates BAs in HIF-1α destabilisation. We also assessed HIF-1α quantitatively in DU-145 cells using an anti-HIF-1α ELISA (Fig. 1d). As expected HIF-1α expression was increased in DMOG treated cells compared to untreated controls. The presence of BAs reduced HIF-1α expression in DMOG treated cells. This supports the conclusion that BAs generally promote HIF-1α destabilisation under hypoxic conditions. Finally, we investigated the expression of hexokinase II (HK II), a downstream transcriptional target of HIF-1α (Fig. 1e). These data showed that HK II expression increased in DMOG treated cells, whereas for both BAs, HK II expression in DMOG treated cells was not statistically different to the untreated control.

Before assessing the impact of BAs on cancer cell phenotypes, it was important to establish whether or not BAs exhibited cytotoxic activity towards these cells. The effects of dihydroxylated BAs was assessed in all cell lines using a lactate dehydrogenase (LDH) cytotoxicity assay (Fig. 2a-c). Compared with exposure of cells to 0.1 % Triton X-100 that was assigned an arbitrary 100 % LDH release value, untreated cells released minimal LDH levels, reflective of uncompromised plasma membranes and retained intracellular LDH. In the presence of BAs, all cell lines released negligible (<10 %) LDH reflecting the minimal impact of BAs on cytotoxicity at this time-point.

We also assessed the impact of BAs on cell morphology using simply phase contrast microscopy (shown for DU-145 in Fig. 2d). This indicated no obvious differences between untreated and BA treated cells. Overall, these assays indicated that minimal cytotoxic effects were observed upon short-term exposure to BAs, which is in the time frame in which motility and invasion assays can be performed.

The DU-145 cell line was chosen as a cell model in which to establish the impact of bile acids on cell phenotypes associated with metastasis, including adhesion, invasion and migratory capacity. DU-145 cells are derived from an androgen-independent prostate cancer and represents an aggressive cancer cell phenotype that has undergone an Epithelial Mesenchymal Transition (EMT) that can be reversed by suppressing regulators of EMT pathways in vitro [18].

We first established whether BAs triggered apoptosis in DU-145 cells using FITC-labelled Annexin V to detect exposed plasma membrane phosphatidylserine (PS) and DAPI nuclear stain. Untreated control cells did not present any PS staining (Fig. 3). In contrast, cells incubated with the pro-apoptotic positive control [19], cinnabinaric acid (150 μM) exhibited extensive PS staining indicating extensive apoptosis induction. BA-treated cells exhibited negligible green punctae indicating that BAs do not induce apoptosis in DU-145 cells. This assay combined with the cytotoxicity data (Fig. 2) demonstrates that short term exposure to BAs has minimal impact on cytotoxicity and apoptosis. Thus the effects of BAs on cell adhesion, migration and invasion could be further investigated without cell death confounding assays.

The in vivo binding of cells to the extracellular matrix (ECM) is essential for cell survival and cell-to-cell communication, but is also essential for the metastatic potential of cancer cells [20]. The effects of BAs on cell adhesion were assessed in vitro by allowing DU-145 cells to adhere to either collagen or Matrigel in the presence of BAs and assessing the relative numbers of attached cells. Untreated and DMOG-treated cells exhibited similar adhesion to collagen (Fig. 4a) or Matrigel (Fig. 4b). Cells treated with CDCA or DCA exhibited considerably reduced cell adhesion compared to controls (Fig. 4a and b). This suggests that BAs reduce the ability of DU-145 cells to adhere to ECMs which may impact on their migration and invasion capacity.

The in vitro invasive and migratory capacity of metastatic cancer cells provide insights into how cells behave in the presence of modulators of cancer progression. Here, the invasion capabilities of DU-145 cells in the presence of BAs were assessed using Matrigel (on 0.8 μm membranes) (Fig. 5a) and directional migration was assessed using Transwell assays (Fig. 5b). DU-145 cell invasion and migration were both reduced approximately 2-fold compared to untreated control cells. This indicates that BAs impair the ability of these cells to invade and migrate efficiently.

To further explore migratory potential, a wound healing assay was performed (Fig. 6). One of the advantages of this assay is that it mimics the migration of wounded cells in vivo [21]. Untreated DU-145 monolayers were wounded at 0 h, fixed, stained and photographed to establish a basal reference wound. Treated cells were then allowed to fill the wound and at 24 h were similarly stained and photographed to assess cell migration relative to both time and the presence of BAs. Compared to the time 0 h reference wound (Fig. 6) both untreated and DMOG treated cells displayed extensive wound closures after 24 h, reflective of similar cell migratory patterns. Cells exposed to BAs exhibited slightly reduced wound healing patterns at 24 h for both CDCA and DCA respectively. These data demonstrate BAs impaired DU-145 cell migratory potential.

To investigate the long-term effects of exposure of cancer cells to BAs, we used clonogenic assays which measure a key cancer cell phenotype; the ability to form colonies when seeded at low density. The clonogenic potential of A-549, MCF-7 and DU-145 cells was assessed in the presence of BAs over a twenty one day period (Fig. 7a-c). The cell survival fraction (an index of clonogenic growth and cell survivability) was reduced for MCF-7 and DU-145 cells (Fig. 7b and c) but was unaltered for A-549 cells (Fig. 7a). For all cell lines tested, untreated and DMOG-treated cells were characterised by high colony density (no measurable differences). Similarly, BAs had no effect on A-549 clonogenicity as evidenced by dense staining in all plates, high colony numbers and no measurable differences. However, BA-treated DU-145 and MCF-7 cells exhibited greatly reduced colony numbers. These data demonstrate that BAs markedly inhibited the clonogenic and survival potential of DU-145 and MCF-7 cells. This suggests that exposure to BAs would limit the survival and metastatic potential of cancer cells under conditions of hypoxia. Importantly, in DU-145 cells, the reduction in colony formation numbers was independent of metabolic activity, cell proliferation and cell viability, which all remained relatively unchanged over time of exposure to BAs (Fig. 8a–c).