Nitric oxide may regulate focal adhesion turnover and cell migration in MDA-MB-231 breast cancer cells by modulating early endosome trafficking

The MDA-MB-231 human invasive breast cancer cell line (ATCC® HTB-26™) was purchased from the American Type Culture Collection cell bank (Manassas, VA). These cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% of foetal bovine serum (Gibco™, A4766801) and 1% v/v penicillin/streptomycin (10,000 units/ml penicillin, 10,000 μg/ml streptomycin, Gibco). For optimal cell growth, these cells were incubated at 37°C in an atmosphere containing 95% air, 5% CO₂ and approximately 90% humidity. These cells were tested regularly for mycoplasma with the aid of an EZPCR Mycoplasma Test kit (Geneflow, K1-0210, UK), following the manufacturer’s instruction.

Nω-Nitro-L-arginine methyl ester hydrochloride (L-NAME) (Abcam, ab120136) and 1400W (Tocris Bioscience, 1415) were used to inhibit all NOS isoforms and iNOS, respectively. PAPA/NO; NOC-15; 1-(3-aminopropyl)-2-hydroxy-3-oxo-1-propyltriazan (PAPA NONOate) (Santa Cruz, 146672-58-4) was used as the NO donor. L-NAME (5 mM), 1400W (2 mM) and PAPA NONOate (50 μM) were diluted in phosphate buffered saline (PBS). Amilorid (50 μM), Pitstop 2 (25 μM) and Dynasore (20 μM) were diluted in dimethyl sulphoxide (DMSO). The vehicle used in control conditions was either PBS or DMSO (0.5%).

MDA-MB-231 breast-cancer cells (1 x 10⁵ cells/ ml) were seeded on a 6-well-plate and were left to incubate overnight under culture conditions mentioned above. These cells were subsequently treated for 24 hours with PBS (vehicle), NO synthase inhibitors such as L-NAME or 1400W or a nitric oxide donor (PAPA NONOate). Live imaging was used to monitor cell migration; an image was taken every 15 minutes for a period of 24 hours. The speed of migration was calculated as a total traveling distance divided by 24 hours.

MDA-MB-231 breast-cancer cells (1 x 10⁵ cells/ ml) were seeded on a 6-well-plate and were left to incubate overnight under the culture condition mentioned above. These cells were subsequently treated for 24 hours with the vehicle, L-NAME (5 mM), 1400W (2 mM) and PAPA NONOate (50 μM). Live-cell imaging was performed to monitor cell migration with a time-lapse microscope (Nikon Eclipse TiE); an image was taken (at 10x magnification) every 15 minutes over a period of 24 hours using the Nikon Eclipse TE200 running NIS elements software and a Nikon DXM1200 camera. The MtrackJ tool of the ImageJ software was used to determine the total distance covered by individual cells. The speed of migration (expressed in μm/hour) was calculated as a total distance covered by individual cells divided by 24 hours.

MDA-MB-231 breast-cancer cells (1 x 10⁵ cells/ ml) were plated onto a 6-well-plate and were grown to 90% confluence. The medium was removed, a 200 μl pipette tip was used to introduce several wound lines by scratching the cell monolayer. Thereafter, cells were washed three times and incubated in DMEM containing either the vehicle, L-NAME, 1400W or PAPA NONOate. The width of the wound was determined immediately after the addition of the DMEM (t = 0 hour) and 24 hours later, using the Nikon Eclipse TiE microscope. Images of the wound lines were taken at 10x magnification. The percentage of covered area of the wound after 24 hours was calculated as (Average initial area, t = 0 hour) – (Average final area, t = 24 hours) x 100.

24-hours prior to transfection, MDA-MB-231 cells were seeded (1 x 10⁴ cells/mm²) onto 35 mm ibidi culture dishes treated with 2 mg/ml collagen (BD Bioscience, 354236). The cells were left to incubate overnight in order to allow their attachment onto the bottom of the plate. These cells were subsequently transfected with 3 μg plasmid DNA consisting of zyxin-mCherry and with the aid of the TurboFect™ transfection reagent (Thermo Scientific, R0532) at a ratio of 1:1 (w/v, DNA: TurboFect). Approximately 16 hours after transfection, the cells were treated with L-NAME, 1400W or the vehicle for 10 minutes. A live-cell imaging was performed at 100x magnification using confocal microscopy (Nikon A1R). Cells were imaged at the low power excitation laser beam at a wavelength of 488 nm and 568 nm. Images were taken every 5 seconds for 10 minutes to monitor the full lifetime of zyxin assembly and disassembly. The average life-time of zyxin was determined with ImageJ.

To differentiate the endocytotic pathways, it is important to perform the transferring and dextran uptake assay. Cells (1 x 10⁵ cells/ml) were seeded in 6 well plates upon which glass coverslips were placed. After an overnight culture, the cell reached approximately 70% confluence, the medium supplemented with foetal bovine serum was replaced with serum free DMEM. These cells were incubated for 3 hours at 37°C in a humidified atmosphere containing 95% air and 5% CO₂ and were subsequently treated with the vehicle (PBS or DMSO), L-NAME (5 mM), 1400W (2 mM), PAPA NONOate (50 μM), Amilorid (50 μM), Pitstop 2 (25 μM) and Dynasore (20 μM) for 48 hours. Thereafter, the medium of the cells was replaced with serum free DEMEM containing either transferrin conjugated to Alex Fluor 546 (Thermo Fisher, T13342) (25 μg/ml) or dextran conjugated to fluorescein (Thermo Fisher, D1820) (500 μg/ml) and was left to incubate for 30 minutes. The cells were washed twice with ice-cold PBS then fixed with 4% paraformaldehyde (Sigma Aldrich) diluted in PBS (4°C). The cells were washed 3 times with ice-cold PBS and were subsequently treated with an acidic buffer (1M Glycine, 150 mM NaCl, pH 3). The acidic washing step was performed to strip away transferrin/dextran that are bound to the cell surface, since endocytosis blocked at an intermediate stage will cause transferrin/dextran accumulation on the plasma membrane which in turn will affect the total fluorescence measured in the cell. The cells were permeabilised for 10 minutes with PBS containing 0.5% Triton X-100 and were washed 3 times with ice-cold PBS. Confocal images were taken at different layers of the cell (z-stacks) with a resolution of 1024×1024 pixels in order to capture transferring or dextran uptake. The brightness & contrast were then adjusted using ImageJ, then the fluorescence intensity of transferrin and dextran was measured. Alexa Fluor 546 channels and laser power were adjusted as the following values: HV: 64, Offset: 0 and Alexa Fluor 546 nm: 6.35 for all the experiments. Using ImageJ, the region of interest was selected around the cell and the relative intensity of transferrin or dextran was quantified and subtracted from the average background fluorescence in order to determine the number and the size of endosomes containing dextran or transferrin.

The cellular localisation of iNOS, eNOS and early endosome markers were investigated by immunostaning. Cells were incubated (at 37°C) overnight in 6-well plates containing glass coverslips. Thereafter, the growth medium was discarded and replaced with media containing the inhibitors or the NO donor depending on the experiments. After 24 hours, the cells were washed in PBS and fixed for 15 minutes in 4% paraformaldehyde (PFA) in PBS at room temperature. To permeabilised these fixed cells, the excess PFA was washed away with PBS and the cells were incubated at room temperature for 10 minutes in PBS containing 0.5% Triton X-100. These cells were subsequently treated with PBS containing 10% goat serum for 30 minutes at room temperature. Antibodies raised against EEA1 (Cell Signalling, #2411S), eNOS (Cell Signalling, #9572S) or iNOS (Abcam, ab3523) diluted to 1:100 in PBS containing 2% goat serum were added to the cells and incubated for 60 minutes at room temperature. Unbound antibodies in excess were washed away with PBS and the cells were transferred into a box where they were protected from light. Secondary antibodies conjugated with Alexa Fluor 546 or 488 (Cell signalling, #44125, #4408S) diluted to 1:200 in PBS containing 2% goat serum were added to the cells and incubated for 30–60 minutes at room temperature. These cells were washed with PBS and were visualised using confocal microscopy.

The number and the size of FAs and endosomes were analysed on fixed cells. All images were acquired with a confocal microscopy (Nikon Eclipse Ti Laser-scanner), at 100x magnification and a resolution of 1024 × 1024 pixels using the Nikon confocal system software. Images taken at depths where both endosomes and FA were both present. The acquired images were processed by subtracting the background; the average number and size of both FA and endosomes were determined with ImageJ. To investigated whether the localisations of FA and endosomes were correlated, the Spearman’s rank correlation coefficient was determined within a selected region of interest using ImageJ.

The potential link between early endosomes and NO were investigated using two methods; Biotin switch assay and western blotting to detect whether early endosomes proteins were S-nitrosylated. S-nitrosylated cysteine residues were detected with the S-nitrosylated Protein Detection kit from (Cayman Chemical, 75816-684), following the manufacturer’s instruction. After cells were washed twice with 5 ml ice-cold PBS, they were removed from the flask then centrifuged at 500g for 5 minutes at 4°C. Cell pellets were resuspended with 0.5 ml buffer A, containing a blocking reagent, and were incubated for 30 minutes at 4°C under mild agitation. 2 ml of ice-cold 70% acetone (diluted in deionised distilled H₂O) was added to the resuspended cells and were left to incubate at -20°C for 1 hour. The samples were subjected to centrifugation at 3000g for 10 minutes (4°C) and the pellets were resuspended and incubated for 1 hour at room temperature in buffer B, containing reducing and labelling reagents. 2 ml of ice-cold 70% acetone (diluted in deionised distilled H₂O) was added to the samples and incubated at -20°C for 1 hour. The pellets were retained following a centrifugation at 3,000g for 10 minutes (4°C). After the resuspension of the pellet in ice-cold PBS (100 μl), the cells were lysed overnight at 4°C with 100 μl of RIPA buffer, containing streptavidin-agarose bead (20 μl). The samples were subsequently centrifuged at 14,000g for 30 seconds (at 4°C) and the pellets were washed three times with 300 μl with RIPA buffer. Following the streptavidin pulldown, the samples were denatured at 95°C (5 minutes) in reduced laemmli buffer containing 2.5% β-mercaptoethanol. Samples treated with 100 μM mercury chloride (HgCl₂) to cleave S-NO bonds in proteins prior to the assay to serve as negative control for the assay. Proteins in the samples were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), then transferred onto a polyvinylidenedifluoride (PVDF) membrane. The PVDF membrane was incubated for 1 hour at room temperature in a blocking solution made up of 5% dry milk solubilised in Tris-Buffered Saline Tween (TBST). The membranes were subsequently washed three times in TBST (10 minutes each time) and were incubated overnight at 4°C under mild agitation with primary antibodies such as anti-Rab5 (Cell Siganlling, #3547), anti-EEA1 (Cell Signalling, #3288), anti-APPL1 (Cell signalling, #3858), anti-H2B (Abcam, ab52484) and anti-GAPDH (Sigma, G8795) diluted in either 5% dry milk or 5% BSA solubilised in TBST (1:1000). The membranes were washed three times in TBST (10 minutes each time) and incubated for 1 hour at room temperature under agitation with horseradish peroxidase-conjugated antibodies raised against mouse (1:2500, in 5% of dry milk) and rabbit (1:2500, in 5% of dry milk) IgGs (Sigma-Aldrich, AP306P, AP307P). The membranes were subsequently washed three times with TBST (10 minutes each time) and an enhanced chemiluminescence solution was added to the membrane. Images of the immunoreaction were acquired using the ImageQuant LAS 4000 imager (GE Healthcare Life Sciences) and Image Quant TL software.

GPS-SNO and IntFOLD tools were used to identify buried and exposed cysteines residue in proteins in order to determine what residue which is most likely to be S-nitrosylated. Three dimensional structural models of human early endosomal proteins such as EEA1, APPL1 and Rab5 were generated with the IntFOLD server [23,24]. GPS-SNO analysis was used identify cysteines residues in each protein. Accessibility of cysteine residues within the protein three-dimensional structure was assumed to determine a potential to be S-nitrosylated. The GPS-SNO S-nitrosylation site prediction algorithm was used to predict the specific cysteine residue that is likely to be S-nitrosylated [25]. The prediction algorithm was applied to APPL1, EEA1, Rab5 and other proteins including H2B and ubiquitin.

The data obtained by the imaging software were statistically analysed with GraphPad prism 5 software (GraphPad software, San Diego, CA). One-way analysis of variance (ANOVA) was performed on data resulting from experiment comprising three or more experimental groups. Following a significant difference identified via ANOVA, a Dunnett’s post-hoc test was used for pairwise comparisons in order to compare each treatment to a single control (vehicle). All results were obtained from at least three independent experiments in triplicate (n=3). A value of p-value ≤ 0.05 was considered statistically significant.

To evaluate the effect of NO on cell migration, a wound healing assay and a cell tracking assay was employed on MDA-MB-231 cells treated with the iNOS inhibitor 1400W or the whole NOS inhibitor L-NAME or the NO donor PAPA NONOate. Our results showed that cells treated with L-NAME or 1400W displayed a reduced covering area 24 hours after scratching (Fig. 1A and 1B). PAPA NONOate showed no effect on wound-healing (Fig. 1A and 1B). As shown in figure 1C, the speed of cell migration was reduced in cells treated with L-NAME (17.1 ± 0.7 μm/hour, p<0.01) and 1400W (19.0 ± 1.3 μm/ hour, p<0.05) compared to that of vehicle-treated cells (23.5 ± 0.8 μm/hour). However, PAPA NONOate exhibited no effect on the speed of cell migration (23.4 ± 0.3 μm/hour, p>0.05). Next, we investigated whether the inhibition of NOS or iNOS, as well as the addition of NO, can affect FA turnover. Cells transfected with zyxin-mCherry displayed a prolonged duration of FA turnover when treated with L-NAME (51.2 ± 2.0 seconds, p<0.05) or 1400W (48.8 ± 2.3 seconds, p<0.05) compared to vehicle-treated cells (36.0 ± 3.8 seconds), while PAPA NONOate had no effect on the FA turnover (38.3 ± 1.7, p>0.05) (Fig. 2A and 2B).

Figure 1

Figure 2

To examine the effect of NOS inhibitors on two endocytotic pathways, namely endocytosis or micropinocytosis, we measured the uptake of transferrin and dextran in MDA-MB-231 cells treated with L-NAME and 1400W. As shown in figure 3A and 3B, the inhibition of micropinocytosis with Amiloride as well as the inhibition of dynamin (Dynasore) and clathryn-dependent endocytosis (Pitstop 2) reduced the uptake of transferrin and dextran. The number of vesicles containing transferrin decreased with Amiloride, Dynasore and Pitstop 2 treatment. Cells that were treated with Amiloride and Pitstop 2 exhibited a decreased number of vesicles containing dextran, while those treated with Dynasore showed an increased number of vesicles containing dextran. Regarding the size of vesicles containing transferrin or dextran, neither Amiloride nor Dynasore nor Pitstop 2 had an effect. Both L-NAME and 1400W did not affect transferrin and dextran uptake, since the intensity of fluorescence of transferrin and dextran did not change in L-NAME (433.0 ± 37.8 arbitrary unit) and 1400W (447.3 ± 3.8 arbitrary unit) treated cells compared to vehicle-treated cells (421.3± 55.2 arbitrary unit) (Fig. 3C and 3D). Similarly, these NOS inhibitors did not change the number of vesicles containing dextran and transferrin as the number of vesicles containing transferrin and dextran in cells treated with L-NAME (22 ± 1) and 1400W (19 ± 2) did not differ from that of vehicle-treated cells (23 ± 2) (Fig. 3C and 3D). The size of vesicles containing transferrin and dextran were found to be comparable across cells treated with L-NAME (1.5 ± 0.1 μm²), 1400W (1.7 ± 0.2 μm²) or the vehicle (1.7 ± 0.1 μm²) (Fig. 3C and 3D).

Figure 3

To examine whether NO is involved in endocytosis regulation through early endosomal compartments, the impact of NOS inhibition and exogenous NO on the number and size of EEA1-positive endosomes and Rab5-positive vesicles were assessed in MDA-MB-231 cells. As shown in figure 4A, the number of EEA1-positive endosomes was found to be lower in cells treated with L-NAME (24 ± 2, p<0.05) and 1400W (20 ± 7, p<0.05) compared to that vehicle-treated cells (44 ± 3). However, PAPA NONOate did not affect the number of EEA1-positive endosomes (41 ± 1, p>0.05). The analysis of the size of EEA1-positive endosomes showed that cells treated with L-NAME (1.0 ± 0.1 μm², p<0.01) or 1400W (1.1 ± 0.1 μm², p<0.01) exhibited an increased size of these endosomes compared to vehicle-treated cells (0.5 ± 0.1 μm²). The size of EEA1-positive endosomes did not differ between cells treated with PAPA NON-Oate (0.8 ± 0.1 μm², p>0.05) and that of vehicle-treated cells (Fig. 4A). Next, investigate the effect of NOS inhibition on the number and size of Rab5-positive vesicles. Our results in figure 4B showed that both L-NAME and 1400W inhibitors exerted no effect on the number of Rab5-positive vesicles since the number of Rab5-positive vesicles was not affected in cells treated with L-NAME (35 ± 1) or 1400W (36 ± 1) compared to vehicle-treated cells (33 ± 1). The size of Rab5-positive vesicles was not affected as a result of L-NAME (1.2 ± 0.1 μm²) or 1400W (1.5 ± 0.2 μm²) treatment, since the size of Rab5-positive vesicles in vehicle-treated cells was found to be 1.4 ± 0.2 μm².

Figure 4

Given that in our previous results, we demonstrated that the inhibition of NOS regulates the number and the size of EEA1-positive endosomes, we subsequently investigated whether the localisation of EEA1-positive endosomes and that of eNOS or iNOS correlates using immunocytochemistry on fixed MDA-MB-231 cells. As shown in figure 5A, the localisations of EEA1 and eNOS were found to be correlated, as the correlation coefficient was found to be 0.64 ± 0.17. Similarly, the correlation coefficient between the localisations of EEA1 and iNOS was 0.63 ± 0.11.

Figure 5

Following the investigation of the localisation of eNOS and iNOS in relations to EEA1, we assessed the S-nitrosylation of proteins implicated in the formation of early endosomes using the biotin switch assay. The Analyses by western blot revealed that non-S-nitrosylation proteins such as H2B or ubiquitin were not detected by the assay (Fig. 5B). Our data showed the early endosomal proteins such as APPL1, EEA1 and Rab5 were found to be S-nitrosylated. To ensure that the S-nitrosothiols group on these proteins were accurately detected, the lysates were treated with HgCl₂ to cleave S-NO bonds prior the biotin switch assay. As a result, HgCl₂ treated lysates showed no immunoreaction in all evaluated proteins (Fig. 5B).

Once it was ascertained that early endosome markers were S-nitrosylated, a bioinformatic analysis that uses the GPS-SNO algorithm was performed on APPL1, EEA1, Rab5 and other proteins including H2B and ubiquitin in order to predict cysteine residues with the potential to be subjected to S-nitrosylation. The results showed that APPL1 presented two potential cysteine sites at position 99 and 603 that could be S-nitrosylation; four cysteine sites at position 46, 255, 894 and 1102 were identified in EEA1; a single cysteine site for potential S-nitrosylation was found in Rab5, at position 212. However, both H2B and ubiquitin expectedly showed no potential sites for S-nitrosylation (Fig. 5B). Next, we used the IntFOLD software to generate three-dimensional models of early endosome proteins. The tool was subsequently used to identify cysteine residues that are easily accessible, therefore most likely to be S-nitrosylated. As shown in figure 6, the three-dimensional structure of EEA1 showed that two cysteine residues, at position 255 and 1102, are hidden within the structure of the protein, thus suggesting that these sites are not likely to be subjected to S-nitrosylation. Concerning the other two cysteine residues, at position 46 and 894, they appear to be exposed, suggesting that these cysteine sites could be S-nitrosylation. The three-dimensional model of Rab5 showed that the cysteine residue, at position 212 was located in an accessible area of the protein, therefore would be likely S-nitrosylated (Fig. 7). However, a three-dimensional model for APPL1 indicated that the location of both cysteine residues was hidden and thus would be unlikely to be S-nitrosylated (Fig. 8).

Figure 6

Figure 7

Figure 8