The effect of cycling hypoxia on MCF-7 cancer stem cells and the impact of their microenvironment on angiogenesis using human umbilical vein endothelial cells (HUVECs) as a model

Culture conditions

The breast cancer cell line (MCF-7) was purchased from the American Type Culture Collection (ATCC, USA), and was expanded as a monolayer in vented 75 cm² cell culture flasks (Membrane Solutions, USA) using RPMI 1640 media (HyClone, Logan, UT, USA). The media was supplemented with heat-inactivated fetal bovine serum (FBS) 10% (v/v) (HyClone, USA), antibiotics (100 µg/ml streptomycin and 100U/ml penicillin) (HyClone, USA), and 1% 2mM L-glutamine (HyClone, Logan, UT, USA). The cells were incubated in a designated incubator (NuAire, Shanghai, China) at (5% CO₂ at 37 °C). All cell culture procedures were performed in sterile conditions under a class II biological safety cabinet (Heal-Force, Shanghai, China). All used materials and disposables were disinfected with ethanol 76% before use.

Mammosphere cultures

To obtain CSCs from MCF-7 cell line and propagate them as mammospheres, MCF-7 cells were expanded in complete growth media described above as a single-cell suspension in a standing flask to enhance the growth of CSCs as they tend to grow in non-adherent conditions. Two days later, remaining cells were collected by centrifugation, washed in phosphate buffer saline, and seeded in six well ultra-low adherence plates (Corning, Corning, NY, USA) at a density of 10,000 cells/well. Cells were expanded in serum-free DMEM/F12 (Sigma-Aldrich) media, supplemented with 20 ng/mL basic fibroblast growth factor (GIBCO, USA), 20 ng/mL epidermal growth factor (MPBio, Santa Ana, CA, USA), 2% B27 (GIBCO, USA), 10 µg/mL insulin (MPBio, Santa Ana, CA, USA), 0.5 µg/mL hydrocortisone (Nacalai, Japan), 0.4% bovine serum albumin (Sigma, St. Louis, MO, USA), and 2 mM L-Glutamine (Biowest Co., Nuaillé, France). This is called mammospheres culture media. Cells were preserved in a humidified incubator at 37 °C, 5% CO₂. Every two days, fresh media (2 mL) were added to each well (without removal of the old one). At day seven, primary mammosphere clusters were counted, then collected and centrifuged to form a pellet. In order to obtain single cells, the pellet was enzymatically disaggregated with trypsin and dispersed by passing through a 40 µm pore-size filter (Millipore, Billerica, MA, USA). Single cells of primary mammospheres were collected for sorting, using beads technique and identified by flow cytometry (detailed below) before and after sorting. Disaggregated cells were cultured in ultra-low adherence flask (Corning, Corning, NY, USA) instead of well plates to obtain secondary mammospheres. Then, the same procedure was applied again to obtain tertiary and quaternary mammospheres.

To calculate the percentage of mammosphere forming efficiency (%MFE), the number of mammospheres was divided by the number of seeded cells and multiplied by hundred. Mammospheres were counted under 10x magnification of an inverted light microscope (Olympus, Tokyo, Japan). Twenty one (21) days after initial seeding in ultra-low attachment and serum-free media, a sufficient mammospheres’ population was established.

Cell viability

Cellular viability was evaluated using the trypan blue method. After de-attachment and good homogenization of cells until there was a single-cell suspension, a sample was taken and diluted in 1:1 ratio with trypan blue (Sigma-Aldrich, Gillingham, UK). Then 10 uL of the diluted mixture was loaded on a hemocytometer (Neubauer Double, Zuzi, Spain) and inspected under an inverted light microscope (Olympus, Tokyo, Japan). Cells with shiny white appearance were considered viable and counted because their cell membrane was impermeable to trypan blue, while blue-colored cells were disregarded. Hypoxic cells’ viability was checked upon every other split and compared to its normoxic counterparts to assess proliferation changes. While mammosphere cells were enzymatically disaggregated into single-cells in order to observe viability (Cassavaugh & Lounsbury, 2011).

CSCs sorting with magnetic beads

CD44⁺/CD24⁻ are the expression biomarkers of interest in MCF-7 derived mammospheres. At day seven of the initial culture of mammospheres in 6-well plate, CSCs enriched with CD44⁺/CD24⁻ were sorted from MCF-7 mammospheres using MagCellect CD24⁻/CD44⁺ Breast Cancer Stem Cell Isolation Kit (R&D System, Cat. MAGH111) according to their protocol. Initially, CD24⁺ cells were labeled and removed magnetically. Then from the CD24⁻ population, CD44⁺ cells were sorted magnetically using a biotinylated human antibody and streptavidin-conjugated magnetic beads as a positive selection model. The efficiency of sorting was assessed by staining recovered cells with fluorochrome-conjugated anti-human CD44⁺ and CD24⁻antibodies. Isolated CD44⁺/CD24⁻ cells were later identified via FACS analysis.

Exposure to hypoxia

To create hypoxic conditions, an anaerobic atmosphere generating system, AnaeroGen Compact (Oxoid, Basingstroke, UK) was used. The AnaeroGen system is used in microbiology area to create the low oxygen conditions needed for the growth of anaerobic bacteria. Here, we adopted the AnaeroGen system to induce hypoxia into CSCs. The system consists of a gas generatingsachets and tightly sealable bags. The sachet reacts promptly upon contact with air and consumes oxygen thus reducing oxygen concentration to less than 1% inside the bag. It has been confirmed in previous studies by using biochemical and electronic testing that this technique reduces oxygen content in plastic bag to below 1% within 30 min (Chenevier-Gobeaux et al., 2013; Mellor et al., 2005). The vented ultra-low attachment 75 cm² tissue culture flasks were placed inside the plastic bags and the sachets were placed inside then sealed. After 21 days of culturing the mammosphere CSCs, they were divided into two groups and exposed to hypoxic conditions intermittently (INTR.) or continuously (CONT). CSCs in the intermittent group were exposed to 10, 20, 30 or 40 hypoxic shots for 8 h three times a week, denoted as (INTR.10, INTR.20, INTR.30, and INTR.40) respectively, while CSCs in the continuous group was exposed to 5, 10, or 15 hypoxic shots for 72 h once per week, denoted as (CONT.5, CONT.10, and CONT.15). Apart from these optimized hypoxic shots, CSCs mammospheres were also incubated in as control, using the same culture media and conditions that were used for expansion of mammospheres detailed above.

Flow cytometry to examine enrichment of CSCs mammosphere with CD44⁺/CD24⁻ under Hypoxia/Normoxia

Flow cytometry was used to identify and characterize CD44⁺/CD24⁻ surface phenotype, which is in direct proportion to stemness, and was performed for parental MCF-7 cells, un-sorted mammospheres, sorted CSCs mammospheres three days after sorting, CSCs mammospheres which were 21 days-old all under normoxic conditions and finally for CSCs^{HY P} exposed to shots of INTR.20, INTR.40, CONT.5 and CONT.15. Mammospheres were dispersed to obtain single-cell populations as described above. The CSCs’ mammosphere pellet was washed in phosphate-buffer saline (PBS) (HyClone, USA) with 2% bovine serum albumin and stained with APC-anti-mouse CD24 and PE-anti-mouse CD44 antibodies (BD Pharmingen, CA). CSCs mammospheres were incubated in ice for 30 min, washed twice with PBS and then fixed in PBS containing paraformaldehyde. The flow cytometric analysis was performed on a BD FACSCalibur system (BD Biosciences, San Jose, CA, USA), and the identification was performed with the BD Cellquest software (BD Biosciences).

Cytotoxicity evaluation using MTT assay

To examine the chemoresistance of MCF-7 parental cells and their derived CSCs, both CSCs^NOR and CSCs^{HY P}, drug resistance towards doxorubicin was used as an indicator. A CellTiter Cell Proliferation Assay Kit^® (Promega, Madison, WI, USA) was used. This assay includes 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) which is a yellow tetrazole reduced by viable cells to purple formazan. Consequently, surviving cells after being exposed to multiple concentrations of doxorubicin can be quantified by spectrophotometry. The MTT proliferation assay was performed on MCF-7 parental cells seeded as a monolayer in an adherent 96-well plate. Moreover, CSCs^{HY P} were dissociated into single-cells, filtered, counted then seeded at a 7 × 10³ cells/well density in a 96-well plate and incubated for 24 h before treatment. Each concentration was tested in triplicate on CSCs exposed to (INTR.10, INTR.20, INTR.30, and INTR.40) or exposed to (CONT.5, CONT.10, and CONT.15). A control experiment was done for sorted and identified CSCs^NOR from mammospheres cultured under normoxic conditions.

Incremental serial dilutions of doxorubicin concentrations from 0.01 µM to 200 µM were used. After 72 h of incubation, old media was aspirated, and 100 µl of fresh media was added to each well. Then, 15 µl of MTT reagent was added to each well and incubated for 4 h. Solubilization/stop solution was added to each well and incubated for an hour. Absorbancy was read using an Elisa reader (Sunrise basic sciences, Austria) at 570 nm wavelength. The half maximal inhibitory concentration (IC₅₀) was calculated using GraphPad PRISM^®7.04 software (GraphPad Software, Inc.).

Endothelial cells isolation from umbilical cords

Human umbilical cords were obtained from two delivering females at the Jordan University Hospital after getting the approval of their Institutional Review Board (IRB), decision number 98/2018 and approval number (67/2018/481) and both patients having signed an informed consent before sample collection.

Enzyme digestion technique was used to isolate endothelial cells from the human umbilical cord. Debris on the outer surface of the cord was wiped using antiseptic solution. Any clots formed inside the vein were removed by slight squeezing. The umbilical vein was washed with RPMI medium (HyClone, Logan, UT, USA) containing 1% penicillin/streptomycin (HyClone, Logan, UT, USA) to remove excess blood and debris using 20 mL syringe inserted into the cord. Then, 0.1% of type I collagenase (Sigma-Aldrich, St. Louis, MO, USA) was put inside the vein where it was incubated for 20 min at 37 °C for digestion. Hereafter, the vein was washed with 20 mL RPMI media (HyClone, Logan, UT, USA) to stop collagenase proteolytic activity and was centrifuged at 1,500 rpm for 5 min. The resultant pellet was resuspended in 5 ml endothelial cell growth medium EGM-2 media (Lonza, Walkersville, MD, USA) which consists of the endothelial cell basal media EBM-2 with 10% FBS and other additives (including VEGF, bFGF- Fibroblast Growth Factor, EGF-Epidermal Growth Factor, and IGF-Insulin Growth Factor). HUVECs were expanded on 0.2% gelatin-coated T25 tissue culture flask and incubated overnight. Next day media was renewed to get rid of cells debris. Only passages earlier than six of HUVECs were used in all experiments.

Collection of conditioned-medium from hypoxic/normoxic CSCs

In order to obtain the CdM, CSCs were seeded as single cells at a density of 10,000 cells/cm² in an ultra-low adherence 75 cm² flasks with serum-free DMEM/F-12 and with supplements (mammospheres culture media detailed above). CSCs mammospheres were starved overnight by washing them twice with serum and supplement free DMEM/F-12 media, then centrifuged at 300 rpm for 10 min. Finally, the media was renewed without serum and without supplements for 8 h for intermittent and 72 h for the continuous hypoxic condition. The resultant CdM from where CSCs grown under hypoxia or normoxia were collected, centrifuged at 300 rpm for 10 min to remove cellular components and filtered through 0.2 µm pore-size filters (Millipore, Billerica, MA, USA). CdM from flasks under normoxic conditions were used as a control. Aliquots of the CdM were stored at −80 °C before being used. CdM were collected from eight sources, from CSCs^NOR, and from CSCs^{HY P} grown under the hypoxic shots (INTR.10, INTR.20, INTR.30, INTR.40, CONT.5, CONT.10, or CONT.15) and were utilized in successive experiments as discussed subsequently.

Capillary-like tube structure formation assay using CSCs’ mammospheres (CdM) on HUVECs

The effect of the factors secreted in the CdM obtained from CSCs^{HY P} and from CSCs^NOR on the angiogenic ability of HUVECs was examined through the formation of capillary tube-like structures by HUVECs, using the in vitro matrigel angiogenesis assay to mimic the microenvironment of CSCs. To eliminate the impact of growth factors in Matrigel, reduced growth factor Matrigel (BD Biosciences, San Jose, CA, USA) was used. HUVECs were incubated for 6 h in endothelial basal medium (EBM-2, Lonza) without serum and supplements for starvation. Briefly, the Matrigel stored at −20 °C was thawed in ice to prevent premature polymerization. Aliquots of 50 µl were plated into each well of pre-chilled 96-well culture plates and were left to polymerize at 37 °C for 2 h. HUVECs were removed from confluent cultures by treatment with Trypsin 0.05% (HyClone, Logan, UT, USA). HUVECs were collected using serum-containing media to stop trypsin effect, then counted, centrifuged and washed with phosphate buffer saline, then resuspended with CdM obtained from the different hypoxic episodes, including intermittent shots (INTR.10, INTR.20, INTR.30, INTR.40) and continuous shots (CONT.5, CONT.10, CONT.15) of hypoxia and from its normoxic counterparts as a control in triplicates. HUVECs were seeded at a density of 2 × 10⁴ cells/well. For quantification of tube formation complex, three primary variables were used to determine the magnitude of tube formation, average of total length of the branched tube, the number of loops and covered area. These variables were measured by Wimasis Image Analysis software.

VEGF quantification by enzyme-linked immunosorbent assay

The CdM that was collected as described above, was used for the determination of VEGF levels secreted by CSC^{HY P} and CSC^NOR using the Enzyme-Linked Immunosorbent kit. Human VEGF quantitative ELISA assay (R&D Systems, Minneapolis, MN, USA) was performed according to manufacturer’s instructions. A quantity of 200ul of the CdM from different CSCs^{HY P} treatments was tested.

Cell migration assay of HUVECs

HUVECs ability to migrate under the treatment of the CdM was investigated using a Transwell plate of 8 um pore-size chamber (Corning, Corning, NY, USA). HUVECs were starved overnight using EBM-2 media serum and supplement free for 6 h, then seeded in the upper chamber (insert) of the 24 well plate at density of 5 × 10⁴ cells in 500 uL of EBM -2 media, each placed into lower wells containing 750 ul of CdM obtained from (INTR.10, INTR.20, INTR.30, INTR.40) intermittent hypoxic shots in addition to continuous hypoxic shots (CONT.5, CONT.10, CONT.15). The same procedure was applied to the control by using serum-free EBM-2 media in the lower well instead of CdM. Then, cells were allowed to migrate for 8 h, then 70% cold ethanol was used for 10 min to fix the chambers. Membranes were stained with 0.5% (w/v) crystal violet (Sigma-Aldrich, St. Louis, MO, USA) for 30 min and afterward washed carefully with water. Cotton swabs were used to remove the cells that were not able to migrate to lower wells. Solubilizing bound crystal violet in methanol 100% (wt./vol) was used to quantify the migrated cells. Cells were inspected under an inverted light microscope (Olympus, Tokyo, Japan) and photographed at 4×  magnification. Each sample was done in triplicate and three fields were measured from each to calculate the migration.

Direct and indirect co-culture of HUVECs and CONT.5 CSCs mammospheres

A co-culture system has been used to better understand how CSCs ^{HY P} regulate their microenvironment and interact with HUVECs as a vascularization model for angiogenesis and migration.

Wound healing assay for HUVECs

In a 6-well plate, HUVECs at a density of 12 × 10⁴ were seeded till confluency. Then they were incubated for 6 h in EBM-2 media without serum to prevent any external factors from contributing to wound closure. After 6 h, using a 200uL micropipette tip, a straight wound-like scratch was made into the cells. Then cells were treated with CSCs^{HY P} CdM obtained from INTR.10, INTR.20, INTR.30, INTR.40, CONT.5, CONT.10 and CONT.15 hypoxic shots in which CSCs^{HY P} were grown, in addition to the control which was treated with CdM obtained from CSCs^NOR. Cells were carefully washed using phosphate buffer saline (HyClone, Logan, UT, USA), before and after making the scratch. Wound healing was inspected under an inverted light microscope (Olympus, Tokyo, Japan) and photographed after 48 h and at the beginning of the experiment. The surface area of the wound could not be calculated by image analysis software because control cells started to die after 16 h.

Wound healing assay for Hypoxic CSCs

In a 6-well plate coated with 1% gelatin, CSCs were seeded at a density of 8 × 10⁴ and incubated overnight, using DMEM/F-12 media without serum to prevent any external factors from contributing to wound closure. The subpopulation of CSCs used was originally cultured under intermittent or continuous hypoxic conditions; INTR.10, INTR.20, INTR.30, INTR.40, CONT.5, CONT.10, and CONT.15 in addition to the control CSCs^NOR originally grown under normoxia. They were all used upon confluency before being seeded in a gelatin-coated 6-well plate to allow CSCs adherence as a 2D monolayer for wound healing assay. Using a 200 uL micropipette tip, a straight wound-like scratch was made. Cells were washed using phosphate buffer saline, before and after making the scratch, and then relevant media was used; DMEM/F-12 without serum. Wound healing was photographed at the beginning of the experiment and after 36 h. The surface area of the wound was measured by image analysis (ImageJ 1.4.3.67 Launcher Symmetry Software).

Statistical analysis

Results were presented as a mean ± SD. Experiments were performed in triplicate with comparable results unless indicated otherwise. An ANOVA test was used to compare the differences between replicates which were considered significant at p ≤ 0.05. GraphPad Prism (version 7.04; San Diego, CA, USA) was used for statistical analysis.

(3D) Mammospheres Derived from MCF-7 cells

Mammosphere culture has been widely used to enrich CSCs. Thereof, MCF-7 mammospheres were cultured in ultra-low adherence plates and flasks, in serum-free media with supplements and successfully formed rigid and compact (3D) structures within 3–7 days (Fig. 2). The mammosphere multicellular spheroidal clumps became larger in size at day 7 compared to earlier days of growth as shown in (Fig. 2) with 30 um scale bar comparison. Moreover, as a quantitative measure, at day seven the mammosphere formation efficiency (MFE) of MCF-7 cells was 0.012% (±0.002%) only and after three weeks the MFE had increased to 5.30% (±0.307%).

Sorting of CSCs From 3D mammospheres using magnetic beads

The magnetic beads separation protocol uses antibodies conjugated to a magnetic bead. CSCs were successfully sorted and isolated using the magnetic beads technique, but their number became sufficient only 3 days after sorting. Thereafter, sorted CSCs were identified via FACS analysis and were used for hypoxic-condition exposure. The expression of CD44⁺/CD24⁻ mammospheres is shown in (Table 1) and their morphology is shown in (Fig. 2).

Enrichment of sorted CSCs mammospheres under Hypoxic/Normoxic conditions

The morphology of CSCs^{HY P} and the control in normoxia CSCs^NOR is shown in (Fig. 3). It is observed that CSCs^{HY P} that underwent CONT.5 and INTR.20 hypoxic shots were denser and larger in number than CSCs^NOR. CSCs^{HY P} that underwent CONT.15 and INTR.40, that is the maximum and final hypoxic shots showed the least growth compared to other hypoxic exposures and CSCs^NOR. The MFE could not be measured for CSCs^{HY P} because long-term hypoxic exposure (INTR.40 and CONT.15) made mammospheres aggregate into clumps, thus, the single mammosphere could not be distinguished and compared to control and to other hypoxic mammospheres.

Identification of CD44⁺/CD24⁻ phenotype content by flow cytometry

Flow cytometry conducted for CSCs after 3 days of sorting showed a percentage of CD44⁺/CD24⁻ expression equal to 81.0 (±7.5%). Three weeks (21 days) after growing in ultra-low attachment flasks, the percentage decreased to approximately 35.5 (±4.5%). In contrast, the parental MCF-7 cells had only 1 (±0.1%) expression and the unsorted mammospheres had 33.2 (±3%) expression. After exposure to hypoxic conditions, flow cytometry results based on CD44⁺/CD24⁻ markers showed that the highest expression occurred in CSCs^{HY P} exposed to INTR.20 shots, 39.8 (±5.2%), and in CSCs^{HY P} exposed to CONT.5, 51.6 (± 6.1%). On the other hand, the lowest percentages were observed in CSCs^{HY P} that underwent or were exposed to INTR.40 shots, 0.3 (±0.01%), and in CSCs^{HY P} exposed to CONT.15, 0.5 (±0.05%). In conclusion, isolated and sorted CD44⁺/CD24⁻ breast cancer cells only temporarily preserve this phenotype and ultimately revert to an equilibrium state in which the expanded sub-population regains the original cell surface profile of the parental cell line, which was deduced because the unsorted mammospheres and the 3-week old sorted CSCs both have similar CD44⁺/CD24⁻ content (33.2% and 35.5%, respectively). The CD44⁺/CD24⁻ percentage in unsorted mammospheres that were grown in ultra-low attachment plates in serum-free media for 7 days before sorting was 33.2 (±3%), which is much higher than the percentage of parental MCF-7 cells 1.0 (±0.1%). The expression of CD44⁺/CD24⁻content for all subpopulations is summarized in (Table 1) and demonstrated in (Fig. 4).

Hypoxic CSCs show higher drug resistance to conventional chemotherapies

To examine whether self-renewing CSCs mammospheres have higher chemoresistance ability, and to assess if hypoxic treatment of CSCs would increase their stemness character and chemoresistance, MTT assay was implemented to test the sensitivity of parental MCF-7 versus the sorted CSCs mammosphere towards doxorubicin. In addition, we assessed the resistance towards doxorubicin acquired by CSCs^{HY P} which underwent intermittent hypoxic shots (INTR.10, INTR.20, INTR.30 INTR.40), and CSCs^{HY P} that underwent continuous hypoxic shots (CONT.5, CONT.10, CONT.15). Then we compared the inhibitory drug concentration (IC₅₀) results for the CSCs^{HY P} versus CSCs^NOR mammospheres (as control) and compared the IC₅₀ value of CSCs^{HY P} versus parental MCF-7 cells, described in Table 2. Overall, the drug resistance of CSCs^NOR was found to be higher (IC₅₀ = 1.90 uM) when compared to the parental MCF-7 cells (IC₅₀ = 0.46 uM) by 4.08 times. Comparing the IC₅₀ values of CSCs^NOR mammospheres with CSCs^{HY P}, we observed significantly increased values under hypoxic conditions at INTR.20 (IC₅₀ = 6.13 uM) and CONT.5 (IC₅₀ = 7.12 uM) with (3.28) and (3.76) fold increases respectively compared to IC₅₀ of CSCs^NOR mammospheres. These increaseswere much higher upon comparing IC₅₀ for CSCs^{HY P} with parental MCF-7. The lowest IC₅₀ values were observed at INTR.40 and CONT.15 which represent the final hypoxic treatments. Taken together, these results support the likelihood that enriching the mammospheres’ culture with CSCs contributed to the higher drug resistance toward doxorubicin when compared to the parental MCF-7.

Hypoxic conditioned medium (CdM) induces HUVECs to form capillary-like tube structures in matrigel

In this study, HUVECs were treated with CdM in which CSCs were expanded under several hypoxic and normoxic conditions, as described in the method section. Our results showed that treating HUVECs with CdM obtained from CSCs^{HY P} resulted in an increased construction of capillary-like tube structures that reached its peak at 8 h of treatment and disappeared eventually in a time-dependent manner. However, HUVECs treated with CdM obtained from CSCs^NOR (control), resulted in a reduced formation of complex tubular structures compared to the ones treated with the CdM obtained from CSCs^{HY P} (Fig. 5). The capillary-like tube structures were evaluated by quantitative analysis using Wimasis Image Analysis. The results showed that the highest increase of branched tube lengths, the number of loops and the covered area was with HUVECs that were cultured with CdM obtained from CSCs^{HY P} culture that underwent INTR.20 and CONT.5 shots at each time point (as demonstrated in Figs. 5 and 6). A bar graph (data combined from three independent trials) (Fig. 6) shows that “the total length of the tubes” underwent a remarkable increase upon the treatment with CdM from INTR.20 and CONT.5, by almost 2.4 fold in both INTR.20 and CONT.5 when compared to the control, moreover, the “covered area” significantly increased when HUVECs were cultured with CdM from CSCs^{HY P} culture that underwent CONT.5 and INTR.20, by 3.5 and 2.7 times respectively. The total number of loops were significantly increased by 6.7 fold in INTR.20 and 8.9 fold in CONT.5 compared to control HUVECs incubated with CdM obtained from the CSCs^NORculture expanded in normoxic state. As shown in Figs. 5 and 6, CdM induced a significant increase in tube formation in HUVECs starting from the treatment with INTR.10 CdM and reaching a maximum at INTR.20 (Fig. 5F), whereas CONT.5 (Fig. 5B) caused the highest increase in tube formation.

Detection of secreted vegf in conditioned-medium (CdM) by ELISA

In order to identify the possible mediators of angiogenesis in breast cancer, a quantitative assay was performed for the main angiogenic factor, VEGF-A. As the CdM obtained from the hypoxic episodes INTR.20, INTR.30 and CONT.5 had the highest VEGF concentrations, INTR.20, 1159 (±90 pg/ml), INTR.30, 1013 (±32 pg/ml) and CONT.5, 934 (±28 pg/ml) are compared to control normoxic CdM 225 (±24 pg/ml) as shown in (Fig. 7).

Effect of conditioned-medium (CdM) on migration of HUVECs

The role of the CdM in promoting migration of HUVECs was investigated using migration assay to confirm whether secretions of CSCs^{HY P} enhance angiogenesis in the tumor microenvironment. HUVECs were tested for their ability to migrate toward CdM from different hypoxic episodes compared to their migration toward endothelial basal medium (EBM-2) serum and supplement free, as a control, using transwell migration assay. The Transwell migration assay is used to create a chemical gradient by putting the CdM in the lower chamber. Comparing the migration of HUVECs from upper chamber across the membrane upon using the EBM-2 without serum and without supplements (control) group versus using the CdM obtained from hypoxic shots, it was found that migration was much higher under the use of the hypoxic CdM (Fig. 8). The assay results showed that there were more purple-stained cells migrating toward the CdM obtained from the CONT.5 shot than the other two episodes (CONT.10 and CONT.15). The CdM obtained from the INTR.20 episode had also more purple-stained cells migrating towards it, compared to the other three episodes (INTR.10, INTR.30, and INTR.40). Interestingly, a significant difference found was that the migrated-cell count in HUVECs towards CdM from CONT.5 vs control was 45,000 (±2.1) vs. 4,500 (±1.5), while the migrated cell count in INTR.20 vs control was 38,000 (±2.6) vs. 4,500 (±1.5). These findings suggest that the secretions presented in the CdM which were obtained from CONT.5 and INTR.20 episodes significantly promoted the migration of HUVECs. The data also showed that the migration capacity of HUVECs increased significantly (10-fold) when cultured using the CdM from CONT.5 shot and (8.4-fold) when cultured using the CdM from INTR.20 shot compared with HUVECs cultured with control (Fig. 8).

Stimulation of HUVECs growth & morphological alterations in co-culture with CSCs^{HY P}

To clarify the interaction between endothelial cells and CSCs^{HY P} in the tumor microenvironment, HUVECs were examined in direct and indirect co-cultures with the CSCs^{HY P} subpopulation that underwent CONT.5 episode. The control of HUVECs was cultured in serum and supplement free media and possessed a ‘teardrop-like’ morphology. In the direct co-culture model, the endothelial cell morphology changed and acquired elongated threadlike shape as a response to the influence of CSCs^{HY P} secretions though the growth media was free from serum and any growth factors as shown in (Figs. 9A–9C).

We also examined the growth of HUVECs indirectly co-cultured with CSCs^{HY P} using a 0.4 uM permeable membrane that was placed between HUVECs and CSCs^{HY P} subpopulation. Indirect co-culture of the subpopulation of CONT.5 CSCs^{HY P} with HUVECs induced sequential morphological changes. As a response to the influence of secretions from CSCs^{HY P} thattrans-passed through the membrane, the morphology of HUVECs began to change at 72 h, forming net-like structures resembling a vascular network, though the growth media was free from serum and any growth factors (Figs. 9D & 9E). Control HUVECs proliferation reached a plateau in 24 h after which cells started to die.

In vitro wound-healing assay of HUVECs

This assay was done to assess how secretions from CSCs^{HY P} in the tumor microenvironment would affect the ability of HUVECs to migrate in response to chemoattractants in the CdM obtained from different hypoxic shots. This model mimics the tumor microenvironment where angiogenic ability of endothelial cells is affected by secretions from CSCs^{HY P}. In the control experiment, HUVECs were treated with CdM obtained from CSCs^NOR. In the presence of the CdM obtained from hypoxic shots, HUVECs were able to migrate and the wound closure was enhanced after 48 h. The wound closure was significantly enhanced by CdM obtained from INTR.20 (Fig. 10B) and CONT.5 (Fig. 10C) shots. Whereas in the case of the control, HUVECs that were treated with CdM media obtained from CSCs^NOR, showed a weaker capacity to migrate. The wound surface area could not be calculated in order to assess wound closure because the HUVECs treated with CdM obtained from CSCs^NOR and from CSCs^{HY P} after long-term exposure (INTR.40 and CONT.15) started to die after 16 h of treatment which prevented further quantitative measurements.

In vitro wound-healing assay of hypoxic CSCs

Due to the significant role that CSCs play in tumor recurrence and metastasis, a wound-healing assay was performed. CSCs were seeded as a (2D) monolayer, scratched, and then cell migration was evaluated. The seeded CSCs were originally expanded and grown in hypoxic conditions, INTR.10, INTR.20, INTR.30, INTR.40, CONT.5, CONT.10, and CONT.15 in addition to the control CSCs in normoxia. The scratched CSCs^{HY P} and CSCs^NOR cultured in DMEM/F-12 without serum were assessed for wound closure ability which was determined by calculating the ratio of the wound surface area at the endpoint (36 h) to its surface area at the starting point of the experiment (zero hour). Images were analyzed by ImageJ software. Subpopulations of CSCs^{HY P} showed greater ability for wound closure compared to CSCs^NOR (Figs. 11 and 12). CSCs^{HY P} from INTR.20 and from CONT.5 manifested the greatest wound closure (Fig. 11). At 36 h, the wound closure of the control was only 40%, while it was 98% and 95% for INTR.20 (Fig. 12C) and CONT.5 (Fig. 12F), respectively.