Hypoxia-induced cancer stemness acquisition is associated with CXCR4 activation by its aberrant promoter demethylation

A normal human lung cell line (BEAS-2B) and four lung cancer cell lines (A549, H292, H226, and H460) were obtained from the American Type Culture Collection (ATCC, Manassas, VA). The BEAS-2B cells were maintained in a DMEM/F12 medium; whereas the A549, H292, H226, and H460 cells were cultured in a RPMI1640 medium. The cultured medium was supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, 100 μg/mL streptomycin, and 250 ng/mL amphotericin B at 37 °C under 5% CO₂ conditions. To investigate the effect of hypoxia, the cells were transferred to the Thermo Forma Anaerobic System Model 1029 (i.e., less than 1% of O₂ partial pressure) (Thermo Electron Corp., Waltham, MA).

CXCR4 siRNA (Santa Cruz Biotechnology, Santa Cruz, CA) was used to reduce the CXCR4 expression. After being cultured in a hypoxic condition for 24 h, the cells were transfected with 300 nM CXCR4 siRNA or a control siRNA using a transfection reagent (Santa Cruz Biotechnology, Santa Cruz, CA) for 6 h. The cells were harvested at 72 h post-transfection and used for the Matrigel invasion and soft agar colony formation assays.

The transcriptomes of the BEAS-2B and A549 cells that were cultured under normoxic or hypoxic conditions for 24 h were analyzed. The illumina-compatible libraries of mRNA were constructed using the TruSeq Stranded mRNA Library Preparation Kit (Illumina Inc., San Diego, CA) according to the manufacturer’s instructions. Samples were sequenced using the HiSeq 2000 Sequencing System (Illumina Inc., San Diego, CA). To estimate the expression levels, the RNA-Seq reads were mapped to the human genome using TopHat (version 1.3.3). The reference genome sequence (hg19, Genome Reference Consortium GRCh37) and annotation data were downloaded from the University of California Santa Cruz (UCSC) website (https://​genome.​ucsc.​edu/). The transcript counts at the gene level were calculated, and the relative transcript abundances were measured in fragments per kilobase of exon per million fragments mapped (FPKM) using Cufflinks software (version 1.2.1).

Incubated cells were lysed in radioimmunoprecipitation assay (RIPA) buffer with protease inhibitors for 20 min. The protein samples (20 μg) were separated on a discontinuous PAGE gel and transferred to a nitrocellulose membrane at 70 V for 2 h. The membrane was then incubated with the E-cadherin, N-cadherin, fibronectin, vimentin, and α-SMA antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) at 4 °C overnight. To investigate the time-dependent manner of CXCR4 expression, cells were harvested at 2, 6, 24, and 48 h after incubation. After the protein extraction, the membranes with 20 μg of protein were incubated with the CXCR4 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) at 4 °C overnight. The target protein was detected using an ECL Kit (Amersham Pharmacia Biotech, Little Chalfont, Buckinghamshire, UK) and X-ray film.

The total RNA was extracted and converted into cDNA using HyperScript RT Mastermix (GeneAll Biotechnology, Seoul, Korea). The cDNA was amplified in a MyCycler Thermal Cycler (Bio-Rad, Hercules, CA) using CXCR4 forward primer: TTCTACCCCAATGACTTGTG and reverse primer: ATGTAGTAAGGCAGCCAACA. The cycling conditions involved 35 cycles for 30 s at 94 °C, 60 °C, and 72 °C. The amplified PCR products were visualized on a 1% agarose gel stained with ethidium bromide.

The Matrigel invasion assay was performed using the Corning Matrigel Matrix (Corning Inc., Tewksbury, MA). After incubation under hypoxic conditions for 24 h, 2.5 × 10⁴ cells were plated onto the Matrigel-coated membrane in Falcon Cell Culture Inserts for 24-well plates (pore size, 8 μm; Corning Inc., Corning, NY) with a 500 μL serum-free medium. Inserts were placed in 24-well culture plates containing 750 μL medium including 5% FBS. After a 24-h incubation, cells that had migrated through the Matrigel were stained with Diff-Quik solution. The cells were captured by a Leica SCN400 Slide Scanner (Leica Microsystems, Buffalo Grove, IL) and the number of invaded cells was counted in five random fields using ImageJ software. The assay was performed in triplicate.

Lung cell lines (1 × 10⁵ cells per well) were seeded in six-well plates and incubated in a hypoxic chamber for 24 h. After cells were detached with trypsin-EDTA, 1 × 10³ cells per well were mixed with 0.7% agarose in 10% FBS and plated on top of 0.8% agarose in a six-well plate. After 3 weeks, colonies were washed with a phosphate-buffered saline (PBS) and stained for 2 h with 0.005% crystal violet. The images were captured using a digital camera and inverted light microscopy.

The lung cell lines were detached to form a single cell suspension and were seeded at densities of 5 × 10³ cells per well in serum-free DMEM/F12 medium supplemented with 1 x B27 (Gibco, Grand Island, NY), 20 ng/mL epidermal growth factor (EGF; PeproTech Inc., Rocky Hill, NJ), and 20 ng/mL fibroblast growth factor (FGF; PeproTech Inc., Rocky Hill, NJ) into Corning Ultra-Low Attachment Products (six-well; Corning Inc., Corning, NY). Experimental groups were incubated under hypoxic conditions for 24 h; whereas, the control groups were maintained under normoxic conditions. All cell lines were then incubated under normoxic conditions for 5 days.

Animal experiments were performed according to guidelines approved by the Institutional Animal Care and Use Committee of the Catholic University Medical School (Approval no. CUMC-2015-0080-01). We used a xenograft tumor model for stemness investigation based on methods described in previous studies which reported the stemness characteristics using cancer cells from animal subcutaneous implantation models [9‐11]. Male BALB/c nude mice (Charles River Laboratories, Yokohama, Japan) were used. After 24 h of incubation of the BEAS-2B, A549, and H226 cells in a normoxic or hypoxic chamber, 1 × 10⁵, 1 × 10⁶, or 5 × 10⁶ cells were injected subcutaneously into each side of the mouse flank: the cells incubated under normoxic conditions were implanted on the left, and the cells incubated under hypoxic conditions were implanted on the right. Tumor development and growth were monitored three times per week. Tumor sizes were measured and calculated according to the formula (L x W x H)/2, where L is the maximal length of the tumor, W is the maximal width perpendicular to L, and H is the maximal height [12]. The mice were sacrificed by CO₂ asphyxiation to obtain tumor tissues after 2 weeks of injections.

For CXCR4 immunohistochemistry, the paraffin sections were immunostained with an anti-CXCR4 antibody (3 μg/ml dilutions; Abcam, Cambridge, UK; #ab58176) [13]. The tissue slides were stained with H&E for comparison with the immunostained tissues, and they were reviewed by three pathologists (TJ Kim, SB Lee, and SH Lee). A semi-quantitative scoring system was used that was based on the summation of scores for staining intensity (0, negative; 1, weak; 2, moderate; 3, strong) and the percentage of positive cells (0, < 10% of tumor cells; 1, 10–25% of tumor cells; 2, 26–50% of tumor cells; 3, > 51% tumor cells).

Immunofluorescence was performed to analyze the representative expression patterns of the epithelial cell phenotype (E-cadherin), the mesenchymal cell phenotype (α-SMA), and a stem cell marker (CXCR4). These cells were triple-stained with E-cadherin, α-SMA, and CXCR4. The stained slides were visualized with a LSM 510 META Laser Scanning Microscope (Carl Zeiss, Jena, Germany) and analyzed with the ZEN 2009 Light Edition software (Carl Zeiss, Jena, Germany).

To determine whether DNA methylation influences gene expression, cells were incubated with 5 μM 5-azacytidine (AZA; DNA methyltransferase inhibitor; Sigma, St. Louis, MO) for 24 h [14].

One microgram of the RNA extracted from cell lines was used to synthesize cDNA using HyperScript™ RT Master mix (GeneAll Biotechnology, Seoul, Korea). The expression of CXCR4 was then quantified using GoTaq® qPCR Master Mix (Promega, Madison, WI) and above CXCR4 primers on Exicycler™ 96 (Bioneer, Daejeon, Korea) with an initial denaturation step of 95 °C for 5 min, followed by 40 cycles for 30 s at 95 °C, 60 °C, and 72 °C. The expression of CXCR4 was normalized to that of GAPDH.

The methylation-specific PCR was performed to analyze CXCR4 CpG sites in bisulfite-modified DNA [14]. The following primer pairs were used: forward, GGACGTAGTTTTTCGGTTTC; and reverse, AACGCGCAAACAATACTTT. The methylation level was normalized with the β-actin gene. A region of β-actin devoid of any CpG dinucleotide was amplified using the following primers: forward, TGGTGATGGAGGAGGTTTAGTAAGT; and reverse, AACCAATAAAACCTACTCCTCCCTTAA. The bisulfite sequencing was preformed using the following primers: forward, AAGATTGGTAGGTGTAAGTG; reverse, TTTCATTTCCTCACTCTCCC.

To examine the effect of hypoxia on the mRNA expression in the BEAS-2B and A549 cells, a transcriptome analysis was performed using next-generation sequencing. Distinct differences in mRNA expression patterns were observed between the cells that were cultured under normoxic and hypoxic conditions (Fig. 1a). To examine the effect of hypoxia on the EMT, various EMT markers were analyzed. Mesenchymal markers (fibronectin, vimentin, α-SMA, slug, snail, and ZEB1) increased more than 2-fold; whereas, the expression of the epithelial marker E-cadherin was reduced 1.2- to 2.3-fold in cells exposed to the hypoxic conditions (Fig. 1b). Among the cancer stem cell candidates, the fold change in the CXCR4 expression was the highest following hypoxia treatment (BEAS-2B 11.88424 and A549 6.338601) (Fig. 1c). The fold changes of the various EMT and stem cell markers are provided in Table 1.

Consistent with the transcriptome analysis, the E-cadherin expression in four lung cell lines (BEAS-2B, A549, H292, and H226) decreased according to the length of time that the cells were exposed to hypoxia. The expression of fibronectin, vimentin, and α-SMA increased; although, the expression levels differed according to the length of exposure to hypoxia (Fig. 2a).

The immunofluorescence analysis revealed that the expression of the epithelial cell marker E-cadherin was lost following hypoxic stimuli; although, the expression of the mesenchymal cell marker α-SMA and stem cell marker CXCR4 was apparent following exposure to hypoxia (Fig. 2b). These results indicate that hypoxic conditions result in a gain of the EMT and stem cell phenotypes. Furthermore, cells incubated in the hypoxic chamber displayed spindle-shaped morphological changes that are consistent with the development of the EMT process.

The time-dependent mRNA and protein expressions of CXCR4 are provided in Fig. 2c. Compared with the normoxic condition, the cells exposed to the hypoxic condition exhibited increased CXCR4 mRNA and protein expressions. The mRNA expressions of CXCR4 in each of the cell lines increased as early as 2 h; although, the protein expressions were definite in 24 or 48 h according to the cell lines.

To investigate whether hypoxia enhances invasion, Matrigel invasion assay was performed. The Matrigel invasion assay demonstrated that hypoxia significantly enhanced transwell invasion in lung cells. Moreover, compared with the hypoxia (+) group, the CXCR4 siRNA transfected cells exhibited a significant decrease in migrated cells (Fig. 3a and b).

The soft agar colony formation assay has been widely used to assess stem cell growth. To compare colony formation ability, cells were seeded at 1 × 10³ cells per well. The colony formation rates of all lung cell lines incubated under hypoxic conditions were higher than those incubated under normoxic conditions. The cells treated with CXCR4 siRNA exhibited decreased numbers and diameters in colony formation compared to the cells exposed only to hypoxia (Fig. 4a). Moreover, the lung cell lines exposed to hypoxia displayed increased sphere formation ability (Fig. 4b).

The tumor-forming potential of the cells cultured under normoxic or hypoxic conditions was compared following an injection into BALB/c nude mice. The ability of the cells to form a tumor was evaluated as the rate of tumor development and size (Fig. 5a). In BEAS-2B or H226 cells injected mice, tumor formation and size were higher in the hypoxic cell groups compared with the normoxic groups (P < 0.01). A549 injected mice showed a tendency to increase although not significant.

To investigate the CXCR4 expression in mouse tumors, immunohistochemistry was performed in tumors injected with the BEAS-2B, A549, and H226 cells. Compared with the tumor tissues injected with normoxic cells, tumor tissues injected with cells exposed to hypoxia displayed strong immunoreactivity of the CXCR4 expression compared with those cultured under the normoxic conditions (Fig. 5b).

To identify whether CXCR4 expression was activated aberrantly, we investigated the expression of CXCR4 after treatment with a DNA methyltransferase inhibitor (AZA). The AZA treated cell lines exhibited significantly increased CXCR4 expression compared with the hypoxia (−) control (n = 4, P < 0.05; Fig. 6a). These results suggest that DNA demethylation was involved in the activation of CXCR4 expression.

The DNA from the BEAS-2B, A549, and H292 cell lines was converted by sodium bisulfite, and the methylation level at the CXCR4 promoter region was compared in normoxic and hypoxic conditions using a methylation-specific PCR. The methylation-specific real-time PCR revealed a lower methylation level at the CXCR4 promoter region in the hypoxic condition compared with the normoxic condition (n = 4, P < 0.05; Fig. 6b).

Bisulfite sequencing was performed using the BEAS-2B, A549, and H292 cell lines to demonstrate the CpG site methylation status of the CXCR4 promoter region. Representative bisulfite sequencing results from 11 CpG sites presented in underlined letters within an 82-bp promoter region of CXCR4 are provided. The CpG sites that were demethylated after hypoxic conditions were demonstrated in red letters (Fig. 6c). The heterozygote C/T double peaks (Y) were observed at multiple sites.