This article is mentioned in:

Introduction

Hypoxia is a reduction in the normal level of tissue oxygen tension and occurs in many disease processes, including cancer. Tumor hypoxia is typically associated with poor patient prognosis, partly because low oxygen levels reduce the effectiveness of radiation therapy, which kills tumor cells by generating reactive oxygen species (1). Hypoxia-inducible factor (HIF) is a transcription factor found in mammalian cells cultured under reduced oxygen tension and plays a key role in the cellular response to hypoxia. HIF regulates the transcription of several genes involved in biological processes, such as angiogenesis, cell proliferation and survival, glucose metabolism, pH regulation and apoptosis (2). Angiogenesis is a crucial regulator of tumor growth and metastases (3). Tumor angiogenesis is regulated by the production of angiogenic stimulators, including vascular endothelial growth factor (VEGF), which is a key regulatory factor in the prognosis of various cancers. Several studies have shown that transforming growth factor-β1 (TGF-β1) is involved in angiogenesis, leading to tumor progression (4). TGF-β1 signaling has been shown in concert with HIF-1α to regulate VEGFA expression (5).

Cancer stem cells (CSCs) typically represent a small fraction of tumor cells that have the ability to self-renew and differentiate into many more mature cancer cells (6). HIF stabilization in hypoxic tumor cells may promote the adoption of stem cell properties, including self-renewal and multipotency, by stimulating the expression or activity of Oct4, Notch, and other critical signaling pathways. Oct4 has been shown to function in a complex with Nanog and Sox2 to activate and repress genes controlling stem cell identity and differentiation (7). KLF4 is highly expressed in CSC-enriched populations in mouse primary mammary tumors and breast cancer cell lines (8). Indeed, CSCs have been shown to be more radioresistant than non-stem cancer cells and are, therefore, believed to be responsible for treatment failure and tumor recurrence (9).

Among females, breast cancer ranks first at age 20–59 years (10). In recent years, the encouraging trend towards earlier detection and the increasing use of systemic adjuvant treatment have improved survival rates; however, nearly half of the breast cancer patients treated for localized disease develop metastasis (11). A lower incidence of breast cancer is associated with the high consumption of phytoestrogens, which are biologically active plant-derived phenolic compounds that structurally mimic the mammalian estrogen 17β-estradiol (12). Basic and preclinical research have focused on resveratrol [RES (trans-3,4′,5-trihydroxystilbene)], a naturally occurring polyphenol enriched in grapes and red wine (13). Resveratrol has recently been shown to function as a cancer chemoprevention agent, an anti-mutagenic agent, and an anti-initiative agent (14).

This study demonstrated that the resveratrol analog HS-1793 inhibits angiogenesis via the regulation of hypoxia and displays hypoxia-induced CSC properties in FM3A mouse breast cancer cells.

Materials and methods

Preparation of HS-1793

To obtain HS-1793, the stilbene double bond present in resveratrol was substituted with a naphthalene ring, as previously described (15). A stock solution was generated in absolute ethanol at 50 mM, and the working dilutions were generated directly in culture media. The control vehicle was culture media containing amounts of ethanol equivalent to those present in HS-1793.

Cell culture conditions

FM3A (murine breast cancer cells) originated from the mammary gland of the C3H/He mouse were grown in RPMI-1640 (Gibco, Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS; Hyclone, Logan, UT, USA), 100 U/ml penicillin/streptomycin (Gibco), in a humidified atmosphere containing 5% CO2 and 37°C. Exponentially growing cells (~70–80% confluence) in complete medium were pretreated for 2 h with different concentrations of HS-1793, followed by exposure to normoxic (5% CO2 at 37°C) or hypoxic conditions (1% O2 at 37°C) in an Anaerobic Glove Box (Forma Scientific, Marietta, OH, USA) for treatment. Environmental hypoxic conditions (1% O2) were achieved in an airtight humidified chamber continuously flushed with a gas mixture containing 5% CO2 and 95% N2.

HS-1793 treatment and assessment of cell viability

HS-1793 was dissolved in EtOH and stored at −80°C until use. FM3A cells were plated onto 60-mm dishes with 1×105 cells for 24 h, and the cells were treated with HS-1793 (0, 1.3, 2.5, 5, 10 and 20 μM) under normoxic or hypoxic conditions. The cells were harvested at 12 or 24 h after treatment and counted using the ADAM automatic cell counter (Digital Bio., Korea). The relative cell count was determined as follows: final count of treated cultures/final count of control cultures × 100.

Hypoxia detection in tumor cells

To validate the efficiency of hypoxic treatment, a hypoxia detection kit (Hypoxyprobe-1; Chemicon, Temecula, CA, USA) was used to identify 2-nitro-imidazole adducts formed in cultured FM3A cells under hypoxia. FM3A cells were cultured on 24-well plates at a density of 1×105 cells/well for 24 h; then, the cells were treated with HS-1793 for 2 h, followed by incubation under normoxic or hypoxic conditions for 12 h. After treatment with 100 μM pimonidazole hydrochloride (1-[(2-hydroxo-3-piperidinyl) propyl]-2-nitroimidazole hydrochloride) under hypoxic conditions for 4 h, single-cell suspensions were prepared with trypsin and spun in clean, fat-free glass slides using a cytocentrifuge. Cytocentrifuged cells were fixed for 30 min in 4% paraformaldehyde (PFA). After washing with phosphate-buffered saline (PBS), the cells were permeabilized with PBS containing 0.5% Triton X-100 for 10 min. Next, the cells were incubated for 30 min at room temperature with FITC-conjugated antibody (1:50) against pimonidazole adducts. Hoechst 33342 was used as a nuclear counterstaining agent. The cells were imaged under Nikon Eclipse TS100 fluorescence microscopy (Nikon, Tokyo, Japan).

Western blot analysis

Protein samples were separated by 10% SDS-PAGE and transferred to a polyvinylidene difluoride membrane. The membrane was allowed to react with anti-HIF-1α (1:200; Novus, USA), anti-VEGF (1:500; Santa Cruz Biotechnology, CA, USA), anti-OCT4 (1:200; Santa Cruz Biotechnology), anti-KLF4 (1:500; Abcam, UK), anti-SOX (1:200; Santa Cruz) or anti-GAPDH (1:2,000; Santa Cruz Biotechnology). Immunostaining with antibodies was performed using the Super-Signal West Pico enhanced chemiluminescence substrate, and detection was performed using LAS-3000PLUS (Fuji Photo Film Co., Kanagawa, Japan).

Cell migration assay

The cells were grown to confluence in 6-well plates for 2 days, and a scrape in the form of a cross was created through the confluent monolayers with a plastic pipette tip. The cells were treated with HS-1793 for 2 h, followed by incubation under normoxic or hypoxic conditions for 24 h. Several wounded areas were marked for orientation, observed, and then imaged using a Nikon Eclipse TS100 microscope (Nikon) at 24 h after the scratch.

Annexin-V binding assay

FM3A cells were pretreated with 2.5 μM HS-1793 for 2 h, followed by incubation under hypoxic conditions. After 12 h, these cells were exposed to γ-IR (0, 2, 4 Gy) using a 137Cs source (BioBeam 8000, STS, Braunschweig, Germany) and cultured under hypoxic conditions for 12 h. The cells were harvested and then washed with PBS. The pellets were resuspended in 1X Annexin V binding buffer (BD Biosciences, Bedford, MA, USA) at a concentration of 1×106 cells/ml. Next, 100 μl of the solution was transferred to a 5-ml culture tube, and 5 μl of Annexin V-FITC and 5 μl 7-AAD were added. The cells were then gently vortexed and incubated for 15 min at RT (25°C) in the dark. Next, 400 μl of 1X binding buffer was added to each tube, and the samples were analyzed by flow cytometry (BD FACSAria, USA) within 1 h.

Animal studies

All experiments were performed on 6-week-old C3H/He female mice obtained from Central Lab. Animal Inc. (Seoul, Korea). The animals were raised under SPF conditions at the Korea Institute of Toxicology, Hospital of Dong-A University according to Good Laboratory Practices OECD guidelines. All animal procedures were performed according to approved protocols (approval no. DIACUC-09-24) from the Institutional Animal Care and Use Committee (IACUC) of Dong-A University and in accordance with recommendations for the proper use and care of laboratory animals. FM3A cells in logarithmic growth phase were used to establish a breast tumor model. Xenografts were generated by subcutaneously injecting 2×106 tumor cells in 50 μl of PBS into the flank adjacent to the right hind limb of 7-week-old female C3H/He mice. When the tumor grew to a size of ~500 mm3, HS-1793 (1.5 mg/kg) was administered as an intraperitoneal (i.p.) single injection for the determination of the hypoxic status and perfusion level in the tumor tissue. When the tumor grew to a size of ~40 mm3 (~10 days), the mice were stratified into groups of 8–10 animals with equal mean tumor volumes, and HS-1793 (0.1, 0.5 and 1.5 mg/kg) was administered by i.p. injection twice a week for 30 days. The tumor sizes were measured with calipers once weekly. The tumor volume was calculated as follows: (width)2 × length × 0.52. After 4 weeks, the mice were euthanized to obtain serum samples and tumor tissues.

ELISA

The serum VEGF concentrations from animals bearing xenografts were determined. The blood was collected from the abdominal vein of tumor-bearing or normal mice at 40 days after the inoculation of tumor cells. Serum samples were obtained by centrifugation at 3,000 rpm for 10 min and stored at −20°C until further analysis. The VEGF concentrations were determined using a quantitative enzyme-linked immunosorbent assay (ELISA, R&D Systems, Minneapolis, MN, USA). The amount of VEGF immunoreactivity was calculated using recombinant mouse VEGF standards present on each microtiter plate. Optical densities were determined at 550 nm using a microtiter plate spectrophotometer (Beckman Coulter detection platform, USA).

Real-time PCR quantification

Total cellular RNA was isolated from tumor tissue using TRIzol reagent (Invitrogen) and used for reverse transcription. Complementary DNA (cDNA) was synthesized from 1 μg of total RNA from each sample using a Maxime RT PreMix (iNtRON) according to the manufacturer’s instructions and used as a template for TaqMan™ real-time quantitative PCR. The relative mRNA levels were quantified using fluorescent TaqMan® technology. PCR primers and probes specific for murine VEGF (assay ID: mM 01281449_m1) and HIF-1α (assay ID: mM 00468869_m1) were obtained as TaqMan Gene Expression arrays (Applied Biosystems, Germany). Glyceraldehyde-3-A phosphate-dehydrogenase (GAPDH) (assay ID: mM 99999915_g1) was used as an internal control. PCR amplifications of HIF-1α, VEGF and GADPH were performed in 20-μl volumes and in separate reactions. The mix contained 1 μl of the 20X ready-to-use primer and probe mix for HIF-1α, VEGF or GAPDH, 2 μl of the undiluted cDNA, and 10 μl of the 2X TaqMan Universal Master Mix (Applied Biosystems). The probe was labeled at the 5′-end with the reporter molecule 6-carboxy-fluorescein (FMA) and at the 3′-end with a non-fluorescent quencher and MGB. PCR assays were performed using a CFX96 Real-Time PCR Detection system (Bio-Rad, Mississauga, ON, Canada). The PCR was initiated with a 2-min incubation at 50°C and then an initial 10-min denaturation at 95°C, followed by a total of 50 cycles of 15-sec denaturation at 95°C, and 1 min of annealing and elongation at 60°C. For each sample, ΔΔCt (crossing point) values were calculated as the Ct of the target gene minus the Ct of the GAPDH gene. Gene expression was derived according to the 2−ΔΔCt method; changes in gene expression are expressed relative to the basal condition.

Determination of the hypoxic status and perfusion level in tumor tissue

As markers of hypoxia, pimonidazole hydrochloride (1-[(2-hydroxo-3-piperidinyl)propyl]-2-nitroimidazole hydrochloride) (16,17) and CCI-103F (1-(2-hydroxy-3-hexafluoroisopropoxy-propyl)-2-nitroimidazole) (18) were used. Pimonidazole hydrochloride (Hypoxyprobe-1 Plus kit; Chemicon) and CCI-103F (Hypoxyprobe F6 kit; Chemicon) are bioreductive chemical probes with an immunorecognizable side chain. The addition of the first electron in the bioreductive activation is reversibly inhibited by oxygen, resulting in futile cycling with a half maximal pO2 of inhibition of ~3 mM Hg and with complete inhibition occurring at ~10 mM Hg (17). Pimonidazole hydrochloride was dissolved in saline, and CCI-103F was dissolved in 10% dimethyl sulfoxide (DMSO) and 90% peanut oil. To determine the effect of HS-1793 on vessel functionality, the double-fluorescent dye technique based on the perfusion markers, Hoechst 33342 (Sigma-Aldrich) and DiOC7 (Invitrogen) were used. Hoechst 33342 and DiOC7 are removed very rapidly from the circulation (half-time of 2 min) and are very stable once bound to DNA. Thus, Hoechst 33342 and DiOC7 specifically label the nuclei and mitochondria of endothelial cells, respectively, and those of the cells adjacent to the vessel walls, thereby delineating the perfused vessels (19,20). Hoechst 33342 and DiOC7 were dissolved in saline.

For the single-injection efficacy of HS-1793 concerning the change in tumor hypoxia, pimonidazole (60 mg/kg) was administered for 4 h 15 min, HS-1793 (1.5 mg/kg) was administered for 2 h 15 min, and CCI-103F (60 mg/kg) was administered for 2 h before animal euthanasia (Fig. 2A). With regard to the single-injection efficacy of HS-1793 and the change in tumor perfusion, Hoechst 33342 (15 mg/kg) was administered for 26 min, HS-1793 (1.5 mg/kg) was administered for 16 min, and DiOC7 (1 mg/kg) was administered for 1 min before animal euthanasia (Fig. 2B). For the 30-day repeat injection efficacy of HS-1793 (0.5, 1 and 1.5 mg/kg), pimonidazole hydrochloride was injected intravenously (i.v.) into the tail vein at a dose of 60 mg/kg, and Hoechst 33342 was injected i.v. at a dose of 15 mg/kg in a total volume of 0.1 ml. Pimonidazole hydrochloride was administered for 2 h, and H33342 was administered for 1 min before animal euthanasia. After animal euthanasia, the tumor specimens were removed and frozen in OCT mounting medium (Sakura Finetek, Torrance, CA, USA). Next, consecutive 4-μm-thick frozen sections were cut on a Shandon Cryotome FSE (Thermo Fisher, Waltham, MA, USA). The sections were then stored at −80°C until staining.

Figure 2 Effect of HS-1793 on TGF-β1 induced HIF-1α and VEGF expression in FM3A cells. The cells were treated with TGF-β1 (4 ng/ml) for the indicated times, and the HIF-1α, VEGF and phosphorylated forms of Smads (p-Smads) were detected by immunoblot assay (A). Dose-dependent induction of HIF-1α and VEGF expression by TGF-β1 under normoxic or hypoxic conditions (B). The cells were treated with different concentrations of HS-1793 for 2 h. Next, the cells were treated with TGF-β1 (4 ng/ml) under normoxic or hypoxic conditions for 12 h. The cell lysates were subjected to immunoblot analysis (C).

IHC staining and fluorescence microscopy in tumor tissues

After thawing, the sections were fixed in cold acetone (4°C) for 30 min. Between all consecutive steps of the staining procedure, the sections were rinsed once for 5 min in 1X Tris-buffered saline with 0.1% Tween-20 (TBS-T). Unless otherwise stated, all antibodies were diluted in PBS with 1% bovine serum albumin (BSA). The sections on slides were treated with rabbit anti-CD31 polyclonal antibody (1:100; Abcam, ab28364) or mouse anti-HIF-1α polyclonal antibody (1:100; Abcam, ab1) as a primary antibody overnight at 4°C. The sections were washed and then treated with a secondary antibody using the Envision Detection kit (Dako, K5007). Pimonidazole and CCI-103F are reductively activated in hypoxic cells and forms stable adducts with thiol groups in proteins, peptides, and amino acids. FITC-MAb1 (fluorescein-conjugated mouse IgG1 monoclonal antibody clone) and rabbit anti-CCI-103F antisera (PAbF6) bind to these adducts, allowing their detection by immunochemical means. The tumor sections were rehydrated and preincubated for 15 min with 1% BSA, immediately followed by incubation with FITC-conjugated anti-pimonidazole antibody (1:50) for 30 min and rabbit anti-CCI-103F antisera (PAbF6) (1:50) for 1 h at room temperature, respectively. For CCI-103F staining, the sections were washed and then treated with Texas Red Anti-Rabbit IgG (Vector, TI-100) (1:100) for 1 h at room temperature. Finally, the tumor sections were mounted with Malinol (Muto Pure Co.) and covered with a cover slip. The tumor sections were stored at 4°C and scanned within 1–2 days after staining. The different fluorescence excitation and emission properties of the Hoechst 33342 and DiOC7 dyes allow for the detection of temporal and spatial fluctuations in perfusion. Examinations were performed under a Nikon Eclipse 80i microscope (Nikon) equipped with Image Pro Plus 7.0 software (Media Cybernetics, Silver Spring, MD, USA).

Analysis of hypoxia and the perfused vasculature

The tumor sections were quantitatively analyzed using a semiautomatic method based on a computerized digital image analysis system. A high-resolution intensified solid-state camera on a fluorescence microscope (Nikon Eclipse 80i, Japan) with a computer-controlled motorized stepping stage was used. Each tumor cross-section was sequentially scanned at ×40 magnification using different filters for the Hoechst 33342 (blue), FITC (green), and Texas Red signals. After each scan, one composite digital image was reconstructed from the individual microscopic fields. The entire scanning procedure thus yielded three composite images from each tumor section that was analyzed with Image Pro Plus 7.0 software (Media Cybernetics). The figures were prepared using Adobe Photoshop 7.0. depending on the structures and markers visualized, different composite images were obtained. For the qualitative analysis of changes caused by oxygen-modifying intervention, both hypoxic and perfusion markers were stained on the same tissue. This process resulted in composite binary images showing the hypoxic markers CCI-103F (Texas Red) and pimonidazole (FITC) and the perfusion markers Hoechst 33342 (blue) and DiOC7 (green). We quantified the staining in two ways: the thresholds were applied to the pimonidazole and CCI-103F images to derive the fractions of the tumor section that were positive for each marker. Threshold values are inevitably arbitrary and subject to the possibility that a different threshold value would yield different results. To investigate this possibility, we exported the image histograms into Excel and generated positive fractions for each threshold value.

Statistical analysis

The results are expressed as the mean ± standard deviation (SD). Significant differences between the treatments and control were evaluated by ANOVA (Dunnett’s test). A P-value <0.05 was defined as statistically significant.

Results

HS-1793 changes the hypoxic condition in FM3A cells

To detect the cytotoxicity of HS-1793 in FM3A cells, we treated the cells with different concentrations of HS-1793 for 12 or 24 h under normoxic or hypoxic conditions, and we then measured the viability of FM3A cells. In Fig. 1A and B, no marked reduction of cell viability ≤2.5 μM HS-1793 was observed under either normoxic or hypoxic conditions in FM3A cells at 12 or 24 h. Thus, we used 0.6, 1.3 and 2.5 μM HS-1793 in subsequent experiments. We explored whether HS-1793 inhibits HIF-1α and VEGF expression relative to angiogenesis under hypoxic conditions of the tumor microenvironment. Pretreatment of FM3A cells with HS-1793 resulted in a dose-dependent decrease in hypoxia-induced HIF-1α and VEGF protein levels (Fig. 1C).

Figure 1 Effect of HS-1793 on the change in the hypoxic condition in FM3A cells. The cells were treated with different concentrations of HS-1793 for 2 h. Next, the cells were cultured under normoxic (A) or hypoxic (B) conditions for 12 or 24 h. Cell viability was determined after the treatment time and is expressed as a percentage of control growing cells. The cell lysates were subjected to immunoblot analysis with anti-HIF-1α and anti-VEGF (C). After HS-1793 treatment, the cells were incubated in the presence of 100 μM pimonidazol hydrochloride for 4 h and were fixed with 4% PFA. The cells were incubated for 30 min at room temperature with FITC-conjugated anti-pimonidazole antibody, and Hoechst 33342 was used as a nuclear counterstaining agent. The cells were imaged under Nikon Eclipse TS100 fluorescence microscopy (x200) (D). The data are expressed as the mean ± SD, and significant differences emerged between the treated groups by ANOVA followed by Dunnett’s test. *P<0.05 compared with the untreated control (0 μM) for the 24-h condition and #P<0.05 compared with untreated control (0 μM) for the 12-h condition.

To investigate whether HS-1793 regulates the hypoxic tumor microenvironment, we evaluated FM3A cells under hypoxic conditions. The cells were treated with HS-1793 under hypoxic conditions for 12 h and added characteristic pimonidazole adducts of hypoxia. In Fig. 1D, the fluorescent expression of pimonidazole was increased under hypoxic conditions compared with that under normoxic conditions. However, HS-1793 reduced the fluorescent expression of pimonidazole dose-dependently.

HS-1793 reduces TGF-β1-induced HIF-1α accumulation and VEGF expression in FM3A cells

HIF-1α plays a crucial role in VEGF induction under hypoxic conditions. To explore the role of TGF-β1 in HIF-1α and VEGF expression under normoxic conditions, we initially evaluated the effect of TGF-β1 on HIF-1α and VEGF expression in FM3A cells. HIF-1α and VEGF expression was upregulated by TGF-β1 at the protein level in a time-dependent manner. An increase in the HIF-1α level was clearly detectable 12 h after treatment, indicating that TGF-β1 enhances the stability of HIF-1α proteins in FM3A cells. To explore the signaling pathways implicated in the TGF-β1 induction of VEGF, we investigated the involvement of the Smad pathway, a crucial mediator of the TGF-β1 signaling cascade. In response to TGF-β1, the levels of Smad2 and Smad 3 phosphorylation were increased in FM3A cells (Fig. 2A). Moreover, FM3A cells were treated with TGF-β1 under normoxic and hypoxic conditions for 12 h. As shown in Fig. 2B, the HIF-1α and VEGF protein levels were markedly increased by TGF-β1. However, pretreatment with HS-1793 in FM3A cells reduced HIF-1α and VEGF expression dose-dependently (Fig. 2C). Therefore, these findings suggest that HS-1793 inhibits angiogenesis via the regulation of HIF-1α and VEGF expression not only under hypoxia but also under the TGF-β1-treated condition.

HS-1793 improves the perfusion level and hypoxic status in tumor tissues in mice

We assessed blood flow fluctuations in HS-1793-injected mice bearing FM3A breast cancer cells using the double-fluorescent dye method based on two perfusion markers, Hoechst 33342 and DiOC7. Fluorescence microscopy revealed similar intensities to Hoechst 33342 and DiOC7 in control FM3A breast tumors. By contrast, HS-1793-treated (1.5 mg/kg) tumors showed uniform uptake of DiOC7 dye, indicative of improved vessel functionality. To determine whether HS-1793 leads to decreased tumor hypoxia, mice bearing tumors were injected with the hypoxia markers pimonidazole and CCI-103F. Pimonidazole and CCI-103F were stained with green (FITC) and red (Texas Red) colors, respectively, on the photomicrograph. At 2-h time point, the staining intensities of the two hypoxia markers within the same range of a control tumor were similar between pimonidazole and CCI-103F. The extensive hypoxia that was present before HS-1793 treatment (1.5 mg/kg) (hypoxic fraction determined by pimonidazole staining) was reduced after HS-1793 treatment (1.5 mg/kg) (hypoxic fraction determined by CCI-103F staining) (Fig. 3C). As shown in Fig. 3D, a significant difference between the untreated and HS-1793-treated images was obtained over a wide range of threshold values.

Figure 3 Changes in tumor perfusion and hypoxia after HS-1793 treatment in FM3A tumor-bearing mice. Experimental setup with the injection schedule for the two sequentially administered hypoxic (A) and perfusion cell markers (B). Composite panoramic sections (original ×40, reduced for panoramic view) in tumors stained for both perfusion (Hoechst 33342, blue; DiOC7, green) and hypoxia markers (Pimonidazole, green; CCI 103F, red) (C). The mean perfusion and hypoxic fractions were exported as image histograms into Excel, and positive fractions were generated for each threshold value (D). Composite panoramic sections (original ×40, reduced for panoramic view) in the tumor tissue (untreated control, 0.5, 1 and 1.5 mg/kg) captured the Hoechst 33342-labeled cells (upper) and pimonidazole-stained cells (down). Tissue regions of a section from a tumor isolated on day 30 (E). The mean Hoechst and pimonidazole fraction were exported as image histograms into Excel, and positive fractions were generated for each threshold value (F and G). #P<0.05, compared with the untreated control.

The mean number of Hoechst 33342-labeled cells represents the blood perfusion of the tumors and indirectly represents the level of hypoxia in the tumors. The lower number of labeled cells suggests impaired blood perfusion and a higher level of hypoxia in the tumors. The Hoechst 33342 stain appeared as a blue color on the photomicrograph (Fig. 3E). The percentages of Hoechst 33342-labeled cells in the untreated control and HS-1793 treated groups (0.5/1.0/1.5 mg/kg) were 13.8 and 17.0/20.9/22.6%, respectively (Fig. 3F). The number of labeled cells in the HS-1793-treated group increased significantly compared with that in the untreated control group. This result suggests increased blood perfusion and slightly decreased tumor hypoxia after HS-1793 treatment.

To examine the effects of HS-1793 treatment on tumor hypoxia, the changes in the extent of the hypoxic areas in FM3A xenografts were assessed by immunohistochemical analysis with pimonidazole. As shown in Fig. 3E, an obvious reduction of the hypoxic areas, that is, the pimonidazole-positive regions, was observed in tumors treated with HS-1793. The proportions of the hypoxic areas to the entire tumor regions were scored as 19.2% (untreated control) and 14.0% (HS-1793 treated mouse, 1.5 mg/kg), respectively, indicating that the proportion of the hypoxic areas after HS-1793 treatment was significantly decreased compared with that in the untreated control (Fig. 3G).

HS-1793 stimulates anti-angiogenesis through HIF-1α suppression

To examine the effect of HS-1793 on HIF-1α expression in tumor tissue, we performed immunohistochemical analysis of HIF-1α (Fig. 4A). Quantitative data showed that HS-1793 treatment in mice bearing FM3A cells (1.5 mg/kg) affected HIF-1α expression (0.6%) compared with the untreated control (1.9%; Fig. 4B). The mRNA levels of HIF-1α following treatment with HS-1793 were determined on day 30 after treatment. The results showed that HIF-1α mRNA expression decreased significantly after HS-1793 (1.5 mg/kg) treatment compared with that in the untreated control (Fig. 4C).

Figure 4 Suppression of HIF-1α expression and anti-angiogenic activity after HS-1793 treatment in FM3A xenografts. C3H/He female mice were inoculated by single s.c. injection of 2×106 FM3A tumor cells in 50 μl of PBS. Mice were treated with HS-1793 (0.5, 1 and 1.5 mg/kg, i.p.) twice a week for 30 days when the tumor grew to a size of ~40 mm3. Typical microscopic fields of FM3A xenografts immunostained for HIF-1α (A). Quantitative data obtained by immunohistochemical analysis (B). Real-time PCR was performed with specific HIF-1α (C). Typical microscopic fields of FM3A xenografts immunostained for CD31 (D). The microvessel density of FM3A was determined as described above (E). The serum VEGF concentrations from animals bearing xenografts were determined using a quantitative enzyme-linked immunosorbent assay (ELISA, R&D Systems) (F). The data are expressed as the mean ± SD of five mice per group. *P<0.05, compared with tumor-free mice. #P<0.05, compared with the untreated control. Tumor regions of a section from a tumor isolated on day 30. Original magnification, ×200.

There are some reports on the anti-angiogenic activity of resveratrol (21). In this study, we examined the anti-angiogenic activity of the resveratrol analog HS-1793 and found that HS-1793 significantly decreased microvessel density (MVD; Fig. 4D). MVD, that is, the percentage of endothelial cells associated with pericytes, is used as a quantitative measure of vascular maturation. Quantitative analysis of pericyte coverage showed 8.29% MVD staining in HS-1793-treated (1.5 mg/kg) FM3A tumors compared with control FM3A tumors (15.68%; Fig. 4E).

The VEGF peptide concentration in mouse serum was significantly higher in untreated control mice (446 pg/ml) than in tumor-free mice (187 pg/ml) but was significantly reduced (P<0.05) in HS-1793-treated mice in a dose-dependent manner (Fig. 4F). Therefore, these findings suggest that HS-1793 inhibits angiogenesis via the downregulation of HIF-1α and VEGF expression under hypoxic conditions of the tumor microenvironment.

To determine the effect of HS-1793 treatment (0.5, 1.0 and 1.5 mg/kg) twice a week for 30 days on tumor growth, changes in the tumor volume were calculated for a period of 30 days following treatment. The tumor volume in the HS-1793-treated mouse group (1.0 and 1.5 mg/kg) was greatly decreased after HS-1793 treatment compared with the untreated control (22). These results indicate that HS-1793 delays tumor growth via the improvement of hypoxia in murine breast cancer.

HS-1793 inhibits hypoxia-induced cancer stem cell properties, including migration and radio-resistance of FM3A cells

Hypoxia increases the number of cells that express CSC markers in bulk populations (23–27). Interestingly, our data showed that HS-1793 decreased the hypoxia induced expression of OCT4, KLF4 and SOX2 protein (Fig. 5A). In other experiments, we investigated the effects of HS-1793 on cell migration under hypoxic conditions. As shown in Fig. 5B, there was minimum migration of FM3A cells under normoxic conditions, while we observed higher levels of FM3A cell migration under hypoxic conditions. However, HS-1793 inhibited hypoxia-induced cell migration in a dose-dependent manner. Accordingly, these results suggest that HS-1793 effectively inhibits the CSC properties of FM3A cells induced by hypoxia. The combinatorial effect of HS-1793 (2.5 μM) and radiation (2 or 4 Gy) in hypoxic FM3A cells was evaluated by the Annexin V/PE apoptosis assay after 12 h of treatment. The 7-AAD-positive cells that also bound Annexin V were defined as late apoptotic cells (Annexin V+, 7-AAD+). In Fig. 5C, Annexin V/PE-positive cells were decreased by radiation treatment under hypoxic conditions compared with that under normoxic conditions. However, pretreatment with HS-1793 increased the percentage of late apoptotic cells (4.6±0.38%) compared with the percentage of apoptotic cells in the radiation-treatment (2.5±0.15%) groups at 1% O2 (hypoxia). These data suggest that the combinatorial treatment of HS-1793 and radiation increases the therapeutic effect in hypoxic FM3A cells.

Figure 5 Effect of HS-1793 on hypoxia-induced cancer stem cell characteristics in FM3A cells. The cells were treated with different concentrations of HS-1793 for 2 h. Next, the cells were cultured under normoxic or hypoxic conditions for 12 h. The cell lysates were subjected to immunoblot analysis with anti-OCT4, anti-KLF4, anti-SOX2 and anti-survivin (A). The cells were cultured in normal growth medium to confluence, and then, a scrape in the form of a cross was generated through the confluent monolayer with a plastic pipette tip. The cells were pretreated with vehicle or different concentrations of HS-1793 for 2 h under normoxic or hypoxic conditions. Several wounded areas were marked for orientation, observed and then imaged by phase-contact microscopy 24 h after the scratch (B). Cells were treated with radiation (0, 2 and 4 Gy) or radiation+HS-1793 (2.5 μM). After 72 h, the percentage of Annexin V/PE-positive cells was quantified using flow-assisted cell sorting (C). The data are expressed as the mean ± SD, and significant differences emerged between treated groups by t-test.

Discussion

Breast cancer remains a common and frequently fatal disease among women. Generally, the process of tumor progression is characterized by rapid cellular growth accompanied by alterations in the microenvironment of tumor cells (28). Critical changes in the cellular microenvironment occur under hypoxic conditions, where oxygen supply is inadequate (29). Breast carcinomas usually support their growth by stimulating blood vessel development (angiogenesis). Blood flow within these new vessels is often chaotic, causing periods of hypoxia followed by reperfusion.

Aggressive and metastatic cancer phenotypes that are associated with resistance to radiation therapy, chemotherapy, and a poor treatment outcome can be generated as a result of the hypoxic environment within the tumor (30). Hypoxia detected in the central regions of solid tumors is a leading cause of angiogenesis, a fundamental determinant of malignant tumor progression, by activation of angiogenic factors (31). A key factor in this process is HIF-1. The overexpression of HIF-1α protein has been shown in many human cancers and their metastases and is closely associated with a more aggressive tumor phenotype, including an advanced tumor grade, an increased vascularity, increased resistance to chemo/radiotherapies, and tumor progression (32). The most characterized HIF-regulated gene is VEGF, which is involved in regulating endothelial cell proliferation and blood vessel formation in both normal and cancer cells (33). TGF-β1 is involved in blood vessel formation, and one of its oncogenic functions is to promote tumor angiogenesis, which is associated with its ability to induce VEGF expression (34). Hypoxia, TGF-β1 and VEGF are important factors of the tumor microenvironment that regulate cancer progression and metastasis (35). Recently, resveratrol remarkably inhibited hypoxia-induced HIF-1α accumulation and VEGF expression in both human tongue squamous cell carcinomas (SCC-9) and hepatoma (HepG2) cells and dramatically suppressed hypoxia-stimulated invasiveness of SCC-9 cells in vitro (36).

The in vitro and in vivo findings from this study suggest that the resveratrol analog HS-1793 modulates the hypoxic status and perfusion levels in mouse breast cancer FM3A cells. Our data show that HS-1793 decreased HIF-1α and VEGF expression under hypoxic conditions in FM3A cells (Fig. 1). FM3A cells were treated with TGF-β1 under normal and hypoxic conditions, and HIF-1α and VEGF protein levels were markedly increased by TGF-β1. However, pretreatment with HS-1793 reduced HIF-1α and VEGF expression in FM3A cells dose-dependently (Fig. 2). When the tumor grew to a size of ~500 mm3, to determine the acute injection efficacy of HS-1793, hypoxia markers (pimonidazole/CCI-103F) and perfusion markers (Hoechst 33342/DiOC7) were used to evaluate changes in tumor hypoxia and perfusion. The results indicated a reduced hypoxic status and enhanced perfusion in the tumor mass. To evaluate the therapeutic effect of HS-1793, mouse breast cancer cells (FM3A) in the logarithmic growth phase were inoculated subcutaneously on the right flank of female C3H/He mice, and the effects of intraperitoneal injection of HS-1793 on HIF-1α and VEGF expression, microvessel density, the change in the hypoxic status and perfusion level, and tumor growth were determined after 4 weeks of treatment. The higher level of serum VEGF in tumor-bearing mice was dose-dependently diminished in the HS-1793-treated groups, a result that was in accordance with the result of HIF-1α expression in the tumor tissue. The tumor tissue of the HS-1793-treated group showed weak expression in immunofluorescent staining of pimonidazole, which specifically binds to thiol-containing proteins in hypoxic cells, and immunohistochemical staining of the endothelial cell marker CD31. However, Hoechst 33342-labeled cells, which represent the blood perfusion of tumor tissue, were markedly increased in the HS-1793-treated groups (Figs. 3 and 4). Finally, HS-1793 delays tumor growth via the improvement of hypoxia in murine breast cancer (22).

CSCs account for a minor fraction of tumor populations; however, those cells might be particularly involved in tumor initiation, proliferation, or metastatic process. HIF-1 has been implicated in the maintenance of CSCs (37,38), and knockdown experiments have concluded that HIFs are required for CSC survival and tumor progression (39). In this study, HS-1793 effectively inhibited hypoxia-induced CSC properties, including the migration of tumor cells under hypoxic conditions in FM3A cells (Fig. 5A and B). Radiation therapy is an effective modality for the treatment of many tumors (40). However, the dose-limiting normal tissue toxicity and radio-resistant tumors are still linked to life-threatening radiation treatment failure (41). The enhanced level of HIFs in tumors can be correlated with the level of oxygen within the tumor and has been shown to correlate with tumor radiation resistance (42,43). IR-induced apoptosis was suppressed under hypoxic conditions compared with that under normoxic conditions. However, HS-1793 enhanced IR-induced apoptosis in hypoxic FM3A cells (Fig. 5C).

In conclusion, this study demonstrated that the resveratrol analog HS-1793 exerts anti-metastatic effects by the inhibition of cell migration via the reduction of HIF-1α and VEGF expression, and HS-1793 inhibits hypoxia-induced cancer stem cell properties and enhances IR-induced apoptosis in hypoxic mouse breast tumor FM3A cells. The results of this study demonstrate that HS-1793 increases the therapeutic efficacy of radiation in hypoxic cells through the modulation of the hypoxic condition.

Acknowledgements

This study was supported by the National Research Foundation of Korea (DIRAMS) grant funded by the Korea government (MSIP) (50590-2016) and the [Basic Science Research Program] through the National Research Foundation of Korea (NRF) (NRF-2013 M2A2A 7043665) funded by the Ministry of Science, ICT & Future Planning.

References