Polyphenol-rich strawberry extract (PRSE) shows in vitro and in vivo biological activity against invasive breast cancer cells

We describe the biological effects of a polyphenol-rich strawberry extract (PRSE), obtained from the “Alba” variety, on the highly aggressive and invasive basal-like breast cancer cell line A17. Dose-response and time-course experiments showed that PRSE is able to decrease the cellular viability of A17 cells in a time- and dose-dependent manner. PRSE effect on cell survival was investigated in other tumor and normal cell lines of both mouse and human origin, demonstrating that PRSE is more active against breast cancer cells. Cytofluorimetric analysis of A17 cells demonstrated that sub-lethal doses of PRSE reduce the number of cells in S phase, inducing the accumulation of cells in G1 phase of cell cycle. In addition, the migration of A17 cells was studied monitoring the ability of PRSE to inhibit cellular mobility. Gene expression analysis revealed the modulation of 12 genes playing different roles in the cellular migration, adhesion and invasion processes. Finally, in vivo experiments showed the growth inhibition of A17 cells orthotopically transplanted into FVB syngeneic mice fed with PRSE. Overall, we demonstrated that PRSE exerts important biological activities against a highly invasive breast cancer cell line both in vitro and in vivo suggesting the strawberry extracts as preventive/curative food strategy.


Results
Phytochemical analysis of PRSE. The nutritional and phytochemical composition of PRSE was characterized by analyzing vitamin C content, phytochemical levels and the total antioxidant capacity (TAC) of the extract. Results are reported in Table 1 and show, according to our previous studies 27 , that "Alba" extract presents high levels of Vitamin C (0.57 mg/g) and phytochemical compounds measured as Total Phenolic Content (TPH) (2.26 mg GAEq/g), Total Flavonoids Content (TFC) (0.61 mg CEq/g) and Total Anthocyanin Content (ACY) (0.46 mg Pg-glcEq/g).
However, TFC and ACY values are similar to those of other strawberry cultivars, while the level of total phenolic compounds results very high if compared to commercial strawberry varieties previously analyzed 27,28 .

Biological effects of PRSE on A17 cells. The effects induced by the extract on the survival of A17 cells
were analyzed in both dose-response and time-course experiments. Cells were treated for 24 h, 48 h or 72 h with concentrations of extract ranging from 0.5 to 5 mg/ml. Interestingly, we found that the extract reduces the survival of A17 cells in a time and dose-dependent manner (Fig. 1a).
Once characterized in A17 cells, the activity of PRSE was investigated also in normal fibroblast cell lines of both murine (NIH-3T3) and human (WI38) origin and in the human breast cancer cell line MCF-7 by dose-response studies at 48 h of treatment. Interestingly, we found that: i) A17 cells show the higher response to PRSE exposure (IC 50 of 1.14 ± 0.29 mg/ml) and, most importantly, that ii) normal cells are significantly less sensitive to PRSE respect to cancer cell lines (mean IC 50 of 3.24 ± 0.14 mg/ml and 1.68 ± 0.77 mg/ml, respectively - Fig. 1b).
Subsequently, possible alterations of cell cycle phase distribution were evaluated by cytofluorimetric analysis of propidium iodide stained cells. A17 cells treated for 48 h with different concentrations of PRSE showed, at the concentrations of 0.5 and 1 mg/ml of PRSE treatments, a reduction of cells in S phase and a concomitant increase of cells in G1 phase of the cell cycle (Fig. 2). Moreover, treatments with higher doses of PRSE (2.5 mg/ml) drastically change the biological response of A17 cells, hindering the cell cycle analysis but showing the appearance of a hypodiploid cellular subpopulation (Fig. 3).
A17 cell line is an intriguing model of study because of its highly aggressive and invasive phenotype 25 . In consideration of the critical role of cellular migration on the multistep process of tissue invasion 29  to investigate the possible interference of PRSE on the mobility of A17 cells by wounding assay. To avoid cell turnover that could mimic the cell migration, cells were starved during this assay. After 48 h of culture, A17 cells not subjected to treatments were able to migrate and partially fill the wound empty area while exposure to PRSE inhibited the cell migration in a dose-dependent manner (Fig. 4a). Notably, the inhibition of wound closure was monitored starting from the lowest concentration of PRSE tested (0.5 mg/ml), and appears almost completely inhibited starting from a concentration of 2.5 mg/ml (Fig. 4a). A quantitative analysis of wound closure was carried out using the NIH Image J software (Fig. 4b).
PRSE modulation of gene expression. The biological effects mediated by the extract on A17 cells were also investigated at the molecular level by analyzing the expression of a panel of genes known to be involved, with different roles, in the cellular migration, adhesion and invasion processes. For this purpose, the level of 84 different mouse gene transcripts was analyzed by applying a quantitative RT-PCR array on A17 cells untreated or treated for 48 h with sub-lethal doses of PRSE. Among the 84 gene transcripts screened, 12 genes showed a regulation exceeding the 2-fold criteria compared to the untreated control (Fig. 5a). Genes down-regulated include the colony stimulating factor 1 (Csf1, − 2. Subsequently, two genes showing significant modulation, Mcam and Nr4a3, were further investigated at protein level by western blotting (Fig. 5b). Densitometric analysis of the bands of the two proteins was performed and normalized to alpha-tubulin expression, confirming the downregulation of Mcam (− 1.89 fold at 0.5 mg/ml and − 4.25 fold at 1 mg/ml of extract) and, at a less extent, of Nr4a3 (− 2.37 fold at 0.5 mg/ml and − 1.54 fold at 1 mg/ml of extract). In addition, caspase-1 levels were also analyzed by western blotting showing the upregulation of the protein at both 0.5 mg/ml (+ 1.89 fold) and 1 mg/ml (+ 4.16 fold) of PRSE (Fig. 5b).
In vivo evaluation of PRSE activity. A17 cells were used to generate an orthotopic model of breast cancer in syngeneic female FVB/N mice. Four weeks mice were fed with 15% strawberry extract-enriched food or regular food as control (ten mice for each group). Upon the 8 th week of age, all the mice underwent tumor challenge while animals of each group continued to receive their diets. After 5 weeks, tumors were withdrawn, measured using a caliper and weighted. Results show a significant reduction in both tumor weight (Fig. 6a) and tumor volume ( Fig. 6b) in mice fed with PRSE.

Discussion
The strawberry (Fragaria x ananassa) was chosen as the model in this study for its high content of antioxidants and bioactive compounds, as well as for its high availability for food industry and fresh consumption 13 . In the last decades, growing interest has been focused on the antioxidant capacity of strawberries so that TAC is considered a quality parameter and an indicator of bioactive compounds present in the fruit. The antioxidant capacity of the strawberry indicates that its consumption could contribute to prevent and reduce oxidative reactions that cause negative effects on human health, playing different roles in the development of chronic diseases and cancers.
Moreover, even if the high TAC of strawberries has been proved, it has also been demonstrated that this parameter is strongly influenced by the strawberry genetic background and is strictly related to the presence of oxygen radical scavengers such as Vitamin C and polyphenols 11,27,28,30,31 . Among polyphenols, anthocyanins are quantitatively the most important phenolic compounds in strawberries 13 , so that more than 25 different anthocyanin pigments have been described from different varieties and selections 32 . Analyses of TPH, ACY and TFC in the "Alba" cultivar in the present study confirm the high content of phytochemicals of this strawberry.
Strawberries are important also for their high content of Vitamin C, which is even higher than that of citrus fruit. Vitamin C content is an essential parameter due to the high number of biological roles that it plays in humans, lowering the incidence of cardio-and cerebro-vascular diseases 33 , of several cancers 34 , and other health disorders such as lead toxicity 35 . When evaluating the Vitamin C content in strawberries, it is important to consider that the molecule is very labile and in adverse conditions undergoes oxidation, depending on several factors such as temperature, water and pH 36 .
The choice of using "Alba" strawberries in this study was also based on the very interesting results obtained in recent years using this cultivar, both in vitro 37,38 and in vivo 39,40 .
As shown above, the "Alba" strawberry extract is able to strongly decrease cellular viability of the highly tumorigenic stromal cell line A17, which is characterized by mesenchymal features and metastasis formation ability 26,41 .  In particular, it is possible to distinguish two different biological behaviors: i) a cytostatic-like effect at low doses of PRSE and ii) an acute toxic effect at higher doses. The capacity of the strawberry extract to activate the apoptotic process, as well as its anti-cancer potential in other cancer models, was already characterized by other authors 19 . However, although these observations are in accordance with the increased hypodiploidy observed in this study at high PRSE doses, the analysis of DNA fragmentation by laddering assay did not show the induction of the apoptotic internucleosomic cleavage of DNA (data not shown). Interestingly, we observed an increased expression level of the precursor form of caspase-1 that, together with the monitored hypodiploidy, could suggest an activation of the apoptotic pathway at a very early stage, and thus below the detection limit of the experimental approaches used. In addition, we found no evidence of caspase-3 modulation and/or activation (data not shown).

Figure 6. Effect of PRSE on A17 tumors in mice.
Ten mice (fed with 15% strawberry extract-enriched food) were housed one mouse/cage, while ten mice (fed with regular food) were housed in two cages with free access to water and food. Upon the 8 th week of age, all the mice underwent tumor challenge with 2 × 10 5 A17 cells. After cells injection mice were orally treated for further 5 weeks, at the end of which tumors were withdrawn and both tumor weight (a) and tumor volume (b) were measured. * * * P < 0.001 by Student's T test.
Scientific RepoRts | 6:30917 | DOI: 10.1038/srep30917 We hypothesized that sub-lethal doses of strawberry extract could be able to inhibit the known invasive potential of A17 breast cancer cells. Thus, the effect of PRSE on the cellular migration, which is known to play a pivotal role in invasion, was first analyzed at a biological level by wounding healing assay. The observation that low doses of "Alba" PRSE were sufficient to block cell migration gave us the rationale to extend the study at the molecular level. The subsequent characterization of the expression levels of genes involved in the cellular migration, adhesion and invasion processes allowed us to obtain an overall view of changes that take place in several pathways. In particular, the down-regulation of a subgroup of genes (Csf1, Mcam, Nr4a3 and Set) supports the anti-invasion effect of the PRSE extract because high expression levels of these genes are often associated to the invasive phenotype of cells. In the context of human breast cancer it has been already reported that: i) high expression of Csf1 is correlated to metastasis formation and thus has been proposed as negative prognostic factor; ii) up-regulation of human Mcam (named Muc18 in humans) promotes motility, invasiveness and tumorigenesis of human breast cancer cells; iii) high expression levels of Nr4a3 were found correlated with increased risk of developing distant metastasis in triple-negative breast cancer patients; iv) Set nuclear proto-oncogene (named also I2pp2a) is one of the genes down-regulated by the mushroom Ganoderma lucidum extract and is involved in the suppression of breast-to-lung cancer metastasis [42][43][44][45] . Interestingly, the modulation of Mcam and Nr4a3 was confirmed also at the protein level, further suggesting a role of these proteins in the effect induced by PRSE on A17 cells. Likewise, we found two genes whose monitored up-regulation, as a consequence of PRSE treatment, can be consistent with an anti-invasion effect: i) Htatip2 (named also Tip30), a putative metastasis suppressor gene which is inversely correlated with lymph node metastasis in breast cancer patients, and ii) Gpnmb (also named Osteoactivin/HGFIN), a gene with a controversial role in the metastatic process 46,47 .
Otherwise, it was not possible to find any correlation between the known activity of the remaining six genes found up-regulated by PRSE (Ccl7, Ctsl, Cxcr4, Itgb3, Mmp19, Mmp3) since their functions are generally associated to the acquisition of invasive features 44,48-51 .
As already stated, the role of strawberry bioactive compounds on cancer prevention seems to involve different mechanisms of action that are still unclear. Previous studies indeed indicate that the biological activity exerted by berries against cancer cells is probably mediated by the synergy between different compounds suggesting the involvement of several molecular pathways. To date, only few studies have tried to investigate the molecular basis of strawberry activity against breast cancer cells. For example, the role of p73 in triggering apoptosis of p53-null cells has been recently suggested 19 . Our study indicates the involvement of other genes known to play key roles in cellular invasion, adhesion and migration.
Notably, cell survival studies show that the effect of the extract is significantly higher in breast cancer cell lines, and in particular A17 cells, with respect to normal cells, suggesting a therapeutic window for PRSE in vivo. This observation is in accordance with previous studies on mice that showed efficacy of strawberry methanolic extract against Ehrlich ascites carcinoma (EAC) 19 . We thus explored the efficacy of PRSE in inhibiting tumor formation in vivo exploiting the intriguing orthotopic breast cancer model generated by injection of A17 cells directly into the mammary gland of mice, finding a strong reduction of both tumor size and tumor volume in mice fed with PRSE from 4 weeks before tumor challenge.
Although the mechanism by which the strawberry extract exerts its biological effect is not completely understood at the molecular level, as well as which component plays the major antineoplastic role, our results suggest an interesting anti-invasive potential of "Alba" PRSE against breast cancer cells both in vitro and in vivo. Further studies will be necessary to unravel the pathways involved in the biological effects of PRSE and to elucidate the molecular basis of strawberry extract action, as well as to shed light on the possible introduction in the diet of "Alba" and, more in general, strawberries as a useful nutrient to limit (or prevent) tumor formation.

Methods
Preparation of PRSE. Strawberry fruits of the "Alba" variety were collected in the experimental fields of the Agricultural Faculty of "Università Politecnica delle Marche" located in Agugliano (AN), in central east Italy (43°31′ 60″ N-13°22′ 60″ E). Fruit samples from the selected variety were hand-picked at the same day-time on different days, corresponding to the ripening times of the selected clone, from the second to the fourth picking. Fruit samples were selected for homogenous fruit, avoiding unripe, wounded or shriveled fruits. Within 2 h after harvest, whole fruits were stored at − 20 °C before analyses. For the evaluation of total antioxidant capacity (TAC), total phenolic content (TPH), total anthocyanin content (ACY) and total flavonoid content (TFC), a methanolic extract was prepared via homogenization. Frozen strawberries were thawed for 60 min at 4 °C. Ten gram aliquots of the fruits were added to 100 mL of the extraction solution, consisting of methanol/milliQ water/concentrated formic acid (80:20:0.1 v/v), and fruits were homogenized using an Ultraturrax T25 homogeniser (Janke & Kunkel, IKA Labortechnik, Staufen, Germany) for 2 min. Extraction was maximized by stirring the suspension for 2 h in the dark at room temperature (RT), then the tubes were centrifuged at 3500 rpm for 15 min, in two sequential times, to sediment solids. Supernatants were filtered through a 0.45 μ m Minisart filter (PBI International, Milan, Italy), transferred to 5.0 ml amber glass vials and stored at -20 °C until analysis. Methanolic extract was concentrated through a rotary evaporator and stored in aliquots at − 80 °C for subsequent experimental procedures. Vitamin C analysis was conducted as described 28 . PRSE analysis. Two methods were used for the determination of the antioxidant capacity (TAC) of strawberry extracts: the Trolox Equivalent Antioxidant Capacity (TEAC) and the Ferric Reducing Antioxidant Power (FRAP). TEAC assay consists in the quantification of strawberry extract free radical scavenging activity against 2,2-azinobis-(3-ethylbenz-thiazoline-6-sulfonate) radical cation (ABTS + ), using Trolox (6-Hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid) as reference standard 52,53 . Briefly, 1 ml of the ABTS + radical solution was mixed with 10 μ l of reagent (strawberry extract or standard). The analysis solution was vortexed for 20 seconds and, after 1-3 min, spectrophotometrically analyzed at 734 nm measuring the color inhibition of the Scientific RepoRts | 6:30917 | DOI: 10.1038/srep30917 ABTS + radical. FRAP assay was carried out according to the protocol proposed by Deighton and coworkers 54 , with slight modifications 55 .
Total phenolic content (TPH) of the strawberry extracts was determined using the Folin-Ciocalteu colorimetric method, as modified by Slinkard and Singleton 56 . Briefly, 100 μ L of sample (milliQ water, water diluted strawberry extracts or gallic acid standard solutions) were added to 500 μ L of Folin-Ciocalteau reagent previously diluted in water (dil. 1/10) and kept at 4 °C, in the dark. The mixture was incubated for 1 to 8 min at RT, then 400 μ L of 0.7 M sodium carbonate were added and the mixture was vortexed. The solution was incubated for 2 h at RT, in the dark and spectrophotometrically analyzed at 760 nm.
The total anthocyanin (ACY) content of the strawberry extracts was determined using a modified pH differential method 57 , while total flavonoid content (TFC) was determined by using a colorimetric method 58,59 .
Vitamin C content was measured through HPLC analysis immediately after the extraction procedure. The HPLC system comprised a Jasco (Jasco Inc., Easton, USA) PU-2089 Plus controller and a Jasco UV-2070 Plus ultraviolet (UV) detector set at absorbance of 260 nm. An isocratic elution with 50 mM potassium phosphate at pH 3.2 was performed by means of a Supelcosil LC8 150 × 4.6 mm HPLC column (Sigma-Aldrich S.r.l. Milan, Italy) 60 . Quantification of the vitamin C was carried out through a comparison with pure vitamin C calibration curve. All the analyses were conducted in triplicate. All the cell lines were grown in a humidified atmosphere at 37 °C and 5% CO 2 as previously described 61,62 .
Cell viability. Dried PRSE, stored in aliquots at − 80 °C, was diluted in culture medium immediately before use. Cells were seeded in triplicate in 6-well plates 16 hours before PRSE addition. After treatment with different concentrations of extract, cell viability was measured by a TC10 automatic cell counter (BioRad, Hercules, CA, USA) and compared with an untreated control to estimate the cell viability. The 50% inhibitory concentration (IC 50 ) value was calculated with CompuSyn software (ComboSyn, Inc., Paramus, NJ, USA) 63 . Data are reported as mean (± SD) resulting from three independent experiments.
Cell cycle analysis. Cell cycle was analyzed using the propidium iodide staining procedure as previously reported 64 . Cells were fixed in ice-cold 70% ethanol and stained using a propidium iodide staining solution (50 mg/ml). Cytofluorimetric acquisitions were carried out with a PAS flow cytometer (Partec, Münster, Germany) and sample analysis conducted using FlowJo 8.6.3 software (Tree Star, Inc., Ashland, OR, USA). Cell cycle percentage values were calculated using a Watson pragmatic model.

Wounding assay.
Wounding assay was performed as previously described on the same cellular model (A17) 65 . Briefly, a linear wound was produced in A17 confluent cellular population by scratching the bottom of each dish with a sterile pipette tip. After wounding, cells were washed with phosphate buffer saline (PBS) and incubated in high glucose DMEM containing 0.5% FBS with different strawberry extract concentrations (0.5, 1, 2.5 and 5 mg/ml). A17 cells were allowed to migrate for additional 48 h after which images acquisition was carried out by a LeitzFluovert FU (Leica Microsystems) microscope. Remaining wound areas were determined using NIH Image J software for calculation of the percentage of wound closure. Analyses were performed in triplicate.
Gene expression analysis and western blotting. The Tumor Metastasis RT 2 Profiler PCR Array (Qiagen, Hilden, Germany), consisting of 84 mouse genes known to be involved in the metastasis formation process, was used to profile the biological response of A17 cells induced by the PRSE. Briefly, total RNA was extracted from treated or untreated A17 cells using RNeasy extraction kit and reverse transcribed into cDNA using an RT 2 First Strand Kit (Qiagen). The cDNA was combined with RT 2 SYBR Green qPCR Master Mix (Qiagen), and equally distributed (25 μ l) to each well of the PCR array plate. Real-time PCR assay and subsequent data collection were performed on a Rotor-Gene 6000 robocycler (Corbett Life Science, Sydney, Australia) 66 . Transcript relative enrichments were calculated following manufacturer's instructions (Qiagen). Experiments were performed in triplicate.
In vivo evaluation of PRSE efficacy. Four weeks old female FVB/N mice were obtained from Animal Care Facilities of the University of Camerino. Ten mice (fed with 15% strawberry extract-enriched food) were Scientific RepoRts | 6:30917 | DOI: 10.1038/srep30917 housed one mouse/cage, while ten mice (fed with regular food) were housed in two cages with free access to water and food, and kept at temperature of 19-22 °C and relative humidity of 45-65% under 12 h/12 h light/dark cycle. Upon the 8 th week of age, all the mice were orthotopically injected with 2 × 10 5 A17 cells. Tumor monitoring was performed twice a week by palpation. After 5 weeks, the tumors were analyzed, after resection, evaluating both weight and volume (two perpendicular diameters -a and b -on each tumor were measured using a caliper and the volumes were calculated by the V = π /6[(a + b)/2] 3 formula). All the experimental procedures carried out in this study were in compliance with the UK Animals (Scientific Procedures) Act 1986 and associated guidelines, EU Directive 2010/63/EU, and were approved by the Ethic Committee on Animal Use of the University of Camerino (protocol number 14/2012).