PHYTOTHERAPY RESEARCH

Phytother. Res. 29: 1964–1973 (2015)

Published online 14 October 2015 in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/ptr.5491

Betanin-Enriched Red Beetroot (Beta vulgaris

L.) Extract Induces Apoptosis and Autophagic
Cell Death in MCF-7 Cells

Laëtitia Nowacki,1,2 Pascale Vigneron,2 Laura Rotellini,2 Hélène Cazzola,1 Franck Merlier,1
Elise Prost,1 Robert Ralanairina,3 Jean-Pierre Gadonna,3 Claire Rossi1* and Muriel Vayssade2*
1
Sorbonne Universités, Université de Technologie de Compiègne, CNRS, Unité Génie Enzymatique et Cellulaire, Centre de
Recherche Royallieu, CS 60319, Compiègne cedex 60203, France
2
Sorbonne Universités, Université de Technologie de Compiègne, CNRS, UMR 7338 Biomécanique et Bioingénierie, Centre de
Recherche Royallieu, CS 60319, Compiègne cedex 60203, France
3
Institut Polytechnique LaSalle Beauvais, Département STAI, rue Pierre Waguet, BP 30313, Beauvais 60026, France

Recent studies have pointed out the preventive role of beetroot extracts against cancers and their cytotoxic activity
on cancer cells. Among many different natural compounds, these extracts contained betanin and its stereoisomer
isobetanin, which belongs to the betalain group of highly bioavailable antioxidants. However, a precise identification
of the molecules responsible for this tumor-inhibitory effect was still required. We isolated a betanin/isobetanin
concentrate from fresh beetroots, corresponding to the highest purified betanin extract used for studying anticancer
activities of these molecules. The cytotoxicity of this betanin-enriched extract was then characterized on cancer
and normal cells and we highlighted the death signalling pathways involved. Betanin/isobetanin concentrate
significantly decreased cancer cell proliferation and viability. Particularly in MCF-7-treated cells, the expressions
of apoptosis-related proteins (Bad, TRAILR4, FAS, p53) were strongly increased and the mitochondrial membrane
potential was altered, demonstrating the involvement of both intrinsic and extrinsic apoptotic pathways.
Autophagosome vesicles in MCF-7-treated cells were observed, also suggesting autophagic cell death upon
betanin/isobetanin treatment. Importantly, the betanin-enriched extract had no obvious effect towards normal
cell lines. Our data bring new insight to consider the betanin/isobetanin mix as therapeutic anticancer
compound, alone or in combination with classical chemotherapeutic drugs, especially in functional p53 tumors.
Copyright © 2015 John Wiley & Sons, Ltd.

Keywords: betanin concentrate; isobetanin; betalains; anticancer activity; 2D and 3D cultures.

pigments (Kanner et al., 2001; Tesoriere et al., 2004,

INTRODUCTION
2013). In the red beetroot species, the predominant
betacyanins are betanin (Fig. 1A), which represents
With surgery and radiotherapy, chemotherapy is a major between 75% and 90% of the total pigments of red
treatment for cancer. Chemical agents act on cancer cells beetroot (Henry, 1996), and isobetanin, its C15
at different levels and trigger cell destruction via the stereoisomer. In addition to the powerful antioxidant
induction of an apoptotic program. However, classic properties of betalains, studies have pointed out their
chemotherapeutic agents often target rapidly dividing preventive role against cancers (Kapadia et al., 1996,
cancer and normal cells, leading to side effects for the 2003; Lechner et al., 2010). Nevertheless, these studies
patients (Rao et al., 2013). As the cancer incidence world- were based on animal diet with red food coloring (E162)
wide tends to increase, the development of new, safer and and other standard beetroot extracts, whose betanin
efficient anticancer agents is a major challenge. content does not exceed 1.2%. Antitumoral activity of
Certain plant-derived compounds exhibit anticancer betalains was also evaluated in vitro on several cancer
activities and have attracted interest in preventing and cell lines, but data obtained by the authors were disparate
treating cancers with over 60% of currently used anti- according to the cell lines and the cytotoxicity assays used
cancer agents derived from natural sources (Cragg and (Reddy et al., 2005; Sreekanth et al., 2007; Paluszczak
Newman, 2005; Gupta and Prakash, 2014). Beetroots et al., 2010). Moreover, the betanin content or the purity
contain both red (betacyanins) and yellow pigments degree of beetroot extracts used has never been precisely
(betaxanthins) known collectively as betalains, which con- defined. In order to dispel any ambiguity, the character-
stitute a class of highly bioavailable natural antioxidant ization of the betanin effect on tumor cells requires a
highly purified betanin concentrate with a precisely
* Correspondence to: Muriel Vayssade, Sorbonne Universités, Université known composition.
de Technologie de Compiègne, CNRS, UMR 7338 Biomécanique et During preclinical testing, in vitro analyses are per-
Bioingénierie, Centre de recherche Royallieu, CS 60319, 60203 Compiègne formed to select compounds with potential antitumor
cedex, France; Claire Rossi, Sorbonne Universités, Université de Technologie activity before in vivo studies in relevant animal models.
de Compiègne, CNRS, Unité Génie Enzymatique et Cellulaire, Centre de
Recherche Royallieu, CS 60319, 60203 Compiègne cedex, France.
Cell culture approach is therefore of the highest impor-
E-mail: [email protected] (Muriel Vayssade); [email protected] tance for the initial screening of molecules. However,
(Claire Rossi) cell lines used are conventionally cultured as monolayers

Received 23 April 2015

Copyright © 2015 John Wiley & Sons, Ltd. Revised 07 September 2015
 IN VITRO ANTITUMOR ACTIVITY OF BETACYANINS 1965

results showed that the betanin concentrate induces

cancer cell growth inhibition, associated with an
apoptotic cell death and an autophagic activity increase
in MCF-7 cells, but has no obvious effect towards
normal cells.

MATERIALS AND METHODS

Betanin extraction and purification. Grinded red beet-

roots (local market) were extracted into ethanol/water
(80/20 v/v) solvent at a solid/liquid ratio of 1/3 (g/mL)
for 1 h under continuous mechanical stirring. The solid
material was separated from the macerate by centrifuga-
tion at 12 000 g for 15 min at 4 °C followed by a filtration
on a polypropylene membrane filter (0.2 μm, Pall Corpo-
ration). The filtrate was concentrated by ethanol evapora-
tion under reduced pressure at 30 °C and then enriched
in betanin and isobetanin by adsorption chromatography
(Stintzing et al., 2002). A glass column was packed with
1 L of activated polymeric adsorbent (Amberlite XAD-
16, Sigma–Aldrich, France). The pH of beetroot extract
was adjusted to 3 with hydrochloric acid (HCl). 100 mL
of extract was passed through the resin column at 1 bed
volume (BV) per hour. After adsorption of the beta-
cyanins, the other extract constituents were eluted using
3 BV of deionized water (resistivity > 18.2 mΩ.cm 1,
MilliQ plus unit, Merk–Millipore) acidified to pH 3 at
2 BV/h. Betanin and isobetanin were eluted with ethanol
at 0.5 BV/h. The collected fraction was concentrated by
ethanol evaporation under vacuum at 30 °C and freeze-
dried. The obtained betanin/isobetanin (Bet./IsoBet.)-
enriched powder was stored at 80 °C.

Extract characterization

Betanin quantification. All betacyanin/betaxanthin

quantifications were performed using the multicompo-
nent photometric method described by J. H. von Elbe
(Von Elbe, 2001). The pigment concentration calcula-
Figure 1. (A) Structure of betanin. The chiral center (15C), which tions were based upon the absorptivity values: A1%
differentiates betanin (15S) and isobetanin (15R stereoisomer), is 1120 for betanin and isobetanin (at 538 nm) and 750
indicated by a star. (B) HPLC chromatogram of pigment-enriched for betaxanthins (at 477 nm).
beetroot extract monitored by UV-Visible absorption at 477 nm.
The retention times were attributed to betanin (13.05 min) and
isobetanin (14.51 min) by mass spectrometry analysis. (C) Positive Protein and total carbohydrate quantifications. Total
ion spray tandem mass spectrum of betanin. The daughter ion of
+
carbohydrate amount present in extracts was deter-
m/z 389 corresponding to positive aglycone ion [betanidin + H] mined using the DuBois method (DuBois et al., 1956).
was obtained by fragmentation of the parent ion of m/z 551
+
assigned to [betanin + H] using a collision energy of 15 eV.
Glucose was used as a standard. The protein quantifica-
tion was performed using the Lowry method (Lowry
et al., 1951). The absence of pigment inferences with
(2D culture) and do not mimic the phenotype of cancer tis- these two colorimetric methods was controlled.
sues because of the lack of cell–cell and cell–extracellular
matrix interactions. 3D models (cell spheroids or aggre- HPLC/ESI-MS analysis. The HPLC system (Infinity
gates) are obtained by preventing cells from attaching 1290, Agilent Technologies, France) was equipped with
to the culture substrata. Based on their ability to mimic diode array detector coupled with a Q-TOF micro hybrid
tissue-like structures more effectively than 2D cell cultures, quadrupole time of flight mass spectrometer (Agilent
the use of 3D in vitro models is believed to be a requisite 6538, Agilent technologies, France). HLPC analyses were
mean into anticancer drug development (Breslin and performed using an analytical scale (150 × 2.1 mm i.d.)
O’Driscoll, 2013). hypersil gold C18 reversed phase column with a particle
This study was carried out to clarify the antitumor size of 3 μm (Thermo Scientific, France). Eluents A and
effect of a betanin-enriched beetroot dried extract. B (LC-MS grade) consisted of 100% water with 0.2%
We used both 2D and 3D cultures to assess the cytotoxic (v/v) formic acid and 100% acetonitrile, respectively.
activity of betanin on cancer and normal cells. Our Betalains separation was achieved within 40 min at 20°C
Copyright © 2015 John Wiley & Sons, Ltd. Phytother. Res. 29: 1964–1973 (2015)
1966 L. NOWACKI ET AL.

at a flow rate of 0.3 mL/min. The elution profile was and 25 μg propidium iodide (PI, Sigma–Aldrich) and in-
0–3 min 100% A, 3–21 min 0–13% B in A (linear gradient), cubated for 15 min protected from light. The stained sam-
21–24 min 13% B, 24–30 min 13–50% B (linear gradient), ples were analysed in Gallios flow cytometer (Beckman
30–33 min 50–100% B, 33–40 min 100% B. Detection Coulter, France). Histograms were analysed using
wavelengths were 477 nm (betaxanthins and betacyanins) Wincycle software (PHOENIX flow systems, USA).
and 538 nm (betacyanins only). Positive ion electrospray
mass spectra were acquired by scan mode, consisted of
scanning from m/z 50 to 1700 at electrospray voltage
3800 V, fragmentor voltage 140 V. Nitrogen was nebulized Apoptosis detection and quantification. To assess early
at 12.0 L/min under 45.0 psi and heated at 350 °C. Betanin stage of apoptosis, cells cultured in 2D and 3D condi-
and isobetanin structural identities were confirmed by tions were treated with 30 μM Bet./IsoBet. mix for 24 h
tandem mass (relative collision energy of 15 eV) spectros- and labelled with Annexin-V-FITC (Beckman Coulter).
copy and NMR studies. The 1H NMR spectrum of betanin For each cell line, a positive control for apoptosis was
was recorded in D2O (data not shown) and corresponds to tested (cells treated with 1 μM staurosporine). The
the previously published data (Stintzing et al., 2004). cytometer was adjusted according to the Annexin-V-
positive cells of these controls. According to the manu-
facturer’s instructions, cells were washed with PBS and
suspended in the binding buffer at 5.105–5.106 cells/mL.
Cell culture and treatment. The mouse melanoma cell line
One microliter of Annexin-V-FITC solution and 5 μL
B16F10 was a generous gift from Dr L. Larue (Institut
of PI solution were added to 100 μL of cell suspension.
Curie, France). Human breast cancer lines (MCF-7
Cells were incubated for 15 min on ice in the dark, and
ATCC® HTB-22, MDA-MB-231 ATCC® HTB-26),
then diluted by adding 400 μL of binding solution. The
human colorectal cells (HT-29 ATCC® HTB-38) and nor-
Annexin-V-positive/IP-negative cell populations were
mal human fibroblasts (MRC-5 ATCC® CCL-171) were
detected with a Gallios flow cytometer and were repre-
obtained from the American Type Culture Collection.
sentative of apoptotic cells.
Human umbilical vein endothelial cells (HUVEC) were
To detect DNA fragmentation, an Apostain binding
purchased from PromoCell.
assay (AbCys) was used. MCF-7 and HUVEC cells
Cells were cultured as monolayers in RPMI 1640
were treated with Bet./IsoBet. (30 μM) or staurosporine
(B16F10), MEM (MCF-7, MDA-MB-231, MRC-5),
(1 μM) for 48 h, fixed with ice-cold methanol 80% (v/v)
DMEM (HT-29), or M199 (HUVEC). All cell culture me-
for 30 min, washed and suspended in 250 μL formamide
dia (Gibco, Invitrogen, France) were supplemented with
for 10 min at 75 °C and for 5 min at room temperature
10% fetal bovine serum (FBS, Gibco, Invitrogen), 2 mM
(RT). Cells were then incubated for 45 min at RT with
L-glutamine (Gibco, Invitrogen), penicillin (100 μg/mL,
a mouse monoclonal antibody to single-stranded DNA
Gibco, Invitrogen) and streptomycin (100 μg/mL, Gibco,
(ssDNA, diluted 1:10 in PBS containing 5% FBS).
Invitrogen). All cell lines were maintained at 37 °C in an
Bound specific antibodies were revealed by incubation
air atmosphere of 10% CO2 (B16F10) or 5% CO2
with Cy3-conjugated antimouse antibody (diluted 1:200
(MCF-7, MDA-MB-231, HT-29, HUVEC, MRC-5).
in PBS containing 1% milk, Jackson ImmunoResearch)
Nunclon polystyrene plates (tPS) were used for 2D
for 30 min at RT. Each sample was counterstained with
cultures. Polyhydroxyethylmethacrylate (polyHEMA,
DAPI (4′,6′-diamidino-2-phenylindole, 1 μg/mL, Sigma–
Sigma–Aldrich) coated polystyrene plates were pre-
Aldrich). The cells were then cytocentrifuged (Shandon
pared as previously described (Velzenberger et al.,
Cytospin) and each slide was mounted in mowiol and ex-
2008) and used for 3D cultures. Cells were seeded at
amined by epifluorescence microscopy (Leica DMI6000).
10,000 cells/cm2 without (controls) or with increasing
Ten fields were randomly chosen from each slide, and
concentrations of Bet./IsoBet. (from 10 to 40 μM), and
the percentage of apoptotic cells in each picture was
cultured for 24–48 h.
calculated as the ratio of the labelling cells to the total
number of cells.

Proliferation assay and cell cycle analysis. In 2D cul-

tures, cells were removed by incubation for 5 min at
37 °C in 0.25% trypsin + 1 mM EDTA solution (Gibco, Apoptotic signalling pathway analysis. Human apopto-
Invitrogen). In 3D cultures, aggregates were collected sis arrays (Proteome Profiler Array, R&D Systems)
by centrifugation (5 min, 200 g) and dissociated using were used to simultaneously detect the relative levels
trypsin-EDTA solution. Trypsin reaction was stopped of expression of 35 apoptosis-related proteins. MCF-7,
by adding half a volume of FBS, and cells were counted MDA-MB-231 and HUVEC cells were treated with
in a Malassez hemocytometer using trypan blue exclu- Bet./IsoBet.-enriched extract for 48 h, rinsed with PBS
sion assay. The proliferative index was estimated as and lysed for 30 min at 4 °C using lysis buffer provided
the ratio between viable cell counts 48 h post-seeding by the manufacturer. Lysates were then centrifuged at
and number of seeded cells. Each proliferative index 14,000 g for 5 min and supernatants were harvested.
was normalised to this one obtained for control cells Protein concentrations for each sample were deter-
and data were expressed as percentages. mined using the Bradford method (Coomassie Protein
For cell cycle analysis, harvested cells were washed Assay Reagent, Pierce, Interchim). 400 μg of cell lysates
twice in PBS containing 5 mM EDTA, and fixed for were incubated on each array overnight at 4 °C. Arrays
45 min at 4 °C in 1 mL 75% ethanol in PBS with 5 mM were incubated for 1 h at RT with a reconstituted detec-
EDTA. Cell lines were washed twice and suspended in tion antibody cocktail, and then with a streptavidin-
PBS, 5 mM EDTA containing 0.1% TritonX100 (Sigma– HRP solution for 30 min. Arrays were revealed with
Aldrich) mixed with 40 μg RNase A (Sigma–Aldrich) a Chemireagent Mix using ChemiDoc (Biorad), and
Copyright © 2015 John Wiley & Sons, Ltd. Phytother. Res. 29: 1964–1973 (2015)
 IN VITRO ANTITUMOR ACTIVITY OF BETACYANINS 1967

each spot was quantified with Quantity One software reveals that they are totally eliminated at this purification
(Biorad). step, no residual complex or simple carbohydrates could
To detect change in the mitochondrial membrane po- be detected in the final betanin concentrate. As betanin
tential, a MitoCapture dye (BioVision) was used. MCF- and isobetanin differ only by the absolute configuration
7 cells were treated with Bet./IsoBet. mix (15 and of their C15 chiral center, we did not distinguish them
30 μM) for 24 h, harvested and then resuspended in for determination of the total pigment amount in the
1 mL diluted MitoCapture reagent provided by the purified extract. The total Bet./IsoBet. content in the
manufacturer. Cells were incubated for 20 min at 37 °C dried concentration was evaluated at 80% by photometric
in a 5% CO2 incubator. After centrifugation (500 g), the quantification. The remaining 20% was attributed to re-
supernatant was discarded, the cells were suspended in sidual proteins by total protein content determination.
1 mL of prewarmed incubation buffer and analysed by Based on HPLC peak area values, the betacyanin com-
epifluorescence microscopy. position was estimated at 64% of betanin and 36% of
isobetanin. In the rest of the article, all the concentrations
given correspond to the total Bet./IsoBet. concentrations
Autophagy assay. Autophagy was detected using the applied to the cells.
Autophagy Detection Kit (Abcam). Briefly, MCF-7
cells were grown 24 h on glass slides, and then were
treated with Bet./IsoBet. (15 and 30 μM) or with an au- Effect of Bet./IsoBet. on cell proliferation and
tophagy inducer, rapamycin (500 nM provided by the morphology
manufacturer) for 24 h. Medium was removed and cells
were washed with a buffer supplied by the manufac- We used both cancer (B16F10, MCF-7, MDA-MB-231,
turer. Cells were incubated with the dye (diluted at HT-29) and normal (HUVEC, MRC-5) cells to assess
1:500 in assay buffer) and DAPI for 20 min at 37 °C. the biological effect of betanin extract. All cells were
Cells were carefully washed and slides were examined cultured as monolayer (2D) on tPS and as aggregates
by epifluorescence microscopy. (3D) on antiadhesive polyHEMA substratum, and then
treated with betanin-enriched extract for 48 h.
As shown in Fig. 2A, Bet./IsoBet. treatment signifi-
cantly decreased the proliferation of B16F10 and
Statistical analysis. All statistical evaluations were per- MCF-7 cells on both surfaces. The half maximal inhibi-
formed using GraphPad InStat software. Continuous tory concentration (IC50) was determined for each cell
variables are expressed as means ± standard deviation. line and was about 25 μM for B16F10 and MCF-7 cells.
As most of the continuous values measured had a No viable cells were observed with 40 μM Bet./IsoBet.
non-Gaussian distribution, nonparametric Dunn and (data not shown). MDA-MB-231 and HUVEC cells
Kruskal–Wallis tests were used for comparisons. A were also sensitive to betanin-enriched extract, but to
value of p < 0.05 was taken as significant. a lower extent (IC50 value was 35 μM) (Fig. 2B). The
cell viabilities measured 48 h post-treatment with 40 μM
Bet./IsoBet. decreased weakly: 35.6% ± 17.2 (2D) and
52.0% ± 19.2 (3D) for MDA-MB-231 and 75.2% ± 26.9
RESULTS (2D) and 71.0% ± 27.4 (3D) for HUVEC cells (data not
shown). For the same concentration range, HT-29 and
Betanin/isobetanin concentrate characterization MRC-5 cells were not sensitive to betanin concentrate:
at 40 μM, cell proliferation was not significantly de-
Fig. 1B shows the chromatogram corresponding to the creased compared with controls (Fig. 2C). Viabilities
pigment-enriched red beetroot dried extract obtained of HT-29 cells treated with 40 μM Bet./IsoBet. for
after purification by adsorption chromatography. The 48 h were 95.2% ± 5.3 (2D) and 96.6% ± 5.2 (3D) (data
HPLC profile monitored at 477 nm combined with not shown). For MRC-5 cells, viabilities remained high
electrospray mass spectrometry in positive mode allows with Bet./IsoBet. treatment: 87.7% ± 9.3 (2D) and
the identification of the betaxanthin and betacyanin 76.8% ± 12.8 (3D).
molecules contained in this final extract (Stintzing et al., Cell morphologies were examined (Fig. 2D): MCF-7
2004). Only two betalains were detected and identified and HUVEC control cells cultured on tPS spread well
as betanin (tR = 13.05 min, [M + H]+ = 551) and isobetanin, and colonized the culture dishes, while cells on
the 15R stereoisomer of betanin (tR = 14.51 min, polyHEMA had a rounded aggregated morphology.
[M + H]+ = 551) (Kujala et al., 2002; Stintzing et al., Betanin-enriched extract reduced MCF-7 cell spread-
2004). The betanin and isobetanin structures were ing on tPS and MCF-7 cell aggregation on polyHEMA,
confirmed by mass tandem spectroscopy. The positive whereas a moderate effect of Bet./IsoBet. concentrate
ion spray mass spectrum (Fig. 1C) shows the daughter was observed on HUVEC cells cultured in 2D and 3D
ion produced by fragmentation of the parent ion of conditions. No morphological change was observed
m/z of 551 [M + H]+ = 551 assigned to betanin or iso- for the HT-29 and MRC-5 cells treated with Bet./
betanin. The fragment ion at the mass charge (m/z) of IsoBet. (data not shown).
389 indicated that this ion is obtained by glucose loss Thus, we identified three groups of cell lines with dif-
and corresponds to the protonated aglycones [betanidin ferent sensitivities to Bet./IsoBet.: B16F10 and MCF-7
+ H]+ or [isobetanidin + H]+ (Castellanos-Santiago and as highly sensitive (no viable cells with 40 μM Bet./
Yahia, 2008). IsoBet.); MDA-MB-231 and HUVEC as weakly sensi-
All the betaxanthins were eluted during beetroot ex- tive (cell viabilities from 35% to 70% with 40 μM Bet./
tract purification on polymeric resin while the betacyanins IsoBet.); HT-29 and MRC-5 as resistant (cell viabilities
remained absorbed. Total carbohydrate quantification about 90% with 40 μM Bet./IsoBet.).
Copyright © 2015 John Wiley & Sons, Ltd. Phytother. Res. 29: 1964–1973 (2015)
1968 L. NOWACKI ET AL.

Figure 2. Effect of Bet./IsoBet. concentrate on cell proliferation and morphology. (A, B and C) Relative proliferation index of cancer and
non-tumorigenic cells treated with Bet./IsoBet. Results are means ± SD of three experiments each performed in triplicate. Asterisks indicate
* ** ***
significant differences in cell proliferation index of treated and control cells (Dunn test): p < 0.05; p < 0.01; p < 0.001. (D) Morphology
of MCF-7 and HUVEC cells grown for 48 h on tPS (2D) or polyHEMA (3D) with 40 μM Bet/IsoBet. Cells were examined under an Olympus
CKX41 inverted microscope. Scale bar = 100 μm.

Effect of Bet./IsoBet. concentrate on cell cycle distribution Table 1. Cell cycle distribution of cells grown on tPS or polyHEMA
in the presence or absence (control) of Bet./IsoBet. extract
To better understand Bet./IsoBet. concentrate effect, we
focused on three human cell lines that represent three Bet./IsoBet.
different and specific phenotypes: p53 wild-type cancer Control (30 μM)
cell line (MCF-7), p53-mutated cancer cell line (MDA- MCF7 2D G1 54.7% ± 1.6 49.1% ± 4.2*
MB-231), and normal cells (HUVEC). The cell cycle S 26.4% ± 2.8 34.2% ± 9.8
progression was studied 48 h post-seeding without G2/M 18.8% ± 4.1 16.7% ± 5.5
(control) or with 30 μM Bet./IsoBet. (Table 1). For each 3D G1 77.0% ± 2.6 68.2% ± 1.8**
cell line, the cell cycles of control cultures on tPS and S 12.4% ± 3.2 17.8% ± 2.2
polyHEMA were significantly different: there were G2/M 10.6% ± 0.7 14.0% ± 0.7***
more cells in G1 phase and fewer in the S and G2/M MDA-MB231 2D G1 62.3% ± 1.9 67.4% ± 1.2
phases in 3D cultures. S 24.8% ± 1.6 19.8% ± 0.9*
Betanin/isobetanin-enriched extract significantly de- G2/M 12.9% ± 0.8 13.5% ± 2.2
creased the percentage of MCF-7 cells in G1 phase 3D G1 74.6% ± 10.0 72.4% ± 10.5
(2D and 3D conditions) and increased percentage of S 22.9% ± 11.6 21.7% ± 13.2
cells in S phase (2D and 3D). For MDA-MB-231 and G2/M 2.5% ± 2.9 5.8% ± 2.7
HUVEC cell lines, Bet./IsoBet. concentrate decreased HUVEC 2D G1 51.2% ± 2.4 58.4% ± 1.9
cell number in S phase (2D). No obvious effect of S 40.0% ± 1.5 30.2% ± 2.5*
Bet./IsoBet. was observed on the 3D MDA-MB-231 G2/M 8.8% ± 1.4 11.7% ± 1.6*
cell cycle repartition. Betanin/isobetanin treatment de- 3D G1 90.1% ± 1.4 80.9% ± 4.5*
creased the 3D HUVEC cell number in G1 phase and S 5.5% ± 1.2 13.1% ± 4.6
increased the percentage of cells in S and G2/M phases. G2/M 4.3% ± 1.3 5.9% ± 1.2

Results are means ± SD of three experiments each performed in

Characterization of MCF-7 cell death induced by Bet./ triplicate. Asterisks indicate significant differences in cell cycle
IsoBet. concentrate distribution of treated and control cells (Dunn test):
*p < 0.05;
As MCF-7 cell viabilities strongly decreased during **p < 0.01;
Bet./IsoBet. treatment, we analysed the nature of cell ***p < 0.001.

Copyright © 2015 John Wiley & Sons, Ltd. Phytother. Res. 29: 1964–1973 (2015)
 IN VITRO ANTITUMOR ACTIVITY OF BETACYANINS 1969

Table 2. Annexin-V labelling of MCF-7, MDA-MB231 and HUVEC cells treated with Bet./IsoBet. or staurosporine

Control Bet./IsoBet. (30 μM) Staurosporine (1 μM)

MCF7 2D 0.59% ± 1.41 23.24% ± 2.72 * 18.75% ± 11.19 *

3D 11.0% ± 6.91 17.29% ± 6.25 29.41 % ± 14.51 *
MDA-MB231 2D 0.29% ± 0.04 0.71% ± 0.15 13.45% ± 2.52 ***
3D 1.94% ± 0.5 4.25 % ± 1.40* 1.79 % ± 0.68
HUVEC 2D 1.44% ± 0.72 4.05% ± 0.86 56.99% ± 10.37 ***
3D 25.06% ± 5.62 38.72% ± 7.8 18.13% ± 7.81

Results are means ± SD of two experiments each performed in triplicate. Asterisks indicate significant differences in Annexin-V labelling of
treated and control cells (Dunn test):
*p < 0.05;
***p < 0.001.

death induced by the pigments. First, we checked apopto- cultured in 2D or 3D with Bet./IsoBet. were not strongly
sis induction and quantification using Annexin-V-FITC labelled with Annexin-V-FITC probe (Table 2). Interest-
labelling and cell analysis by flow cytometry (Table 2, ingly, HUVEC cells were found to be non-apoptotic when
Fig. 3A). Our results showed that Bet./IsoBet. treat- cultured in 2D and treated with Bet./IsoBet., whereas
ment significantly increased the percentage of Annexin- staurosporine treatment induced significantly apoptosis
V-positive/PI-negative MCF-7 cells, as the positive con- (Table 2, Fig. 3A). Anchorage-independent culture of
trol staurosporine. MCF-7 3D cell culture promoted HUVEC cells triggered Annexin-V labelling and the
phosphatidylserine externalization, and the percentage percentage of apoptotic cells was slightly increased with
of Annexin-V-positive cells was found to be increased Bet./IsoBet. treatment. To confirm apoptosis induction
with Bet./IsoBet. or staurosporine. MDA-MB-231 cells in MCF-7 cells treated with Bet./IsoBet. concentrate, we

Figure 3. Apoptosis detection in cells treated with Bet./IsoBet. concentrate. (A) Phosphatidylserine flipping in MCF-7 and HUVEC cells treated
with Bet./IsoBet. (30 μM) or staurosporine (1 μM) assessed by Annexin-V binding and flow cytometry. Flow cytometry profile represents
Annexin-V-FITC staining in x axis and propidium iodide in y axis. (B) Apostain labelling of MCF-7 and HUVEC cells treated with Bet./IsoBet.
(30 μM) or staurosporine (1 μM). Scale bar = 100 μm.

Copyright © 2015 John Wiley & Sons, Ltd. Phytother. Res. 29: 1964–1973 (2015)
1970 L. NOWACKI ET AL.

Figure 4. Apoptosis characterization. (A) Representative proteome arrays of MCF-7 cells obtained without or with Bet./IsoBet. treatment
(30 μM) for 48 h. Major differences in protein expression levels are indicated with black arrows. (B) Quantitative analysis of protein expressions
in MCF-7, MDA-MB231 and HUVEC cells cultured in 2D. Data represent the ratio of the expression level obtained after treatment normalised to
*
the expression level in untreated cells. The observed increase corresponds to the level of pro-caspase 3 protein. As MCF-7 cell line is known to
be deficient in caspase 3 activity, we did not comment on this result.

Figure 5. Mitochondrial potential analysis in MCF-7 cells treated with Bet./IsoBet., staurosporine or curcumin. In untreated and live cells, the dye
accumulates and aggregates in the mitochondria giving off a bright red fluorescence. With altered mitochondrial transmembrane potential, the
dye remains in cytoplasm in its monomer form fluorescing green. Scale bar = 100 μm.

Copyright © 2015 John Wiley & Sons, Ltd. Phytother. Res. 29: 1964–1973 (2015)
 IN VITRO ANTITUMOR ACTIVITY OF BETACYANINS 1971

Figure 6. Vesicle production in MCF-7 cells treated with Bet./IsoBet. concentrate. Untreated cells do not display green staining while
betanin- or rapamycin-treated cells display intense punctuate structures. Scale bar = 100 μm.

analysed DNA fragmentation using an antibody recogniz- to study new anticancer agents because they are believed
ing single-stranded DNA. After culture for 48 h with Bet./ to bridge the gap between in vitro 2D assessment and
IsoBet. (30 μM) or staurosporine (1 μM), about 80% of animal models (Breslin and O’Driscoll, 2013). Cell cyto-
MCF-7 cells were labelled (Fig. 3B). Only 1.84% of toxicity analysis revealed that p53 wild-type cancer cell
HUVEC cells treated with betanin-enriched extract were lines (B16F10, MCF-7) were highly sensitive to 40 μM
found to be positive (Fig. 3B), whereas 29.03% were Bet./IsoBet. mix (assessed by proliferation inhibition
labelled with the positive control staurosporine. and low cell viabilities), whereas cancer cell lines ex-
Proteome profiler arrays allowed us to detect the ex- pressing mutated p53 were less (MDA-MB-231) or
pression levels of apoptosis-related proteins in MCF-7, not sensitive (HT-29) for the same concentration range.
MDA-MB-231 and HUVEC cells treated with 30 μM The betanin-enriched extract effect was similar whatever
Bet./IsoBet. Typical results of obtained arrays are the culture conditions, 2D or 3D. This shows that Bet./
shown in Fig. 4A. The most important differences in IsoBet. concentrate inhibits aggregated cancer cell
protein expressions were found in sensitive MCF-7 cells: proliferation, a cell structure known as more resistant
we observed a strong induction of Bad, TRAILR1/DR4, to apoptosis induction (Hirschhaeuser et al., 2010). In
Fas/TNFRSF6/CD95 and phospho-p53 (S392) proteins this way, B16F10 and MDA-MB-231 are metastatic
in betanin-treated cells compared with control cells cell lines, resistant to anoikis when cultured under
(Fig. 4B, Table I supplementary data). No significant anchorage-independent conditions because of the acti-
variation in protein expression was observed in MDA- vated ERK signalling pathway (Goundiam et al., 2010;
MB-231 cells. In HUVEC cells, Bet./IsoBet. treatment Fukazawa et al., 2002). For these reasons, identifying
induced a moderate overexpression of p21 and p27 pro- an efficient molecule on anoikis-resistant cancer cells
teins (Table I supplementary data). is a very hopefully result. Cell proliferation inhibition
The effect of Bet./IsoBet. on the mitochondrial mem- by Bet./IsoBet. was confirmed with affected cell cycle.
brane potential was checked in MCF-7 cells using In MCF-7 cells, Bet./IsoBet. extract decreased the G1
MitoCapture dye (Fig. 5). Our data showed an increase cell number and promoted S phase increase, a behavior
in green fluorescence in Bet./IsoBet.-treated cells, in a previously described in MCF-7 cells treated with resvera-
same extent than positive control (curcumin 50 μM). trol or riproximin (Joe et al., 2002; Pervaiz et al., 2015).
To assess another cell death, autophagy, MCF-7 cells Betanin/isobetanin extract may modify the expression of
treated with Bet./IsoBet. or rapamycin (an autophagy cell cycle regulators, such as resveratrol and riproximin,
inducer) were analysed for lysosomal vacuole formation which downregulate cyclin A2 and cyclin B1 levels in
(Fig. 6). We observed positive labelling in MCF-7- MCF-7 cells. 2D cultured MDA-MB-231 cells were
treated cells, indicating an autophagic activity increase. blocked in G1 phase during treatment, but the molecules
did not significantly affect the cell cycle repartition when
cells were cultured as aggregates, probably because of
the G1 arrest induced by the culture configuration.
DISCUSSION Kapadia et al. (2011, 2013) have previously evaluated
the cytotoxic effect of a red beetroot extract in MCF-7 cell
We described that a highly Bet./IsoBet.-enriched con- line and the IC50 value they obtained was 600 μM (after
centrate produced from red beetroots inhibits cancer 72 h of exposure). Reddy et al. (2005) also observed a
cell proliferation and induces MCF-7 cell death but has growth inhibition of MCF-7 cells treated with a betanin
no obvious effect towards normal cells. concentrate for 48 h (IC50 value was 294 μM). Our data
Betanin and isobetanin are the most predominant clarify these previous studies. The betanin purification pro-
betalains in red beetroot. We were able to obtain a cess we applied to the crude beetroot extract allowed us to
pigment-enriched dried extract from beetroots that con- obtain a significant MCF-7 growth inhibition associated
tain only Bet./IsoBet. as coloring agents. This concentrate with cell death for very low concentrations (below 40 μM).
consists of 80% of Bet./IsoBet. mixture (of which betanin As MCF-7 cell viability strongly decreased when
accounts for 64% and isobetanin for 36%) and corre- treated with Bet./IsoBet. mix, the nature of cell death
sponds, to our knowledge, to the highest purified betanin was analysed. Using different methods, we showed that
extract used for studying its anticancer activities. To betanin treatment induced apoptosis in 2D MCF-7 cells.
better evaluate the cytotoxicity of the betanin-enriched The expressions of apoptosis-related proteins (Bad,
extract, we used two different culture conditions: classical TRAILR4, FAS, phosphorylated p53) were strongly
monolayer culture (2D) and cells cultured as aggregates increased and the mitochondrial membrane potential
on antiadhesive substratum (3D) (Velzenberger et al., was clearly altered. Taken together, all these data sug-
2008). These 3D culture models are now commonly used gest both mitochondrial and death-receptor pathway
Copyright © 2015 John Wiley & Sons, Ltd. Phytother. Res. 29: 1964–1973 (2015)
1972 L. NOWACKI ET AL.

involvement, and a p53-dependent response in MCF-7 believed to be a realistic prospect for cancer therapy
cells treated with the Bet./IsoBet. concentrate. Betanin (Bincoletto et al., 2013).
extract (30 μM) only induced a moderate apoptosis in To fully characterize the biological effect of betanin, we
aggregated MCF-7 cells, a 3D configuration often pro- also evaluated its cytotoxicity on normal cells. Impor-
moting cell death resistance. Recently, Gong et al. (2015) tantly, normal cell lines (HUVEC, MRC-5) were found
have compared the cytotoxicity effect of the doxorubicin to be weakly or not sensitive to betanin C15 stereoisomer
on MCF-7 cells cultured as monolayer or spheroids. Their mix for the same concentration range used on cancer
data show that 3D cultured MCF-7 cells were less sensi- cells. Particularly, HUVEC endothelial cells slowed down
tive than their 2D counterpart. According to our data on their proliferation rate upon betanin exposure, remaining
cell proliferation and viability, we would probably observe viable for higher concentrations than MCF-7 cells. When
an apoptosis induction on 3D MCF-7 cells with higher cultured as monolayer and treated with the Bet./IsoBet.
concentrations of Bet./IsoBet. mix. extract, HUVEC cells accumulated in the G1 phase and
Intrinsic pathway activation upon betanin exposure the percentage of cells in the DNA replicative phase
was previously observed by Sreekanth et al. (2007) in decreased, confirming the cell proliferation data. 3D
K562 cells with mitochondrial membrane potential de- betanin-treated HUVEC cells progressively accumulated
crease and cytochrome c release. A significant activa- in the S and G2/M phases. Moreover the Bet./IsoBet.
tion of caspase 9 and effectors (caspases 3 and 7) was treatment triggered an overexpression of p21 and p27
also reported in lung cancer cells treated with 400 μM proteins, cyclin-dependent kinase inhibitors involved in
betanin for 48 h (Zhang et al., 2013). Our results confirm cell cycle arrest. However, the HUVEC cell proliferation
these previous studies about the mitochondrial pathway inhibition by the betanin concentrate was not associated
involvement. Moreover, we show that Bet./IsoBet. mix in- with apoptosis or autophagy induction: this suggests that
duces up-regulation of death receptors and p53 activa- undesired side effects could be limited in therapies inte-
tion, as curcumin does on MCF-7 cells (Choudhuri et al., grating these molecules. Indeed, the preliminary data
2002; Mohankumar et al., 2014). P53 is a well-known im- we performed on the non-tumorigenic MCF-12F cell
portant actor in response to cellular stresses (Goh et al., line confirmed the harmless interaction of the mole-
2011). The mutated p53 status in MDA-MB-231 and cules on normal epithelial breast cells (data not shown).
HT-29 cells (Bartek et al., 1990; Rodrigues et al., 1990) In conclusion, we reported that a purified Bet./IsoBet.
could explain in part the weak betanin cytotoxicity for concentrate produced from red beetroots is cytotoxic at
these cell lines. Particularly, HT-29 cells express a mutant low concentrations for cancer cells expressing functional
p53 protein that gains function, promoting cell prolifera- p53 but have no or moderate effect on normal cells,
tion and chemoresistance (van Oijen and Slootweg, suggesting limited in vivo side effects. In p53 wild-type
2000). In this way, Arafa et al. (2013) have recently com- cancer cells, Bet./IsoBet. extract induces apoptosis (via
pared HT-29 and MDA-MB-231 sensitivity with novel extrinsic and intrinsic activation pathways) and autoph-
quinoline-based compounds. They found that the HT-29 agy. Moreover, betanin-enriched extract inhibits some
cell line was more refractory to the cytotoxic activity of p53-mutated cancer cell growth, without inducing
most compounds than MDA-MB-231 cells. In our work, apoptosis. Our data bring new insight to consider
HT-29 cells were also resistant to Bet./IsoBet. effect, Bet./IsoBet. as therapeutic anticancer molecules alone
whereas molecules slowed down MDA-MB-231 cell or in combination with classical chemotherapeutic drugs,
proliferation without inducing apoptosis. Therefore, low especially in functional p53 tumors.
concentrations of betanin-enriched extract appear to be
efficient on a wild-type p53 cancer cells to induce cell
death (MCF-7) and inhibit a p53-mutated cancer cell pro- Acknowledgements
liferation (MDA-MB-231), but have no effect on another
We thank the European Union (co-funding of equipment within the
p53-mutated cell line (HT-29). CPER 2007–2013 and FEDER) and the Conseil Régional of Picardie
As some natural-derived compounds have already for financial support (BetOX project). The English text was edited
been shown to induce cell death via an autophagic process by Kayla Belanger.
(Zhang et al., 2012), we investigated and revealed lyso-
somal vacuole formation in MCF-7 cells upon betanin
treatment. We are the first to describe that Bet./IsoBet. Conflict of Interest
mix also induces an autophagic response in cells. The
identification of molecules that target autophagy is The authors have no conflict of interest to disclose.

REFERENCES
Arafa RK, Hegazy GH, Piazza GA, Abadi AH. 2013. Synthesis Castellanos-Santiago E, Yahia EM. 2008. Identification and quantifica-
and in vitro antiproliferative effect of novel quinoline- tion of betalains from the fruits of 10 mexican prickly pear cultivars
based potential anticancer agents. Eur J Med Chem 63: by high-performance liquid chromatography and electrospray ion-
826–832. ization mass spectrometry. J Agric Chem 56: 5758–5764.
Bartek J, Iggo R, Gannon J, Lane DP. 1990. Genetic and immuno- Choudhuri T, Pal S, Agwarwal ML, Das T, Sa G. 2002. Curcumin
chemical analysis of mutant p53 in human breast cancer cell induces apoptosis in human breast cancer cells through p53-
lines. Oncogene 5: 893–899. dependent Bax induction. FEBS Lett 512: 334–340.
Bincoletto C, Bechara A, Pereira GJ, et al. 2013. Interplay between Cragg G, Newman DJ. 2005. Plants as a source of anti-cancer
apoptosis and autophagy, a challenging puzzle: new perspec- agents. J Ethnopharmacol 100: 72–79.
tives on antitumor chemotherapies. Chem Biol Interact 206: DuBois M, Gilles KA, Hamilton JK, Rebers PA, Smith F. 1956.
279–288. Colorimetric method for determination of sugars and related
Breslin S, O’Driscoll L. 2013. Three-dimensional cell culture: substances. Anal Chem 28: 350–356.
the missing link in drug discovery. Drug Discov Today 18: Fukazawa H, Noguchi K, Murakami Y, Uehara Y. 2002. Mitogen-acti-
240–249. vated protein/extracellular signal-regulated kinase kinase (MEK)

Copyright © 2015 John Wiley & Sons, Ltd. Phytother. Res. 29: 1964–1973 (2015)
 IN VITRO ANTITUMOR ACTIVITY OF BETACYANINS 1973

inhibitors restore anoikis sensitivity in human breast cancer cell gene expression in MCF7 breast cancer cells. Toxicol Lett
lines with a constitutively activated extracellular-regulated 192: 119–125.
kinase (ERK) pathway. Mol Cancer Ther 1: 303–309. Pervaiz A, Zepp M, Adwan H, Berger MR. 2015. Riproximin modu-
Goh AM, Coffill CR, Lane DP. 2011. The role of mutant p53 in lates multiple signaling cascades leading to cytostatic and
human cancer. J Pathol 223: 116–126. apoptotic effects in human breast cancer cells. J Cancer Res
Gong X, Lin C, Cheng J, et al. 2015. Generation of multicellular Clin Oncol doi: 10.1007/s00432-015-2013-3.
tumor spheroids with microwell-based agarose scaffolds for Rao B, Lain S, Thompson AM. 2013. P53-based cyclotherapy:
drug testing. PLoS One 10(6): 1–18. doi: 10.1371/journal. exploiting the ’guardian of the genome’ to protect normal cells
pone.0130348. from cytotoxic therapy. Br J Cancer 109: 2954–2958.
Goundiam O, Nagel MD, Vayssade M. 2010. Growth and survival Reddy KM, Alexander-Lindo RL, Nair MG. 2005. Relative inhibition
signalling in B16F10 melanoma cells in 3D culture. Cell Biol of lipid peroxidation, cyclooxygenase enzymes, and human
Int 34: 385–391. tumor cell proliferation by natural food colors. J Agric Food
Gupta C, Prakash D. 2014. Phytonutrients as therapeutic agents. Chem 53: 9268–9273.
J Complementary Integr Med 11: 151–169. Rodrigues NR, Rowan A, Smith ME, et al. 1990. p53 mutations in
Henry BS. 1996. Natural Food Colours. In Natural Food Colorants; colorectal cancer. Proc Natl Acad Sci U S A 87: 7555–7559.
Hendry GA, Houghton J. Eds.; Chapman & Hall: London, Sreekanth D, Arunasree MK, Roy KR, Reddy C, Reddy GV,
UK;40–79. Reddanna P. 2007. Betanin a betacyanin pigment purified from
Hirschhaeuser F, Menne H, Dittfeld C, West J, Mueller-Klieser W, fruits of Opunita ficus-indica induces apoptosis in human
Kunz-Schughart LA. 2010. Multicellular tumor spheroids: an chronic myeloid leukemia cell line K562. Phytomedicine 14:
underestimated tool is catching up again. J Biotechnol 148: 3–15. 739–746.
Joe AK, Liu H, Suzui M, Vural ME, Xiao D, Weinstein IB. 2002. Stintzing FC, Schieber A, Carle R. 2002. Identification of
Resveratrol induces growth inhibition, S-phase arrest, apopto- betalains from yellow beet (Beta vulgaris L.) and cactus pear
sis, and changes in biomarker expression in several human [Opuntia ficus-indica (L.) Mill.] by high-performance liquid
cancer cell lines. Clin Cancer Res 8: 893–903. chromatography-electrospray ionization mass spectrometry.
Kanner J, Harel S, Granit R. 2001. Betalains-a new class of dietary J Agric Food Chem 50: 2302–2307.
cationized antioxidants. J Agric Food Chem 49: 5178–5185. Stintzing FC, Conrad J, Klaiber I, Beifuss U, Carle R. 2004.
Kapadia GJ, Tokuda H, Konoshima T, Nishino H. 1996. Chemopre- Structural investigations on betacyanin pigments by LC
vention of lung and skin cancer by Beta vulgaris (beet) root ex- NMR and 2D NMR spectroscopy. Phytochemistry 65:
tract. Cancer Lett 100: 211–214. 415–422.
Kapadia GJ, Azuine MA, Sridhar R, et al. 2003. Chemoprevention Tesoriere L, Allegra M, Butera D, Livrea MA. 2004. Absorption,
of DMBA-induced UV-B promoted, NOR-1-induced TPA pro- excretion, and distribution of dietary antioxidant betalains in
moted skin carcinogenesis, and DEN-induced phenobarbital LDLs: potential health effects of betalains in humans. Am J
promoted liver tumors in mice by extract of beetroot. Clin Nutr 80: 941–945.
Pharmacol Res 47: 141–148. Tesoriere L, Gentile C, Angileri F, et al. 2013. Trans-epithelial trans-
Kapadia GJ, Azuine MA, Rao GS, Arai T, Iida A, Tokuda H. 2011. port of the betalain pigments indicaxanthin and betanin across
Cytotoxic effect of the red beetroot (Beta vulgaris L.) extract Caco-2 cell monolayers and influence of food matrix. Eur J
compared to doxorubicin (Adriamycin) in the human prostate Nutr 52: 1077–1087.
(PC-3) and breast (MCF-7) cancer cell lines. Anticancer van Oijen MG, Slootweg PJ. 2000. Gain-of-function mutations
Agents Med Chem 11: 280–284. in the tumor suppressor gene p53. Clin Cancer Res 6:
Kapadia GJ, Rao GS, Ramachandran C, Iida A, Suzuki N, Tokuda H. 2138–2145.
2013. Synergistic cytotoxicity of red beetroot (Beta vulgaris L.) Velzenberger E, Vayssade M, Legeay G, Nagel MD. 2008. Study of
extract with doxorubicin in human pancreatic, breast and pros- cell behaviour on a cellulose anti-adhesive substratum. Cellulose
tate cancer cell lines. J Complementary Integr Med 10: 1–10. 15: 347–357.
Kujala TS, Vienola MS, Klika KD, Lopokonen JM, Pihlaja K. 2002. Von Elbe. 2001. Betalains. In Currents Protocols in Food Analytical
Betalain and phenolics composition of four beetroot (Beta Chemistry, John Wiley & Sons, Inc.: New York; F 3.1.1–F3.1.7.
vulgaris) cultivars. Eur Food Res Technol 214: 505–510. Zhang Q, Pan J, Wang Y, Lubet R, You M. 2013. Beetroot red
Lechner JF, Wang LS, Rocha CM, et al. 2010. Drinking water with (betanin) inhibits vinyl carbamate- and benzo(a)pyrene-in-
red beetroot food color antagonizes esophageal carcinogene- duced lung tumorigenesis through apoptosis. Mol Carcinog
sis in N-nitrosomethylbenzylamine-treated rats. J Med Food 52: 686–691.
13: 733–739. Zhang X, Chen LX, Ouyang L, Cheng Y, Liu B. 2012. Plant natural
Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. 1951. Protein compounds: targeting pathways of autophagy as anti-cancer
measurement with the folin phenol reagent. J Biol Chem therapeutic agents. Cell Prolif 45: 466–476.
193: 265–275.
Mohankumar K, Pajaniradje S, Sridharan S, et al. 2014. Mechanism
of apoptotic induction in human breast cancer cell, MCF-7, by
an analog of curcumin in comparison with curcumin—an
SUPPORTING INFORMATION
in vitro and in silico approach. Chem Biol Interact 210: 51–63.
Paluszczak J, Krajka-Kuzniak V, Baer-Dubowska W. 2010. The Additional supporting information may be found in the
effect of dietary polyphenols on the epigenetic regulation of online version of this article at the publisher’s web-site.

Copyright © 2015 John Wiley & Sons, Ltd. Phytother. Res. 29: 1964–1973 (2015)