Antimicrobial cyclic decapeptides with anticancer activity

Article abstract of DOI:10.1016/j.peptides.2010.07.027

Antimicrobial peptides have been considered as potential candidates for cancer therapy. We report here the cytotoxicity of a library of 66 antibacterial cyclodecapeptides on human carcinoma cell lines, and their effects on apoptosis [as assessed by cleava

Full text of DOI:10.1016/j.peptides.2010.07.027

Peptides 31 (2010) 2017–2026
Contents lists available at ScienceDirect
Peptides
journal homepage: www.elsevier.com/locate/peptides
Antimicrobial cyclic decapeptides with anticancer activity
Lidia Feliua,1, Glòria Oliverasb,c,1, Anna D. Ciraca,d, Emili Besalúd, Cristina Rosésa,
Ramon Colomere, Eduard Bardajía, Marta Planasa,∗, Teresa Puigb,c,∗∗
a LIPPSO, Department of Chemistry, University of Girona, Campus Montilivi, E-17071 Girona, Spain
b Catalan Institute of Oncology (ICO) - Girona Biomedical Research Institute (IdIBGi), Josep Trueta Hospital, E-17001, Girona, Spain
c Faculty of Science and Medicine, University of Girona, Campus Montilivi, E-17071 Girona, Spain
d Institute of Computational Chemistry, University of Girona, Campus Montilivi, E-17071 Girona, Spain
e MD Anderson Espa˜na Hospital, Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 23 June 2010
Received in revised form 30 July 2010
Accepted 30 July 2010
Available online 11 August 2010
Antimicrobial peptides have been considered as potential candidates for cancer therapy. We report here
the cytotoxicity of a library of 66 antibacterial cyclodecapeptides on human carcinoma cell lines, and
their effects on apoptosis [as assessed by cleavage of poly(ADP-ribose) polymerase (PARP)] and cell sig-
naling proteins (p53 and ERK1/2) in cultured human cervical carcinoma cells. A design of experiments
approach permitted to analyze the results of a subset of 16 peptides and define rules for high anticancer
activity against MDA-MB-231 breast carcinoma cells. Eight peptides were identified with IC50 values rang-
ing from 18.5 to 57.5 ␮M against the five cell lines tested, being HeLa cells the most sensitive. Among
these sequences, BPC88, BPC96, BPC98, and BPC194 displayed specificity and high cytotoxicity against
HeLa cells (IC50 of 22.5–38.5 ␮M), showed low hemolytic activity and low cytotoxicity to non-malignant
fibroblasts, and were stable to proteases in human serum. Induction of apoptosis by these peptides was
observed and the apoptotic effect of BPC88 and BPC96 caused a marked decrease on the activated form of
ERK1/2 kinase and an induction of p53. We further showed that BPC96 at low doses synergized the cyto-
toxic effect of cisplatin. These findings suggest that cyclic decapeptides may represent novel anticancer
agents providing a new strategy in cancer therapy.
Keywords:
Apoptosis
Cervical carcinoma
Cyclopeptides
Pharmacological synergism
Design of experiments
© 2010 Elsevier Inc. All rights reserved.
1. Introduction
through a mechanism in which the presence of a net positive
charge and the ability to assume an amphipatic structure are cru-
Antimicrobial peptides have emerged as potential candidates
for cancer therapy. They are cytotoxic not only against a wide
range of bacteria, fungi, enveloped viruses and protozoa, but
also against different types of human cancer cells, such as those
from breast, bladder, ovarian and lung [15,17,23,25,31,46]. More-
over, antimicrobial peptides have a mode of action and cellular
targets that are different from those of currently available anti-
cancer drugs, and some of them show low eukaryotic cytotoxicity
[15,17,23,25,31,46].
cial peptide structural requirements [3,5,14,15,23,25,31,37]. The
positively charged amino acids interact electrostatically with the
negatively charged molecules present on the cell surface while the
amphipaticity favors their insertion into the membrane. Antimicro-
bial peptides may kill cancer cells via membrane permeation, either
by a detergent-like disruption of the cell membrane into peptide
coated vesicles or by formation of transient transmembrane pores.
Based on this mode of action, these peptides are unlikely to cause
rapid emergence of resistance because it would require significant
alteration of membrane composition, which is difficult to occur
[33,45]. In addition, there is increasing evidence that apart from
membrane damage, other mechanisms may be involved including
intracellular targets, such as mitochondria and nucleic acids which
may induce apoptosis in cancer cells [15,25,31].
Some antimicrobial peptides with anticancer activity do not
show significant cytotoxicity against untransformed proliferating
cells at peptide concentrations that are able to kill cancer cells
[15,23,25,31]. Parameters that could account for this selectivity
involve the higher net negative charge and membrane fluidity of
cancer cells as compared to normal cells. The higher negative charge
of cancer cells is due to a higher expression of anionic molecules
Several modes of action have been suggested for the anticancer
activity of antimicrobial peptides. They affect cell membranes
∗
Corresponding author at: LIPPSO, Department of Chemistry, University of Girona,
Campus Montilivi, E-17071 Girona, Spain. Tel.: +34 972 418274;
fax: +34 972 418150.
∗∗
Corresponding author at: Catalan Institute of Oncology (ICO) - Girona Biomedical
Research Institute (IdIBGi), Josep Trueta Hospital, E-17001, Girona, Spain.
Tel.: +34 972 940282; fax: +34 972 944854.
E-mail addresses: marta.planas@udg.edu (M. Planas), mtpuig@iconcologia.net
(T. Puig).
1
These two authors contributed equally to this work.
0196-9781/$ – see front matter © 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.peptides.2010.07.027

2018
L. Feliu et al. / Peptides 31 (2010) 2017–2026
such as phosphatidylserine and O-glycosylated mucins on the
outer membrane leaflet, favoring the electrostatic interactions with
cationic antimicrobial peptides. The increased membrane fluidity
facilitates cancer cell membrane destabilization by membrane-
bound antimicrobial peptides.
lès, Spain). Solvents for reverse-phase high-performance liquid
chromatography (RP-HPLC) were obtained from Scharlau (Sent-
menat, Spain). Cisplatin (Pharmacia Spain S.A., Barcelona, Spain)
was provided by the Girona Division of Catalan Institute of Oncol-
ogy Hospital Pharmacy (ICO, Hospital Josep Trueta, Girona, Spain)
and was stored at 4 ◦C. Human serum was supplied by Banc
de Sang i Teixits of Hospital Josep Trueta and was stored at
20 ◦C. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro-
mide (MTT) and monoclonal anti-␤-actin mouse antibody (clone
AC-15) were purchased from Sigma (St. Louis, MO, USA). Rab-
bit polyclonal antibodies against PARP and mouse monoclonal
antibodies against phospho-ERK1/2 were from Cell Signaling Tech-
nology (Beverly, MD, USA). Mouse monoclonal antibody against
p53 (clone PAb 1801) was from Neomarkers (Fremont, CA, USA).
Peroxidase conjugated secondary antibody was from Calbiochem
(San Diego, CA, USA).
Although peptides with significant biological activities have
been described, their therapeutic use is mainly limited by their
low stability towards protease degradation [14,40]. Peptide cycliza-
tion constitutes a chemical strategy to overcome this limitation.
In particular, cyclic peptides have reduced conformational free-
dom which, apart from increasing their stability to proteolysis,
makes them potent and specific binding ligands to macromolecu-
lar receptors [7,9,21]. In recent years, there has been much interest
in the development of cyclic peptides through modification of nat-
ural products or by de novo design, leading to the identification
of compounds with improved or novel pharmacological activities
[8,12,21]. For instance, potent cyclic antitumor peptides have been
studied, such as Kahalalide F which is currently under clinical trials
for several types of cancer [11,29], and novel anticancer peptide-
based immunotherapies, such as peptide-based vaccines used in
cancer clinical trials [27].
The process involved in the development of lead candidates
is time-consuming and limited by the number of individual com-
pounds that can be prepared. Synthesis of libraries of small cyclic
peptides by combinatorial chemistry constitutes a suitable strategy
for the rapid identification of effective compounds with pharma-
cological activity [10,28,34,38,39]. In fact, such an approach has
led to the discovery of cyclic peptides as ligands targeted to dif-
ferent types of cancer cells [13,18,24,43,47]. However, a limitation
of combinatorial chemistry to optimize molecular properties is the
difficulty in determining cooperative effects among the molecular
substitutions. Design of experiments (DOE) [4] constitutes a gen-
eral statistical methodology able to grasp simultaneous, synergic
and non-linear effects among experimental factors and to eluci-
date inner rules governing the system’s behavior in order to assist
an investigation course. This methodology has been successfully
applied in the peptide design and activity prediction [1,28].
We have recently designed and synthesized a library of cyclic
decapeptides with general structure c(X5-Phe-X3-Gln) where X is
Lys or Leu that display high antibacterial activity and low hemolytic
activity [28]. Based on these properties, we decided to evaluate the
anticancer activity of these peptides. We first explored the cyto-
toxicity of the cyclic decapeptide library on human breast cancer
cells. A DOE approach was applied to determine the cooperative
effects among the molecular substitutions and get the associated
statistical significance. The most active peptides were then tested
for cytotoxicity on five different human carcinoma cell lines. Pep-
tides markedly selective towards HeLa cells by inducing apoptosis,
as evidenced by caspase activation, were identified. For the peptide
with the optimal biological profile, we also evaluated its ability to
synergize with cisplatin in HeLa cells.
2.2. Synthesis of the cyclic decapeptide library
The MBHA resin (3.5 g, 0.4 mmol/g) was swollen with
CH2Cl2 (1× 20 min) and DMF (1× 20 min), and washed with
piperidine–DMF (3:7, 1× 5 min), DMF (6× 1 min) and CH2Cl2
(3× 1 min). Then, the resin was treated with Fmoc-Rink linker
(5 equiv.), N-[1H-benzotriazol-1-yl)(dimethylamino)methylene]-
N-methylmethanaminium hexafluorophosphate N-oxide (HBTU)
(5 equiv.) and DIEA (10 equiv.) in DMF overnight. After this time,
the resin was washed with DMF (6× 1 min), CH2Cl2 (3× 1 min) and
diethyl ether (3× 1 min), and air-dried.
The linear peptide precursors were prepared on an ACT Multi-
ple Peptide Synthesizer Apex 396 S (Aaptec). The above resin was
split into 66 wells (50 mg). Couplings were conducted using the
corresponding Fmoc-protected amino acid (0.6 M), HBTU (0.6 M),
1-hydroxybenzotriazole (HOBt) (0.6 M) and DIEA (1.2 M) in DMF
for 1 h. Fmoc removal was performed with piperidine–DMF (3:7,
5 + 10 min). After each coupling and deprotection step the resin was
washed with DMF (5× 1 min).
Once the chain assembly was completed, next steps were
performed manually in 66 polypropylene syringes fitted with a
polyethylene porous disk. The C-terminal allyl ester was cleaved
by treatment with Pd(PPh3)4 (3 equiv.) in CHCl3–acetic acid–NMM
(37:2:1) under nitrogen for 3 h, and the resin was washed with
tetrahydrofuran (3× 2 min), DMF (3× 2 min), DIEA–CH2Cl2 (1:19,
3× 2 min), sodium N,N-diethyldithiocarbamate (0.03 M in DMF,
3× 15 min), DMF (10× 1 min) and CH2Cl2 (3× 2 min). Following
Fmoc removal with piperidine–DMF (3:7, 2 + 10 min), cyclization
was carried out by treating the resin with benzotriazol-1-yl-N-
oxytris(pyrrolidino)phosphonium hexafluorophosphate (PyBOP)
(5 equiv.), HOBt (5 equiv.), and DIEA (10 equiv.) in NMP at 25 ◦C
for 24 h. Following washes with NMP (6× 1 min) and CH2Cl2 (3×
1 min), cyclodecapeptides were cleaved from the resin by treat-
ment with TFA–H2O–TIS (95:2.5:2.5) for 1 h. The cleavage cocktail
was then removed using a Thermo Savant SPD121P SpeedVac
concentrator. After diethyl ether extraction, cyclic peptides were
dissolved in H2O, lyophilized, and analyzed by analytical RP-HPLC
performed at 1.0 ml/min using a Kromasil (4.6× 40 mm; 3.5 ␮m
particle size) C18 reverse-phase column. Linear gradients of 0.1%
aqueous TFA and 0.1% TFA in CH3CN were run from 0.98:0.02 to
0:1 over 7 min with UV detection at 220 nm. Final products were
obtained in ∼90% HPLC purity and were confirmed by electrospray
ionization mass spectrometry.
2. Materials and methods
2.1. Chemicals, reagents and antibodies
The 9-fluorenylmethoxycarbonyl (Fmoc)-protected amino acid
derivatives, coupling reagents, and 4-methylbenzhydrylamine
(MBHA) resin hydrochloride (0.4 mmol/g) were obtained from
Iris Biotech (Marktredwitz, Germany). Trifluoroacetic acid (TFA),
N,N-dimethylformamide (DMF), N-methyl-2-pyrrolidinone (NMP),
Pd(PPh3)4, sodium N,N-diethyldithiocarbamate, triisopropylsi-
lane (TIS), and CHCl3 were from Sigma–Aldrich Corporation
(Madrid, Spain). Piperidine, N-methylmorpholine (NMM), and N,N-
diisopropylethylamine (DIEA) were purchased from Fluka (Buchs,
Switzerland). Acetic acid was from Panreac (Castellar del Val-
2.3. Cell lines and cell culture
MDA-MB-231 human breast Caucasian adenocarcinoma and
Panc-1 human pancreatic carcinoma cells were obtained from the
ATCC (American Type Culture Collection Rockville, MD, USA). A431

L. Feliu et al. / Peptides 31 (2010) 2017–2026
2019
human epidermoid carcinoma, HeLa human cervical carcinoma and
2.6. Hemolytic activity
HepG2 human hepatocellular carcinoma cells were obtained from
Eucellbank (Universitat de Barcelona, Spain). Cells were routinely
grown in Dulbecco’s modified Eagle’s medium (DMEM, Gibco,
Berlin, Germany) containing 10% fetal bovine serum (FBS, Bio-
Whittaker, Walkersville, MD, USA), 1% l-glutamine, 1% sodium
pyruvate, 50 U/ml penicillin and 50 ␮g/ml streptomycin (Gibco).
The cells remained free of Mycoplasma and were propagated in
adherent culture according to established protocols. Cells were
maintained at 37 ◦C in a humidified atmosphere of 95% air and 5%
CO2.
The data corresponding to the hemolytic activity of peptides
was previously reported by Monroc et al. [28]. It was evaluated at
375 ␮M by determining the hemoglobin release from erythrocyte
suspensions of fresh human blood (5%, v/v) using absorbance at
540 nm.
2.7. Design of experiments (DOE)
DOE was applied over a set of 16 cyclodecapeptides of general
structure c(X1X2X3X4LysPheLysLysLeuGln) for their growth inhi-
bition percentage of MDA-MB-231 breast carcinoma cells at 40 ␮M
peptide concentration. Two residues (Leu or Lys) could be present
at every molecular position coded with the letter X. Hence, every
substitution point constitutes a two-level factor and a total of 24
structures were studied, being the calculation a full factorial design.
All the molecular activity determinations were carried out in trip-
licate and in a random order. DOE calculations allowed not only to
detect structural rules for high activity but also it helped to deter-
mine that these rules were statistically significant. This information
was graphically represented in an effect plot. Data manipulations
were carried out with the MINITAB program (MINITAB version 14
for Windows. Minitab Inc., State College, PA, 2004).
2.4. Cell growth inhibition assay
Peptides were dissolved in sterile Phosphate Saline Buffer
(PBS) to get a final stock solution of 9600 ␮M. Sensitivity was
determined using a standard colorimetric MTT assay. Briefly,
human cancer cells were plated out at a density of 7 × 103
cells/100 ␮l/well in 96-well microtiter plates and allowed an
overnight period for attachment. Then, the medium was removed
and fresh medium along with the corresponding peptide con-
centration (1–60 ␮M) were added to the cultures. For the
drug-combination assay, a given concentration of peptide BPC96
(5–30 ␮M), of cisplatin (1.5, 2.5, 4.5 and 6.5 ␮М) or of BPC96
(5–30 ␮M) plus a fixed cisplatin dose (1.5, 2.5, 4.5 and 6.5 ␮М)
were added to the cultures. Agents were not renewed dur-
ing the 48 h of cell exposure, and control cells without agents
were cultured under the same conditions. Following treatment,
cells were fed with drug-free medium (100 ␮l/well) and MTT
solution (10 ␮l, 5 mg/ml in PBS), and incubation was prolonged
for 3 h at 37 ◦C. After carefully removing the supernatants,
the MTT-formazan crystals formed by metabolically viable cells
were dissolved in dimethyl sulphoxide (DMSO, 100 ␮l/well) and
the absorbance was measured at 570 nm in a multi-well plate
reader (Model Anthos Labtec 2010 1.7 reader). Using control
optical density (OD) values (C), test OD values (T), and time
zero OD values (T0), the peptide concentration that caused 50%
growth inhibition (IC50 value) was calculated from the equation,
100 × [(T − T0)/(C − T0)] = 50.
2.8. Analysis of apoptosis
Apoptosis and induction of caspase activity were evaluated by
Western blotting analysis of PARP cleavage. Following treatment of
HeLa cells with the corresponding selected peptide (20 ␮M) during
12 h, cells were harvested with trypsin–EDTA, washed twice with
PBS, lysed in a lysis buffer (1 mM EDTA, 150 mM NaCl, 100 ␮g/ml
PMSF, 50 mM Tris–HCl (pH 7.5)) and kept at 4 ◦C while they were
routinely mixed every 5 min on a vortex during 30 min. A sample
was taken for measurement of protein content by the Lowry-based
BioRad assay. Total protein extracts were immunoblotted using
4–12% sodium dodecyl sulphate-polyacrylamide gel (SDS-PAGE,
Invitrogen, Van Alley Way, CA, EUA), transferred to nitrocellulose
membranes and blocked for 1 h at room temperature in block-
ing buffer (2.5% powdered-skim milk in PBS-T [10 mM Tris–HCl
(pH 8.0), 150 mM NaCl and 0.05% Tween-20]) to prevent non-
specific antibody binding. Blots were incubated overnight at 4 ◦C
with the primary antibody (PARP) diluted in blocking buffer. After
washes in PBS-T (3× 5 min), blots were incubated for 1 h with
the corresponding peroxidase conjugated secondary antibody and
developed employing a commercial kit (West Pico chemilumines-
cent substrate). Blots were reproved with an antibody for ␤-actin
as control for protein loading and transfer.
2.5. Isobologram analysis of drug interactions
The interaction between BPC96 and cisplatin was evaluated by
the isobologram method as we have previously published [26,35].
Briefly, the concentration of one agent producing a 30% inhibitory
effect (IC30) is plotted on the horizontal axis, and the concen-
tration of another agent producing the same degree of effect
is plotted on the vertical axis; a straight line joining these 2
points represents zero interaction (addition) between 2 agents. The
experimental isoeffect points were the concentrations (expressed
relative to the IC30 concentrations) of the 2 agents that when
combined kill 30% of the cells. When the experimental isoeffect
points felt below that line, combination effect of the 2 drugs was
considered to be supra-additive or synergistic, whereas antago-
nism occurred if the experimental isoeffect points lay above it.
A quantitative index of these interactions was provided by the
equation Ix = (A/a) + (B/b), being a and b the respective concen-
trations of BPC96 and cisplatin required to produce a fixed level
of inhibition (IC30) when administered alone; A and B, the con-
centrations required for the same effect when the drugs were
administered in combination; and Ix, a drug interaction index.
Values of Ix < 1 indicate synergy, a value of 1 represents addi-
tion, and values >1 indicate antagonism. To estimate Ix, isoboles
were used only when intercept data for both axes were avail-
able.
2.9. Immunoblot analysis of p53 and phospho-ERK1/2
The effect of selected peptides on the tumor suppressor factor
p53 and on phospho-ERK1/2 was investigated. Following treatment
of HeLa cells with the corresponding selected peptide (20 ␮M) dur-
ing 12 or 24 h, cells were lysed in a buffer containing 1 mM EDTA,
150 mM NaCl, 100 ␮g/ml PMSF, 50 mM Tris–HCl (pH 7.5), protease
and phosphatase inhibitor cocktails. A sample was taken for mea-
surement of protein content by the Lowry-based BioRad assay.
Total protein extracts were immunoblotted using 4–12% SDS-PAGE,
transferred to nitrocellulose membranes and blocked for 1 h at
room temperature in blocking buffer (2.5% powdered-skim milk in
PBS-T [10 mM Tris–HCl (pH 8.0), 150 mM NaCl and 0.05% Tween-
20]) to prevent non-specific antibody binding. Blots were incubated
overnight at 4 ◦C with the corresponding primary antibody diluted
in blocking buffer. After washes in PBS-T (3× 5 min), blots were
incubated for 1 h with the corresponding secondary antibody cho-

2020
L. Feliu et al. / Peptides 31 (2010) 2017–2026
Scheme 1. General synthetic route for the synthesis of the cyclic decapeptides depicted for BPC96.
sen according to the species of origin of the primary antibody (p53
and phospo-ERK1/2), and developed employing a commercial kit
(West Pico chemiluminescent substrate). Blots were reproved with
an antibody for ␤-actin as control for protein loading and transfer.
2.10. Stability of peptides in serum
The stability of selected peptides was evaluated in human
serum. Each peptide (2.5 mg) was exposed to 25% aqueous filtered
human serum (1 ml) at 37 ◦C. After 30 min, 1, 1.5 and 2 h exposure,
aliquots (200 ␮l) were removed and proteins were precipitated
with acetonitrile (200 ␮l). Samples were cooled to 0 ◦C for 15 min
and centrifuged (11,000 rpm, 5 min). The supernatant was frozen,
freeze-dried and analyzed by matrix-assisted laser desorption ion-
ization with time-of-flight analysis (MALDI-TOF, ULTRAflex TOF,
Bruker) using an ␣-cyano-4-hydroxycinnamic acid matrix.
3. Results
3.1. Library design and synthesis
Fig. 1. Effects plot obtained when single and two-residues interactions are being
considered. The statistically significant and dominant factors combinations are rep-
resented by squares, labeled as single (X1, X4, X3) and double interactions (X1X3).
The library comprised 66 cyclic decapeptides which incorpo-
rated a Phe and a Gln residue at positions 6 and 10, respectively,
and alternating Lys and Leu at the other positions (Table 1). Among
them, 50 peptides incorporated all combinations of three Leu
and five Lys, and 16 cyclopeptides consisted of the substructure
Lys5PheLysLysLeuGln10 and all possible combinations of Leu and
Lys at positions 1–4. This cyclic decapeptide library was synthe-
sized by carrying out solid-phase synthesis of linear sequences
on an ACT Multiple Peptide Synthesizer Apex 396 S, followed
by manually on-resin cyclization. A three-dimensional orthogonal
Fmoc/tButyl/Allyl strategy was used [20]. Side-chain protection for
Lys was as tert-butyl carbamate (Boc). A Fmoc-Glu-OAl residue was
introduced as trifunctional amino acid to allow peptide anchor-
ing onto the resin, which resulted in a Gln after peptide cleavage
from the solid support. Peptides were obtained with ∼90% purity as
determined with high-performance liquid chromatography (HPLC).
Electrospray ionization mass spectrometry was used to confirm
peptide identity. The general strategy for the synthesis of the pep-
tide library is depicted in Scheme 1 for BPC96.
peptides (BPC88, BPC94, BPC96, BPC98, BPC184, BPC194, BPC198,
and BPC202) have the substructure Lys5PheLysLysLeuGln10 as com-
mon structural feature. The most active peptides were BPC94 and
BPC184, with a growth inhibition percentage of 95 and 88%, respec-
tively.
The above eight peptides belong to a set of sixteen sequences
that were previously designed to be analyzed by design of experi-
ments (DOE) [28]. We decided to apply this methodology to study
the influence of the residues X1 to X4 in the general structure
c(X1X2X3X4LysPheLysLysLeuGln) on the growth inhibition per-
centage of the MDA-MB-231 cells. Representative data series are
shown in Table 2.
A full two-level factorial design was performed. In a first calcu-
lation, only first order (single residues) and two-order interactions
(pairs of residues) were analyzed. Results are represented by means
of the effects plot (Fig. 1), a probabilistic scale graph commonly used
to represent the influential factors obtained from DOE calculations.
The points depicted in plots represent calculated effects (increase
or decrease of the measured activity) associated to different fac-
tor combinations (single, double or higher order interactions). This
magnitude, once expressed in standardized deviation units, consti-
tutes the abscissas axis variable. The vertical axis variable is the left
queue cumulated probability to find a data point in the sequence of
points once sorted by its effect magnitude. The straight line in Fig. 1
shows the ideal fitting line to which the points should approach if a
random Gaussian distribution is followed. The black points around
this line represent effects that are not significant. The points that
3.2. Cytotoxic effect of cyclic peptides against a breast carcinoma
cancer cell line and design of experiments (DOE)
The anticancer activity of the full cyclic decapeptide library was
first screened in MDA-MB-231 human breast adenocarcinoma cells
by determining the growth inhibition percentage at a peptide con-
centration of 40 ␮M by an MTT assay. Growth inhibition values
are summarized in Table 1. Eleven peptides produced an inhibi-
tion percentage of breast carcinoma cell growth between 20 and
50%, and eight sequences caused ≥50% inhibition, indicating that
these peptides exhibit a potential anticancer effect. All the latter

L. Feliu et al. / Peptides 31 (2010) 2017–2026
2021
Table 1
Antitumor activity against MDA-MB-231 cells and hemolysis of the cyclic decapeptide library.
Peptide
Growth inhibition (%)a
Hemolysis (%)b
Code
Sequence
BPC10L
BPC58
BPC60
BPC62
BPC64
BPC66
BPC68
BPC70
BPC72
BPC74
BPC76
BPC78
BPC80
BPC82
BPC84
BPC86
BPC100
BPC102
BPC104
BPC106
BPC108
BPC110
BPC112
BPC114
BPC116
BPC118
BPC120
BPC122
BPC124
BPC126
BPC128
BPC130
BPC132
BPC134
BPC136
BPC138
BPC140
BPC142
BPC144
BPC146
BPC148
BPC150
BPC152
BPC154
BPC156
BPC158
BPC160
BPC162
BPC164
BPC166
BPC88
c(Lys-Leu-Lys-Leu-Lys-Phe-Lys-Leu-Lys-Gln)
c(Lys-Lys-Lys-Lys-Lys-Phe-Leu-Leu-Leu-Gln)
c(Lys-Lys-Lys-Lys-Leu-Phe-Lys-Leu-Leu-Gln)
c(Lys-Lys-Lys-Leu-Lys-Phe-Lys-Leu-Leu-Gln)
c(Lys-Lys-Leu-Lys-Lys-Phe-Lys-Leu-Leu-Gln)
c(Lys-Leu-Lys-Lys-Lys-Phe-Lys-Leu-Leu-Gln)
c(Leu-Lys-Lys-Lys-Lys-Phe-Lys-Leu-Leu-Gln)
c(Lys-Lys-Lys-Lys-Leu-Phe-Leu-Lys-Leu-Gln)
c(Lys-Lys-Lys-Leu-Lys-Phe-Leu-Lys-Leu-Gln)
c(Lys-Lys-Leu-Lys-Lys-Phe-Leu-Lys-Leu-Gln)
c(Lys-Leu-Lys-Lys-Lys-Phe-Leu-Lys-Leu-Gln)
c(Leu-Lys-Lys-Lys-Lys-Phe-Leu-Lys-Leu-Gln)
c(Lys-Lys-Lys-Leu-Leu-Phe-Lys-Lys-Leu-Gln)
c(Lys-Lys-Leu-Lys-Leu-Phe-Lys-Lys-Leu-Gln)
c(Lys-Leu-Lys-Lys-Leu-Phe-Lys-Lys-Leu-Gln)
c(Leu-Lys-Lys-Lys-Leu-Phe-Lys-Lys-Leu-Gln)
c(Lys-Lys-Lys-Lys-Leu-Phe-Leu-Leu-Lys-Gln)
c(Lys-Lys-Lys-Leu-Lys-Phe-Leu-Leu-Lys-Gln)
c(Lys-Lys-Leu-Lys-Lys-Phe-Leu-Leu-Lys-Gln)
c(Lys-Leu-Lys-Lys-Lys-Phe-Leu-Leu-Lys-Gln)
c(Leu-Lys-Lys-Lys-Lys-Phe-Leu-Leu-Lys-Gln)
c(Lys-Lys-Lys-Leu-Leu-Phe-Lys-Leu-Lys-Gln)
c(Lys-Lys-Leu-Lys-Leu-Phe-Lys-Leu-Lys-Gln)
c(Lys-Leu-Lys-Lys-Leu-Phe-Lys-Leu-Lys-Gln)
c(Leu-Lys-Lys-Lys-Leu-Phe-Lys-Leu-Lys-Gln)
c(Lys-Lys-Leu-Leu-Lys-Phe-Lys-Leu-Lys-Gln)
c(Leu-Lys-Lys-Leu-Lys-Phe-Lys-Leu-Lys-Gln)
c(Lys-Leu-Leu-Lys-Lys-Phe-Lys-Leu-Lys-Gln)
c(Leu-Lys-Leu-Lys-Lys-Phe-Lys-Leu-Lys-Gln)
c(Leu-Leu-Lys-Lys-Lys-Phe-Lys-Leu-Lys-Gln)
c(Lys-Lys-Lys-Leu-Leu-Phe-Leu-Lys-Lys-Gln)
c(Lys-Lys-Leu-Lys-Leu-Phe-Leu-Lys-Lys-Gln)
c(Lys-Leu-Lys-Lys-Leu-Phe-Leu-Lys-Lys-Gln)
c(Leu-Lys-Lys-Lys-Leu-Phe-Leu-Lys-Lys-Gln)
c(Lys-Lys-Leu-Leu-Lys-Phe-Leu-Lys-Lys-Gln)
c(Lys-Leu-Lys-Leu-Lys-Phe-Leu-Lys-Lys-Gln)
c(Leu-Lys-Lys-Leu-Lys-Phe-Leu-Lys-Lys-Gln)
c(Lys-Leu-Leu-Lys-Lys-Phe-Leu-Lys-Lys-Gln)
c(Leu-Lys-Leu-Lys-Lys-Phe-Leu-Lys-Lys-Gln)
c(Leu-Leu-Lys-Lys-Lys-Phe-Leu-Lys-Lys-Gln)
c(Lys-Lys-Leu-Leu-Leu-Phe-Lys-Lys-Lys-Gln)
c(Lys-Leu-Lys-Leu-Leu-Phe-Lys-Lys-Lys-Gln)
c(Leu-Lys-Lys-Leu-Leu-Phe-Lys-Lys-Lys-Gln)
c(Lys-Leu-Leu-Lys-Leu-Phe-Lys-Lys-Lys-Gln)
c(Leu-Lys-Leu-Lys-Leu-Phe-Lys-Lys-Lys-Gln)
c(Leu-Leu-Lys-Lys-Leu-Phe-Lys-Lys-Lys-Gln)
c(Lys-Leu-Leu-Leu-Lys-Phe-Lys-Lys-Lys-Gln)
c(Leu-Lys-Leu-Leu-Lys-Phe-Lys-Lys-Lys-Gln)
c(Leu-Leu-Lys-Leu-Lys-Phe-Lys-Lys-Lys-Gln)
c(Leu-Leu-Leu-Lys-Lys-Phe-Lys-Lys-Lys-Gln)
c(Lys-Lys-Leu-Leu-Lys-Phe-Lys-Lys-Leu-Gln)
c(Lys-Leu-Lys-Leu-Lys-Phe-Lys-Lys-Leu-Gln)
c(Leu-Lys-Lys-Leu-Lys-Phe-Lys-Lys-Leu-Gln)
c(Lys-Leu-Leu-Lys-Lys-Phe-Lys-Lys-Leu-Gln)
c(Leu-Lys-Leu-Lys-Lys-Phe-Lys-Lys-Leu-Gln)
c(Leu-Leu-Lys-Lys-Lys-Phe-Lys-Lys-Leu-Gln)
c(Lys-Leu-Leu-Leu-Lys-Phe-Lys-Lys-Leu-Gln)
c(Lys-Lys-Lys-Leu-Lys-Phe-Lys-Lys-Leu-Gln)
c(Leu-Leu-Leu-Lys-Lys-Phe-Lys-Lys-Leu-Gln)
c(Leu-Lys-Lys-Lys-Lys-Phe-Lys-Lys-Leu-Gln)
c(Leu-Lys-Leu-Leu-Lys-Phe-Lys-Lys-Leu-Gln)
c(Lys-Lys-Leu-Lys-Lys-Phe-Lys-Lys-Leu-Gln)
c(Leu-Leu-Lys-Leu-Lys-Phe-Lys-Lys-Leu-Gln)
c(Lys-Leu-Lys-Lys-Lys-Phe-Lys-Lys-Leu-Gln)
c(Leu-Leu-Leu-Leu-Lys-Phe-Lys-Lys-Leu-Gln)
c(Lys-Lys-Lys-Lys-Lys-Phe-Lys-Lys-Leu-Gln)
0
0
0
0
0
0.11
0
0.07
0.08
84 6.9
0.5
72 5.5
0.6
10 2.0
22 4.0
6
6
4
0
0.5
19 0.5
2.0
4
2
7
22 0.08
14 0.05
13 0.05
0.11
25 0.03
27 0.11
0.21
0.10
23 0.08
11 0.08
17 0.10
0.09
14 0.15
0.06
17 0.05
0.24
36 1.7
13 1.9
0
0.4
9
19 0.7
36 2.5
45 3.5
8
6
8
0.9
1.8 0.8
26 3.0
15 2.2
0
2
9
1
0.1
0.2
0.4
0.1
9
9
24 1.8
13 0.9
5
19 0.01
12 0.02
6
8
4
3
4
1
6
4
2
0.6
1.6
1.3
0.1
0.6
0.3
3.0
0.9
0.3
7
4
8
0.08
0.07
0.03
22 0.06
15 0.20
18 0.18
0.15
29 0.05
3
41 1.8
0.9
0
0
0
5
23 1.1
14 1.6
7
0.17
9
3
1
4
2
1.9
0.5
0.2
0.2
0.1
12 0.19
19 0.09
11 0.22
0
28 0.05
11 1.7
1
2
0.26
0.13
20 0.26
16 0.31
16 0.15
3
9
3
2
4
0
0.4
0.7
0.4
0.8
0.5
0.1
5
0.24
50 0.10
23 0.17
13 0.08
95 0.01
59 0.08
60 0.07
88 0.01
22 0.09
14 0.01
0
20 0.05
70 0.01
18 0.09
56 0.06
11 0.27
63 0.02
33 3.3
30 4.1
BPC90
BPC92
BPC94
BPC96
7
0.7
73 1.6
32 7.2
36 3.7
89 5.3
BPC98
BPC184
BPC186
BPC188
BPC190
BPC192
BPC194
BPC196
BPC198
BPC200
BPC202
0
0.4
87 6.5
0
49 7.7
17 1.7
47 7.2
14 1.4
71 11.7
2
0.2
a
Percentage of MDA-MB-231 human breast cancer cell growth inhibition at 40 ␮M dose plus standard deviation.
Percentage of erythrocytes hemolysis at 375 ␮M plus confidence interval.
b
move far away from the ideal line (squares) reveal which kind of
first rule accounted for eight molecules, six of which produced
a growth inhibition percentage ≥50% (BPC88, BPC94, BPC184,
BPC194, BPC198, and BPC202). Next residue crucial on the activity
was X4, and a Lys was also preferred. This second rule led to select
interactions are statistically significant.
DOE calculations indicated that X1 exerted the greatest influ-
ence on the activity (Fig. 1), being a Lys the optimal residue. This

2022
L. Feliu et al. / Peptides 31 (2010) 2017–2026
Table 2
Data used for DOE analysis (changes of residues X1 to X4 in the cyclic decapeptides c(X1X2X3X4LysPheLysLysLeuGln) and growth inhibition of MDA-MB-231 cells).
Residues
X1
Peptide
Growth inhibition (%)a
X2
X3
X4
Leu
Lys
Leu
Lys
Leu
Lys
Leu
Lys
Leu
Lys
Leu
Lys
Leu
Lys
Leu
Lys
Leu
Leu
Lys
Lys
Leu
Leu
Lys
Lys
Leu
Leu
Lys
Lys
Leu
Leu
Lys
Lys
Leu
Leu
Leu
Leu
Lys
Lys
Lys
Lys
Leu
Leu
Leu
Leu
Lys
Lys
Lys
Lys
Leu
Leu
Leu
Leu
Leu
Leu
Leu
Leu
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
BPC200
BPC184
BPC192
BPC88
BPC196
BPC90
7.2
88.0
24.2
42.4
15.3
35.0
19.8
16.4
13.6
95.1
54.3
70.5
55.9
52.0
0.0
0.0
88.6
18.2
51.2
12.7
16.5
11.2
21.6
14.6
93.6
56.9
68.9
58.6
57.5
0.0
29.5
87.4
17.4
57.5
26.2
16.4
8.9
27.1
13.8
94.9
65.9
70.6
66.5
59.8
0.0
BPC92
BPC186
BPC188
BPC94
BPC96
BPC194
BPC98
BPC198
BPC190
BPC202
63.8
62.8
61.7
a
Percentage of MDA-MB-231 human breast cancer cell growth inhibition at 40 ␮M dose. Data are obtained from three independent experiments as required by DOE.
six active compounds (BPC94, BPC96, BPC98, BPC194, BPC198, and
BPC202) out of eight. A third rule referred to X3, being Leu the
preferred residue and selecting five active peptides (BPC88, BPC94,
BPC96, BPC184, and BPC194). Finally, a double interaction X1X3 was
worth of mention, being the optimal combination the Lys-Leu pair.
This significant effect pointed to consider four active compounds,
BPC88, BPC94, BPC184, and BPC194. The simultaneous application
of these four rules accounted for two of the most active peptides:
BPC94 and BPC194. These four effects were statistically significant
with p values much lesser than 0.001. For all cases, the respective
standardized effects, in terms of t-Student variable, were greater
than 3 units with absolute values of 6.50, 4.12, 3.69, and 3.04 for
the effects X1, X4, X3, and X1X3, respectively (Fig. 1).
The numerical analysis revealed that peptides mainly showed
individual effects on substitution. Cooperative effects were stated
only for the X1X3 interaction, but it was the last rule in the hierar-
chic path. A second calculation including all the possible orders
of interaction effects showed that the triple interaction X1X2X3
was also important, being Lys-Leu-Leu the preferred residues. This
interaction pointed to the two most active structures BPC94 and
BPC184. This proposal was consistent with the above single-effect
preferences and, together with the Lys substitution at position X4,
led to identify the most active peptide BPC94.
In view of the above results, we checked the degree of coherence
among the data and the prediction capacity. A simpler fractional
factorial design was performed considering only a subset of eight
molecules (BPC184, BPC186, BPC188, BPC190, BPC192, BPC194,
BPC196, and BPC198). Deliberately, the most active compound
BPC94 was not included in the calculations. This partial data
revealed that many single effects and some two-factor interactions
were relevant. The most preferential rule (t value greater than 30.0)
was to set X1 = Lys, followed by the condition X3 = Leu (t = 15.69).
These two rules led to identify the active peptides BPC184 and
BPC194 (both were included into the calculations). They also pre-
dicted the design of other compounds which were not present in
the analyzed data set, i.e., the result gave a clue to consider peptides
BPC88 and BPC94. Additionally, the next relevant rule suggested to
set X2 = Leu (t = 10.65), definitively pointing out to peptide BPC94.
The remaining rules were also coherent with this choice. These
results showed that DOE is a useful tool in molecular design.
cell line and also against HeLa cervical, HepG2 hepatocellular, A431
epidermoid, and Panc-1 pancreatic carcinoma cells (Table 3). Pep-
tides were active at the concentrations tested against all cell lines
with IC50 values ranging from 18.5 to 57.5 ␮M, except for BPC198 in
HepG2. In these experiments, BPC94 was active against all cell lines,
with an IC50 comprised between 18.5 and 23 ␮M. BPC88 was also
effective although at higher concentrations (IC50 of 22.5–32.5 ␮M),
as were peptides BPC96, BPC98, BPC184 and BPC194, with IC50
values from 24.5 to 51 ␮M. Interestingly, in HeLa cells except for
BPC202 (IC50 of 54 ␮M) and BPC98 (IC50 of 38.5 ␮M), the other
peptides exhibited an IC50 ≤ 35 ␮M.
3.4. Hemolysis and growth inhibition assay of peptides in normal
mammalian cells
The hemolytic activity of the cyclic peptides was evaluated at
375 ␮M (Tables 1 and 3). Results showed that two of the eight
selected peptides, BPC94 and BPC184, showed significant hemol-
ysis (73 and 89%, respectively). In contrast, the other peptides
exhibited a low hemolytic activity, ranging from 2 to 36% (Table 3).
The activity on non-malignant human cells (fibroblasts N1) was
assessed for peptides that displayed high cytotoxic activity against
HeLa cells and also low hemolysis: BPC88, BPC96, BPC98, BPC194
and BPC198. Peptides BPC94 and BPC184 were not included due to
their high hemolysis and BPC202 due to its low activity against HeLa
cells. Results showed that BPC198 was highly cytotoxic in non-
malignant cells (IC50 of 7.5 ␮M in N1 vs. IC50 of 35 ␮M in HeLa cells).
By contrast, the IC50 values of BPC88, BPC96, BPC98 and BPC194 in
non-malignant fibroblasts were up to two fold their IC50 values in
malignant HeLa cells. This trend is particularly important for BPC96,
which only caused a letality of 13% in non-malignant cells (data not
shown) at its IC50 value in HeLa cells (24.5 ␮M).
3.5. Stability of peptides in human serum
The stability in human serum of the cyclic decapeptides BPC88,
BPC96, BPC98, BPC194 and BPC198 was tested next (see Supporting
Information). After exposure to 25% aqueous human serum at dif-
ferent time intervals, the presence of peptide was analyzed by
MALDI-TOF. Results showed that all five cyclic peptides were stable
in human serum. The least stable was BPC88 which after a 30 min
exposure was almost completely degraded. In contrast, BPC96 and
BPC98 were still detected up to 1.5 and 1 h, respectively. The most
stable peptides were BPC194 and BPC198 which were observed
after 2 h.
3.3. Cytotoxic effects of cyclic peptides against a panel of human
cancer cells
The eight best peptides in terms of cancer growth inhibition of
MDA-MB-231 cells were selected to determine the IC50 against this

L. Feliu et al. / Peptides 31 (2010) 2017–2026
2023
Table 3
Cytotoxicity (IC50
a
)
of selected cyclic decapeptides against a panel of human carcinoma cells and non-malignant fibroblasts (N1), and hemolytic activity.
Code
MDA-MB-231
HeLa
HepG2
A431
Panc-1
Hemolysisb
N1
BPC88
BPC94
BPC96
BPC98
BPC184
BPC194
BPC198
BPC202
31.2
22.0
40.0
40.7
32.2
32.5 0.5
53.2 13
40.5
5
0
7
3
7
22.5
23.0
24.5 0.7
38.5
30.5 12
29.5
35.0
54.0
0
3
32.5
20.5
34.5
44.0
42.2
46.0
>60
4
6
2
3
2
3
28.0
21.5
35.0
47.5
24.5 0.7
50.0 10
3
8
7
4
32.5 11
33 3.3
73 1.6
32 7.2
36 3.7
89 5.3
17 1.7
14 1.4
42.0 7.1
ndc
39.0 9.9
58.5 2.1
nd
56.5 0.7
7.5 0.7
nd
18.5
51.0
8
6
4
44.5 0.7
41.5 11
2
1
6
40.0
57.5
3
6
49.5
35.5
6
2
9
46.0
8
43.5 0.7
2
0.2
a
Cells were treated for 48 h. Cytotoxicity (␮M) was determined using the MTT assay. Each value represents the mean SD from three independent experiments performed
in triplicate.
b
Percentage of erythrocytes hemolysis at 375 ␮M plus confidence interval.
Not determined.
c
3.6. Analysis of apoptosis induced by peptides
co-treated with 1.5 ␮M cisplatin plus 5 ␮M BPC96 increased to 25%
(Fig. 3).
The cell death mechanism displayed by peptides BPC88, BPC96,
BPC98, BPC194 and BPC198 was examined using HeLa cells. First,
apoptosis and induction of caspase activity were checked by
Western blotting analysis showing cleavage of poly-ADP-ribose
polymerase (PARP) (Fig. 2A). Treatment of HeLa cells with pep-
tides at 20 ␮M during 12 h induced a marked increase on the
levels of the PARP cleavage product (89 kDa band) that, com-
pared to untreated cells, ranged between 32- (BPC88) and 684-fold
(BPC194).
We further examined the degree of synergistic interaction
between BPC96 and cisplatin in HeLa cells. The combined effect
was analyzed by the isobole method, using a series of isobologram
transformations of multiple dose-response curves at an effect level
of 30% (IC30) [26,35]. Based on the median interaction index value
(Ix), co-treatment of HeLa cells with BPC96 plus cisplatin resulted
in a marked synergistic interaction (Ix = 0.74 0.17).
4. Discussion
3.7. Effect of peptides on the expression levels of cellular proteins
p53 and phospho-ERK1/2
Several natural antimicrobial peptides are known to exhibit
anticancer activity. Particularly those with no or little hemolytic
activity can be considered as lead molecules for the develop-
ment of new anticancer agents [15,23,25]. We have recently
documented the antibacterial activity of a library of 66 de novo
designed cyclic decapeptides with general structure c(X5-Phe-
X3-Gln) where X is Lys or Leu [28]. Best peptides were active
against the phytopathogenic bacteria Erwinia amylovora, Pseu-
domonas syringae and Xanthomonas vesicatoria (3.1–25 ␮M) and
displayed low hemolysis (∼15% at 375 ␮M). In the present study,
we examined the anticancer activity of this peptide library.
Antitumor activity of the cyclic decapeptide library against
MDA-MB-231 human breast adenocarcinoma cells correlated with
the previously observed antimicrobial activity against the above
bacteria [28]. In both cases, the most active compounds were
identified within a subset of 16 peptides that share the substruc-
ture Lys5PheLysLysLeuGln10. The most potent antitumor peptides
caused a growth inhibition percentage of MDA-MB-231 cells
between 70 and 95%.
We also examined the effect of peptides BPC88, BPC96, BPC98,
BPC194 and BPC198 on the two related mediators involved on
two different targets: the oncogen p53 and the ERK1/2, a Mitogen-
Activated Protein Kinase pathway (Fig. 2B and C). Treatment of HeLa
cells with BPC98, BPC194 and BPC198 at 20 ␮M for 12 h had little
effect on the cellular protein levels of the tumor suppressor factor
p53 (Fig. 2B). In contrast, incubation of cells with BPC88 and BPC96
at 20 ␮M for 12 h noticeably increased the protein levels of p53
compared to untreated cells (1.8- and 1.5-fold, respectively). Con-
cerning ERK1/2-related pathway, incubation of cells with BPC88,
BPC96, and BPC198 at 20 ␮M for 12 and 24 h substantially blocked
this survival pathway based on the decrease in the activated form
of ERK1/2 (p-ERK1/2) compared to untreated cells (Fig. 2C). During
this experiment, there was no significant change in the total level
of ERK1/2 protein as assessed by Western blotting analysis (data
not shown).
3.8. Cytotoxicity of BPC96 and cisplatin co-treatment in HeLa cells
In the present study, DOE was a useful tool to define general
rules leading to the selection of optimal compounds against MDA-
MB-231 cells. This approach has been previously used in peptide
chemistry and allows to simultaneously study all the effects of each
residue in the peptide sequence and to analyze and discriminate
single or multiple interactions between them. Recently, the anal-
ysis using DOE of the peptide subset with the general structure
c(X1X2X3X4LysPheLysLysLeuGln), with X being Lys or Leu, led to
the identification of the optimal amino acids at positions 1–4 for
antibacterial activity [28]. It was found that the general rule to fol-
low was to set X2 =/ X3. In contrast, the numerical analysis of the
antitumor activity of the same peptide subset showed that sub-
stitutions at X1, X3 and X4 had the highest influence on activity.
The optimal residues found were Lys at positions 1 and 4, and Leu
at position 3. These rules led to the identification of the peptides
with the best antibacterial (BPC194 and BPC198) and antitumor
(BPC94 and BPC184) activities. Therefore, these results suggest that
cyclic decapeptides with selective antibacterial or antitumor activ-
ity could be designed.
Peptide BPC96 was selected to study its ability to synergize
with cisplatin in HeLa cells. Cells were co-treated for 48 h with
different concentrations of BPC96 (5–30 ␮M) along with a fixed
cisplatin dose (1.5, 2.5, 4.5, and 6.5 ␮M). We previously per-
formed dose-response experiments (MTT assay) of cisplatin in HeLa
cells to determine these fixed cisplatin concentrations required to
perform the combination experiments (data not shown). The per-
centage of surviving cells after treatment with 1.5, 2.5, 4.5, and
6.5 ␮M cisplatin was of 94 2, 72 3, 50 7, 45 6 and 40 4%,
respectively (Fig. 3). The co-treatment of HeLa cells with BPC96
(5–30 ␮M) and a fixed dose of cisplatin showed a strong syner-
gistic cytotoxicity in all cisplatin concentrations assayed (Fig. 3).
Interestingly, this synergistic effect was even significant when
the lowest BPC96 concentration tested (5 ␮M) was combined
with the lowest cisplatin dose assayed (1.5 ␮M). In summary, the
letality of HeLa cells treated with 5 ␮M BPC96 or with 1.5 ␮M
cisplatin was 8 and 5%, respectively, while letality of HeLa cells

2024
L. Feliu et al. / Peptides 31 (2010) 2017–2026
Fig. 2. Effect of cyclic peptides on the activity of caspase, p53 and ERK1/2 in HeLa cells (Western blotting and densitometry). (A) Effect of peptides BPC88, BPC96, BPC98,
BPC194 and BPC198 on caspase activity by PARP cleavage. HeLa cells were treated with peptides (20 ␮M) for 12 h, and equal amounts of lysates were immunoblotted with
anti-PARP antibody to identify the 89 kDa band (cleavage product). Blots were reproved for ␤-actin as loading control. Gels shown are representative of those obtained
from two independent experiments. (B) Effect of BPC88, BPC96, BPC98, BPC194 and BPC198 on p53 levels. HeLa cells were treated with peptides (20 ␮M) for 12 h, and equal
amounts of lysates were subjected to Western blot analyses with an anti-p53 antibody. Gels shown are representative of those obtained from two independent experiments.
Blots were reproved for ␤-actin as loading control. (C) Effect of BPC88, BPC96, BPC98, BPC194 and BPC198 on p-ERK1/2. HeLa cells were treated with peptides (20 ␮M) for
12 and 24 h, and equal amounts of lysates were subjected to Western blot. Activation of ERK1/2 was analyzed by assessing the phosphorylation status of ERK1/2 using the
corresponding phospho-specific antibody (phospo-ERK1/2). Blots were reproved for ␤-actin as loading control. Gels shown are representative of those obtained from two
independent experiments.
The DOE approach was also considered for the above subset
of peptides but for the hemolytic activity [28]. It was found that
the simultaneous presence of a Lys at positions 2 and 3 accounted
for many of the less hemolytic peptides (BPC92, BPC186, BPC190,
and BPC202). This rule was not exactly coincident with the pep-
tide sequence pattern associated to high antitumor activity. This
was also observed when comparing the rules for antibacterial
and hemolytic activities. However, except for peptides BPC94 and
BPC184, many of the compounds with the highest antitumor activ-
ity (BPC88, BPC96, BPC98, BPC194, BPC198, and BPC202) displayed
low levels of hemolysis.
A variability in peptide cytotoxicity was observed when they
were tested against different types of cancer cells. Among the
human cancer cell lines tested, data pointed to HeLa cells as the
most sensitive cell line to the selected peptides. This different sen-
sitivity may be attributed to differences in cell membrane composi-
tion, fluidity, and surface area between these cell lines [15]. Notably,
cyclic peptides selective against HeLa cells exhibited low toxicity
against erythrocytes and normal mammalian cells. Peptides BPC88,
BPC96, BPC194, and BPC198 did not show a significant hemolytic
activity, even at a much higher concentration than the IC50 values
against all tumor cell lines tested (14–33% hemolysis at 375 ␮M).
Moreover, BPC88, BPC96, and BPC194 were less cytotoxic against
non-malignant fibroblasts N1 than against tumor cells. In partic-
ular, a mortality of only 13% in N1 cells was observed for BPC96
at its IC50 value towards HeLa cells (24.5 ␮M, data not shown). The
basis of the selective killing of tumor cells by antimicrobial peptides
has been attributed to their mode of action [15,23,25,31,36,42,45].
While much is known about the general mechanism by which
antimicrobial peptides are active against microorganisms, less is
known concerning their anticancer mechanism of action. Gener-
ally, antimicrobial peptides function by disrupting cell membranes,
being the positive charge of the peptides and the different compo-
sition of the cell membrane bilayers the most critical factors for
the activity. Due to the fact that the amount of phosphatidylserine
located in the outer leaflet of the membrane is higher in cancer cells
than in normal cells, the former are more susceptible to the lytic
action of antimicrobial peptides [36].
Susceptibility to enzymatic cleavage is one of the limitations of
using peptides as therapeutic agents. Cyclization has been applied
to improve peptide stability against metabolic degradation. In a
previous report, we observed this tendency when analyzing the

L. Feliu et al. / Peptides 31 (2010) 2017–2026
2025
insights into the molecular mechanisms underlying the observed
cytotoxicity.
A synergistic effect between low doses of BPC96 and cisplatin
was strongly evident. This effect was observed even at the lowest
cisplatin concentration assayed. These synergistic results are
promising, since there is a lack of specific and non-toxic therapies
to treat human cervical carcinomas [2,6]. Current therapies use
high doses of cisplatin causing important undesirable collateral
effects such as nephrotoxicity [44]. The synergism between antimi-
crobial peptides and existing chemotherapeutic agents, which may
decrease the required dosage of these drugs to achieve a positive
response and thereby reduce drug-related side effects, has been
previously documented [25]. In fact, it has been described that
antimicrobial peptides with anticancer activity would supplement
rather than replace the conventional chemotherapeutics. For
example, cecropin A enhances the cytotoxic effect of 5-fluorouracil
or cytarabine on lymphoblastic leukemia cells [16]. A similar
additive inhibitory effect on the growth of small cell lung cancer
cells has been demonstrated when synthetic magainin analogues
are used in combination with cisplatin or etoposide [30].
5. Conclusions
Taken together, a number of our designed cyclic decapeptides
displayed a favorable profile for future therapeutic development
against cancer. In particular, BPC96 exhibited a marked anticancer
activity, high stability to protease degradation, induced apoptosis
in human cancer cells, and showed synergistic antitumor effects
when combined with cisplatin. These findings suggest that BPC96
may represent a novel therapeutic strategy for the treatment of
cervical carcinoma. Future experiments will investigate the efficacy
of BPC96 in a pre-clinical model of HeLa cisplatin resistant cells and
the anticancer efficacy and long-term toxicity of the combination
treatments in vivo using a murine cervical tumor model.
Fig. 3. Cytoxicity in HeLa cells following BPC96 and cisplatin (CP) combination treat-
ment. HeLa cells were treated with different concentrations of BPC96 (5–30 ␮М) or
with different combinations of BPC96 (5–30 ␮М) and a fixed cisplatin concentration
(1.5, 2.5, 4.5 and 6.5 ␮М) for 48 h. Black circles represent the percentage of surviving
cells after 48 h in cisplatin treatment, which was determined using the MTT assay.
Results are expressed as percentage of surviving cells from three independent exper-
iments performed in triplicate. The interaction index (Ix) for the two-drug effect
in HeLa cells was calculated using isobologram analysis. The Ix parameter indicate
whether the doses of the two drugs required to produce a given degree of cytotoxi-
city are greater than (Ix > 1 or antagonism) equal to (Ix = 1 or additivism) or less than
(Ix < 1 or synergism) the doses that would be required if the effect of two agents were
strictly additive. Ix values for the two-drug treatment were obtained from duplicate
studies. * (p < 0.05) Indicate the level of statistical significance of the Ix compared
with an Ix of 1.
Acknowledgments
Financial support was provided by grant from the University
of Girona (Grant for R+D Projects on Health Sciences). We are
also grateful to the Serveis Tècnics de Recerca of the University of
Girona for their support with the mass spectrometry analysis. We
acknowledge the Banc de Sang i Teixits of Josep Trueta Hospital of
Girona for supplying us the human serum and to Girona Division of
Catalan Institute of Oncology (ICO) Hospital Pharmacy for providing
us the cisplatin. E. Besalú thanks the project CTQ2009-09370 from
the Ministry of Science and Innovation of the Spanish Government.
stability of BPC10L to protease degradation. This cyclic peptide was
degraded significantly slower than its linear counterpart [28]. In the
present study, the stability of cyclic peptides in human serum was
checked. Notably, the cyclic peptides BPC96 and BPC194 with an
optimal biological profile in terms of anticancer activity, hemoly-
sis and cytotoxicity against non-malignant fibroblasts also showed
good stability in human serum.
Appendix A. Supplementary data
Apart from disrupting cell membranes, it has been reported
that antimicrobial peptides are endowed with cytotoxic mech-
anisms in cancer cells involving apoptosis via mitochondrial
pathway [15,22,31]. This mode of action was observed for the
above selected peptides, which induced PARP cleavage indicat-
ing their apoptotic activity in HeLa cells. The anticancer activity
of these peptides was accompanied by changes in cell growth and
proliferation signaling pathways. BPC96 exhibited an optimal bal-
ance between PARP cleavage, activation of p53 and blockage of
ERK1/2. On the other hand, peptides had different effects on the
proliferation signaling pathways. In particular, BPC88 and BPC96
induced apoptosis in HeLa cells through activation of p53 medi-
ated by ERK1/2 kinases. These findings are consistent with previous
reports showing that ERK1/2 kinases target the transcriptional reg-
ulatory domain of p53 responding to several anticancer drugs,
such as cisplatin [19,32,41]. In contrast, the induction of apop-
tosis by BPC198 was dependent on ERK1/2 kinases but was not
related with the p53 pathway. Further molecular studies with
these peptides are needed with the objective of gaining additional
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.peptides.2010.07.027.
References
[1] Barroso JM, Besalú E. Design of experiments applied to QSAR: ranking a set
of compounds and establishing a statistical significance test. J Mol Struct
(Theochem) 2005;727:89–96.
[2] Bar-Sela G, Epelbaum R, Schaffer M. Curcumin as an anti-cancer agent: review
of the gap between basic and clinical applications. Curr Med Chem 2009 [Epub
ahead of print].
[3] Bechinger B. Structure and function of membrane-lytic peptides. Crit Rev Plant
Sci 2004;23:271–92.
[4] Box GEP, Hunter WG, Hunter JS. Statistics for experimenters. An introduction
to design, data analysis, and model building. New York: John Wiley & Sons, Inc.;
1978.
[5] Brogden KA. Antimicrobial peptides: pore formers or metabolic inhibitors in
bacteria? Nat Rev Microbiol 2005;3:238–50.
[6] Chung SH, Lambert PF. Prevention and treatment of cervical cancer
in mice using estrogen receptor antagonists. Proc Natl Acad Sci USA
2009;106:19467–72.

2026
L. Feliu et al. / Peptides 31 (2010) 2017–2026
[7] Cowell SM, Lee YS, Cain JP, Hruby VJ. Exploring Ramachandran and Chi space:
conformationally constrained amino acids and peptides in the design of bioac-
tive polypeptide ligands. Curr Med Chem 2004;11:2785–98.
[8] Cruz LJ, Luque-Ortega JR, Ribas L, Albericio F, Kahalalide F. An antitumor dep-
sipeptide in clinical trials, and its analogues as effective antileishmanial agents.
Mol Pharmaceut 2009;6:813–24.
[9] Dathe M, Nikolenko H, Klose G, Bienert M. Cyclization increases the
antimicrobial activity and selectivity of arginine- and tryptophan-containing
hexapeptides. Biochemistry 2004;43:9140–50.
[28] Monroc S, Badosa E, Besalú E, Planas M, Bardají E, Montesinos E, et al.
Improvement of cyclic decapeptides against plant pathogenic bacteria using
a combinatorial chemistry approach. Peptides 2006;27:2575–84.
[29] Morita H, Takeya K. Bioactive cyclic peptides from higher plants. Heterocycles
2010;80:739–64.
[30] Ohsaki Y, Gazdar AF, Chen H-C, Johnson BE. Antitumor activity of magainin
analogues against human lung cancer cell lines. Cancer Res 1992;52:3534–8.
[31] Papo N, Shai Y. Host defense peptides as new weapons in cancer treatment.
Cell Mol Life Sci 2005;62:784–90.
[10] Duléry V, Uhlich NA, Maillard N, Fluxá VS, Garcia J, Dumy P, et al. A cyclode-
capeptide ligand to vitamin B12. Org Biomol Chem 2008;6:4134–41.
[11] Faircloth G, Cuevas C, Kahalalide F. ES285: potent anticancer agents from
marine moluscs. Prog Mol Subcell Biol 2006;43:363–79.
[12] Frecer V, Ho B, Ding JL. De novo design of potent antimicrobial peptides. Antimi-
crob Agents Chemother 2004;48:3349–57.
[13] Garcia-Martin F, Cruz LJ, Rodriguez-Mias RA, Giralt E, Albericio F. Design and
synthesis of FAJANU: a de novo C2 symmetric cyclopeptide family. J Med Chem
2008;51:3194–202.
[14] Hancock REW, Sahl HG. Antimicrobial and host defense peptides as new anti-
infective therapeutic strategies. Nature Biotechnol 2006;24:1551–7.
[15] Hoskin DW, Ramamoorthy A. Studies on anticancer activities of antimicrobial
peptides. Biochim Biophys Acta 2008;1778:357–75.
[32] Park SA, Park HJ, Lee BI, Ahn YH, Kim SU, Choi KS. Bcl-2 blocks cisplatin-induced
apoptosis by suppression of ERK-mediated p53 accumulation in B104 cells. Mol
Brain Res 2001;93:18–26.
[33] Peschel A, Sahl HG. The co-evolution of host cationic antimicrobial peptides
and microbial resistance. Nat Rev Microbiol 2006;4:529–36.
[34] Qin CH, Bu X, Zhong X, Ng NLJ, Guo Z. Optimization of antibacterial cyclic
decapeptides. J Comb Chem 2004;6:398–406.
[35] Ropero S, Menéndez JA, Vázquez-Martín A, Montero S, Cortés-Funes H, Colomer
R. Trastuzumab plus tamoxifen: anti-proliferative and molecular interactions
in breast carcinoma. Breast Cancer Res Treat 2004;86:125–37.
[36] Schröder-Borm H, Bakalova R, Andrä J. The NK-lysin derived peptide NK-2 pref-
erentially kills cancer cells with increased surface levels of negatively charged
phosphatidylserine. FEBS Lett 2005;579:6128–34.
[16] Hui L, Leung K, Chen HM. The combined effects of antibacterial peptide cecropin
A and anticancer agents on leukemia cells. Anticancer Res 2002;22:2811–6.
[17] Jenssen H, Hamill P, Hancock REW. Peptide antimicrobial agents. Clin Microbiol
Rev 2006;19:491–511.
[18] Jiang S, Liao C, Bindu L, Yin B, Worthy KW, Fisher RJ, et al. Discovery of thioether-
bridged cyclic pentapeptides binding to Grb2-SH2 domain with high affinity.
Bioorg Med Chem Lett 2009;19:2693–8.
[37] Shai Y. Mode of action of membrane active antimicrobial peptides. Biopolymers
2002;66:236–48.
[38] Takahashi T, Nagamiya H, Doi T, Griffiths PG, Bray AM. Solid phase library syn-
thesis of cyclic depsipeptides: aurilide and aurilide analogues. J Comb Chem
2003;5:414–28.
[39] Ueda S, Oishi S, Wang Z, Araki T, Tamamura H, Cluzeau J, et al. Structure-activity
relationships of cyclic peptide-based chemokine receptor CXCR4 antagonists:
disclosing the importance of side-chain and backbone functionalties. J Med
Chem 2007;50:192–8.
[19] Johnson CL, Lu D, Huang J, Basu A. Regulation of p53 stabilization by DNA
damage and protein kinase. Mol Cancer Ther 2002;1:861–7.
[20] Kates SA, Solé NA, Johnson CR, Hudson D, Barany G, Albericio F. A novel, conve-
nient, three-dimensional orthogonal strategy for solid-phase synthesis of cyclic
peptides. Tetrahedron Lett 1993;34:1549–52.
[21] Katsara M, Tselios T, Deraos S, Deraos G, Minos-Timotheos M, Lazoura E,
et al. Round and round we go: cyclic peptides in disease. Curr Med Chem
2006;13:2221–32.
[22] Kim S, Kim SS, Bang Y-J, Kim S-J, Lee BJ. In vitro activities of native and design
peptide antibiotics against drug sensitive and resistant tumor cell lines. Pep-
tides 2003;24:945–53.
[23] Leuschner C, Hansel W. Membrane disrupting lytic peptides for cancer treat-
ments. Curr Pham Design 2004;10:2299–310.
[40] Vlieghe P, Lisowski V, Martinez J, Khrestchatisky M. Synthetic therapeutic pep-
tides: science and market. Drug Discov Today 2010;15:40–56.
[41] Woessmann W, Chen X, Borkhardt A. Ras-mediated activation of ERK by
cisplatin induces cell death independently of p53 in osteosarcoma and neu-
roblastoma cell lines. Cancer Chemother Pharmacol 2002;50:397–404.
[42] Won H-S, Seo M-D, Jung S-J, Lee S-J, Kang S-J, Son W-S, et al. Structural determi-
nants for the membrane interaction of novel bioactive undecapeptides derived
from gaegurin 5. J Med Chem 2006;49:4886–95.
[43] Yao N, Xiao W, Wang X, Marik J, Park SH, Takada Y, et al. Discovery of targeting
ligands for breast cancer cells using the one-bead one-compound combinatorial
method. J Med Chem 2009;52:126–33.
[24] Liu T, Joo SH, Voorhees JL, Brooks CL, Pei D. Synthesis and screening of a cyclic
peptide library: discovery of small-molecule ligands against human prolactin
receptor. Bioorg Med Chem 2009;17:1026–33.
[25] Mader JS, Hoskin DW. Cationic antimicrobial peptides as novel cytotoxic agents
for cancer treatment. Expert Opin Investig Drugs 2006;15:933–46.
[26] Menendez JA, Lupu R, Colomer R. Inhibition of tumor-associated fatty acid
synthase hyperactivity induces synergistic chemosensitization of HER-2/neu-
overexpressing human breast cancer cells to docetaxel (taxotere). Breast
Cancer Res Treat 2004;84:183–95.
[44] Yao X, Panichpisal K, Kurtzman N, Nugent K. Cisplatin nephrotoxicity: a review.
Am J Med Sci 2007;334:115–24.
[45] Yeaman MR, Yount NY. Mechanisms of antimicrobial peptide action and resis-
tance. Pharmacol Rev 2003;55:27–55.
[46] Zhang L, Falla TJ. Antimicrobial peptides: therapeutic potential. Expert Opin
Pharmacother 2006;7:653–63.
[47] Zhang Y, Zhou S, Wavreille A-S, DeWille J, Pei D. Cyclic peptidyl inhibitors of
Grb2 and tensin SH2 domains identified from combinatorial libraries. J Comb
Chem 2008;10:247–55.
[27] Mocellin S, Pilati P, Nitti D. Peptide-based anticancer vaccines: recent advances
and future perspectives. Curr Med Chem 2009;16:4779–96.