Biochemical Pharmacology 86 (2013) 1254–1262

Contents lists available at ScienceDirect

Biochemical Pharmacology
journal homepage: www.elsevier.com/locate/biochempharm

Role of glycosylation in the anticancer activity of antibacterial peptides

against breast cancer cells
Yang-Yang Han, Hong-Yan Liu, Dong-Ju Han, Xi-Cui Zong, Shuang-Quan Zhang,
Yu-Qing Chen *
Jiangsu Province Key Laboratory for Molecular and Medical Biotechnology, Life Sciences College, Nanjing Normal University, Nanjing 210000, China

A R T I C L E I N F O A B S T R A C T

Article history: Antibacterial peptides (ABPs) with cancer-selective toxicity have received much more attention as
Received 27 May 2013 alternative chemotherapeutic agents in recent years. However, the basis of their anticancer activity
Accepted 7 August 2013 remains unclear. The modification of cell surface glycosylation is a characteristic of cancer cells. The
Available online 17 August 2013
present study investigated the effect of glycosylation, in particular sialic acid, on the anticancer activity
of ABPs. We showed that aurein 1.2, buforin IIb and BMAP-28 m exhibited selective cytotoxicity toward
Keywords: MX-1 and MCF-7 breast cancer cells. The binding activity, cytotoxicity and apoptotic activity of ABPs
Antibacterial peptide
were enhanced by the presence of O-, N-glycoproteins, gangliosides and sialic acid on the surface of
Glycoproteins
breast cancer cells. Among N-, O-glycoproteins and ganglioside, O-glycoproteins almost had the
Gangliosides
Sialic acid strongest effect on the binding and cytotoxicity of the three peptides. Further, up-regulation of hST6Gal1
Selectivity in CHO-K1 cells enhanced the susceptibility of cells to these peptides. Finally, the growth of MX-1
Breast cancer xenograft tumors in mice was significantly suppressed by buforin IIb treatment, which was associated
with induction of apoptosis and inhibition of vascularization. These data demonstrate that the three
peptides bind to breast cancer cells via an interaction with surface O-, N-glycoproteins and gangliosides.
Sialic acids act as key glycan binding sites for cationic ABP binding to glycoproteins and gangliosides.
Therefore, glycosylation in breast cancer cells plays an important role in the anticancer activity of ABPs,
which may partly explain their cancer-selective toxicity. Anticancer ABPs with cancer-selective
cytotoxicity will be promising candidates for anticancer therapy in the future.
ß 2013 Elsevier Inc. All rights reserved.

1. Introduction selectivity for tumor cells, represent a major limitation to this

type of therapy. Hence, the identification of more specific targets
Cancer is a leading cause of death worldwide, accounting for expressed within a certain tumor type is an important goal of
approximately 1/8 of all deaths, and breast cancer is the most anticancer research [2].
common cancer among women worldwide [1]. Chemotherapy is Antibacterial peptides, which are effective molecules of innate
currently the standard treatment in most cancers. However, immunity, have recently been shown to have anticancer activity.
resistance as well as potential toxicity and the many side effects More than 100 natural and synthetic animal ABPs have been
of chemotherapeutics, which are mainly due to insufficient shown to possess the ability to kill cancer cells in vitro or in vivo.
Therefore, they have gained much attention as potential anti-
cancer agents [3–5]. Anticancer ABPs can be assigned to two broad
Abbreviations: ABPs, Antibacterial peptides; BnGalNac, benzyl-2-acetamido-2- categories based on their target specificity: (1) ABPs that are highly
deoxy-a-D-galactopyranoside; DAPI, 40 ,6-diamidino-2-phenylindole; DMEM, Dul- potent against cancer cells but not against normal mammalian
becco’s modified Eagle’s medium; FBS, fetal bovine serum; FITC, fluorescein cells and (2) ABPs that are cytotoxic against both cancer cells and
isothiocyanate; FITC-SNA, FITC labeled Sambucus nigra agglutinin; GAPDH,
normal mammalian cells. ABPs with cancer-selective toxicity have
glyceraldehyde-3-phosphate dehydrogenase; H&E, hematoxylin and eosin;
hST6Gal1, human b-galactoside; a2, 6-sialyltransferase; L-PPMP, phenyl-2-
received much attention as alternative chemotherapeutic agents to
hexadecanoyamino-3-morpholino-1-propanol hydrochloride; MTT, methyl thia- overcome the limitations of current drugs. Among these, buforin
zolyl tetrazolium; MVD, mean vascular density; PARP, poly ADP-Ribose polymer- IIb (21 residues) and aurein 1.2 (13 residues) are attractive
ase; PBS, phosphate-buffered saline; TACAs, tumor-associated carbohydrate candidates because they display selective cytotoxicity against
antigens; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP-biotin
most cancer cell lines [6,7]. BMAP-28 m (also named Mu-5 BMAP),
nick end labeling.
* Corresponding author. an amino acid substituted analog of BMAP-28, exhibited signifi-
E-mail address: yuqingchen515@yahoo.com (Y.-Q. Chen). cantly reduced cytotoxicity against 3T3 cells and erythrocytes [8].

0006-2952/$ – see front matter ß 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.bcp.2013.08.008
 Y.-Y. Han et al. / Biochemical Pharmacology 86 (2013) 1254–1262 1255

Several mechanisms have been proposed to explain the anticancer purchased from Sigma–Aldrich (St. Louis, MO, USA), FITC-SNA
activity of ABPs, including cell membrane lytic effect, activation of were purchased from Vector Laboratories (Burlingame, CA, USA).
intrinsic pathways of apoptosis via mitochondrial membrane Caspase-9, caspase-8, PARP and GAPDH were purchased from Cell
disruption, suppression of tumor cell invasion by inhibiting Signaling Tech (Danvers, MA, USA). Terminal deoxynucleotidyl
matrix metalloproteinase 9 expressions or inhibition of blood transferase was purchased from Roche (Shanghai, China). Annexin
vessel development [9–11]. Regardless of the underlying anti- V/FITC Kit was purchased from BD Biosciences Pharmingen (San
cancer mechanism, the activity of these molecules requires Diego, CA, USA). Plasmid pcDNA3.0 and Lipofectamine 2000 were
peptide-membrane interaction, and the plasma membrane is the purchased from Invitrogen (CA, USA). All other reagents were made
major target of these anticancer ABPs. Upon peptide interaction in China and were of analytical grade.
with the membrane, several factors modulate the activity and
selectivity. 2.2. Cell and culture conditions
The cell membrane of eukaryotes differs significantly from
prokaryotic membranes in lipid composition, and this aspect is All cell lines used in this research were obtained from the
particularly important for the bacteria-selective activity of ABPs Shanghai Institute of Biochemistry and Cell Biology, Chinese
[12]. Acidic lipids, such as phosphatidylserine (PS), phosphati- Academy of Sciences. Chinese hamster ovary (CHO-K1) cells and
dylglycerol (PG) and cardiolipin (CL), are abundant on the outer the breast cancer cell line MCF-7 were maintained in DMEM, and
leaflets of the bacterial membrane, whereas the outer leaflets of the breast cancer cell line MX-1 was maintained in RPMI 1640
mammalian membranes are typically made up of zwitterionic medium, both media were supplemented with 10% FBS. Mouse
phospholipids. Indeed, one of the major differences between hepatocytes were isolated from the livers of adult mice and
cancer and non-cancer cell surfaces is the exposure of the maintained in RPMI 1640 medium supplemented with 10% FBS.
negatively charged molecules. Certain tumor cells show an Cells were cultured at 37 8C in a 5% CO2 incubator.
elevated expression of PS, a negatively charged phospholipid,
on the outer membrane leaflet, which presumably increases 2.3. Sialidase and inhibitor treatment
the susceptibility of these cells to ABPs [13,14]. However,
most cancer cells generally contain a small amount of this lipid MCF-7 cells and MX-1 cells (1  106) were maintained in
(3–9%) in the outer leaflet compared to that in normal cells, phenol red-free, serum-free medium and pretreated as follows:
which does not enough to explain the selective binding and 0.1 U/ml sialidase for 30 min, 2.5 mg/ml of tunicamycin for 48 h,
activity [15]. 1 mM of L-PPMP for 48 h, or 2 mM of BnGalNac for 48 h. Inhibitor-
Glycosylation is one of the most frequent occurring co- or post- treated and untreated cells were cultured in 96-well plates
translational modifications of proteins and lipids in cells and more (1  104 cells/well) in FBS-free DMEM for MCF-7 cells, or RPMI
than 50% of known protein sequences can potentially be 1640 for MX-1 cells.
glycosylated [16]. Many cell surface proteins carry polysaccharide
moieties which are either used as signaling devices during the 2.4. Cytotoxicity assay
biosynthetic pathway (e.g. N-linked glycosylation) or are involved
in the extra-cellular matrix function of proteins (O-linked Cells were cultured in 96-well plates (1  105 cells/well) in
glycosylation). The modification of cell surface glycosylation medium supplemented with 2% FBS, and then were treated with
occurs during carcinogenesis, which is known to be a common different concentrations of aurein1.2, buforin IIb and BMAP-28 m
feature of cancer cells, reflecting cancer-specific changes in the for 24 h. The MTT reagent was added and cells were further
biosynthesis of glycans such as glycosyltransferases and glycosides incubated for 4 h. After dissolving the formazan precipitate by
[17,18]. In breast cancer, glycosylation changes often lead to the dimethyl sulfoxide, the absorbance was measured at a test
expression of TACAs and these antigens are usually associated with wavelength of 490 nm with a reference wavelength of 570 nm.
cell adhesion, migration, proliferation and tumor growth [19]. To The lethality of treated cells was normalized to that of the PBS
the best of our knowledge, few studies have attempted to examine control cells set as zero.
the relationship between glycosylation and the anticancer activity
of ABPs. 2.5. Hemolytic activity assay
In the present study, we investigated the effect of glycosylation
on the anticancer activity of ABPs in breast cancer with particular Erythrocytes were isolated from fresh mouse blood cells by
emphasis on sialylated oligosaccharides of glycoproteins and centrifugation at 1000  g for 10 min and washed three times with
glycolipids to further our understanding of the selectivity of ABPs PBS. Erythrocytes were treated with aurein 1.2, buforin IIb or
against cancer cells. Our results suggest that sialic acids, together BMAP-28 m for 1 h at 37 8C, followed by centrifugation at 1000  g
with glycoproteins or gangliosides may play an important role in for 5 min. The absorbance of the supernatants was measured at
the anticancer activity of ABPs and could contribute to their 414 nm. Zero hemolysis was determined in PBS (APBS) and 100%
selectivity. hemolysis was determined in 0.1% (v/v) Triton X-100 (ATriton).
Percent hemolysis was calculated: (Apeptide  APBS)/(ATritonX-
2. Materials and methods 100  APBS)  100%.

2.1. Peptide synthesis, labeling and reagents 2.6. Measurement of surface sialic acid density and peptide binding

Aurein 1.2, buforin IIb and BMAP-28 m were synthesized using MX-1, MCF-7, mouse erythrocytes and hepatocytes were used
solid-phase methodology at GL Biochemistry Corporation (Shang- to measure surface sialic acid density and peptide binding by flow
hai, China). Purification by preparative RP-HPLC gave final cytometry. Cells at concentrations above 1  106 cells/ml were
products deemed >95% pure. Selective N-terminal fluorescein incubated with 5 mg/ml of FITC-SNA for 45 min at 4 8C. The binding
labeling of the peptide was performed with FITC and deemed >95% activities of peptides were assessed using FITC labeled aurein1.2,
homogeneous. buforin IIb and BMAP-28 m at 37 8C for 30 min. After incubation
Sialidase from Vibrio cholerae, L-PPMP, BnGalNac, MTT, DAPI, and washing with PBS, the cells were analyzed by flow cytometry
Rhodamine 123, CD31 polyclonal goat anti-mouse antibody were at 488 nm excitation.
1256 Y.-Y. Han et al. / Biochemical Pharmacology 86 (2013) 1254–1262

2.7. Annexin V/propidium iodide staining and tumor dimensions were measured every 3 days and the tumor
volume (mm3) was calculated using the formula
Apoptosis was evaluated using two-color flow cytometric V = length  width2  0.5. In each treatment group, the mice were
analysis with propidium iodide (PI) and Annexin V labeling using sacrificed and tumor tissues were carefully removed, cleaned and
an annexin V/FITC Kit. Briefly, MCF-7 and MX-1 cells (2.5  106 weighed.
cells) were pretreated with sialidase, tunicamycin, L-PPMP or
BnGalNac as described above, then treated with aurein1.2, buforin 2.12. Immunohistochemistry, fluorescence staining and TUNEL assay
IIb or BMAP-28 m at concentrations of 16 mM, 8 mM, and 16 mM
for MCF-7 cells; and 8 mM, 4 mM, and 8 mM for MX-1 cells, Xenograft tissue was fixed in 10% formaldehyde and embedded
respectively. After incubation for 12 h, the cells were washed twice in paraffin. Paraffin sections (4 mm) were dewaxed in xylene and
with PBS and resuspended in 500 ml of binding buffer. FITC- rehydrated through a graded series of ethanol. Tissue sections
conjugated Annexin V and PI were added to the cell suspension, were subjected to immunohistochemical staining by routine H&E
and the mixture was incubated for 10 min on ice in the dark. The and CD-31 staining. Briefly, following antigen retrieval, tissues
stained cells were analyzed by flow cytometry and the percentage were incubated with CD31 polyclonal goat anti-mouse antibody in
of apoptotic cells was calculated using Cell Quest software. a humidity chamber overnight at 4 8C. Then the CD31 goat anti-
mouse primary antibody was bridged by FITC-labeled anti-goat
2.8. DAPI staining and fluorescence double staining IgG. MVD was quantified by measuring pixels in 10 consecutive
fields at 200 magnification. The sialic acid content and buforin IIb
MCF-7 cells were cultured on chamber slides and treated with binding activity were compared between normal mouse tissues
buforin IIb (8 mM) for 24 h. Apoptotic nuclei were detected by and xenograft tissues using paraffin sections. FITC-SNA (15 mg/ml)
2 mg/ml of DAPI staining. Images of DAPI fluorescence were or FITC labeled buforin IIb (2 mM) was incubated at 37 8C in the
collected using a Nikon fluorescence microscope. MCF-7 cells were dark overnight.
also incubated with FITC labeled buforin IIb for 6 h in the dark, then TUNEL staining was performed to assess cell death in tumors as
harvested, rinsed three times with PBS and treated with 30 nM described previously. Paraffin sections were dewaxed in xylene
Rhodamine 123 for 30 min in the dark. Cells were washed and the and rehydrated through a graded series of ethanol. After
distribution of fluorescence was immediately analyzed by confocal permeabilizing by Triton X-100, the slides were treated with
laser scanning microscopy. Optical excitation was carried out with proteinase K for 30 min at 37 8C and then incubated in terminal
a 488 nm argon laser beam for the FITC signal and 525 nm for the deoxynucleotidyl transferase for 1 h at 37 8C in the dark. The slides
Rhodamine 123 signal. were also stained with DAPI.

2.9. Western blot analysis 2.13. Statistical analysis

MCF-7 cells (5  106 cells) were pretreated with sialidase, All experiments were repeated at least three times. Results
tunicamycin, L-PPMP or BnGalNac as described above, and then were expressed as mean values  standard error (mean  SE).
incubated with buforin IIb (8 mM) for 12 h. After treatment, cells Significance differences between the treatments and the respective
were harvested in cell lysis buffer, followed by centrifugation at controls were determined based on Student’s t test. A level of p < 0.05
15,000  g for 10 min. Supernatants were collected, electrophor- was considered to be significant.
esed on 12% SDS–PAGE gels, and transferred onto nitrocellulose
membranes. Membranes were blocked with 5% BSA, and probed 3. Results
with polyclonal antibodies against caspase-9, caspase-8, PARP and
GAPDH. Immunoblotting was performed according to Cell Signal- 3.1. Selectivity of ABPs against breast cancer cells vs. normal cells
ing Technology–recommended procedures. The filter was visual-
ized by the Odyssey infrared imaging system. The IC50 of aurein 1.2, buforin IIb and BMAP-28 m was less than
8 mM in MCF-7 cells and less than 20 mM in MX-1 cells (Fig. 1A). On
2.10. Recombinant constructs and transfection of CHO-K1 cells the other hand, at a concentration of 60 mM, all three peptides only
displayed 20% cytotoxicity to mice hepatocytes. At their IC50
ST6Gal1 cDNA (GeneBank accession no. NM173216.2) was concentration, the three peptides showed lower hemolytic activity,
obtained from HepG2 cells by RT-PCR, and inserted into the vector especially for buforin IIb and BMAP-28 m (Fig. 1B). The relative
pcDNA 3.0. Recombinant constructs were transfected into CHO-K1 amount of sialic acid on the surface of MX-1 cells, MCF-7 cells,
cells using Lipofectamine 2000 according to the manufacturer’s hepatocytes and erythrocytes was determined by FITC-SNA, a
instructions. Changes in sialic acid levels were evaluated by specific lectin that binds to a-2,6 linked sialic acid. Flow cytometric
analyzing the FITC-SNA fluorescence by flow cytometry. The analysis showed that the a-2,6 linked sialic acid content in breast
cytotoxicity of aurein1.2, buforin IIb and BMAP-28 m to trans- cancer cells (MCF-7 and MX-1) was significantly higher than that
fected CHO-K1 cells was measured by the MTT assay. in normal hepatocytes and erythrocytes (Fig. 1C). These data
suggest that the breast cancer cells were more sensitive to these
2.11. In vivo assessment of the antitumor effect of buforin IIb in tumor three peptides than normal cells and the content of a-2,6 sialylated
xenografts glycoproteins and glycolipids on the surface of cells was much
higher in breast cancer cells than in normal cells.
MX-1 cells (2  106) resuspended in 0.1 ml of PBS were
transplanted subcutaneously into the lateral flanks of nude mice 3.2. Glycosylation enhances the binding of ABPs to the surface of cells
(BALB/c). When the tumors measured approximately 100 mm3, the
mice were randomly divided into three groups (five mice in each We determined the effect of glycoproteins and glycolipids on
group) and treated for 15 days as follows: group 1, buforin IIb at a the binding of ABPs to the surface of breast cancer cells. Four
dose of 2.5 mg/kg; group 2, buforin IIb at a dose of 5 mg/kg; and treatments were used in our study: sialidase, tunicamycin,
group 3, equivalent volumes of PBS. All compounds were injected BnGalNac and L-PPMP. Sialidase is a specific enzyme that
through the tail vein of mice on days 1, 4, 8, and 12. Body weight hydrolyses a-(2!3), a-(2!6) and a-(2!8) glycosidic linkages
 Y.-Y. Han et al. / Biochemical Pharmacology 86 (2013) 1254–1262 1257

Fig. 1. Peptide selectivity and the content of sialic acid. (A) Concentration response curves of ABPs for growth inhibition against MCF-7, MX-1 and mice hepatocytes by MTT
assay. (B) Hemolytic activity of peptides. (C) The content of sialic acid in MCF-7, MX-1, hepatocytes and erythrocytes by flow cytometry analysis. Cells were treated with FITC-
SNA (5 mg/ml) for 45 min and then FITC fluorescence was analyzed. Error bars represent standard deviation from mean cell viability as determined by three independent
experiments. All assays were performed as described in Section 2. Results are mean  SE of 4–6 different experiments.

of terminal sialic residues in oligosaccharides. Tunicamycin is a 3.4. Glycosylation affects apoptosis induced by ABPs
nucleoside antibiotic that blocks the synthesis of all N-linked
glycoproteins (N-glycans). BnGalNac is a specific inhibitor of O- The apoptotic activity of aurein 1.2, buforin IIb and BMAP-28 m
glycosylation and L-PPMP is a ganglioside biosynthesis inhibitor. in both MX-1 and MCF-7 cells was detected by annexin V-FITC/PI
The binding activity of aurein 1.2, buforin IIb and BMAP-28 m staining (Fig. 3A). We further investigated the effect of glycosyla-
differed between MX-1 and MCF-7 cells (Fig. 2A). A decline in the tion on ABP-induced apoptosis. Treatment of cells with sialidase,
binding activity of these peptides, as determined by changes in tunicamycin, BnGalNac and L-PPMP reduced the apoptotic rate
fluorescence intensity, was observed after treatment of cells with induced by aurein 1.2, buforin IIb and BMAP-28 m. In particular,
sialidase or the inhibitors listed above. Following the effect of MX-1 cells treated with glycosylation inhibitors showed 50%
sialidase on the binding activity to aurein 1.2, inhibition of O- reduction in the rate of apoptosis induced by buforin IIb treatment
glycosylation showed the strongest inhibitory effect on the (Fig. 3B). Sialic acid removal had the strongest inhibitory effect on
binding activity of buforin IIb and BMAP-28 m in both MX-1 and the apoptotic activity induced by most peptides, followed by N-
MCF-7 cells (Fig. 2B). Except for the binding of buforin IIb in MCF-7 linked glycoprotein inhibitors.
cells, inhibition of ganglioside biosynthesis had a stronger The obvious apoptosis of MCF-7 cells induced by buforin IIb was
inhibitory effect on binding activity than inhibition of N- also examined by DAPI staining (Fig. 3C). Cellular uptake of buforin
glycosylation in breast cancer cells. These results suggest that IIb was determined using FITC-labeled buforin IIb and Rhodamine
surface glycoproteins and gangliosides mediate the binding of 123, a specific probe for the localization of mitochondria in living
aurein 1.2, buforin IIb and BMAP-28 m to the surface of breast cells. The results showed that FITC labeled buforin IIb mainly
cancer cells. localized to the plasma membrane and mitochondria (Fig. 3D).
Buforin IIb also induced the activation of pro-caspase 9 (a caspase
3.3. Glycosylation promotes the cytotoxicity of ABPs that initiates mitochondria-dependent apoptosis) and the cleavage
of PARP (a caspase 3 substrate) in MCF-7 cells as determined by
Aurein 1.2, buforin IIb and BMAP-28 m exhibited cytotoxicity Western blot analysis (Fig. 3E), indicating that the effect of buforin
to MX-1 and MCF-7 cells in a dose-dependent manner and IIb may be mediated by the mitochondria-dependent apoptosis
inhibition of glycosylation also showed certain cytotoxicity to pathway in MCF-7 cells. We then assessed the effect of
both MX-1 and MCF-7 cells (Fig. 2C). However, when eliminating glycosylation inhibitors on the apoptotic pathway induced by
the cytotoxicity of inhibitors or using sialidase alone, inhibition buforin IIb. Although treatment with sialidase, tunicamycin, L-
of glycosylation reversed the cytotoxic effect of ABPs to PPMP or BnGalNac alone induced the activation of pro-caspase 9
different degrees. The effect of sialidase or inhibitors on the and the cleavage of PARP to a certain extent, a reduction in cleaved
cytotoxicity of peptides was stronger in MX-1 cell than in caspase 9 and PARP induced by buforin IIb was observed after
MCF-7 cells. In MCF-7 cells, the effect was as follows: treatment with inhibitors of O-, N-glycosylation, and ganglioside
sialidase > BnGalNac > Tunicamycin> L-PPMP, indicating that synthesis or sialic acid removal. These results indicate that
the sialic acid on the surface of MCF-7 cells has strongest effect glycosylation participates in the apoptotic pathway induced by
on the anticancer activity of aurein 1.2, buforin IIb and BMAP- buforin IIb.
28 m, followed by O-glycosylation. In MX-1 cells, the effect of O-
glycosylation was strongest for both aurein 1.2 and buforin IIb. 3.5. Up-regulated sialic acid increases the susceptibility of normal
Inhibition of O-glycosylation decreased the lethality induced by cells to ABPs
aurein 1.2, buforin IIb and BMAP-28 m by approximately 28% to
57% in MX-1 cells. Blocking the synthesis of all N-glycoproteins The effect of sialic acid on the cytotoxicity of aurein 1.2, buforin
resulted in a significant reduction in the lethality of ABPs, IIb and BMAP-28 m was assessed by upregulating sialic acid in
whereas the effect was lower than that of O-glycoprotein CHO-K1 cells. Cells transfected with hST6Gal1 showed higher
inhibitors, but stronger than inhibition of gangliosides, except fluorescence intensity than empty vector-transfected CHO-K1 cells
for BMAP-28 m in MX-1 cells. Inhibition of gangliosides resulted (107.19 vs. 46.76), which indicated a higher amount of bound SNA.
in a reduction of cytotoxicity of less than 23% for both aurein 1.2 Therefore, the transfection of hST6Gal1 upregulated the expression
and buforin IIb in MX-1 cells. Thus, the surface O-, N- of sialic acid on the CHO-K1 cell surface (Fig. 4A). All peptides
glycoproteins and gangliosides promotes the cytotoxicity of showed relative weak cytotoxicity against normal CHO-K1 cells.
aurein 1.2, buforin IIb and BMAP-28 m. However, in cells with up-regulated sialic acid, the lethality of
1258 Y.-Y. Han et al. / Biochemical Pharmacology 86 (2013) 1254–1262

Fig. 2. Effects of glycosylation on the binding and cytotoxicity of ABPs in breast cancer cells. (A) ABP fluorescence intensity on the surface of MX-1 and MCF-7 cells after
glycosylation inhibition. Cells were pretreated with the following treatments: sialidase (0.1 U/ml, 30 min), tunicamycin (2.5 mg/ml, 48 h), L-PPMP (1 mM, 48 h), BnGalNac
(2 mM, 48 h). Then cells were incubated with FITC labeled aurein1.2 (4 mM), buforin IIb (2 mM) and BMAP-28 m (4 mM) for 30 min at 37 8C and analyzed by flow cytometry.
(B) Reduced fluorescence density induced by the treatments in A. (C) Effect of glycosylation inhibition on the cytotoxicity of ABPs determined by MTT assay. Results are
mean  SE of 3 different experiments. *p < 0.05, significant difference comparing the lethality of peptides after glycosylation inhibition treatment to untreated controls.

buforin IIb increased to 92% at 40 mM, which was 28% higher than treatment group (Fig. 5D). In the TUNEL assay, apoptotic cells were
in controls. The lethality of aurein 1.2 increased from 9.89% to stained green and DAPI (blue) was used to show live cells. Control
19.99% at 20 mM, while the lethality of BMAP-28 m increased from tumors showed few TUNEL positive cells, whereas most cells in
32% to 50% at 50 mM (Fig. 4B). Thus, the cells with higher content of sections from tumors treated with 5 mg/kg buforin IIb stained
sialic acid were more susceptible to aurein 1.2, buforin IIb and positive. Histologic sections were probed against CD31, a well-
BMAP-28 m treatments at the same concentrations than the cells established marker for endothelial cells used as an indicator of the
with normal level of sialic acid. degree of vascularization, and the results showed an increased
number of CD31+ cells among the untreated tumor sections,
3.6. Binding and anticancer activity of buforin IIb against MX-1 whereas fewer CD31+ cells were detected in tumors treated with
tumors in vivo 5 mg/kg buforin IIb. Quantification of MVD showed a significant
reduction in MVD in xenografts treated with Buforin IIb (4.15  0.93
The in vivo anticancer activity of buforin IIb was assessed in an vessels/mm2 versus 14.67  0.95 vessels/mm2, p < 0.05).
MX-1 xenograft mouse model. All animals survived the experi- Staining of histologic sections with FITC-SNA showed that FITC
mental period, and there was no significant loss of body weight in fluorescence was significantly lower in mammary epithelial tissue
mice treated with buforin IIb in comparison with the controls from normal mice than in MX-1 xenograft tumors (Fig. 5E), and
(Fig. 5A). On the other hand, buforin IIb treatment significantly FITC-labeled buforin IIb was found mostly in the tumor tissue.
suppressed the growth of xenograft tumors (Fig. 5B and C). H&E Thus, after binding to the MX-1 xenograft tumors cells, buforin IIb
staining showed that buforin IIb induced nuclear shrinkage in the induces extensive apoptosis.
 Y.-Y. Han et al. / Biochemical Pharmacology 86 (2013) 1254–1262 1259

Fig. 3. Effects of glycosylation on apoptosis induced by aurein 1.2, buforin IIb and BMAP-28 m in breast cancer cells. MX-1 and MCF-7 cells were pretreated with different
inhibitors, and then treated with peptides. (A) Apoptosis analyses by annexin v-FITC/PI staining. (B) Reduced apoptosis rate calculated from A. (C) Buforin IIb-induced
apoptosis of MCF-7 cells as detected by DAPI staining. Cells were treated with PBS (left) or buforin IIb 8 mM (right) for 12 h. (D) Intracellular distribution of buforin IIb in MCF-
7 cells. Cells were incubated with 8 mM of FITC labeled buforin IIb for 6 h and stained with Rh123 (30 nM) for 30 min. Images show the Rh 123 fluorescence channel (left), FITC
channel (middle), and merged color images (right). (E) The effect of glycosylation on the activation of caspase-8, -9, and on cleavage of PARP induced by buforin IIb in MCF-7
cells. Cytosolic lysates were harvested after treatment for 12 h, and analyzed by Western blotting.

4. Discussion cells significantly promoted the binding activity of these peptides.

Inhibition of O-linked glycosylation reduced the binding activity of
We found that aurein 1.2, buforin IIb and BMAP-28 m exhibited ABPs to the greatest extent. O-linked glycans are found on
relatively higher cytotoxicity against breast cancer cells than approximately 30% of all cell surface proteins and, in particular, on
against normal cells. The selectivity of these peptides could be the membrane-bound mucins overexpressed in cancer [20,21].
related to the presence of specific molecules on the surface of Therefore, the large amount of O-glycoproteins, which were
breast cancer cells. We hypothesized that some glycosylated probably related to membrane-bound mucins, contributed largely
molecules may play a significant role in the anticancer activity of to the binding activity of ABPs to breast cancer cells. We found that
these peptides. Therefore, we first examined the effect of N-glycoproteins also promote the binding activity of peptides to
glycoproteins and glycolipids on the binding activity of ABPs to breast cancer cells. Angiogenesis is essential for progression of
the surface of breast cancer cells. We found that O-, N- breast cancer and other solid tumors, and metastasis is associated
glycoproteins and gangliosides on the surface of breast cancer with N-glycan expression; furthermore, high mannose type
1260 Y.-Y. Han et al. / Biochemical Pharmacology 86 (2013) 1254–1262

binding activity and anticancer activity. High content of glyco-

proteins and gangliosides could provide more binding targets for
peptides. Because of aberrant glycosylation, mucin glycoconju-
gates (the main class of O-glycosylated glycoproteins) and total
ganglioside levels are commonly significantly higher in breast
tumor tissue than in normal mammary tissue [26,27]. This may
lead to higher binding level of aurein 1.2, buforin IIb and BMAP-
28 m in breast cancer cells than in non-cancer cells. Indeed, in vivo
experiments confirmed that the binding levels of buforin IIb was
higher in mouse MX-1 xenograft tumors than in normal mammary
epithelial tissue. Therefore, the higher sensitivity of breast cancer
cells to ABPs, which is the result of a higher content of
glycoproteins and ganglioside, may partly explain the selectivity
of the tested peptides.
Among the glycans, a particularly important molecule is sialic
acid. It is negatively charged and usually found on the terminal
residue of O-, N-glycoproteins and gangliosides. We found that
removal of sialic acid from glycoproteins and gangliosides in MCF-
7 and MX-1 cells resulted in a decrease in binding, cytotoxicity and
apoptosis promoting activity of aurein 1.2, BMAP28 m and buforin
IIb. A similar result was reported in a study, in which sialidase-
treated U937 cells were less susceptible to permeabilization by
PMAP peptides [28]. Up-regulated expression of ST6Gal1, a key
enzyme responsible for the terminal sialylation of various
glycoproteins or glycolipids, enhanced the cytotoxicity of aurein
1.2, BMAP28 m and buforin IIb in CHO-K1 cells. Therefore, the sialic
Fig. 4. Effect of hST6Gal1 expression on the cytotoxicity of aurein 1.2, buforin IIb acid of glycoproteins and gangliosides acts as the key sugar
and BMAP-28 m in CHO-K1 cells. (A) The content of sialic acid was determined by mediating the binding and cytotoxicity of cationic ABPs. These
FITC-SNA. Cells were incubated with PBS (a) and FITC-SNA (b and c) for 45 min, and sialyl-glycoproteins or gangliosides may partly explain the effect of
then 106 cells were analyzed by flow cytometry. (B) The cytotoxicity of peptides
glycoproteins and gangliosides on the binding, cytotoxicity and
against the hST6Gal1-expressed CHO-K1 cells was assessed by the MTT assay.
Results are mean  SE of 3 different experiments. *p < 0.05. apoptosis promoting activity of cationic ABPs in breast cancer cells
in our study. A recent report showed that the cationic amphipathic
KLK peptide led to the accumulation of sialylated membrane
N-linked glycans are elevated in breast cancer cells [22,23]. In proteins both in cancer cells and normal cells by electrostatically
addition to glycoproteins, the presence of gangliosides on the interacting with negatively charged sialic acid moieties when used
surface can also contribute to the binding activity of ABPs, and its at very high concentrations [29]. Our data strongly suggest that
effect is stronger than that of N-glycosylation in breast cancer cells. sialic acid moieties of glycoproteins or gangliosides on the surface
Gangliosides are typically anchored in the outer leaflet of the of breast cancer cells provide important binding sites for cationic
plasma membrane and their sialoglycans, which extend outward ABPs and thus play a role in the peptide-membrane interaction.
from the cell surface, may participate in intermolecular interac- Compared to erythrocytes, hepatocytes and CHO-K1 cells, the
tions [24]. Gangliosides on the surface of Jurkat leukemia cells are content of a-2,6 linked sialic acid of glycoproteins or gangliosides
specific binding sites for buforin IIb and allow buforin IIb to was significantly higher in MCF-7 and MX-1 breast cancer cells.
distinguish cancer cells from normal cells [7]. Hence, O-, N- This was confirmed in histologic sections of MX-1 xenograft
glycoproteins and gangliosides on the surface are binding target tumors, in which the content of a-2,6 linked sialic acid was
proteins of ABPs to breast cancer cells. significantly higher than in mammary epithelial tissue. The higher
Further study showed that glycoproteins and gangliosides also content of a-2,6 linked sialic acid may be the result of an increase
significantly enhanced the cytotoxicity and apoptotic activity of in a-2,6-sialylation in breast cancer cells. Although the a-2,3
ABPs in breast cancer cells. They are involved in ABP-mediated linked sialic acid content was not detected in our study, a general
apoptosis and cytotoxic effects in breast cancer cells. We found increase in sialylation is often found in cell surface glycoproteins
that buforin IIb can induce caspase-dependent apoptosis in MCF-7 and glycolipids of malignant cells, including breast cancer cells,
cells, and the same apoptotic pathway induced by buforin IIb was which could significantly increase the potential for invasion and
also found in Jurcat cells [7]. Glycoproteins and gangliosides also metastasis [16,30]. The higher content of negatively charged sialic
participated in the caspase-dependent apoptosis induced by acid residues on the surface of breast cancer cells provides more
buforin IIb. This effect is likely to be the result of the enhanced binding sites for cationic ABPs, and thus plays an important role in
binding activity of ABPs induced by glycoproteins and gangliosides the cancer-selective toxicity of our tested peptides. In addition to
in breast cancer cells. Among N-, O-glycoproteins and gangliosides, sialic acid, sulfated carbohydrates are another type of glycans that
O-glycoproteins had the strongest effect on the cytotoxicity of the bear a net negative charge. It has been recently shown that
three peptides, which is in accordance with the above described sulphated glycans on the surface are primary target structures of
effect of O-glycoproteins on the binding activity of ABPs. These the NK-2 peptide against PC-3 prostate cancer cells [31]. However,
data also indicated a certain positive correlation between the other reports showed that the cytotoxic activity of small lytic
binding activity and cytotoxicity. peptides was increased or not affected by cell surface heparan
The plasma membrane is the first binding target of almost all sulfate [32].
antibacterial peptides and differences in binding activity contrib- Aurein 1.2, buforin IIb and BMAP-28 m showed obvious
ute largely to their anticancer activity and selectivity [25]. Upon cytotoxicity in breast cancer cells. Especially for buforin IIb, it
aurein 1.2, buforin IIb and BMAP-28 m interaction with the exerted promising anticancer activity in vivo in a MX-1 xenograft
membrane, both glycoproteins and ganglioside modulated their mouse model. Advances in cancer therapy have produced a new
 Y.-Y. Han et al. / Biochemical Pharmacology 86 (2013) 1254–1262 1261

Fig. 5. Effect of buforin IIb on MX-1 xenograft tumors. MX-1 bearing nude mice (n = 5) were treated with buforin IIb (2.5 mg/kg, 5.0 mg/kg) administered in 4 injections. Body
weight (A), tumor volume (B) and tumor weight (C) were determined. On day 15 post injection, tumors were carefully excised. Paraffin sections were prepared for H&E
staining, immunohistochemical staining using anti-CD31 antibodies, TUNEL staining (D). Paraffin sections were also analyzed by FITC-SNA and FITC labeled buforin IIb
staining (E). Representative images from both groups are shown. Results are mean  SE of 5 different experiments. *p < 0.05.

generation of innovative targeted cancer therapies to avoid the activity of aurein 1.2, buforin IIb and BMAP-28 m. Furthermore, the
side effects of traditional chemotherapy. Currently, several sialic acid moieties of glycoproteins or gangliosides on surface
molecules over-expressed have been explored as potential targets provide important glycan binding sites for cationic ABPs and thus
in breast cancer therapy, such as hormone receptor, human play a role in the peptide–membrane interaction. Anticancer ABPs
epidermal growth receptor 2 (HER2) [33]. Among them, the with cancer-selective cytotoxicity will be promising candidates for
treatment of HER2-overexpressing breast cancers is the most anticancer therapy in the future.
advanced in terms of personalized breast cancer therapy [34,35].
TACAs, especially membrane-bound mucins, sialyl-Lewis antigens, Acknowledgments
GM3, and GD3, which are the result of aberrant glycosylation, have
been the targets of interest for possible immunotherapy of cancer This work was financially supported by a grant of the National
[36,37]. In our study, we found that antibacterial peptides target Natural Science Foundation of China (Grant No. 30900743), the
glycoproteins and gangliosides, especially sialyl-glycoproteins and Natural Science Foundation of Jiangsu Province of China (Grant No.
gangliosides in MCF-7 and MX-1 breast cancer cells. Besides MCF-7 BK2011368) and the Priority Academic Program Development of
and MX-1 cells, breast tumor tissues and many other breast cancer Jiangsu Higher Education Institutions (PAPD).
cells such as BT20, T47D, MDA-MB-231 cells also show increased
expression of glycoproteins (especially mucins) and gangliosides References
[38]. This suggests that antibacterial peptides are promising
candidates for targeting glycosylated molecules in breast person- [1] Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics.
CA Cancer J Clin 2011;61:69–90.
alized cancer therapy. Further study is needed to further enhance
[2] Riedl S, Zweytick D, Lohner K. Membrane-active host defense peptides –
selectivity and diminish the side effects of these agents. challenges and perspectives for the development of novel anticancer drugs.
In conclusion, our results clearly demonstrate that O-, N- Chem Phys Lipids 2011;164:766–81.
glycoproteins and gangliosides on the surface of breast cancer cells [3] Papo N. Shai Y.Host defense peptides as new weapons in cancer treatment. Cell
Mol Life Sci 2005;62:784–90.
are specific binding molecules for aurein 1.2, buforin IIb and BMAP- [4] Hoskin DW, Ramamoorthy A. Studies on anticancer activities of antimicrobial
28 m. Glycosylation may play an important role in the anticancer peptides. Biochim Biophys Acta 2008;1778:357–75.
1262 Y.-Y. Han et al. / Biochemical Pharmacology 86 (2013) 1254–1262

[5] Schweizer. F.Cationic amphiphilic peptides with cancer-selective toxicity. Eur [22] Banerjee DK. N-glycans in cell survival and death: cross-talk between glyco-
J Pharmacol 2009;625:190–4. syltransferases. Biochim Biophys Acta 2012;1820:1338–46.
[6] Dennison SR, Harris F, Phoenix DA. The interactions of aurein 1.2 with cancer [23] De Leoz ML, Young LJ, An HJ, Kronewitter SR, Kim J, Miyamoto S, et al. High-
cell membranes. Biophys Chem 2007;127:78–83. mannose glycans are elevated during breast cancer progression. Mol Cell
[7] Lee HS, Park CB, Kim JM, Jang SA, Park IY, Kim MS, et al. Mechanism of Proteomics 2011;10:1–9.
anticancer activity of buforin IIb, a histone H2A-derived peptide. Cancer Lett [24] Lopez PH, Schnaar RL. Gangliosides in cell recognition and membrane protein
2008;271:47–55. regulation. Curr Opin Struct Biol 2009;19:549–57.
[8] Ahmad A, Asthana N, Azmi S, Srivastava RM, Pandey BK, Yadav V, et al. [25] Chen YQ, Min C, Sang M, Han YY, Ma X, Xue XQ, et al. A cationic amphiphilic
Structure-function study of cathelicidin-derived bovine antimicrobial peptide peptide ABP-CM4 exhibits selective cytotoxicity against leukemia cells. Pep-
BMAP-28: design of its cell-selective analogs by amino acid substitutions in tides 2010;31(8):1504–10.
the heptad repeat sequences. Biochim Biophys Acta 2009;1788:2411–20. [26] Mukhopadhyay P, Chakraborty S, Ponnusamy MP, Lakshmanan I, Jain M, Batra
[9] Zaiou M. Multifunctional antimicrobial peptides: therapeutic targets in several SK. Mucins in the pathogenesis of breast cancer: implications in diagnosis,
human diseases. J Mol Med (Berl) 2007;85:317–29. prognosis and therapy. Biochim Biophys Acta 2011;1815:224–40.
[10] Park JH, Jeong YJ, Park KK, Cho HJ, Chung IK, Min KS, et al. Melittin suppresses [27] Marquina G, Waki H, Fernandez LE, Kon K, Carr A, Valiente O, et al.
PMA-induced tumor cell invasion by inhibiting NF-kappaB and AP-1-depen- Gangliosides expressed in human breast cancer. Cancer Res 1996;56:
dent MMP-9 expression. Mol Cells 2010;29:209–15. 5165–71.
[11] Jo M, Park MH, Kollipara PS, An BJ, Song HS, Han SB, et al. Anti-cancer effect of [28] Risso A, Zanetti M, Gennaro R. Cytotoxicity and apoptosis mediated by two
bee venom toxin and melittin in ovarian cancer cells through induction of peptides of innate immunity. Cell Immunol 1998;189:107–15.
death receptors and inhibition of JAK2/STAT3 pathway. Toxicol Appl Pharma- [29] Weghuber J, Aichinger MC, Brameshuber M, Wieser S, Ruprecht V, Plochberger
col 2012;258:72–81. B, et al. Cationic amphipathic peptides accumulate sialylated proteins and
[12] Teixeira V, Feio MJ, Bastos M. Role of lipids in the interaction of antimicrobial lipids in the plasma membrane of eukaryotic host cells. Biochim Biophys Acta
peptides with membranes. Prog Lipid Res 2012;51:149–77. 2011;1808:2581–90.
[13] Schröder-Borm H, Bakalova R, Andrä J. The NK-lysin derived peptide NK-2 [30] Miyagi T, Wada T, Yamaguchi K, Hata K. Sialidase and malignancy. A minire-
preferentially kill cancer cells with increased surface levels of negatively view. Glycoconj J 2004;20:189–98.
charged phosphatidylserine. FEBS Lett 2005;579:6128–34. [31] Gross S, Andrä J. Anticancer peptide NK-2 targets cell surface sulphated
[14] Iwasaki T, Ishibashi J, Tanaka H, Sato M, Asaoka A, Taylor D, et al. Selective glycans rather than sialic acids. Biol Chem 2012;393:817–27.
cancer cell cytotoxicity of enantiomeric 9-mer peptides derived from beetle [32] Fadnes B, Uhlin-Hansen L, Lindin I, Rekdal Ø. Small lytic peptides escape the
defensins depends on negatively charged phosphatidylserine on the cell inhibitory effect of heparan sulfate on the surface of cancer cells. BMC Cancer
surface. Peptides 2009;30:660–8. 2011;11:116.
[15] Papo N, Shai Y. New lytic peptides based on the D,L-ABP hipathic helix motif [33] Nowsheen S, Aziz K, Panayiotidis MI, Georgakilas AG. Molecular markers for
preferentially kill tumor cells compared to normal cells. Biochemistry cancer prognosis and treatment: have we struck gold. Cancer Lett 2012;327:
2003;42:9346–54. 142–52.
[16] Wang PH. Altered sialylation and its roles in gynecologic cancers. J Cancer Mol [34] Marmé F, Schneeweiss A. Personalized therapy in breast cancer. Onkologie
2006;2:107–16. 2012;35:28–33.
[17] Ohyama C. Glycosylation in bladder cancer. Int J Clin Oncol 2008;13:308–13. [35] Harbeck N, Salem M, Nitz U, Gluz O, Liedtke C. Personalized treatment of early-
[18] Lityńska A, Przybyło M, Hoja-Łukowicz D, Pocheć E, Kremser M, Ciołczyk- stage breast cancer: present concepts and future directions. Cancer Treat Rev
Wierzbicka D, et al. Protein glycosylation as marker of melanoma progression. 2010;36:584–94.
Curr Cancer Thera Rev 2008;4:144–8. [36] Raina D, Ahmad R, Joshi MD, Yin L, Wu Z, Kawano T, et al. Direct targeting of the
[19] Cazet A, Julien S, Bobowski M, Burchell J, Delannoy P. Tumour-associated mucin 1 oncoprotein blocks survival and tumorigenicity of human breast
carbohydrate antigens in breast cancer. Breast Cancer Res 2010;12:204–16. carcinoma cells. Cancer Res 2009;69:5133–41.
[20] Lavrsen K, Madsen CB, Rasch MG, Woetmann A, Odum N, Mandel U, et al. [37] Heimburg-Molinaro J, Lum M, Vijay G, Jain M, Almogren A, Rittenhouse-
Aberrantly glycosylated MUC1 is expressed on the surface of breast cancer Olson K. Cancer vaccines and carbohydrate epitopes. Vaccine 2011;29:
cells and a target for antibody-dependent cell-mediated cytotoxicity. Glyco- 8802–26.
conj J 2013;30:227–36. [38] Walsh MD, Luckie SM, Cummings MC, Antalis TM, McGuckin MA. Heteroge-
[21] Beatson RE, Taylor-Papadimitriou J, Burchell JM. MUC1 immunotherapy. neity of MUC1 expression by human breast carcinoma cell lines in vivo and in
Immunotherapy 2010;2:305–27. vitro. Breast Cancer Res Treat 1999;58:253–64.