Journal of Dermatological Science 62 (2011) 22–35

Benzo[c]phenanthridine alkaloids exhibit strong anti-proliferative activity inmalignant melanoma cells regardless of their p53 status

Jindriska Hammerova a, Stjepan Uldrijan a, Eva Taborska b, Iva Slaninova a,*a Department of Biology, Faculty of Medicine, Masaryk University, Kamenice 5, Building A6, 62500 Brno, Czech Republicb Department of Biochemistry, Faculty of Medicine, Masaryk University, Kamenice 5, Building A16, 62500 Brno, Czech Republic

A R T I C L E I N F O

Article history:

Received 29 September 2010

Received in revised form 11 January 2011

Accepted 14 January 2011

Keywords:

Apoptosis

Benzo[c]phenanthridine alkaloids

DNA damage

Melanoma

p53

A B S T R A C T

Background: Search for new substances with antiproliferative activity towards melanoma cells is

important since malignant melanoma is notoriously resistant to conventional chemotherapy.

Benzo[c]phenanthridine alkaloids (BAs) are natural products with significant anti-proliferative

activities, therefore they are considered as agents promising for cancer therapy.

Objectives: The effects of five BAs (sanguinarine, chelerythrine, chelidonine, sanguilutine, and

chelilutine) on human malignant melanoma cell lines were compared. The study focused on BAs

effects on DNA, anti-apoptotic and p53 protein levels; and the involvement of p53 in cellular responses

to alkaloids treatment.

Methods: Melanoma cell lines, two wild types and two with dysfunctional p53 derived from one of them

were used. The mechanism of anti-proliferative and pro-apoptotic effects and the effect on DNA was

investigated using MTT assay, flow cytometry, Western blot analysis, fluorescence and electron

microscopy.

Results: All tested alkaloids exhibit strong anti-proliferative activity. CHL, CHE and SA induced apoptosis,

which was probably mediated by decreasing levels of anti-apoptotic proteins (Bcl-xL, Mcl-1, XIAP) and

was accompanied by mitochondrial membrane potential decrease as well as caspase-3 and PARP

cleavage. Although all alkaloids caused DNA damage, which was demonstrated by induction of H2AX

phosphorylation, none of the tested alkaloids stabilised p53 and their toxicity in cells with non-

functional p53 was comparable to wild type cells.

Conclusion: Despite the profound similarity of BAs molecular structures, it is clear that the mechanism of

cell death induction is different for each alkaloid. Our results indicate that BAs could be effective in

malignant melanoma treatment, including tumours which have lost wild type p53.

� 2011 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights

reserved.

Contents lists available at ScienceDirect

Journal of Dermatological Science

journa l homepage: www.e lsev ier .com/ jds

1. Introduction

Over the last decades, cutaneous melanoma has shownincreasing incidence rates and, in Caucasian populations, it hasbecome the cancer with the fastest growing incidence rate. Cancerstatistics in the United States and central Europe revealed morethan a threefold increase in the incidence rates of malignantmelanoma between 1970 and 2000. Cohort studies from severalcountries indicate that the trend of increasing incidence rates willcontinue in the future [1].

Inoperable metastatic melanoma is difficult to treat because itis usually resistant to conventional chemotherapy. The mediansurvival time is only 6 months, and 5-year survival rates are less

* Corresponding author. Tel.: +420 549496985; fax: +420 549491327.

E-mail addresses: ipokorna@med.muni.cz, islaninova@seznam.cz (I. Slaninova).

0923-1811/$36.00 � 2011 Japanese Society for Investigative Dermatology. Published b

doi:10.1016/j.jdermsci.2011.01.006

than 5% [2]. Dacarbazine, temozolomide and cisplatin are used assingle chemotherapeutics agents in melanoma treatment; al-though, none of them produces response rates of more than 25%.The most widely used chemotherapy combinations includecisplatin, vinblastine, and dacarbazine with response rates rangingfrom 30% to 50% [3]. Therefore, identification of new compoundsthat can stop the proliferation of malignant melanoma cellsremains an important target in the field of cancer research.Secondary plant metabolites belong also to the group of moleculeswith antiproliferative potential, among them are benzo[c]phenan-thridine alkaloids (BAs).

BAs are a relatively small group of isochinoline alkaloids, whichoccur in many plant species of the families Papaveraceae,Fumariaceae, Ranunculaceae and Rutaceae. The most common areBAs with quaternary nitrogen atom whose main representativesare commercially available sanguinarine (SA) and chelerythrine(CHE). Structurally very similar, but with fully hydrogenated rings

y Elsevier Ireland Ltd. All rights reserved.

[()TD$FIG]

NCH3

R5

R4

R3

R2

R1

R6R6R5 R4 R3 R2 R1

Sanguinarine -OCH2 -OCHO- 2 H H O-

Chelerythrine -OCH2 OCHO- 3 OCH 3 H H

Sanguilutine OCH3 OCH3 OCH 3 OCH 3 OCH3 H

Chelilutine -OCH2 OCHO- 3 OCH 3 OCH3 H

Chelidonine

Fig. 1. Chemical structure of tested BAs.

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–35 23

B and C is alkaloid chelidonine (CHLD) that is the main alkaloid ofgreater celandine (Chelidonium majus; Fig. 1). The lesser known BAsare sanguilutine (SL) and chelilutine (CHL), which are studied inthis work and which were also studied in our previous works [4,5]and sanguirubine, chelirubine and macarpine which were thesubjects of our previous study [6]. All of last mentioned alkaloidscontain quaternary nitrogen. They receive little attention and assuch, little is known about their biological actions. Herbal extractsfrom many plants containing SA, CHE and CHLD are often used infolk medicine for their antifungal, antimicrobial and anti-inflammatory effects. Mixtures of SA and CHE are used as activeingredients of toothpastes or mouthwashes thanks to theirantiplaque effect [7,8]. Ukrain, a semi-synthetic derivative ofCHLD has been used as a potent anticancer drug [9].

In several publications it was described that BAs are able toinduce apoptosis, which is an advantageous property for sub-stances used in cancer therapy. Tumour cells often have defects inapoptotic pathways, leading to increased resistance to chemo-therapy. One of the mechanisms of the apoptosis induction by BAsmay be the inhibition of members of Bcl-2 antiapoptotic proteinfamily, as has been suggested by several authors [10–15].However, the exact mechanism of BAs action is not known.Quaternary BAs, as a result of their chemical structure (planarmolecule, iminium bond), are able to interact with various cellularcomponents, including DNA [16,17]. On the other hand, the DNA-binding capacity of chelidonine was shown to be negligible,despite its ability to induce apoptosis [18–20].

The known ability of quaternary alkaloids to interact with DNAevokes the question to what extent they cause DNA damage andhow this damage affects the cell response. In response to DNAdamage p53 plays a key role. Tumour suppressor p53 as acomponent of stress response plays a key role in protection againstcancer development [21]. Loss or mutation of p53 are found inabout 50% of human cancers, and increased levels of its negativeregulators MDM2 and MDM4 (MDMX) in a large proportion of theremaining cancers [22]. p53 is extremely sensitive to even lowlevels of DNA damage. However it was suggested that the responseto oncogene activation, but not to DNA damage is crucial for p53-mediated tumour suppression. Other studies indicate thatgenotoxic stress is a key signal to activate p53 ([21] and citationstherein). Tumour cells with mutated or non-functional p53 aremore resistant to the effects of commonly used chemotherapy, andidentification of substances with cytostatic effects, independent offunctional p53, could lead to more efficient treatments for a largeproportion of human cancers.

The objective of our present work was to determine and comparethe antiproliferative effects of five BAs – three major BAs (SA, CHEand CHLD) and two of their minor derivatives (SL and CHL) – inhuman malignant melanoma cell lines with both functional andinactivated p53 and to study the mechanism of action with regard toinduction of DNA damage, apoptosis and the role of tumoursuppressor p53 in the cellular response to the tested BAs.

2. Materials and methods

2.1. Drugs and reagents

MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazoliumbromide], dimethyl sulphoxide (DMSO), propidium iodide (PI),camptothecine, z-VAD-fmk, JC-1 (5,5 V, 6,6 V-tetrachloro-1,1 V, 3,3V-tetraethyl-imidacarbocyanine iodide); 2� Laemmli buffer,Triton X-100, paraformaldehyde and DAPI (4,6-diamino-2-pheny-lindole) were obtained from Sigma–Aldrich, Inc. DMEM medium(Dulbecco’s Modification of Eagle’s Medium), foetal bovine serum,glutamine, penicillin and streptomycin were obtained from PAALaboratories (Pasching, Austria). RPMI and DMEM/F12 1:1 (GibcoDulbecco’s Modification of Eagle’s Medium DMEM/F12 (Ham) 1�)media, Lipofectamine 2000 transfection reagent and G418sulphate were obtained from Invitrogen, U.K. Doxorubicin wasobtained from AppliChem GmbH (Darmstadt, Germany). Vecta-shield mounting medium was obtained from Vector Laboratories,Inc., USA.

2.2. Alkaloids

Sanguinarine (SA) chloride, chelerythrine (CHE) chloride,chelidonine (CHLD) chloride, sanguilutine (SL) chloride andchelilutine (CHL) chloride were isolated at the Department ofBiochemistry of our Faculty. SA and CHE were isolated fromDicranostigma lactucoides (Hook. F. et Thoms.) [23]; SL and CHLwere isolated from Sanguinaria canadensis (L.) [24] and CHLD wasisolated from C. majus [25] (Fig. 1). Alkaloids were dissolved indistilled water at a concentration 1 mg/ml (stock solutions) andstored at �20 8C.

2.3. Cell lines and culture conditions

Malignant melanoma cell lines were used in the experiments.A-375 and SK-MEL-2 cells have wild type p53. We have alsocreated and used two cell lines with dysfunctional p53 derived

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–3524

from the line A-375. A-375-p53DD in which p53 function wasblocked at the protein level by protein–protein interactions and A-375-p53sh in which p53 expression was inhibited at the mRNAlevel, using shRNA (short hairpin RNA) technology.

A-375-p53DD cells were derived from A-375 parental cell lineby stable transfection with plasmid pCMVneoDDp53 coding for adominant-negative truncated p53 protein which consists of theamino acid residues 1–14 and 302–390 of mouse p53 expressedunder the control of the CMV promoter [26]. Transfections wereperformed using Lipofectamine 2000 transfection reagent andstable transfectants were selected with G418 sulphate (1 mg/ml)in the growth medium. As previously observed in another wildtype p53-expressing cancer cell line (MCF7) [27], the expression ofthe DD miniprotein led to stabilisation of endogenous wild typep53 in stably transfected cells. However, p53 stabilised in DD-expressing cells is unable to efficiently induce expression of p53-target genes in response to p53-activating stimuli such as DNAdamage or transcription inhibition.

In A-375-p53sh cells expression of p53 was suppressed by astable transfection with pSUPER.p53 vector (OligoEngine) (expres-sing shRNA), using Lipofectamine 2000 reagent. Cells were treatedwith puromycin (5 mg/ml) 24 h after transfection to establish astable cell line expressing p53 shRNA. To determine the amount ofp53 protein expressed, a Western blot analysis was performedusing lysates of cells treated with doxorubicin (1 mg/ml) andcamptothecine (10 mM) for 24 h. The analysis revealed a substan-tial decrease in p53 protein levels in A-375-p53sh cells comparedto A-375 cells.

The A-375 cell line was purchased from the European Collectionof Animal Culture (ECACC, Salisbury, U.K.), SK-MEL-2 was kindlyprovided by Dr. Hajduch (Palacky University, Olomouc, CzechRepublic). The A-375 and A-375-p53sh cell lines were grown inRPMI medium, the A-375-p53DD cell line was grown in DMEMmedium, and the SK-MEL-2 cell line was grown in DMEM/F12 1:1medium. In each case, the growth medium was supplemented with10% foetal bovine serum, 2 mM glutamine, 100 IU/ml penicillinand 100 mg/ml streptomycin. All cultures were grown at 37 8Cunder 5% CO2 in a high-humidity-atmosphere and subculturedthree times a week.

2.4. Cytotoxicity assays

2.4.1. MTT

MTT assays were performed in 96-well plates (Nunc A/S,Rockilde, Denmark). Individual wells were seeded with 200 ml ofcells (1 � 104/ml) and incubated at 37 8C in a humidifiedatmosphere (5% CO2). After 24 h of incubation, the medium wasreplaced with medium containing alkaloids at concentrations(0.01, 0.03, 0.1, 0.3, 1, 1.5, 2 and 3). The MTT assay was performed48 h after alkaloid addition: 20 ml of MTT (5 mg/ml in PBS) wasadded to each well and incubated for 4 h under culture conditions.At the end of the incubation period the medium was removed, theformazan crystals were dissolved in 200 ml DMSO, and the opticaldensity was measured at 570 nm using a SLT Spectra ShellMicroplate Reader (SLT-Lab Instruments GmBH, Saltsburg,Austria). The anti-proliferative effect was expressed as the IC50

(i.e. the concentration that induces a 50% inhibition of metabolicactivity compared to controls). Each concentration of eachcompound was examined in four replicate wells. The experimentswere independently repeated three times.

2.4.2. Cell viability assay

The cell viability assay is based on the exclusion of propidiumiodide (PI) by intact viable cells. Melanoma cells grown on 24-welltissue culture test plates (Orange Scientific, Belgium) to �80–90%confluence were treated for 48 h with alkaloids at concentrations

0.1, 0.5, 1, and 3 mg/ml in 0.5 ml medium/well. Thereafter, cellswere harvested, washed in PBS and PI (1 mg/ml PBS) was added.Immediately after PI addition, the percentage of dead (PI-stained)cells was detected using a Cytomics FC 500 flow cytometry system(Beckman Coulter, Inc., CA, USA) at channel FL3 (emission at620 nm). 10,000 cells per sample were analyzed.

2.4.3. The effect of inhibitor of caspases, z-VAD-fmk

To determine if activation of caspases is an essential step in thecell death induced by the studied alkaloids (0.5, 0.7, 1, 2, 3 and4 mg/ml), A-375 cells were pre-treated with 20 mM z-VAD-fmk, acell permeable broad spectrum caspase inhibitor, for 3 h and thentreated with BAs for an additional 45 h. Cell viability was assessedusing PI exclusion assay (Section 2.4.2).

2.4.4. Growth curves – xCELLigence

The xCELLigence System (Roche Diagnostics GmbH, Mannheim,Germany) was used to monitor the dynamics of cytotoxic effects ofBAs on A-375 melanoma cells. The xCELLigence System is amicroelectronic biosensor system for real-time, label free cellularanalysis. This system allows monitoring of cellular events in realtime by measuring electrical impedance across interdigitatedmicro-electrodes on the bottom of special tissue culture plates.2000 cells per well (in 100 ml) were seeded into each well of an E-plate 16 (Roche Diagnostics GmbH, Mannheim, Germany). Theplate was placed into an analyzer holder, which was then insertedinto an incubator. After 24 h the medium was exchanged formedium containing BAs at concentrations of 0.1, 0.3, 0.5, 1 or1.5 mg/ml. Each sample was in duplicate. Data were collected atthe following intervals: every one minute for the first hour ofexperiment, every 15 min for the next 23 h; after alkaloid addition,every 1 min for 24 h and every 30 min for the remainder of theexperiment. Total duration of the experiment was 72 h. RTCASoftware 1.0 was used for both data acquisition and data analysis.Data are presented as a dimensionless parameter called the CellIndex (CI; derived as a relative change in measured electricalimpedance).

2.5. Assessment of mitochondrial membrane potential

JC-1 was used to measure the mitochondrial membranepotential (MMP). JC-1 is a cationic dye that exhibits potential-dependent accumulation in mitochondria, indicated by a fluores-cence emission shift from green (525 � 10 nm) to red(610 � 10 nm). Melanoma cells (1 � 105 cells/ml) growing in 4-cmdishes, in a total volume of 2 ml, were treated for 24 h with alkaloids(1, 2 and 3 mg/ml) or cytostatic agents doxorubicin (1 mg/ml) orcamptothecin (10 mM), which was used as a positive controls. Aftertreatment, cells were collected, washed in PBS and stained with JC-1(2 mg/ml) at 37 8C for 20 min and then analyzed using the CytomicsFC 500 flow cytometer. Green fluorescence was analyzed using theFL1 channel (530 nm) and red fluorescence using the FL2 channel(585 nm). 10,000 cells per sample were analyzed. In healthy non-apoptotic cells, JC-1 is present as a green monomer in the cytosol andas red aggregates in the mitochondria. Whereas in apoptotic cellswith altered mitochondrial membranes, JC-1 is present only in themonomeric green form in the cytosol. Therefore, cells with increasedgreen fluorescence (FL1 channel) were assessed as apoptotic. Theexperiments were repeated three times.

2.6. SDS-PAGE and Western blot analysis

Cells (�1 � 105 cells/ml) were cultivated in the presence of thealkaloids (0.5–3 mg/ml) or positive controls (doxorubicin (1 mg/ml) or camptothecin (10 mM)) in 12-well tissue culture plates(0.5 ml medium/well) for 0.5, 1, 2, 3, 6, 8 or 12 h. After treatment,

[()TD$FIG]

Fig. 2. Growth curves obtained on xCELLigence System showing kinetics of treatment of A-375 cells with individual alkaloids at various concentrations. SA (0.1, 0.3, 0.5, 1, and

1.5 mg/ml) (A); CHLD (0.1, 0.5, 1, and 1.5 mg/ml) (B); SL (0.1, 0.3, and 0.5 mg/ml) (C); CHE (0.1, 0.5, 1, 1.5, and 2 mg/ml) (D); CHL (0.1, 0.3, 0.5, and 1 mg/ml) (E). SA, CHLD and SL

show no toxicity at one concentration and high toxicity at a concentration only slightly higher (2–5 multiple). The effect of CHE and CHL was more gradual; increasing

concentrations led to a smooth, continual decline in the CI (D and E). CI (Cell Index; derived as a relative change in measured electrical impedance). The scale of CI is 0.5–3.0 for

SA and CHE, 0.5–3.5 for SL and CHL and 0.5–4.0 for CHLD.

[()TD$FIG]

20

40

60

80

100

of cells

with d

ecre

ased M

MP

*

* *

** *

*

*

* * *

**

**

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–35 25

the cells were lysed in 2� Laemmli buffer. Total cell lysates wereseparated by electrophoresis on 10, 12.5 or 15% SDS-polyacryl-amide gels and then transferred to a polyvinylidene difluoridemembrane (PVDF; Millipore, Billerica, MA, USA). The membranewas blocked with 5% non-fat milk in TBS (10 mM Tris–HCl, 100 nMNaCl and 0.05% Tween 20; pH = 7.4) for 1 h at room temperature.After blocking, the membrane was incubated with primaryantibodies (using dilutions recommended by the manufacturer)overnight at 4 8C. The following primary antibodies were used:rabbit monoclonal anti-cleaved caspase-3 and anti-Bcl-xL, mousemonoclonal anti-cleaved PARP, rabbit polyclonal anti-XIAP (CellSignalling Technology, Inc.), anti-Mcl-1 (Sigma–Aldrich, Inc.), andmouse monoclonal anti-gH2AX (Ser139) (Biolegend, Europe,Netherlands). Mouse monoclonal anti-p53 (DO-1; staining totalp53), anti-PCNA (PC10) [28] and anti-p21 [29] antibodies werekindly provided by Dr. Vojtesek (Masaryk Memorial CancerInstitute, Brno). After incubation with the appropriate primaryantibody, sheep anti-mouse (GE Healthcare, U.K.) or donkey anti-rabbit (Santa Cruz Biotechnology, Inc.) secondary horseradishperoxidase conjugated antibodies were applied (1 h; roomtemperature). Visualization was done on blue sensitive MedicalX-ray film (Agfa, Belgium) using an enhanced chemiluminescencekit (ECL, Amersham Life Science, Inc.). To confirm equal loading, thesame blot was stripped and reprobed with anti-PCNA antibody.Every experiment was repeated three times. Densities of bandswere calculated with Image J software (http://rsbweb.nih.gov/ij/index.html). The measured values for the studied protein levelswere related to PCNA internal loading control. Ratio of banddensities of protein of interest normalized to corresponding PCNAand control normalized to corresponding PCNA was calculated.

Table 1The IC50 values of alkaloids.

Alkaloid/cell line IC50 (mg/ml)

A-375 A-375-p53DD A-375-p53sh SK-MEL-2

SA 0.434� 0.063 0.416�0.061 0.296� 0.004 0.648� 0.017

CHE 1.115� 0.054 1.470�0.184* 0.832� 0.065 >3

SL 0.188� 0.053 0.232�0.017 0.222� 0.038 0.991� 0.070

CHL 0.521� 0.084 0.860�0.186* 0.634� 0.004 1.967� 0.047

CHLD 0.910� 0.017 0.634�0.009* 0.772� 0.045* >3

Data are means� S.E.M. of three independent experiments.* p<0.05 A-375 vs. A-375-p53DD or A-375 vs. A-375-p53sh.

2.7. Fluorescence microscopy

2.7.1. DAPI staining of nuclei

A-375 cells, grown on slides treated with SA or CHE at aconcentration of 1.5 mg/ml for 24 h, were fixed with 3%paraformaldehyde in PBS for 20 min at room temperature andafter washing in PBS, stained with DAPI (1 mg/ml) for 1 h at roomtemperature in the dark. The slides were mounted in Vectashieldmounting medium. Microscopic analysis was performed using anOlympus FluoViewTM 500 confocal laser scanning fluorescencemicroscope (Olympus C&S Ltd., Prague, Czech Republic).

2.7.2. Immunofluorescence of gH2AX

A-375 cells, grown 24 h on slides in the presence of SA (0.5–2 mg/ml) or CHE (3 mg/ml), were fixed with 3% paraformaldehydein PBS for 20 min at room temperature, permeabilized with 0.2%Triton X-100 in PBS for 5 min and preincubated 20 min in 0.5% BSAin PBS. The cells were stained with primary anti-Ser139-phosphorylated H2AX antibody for 120 min. DyLightTM594 conju-

0

Ctr

l

SA

1

SA

2

SA

3

CH

E1

CH

E2

CH

E3

CH

LD

1

CH

LD

2

CH

LD

3

SL1

SL2

SL3

CH

L1

CH

L2

CH

L3

DO

XO

CA

MP

T

%

Fig. 3. Determination of changes in mitochondrial membrane potential of A-375

cells after 24 h treatment with alkaloids at concentrations 1, 2 and 3 mg/ml. Two-

dimensional flow cytometry analysis using the FL1 channel (green, 530 nm) and FL2

channel (red, 585 nm) was used. In non-apoptotic cells, JC-1 is present as a green

monomer in the cytosol and red aggregates in the mitochondria, whereas in

apoptotic cells with altered mitochondrial membrane, JC-1 is present only in

monomeric green form. Graph shows percentage of cells with increased green

fluorescence that was assessed as apoptotic. Data are means � S.E.M. of at least three

independent experiments. *p < 0.05 vs. untreated control. Doxorubicin (1 mg/ml;

DOXO) or camptothecin (10 mM; CAMPT) was used as positive controls.

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–3526

gated AffiniPure Donkey Anti-Mouse IgG (H + L; Jackson ImmunoResearch Laboratories, Inc., USA) was used as the secondaryantibody. DAPI at final concentration of 1 mg/ml was added to thesolution of the secondary antibody for visualization of nuclei. Theslides were mounted and observed as described in the previoussection.

2.8. Electron microscopy

Cells were processed as was described in our previous paper [6].Briefly, after 24-h treatment with SA or CHE, at a concentration of1.5 mg/ml, cells were harvested, washed three times in 0.1 Mcacodylate buffer (pH = 6.98), and fixed in 3% glutaraldehydesolution in 0.1 M cacodylate buffer (containing 0.2 M saccharose)for 2.5 h, and post-fixed in 1% osmium tetroxide in 0.1 Mcacodylate buffer for 2.5 h at room temperature. The sectionswere stained with 2.5% uranyl acetate (6 min) and lead citrate

[()TD$FIG]

Fig. 4. Western blot analysis of cleaved caspase-3 (c-cas-3) and cleaved PARP (c-PARP) of A

treatment with BAs at concentrations 0.1, 1, 2, and 3 mg/ml. Cleaved caspase-3 was dete

after 6 h treatment (C–E and G). No significant difference in caspase-3 or PARP cleavage w

a 15% (caspase-3) and 10% (PARP) SDS gels. PCNA was used as a loading marker. Num

corresponding PCNA and control normalized to corresponding PCNA.

(3 min), and observed with a transmission electron microscopePhilips Morgagni (FEI Company, Eindhoven, Netherlands).

2.9. Statistical analysis

At least three independent experiments were performed underidentical conditions. Results were analyzed using Student’s t-testand the threshold for statistical significance was set at p < 0.05.

3. Results

3.1. Cytotoxic effect of BAs on melanoma cell lines

3.1.1. Antiproliferative activities

Anti-proliferative activity of BAs after 48 h of treatment wastested using a colorimetric MTT assay, which detects metabolicallyactive cells (specifically, the activity of mitochondrial succinate

-375 (A, C, E, and H–J), A-375-p53DD (B and D) and A-375-p53sh (F and G) cells after

cted after 3 h (A, B, and F) and after 0.5, 1 and 2 h (H–J) treatment and cleaved PARP

as observed between tested melanoma cell lines. Total cell lysates were separated on

eric values represent ratio of band densities of protein of interest normalized to

0

20

40

60

80

100

120

Ctr

l

SA

0.5

SA

0.7

SA

1

SA

2

CH

E1

CH

E2

CH

E3

CH

LD

1

CH

LD

2

CH

LD

3

CH

L1

CH

L2

CH

L3

CH

L4

SL0.5

SL1

SL2

% o

f dead c

ells

Alkaloid

Alkaloid + z-VAD-fmk*

* **

*

*

*

***

Fig. 5. The effect of z-VAD-fmk inhibitor of caspases on alkaloid cytotoxicity. The

viability of A-375 cells was studied using flow cytometry analysis of PI stained cells.

Comparison of the percentage of the dead cells pretreated for 3 h with z-VAD-fmk

(20 mM) and then treated with BAs (0.5, 0.7, 1, 2, 3, and 4 mg/ml) for additional 45 h

and cells incubated directly with alkaloid. The effect of all alkaloids with the

exception of SL was significantly affected with z-VAD-fmk pretreatment. The most

pronounced effect (40% decrease of the cell death) was apparent after treatment

with CHL at concentration 4 mg/ml. Data are means � S.E.M. of three independent

experiments. *p < 0.05 pretreated vs. non-pretreated cells.

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–35 27

dehydrogenase). Reduction in metabolic activity is proportional tothe toxicity of the tested alkaloids. Treatment with all individualalkaloids (0.1–3 mg/ml) resulted in a dose-dependent decrease inthe metabolic activity of cells. Table 1 shows the IC50 values of testedalkaloids on individual cell lines. SK-MEL-2 cells proved to be theleast sensitive to alkaloids effects. The IC50 values of CHE and CHLDwere higher even than the highest tested concentration (3 mg/ml).Only slight differences in the IC50 values of the individual alkaloidswere observed in the A-375, A-375-p53sh and A-375-p53DD celllines. Statistical analysis revealed lower toxicity of CHE and CHL onA-375-p53DD than on A-375 cells and higher toxicity of CHLD on A-375-p53DD and A-375-p53sh than on A-375 cells.

The results indicate that cellular lineage plays more importantrole in determining the response of a particular tumour toindividual BAs than solely the activity of p53 protein.

3.1.2. xCELLigence – growth curves

The xCELLigence System, contrary to end point assays such asMTT, allowed us to observe the dynamics of A-375 melanoma cellsresponse to alkaloid treatment. The obtained growth curves (Fig. 2)show that the kinetics of responses to individual alkaloids weredifferent and were dependent on the concentrations of alkaloids.The results of SA, CHLD and SL treatment show significantdifferences in toxicity of these alkaloids in relation to the relativelysmall differences in the concentration of the alkaloid. SA atconcentration 0.1 mg/ml had almost no effect; while at aconcentration 0.3 mg/ml the curves stopped increasing and evenshowed a slight decrease in the CI (Fig. 2A). However, immediatedecreases in the CI can be mediated by detaching the cells from thesurface of the plate. Despite this, using the PI exclusion assay, wefound that detached cells die within 2 h (data not shown). CHLDshowed similar dynamics at concentrations 0.1 mg/ml vs. 0.5 mg/ml (Fig. 2B). In response to SL treatment, a significant reduction inthe CI was observed at concentrations of 0.1 mg/ml and cell growthstopped completely at a concentration of 0.5 mg/ml (Fig. 2C). Theeffect of CHE and CHL was more gradual; increasing concentrationsled to a smooth, continual decline in the CI (Fig. 2D and E).

3.2. Apoptosis induction

3.2.1. Effect of alkaloids on mitochondrial membrane potential

We used a dual-emission probe JC-1 to study changes in themitochondrial membrane potential of A-375 cells in response to24 h treatment with alkaloids at concentrations 1, 2 and 3 mg/ml.Anticancer drugs and apoptosis inductors, camptothecin (10 mM)or doxorubicin (1 mg/ml), were used as positive controls. About15% of the untreated control cells showed decreased MMP. Themajority of camptothecin- or doxorubicin-treated cells showed adecrease in MMP, 86% and 76%, respectively. The percentage ofcells with decreased MMP ranged from about 70% (SA 1 mg/ml) to96 and 98% (SA 1.5 mg/ml and 3 mg/ml, respectively) after SAtreatment and 21% (CHE 1 mg/ml), 19% (CHE 1.5 mg/ml) and 48%(CHE 3 mg/ml) in response to CHE treatment. CHLD, at theconcentration of 1 and 1.5 mg/ml, resulted in about 50% of cellshaving decreased MMP and at concentration of 3 mg/ml, 62% ofcells had a decreased MMP. After CHL treatment 32% (CHL 1 mg/ml), 29% (CHL 1.5 mg/ml) and 64% (CHL 3 mg/ml) cells showeddecreased MMP. About 30% (1 and 1.5 mg/ml) and 63% (3 mg/ml)cells with decreased MMP were detected in response to SLtreatment (Fig. 3). Similar dose-dependent reduction of MMP wasalso detected in alkaloid-treated A-375-p53DD and SK-MEL-2 cells(data not shown).

3.2.2. Immunodetection of cleaved caspase-3 and cleaved PARP

The results of Western blot analysis showed increased levels ofboth cleaved caspase-3 (Fig. 4A, B, and F) and cleaved PARP (Fig. 4C,

D, E, and G) after 3 and 6 h, respectively. The cells were treated withBAs at concentrations 0.1, 1, 2 and 3 mg/ml. Treatment of A-375, A-375-p53DD and A-375-p53sh melanoma cell lines with SA at theconcentrations of 1, 2 and 3 mg/ml and CHE and CHL at theconcentration of 3 mg/ml resulted in increase of both caspase-3and PARP cleavage (Fig. 4). In contrast, CHLD and SL, at testedconcentrations, did not induce caspase-3 or PARP cleavage.Similarly, in alkaloid-treated SK-MEL-2 melanoma cells we sawneither caspase-3 nor PARP cleavage induction with any of thetested alkaloids even at concentrations as high as 4 or 5 mg/ml(data not shown).

In order to study the dynamics of BAs-induced apoptosis, welooked for the markers of apoptosis in A-375 cell lysates at shortertime points post treatment (0.5, 1, 2 and 3 h; Fig. 4H–J). Slightcaspase-3 cleavage was observed after 0.5 h of treatment with CHLand CHE at the concentration of 3 mg/ml. After 1 and 2 h oftreatment caspase-3 cleavage was detected in response to SA, CHEand CHL (3 mg/ml; Fig. 4H–J). Interestingly, no or only slightcaspase-3 cleavage was detected in response to doxorubicintreatment at any of the tested time periods, which indicates, thatBAs could induce apoptosis more rapidly than doxorubicin, DNAdamaging chemotherapeutic agent.

3.2.3. The effect of z-VAD-fmk inhibitor of caspases on alkaloid

cytotoxicity

Pretreatment with the caspase inhibitor z-VAD-fmk decreasedthe toxicity of all tested alkaloids with the exception of SL (Fig. 5).The alkaloids were tested at concentrations (0.5, 0.7, 1, 2, 3 and4 mg/ml). The percentage of dead cells after preincubation with z-VAD-fmk decreased in SA- and CHLD-treated cultures at all testedconcentrations by 15–22%. The most pronounced effect (40%decrease in cell death) was apparent after treatment with CHL at aconcentration of 4 mg/ml. CHE was non-toxic at concentrations of1 and 2 mg/ml; its toxic effect, at a concentration of 3 mg/ml, wasdiminished by 21.7% after z-VAD-fmk pretreatment. Interestingly,the cytotoxicity of SL, the most toxic of the tested alkaloids, was notaffected by z-VAD-fmk pretreatment, indicating that somealkaloids might be able to induce caspase-independent cell death.

3.2.4. Role of the Bcl-2 family proteins and inhibitor of apoptosis

(XIAP) in alkaloid-induced apoptosis

The alkaloids were tested at concentrations 0.1, 0.5, 1, 1.5, 2 and3 mg/ml. SA and CHE treatment (6 h) led to a reduction of Bcl-xLprotein levels in all tested cell lines (Fig. 6A–C), while CHLD hadeffect at concentration 1 mg/ml on SK-MEL-2 (Fig. 6C) and A-375-[()TD$FIG]

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–3528

p53sh (Fig. 6J) cells and at concentration 2 mg/ml and higher on A-375 and A-375-p53DD cells (data not shown). SK-MEL-2 cellsshowed, as in previous experiments, the least sensitivity, andalkaloids other than SA and CHE had effects only at higherconcentrations (2–3 mg/ml). SL and CHL induced decreases in Bcl-xL levels in all tested cell lines. Interestingly, SL showed an effecteven at a concentration of 0.1 mg/ml (Fig. 6G–K). Similarly as Bcl-xL, a reduction in Mcl-1 protein levels was observed in response to

[()TD$FIG]

Fig. 6. Western blot analysis of Bcl-xL, Mcl-1 and XIAP proteins. Protein levels of Bcl-xL

treatment with BAs (0.1, 0.5, 1, 1.5, 2, and 3 mg/ml) of A-375 (A, D, G, L, and M), A-375-p

lysates were separated on a 10% SDS gels. PCNA was used as a loading marker. Num

corresponding PCNA and control normalized to corresponding PCNA.

all tested alkaloids at concentrations of 2 and 3 mg/ml in all celllines (Fig. 6D–K) with the exception of CHE and CHLD in SK-MEL-2cell line (data not shown). SL was active also at concentration0.1 mg/ml in A-375-p53DD and SK-MEL-2 cells (Fig. 6H and I).Similar results for both protein levels were also obtained after 8and 12 h of alkaloid treatment (data not shown).

Three-hour treatment with SA, CHE, CHL and SL (1 mg/ml)caused decrease in the levels of XIAP, while CHLD (1 mg/ml) had no

(A–C and G–K), Mcl-1 (D–K) and XIAP (L and M) were detected after 6 h and 3 h (L)

53DD (B, E, and H), SK-MEL-2 (C, F, and I) and A-375-p53sh (J and K) cells. Total cell

eric values represent ratio of band densities of protein of interest normalized to

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–35 29

effect (Fig. 6L). SA at concentrations 0.1, 0.5 and 1 mg/ml and CHEat a concentration of 1 mg/ml caused decreased levels of XIAP alsoafter 6 h (Fig. 6L) and after 12 h (data not shown) of treatment ofthe A-375 cell line.

3.3. The effect of alkaloids on DNA

3.3.1. Chromatin condensation – fluorescence and electron

microscopy

DAPI staining of nuclei showed that 24 h treatment of A-375melanoma cells with SA (1.5 mg/ml) resulted in chromatincondensation, which was detectable in almost all cells (Fig. 7B).Interestingly, we did not observe similar chromatin condensationafter treatment with CHE, at this concentration (data not shown).We confirmed the results of fluorescence microscopy usingelectron microscopy. Chromatin condensation and mitochondriaswelling (Fig. 7D and E) were apparent following SA (1.5 mg/ml)treatment.

3.3.2. Induction of H2AX phosphorylation

The fact that BAs interact with DNA together with the abovementioned results led us to test whether BAs treatment causesnuclear DNA damage. Using Western blot analysis and fluores-cence microscopy, we looked for phosphorylated histone H2AX(gH2AX) as a marker of DNA damage. All tested BAs increasedgH2AX levels. After 6 h treatment with SA, at all used concentra-tions all tested cell lines showed a significant increase in gH2AX.CHE had similar effect as SA at concentrations 2 and 3 mg/ml in A-375 and at concentration 3 mg/ml in A-375-p53sh (Fig. 8D) and atconcentration 1 mg/ml in A-375-p53DD cell line (Fig. 8A–C). Wedetected slight positivity of gH2AX staining in response to CHLD[()TD$FIG]

Fig. 7. Chromatin condensation in A-375 cells after 24 h treatment with SA. Fluorescen

sections (C–E) – control (A and C), cells after treatment with SA at concentration 1.5 mg

(D and E) are apparent. n – nucleus; nu – nucleolus; m – mitochondria; er – endoplasm

1 mm (E).

treatment at concentration 1 mg/ml in A-375, A-375-p53DD and atconcentrations 2 and 3 mg/ml in A-375-p53sh cells. SL and CHLwere tested on A-375 and A-375-p53sh cells. SL at concentration3 mg/ml and CHL at concentration 2 mg/ml have slight effect, whileCHL at concentration 3 mg/ml had strong effect on H2AXphosphorylation (Fig. 8D).

In order to determine whether H2AX phosphorylation precedescaspase-3 activation (Section 3.2.2), we detected gH2AX aftershorter treatments (0.5, 1 and 2 h; Fig. 8E–H) of A-375 cells. We didnot observe a significant increase of gH2AX signal after 0.5 h oftreatment with any alkaloid (Fig. 8E and H). After 1 and 2 h oftreatment we detected a weak gH2AX signal in response to SA, SLand CHLD (3 mg/ml) and a strong signal in response to CHE and CHL(3 mg/ml) treatment (Fig. 8F and H); doxorubicin (1 mg/ml) used asa positive control caused slight gH2AX positivity after 0.5 h andstrong positivity after longer periods of treatment (Fig. 8E–I).

The phosphorylation of H2AX, in the alkaloid-treated A-375cells was also observed using immunofluorescence. Treatmentwith SA, at the concentrations of 1 mg/ml and higher, resulted inthe formation of gH2AX foci in nuclei (Fig. 8M–O). We alsoobserved the formation of gH2AX foci in nuclei after treatmentwith CHE at the concentration of 3 mg/ml, but not at lowerconcentrations (data not shown).

3.4. The role of p53 in cellular response to BAs

3.4.1. The effect of alkaloids on the cellular p53 and p21 levels

While doxorubicin (1 mg/ml; used as a positive control)treatment led to a significant increase in p53 levels, comparedto untreated controls, twelve-hour treatment with all tested BAs(1, 1.5, 2 and 3 mg/ml) failed to produce a significant increase in

ce microscopy of DAPI stained cells (A and B) and electron microscopy of ultrathin

/ml (B, D, and E). Chromatin condensation (B, D, and E) and mitochondria swelling

ic reticulum; ch – condensed chromatin. Bars – 20 mm (A and B); 5 mm (C and D);

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–3530

p53 levels in either A-375 cells (Fig. 9A and B) or SK-MEL-2 cells(Fig. 9E and F). Only SA (1, 1.5 and 2 mg/ml) and CHLD (1, 1.5 and2 mg/ml) treatments led to a slight increase in p53 level, howeverto a much lesser extent than doxorubicin (Fig. 9A). Quitesurprisingly, SA, CHE, SL and CHL treatments of A-375 (Fig. 9Aand B) and SA, SL and CHL treatments of SK-MEL-2 cells (Fig. 9E andF) produced decrease in the levels of p53. As expected, alkaloidshad none or only slight effect on p53 levels in A-375-p53DD(Fig. 9C and D) and in A-375-p53sh almost no p53 protein could bedetected (Fig. 9G and H).

[()TD$FIG]

Fig. 8. Detection of phosphorylated histone H2AX (gH2AX). Western blot detection of gp53sh (D), and after 0.5 h (E), 1 h (F) and 2 h (G) treatment of A-375 with SA, CHE and DOX

tested at concentrations of 1, 2 and 3 mg/ml, or doxorubicin (DOXO) at concentration 1 m

marker. Numeric values represent ratio of band densities of protein of interest norm

Fluorescence microscopy of control (J–L) A-375 cells and cells treated 24 h with SA (1 m(L and O). Bars = 50 mm.

Similarly, in contrast to doxorubicin, which was used as apositive control, alkaloid treatment failed to produce a significantincrease in the levels of p53 target and cyclin dependent kinaseinhibitor p21 in A-375 cells (Fig. 9I and J).

3.4.2. Absence of functional p53 does not have a negative impact on

cell death induction with BAs

Compared with A-375, we did not observe any significantincrease of resistance of either A-375-p53DD or A-375-p53sh (cellslacking functional p53) to alkaloid treatment. Conversely, A-375-

H2AX after 6 h treatment of A-375 (A), A-375-p53DD (B), SK-MEL-2 (C) and A-375-

O and after 0.5 h, 1 h and 2 h (H) and 3 h and 6 h (I) with SL, CHL, and CHLD. BAs were

g/ml. Total cell lysates were separated on a 15% SDS gels. PCNA was used as a loading

alized to corresponding PCNA and control normalized to corresponding PCNA.

g/ml; M–O). DAPI staining (J and M); immunostaining of gH2AX (K and N); merge

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–35 31

p53DD cells showed higher sensitivity to treatment with alkaloidsat lower concentrations than parental A-375 cell line; however, atconcentrations of 1 mg/ml and higher, sensitivity to BAs seems tobe almost the same for both parental and p53 deficient cells(Fig. 10A). Similarly, A-375-p53sh showed higher sensitivity uponSL (1 mg/ml) and CHL (3 mg/ml) treatment than A-375, while SA(1 mg/ml) was slightly more toxic for A-375-p53sh. Sensitivity ofboth cell lines to other alkaloids was comparable (Fig. 10B). Theabove mentioned results suggest that p53 function does notsignificantly influence the sensitivity of melanoma cells to BAstreatment.

4. Discussion

BAs have been evaluated as potent anti-proliferative and pro-apoptotic agents by many authors and on various types of cancercells, including melanoma cells [12,30,31]. However, the mecha-nism of their action has not been fully elucidated yet and only themajor alkaloids such as SA, CHE and to a lesser extent CHLD havebeen studied.

[()TD$FIG]

Fig. 9. Western blot analysis of p53 and p21. Protein levels of p53 (A–H) and p21 (I and J) a

375 (A, B, I, and J), A-375-p53DD (C and D), SK-MEL-2 (E and F) and A-375-p53sh (G and H

with any tested alkaloid. On the contrary, doxorubicin used as a positive control, stabilise

samples. In A-375-p53sh, almost no p53 protein was detected. Total cell lysates were s

represent ratio of band densities of protein of interest normalized to corresponding PC

In contrast to previous studies, which dealt with only onealkaloid (mainly SA), our study evaluated the effect of three majoralkaloids (SA, CHE, and CHLD) and two minor alkaloids (SL andCHL) on human malignant melanoma cell lines with bothfunctional and non-functional tumour suppressor p53. While SA,CHE, SL, and CHL have quaternary nitrogen atom in their molecule,CHLD is a tertiary hexahydro-benzophenanthridine alkaloid. Ourstudy is focused on the mechanisms associated with anti-proliferative activity of BAs, including their possible effects onDNA, anti-apoptotic proteins and the p53 tumour suppressorpathway.

The possible relationship between p53 status and cellularresponse to BAs treatment is not fully understood yet. Similarly,the extent of DNA damage and its role in the induction of cell deathin BA-treated cells is not entirely clear. Ahsan et al. [14] observedthat antiproliferative and proapoptotic effects of SA on pancreaticcarcinoma cells, involved changes in the balance between pro- andanti-apoptotic BCL2 family proteins and were accompanied by adecrease in the levels of p53, but increased phosphorylation of thisprotein. Despite the fact that these cells had a p53 mutation, whichcould cause a loss of its transcription activity, SA probably restored

fter 12 h BAs (1, 1.5, 2, and 3 mg/ml) or doxorubicin (DOXO; 1 mg/ml) treatment of A-

) cells. Significant stabilisation of neither p53 nor p21 was detected after treatment

d p53 in both A-375 and SK-MEL-2 cells. In A-375-p53DD, p53 level was stable at all

eparated on a 12.5% SDS gels. PCNA was used as a loading marker. Numeric values

NA and control normalized to corresponding PCNA.

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–3532

the function of p53 by inducing its phosphorylation. On the otherhand, Matkar et al. [32] concluded that apoptosis of coloncarcinoma cells induced by SA was p53- and DNA damage-independent and that observed DNA fragmentation was the resultof apoptosis, not its cause. Similarly, other authors failed to detectincreased level of p53 after SA treatment [13,33]. NK314, a novelsynthetic benzo[c]phenanthridine, that is currently in clinicaltrials, inhibited topoisomerase IIa and induced G2 cell cycle arrestindependently of p53 status [34]. However, NK314 (100 nmol/l)treatment was associated with H2AX phosphorylation, which wasdetected especially in the G2 population.

Using fluorescence and electron microscopy, we observedchromatin condensation in most of the cells treated with SA and insome cells treated with CHE. SA-induced chromatin aggregationwas also observed by Selvi et al. [35] by confocal and atomic forcemicroscopy. To assess the extent of DNA damage, associated withBAs treatment, we detected gH2AX, i.e. the phosphorylation ofhistone H2AX, which is a sensitive method for detecting DNAdouble strand breaks. In our experiments, we found that all testedBAs caused increase of gH2AX to a various extent. We were able todetect both gH2AX and caspase-3 cleavage at the same time upon

[()TD$FIG]

0

20

40

60

80

100

Ctr

l

SA

0.1

SA

0.5

SA

1

SA

3

CH

E0.1

CH

E0.5

CH

E1

CH

E3

CH

LD

0.1

% o

f dead c

ells

A-375

A-375-p53DD

A

**

*

*

*

** *

*

0

20

40

60

80

100

Ctr

l

SA

0.1

SA

0.5

SA

1

SA

3

CH

E0

.1

CH

E0

.5

CH

E1

CH

E3

CH

LD

0.1

% o

f dead c

ells

A-375

A-375-p53sh

B

*

Fig. 10. Flow cytometry analysis of PI stained A-375, A-375-p53DD and A-375-p53sh sho

with A-375 no basic difference in BAs toxicity to A-375-p53DD (A) or A-375-p53sh (B) w

375 vs. A-375-p53DD (A) or A-375 vs. A-375-p53sh (B).

SA (1 h) and CHE (0.5 h) treatment. Upon CHL treatment caspase-3cleavage was detected even earlier than gH2AX (0.5 h vs. 1 h). Inview of these results it is possible that in the case of CHL apoptosisinduction preceded DNA damage. In the case of other alkaloidstreatment we were not able to accurately determine whetherapoptosis is a consequence or the cause of DNA strand breaks. Inthe case of doxorubicin (used as a positive control), the gH2AX wasdetected already after 0.5 h and probably preceded caspase-3cleavage, which we had not detected even after 3 h of treatment,suggesting that initiation of apoptosis follows induction of DNAdamage associated with doxorubicin treatment. These resultsshow the different dynamics of cellular effects of individual drugs,even those as closely related substances, as various BAs.

Quaternary BAs exhibit pH-dependent structural equilibriumbetween the iminium (charged) and alkanolamine (neutral) formsdependent on pKa values of individual alkaloids [36]. QuaternaryBAs at their iminium forms are highly reactive molecules sensitiveto attack by nucleophiles and interacting with various proteins andenzymes, mostly through the interaction with SH (mercapto)groups of cysteines [37]. Interactions of SA and CHE with widespectrum of enzymes, which are involved in signal transduction

CH

LD

0.5

CH

LD

1

CH

LD

3

SL0.1

SL0.5

SL1

SL3

CH

L0.1

CH

L0.5

CH

L1

CH

L3

*

*

*

*

*

* *

*

*C

HL

D0

.5

CH

LD

1

CH

LD

3

SL0.1

SL0.5

SL1

SL

3

CH

L0

.1

CH

L0

.5

CH

L1

CH

L3

*

*

*

wing cell viability after 48 h treatment with BAs (0.1, 0.5, 1, and 3 mg/ml). Compared

as observed. Data are means � S.E.M. of three independent experiments. *p < 0.05 A-

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–35 33

pathways leading to cell proliferation or cell death, were describedby many authors (for review see [7,8,36,37]) and this couldcontribute to multiple cellular effects.

As already mentioned, another important property of quater-nary BAs is their ability to interact with DNA and chromatin. Theintercalative complexation of sanguinarine to DNA was describedby several scientific groups (for review see [36]). SA and CHE havehigher binding affinity for G-C rich sequences than A-T richsequences. The interaction of SA with DNA is based on noncovalentcomplexation based on week intermolecular forces [38]. Stiborovaet al. [39] demonstrated covalent DNA modifications by SA in an invitro experiment after metabolic activation of the alkaloid. In spiteof the fact that all tested quaternary BAs can interact with DNA invitro [17], in living cells only SA and to a lesser extent CHE werelocalised in nuclei, which was observed thanks to their fluores-cence [4,16,40].

Nevertheless, the effect of SA on chromatin is more complex.Not only direct interactions with DNA, but also effect on histonemodifications [35] and topoisomerase II relocalisation [41] wereobserved. It was found that SA inhibits histone methylation andacetylation that can lead to the modulation of global geneexpression [35]. The possibility that BAs can affect gene expressioncould explain some effects of BAs such as observed decrease inantiapoptotic protein levels.

It has been suggested that the cytotoxic activity of SA and CHEcorrelates with their DNA intercalating properties and with theinduction of DNA strand breaks, because the non-intercalatingalkaloid CHLD was less toxic [18]. However, the same groupcompared the DNA intercalating effects of SA with non-intercalat-ing CHLD and concluded that the apoptogenic activity of thesealkaloids was independent of their DNA-damaging properties [20].In addition to its DNA binding, DNA fragmentation has been shownafter SA treatment in DNA fragmentation assays [5,10,42,43].Another example for alkaloid-specific cellular effects is providedby non-intercalating tertiary hexahydro-benzophenanthridinealkaloid CHLD, which is able to inhibit tubulin polymerisationand cause cell cycle arrest at G2/M phase that is probably the mainmechanism of its action [44].

Even though we have detected signs of DNA damage, we did notobserve significant increases in p53 levels after treatment with thetested BAs in our experiments. We also evaluated p21 (CDKinhibitor and p53 transcriptional target) levels in order to assessthe activity of p53, but no increase in p21 was detected. We havecreated two cell lines with defective p53 derived from the line A-375. A-375-p53DD in which p53 transcriptional activity is blockedat the protein level by protein–protein interactions and A-375-p53sh in which p53 function is downregulated at the mRNA levelby a specific short hairpin RNAs (shRNA). Results presented inFig. 10 suggest that the cancer cells lacking functional p53 (A-375-p53DD and A-375-p53sh) exhibit similar and possibly evenslightly higher sensitivity to lower BAs concentrations than wildtype p53-expressing cells (A-375). In the case of A-375-p53DDcells, in addition to the activity of p53 as transcription factor, theprotein can be translocated to the mitochondria and directlyinduce transcription independent cell death. It might be possiblethat high levels of transcriptionally inactive p53 stabilised bybinding to p53DD can sensitize the A-375-p53DD cells to sometypes of stress via this pathway. Our results indicate that p53 doesnot play an essential role in cellular responses to BAs treatmentand that BAs can inhibit the proliferation of cancer cells lackingfunctional p53 with efficiency comparable to that achieved in wildtype p53 expressing cells. We previously reported that theindependence of apoptosis induction on functional p53 upon SA,CHE, sanguirubine, chelirubine, macarpine [6], SL and CHL [5]treatment in HL60 leukaemia cells, which lack both p53 alleles. Theresults of our studies were also supported by Matkar et al. [32],

who observed a comparable effect of SA on cells bearing wild typeand mutated p53.

The ability of SA and CHE [10,14,30,45–47] and some minor BAs[5,6] to induce apoptosis has been demonstrated using varioustypes of cells. Apoptosis induced by BAs may be mediated by bothcaspase-9-dependent mitochondrial pathways [12,19], or thedeath receptor pathway, in which caspase-8 is activated [10,48].To assess the dependency of the cytotoxic effects of the tested BAson the activity of caspases and thereby determine whether theobserved cell death was mediated by apoptosis, we studied theviability of BAs treated cells that had been preincubated with thecaspase inhibitor z-VAD-fmk. We found that the toxicity of SA,CHE, CHLD and CHL was significantly reduced by z-VAD-fmkpreincubation, while SL toxicity was not. Even though we did notdetect significant cleavage of caspase-3, or its substrate PARP, aftertreatment with CHLD, the decrease in toxicity following z-VAD-fmk pretreatment suggests that the cytotoxicity of CHLD may be atleast partly mediated by caspases. Taken together, our resultsindicate that apoptosis and caspase activation contribute to theeffects of CHL, CHE, SA and possibly CHLD, while the effects of SLmay be caspase-independent. Again, these results also suggest thatthe mechanisms of action of individual BAs can be quite different.While SA-, CHE- and CHL-treated cells exhibit classic character-istics of the induction of apoptosis, i.e. decreased MMP, caspase-3and PARP cleavage and decreased toxicity after z-VAD-fmkpretreatment; CHLD-treated cells showed only decreased MMPand z-VAD-fmk pretreatment decreased the toxicity of this drug. SLtreatment only led to decreased MMP.

High levels of expression of pro-survival members of the Bcl-2protein family or low levels of pro-apoptotic members in humancancer cells have frequently been found to be reliable prognosticindicators of a poor response to chemotherapy. That is why thesearch for substances that can reduce levels of anti-apoptoticproteins is critical. High throughput screening of 107,423 extractsderived from natural products identified CHE as an inhibitor of Bcl-xL-Bak Bcl-2 homology 3 (BH3) domain binding [11]. Zhang et al.[49] found that CHE and SA bind to the BH groove and BH1 region,respectively, of Bcl-xL. This is different from BH3 binding that istargeted by other known inhibitors of Bcl-xL.

Increased levels of pro-apoptotic Bcl-2 family members (Bax,Bak, Bad, Bid; [13,14,50]), decreased the levels of anti-apoptoticBcl-2 family members (Bcl-2, Bcl-xL, Mcl-1; [10–15]) andincreased levels of Apaf-1 [10] have been described after BAs(SA and/or CHE) treatment. Our results obtained with a widerspectrum of BAs are in agreement with the above-mentionedresults showing decreased levels of the anti-apoptotic protein Bcl-xL. In addition to that, we demonstrated a decrease in the levels ofanti-apoptotic Mcl-1 and XIAP in BAs-treated melanoma cells.

The results of this study show that despite the induction ofH2AX phosphorylation by all alkaloids, no or slight p53 proteinstabilisation is apparent. It is possible that although BAs interactwith DNA, they do not cause the immediate DNA damage and thatthe observed induction of gH2AX is mostly the result of DNAcleavage in apoptosis. We propose that mechanism other than DNAdamage rapidly triggers apoptosis upon BAs treatment. It could bea direct effect on mitochondria leading to cytochrome c release andtriggering of apoptosis. This is possibly supported by the above-mentioned results showing changes in pro- and anti-apoptoticproteins levels. Another possibility should be a direct interaction ofBAs with mitochondria affecting their membrane permeability.Serafim et al. [12] showed that SA affects the mitochondrialrespiratory chain and has dual effects on mitochondrial calciumloading capacity. They concluded that SA acts as a DNA damagingagent, showing also collateral damage to mitochondrial bioener-getics. Kaminskyy et al. [19] described impaired mitochondria inresponse to SA and CHE treatment, but not CHLD treatment. This is

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–3534

in agreement with our results showing proapoptotic effect of SAand CHE, while CHLD induced apoptosis to a much lesser extent.Our previous results analysing the localisation of BAs in living cellsshowed CHL, SL and partially CHE and SA in cytoplasmic structuressome of which could be mitochondria especially at the case of CHLand SL [4,16].

Different effects of CHLD in comparison to quaternary BAs arenot surprising given the fundamental difference in its molecularstructure based on the absence of quaternary nitrogen. On theother hand, the substantial difference in the effect of chemicallyrelated molecules such as SL and CHL was quite surprising. CHLappears to be the strongest inductor of apoptosis, while SL,though the most toxic, does not induce apoptosis in melanomacells at the tested concentrations and probably kills cells byactivating another type of the cell death that does not requireactivity of caspases. In our opinion the properties of these minorBAs are interesting and deeper analysis of their mechanisms ofaction will bring new interesting findings. From the comparisonof their molecular structure in relation to their toxicity, it isclear that a high content of methoxy groups leads to highertoxicity (SL, CHL). We have learned earlier that alkaloidscontaining methylenedioxy groups, especially in positions C7and C8 (R3 and R4 – Fig. 1) have more planar molecule andhigher capacity to intercalate into DNA [16]. On the other hand,the results of this study indicate that the toxicity of BAs andespecially their ability to induce apoptosis cannot be simplylinked to their ability to interact with DNA.

In addition to the chemical structure of alkaloid, its concentra-tion can also play a decisive role in the mechanism of the cell deathinduction. Recently, it was made clear that not only apoptosis, butalso other types of non-apoptotic programmed cell death can playrole in the cellular responses to BAs treatment. Bimodal cell deathdistinguishing two distinct concentration-dependent modes of celldeath, apoptosis (lower concentrations of BAs; caspase 3 and PARPpositive) and oncosis (higher concentrations of BAs; caspase 3 andPARP negative), was described by several authors after treatmentof various cancer cells with SA [18,30,47,50,51]. The ability of BAsto induce various types of the cell death can explain the observedlower rate of apoptotic features in response to the treatment withmore toxic alkaloids such as SA and SL that can also induce oncosisor necrosis.

In this study, we have compared the effect of five BAs onmalignant melanoma cell lines and demonstrated that BAs havestrong anti-proliferative effects. Despite the profound similari-ties in the molecular structures of various BAs, it is clear that themechanism associated with cell death induction is different foreach alkaloid. SL, which was identified as the most toxic of thetested BAs, was the weakest inductor of apoptosis, indicating thatit probably causes necrosis or other type of cell death which isindependent of caspases. The most potent inductors of apoptosisare CHL and CHE. SA, though it has a greater antiproliferativeactivity than CHE and CHL, induces apoptosis to a lesser extent.Its effect probably causes a several different types of the celldeath. Although quaternary BAs represent substances thatinteract with DNA, induction of cell death is independent offunctional p53 and the mechanism of cell death induction by alltested BAs, apparently does not seem to require p53 activity.Apoptosis is probably induced through weakening of pro-survival mechanisms, such as lowering the levels of anti-apoptotic protein (Bcl-xL, Mcl-1 and XIAP) and by direct effectof BAs on mitochondria.

Taken together, our results suggest that BAs represent a groupof drugs which are able to affect the growth and viability ofmalignant melanoma cells via diverse molecular mechanisms andhave therefore a high potential for the development of moreefficient anti-melanoma therapies.

Acknowledgements

The authors thank Roche s.r.o. (www.roche-diagnostics.cz) forthe kind loan of xCELLigence analyser, Mrs. Z. Prokopova forlaboratory assistance and Thomas Secrest for language correction.This work was supported by grants from the Czech ScienceFoundation (525/08/0819) and from the Ministry of Education ofthe Czech Republic (VZ MSM0021622415 and LC06077).

References

[1] Garbe C, Leiter U. Melanoma epidemiology and trends. Clin Dermatol2009;27:3–9.

[2] Gray-Schopfer V, Wellbrock C, Marais R. Melanoma biology and new targetedtherapy. Nature 2007;445:851–7.

[3] Jilaveanu LB, Aziz SA, Kluger HM. Chemotherapy and biologic therapies formelanoma: do they work? Clin Dermatol 2009;27:614–25.

[4] Slaninova I, Taborska E, Bochorakova H, Slanina J. Interaction of benzo[c]phe-nanthridine and protoberberine alkaloids with animal and yeast cells. Cell BiolToxicol 2001;17:51–63.

[5] Slunska Z, Gelnarova E, Hammerova J, Taborska E, Slaninova I. Effect ofquaternary benzo[c]phenanthridine alkaloids sanguilutine and chelilutineon normal and cancer cells. Toxicol In Vitro 2010;24:697–706.

[6] Slaninova I, Slunska Z, Sinkora J, Vlkova M, Taborska E. Screening of minorbenzo[c]phenanthridine alkaloids for antiproliferative and apoptotic activi-ties. Pharm Biol 2007;45:131–9.

[7] Simanek V, Vespalec R, Sedo A, Ulrichova J, Vicar J. Quaternary Benzo(c)phe-nanthridine alkaloids – biological activities. In: Schneider MP, editor. Chemi-cal probes in biology. Netherlands: Kluwer Academic Publishers; 2003. p. 245–54.

[8] Zdarilova A, Malikova J, Dvorak Z, Ulrichova J, Simanek V. Quaternary isoquino-line alkaloids sanguinarine and chelerythrine. In vitro and in vivo effects. ChemListy 2006;100:30–41.

[9] Ernst E, Schmidt K. Ukrain – a new cancer cure? A systematic review ofrandomised clinical trials. BMC Cancer 2005;5:69.

[10] Adhami VM, Aziz MH, Mukhtar H, Ahmad N. Activation of prodeath Bcl-2family proteins and mitochondrial apoptosis pathway by sanguinarine inimmortalized human HaCaT keratinocytes. Clin Cancer Res 2003;9:3176–82.

[11] Chan SL, Lee MC, Tan KO, Yang LK, Lee ASY, Flotow H, et al. Identificationof chelerythrine as an inhibitor of BclXL function. J Biol Chem 2003;278:20453–6.

[12] Serafim TL, Matos JAC, Sardao VA, Pereira GC, Branco AF, Pereira SL, et al.Sanguinarine cytotoxicity on mouse melanoma K1735-M2 cells – nuclear vs.mitochondrial effects. Biochem Pharmacol 2008;76:1459–75.

[13] Reagan-Shaw S, Breur J, Ahmad N. Enhancement of UVB radiation-mediatedapoptosis by sanguinarine in HaCaT human immortalized keratinocytes. MolCancer Ther 2006;5:418–29.

[14] Ahsan H, Reagan-Shaw S, Breur J, Ahmad N. Sanguinarine induces apoptosis ofhuman pancreatic carcinoma AsPC-1 and BxPC-3 cells via modulations in Bcl-2 family proteins. Cancer Lett 2007;249:198–208.

[15] Han MH, Yoo YH, Choi YH. Sanguinarine-induced apoptosis in human leuke-mia U937 cells via Bcl-2 downregulation and caspase-3 activation. Chemo-therapy 2008;54:157–65.

[16] Slaninova I, Slanina J, Taborska E. Quaternary benzo[c]phenanthridine alka-loids – novel cell permeant and red fluorescing DNA probes. Cytometry Part A2007;71A:700–8.

[17] Urbanova J, Lubal P, Slaninova I, Taborska E, Taborsky P. Fluorescence proper-ties of selected benzo[c]phenantridine alkaloids and studies of their interac-tion with CT DNA. Anal Bioanal Chem 2009;394:997–1002.

[18] Kaminskyy VO, Lootsik MD, Stoika RS. Correlation of the cytotoxic activity offour different alkaloids, from Chelidonium majus [greater celandine], with theirDNA intercalating properties and ability to induce breaks in the DNA of NK/Lymurine lymphoma cells. Cent Eur J Biol 2006;1:2–15.

[19] Kaminskyy V, Kulachkovskyy O, Stoika R. A decisive role of mitochondria indefining rate and intensity of apoptosis induction by different alkaloids.Toxicol Lett 2008;177:168–81.

[20] Philchenkov A, Kaminskyy V, Zavelevich M, Stoika R. Apoptogenic fromactivity of two benzophenanthridine alkaloids Chelidonium majus L. doesnot correlate with their DNA damaging effects. Toxicol In Vitro 2008;22:287–95.

[21] Vousden KH, Lane DP. p53 in health and disease. Nat Rev Mol Cell Biol2007;8:275–83.

[22] Toledo F, Wahl GM. Regulating the p53 pathway: in vitro hypotheses, in vivoveritas. Nat Rev Cancer 2006;6:909–23.

[23] Slavik J, Slavikova L. Alkaloide der Mohngewachse [Papaveraceae]. 17. UberNeue Alkaloide aus Sanguinaria-Canadensis L.. Collect Czech Chem Commun1960;25:1667–75.

[24] Taborska E, Veznik F, Slavikova L, Slavik J. Alkaloids of Papaveraceae. 67.Quaternary alkaloids of 3 species of Dicranostigma Hook-F et Thoms. CollectCzech Chem Commun 1978;43:1108–12.

[25] Slavik J, Slavikova L, Brabenec J. Alkaloide der Mohngewachse (Papaveraceae).30. Uber Weitere Alkaloide aus der Wurzel Von Chelidonium Majus l.. CollectCzech Chem Commun 1965;30:3697–704.

J. Hammerova et al. / Journal of Dermatological Science 62 (2011) 22–35 35

[26] Shaulian E, Haviv I, Shaul Y, Oren M. Transcriptional repression by the C-terminal domain of P53. Oncogene 1995;10:671–80.

[27] Kotala V, Uldrijan S, Horky M, Trbusek M, Strnad M, Vojtesek B. Potentinduction of wild-type p53-dependent transcription in tumour cellsby a synthetic inhibitor of cyclin-dependent kinases. Cell Mol Life Sci2001;58:1333–9.

[28] Vojtesek B, Bartek J, Midgley CA, Lane DP. An immunochemical analysis of thehuman nuclear phosphoprotein-P53 – new monoclonal-antibodies and epi-tope mapping using recombinant-P53. J Immunol Methods 1992;151:237–44.

[29] Fredersdorf S, Milne AW, Hall PA, Lu X. Characterization of a panel of novel anti-p21Waf1/Cip1 monoclonal antibodies and immunochemical analysis ofp21Waf1/Cip1 expression in normal human tissues. Am J Pathol 1996;3:825–35.

[30] Kemeny-Beke A, Aradi J, Damjanovich J, Beck Z, Facsko A, Berta A, et al.Apoptotic response of uveal melanoma cells upon treatment with chelidonine,sanguinarine and chelerythrine. Cancer Lett 2006;237:67–75.

[31] De Stefano I, Raspaglio G, Zannoni GF, Travaglia D, Prisco MG, Mosca M, et al.Antiproliferative and antiangiogenic effects of the benzophenanthridine alka-loid sanguinarine in melanoma. Biochem Pharmacol 2009;78:1374–81.

[32] Matkar SS, Wrischnik LA, Hellmann-Blumberg U. Sanguinarine causes DNAdamage and p53-independent cell death in human colon cancer cell lines.Chem Biol Interact 2008;172:63–71.

[33] Lee BY, Lee SJ, Park SS, Kim SK, Kim SR, Jung JH, et al. Sanguinarine-inducedG[1]-phase arrest of the cell cycle results from increased p27KIP1 expressionmediated via activation of the Ras/ERK signaling pathway in vascular smoothmuscle cells. Arch Biochem Biophys 2008;471:224–31.

[34] Guo L, Liu XJ, Nishikawa K, Plunkett W. Inhibition of topoisomerase II alphaand G[2] cell cycle arrest by NK314, a novel benzo[c]phenanthridine currentlyin clinical trials. Mol Cancer Ther 2007;6:1501–8.

[35] Selvi BR, Pradhan SK, Shandilya J, Das C, Sailaja BS, Shankar GN, Gadad SS,Reddy A, Dasgupta D, Kundu TK. Sanguinarine interacts with chromatin,modulates epigenetic modifications, and transcription in the context of chro-matin. Chem Biol 2009;16:203–16.

[36] Maiti M, Kumar GS. Molecular aspects on the interaction of protoberberine,benzophenanthridine, and aristolochia group of alkaloids with nucleic acidstructures and biological perspectives. Med Res Rev 2007;27:649–95.

[37] Dvorak Z, Kuban V, Klejdus B, Hlavac J, Vicar J, Ulrichova J, et al. Quaternarybenzo[c]phenanthridines sanguinarine and chelerythrine: a review of investi-gations from chemical and biological studies. Heterocycles 2006;68:2403–22.

[38] Vlckova M, Kuban V, Vicar J, Simanek V. Capillary zone electrophoretic studiesof interactions of some quaternary isoquinoline alkaloids with DNA consti-tuents and DNA. Electrophoresis 2005;26:1673–9.

[39] Stiborova M, Simanek V, Frei E, Hobza P, Ulrichova J. DNA adduct formationfrom quaternary benzo[c]phenanthridine alkaloids sanguinarine and cheler-ythrine as revealed by the P-32-postlabeling technique. Chem Biol Interact2002;140:231–42.

[40] Slaninova I, Slanina J, Taborska E. Fluorescence properties of quaternarybenzo[c]phenanthridine alkaloids and their use as supravital DNA probes.Chem Listy 2008;102:427–33.

[41] Holy J, Lamont G, Perlina E. Disruption of nucleocytoplasmic traffickingof cyclin D1 and topoisomerase II by sanguinarine. BMC Cell Biol 2006;7:13.

[42] Chang MC, Chan CP, Wang YJ, Lee PH, Chen LI, Tsai YL, et al. Induction ofnecrosis and apoptosis to KB cancer cells by sanguinarine associated withreactive oxygen species production and mitochondrial membrane depolari-zation. Toxicol Appl Pharmacol 2007;218:143–51.

[43] Kim S, Lee TJ, Leem J, Choi KS, Park JW, Kwon TK. Sanguinarine-inducedapoptosis: generation of ROS, down-regulation of Bcl-2, c-FLIP, and synergywith TRAIL. J Cell Biochem 2008;104:895–907.

[44] Panzer A, Joubert AM, Bianchi PC, Hamel E, Seegers JC. The effects of cheli-donine on tubulin polymerisation, cell cycle progression and selected signaltransmission pathways. Eur J Cell Biol 2001;80:111–8.

[45] Malikova J, Zdarilova A, Hlobilkova A, Ulrichova J. The effect of chelerythrineon cell growth, apoptosis, and cell cycle in human normal and cancer cells incomparison with sanguinarine. Cell Biol Toxicol 2006;22:439–53.

[46] Debiton E, Madelmont JC, Legault J, Barthomeuf C. Sanguinarine-inducedapoptosis is associated with an early and severe cellular glutathione depletion.Cancer Chemother Pharmacol 2003;51:474–82.

[47] Ding ZH, Tang SC, Weerasinghe P, Yang XL, Pater A, Liepins A. The alkaloidsanguinarine is effective against multidrug resistance in human cervical cellsvia bimodal cell death. Biochem Pharmacol 2002;63:1415–21.

[48] Yin HQ, Kim YH, Moon CK, Lee BH. Reactive oxygen species-mediated induc-tion of apoptosis by a plant alkaloid 6-methoxydihydrosanguinarine in HepG2cells. Biochem Pharmacol 2005;70:242–8.

[49] Zhang YH, Bhunia A, Wan KF, Lee MC, Chan SL, Yu VCK, et al. Chelerythrine andsanguinarine dock at distinct sites on Bcl[XL] that are not the classic BH3binding cleft. J Mol Biol 2006;364:536–49.

[50] Weerasinghe P, Hallock S, Tang SC, Liepins A. Role of Bcl-2 family proteins andcaspase-3 in sanguinarine-induced bimodal cell death. Cell Biol Toxicol2001;17:371–81.

[51] Vrba J, Dolezel P, Vicar J, Modriansky M, Ulrichova J. Chelerythrine anddihydrochelerythrine induce G1 phase arrest and bimodal cell death in humanleukemia HL-60 cells. Toxicol In Vitro 2008;22:1008–17.