Evaluation of cytotoxic properties of organometallic ferrocifens on melanocytes, primary and metastatic melanoma cell lines

Journal of Inorganic Biochemistry 102 (2008) 1980–1985

Contents lists available at ScienceDirect

Journal of Inorganic Biochemistry journal homepage: www.elsevier.com/locate/jinorgbio

Evaluation of cytotoxic properties of organometallic ferrocifens on melanocytes, primary and metastatic melanoma cell lines Q. Michard a, G. Jaouen b, A. Vessieres b, B.A. Bernard a,* a b

L’Oréal Recherche, 90 rue du Général Roguet, 92583 Clichy Cedex, France ENSCP, Laboratoire de Chimie et Biochimie des Complexes Moléculaires, UMR CNRS 7576, 11, rue Pierre et Marie Curie, 75231 Paris Cedex05, France

a r t i c l e

i n f o

Article history: Received 29 April 2008 Received in revised form 22 July 2008 Accepted 24 July 2008 Available online 3 August 2008 Keywords: Melanoma cancer Organometallics Ferrocene Iron Oxidative stress

a b s t r a c t Malignant melanoma is one of the most severe forms of skin cancer, and chemotherapeutic agents currently in use are poorly effective in curing the disease. Here we describe the properties of two organometallic ferrocenyl derivatives, ferrocifen (Fc-OH-Tam) and ferrociphenol (Fc-diOH) that show a specific antiproliferative effect on melanoma cells. After a short incubation period, Fc-OH-Tam is highly cytotoxic on melanoma cells but less toxic on melanocytes. Fc-diOH is slightly toxic at a high concentration but no discrepancy is observed between malignant and normal cells. After a long incubation time the latter is highly toxic for malignant cells but not for normal cells while the former was very highly toxic for primary malignant cells and significantly less toxic for normal cells. We also found that oxidative stress is not implicated in the mechanism of cytotoxicity, since both derivatives neither induce reactive oxygen species (ROS) level in melanocytes nor in melanoma cells. Finally, investigation on hair follicle growth revealed that the two organometallic derivatives induced an irreversible ejection of the hair shaft, thus predicting a potential hair loss side effect if used as a chemotherapeutic treatment. Ó 2008 Elsevier Inc. All rights reserved.

1. Introduction Melanoma represents only 4% of skin cancers, but is one of the most severe forms, resulting in 65% of all patient deaths. Early detection of melanoma carries an excellent prognosis, since it can be cured by surgical resection. However, metastatic malignant melanoma does not respond to current therapies, and therefore has a very poor prognosis, with a median survival rate of 6 months and 5-year survival rate of less than 5% [1–5]. To date, the only FDA (USA)-approved chemotherapy for melanoma is the alkylating agent Dacarbazine, which gives clinical response in 5–10% of cases and cures approximately 1% of patients [6]. In addition, many different immunotherapies have been tested, but so far none of these approaches has received FDA regulatory approval. Interferon-a (IFN-a) is the most commonly used adjuvant immunotherapy for advanced melanoma, although its efficacy is still a matter of debate. High-dose interleukin-2 (IL-2) has also been approved, but still presents low efficiency with side toxic effects [7–9]. Consequently, the specificity of melanoma biology has rapidly become a crucial point for understanding the progression of the disease and its resistance to classical chemotherapies. These investigations have led to the identification of various essential cell * Corresponding author. Fax: +33 1 47 56 80 46. E-mail address: [email protected] (B.A. Bernard). 0162-0134/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2008.07.014

signaling pathways that are involved in melanoma initiation and progression, and have given rise to the discovery of new strategies for treatment as well as new drug designs [7,10,11]. Among them, the most recent strategies are focused on the Ras/Raf/MEK as well as the Ras/PI(3)K/Akt signaling pathways, both involved in melanoma progression, survival and motility. Moreover, there is now a growing consensus that there may be some functional redundancy between these multiple pathways, and that two or more signaling pathways should be targeted simultaneously to overcome drug resistance [12]. Ferrocifen (Fc-OH-Tam), a ferrocenyl derivative of hydroxytamoxifen, the active metabolite of tamoxifen, which is widely prescribed for the treatment of hormone-dependent breast cancer, and ferrociphenol (Fc-diOH) (Chart 1), have been recently reported to possess a strong antiproliferative effect (IC50 values 0.5 lM) in vitro on both hormone-dependent (MCF-7, estrogen receptor positive; ER+) and independent (MDA-MB-231, estrogen receptor negative; ER) breast cancer cells [13–17]. The antiproliferative effect observed on ER+ cells with Fc-OH-Tam seems to be both antiestrogenic and cytotoxic, while the effect observed on ER cells with Fc-OH-Tam and Fc-diOH can be linked only to the cytotoxicity of the compounds. The precise mechanism underlying this effect is not yet fully understood and is still discussed [18,19]. One proposed mechanism is based on two successive intracellular oxidations: the first one arises through oxidation of Fe2+ to Fe3+ ions and is followed by removal of phenolic proton; the second one

Q. Michard et al. / Journal of Inorganic Biochemistry 102 (2008) 1980–1985

1981

2.3. Cell viability assay Cells (5  104) were seeded in 24-well plates and left to grow overnight. The following day, cells were treated with increasing concentrations of ferrocifenyl derivative (1–20 lM) in triplicate. After 24 h, cell viability was measured using the Cell Proliferation kit II (XTT) (Roche Diagnostics, Basel, Switzerland). Results were representative of at least three independent experiments, and were expressed as percentage of the value observed with no drug treatment. T-Tukey tests were performed to compare every condition with each other.  means p-value <0.05. Chart 1. Chemical structure of the two tested ferrocenyl derivatives, ferrocifen (FcOH-Tam) and ferrociphenol (Fc-diOH).

involves the formation of a quinone methide, a reactive species able to react, via a Michael addition with macromolecules such as DNA or proteins, resulting in cell death. Such a mechanism, supported by electrochemical studies [20], requires a ferrocenyl/double bond/phenol ring moiety for the expression of a strong antiproliferative effect. The production of reactive oxygen species (ROS) in the Fenton reaction has also been proposed as a mechanism of cytotoxicity for these type of molecules [21] and an increased production of ROS has been associated with induction of apoptosis by tamoxifen [22]. Here we report the evaluation of the potential anticancer activity of the two ferrocenyl derivatives on a panel of four cell lines, a normal cell line from neonatal human epidermal melanocytes (NHEM) and three malignant melanoma cell lines including two from primary tumors (radial growth phase WM35 and vertical growth phase WM793) and one metastatic cell line (WM9) [7,23–25]. In order to gain a better insight into the mechanism underlying the toxicity of tested derivatives, we next measured the production of reactive oxygen species (ROS) they induced in these cell lines. Finally, since the two tested compounds could represent new anticancer drug candidates, we also evaluated their potential for inducing hair loss by testing in vitro their effect on the growth of human hair follicle.

2.4. Cell proliferation assay Cells (104) were seeded in 24-well plates and treated with ferrocenyl derivative (final concentration 0.1 or 1 lM) in sextuplate. After 96 h of drug contact, cell proliferation was measured by Alamar Blue assay. Data were representative of three independent experiments. T-Tukey tests were performed to compare every condition with each other.  means p-value <0.05. 2.5. Measurement of intracellular reactive oxygen species To analyze intracellular ROS levels in living cells, we used the fluorescent probe 5-(and -6)-carboxy-20 ,70 -dichlorodihydrofluorescein diacetate (H2DCFDA, C2938, Molecular Probes Inc., Eugene, OR). Cells (4  104) were seeded in 48-well plates and left to grow overnight. At 24 h, medium was replaced with a solution of 10 lM H2DCFDA in PBS (with calcium and magnesium) and cells were incubated at 37 °C for 20 min. Dye solution was then replaced by medium, and cells were left to recover at 37 °C for 30 min, before incubation with hydrogen peroxide (250 lM) and ferrocifenyl derivative (1–20 lM). DCF fluorescence was measured with a Fluoroscan, using an excitation wavelength of 485 nm and a 538 nm emission filter, after 20 min, 1 h 30 min and 3 h incubation times. T-Tukey tests were assessed to compare all conditions with each other.  indicates a p-value <0.05. 2.6. Human hair follicle culture in vitro

2. Materials and methods 2.1. Chemicals The ferrocenyl derivatives Fc-OH-Tam and Fc-diOH were synthesized as previously described [16,17]. Stock solutions (1 mM) were prepared in DMSO. They were stable at 4 °C for several months.

Individual human terminal scalp hair follicles were isolated as previously described [27]. Each isolated hair follicle with its perifollicular external connective tissue sheath was incubated at 37 °C in a humidified atmosphere with 5% CO2/95% air. Hair follicles were maintained free floating in William’s E culture medium supplemented with 2 mM L-glutamine, 10 lg/ml insulin, 10 ng/ ml hydrocortisone and antibiotics (GIBCO, Bethesda, MD, USA) [28].

2.2. Cell culture 2.7. Western blotting Radial growth phase WM35 and vertical growth phase WM793 primary melanoma cells together with the metastatic WM9 melanoma cell line were kindly provided by Dr. M. Herlyn (Wistar Institute, Philadelphia, Pennsylvania, USA) [23,25]. WM35, WM9, as well as the TRP-2 overexpressing WM35 derived clone WM35-C2 [26] were cultured in RPMI 1640 medium (GIBCO, Bethesda, MD, USA) containing 5% fetal bovine serum (FBS), 1% L-glutamine and 0.5% penicillin–streptomycin (GIBCO, Bethesda, MD, USA). WM793 cells were cultured in MCDB 153 supplemented with 20% L-15 medium (GIBCO, Bethesda, MD, USA), 2% FBS, 5 lg/ml insulin and 1.68 mM CaCl2. Neonatal human epidermal melanocytes (NHEM) (Promocell, Heidelberg, Germany) were cultured in M2 medium (Promocell, Heidelberg, Germany). Cells were all maintained at 37 °C in a humidified atmosphere with 5% CO2/95% air.

Cells were lysed on ice in Tris–HCl 10 mM (pH 7.2), 2% SDS, 1% Triton X100, 10% glycerol, 1% protease and phosphate inhibitors. Eight microgram protein samples were loaded on NuPAGE 4–12% Bis–Tris gel (Invitrogen, Carlsbad, CA, USA) for electrophoresis. Proteins were then transferred onto Hybond-ECL nitrocellulose membranes (Amersham Biosciences, Piscataway, NJ, USA). After blocking with non-fat dry milk, membranes were incubated with rabbit polyclonal antibody against TRP-2 (apep8h, 1:1500, generous gift from V. Hearing), and mouse monoclonal antibody against Vimentin (VIM 3B4, 1:25,000, Cymbius Technology, Hofheim, Germany). After incubating with secondary antibodies, detection was carried out using the ECL-plus chemiluminescence reagent (Amersham Biosciences, Piscataway, NJ, USA).

1982

Q. Michard et al. / Journal of Inorganic Biochemistry 102 (2008) 1980–1985

3. Results 3.1. Effect of a short incubation period with high doses of ferrocenyl derivatives on cell viability and morphology of melanocytes and melanoma cells Neonatal human epidermal melanocytes (NHEM) and the three melanoma cells (WM9, WM35, WM793) were incubated for 24 h in the presence of various doses (1–20 lM) of Fc-OH-Tam and FcdiOH. The results on cell viability are shown in Fig. 1A and C. Incubation with Fc-OH-Tam clearly induced a disparity of sensitivity between normal and malignant cells, primary melanoma cells WM35 showing a particularly high mortality rate (IC50  2 lM) while melanocytes still maintained a 100% cell viability at a 2 lM concentration. At 5 lM, all melanoma cells showed a drastic sensitivity to the compound (cell viability 20%), whereas melanocytes kept a cell viability of nearly 50%. At this concentration, melanoma cells were round shaped, had lost adherence and floated dead in the medium (Fig. 1B compare b, c, d, respectively to f, g, h) whereas melanocytes still retained their original morphology with no sign of suffering (Fig. 1B compare a–e). So, in this cell viability test Fc-OH-Tam clearly exhibited a selective toxicity for melanoma cell lines. In the same test, the second ferrocenyl derivative, Fc-diOH, was far less cytotoxic, with a significant antiproliferative effect of about 50% arising only at the high concentration of 5 lM, and on the primary melanoma cell line WM793 only (Fig. 1C). At 10 lM, Fc-diOH became toxic on three out of the four cell lines, the two primary melanoma cells and normal melanocytes, but it was significantly less toxic on the metastatic melanoma cells (WM9). So under these tested conditions (short incubation time, high concentration) FcOH-Tam was more cytotoxic and more selective for malignant cells than Fc-diOH.

3.2. Effect of a long incubation time with low doses of ferrocenyl derivatives on the proliferation of melanocytes and melanoma cells The antiproliferative effect of the two tested derivatives was next studied after a 96 h incubation time in the presence of low concentration (0.1 and 1 lM) of compound (Fig. 2). At 0.1 lM, Fc-OH-Tam had almost no effect on any of the four cell types (cell proliferation around 90% of the control) while at 1 lM, it exhibited strong cell growth inhibition that selectively and significantly affected melanoma cells compared to normal melanocytes (Fig. 2A). Actually, at this concentration the three melanoma cells (WM9, WM35 and WM793) showed a dramatic growth inhibition while the growth of melanocytes was only slightly affected (18– 35% vs. 80% of surviving cells, respectively). Turning to Fc-diOH, a low concentration of 0.1 lM induced strong cell growth inhibition by about 50% in the two primary melanoma cells (WM35 and WM793) but not in melanocytes and metastatic cells WM9 (Fig. 2B). At 1 lM, Fc-diOH showed a dramatic antiproliferative effect on all cell lines including normal melanocytes which confirmed that ferrociphenol seemed to be less likely to discriminate between normal and malignant cells than ferrocifen. Under these conditions involving low concentration and long incubation time, the inversion of toxic potency between the two compounds should be noted. Fc-diOH here being clearly more toxic than Fc-OH-Tam (compare Fig. 1A–C to Fig. 2A and B). 3.3. Effect of ferrocenyl derivatives on the induction of reactive oxygen species (ROS) in melanocytes and melanoma cells We next analyzed the potential induction of intracellular ROS in the different cell lines [18] using the commonly used H2DCF-DA fluorescent probe which is easy to setup, known for its stability and high sensibility to detect hydroperoxides production levels

Fig. 1. Melanocyte (NHEM), primary (WM35 and WM793) and metastatic (WM9) melanoma cell survival after a 24 h incubation with high dose (1–10 lM) of ferrocenyl derivatives. (A) Cell viability assay after incubation with Fc-OH-Tam.  means p-value <0.05. (B) Phase contrast microscope examination of melanocytes (a; e), WM9 (b; f), WM793 (c; g), WM35 (d; h) melanoma cells, non-treated (a–d) or after 24 h incubation with 5 lM of Fc-OH-Tam (e–h). (C) Cell viability assay after incubation with Fc-diOH.  means p-value <0.05.

Q. Michard et al. / Journal of Inorganic Biochemistry 102 (2008) 1980–1985

1983

Fig. 2. Cell viability assays after a 96 h incubation with low doses (0.1–1lM) of FcOH-Tam (A) or Fc-diOH (B).  means p-value <0.05.

[29]. In this test, cells were incubated with H2DCF-DA, and DCFH was generated by cell esterases located in cell membranes [30,31]. DCFH was then able to react with ROS present in the cell and the fluorescent intensity of DCF, the ROS-oxidized product, was measured after 20 min, 1 h 30 min and 3 h incubation time at non-lethal (1 lM) and lethal (10 lM) concentrations of ferrocenyl derivative. Incubation in the presence of 250 lM H2O2 was used as a positive control. As shown in Fig. 3, the intracellular ROS level in untreated cells remained approximately constant in all cell lines, while the ROS level induced by H2O2 varied greatly between the cell lines. It was negligible on melanocytes, significant on WM793 and very high on WM9. Surprisingly, neither ferrocenyl derivative induced any increase in ROS signal whatever the cell line, even after incubating for 3 h. In fact the ferrocifenyl derivatives slightly inhibited the natural, progressive, time-dependent production of ROS since a significant difference in ROS signal between untreated cells and cells treated with 10 lM Fc-OH-Tam or Fc-diOH appeared after incubating for 3 h, especially in melanocytes and WM793. After 24 h, deleterious effects of both compounds were found by microscopic examination and cell viability assays (data not shown).

Fig. 3. Detection of ROS with the fluorescent probe H2DCFDA in melanocytes and melanoma cells, after incubating for (A) 20 min, (B) 1 h 30 min (C) 3 h with 1 or 10 lM of Fc-OH-Tam or Fc-diOH. H2O2 (250 lM) was used as positive control.  means p-value <0.05 for a T-student test between the corresponding signal and mock signal.

3.5. Effect of the ferrocenyl derivatives on hair follicle morphology in vitro

3.4. Effect of ferrocenyl derivatives in the presence of TRP-2 TRP-2 is a protein that had been involved in melanoma resistance to oxidative stress and cisplatin [32,33]. In particular, overexpression of the protein in the melanoma cell line WM35 has been shown to be beneficial to cell management of oxidative stress generated by hydrogen peroxide or paraquat [26]. Here we submitted the TRP-2 overexpressing clone WM35-C2 (Fig. 4A) to the tested ferrocenyl derivatives, and compared its response with that of wild type cells WM35-wt (Fig. 4B–D). The results were as follows: overexpression of TRP-2 significantly reduced stress sensitivity to hydrogen peroxide (Fig. 4B) but had no effect on the toxicity of ferrocenyl derivatives since differences in viability of WM35-wt and WM35-C2 cells incubated with Fc-OH-Tam (Fig. 4C) or Fc-diOH (Fig. 4D) were statistically insignificant. Therefore, TRP-2 expression did not provide any benefit to the management of the toxicity of the tested compounds on the melanoma cell line WM35.

Since hair loss is a common side effect of chemotherapy, we investigated human hair follicle sensitivity to both derivatives. Either 0.1 or 1 lM ferrocenyl derivative was added to the medium where hair follicles were grown in vitro [27]. Medium was changed every 3 days, and fresh solution of the derivative was added each time. After 9 days of culture, pictures of hair follicles were captured under inverted microscope with a digital camera. Results are shown in Fig. 5. Control hair follicle morphology was very well preserved, the newly formed hair shaft showing a normal, regular structure. Hair follicles treated with 0.1 lM of either derivative seemed to have grown normally, and still showed a healthy structure, with all the compartments clearly identifiable. Nevertheless, one should note the thinning of the late synthesized hair shaft (arrows) reflecting some delayed injury during culture. At 1 lM, both tested compounds induced an irreversible ejection of the hair shaft (arrows), and a complete loss of hair bulb structure reminiscent of anagen to catagen transition [34]. This preliminary experiment

1984

Q. Michard et al. / Journal of Inorganic Biochemistry 102 (2008) 1980–1985

Fig. 4. Role of TRP-2 in the management of ferrocenyl derivatives induced toxicity. (A) Western blot analysis of TRP-2 expression level in WM35-wt (wild type cells) and in WM35-C2 cells (cells overexpressing TRP-2); (B) cell viability after 1 h incubation with H2O2; (C) Fc-OH-Tam; (D) Fc-diOH.

Fig. 5. Effect of the ferrocenyl derivatives on the in vitro growth of hair follicle after incubating for 9 days with 0.1 or 1 lM of Fc-OH-Tam or Fc-diOH. Pictures are representative of the observation made in 12 hair follicles per condition, treated independently.

suggested that the ferrocenyl derivatives exhibited a high toxicity toward tissues with a high proliferative activity. 4. Discussion The anticancer properties of ferrocifen and ferrociphenol were assessed on three malignant melanoma cell lines, each representing a model of distinct stages of the disease (radial, vertical, metastatic), and on normal melanocytes. Two different pharmacological approaches were considered: (i) short incubation time (24 h), high concentration of tested compounds; (ii) long incubation time (96 h), low concentration of tested compounds. In this pharmaco-

logical study, the two compounds showed different behavior. After 24 h incubation, ferrocifen was significantly more cytotoxic than ferrociphenol and also more liable to discriminate between normal and malignant cells. After a 4 day incubation, ferrocifen was still more cytotoxic on melanoma cells than on melanocytes but cytotoxicity of ferrociphenol turned out to be higher than that of ferrocifen, especially on the two primary melanoma cell lines, WM793 et WM35. In fact, IC50 values found for Fc-diOH on these two cell lines were around 0.1 lM, i.e., lower than the values reported on breast cancer cells MCF-7 and MDA-MB-231 [15,17,35,36]. The two compounds thus showed totally different patterns of toxicity, Fc-OH-Tam being more potent after a short incubation time in high concentration, while the cytotoxic effect of Fc-diOH seemed to require a lower concentration over a longer time to find expression. As these two set of experiments are completely different one can suggest that Fc-OH-Tam possesses a strong ability to induce apoptosis, while the activity of Fc-diOH is more likely associated with its oxidative metabolism presumably leading to the formation of its corresponding quinone methide. Nevertheless both compounds failed to induce H2DCFDAdetectable ROS levels in melanocytes and melanoma cells. This result was rather puzzling at first glance since significant ROS production was found to rapidly take place following 1 lM incubation with both ferrocenyl derivatives in breast cancer cells (R.A. Toillon, in preparation). Nevertheless a direct correlation between ROS production and a cytotoxic effect has not been demonstrated so far. It should be mentioned however that oxidation of biological components in cells can occur via many different pathways for example through cytochrome P450. Bolton et al. have shown that tamoxifen can be oxidized to quinone methide via this pathway [37] and we have also found that ferrocifen and ferrociphenol are actively metabolized to their quinone methides by rat liver microsomes or recombinant human cytochromes (P. Dansette, personal communication). Therefore, ROS production is not a prerequisite for quinone methides formation in cells. On the

Q. Michard et al. / Journal of Inorganic Biochemistry 102 (2008) 1980–1985

other hand, considering that melanocytes have to endure constant oxidative aggression induced by UV radiation or o-quinone toxicity through melanin synthesis, it is quite realistic to imagine that these cells and their malignant derivatives might have developed specific metabolic pathways dedicated to the efficient scavenging of ROS, a necessary condition for their survival. Finally, a significant amount of the alpha form of the estrogen receptor has been found in melanocytes [38], and estradiol and the pure antiestrogen ICI 164384 induce a stimulation of tyrosine kinase activity in melanocytes which is mediated via a non-genomic pathway [39]. As tested ferrocenyl derivatives possess a good affinity for the estrogen receptor, an antiproliferative effect via an hormonal pathway could also be considered. However, it seems rather unlikely as an antiproliferative effect of antiestrogens on melanocytes has never been described. In conclusion, the elucidation of mechanism of action of the ferrocenyl compounds still remains an open question that needs further investigations. Because ferrocifenyl derivatives may be good candidates for melanoma cancer therapy and TRP-2 melanogenic protein has been involved in cell detoxification processes against particular chemotherapeutic agents as well as oxidative stress [26,32,33,40,41], we next explored the toxic potential of Fc-OHTam and Fc-diOH towards TRP-2 overexpressing WM35 melanoma cell line. Interestingly, TRP-2 overexpression did not provide any resistance to tested compounds suggesting they might have a broad toxic spectrum with respect to melanomas. Finally, chemotherapy-induced alopecia is a common side effect of cancer treatment. It is attributable to the high proliferative rate of follicular matrix cells, which makes the hair follicle highly sensitive to cytotoxic effects of many chemotherapeutics. We thus next explored the potential toxicity of Fc-OH-Tam and Fc-diOH at low concentration on hair follicles in vitro. At 0.1 lM, no major changes in the hair follicle appeared. But at 1 lM, both compounds produced distinct changes in hair follicle morphology, reminiscent of anagen to catagen transition [34]. This preliminary experiment therefore predicted hair loss as a potential side effect of treatment with these compounds that should be considered within the framework of future clinical evaluation. Acknowledgements The authors are grateful to Dr. S. Commo for his help and advice throughout the study. The authors thank Dr. C. Bouillon and Dr. E.A. Hillard for their critical comments on the manuscript. References [1] D.L. Cummins, J.M. Cummins, H. Pantle, M.A. Silverman, A.L. Leonard, Mayo Clin. Proc. 81 (2006) 500–507. [2] I. Ahmed, Mayo Clin. Proc. 72 (1997) 356–361. [3] C. Garbe, P. Buttner, J. Bertz, G. Burg, B. d’Hoedt, Cancer 75 (1995) 2484–2491.

1985

[4] A. Barth, L.A. Wanek, D.L. Morton, J. Am. Coll. Surg. 181 (1995) 193–201. [5] C.M. Balch, S.J. Soong, J.E. Gershenwald, J.F. Thompson, D.S. Reintgen, J. Clin. Oncol. 19 (2001) 3622–3634. [6] L. Serrone, M. Zeuli, F.M. Sega, F. Cognetti, J. Exp. Clin. Cancer Res. 19 (2000) 21–34. [7] V. Gray-Schopfer, C. Wellbrock, R. Marais, Nature 445 (2007) 851–857. [8] A.A. Tarhini, S.S. Agarwala, Dermatol. Ther. 19 (2006) 19–25. [9] J.M. Kirkwood, S. Moschos, W. Wang, Clin. Cancer Res. 12 (2006) 2331s–2336s. [10] E. Atallah, L. Flaherty, Curr. Treat. Option. Oncol. 6 (2005) 185–193. [11] B. Kasper, V. D’Hondt, P. Vereecken, A. Awada, Crit. Rev. Oncol. Hematol. 62 (2007) 16–22. [12] K.S. Smalley, N.K. Haass, P.A. Brafford, M. Lioni, K.T. Flaherty, Mol. Cancer Ther. 5 (2006) 1136–1144. [13] E.A. Hillard, P. Pigeon, A. Vessieres, C. Amatore, G. Jaouen, Dalton Trans. (2007) 5073–5081. [14] G. Jaouen, S. Top, A. Vessieres, G. Leclercq, J. Quivy, CR Acad. Sci. Paris Serie IIc (2000) 89–93. [15] G. Jaouen, S. Top, A. Vessieres, in: G. Jaouen (Ed.), Bioorganometallics, WileyVCH, Weinheim, 2006, pp. 65–95. [16] S. Top, A. Vessieres, G. Leclercq, J. Quivy, J. Tang, J. Vaissermann, M. Huché, G. Jaouen, Chem. Eur. J. 9 (2003) 5223–5236. [17] A. Vessieres, S. Top, P. Pigeon, E.A. Hillard, L. Boubeker, J. Med. Chem. 48 (2005) 3937–3940. [18] D. Osella, M. Ferrali, P. Zanello, F. Laschi, M. Fontani, Inorg. Chim. Acta 306 (2000) 42–48. [19] D. Osella, H. Mahboobi, D. Colangelo, G. Cavigiolio, A. Vessieres, Inorg. Chim. Acta (2005) 1993–1998. [20] E. Hillard, A. Vessieres, L. Thouin, G. Jaouen, C. Amatore, Angew. Chem. Int. Ed. Engl. 45 (2006) 285–290. [21] G. Tabbi, C. Cassino, G. Cavigiolio, D. Colangelo, A. Ghiglia, J. Med. Chem. 45 (2002) 5786–5796. [22] A. Kallio, A. Zheng, J. Dahllund, K.M. Heiskanen, P. Harkonen, Apoptosis 10 (2005) 1395–1410. [23] M. Herlyn, G. Balaban, J. Bennicelli, D. Guerry, R. Halaban, J. Natl. Cancer Inst. 74 (1985) 283–289. [24] W.H. Clark Jr., M.J. Mastrangelo, A.M. Ainsworth, D. Berd, R.E. Bellet, Adv. Cancer Res. 24 (1977) 267–338. [25] M. Herlyn, J. Thurin, G. Balaban, J.L. Bennicelli, D. Herlyn, Cancer Res. 45 (1985) 5670–5676. [26] Q. Michard, S. Commo, J.P. Belaidi, A.M. Alleaume, J.F. Michelet, Free Radical Biol. Med. 44 (2008) 1023–1031. [27] S. Thibaut, C. Collin, L. Langbein, J. Schweizer, B. Gautier, Exp. Dermatol. 12 (2003) 160–164. [28] M.P. Philpott, M.R. Green, T. Kealey, J. Cell Sci. 97 (Pt 3) (1990) 463–471. [29] G. Bartosz, Clin. Chim. Acta 368 (2006) 53–76. [30] D.M. Hockenbery, Z.N. Oltvai, X.M. Yin, C.L. Milliman, S.J. Korsmeyer, Cell 75 (1993) 241–251. [31] A. Gomes, E. Fernandes, J.L. Lima, J. Biochem. Biophys. Methods 65 (2005) 45– 80. [32] B.J. Pak, Q. Li, R.S. Kerbel, Y. Ben David, Melanoma Res. 10 (2000) 499–505. [33] B.J. Pak, J. Lee, B.L. Thai, S.Y. Fuchs, Y. Shaked, Oncogene 23 (2004) 30–38. [34] S. Commo, B.A. Bernard, Br. J. Dermatol. 137 (1997) 31–38. [35] G. Jaouen, S. Top, A. Vessieres, G. Leclercq, M.J. McGlinchey, Curr. Med. Chem. 11 (2004) 2505–2517. [36] A. Nguyen, A. Vessieres, E.A. Hillard, S. Top, P. Pigeon, Chimia (2007) 725–761. [37] P.W. Fan, F. Zhang, J.L. Bolton, Chem. Res. Toxicol. 13 (2000) 45–52. [38] S.H. Jee, S.Y. Lee, H.C. Chiu, C.C. Chang, T.J. Chen, Biochem. Biophys. Res. Commun. 199 (1994) 1407–1412. [39] S.D. McLeod, M. Ranson, R.S. Mason, J. Steroid Biochem. Mol. Biol. 49 (1994) 9– 14. [40] G. Liu, Y. Akasaki, H.T. Khong, C.J. Wheeler, A. Das, Oncogene 24 (2005) 5226– 5234. [41] E. Nishioka, Y. Funasaka, H. Kondoh, A.K. Chakraborty, Y. Mishima, Melanoma Res. 9 (1999) 433–443.