Targeting Bcl-2/Bcl-XL Induces Antitumor Activity in Uveal Melanoma Patient-Derived Xenografts

Purpose Uveal melanoma (UM) is associated with a high risk of metastases and lack of efficient therapies. Reduced capacity for apoptosis induction by chemotherapies is one obstacle to efficient treatments. Human UM is characterized by high expression of the anti-apoptotic protein Bcl-2. Consequently, regulators of apoptosis such as Bcl-2 family inhibitors may constitute an attractive approach to UM therapeutics. In this aim, we have investigated the efficacy of the Bcl-2/Bcl-XL inhibitor S44563 on 4 UM Patient-Derived Xenografts (PDXs) and derived-cell lines. Experimental Design Four well characterized UM PDXs were used for in vivo experiments. S44563 was administered alone or combined with fotemustine either concomitantly or after the alkylating agent. Bcl-2, Bcl-XL, and Mcl-1 expressions after S44563 administration were evaluated by immunohistochemistry (IHC). Results S44563 administered alone by at 50 and 100 mg/kg i.p. induced a significant tumour growth inhibition in only one xenograft model with a clear dose effect. However, when S44563 was concomitantly administered with fotemustine, we observed a synergistic activity in 3 out of the 4 tested models. In addition, S44563 administered after fotemustine induced a tumour growth delay in 2 out of 3 tested xenografts. Finally, IHC analyses showed that Bcl-2, Bcl-XL, and Mcl-1 expression were not modified after S44563 administration. Conclusion The novel anti-apoptotic experimental compound S44563, despite a relative low efficacy when administered alone, increased the efficacy of fotemustine in either concomitant or sequential combinations or indeed subsequent to fotemustine. These data support further exploration of potential therapeutic effect of Bcl-2/Bcl-xl inhibition in human UM.


Introduction
Uveal melanomas (UM) represent the most frequent intraocular tumour in adult patients. Whereas the 5-year overall survival rate of localized disease is greater than 70%, the prognosis drops dramatically in cases of metastases [1]. Up to 50% of patients will develop metastases within a median time of 2.4 years [2]. The overall survival is poor, and the majority of patients will succumb to their cancer. Systemic therapy with alkylating agents, i.e. fotemustine, dacarbazine, or temozolomide, have shown only modest efficacy [3]. Consequently, because of the limited efficacy of current treatments, new therapeutic strategies need to be developed.
One of the primary means by which UM cells evade treatmentinduced apoptosis is by up-regulation of members of the prosurvival Bcl-2 family proteins such as Bcl-2 and Bcl-XL [4,5]. Indeed, human uveal melanomas are characterized by a high average percentage of Bcl-2 positive cells of 82% (range: 44%-100%) [6][7][8][9][10][11][12], but without any prognostic impact [6,9,10,12]. This observation was also confirmed by the present group in a panel of primary human UM xenografts obtained from patient's tumors, where Bcl-2 was shown to be expressed in almost all studied patient's tumours and corresponding xenografts [13].
A variety of approaches to target these anti-apoptotic oncoproteins have been pursued in order to try and restore the natural process of programmed cell death [14], notably bcl-2 anti-sense oligonucleotides such as Oblimersen (GenasenseH) with contrasted positive [15,16] or negative [17][18][19] impact in randomized clinical trials performed in cutaneous melanoma, chronic lymphoid leukemia, multiple myeloma, and prostate cancer patients. Another potential therapeutic approach consists of using small molecules that occupy the BH3 binding groove of antiapoptotic Bcl-2 family members (BH3 mimetics), including ABT-737 [20], ABT-263 (NavitoclaxH) [21], and GX15-070 (ObatoclaxH) [22]. These drugs disrupt Bcl-2/Bcl-X L interactions with pro-death proteins (e.g., Bim), leading to the initiation of apoptosis. In human tumor cells, BH3 mimetics induce Bax translocation, cytochrome c release, and subsequent apoptosis. In human xenograft models of small-cell lung cancer, multiple myeloma, lymphoblastic leukaemia, and aggressive B-cell lymphoma, Bcl-2/ Bcl-X L inhibitors were previously shown to significantly enhance the efficacy of clinically relevant therapeutic regimens [21,[23][24][25]. As relatively few clinical studies using these new compounds have been reported [26,27], no randomized clinical data are already available.
In contrast to the high number of reports in various hematologic and solid tumors including cutaneous melanoma, only one in vitro study has evaluated Bcl-2 targeting in uveal melanoma cells, showing synergistic effect with chemotherapy and multi-drug resistance reversion [28]. Based on this observation and considering the high Bcl-2 expression in UM and the occurrence of liver metastases in its natural history, we have evaluated the therapeutic potential of a new experimental Bcl-2/Bcl-X L inhibitor, S44563, in different uveal melanoma Patient-Derived Xenografts (PDXs), and derived-cell lines, alone or in combination with fotemustine.

Ethics statement
Before PDX establishment, all patients had previously given their verbal informed consent for experimental research on residual tumor tissue available after histophatologic and cytogenetic analyses. Those PDXs establishments have been performed after approval of the ethics committee of the Institut Curie. According to the French rules and the ethics committee of the Institut Curie, a written consent from patients for obtaining residual tumor tissues is not required. In case of patient refusal that could be orally expressed or written, residual tumor tissues are not collected. All issues that patients want to be discussed could be raised at any time during any medical consultations. This procedure was approved by ethics committees. This research was not conducted outside of our country. Studies have been performed in compliance with the recommendations of the French Ethical Committee and under the supervision of authorized investigators. The experimental protocol and animal housing were in accordance with institutional guidelines as put forth by the French Ethical Committee (Agreement C75-05 -18, France), and the ethics committee of the Institut Curie that approved this project.

Compounds
S44563 is a new Bcl-2/bcl-X L inhibitor. It potentially inhibits the anti-apoptotic proteins Bcl-2 and Bcl-X L by binding to their respective BH3 binding groove. S44563 was synthesized at IdRS (France) as described (initial patent application filed in France 02/ 02/2007 under Nu07/00741). S44563 (C 44 H 47 CIN 6 0 5 S 2 ) has a molecular weight of 912.4 (chemical structure is presented in Figure S1). S44563-2 was synthesized at Institut de Recherches Servier. The purity of the batches ranges between 97.8% and 99.9%, and the optical purity was 99%. Fotemustine (MuphoranH) was kindly provided from IRIS (France).

Fluorescence polarisation assays
A fluorescent BH3 peptide (BH3 motif from the protein Puma) was incubated with either recombinant Bcl-2 or Bcl-X L protein (20 mM Na 2 HPO 4 pH 7.4, 50 mM NaCl, 1 mM EDTA, 0.05% pluronic acid). Upon binding, the fluorescence emitted by the BH3 peptide becomes polarised. The inhibition of the interaction between the fluorescent BH3 peptide and Bcl-2 or Bcl-X L is followed by measuring a decrease in fluorescence polarisation. In these assays, Bcl-2 or Bcl-X L concentration was 100 nM, the fluorescent BH3 peptide concentration was 15 nM and S44563-2 was titrated from 1 nM to 100 mM.

Uveal melanoma PDXs
Three UM cell lines have been used, all derived from UM PDXs established from intraocular lesion (MP41, MM26, and MM66). Four UM PDXs obtained from patients after enucleation (MP41 and MP77) or liver metastasis surgery (MM26 and MM66), established, characterized and maintained into SCID mice have been used [13,29]. The principal characteristics of these 4 xenografts are presented in the Table S1. We could not use each model in both in vitro and in vivo experiments, because some xenografts could not be maintained in prolonged in vitro cultures.
Cell viability was determined with WST-1 assay (Roche, Indianapolis, USA). The culture medium was replaced with a medium containing a WST-1 reagent, and 1 h later, the absorbance in the well was determined at 450 nm with a reference wavelength of 620 nm by using a microplate reader TECAN Infinite M200 (TECAN Deutschland GmbH, Crailsheim, Germany, Software Magellan TM ). The percentage of viability was evaluated by DOD (540-620) treated/DOD (540-620) control6100. The 50% growth inhibitory concentration (IC 50 ) (50% inhibitory concentration) was then defined for each drug and cell line.
For determination of apoptosis, mitochondrial transmembrane potential (DYm) and cell cycle, cells were incubated with S44563 at a concentration of 17 and 34 mM for 24 h. The detection of apoptosis and DYm alterations was done by the protocol AbC Annexin 5/FITC kit (AbCys, Paris, France) and BD TM MitoScreen kit (BD Biosciences, San Jose, CA, USA) respectively with DAPI (1/200, Invitrogen, Carlsbad, CA, USA). The study of the cell cycle by propidium iodure (1/100, Sigma-Aldrich Chemie, Saint-Quentin Fallavier, France) was performed after fixing cells with ethanol 70% and permeabilization with RNase A (1/100, Roche, Indianapolis, IN).
All results were statistically analyzed by two-way ANOVA test with Bonferroni post-test versus control group. Data were represented as mean 6 standard error of mean (n = 3-5 replicates). p,0.05 were considered significant (*).

In vivo experiments
SCID mice, 4 to 6 weeks old, bred at Curie Institute, were used for in vivo experiments. Fragments of 30-60 mm 3 were grafted subcutaneously into the interscapular fat pad. When tumors reached a size of 60-200 mm 3 , mice were randomly assigned to the control or treatment groups and treatment was started on day 1. Between 8 to 12 mice per group were included in in vivo experiments. Tumor growth was evaluated by measurement of two perpendicular diameters of tumors with a caliper, twice a week. Individual tumor volume, relative tumor volume (RTV), and tumor growth inhibition (TGI) were calculated according to standard methodology [13]. Statistical significance of TGI was calculated by the paired Student's t-test. Xenografted mice were sacrificed when their tumor reached a volume of 2500 mm 3 .
In all in vivo experiments, a relative tumor volume variation (RTVV) of each S44563-treated mouse was calculated from the following formula: [(Vt/Vc)21], where Vt is the volume of the treated mouse and Vc the median volume of the corresponding control group at a time corresponding to the end of treatment.

Immunohistochemical studies
All immunohistochemistry protocols were realized automatically on a Ventana TM Discovery XT machine with neutral buffered formalin-fixed (4%) tissue sections of paraffin-embedded tumor biopsies. In each case, 5-mm sections were pre-treated with standard CC1 reagent (Ventana TM Tris/Borate/EDTA buffer pH 8-48 min-95uC) for Bcl-2,) or mild CC1 (28 min detections on MM26 and MP41 tumours the chromogenic system was NBT/BCIP and sections were counter-stained with Nuclear Fast Red. All the incubations were performed at 37uC with the exception of pre-treatments (95uC).
Intensity and percentage of positive stained (DAB) cells were then quantified by an anatomopathologist (Dr J.F. Boivin). Two parameters were used to evaluate Bcl-2, Bcl-X L , and Mcl-1 expression: semi-quantitative visual grading system of the intensity of staining (2, negative; +, weak; ++ modest; +++ strong), and the percentage of labelled tumor cells. To integrate these two parameters, a H-score (histological score) was calculated as Hscore = tumor cell staining percentage6staining intensity. One parameter (percentage of labeled tumor cells) was used to evaluate cleaved Caspase 3.

Results
Inhibition of Bcl-2 and Bcl-X L binding to a BH3 peptide by S44563 Inhibition of Bcl-2 and Bcl-X L was investigated by fluorescence polarisation (FP) assays using recombinant Bcl-2 or Bcl-X L and a fluorescent Puma BH3 peptide. The Figures 1A and 1B show the inhibition of Bcl-2 and Bcl-X L respectively by S44563. S44563 is a potent inhibitor of Bcl-2 and Bcl-X L . Its IC 50s in a Bcl-2/F-PumaBH3 interaction assay and a Bcl-X L /F-PumaBH3 interaction assay were 131 nM [123 nM;135 nM] and 140 nM [133 nM;150 nM], respectively.

Bcl-2, Bcl-X L , and Mcl-1 expression in uveal melanoma PDXs
The determination of the expression of Bcl-2, Bcl-X L , and Mcl-1 was performed in 3 to 5 different tumors for each model. The expression of Bcl-2 was high for 3 UM xenografts used for in vivo experiments (MP41, MP77, and MM26), where the median Hscore ranged between 150 and 175, and low for the MM66 model for which the median Bcl-2 H-score was lower than 5 (Table S2) (Figure 2). In contrast, Bcl-X L was highly expressed in the MP41 xenograft (H-score of 175) and not or slightly expressed in the 3 remaining models (Table S1) ( Figure 2). Finally, Mcl-1 was expressed in a high proportion of tumor cells of all studied xenografts, except for the MM66 xenograft, but with a relative weak intensity (Table S1) (Figure 2). A global IHC score was then defined, including Bcl-2, Bcl-X L , and Mcl-1 protein expression, as: (Bcl-2 score)+(Bcl-X L score)/Mcl-1 score. The median global score differed among the 4 studied models, i.e. 3

In vitro studies
For the 3 uveal melanoma xenograft-derived cell lines MP41, MM26, and MM66 incubated with increased concentrations of S44563, 17 mM and 34 mM for 24 hours, the IC50s were 4 mM, 7 mM, and 6 mM for the MP41, MM26, and MM66 cell lines, respectively ( Figure S2). Similarly, to evaluate apoptosis induction, the MP41, MM26, and MM66 cell lines were incubated with 17 mM and 34 mM S44563 for 24 hours, after which the samples were labeled with DAPI/annexin V-FITC or JC1. We observed apoptosis induction and a drop of the DYm in the 3 UM cell lines, although the effects were less important in the MM26 cell line ( Figure 3A and 3B). Similarly, in the two UM cell lines MP41 and MM66 that are more sensitive to S44563, we observed a significant decrease of the G2/M phases which were not modified in the less sensitive MM26 cell line ( Figure 3C).
To study the impact on autophagy induction of S44563, the 3 UM cell lines were cultured with non-apoptotic doses of S44563 (2 and 4 mM for 24 hours). The amount of active (cleaved form) and inactive (non-cleaved form) protein LC3 was measured by western blotting. We observed that both the proportion and the total amount of cleaved LC3 protein relative to b-actin, were not significantly modified after S44563 administration in the 3 studied UM cell lines ( Figure 3D).

In vivo experiments
In the 4 UM PDXs, S44563 administered alone at 50 and 100 mg/kg, for 2 cycles of treatment (5 days a week for 2 weeks and 1 week off, each cycle), induced a significant antitumor effect in one model (MP41), with a dose-dependent response and an optimal TGI of 61% at day 53 after start of treatment (Table 1) ( Figure 4A). To evaluate responses to S44563 observed in the 4 models according to individual mouse variability, we decided to consider each mouse as one tumor-bearing entity. In this, when relative tumor volume variation [(RTVV)21] of each S44563treated mouse was calculated, regardless of the doses of S44563, we observed that 68% of all S44563-treated mice (59/87) had a negative ratio compared to control groups, and that 22% had a ratio lower than 50%, which constitute the threshold to consider a  Bcl-2/Bcl-X L Targeting in Uveal Melanoma PLOS ONE | www.plosone.org model as responding. Moreover, the overall response rate (RTVV#50%) was 11% for the 50 mg/kg per injection of S44563, and 32% for the 100 mg/kg dose (p,0.05), confirming the dose-dependent efficacy of the Bcl-2/Bcl-X L inhibitor ( Figures 3A and S3B).
We then evaluated administration of both S44563 and fotemustine (Tables 1 and 2). In the MP41 xenograft, a tumor growth delay was observed in the 3 groups of combined treatments, i.e. concomitant fotemustine/S44563, fotemustine first then S44563 since day 43, and concomitant fotemustine/S44563 then S44563 since day 43 (data not shown). The median OS was 44 days for the control group, 67 days for the S44563 group, 72 days for the fotemustine group, and 102 days for the fotemus-tine+S44563 (whatever the schedule of combination) (p,10 23 ) ( Figure 4B). In the MP77 xenograft, concomitant S44563 (100 mg/kg per injection) and fotemustine (15 mg/kg) induced a more significant TGI at day 35 after start of treatment than fotemustine alone (p = 0.04) ( Figure 4C). Moreover, an increased median overall survival was observed when S44563 was administered in an adjuvant setting after fotemustine-induced complete remission. In the MM26 xenograft, concomitant S44563 (50 mg/ kg per injection) and fotemustine (30 mg/kg) slightly improved the TGI in comparison to fotemustine alone, but significantly improved the complete remission rate (p,0.02) ( Table 2). In contrast, because of the occurrence of toxic deaths when S44563 was administered after fotemustine since day 43, we have not observed prolonged survival. This toxicity was observed in the only one MM26 model and could not be clearly explained due to its restrictive occurrence in the MM26 UM PDX. Finally, in the MM66 xenograft, concomitant S44563 (50 mg/kg per injection) and fotemustine (30 mg/kg) significantly enhances efficacy compared to treatment with either agent alone (p,0.003) ( Figure 4D).

Correlations between IHC expressions and ORR to S44563
We then evaluated correlations between histopathological analyses and response to S44563. We first showed that high Bcl-2 expressing tumors could be either sensitive or resistant to S44563 administered alone (Figure 2 and Table 1). A similar observation could be done for high Mcl-1 expressing tumors. In contrast, only MP41 xenograft with a high Bcl-X L expression was sensitive to S44563, whereas the 3 other models characterized by a low Bcl-X L expression were resistant to S44563 administered alone. However, because of the limited number of tested xenografts, it was not possible to determine the relation between the global pathological score, defined as: (Bcl-2 score)+(Bcl-X L score)/Mcl-1 score, and responses to S44563.

Immunohistochemical studies after S44563 administration
In order to evaluate the impact of S44563 on the expression of Bcl-2, Bcl-X L , and Mcl-1 proteins, two to four tumor samples have . After S44563 treatment, the amount of active and inactive protein LC3 was determined and reported to ß-actine. The quantity of total LC3 protein was calculated to study the activation of LC3. Results were presented as percentage and total amount of active cleaved LC3, relative to b-actin. A two-way ANOVA with Bonferroni post-test was then performed (* means a p,0.05). doi:10.1371/journal.pone.0080836.g003 Bcl-2/Bcl-X L Targeting in Uveal Melanoma PLOS ONE | www.plosone.org been collected at the time of mice sacrifice for immunohistochemical studies. IHC analyses showed that Bcl-2, Bcl-X L , and Mcl-1 expressions were not modified after S44563 administration (Table S3) ( Figure S4). Surprinsingly, we have observed that fotemustine alone induced a dramatic increase of Bcl-2 in the MM66 xenograft, and an increase of the Mcl-1 protein in the MP41 and MM66 xenografts. Consequently, the expression of both Bcl-2 and Mcl-1 was highly increased after concomitant administration of fotemustine and S44563 in the MM66 xenograft ( Figure 5).
In order to evaluate in vivo apoptosis induction, we have studied cleaved caspase-3 expression before and after S44563 +/2 fotemustine administration. Tumor samples have also been collected at the time of mice sacrifice. We did not detect significant increase of activated caspase-3 in the four treated UM xenografts (data not shown). Similarly, we did not observe LC3-positive tumor cells before and after treatments (data not shown), as in in vitro experiments, except in one MM66 tumor among five xenografts treated by S44563 (50 mg), where a unique islet of about 20 positive cells has been observed.

Discussion
Using four well-characterized uveal melanoma PDXs, we have shown that a specific Bcl-2/Bcl-X L inhibitor only had a moderate efficacy when administered alone, except in one model (MP41)  with a clear dose-dependent in vivo efficacy. However, we have shown that this Bcl-2/Bcl-X L inhibition significantly increased the antitumor activity of fotemustine in these preclinical models. Indeed, in 3 out of the 4 tested xenografts (MP41, MP77, and MM66), S44563 increased tumor growth inhibition when combined to fotemustine in comparison to fotemustine alone. It also increased the complete remission rate (MM26 model) and the overall survival (MP41 model). Overall, these data strongly suggest that targeting Bcl-2/Bcl-X L in combination with fotemustine significantly improves, in human UM xenografts, response to the alkylating agent administered alone. Hence, our treatment combination significantly enhanced efficacy compared to treatment with either agent alone, confirming positive data reported with the BH3-mimetic ABT-737 and various chemotherapies such as fotemustine [31], temozolomide [31,32], melphalan [33], cytosine arabinoside [34], gemcitabine [35], actinomycine D [36], carboplatin [37], and others, or with Bcl-2 antisense and cyclophosphamide [38] or chlorambucil [39]. In the MM66 xenograft, this observation could be explained by the fact that fotemustine highly increased Bcl-2 expression that could be therefore more efficiently inhibited. In few reports, it has been shown that this effect was p53-independent but related to the pro-apoptotic protein Noxa [31,32,40], which inactivates other Bcl-2-family members such as Mcl-1 or A1 proteins [40]. The balance of Bcl-2/Mcl-1 proteins has also been largely shown to impact the efficacy of ABT-737 [34][35][36]41,42], suggesting that modulation of Noxa and Mcl-1 could be used as a strategy for sensitizing tumor cells to ABT-737 [31,43]. Another predictive marker could be Bcl-X L expression and indeed in our in vivo experiments, the best responding xenografts were characterized by higher Bcl-X L expression, suggesting that specifically targeting the Bcl-X L protein may be an efficient therapeutic approach by itself. Such an observation is confirmed by the report of Jain and Meyer-Hermann who showed that inhibition of Bcl-X L may decrease the reparation of DNA damages induced by carboplatin [37].
One other mechanism leading to the synergistic effect between a Bcl-2 inhibitor and temozolomide could be the specific induction of autophagic cell death, as recently reported [44,45]. In our in vitro experiments testing the effect of S44563 in 3 human UM cell lines, S44563 did not modify the proportion and the total amount of Taking together, our results clearly demonstrate that Bcl-2/Bcl-X L targeting is a promising therapeutic approach for uveal melanoma. Despite a modest efficacy when used alone, S44563 significantly increased efficacy of fotemustine. Even intra-arterial hepatic administration of this alkylating agent remains poorly efficient in UM patients, both in an adjuvant setting after proton beam irradiation [46] and in metastatic evolution [47]. It is therefore required to improve its efficacy and reverse mechanisms of resistance. Targeting specific members of the Bcl-2 family, such as Bcl-2 and Bcl-X L proteins, could be of high clinical interest. In our in vivo experiments, we have evaluated this targeted approach in relevant preclinical models and showed that both in concomitant and differed administration, S44563 were able to increase the effect of fotemustine. Our data therefore suggest that S44563 could probably be evaluated in both adjuvant and metastatic settings of UM patients. In the first clinical situation, genomic features of UM tumors clearly defined prognostic outcome of the patients [48,49], allowing a relevant discrimination and selection of poor prognosis cases, and, in the second clinical situation, the overall prognosis remains very poor. Altogether, our results emphasize a potential therapeutic interest for targeting Bcl-2/ Bcl-X L UM patients.