Antitumor Activity and Mechanism of Action of the Hormonotoxin, an LHRH Analog Conjugated to Dermaseptin-B2, a Multifunctional Antimicrobial Peptide

Prostate cancer represents the most common cancer in men. For patients with advanced or metastatic form, treatments will be able to slow down the progression but cannot cure it even with the used of new targeted therapies. In this context, the development of innovative drugs resulting from the exploration of biodiversity could open new therapeutic alternatives. Dermaseptin-B2 (DRS-B2), a natural multifunctional antimicrobial peptide isolated from the Amazonian frog skin, has been reported to possess antitumor and antiangiogenic activities. To improve DRS-B2 pharmacological properties and target prostate tumor cells more specically, we have developed a chimeric molecule, called Hormonotoxin (H-B2) which is composed of a DRS-B2 combined with a hormonal analog, d-Lys 6 -LHRH, to target LHRH-Receptor which is overexpressed in more than 85% of prostate cancers. In vitro H-B2 has a signicant antiproliferative effect on the PC3 tumor cell line, with an IC 50 value close to that of DRS-B2. The antitumor activity of HB2 was conrmed in vivo in mouse model xenografted with PC3 tumors and appears to be better tolerated than DRS-B2. Biophysical experiments showed that the addition of the hormonal analog to DRS-B2 did not alter either its secondary structure or its biological activity. Combination of different experimental approaches indicated that H-B2 induces cell death by an apoptotic mechanism whereas DRS-B2 was shown to induce it by necrosis. These results could explain the H-B2 less toxicity compared to DRS-B2. H-B2 represents a promising targeting approach for cancer therapy. The mechanism of antitumor action of these two peptides was addressed by studying, on the one hand, their bioactive structure by three spectroscopic approaches, CD, uorescence, and 2D-NMR in the presence of micelles mimicking the plasma membranes of the target cells and, on the other hand, by measuring the cell viability after double annexin V-FITC and PI staining, the cytotoxic activity by measuring the cytoplasmic LDH released, and the DNA fragmentation induced by these peptides. Cells were seeded at a density of 5×10 3 cells/well in 96-multiwell plates in complete medium and incubated for 24 h at 37˚C in a controlled humidied 7% CO 2 environment. Cells were then treated with DRS-B2 or HB2 as indicated, for 48 h. Cell viability was measured using the 3-(4,5-dimethylthiazol2-yl)-diphenyltetrazolium bromide (MTT) dye method (Sigma, Saint Quentin Fallavier, France) according to the manufacturer’s instructions. Each experiment was performed in triplicate and at least three independent experiments as previously described by Dos santos et al. 5 . IC 50 values were determined by GraphPad Prism 5.0 (GraphPad Software, USA).


Introduction
Prostate cancer (PCa) is the second most frequent cancer diagnosis made in men and the fth leading cause of death worldwide in 2018 1 . For most men with PCa, their disease will follow an indolent course.
The 5-year survival rates are encouraging: 98% and 83% in the USA and Europe, respectively 2 . Localized PCa may be cured with surgery or radionuclide therapy, however, the disease recurs in approximately 20 to 30% of men treated for localized PCa and advanced disease is associated with poor outcomes.
Concerning chemotherapy, the current standard treatment for hormone sensitive metastatic PCa is androgen deprivation by luteinizing hormone-releasing hormone (LHRH) agonists or antagonists that induce castration by blocking the gonadotropic axis. Although most men with metastatic PCa initially respond to this Androgen-Deprivation Therapy (ADT), inevitably their cancer progresses on this treatment to a disease state known as castration-resistant prostate cancer (CRPC). Despite the appearance of a panel of new therapies in this indication since 2010, such as second-generation hormone therapy (abiraterone, enzalutamide), taxane-based chemotherapy (docetaxel, cabazitaxel), immunotherapy (sipuleucel-T), targeted therapy (ipilimumab, an anti-CTLA-4 antibody) and metabolic radiotherapy (Ra-223) for men with bone metastasis, the median survival for men with metastatic castration-resistant prostate cancer (mCRPC) is less than 2 years 2 . In this context, the development of innovative therapies therefore represents a major challenge in the hope of offering to patients a more effective treatment combining low toxicity, reduced side effects and resistance associated with conventional therapies.
Among these new molecules, those from biodiversity could represent a particular interest. A group of interesting peptides from natural sources are antimicrobial peptides (AMPs) 3 . Indeed, in recent years, an increasing number of articles show that these AMPs are in fact multi-functional peptides such as anticancer agents, immuno-modulators, chemokines, vaccine adjuvants, or regulators of innate defense [4][5][6] .
From this group of AMPs, the cationic antimicrobial peptides (CAPs) offer a concrete development path derived from biodiversity. Characterized for nearly thirty years, they were initially studied for their antimicrobial virtues. Indeed, a growing number of studies have shown that some of these peptides, which are toxic to bacteria but not to normal mammalian cells, have a broad spectrum of cytotoxic activity against cancer cells. Electrostatic interactions between the positive charges carried by CAPs and the anionic components of cell membranes are considered to be the major elements involved in the selective destruction of cancer cells 7 . Among CAPs, dermaseptins B2 and B3 (DRS-B2 and DRS-B3) are 2 natural antimicrobial peptides isolated from the skin of an Amazonian tree frog of the genus Phyllomedusa bicolor 8, 9 . Our team has initially reported signi cant antitumor activity of DRS-B2 and DRS-B3 on various human cell lines including prostate cancer while they have no effect on normal cells 5,10 . Furthermore, the anti-tumor activity of DRS-B2 was con rmed in vivo in PC3 prostatic tumor cell line xenografted in athymic mice model. Treatment with DRS-B2 at 2.5mg/kg body weight six times a week in peritumoral reduced the tumor growth of 50% after 35 days. Concerning the mechanism of action of DRS-B2, our previous studies suggested a rapid mechanism of cell death, with aggregation at the plasma membrane of cancer cells and penetration into the cytoplasm and the nucleus 5,11 .
These polycationic peptides, that appear to interact and speci cally cross the membrane of the tumor cells, and with no effect on normal cells represent an innovative technological platform for the development and design of original molecules that can be used in the targeted treatment of cancers resistant to current therapies. However, its in-vivo use could question the problem of signi cant toxicity regarding the doses used. To reduce the toxicity of pharmacological molecule, the targeted therapy could represent a promising way. The concept of hormonotoxin (H-B2) was born from this problem, based on tumor targeting by associating DRS-B2 with a hormone (H) whose receptor is over-expressed on the tumor surface. This concept is related to the immunotoxin approach, which combines a toxin with a monoclonal antibody 12 . The speci c interaction of the ligand with its receptor would ideally allow targeting tumor cells while optimizing the interaction of the peptide with the membrane. The advantage of peptides is the simplicity of their production by chemical synthesis. Numerous studies have shown that many membrane receptors are over-expressed on the surface of cancer cells and that there are natural or synthetic peptide ligands (agonists or antagonists) that bind to them with very good a nity and selectivity 13 . This is the case of the luteinizing hormone-releasing hormone receptor (LHRH-R), whose expression rate is greater than 80% in endometrial, ovarian and prostate cancer cells 14 . Available data strongly suggest that about 91% of prostate cancers express LHRH-R high-a nity binding sites 15 . In these cancers, in vitro proliferation may be inhibited by the analogues of LHRH in a dose-and timedependent manner [15][16][17] . As a result, LHRH-R appears to be an ideal target for the development of personalized therapeutic treatment of various human cancers. An example is the fusion of LHRH or its analogues with various bacterial and plant toxins that have been used to target and kill cancer cells expressing LHRH receptors [18][19][20][21][22] .
In this paper we reported the design and synthesis of the chimeric peptide H-B2, a molecule composed of dermaseptin-B2 coupled to an LHRH analogue, and studies of its structure and biological activities in vitro and in vivo, as well as the deciphering of its antitumor mechanism of action.

Results
The Hormonotoxin H-B2 induces a signi cant antiproliferative effect on the hormone-resistant prostate tumor cell line PC3 and has a low hemolytic activity.
Preliminary experiments were performed to investigate the existence of a correlation between LHRH receptor expression and the effect of H-B2. For this purpose, we used four different prostate cell lines (PC3, DU145, 22RV1 and BPH-1) which are presented in Table 1. The PC3, DU145, and 22Rv1 cell lines were used as models for androgen-independent prostate cancer. PC3 and DU145 were derived from bone and brain metastasis, respectively. 22Rv1 cells were originally derived from a primary site of an advanced PCa that was serially transplanted in nude mice and BPH-1 is a prostatic hyperplasic cell line. Western blot analysis indicated that LHRH-R is strongly expressed by metastatic PC3 and DU145 cells, but its expression is very low in 22Rv1 and absent in BPH-1 cells (Fig. 2A)  cancer cells that express LHRH-R compared to BPH-1 cells that do not express LHRH receptor (Fig. 2B, 2C). Since PC3 cells present a relatively good sensitivity to H-B2 and DRS-B2, we decide to use this cell line for further investigation concerning the in vivo study of H-B2 and its antiproliferative mechanism of action.
To assess the functional role of LHRH-R on the H-B2 antiproliferative activities, LHRH-R was knocked down in PC3 cells. The expression level of LHRH-R was validated by WB analysis (Fig. 3A). The viability test showed that knockdown of LHRH-R in PC3 cells slightly decreased the sensitivity of PC3 cells to H-B2 as compared to si-control or parental PC3 cells who express LHRH-R (Fig. 3B). The calculated IC 50 was 3.73 µM for siLHRH-R PC3 compared to 2.95 µM for si-control PC3 and 3.05 µM for PC3 parental cells.
Collectively, these results indicated that the antiproliferative activity of H-B2 is higher in cells which express LHRH receptor, such as prostate cancer cells.
Since CAPs could present hemolytic activity, H-B2 and DRS-B2 were tested on human red blood cells. The hemolytic activity was determined by incubation of DRS-B2 or H-B2 at different doses with human erythrocytes. In parallel, the erythrocytes were incubated in the presence of 0.2% (v/v) Triton representing the positive control (100% hemolysis). Results presented in Fig. 4 show that DRS-B2 and H-B2 have a low hemolytic activity. Indeed, while some cationic antimicrobial peptides show hemolytic activity, DRS-B2 and H-B2 do not cause more than 10% hemolysis at the highest doses tested (50 µM) and could be considered as low cytotoxicity peptides. This nding is encouraging and essential for in vivo testing and for potential therapeutic use.
Hormonotoxin H-B2 signi cantly inhibits tumor growth in xenografted mice without measurable side effects.
To investigate the antitumor e cacy of the H-B2, we performed in vivo experiments using xenografting PC3 tumor cells in nude mice. Fourteen days after injection of PC3 cells, tumors of around 100 mm 3 were developed. The mice were then randomized into four groups (n = 6 mice/group) and treated by IP injection of 2.5 and 5 mg/kg of H-B2 or 2.5 mg/kg of DRS-B2 or vehicle, 3 times per week for ve weeks. The results showed that H-B2 has inhibited PC3 tumor growth in a dose-dependent manner (Fig. 5). The inhibition of the tumor growth was about 35% and 54% when mice were treated with H-B2 at 2.5 mg/kg and 5 mg/kg, respectively ( Fig. 5A). At the dose of 2.5 mg/kg, H-B2 has a better antitumor effect (35%) than DRS-B2 (26%) (Fig. 5A). Similar results were obtained by analyzing the weights of tumors harvested from tumor-bearing mice (Fig. 5B).
To further de ne the effect of H-B2 on tumor proliferation in vivo, Ki67 labeling was performed on tumor sections from each group. As shown in Fig. 5C, the proliferation index was signi cantly inhibited in H-B2treated mice (35% inhibition at 2.5 mg/kg and 68% at 5 mg/kg) as compared to the control group.
Taken together, these results indicate that HB2 has a better antitumor effect than DRS-B2 and less toxicity in mice (data not shown).

Mechanism of antitumor action of the Hormonotoxin H-B2 in comparison with DRS-B2
The mechanism of antitumor action of these two peptides was addressed by studying, on the one hand, their bioactive structure by three spectroscopic approaches, CD, uorescence, and 2D-NMR in the presence of micelles mimicking the plasma membranes of the target cells and, on the other hand, by measuring the cell viability after double annexin V-FITC and PI staining, the cytotoxic activity by measuring the cytoplasmic LDH released, and the DNA fragmentation induced by these peptides.
Addition of the hormonal analogue (d-Lys 6 -LHRH) to the DRS-B2 did not alter the secondary structure of the hormonotoxin H-B2.
Preliminary indications of the peptides secondary structures obtained by CD measurements in PBS and in zwitterionic detergent (DPC) at different concentrations are shown in Fig. 6. The CD spectrum of DRS-B2 and H-B2 in PBS buffer showed that these peptides have very little ordered structure. However, there is a clear change in the spectra of these two peptides as soon as the critical micellar concentration (CMC) of DPC is reached ( Fig. 6A.1 for DRS-B2 and 6B.1 for H-B2) that indicates an enhanced helical content, with minima at 208 and 222 nm. Deconvolution of the spectra allows us to quantify the relative proportions of the secondary structures of our peptides. Thus, in the absence of a micelle in the medium, the observed αhelix ratio is about 11% for DRS-B2 and 10% for HB2 as shown in Fig. 6A.2 and 6B.2, respectively. When the concentration of DPC is increased to 5 mM, the α-helicity rates then reach 25% and 55% for H-B2 and DRS-B2 respectively.
The presence of a W residue in the sequence of DRS-B2 and H-B2 allows us to study the in uence of the microenvironment on the structure of these two peptides. Indeed, the uorescence emission maximum of W is between 320 and 355 nm when excited at 290 nm, and the maximum emission wavelength re ects the exposure of W to the solvent. This uorescence is measured in aqueous solution (PBS 1x) for observation in a non-structural environment (the peptides does not form an α-helix in water) and in micellar solution to study the effect of a lipid-like microenvironment ( Fig. 6A.3 and 6B.3). We observe that beyond 1 mM, which is the CMC of DPC, the uorescence emission maxima of DRS-B2 and H-B2 shift to shorter wavelengths ("blue shift") and show a strong increase in uorescence intensity (hyperchromic shift). These spectral changes re ect a change from hydrophilic to hydrophobic environment that can be explained either by the burial of the W residue within the hydrophobic layers of DPC micelles, or by burying the W after folding of the peptide following its conformational changes.
Finally, the structure of H-B2 was investigated using NMR in micellar DPC solution. This technique allows a direct analysis of peptide conformation and exibility at the residue level. For each residue of the peptide chain, the chemical shift deviation (CSD) of the Hα protons were calculated (Fig. 7A). This corresponds to the difference between the observed chemical shift of the Hα protons and the "random coil" chemical shift ( The cell death induced by H-B2 and DRS-B2 uses two different mechanisms that are related to apoptosis and necrosis, respectively.
To obtain more information about how H-B2 acts on PC3 tumor cells, the anti-proliferative mechanism of action of this peptide was studied in comparison with that of DRS-B2. Since the known anti-proliferative effect of DRS-B2 on tumor cells resembled more to a membrane killing-like effect, experiments concerning the cytotoxic effect, the induction of apoptosis or necrosis were investigated with H-B2 in comparison with DRS-B2.
A key feature of cells undergoing apoptosis, necrosis, and other forms of cellular damage could be analyzed by measuring the activity of cytoplasmic lactate dehydrogenase (LDH) released by damaged cells. Previously, we have shown that DRS-2 increased this release 11 . Thus, PC3 cells were treated with different concentrations (1, 5, 7.5 and 10 µM) of H-B2 or DRS B2 for 24 h and the amount of released cytoplasmic LDH into the medium was measured (Fig. 8). Cells treated with Triton X100 0.9% (v/v) were used as an internal positive control. The results show a low release of cytoplasmic LDH when the PC3 cells are treated with 1 µM of H-B2 or DRS-B2 with 15 and 10% release respectively. The LDH release is maximal upon addition of 5 µM or higher doses of these peptides. However, this maximum release was limited to 65-70%. These data suggested that the cytotoxic effect of H-B2 on PC3 cells is comparable to that of DRS-B2.
To further characterize the effect of H-B2 and DRS-B2 on cell death in PC3 cells, we rst examined the viability of cells treated with these peptides by ow cytometry using annexin V-FITC and PI double staining of the cells. As shown in Fig. 9, ow cytometry analysis showed that H-B2 promoted cell apoptosis in a dose-dependent manner. After being treated with H-B2 for 24 h, 4%, 15%, and 63% of double staining A-V + /PI − which corresponds to early apoptotic cells were found in PC3 cells treated with 1 µM, 2.5 µM, and 5 µM, respectively. These values were signi cantly higher than that of untreated cells  (Fig. 10). PC3 cells treated with 10 nM Taxotere or 1 µM of Staurosporin as positive controls show 96% and 86% DNA fragmentation, respectively (Fig. 10). These data con rm that the mechanism of PC3 cell death induced by H-B2 is different from that induced by DRS-B2, which is necrotic, and could rather correspond to an apoptotic mechanism, in agreement with the ow cytometry results described previously.

Discussion
With 1.1 million prostate cancers diagnosed worldwide in 2012, of which approximately 6% are metastatic from the start, the development of treatments for mCPRC is a global necessity with a promising market. The development of treatments for the management of castration-resistant metastatic patients is necessary because currently available treatments only allow a few months of survival gain mainly due to problems of drug resistance of cancer cells. In this context, the use of natural or synthetic peptides could offer interesting therapeutic approaches. To limit peripheral toxicity and increase local concentration, numerous cytotoxic molecules conjugated with peptide hormones such as LHRH or somatostatin whose receptors are widely overexpressed on the surface of tumor membranes have been developed [23][24][25][26][27] . Currently, only one cytotoxic peptide is in clinical development (Phase 2 completed), AEZS-108 (zoptarelin doxorubicin) 28 .
Starting from our previous studies on the anticancer activity of the CAP dermaseptin-B2 (DRS-B2), a natural peptide issued from the biodiversity, we have developed a synthetic chimeric H-B2 peptide combining DRS-B2 and LHRH peptides 5,10,11 . The use of LHRH was justi ed by the fact that 86% of prostate cancers express the LHRH receptor. A peptide targeting this receptor therefore should increase the speci city for cancer cells and decrease the in vivo toxicity of DRS-B2.
At the structural level, the coupling of the peptide hormone (H) LHRH to the C-terminus of DRS-B2 does not modify the structure of DRS-B2 nor its conformational behavior in a membrane environment. The characteristics essential to the cytotoxic mode of action of these polycationic peptides and to the targeted molecular interaction are therefore preserved. CAPs are supposed to act on cancer cells in which the outer layer of the plasma membrane is highly negatively charged as for bacteria plasma membranes.
In the case of cancer cells, the negatively charged membrane is due to the presence of phosphatidylserine, negatively charged mucin proteins or highly sulfated GAGs (5,25). After binding dermaseptin peptides accumulate in a carpet like manner on the outside of a lipid bilayer until a threshold concentration is reached, causing them to form pores in which the peptides are inserted with the phospholipid headgroups of the membrane. We have shown that H-B2 was structured as a continuous helix in its "toxin" portion (DRS-B2) and that the "hormone" ligand portion remains free to interact with LHRH-R.
In vitro, H-B2 was slightly more effective than DRS-B2 on the proliferation of the various prostate cancer cell lines tested. Even if the speci city of interaction with LHRH-R could not be demonstrated physically, the knock down expression of the LRHR receptor by RNA interference on PC3 cells slightly decreases the e ciency of H-B2, thus pointing out that addition of the peptide hormone to the DRS-B2 was bene t. We could also mention that cells expressing high level of LHRH receptor like PC3 and DU145 presented a modest but signi cative gain in IC 50 when compared with those of DRS-B2 alone. Targeting cancer cells with an overexpressed receptor like LHRH is a strategy that could permit to improve the e ciency of a peptide like DRS-B2 to gain in its ED 50 . However, in the case of H-B2, this gain is low. We could imagine that even if the LHRH receptor is overexpressed on PC3 cells, its concentration at the surface of the plasma membrane is too low to guide enough DRS-B2 to membrane to complete its carpet like structure to further form pores in the plasma membrane. Since saturation of the LHRH receptor by H-B2 could not be su cient to kill cells, therefore, additional part of DRS-B2 in H-B2 structure could be necessary to do it.
In vivo H-B2 inhibited tumor growth by more than 50% and signi cantly decreased proliferation without any major side effects. Since the addition of the hormone peptide to the DRS-B2 permits a better tolerance when injected in mice, we could conclude that despite signi cant improvement in IC 50 on the cell line, the hormone peptide improved the in vivo activity of DRS-B2 since we could treat mice with 5 mg/kg H-B2 without side effects instead of 2.5 mg/kg with DRS-B2. It is interesting to note that the treatment of prostate tumor cells with H-B2 is globally well tolerated. We did not observe any major abnormalities in the blood workup after H-B2 injection. Importantly, at a dose of 10 mg/kg, DRS-B2 is lethal while H-B2 at the same dose was well tolerated in toxicity tests, with no behavioral changes in the mice or weight loss. The design of the H-B2 chimeric peptide thus made it possible to circumvent the toxicity of DRS-B2 while maintaining its anti-tumor e cacy.
Concerning the mechanism of action, the combination of different experimental approaches used in this study such as cell viability by ow cytometry, cytotoxicity by cytoplasmic LDH release, and DNA fragmentation by TUNEL essay allowed us to show that the mechanism of cell death induced by H-B2 is probably different from that of DRS-B2 and might be similar to apoptosis. Indeed, we showed that both peptides have the same effect on cytoplasmic LDH release but a distinctly different cell labeling when cell viability is analyzed by ow cytometry and DNA fragmentation by TUNEL assays. The differences between the effect of the two peptides observed in the TUNEL assay show that H-B2 induces DNA fragmentation up to 70% in contrast to DRS-B2 which is less than 10%. These results are also in agreement with those obtained in 2012 by Van Zoggel et al who showed a double Annexin V+/PI+ labeling of PC3 cells treated with DRS-B2, suggesting a necrotic cell death mechanism 11 .
Apart from DRS-B2 previously studied in the laboratory, two new members of the dermaseptin family have been recently reported to exhibit antitumor activities: Dermaseptin-PP from Phyllomedusa palliata and Dermaseptin-PT9 from Phyllomedusa tarsius 29,30 . Both exhibited antiproliferative activity against various human tumor cells, rapid LDH release activity, and an apoptosis-like cell death mechanism. Among these dermaseptins, the DRS-B2 presented the higher ED 50 on various tumor cells.
Finally, the study of the hemolytic effect of the different peptides allowed us to observe that both DRS-B2 and H-B2 have a low hemolytic activity, which is encouraging in the perspective of a therapeutic approach. Further in vivo studies on mice could be necessary to obtain more pharmacokinetic and toxicology information to complete the study.

Experimental Procedures
Peptide synthesis.
The LHRH analogue containing a dK residue (d for D con guration of the Lys residue) in position 6 (peptide A: pEHWSY(dK)LRG-amide) and DRS-B2 (peptide B: GLWSKIKEVGKEAAKAAAKAAGKAALGACSEAV-acid) were synthesized by X'PROCHEM Company (Lille, France). The chimeric peptide H-B2 (Fig. 1A) composed of the peptide A grafted by its epsilon NH 2 of the dK 6 residue with the COOH-terminal of the peptide B is determined to be >95% pure by RP-HPLC (Fig. 1B) and its molecular weight determined by ESI-mass spectrometry is [M+4H]/4=1105.2 (Fig. 1C). The peptide is dissolved in sterile water for experimental use.
The helical structure of HB2 and DRS-B2 were analyzed by CD spectroscopy using a Jobin Yvon CD6 dichrograph linked to a PC micro-processor as described 31 . Brie y, measurements were calibrated with (+)-10-camphorsulfonic acid and performed with 10 mM HB2 or DRS-B2 diluted in PBS alone or with increasing concentrations of dodecylphosphocholin (DPC) (10, 30, 100, 1000 and 5000 μM) at 25 C using a quartz cuvette (Hellma) with a path length of 0.1 cm. Spectra, recorded in 1 nm steps, were averaged over ve scans, and corrected for the base line. The CD spectra were deconvoluted using CDNN Software 32 . Circular dichroism measurements are reported as Δε/n, where Δε is the dichroic increment (M -1 cm -1 ) and n is the number of residues in the peptide. The α-helix content of peptides was obtained using the relation: Pα = -[Δε 222nm x10] (Pα: percentage of α-helix; Δε 222nm : dichroic increment per residue at 222 nm) 33 .
Fluorescence of tryptophan-containing peptides performed as previously described by Dos Santos et al. 5 Emission spectra were recorded on a Jobin-Yvon Fluoromax II instrument (HORIBA Jobin-Yvon, France) equipped with an Ozone-free 150 W xenon lamp. The excitation wavelength was 290 nm, and the emission spectra were acquired at 300-360 nm. At least ve measurements for each titration point were recorded with an integration time of 1 s. The HB2 and DRS-B2 concentration in PBS was 2 μM, and the DPC concentration varied from 0 to 1000 μM. Tryptophan uorescence was determined by subtracting spectra without peptide.
Cells were seeded at a density of 5×10 3 cells/well in 96-multiwell plates in complete medium and incubated for 24 h at 37˚C in a controlled humidi ed 7% CO 2 environment. Cells were then treated with DRS-B2 or HB2 as indicated, for 48 h. Cell viability was measured using the 3-(4,5-dimethylthiazol2-yl)diphenyltetrazolium bromide (MTT) dye method (Sigma, Saint Quentin Fallavier, France) according to the manufacturer's instructions. Each experiment was performed in triplicate and at least three independent experiments as previously described by Dos santos et al. 5 . IC 50 values were determined by GraphPad Small Interference RNA (siRNA) Assay.
PC3 cell silencing of LHRH-R were performed using Lipofectamine RNaiMax (Invitrogen) according to the manufacturer's protocol as previously described by Elahouel et al. 34 . Predesigned siRNAs of either nontargeting siRNA (GGCUACGUCCAGGAG CGCACCTT) as a control or target-speci c siRNAs (AAGCAUGGAUUGGAUCAGUAATT) to knock-down LHRH-R were obtained from Euro ns Genomics.
Brie y, cells were plated overnight at 50-60% con uence, then transfected with 10 nM of either nontargeting siRNA or LHRH-R target-speci c siRNAs, in serum-and antibiotic-free Opti-MEM medium (Invitrogen). After 24 h, transfected cells were treated with HB2 as indicated for 24 h. In addition, transfection e ciency was evaluated by Western blot analysis 24-48 h after transfection.
Western blotting was performed as previously described by Elahouel et al. 34  . Immunocomplexes were revealed using a chemiluminescence detection system kit "SuperSignal™ West Pico PLUS" (Thermoscienti c) and visualized using the GBox system (Syngene).
The hemolytic activity of H-B2 and DRS-B2 was determined using fresh human erythrocytes from a healthy donor that was prepared as follows: 4 mL of whole blood was collected and centrifuged at 900 g for 10 min at 4°C to separate plasma from erythrocytes. The pellet was then rinsed with PBS pH 7.4 and centrifuged at 900 g for 10 min at 4°C. After counting and making a red cell solution at 4 × 10 8 erythrocytes/mL (diluted in PBS pH 7.4), 50 µL of a diluted peptide solution was added in cascade to which 50 µL of the erythrocyte solution was added (made in triplicate). After 1 hour of incubation at 37°C, the tubes were centrifuged at 12 000 rpm for 15 seconds at 4°C. The supernatant was then recovered. The hemoglobin present in the erythrocytes was determined in the supernatants via a plate reader at 450 nm. A parallel incubation in the presence of 0.2% (v/v) Triton was carried out to determine the absorbance associated with 100% hemolysis.
Tumor xenograft Studies were performed as previously described by Van Zoggel et al. 11 . PC3 (2 × 10 6 ) cells were injected subcutaneously into the right ank of 4-week-old male NMRI nude mice (Janvier, Le Genest-Saint-Isle, Immunohistochemical analysis of the tumors were performed as previously described by Van Zoggel et al. 11 . The xed tumors were embedded in para n and then 6 µm sections were prepared. The tumors sections were depara nized, antigen unmasking was performed, and endogenous peroxidase activity was inactivated with a 2% hydrogen peroxide solution for 10 minutes. Unspeci c staining was blocked using Power Block Universal reagent (Biogenex Laboratories/Microm Microtech, Francheville, France) for 10 min at 37°C. Tissues were then incubated 2 hours at room temperature with anti-human Ki67 antibody (Mouse monoclonal, M7240, Dako, 1:50) for 2 hours. Immuno-complexes were revealed using HRP conjugated secondary antibodies and the DAB substrate. Tissues were then counterstained with hematoxylin and cover slipped with Mowiol mounting medium. Quanti cation of Ki67 positive stained cells was quanti ed by image J software analysis on the whole tumor section.
The cytoplasm LDH release assay was performed as previously described 11       Hemolytic activity of DRS-B2 and H-B2 on human erythrocytes. Human erythrocytes were cultured with different concentrations of DRS-B2 or H-B2 for 1 h at 37°C. Cells treated with 0.2% Triton X00 were used as a positive control and correspond to 100% of the hemolytic activity used as reference.