Bombesin Analogue-Mediated Delivery Preferentially Enhances the Cytotoxicity of a Mitochondria-Disrupting Peptide in Tumor Cells

Tumor-homing peptides that recognize specific markers on tumor cells have shown potential as drug carriers for targeted cancer therapy. Bombesin receptors are frequently overexpressed or ectopically expressed in a wide range of human tumors. Bombesin and its analogues have been widely used as drug carriers for tumor imaging and tumor therapy. However, the cargos used in previous studies, including radioactive and chemotherapeutic agents, are usually small molecules. Mitochondrial-disrupting peptides depolarize the mitochondria and trigger apoptosis after entering tumor cells. We are interested in whether the bombesin analogue, Bn(6–14), which contains a bombesin receptor-binding motif, can specifically deliver the mitochondria-disrupting peptide, B28, to tumor cells. To this end, we created a chimeric peptide, B28Bn(6–14), by conjugating B28 to Bn(6–14) at its N-terminus. The cytotoxicity of B28Bn(6–14) in tumor cells was much stronger than unconjugated B28. The IC50 values of B28Bn(6–14) in tumor cells (1.7–3.5 µM) were approximately 10 times lower than B28. However, conjugation of B28 to Bn(2–7), which lacks the bombesin receptor-binding motif, did not increase its cytotoxicity. In addition, the IC50 values of B28Bn(6–14) in tumor cells (1.7–3.5 µM) was 3–10 times lower than in normal cells (10.8–16.8 µM). We found that selective binding of B28Bn(6–14) to tumor cells is Bn(6–14)-dependent. Upon entering the tumor cell, B28Bn(6–14) accumulated in the mitochondria and triggered caspase-dependent apoptosis. Intratumoral and intraperitoneal administration of B28Bn(6–14) substantially suppressed the growth of DU145 tumor xenografts in mice. These results demonstrate that Bn(6–14) is able to deliver the mitochondria-disrupting peptide to tumor cells, and B28Bn(6–14) should be further developed as novel anti-cancer agent.


Introduction
Traditional chemotherapy usually has very limited selectivity toward tumor tissues and frequently induces the emergence of multiple drug resistance due to the requirement for high drug doses [1]. Developing strategies to exhibit selective toxicity toward tumor cells relative to normal cells is currently one of the major challenges in anticancer therapy. Targeted delivery of anticancer agents to malignant cells based on tumor biomarkers has the potential to increase therapeutic efficacy while decreasing doselimiting side effects [2,3]. Tumor-homing peptide ligands represent a promising approach for the specific delivery of diagnostic and therapeutic agents, as the ligands show a strong affinity toward biomarker receptors overexpressed on tumor cells or tumor vasculature [4,5].
One strategy for targeted drug delivery by using tumor-homing peptides is the coadministration of drugs and the peptides as separate entities without conjugation. After the tumor-homing peptide selectively accumulates in tumor tissues, an additional motif in the peptide, such as CendR, induces leakage of the tumor vasculature by affecting the integrity of angiogenic endothelial cells and triggers the targeted delivery of the bystander drugs into tumor tissues [6,7]. On the other hand, most tumor-homing peptides, as leader moieties, can be conjugated to diverse cargos, including cytotoxic drugs, imaging agents, and various nanoparticles, for tumor diagnosis and targeted treatment. Based on conjugation, many tumor-homing peptide-directed agents have been used in the clinic or are undergoing clinical trials [4,5,8,9]. For instance, radiolabeled somatostatin analogues are currently used for cancer imaging and therapy. Among these analogues, 111 In-penetreotide based somatostatin receptor scintigraphy is a standard clinical procedure to determine the localization of neuroendocrine tumors [9,10]. However, the overexpression of somatostatin receptors is limited to neuroendocrine tumors [11].
Bombesin, which is an amidated tetradecapeptide isolated from frog skin, is another attractive vehicle for tumor-targeting delivery. Bombesin shares the same, or a similar, seven C-terminal amino acid sequence with gastrin-releasing peptide and neuromedin B, respectively. Therefore, the bombesin receptor family in mammals is comprised of gastrin-releasing peptide receptor (GRPR), neuromedin B receptor (NMBR), and bombesin receptor subtype 3 (BRS-3) [12]. These bombesin receptors, especially GRPR, are frequently overexpressed or ectopically expressed in many common malignancies, including lung cancer, prostate cancer, breast cancer, pancreatic cancer, head/neck cancer, colon cancer, uterine cancer, ovarian cancer, renal cell cancers, glioblastomas, neuroblastomas, gastrointestinal carcinoids, intestinal carcinoids, and bronchial carcinoids. Thus, there is special interest in developing bombesin receptor-mediated agents to treat these tumors [8,12]. Currently, numerous radiolabeled bombesin analogues are undergoing investigation for tumor imaging and radiotherapy. Some 99m Tc or 68 Ga-labeled analogues were tested in healthy volunteers or patients for diagnostic purposes [8]. In addition, a few nonradiolabeled analogues that were constructed by conjugating bombesin analogues to chemotherapeutic agents, such as camptothecin, doxorubicin, and paclitaxel, have successfully increased the selectivity or efficacy of these drugs in preclinical studies [13,14,15].
Previous studies demonstrated that peptide fragments containing residues 7-9 in the C terminus of bombesin show high affinity toward bombesin receptors [16]. These bombesin analogues have been widely studied as vehicles of tumor-imaging and targeted therapy agents. However, the cargos that have been used in these studies primarily include small molecule radiolabeled and chemotherapeutic agents [8,17,18,19]. Limited biomolecules, such as marine toxin, diphtheria toxin and nanoparticles loading siRNA, have also been fused to bombesin analogues for targeted delivery [8]. Mitochondria are considered to be the powerhouse of the cells and one of the crucial signal regulators for cell survival and death [20]. Mitochondria-disrupting peptides can efficiently activate mitochondrial membrane permeabilization (MMP) and disruption, and trigger apoptosis after being delivered into tumor cells by drug carriers, such as tumor cell-selective peptides or antibodies [21,22]. Because C-terminal fragments of bombesin containing the receptor-binding motif have been used as vehicles for small molecules, we sought to determine whether these bombesin analogues could be used for targeted delivery of a mitochondriadisrupting peptide.

Cell Culture
Human prostate cancer cells (DU145 and PC-3), human breast cancer cells (MCF-7), human skin fibroblast cells (HSF), human lung fibroblast cells (MRC-5), and human prostate stromal cells (PrSC) were from the American Type Culture Collection. Human prostatic smooth muscle cells (SMC) and human breast cancer cells (MDA-MB-435S) were obtained from the Cell Bank of the Chinese Academy of Science. All of the cells were grown in RPMI 1640 supplemented with 10% FBS, 2 mmol/L L-glutamine, 100 U/mL penicillin, and 100 mg/mL streptomycin at 37uC with 5% CO 2 .

Circular Dichroism Analysis of Peptide Secondary Structure
Each peptide, at the concentration of 0.5 mg/ml, was dissolved in 2 mM phosphate buffer, pH 7.4. The circular dichroism (CD) spectra of the peptides were recorded from 190 nm to 400 nm on a Model 400 Circular Dichroism Spectrophotometer (Aviv Biomedical, Inc.) at 25uC. Background scans were collected in buffer alone and subtracted from the peptide scan. Each sample was performed in triplicate, and the CD spectrum was obtained with the calculated average.

Cytotoxicity Assay
Cells were seeded at a density of 1610 4 per well in a 96-well plate, were grown overnight and were co-incubated at 37uC for 1 h with increasing concentrations of the peptide in 100 ml RPMI 1640 supplemented with 2% FBS. Subsequently, the peptide solution was replaced by 100 ml fresh medium. To determine cell viability, 10 ml CCK-8 was added to the well, and the cultures were incubated for an additional 2-4 h. The absorbance was then measured at 450 nm (with a reference wavelength at 620 nm), and cell viability was calculated as the percentage of the value relative to the absorbance of the cells incubated without peptide. The IC 50 values of the peptides were calculated from the respective cell viability curves. Simultaneously, the LIVE/DEAD BacLight bacterial viability kit was used to visualize the cytotoxicity of the peptides. Combined incubation with SYTO 9 (15 mM) and propidium iodide (PI, 2.5 mg/ml) stained the living cells fluorescent green and the dead cells fluorescent red due to their distinct cell membrane permeability.

Peptide Cellular Uptake and Localization in Mitochondria
To quantitate the cellular uptake of the peptide by fluorescenceactivated cell sorting (FACS), cells were detached with a nonenzymatic cell dissociation solution (Sigma Aldrich), washed with culture medium and suspended in PBS containing 0.5% FBS. Subsequently, 1.5610 5 cells were treated with 300 ml of 5 mM FITC-labeled peptides at 37uC for 30 min. The cells were collected by centrifugation, washed twice with PBS containing 0.5% FBS and subjected to FACS analysis. An unrelated FITClabeled goat anti-rabbit IgG was used as a negative control. To further examine the cellular localization of the peptide to the mitochondria, cells were plated at a density of 3610 4 cells/cm 2 and incubated with 5 mM FITC-labeled peptides in PBS containing 0.5% FBS at 37uC for 15-30 min. The mitochondrion-selective fluorescent probe, Mitotracker CMXRos (10 ng/ml), was added for the last 15 min of the incubation. After quenching the cell surface-bound FITC peptide with 200 mg/ml trypan blue, as described previously [21], the cells were gently washed with PBS containing 0.5% FBS and observed under a fluorescence microscope.

Mitochondrial Depolarization Assessment
Cells were detached, washed with medium, and 1.5610 5 cells were suspended in 300 ml of RMPI 1640 supplemented with 10% FBS. Then, the cells were treated with the peptide for the indicated period of time and stained with 2 mM JC-1 dye for 15 min. After being washed twice with PBS, the cells were analyzed by FACS to evaluate the change in mitochondrial transmembrane potential. JC-1 forms red fluorescent J-aggregates upon localization in healthy mitochondria, whereas the dispersed monomeric form of the dye in the cytoplasm fluoresces green [26]. Hence, the loss of mitochondrial transmembrane potential results in a decrease in the ratio of red to green fluorescence. To visualize the mitochondria under a fluorescence microscope, attached cells, at a density of 3610 4 cells/cm 2 , were treated with the peptide over a time course. Subsequently, JC-1 and DAPI were used to identify the mitochondria and nuclei, respectively.

Detection of Apoptosis and Necrosis
Phosphatidylserine exposed on the outer membrane reflects the early stages of apoptosis and can be detected by FITC-Annexin V and PI staining. The cells were detached and washed with medium. One-hundred and fifty thousand cells were suspended in 300 ml of RMPI 1640 supplemented with 10% FBS and treated with the peptide at the indicated concentration and time points. Then, the cells were subjected to dual staining with FITC-Annexin V and PI, according to the manual. The cells were observed under a fluorescence microscope and subjected to FACS analysis. Cells showing Annexin V2/PI-, Annexin V+/PI-, and Annexin V+/ PI+ signals are considered to be living, early stage apoptotic, and either late stage apoptotic or necrotic, respectively. To assess DNA damage during apoptosis, the TUNEL assay was performed in accordance with the manufacturer's instruction. Briefly, after treatment for 30 min-1 h with the peptide, cells were washed twice with PBS, fixed with 4% paraformaldehyde for 10-15 min at room temperature, permeabilized with 0.1% Triton X-100 for 2 min on ice, incubated in the reaction mixture and observed under a fluorescence microscope. FITC labeled nucleotides incorporated into the nuclei reflects DNA degradation. For investigating the activation of caspases, cells were preincubated with the Pan-caspase inhibitor, Z-VAD-FMK (100 mM), for 2 h before peptide treatment in the cytotoxicity assay. Immunofluorescence staining of caspase-3 was performed by using the Alexa Fluor 488-conjugated antibody against human cleaved caspase-3. The activation of caspase-3 was further verified by using a caspase-3 colorimetric assay kit (Genescript, Nanjing, China) with TNFrelated apoptosis-inducing ligand (TRAIL) as a positive control. The activity of caspase-3 was presented as fold of peptide-treated cells vs untreated cells.
Necrosis was evaluated with the Cytotox 96 non-radioactive cytotoxicity assay kit by detecting the release of lactate dehydrogenase (LDH) from necrotic cells. Ten-thousand cells were plated in a 96-well plate and treated with the peptide in 100 ml of RMPI 1640 supplemented with 2% FBS for 10 min-2 h, followed by LDH detection with 50 ml of culture supernatant from the treated cells, according to the manufacturer's instruction. The LDH level was presented as the percentage relative to that of cells treated with 0.9% Triton X-100.

Peptide-induced Hemolysis Analysis
Hemolysis was assessed according to a procedure described previously, with some modifications. Human erythrocytes from healthy donors were isolated by low-density (Ficoll-Paque PLUS, 1.077 g/mL, GE Healthcare) gradient centrifugation, washed three times with PBS and suspended in PBS at the concentration of 4% (volume ratio). In triplicate, erythrocyte suspensions were treated with the peptide in a 96-well plate at 37uC from 30 min to 16 h, and the absorbance was measured at 540 nm [27]. Hemolysis was presented as the percentage of the value relative to the absorbance of erythrocytes treated with 0.1% Triton X-100. In Vivo Tumor Xenograft Models All of the protocols used for in vivo experiments were approved by the University Animal Care and Use Committee. Four to sixweek-old male BALB/c nu/nu mice were purchased from the University Animal Production Center. Seven million DU145 cells suspended in 200 ml PBS were subcutaneously injected into the left flank region of mice. On day 7-9 post inoculation, the mice were randomized into three groups (n = 5) and intraperitoneally (i.p.) received 15 mg/kg B28Bn (6)(7)(8)(9)(10)(11)(12)(13)(14), B28 or an equivalent volume of PBS once daily for 7 consecutive days. Alternatively, the mice intratumorally received 5 mg/kg drug once daily for 5 consecutive days. The tumor volume was calculated as length6width 2 60.5. At the end of i.p. therapy, the animals were sacrificed and the major organs were excised, paraffin-embedded, sectioned, and stained with hematoxylin/eosin (H&E) for the examination of histopathologic architecture. To evaluate the cytotoxicity on tumor cells in vivo, a single dose of 50 mg of peptide in 50 ml of PBS was injected into the tumor graft (200-300 mm 3 ), according to the method described previously [21]. At 24 h post-injection, the animals were sacrificed, and the tumor tissues were routinely stained with H&E.

Statistical Analysis
Each experimental condition was performed in triplicate or repeated at least three times. The data are presented as the mean6SD. Student's t test was used to evaluate differences in the cell viability in caspase inhibition assays. One-way ANOVA was applied to compare differences in tumor growth in vivo. Significant differences were demonstrated to exist at P,0.05.

B28Bn(6-14) Induces Caspase-dependent Apoptosis in Tumor Cells
When the cytoplasm is intact, exposure of phosphatidylserine (PS) on the outer leaflet of plasma membrane indicates cellular apoptosis. Annexin V binds to PS with high affinity. Cellimpermeable PI can only enter the cells when the plasma membrane is compromised. Therefore, Annexin V+/PI-cells were considered to be in the early stages of apoptosis [30]. After treatment with 5 mM B28Bn(6-14) and dual staining with Annexin V and PI, numerous DU145 cell were Annexin V+/PI-when visualized under a fluorescence microscope ( Figure 4A). FACS analysis demonstrated that the percentage ratios of early apoptotic DU145 cells increased in a time-and peptide concentrationdependent manner. After treatment with 5 mM B28Bn(6-14) for 0, 1, 5, and 15 min, the percentages of early apoptotic DU145 cells were 3.63, 36.99, 32.94, and 35.36, respectively ( Figure 4B, upper panel). Similarly, after treatment with 0, 2.5, 5, and 10 mM B28Bn(6-14) for 10 min, the percentage ratios of early apoptotic DU145 cells were 3.80, 18.05, 39.19, and 43.18, respectively ( Figure 4B, lower panel).
The mitochondrial membrane of bacterial origin and is largely negatively charged [37]. The mammalian plasma membrane usually consists of neutral phospholipids and cholesterol, even though cancer cell membranes carry partial net negatively charged molecules. Therefore, the mammalian plasma membrane is relatively resistant to disruption by cationic antimicrobial peptides, including B28 and its chimeric peptides [23,36]. In addition, the invalidation of the apoptotic response in cancer has been reported in experimental and clinical studies due to a variety of strategies, particularly the inactivation of MMP to limit or circumvent apoptosis in tumor cells [20,38]. Strategies, including those used in this study, that disrupt the mitochondrial membrane may bypass this mechanism and overcome drug resistance [20,22,39].
Finally, we evaluated the anti-cancer efficacy of B28Bn(6-14) in vivo. Intratumoral and intraperitoneal administration of B28Bn (6)(7)(8)(9)(10)(11)(12)(13)(14) significantly suppressed the growth of prostate tumors in a mouse model without any obvious side effects ( Figure 5). On the other hand, the suppressive effect on tumor growth was decreased with intraperitoneal administration compared to intratumoral injection. One crucial reason for this might be the short half-life of the peptide, which is comprised of only the natural form of amino acids in vivo. This indicates that strategies such as PEGylation and the incorporation of unnatural amino acids into the peptide should be considered to prolong the half-life of B28Bn (6)(7)(8)(9)(10)(11)(12)(13)(14) in vivo [11,40]. Although many cancer biomarkers have been applied in developing anti-cancer drugs, less progress has been made in the development of clinical tools to predict the applicability of these targets in patients [41]. Bombesin receptors are not always overexpressed in individual tumor patients [8,12,33]. Further development of radiolabeled bombesin analogues that already exist for tumor imaging in humans may be helpful to select for patients who will be suitable for clinical trials and benefit from bombesin-directed therapeutic agents, such as B28Bn (6)(7)(8)(9)(10)(11)(12)(13)(14).