Purpose: Osteosarcoma represents the most common malignant primary bone tumor in childhood; however, the survival rate has remained unchanged for the past 20 years. To improve existing diagnosis and treatment methods and broaden the spectrum of imaging agents that can be used for early detection and assessment of tumor response to therapy, we performed a phage display–based screening for peptide sequences that bind specifically to osteosarcoma cells.
Experimental Design: From the Ph.D.-12 phage display peptide library composed of 2.7 × 109 different displayed peptides, one peptide was enriched after four rounds of in vitro selection in 143B osteosarcoma tumor cells with 293T human embryonic kidney cells as a control. Both the peptide and the phage clone displaying the peptide were conjugated with fluorescent dyes for in vitro cell and ex vivo tumor tissue stainings. The peptide was further labeled with 18F for positron emission tomography imaging studies. Cell uptake and efflux and ex vivo biodistribution were also done with 18F-labeled osteosarcoma specific peptide.
Results: ASGALSPSRLDT was the dominant sequence isolated from biopanning and named as OSP-1. OSP-1 shares a significant homology with heparinase II/III family protein, which binds and reacts with heparan sulfate proteoglycans. The fluorescence staining showed that FITC-OSP-1-phage or Cy5.5-OSP-1 had high binding with a panel of osteosarcoma cell lines, much less binding with UM-SCC1 human head and neck squamous cell carcinoma cells, and almost no binding with 293T cells, whereas the scrambled peptide OSP-S had virtually no binding to all the cell lines. 18F-OSP-1 had significantly higher accumulation in 143B tumor cells both in vitro and in vivo than 18F-OSP-S. 18F-OSP-1 also had higher uptake in 143B tumors than in UM-SCC-1 tumors.
Conclusions: Our data suggest that OSP-1 peptide is osteosarcoma specific, and the binding site of OSP-1 might be related to heparan sulfate proteoglycans. Appropriately labeled OSP-1 peptide has the potential to serve as a novel probe for osteosarcoma imaging. Clin Cancer Res; 16(16); 4268–77. ©2010 AACR.
Osteosarcoma is the most common nonhematologic primary malignant neoplasm of the bone, and the disease is mainly developed in young patients between 10 and 25 years old. Early detection and differentiation of osteosarcoma from osteoid osteoma, aneurysmal bone cyst, or infectious or inflammatory processes would be of great help for better control of this malignant disease. In this study, we performed in vitro peptidic phage screening and identified a new 12-mer peptide (OSP-1) with high binding affinity to osteosarcoma cells. 18F-labeled OSP-1 allowed successful noninvasive positron emission tomography imaging of osteosarcoma tumors in the athymic nude mouse model, indicating OSP-1 imaging as a promising strategy for early detection of osteosarcoma.
Osteosarcoma, the most common nonhematologic primary malignant neoplasm of the bone, is characterized by the development of bone or osteoid substance by the tumor cells (1). The disease is mainly developed in young patients between 10 and 25 years old and it is one of the most frequent causes of cancer-related deaths in childhood (2). Approximately 25% of osteosarcoma metastasize, typically to the lung. Despite effective surgical removal of the primary tumor and aggressive chemotherapy, the rate of long-term survival is only 15% to 20% because of pulmonary metastases, nonresponsiveness to therapy, or disease relapse (3). Clinically, patients with osteosarcoma are usually diagnosed by a palpable mass on physical examination and a characteristic radiographic lesion. Laboratory work can disclose elevations in alkaline phosphatase, lactate dehydrogenase, and erythrocyte sedimentation rate (4). However, early detection and differentiation of osteosarcoma from osteoid osteoma, aneurysmal bone cyst, or infectious or inflammatory processes would be of great help for better control of this malignant disease.
Compared with traditional methods, molecular imaging usually exploits specific molecular probes as well as intrinsic tissue characteristics as the source of image contrast and provides the potential for understanding the integrative biology, earlier detection, and characterization of disease and evaluation of treatment (5). Imaging probes with high affinity and specificity would be the key to successful molecular imaging. Phage display technology represents a high-throughput combinatorial technique for screening billions of random fusion peptide ligands against multiple targets on the surface or located within cancer cells and targets on tumor blood vessels (6). The distinctive advantage of this technique is that targets may be unknown and nonimmunogenic yet may serve as a delineating character for a particular cell type or tumor type (7). Meanwhile, these peptide ligands can be conjugated with imaging agents or therapeutic drugs and may be a promising tool for affinity-based targeted delivery of imaging agents and drugs (8). For example, RGD peptide has been screened with phage display to target integrin proteins (9). RGD peptide and its derivatives have been widely applied to image integrin expression and tumor angiogenesis after being labeled with various radioisotopes (10). Thus far, no specific probe has been developed to image osteosarcoma.
In this study, we performed in vitro peptidic phage screening using human osteosarcoma 143B cells as selecting target with the protocol (Supplementary Fig. S1). A new 12-mer peptide (OSP-1) was identified after four rounds of biopanning, with high binding affinity to 143B cells. The sequence of OSP-1 was found to be almost identical to the 17 to 28 residues of heparinase II/III. This enzyme binds and cleaves cell-surface heparan sulfate proteoglycans (HSPG), which are overexpressed on osteosarcoma cells. 18F-labeled OSP-1 allowed successful noninvasive positron emission tomography (PET) imaging of osteosarcoma tumors in the athymic nude mouse model, indicating OSP-1 imaging as a promising strategy for early detection of osteosarcoma. To our best knowledge, this is the first PET imaging of osteosarcoma with specific small molecular peptide probe.
Materials and Methods
Cell lines and animal models
The 143B, G292, MG-63, U-2 OS, and Saos-2 human osteosarcoma cell lines, UM-SCC1 human head and neck squamous carcinoma cell line, and 293T human embryonic kidney cell line were purchased from the American Type Culture Collection. 143B cells were grown in RPMI 1640 supplemented with 10% (v/v) fetal bovine serum (Invitrogen) and 0.015 mg/mL 5-bromo-2′-deoxyuridine at 37°C in an atmosphere containing 5% CO2. U-2 OS, MG-63, G292, and Saos-2 cells were grown in DMEM (Invitrogen) supplemented with 15% (v/v) fetal bovine serum at 37°C in an atmosphere containing 5% CO2. UM-SCC1 and 293T cells were grown in DMEM (Invitrogen) supplemented with 10% (v/v) fetal bovine serum at 37°C in an atmosphere containing 5% CO2. The 143B tumor model was generated by s.c. injection of 5 × 106 cells into the left front flank of female athymic nude mice (Harlan Laboratories). The UM-SCC1 tumor model was established by injection of 5 × 106 cells into the right front flank of the same mice 2 weeks before 143B cell inoculation. The mice were used for microPET studies when the tumor volume reached about 300 mm3 (about 1-2 weeks for 143B and about 3-4 weeks for UM-SCC1). All animal studies were conducted in accordance with the principles and procedures outlined in the Guide for the Care and Use of Laboratory Animals (11) and were approved by the Institutional Animal Care and Use Committee of Clinical Center, NIH.
Selection of tumor cell binding peptides
For biopanning, a linear 12-amino-acid peptide library (Ph.D.-12 phage display peptide library, New England Biolabs, Inc.) was used. Each selection round was conducted as follows: For negative selection, 1 × 1011 plaque-forming units were added to 293T cells. The supernatant was then transferred to 143B cells for positive selection. After 1 hour, the cells were washed five times with PBS plus 1% bovine serum albumin (BSA) to remove unbound phage particles. The cells and the bound phages were then incubated with E. coli host strain ER2738 to be amplified according to the manufacturer's protocol. After four rounds of screening, 20 random phage clones were selected for DNA sequencing. The amino acid sequences of displayed peptides were deduced from the DNA sequence. The dominant peptide sequence, ASGALSPSRLDT, was identified and named as OSP-1 for further experiments (Fig. 1).
Dye labeling of peptides and phage particles
The peptide OSP-1 and its scrambled peptide OSP-S (DLPSRTSALASG) were synthesized using a peptide synthesizer and purified with high-performance liquid chromatography (HPLC). For peptide labeling, OSP-1 (or OSP-S, 0.5 mg, 0.4 μmol) and diisopropylethyalamine (DIPEA; (7.5 μL) were added to a solution of Cy5.5 N-hydroxysuccinimide ester (0.5 mg, 0.4 μmol) in N,N-dimethylformamide (DMF; 150 μL); the reaction mixture was stirred at room temperature for 2 hours and quenched with 10 μL of acetic acid. The crude product was purified by reserve-phase HPLC on a semipreparative C-18 column. The desired fractions containing Cy5.5-OSP-1 (or Cy5.5-OSP-S) were collected and lyophilized to give a green fluffy powder (yield: 57%; >99% purity). The identity of the products was confirmed by TOF-MS ES+: Cy5.5-OSP-1, m/z 1,035.79 for [M+H/2] (C89H123N17O32S4, calculated [MW] 2,068.73), and Cy5.5-OSP-S, m/z 1,035.75 for [M+H/2] (C89H123N17O32S4, calculated [MW] 2,068.73), respectively.
OSP-1 phages (1 × 1012 plaque-forming units) were resuspended in 100 μL of a 0.3 mol/L NaHCO3 (pH 8.6) solution containing 0.25 mg/mL FITC. The phage/fluorochrome reaction was allowed to continue for 1 hour at room temperature in the dark. Subsequent to incubation, the volume of the labeled OSP-1-phage was brought up to 1 mL with DPBS, and the OSP-1-phage was then polyethylene glycol–precipitated twice and dialyzed extensively against 50 mmol/L Tris-HCl and 150 mmol/L NaCl (pH 7.5; TBS) to remove excess FITC (12).
Live 143B, G292, MG-63, U-2 OS, Saos-2, UM-SCC1, and 293T cells were blocked with 10% BSA for 60 minutes at 37°C and then stained with FITC-OSP-1-phage, Cy5.5-OSP-1, or Cy5.5-OSP-S (100 nmol/L) for 60 minutes at 37°C or at room temperature in the dark. After five washing steps, fixed cells were mounted with 4′,6-diamidino-2-phenylindole (DAPI)–containing mounting medium and observed with an epifluorescence microscope (Olympus, X81).
Frozen 143B and UM-SCC1 tumor tissue slices (8-10 μm) from the tumor-bearing nude mice were fixed with cold acetone for 20 minutes and dried in air for 30 minutes at room temperature. After blocking with 1% BSA for 30 minutes, the sections were incubated with Cy5.5-OSP-1 or Cy5.5-OSP-S (100 nmol/L) for 60 minutes at room temperature in the dark. After five washing steps, the slices were mounted with DAPI-containing mounting medium under an epifluorescence microscope (Olympus, X81). Each experiment was done in duplicate and repeated twice.
To determine the cell surface expression pattern of HSPG receptor, 143B, U-2 OS, MG-63, G292, Saos-2, UM-SCC1, and 293T cells were fixed with cold alcohol for 20 minutes. After blocking with 10% BSA for 30 minutes, fixed cells were incubated with mouse anti-heparin/heparan sulfate monoclonal antibody (1:300; Millipore) for 1 hour at room temperature and then visualized with Cy3-conjugated donkey anti-mouse secondary antibody (1:300; Jackson ImmunoResearch Laboratories). To confirm that OSP-1 peptide binds to HSPG receptor, fixed 143B cells were blocked with 10% BSA for 30 minutes and then incubated with OSP-1 (10 μmol/L) for 1 hour at room temperature, followed by mouse anti-heparin/hepararan sulfate monoclonal antibody (1:300; Millipore) for 1 hour at room temperature, and then visualized with Cy3-conjugated donkey anti-mouse secondary antibody (1:300; Jackson ImmunoResearch Laboratories). After five washing steps, fixed cells were mounted with DAPI-containing mounting medium and observed with an epifluorescence microscope (Olympus, X81).
To a solution of 2-fluoropropionic acid (92 μg, 1 μmol) in DMF (9.2 μL) was added a solution of O-(N-succinimidyl)-1,1,3,3-tetramethyluronium tetrafluoroborate (0.3 mg, 1 μmol) in DMF (30 μL) and DIPEA (10 μL). The reaction mixture was heated at 60°C for 20 minutes and was added to a solution of OSP-1 (or OSP-S, 1.2 mg, 1 μmol) in DMF (120 μL). The reaction mixture was heated for another 20 minutes at 60°C and quenched with 20 μL of acetic acid. The crude product was purified by reserve-phase HPLC on a semipreparative C-18 column. The desired fractions containing FP-OSP-1 (or FP-OSP-S) conjugate were collected and lyophilized to give a white fluffy powder (yield: >95%; purity >99%). The identity of the products was confirmed by TOF-MS ES+: FP-OSP-1, m/z 1,248.65 for [MH]+ (C51H87FN15O20, calculated [MW] 1,248.32), and FP-OSP-S, m/z 1,248.67 for [MH]+ (C51H87FN15O20, calculated [MW] 1,248.32), respectively.
The 18F labeling precursor 4-nitrophenyl 2-18F-fluoropropionate (18F-NFP) was synthesized as previously reported (13). OSP-1 and OSP-S labeling was as follows: OSP-1 (or OSP-S, 500 μg) dissolved in 150 μL of anhydrous DMSO was added to dried 18F-NFP in a 1-mL reaction vial, followed by addition of 20 μL of DIPEA. The reaction mixture was allowed to stand at room temperature for 30 minutes and quenched with 800 μL of 5% aqueous acetic acid solution. The labeled peptide was purified by reserve-phase HPLC on a semipreparative C-18 column. The desired fractions containing 18F-FP-OSP-1 (or 18F-FP-OSP-S) were collected and diluted with 20 mL of water. After trapping with a C-18 cartridge preactivated with 5 mL of ethanol and 10 mL of water, the product was washed with 2 mL of water and eluted with 2 mL of ethanol. The ethanol solution was blow-dried with a slow stream of N2 at 60°C. The 18F-labeled peptide was redissolved in PBS solution and passed through a 0.22-μm Millipore filter into a sterile multidose vial for in vitro and in vivo experiments. The labeling yield was 20% after the unlabeled peptide was efficiently separated from the product. The specific activity was estimated to be ∼37 TBq/mmol on the basis of the labeling agent 18F-NFP.
Cell uptake and efflux
The cell uptake studies were done as we have previously described with some modifications (14). Briefly, 143B, UM-SCC1, or 293T cells were seeded into 12-well plates at a density of 5 × 105 per well and incubated (∼37 kBq/well) with 18F-labeled tracers at 37°C for 15, 30, 60, and 120 minutes. The tumor cells were then washed three times with chilled PBS and harvested by trypsinization with 0.25% trypsin/0.02% EDTA (Invitrogen). Cell suspensions were collected and measured in a gamma-counter (Packard). Cell uptake was expressed as percentage of decay-corrected total input radioactivity. Experiments were done twice with triplicate wells. For efflux studies, 18F-labeled tracers (∼37 kBq/well) were first incubated with 143B, UM-SCC1, or 293T cells in 12-well plates for 2 hours at 37°C to allow internalization. Then, cells were washed twice with PBS and incubated with cell culture medium for 15, 30, 60, and 120 minutes. After washing three times with PBS, cells were harvested by trypsinization with 0.25% trypsin/0.02% EDTA. The cell suspensions were collected and measured in a gamma-counter. Experiments were done twice with triplicate wells. Data are expressed as percent added dose after decay correction.
PET scans and image analysis were done using an Inveon microPET scanner (Siemens Medical Solutions). Each 143B and UM-SCC1 tumor–bearing mouse was injected in a tail vein with 3.7 MBq (100 μCi, 0.1 nmol in 100 μL) of 18F-OSP-1 or 18F-OSP-S under isoflurane anesthesia (n = 6 per group). For static PET, 5-minute scans were acquired at 30 minutes, 1 hour, and 2 hours after injection. The images were reconstructed using a two-dimensional ordered subsets expectation maximum algorithm and no correction was applied for attenuation or scatter. For each microPET scan, regions of interest (ROI) were drawn over the tumor, normal tissue, and major organs using the vendor software ASI Pro 126.96.36.199 on decay-corrected whole-body coronal images. The maximum radioactivity concentrations (accumulation) within a tumor or an organ were obtained from mean pixel values within the multiple ROI volume and then were converted to megabecquerels per milliliter using a conversion factor. These values were then divided by the administered activity to obtain (assuming a tissue density of 1 g/mL) an image ROI-derived percent injected dose per gram (%ID/g), according to the following formulas: Cf = MBq/mL/MV; %ID/g = 100 × MV/Cf/total activity in MBq (Cf, conversion factor; MBq, megabecquerel; MV, mean value of pixels in ROI).
Ex vivo biodistribution
Female athymic nude mice bearing 143B and UM-SCC1 xenografts were injected with 0.925 MBq (25 μCi, 25 pmol in 100 μL) of 18F-OSP-1 or 18F-OSP-S to evaluate the distribution of the tracers in the tumor tissues and major organs. At 2 and 4 hours after injection of the tracer, the tumor-bearing mice were sacrificed and dissected. Blood, tumor, major organs, and tissues were collected and wet-weighed. The radioactivity in the wet whole tissue was measured with a gamma-counter (Packard). The results are presented as percentage injected dose per gram of tissue (%ID/g). For each mouse, the radioactivity of the tissue samples was calibrated against a known aliquot of the injectate and normalized to a body mass of 20 g. Values were expressed as mean ± SD for groups of four animals (n = 4 per group).
Quantitative data were expressed as means ± SD. Means were compared using one-way ANOVA and Student's t test. P < 0.05 was considered statistically significant.
Identification of the 143B-binding peptide OSP-1
Four selection rounds were performed on the 143B cell line with a linear 12-amino-acid peptide library to allow for enrichment of tumor cell binding or internalizing phages. Before each selection round, a negative selection with the 293T human embryonic kidney cell line was done to subtract phages that bound to nontumor cells. After the final selection round, the ssDNAs of 20 clones were sequenced and analyzed. One sequence was found to be enriched in 65% of all sequenced clones and termed OSP-1. OSP-1 is a 12-amino-acid long linear peptide comprising the sequence ASGALSPSRLDT. To provide a control peptide, we reshuffled the amino acid sequence of OSP-1 and synthesized a scramble peptide, DLPSRTSALASG, named as OSP-S.
To determine the osteosarcoma cell binding ability of OSP-1 in vitro, live 143B, G292, MG-63, U-2 OS, Saos-2, UM-SCC1, and 293T cells were stained with FITC-OSP-1-phage, Cy5.5-OSP-1, or Cy5.5-OSP-S (100 nmol/L). As shown in Fig. 2, FITC-OSP-1-phage showed abundant binding to 143B cells, as indicated by strong fluorescent signal, and almost no binding to 293T cells (Fig. 2A). Similarly, C5.5-OSP-1 peptide showed strong binding to all five osteosarcoma cell lines tested (143B, G292, MG-63, U-2 OS, and Saos-2), much less binding to UM-SCC1 cells, and no binding to 293T cells. Moreover, the scramble peptide Cy5.5-OSP-S exhibited little binding to all the cell lines (Supplementary Fig. S2). Ex vivo staining with Cy5.5-OSP-1 showed that the fluorescent signal distributed strongly and diffusively among the whole 143B tumor sections. We did not observe any preferred localization region. No apparent staining signals were observed in UM-SCC1 tumor sections and no binding of Cy5.5-OSP-S to 143B and UM-SCC1 tumors was observed (Fig. 2B). These in vitro and ex vivo staining data support the high affinity and specificity of the identified OSP-1 peptide to 143B osteosarcoma.
As the OSP-1 peptide (ASGALSPSRLDT) shares a significant homology with the 17 to 28 amino acid residues of heparinase II/III family protein (AGGTVTPARLDT), which binds and reacts with HSPG, we first checked the expression pattern of HSPG receptor in five osteosarcoma cell lines (143B, G292, MG-63, U-2 OS, and Saos-2) and compared it with those in the UM-SCC1 head and neck cancer cell line and the 293T cell line by heparin/heparan sulfate antibody staining. It was found that proteoglycans are abundantly expressed by all the tested osteosarcoma cell lines but not by UM-SCC1 and 293T cells (Fig. 3A; Supplementary Fig. S3). We then checked the HSPG receptor specificity of OSP-1 peptide. As shown in Fig. 3B, preincubation with OSP-1 peptide abrogated the binding of heparin/heparan sulfate antibody to 143B osteosarcoma cells.
Cell uptake and efflux of OSP-1
We further developed OSP-1 as a PET imaging tracer by labeling it with a positron emitting radioisotope, 18F (t1/2 = 108 minutes). The labeling and structure of 18F-OSP-1 and 18F-OSP-S are shown in Fig. 1. Both 18F-FP-OSP-1 and 18F-FP-OSP-S were characterized in vitro by cell uptake and efflux assay in 143B, UM-SCC1, and 293T cells. 18F-FP-OSP-1 had significantly higher cell uptake in 143B cells than in UM-SCC1 and 293T cells (Fig. 4A). UM-SCC1 was between those of 143B and 293T with moderate cell uptake of 18F-FP-OSP-1. 18F-FP-OSP-1 uptake in 143B cells reached the maximum of 0.52% of total input radioactivity after 120 minutes. In contrast with 18F-FP-OSP-1, 18F-FP-OSP-S showed only 0.02% of total input radioactivity in 143B. Although decreasing along with time, the cell retention of 18F-FP-OSP-1 in the three cell lines also showed the order of 143B > UM-SCC1 > 293T. 18F-FP-OSP-1 in UM-SCC1 and 293T cells was undetectable after 60 minutes. 18F-FP-OSP-S in 143B cells was undetectable after 15 minutes (Fig. 4B).
Next, we performed in vivo microPET imaging with 18F-FP-OSP-1 and 18F-FP-OSP-S as imaging agents. Representative coronal microPET images of 143B and UM-SCC1 tumor–bearing mice (n = 6 per group) at different times after i.v. injection of 3.7 MBq (100 μCi) of 18F-FP-OSP-1 or 18F-FP-OSP-S are shown in Fig. 5A. The 143B tumors were clearly visible with high contrast in relation to the contralateral UM-SCC1 tumors at all time points measured from 30 to 120 minutes after injection of 18F-FP-OSP-1. Prominent uptake of 18F-FP-OSP-1 was also observed in the kidneys at early time points, suggesting that this tracer is mainly excreted through the renal-urinary route. Tumor and major organ activity accumulation in the microPET scans was quantified by measuring the ROIs that encompassed the entire organ on the coronal images. The 143B tumor uptake of 18F-FP-OSP-1 was significantly higher than that by the UM-SCC1 tumor at all time points examined (P < 0.01; Supplementary Fig. S4A). At 120 minutes after injection, the uptake in 143B tumors was 1.43 ± 0.14% ID/g versus 0.75 ± 0.12% ID/g in UM-SCC1 tumors. The in vivo 143B tumor binding specificity of 18F-FP-OSP-1 was also confirmed by comparison with 18F-FP-OSP-S studies. 18F-FP-OSP-S showed significantly lower uptake than 18F-FP-OSP-1 in 143B tumor at all time points (P < 0.01). At 30, 60, and 120 minutes after injection, a 5.9-, 3.6-, and 3.9-fold uptake of 18F-FP-OSP-1 compared with 18F-FP-OSP-S within 143B tumor was observed (Supplementary Fig. S4B). The tumor-to-nontumor (T/NT) ratios of 18F-FP-OSP-1 and 18F-FP-OSP-S at 120 minutes were calculated and are compared in Fig. 5B and C. The tumor/kidney, tumor/liver, and tumor/muscle ratios of 18F-FP-OSP-1 in 143B were significantly higher than those in UM-SCC1 (P < 0.05). By contrast, the T/NT ratios of 18F-FP-OSP-S in 143B and UM-SCC1 tumors were not significantly different from each other. In addition, the tumor/muscle ratio of 18F-FP-OSP-1 in 143B tumors was significantly higher than that of 18F-FP-OSP-S (P < 0.05).
To validate the microPET quantification, we also performed biodistribution studies. Female nude mice bearing 143B and UM-SCC1 xenografts (n = 4 per group) were injected i.v. with 0.925 MBq (25 μCi) of 18F-FP-OSP-1 and then sacrificed at 2 or 4 hours after injection of the tracer. The data are expressed as the percentage injected dose per gram of tissue (%ID/g) in Fig. 6. 18F-FP-OSP-1 uptake in 143B tumors was significantly higher than that in UM-SCC1 tumors (P < 0.05), which is consistent with the microPET data. To further determine the in vivo binding specificity of 18F-OSP-1, we also injected 18F-OSP-S into 143B and UM-SCC1 tumor–bearing mice (n = 4 per group) at a dose of 0.925 MBq (25 μCi). In 143B tumor, a significantly lower uptake of 18F-OSP-S was seen compared with uptake of 18F-OSP-1 (Supplementary Fig. S5A; 1.106 ± 0.096 ID/g versus 2.506 ± 0.002% ID/g, P < 0.01) at the 120-minute time point, indicating the tumor targeting specificity of 18F-OSP-1 in the 143B tumor model. The T/NT ratios of 18F-FP-OSP-1 and 18F-FP-OSP-S at 120 minutes were calculated from the biodistribution data and are compared in Supplementary Fig. S5B. Consistent with the quantification from PET imaging, the T/NT ratios of 18F-FP-OSP-1 in 143B tumors were also significantly higher than those in UM-SCC1 tumors (P < 0.05). The tumor muscle ratio of both 18F-FP-OSP-1 and 18F-FP-OSP-S quantified from ex vivo biodistribution data (3.49 ± 0.35 and 1.74 ± 0.05) is comparable with that from PET imaging (3.77 ± 0.21 and 1.54 ± 0.18).
In view of the high level of malignancy and metastatic ability of osteosarcoma, early detection with high sensitivity and specificity would greatly help tumor control. Phage display is a successful tool for identifying novel peptides with high specificity to tumor cells or tumor blood vessels (6, 15). In this study, by applying a 12-mer peptide library to human osteosarcoma 143B cell lines, we were able to identify the binding peptide OSP-1. This peptide was found to be enriched in 65% of all sequenced clones. OSP-1 showed specific binding to osteosarcoma cell lines in vitro. The specificity was supported by significantly lower binding of OSP-1 to nonosteosarcoma cancer cells and normal cells and much lower cell binding in vitro of the scramble peptide OSP-S.
The in vivo behavior of 18F-FP-OSP-1 was tested in animal experiments using microPET. 18F-FP-OSP-1 showed enhanced tumor uptake versus 18F-FP-OSP-S in 143B xenograft animal models. The tumor-to-muscle ratio for 18F-FP-OSP-1 was 3.8 compared with a ratio of 1.5 for 18F-FP-OSP-S at 120 minutes postinjection, indicating that OSP-1 is a promising molecule for specific imaging of osteosarcoma. The specificity for osteosarcoma of OSP-1 was further shown with animals bearing UM-SCC1 human head and neck squamous carcinoma tumors. The much lower tumor-to-muscle ratio of 18F-FP-OSP-1 in UM-SCC1 tumors indicated that OSP-1 has the potential to differentiate osteosarcoma and other malignant tumors. The biodistribution data showed that the tracer uptake of 18F-FP-OSP-1 in 143B tumors at 120 minutes was significantly higher compared with that in UM-SCC1 tumors and with the uptake of 18F-FP-OSP-S in 143B tumors. Although the tumor uptake value acquired by ex vivo biodistribution assay is higher than that quantified based on PET scan, the tumor/non tumor ratios are consistent, which shows the feasibility to detect osteosarcoma with PET imaging using 18F-FP-OSP-1 as a specific tracer.
We noticed the different pharmacokinetics of 18F-FP-OSP-1 and the scramble peptide 18F-FP-OSP-S, with OSP-1 excreted mainly through the renal-urinary route and the scramble peptide through the biliary route. The exact reason for this is not clear because the amino acid components are identical with these two peptides and both peptides are metabolically stable. We speculated that OSP-1 and OSP-S form different conformations due to varying order of the same amino residues, resulting in different hydrophobicities. The exact binding target is usually unknown when using whole cells as the selecting object during phage display process. A protein database search for OSP-1 revealed that this peptide shares a significant homology with heparinase II/III family protein, which binds and reacts with HSPGs (16, 17). Because proteoglycans are abundantly expressed by osteocytes and osteogenic tumor cells, which rely on proteoglycans in cell attachment processes and mineralization (18), it is highly reasonable to surmise that the binding targets of OSP-1 on the surface of osteosarcoma cells are HSPGs. Our data confirmed that osteosarcoma cells overexpress HSPGs and that anti-heparin/heparan sulfate antibody binding to osteosarcoma cells can be blocked by excess amount of OSP-1 peptide.
In conclusion, the novel linear 12-mer peptide OSP-1 showed significant ability to bind to osteosarcoma. Due to its high and specific binding to osteosarcoma cells in vitro and in vivo, OSP-1 can be used for coupling with radioactive isotopes and fluorescence and anticancer agents and has potential to be used for the diagnosis and treatment of osteosarcoma.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
We thank Dr. Henry S. Eden for proofreading the manuscript.
Grant Support: Intramural Research Program of the National Institute of Biomedical Imaging and Bioengineering, NIH.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.