556
Bioconjugate Chem. 2010, 21, 556–562
A Phospholipid-PEG2000 Conjugate of a Vascular Endothelial Growth Factor Receptor 2 (VEGFR2)-Targeting Heterodimer Peptide for Contrast-Enhanced Ultrasound Imaging of Angiogenesis R. Pillai,† E. R. Marinelli,† H. Fan,† P. Nanjappan,† B. Song,† M. A. von Wronski,† S. Cherkaoui,‡ I. Tardy,‡ S. Pochon,‡ M. Schneider,‡ A. D. Nunn,† and R. E. Swenson*,† The Ernst Felder Laboratories, Bracco Research USA, 305 College Road East, Princeton, New Jersey 08540, and Bracco Research SA, Route de la Galaise, 31, CH-1228 Plan-les-Ouates, Geneva, Switzerland. Received December 21, 2009; Revised Manuscript Received January 26, 2010
The transition of a targeted ultrasound contrast agent from animal imaging to testing in clinical studies requires considerable chemical development. The nature of the construct changes from an agent that is chemically attached to microbubbles to one where the targeting group is coupled to a phospholipid, for direct incorporation to the bubble surface. We provide an efficient method to attach a heterodimeric peptide to a pegylated phospholipid and show that the resulting construct retains nanomolar affinity for its target, vascular endothelial growth factor receptor 2 (VEGFR2), for both the human (kinase insert domain-containing receptor - KDR) and the mouse (fetal liver kinase 1 - Flk-1) receptors. The purified phospholipid-PEG-peptide isolated from TFA-based eluents is not stable with respect to hydrolysis of the fatty ester moieties. This leads to the time-dependent formation of the lysophospholipid and the phosphoglycerylamide derived from the degradation of the product. Purification of the product using neutral eluent systems provides a stable product. Methods to prepare the lysophospholipid (hydrolysis product) are also included. Biacore binding data demonstrated the retention of binding of the lipopeptide to the KDR receptor. The phospholipid-PEG2000-peptide is smoothly incorporated into gas-filled microbubbles and provides imaging of angiogenesis in a rat tumor model.
INTRODUCTION Molecular imaging is unprecedented in its ability to noninvasively provide in vivo information for diagnosis, prognosis, and monitoring of disease. Of the common modalities, ultrasound imaging is noteworthy for its cost effectiveness, portability, and ease of use. The injection of small (2-5 µm diameter) echogenic gas-filled unilamellar vesicles, commonly referred to as microbubbles, allows improved image definition due to the administered microbubble contrast agent. Recently, the use of targeting groups on the surface of these microbubbles has allowed the monitoring of biochemical processes in the body in real time with sensitivity comparable to that of nuclear medicine (1). One of the important biological processes for imaging is angiogenesis. Tumor growth requires the development of new blood vessels, which is correlated with aggressiveness (2). Microbubbles targeted to Rvβ3 (3) or via the targeting tripeptide RRL (4) have been used to image angiogenesis. However, the receptor tyrosine kinase VEGFR2, or KDR, has been shown to be one of the major mediators of angiogenesis (5). Ultrasound bubbles bearing antibodies to VEGFR2 have also shown good imaging of angiogenesis (6). For this reason, we focused on developing a peptide-based targeting agent for imaging the KDR receptor with ultrasound bubbles. We previously described the use of phage display to generate two families of KDR-binding peptides. Each family of peptides * Author to who correspondence should be addressed. Dr. Rolf E. Swenson, Discovery Chemistry,The Ernst Felder Laboratories, Bracco Research USA, 305 College Road East, Princeton, NJ 08540; Tel. 609514-2526; Fax: 609-514-2446; e-mail:
[email protected]. † Bracco Research USA. ‡ Bracco Research SA.
was found to bind to a distinct, noncompeting epitope on KDR as evidenced by the finding that members of each family could not displace the KDR-binding of members of the other family. This was exploited by preparation of heterodimer peptides containing one member of each family of peptides. In binding and functional assays, these constructs were superior to individual peptides and homodimer constructs based on each family of peptides. Chemical optimization of a peptide selected from each family and their linkage provided a potent heterodimer peptide ligand for KDR (7). A versatile chemical approach for preparation of a range of heterodimeric peptides using disuccinimidyl glutarate (DSG) was also developed (8). Targeted microbubbles for research applications are routinely prepared by the conjugation of biotinylated antibodies, peptides, or small molecules to preformed streptavidin-bound biotinylated bubbles (9). Commercial availability of streptavidin-coated bubbles (Target-Ready MicroMarker bubbles, VisualSonics, Toronto, Ontario, Canada) has facilitated adoption of this technique for research studies. The use of streptavidin-coated microbubbles bearing biotinylated targeting vectors in humans is undesirable due to the antigenicity of streptavidin (10). Hence, for clinical product applications, covalent attachment of the targeting group to the phospholipid of the bubble surface is optimal. A number of approaches for preparation of targeting microbubbles in this manner have been offered in the literature, including attachment of thiolated targeting groups to maleimide or bromoacetylated bubbles (11), carbamate formation by reaction of amines of targeting groups with mixed alkoxynitrophenyl carbonates (12), amide coupling reactions (13), chemical ligation approaches (14), and preparation of artificial lipids (15). Success with these strategies has been limited by inconsistent yields, undesired side reactions, deactivation of coupling partners, or concerns about toxicity of unnatural
10.1021/bc9005688 2010 American Chemical Society Published on Web 02/19/2010
Phospholipid Peptide for Ultrasound Imaging
lipid conjugates. Therefore, the preparation of targeted bubbles still provides challenges. Additional complexity results from the need to incorporate an appropriate spacer group to separate the targeting group from the lipid surface of the bubble. The most commonly used spacers are PEGs (poly(ethylene glycols)) of different chain lengths. Other issues include the incompatibility of the phospholipid ester groups with the conditions of peptide synthesis, particularly the TFA-mediated cleavage/deprotection step. This paper provides a method for the preparation of a viable product for targeted ultrasound clinical imaging.
EXPERIMENTAL PROCEDURES Materials. All common reagents were commercially available and used as received. Disuccinimidyl glutarate (DSG) and 1,2distearoyl-sn-glycero-3-phospho-ethanolamine-N-[amino (poly(ethylene glycol))2000] ammonium salt, [DSPE-PEG2000-NH2] were obtained from Pierce Chemical Co. (Rockford, IL) and Avanti Polar Lipids (Alabaster, AL), respectively. Fmoc-GlyGly-Gly-OH was prepared in-house from triglycine by reaction with N-hydroxysuccinimidyl-9-fluorenylmethyl carbonate (FmocOSu). AG MP-50 ion-exchange resin was obtained from BioRad (Hercules, CA). Analytical HPLC. For routine work, analytical HPLC data were obtained using a Shimadzu LC-10AT VP dual-pump gradient system employing a Waters XTerra MS-C18 50 mm × 4.6 mm column, (5 µm particle, 120 Å pore) and gradient or isocratic elution systems using H2O (0.1% TFA) as eluent A and CH3CN (0.1% TFA) or CH3CN-CH3OH 1:1 (v/v) (0.1% TFA) as eluent B. Detection of compounds was accomplished using a UV detector at 220 and 254 nm. Analytical and purity studies of the phospholipid-PEG2000-peptide derivatives were conducted employing a YMC C4 (250 mm × 4.6 mm, 5 µm particle, 300 Å pore) column or on a Zorbax 300 SB-C3 (150 mm × 3 mm, 3.5 µm particle, 300 Å pore) column and gradient elution using H2O (10 mM NH4OAc) as eluent A and CH3CN-H2O 9:1 (v/v) (10 mM NH4OAc) as eluent B. Detection of compounds was accomplished with a SEDEX 55 evaporative light scattering detector (ELSD) and a UV detector. Preparative HPLC. Preparative HPLC was conducted on a Shimadzu LC-8A dual-pump gradient system equipped with a SPD-10AV UV detector fitted with a preparative flow cell. The solution containing the crude peptide was loaded onto a reversed-phase C18, C4, or C3 column, depending on the compound characteristics, using a third pump attached to the preparative Shimadzu LC-8A dual-pump gradient system. After the solution of the crude product mixture was applied to the preparative HPLC column, the reaction solvents and solvents employed as diluents, such as DMF or DMSO, were eluted from the column at low percentage organic phase composition. Then, the desired product was eluted using a gradient elution of eluent B into eluent A. Product-containing fractions were combined on the basis of their purity, as determined by analytical HPLC and mass spectral analysis. The combined fractions were lyophilized to provide the desired product. Mass Spectroscopy. Mass spectral data were obtained from MScan Inc. (West Chester, PA) or in-house on an Agilent LCMSD 1100 mass spectrometer. Indirect CE-UV for the Simultaneous Analysis of Trifluoroacetate and Acetate Ions. The simultaneous determination of trifluoroacetate and acetate was carried out by capillary electrophoresis (CE) with indirect UV detection at 232 nm. CE data were generated using a PA800 Beckman Coulter system (Fullerton, CA, USA) equipped with an on-column diode-array detector, an autosampler, and a power supply able to deliver up to 30 kV. The separations were performed on a 75 µm (i.d.) × 60 cm (total length)/50 cm (length to the detector) fused silica
Bioconjugate Chem., Vol. 21, No. 3, 2010 557
capillary. A Beckman 32 Karat software package was used for instrument control, data acquisition, and data handling. CeofixAnions5 kit (Analis, Namur, Belgium), including a pyridine-dicarboxylic acid buffer at pH 5.6, was used for the quantification of trifluoroacetate and acetate. First, the capillary was rinsed with the initiator buffer, which contains a polycation, to form a positive dynamic coating and invert the electroosmotic flow. Second, the capillary was filled with the accelerator background electrolyte containing a chromophore which absorbs at 232 nm. For quantitative purposes, propionic acid was used as an internal standard. After sample injection by pressure, separation was carried out by applying a high voltage. All experiments were carried out in anionic mode (cathode at the inlet and anode at the outlet). The capillary was thermostatted at 25 °C and a constant voltage of 30 kV (-50 µA) for 6 min, with an initial ramping of 1 min, was applied during analysis. Samples were kept at ambient temperature in the autosampler and injection (∼20 nL injection volume) was achieved by lowpressure mode for 15 s at 0.5 psi. Both standard and test solutions were prepared in DMSO-50 mM aqueous NaOH 1:4 (v/v). Before its first use, the capillary was flushed with 0.1 M NaOH for 1 min, paused for 4 min, then rinsed 0.5 min with 0.1 M NaOH followed by H2O for 1 min. This initialization procedure was also employed after prolonged storage. Between analyses, the capillary was flushed with 0.1 M NaOH and H2O (20 psi, 0.5 min each). Before each run, the capillary was pressure-rinsed and filled with the initiator and the accelerator solutions. To achieve reproducible and accurate results, the CEofix buffer solutions were replaced every 20 runs and the first runs of a day were systematically discarded. When not in use, the capillary was washed with 0.1 M NaOH, then H2O, and stored dry. Preparation of Heterodimer Peptide Conjugate 1 Having Low TFA Levels. Heterodimer peptide 1 (Scheme 1) was synthesized as a TFA salt following procedures described previously (8). Compound 1 as the TFA salt, [120 mg in 80 mL of CH3CN-H2O 3:7 (v/v)] was loaded onto an AG MP50 column and the column was washed with the same eluent until the conductivity was stable at 1 µS/cm. A step gradient elution of the column with NH4OAc in CH3CN-H2O 3:7 (v/v) was conducted using 100 mL aliquots of 200 mM and 400 mM salt followed by 200 mL aliquots of 600 mM and 800 mM salt. The compound eluted at ca. 600 mM salt. The pure product-containing fractions were pooled and lyophilized. The resulting isolate was dissolved in CH3CN-H2O 3:7 (v/v) and lyophilized several times to constant weight. This gave 104 mg (86.7% yield) of the pure material 1 as the acetate salt. HPLC: tR ) 5.2 min; assay >99% (area %); column, Waters XTerra MS-C18, 50 mm × 4.6 mm, 5 µm particle, 120 Å pore; eluents: A, H2O (0.1% TFA); B, CH3CN (0.1%TFA); elution, initial condition 20% B, linear gradient 20-60% B over 8 min; flow rate, 3 mL/min; detection, UV at 220 nm; MS, API-ES neg. ion; m/z ) 1816.3 [M-3H]/3, 1362.0 [M-4H]/4, 1089.2 [M-5H]/5; CE analysis (counterion % wt/wt) TFA 0.2%; acetate 0.15%. Preparation and Purification of the Phospholipid-PEG2000Peptide Conjugate 3 as Its Acetate Salt from Heterodimer Peptide 1 Acetate Salt (Scheme 1). A solution of the heterodimer peptide (1, acetate salt) (0.108 g, 0.02 mmol) in anhydrous DMF (1.0 mL) was added to a solution of disuccinimidyl glutarate (DSG) (50.0 mg, 0.15 mmol) in anhydrous DMF (0.5 mL) containing diisopropylethylamine (DIEA) (0.07 mL, 0.4 mmol, 20 equiv) with vigorous stirring over a period of 2 min and the reaction mixture was stirred for 30 min. HPLC analysis on a Waters XTerra C-18 column and MS showed the disappearance of the starting material
558 Bioconjugate Chem., Vol. 21, No. 3, 2010
Pillai et al.
Scheme 1. Preparation of DSPE-PEG2000-Heterodimer Peptide Conjugate 3
and formation of a new product. The solution was concentrated at reduced pressure. The viscous material was stirred with 10 mL of ethyl acetate for 10 min. The precipitated solid, the glutaric acid monoamide mono-NHS ester 2 of the heterodimer peptide 1, was centrifuged to a pellet. The ethyl acetate was decanted, and this process was repeated three more times and the solid was finally washed with hexane (10 mL) and dried under vacuum (20 mmHg) for 20 min. A solution of DSPE-PEG2000-NH2-ammonium salt (0.055 g, 0.02 mmol, 1 equiv) in anhydrous DMF (0.5 mL) was added to a solution of heterodimer intermediate 2 in anhydrous DMF (1.0 mL) containing DIEA (0.07 mL, 0.4 mmol, 20 equiv), and the mixture was stirred at ambient temperature. The reaction was monitored by analytical HPLC and was complete in 24 h. The reaction mixture was then diluted with H2O (15 mL), and the resulting solution was loaded onto a reversed-phase C4 preparative column (Kromasil Prep C4, 250 × 20 mm, 10 µm particle, 300 Å pore), which had been pre-equilibrated with a buffer containing 20% of CH3CN-H2O 9:1 (v/v) containing 10 mM ammonium acetate (NH4OAc) (eluent B) in 10 mM aqueous NH4OAc
(eluent A) at a flow rate of 25 mL/min. During the application of the sample solution to the column, the flow of the equilibrating eluent from the preparative HPLC system was stopped. After the sample solution was loaded, the flow of the equilibrating eluent was reinitiated and the column was eluted with the equilibrating buffer until the DMF used in the reaction was removed. The column was then eluted with a gradient of 30-100% B over a period of 40 min. Fractions (15 mL) were collected using UV (220 nm) as an indicator of product elution. Fractions were checked for purity on an analytical HPLC system (YMC C3, 150 mm × 4.6 mm, 5 µm particle, 300 Å pore) using UV at 220 nm and an ELS detector (temp 50 °C, pressure 3.2 psi, sensitivity 10). The latter detector was employed to detect any trace amount of unreacted DSPEPEG2000-NH2 which is nearly UV silent at 220 nm. The fractions containing the pure product were pooled and lyophilized. The product was redissolved in H2O-CH3CN 4:1 (v/v) and freeze-dried to afford 94 mg (62% yield) of the required compound 3. HPLC: tR 9.32 min; assay >99% (area %); column, YMC C4, 50 mm × 4.6 mm, 5 µm particle, 300 Å pore; eluents:
Phospholipid Peptide for Ultrasound Imaging
Bioconjugate Chem., Vol. 21, No. 3, 2010 559
Scheme 2. Hydrolysis of Compound 3 to Compounds 4 and 5
A, H2O (0.1% TFA); B, CH3CN-MeOH 1/1 (v/v) (0.1% TFA); elution, initial condition 70% B, linear gradient 70-100% B over 20 min; flow rate, 3.0 mL/min; detection, UV at 220 nm and ELS (temp 50 °C, pressure 3.2 psi, sensitivity 10); CE analysis (counterion % wt/wt): % wt TFA, 0.3%; % wt acetate, 0.4%. MALDI-TOF MS: m/z (%) ) 8367.7 (100), [M+NH4++CH3CN]+calc ) 8367.6. Preparation of Lysophospholipid 4 and Phosphoglyceryl Ester 5 from Phospholipid-PEG2000-Peptide Conjugate 3 by Acid Hydrolysis (Scheme 2). TFA (300 µL) was added to a solution of compound 3 (50 mg) in 30 mL of CH3CN-H2O 1:2 (v/v), and the mixture was stirred at room temperature for 7 days. The mixture was purified by preparative HPLC on a YMC C4 (250 mm × 10 mm) column using 10 mm NH4OAc in H2O as eluent A and 9:1 CH3CN-H2O containing 10 mM NH4OAc as eluent B. A gradient 10-100% B over a period of 40 min at a flow rate of 10.0 mL/min with UV detection at 220 nm was employed. Fractions were analyzed by HPLC, and those containing pure compound were combined and lyophilized to afford pure products. Compounds 4 (9.8 mg) and 5 (3.2 mg) were isolated. HPLC: tR Compound 4, 11.3 min; tR Compound 5, 4.09 min; column, Agilent C3, 150 mm × 3 mm, 3.5 µm particle, 300 Å pore; eluents: A, H2O (10 mM NH4OAc); B, CH3CN-H2O 9:1 (v/v) (10 mM NH4OAc); elution, linear gradient, 50-100% B over 30 min; flow rate, 1.0 mL/min; detection, ELSD, temp 50 °C, pressure 3.2 psi, sensitivity 10. Under these conditions, tR of compound 3 was 19.0 min. MALDI-TOF MS: Compound 4, m/z (%) ) 8069.0 (100),
[M+Na]+calc ) 8065.1; Compound 5, m/z (%) ) 7801.5 (100), [M+Na]+calc ) 7798.6. Preparation of Lysophospholipid 4 by Enzymatic Hydrolysis of Phospholipid-PEG2000-Peptide Conjugate 3 with Phospholipase-A2. A solution of compound 3 (40 mg) in 20 mL of CH3CN-1.0 M aqueous HEPES buffer (pH 8.5) 1:2 (v/v) was stirred with a solution of Phospholipase-A2 (16) from honey bee venom (Sigma, 40 units in 25 µL HEPES buffer) at room temperature for 1 h. HPLC analysis indicated the complete conversion of the starting material to the lysophospholipid compound 4. The crude reaction mixture was treated with 20 mg of hydroxylamine in 2.0 mL of ethanol for 30 min. The resulting solution was purified by preparative HPLC as described above. Fractions were analyzed by HPLC, and those containing the pure product were combined and lyophilized to afford 28 mg (76% yield) of compound 4, which, in HPLC analysis, coeluted with the material obtained by acid hydrolysis. MALDITOF MS: Compound 4, m/z (%) ) 8062.0 (100), [M+Na]+calc ) 8065.1. Surface Plasmon Resonance (SPR) Measurements of Binding of Compounds 1 and 3 to Human and Mouse VEGFR2. SPR measurements of the binding kinetics of compound 1 and compound 3 to human and mouse VEGFR2/ Fc (R&D Systems, Minneapolis, MN) were carried out on a Biacore X100 instrument (General Electric, Piscataway, NJ). Protein A (Zymed, South San Francisco, CA) was covalently coupled to a CM5 sensor chip (General Electric) using an amine coupling kit (General Electric). For compound 1, the sample
560 Bioconjugate Chem., Vol. 21, No. 3, 2010
and running buffer was PBS with 0.005% P20 (General Electric). For compound 3, the buffer was PBS with the amount of P20 increased to 0.1% to block nonspecific interactions with the sensor chip. For each measurement cycle, human or mouse VEGFR2/Fc was immobilized on the Protein A-conjugated sensor chip at 5 µL/min, then varying concentrations of the compound solution were injected for 3 min at 30 µL/min. Dissociation of the compound from its binding target was measured for 10 min after the end of each injection. The chip was regenerated between cycles with serial injections of 10 mM glycine (pH 2) and 0.025% SDS in H2O. The sensorgrams obtained for each compound were fit globally to a bivalent binding model using Biacore X100 Evaluation Software v 1.0. Preparation of VEGFR2-Targeted Microbubbles for Ultrasound Imaging. Compound 3 was incorporated in the phospholipid membrane of microbubbles as previously described (17). A 2 mL aliquot of a 5% glucose solution (B. Braun Medical, Sempach, Switzerland) was added to the resulting lyophilized product (BR55). After dissolution of the cake, a suspension of VEGFR2-targeted microbubbles was obtained. In Vivo Ultrasound Imaging. Ultrasound imaging of VEGFR2-targeted microbubbles (BR55) was performed in an orthotopic breast tumor model in rats as previously described (17). Briefly, tumor was induced by injection of 13762 MAT B III rat mammary adenocarcinoma cells (5 × 105 in 100 µL of culture medium) (American Tissue Culture Collection, Manassas, VA) in the mammary fat pad of female Fischer 344 rats (Charles River Laboratories, France). Ultrasound imaging was performed after eight days of tumor growth using a Sequoia 512 scanner (Siemens Medical Systems, Issaquah, WA) equipped with a 15L8 linear transducer (transmit frequency, 7 MHz). BR55 (2.4 × 107 microbubbles) was injected intravenously, and intermittent tumor imaging (T ) 0 s, T ) 20 s, T ) 10 min) was performed using contrast pulse sequencing (CPS) mode at low acoustic power (MI ) 0.25) to follow the accumulation of the targeted microbubbles in the tumor. After the 10 min imaging time point, microbubbles present in the tumor were destroyed by increasing the acoustic power (MI ) 1.9) and the tumor was imaged immediately thereafter (T ) 11 min).
RESULTS Synthesis, Purification, Characterization, and Stability Studies of Phospholipid-PEG2000-Peptide Conjugate 3. We have shown that the affinity of a heterodimer obtained by linking of two phage-derived noncompetitive VEGFR2 (KDR) binding peptides is significantly enhanced vs that of the constituent monomer peptides (8). The heterodimer peptide 1 was reacted with DSG (disuccinimidyl glutarate) to provide 2, the glutaric acid mono-NHS ester monoamide of the heterodimer peptide 1. The remaining DSG was removed by trituration of compound 2 with ethyl acetate, pelleting and decanting the DSG-rich supernatant. Then, the heterodimer peptide glutaric acid monoamide mono-NHS ester 2 was reacted with DSPE-PEG2000NH2 to provide the DSPE-PEG2000-NH-glutaryl heterodimer peptide 3 (Scheme 1). The crude product was purified by preparative HPLC on a C3 column using an eluent system containing 0.1% TFA. It was observed that the isolated product, bearing TFA, was not stable and, upon standing in solution or as a lyophilizate, undergoes a time-dependent degradation to a lysophospholipid 4 via hydrolysis of one of the stearoyl groups. The formation of compound 4 could be demonstrated by accelerated hydrolysis in the presence of higher concentrations of TFA than that obtained by dissolution of the isolated product 3. Lysophospholipid 4 was also obtained by enzymatic hydrolysis of compound 3 using Phospholipase A2 (Scheme 2). Both methods
Pillai et al.
Figure 1. HPLC chromatogram of a mixture containing heterodimer peptide 1, its DSPE-PEG2000 conjugate 3, degradation products 4 and 5 and DSPE-PEG2000-NH2. Conditions: column, Agilent C3, 150 mm × 3 mm, 3.5 µm particle, 300 Å pore; eluents; A H2O (10 mM NH4OAc), B CH3CN-H2O 9:1 (v/v) (10 mM NH4OAc); gradient 50-100% B over 30 min, flow rate 1.0 mL/min; detection, ELSD, temp 50 °C, pressure 3.2 psi, sensitivity 10. 5 ) peptide phosphoglyceryl ester, 1 ) heterodimer peptide, 3 ) DSPE-PEG2000-Glut-heterodimer peptide 1 conjugate, 4 ) lysophospholipid.
provided a product which displayed the same HPLC retention time and mass spectrum as the initially detected lysophospholipid 4. Lysophospholipid formation could be eliminated by conversion of the trifluoroacetate salt of heterodimer peptide 1 to the stable acetate salt by ion exchange chromatography on AG MP50 ion-exchange resin and use of this material for the conjugation reaction that provides compound 3. The decrease in TFA content could be quantified with capillary electrophoresis. The DSPE-PEG2000-NH-glutaryl-heterodimer peptide acetate salt was prepared using DSG followed by preparative HPLC purification using an eluent system buffered with ammonium acetate (Scheme 1) instead of TFA. The HPLC analysis conditions for the separation of the desired product, as well as intermediates and side products, are shown in Figure 1. Studies showed that the material isolated using this procedure is very stable in the solid state and in solution. As an example, aqueous solutions of the isolated material were stable for 21 days at ambient temperature. After isolation of the acetate salt of 3, we were surprised to find that this material has a very high affinity for TFA, perhaps due to the PEG linker. Increases in TFA content were obtained from lyophilization with other TFA-containing peptides or during preparative HPLC with ammonium acetate buffer on a column that had previously been used with TFA-containing eluent systems. Binding Studies of Heterodimer Peptide 1 and PhospholipidPEG2000-Peptide Conjugate 3. SPR analysis of compound 1 binding to human VEGR-2/Fc indicated that the peptide bound tightly with a KD of 0.22 nM (Table 1). This is somewhat lower than the KD of 0.5 nM reported for this compound earlier (7), most likely due to the present method of immobilizing VEGFR2/ Fc with Protein A, which may better orient the target for binding than did direct amine coupling to the sensor chip. The binding of compound 1 to mouse VEGFR2/Fc was almost 8-fold weaker, with a KD of 1.6 nM indicating a partial binding preference for the human protein. The phospholipid-PEG2000 conjugate 3 bound human VEGFR2/Fc with a KD of 4.8 nM (Table 1). The
Phospholipid Peptide for Ultrasound Imaging
Bioconjugate Chem., Vol. 21, No. 3, 2010 561
Table 1. Surface Plasmon Resonance Analysis of Heterodimeric Peptides Binding to Human and Mouse VEGFR2/Fc
cmpd 1 cmpd 3 a
kon (M-1 s-1) × 106
Human VEGFR2/Fc koff (s-1) × 10-4
KD (pM)
na
kon (M-1 s-1) × 106
Mouse VEGFR2/Fc koff (s-1) × 10-4
KD (pM)
na
1.2 ( 0.2 0.024 ( 0.007
2.6 ( 0.3 1.1 ( 0.01
220 ( 20 4800 ( 1,500
3 2
0.36 ( 0.07 ndb
5.5 ( 1.1 ndb
1600 ( 570 ndb
2 -
n, number of determinations. b nd, not determined.
Figure 2. Ultrasound images of MAT B III tumors in CPS mode. Tumor, green outline; nontumoral vasculature, red outline. A: Baseline image before injection of BR55 - the white areas are high intensity reflections of hyperechoic structures. B: At 20 s after injection of BR55 (peak intensity), the circulating microbubbles delineate the vasculature feeding the tumor and the tumor vasculature. C: At 10 min after injection of BR55, freely circulating microbubbles are cleared from the circulation, while the VEGFR2-targeted microbubbles are adherent to the tumor vasculature. D: At 11 min, imaging after destruction of bubbles in the tumor with high acoustic pressure followed by a waiting period for refilling of the tumor with residual circulating bubbles, demonstrated that >95% of the intensity of the image in panel C was due to tumor-bound microbubbles. Depth ) 20 mm.
difference in KD relative to compound 1 was entirely due to a much slower on-rate. Since a higher concentration of P20 detergent had to be used when testing compound 3 to prevent excessive nonspecific binding, it is possible that the kinetic values obtained underestimate the binding to human VEGFR2/ Fc. Regardless, it is clear that compound 3 retained potent human VEGFR2 binding capability. Ultrasound Imaging Results with VEGFR2-Targeted Microbubbles (BR55). BR55 was evaluated in a group of five rats bearing MAT B III mammary tumors of approximately 5-8 mm in diameter (Figure 2). Prior to injection of BR55, ultrasound imaging of the tumor and surrounding area displayed
a dark field with small areas of high intensity (panel A); these are artifacts due to highly echogenic anatomical structures (18). Following injection of BR55, the strong signal enhancement, due to circulating microbubbles both in the vasculature feeding the tumor and in the intratumoral vasculature, indicated good perfusion of the lesion (panel B). Ten minutes after injection, when microbubbles are largely cleared from the circulation, the signal from the tumor was still enhanced (panel C), indicating that microbubbles were bound to the tumoral endothelium (19). Contrast enhancement in the tumor was reduced to background values following the application of a burst of high acoustic pressure and a 1 min waiting period to allow residual circulating
562 Bioconjugate Chem., Vol. 21, No. 3, 2010
microbubbles to perfuse the tumor (panel D). This confirmed that >95% of the tumoral signal intensity obtained in panel C was due to targeted microbubbles.
DISCUSSION The preparation of targeted ultrasound bubbles extends the value of contrast enhanced ultrasound. Targeted bubbles are routinely formed in a lab environment by conjugating biotinylated antibodies, peptides or small molecules to streptavidincoated bubbles (e.g., Target Ready MicroMarker bubbles, VisualSonics). This approach is ideal for a lab environment but not acceptable for human clinical trials, due to the antigenicity of streptavidin (10) and the need to conjugate the targeting group to the bubble at the time of use. For clinical use, there are obvious advantages to direct incorporation of the peptide into the lipid shell of the bubbles. Although a number of different methods for covalent attachment of peptides to lipids have been published, we have found that the coupling of peptides to DSPE-PEG2000 amine with disuccinimidyl glutarate is a mild high-yield process. Although other lengths of PEG spacers were evaluated, the PEG2000 linker was optimal and provided a phospholipid-peptide conjugate with good binding as indicated by surface plasmon resonance studies. In order to minimize TFA in the product, it proved advantageous to use the acetate salt of the heterodimer peptide 1 in the synthesis of the phospholipid-peptide 3. Preparative-HPLC employing a dedicated TFA-free system, ammonium acetate buffered eluents, and analysis of collected fractions using an ELS detector was ideal for obtaining a pure and stable product. Adventitious contamination of the isolated fractions with TFA was further prevented by lyophilization of the pooled productcontaining fractions on a TFA-free lyophilization unit. This resulting pure DSPE-PEG2000-heterodimer peptide conjugate 3 could then be incorporated into microbubbles which provided images of active angiogenesis in a rat tumor model. Compound 3 is suitable for further advanced animal studies, as well as clinical studies. The methods for synthesis, purification, and analysis described herein should prove applicable to other phospholipid-PEG linked peptides or antibodies for targeted ultrasound clinical studies. Supporting Information Available: HPLC chromatogram of purified compound 3, MALDI-TOF mass spectra of compounds 3 and 4. This material is available free of charge via the Internet at http://pubs.acs.org.
LITERATURE CITED (1) Klibanov, A. L. (2005) Ligand-carrying gas-filled microbubbles: ultrasound contrast agents for targeted molecular imaging. Bioconjugate Chem. 16, 9–17. (2) Carmeliet, P. (2005) Angiogenesis in life, disease and medicine. Nature 438, 932–936. (3) Ellegala, D. B., Leong-Poi, H., Carpenter, J. E., Klibanov, A. L., Kaul, S., Shaffrey, M. E., Sklenar, J., and Lindner, J. R. (2003) Imaging tumor angiogenesis with contrast ultrasound and microbubbles targeted to Rvβ3. Circulation 108, 336–341. (4) Weller, G. E. R., Wong., M. K. K., Modzelewski, R. A., Lu, E., Klibanov, A. L., Wagner, W. R., and Villanueva, F. S. (2005) Ultrasonic imaging of tumor angiogenesis using contrast microbubbles targeted via the tumor-binding peptide ArginineArginine-Leucine. Cancer Res. 65, 533–539. (5) Ferrara, N., Gerber, H. P., and LeCouter, J. (2003) The biology of VEGF and its receptors. Nat. Med. 9, 669–676.
Pillai et al. (6) Willmann, J. K., Cheng, Z., Davis, C., Lutz, A. M., Schipper, M. L., Nielsen, C. H., and Gambhir, S. S. (2008) Targeted microbubbles for imaging tumor angiogenesis: assessment of whole-body biodistribution with dynamic micro-PET in mice. Radiology (Oak Brook, IL, U. S.) 249, 212–219. (7) Shrivastava, A., von Wronski, M. A., Sato, A. K., Dransfield, D. T., Sexton, D., Bogdan, N., Pillai, R., Nanjappan, P., Song, B., Marinelli, E., DeOliveira, D., Luneau, C., Devlin, M., Muruganandam, A., Abujoub, A., Connelly, G., Wu, Q. L., Conley, G., Chang, Q., Tweedle, M. T., Ladner, R. C., Swenson, R. E., and Nunn, A. D. (2005) A distinct strategy to generate high-affinity peptide binders to receptor tyrosine kinases. Prot. Eng. Des. Sel. 18, 417–424. (8) Pillai, R., Marinelli, E. R., and Swenson, R. E. (2006) A flexible method for preparation of peptide homo- and heterodimers functionalized with affinity probes, chelating ligands, and latent conjugating groups. Biopolymers 84, 576–585. (9) Lindner, J. R., Song, J., Christiansen, J., Klibanov, A. L., Xu, F., and Ley, K. (2001) Ultrasound assessment of inflammation and renal tissue injury with microbubbles targeted to P-selectin. Circulation 104, 2107–2112. (10) Meyer, D. L., Schultz, J., Lin, Y., Henry, A., Sanderson, J., Jackson, J., Goshorn, S., Rees, A. R., and Graves, S. S. (2001) Reduced antibody response to streptavidin through site-directed mutagenesis. Protein Sci. 10, 491–503. (11) Kirpotin, D., Park, J. W., Hong, K., Zalipsky, S., Li, W. L., Carter, P., Benz, C. C., and Papahadjpoulos, D. (1997) Sterically stabilized anti-HER2 immunoliposomes: design and targeting to human breast cancer cells in vitro. Biochemistry 36, 66–75. (12) Torchilin, V. P., Levchenko, T. S., Lukyanov, A. N., Khaw, B. A., Klibanov, A. L., Rammohan, R., Samokhin, G. P., and Whiteman, K. R. (2001) p-Nitrophenylcarbonyl-PEG-PE-liposomes: fast and simple attachment of specific ligands, including monoclonal antibodies, to distal ends of PEG chains via p-nitrophenylcarbonyl groups. Biochim. Biophys. Acta, Biomembr. 1511, 397–411. (13) Maeda, N., Takeuchi, Y., Takada, M., Namba, Y., and Oku, N. (2003) Synthesis of angiogenesis-targeted peptide and hydrophobized polyethylene glycol conjugate. Bioorg. Med. Chem. Lett. 14, 1015–1017. (14) Grogan, M. J., Kaizuka, Y., Conrad, R. M., Groves, J. T., and Bertozzi, C. R. (2005) Synthesis of lipidated green fluorescent protein and its incorporation in supported lipid bilayers. J. Am. Chem. Soc. 127, 14383–14387. (15) Schumann, P. A., Christiansen, J. P., Quigley, R. M., McCreery, T. P., Sweitzer, R. H., Unger, E. C., Lindner, J. R., and Matsunaga, T. O. (2002) Targeted-microbubble binding selectively to GPIIb IIIa receptors of platelet thrombi. InVest. Radiol. 37, 587–593. (16) Dennis, E. A. (1987) The Regulation of Eicosanoid Production: Role of Phospholipases and Inhibitors. Bio/Technology 5, 1294–1300. (17) Pochon, S., Tardy, I., Bussat, P., Bettinger, T., Brochot, J., Von Wronski, M., Passantino, L., and Schneider, M. (2010) BR55: A lipopeptide-based VEGFR2-targeted ultrasound contrast agent for molecular imaging of angiogenesis. InVest. Radiol. 45, 89–95. (18) Farrow, C. S. (1996) How ultrasound works. Small Animal Imaging (Green, R. W. , Ed.) pp 19-27, Chapter 2, Lippincott, Williams and Williams, Philadelphia, PA. (19) Lindner, J. R., Song, J., Xu, F., Klibanov, A. L., Singbartl, K., Ley, K., and Kaul, S. (2000) Noninvasive ultrasound imaging of inflammation using microbubbles targeted to activated leukocytes. Circulation 102, 2745–2750. BC9005688
