Targeting vascular endothelium with avidin microbubbles

Ultrasound in Med. & Biol., Vol. 31, No. 9, pp. 1279 –1283, 2005 Copyright © 2005 World Federation for Ultrasound in Medicine & Biology Printed in the USA. All rights reserved 0301-5629/05/$–see front matter

doi:10.1016/j.ultrasmedbio.2005.06.001

● Technical Note TARGETING VASCULAR ENDOTHELIUM WITH AVIDIN MICROBUBBLES GRZEGORZ KORPANTY,* PAUL A. GRAYBURN,‡ RALPH V. SHOHET,† and ROLF A. BREKKEN* *Hamon Center for Therapeutic Oncology Research, Departments of Surgery and Pharmacology and †Division of Cardiology, University of Texas Southwestern Medical School, Dallas, TX, USA; and ‡Baylor University Medical Center, Baylor Heart and Vascular Institute, Dallas, TX, USA (Received 11 January 2005, revised 8 May 2005, in final form 24 May 2005)

Abstract—Targeting microbubbles (MBs) to specific vascular beds enables contrast ultrasound to be used for molecular imaging. There are several methods for attaching targeting moieties to the surface of MBs. In the present study, we demonstrate that avidin (Av) can be incorporated into the shell of perfluorocarbon-exposed sonicated dextrose albumin (PESDA) MBs (Av-PESDA-MBs) and serve as an anchor that links Av-PESDA-MBs to biotinylated monoclonal antibodies (mAbs). This novel linking strategy was used to conjugate Av-PESDA-MBs to mAbs specific for endoglin (CD105) or a control IgG. MBs targeted to CD105 specifically bound to endothelial cells, but not to fibroblasts, in vitro but Av-PESDA-MBs conjugated with the control IgG did not specifically target either cell type. We conclude that Av-PESDA-MBs represent a novel and attractive tool to conjugate MBs with biotinylated mAbs for the purposes of vascular targeting and molecular imaging. (E-mail: [email protected]) © 2005 World Federation for Ultrasound in Medicine & Biology. Key Words: Contrast ultrasound, Targeted microbubbles, Vascular endothelium, Molecular imaging, Avidinbiotin system.

tions provide an efficient method for noncovalent coupling of molecules. In general, when the avidin-biotin system is used to link a targeting agent to MBs, the MB shell consists of a biotinylated compound. The biotin-containing MBs are incubated with avidin (Av), washed and then incubated with a biotinylated targeting moiety, resulting in a biotinavidin-biotin bridge between the MBs and the targeting moiety (Klibanov 1999). In the present study, we hypothesized that Av can be incorporated into the shell of perfluorocarbon-exposed dextrose albumin MBs by sonication (PESDA) and retain its capacity for coupling biotinylated monoclonal antibodies (mAbs). This new method, thus, requires only one incubation and two wash steps vs. two incubations and three wash steps that are needed when biotinylated derivates are incorporated into shells of MBs (Klibanov 1999). To demonstrate the efficacy of this novel system, we chose to target endoglin (CD105), a glycoprotein expressed on the surface of endothelial cells.

INTRODUCTION Microbubbles (MBs) have been approved by the US Food and Drug Administration as IV contrast agents for nonspecific and passive enhancement of ultrasound (US) images (Bruce et al. 2004). MBs can be targeted to specific vascular beds using targeting moieties (antibodies, peptides) that bind to vascular endothelium. A number of studies have demonstrated that contrast US using targeted MBs is useful in distinguishing normal and pathologic tissue (Lindner et al. 2000; Schumann et al. 2002; Villanueva et al. 1998). MBs have been targeted to vascular endothelium in inflamed tissue, tumors and to platelets in thrombotic vessels (Ellegala et al. 2003; Schumann et al. 2002; Villanueva et al. 1998). The targeting agent can be attached covalently to the MB shell with the use of chemical bifunctional linkers or noncovalently by the use of an avidin-biotin system (Klibanov 1999; Villanueva et al. 1998). Avidin-biotin interac-

MATERIALS AND METHODS

Address correspondence to: Rolf A. Brekken, Ph.D., Hamon Center for Therapeutic Oncology Research, Departments of Surgery and Pharmacology, UT-Southwestern Medical Center, 6000 Harry Hines Blvd., Dallas, TX 75390-8593 USA. E-mail: rolf. [email protected]

Cell culture Mouse fibroblasts (3T3) and polyoma middle Ttransformed mouse brain capillary endothelial cell line 1279

1280

Ultrasound in Medicine and Biology

(bEND.3) were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen, Carlsbad, CA, USA) containing 5% fetal bovine serum (FBS) (Gemini Bio-Products, Woodland, CA, USA). The hybridomas MJ7/18 (obtained from the Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA, USA) and AFRC Mac 48 (Mac 48) (European Collection of Animal Cell Cultures) were grown in Iscove’s modified Dulbecco’s medium (Invitrogen) with 5% FBS. All cell lines were grown at 37 °C in humidified 5% CO2 atmosphere. Monoclonal antibody production, purification and modification Rat mAbs against mouse CD105, MJ7/18 (Ge and Butcher 1994) and the isotype-matched control rat mAbs against a plant antigen (phytochrome), Mac 48, were isolated from tissue culture supernatant by protein A chromatography using Pierce ImmunoPure binding/elution buffering system (Pierce Biotechnology, Rockford, IL, USA) and dialyzed against phosphate-buffered saline (PBS). Purity was judged by sodium dodecyl sulfate (SDS)-gel electrophoresis followed by staining with Coomassie brilliant blue R-250. Purified mAbs were biotinylated by incubation with a 10-fold molar excess of biotinamidocaproate N-hydroxysuccinimide (NHS)-ester (Sigma, St. Louis, MO, USA) for 1 h at room temperature. The reaction was stopped by adding tricine (tris/ glycine) (Sigma) to 10 mM. Free biotin was removed by size exclusion column chromatography using a PD-10 column (Amersham Biosciences, Piscataway, NJ, USA). For imaging and flow cytometry, biotinylated mAbs were labeled with fluorescein isothiocyanate (FITC) using EZ-Label™ fluorescein protein labeling kit (Pierce) according to the manufacturer’s protocol. The same labeling protocol was used for conjugation of bovine serum albumin (BSA) with FITC. Production of avidin microbubbles PESDA-MBs were prepared as described by Porter et al. (2001). Avidin containing MBs (Av-PEDSA-MBs) were generated by sonicating a solution of 5% bovine serum albumin (BSA) (Calbiochem, San Diego, CA, USA), 0.5% fluorescein-conjugated avidin (FITC-Av) (Pierce) and 5% dextrose (Sigma) in PBS. FITC-Av was mixed with BSA at eight different weight ratios (FITCAv:BSA), 1:1000, 1:500, 1:250, 1:125, 1:62.5, 1:31.25, 1:25 and 1:12.5. Three volume parts of 5% dextrose (1.5 mL) and one part of BSA and FITC-Av mixture (0.5 mL) were drawn up into a 3-mL syringe that was connected with three-way stopcock to another 3-mL syringe with 2 mL of perfluoropropane gas (C3F8) (Air Products, Inc., Allentown, PA, USA). The solution of dextrose, BSA and FITC-Av was mixed back and forth between these

Volume 31, Number 9, 2005

two syringes 20 times and placed into a 15-mL conical polypropylene tube and sonicated at 20 kHz for 16 s using an ultrasonic processor XL2020 (Heat Systems Inc., Farmingdale, NY, USA). Immediately after that, the tube was placed on ice for 30 min, to allow MBs to separate by flotation. The concentration of MBs ranged from 1 to 3 ⫻ 109 per mL and the mean diameter ranged from 1.1 to 1.4 ⫾ 0.13 ␮m when assessed by Multisizer™ 3 Coulter counter (Beckman Coulter Inc., Fullerton, CA, USA). Flow cytometry was used to demonstrate that FITC-Av was incorporated into the MB shells (AvPESDA-MBs). Flow cytometry Flow cytometry was used to analyze the various MB constructs. For example, flow cytometry was used to demonstrate that FITC-Av was incorporated into the MB shell and also to show that FITC-conjugated biotinylated IgG bound to Av-PEDSA-MBs (see below). In the latter case, the Av-PEDSA-MBs were produced using unlabeled Av. In each case, 100 ␮L of MB were diluted in 2 mL of PBS and samples were analyzed for fluorescence using Becton Dickinson FACSCalibur™ system (Becton Dickinson, Franklin Lakes, NJ, USA). Data were analyzed using CellQuest software (Becton Dickinson). A total of 10,000 events were captured for each sample in duplicate. Conjugation of biotinylated mAbs to Av-PESDA-MBs Av-PESDA-MBs were prepared as described above, except that 1% avidin (NeutrAvidin™, Pierce) was used instead of 0.5% Av and it was not labeled with FITC. Before incubation with the labeled IgG, the MBs were washed to remove free Av and BSA, as described by Takalkar et al. (2004). Different amounts of FITCconjugated biotinylated IgG (MJ7/18 or Mac 48) were incubated with 100 ␮L of washed Av-PESDA-MBs (approximately 108 MBs). Excess unbound antibody was removed by washing with PBS. IgG-linked MBs were analyzed by flow cytometry, as described above. In vitro cell binding assay Mouse fibroblasts (3T3) that express low levels of CD105 (St-Jacques et al. 1994) and mouse endothelial cells (bEND.3) that express CD105 were grown on glass slides, washed once with ice-cold PBS, fixed in 3% paraformaldehyde in DMEM and washed with PBS again. Fixation does not influence the expression level of CD105, but enhances adherence of cells to the glass slides. A total of 100 ␮L of Av-PEDSA-MBs produced using FITC-labeled BSA (FITC-BSA) and unlabeled Av linked to unlabeled biotinylated IgG (MJ7/18 or Mac 48) were placed on an empty slide. A slide with fixed cells was placed on the top of the slide with the MBs. After 15

Targeting avidin microbubbles ● G. KORPANTY et al.

Table 1. Av-PESDA-MBs conjugated to biotinylated MJ7/18 bind to endothelial cells

1281

er’s post hoc test and considered to be significant at p ⬍ 0.05.

Experimental groups Antibody MJ7/18 Mac 48 MJ7/18 Mac 48

Cells

Pixels reaching threshold (%)

bEND.3 bEND.3 3T3 3T3

38.1 ⫾ 19.3* 0.06 ⫾ 0.04 0.02 ⫾ 0.01 0.04 ⫾ 0.02

bEND.3 and 3T3 cells were grown on glass slides, fixed and incubated with Av-PESDA-MBs that were produced with of FITC-BSA and avidin conjugated with either biotinylated MJ7/18 or Mac 48. Fluorescence intensity was determined by using MetaMorph software. The fluorescence threshold was set by incubating control untargeted Av-PESDA-MBs with bEND.3 and 3T3 cells; *p ⬍ .0001.

min of static exposure, slides with cells were washed 3 times with PBS and mounted with propidium iodidecontaining medium (Vectashield, Vector Laboratories Inc., Burlingame, CA, USA). Green fluorescence (FITCBSA) was detected using fluorescent microscopy (Nikon Eclipse E600). Images were recorded using a Cool-Snap HQ camera (Photometrix, Roper Scientific, Tucson, AZ, USA) and MetaMorph (Universal Imaging Corporation, Downingtown, PA, USA) software. As a control, we used MBs targeted with an IgG of irrelevant specificity, Mac 48. The fluorescent signal was quantified using MetaMorph software. For quantification purposes, AvPESDA-MBs (untargeted and targeted, both made with FITC-BSA) were used and the same exposure time was used to capture all images. The fluorescence intensity of cells incubated with untargeted MBs was set as the threshold (background fluorescence). The percentage of pixels over the threshold value (background) was calculated by taking the mean value for six random fields for each cell line used (Table 1). The values displayed represent the mean relative fluorescence intensity and are expressed as the percentage of pixels above the threshold value. The cells were imaged at a total magnification of 100x (Table 1) or 1000x (Fig. 3), which equals approximately 4.1 mm2 and 0.04 mm2, respectively. The number of cells at a total magnification 100x is approximately 500 for bEND.3 and 500 for 3T3 cells. The number of confluent coverslips analyzed for each cell line was 20 (n ⫽ 20). Half of the slides from each cell line (n ⫽ 10) were incubated with microbubbles conjugated to MJ7/18 and half (n ⫽ 10) with microbubbles conjugated to the control antibody. Statistical analysis Data were analyzed with Statview software (SAS, Cary, NC, USA). The results are expressed as mean ⫾ 1 SD. Differences were analyzed by ANOVA with Fish-

RESULTS Production of avidin MBs and conjugation with biotinylated mAbs Sonication of albumin, FITC-Av and dextrose in the presence of perfluoropropane gas resulted in incorporation of FITC-Av into the shell of PESDA-MBs (AvPESDA-MBs). The percentage of Av-PESDA-MBs formed during sonication was dependent on the concentration of Av, as shown in Fig. 1. Each MB preparation had a similar mean diameter. The exact number of Av molecules incorporated into the shell of single MB is not clear, but we estimate that, for the MBs made with a 1:31.25 ratio of Av:BSA, the number to be between 105–106, based on the concentration of Av used and the mean concentration of MBs in solution. Av-PESDA-MBs were successfully conjugated with biotinylated mAbs. The percentage of antibody linked Av-PESDA-MBs was dependent on both Av:BSA ratio and the total amount of mAbs that were incubated with the MBs. To determine the optimal conditions (i.e., using the smallest possible amounts of both Av and mAbs) for linking biotinylated mAbs to Av-PESDAMBs, various concentrations of biotinylated mAbs were incubated with the Av-PESDA-MBs that were made with three different concentrations of Av. The percentage of MBs linked to FITC-conjugated biotinylated mAbs was determined by flow cytometry. The best yield of mAb-linked MBs was achieved by using a ratio of

Fig. 1. Analysis of Av-PESDA-MBs by flow cytometry. MBs were prepared by sonicating BSA, dextrose and FITC-Av in the presence of perfluoropropane gas. The incorporation of FITC-Av into the shell of MBs was concentration-dependent. Mean diameter of MBs was determined with Multisizer™ 3 Coulter counter and varied between 1.0 to 1.4 ⫾ 0.13 ␮m.

1282

Ultrasound in Medicine and Biology

Fig. 2. Flow cytometric analysis of Av-PESDA-MBs linked to FITC-labeled and biotinylated MJ7/18 antibody. The ratio of Av:BSA used to make the MBs is indicated. The optimal ratio of Av:BSA was 1:31.25 and the concentration of antibody that gave the highest linking to MBs was 0.8 ␮M. Using these conditions, 66.5 ⫾ 0.2% of all MBs were linked to MJ7/18.

1:31.25 Av (6.7 ␮M) to BSA (189.4 ␮M) and by subsequent incubation with 0.8 ␮M of antibody. These experimental settings resulted in 66.5% ⫾ 0.2 of MBs in solution with antibody bound to their shells (Fig. 2). Targeting Av-PESDA-MBs to endothelial cells in vitro MJ7/18 or Mac 48 linked to Av-PESDA-MBs made with FITC-BSA were incubated with either bEND.3 or 3T3 cells, which express high or low levels of CD105, respectively (St-Jacques et al. 1994). Figure 3a shows MJ7/18 conjugated Av-PESDA-MBs bound to the surface of endothelial cells, but not 3T3 cells (Fig. 3c). Av-PESDA-MBs linked to Mac 48 did not bind to either bEND.3 or 3T3 cells (Fig. 3b, d). In addition, there was no binding of untargeted Av-PESDA-MBs to either 3T3 or bEND.3 cells (data not shown). Specific binding was quantified (Table 1) by determining the percentage of pixels that exceeded a background threshold value.

Volume 31, Number 9, 2005

The introduction of MBs as US contrast agents has enhanced tissue characterization and edge detection, and targeting MBs with specific ligands permits detection of disease-specific proteins, receptors or antigens (biomarkers) (Dayton and Ferrara 2002). Because MBs are true intravascular tracers, their application in molecular imaging is well-suited for targets expressed on vascular endothelium (Lanza and Wickline 2003). The utility of MBs targeted to sites of vascular pathology has been demonstrated previously using animal models of inflammation, atherosclerosis, thrombosis and tumor angiogenesis (Lindner et al. 2000; Schumann et al. 2002; Ellegala et al. 2003). There are several strategies to target MBs. Attachment of plain untargeted MBs to vascular endothelium or blood cells can be based on the chemical or electrostatic properties of the MB shells (Lindner et al. 2000). MB targeting can be also mediated by mAbs, which provide enhanced specificity and offer greater flexibility for target selection. We assessed the binding of specific and control antibody-conjugated MBs to endothelial cells (bEND.3) and fibroblasts (3T3). Av-PESDA-MBs conjugated to mAb MJ7/18 bound specifically to endothelial cells, but not to fibroblasts (Fig. 3; Table 1), but untargeted and control IgG-targeted MBs did not bind to either bEND.3 or 3T3 cells. Because the binding efficacy of MBs targeted with MJ7/18 was assessed under “static” conditions and in vitro, further studies on the dynamic behavior of the MBs under in vivo conditions are needed. One

DISCUSSION In this study, we demonstrate that Av can be incorporated into the shell of MBs during sonication and that it can serve as a direct ligand for biotinylated mAbs. We further show ligand-directed accumulation of MBs on endothelial cells in vitro. This study describes a novel method of linking targeting moieties to MBs that is easier to perform when compared with previously described protocols that make use of the biotin-avidin system (Klibanov 1999).

Fig. 3. MJ7/18-linked MBs bind to endothelial cells. (a), (b) bEND.3 and (c), (d) 3T3 cells were grown on glass slides, fixed and incubated with Av-PESDA-MBs (made with FITC-BSA) linked to either biotinylated MJ7/18 (a), (c) or Mac 48 (b), (d). Nuclei are stained red with propidium iodine. MBs conjugated to MJ7/18 adhere to (a) bEND.3 endothelial cells but not to (c) 3T3 cells, and Mac 48-linked MBs do not bind to (b) bEND.3 or (d) 3T3 cells.

Targeting avidin microbubbles ● G. KORPANTY et al.

of the possible limitations of the avidin-biotin targeting system used in vivo is host immune response to Av. An emerging clinical application of US is to detect and monitor the onset and progression of angiogenesis (new blood vessel formation) together with noninvasive blood flow assessment in tumors (Krix et al. 2003). Activated and proliferating endothelial cells lining tumor vessels express CD105, a 180-kDa transmembrane protein that serves as an accessory receptor for transforming growth factors beta 1 and 3 (TGF-␤1, -3) (Li et al. 2000; Brekken at al. 2002). Intratumoral microvascular density determined using mAbs to CD105 has been found to be an independent prognostic indicator, with increased microvascular density correlating to shorter survival of cancer patients (Duff et al. 2003). Thus, CD105 is a promising target for tumor imaging and prognosis and may have therapeutic potential in patients with solid tumors and other angiogenic diseases. We anticipate that CD105-targeted Av-PESDA-MBs combined with US imaging will enable noninvasive characterization of tumor vasculature in vivo. Acknowledgements—The authors thank Peregrine Pharmaceuticals Inc. for financial support of the study. The hybridoma MJ7/18, developed by Eugene C. Butcher and Mark Jutila, was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of NICHD and maintained by The University of Iowa, Department of Biologic Sciences (Iowa City, IA 52242 USA). The authors thank Bonnie Darnell from the Department of Pathology at UTSW for her excellent assistance with flow cytometry.

REFERENCES Brekken RA, Li C, Kumar S. Strategies for vascular targeting in tumors. Int J Cancer 2002;100:123–130. Bruce M, Averkiou M, Tiemann K et al. Vascular flow and perfusion imaging with ultrasound contrast agents. Ultrasound Med Biol 2004;30:735–743.

1283

Dayton PA, Ferrara KW. Targeted imaging using ultrasound. J Magn Reson Imaging 2002;16:362–377. Duff SE, Li C, Garland JM, Kumar S. CD105 is important for angiogenesis: Evidence and potential applications. FASEB J 2003;17: 984 –992. Ellegala DB, Leong-Poi H, Carpenter JE, Klibanov AL, Kaul S, Shaffrey ME, Sklenar J, Lindner JR. Imaging tumor angiogenesis with contrast ultrasound and microbubbles targeted to alpha(v)beta3. Circulation 2003;108:336 –341. Ge AZ, Butcher EC. Cloning and expression of cDNA encoding mouse endoglin, and endothelial cell TGF-beta ligand. Gene 1994;138: 201–206. Klibanov AL. Targeted delivery of gas-filled microspheres, contrast agents for ultrasound imaging. Adv Drug Deliv Rev 1999;37:139 – 157. Krix M, Kiessling F, Vosseler S, et al. Sensitive noninvasive monitoring of tumor perfusion during antiangiogenic therapy by intermittent bolus-contrast power Doppler sonography. Cancer Res 2003; 63:8264 – 8270. Lanza GM, Wickline SA. Targeted ultrasonic contrast agents for molecular imaging and therapy. Curr Probl Cardiol 2003;28:625– 653. Li C, Hampson IN, Hampson L, et al. CD105 antagonizes the inhibitory signaling of transforming growth factor beta1 on human vascular endothelial cells. FASEB J 2000;14:55– 64. Lindner JR, Song J, Xu F, et al. Noninvasive ultrasound imaging of inflammation using microbubbles targeted to activated leukocytes. Circulation 2000;102:2745–2750. Porter TR, Hiser WL, Kricsfeld D, et al. Inhibition of carotid artery neointimal formation with intravenous microbubbles. Ultrasound Med Biol 2001;27:259 –265. Schumann PA, Christiansen JP, Quigley RM, McCreery TP, Sweitzer RH, Unger EC, Lindner JR, Matsunaga TO. Targeted-microbubble binding selectively to GPIIb IIIa receptors of platelet thrombi. Invest Radiol 2002;37:587–593. St-Jacques S, Cymerman U, Pece N, Letarte M. Molecular characterization and in situ localization of murine endoglin reveal that it is a transforming growth factor-beta binding protein of endothelial and stromal cells. Endocrinology 1994;134:2645–2657. Takalkar AM, Klibanov AL, Rychak JJ, Lindner JR, Ley K. Binding and detachment dynamics of microbubbles targeted to P-selectin under controlled shear flow. J Control Release 2004;96:473– 482. Villanueva FS, Jankowski RJ, Klibanov S, et al. Microbubbles targeted to intercellular adhesion molecule-1 bind to activated coronary artery endothelial cells. Circulation 1998;98:1–5.