Detection of Vascular Adhesion Molecule-1 Expression Using a Novel Multimodal Nanoparticle
http://www.100md.com
循环研究杂志 2005年第2期
the Center for Molecular Imaging Research, Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Charlestown.
Abstract
Endothelial vascular adhesion molecule-1 (VCAM-1) is a critical component of the leukocyteeCendothelial adhesion cascade, and its strict temporal and spatial regulation make it an ideal target for imaging and therapy. The goal of this study was to develop novel VCAM-1eCtargeted imaging agents detectable by MRI and fluorescence imaging using phage displayeCderived peptide sequences and multimodal nanoparticles (NPs). We hypothesized that VCAM-1eCmediated cell internalization of phage displayeCselected peptides could be harnessed as an amplification strategy to chaperone and trap imaging agents inside VCAM-1eCexpressing cells, thus improving target-to-background ratios. To accomplish our goal, iterative phage display was performed on murine endothelium under physiological flow conditions to identify a family of VCAM-1eCmediated cell-internalizing peptides. One specific sequence, containing the VHSPNKK motif that has homology to the -chain of very late antigen (a known ligand for VCAM-1), was shown to bind VCAM-1 and block leukocyteeCendothelial interactions. Compared with VCAM-1 monoclonal antibody, the peptide showed 12-fold higher target-to-background ratios. A VHSPNKK-modified magnetofluorescent NP (VNP) showed high affinity for endothelial cells expressing VCAM-1 but surprisingly low affinity for macrophages. In contrast, a control NP without VCAM-1eCtargeting sequences showed no affinity for endothelial cells. In vivo, VNP successfully identified VCAM-1eCexpressing endothelial cells in a murine tumor necrosis factor-eCinduced inflammatory model and colocalized with VCAM-1eCexpressing cells in atherosclerotic lesions present in cholesterol-fed apolipoprotein E apoEeC/eC mice. These results indicate that: (1) small peptide sequences can significantly alter targeting of NPs, (2) the used amplification strategy of internalization results in high target-to-background ratios, and (3) this technology is useful for in vivo imaging of endothelial markers.
Key Words: cardiovascular diseases cell adhesion molecules imaging inflammation vasculature
Introduction
Cardiovascular diseases are the leading cause of death in the developed world and represent an immense clinical burden. Atherosclerosis is the major contributor to the pathogenesis of acute myocardial and cerebral thrombosis,1 with 500 000 deaths annually in the United States alone. Although significant advances have been made in the primary prevention of cardiovascular disease, risk stratification remains suboptimal in large segments of the population. It is now widely accepted that the degree of coronary artery stenosis is not predictive of the risk of plaque rupture and subsequent thrombosis at that location. Clinical risk scores, such as that derived from the Framingham Heart Study, are useful but may lose predictive value in the large segments of the population at intermediate risk.2 Despite the addition of new measurements to stratify clinical risk in individuals at intermediate risk, such as serum C-reactive protein or the detection of coronary calcification by computed tomography, there remains a need for novel molecular imaging probes to better define cardiovascular risk and guide therapy more rationally.
Atherosclerosis is a chronic, progressive inflammatory disease characterized by a specific series of cellular and molecular events,3 from early inflammatory lesions ("fatty streaks"), comprised of monocyte-derived macrophages and T lymphocytes,4 to development of mature plaque and potentially rupture and thrombosis. The earliest molecular changes occur on endothelial surfaces of the aorta and vasa vasorum and are key contributors to the initiation, progression, and thrombotic complications of atherosclerosis.1 Although lipid-lowering therapies have demonstrated direct clinical benefit, there remains a dramatic need for novel diagnostic tools to noninvasively assess the benefit of therapies and identify early molecular changes associated with atherosclerosis, and for new targeted and specific therapies.
The strict temporal and spatial regulation of inducible endothelial adhesion molecules and their critical function in atherosclerosis makes them ideal targets for the next generation of diagnostics and therapeutics. Vascular adhesion molecule-1 (VCAM-1) is upregulated on the endothelium under inflammatory conditions including atherosclerosis and cardiac allograft rejection5eC8 and appears on the endothelial cell surface of atheroprone areas before the onset of visible disease.6 Its importance in the initiation of atherogenesis is demonstrated by delayed lesion development in mice carrying a mutation that hinders VCAM-1 function.9 Comparable to mouse atherosclerosis, VCAM-1 expression is induced early in human atheroma and is an important element in the inflammatory component of atherosclerosis, contributing to monocyte and lymphocyte recruitment from adventitial vessels and the arterial lumen.3,10,11 Although several approaches have been described to image VCAM-1 expression using radiolabeled antibodies (Abs),12,13 these agents usually result in modest target-to-background ratios, limiting their use for in vivo cardiovascular imaging.
We determined that additional amplification steps would be required to image VCAM-1 expression directly in the endothelium. Thus, we developed a modified phage display approach to screen for specific peptide sequences that would be internalized by VCAM-1eCexpressing endothelial cells under physiological flow conditions. One specific sequence (VHSPNKK), with homology to the -chain of very late antigen (VLA4; a known ligand for VCAM-1), was shown to bind VCAM-1 and to block leukocyteeCendothelial interactions. A derived multivalent nanoparticle (NP) carrying the VHSPNKK peptide successfully identified VCAM-1eCexpressing endothelial cells in a murine tumor necrosis factor- (TNF-)eCinduced inflammatory model via confocal microscopy, and furthermore, it allowed in vivo detection of atherosclerotic lesions in apolipoprotein EeC/eC (apoEeC/eC) mice via MRI.
Materials and Methods
Antibodies
Anti-mouse CD31 (clone MEC 13.3), anti-CD16/CD32 (FcR block), anti-mouse CD106 (VCAM-1, clone 429), and anti-mouse CD106-fluorescein isothiocyanate (FITC) VCAM-1, clone 429) were from BD Pharmingen (San Diego, Calif), and anti-mouse VCAM-1 (clone M/K 2.7) was from American Type Culture Collection (Manassas, Va). Biotinylated anti-rat and anti-rabbit IgG (heavy and light chain [H+L]) Texas Red or Cy3 anti-rat IgG (H+L) were from Jackson ImmunoResearch (West Grove, Pa), and anti-FITC/Oregon GreeneChorseradish peroxidase (HRP) was from Molecular Probes (Eugene, Ore). BiotineCanti-M13 rabbit IgG was from AbCam, and goat F(ab')2 anti-rat IgG (H+L)-phycoerythrein and Texas Red were from Caltag Laboratories (Burlingame, Calif).
Mice
C57BL/6 wild-type mice (7 to 9 weeks of age) from Taconic and apoEeC/eC (The Jackson Laboratory) were maintained in approved pathogen-free institutional housing facilities. All experiments were performed according to institutional guidelines. Animals were anesthetized for imaging procedures using isoflurane inhalation (1% to 2% isoflurane; 1 L O2). Animals were euthanized by CO2 asphyxiation as approved by the panel on euthanasia at the American Veterinary Association.
Endothelial Cell Culture
Murine cardiac endothelial cells (MCECs) or murine lung endothelial cells (MLECs) were isolated using previously published methods14 and used at passages 1 to 3.
Phage-Display Library Selection
The parallel plate-flow chamber has been described in detail.15 Smith and Petrenko produced a detailed rationale of phage display selection and negative selection or depletion procedures.16 Phage selection and negative depletion were performed using strain-matched MCECs and MLECs. 1010 PFU of phage, displaying random disulfide-constrained 7-aa peptides (C7C-PhD; New England Biolabs), were drawn across MCEC monolayers at 37°C for 10 minutes at 0.52 mL/min (estimated wall shear stress 1.0 dyne/cm2). MCECs were then incubated at 37°C for an additional 15 minutes to allow time for VCAM-1eCmediated internalization (t1/2 11.2 minutes).17 Extracellular-restricted phage were removed with 0.2 mol/L glycine, pH 2.2 (3x8 minutes). Internalized phage were recovered by lysis with 0.1% triethanolamine (Sigma) in PBS, pH 7.4, (4 minutes; room temperature). Extracts were neutralized with 500 e蘈 of 0.5 mol/L Tris-HCl, pH 9.0. To deplete phage that bound to identical markers present on MCECs and MLECs, the phage pool isolated after one round of selection was subtracted by three rounds of successive incubation at 37°C for 30 minutes with confluent monolayers of strain-matched MLECs that do not express VCAM-1. The phage that were internalized by MCECs but not MLECs were amplified by Escherichia coli, titered, subjected to three additional rounds of positive selection, and individual clones selected for ELISA assay and sequencing.
ELISA Assay of VCAM-1 Binding
Protein A (20 e/mL) on Nunc Maxisorp plates (Fisher; 4°C overnight) was incubated sequentially at room temperature with FceCVCAM-1 (10 e/mL; 1 hour) and phage clones (1010 PFU; 1 hour), washed with PBS containing 0.1% Tween-20, incubated with biotinylated anti-M13 Ab (1:40; 1 hour), detected with streptavidin-HRP (1:500), and developed with tetra methyl benzidine, and absorbance650 was determined (Emax; Molecular Devices).
Synthesis of FITC Peptides and Fluorescent Magnetic NP Conjugates
The identified peptide sequence "CVHSPNKKC" and corresponding magnetofluorescent NPs were synthesized as described in the online data supplement (available at http://circres.ahajournals.org). The following compounds were synthesized in bulk: VCAM-1 peptide (VP): CVHSPNKKCGGSK(FITC)GK; control peptide (CP): CPKNVSKHCGGSK(FITC)GK; VCAM-1 NP (VNP): CVHSPNKK-CGGSK(FITC)GK(CLIO-Cy5.5); control NP (CNP): CPKNVSK-HCGGSK(FITC)GK(CLIO-Cy5.5); and CLIO-Cy5.5 (NP): CLIO-Cy5.5 backbone (no peptide).
Fluorescence Microscopy and Flow Cytometry
MCECs were incubated with 1 eol/L of either VP or CP (1 hour; 37°C), washed 3x with PBS, and visualized by fluorescence microscopy (Nikon Eclipse TE2000-S; Insight QE; x40 objective). For confocal microscopy, MCECs were incubated with 1 eol/L of VP as above, fixed with 2% paraformaldehyde, stained with antieCVCAM-1 (10 e/mL; 1 hour), and detected with Texas RedeCconjugated secondary Ab (1:100; 30 minutes). Stained cells were analyzed on an LSM 5 PASCAL confocal microscope (Zeiss). For competition experiments, peptides were preincubated (30 minutes) with 5x murine FceCVCAM-1 or FceCICAM-1 (R & D Systems). MCECs were then detached, stained for cell surface VCAM-1, and analyzed via flow cytometry (10 000 cells/sample) on a Becton Dickinson FACSCalibur. For comparison of VP with antieCVCAM-1eCFITC, cells were incubated with equimolar concentrations (0.5 eol/L) of VP or antieCVCAM-1eCFITC for 2 hours at 37°C and analyzed by flow cytometry.
In Vitro Blocking Studies Under Flow Conditions
MCECs were activated for 5 hours with murine TNF- (mTNF-; R & D Systems) and incubated for 2 hours with either 25 e/mL M/K 2.7 antieCVCAM-1 Ab, 100 eol/L VP, or vehicle control. Strain-matched mononuclear cells (1x106 cells/mL in DPBS+eC0.1% BSA) were drawn across MCECs for 5 minutes at 0.52 mL/min. LeukocyteeCendothelial interactions were determined from 6 to 8 high-power fields at 5 minutes.
Targeting of VNP to Endothelial Cells
The isolation of primary MCECs was performed as described previously. Primary mouse macrophages were isolated as presented in the online data supplement. Cells were incubated with either VNP or NP (4.5 eol/L Cy5.5; 4 hours; 37°C) and analyzed by flow cytometry as described previously.
Immunohistochemistry and Intravital Confocal Microscopy
C57BL/6 mice (n=10) were injected subcutaneously in the right ear with 5 ng/50 e蘈 mTNF- in normal saline. After 24 hours, animals were anesthetized via inhaled isofluorane and injected intravenously (tail vein) with 10 nmol/L fluorochrome of VP (n=2), VNP (n=5), or CNP (n=3). Perfused vessels were located using bright-field illumination and fluorescence signal because of circulating peptide/conjugate. Intravital confocal microscopy was performed using a Nikon EF600N equipped with a BioRad Radiance 2100 confocal and x40 water immersion objective at 0 hours, 4 hours, and 24 hours after agent administration. Image acquisition was performed simultaneously in two fluorescence channels (photo multiplier tube [PMT]1: peptide-FITC; PMT3: peptide-NP) in several fields from both ears of each animal using built-in compensation routines to exclude channel bleed-through (Laser Sharp software; BioRad), and 3D reconstructions were performed using Amira (TGS). Subsequently, ears were removed for histological analysis. Adjacent serial frozen sections were stained for the presence of CD31, VCAM-1, or injected agent. Digital images were taken using a Nikon Eclipse E400 upright microscope (x40 objective) equipped with an Insight color camera.
MRI of VCAM-1 Expression
We performed in vivo MRI of 1-year-old apoEeC/eC mice with extensive lesion burden that had been cholesterol-fed for 3 months (n=3). T1- (repetition time [TR] 50; echo time [TE] 2; 20 number of excitations [NEX] 10 minutes) and T2-weighted (TR 2000; TE 20 to 200; NEX 6 minutes) spin echo images and gradient echo sequences were acquired before and 24 hours after intravenous administration of VNP or NP to apoEeC/eC mice or VNP to C57BL/6 mice (5 mg Fe/kg body weight). In addition, we obtained one bright-blood angiographic sequence (TR 50; TE 2; 20 NEX 10 minutes) at the end of the imaging session using gadolinium-protected graph copolymer (Gd-PGC)18 (0.02 mmol/L Gd/kg body weight) to better outline the vascular lumen. The blood pressure and heart rate of animals before injection and after injection were monitored. Subsequently, aortas were removed for histological analysis. Serial frozen sections were stained for the presence of VCAM-1 (green), nuclei (4',6-diamidino-2-phenylindole [DAPI]; blue), or injected agent (red).
Results
Selection of Phage Display Library Clones
In cell culture, MCECs exhibit high constitutive expression of VCAM-1 compared with strain-matched MLECs,19 making this an ideal target for positive selection of VCAM-1eCtargeted phage. Phage selection (4 rounds) resulted in a 330-fold increase in the ratio of cells internalized to extracellular bound phages (Figure 1A). Sixty clones were sequenced and screened for VCAM-1 binding. Phage clone 9 exhibited highest binding to VCAM-1 (Figure 1B). The peptide sequences of VCAM-1eCbinding clones were used to search the online National Center for Biotechnology Information database using the basic local alignment search tool (BLAST) algorithm (Table). Interestingly, clones 9 and 20 share sequence homology to the -chain of VLA4, the VCAM-1 ligand expressed on circulating hematopoietic cells.20 We synthesized the corresponding peptide (termed VP) for clone 9, "CVHSPNKKC," incorporating a C-terminal GGSKGK peptide extension for fluorescein attachment and NP conjugation.
Specificity of VCAM-1eCTargeting VP
Fluorescent VP was rapidly internalized by target MCECs when incubated at 37°C (but not at 4°C), showing punctate intracellular staining (Figure 2A). Confocal microscopy of MCECs incubated with VP (green) and stained for the presence of VCAM-1 (red) demonstrated intracellular accumulation of VP (Figure 2C). Unlike VP, the scrambled CP did not bind to MCECs, confirming the specificity of the peptide (Figure 2B). When cells were preincubated with FceCVCAM-1, no binding of either VP (Figure 2D) or CP (Figure 2E) could be observed. In subsequent quantitative flow cytometry analysis, >95% of MCECs were labeled with VP and cellular uptake correlated with VCAM-1 expression (Figure 2F). Finally, preincubation of VP with FceCVCAM-1 (Figure 2G) significantly inhibited VP binding and uptake. In contrast, preincubation of VP with FceCICAM-1 did not affect VP binding to MCECs (Figure 2G.) Together, these data indicate that VP is specifically internalized by MCECs via VCAM-1.
In functional analysis, preincubation of TNF-eCactivated MCECs with either function-blocking monoclonal Ab (mAb) M/K 2.7 or VP significantly inhibited mononuclear cell recruitment under flow conditions. VP reduced mononuclear cell accumulation by 68%, compared with 52%, via mAb blocking, suggesting that the VP is as effective at preventing VCAM-1eCmediated leukocyteeCendothelial interactions (Figure 2H). These data suggest that VP is effective at blocking leukocyteeCendothelial interactions and the mechanism may be attributable to binding of the peptide to the VLA4 adhesion domain of VCAM-1 or may be the result of cell surface depletion of VCAM-1 by peptide binding.
Targeting of Magnetofluorescent NPs to VCAM-1
Whereas Abs have been used to image VCAM-1 expression, the modest target-to-background ratios obtained using these agents often limit their use in vivo for cardiovascular imaging. To test the comparative amplification potentially afforded by cell-internalizing affinity ligands versus cell surface labeling with specific mAb, we directly compared FITC-labeled VCAM-1 and VP by FACS analysis using equimolar concentrations. VP demonstrated a >12-fold higher uptake compared with the Ab (Figure 2I).
To permit evaluation of VCAM-1 expression via fluorescence or MRI and to develop a VCAM-1eCtargeted imaging agent, we coupled VP multivalently to a superparamagnetic fluorescent NP termed VNP. In vitro flow cytometric analysis demonstrated that VNP retained the ability of the peptide to bind specifically to MCECs (Figure 3A) and in fact showed an 10-fold higher uptake by target cells compared with uptake of the negative control particle (CNP).
Because an NP-based in vivo imaging agent will contact multiple cell types, we also sought to determine the specificity of VNP for endothelial cells as opposed to macrophages, which have been shown to accumulate iron oxide NPs. In vitro flow cytometric analysis demonstrated that VNP preferentially accumulated in endothelial cells (Figure 3B). Further, VNP had an 11-fold higher accumulation in endothelial cells compared with its accumulation in macrophages, consistent with the known pattern of VCAM-1 expression (Figure 3B). In contrast, endothelial cells lacked NP accumulation, suggesting VHS peptide labeling of NP changes the cellular specificity of the NP.
Intravital Confocal Microscopy of VCAM-1eCExpressing Microvasculature
To create a biologically monovariant in vivo mouse model, we next induced acute inflammation (24 hours) in mouse ears by injecting TNF- to upregulate VCAM-1.12 When VP was injected intravenously into this mouse model of inflammation, it exhibited very fast kinetics (as determined via intravital confocal microscopy), with a vascular T1/2 of 110 seconds, rapid extravasation into the interstitium, and near-complete renal clearance within 10 minutes. In contrast, injection of VNP, CNP, or NP resulted in no apparent extravasation of the agents within 20 minutes after injection, as observed via intravital confocal microscopy. VNP, CNP, and NP were cleared from the circulation within 4 hours after injection, with only VNP remaining bound to the vasculature. At 4 hours after VNP injection, near-infrared fluorescence was highly elevated in the vasculature of the inflamed ear, whereas no significant background fluorescence was observed (Figure 4A). In the control ear of the same animal, there was little detectable signal (Figure 4B). In control animals injected with CNP, an initial fluorescence signal throughout the circulation immediately after administration was observed, but no significant fluorescence in either ear after 4 hours was determined. Control animals that received NP alone had cleared the agent from the circulation by 4 hours after injection (data not shown). At 24 hours, significant amounts of VNP could still be visualized in the vessel walls of mTNF-eCtreated ears (Figure 4C and 4E), whereas it was absent in the untreated ears (Figure 4D). In further time-lapse analysis (Figure 4F), the fluorescence profile of VNP did not change during a 5-minute observation, confirming that the probe was indeed retained at specific locations within the vessel wall and not free flowing in the circulation.
Histological analysis of mTNF-eCtreated ears (Figure 5A through 5F) revealed the upregulation of VCAM-1 in endothelial cells (Figure 5B and 5E). In contrast, endothelial cells from untreated ears did not show appreciable VCAM-1 staining (Figure 5H). Anti-FITC Ab staining revealed the presence of the VNP agent in endothelial cells from inflamed ears (Figure 5C) but not in control ears (Figure 5I). Endothelial cells from CNP-injected animals were also devoid of anti-FITC staining (Figure 5F), confirming the specificity of VNP. Together, these data confirm the in vivo delivery and accumulation of VNP, but not CNP or NP, in the endothelium of acutely inflamed tissue.
MRI of Atherosclerotic Lesions in ApoEeC/eC Mice
To determine whether VCAM-1 expression could be detected in vivo in atherosclerotic lesions, we injected VNP into cholesterol-fed apoEeC/eC mice. In this model, there is extensive neovascularization and VCAM-1 expression in vast areas throughout the lesion5 (Figure 6G). Injection of agents did not result in an increase of blood pressure or heart rate. After intravenous administration of VNP, we detected extensive signal intensity decreases associated with iron oxide NP accumulation in atherosclerotic lesions (Figure 6A and 6C), in particular in the aortic arch and around the take-off of large vessels. Using a spin echo sequence with a TE of 20 milliseconds, the signal-to-noise ratio of plaques decreased from 20.1±1.1 to 9.1±0.9 (aorta/muscle ratio before VNP administration 0.82 versus post-VNP administration 0.35). In areas where lesion development caused crescent-shaped aortic wall thickening, postadministration images showed darkening of the affected aortic wall (Figure 6B and 6C). These in vivo changes in magnetic resonance (MR) signal correlated with ex vivo MRI and macroscopic fluorescence imaging of the excised aortas (Figure 6D and 6E). NP administered to apoEeC/eC animals at equimolar amounts failed to accumulate in the aorta (Figure 7). Further, injection of VNP into C57BL/6 animals (wild type) also did not accumulate in a significant amount in the aorta (Figure 7). To confirm the presence of atherosclerotic lesions and to corroborate VNP colocalization with VCAM-1 expression, we performed histological evaluation (Figure 6F) and multifluorescence immunohistochemistry. Fluorescence microscopy confirmed extensive colocalization of VNP (Cy.5.5 channel; Figure 6H) and VCAM-1 expression (antieCVCAM-1eCFITC channel (Figure 6G). In contrast, there was little accumulation of VNP in macrophages, reminiscent of the in vitro data shown in Figure 3B.
Discussion
Pathophysiologically inducible endothelial targets provide an attractive paradigm for the development of targeted and disease-specific molecular imaging agents, attributable largely to their immediate accessibility via the circulation. The strict temporal and spatial regulation of VCAM-1 in the development and progression of atherosclerosis5eC9 makes it a prime candidate for development of the next generation of targeted diagnostics and therapeutics.
Several approaches have been described previously using radiolabeled Abs to detect VCAM-1 expression;12,13 however, this approach has resulted in very modest target-to-background ratios, limiting its use for in vivo cardiovascular imaging. To overcome these limitations, we commenced development of an agent that not only decorated the endothelial surface but was internalized by endothelial cells ("biological amplification through intracellular trapping").21 We used phage display to identify peptide sequences that bound under physiological flow conditions and were capable of VCAM-1eCmediated endothelial cell internalization. Using this approach, we identified a novel VCAM-1eCspecific cell internalizing peptide that results in higher target-to-background ratios than monoclonal Abs and allows in vivo imaging of inflammation and cardiovascular disease.
Our unique approach resulted in the identification of a small panel of peptides forming 42 consensus groups, with a subset that bound specifically to the extracellular domain of VCAM-1 (Table). One of these peptides, VHSPNKK, with the highest binding to VCAM-1, was homologous to VLA4, the ligand for VCAM-1. Direct intravenous injection of the fluorescently labeled selected peptide (VP), resulted in rapid leakage into the interstitium, and the resulting high background fluorescence in the surrounding tissue eliminated the possibility of using the ligand for intravital work, although it remains a highly specific and valuable agent for in vitro assays. The derived multivalent (chemical amplification) magnetofluorescent imaging agent (VNP) exhibited improved in vivo pharmacokinetics, retained specificity for VCAM-1, and in vivo accumulated in the vessel wall, remaining detectable for at least 24 hours by noninvasive optical and MRI modalities, confirming the ability of the peptide to shuttle large molecules across the plasma membrane.
A number of previous publications have identified peptide sequences that bind to atherosclerotic plaque or are potentially relevant to targeting cell surface receptors.22eC25 In particular, Liu et al25 used phage display to interrogate the surface of atherosclerotic lesions in apoEeC/eC mice and identified consensus sequences that bound to endothelial cell surface glucose-regulated protein 78 and a homologue of tissue inhibitor of metalloproteinase-2 (TIMP-2) that bound to mature lesions in vivo. In the present study, we identified three peptide sequences with similarity to TIMP-2; however, no other similar sequences were isolated. The difference in these results could be explained by the fact that Liu et al examined the binding of phage in vivo in apoEeC/eC mice with advanced lesion development,25 where proteins other than VCAM-1 may be more abundant on the endothelial cell surface.
The data presented here demonstrate for the first time that endothelial targets can be visualized by in vivo MRI without the use of monoclonal Abs. Our approach has far-reaching implications in the design of future imaging agents by allowing the visualization of nonabundant targets through intracellular accumulation. The first generation of nontargeted magnetic NPs are in clinical use,26 and thus, it is conceivable that targeted NPs could provide significant benefits in specificity for imaging and therapeutics. Our in vitro and in vivo results suggest that conjugation of the magnetofluorescent NP with phage displayeCderived VCAM-1eCspecific peptides changes the cellular specificity of the parent NP and allows the targeted imaging of activated endothelium. Indeed, we demonstrate that VCAM-1 expression can be visualized by high-resolution intravital confocal microscopy and MRI. This could have a number of important clinical implications. For example, VNP could be used to noninvasively monitor the efficacy of anti-inflammatory or lipid-lowering drugs by examining the level of VCAM-1 in vessels. Furthermore, the agent or similarly developed probes could be used for combined or simultaneous imaging and targeted drug delivery. An intriguing application would be to use the VP domain to shuttle larger therapeutic proteins to sites of inflammation in a targeted fashion. For example, targeting of doxorubicin to angiogenic cells through peptide conjugation increases the efficacy-to-toxicity ratio by 14-fold.27 Ultimately, the developed molecular framework could be applied to other targets, making it a useful scaffold for the development of novel targeted therapeutics or diagnostic tools.
Acknowledgments
This work was supported by a grant from the Donald W. Reynolds Foundation Cardiovascular Clinical Research Program (R.W. and J.R.A.), National Institutes of Health (NIH) grants CA86355 and CA92782 (R.W.) and EB000662 (L.J.), and NIH training grant CA09362 (K.A.K.). The authors would like to thank Dr Peter Libby, Brigham and Women’s Hospital, for critical review and many helpful discussions, Pratik Patel for assistance with confocal imaging and 3D reconstructions, Dr Elena Aikawa for help with immunohistochemistry, Dr Frederick Reynolds for assistance with peptide synthesis, Drs Farouc Jaffer and Umar Mahmood for MR pulse sequence optimizations, Dr Alex Bogdanov for Gd-PGC, and Dr Matthias Nahendorf MD for aorta dissection.
These authors contributed equally to this work.
References
Libby P. Inflammation in atherosclerosis. Nature. 2002; 420: 868eC874.
Libby P, Ridker PM. Inflammation and atherosclerosis: role of C-reactive protein in risk assessment. Am J Med. 2004; 116 (suppl 6)A: 9SeC16S.
Libby P. The vascular biology of atherosclerosis. In Brunwald E, Zipes DP, Libby P, eds. Heart Disease: A Textbook of Cardiovascular Medicine, 6th Edition. Philadelphia, PA: WB Saunders; 2001: 995eC1009.
Stary HC, Chandler AB, Glagov S, Guyton JR, Insull W Jr, Rosenfeld ME, Schaffer SA, Schwartz CJ, Wagner WD, Wissler RW. A definition of initial, fatty streak and intermediate lesions of atherosclerosis. A report from the Committee on Vascular Lesions of the Council on Arteriosclerosis, American Heart Association. Circulation. 1994; 89: 2462eC2478.
Iiyama K, Hajra L, Iiyama M, Li H, DiChiara M, Medoff BD, Cybulsky MI. Patterns of vascular cell adhesion molecule-1 and intercellular adhesion molecule-1 expression in rabbit and mouse atherosclerotic lesions and at sites predisposed to lesion formation. Circ Res. 1999; 85: 199eC207.
Cybulsky MI, Gimbrone MA Jr. Endothelial expression of a mononuclear leukocyte adhesion molecule during atherogenesis. Science. 1991; 251: 788eC791.
Orosz CG. Endothelial activation and chronic allograft rejection. Clin Transplant. 1994; 8: 299eC303.
Koskinen PK, Lemstrom KB. Adhesion molecule P-selectin and vascular cell adhesion molecule-1 in enhanced heart allograft arteriosclerosis in the rat. Circulation. 1997; 95: 191eC196.
Cybulsky MI, Iiyama K, Li H, Zhu S, Chen M, Iiyama M, Davis V, Gutierrez-Ramos JC, Connelly PW, Milstone DS. A major role for VCAM-1, but not ICAM-1, in early atherosclerosis. J Clin Invest. 2001; 107: 1255eC1262.
Davies MJ, Gordon JL, Gearing AJ, Pigott R, Woolf N, Katz D, Kyriakopoulos A. The expression of the adhesion molecules ICAM-1, VCAM-1, PECAM, and E-selectin in human atherosclerosis. J Pathol. 1993; 171: 223eC229.
O’Brien KD, Allen MD, McDonald TO, Chait A, Harlan JM, Fishbein D, McCarty J, Ferguson M, Hudkins K, Benjamin CD, et al. Vascular cell adhesion molecule-1 is expressed in human coronary atherosclerotic plaques. Implications for the mode of progression of advanced coronary atherosclerosis. J Clin Invest. 1993; 92: 945eC951.
McHale JF, Harari OA, Marshall D, Haskard DO. Vascular endothelial cell expression of ICAM-1 and VCAM-1 at the onset of eliciting contact hypersensitivity in mice: evidence for a dominant role of TNF-alpha. J Immunol. 1999; 162: 1648eC1655.
Sans M, Fuster D, Vazquez A, Setoain FJ, Piera C, Pique JM, Panes J. 123Iodine-labeled anti-VCAM-1 antibody scintography in the assessment of experimental colitis. Eur J Gastroenterol Hepatol. 2001; 13: 31eC38.
Allport JR, Lim Y-C, Shipley JM, Senior RM, Shapiro SD, Matsuyoshi N, Vestweber D, Luscinskas FW. Neutrophils from MMP-9- or neutrophil elastase-deficient mice show no defect in transendothelial migration under flow in vitro. J Leukoc Biol. 2002; 71: 821eC828.
Shen J, Luscinskas FW, Connolly A, Dewey CF Jr, Gimbrone MA Jr. Fluid shear stress modulates cytosolic free calcium in vascular endothelial cells. Am J Physiol. 1992; 262: C384eCC390.
Smith GP, Petrenko VA. Phage display. Chem Rev. 1997; 97: 391eC410.
Ricard I, Payet MD, Dupuis G. VCAM-1 is internalized by a clathrin-related pathway in human endothelial cells but its alpha 4 beta 1 integrin counter-receptor remains associated with the plasma membrane in human T lymphocytes. Eur J Immunol. 1998; 28: 1708eC1718.
Lewin M, Bredow S, Sergeyev N, Marecos E, Bogdanov A Jr, Weissleder R. In vivo assessment of vascular endothelial growth factor-induced angiogenesis. Int J Cancer. 1999; 83: 798eC802.
Lim Y-C, Garcia-Cardena G, Allport JR, Zervoglos M, Connolly AJ, Gimbrone MA Jr, Luscinskas FW. Heterogeneity of endothelial cells from different organ sites in T-cell subset recruitment. Am J Pathol. 2003; 162: 1591eC1601.
Holzmann B, Weissman IL. Integrin molecules involved in lymphocyte homing to Peyer’s patches. Immunol Rev. 1989; 108: 45eC61.
Weissleder R, Ntziachristos V. Shedding light onto live molecular targets. Nat Med. 2003; 9: 123eC128.
Molenaar TJ, Twisk J, de Haas SA, Peterse N, Vogelaar BJ, van Leeuwen SH, Michon IN, van Berketl TJ, Kuiper J, Biessen EA. P-selectin as a candidate target in atherosclerosis. Biochem Pharmacol. 2003; 66: 859eC866.
Houston P, Goodman J, Lewis A, Campbell CJ, Braddock M. Homing markers for atherosclerosis: applications for drug delivery, gene delivery and vascular imaging. FEBS Lett. 2001; 492: 73eC77.
Winter PM, Morawski AM, Caruthers SD, Fuhrhop RW, Zhang H, Williams TA, Allen JS, Lacy EK, Robertson JD, Lanza GM, Wickline SA. Molecular imaging of angiogenesis in early-stage atherosclerosis with alpha(v)beta3-integrin-targeted nanoparticles. Circulation. 2003; 108: 2270eC2274.
Liu C, Bhattacharjee G, Boisvert W, Dilley R, Edgington T. In vivo interrogation of the molecular display of atherosclerotic lesion surfaces. Am J Pathol. 2003; 163: 1859eC1871.
Harisinghani MG, Barentsz J, Hahn PF, Deserno WM, Tabatabaei S, van de Kaa CH, de la Rosette J, Weissleder R. Noninvasive detection of clinically occult lymph-node metastases in prostate cancer. N Engl J Med. 2003; 348: 2491eC2499.
Arap W, Pasqualini R, Ruoslahti E. Cancer treatment by targeted drug delivery to tumor vasculature in a mouse model. Science. 1998; 279: 377eC380.(Kimberly A. Kelly, Jennif)
Abstract
Endothelial vascular adhesion molecule-1 (VCAM-1) is a critical component of the leukocyteeCendothelial adhesion cascade, and its strict temporal and spatial regulation make it an ideal target for imaging and therapy. The goal of this study was to develop novel VCAM-1eCtargeted imaging agents detectable by MRI and fluorescence imaging using phage displayeCderived peptide sequences and multimodal nanoparticles (NPs). We hypothesized that VCAM-1eCmediated cell internalization of phage displayeCselected peptides could be harnessed as an amplification strategy to chaperone and trap imaging agents inside VCAM-1eCexpressing cells, thus improving target-to-background ratios. To accomplish our goal, iterative phage display was performed on murine endothelium under physiological flow conditions to identify a family of VCAM-1eCmediated cell-internalizing peptides. One specific sequence, containing the VHSPNKK motif that has homology to the -chain of very late antigen (a known ligand for VCAM-1), was shown to bind VCAM-1 and block leukocyteeCendothelial interactions. Compared with VCAM-1 monoclonal antibody, the peptide showed 12-fold higher target-to-background ratios. A VHSPNKK-modified magnetofluorescent NP (VNP) showed high affinity for endothelial cells expressing VCAM-1 but surprisingly low affinity for macrophages. In contrast, a control NP without VCAM-1eCtargeting sequences showed no affinity for endothelial cells. In vivo, VNP successfully identified VCAM-1eCexpressing endothelial cells in a murine tumor necrosis factor-eCinduced inflammatory model and colocalized with VCAM-1eCexpressing cells in atherosclerotic lesions present in cholesterol-fed apolipoprotein E apoEeC/eC mice. These results indicate that: (1) small peptide sequences can significantly alter targeting of NPs, (2) the used amplification strategy of internalization results in high target-to-background ratios, and (3) this technology is useful for in vivo imaging of endothelial markers.
Key Words: cardiovascular diseases cell adhesion molecules imaging inflammation vasculature
Introduction
Cardiovascular diseases are the leading cause of death in the developed world and represent an immense clinical burden. Atherosclerosis is the major contributor to the pathogenesis of acute myocardial and cerebral thrombosis,1 with 500 000 deaths annually in the United States alone. Although significant advances have been made in the primary prevention of cardiovascular disease, risk stratification remains suboptimal in large segments of the population. It is now widely accepted that the degree of coronary artery stenosis is not predictive of the risk of plaque rupture and subsequent thrombosis at that location. Clinical risk scores, such as that derived from the Framingham Heart Study, are useful but may lose predictive value in the large segments of the population at intermediate risk.2 Despite the addition of new measurements to stratify clinical risk in individuals at intermediate risk, such as serum C-reactive protein or the detection of coronary calcification by computed tomography, there remains a need for novel molecular imaging probes to better define cardiovascular risk and guide therapy more rationally.
Atherosclerosis is a chronic, progressive inflammatory disease characterized by a specific series of cellular and molecular events,3 from early inflammatory lesions ("fatty streaks"), comprised of monocyte-derived macrophages and T lymphocytes,4 to development of mature plaque and potentially rupture and thrombosis. The earliest molecular changes occur on endothelial surfaces of the aorta and vasa vasorum and are key contributors to the initiation, progression, and thrombotic complications of atherosclerosis.1 Although lipid-lowering therapies have demonstrated direct clinical benefit, there remains a dramatic need for novel diagnostic tools to noninvasively assess the benefit of therapies and identify early molecular changes associated with atherosclerosis, and for new targeted and specific therapies.
The strict temporal and spatial regulation of inducible endothelial adhesion molecules and their critical function in atherosclerosis makes them ideal targets for the next generation of diagnostics and therapeutics. Vascular adhesion molecule-1 (VCAM-1) is upregulated on the endothelium under inflammatory conditions including atherosclerosis and cardiac allograft rejection5eC8 and appears on the endothelial cell surface of atheroprone areas before the onset of visible disease.6 Its importance in the initiation of atherogenesis is demonstrated by delayed lesion development in mice carrying a mutation that hinders VCAM-1 function.9 Comparable to mouse atherosclerosis, VCAM-1 expression is induced early in human atheroma and is an important element in the inflammatory component of atherosclerosis, contributing to monocyte and lymphocyte recruitment from adventitial vessels and the arterial lumen.3,10,11 Although several approaches have been described to image VCAM-1 expression using radiolabeled antibodies (Abs),12,13 these agents usually result in modest target-to-background ratios, limiting their use for in vivo cardiovascular imaging.
We determined that additional amplification steps would be required to image VCAM-1 expression directly in the endothelium. Thus, we developed a modified phage display approach to screen for specific peptide sequences that would be internalized by VCAM-1eCexpressing endothelial cells under physiological flow conditions. One specific sequence (VHSPNKK), with homology to the -chain of very late antigen (VLA4; a known ligand for VCAM-1), was shown to bind VCAM-1 and to block leukocyteeCendothelial interactions. A derived multivalent nanoparticle (NP) carrying the VHSPNKK peptide successfully identified VCAM-1eCexpressing endothelial cells in a murine tumor necrosis factor- (TNF-)eCinduced inflammatory model via confocal microscopy, and furthermore, it allowed in vivo detection of atherosclerotic lesions in apolipoprotein EeC/eC (apoEeC/eC) mice via MRI.
Materials and Methods
Antibodies
Anti-mouse CD31 (clone MEC 13.3), anti-CD16/CD32 (FcR block), anti-mouse CD106 (VCAM-1, clone 429), and anti-mouse CD106-fluorescein isothiocyanate (FITC) VCAM-1, clone 429) were from BD Pharmingen (San Diego, Calif), and anti-mouse VCAM-1 (clone M/K 2.7) was from American Type Culture Collection (Manassas, Va). Biotinylated anti-rat and anti-rabbit IgG (heavy and light chain [H+L]) Texas Red or Cy3 anti-rat IgG (H+L) were from Jackson ImmunoResearch (West Grove, Pa), and anti-FITC/Oregon GreeneChorseradish peroxidase (HRP) was from Molecular Probes (Eugene, Ore). BiotineCanti-M13 rabbit IgG was from AbCam, and goat F(ab')2 anti-rat IgG (H+L)-phycoerythrein and Texas Red were from Caltag Laboratories (Burlingame, Calif).
Mice
C57BL/6 wild-type mice (7 to 9 weeks of age) from Taconic and apoEeC/eC (The Jackson Laboratory) were maintained in approved pathogen-free institutional housing facilities. All experiments were performed according to institutional guidelines. Animals were anesthetized for imaging procedures using isoflurane inhalation (1% to 2% isoflurane; 1 L O2). Animals were euthanized by CO2 asphyxiation as approved by the panel on euthanasia at the American Veterinary Association.
Endothelial Cell Culture
Murine cardiac endothelial cells (MCECs) or murine lung endothelial cells (MLECs) were isolated using previously published methods14 and used at passages 1 to 3.
Phage-Display Library Selection
The parallel plate-flow chamber has been described in detail.15 Smith and Petrenko produced a detailed rationale of phage display selection and negative selection or depletion procedures.16 Phage selection and negative depletion were performed using strain-matched MCECs and MLECs. 1010 PFU of phage, displaying random disulfide-constrained 7-aa peptides (C7C-PhD; New England Biolabs), were drawn across MCEC monolayers at 37°C for 10 minutes at 0.52 mL/min (estimated wall shear stress 1.0 dyne/cm2). MCECs were then incubated at 37°C for an additional 15 minutes to allow time for VCAM-1eCmediated internalization (t1/2 11.2 minutes).17 Extracellular-restricted phage were removed with 0.2 mol/L glycine, pH 2.2 (3x8 minutes). Internalized phage were recovered by lysis with 0.1% triethanolamine (Sigma) in PBS, pH 7.4, (4 minutes; room temperature). Extracts were neutralized with 500 e蘈 of 0.5 mol/L Tris-HCl, pH 9.0. To deplete phage that bound to identical markers present on MCECs and MLECs, the phage pool isolated after one round of selection was subtracted by three rounds of successive incubation at 37°C for 30 minutes with confluent monolayers of strain-matched MLECs that do not express VCAM-1. The phage that were internalized by MCECs but not MLECs were amplified by Escherichia coli, titered, subjected to three additional rounds of positive selection, and individual clones selected for ELISA assay and sequencing.
ELISA Assay of VCAM-1 Binding
Protein A (20 e/mL) on Nunc Maxisorp plates (Fisher; 4°C overnight) was incubated sequentially at room temperature with FceCVCAM-1 (10 e/mL; 1 hour) and phage clones (1010 PFU; 1 hour), washed with PBS containing 0.1% Tween-20, incubated with biotinylated anti-M13 Ab (1:40; 1 hour), detected with streptavidin-HRP (1:500), and developed with tetra methyl benzidine, and absorbance650 was determined (Emax; Molecular Devices).
Synthesis of FITC Peptides and Fluorescent Magnetic NP Conjugates
The identified peptide sequence "CVHSPNKKC" and corresponding magnetofluorescent NPs were synthesized as described in the online data supplement (available at http://circres.ahajournals.org). The following compounds were synthesized in bulk: VCAM-1 peptide (VP): CVHSPNKKCGGSK(FITC)GK; control peptide (CP): CPKNVSKHCGGSK(FITC)GK; VCAM-1 NP (VNP): CVHSPNKK-CGGSK(FITC)GK(CLIO-Cy5.5); control NP (CNP): CPKNVSK-HCGGSK(FITC)GK(CLIO-Cy5.5); and CLIO-Cy5.5 (NP): CLIO-Cy5.5 backbone (no peptide).
Fluorescence Microscopy and Flow Cytometry
MCECs were incubated with 1 eol/L of either VP or CP (1 hour; 37°C), washed 3x with PBS, and visualized by fluorescence microscopy (Nikon Eclipse TE2000-S; Insight QE; x40 objective). For confocal microscopy, MCECs were incubated with 1 eol/L of VP as above, fixed with 2% paraformaldehyde, stained with antieCVCAM-1 (10 e/mL; 1 hour), and detected with Texas RedeCconjugated secondary Ab (1:100; 30 minutes). Stained cells were analyzed on an LSM 5 PASCAL confocal microscope (Zeiss). For competition experiments, peptides were preincubated (30 minutes) with 5x murine FceCVCAM-1 or FceCICAM-1 (R & D Systems). MCECs were then detached, stained for cell surface VCAM-1, and analyzed via flow cytometry (10 000 cells/sample) on a Becton Dickinson FACSCalibur. For comparison of VP with antieCVCAM-1eCFITC, cells were incubated with equimolar concentrations (0.5 eol/L) of VP or antieCVCAM-1eCFITC for 2 hours at 37°C and analyzed by flow cytometry.
In Vitro Blocking Studies Under Flow Conditions
MCECs were activated for 5 hours with murine TNF- (mTNF-; R & D Systems) and incubated for 2 hours with either 25 e/mL M/K 2.7 antieCVCAM-1 Ab, 100 eol/L VP, or vehicle control. Strain-matched mononuclear cells (1x106 cells/mL in DPBS+eC0.1% BSA) were drawn across MCECs for 5 minutes at 0.52 mL/min. LeukocyteeCendothelial interactions were determined from 6 to 8 high-power fields at 5 minutes.
Targeting of VNP to Endothelial Cells
The isolation of primary MCECs was performed as described previously. Primary mouse macrophages were isolated as presented in the online data supplement. Cells were incubated with either VNP or NP (4.5 eol/L Cy5.5; 4 hours; 37°C) and analyzed by flow cytometry as described previously.
Immunohistochemistry and Intravital Confocal Microscopy
C57BL/6 mice (n=10) were injected subcutaneously in the right ear with 5 ng/50 e蘈 mTNF- in normal saline. After 24 hours, animals were anesthetized via inhaled isofluorane and injected intravenously (tail vein) with 10 nmol/L fluorochrome of VP (n=2), VNP (n=5), or CNP (n=3). Perfused vessels were located using bright-field illumination and fluorescence signal because of circulating peptide/conjugate. Intravital confocal microscopy was performed using a Nikon EF600N equipped with a BioRad Radiance 2100 confocal and x40 water immersion objective at 0 hours, 4 hours, and 24 hours after agent administration. Image acquisition was performed simultaneously in two fluorescence channels (photo multiplier tube [PMT]1: peptide-FITC; PMT3: peptide-NP) in several fields from both ears of each animal using built-in compensation routines to exclude channel bleed-through (Laser Sharp software; BioRad), and 3D reconstructions were performed using Amira (TGS). Subsequently, ears were removed for histological analysis. Adjacent serial frozen sections were stained for the presence of CD31, VCAM-1, or injected agent. Digital images were taken using a Nikon Eclipse E400 upright microscope (x40 objective) equipped with an Insight color camera.
MRI of VCAM-1 Expression
We performed in vivo MRI of 1-year-old apoEeC/eC mice with extensive lesion burden that had been cholesterol-fed for 3 months (n=3). T1- (repetition time [TR] 50; echo time [TE] 2; 20 number of excitations [NEX] 10 minutes) and T2-weighted (TR 2000; TE 20 to 200; NEX 6 minutes) spin echo images and gradient echo sequences were acquired before and 24 hours after intravenous administration of VNP or NP to apoEeC/eC mice or VNP to C57BL/6 mice (5 mg Fe/kg body weight). In addition, we obtained one bright-blood angiographic sequence (TR 50; TE 2; 20 NEX 10 minutes) at the end of the imaging session using gadolinium-protected graph copolymer (Gd-PGC)18 (0.02 mmol/L Gd/kg body weight) to better outline the vascular lumen. The blood pressure and heart rate of animals before injection and after injection were monitored. Subsequently, aortas were removed for histological analysis. Serial frozen sections were stained for the presence of VCAM-1 (green), nuclei (4',6-diamidino-2-phenylindole [DAPI]; blue), or injected agent (red).
Results
Selection of Phage Display Library Clones
In cell culture, MCECs exhibit high constitutive expression of VCAM-1 compared with strain-matched MLECs,19 making this an ideal target for positive selection of VCAM-1eCtargeted phage. Phage selection (4 rounds) resulted in a 330-fold increase in the ratio of cells internalized to extracellular bound phages (Figure 1A). Sixty clones were sequenced and screened for VCAM-1 binding. Phage clone 9 exhibited highest binding to VCAM-1 (Figure 1B). The peptide sequences of VCAM-1eCbinding clones were used to search the online National Center for Biotechnology Information database using the basic local alignment search tool (BLAST) algorithm (Table). Interestingly, clones 9 and 20 share sequence homology to the -chain of VLA4, the VCAM-1 ligand expressed on circulating hematopoietic cells.20 We synthesized the corresponding peptide (termed VP) for clone 9, "CVHSPNKKC," incorporating a C-terminal GGSKGK peptide extension for fluorescein attachment and NP conjugation.
Specificity of VCAM-1eCTargeting VP
Fluorescent VP was rapidly internalized by target MCECs when incubated at 37°C (but not at 4°C), showing punctate intracellular staining (Figure 2A). Confocal microscopy of MCECs incubated with VP (green) and stained for the presence of VCAM-1 (red) demonstrated intracellular accumulation of VP (Figure 2C). Unlike VP, the scrambled CP did not bind to MCECs, confirming the specificity of the peptide (Figure 2B). When cells were preincubated with FceCVCAM-1, no binding of either VP (Figure 2D) or CP (Figure 2E) could be observed. In subsequent quantitative flow cytometry analysis, >95% of MCECs were labeled with VP and cellular uptake correlated with VCAM-1 expression (Figure 2F). Finally, preincubation of VP with FceCVCAM-1 (Figure 2G) significantly inhibited VP binding and uptake. In contrast, preincubation of VP with FceCICAM-1 did not affect VP binding to MCECs (Figure 2G.) Together, these data indicate that VP is specifically internalized by MCECs via VCAM-1.
In functional analysis, preincubation of TNF-eCactivated MCECs with either function-blocking monoclonal Ab (mAb) M/K 2.7 or VP significantly inhibited mononuclear cell recruitment under flow conditions. VP reduced mononuclear cell accumulation by 68%, compared with 52%, via mAb blocking, suggesting that the VP is as effective at preventing VCAM-1eCmediated leukocyteeCendothelial interactions (Figure 2H). These data suggest that VP is effective at blocking leukocyteeCendothelial interactions and the mechanism may be attributable to binding of the peptide to the VLA4 adhesion domain of VCAM-1 or may be the result of cell surface depletion of VCAM-1 by peptide binding.
Targeting of Magnetofluorescent NPs to VCAM-1
Whereas Abs have been used to image VCAM-1 expression, the modest target-to-background ratios obtained using these agents often limit their use in vivo for cardiovascular imaging. To test the comparative amplification potentially afforded by cell-internalizing affinity ligands versus cell surface labeling with specific mAb, we directly compared FITC-labeled VCAM-1 and VP by FACS analysis using equimolar concentrations. VP demonstrated a >12-fold higher uptake compared with the Ab (Figure 2I).
To permit evaluation of VCAM-1 expression via fluorescence or MRI and to develop a VCAM-1eCtargeted imaging agent, we coupled VP multivalently to a superparamagnetic fluorescent NP termed VNP. In vitro flow cytometric analysis demonstrated that VNP retained the ability of the peptide to bind specifically to MCECs (Figure 3A) and in fact showed an 10-fold higher uptake by target cells compared with uptake of the negative control particle (CNP).
Because an NP-based in vivo imaging agent will contact multiple cell types, we also sought to determine the specificity of VNP for endothelial cells as opposed to macrophages, which have been shown to accumulate iron oxide NPs. In vitro flow cytometric analysis demonstrated that VNP preferentially accumulated in endothelial cells (Figure 3B). Further, VNP had an 11-fold higher accumulation in endothelial cells compared with its accumulation in macrophages, consistent with the known pattern of VCAM-1 expression (Figure 3B). In contrast, endothelial cells lacked NP accumulation, suggesting VHS peptide labeling of NP changes the cellular specificity of the NP.
Intravital Confocal Microscopy of VCAM-1eCExpressing Microvasculature
To create a biologically monovariant in vivo mouse model, we next induced acute inflammation (24 hours) in mouse ears by injecting TNF- to upregulate VCAM-1.12 When VP was injected intravenously into this mouse model of inflammation, it exhibited very fast kinetics (as determined via intravital confocal microscopy), with a vascular T1/2 of 110 seconds, rapid extravasation into the interstitium, and near-complete renal clearance within 10 minutes. In contrast, injection of VNP, CNP, or NP resulted in no apparent extravasation of the agents within 20 minutes after injection, as observed via intravital confocal microscopy. VNP, CNP, and NP were cleared from the circulation within 4 hours after injection, with only VNP remaining bound to the vasculature. At 4 hours after VNP injection, near-infrared fluorescence was highly elevated in the vasculature of the inflamed ear, whereas no significant background fluorescence was observed (Figure 4A). In the control ear of the same animal, there was little detectable signal (Figure 4B). In control animals injected with CNP, an initial fluorescence signal throughout the circulation immediately after administration was observed, but no significant fluorescence in either ear after 4 hours was determined. Control animals that received NP alone had cleared the agent from the circulation by 4 hours after injection (data not shown). At 24 hours, significant amounts of VNP could still be visualized in the vessel walls of mTNF-eCtreated ears (Figure 4C and 4E), whereas it was absent in the untreated ears (Figure 4D). In further time-lapse analysis (Figure 4F), the fluorescence profile of VNP did not change during a 5-minute observation, confirming that the probe was indeed retained at specific locations within the vessel wall and not free flowing in the circulation.
Histological analysis of mTNF-eCtreated ears (Figure 5A through 5F) revealed the upregulation of VCAM-1 in endothelial cells (Figure 5B and 5E). In contrast, endothelial cells from untreated ears did not show appreciable VCAM-1 staining (Figure 5H). Anti-FITC Ab staining revealed the presence of the VNP agent in endothelial cells from inflamed ears (Figure 5C) but not in control ears (Figure 5I). Endothelial cells from CNP-injected animals were also devoid of anti-FITC staining (Figure 5F), confirming the specificity of VNP. Together, these data confirm the in vivo delivery and accumulation of VNP, but not CNP or NP, in the endothelium of acutely inflamed tissue.
MRI of Atherosclerotic Lesions in ApoEeC/eC Mice
To determine whether VCAM-1 expression could be detected in vivo in atherosclerotic lesions, we injected VNP into cholesterol-fed apoEeC/eC mice. In this model, there is extensive neovascularization and VCAM-1 expression in vast areas throughout the lesion5 (Figure 6G). Injection of agents did not result in an increase of blood pressure or heart rate. After intravenous administration of VNP, we detected extensive signal intensity decreases associated with iron oxide NP accumulation in atherosclerotic lesions (Figure 6A and 6C), in particular in the aortic arch and around the take-off of large vessels. Using a spin echo sequence with a TE of 20 milliseconds, the signal-to-noise ratio of plaques decreased from 20.1±1.1 to 9.1±0.9 (aorta/muscle ratio before VNP administration 0.82 versus post-VNP administration 0.35). In areas where lesion development caused crescent-shaped aortic wall thickening, postadministration images showed darkening of the affected aortic wall (Figure 6B and 6C). These in vivo changes in magnetic resonance (MR) signal correlated with ex vivo MRI and macroscopic fluorescence imaging of the excised aortas (Figure 6D and 6E). NP administered to apoEeC/eC animals at equimolar amounts failed to accumulate in the aorta (Figure 7). Further, injection of VNP into C57BL/6 animals (wild type) also did not accumulate in a significant amount in the aorta (Figure 7). To confirm the presence of atherosclerotic lesions and to corroborate VNP colocalization with VCAM-1 expression, we performed histological evaluation (Figure 6F) and multifluorescence immunohistochemistry. Fluorescence microscopy confirmed extensive colocalization of VNP (Cy.5.5 channel; Figure 6H) and VCAM-1 expression (antieCVCAM-1eCFITC channel (Figure 6G). In contrast, there was little accumulation of VNP in macrophages, reminiscent of the in vitro data shown in Figure 3B.
Discussion
Pathophysiologically inducible endothelial targets provide an attractive paradigm for the development of targeted and disease-specific molecular imaging agents, attributable largely to their immediate accessibility via the circulation. The strict temporal and spatial regulation of VCAM-1 in the development and progression of atherosclerosis5eC9 makes it a prime candidate for development of the next generation of targeted diagnostics and therapeutics.
Several approaches have been described previously using radiolabeled Abs to detect VCAM-1 expression;12,13 however, this approach has resulted in very modest target-to-background ratios, limiting its use for in vivo cardiovascular imaging. To overcome these limitations, we commenced development of an agent that not only decorated the endothelial surface but was internalized by endothelial cells ("biological amplification through intracellular trapping").21 We used phage display to identify peptide sequences that bound under physiological flow conditions and were capable of VCAM-1eCmediated endothelial cell internalization. Using this approach, we identified a novel VCAM-1eCspecific cell internalizing peptide that results in higher target-to-background ratios than monoclonal Abs and allows in vivo imaging of inflammation and cardiovascular disease.
Our unique approach resulted in the identification of a small panel of peptides forming 42 consensus groups, with a subset that bound specifically to the extracellular domain of VCAM-1 (Table). One of these peptides, VHSPNKK, with the highest binding to VCAM-1, was homologous to VLA4, the ligand for VCAM-1. Direct intravenous injection of the fluorescently labeled selected peptide (VP), resulted in rapid leakage into the interstitium, and the resulting high background fluorescence in the surrounding tissue eliminated the possibility of using the ligand for intravital work, although it remains a highly specific and valuable agent for in vitro assays. The derived multivalent (chemical amplification) magnetofluorescent imaging agent (VNP) exhibited improved in vivo pharmacokinetics, retained specificity for VCAM-1, and in vivo accumulated in the vessel wall, remaining detectable for at least 24 hours by noninvasive optical and MRI modalities, confirming the ability of the peptide to shuttle large molecules across the plasma membrane.
A number of previous publications have identified peptide sequences that bind to atherosclerotic plaque or are potentially relevant to targeting cell surface receptors.22eC25 In particular, Liu et al25 used phage display to interrogate the surface of atherosclerotic lesions in apoEeC/eC mice and identified consensus sequences that bound to endothelial cell surface glucose-regulated protein 78 and a homologue of tissue inhibitor of metalloproteinase-2 (TIMP-2) that bound to mature lesions in vivo. In the present study, we identified three peptide sequences with similarity to TIMP-2; however, no other similar sequences were isolated. The difference in these results could be explained by the fact that Liu et al examined the binding of phage in vivo in apoEeC/eC mice with advanced lesion development,25 where proteins other than VCAM-1 may be more abundant on the endothelial cell surface.
The data presented here demonstrate for the first time that endothelial targets can be visualized by in vivo MRI without the use of monoclonal Abs. Our approach has far-reaching implications in the design of future imaging agents by allowing the visualization of nonabundant targets through intracellular accumulation. The first generation of nontargeted magnetic NPs are in clinical use,26 and thus, it is conceivable that targeted NPs could provide significant benefits in specificity for imaging and therapeutics. Our in vitro and in vivo results suggest that conjugation of the magnetofluorescent NP with phage displayeCderived VCAM-1eCspecific peptides changes the cellular specificity of the parent NP and allows the targeted imaging of activated endothelium. Indeed, we demonstrate that VCAM-1 expression can be visualized by high-resolution intravital confocal microscopy and MRI. This could have a number of important clinical implications. For example, VNP could be used to noninvasively monitor the efficacy of anti-inflammatory or lipid-lowering drugs by examining the level of VCAM-1 in vessels. Furthermore, the agent or similarly developed probes could be used for combined or simultaneous imaging and targeted drug delivery. An intriguing application would be to use the VP domain to shuttle larger therapeutic proteins to sites of inflammation in a targeted fashion. For example, targeting of doxorubicin to angiogenic cells through peptide conjugation increases the efficacy-to-toxicity ratio by 14-fold.27 Ultimately, the developed molecular framework could be applied to other targets, making it a useful scaffold for the development of novel targeted therapeutics or diagnostic tools.
Acknowledgments
This work was supported by a grant from the Donald W. Reynolds Foundation Cardiovascular Clinical Research Program (R.W. and J.R.A.), National Institutes of Health (NIH) grants CA86355 and CA92782 (R.W.) and EB000662 (L.J.), and NIH training grant CA09362 (K.A.K.). The authors would like to thank Dr Peter Libby, Brigham and Women’s Hospital, for critical review and many helpful discussions, Pratik Patel for assistance with confocal imaging and 3D reconstructions, Dr Elena Aikawa for help with immunohistochemistry, Dr Frederick Reynolds for assistance with peptide synthesis, Drs Farouc Jaffer and Umar Mahmood for MR pulse sequence optimizations, Dr Alex Bogdanov for Gd-PGC, and Dr Matthias Nahendorf MD for aorta dissection.
These authors contributed equally to this work.
References
Libby P. Inflammation in atherosclerosis. Nature. 2002; 420: 868eC874.
Libby P, Ridker PM. Inflammation and atherosclerosis: role of C-reactive protein in risk assessment. Am J Med. 2004; 116 (suppl 6)A: 9SeC16S.
Libby P. The vascular biology of atherosclerosis. In Brunwald E, Zipes DP, Libby P, eds. Heart Disease: A Textbook of Cardiovascular Medicine, 6th Edition. Philadelphia, PA: WB Saunders; 2001: 995eC1009.
Stary HC, Chandler AB, Glagov S, Guyton JR, Insull W Jr, Rosenfeld ME, Schaffer SA, Schwartz CJ, Wagner WD, Wissler RW. A definition of initial, fatty streak and intermediate lesions of atherosclerosis. A report from the Committee on Vascular Lesions of the Council on Arteriosclerosis, American Heart Association. Circulation. 1994; 89: 2462eC2478.
Iiyama K, Hajra L, Iiyama M, Li H, DiChiara M, Medoff BD, Cybulsky MI. Patterns of vascular cell adhesion molecule-1 and intercellular adhesion molecule-1 expression in rabbit and mouse atherosclerotic lesions and at sites predisposed to lesion formation. Circ Res. 1999; 85: 199eC207.
Cybulsky MI, Gimbrone MA Jr. Endothelial expression of a mononuclear leukocyte adhesion molecule during atherogenesis. Science. 1991; 251: 788eC791.
Orosz CG. Endothelial activation and chronic allograft rejection. Clin Transplant. 1994; 8: 299eC303.
Koskinen PK, Lemstrom KB. Adhesion molecule P-selectin and vascular cell adhesion molecule-1 in enhanced heart allograft arteriosclerosis in the rat. Circulation. 1997; 95: 191eC196.
Cybulsky MI, Iiyama K, Li H, Zhu S, Chen M, Iiyama M, Davis V, Gutierrez-Ramos JC, Connelly PW, Milstone DS. A major role for VCAM-1, but not ICAM-1, in early atherosclerosis. J Clin Invest. 2001; 107: 1255eC1262.
Davies MJ, Gordon JL, Gearing AJ, Pigott R, Woolf N, Katz D, Kyriakopoulos A. The expression of the adhesion molecules ICAM-1, VCAM-1, PECAM, and E-selectin in human atherosclerosis. J Pathol. 1993; 171: 223eC229.
O’Brien KD, Allen MD, McDonald TO, Chait A, Harlan JM, Fishbein D, McCarty J, Ferguson M, Hudkins K, Benjamin CD, et al. Vascular cell adhesion molecule-1 is expressed in human coronary atherosclerotic plaques. Implications for the mode of progression of advanced coronary atherosclerosis. J Clin Invest. 1993; 92: 945eC951.
McHale JF, Harari OA, Marshall D, Haskard DO. Vascular endothelial cell expression of ICAM-1 and VCAM-1 at the onset of eliciting contact hypersensitivity in mice: evidence for a dominant role of TNF-alpha. J Immunol. 1999; 162: 1648eC1655.
Sans M, Fuster D, Vazquez A, Setoain FJ, Piera C, Pique JM, Panes J. 123Iodine-labeled anti-VCAM-1 antibody scintography in the assessment of experimental colitis. Eur J Gastroenterol Hepatol. 2001; 13: 31eC38.
Allport JR, Lim Y-C, Shipley JM, Senior RM, Shapiro SD, Matsuyoshi N, Vestweber D, Luscinskas FW. Neutrophils from MMP-9- or neutrophil elastase-deficient mice show no defect in transendothelial migration under flow in vitro. J Leukoc Biol. 2002; 71: 821eC828.
Shen J, Luscinskas FW, Connolly A, Dewey CF Jr, Gimbrone MA Jr. Fluid shear stress modulates cytosolic free calcium in vascular endothelial cells. Am J Physiol. 1992; 262: C384eCC390.
Smith GP, Petrenko VA. Phage display. Chem Rev. 1997; 97: 391eC410.
Ricard I, Payet MD, Dupuis G. VCAM-1 is internalized by a clathrin-related pathway in human endothelial cells but its alpha 4 beta 1 integrin counter-receptor remains associated with the plasma membrane in human T lymphocytes. Eur J Immunol. 1998; 28: 1708eC1718.
Lewin M, Bredow S, Sergeyev N, Marecos E, Bogdanov A Jr, Weissleder R. In vivo assessment of vascular endothelial growth factor-induced angiogenesis. Int J Cancer. 1999; 83: 798eC802.
Lim Y-C, Garcia-Cardena G, Allport JR, Zervoglos M, Connolly AJ, Gimbrone MA Jr, Luscinskas FW. Heterogeneity of endothelial cells from different organ sites in T-cell subset recruitment. Am J Pathol. 2003; 162: 1591eC1601.
Holzmann B, Weissman IL. Integrin molecules involved in lymphocyte homing to Peyer’s patches. Immunol Rev. 1989; 108: 45eC61.
Weissleder R, Ntziachristos V. Shedding light onto live molecular targets. Nat Med. 2003; 9: 123eC128.
Molenaar TJ, Twisk J, de Haas SA, Peterse N, Vogelaar BJ, van Leeuwen SH, Michon IN, van Berketl TJ, Kuiper J, Biessen EA. P-selectin as a candidate target in atherosclerosis. Biochem Pharmacol. 2003; 66: 859eC866.
Houston P, Goodman J, Lewis A, Campbell CJ, Braddock M. Homing markers for atherosclerosis: applications for drug delivery, gene delivery and vascular imaging. FEBS Lett. 2001; 492: 73eC77.
Winter PM, Morawski AM, Caruthers SD, Fuhrhop RW, Zhang H, Williams TA, Allen JS, Lacy EK, Robertson JD, Lanza GM, Wickline SA. Molecular imaging of angiogenesis in early-stage atherosclerosis with alpha(v)beta3-integrin-targeted nanoparticles. Circulation. 2003; 108: 2270eC2274.
Liu C, Bhattacharjee G, Boisvert W, Dilley R, Edgington T. In vivo interrogation of the molecular display of atherosclerotic lesion surfaces. Am J Pathol. 2003; 163: 1859eC1871.
Harisinghani MG, Barentsz J, Hahn PF, Deserno WM, Tabatabaei S, van de Kaa CH, de la Rosette J, Weissleder R. Noninvasive detection of clinically occult lymph-node metastases in prostate cancer. N Engl J Med. 2003; 348: 2491eC2499.
Arap W, Pasqualini R, Ruoslahti E. Cancer treatment by targeted drug delivery to tumor vasculature in a mouse model. Science. 1998; 279: 377eC380.(Kimberly A. Kelly, Jennif)