Effect of PEGylation on Ligand-Based Targeting of Drug Carriers to the

Aug 6, 2013 - Anita K. Bakrania , Bhavesh C. Variya , Snehal S. Patel ... Catherine A Fromen , Mariana Carasco-Teja , Rhima M Coleman , Omolola Eniola...
9 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/Langmuir

Effect of PEGylation on Ligand-Based Targeting of Drug Carriers to the Vascular Wall in Blood Flow Peter J. Onyskiw and Omolola Eniola-Adefeso* Department of Chemical Engineering, University of Michigan, Ann Arbor, Michigan 48109, United States S Supporting Information *

ABSTRACT: The blood vessel wall plays a prominent role in the development of many life-threatening diseases and as such is an attractive target for treatment. To target diseased tissue, particulate drug carriers often have their surfaces modified with antibodies or epitopes specific to vascular wall-expressed molecules, along with poly(ethylene glycol) (PEG) to improve carrier blood circulation time. However, little is known about the effect of poly(ethylene glycol) on carrier adhesion dynamicsspecifically in blood flow. Here we examine the influence of different molecular weight PEG spacers on particle adhesion in blood flow. Anti-ICAM-1 or Sialyl Lewisa were grafted onto polystyrene 2 μm and 500 nm spheres via PEG spacers and perfused in blood over activated endothelial cells at physiological shear conditions. PEG spacers were shown to improve, reduce, or have no effect on the binding density of targeted-carriers depending on the PEG surface conformation, shear rate, and targeting moiety.

■

orientation than higher molecular weight chains.6,7 Surfaces with PEG molecular weights greater than 3.4 kDa with 2−5 mol % surface coverage have been shown to be sufficient conditions for preventing protein adsorption.4,8 However, to date, it is not clear what effect PEGylation would have on the adhesion of VTCs to the vascular wall in human blood flow. Overall, the effectiveness of VTCs is dependent on the following: the carrier’s (1) ability to migrate from the center of bulk blood flow and localize to the blood vessel wall and (2) capacity to specifically target and adhere to the endothelium in the diseased tissue. VTC localization to the blood vessel wall is dependent on carrier parameters such as size, shape, and hemodynamics.9−12 Site-specificity typically requires modifying the carrier’s surface with adhesive ligands whose counterreceptors are upregulated during a diseased state. For PEGylated VTCs this would often entail attaching the adhesive ligand to the free end of the PEG spacer, which has been shown to alter the receptor−ligand dynamics. For example, the binding distance of a biotin−avidin system was previously shown to increase with PEGylated biotin and was a function of the PEG molecular weight.13,14 PEG molecular weight and density was also shown to influence ligand induced cellular adhesion and internalization.15,16 Similarly, PEGylation was found to improve the binding affinity of ligands to their counter receptors in a low shear saline laminar flow system via decreasing bond stress and increasing the frequency of binding.17 However, little work has been done to examine

INTRODUCTION Localized delivery of therapeutics via targeted drug carrier systems remains attractive for the treatment of many chronic life-threatening diseases. The vascular endothelium, the monolayer of endothelial cells (ECs) lining the internal lumen of blood vessels, is attractive for disease targeting because of its direct contact with circulating blood, prevalence throughout the body, and active role in diseases such as atherosclerosis and cancer. A major challenge for systemic circulating vascular-targeted carriers (VTCs) is the avoidance of the immune response: specifically, the avoidance of plasma protein adsorption onto the VTC’s surface which initiates systemic clearance. To date, the most common method used to prevent protein adsorption is PEGylationthe modification of a carriers’ surface with poly(ethylene glycol) (PEG). The hydrophilic nature of PEG helps inhibit plasma proteins from adsorbing onto the carrier’s surface, resulting in an increase in the carrier’s systemic circulation time.1 Increasing VTC circulation time in turn improves the passive targeting ability of drug carriers to areas such as cancerous tumors, the lungs, liver, and the brain.2,3 Many of the passive benefits from PEGylation, including increased circulation time, are dependent on the orientation of the PEG corona.4,5 Improved circulation time is typically achieved when the PEG spacers are oriented in a “brush” conformation where the PEG chains protrude away from the carrier’s surface rather than the “mushroom” orientation where the chains are entangled and remain close to the carrier’s surface. Achieving the brush versus mushroom conformation is dependent on both the PEG molecular weight and grafting density, with lower molecular weight chains requiring higher surface densities to achieve the “brush” © 2013 American Chemical Society

Received: June 8, 2013 Revised: August 5, 2013 Published: August 6, 2013 11127

dx.doi.org/10.1021/la402182j | Langmuir 2013, 29, 11127−11134

Langmuir

Article

mixture was then incubated at room temperature for 20 h. After conjugation, the PEGylated spheres were thoroughly washed and stored in 50 mM phosphate buffered saline (PBS). For ligand conjugation, biotinylated Sialyl Lewisa (sLea) (Glycotech) or biotinylated human-anti-ICAM-1-mouse-IgG-1 (aICAM-1) (R&D Systems) was premixed with 20 μg/mL NeutrAvidin (Thermo Scientific) at an equal volume ratio for 20 min followed by incubation with PEGylated spheres (100 μL total volume) for 45 min at room temperature. Ligated spheres were washed and stored in PBS buffer containing calcium and magnesium ions (Gibco, Life Technologies) and 1% bovine-serum-albumin (BSA) (Sigma Aldrich). For ligand coated non-PEGylated spheres, polystyrene carboxylated spheres were conjugated with NeutrAvidin as previously described.9 BiotinylatedsLea or aICAM-1 was then incubated with the NeutrAvidin conjugated spheres for 45 min at room temperature, after which spheres were washed and stored in PBS buffer containing calcium and magnesium ions and 1% BSA. Determination of PEG and Ligand Surface Density. The surface densities of PEG and targeting ligands on spheres were characterized via flow cytometry (BD FACSCalibur). For measurement of microsphere PEG-biotin surface density, PEGylated spheres were stained with avidin-FITC (Thermo Scientific) at 10 μg/mL for 20 min at room temperature. For FITC-loaded nanospheres, antibiotin-PE was used for PEG-biotin surface density measurements. For measurement of ligand surface density, anti-cutaneous-lymphocyte antigen-PE (aCLA-PE) (Miltenyi Biotec) or goat-anti-mouse IgGFITC (Jackson ImmunoResearch) were used to label sLea or aICAM1, respectfully. Goat-anti-mouse IgG-PE was used to determine the site density of aICAM-1 on nanospheres. Fluorescent intensities were converted to surface densities via a standard calibration curve, fluorescein-to-protein ratio, and microsphere surface area. sLea and aICAM-1 on particles were matched to densities of 1000 and 2600 sites/μm2, respectfully, which previously resulted in sufficient adhesion for non-PEGylated spheres of the same sizes and specific flow conditions explored in this work.10,18 Provided that Nσ1/3 > 15, the length of the PEG corona can be estimated using the following scaling law:

the effect of PEG spacers on the adhesion of VTCs to the endothelium in blood under physiological flow conditions. Hemodynamics significantly alters VTC localization and adhesion to the vascular wall in blood flow when compared to that in saline systems. For example, neutrophils have been shown to negatively affect the adhesion dynamics of 5 μm spheres but have no effect on the adhesion of 2 μm spheres under the same hemodynamic conditions.10,18 The blood hematocrit (volume fraction of red blood cells in blood) plays a significant role in determining the optimum carrier size for efficient targeting.18 Red blood cells (RBCs) align themselves in the center of flow resulting in a RBC-free plasma layer adjacent to the endothelium. The RBC core then acts as a protective barrier from shear forces and alters the adhesion trend of VTCs from that in saline flow.9 However, VTCs must first migrate out of the RBC core before they can successfully interact with the endothelium. The ability of VTCs to migrate through the RBC core has been shown to depend on carrier size and shape.12,18 While previous work has demonstrated improved VTC adhesion characteristics with PEGylation in low shear buffer flow, it is unclear if the improved dynamics is of relevance to particle adhesion in blood flow.17 In this work, we examine the effect of different size PEG spacers (2.3 kDa, 5.5 kDa, and 10 kDa) on the adhesion of 2 μm and 500 nm polystyrene spheres to activated human umbilical vein endothelial cell (HUVEC) monolayer from human blood at physiological shear rates. The effect of PEG molecular weight and surface density was examined using two receptor−ligand systems commonly explored for targeting endothelial inflammation; E-selectin/Sialyl Lewisa (sLea) and intercellular adhesion molecule-1 (ICAM-1)/anti-intercellular adhesion molecule-1 (aICAM-1).19 Our results show that PEG spacers improved, diminished, or did not affect the adhesion efficacy of VTCs in blood depending on the PEG size and surface density, and the nature of the specific receptor−ligand kinetics employed.

■

L = Naσ 1/3

(1)

where N is the degree of polymerization of the PEG chain (obtained by dividing the total molecular weight by the molecular weight of 1 PEG monomer), a is the length of one PEG monomer (0.35 nm), and σ is the dimensionless surface density of PEG on the sphere.21 The PEG density was made dimensionless by normalizing the surface density (PEG chains/nm2) by 1 nm2, in order to maintain the same length scale as the PEG chain’s estimated Flory’s radius, Rf = aN0.64 as described by Murat and Grest.17,21 Flow Adhesion Assay. Laminar flow assays were conducted as previously described.11,18 Briefly, an activated HUVEC monolayer on a glass coverslip was attached to a parallel plate flow chamber (PPFC; Glycotech) containing a straight rectangular gasket (1 cm × 2 cm × 254 μm). Blood containing ligand-coated spheres (5 × 105 particles/ mL) was perfused over the activated monolayer for 5 min at wall shear rates (WSR, γw) of 200 s−1, 500 s−1, and 1000 s−1 by adjustment of the volumetric flow rate through the channel via the following equation:

METHODS

Preparation of Endothelial Cell Monolayer. Human umbilical vein endothelial cells (HUVECs) used in all adhesion assays were isolated from fresh umbilical cords (Mott Children’s Hospital, Ann Arbor, MI) via collagenase perfusion method.20 Endothelial cell monolayer was prepared as previously described where HUVECs were cultured onto 30 mm glass coverslips coated with 1% gelatin (Sigma Aldrich) cross-linked with 0.5% gluteraldehyde (Fisher) and 0.1 M glycine (Sigma Aldrich).10,18 Coverslips with cultured cells were incubated at 37 °C and 5% CO2 to allow growth to confluency. Confluent monolayers were activated with 1 ng/mL interleukin-1β (IL-1β) (Fitzgerald Industries) for 4 or 24 h to induce E-selectin or ICAM-1 expression, respectfully, prior to use in adhesion assays. Preparation of Human Blood. Fresh human blood used in all assays was obtained via venipuncture according to a protocol approved by the University of Michigan Internal Review Board and in line with the standards set by the Helsinki Declaration.18 Appropriate informed consent was obtained from human subjects. Venous blood was collected from healthy adults into a syringe containing citric acidsodium citrate-dextrose (ACD) as anticoagulant. Particle Preparation. Amine-PEG-biotin was covalently coupled to carboxylated polystyrene micro- and nanospheres via carboiimide chemistry. Carboxylated micro- (2.07 μm; Bangs Laboratories) and nanospheres (520 nm; Polysciences Inc.) were premixed with 0.2−5 mg/mL of amine-PEG-biotin (Laysan Bio, Inc.) in MES (Sigma Aldrich) buffer for 15 min; after which, equal volume (75 mg/mL) of N-(3-dimethilaminopropyl)-N′-ethylcarboiimide hydrochloride (EDAC) (Sigma Aldrich) was added and the pH adjusted to 9. The

γW =

6Q −1 ;s h2w

(2)

where h is the channel height (254 μm), w the channel width (1 cm), and Q the volumetric flow rate. After 5 min of perfusion, particle adhesion was assessed via optical imaging using a Nikon TE-2000-S inverted microscope with a digital camera (Photometrics CoolSNAP EZ with a Sony CCD sensor). Results were imaged and analyzed via Metamorph analysis software. Adhesion density was normalized based on the imaged surface area (0.152 mm2). Data is plotted with standard error bars and differences in adhesion density was analyzed using oneway ANOVA with Tukey post-test with α = 0.01 (Prism GraphPad). 11128

dx.doi.org/10.1021/la402182j | Langmuir 2013, 29, 11127−11134

Langmuir

■

Article

lack of a significance difference in adhesion between sLeaspheres of different PEG molecular weights was a result of the differences in the actual PEG density on particle surface, we evaluated the adhesion of 5.5 kDa and 10 kDa PEGylated sLeaspheres at a fixed PEG brush density of ∼5200 PEG chains/ μm2 and a fixed sLea density of 1000 sites/μm2. As shown in Figure 1b, the adhesion density for the 5.5 kDa PEG sLeaspheres was not significantly different from that of 10 kDa PEG sLea-spheres at the fixed PEG density (p = 0.1097 and p = 0.0389 for 200 and 1000 s−1, respectfully). Overall, the adhesion of both non-PEGylated and PEGylated sLea-spheres increased with increasing channel WSR in the PPFC. This observation is expected for sLea-mediated adhesion of 2 μm spheres in blood flow for the ligand density used based on our previous publication and is due to the normal force impacted by the RBCs aiding the particles at the wall in overcoming the increase in shear force associated with an increase in WSR, i.e., an increase in the WSR in the absence of a corresponding increase in disruptive force results in higher adhesion.9 Also, control experiments with untargeted PEGylated spheres resulted in no adhesion to the EC monolayer (Figure S1), confirming that the actual particle adhesion shown in Figure 1 is mediated by the interaction of targeting ligand (sLea) with its EC-expressed receptor(s). To investigate the effect of PEG conformation on the adhesion of spheres, 5.5 kDa PEG was conjugated to the surfaces of 2 μm spheres at either a low or high surface density of approximately 3500 ± 300 or 11 000 ± 200 chains/μm2, which are estimated to correspond to a mushroom (chain overlap = 40%) or a more extended brush conformation (chain overlap = 72%), respectfully. At 200 s−1, there was no difference in adhesion between non-PEGylated and PEGylated sLeaspheres for both PEG conformations at this PEG molecular weight (Figure 2a). At 500 s−1, the adhesion of the PEGMushroom spheres was slightly lower than the adhesion of the non-PEGylated spheres, but there was no significant difference in adhesion between non-PEGylated spheres and the PEGBrush spheres. For the 1000 s−1 WSR, the adhesion of PEGMushroom spheres was significantly lower than both the nonPEGylated and PEG-Brush spheres. There was no significant difference in adhesion between the non-PEGylated and the PEG-Brush spheres. A similar assay with 2.3 kDa PEGylated sLea-spheres shows the same trend where microspheres with a low PEG density of ∼12 800 PEG chains/μm2 (chain overlap = 44%; PEG-Mushroom) displayed a significantly lower adhesion at 1000 s−1 relative to particles with a high PEG density of 26

RESULTS Effect of PEG on sLea Mediated Adhesion. In order to investigate the effect of PEG molecular weight on particle adhesion to activated ECs, 2.3 kDa, 5.5 kDa, and 10 kDa PEG spacers were conjugated onto 2 μm spheres at surface densities that correspond to an extended brush conformation for each of the PEG sizes. PEG brush conformation was estimated using the Flory’s radius, Rf. A brush conformation occurs when the ratio of the Rf and the distance between adjacent PEG chains (S, estimated based on the PEG surface density) is greater than 0.5, i.e., Rf/S > 1/2.17 An Rf of 4.4 nm, 7.7 nm, and 11.3 nm with chain overlaps of 63%, 92%, and 72% were used to achieve an extended brush conformation for the 2.3 kDa, 5.5 kDa, and 10 kDa PEGylated microspheres, respectfully. The approximated lengths of the resulting brush corona are shown in Table 1 as estimated by eq 1. It is worth noting that it is possible that Table 1. PEG Density, Approximate PEG Chain Length (As Estimated by eq 1), and Rf of 2 μm PEGylated Spheres in the Brush Conformation molecular weight PEG Density(chains/ μm2) Approximate PEG Brush Chain Length (nm) Rf (nm)

2.3 kDa

5.5 kDa

10 kDa

25 000 ± 2300

18 440 ± 50

5200 ± 300

5.3

11.6

13.8

4.4

7.7

11.3

the use of multivalent probes may underestimate the PEG surface densities used in the chain overlap calculations, particularly for larger PEG chains that have more mobility on the surface and hence a higher chance of multiple chains reacting with a single multivalent probe. However, there was good agreement in the measured site densities for the high molecular weight PEG chains on 2 μm spheres when probed with the avidin-FITC versus the anti-biotin-PE probe (Table S1). Figure 1a shows the adhesion of sLea-coated spheres with PEG spacers in the brush conformation as a function of channel WSR for a fixed sLea (targeting ligand) density of 1000 sites/ μm2. There was no significant difference in the blood flow adhesion density between the non-PEGylated and PEGylated sLea-spheres for all PEG molecular weights at each of the WSRs evaluated. The PEG brush densities used for assays in Figure 1a were set based on the maximum average density achievable on the 2 μm spheres for each PEG size. To determine whether the

Figure 1. (a) Adhesion of sLea (1000 sites/μm2) PEGylated 2 μm spheres to activated ECs in laminar whole blood flow as a function of PEG size and channel wall shear rate. (b) Adhesion of 5.5 kDa and 10 kDa-PEGylated 2 μm spheres in laminar whole blood flow for a fixed PEG density of 5200 PEG chains/μm2 and sLea density of 1000 sites/μm2. 11129

dx.doi.org/10.1021/la402182j | Langmuir 2013, 29, 11127−11134

Langmuir

Article

Figure 2. Adhesion of (a) 5.5 kDa PEG-Mushroom (3500 PEG chains/μm2) and PEG-Brush (11 000 PEG chains/μm2) and (b) 2.3 kDa PEGMushroom (12 800 PEG chains/μm2) and PEG-Brush (26 800 PEG chains/μm2) 2 μm spheres targeted with 1000 sLea/μm2 to activated ECs in laminar whole blood flow as a function of channel wall shear rate (WSR). * indicates p < 0.01 when compared to the adhesion of non-PEGylated spheres at the same WSR and # indicates p < 0.01 when compared to the adhesion of PEG-Mushroom spheres (of the same molecular weight) at the same WSR.

Figure 3. (a) Adhesion of aICAM-1 (2600 sites/μm2) PEGylated microspheres to activated ECs in laminar whole blood flow as a function of PEG brush length at different wall shear rates (WSR). * indicates p < 0.01 when compared to the adhesion of non-PEGylated spheres at the same WSR and ** indicates p < 0.01 when compared to the adhesion of 2.3 kDa PEG spheres at the same WSR. (b) Adhesion density normalized to the number of particles perfused at a given wall shear rate (Binding Efficiency). # indicates p < 0.01 compared to the adhesion of the same particle at 200 s−1 and ## indicates p < 0.01 compared to the adhesion of the same particle at 500 s−1.

PEG chains/μm2 is used at a fixed sLea density of 1000 sites/ μm2 (data not shown). To determine whether the higher adhesion of PEGylated aICAM-1-spheres with increasing channel WSR is due to improved adhesion kinetics, particle adhesion density was normalized to the total number of particles perfused through the flow channel at a fixed WSR, i.e., fraction of total perfused particle that was bound (adhesion efficiency). Figure 3b shows the adhesion efficiency of PEGylated and non-PEGylated spheres at different WSRs as a function of PEG molecular weight (Table 1). For non-PEGylated spheres the adhesion efficiency significantly decreased with increasing WSR (compared to the efficiency of the same particles at 200 s−1). For all PEGylated spheres the adhesion efficiency at 500 s−1 was not significantly different from the efficiency at 200 s−1, suggesting improved adhesion kinetics relative to the nonPEGylated particles; however, the adhesion efficiency of all PEGylated spheres at 1000 s−1 was significantly lower than the efficiencies of the same spheres at both 200 s−1 and 500 s−1. To determine the effect of PEG conformation on ICAM-1 targeting, 5.5 kDa PEG spacers were conjugated to the surface of 2 μm spheres at the low or high PEG surface density estimated to be in a mushroom (40% chain overlap) or a more extended brush conformation (72% chain overlap), respectfully, as was used for the sLea assays. There was no significant difference in adhesion between the non-PEGylated and PEGylated aICAM-1-spheres with both conformations at 200 s−1 (Figure 4). At 500 s−1, the adhesion of non-PEGylated

800 PEG chains/μm2 (chain overlap = 63%; PEG-Brush) and ones that were non-PEGylated (Figure 2b). Effect of PEG on aICAM-1 Mediated Adhesion. To investigate whether targeting ligand type influences the adhesion of PEGylated spheres, 2 μm PEGylated and nonPEGylated spheres were conjugated with approximately 2600 sites/μm2 of anti-ICAM-1 (aICAM-1) antibody to target ICAM-1an immunoglobulin super family protein upregulated during chronic inflammation. Figure 3a shows the adhesion to activated ECs in blood flow of aICAM-1 spheres containing 2.3 kDa, 5.5 kDa, and 10 kDa PEG spacers in an extended brush conformation at the respective densities reported in Table 1 as a function of channel WSR. At 200 s−1, there was no significant difference in adhesion between the PEGylated and non-PEGylated aICAM-1-spheres. At 500 s−1, the adhesion of non-PEGylated aICAM-1-spheres was significantly lower than the adhesion of all PEGylated spheres and there was no significant difference in adhesion between the different PEGylated spheres. At 1000 s−1, no adhesion was observed for non-PEGylated aICAM-1 spheres and the level of adhesion observed for the 2.3 kDa PEG-Brush spheres (5.3 nm in approximate PEG brush length) was significantly lower than that for both the 5.5 kDa and 10 kDa PEG-Brush spheres (11.6 and 13.8 in approximate PEG brush lengths, respectfully). No difference in adhesion was observed between the 5.5 kDa and 10 kDa PEGylated spheres. As with sLea-targeting, we find no significant difference in the adhesion between the 5 and 10 kDa PEG microspheres when a fixed PEG brush density of ∼5200 11130

dx.doi.org/10.1021/la402182j | Langmuir 2013, 29, 11127−11134

Langmuir

Article

PEGylated spheres with 2600 aICAM-1 sites/μm2 at 1000 s−1 WSR. Effect of PEGylation on Nanosphere Adhesion. To determine whether the effect of PEG spacers on particle adhesion is affected by particle size, we investigated the adhesion of 500 nm spheres conjugated with a 5.5 kDa PEG spacer oriented in the brush conformation (∼16 000 PEG chains/μm2, 11.0 nm approximate brush length) and targeted with either sLea or aICAM-1 (1000 sites/μm2 and 2600 sites/ μm2, respectfully). As shown in Figure 5a, there was no significant difference in adhesion between PEGylated and nonPEGylated sLea-spheres at the three WSRs explored, similar to observations with the 2 μm sLea-spheres. The adhesion of both PEGylated and non-PEGylated 500 nm sLea-spheres increased with increasing WSR between 200 and 500 s−1; however, there was no significant increase in adhesion for both nanosphere types between 500 and 1000 s−1. Increasing the PEG size on the sLea-targeted spheres to 10 kDa did not significantly alter their adhesion compared to both the non-PEGylated and 5.5 kDa PEGylated nanospheres (Figure S2). When targeted with aICAM-1, there was no significant difference between the adhesion of PEGylated and non-PEGylated 500 nm spheres in blood flow at 200 s−1 (Figure 5b), similar to the observation with sLea spheres. However, at 500 and 1000 s−1 the adhesion of 5.5 kDa PEGylated spheres was significantly higher than that of non-PEGylated spheres. There was no significant difference in adhesion between the PEGylated spheres at 500 and 1000 s−1. Overall, the adhesion of the 500 nm spheres was 4- to 17fold lower than the adhesion of the 2 μm spheres at the same blood particle concentration for similar PEG conditions and the WSRs evaluated.

Figure 4. Adhesion of 5.5 kDa PEG-Mushroom (3500 PEG chains/ μm2) and PEG-Brush (11 000 PEG chains/μm2) 2 μm spheres targeted with 2600 aICAM-1 sites/μm2 to activated ECs in laminar whole blood flow as a function of channel wall shear rate (WSR). * indicates p < 0.01 when compared to the adhesion of non-PEGylated spheres at the same WSR, ** indicates p < 0.01 when compared to the adhesion of PEG-Mushroom spheres at the same WSR, and # indicates p < 0.01 when compared to the adhesion of non-PEGylated spheres with 4400 aICAM-1 sites/μm2 and the same WSR.

spheres was significantly lower than the adhesion of both PEGylated spheres, and the adhesion of the PEG-Mushroom spheres was significantly lower than the adhesion of PEG-Brush spheres. At 1000 s−1, there was no adhesion observed for nonPEGylated spheres and minimal adhesion for the mushroomoriented PEG spheres. The adhesion of the PEG-Brush spheres at 1000 s−1 was significantly higher than that of the low density PEG-Mushroom spheres. To determine if improving the adhesion kinetics for nonPEGylated spheres (via increasing aICAM-1 site density) would increase their adhesion density to the level of PEGylated spheres, the adhesion of non-PEGylated spheres with 4400 sites/μm2 aICAM-1 density (Figure 4; solid black bars) was compared to the adhesion of non-PEGylated and PEGylated spheres with 2600 aICAM-1 sites/μm2. At 200 s−1, the adhesion density of the higher aICAM-1-density nonPEGylated spheres was significantly higher than all the spheres with 2600 aICAM-1 sites/μm2. At 500 s−1 WSR, the adhesion of the higher aICAM-1 density spheres was significantly higher than both the PEG-Mushroom and non-PEGylated low aICAM-1 density spheres; however, there was no significant difference in the adhesion density between the PEG-Brush spheres and the higher aICAM-1 density non-PEGylated spheres. At 1000 s−1, the adhesion of the high aICAM-1 density spheres was significantly lower than the PEG-Brush spheres, and there was no significant difference in adhesion between the higher aICAM-1 density-spheres and the PEGMushroom spheres. No adhesion was observed for non-

■

DISCUSSION The addition of PEG onto a carrier’s surface is an effective strategy for sterically reducing protein adsorption and increasing systemic circulation time.2,4 However, it is unclear if the addition of PEG spacers has an effect on ligand-based adhesion of VTCs to the vascular wall in blood flow at physiological WSRs. In this work we investigated the effects of (1) PEG corona size for the case where PEG chains were in the more extended brush conformation and (2) PEG conformation (i.e., high versus low PEG surface density) on VTC adhesion for both sLea and aICAM-1 targeted spheres. Overall, we show that the addition of PEG spacers did not hinder particle adhesion in human blood flow when the VTCs are grafted with a high PEG surface density approximated to be in an extended brush conformation. However, for blood flow at high WSR (1000 s−1), a low-density PEGylation, where the PEG chains are estimated to be oriented in a more entangled mushroom

Figure 5. Adhesion of (a) sLea (1000 sites/μm2)-PEGylated and (b) aICAM-1 (2600 sites/μm2)-PEGylated 500 nm spheres to activated ECs in laminar whole blood flow at different wall shear rates (WSR). * Indicates p < 0.01 when compared to non-PEGylated spheres at a fixed WSR. 11131

dx.doi.org/10.1021/la402182j | Langmuir 2013, 29, 11127−11134

Langmuir

Article

conformation based on the chain overlap less than 1/2) is likely due to poor ligand presentation and steric repulsive forces as a result of the PEG corona length being similar to its equilibrium length, Rf. When the PEG spacers are oriented in a more extended brush conformation (high PEG surface density), where the size of the brush corona is estimated to be greater than its corresponding Rf, the targeting ligand is easily accessible for adhesion with its counter receptor, resulting in an adhesion density equivalent to or greater than the nonPEGylated spheres (depending on the targeting ligand) (Figures 2 and 4). For both the 2.3 kDa and 5.5 kDa PEGspheres at low densities, the size of the PEG corona is equivalent to or slightly shorter than its corresponding Rf and hence a higher degree of PEG chain entanglement is anticipated. Higher chain entanglement may then result in poor ligand presentation (i.e., decreased effective valency), which leads to a decrease in particle adhesion as observed for both targeting ligand systems evaluated. Indeed, our findings are in agreement with previous works with bimodal systems (two PEG spacers of different lengths on a carrier’s surface) that have shown that when the targeting ligand is attached to a spacer that is significantly shorter than the Rf of the adjacent spacer, adhesion in static conditions is inhibited.16,25 However, it may be that improving the ligand valency, via increasing ligand density, on the low-density, mushroom-oriented PEGspheres would be sufficient to increase particle adhesion to the same level as the non-PEGylated and PEG-Brush spheres, although our unpublished observation of 2.3 kDa PEGMushroom spheres show that a 1.4-fold higher sLea density on these spheres did not increase their adhesion to the level of the non-PEGylated spheres shown in Figure 2b, which suggests a further increase in ligand density may be required. Increasing ligand density substantially may, however, affect tissue selectivity as was recently demonstrated by Zern et al., where nanospheres (∼200 nm) with a relatively higher aICAM-1 ligand density of 200 antibodies/nanoparticle showed lower selectivity toward diseased pulmonary endothelium (in vivo) compared to a lower aICAM-1 ligand density of 50 antibodies/ nanoparticle.26 Incidentally, Cruz et al. previously showed that the cellular internalization of 200 nm spheres targeted to dendritic cells with ligands grafted on a PEG spacer was most efficient (∼70% of DC internalizing particles) when PEG sizes between 2 and 3 kDa are used compared to PEG sizes between 6 and 20 kDa (∼40% of DC internalizing particles), suggesting that the shorter PEG sizes are preferable for situations where cellular internalization is critical.15 In this case, care must be taken to ensure that the amount of PEG grafted on particles is sufficient (i.e., Rf/S >1/2), to achieve all the targeting ligand in the appropriate orientation for targeting applications that involve blood flow adhesion prior to cellular internalization, since a higher PEG density is required to cross this threshold with smaller PEG sizes (e.g., ∼16 000 PEG chains/μm2 for 2.3 kDa versus ∼5000 PEG chains/μm2 for 5.5 kDa PEG by our estimation). We were also curious to see if the influence of PEG was particle size dependent by evaluating the effect of PEG spacers on the adhesion of 500 nm spheres under similar conditions as 2 μm spheres. Nanospheres, rather than larger microspheres, were explored due to attractiveness for use as VTCs because of their ability to avoid capillary entrapment, thus potentially improving targeting efficiency. The addition of a PEG in an extended brush conformation did not significantly improve the adhesion density for sLea-targeted 500 nm spheres (Figures 5

conformation (chain overlap 2 kDa) at a sufficiently high density needed to achieve the extended or brush conformation is important since VTC adhesion from blood flow may be negatively impacted, relative to nonPEGylated VTCs bearing the same ligand density, when lower PEG densities that result in a more entangled or mushroom

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by a grant from the American Heart Association (0735043N) to O.E.A.

■

REFERENCES

(1) Wattendorf, U.; Merkle, H. P. Pegylation as a Tool for the Biomedical Engineering of Surface Modified Microparticles. J. Pharm. Sci. 2008, 97, 4655−4669. (2) Choi, K. Y.; Min, K. H.; Yoon, H. Y.; Kim, K.; Park, J. H.; Kwon, I. C.; Choi, K.; Jeong, S. Y. Pegylation of Hyaluronic Acid Nanoparticles Improves Tumor Targetability in Vivo. Biomaterials 2011, 32, 1880−1889. (3) Calvo, P.; Gouritin, B.; Chacun, H.; Desmaele, D.; D’Angelo, J.; Noel, J.-P.; Georgin, D.; Fattal, E.; Andreux, J. P.; Couvreur, P. LongCirculating Pegylated Polycyanoacrylate Nanoparticles as New Drug Carrier for Brain Delivery. Pharm. Res. 2001, 18, 1157−1166. (4) Gref, R.; Lück, M.; Quellec, P.; Marchand, M.; Dellacherie, E.; Harnisch, S.; Blunk, T.; Müller, R. H. ’Stealth’ Corona-Core Nanoparticles Surface Modified by Polyethylene Glycol (Peg): Influences of the Corona (Peg Chain Length and Surface Density) and of the Core Composition on Phagocytic Uptake and Plasma Protein Adsorption. Colloids Surf., B 2000, 18, 301−313. (5) Unsworth, L. D.; Sheardown, H.; Brash, J. L. Protein-Resistant Poly(Ethylene Oxide)-Grafted Surfaces: Chain Density-Dependent Multiple Mechanisms of Action. Langmuir 2008, 24, 1924−1929. (6) Kenworthy, A. K.; Hristova, K.; Needham, D.; McIntosh, T. J. Range and Magnitude of the Steric Pressure between Bilayers Containing Phospholipids with Covalently Attached Poly(Ethylene Glycol). Biophys. J. 1995, 68, 1921−1936. (7) Noppl-Simson, D. A.; Needham, D. Avidin-Biotin Interactions at Vesicle Surfaces: Adsorption and Binding, Cross-Bridge Formation, and Lateral Interactions. Biophys. J. 1996, 70, 1391−1401. (8) Gombotz, W. R.; Guanghui, W.; Horbett, T. A.; Hoffman, A. S. Protein Adsorption to Poly(Ethylene Oxide) Surfaces. J. Biomed. Mater. Res. 1991, 25, 1547−1562.

11133

dx.doi.org/10.1021/la402182j | Langmuir 2013, 29, 11127−11134

Langmuir

Article

(9) Charoenphol, P.; Huang, R. B.; Eniola-Adefeso, O. Potential Role of Size and Hemodynamics in the Efficacy of Vascular-Targeted Spherical Drug Carriers. Biomaterials 2010, 31, 1392−1402. (10) Charoenphol, P.; Mocherla, S.; Bouis, D.; Namdee, K.; Pinsky, D. J.; Eniola-Adefeso, O. Targeting Therapeutics to the Vascular Wall in Atherosclerosis - Carrier Size Matters. Atherosclerosis 2011, 217, 364−370. (11) Namdee, K.; Thompson, A. J.; Phapanin, C.; Eniola-Adefeso, O. Margination Propensity of Vascular-Targeted Spheres from Blood Flow in a Microfluidic Model of Human Microvessels. Langmuir 2013, 29, 2530−2535. (12) Thompson, A. J.; Mastria, E. M.; Eniola-Adefeso, O. The Margination Propensity of Ellipsoidal Micro/Nanoparticles to the Endothelium in Human Blood Flow. Biomaterials 2013, 34, 5863− 5871. (13) Wong, J. Y.; Kuhl, T. Dynamics of Membrane Adhesion: The Role of Polyethylene Glycol Spacers, Ligand-Receptor Bond Strength, and Rupture Pathway. Langmuir 2008, 24, 1225−1231. (14) Jeppesen, C.; Wong, J. Y.; Kuhl, T. L.; Israelachvili, J. N.; Mullah, N.; Zalipsky, S.; Marques, C. M. Impact of Polymer Tether Length on Multiple Ligand-Receptor Bond Formation. Science 2001, 293, 465−468. (15) Cruz, L. J.; Tacken, P. J.; Fokkink, R.; Figdor, C. The Influence of Peg Chain Length and Targeting Moiety on Antibody-Mediated Delivery of Nanoparticle Vaccines to Human Dendritic Cells. Biomaterials 2011, 32, 6791−6803. (16) Gabizon, A.; Horowitz, A. T.; Goren, D.; Tzemach, D.; Mandelbaum-Shavit, F.; Qazen, M. M.; Zalipsky, S. Targeting Folate Receptor with Folate Linked to Extremities of Poly(Ethylene Glycol)Grafted Liposomes: In Vitro Studies. Bioconjugate Chem. 1999, 10, 289−298. (17) Ham, A. S.; Klibanov, A. L.; Lawrence, M. B. Action at a Distance: Lengthening Adhesion Bonds with Poly(Ethylene Glycol) Spacers Enhances Mechanically Stressed Affinity for Improved Vascular Targeting of Microparticles. Langmuir 2009, 25, 10038− 10044. (18) Charoenphol, P.; Onyskiw, P. J.; Carrasco-Teja, M.; EniolaAdefeso, O. Particle-Cell Dynamics in Human Blood Flow: Implications for Vascular-Targeted Drug Delivery. J. Biomech. 2012, 45, 2822−2828. (19) Chacko, A.-M.; Hood, E. D.; Zern, B. J.; Muzykantov, V. R. Targeted Nanocarriers for Imaging and Therapy of Vascular Inflammation. Curr. Opin. Colloid Interface Sci. 2011, 16, 215−227. (20) Burns, A. R.; Bowden, R. A.; MacDonell, S. D.; Walker, D. C.; Odebunmi, T. O.; Donnachie, E. M.; Simon, S. I.; Entman, M. L.; Smith, C. W. Analysis of Tight Junctions During Neutrophil Transendothelial Migration. J. Cell Sci. 2000, 113, 45−57. (21) Murat, M.; Grest, G. S. Structure of a Grafted Polymer Brush: A Molecular Dynamics Simulation. Macromolecules 1989, 22, 4054− 4059. (22) Chang, K.-C.; Hammer, D. A. The Forward Rate of Binding of the Surface-Tethered Reactants: Effect of Relative Motion between Two Surfaces. Biophys. J. 1999, 76, 1280−1292. (23) Brunk, D. K.; Hammer, D. A. Quantifying Rolling Adhesion with a Cell-Free Assay: E-Selectin and Its Carbohydrate Ligands. Biophys. J. 1997, 72, 2820−2833. (24) Eniola, O. A.; Willcox, P. J.; Hammer, D. A. Interplay between Rolling and Firm Adhesion Elucidated with a Cell-Free System Engineered with Two Distinct Receptor-Ligand Pairs. Biophys. J. 2003, 85, 2720−2731. (25) Needham, D.; Kim, D. H. Peg-Covered Lipid Surfaces: Bilayers and Monolayers. Colloids Surf., B 2000, 18, 183−195. (26) Zern, B. J.; Chacko, A.-M.; Liu, J.; Greineder, C. F.; Blankemeyer, E. R.; Radhakrishnan, R.; Muzykantov, V. Reduction of Nanoparticle Avidity Enhances the Selectivity of Vascular Targeting and PET Detection of Pulmonary Inflammation. ACS Nano 2013, 7, 2461−2469.

11134

dx.doi.org/10.1021/la402182j | Langmuir 2013, 29, 11127−11134