Development and Characterization of PLGA-Based Multistage Delivery System for Enhanced Payload Delivery to Targeted Vascular Endothelium

Abstract

Vascular-targeted drug delivery remains an attractive platform for therapeutic and diagnostic interventions in human diseases. This work focuses on the development of a poly-lactic-co-glycolic-acid (PLGA)-based multistage delivery system (MDS). MDS consists of two stages: a micron-sized PLGA outer shell and encapsulated drug-loaded PLGA nanoparticles. Nanoparticles with average diameters of 76, 119, and 193 nm are successfully encapsulated into 3–6 μm MDS. Sustained in vitro release of nanoparticles from MDS is observed for up to 7 days. Both MDS and nanoparticles arebiocompatible with human endothelial cells. Sialyl-Lewis-A (sLe^A) is successfully immobilized on the MDS and nanoparticle surfaces to enable specific targeting of inflamed endothelium. Functionalized MDS demonstrates a 2.7-fold improvement in endothelial binding compared to PLGA nanoparticles from human blood laminar flow. Overall, the presented results demonstrate successful development and characterization of MDS and suggest that MDS can serve as an effective drug carrier, which can enhance the margination of nanoparticles to the targeted vascular wall.

Toc

1. Introduction

Targeted drug delivery has been a promising approach for delivering payloads to targeted tissues with improved therapeutic efficacy while reducing undesired effects to non-targeted sites. In the past decades, numerous efforts have developed various types and designs of drug carriers in an attempt to optimize their delivery efficiency suitable for specific diseased conditions.[1–4] Among the design criteria, physicochemical properties of drug carriers have demonstrated to be a significant factor determining the efficiency of drug delivery systems.[5–7]

For vascular-targeted drug delivery, nano-sized carriers have been commonly employed due to their injectable nature through parenteral routes without occluding needles and capillaries, and their abilities to penetrate into deeper tissues of certain diseases with compromised vascular walls such as cancer and inflammation.[8,9] These nanoparticles, however, often experienced limited localization to the vascular wall from dense blood flow and resulted in their accumulation in blood circulation, filtering organs, and non-targeted tissues.[10,11] Evidence has been shown that these limitations are a significant barrier for clinical translation.[5,12,13] On the other hand, micron-sized particles have shown to favor localization to the inflamed vascular wall but due to their poor ability to penetrate through the cellular walls, their functionality as drug carriers has been limited.[5]

Over a decade ago, Tasciotti et al., introduced a multistage delivery system (MDS) based on mesoporous silicon particles in an attempt to improve the overall drug delivery efficiency.[14] Since then, several multistage delivery systems (MDS), including liposomes,[15] PEGylated micelles,[16] and silica-based carriers[14,17–20] have been reported. The outer shell of MDS is typically designed to favor vascular margination while the inner stage nanoparticles are designed to favor intracellular penetration and controlled drug release. However, most of the reported MDS particles are intended for sustained release of cancer therapeutics, often present extended degradation profiles, and having average diameters not optimal for vascular targeting (over 50 μm).

Several silicon-based MDS have been designed to target the vascular endothelium.[19,21,22] For instance, Serda et al., developed a 3 μm discoidal silicon MDS loaded with superparamagnetic iron oxide nanoparticles (SPIONs).[19] By conjugating the outer shell of the MDS with endothelial specific antibodies to vascular endothelial growth factor receptor 2 and platelet endothelial cell adhesion molecules, the MDS was shown to specifically target human umbilical vein endothelial cells. Although the discoidal shape and the average diameter of the MDS might enhance vascular margination, this approach relied on covalent attachment of the loaded nanoparticles to the silicon network, which could limit the loading capacity of the carrier. In addition, the payload was encapsulated by absorption into the silicon mesh, which means that only nanoparticles smaller than the pore size can be loaded into the MDS, and their release from the MDS is mainly determined by the pore’s diameter.

In this study, MDS was developed from poly-lactic-co-glycolic-acid (PLGA), an FDA-approved biodegradable and biocompatible polymer. PLGA-based carriers offer several advantages. PLGA’s biocompatibility has been extensively studied. The fabrication process for PLGA is simple, scalable, and can be optimized for controlled payload release. In addition, the tunable biodegradability and capability for PLGA to encapsulate hydrophilic, hydrophobic molecules and protein drugs, without a need of attachment or absorption technique, allow various compounds to be simultaneously delivered in a controlled manner.

Herein, the MDS synthesized is composed of a PLGA micron-sized outer shell that contains payload-encapsulated PLGA nanoparticles. Each layer of the MDS was engineered to overcome the physiological barriers and maximize the targeting ability associated with each delivery stage. The outer shell of MDS was designed to have a diameter from 3–6 μm because this size range has evidenced to optimally localize to a specific site at the vascular wall in human blood flow.[5,10,23–25] The size of the nanospheres was varied from 75–200 nm. Nanoparticles in this size range were expected to carry sufficient payloads for effective actions while adequately penetrating through lesion-prone endothelium for intracellular targeting.[26] To enable a specific targeting to inflamed endothelium, the MDS microparticles were conjugated with biotinylated Sialyl Lewis A (sLe^A), a ligand that binds to E-selectin which is an inflammatory molecule on inflamed endothelial cells. Similarly, the nanoparticles were coated with chitosan through an electrostatic adsorption and can be conjugated with biotinylated ligands to target a specific receptor. To our knowledge, this is the first study developing PLGA-in-PLGA MDS system. The fabrication process and characterization of MDS, including the payload release profile and enhanced ability for vascular margination, have been investigated. The developed PLGA-based MDS carrier could provide a platform to potentially enhance vascular wall localization and tissue penetration of drug-loaded nanoparticles.

2. Results and Discussion

2.1. Synthesis of PLGA Nanoparticles

PLGA-nanoparticles were formed and loaded with DiO, a hydrophobic fluorescent dye, using an oil-in-water emulsion process. Size of nanoparticles was primarily controlled by the shear stress applied during the emulsification. By increasing the output intensity of the sonicator probe from 50% to 100%, nanoparticles were fabricated in three size ranges: 76 ± 18 nm, 119 ± 16 nm, and 193± 32 nm (Figure S1, Supporting Information) with approximate DiO encapsulation efficiencies of 54.1%, 59.4%, and 66.8%, respectively. A desired size range can be chosen to fit an application need. Generally, smaller nanoparticles are suitable for targeting deeper tissues while larger nanoparticles are favorable when a higher capacity of payloads is required.

Prior to loading into the MDS outer shell, the surface of the nanoparticles was coated with chitosan to protect their integrity from solvent dissolution during the MDS fabrication. Positively charged chitosan was deposited on the negatively charged PLGA surface through an ionic bond. The concentration of chitosan was experimentally optimized by estimating the resistance of coated particles to the exposure with organic solvent. The chitosan coating on nanoparticles was controlled by coating a fixed concentration of PLGA nanoparticles (5 × 10¹¹ particles mL⁻¹) with a fixed concentration of chitosan (0.1 mg mL⁻¹) for all particle sizes. The thickness of the chitosan coating was analyzed by characterizing the average diameters of nanoparticles before and after the chitosan coating with nanoparticle tracking analysis (NTA). The change in the average diameter suggested a deposition of a thin layer of chitosan (7 nm on average) on the surface of the 119 nm PLGA nanoparticles (Table S1, Supporting Information). Incubation of nanoparticles in a 0.1 mg mL⁻¹ chitosan solution showed to render their zeta potential from approximately −21.6 mV (uncoated) to +20.5 mV and delay the release of the loaded dye in the first 24 h (Figures S2, Supporting Information). After chitosan coating, the chitosan-PLGA nanoparticles were coupled with biotinylated sLe^A ligands via EDAC-carbodiimide chemistry for specific targeting.[5]

2.2. Synthesis of PLGA-Based MDS

PLGA-based MDS was fabricated using a modified water-in-oil-in-water emulsion solvent evaporation technique (Figure 1). The internal structure and size of MDS were primarily controlled by adjusting the emulsification speeds during the first and second emulsion processes. Specifically, the internal structure of MDS could be regulated by controlling the ratio of the inner and the outer droplet diameters (D_in/D_out) during the double emulsification.[27] In general, if D_in is only a few times smaller than D_out (i.e., relatively high D_in/D_out), MDS particles with microcapsule structures are formed. On the other hand, if D_in is many times smaller than D_out (i.e., relatively low D_in/D_out), MDS are formed with a honeycomb structure (Figure S3, Supporting Information). For vascular-targeted drug delivery, MDS with a microcapsule structure is desired because it offers more space for payloads and the structure allows the payloads to be released promptly once the MDS localized to a target site. The D_in/D_out can be directly controlled by adjusting the ratio of the first and second emulsification speeds (RPM₁/RPM₂), while keeping other parameters (i.e., volumes of water and oil phases, the amount of surfactant and PLGA, and size and number of loaded nanoparticles) constant. The droplet diameters are typically decreased as the emulsification speed increases. Likewise, the D_in/D_out decreases with increasing RPM₁/RPM₂. Figure 2 demonstrates that the RPM₁/RPM₂ from 1.5–2 (i.e., high D_in/D_out) yields the microcapsule structure while as the RPM₁/RPM₂ increased to 3 (i.e., low D_in/D_out), the honeycomb structure was formed.

Figure 1.

Synthesis of MDS by a double emulsion solvent evaporation method.

Figure 2.

Fluorescence microscopy images of MDS synthesized using various combinations of the first and the second emulsion homogenization speeds: A) 1st: 12,500 rpm, 2nd: 7,800 rpm (RPM₁/RPM₂ = 1.6), B) 1st: 9,250 rpm, 2nd: 6,250 rpm (RPM₁/RPM₂ = 1.5), C) 1st: 6,250 rpm, 2nd: 3,125 rpm (RPM₁/RPM₂ = 2), and D) 1st: 9,250 rpm, 2nd: 3,125 rpm (RPM₁/RPM₂ = 3). MDS were loaded with 119 nm PLGA nanoparticles.

After the desired internal structure of MDS could be formed, the fabrication protocol was further optimized to produce MDS with desired average diameters of 3–5 μm. The size range was selected because it has evidenced to optimally localize to a specific site at the vascular wall in human blood flow.[5,10,23–25] Among the fabricating parameters, the second emulsification speed (RPM₂) was shown to primarily determine the MDS outer diameter. As the RPM₂ increased from 3125 to 7800 rpm (for D_in/D_out between 1.5 and 2), the average diameters of MDS decreased from 18.4 to 4.2 μm (Table 1). When the second emulsification speed was fixed at 3125 rpm and the first emulsification speed was varied from 6250 to 9250 rpm, the structure of MDS was changed from the microcapsule to the honeycomb while the MDS size was not significantly altered (Table 1 and Figure 2).

Table 1.

Fabrication parameters and characteristics of different formulations of PLGA-based MDS particles. All MDS formulations were formed with fixed volumes of water and oil phases, fixed amount of surfactant and PLGA, and fixed number of loaded nanoparticles with a size of 119 nm. Values = mean ± SD (300 <n < 3000 (n varies with batches)).

Internal structure	1st emulsion speed [rpm]	2nd emulsion speed [rpm]	MDS diameter [μm]	Nanoparticles per MDS	Encapsulation efficiency [%]
Microcapsule	12 500	7800	4.2 ± 1.4	9.8 ± 5.9	2.8 ± 0.3
Microcapsule	9250	6250	9.1 ± 2.6	14.9 ± 11.7	2.3 ± 0.5
Microcapsule	6250	3125	18.4 ± 6.9	34.1 ± 19.7	1.1 ± 0.2
Honeycomb	9250	3125	20.0 ± 2.8	10.9 ± 5.1	0.3 ± 0.04

Different MDS size was demonstrated to carry a different payload capacity. The number of nanoparticles encapsulated within the microcapsule MDS was shown to be linearly related to the diameter of the MDS shell (Figure 3A). MDS with average diameters of 4.2 μm contained an average of eight nanoparticles while the 18.4 μm shell entrapped an average of 35 nanoparticles. It is worth noting that this is not the maximum loading capacity for a given size of MDS. It is expected that increasing the number of nanoparticles added to the first emulsion would result in MDS with a higher number of loaded nanoparticles. Table 1 also demonstrates the encapsulation efficiencies, a term often used to characterize the loading capacity. The encapsulation efficiencies, in contrast, were found to decrease with the MDS diameter. This is because a mass of the PLGA polymer was kept constant during the protocol optimization. Thus, a lower number of MDS was formed with a larger MDS size and resulted in the overall lower encapsulation efficiencies. It is also not surprising that the honeycomb MDS had about three times lower loading capacity compared to that of the microcapsule MDS with a similar diameter. This is likely because the structure of the former consists of polymer matrices while the latter has a hollow structure.

Figure 3.

A) Number of loaded PLGA nanoparticles versus MDS diameter. B) MDS release profile in DPBS+ and human plasma at 37 °C. C) Average diameter and D) Zeta potential of PLGA nanoparticles (119 nm) that were released from MDS incubated in a saline buffer at 37 °C. Value = mean ± SD (n = 3). ns = no significant difference at 95% confidence.

After the fabrication optimization, MDS was formed with a microcapsule structure having a smooth surface with an average particle diameter of 4.2 ± 1.4 μm (Figure 4). The particle size distribution obtained from bright field microscopy was in line with the scanning electron microscope (SEM) results. Since MDS could be used to deliver payloads suitable for different diseased conditions, a study was performed to encapsulate different nanoparticle sizes into the MDS. Figure 4 illustrates that fluorescent DiO-loaded PLGA nanoparticles with average diameters of 76, 119, and 193 nm were successfully loaded into the PLGA MDS. There was no significant difference in the average diameter of MDS loaded with three different sizes of nanoparticles. Sub-resolution nanoparticles (smaller than 200 nm) showed as blurry fluorescent dots enabling their quantification (Figure S4, Supporting Information). Videos recorded using a fluorescence microscope indicated low aggregation and homogenous distribution of the entrapped DiO nanoparticles (Video S1, Supporting Information). Encapsulation efficiencies of 6.0%, 2.8%, and 1.6% were calculated for MDS loaded with PLGA nanoparticles with average diameters of 76, 119, and 193 nm, respectively.

Figure 4.

Fluorescent microscope images of MDS loaded with A) 76 nm, C) 119 nm, and E) 193 nm-DiO loaded nanoparticles. SEM cross-section images of MDS loaded with B) 76 nm, D) 119 nm, and F) 193 nm DiO-loaded nanoparticles.

2.3. Nanoparticle Release from MDS

The release profiles of nanoparticle payloads from MDS were examined in saline and human plasma (Figure 3B). Similar to previous works, MDS exhibited a biphasic response with a faster release in the first 12 h followed by a sustained release up to 7 days in both media.[28,29] The faster release of nanoparticles in phase I was observed likely because some nanoparticles were encapsulated close to the MDS surface, and the formation of cracks could lead to the outer shell disintegration. In phase II, nanoparticles may be released through MDS erosion. The zero-order release profile in the phase II could cause by the interaction between chitosan-coated nanoparticles and the MDS shell which may inhibit the fast release of nanoparticles.[30] As expected, the payloads were released at a significantly higher rate in the human plasma compared to the saline buffer due to a faster degradation of PLGA in the presence of plasma enzymes.[31–33] For a certain application that requires a payload delivery at a different time frame, the MDS degradation rate could be further modified by using PLGA polymers with a different molecular weight and/or a different lactic to glycolic acid monomer ratio. Typically, PLGA polymers with lower molecular weights and/or higher lactic acid to glycolic acid monomer ratios (i.e., 25:75) can accelerate the release rate.[1,30,34]

The released nanoparticles were found to retain similar sizes likely because they resided within the MDS shell and were not exposed to the medium, and the chitosan coated layer could provide additional protection (Figure 3C). However, their surface charges showed a shift from + 20.5 mV to −14.4 mV (Figure 3D). These charge changes could be due to multiple reasons such as the generation of negatively charged blank nanoparticles during MDS preparation, the presence of small MDS shell debris on the supernatant (−25.1 mV charged PLGA-shell) or simply by detachment of chitosan layers from nanoparticles surface due to shear stress applied during MDS fabrication.

2.4. Cytotoxicity Study

Different concentrations of MDS and chitosan-coated nanoparticles were evaluated for their biocompatibility with human endothelial cells. Figure 5 shows that MDS at all concentrations studied (0.01–1 mg mL⁻¹) do not cause toxicity to the endothelial cells for up to 72 h. No significant difference in cell viability of MDS-treated cells and untreated cells was observed. Similarly, the released chitosan-coated nanoparticles were also found to have no influence on the cell viability. Our results agree with other existing works which reported that PLGA particles are biocompatible with a safe concentration up to 25 mg mL⁻¹.[35–37]

Figure 5.

Cell viability of HUVECs treated with MDS at different concentrations. Blank media (containing no particles) is used as a control. Value = mean ± SD (n = 3). ns = no significant difference at 95% confidence.

2.5. Functionalization of MDS and PLGA Nanoparticles

For targeting inflamed endothelium, MDS and PLGA nanoparticles were labeled with sLe^A, a fast on-rate and off-rate targeting ligand for E-selectin on endothelium, using EDAC-carbodiimide and avidin-biotin chemistry. The amount of sLe^A on the MDS surface was quantified by flow cytometry (Figure 6A) and the presence on the nanoparticle surface was confirmed by an ELISA assay (Figure 6B). The estimated ligand density on microparticles and nanoparticles are 1370 site μm⁻² and 3000 site μm⁻², respectively.

Figure 6.

A) Flow cytometry histogram of sLe^A functionalized MDS labeled with PE-secondary antibodies (blue), and uncoated MDS (gray). B) Fluorescence intensity of an ELISA assay for functionalized MDS and PLGA nanoparticles.

2.6. In-Vitro Flow Adhesion of MDS and Nanoparticles

Targeting and delivery efficiency of MDS and nanoparticles to inflamed endothelial cells were studied in a parallel plate flow chamber (PPFC) (Figure 7A,B). In this study, particle localization was investigated in human laminar blood flow at a wall shear rate of 100 s⁻¹. To allow a direct comparison on the targeting efficacy, both MDS and nanoparticles were labeled with sLe^A. Both MDS and nanoparticles were shown to adhere specifically to the cytokine-activated (inflamed) endothelial cells. Their bindings on the inactivated endothelial cells were minimal, suggesting that the sLe^A-particles have a specific interaction with E-selectin receptors on the inflamed endothelium (Figure 7C,D). As expected, MDS demonstrated a significantly higher localization and adhesion to inflamed endothelial cells relative to nanoparticles. The MDS binding density was 2.7-fold to that of nanoparticles, despite the higher ligand density of nanoparticles relative to MDS. In previous works, particle size was suggested to be a primary factor governing the difference in particle adhesion between nanoparticles and microparticles.[5] Thus, it is not anticipated that the higher ligand density of nanoparticles, if not in favor, would negatively result in the low binding of nanoparticles relative to microparticles.

Figure 7.

A) Circular parallel plate flow chamber. B) Experimental setup used to test particle binding under laminar flow conditions. C) Phase contrast microscopy of MDS particles attached to an IL1-β-activated human umbilical vein endothelial cells (HUVECs) monolayer. D) Adhesion of sLe^A or avidin-coated (control) MDS and PLGA nanoparticles suspended in reconstituted blood (40% RBC) to an IL1-β-activated HUVECs monolayer under laminar flow in a parallel plate flow chamber (wall shear rate = 100 s⁻¹). Value = mean ± SD (n = 3).

The higher binding efficiency of the MDS compared to the nanoparticles agree with our previous findings that the microparticles outperformed nanoparticles in blood flow margination.[5,10] Specifically, Charoenphol et al.[10] reported that microparticles in the size range of 3–6 μm is optimal for vascular-adhesion under typical flow profiles and conditions. Although the adhesion of MDS and nanoparticles in this work was tested in the reconstituted blood flow (red blood cells (RBCs) in saline buffer), the presence of white blood cells (WBCs) in physiological conditions is expected to have minimal impact on the margination of MDS with this size range (4.2 ± 1.4 μm).[23] Once MDS localizes to the endothelium, it was shown that WBCs could disrupt the bound microparticles (5–6 μm) on the endothelium but not interfere or compete with them during their margination.[23] Particle removal on the endothelium can be mitigated by simply increasing the ligand density or using adequate ligand density. In addition to particle size, MDS is expected to have different properties than solid PLGA particles. Although MDS properties were not fully characterized, MDS is expected to have a low modulus and low density due to the capsule-like structure, compared to typical solid PLGA particles. Low modulus particles were shown to have increased circulation time in vivo,[38–41] be able to avoid filtration and phagocytosis,[39,40,42–44] and enhance vascular margination at low to intermediate wall shear rate in vitro and in vivo.[6] Although high density is shown to favor particle margination in blood flow,[7] the density of particle must be high enough to see the differences. For instance, silica particles (with a density of 2.0 g mL⁻¹) demonstrated a higher binding in blood flow (with a density of 1.06 g mL⁻¹) compared to that of polystyrene particles (with a density of 1.05 g mL⁻¹).[7] Thus, we expect that the reduced density of PLGA MDS, with a density of typical PLGA particles in a range of 1.21–1.29 g mL⁻¹, does not significantly affect their marginations. However, further studies on the margination of PLGA with varying density will need to be explored. Our future work will also include testing the MDS functionality with an in vitro assay that allows the observation of nanoparticle release from MDS.

3. Conclusions

In summary, this work has demonstrated the development and characterization of PLGA-based multistage delivery system that could offer improved delivery efficacy of nanoparticles to the targeted vascular endothelium. Overall, this engineered MDS offers a number of advantages as a drug delivery microcarrier. Its low toxicity, tunable size, sustained degradation, and release profiles, and enhanced margination and binding capability to a target site make MDS a promising tool for early detection of complex diseases and an effective drug carrier for robust treatments with minimal side effects.

4. Experimental Section

Materials:

Poly lactic-co-glycolic acid (50:50 and 75:25 PLGA) was acquired from Evonik Industries. Poly ethylene-maleic acid (PEMA), poly vinyl alcohol (PVA), chitosan, Dulbecco’s phosphate buffered saline (DPBS), and N-(3-Dimethylaminopropyl)-N-ethyl carbodiimide (EDAC) were purchased from Sigma–Aldrich. DiO and NeutrAvidin protein were acquired from Thermo Fisher. Sialyl Le^a-PAA-biotin (SLe^a) was acquired from GlycoTech. Cutaneous lymphocyte-associated antigen (CLA-PE) was obtained from Miltenyi Biotec. Fluorescein rabbit antimouse IgG-1 was purchased from Jackson Immunoresearch. MESF calibration bead was purchased from Bangslab. IL-1β was acquired from Fitzgerald. Human umbilical vein endothelial cells (HUVECs) and endothelial growth medium-2 (EGM-2 medium) were purchased from Lonza. Colorimetric cell counting kit-8 (CCK-8) reagent was obtained from Dojindo Molecular Technologies. All reagent grade organic solvents were purchased from VWR.

Fabrication and Characterization of Chitosan-Coated PLGA Nanospheres:

DiO-loaded PLGA nanoparticles were prepared by an oil-in-water emulsion solvent evaporation method.^[45–47] Briefly, an oil phase was prepared by dissolving 0.5 mg of selected chromophore (DiO) and 20 mg of PLGA polymer (75:25 PLGA) in 2 mL dichloromethane (DCM).

Subsequently, the oil phase was added dropwise into 10 mL of continuous water phase containing poly-vinyl alcohol (PVA) and poly-ethylene-maleic acid (PEMA) dissolved in DI water. During the emulsification process, the oil phase was broken into small nanodroplets by high energy shear stress produced by either a homogenizer (FSH 125, Fisherbrand) or a 20 kHz sonicator tip (Q125, QSonica). The emulsion mixture was stirred under a fume hood for 3 h to ensure complete evaporation of organic solvent and solidification of nanoparticles. To remove the remaining surfactants in the water phase from particles, spheres were washed several times in DI water and collected via ultracentrifugation. Serial centrifugation was used to narrow down the particle size distribution.

Subsequently, the formed PLGA nanoparticles containing DiO were coated with chitosan to protect their integrities during the following emulsification process. Briefly, a fresh chitosan solution was prepared by dissolving chitosan in a sodium acetate buffer (pH 4.5). Nanoparticles were suspended in a 0.1 mg mL⁻¹ chitosan solution overnight. After the incubation, the particles were extensively washed to remove unbound chitosan and recovered via ultracentrifugation.

Particle size distribution and concentration were determined using a nanoparticle tracking analysis system (NTA, NanoSight LM10, Malvern Instruments). Nanoparticle surface charge was characterized by zeta potential (Zetasizer Nanoseries, Malvern Instruments). Morphology of PLGA particles was verified by scanning electron microscopy (SEM Neoscope JCM-5000, Nikon). Encapsulation efficiency was determined by measuring light absorption for DiO nanoparticles at 484 nm, and comparing against calibration solutions of known concentrations using a microplate reader (Synergy HTX, BioTek). Encapsulation efficiency was defined as the ratio of total DiO obtained from dissolved particles divided by the initial amount of DiO used during the particle fabrication process. The release of DiO from nanoparticles was determined by incubating particles in phosphate buffered saline at 37 °C and periodically measuring the light absorption of the released DiO suspended in the supernatant using the microplate reader.

Fabrication and Characterization of PLGA-Based MDS:

MDS microcapsules were prepared by a water-in-oil-in-water emulsion solvent evaporation method. Briefly, the chitosan coated PLGA nanoparticles were suspended in a 500 μL of 3% PVA-PEMA (95:5) water phase. The nanoparticle suspension at a fixed concentration was then added into an oil phase containing 30 mg of PLGA (50:50) dissolved in 5 mL of ethyl acetate. The first emulsion was formed via homogenization at 9250 rpm for 20 s (FSH 125, Fisherbrand). The emulsion mixture was transferred into 30 mL of the PVA–PEMA water phase and homogenized at 7800 rpm for 2 min. The double emulsion mixture was then transferred into 60 mL of the PVA–PEMA water phase and stirred at 1200 rpm under a fume hood for 4 h to ensure complete evaporation of the organic solvent. The formed MDS spheres were washed and collected via serial centrifugation (5810R, Eppendorf).

MDS morphology and average diameter were characterized through bright field microscopy (Eclipse, TS2R, Nikon) and scanning electron microscopy (SEM Neoscope JCM-5000, Nikon). A fluorescence microscope (Eclipse, TS2R, Nikon) was used to record bright field and fluorescence images of DiO-loaded PLGA nanoparticles entrapped into PLGA-based MDS. Videos and images were recorded and analyzed to quantify the number of nanoparticles inside of MDS shells. A scanning electron microscope (SEM Neoscope JCM-5000, Nikon) was used to verify the morphology and size of loaded nanoparticles.

To simplify protocol optimization, the number of nanoparticles added to the first emulsion during the MDS fabrication process was fixed at a total of 5 × 10¹¹ PLGA nanospheres. The average number of nanoparticles loaded into each MDS was visually estimated from at least 10 recorded fluorescence microscopy images. Each image contained approximately 10–100 particles. Encapsulation efficiency was estimated as the percentage of the ratio of total encapsulated nanoparticles divided over the initial number of nanoparticles.

MDS Release Study:

The release of nanoparticles from MDS was studied in phosphate buffered saline and human plasma under a continuous rotation at 40 rpm at 37 °C for up to 7 days. At a desired timepoint, the MDS sample was centrifuged to collect the supernatant and resuspended in fresh DPBS+ or plasma to maintain sink conditions throughout the release study. The collected supernatants were analyzed via a nanoparticle tracking analysis system (NTA) to determine the number of nanoparticles released. Blank plasma and DPBS+ (with no MDS) were also subjected to the similar continuous rotation at 37 °C as controls. The sample from the controls was collected at every time point and used as NTA background readings.

Isolation of Red Blood Cells (RBCs) and Cell-Free Plasma from Human Whole Blood:

Human blood was collected via venipuncture from healthy adult donors according to a protocol (IRB2017-0450D) approved by the Texas A&M University Internal Review Board and in line with the standards set by the Helsinki Declaration. All donors provided informed signed consent prior to blood donation. Venous blood was obtained into a 30 mL syringe containing acetate-citrate-dextrose, ACD, as an anticoagulant. The collected whole blood was centrifuged at 500 g for 10 min and was separated into two layers. Cell-free plasma was obtained by centrifugation of the supernatant layer at 2500 g for 10 min to remove platelets and leukocytes. RBCs were collected from the bottom layer and washed with DPBS prior to resuspending in DPBS at a desired concentration (40% v/v or 40% Hct).

Culture of Human Umbilical Vein Endothelial Cells:

HUVECs were pooled and cultured following Lonza’s recommended protocols. Briefly, HUVECs were cultured under sterile conditions until reaching confluence in a T75 flask pretreated with gelatin (0.2% w/v). HUVECs cells were maintained in endothelial growth medium-2 containing EBM-2 basal medium supplemented with human epidermal growth factor (hEGF), vascular endothelial growth factor (VEGF), R3-insulin-like growth factor-1 (R3-IGF-1), ascorbic acid, hydrocortisone, human fibroblast growth factor-beta (hFGF-β), heparin, fetal bovine serum (FBS), and gentamicin/amphotericin-B (GA), in a humidified 5% CO₂ incubator.

MDS and Nanoparticle Cytotoxicity Study:

HUVECs were harvested via trypsinization and seeded at a 5 × 10⁵ cell density in a 96-well microplate. After an overnight incubation, cells were treated with various concentrations of MDS loaded with chitosan coated nanoparticles and incubated for 4, 12, 24, and 72 h. Untreated HUVEC cells incubated under the same conditions were used as a control. After a desired incubation time point, 10 μL of colorimetric cell counting kit-8 (CCK-8) reagent was added to each well and the cells were further incubated for 4 h at 37 °C. Subsequently, the cells were measured for an absorption at 450 nm using a plate reader. The cell viability was determined by comparing the recorded absorption intensity to a CCK-8 calibration curve, generated according to the manufacture recommendations.

Preparation and Characterization of Ligand-Conjugated MDS and Nanoparticles:

MDS capsules and nanoparticles were functionalized with targeting ligands via EDAC-carbodiimide chemistry and avidin-biotin linkages as previously described.[5] Briefly, the carboxylated surface of MDS capsules or nanoparticles was initially modified with NeutrAvidin proteins in EDAC and MES buffer. Subsequently, biotinylated human sLe^a was added to the NeutrAvidin conjugated particles. The functionalized particles were washed in DPBS, collected by centrifugation, and stored at 4 °C in the dark until use.

For MDS capsules, the site density of sLe^a on the MDS surface was determined by flow cytometry (FACSCalibur, Becton, Dickinson and Company). sLe^a-MDS were stained with human PE-labeled CLA secondary antibodies (antiCLA-PE). Unstained MDS, avidin-coated MDS incubated with the secondary antibodies, and isotype controls were used as controls. A number of ligand density was calculated using MESF PE and FITC calibration beads.

For nanoparticles, the ligand site density was evaluated by an ELISA assay.[48] Avidin conjugated PLGA nanoparticles were incubated with biotin-labeled horseradish peroxidase (B-HRP) for 30 min. Particles were then washed with PBS to remove unbounded B–HRP and then transferred into an opaque 96-well plate. Amplex Red fluorescent peroxidase was added to samples. After a 10-min reaction, the fluorescent intensity of untreated (control) and amplex red treated nanoparticles were measured using a microplate reader at 544 nm excitation/590 nm emission. MDS with a desired ligand site density was evaluated with the ELISA assay and used as a reference for estimating the ligand site density of nanoparticles relative to MDS.

In Vitro Flow Adhesion Experiment:

A circular parallel plate flow chamber (PPFC) equipped with a straight channel (GlycoTech, Gaithersburg, MD) was used for in vitro flow adhesion assays. Specifically, a silicon rubber gasket with rectangular cutout was attached to a flow chamber deck. The rectangular cutout in the gasket defined the flow channel and the gasket thickness dictated the channel height. Flow experiment setup was as described in ref. [10] with minor modifications.

An endothelial cell monolayer was prepared by seeding HUVECs onto 30 mm glass cover slips precoated with 1% gelatin cross-linked with glutaraldehyde, and incubated at 37 °C in a 5% CO₂ incubator until confluence, as previously described.[10] A confluent monolayer of HUVECs was confirmed using brightfield microscopy prior to all flow adhesion experiments. The HUVEC monolayer was then activated with IL-1β (10 ng mL⁻¹) for 4 h and attached to the bottom of a circular parallel plate flow chamber (GlycoTech) to form the bottom wall of the flow channel. Next, PLGA-based MDS particles or PLGA nanoparticles coated with sLe^a were suspended in PBS saline or reconstituted blood (40% v/v of human red blood cells suspended in PBS) and introduced into the chamber with a laminar flow regulated through a programmable syringe pump (LEGATO 110, KD Scientific). To determine the effect of particle size on their ability to navigate in a laminar flow and marginate to the vascular wall (with no other extraneous variables), the number of both MDS and PLGA nanoparticles was fixed at 1 × 10⁶ particles mL⁻¹ to enable a fair comparison of their binding efficiency. The wall shear rate (γ_w) in the flow channel was controlled to be 100 s⁻¹ by adjusting the volumetric flow rate (Q) in accordance to Equation (1):

$${\gamma }_{\text{w}}=\frac{6Q}{h^{2}w}$$

where Q is the volumetric flow rate (mL min⁻¹), h is the channel height (254 μm), and w is the channel width (1 cm).

Flow experiments were observed and recorded via a digital camera (Eclipse, TS2R, Nikon) connected to an inverted light microscope (Eclipse, TS2R, Nikon). The microscope was located inside a temperature-controlled incubator in which a temperature was maintained at 37 °C for all experiments. Adhesion of functionalized particles to activated HUVEC cells was evaluated after 5 min of flow. The number of particles bound to endothelial cells was quantified and normalized by the area covered by the camera’s field of view. Data was presented as an average over at least three experiments and includes at least five fields of view per experiment.

Statistical Analysis:

All results are expressed as data mean with a standard deviation (n = 3). Statistical analysis of all quantitative data was performed via Two-way ANOVA with Post-Hoc Tukey analysis using GraphPad Prism (v 8.0.2). Statistical significance is demonstrated as * = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001.