Engineering Delivery of Nonbiologics Using Poly(lactic-co-glycolic acid) Nanoparticles for Repair of Disrupted Brain Endothelium.

Traumatic brain injury (TBI) is known to alter the structure and function of the blood-brain barrier (BBB). Blunt force or explosive blast impacting the brain can cause neurological sequelae through the mechanisms that remain yet to be fully elucidated. For example, shockwaves propagating through the brain have been shown to create a mechanical trauma that may disrupt the BBB. Indeed, using tissue engineering approaches, the shockwave-induced mechanical injury has been shown to modulate the organization and permeability of the endothelium tight junctions. Because an injury to the brain endothelium typically induces a high expression of E-selectin, we postulated that upregulation of this protein after an injury can be exploited for diagnosis and potential therapy through targeted nanodelivery to the injured brain endothelium. To test this hypothesis, we engineered poly(lactic-co-glycolic acid) (PLGA) nanoparticles to encapsulate therapeutic nonbiologics and decorated them with ligands to specifically target the E-selectin. A high level of the conjugated nanoparticles was found inside the injured cells. Repair of the injury site was then quantitatively measured and analyzed. To summarize, exploiting the tunable properties of PLGA, a targeted drug delivery strategy has been developed and validated, which combines the specificity of ligand/receptor interaction with therapeutic reagents. Such a strategy could be used to provide a potential theragnostic approach for the treatment of modulated brain endothelium associated with TBI.

Introduction

The blood–brain barrier (BBB) is a unique, selective barrier formed by the brain endothelial cells (BECs). They form complex tight junctions that include structural proteins such as zonula occludens protein 1 (ZO-1 and ZO-2). It has been shown that ZO-1 protein anchors occludins and claudins to the actin cytoskeleton in the cell[1] and allows for paracellular transport to be modulated in response to different stimuli.[2] Another key protein, vascular endothelial cadherin (VE-cad), is an adherent junction protein specific to endothelial cells.[3] Claudin-5 is also a commonly used marker of tight junction formation in monolayers of brain microvascular endothelial cells.[4] Among these tight junction proteins, ZO-1 has been implicated as an important scaffold protein and contains multiple domains that bind a diverse set of junction proteins.[5]

A model with the capability to study structural and physiological alterations in the BECs’ tight junctions in response to mechanical injury has been developed and characterized.[6] Shockwaves were introduced, and micronsized bubbles were formed by propagating shockwaves. The collapse of highly pressurized microbubbles near a monolayer of BECs generated a mechanical injury to the cells. This process is often referred to as microcavitation. Using brain phantoms, microcavitation has been repeatedly demonstrated and may also be directly observed in vivo.[7] The brain injuries associated with mild traumatic brain injury (mTBI) are often not properly diagnosed although sophisticated imaging modalities and detection techniques are available, posing clinical challenges to treat mTBI. It could be attributed to the size of brain lesions that are below the resolution limit of various imaging modalities. Indeed, disruption to the brain endothelium induced by blast shockwaves can be <1 mm in diameter.[6] Further, amplification of pathophysiological responses may be accumulated over time. This leads to an engineering challenge to develop diagnostic and therapeutic techniques to identify the injured BECs and deliver reparative drugs to repair the lesion and restore the BBB structure and functionality. The major goal of our study is to specifically target (diagnostic) and repair (therapeutic) the injured brain endothelium, or stated in another way, the development of a theragnostic treatment.

Nanoparticles (NPs) can function as both diagnostic and therapeutic agents (i.e., theragnostic)[8] and may be used as tools for delivering therapeutic drugs to the specific targets, hence minimizing systemic administration of drugs and potential toxicity. In addition, biocompatible and conjugated NPs may avoid being removed from the circulation and thereby enhance the number of NPs that can reach the intended target.[9] There is a wide spectrum of applications of nanotechnology in pharmaceutical formulations, which is impacting the scientific landscape of prevention, diagnosis, and treatment of diseases. Various pharmaceutical nanotechnology-based systems such as liposomes, carbon nanotubes, quantum dots, dendrimers, polymeric nanoparticles, metallic nanoparticles, and others have brought about revolutionary changes in drug delivery.[10,11] Nanosized objects can be transformed in numerous ways to alter their characteristics. Active targeting by modifying the NP surface with peptides or antibodies has been shown to improve the cell specificity.[12] As an example, poly(lactic-co-glycolic acid) (PLGA) NPs could be engineered, decorated, and loaded with drugs to specifically target the injured tissue and induce repair and restore the tissue’s structural and functional integrity.[13] In addition, NP-based strategies have been developed to bypass the blood–brain barrier and transport molecules into the brain to prevent secondary and long-term damage associated with TBI.[14−16] In this study, PLGA NPs were fabricated, loaded with drugs, and conjugated with P-selectin glycoprotein ligand-1 (PSGL-1) to specifically target the injured tissue and induce repair and restore the tissue’s structural and functional integrity. PSGL-1 is a glycoprotein found on white blood cells and endothelial cells. PSGL-1 can bind to all three members of the family (P-selectin, E-selectin, and L-selectin) that is part of the broader family of cell adhesion molecules. PSGL-1 has been tested in clinical trials on patients with pediatric asthma, multiple trauma, and cardiac indications.[17,18] Interestingly, the P-selectin expression peaks relatively early after activation (∼10 min), while the E-selectin expression continue to increase even after 6 h.[19,20] Therefore, conjugating NPs with PSGL-1 is expected to lead to an early phase binding of the NPs to the cells via P-selectin, followed by more sustained binding to E-selectin.

In this paper, we present experimental results that show the cell adhesion molecules (e.g., E-selectin) are upregulated in response to shockwave-induced lesions in the tissue-engineered brain endothelium model. The injured brain endothelium can be repaired by delivering PLGA NPs specifically designed to bind to E-selectin on the injured endothelial cells, internalized and release the contents of two nonbiologics for structural and functional restoration. The surfactant poloxamer 188 (P188) and antioxidant N-acetylcysteine (NAC), both FDA-approved, were encapsulated into the NPs to repair the injured endothelial cells. It appears evident that the NPs conjugated with specific ligand and loaded with P188 and NAC can target, diagnose, deliver therapeutic drugs, and enhance the restoration of the brain endothelium. Moreover, the biosystem we developed for this study to validate the efficacy of nanodelivery to repair the brain endothelium is easy to engineer and reproducible. It may provide an inexpensive platform to rapidly screen potential theragnostic nanodelivery of drugs and reagents to restore the brain endothelium.

Results and Discussion

The overall experimental strategy and aims of the current study are outlined in Figure . We developed an in vitro brain endothelium model to create shockwave-induced, localized submillimeter-sized lesion in which cells were detached by the mechanical trauma. The injured endothelial cells around the periphery of the lesion were first identified and then targeted by engineering NPs. The PLGA NPs were decorated with specific ligands to E-selectin and encapsulated with at least two nonbiologic compounds (P188 and NAC) that have been proven to demonstrate reparative effects. We quantified the extent of endothelium repair by examining the closure of the lesion, at least the partial reformation of the tight junctions, and then measuring the permeability of tracer molecules (10 kDa) through the brain endothelium model to test the biotransport functionality.

Figure 1

Graphical illustration demonstrating the proposed theragnostic approach. Injuries were simulated by either a mechanical trauma or using an inflammatory factor for comparison purposes. Injured endothelium upregulates the E-selectin expression that can be targeted by decorating nanoparticles with specific ligands and deliver drugs to the injured endothelial cells to restore the structural and functional integrity of the brain endothelium.

E-Selectin Expression

Targeting injured inflammatory endothelial cells was probed by validating the upregulation of E-selectin. BECs were incubated with an inflammatory factor (TNF-α) to generate inflammatory responses and used it as a model for the disruption of the brain endothelium. The rationale was two-fold. First, a trauma was generated using a powerful cytokine and its effects on the brain endothelium can then be determined and compared to those induced by a mechanical trauma (microcavitation). Second, suitable cell surface molecules would have to be identified for targeting and delivery of some compounds with potential reparative capability. After a treatment of BECs with TNF-α (10 ng/mL) for 4 h, time-dependent expressions of both E- and P-selectin were monitored because the selectins are expressed preferentially in the injured, inflamed, diseased, or disrupted endothelial cells.[21,22] We observed a high expression of P-selectin within an hour after the TNF-α administration, but E-selectin expression was not readily evident (Figure C). After 3 h, we detected higher expressions of both E- and P-selectin (Figure E,F). However, after 6 h, the expression of P-selectin was visibly diminished (Figure G), but the expression of E-selectin still remained elevated (Figure H) and persisted up to 24 h (Figure J). Based on these results, E-selectin appears as a more suitable candidate and therefore was chosen as a molecular target of injured endothelial cells. Because therapeutic interventions for brain trauma are typically administered at later stages, it may be clinically more relevant to focus on the markers that may be detected at earlier stages but are sustained until later stages. The expression of E-selectin appears to satisfy these criteria, providing feasibility to targeted delivery of NPs to the injured endothelial cells. Next, the E-selectin expression was then examined in response to a mechanical trauma (microcavitation). One of the characteristics of this particular mechanical injury was shown to create a lesion in which the cells are detached following the collapse of highly pressurized microbubbles.[6,23,24] This mechanically induced event is explicitly demonstrated by the formation of approximately a circular area (∼100 μm diameter; Figure L) in which endothelial cells are detached. There are two observations to note. First, blood cells, toxins, and large molecules from the circulating blood can easily diffuse through the lesion and could be responsible for the sequelae associated with blast TBI. Second, the endothelial cells around the periphery of the lesion were still attached to the substrate and showed a high expression of E-selectin that remained noticeably upregulated after 3 h after the mechanical trauma (Figure M). The fluorescence image results were confirmed by western blot analysis that showed an upregulation of E-selectin (Figure N). These results were important to validate that E-selectin is a viable target for nanodelivery to potentially treat the mechanically induced lesion.

Figure 2

Fluorescence images of time-dependent E- and P-selectin expressions post TNF-α stimulation and in response to microcavitation and western blot protein analysis of E-selectin after blast exposure. (A, B, K) Control cells cultured in DMEM with 1% serum showed essentially no P- or E-selectin expression. (C–J) After a 4 h stimulation with TNF-α (10 ng/mL), the mouse BECs were further incubated in fresh 1% serum medium for 1 (C, D), 3 (E, F), 6 (G, H), and 24 h (I, J), and E- and P-selectins were visualized. After exposure to the blast-induced mechanical trauma, cells were incubated in fresh 1% serum medium for (L) 1 or (M) 3 h, and E-selectin was immunofluorescently stained. Nuclei were visualized blue using DAPI. Bars = 50 μm. (N) Western blot protein analysis showing an upregulated E-selectin protein level. Full-length western blot is included in Supporting Information, Figure S1.

We next tested the potential of reparative effects of two nonbiologics. One is a triblock biocompatible polymer, known as poloxamers (P188), that was approved by FDA for numerous applications.[25−28] Because of the structural resemblance of P188 to the phospholipid bilayer, it has been shown to reseal the damaged cell membrane and facilitate the repair of injured endothelial cells.[29] The other compound is NAC, which is also an FDA-approved antioxidant that is associated with health benefits.[30,31] Both P188 and NAC are nonbiologic and have been shown to attenuate the potential harmful effects of oxidative stress. In order to demonstrate and establish the effects of P188 and NAC in response to oxidative stress, BECs were stimulated by TNF-α (10 ng/mL) for 4 h and then treated with NAC or the combination of P188 + NAC. Using a fluorophore (MitoSOX) that preferentially binds to superoxide, we recorded and quantified the extent of oxidative stress in response to TNF-α. Oxidative stress is an important determinant of endothelial injury,[32] which influences a number of cellular responses by activating several intracellular signaling cascades in endothelial cells, leading to the progression of vascular diseases.[33] As illustrated in Figure , there was essentially no superoxide production in control cells (Figure A). TNF-α induced a significant increase in oxidative stress and, if left untreated, the superoxide level could persist (Figure B). Treatment of the cells with either NAC alone (Figure C) or the combination of NAC + P188 (Figure D) showed a noticeable attenuation of superoxide. As expected, NAC was found to be more effective than P188. More interestingly, when the combination of P188 + NAC was applied, oxidative stress within the cells was diminished only negligibly (Figure E). This result suggests that, although P188 can also reduce the reactive oxygen species (ROS) levels as has been demonstrated,[6] an antioxidant would act faster to suppress the oxidative stress, and then reparative effects of P188 are expected to follow (shown below). This sequence of events may be engineered to accommodate short- and long-term beneficial responses. Another important role of BECs is to regulate biotransport properties across the blood brain barrier. The mouse primary BECs used in this study were characterized by visualizing tight junction proteins (ZO-1). Tight junctions in control cells are clearly demonstrated, showing tightly packed ZO-1 proteins (Figure F) and a well-defined monolayer of BECs. When the cells were exposed to TNF-α, the tight junction proteins (ZO-1) diminished considerably and disorganized (Figure G). However, when the cells were treated with the NAC + P188 cocktail, the tight junctions appeared to have been partially but noticeably restored (Figure H). This is consistent with the restoration and repair of tight junctions by P188 in response to a mechanical trauma and recovery of the permeability through the brain endothelium.[6] Disrupted tight junctions in response to inflammatory insult may be reversed by treating the cells with the combination of NAC + P188.

Figure 3

Superoxide level was visualized using MitoSOX after incubation of cells for 30 min at room temperature, and ZO-1 found in the tight junction was imaged. (A) Control cells showing essentially no superoxide expression. (B) High level of superoxide in response to TNF-α, if left untreated. (C) NAC treatment suppresses the superoxide expression. (D) When treated simultaneously with NAC + P188, the diminished superoxide level remained unchanged. Nuclei were stained with DAPI (blue). Bars = 50 μm for the panels A through D. (E) Quantitative analysis of superoxide expressions. Data represent mean ± SD. * p < 0.05 when compared to the control. n = 6. (F) ZO-1 expression in control BECs. (G) Image of disrupted ZO-1 distribution and expression in response to TNF-α. (H) Potential therapeutic efficacy of NAC + P188 is demonstrated in partially restoring tight junction after exposure to an inflammatory disruption. Bar = 50 μm for the panels F through H.

Having validated the effects of P188 and NAC on the attenuation of oxidative stress and restoration of tight junctions, the next challenge was to encapsulate them into nanoparticles for targeted delivery to injured endothelial cells only. We used a double emulsion method[34,35] to generate PLGA nanoparticles. PLGA has been used in a host of FDA-approved therapeutic devices with proven biocompatibility and extensively studied as delivery vehicles for DNA, drugs, proteins, and peptides.[36] Also, the physical properties of the polymer can be tuned by controlling the molecular weight and the ratio of lactide to glycolic acid. Its concentration can be manipulated to achieve the desired dosage and release interval.[37,38] Fabricated PLGA NPs were characterized under four different conditions using transmission electron microscopy (TEM) and dynamic light scattering. To prepare the NPs for this study, PLGA-50-50 (24,000–30,000 MW) was used for fabrication. TEM images indicated that all NPs maintained a uniform size (Figure A–D) and ζ potential (Figure E). The loading efficiency values for P188 and NAC were also determined to be 85 and 28% by dissolving the NPs and measuring the contents for the estimation of loading efficiency. The low loading efficiency of NAC, when compared to that of P188, is attributed to factors such as the molecular weight of NAC (∼40× less than that of P188), which is expected to cause rapid diffusion from the nanoparticles.39

Figure 4

TEM images and characterization of PLGA NPs. Transmission electron microscopy images of PLGA NPs: (A) blank PLGA NPs, (B) P188-loaded PLGA NPs, (C) NAC-loaded NPs, and (D) P188 + NAC-loaded PLGA NPs. (E) Characterization of the physical properties of PLGA NPs under different loading conditions. No statistical difference was observed among the NP sizes under four loading conditions. Data represent mean ± SD, n = 6.

Release kinetics of P188 and NAC from PLGA NPs were next determined. The amount of drug loaded in the delivery particle plays an important role in the rate and duration of drug release. It is speculated that particles with a higher drug content possess a larger initial burst release than those with a lower content because of their smaller polymer to drug ratio.[40] Hydrophobic interactions and rapid degradation of particles have also been shown to play a role in the burst release.[41] For P188, most of the release took place after day 3 and then sustained from day 7 until day 28 (Figure ). NAC release was rapid as most of its release was achieved within 24 h and approached near 100% release in 7 days; hence, no additional measurements were performed. The faster release rate of NAC can be attributed to its molecular weight and the porosity of NPs.[39,42] However, the rapid release of NAC may be essential for the aims of this study because it can induce early attenuation of ROS before the onset of P188 therapeutic potential that might include promoting cell migration and proliferation and therefore spanning different time scales to trigger reparative machineries inside the cell. A linear regression analysis showed that the time required to reach a half of the maximum release for NAC was ∼7.7 h, whereas that for P188 was ∼5 h. Although P188 has a larger molecular weight, the loading efficiency was greater than that of NAC (85% vs 28%). Thus, the initial burst release of P188 was not unexpected.

Figure 5

Time-dependent release of NAC or P188. (A) Release of NAC (circles) and that of P188 (squares) from NPs were monitored as a function of time. (B) First several data points were plotted separately to determine the time required to reach the half of maximum release. Linear regression was used to fit the data to a line originating from the origin. Data represent mean ± SD, n = 6. Some error bars were smaller than the symbol sizes used to illustrate the release kinetics of two nonbiologics.

Conjugation, Binding, and Cellular Uptake

To validate conjugation and cellular uptake of conjugated NPs, BECs were incubated with NPs that were conjugated with PSGL-1 (specific ligand against E-selectin). The conjugation efficiency was calculated based on the previous method[43] and determined to be 41.86 ± 0.13%. The binding and internalization of PSGL-1-conjugated PLGA NPs into TNF-α activated BECs were next determined. Fluorescently labeled (FITC) PSGL-1conjugated PLGA NPs were internalized by activated endothelial cells (Figure A), whereas TNF-α activated cells but incubated with unconjugated PLGA NPs had a much diminished, detectable internalization (Figure B). To determine whether the resulting conjugated FITC-PLGA NPs retain the target specificity to the E-selectin-expressing cells, we incubated nontargeting FITC-PLGA-NPs with activated and inactivated endothelial cells for 6 h and detected a very small number of FITC-PLGA-NPs in the endothelial cells (Figure C,D). Protein concentrations were used to normalize and quantify the internalization of NPs (Figure E). A time-dependent study was carried out next to determine the duration of time needed for most NPs to be internalized. The results showed that the longer the incubation time, the higher the number of NPs internalized, as expected. Nearly 80% of the conjugated NPs were internalized after 6 h, and it approached 100% internalization after a 24 h incubation (Supporting Information, Figure S2). The reparative effect of the conjugated P188 + NAC NPs in response to mechanically induced lesion was next tested. In control cells that were not exposed to microcavitation, only a small number of NPs was visualized (Figure F). When exposed to mechanical trauma, the cells at the periphery of the cell detachment area showed an increased level of NPs in these cells, suggesting an abundance of internalized (P188 + NAC)-encapsulated NPs (Figure H). This represents a set of confirmatory experiments that again highlights the specificity of NPs targeting the injured endothelial cells.

Figure 6

Internalization and targeting efficiency studies of conjugated and unconjugated PLGA NPs. (A–D) Fluorescence images (green) showing the specificity of PSGL-1 conjugated PLGA NPs to either inactivated or TNF-α activated BECs. (E) Quantitative analysis of the specificity of PSGL-1-conjugated PLGA NPs. Data represent mean ± SD of three to six independent experiments. * p < 0.05 compared to the solid black bar. (F–H) Targeting efficiency study of conjugated PLGA NPs to endothelial cells exposed to mechanical trauma. (F) Control cells incubated with PSGL-1-conjugated PLGA NPs for 1 h. (G) Endothelial cells exposed to mechanical trauma and then incubated with PSGL-1-conjugated PLGA NPs for 1 h. The area of detached cell is apparent and indicated by a dotted line for better illustration. (H) Quantitative analysis of the specificity of PSGL-1-conjugated PLGA NPs. Data represent mean ± SD of three to six independent experiments. * p < 0.05.

The important question was whether NPs loaded with the two nonbiologics can repair the lesion. Monolayers of BECs were preincubated with a green cell tracker and exposed to the mechanical trauma. Again, the in vitro wound was visibly recognizable (Figure A–C). These cells were (1) left untreated (control), (2) treated with the conjugated P188 + NAC NPs for 12 h, or (3) treated with P188 + NAC free drug for 12 h without using NPs (positive control). Fluorescence images were then recorded (Figure D–F) and compared to those recorded at the initial time point. To accentuate the effects, images were magnified and the in vitro wound closure was determined. Image analyses indicate that a 99% closure of the lesion was achieved by applying the conjugated P188 + NAC NPs, as compared to a 93% closure using the free drug delivery method (Figure J); however, there was no statistically significant difference. If left untreated, the wound closure was only 28%. This provides concrete evidence that the repair of a mechanically induced submillimeter lesion is plausible by delivering decorated NPs to the injured endothelial cells. While NAC was fast-acting to suppress oxidative stress (see Figure ), P188 can promote and enhance cell proliferation and migration into the lesion. It should be noted that P188 is not absorbed by the gut and typically delivered intravenously. Two alternative delivery mechanisms can now be contemplated. First, NPs may be delivered via an intranasal pathway, which can bypass the blood–brain barrier. Second, because the goal of this study is to target the injured endothelial cells, a noninvasive delivery of NPs containing NAC and P188 may be envisioned through the formulation of indigestible capsules for easy administration and storage.

Figure 7

Closure of mechanically induced lesion. In response to a mechanical trauma, the PLGA NPs loaded with P188 + NAC induced proliferation/migration to repair the lesion. Results were compared to positive (P188 + NAC free drug) and negative control experiments (no drug). (A–C) Injury sites prior to treatment. (D–F) After 12 h of treatment, the lesion appears to have been repopulated with cells. (G–I) Magnified images that correspond to the panels D through F. (J) Quantitative analysis of the lesion closure. The conjugated P188 + NAC NPs and P188 + NAC free drug showed a wound closure that was statistically indistinguishable. Data represent mean ± SD. * p < 0.05 when compared to the control (n = 3).

Therapeutic Effect of Conjugated P188 + NAC NPs on BEC Permeability

The therapeutic potentials of conjugated P188 + NAC NPs were further validated by measuring the permeability of large molecules (e.g., 10 kDa Dextran) across the brain endothelium model exposed either to TNF-α or mechanical microcavitation. The diffusion chamber was custom-designed and engineered for permeability studies and has been described in detail elsewhere.[6] The rationale was that, since the mechanical trauma creates a lesion of ∼100 μm in diameter, toxins and blood cells from circulation can easily diffuse through the lesion and adversely affect the brain tissue. Since10 kDa tracer molecules are relatively large, the permeability should be negligible in the control brain endothelium (Figure ). Interestingly, either the mechanical trauma or the TNF-α cytokine increased the permeability significantly by several folds. This apparent leaky endothelium was restored by the treatment of NAC + P188 NPs for 12 h. This provides yet another biophysical evidence that the conjugated and encapsulated NPs restore the functionality of the brain endothelium in addition to the demonstrated structural restoration (see Figure ). Collectively, our results validate the diagnostic and therapeutic potential of administrating conjugated P188 + NAC NPs to the brain endothelium and repair of the disrupted blood–brain barrier.

Figure 8

Permeability of brain endothelium. The permeability coefficient (P) of large 10 kDa Dextran was measured. The increase in the permeability is correlated with the (A) mechanical trauma or (B) TNF-α cytokine insult. When compared to the control (EC monolayer), the permeability was significantly increased. Treatment of injured cells with the combination of P188 + NAC NPs for 12 h restored the permeability. Data represent mean ± SD. * p < 0.05, (n = 3).

Conclusions

The development of targeted therapeutic approaches that have been demonstrated in the current study provide a means to address an unmet clinical challenge for the treatment of traumatic brain injuries (see Figure ). Our findings validated the use of fluorescent NPs to visualize internalization in activated endothelial cells in response to a cytokine or mechanical trauma. We reported for the first time that a novel combination of P188 + NAC has efficacious effects on the repair of brain endothelium. The combination delivery to the injured cells appears to promote cell proliferation and migration into a lesion and reestablished the structural and functional integrity of the brain endothelium. Taken together, these results establish a platform for the development of a novel theragnostic treatment to regenerate the disrupted brain endothelium using nonbiologics. The platform is also expected to contribute to the additional development and validation of therapies for brain pathologies.

Experimental Section

Preparation of P188-, NAC-, and P188-NAC-Loaded PLGA NPs

Drugs were loaded to PLGA NPs by the standard double emulsion method, as previously discussed with minor modifications.[34,35,44] Briefly, the drug solution (water phase-w1) was formed by adding either 10 mg of P188 or 10 mg of NAC or combination dissolved in 100 μL of DI water, emulsifying into 2.5% (w/v) PLGA solution prepared in 1.75 mL of chloroform, and sonicating (∼30 s to 1 min at 10 to 20 W). This primary emulsion was added drop wise into 12 mL of 5% (w/v) PVA (85–88% hydrolyzed, 13 kDa) and sonicated again at 30 W for 5 min. After the overnight stirring to evaporate organic solvents, the nanoparticles were washed and isolated by centrifugation at 15,000 rpm for 30 min, and P188 + NAC-loaded NPs were collected via freeze-drying.

NPs Size, ζ Potential, and Morphology Analysis

One milligram of each prepared NPs was suspended in DI water, mixed, and measured. The mean size, size distribution, and ζ potential of prepared NPs were measured by dynamic light scattering using a ZetaPALS zeta potential analyzer (Brookhaven Instruments, Holtsville, NY). The average effective diameter of the nanoparticles was reported. The data were averages of three measurements. The morphology of the NPs was examined by TEM. Briefly, particles were dropped onto a carbon-coated-on lacey support film and allowed to dry before characterization.

Drug Encapsulation Efficiency

To find the amount of P188 and NAC within the NPs, P188 was fluorescently modified (TAMRA-P188)[6] and loaded into the NPs. The fluorescence intensity of TAMRA-P188 in the supernatant was quantified over time using a plate reader. NAC was measured with high-performance liquid chromatography (HPLC). The results were converted into concentrations, and the loading efficiency values for P188 and NAC were 85 and 28%, respectively, and calculated using the following equation.

In Vitro Drug Release

For the release studies, 1 mg each from all prepared NPs was dispersed in 1 mL of release medium (pH 7.4) each to form a suspension. Each suspension was in three replicates in 1.5 mL centrifuge tubes and incubated at 37 °C. The release buffers were collected after centrifugation at 15,000 rpm for 30 min. The release buffers were replaced with fresh buffer at different periods and subjected to analysis using HPLC. The cumulative release of the model drug from each loaded NPs was plotted against time.

Cell Culture Technique and Characterization

BALB/c mouse primary brain microvascular endothelial cells (MPBMECs, Cell Biologics) were grown in endothelial basal medium-2 (EBM-2) with EGM-2 kit (Lonza) in a flask coated with a gelatin-based coating solution. MPBMECs (6.6 × 104 cells/cm2) were directly seeded unto the cover glass coated with fibronectin (lμl/mL) and cultured with EBM-2. Confluent cells were characterized by the formation of tight junction.

Brain Endothelial Cell Viability

The viability of cells following a 24 h exposure to blank PLGA NPs (blank NPs), P188, NAC, and P188 + NAC NPs was assessed with the MTT assay. It measures the metabolic conversion of the MTT salt (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolinum bromide) by active mitochondrial dehydrogenases. Briefly, after exposure of BECs to blank NPs, P188, NAC, and P188 + NAC NPs, cells were treated with MTT (5 mg/mL) for 1 h at 37 °C. The formazan product generated was solubilized by the addition of 20% sodium dodecyl sulfate and 50% N,N-dimethylformamide and quantified by measuring its absorbance at 490 nm. Untreated cells were taken as control with 100% viability. The resulting sample absorbance was used to calculate cell viability. The results of cell viability experiments were analyzed and shown in Supporting Information, Figure S3.

Endothelial Cells Activation

Confluent cells were exposed to TNF-α (10 ng/mL) for 1 to 24 h and then stained for protein expression. Images were acquired to investigate the E- and P-selectin expression after 1, 3, 6, and 24 h. This period was chosen based on immunofluorescence observations that the earliest expression of inflammatory proteins on activated endothelial cells occurs between 1 to 6 h, and the sustained expression of the selectins after 24 h was of interest to complete this study.

Immunostaining for E- and P-Selectin

Inflammatory markers in the BECs were visualized using E- and P-selectin antibodies. Briefly, cells were activated by mechanical (shockwave) and cytokine (TNF-α) induction, fixed with 4% paraformaldehyde in PBS, permeabilized with 0.1% Triton X-100 for 3 min, and blocked with 3% of bovine serum albumin (BSA) for 1 h. Cells were incubated with the E- or P-selectin primary antibodies (1:500, 4.3 μg/mL, Santa Cruz, SC137054 and SC271267, respectively) and subsequently incubated with Alexa Fluor 488-conjugated goat anti-mouse IgG (1:1000, 2 μg/mL, Santa Cruz, SC516167). Nuclei were counter-stained with DAPI.

Measurement of Mitochondrial Reactive Oxygen Species (ROS)

Superoxide was measured using a MitoSOX Red mitochondrial superoxide indicator, according to the manufacturer’s instructions. Briefly, BECs were cultured on glass cover slides and exposed to TNF-α for 4 h before treating with P188, NAC, or P188 + NAC for 12 h. After treatments, the cells were incubated with 1 μM MitoSOX reagent working solution for 30 min at room temperature. After washing three times with PBS, images were taken with a fluorescence microscope and fluorescence intensities were calculated using the Nikon ND2 software.

Conjugation of PSGL-1 to PLGA NPs

The functional ligand PSGL-1 was bound to the surface of the PLGA NPs via the EDC-NHS chemistry, as described elsewhere with modification.[45] Briefly, 120 mg of EDC was added to 180 mg of NHS and dissolved in 5 mL of MES buffer (0.1 M, pH 4.75). PLGA NPs (20 mg) were resuspended in the EDC-NHS solution and rotated for 2 h at room temperature. After 2 h of incubation at room temperature with rotation, the resulting NPs were ultracentrifuge at 15,000 rpm for 30 min at 4 °C, washed three times with PBS, and then resuspended in PBS (1 mg/mL). Eighty microliters of 500 μg/mL mouse PSGL-1 antibody (Sino Biological, 50770-MCCH) was added to the NP solution and incubated for 24 h at 4 °C. After incubation, the NPs were collected by ultracentrifugation at 15,000 rpm for 30 min. The supernatant was used to determine the peptide conjugation efficiency. Pellets were resuspended in DI water, freeze-dried, and stored for use. The supernatant was quantified using bicinchoninic acid protein assays following manufacturers instructions, and unconjugated peptides (PSGL-1) were analyzed based on BSA standards.

Cellular Internalization of Conjugated FITC-Labeled PLGA NPs

Cells (2 × 104 cells per well) were cultured in a 96-well plate in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 2% fetal bovine serum (FBS) and 1% penicillin at 37 °C for 24 h. FITC-loaded NPs (1 mg/mL) were used for the observation of cellular internalization at 37 °C for 6 h. Conjugated FITC NPs were added to two columns of the 96-well plate (one column activated with TNF-α and the other column inactivated), and unconjugated FITC NPs were also added to activated and inactivated ECs. After internalization, the wells were washed three times with PBS and then lysed by incubating with Triton 1X for 1 h. An aliquot from each well was collected and analyzed for fluorescence intensity and protein absorption with a microplate reader with excitation and emission wavelength at 430 and 485 nm, respectively.

In Vitro Injury Model and Comparative Assessment of P188 and NAC

Primary mouse endothelial cells were seeded in 24-well plates (10 × 103 cells/well) and grown until confluence in complete endothelial cell media. Then, a straight scratch was made with a P200 pipette tip to simulate a wound (in vitro scratch wound model). The cell debris was removed by washing with PBS. Wound size images were taken at time, t = 0. The wounds were exposed to different single treatments: P188 (500 μM), NAC (0.82 mg/mL), and in combination (P188 + NAC) at the recommended therapeutic concentration in DMEM with 1% serum for 12 h. The cells without drug were used as the negative control, and cells treated with 10% FBS were used as a positive control. The closure of the scratch wound was observed under a microscope (5× magnification) 12 h after wounding. To quantify the closure of the scratch wound, the difference between the wound width at time 0 and 12 h was determined. The scratch wound area was measured using the ImageJ software (National Institutes of Health, Bethesda, MD, USA). The wound closure efficiency was expressed as a percentage of the wound area normalized by the initial wound area using the equation below, and the results are shown in Supporting Information, Figure S4.

Measurement of Endothelial Cell Permeability

Details of the custom-designed diffusion chamber have been provided elsewhere.[6] FITC Dextran of molecular weight of 10 kDa was used. The fluorescent intensity calibration standard was obtained by serial dilution of fluorescent Dextrans. Concentrations in the collected samples were determined using a calibration curve. The permeability coefficient was then calculated using[6]where P is the permeability coefficient, C2a and C1a are the concentrations in the abluminal chamber at different time intervals, Va is the volume of the abluminal chamber, A is the surface area of the membrane, Δt is the duration of the steady-state flux, and Co is the concentration in the luminal chamber. The expression (C2a – C1a)/Δt is considered as the slope of the diffusion curve over time.