Glycol chitosan, a derivative of chitosan, can be hydrophobically modified by 5ß-cholanic acid to impart amphiphilic properties that enable the self-assembly into nanoparticles in aqueous media at neutral pH. This nanoparticle system has shown initial success as a therapeutic agent in several model cell culture systems, but little is known about its stability against enzymatic degradation. The goal of this research was therefore to investigate the physicochemical properties of hydrophobically modified glycol chitosan nanoparticles (CNP) under exposure to lysozyme, a ubiquitous mammalian enzyme. Dynamic light scattering (DLS) revealed that the CNP vehicles had an average hydrodynamic diameter of 288.6 nm. These nanoparticles were able to encapsulate bovine serum albumin (BSA) at 87-90% efficiency. Varying the initial CNP: BSA mass ratio affected both particle size and total protein release, indicating a difference in complexation between the negatively charged protein and the positively charged chitosan polymer when more cargo is loaded. After 3 hours of exposure to lysozyme, CNP vehicles degraded to 10-150 nm particles, whereas BSA-loaded CNP degraded more extensively to predominantly 10-20 nm particles. Interesting, however, the encapsulated BSA was not fragmented as shown by SDS-PAGE. Minimal cytotoxicity of the CNP vehicles was observed in both rat osteosarcoma (ROS 17/2.8) and murine osteoblast-like (MC3T3-E1) cells, which internalized the nanoparticles within 2 hours. Taken together, these data suggest that hydrophobically modified glycol chitosan nanoparticles are a promising smart material for use in controlled release drug delivery systems.

Drug delivery; Glycol Chitosan Nanoparticles; Self-assembled

The success of cancer drug delivery systems largely depends on their ability to accumulate in the target tissues with minimal loss while in the blood circulation, and the controlled release of the therapeutic inside the tumor cells [1,2]. For example, reducing the size of the vehicles from micro- to nano-sized (100-200 nm) can slow down the rapid clearance from the circulation by macrophages of the reticuloendothelial system (RES) following intravenous administration [3].To achieve targeted delivery, various delivery systems have utilized the tumor-homing characteristic of nanocarriers [4,5]. This passive strategy exploits the enhanced permeability and retention (EPR) of the tumor microenvironment due to its rapid growth that prevents the formation of fully functional vasculature and proper lymphatic drainage [1,2,6]. To promote site-specific targeted release of therapeutics, one strategy is to design stimuli responsive micelles that are very sensitive to intracellular pH [7]and enzymatic activity [8]. Often this increases the anticancer efficacy of the payload, thus reducing the therapeutic dose and the possibility of side effects [9].

Chitosan has emerged as a prominent nanomaterial for biomedical applications due to its biocompatibility, biodegradability, abundant availability, and low cost [10-12]. Native chitosan has poor water solubility at physiological pH and is limited in its biomedical application [13,14]. This can be improved by chemical modification, such as by introducing a hydrophobic or a hydrophilic moiety [15], an alkyl group [16], cholesterol [17], or poly(ethylene glycol) (PEG) [18-24].

Recently, hydrophobic modification of glycol chitosan, a derivative of chitosan, has been shown to assist in the formation of stable amphiphilic nanocomplexes that are capable of entrapping both hydrophilic peptides [25] and low water soluble drugs [4,26-30].These nanoparticles were successfully used to demonstrate selective accumulation in tumors of mice using a near-infrared fluorescent (NIRF) label [31].

Physiological enzymes can rapidly depreciate system drug circulation, and therefore the in vivo stability of chitosanbased drug carriers is an important criterion for predicting their therapeutic efficacy.Elevated serum levels of lysozyme, a universal mammalian enzyme, have been observed in tumorbearing rats as well as patients with solid tumors [32].Because lysozyme breaks down the sugar linkages of polysaccharide chains by hydrolysis, it may also degrade chitosan-based carriers [10,33].However, the effect of lysozyme on these nanoparticles and their payload is not well known and is the focus of this study.

Glycol chitosan (250 kDa molecular weight, degree of deacetylation> 60%), 5ß-cholanic acid, N-(3-dimethylaminopropyl)- N'-ethylcarbodiimide hydrochloride (EDC), N-hydroxysuccinimide (NHS), lysozyme from chicken egg white, and bicinchonic acid (BCA) assay were purchased from Sigma- Aldrich (St. Louis, MO). Monoreactive hydroxysuccinimide ester of Cyanine 5.5 (Cy5.5-NHS) was obtained from Lumiprobe (Hallandale Beach, FL). Anhydrous dimethyl sulfoxide (DMSO) and bovine serum albumin (BSA) were purchased from EMD Chemicals Inc. (Darmstadt, Germany). Analytical grade methanol was acquired from Pharmco-AAPER (Brookfield, CT). Rat osteosarcoma (ROS) 17/2.8 cells were kindly provided by Professor Miriam Rafailovich and subclone 4 mouse calvaria-derived (MC3T3-E1) cells were obtained from ATCC (Manassas, VA). Fluoromount-G was purchased from Southern Biotech (Birmingham, AL).

Water-soluble glycol chitosan (500 mg) was first dissolved in deionized (DI) water (60 mL). The carboxylic acids groups of 5ß-cholanic acid were then activated with 1.5 equivalents of N-hydroxysuccinimide (NHS) and N- (3-Dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride (EDC) dissolved in methanol (60 mL). The hydrophobic modification of glycol chitosan was carried out by slowly combining to two solutions under stirring for 24 hours at room temperature to permit the 5ß-cholanic acid to conjugate to the chitosan through the formation of amide linkages. The solution was dialyzed in a dialysis membrane (10 kDa molecular weight cut off) for 24 hours against a water/methanol mixture (1:4 v/v) and the following 24 hours against DI water. The purified solution was lyophilized and ground into a fine powder. The conjugates were suspended in either DI water or phosphate- buffered saline (PBS) at a concentration of 1 mg/mL and probe-type sonicated (S-450D Sonifier, Branson Ultrasonics, Danbury, CT)at 90 W for two minutes. This sonication step was repeated three times to ensure that the self-assembly of the conjugates into nanoparticles.

Cy5.5 is one of the most prominent near infrared dyes which can be used for noninvasive imaging of live animals, but to some extent its usage is restricted by insufficient hydrophilicity [34]. To fluorescently label the hydrophobically modified glycol chitosan, Cy5.5-NHS (1 mg) was first dissolved in dimethyl sulfoxide (DMSO, 250 µL) and subsequently added dropwise to a bulk solution of 5ß-cholanic acid-conjugated glycol chitosan. The resulting vividly blue solution was then stirred for 6 hours at room temperature, shielded from light. The Cy5.5-labeled conjugate solution was then dialyzed (10 kDa molecular weight cut off) in DI water to remove unreacted Cy5.5 molecules. Following 2 days of dialysis, the solution was lyophilized and ground to produce a blue powder. In order to protect the fluorescent properties of the Cy5.5 dye, these processes was performed in the dark. The Cy5.5-labeled conjugates were then probe-type sonicated as described above to form self-assembled fluorescent nanoparticles. Prior to cell culture, the nanoparticle solutions were passed through a 0.2 µm syringe filter to remove aggregates and biological contaminants.

The size and zeta potential of the chitosan nanoparticles (CNP) were measured at 25^ºC using dynamic light scattering (DLS) with a 633 nm laser (Malvern Zetasizer Nano - ZS; Malvern, Worcestershire, UK). The zeta potential is the measure of surface charge of nanoparticles and is a particularly influential parameter to understand the electrostatic interaction between nanoparticles. This characteristic is also suggestive of a nanoparticle's affinity to penetrate negatively charged biological membranes. CNP suspensions prepared in water at a concentration of 1 mg/mL were further sonicated in a water bath for 10 minutes at room temperature prior to DLS measurements. Additionally, a separate bath of CNP suspension was stored in water at 4^ºC and particle size was measured using the DLS over the course of 10 days to determine the size stability. Average values were calculated from a minimum of three measurements. Standard error was calculated by dividing the standard deviation by the square root of n, the sample size.

The model protein, Bovine Serum Albumin (BSA), was used to evaluate the encapsulation efficiency of the hydrophobically modified glycol chitosan nanoparticles (CNP). Separate solutions of BSA and chitosan conjugates were first prepared in PBS. Two CNP:BSA mass loading ratios were used: 10:1 and 10:2. Immediately after combining, the BSA and chitosan mixture was probe-type sonicated to form self-assembled nanoparticles. Triplicates of each sample were prepared and analyzed.

To calculate the loading efficiency of BSA, each BSA-CNP solution was first centrifuged at 3,000 RPM for 30 min and the supernatant was separated from the pellet. The concentration of BSA in the supernatant, which represented the total amountof protein that was not loaded into the nanoparticles, was determined using a bicinchoninic acid assay (BCA assay) (QuantiPro T, Sigma-Aldrich). Absorbance at 562 nm was measured with a Microplate reader (ELx800, BIO-TEK, Winooski, VT). The loading efficiency was calculated using the following equation:

Because lysozyme levels throughout the body range in concentration from 0.8 mg/mL to 1.7 mg/mL [35,36], we decided to examine the degradation of CNPs at the highest level and prepared a 1.7 mg/mL solution (pH 7.2) from a 10mg/mL stock solution. A lysozyme inactivation solution was prepared containing 4% sodium dodecyl sulfate (SDS), 1 M Tris-HCl (pH = 6.8), and 20% glycerol in DI water. Lysozyme solution was added to suspensions of CNPs to a final volume of 3 mL, and the mixture was incubated for 3 hours at 37ºC in a rotating incubating shaker at 55 RPM (Excella E24, New Brunswick Scientific, Enfield, CT). Particle size was measured with DLS after 0min, 30min, 60min, 120min, 150min, and 180min of incubation at room temperature. A 0.5 mL aliquot of each sample was harvested at each time point and placed in an eppendorf tube with 0.5mL of inactivation solution prior to being centrifuged at 10,000 g for 10 min. The supernatants were then collected for analysis in the DLS. The total volume of each sample was kept at 3 mL between time points to maintain sink conditions.

CNP vehicles (1 mg/mL) and BSA-CNPs prepared at both 10:1 and 10:2 mass loading ratios were incubated with either water or lysozyme (1.7 mg/mL) for 3 hours and immediately deposited onto 300-mesh copper TEM grids coated with a Lacey film. The air dried grids were imaged on a JEOL 1400 TEM (Peabody, MA) at an accelerating voltage of 60 keV.

Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was used to evaluate the integrity of BSA-loaded CNPs after exposure to either water or lysozyme. BSA loading at both 10:2 and 10:1 were evaluated. Samples were loaded into 10% pre-cast acrylamide gels and run under a constant voltage of 90 V.

Rat osteosarcoma (ROS 17/2.8) and mouse calvaria-derived (MC3T3-E1) cell lines were maintained in a-modified minimum essential medium (a-MEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin streptomycin (Pen- Strep) at 37ºC with a humidified 5% CO2 atmosphere. The cells were seeded onto glass coverslips (12 mm diameter, #1) in 24-well tissue culture plates (Becton Dickinson, Franklin Lakes, NJ) at a density of 10,000 cells/cm2 and medium was refreshed every two days. After attaching for 24 hours, the cells were rinsed twice with serum-free a-MEM to remove any remnants of FBS, and incubated for 2 hours with serum-free a-MEM containing CNPs at the following concentrations: 10 µg/mL, 100 µg/mL, and 1000 µg/mL. The cells were then given fresh a-MEM supplemented with 10% FBS and 1% Pen-Strep and maintained at 37ºC (5% CO2, humidified) for up to 72 hours.

After 4 hr, 24 hr, 48 hr and 72 hr of incubation, the cells were fixed with 3.7% paraformaldehyde and rinsed two times with PBS before being fixing with 3.7% formaldehyde. Cells were rinsed two additional times with PBS and stained with the nuclear dye, 4',6-diamidino-2-phenylindole (DAPI). The area density of the cells was determined by enumerating DAPIstained nuclei on an inverted fluorescence microscope (Olympus IX51). Cell viability was calculated as the number of cells in the treatment group divided by the number of cells in the untreated control group, which consisted of cells that were given only a-MEM (without nanoparticles).

BSA was first loaded into the Cy5.5-HGC conjugates at the mass loading ratio of 10:2 or 10:1 (CNP: BSA) by probe sonication. The final concentration of the BSA-loaded Cy5.5-CNP suspension was 1 mg/mL in a-MEM. After centrifugation at 4,000 RPM, the supernatant was aspirated to remove unbound BSA and the loaded nanoparticles were resuspended in serumfree a-MEM and filtered with both 0.8µm and 0.2µm syringe filters to remove aggregates and biological contaminants.

ROS 17/2.8 and MC3T3-E1 cells were seeded at a density of 10,000 cells/cm2 into 24-well tissue culture plates (Becton Dickinson) containing glass coverslips. The cell lines were given a-MEM supplemented with 10% FBS and 1% Pen-Strep and were allowed to attach in 37^ºC incubator (5% CO2, humidified) for 24 hours. The medium was then removed and the cells were rinsed with serum-free a-MEM. BSA-loaded Cy5.5-CNPs were then added to the medium so that the final concentration was 10 µg/mL, 100 µg/mL or 1000 µg/mL. Cells were incubated with the nanoparticle solution for 2 hours.

Immediately after incubation in the CNP solution for 2 hours, the ROS and MC3T3-E1 cells were fixed with 3.7% formaldehyde, rinsed with PBS and stained with DAPI. The glass coverslips were mounted onto glass microscope slides using Fluoromount- G. Fluorescence images were captured on a confocal laser scanning microscope (Olympus IX81 Fluoview FV1000).

Mean particle diameter of the chitosan nanoparticle (CNP) vehicles suspended in water at room temperature was 288.6 ± 21.8 nm (Table 1), respectively, similar to previously reported values [4,31,37]. The size of these nanoparticles was very stable and did not change significantly after 10 days of storage at 4^ºC (data not shown). Loading BSA onto the CNPs increased the average diameter of nanoparticles by 66% at the 10:1 CNP: BSA mass ratio (to 478.8 ± 9.6 nm) and by 30% at the 10:2 CNP:BSA mass ratio (to 374.8 ± 6.0 nm) (Table 1). The encapsulation efficiency was very similar at both BSA loading ratios, ranging from 87% to 90% (Table 1).

TEM images showed that both unloaded CNP vehicles and BSA-loaded CNPs suspended in water generally had a globular morphology, and that some of the nanoparticles also appeared to be aggregated (Figure. 1c and 1e).These TEM images also confirmed that nanoparticles loaded with with higher protein content (10:2 CNP: BSA mass ratio) were smaller in size than those loaded with lower protein content (10:1 CNP:BSA mass ratio). Previous studies have also reported a similar phenomenon when the RGD peptide was loaded into the same type of nanoparticles [25]. One possible explanation could be that in the case of higher protein content, there is an increase in electrostatic association between its negatively charged carboxylic and the positively charged amine groups on the chitosan backbone. As a result, chitosan is complexed to more strongly to the payload and the overall particle is more compact in size.

Average zeta potential of the CNP vehicles was 13.2 mV (Table 1), which is slightly lower than reported values for 359 nm nanoparticles [38]. The zeta potential of BSA-loaded CNPs was ~13 mV for both loading ratios (Table 1). Because of this similarity, we believe that the larger size of BSA-CNP resulted from the encapsulation of the BSA, and not from aggregation of protein on the outer surface of the nanoparticles.

Protein release from BSA-loaded CNP nanoparticles prepared at the two loading ratios showed an overall stable rate of release (up to only 3% after 7 days) (Figure. 2), indicating the existence of a strong complexation between BSA and chitosan. The rate of release did not vary greatly, as the slopes of both profiles are very similar. We also noticed that ~50% more protein was released from the 10:1 CNP:BSA nanoparticles (3.1%) than from the 10:2 CNP:BSA nanoparticles (2.1%) at the end of the 7-day period (Figure. 2). Taken together, these data suggest that the higher protein release observed for the larger nanoparticles was likely due to a looser packing of the protein at the lower loading ratio.

TEM images showed that exposure to lysozyme greatly reduced the size of the CNPs. After 3 hr, CNPs vehicles appeared as spherical moieties ranging from 10-150 nm in diameter (Figure. 1b). This decrease in particle size is similar to previous reports on degradation of unmodified chitosan in the presence of lysozyme [39] and was expected, as lysozyme is known to hydrolyze the glycosidic linkages between consecutive sugar residues of chitosan [39]. BSA-loaded CNPs became transparent moieties in the presence of lysozyme and contained very small particles approximately10-20 nm in size (Figure. 1d and 1f).

Dynamic light scattering data showed a single peak in the range of 100 nm - 1000 nm for CNP vehicles exposed only to water ("t=0 min" in Figure. 3). This peak is in agreement with the TEM images of the CNP vehicles suspended in water (Figure. 1a). As the nanoparticles underwent exposure to lysozyme, the height of this peak gradually decreased, suggesting that chitosan was being degraded. At the same time, a new peak in the range of 2nm - 10nm began to emerge, whose height and width both increased with incubation time, indicating an increase in the population of low molecular weight fragments. Another peak in the range of 10nm - 50nm also appeared, but its presence was not consistent and only appeared for samples incubated in lysozyme for 30min, 60min or 150min.

SDS-PAGE data (Figure. 4) showed the sample remained in the well when it contained only chitosan (lanes 1,6), which was anticipated because only the negatively charged BSA was expected to migrate toward the positive electrode. Free BSA contained a very strong band around 55 kDa, as well as numerous high molecular weight fragments ranging in size from >170 kDa to 110 kDa (lane 2). When BSA was encapsulated in CNP, there was clearly an absence of fragments in the gel when either water (lanes 3-4) or lysozyme (lanes 8,9) was used as solvent, regardless of the CNP:BSA loading ratio. This indicated that encapsulated BSA was not easily degraded in both cases. When samples were incubated in lysozyme, a strong band appeared at ~15 kDa, which corresponded to the molecular weight of the enzyme (lanes 5-9).

Viability data of both osteosarcoma (ROS 17/2.8) and osteoblast- like (MC3T3-E1) cells showed that the CNP vehicles generally were not cytotoxic in the range of 0.01-1mg/mL, as evidenced by the high cell viability (>90%) in both cell lines within the first 72 hours after exposure (Figure. 5). To ascertain cellular uptake of the nanoparticles, we collected fluorescence micrographs using a confocal laser scanning microscope and observed that Cy5.5-labeled CNP vehicles were successfully internalized by both the ROS and MC3T3-E1 cells within 2 hours (Figure. 6). The nanoparticles also appeared to be distributed more closely around the nucleus of the ROS cells (Figure. 6c) compared to the MC3T3 -E1 cells, in which there was a more widespread accumulation of the nanoparticles throughout the cytoplasm (Figure. 6f).

Glycol chitosan was hydrophobically modified with 5ß-cholanic acid and self-assembled into chitosan nanoparticles (CNPs) with an average hydrodynamic diameter of 288.6 nm. The CNPs were able to encapsulate bovine serum albumin (BSA) at 87-90% efficiency. Initially varying the CNP:BSA mass ratio upon loading affected both particle size and total protein release, which indicates a difference in complexation between the negatively charged protein and the positively charged chitosan polymer. After 3 hours of exposure to lysozyme, CNP vehicles degraded to 10-150 nm particles, whereas BSA-loaded CNP degraded more extensively to 10-20 nm particles, however the encapsulated BSA was not fragmented. Minimal cytotoxicity of the CNP vehicles was observed in both rat osteosarcoma (ROS 17/2.8) and murine osteoblast-like (MC3T3-E1) cells, which internalized the nanoparticles within 2 hours. Taken together, these data suggest that hydrophobically modified glycol chitosan nanoparticles are a promising smart material for use in controlled release drug delivery systems.

We would like to thank the staff of the Center for Functional Nanomaterials, Brookhaven National Laboratory, especially Dr. Dmytro Nykypanchuk for his help with the DLS and zeta potential experiments. Research carried out in part at the Center for Functional Nanomaterials, Brookhaven National Laboratory, which is supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under Contract No. DEAC02- 98CH10886.