The molecular structures of chitosan, TPP, and DS are shown in (Figure 2). The preparation of chitosan nanoparticles was conducted by adopting well-established protocols 26 27 28 . The electrostatic interaction of positively charged amine moieties in hydralazine and chemically available functional groups of polyanions, such as phosphoric acid in TPP and the sulfate group in DS, is critical to facilitate the encapsulation of hydralazine inside chitosan nanoparticles. The morphological or physical phenomena of hydralazine encapsulation within nanoparticles were characterized by TEM and zeta potential/particle size analyzer as described above. In contrast to solid particles, the chitosan particles are not clearly identified once inside the cell due to some particle-formed aggregates which appeared as large agglomerates. We plan to work next with flow cytometry that could perhaps offer a better opportunity to observe the internalization of fluorescent-conjugated chitosan nanoparticles.

In all experiments, chitosan nanoparticles were formed at an equivalent mass ratio of chitosan to polyanion due to the fact that high or low concentrations of chitosan compared to polyanions tends to decrease the encapsulation efficiency and/or promote aggregation of particles 29 . Unloaded chitosan nanoparticles were measured to have diameters in the range of ~250 nm ~300 nm (Figure 3C). The incorporation of hydralazine caused a slight increase in the mean diameter of chitosan nanoparticles, resulting in an approximately 300 nm ~350 nm range in diameter. The surface charge of unloaded chitosan nanoparticles ranged from 10.78 ± 1.54 mV to -7.16 ± 3.69 mV for Chi-TPP and Chi-DS, respectively. The number of negatively charged groups of the polyanions, TPP and DS, was responsible for this difference, where DS (MW 9,000 ~ 20,000 Da) would possess the predominant amount of sulfate groups per mole compared to the amounts of phosphoric acid of TPP (MW 368 Da) at experimental conditions (pH 3~4). Positively charged hydralazine loading did slightly increase these values, corresponding to 14.51 ± 2.58 mV and -4.84 ± 1.38 mV for Chi-TPP/Hy and Chi-DS/Hy, respectively. Analysis of particle morphology revealed that Chi-TPP nanoparticles exhibited a well-defined spherical shape with a solid and consistent structure (Figure 3A). On the contrary, Chi-DS nanoparticles showed a clustered spherical structure (Figure 3B). The opposite charge of hydralazine and polyanions significantly contributes to concomitant increase in the efficiency of encapsulation, resulting in 15.8% and 23.5% for Chi-TPP/Hy and Chi-DS/Hy, respectively. It is worthy to note that hydralazine entrapment was increased by approximately 35% due to the association with DS compared to TPP. This is possibly attributed to the presence of sufficient negative charge densities in DS, which facilitated the encapsulation of appreciable quantities of hydralazine through the complexation process with chitosan.

The release profile of hydralazine from Chi-TPP and Chi-DS nanoparticles was observed in Krebs' solution over 5 days duration. Apparently, both types of particles showed burst release of 15 ~ 30% of the total determined concentration at 5 hr, which declined to a constant, but slow, release rate over several days, as shown in (Figure 4). It was likely that the rapid release was caused by desorption of hydralazine loosely attached to the surface of chitosan nanoparticles. The relative small size of the hydralazine molecule would not be interfered by the diffusion/dissociation process stemming from the core or the pores of nanoparticles. In addition, the high affinity and hydrophilic nature of chitosan with Kreb's medium provides space for penetration inside the particles to be able to dissolve the entrapped drug. It should be noted that the molecular size of drug and ionic interaction with polyanions would be a major consideration in deciding the rate of drug release.

Initially, functional tests (MTT and LDH exclusion tests) were used to evaluate the response to different chitosan formulations in PC 12 cells suffering from acrolein exposure. We compared the effectiveness of these formulations using entire cell populations in six different groups: 1) a control group, 2) a 100 μM acrolein-exposed group, 3) a 100 μM acrolein-exposed and Chi-DS nanoparticle - treated group, 4) a 100 μM acrolein-exposed and Chi-DS/Hy nanoparticle - treated group, 5) a 100 μM acrolein-exposed and Chi-TPP nanoparticle - treated group, and 6) a 100 μM acrolein-exposed and Chi-TPP/Hy nanoparticle - treated group. First, we tested whether chitosan nanoparticles were capable of sufficiently reducing acrolein cytotoxicity using the MTT assay (Figure 5). In all experiments, the various types of nanoparticles were added to the cell medium after 15 min following the exposure to acrolein. Acrolein-induced cytotoxicity was expressed as a percentage of MTT activity, which is, as explained above, an indicator of aberrant mitochondria functioning. Thus, MTT activity subsequent to 100 μM acrolein exposure was decreased to 3.5 ± 2.9% of controls values. In contrast, the treatment of poisoned cells with Chi-TPP or Chi-DS increased survival to 44.5 ± 12.6%, and 37.8 ± 9.6% of controls after 5 hr, respectively (P < 0.05). Even more significant increase was observed after application of hydralazine-entrapped chitosan nanoparticles. Treatment with Chi-DS/Hy or Chi-TPP/Hy showed very significant increase in cell viability, corresponding to 120.5 ± 26.6% and 108.6 ± 17.1% of control values, respectively (P < 0.001). Hydralazine-loaded chitosan nanoparticles improved cellular oxidative metabolism to the greatest extent - and even better than control value.

As a complementary test, we observed the release of LDH from PC 12 cells following acrolein attack. LDH results were essentially consistent with the values obtained from MTT assays. The cells exposed to 100 μM acrolein caused significant LDH increase, reaching 178 ± 25.4% of control values (Figure 6). Conversely, LDH release was reduced to 117 ± 18% and 115 ± 10% of control values by post-treatment with Chi-DS or Chi-TPP, respectively (P < 0.01). Remarkably, hydralazine-loaded chitosan nanoparticles, Chi-DS/Hy and Chi-TPP/Hy, caused even more reduction in LDH release to a level close to - or even below that of controls (untreated cells; 100 ± 2% of control values), corresponding to 105 ± 18% and 96 ± 19% of control values, respectively (P < 0.001). Additionally, we examined cell mortality using the live-dead cell assay. Consistent with all results described above, control cells in culture (with a characteristic 90 ± 7% survival) dramatically fell to 29.8 ± 4% at 5 hr when exposed to acrolein. Treatment of these poisoned cultures with different types of chitosan nanoparticles improved survival to 60 ~ 70% (Figure 7).

The addition of hydralazine-loaded chitosan nanoparticles to acrolein poisoned cultures was able to inhibit or interfere with the generation of ROS and the associated process of membrane LPO. The levels of ROS increased approximately four-fold after exposure of cells to 100 μM acrolein, corresponding to 120 ± 14% of control values (37 ± 19% (Figure 8A). Remarkably, the addition of chitosan nanoparticles reduced the level of ROS up to 39 ± 14% of control values - indeed close to that of control groups (38 ± 21%, P < 0.01). Consistent with this striking result, the level of LPO was also significantly increased to 305 ± 22% of control values after acrolein-induced injury (Figure 8B). Also consistent with previous results, the LPO level corresponded to 91 ± 5% (P < 0.001) of control values upon treatment with Chi-DS/Hy, which was even lower than that of the control group.

Chitosan, a biocompatible and biodegradable polymer, has been widely tested in a variety of fields for the purpose of developing treatments as diverse as wound healing, lung surfactant additives, and tissue engineering 30 31 32 . Additionally, substantial efforts have been devoted for the development and application of chitosan nanoparticles as vehicles for drug delivery 26 33 34 35 . In this regard, therapeutic efficacy based on chitosan nanoparticle systems confers several advantages: i) nanospheres composed of highly concentrated chitosan molecules could serve as a "fusogen". Previous studies have demonstrated the effect of chitosan derivatives on the fusion of small dipalmitoyl phosphatidylcholine (DPPC) bilayers by inducing the formation of large phospholipid aggregates 36 37 38 , ii) in contrast to better known fusogens including PEG and Poloxamers, chitosan nanoparticles of small size and hydrophilic character are also able to cross the cell membrane to deliver drugs or other molecules to the cytosol while avoiding the rapid uptake by the reticuloendothelial system (RES), as shown in (Figure 9) 39 . It is particularly important that drugs carried inside nanoparticles would be protected from denaturing, enzymatic degradation, or the general biochemical environment, and iii) the encapsulation and release of drugs would be controllable by manipulating several characteristics of the particles such as size, surface functionalization, and loading concentration. Moreover, the preference of the chitosan moiety to target compromised membranes as discussed above further shows the feasibility for the selective localization of a large amount of the drug at a specific and desired site.

Recently, we have learned that chitosan, as a component of an injectable solution, can serve as a novel therapeutic agent after severe compression injury of guinea pig spinal cord (unpublished observations). In such injuries, progressive destruction of cells and tissues occurs after mechanical trauma, however critical anatomy and function was restored by the injection of chitosan. Chitosan accumulation typically occurs around the defect area in cells and tissues by hydrophobic interactions. Conversely, at intact membranes, high surface densities of lipid moieties inhibit the penetration of chitosan 36 37 38 . The mechanism underlying chitosan-mediated membrane sealing and specific targeting of injured membranes is still being investigated; however a clear demonstration of the latter observation is significantly persuasive that this approach holds unexpected merit in dealing with cell and tissue trauma.

The evidence presented in this report suggests the potential usefulness of chitosan - as a nanoparticle base likely due to its two-pronged attack on cell damage. We emphasize that chitosan nanoparticles themselves are capable of restoring cell viability by first mediating the sealing of damaged membrane. We additionally showed the application of chitosan achieved neuroprotection by interfering with the generation of ROS and LPO. This is also due to its membrane reconstruction properties rather than an ability to directly scavenge these toxins. However upon conjugation with hydralazine, its potential therapeutic effects are dramatically enhanced. Chitosan treatment alone did not provide neuroprotection after exposure to acrolein - even in high concentrations (unpublished observations). This fact suggested the inclusion of hydralazine in the chitosan nanoparticle to directly provide this function.

It is widely accepted that traumatic injuries often deteriorate cell membrane integrity through multiple mechanisms involving several biological processes and biochemistries. Steady and progressive membrane rupture and concomitant loss of intracellular contents initially causes serious derangement of ionic species. Especially, the unregulated influx of Ca²⁺ into cytoplasm plays a key role in destabilizing and collapsing the cytoarchitecture as well as membrane integrity by activating numerous catabolic physiological and biochemical processes such as ROS mediated LPO. There is little evidence that free radicals (with half lives less than seconds - even fractions of a second) are directly toxic to cells. It is the downstream end product of ROS stimulated LPO that produces actual toxins such as a variety of aldehydes - acrolein being the most potent. Compared to free radicals, acrolein has a half-life many orders of magnitude longer 40 . Indeed, highly reactive acrolein actively forms biomolecule conjugates with proteins, DNA, and glutathione, which further stimulates the generation of ROS and LPO. The cytotoxicity of acrolein is known to be concentration-dependent. After exposure to 75 - 100 μM acroleoin, a majority of the PC 12 cell population was dead within 4 hr 41 . In previous reports, we have shown that the toxicity of acrolein can be significantly diminished after application of an acrolein-trapping agent such as hydralazine. Hydralazine, a well-known nucleophile drug, is able to immediately inhibit acrolein cytotoxicity by forming "hydrazones" through a scavenging reaction (Figure 1) 42 43 44 .

There are many studies to take advantage of colloidal particles as drug carriers. The use of colloidal drug carriers can facilitate not only site-specific drug delivery in a designated region but also facilitate drug bioavailability and therapeutic efficacy. On the basis of these considerations, a variety of biocompatible and biodegradable materials such as synthetic or natural polymers, lipids, or solid (metal, semiconductor, magnetic, or insulator) components have been intensively investigated 45 46 .

First, liposome-based colloidal carriers have already commercialized. Despite these advances, liposomes still have some technical weakness with regard to the limitations in reproducibility, stability, and low drug encapsulation efficacy 47 48 . Solid particles such as silica particles have been proposed as alternative drug carriers that could potentially overcome such conventional problems. Silica particles with well-established synthesis methods are easily incorporated with other substances through surface modification, bioconjugation, and entrapment 49 .

On the other hand, the use of biodegradable polymeric drug carriers, either synthetic or natural, shows great potential in controlled drug release by tuning their biodegradability. Of these, chitosan is the one of the most extensively studied - in particular, since it is composed of reactive hydroxyl and amine groups, it tends to adhere well to negatively charged surfaces, thus increasing cellular transport 50 .

Because it is an antihypertensive drug, the systemic exposure of hydralazine will provoke problems related to blood pressure management in hypotensive trauma patients 51 52 . The use of our particulated system is capable of significantly reducing systemic exposure of drug since it is delivered to the intended site of action. In spite of its potent therapeutic effects reported here in vitro, our major challenge is to first, enhance the entrapment efficacy of hydralazine in the chitosan nanoparticle, and second, move the improved system to animal models of spinal cord and brain injury 12 19 .

The first issue arises due to the existence of charge repulsion between cationic chitosan polymer (pKa = 6.5) and positively charged hydralazine (pKa = 7.1). In an attempt to enhance hydralazine encapsulation, two different kinds of polyanions were employed. Interestingly, the encapsulation of hydralazine was particularly dependent on the ionic interaction formed between chitosan and polyanions, where the amount of negatively charged groups of TPP and DS existed per mole was critical for hydralazine association. As a consequence, Chi-TPP formulation showed entrapment efficiency of 15.8% with slightly increased positively charged surface property, whereas Chi-DS exhibited an entrapment ratio of 23.5% with negative surface, characteristics as shown in (Figure 3c). In both cases, after entrapment of hydralazine, the nanoparticle formula showed some increase in diameter and zeta potential value, indicating the presence of drug through inter and intramolecular cross-linking. Both particles, Chi-TPP and Chi-DS, displayed similar trends in release behavior, where an initial burst release could likely be a result of the desorption of hydralazine (which is mostly attached on the surface of particles) and later progressive degradation of chitosan. Such a reduced release could be attributed to an increased viscosity resulting from the concentration ratio of chitosan and the polyanion, where a dense layer of chitosan might cause a greater density of crosslinking and consequently interfere with a swelling effect.

In summary, moving forward with this line of investigation requires exploring the stability of drug-loaded chitosan nanoparticles on the basis of pH or chemical crosslinking for a variety of compounds of interest to us. Various molecular weight chitosan samples or water-soluble chitosan molecules will be employed in this ongoing research. The second challenge requires a diversion towards toxicology studies to decipher any potential side effects or unforeseen problems associated with the chitosan particles or their cargo of hydralazine. Subsequently, we have much experience in the development of behavioral, physiological, and anatomical models of SCI and TBI, as well as access to veterinary clinical models of neurotrauma to move these notions forward.

Chitosan with 85% deacetylation degree and of medium weight (Chi, M.W. 200,000 Da), dextran sulfate (DS, M.W. 9,000 ~ 20,000 Da), and sodium tripolyphosphate (TPP, M.W. 367.8 Da) were purchased from Fluka/Sigma-Aldrich. Two kinds of chitosan particles were synthesized: Chi-DS and Chi-TPP. Briefly, Chi-DS was prepared by complexation of Chi and DS as described previously, where chitosan was dissolved at 0.10% (w/v) with a 1% aqueous acetic acid solution while DS was prepared in deionized water at the concentration of 0.5 mg/ml 26 28 . Equivalent volumes of chitosan and the DS solution were mixed by magnetic stirring at room temperature. Once the nanoparticle suspension started to form, the mixture was stirred for another 20 min. The formation of Chi-TPP nanoparticles was initiated by ionic gelation mechanism based on the interaction of cations and anions 27 33 . Chi-TPP nanoparticles were formed spontaneously when equal volume of Chi (1.75 mg/ml) and TPP (2 mg/ml) solution were prepared and stirred at room temperature. Hydralazine-loaded chitosan nanoparticles were then immediately prepared by incorporating equivalent volume of a Chi acidic solution (1.75 mg/ml) and an aqueous TPP solution (2 mg/ml) or aqueous DS solution (0.5 mg/ml) containing hydralazing (1 mg/ml) while stirring with a magnetic bar.

Particle size and zeta potential measurements were carried out with a zeta-potential/particle size analyzer (Zetasizer). To begin, samples were diluted in deionized water and measured in an automatic mode. All measurements were performed in three ~ five repetitions. The morphology of chitosan nanoparticles was observed by transmission electron microscopy (JEOL 2000FX).

The amount of hydralazine encapsulated in the chitosan nanoparticles was measured by UV spectrometry following centrifugation of the samples at 15000 rpm for 30 min. The difference between the total amount of hydralazine used for the formation of chitosan nanoparticle loaded with hydralazine and untrapped hydralazine in the supernatant solution was calculated to assess the efficiency of encapsulation.

To observe the release behavior of hydralazine from chitosan nanoparticles, a modified Krebs' solution (pH 7.2) that contained 124 mM NaCl, 2 mM KCl, 1.2 mM KH₂PO₄, 1.3 mM MgSO₄, 2 mM CaCl₂, 26 mM NaHCO₃ was used. The release of hydralazine suspended in this Krebs' solution was observed as a function of the concentration of incorporated hydralazine. The released hydralazine was extracted at a different time-interval and centrifuged to permit measurement by UV spectroscopy. The concentration was then calculated by linear equation to determine the hydralazine release curve.

PC 12 cells with a density of 1 × 10⁶ cells/mL were grown in Dulbecco's modified eagle's medium (DMEM; Invitrogen) supplemented with 12.5% horse serum, 2.5% fetal bovine serum, 50 U/ml penicillin, and 5 mg/ml streptomycin - at an incubator setting of 5% CO₂ and 37°C. After trypsinization and centrifugation, cell pellets were resuspended in Hank's balanced salt solution (HBSS) for the exposure to acrolein and subsequent treatment with chitosan nanoparticles. Acrolein (100 μM) was prepared fresh daily in PBS solution. The Chi-TPP/Hy or Chi-DS/Hy suspensions were applied at a concentration of 20 μl/ml in medium with a delay of 15 min after the application of acrolein.

The reduction of nicotinamide adenine dinucleotide (NAD) by the presence of LDH results in the formation of a tetrazolium dye. The interpolation of the production of the dye can permit determination of the amount of LDH by standard spectrophotometric techniques. We use the LDH assay (Sigma) to assess cell membrane integrity after exposure of cells to the toxin acrolein. The amount of release of LDH from the cytoplasmic compartment to the supernatantis indicative of membrane permeability. Ordinarily, there would be little to no loss of LDH from intact cells. However in this assay, the necessary handling of cells results in some "background" loss of LDH which must be taken into account relative to the damage produced by acrolein-mediated attack on membranes. The background absorbance measured at 660 nm was subtracted from the reading at 492 nm. Cells were grown in 12-well plates at a density of 1 × 10⁶ cells/well in HBSS were used for LDH release into the medium and determination of total LDH, respectively.

Upon exposure to acrolein, the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide tetrazolium (MTT, Sigma) assay was used to determine the overall viability of the sample cells as it is an indicator of oxidative metabolism. The assay assesses the activity of a mitochondrial dehydrogenase enzyme which is detectable only from viable cells by the color change of tetrazolium rings from pale yellow MTT to dark blue formazan crystals. Quantification of these shifts was evaluated by spectrophotometric measurement. PC 12 cells were seeded in 12-well plates at 1 × 10⁶ cells/well in HBSS. MTT reconstituted in PBS was added to each well. After incubation, formazan crystals were pelleted by centrifugation, and dissolved in a MTT solubilization solution. The absorbance was read at 550 nm minus the background at 660 nm.

To assess the impact of acrolein exposure on living cells, the determination of the number of live and dead cells in the total cell population was performed by the application of a two-color fluorescence assay (Biovision Inc.). This assay can selectively stain the live and dead cells using calcein AM (Cal AM) and ethidium homodimer (EthD-1). Cal AM typically crosses the cell membrane into living cells and consequently the cleaved calcein fluorophore is released which possess a strong green fluorescence with an emission wavelength of ~525 nm and an excitation wavelength of ~485 nm. In contrast, EthD-1 can only diffuse into dead or dying cells and produces a red fluorescence at ~625 nm when excited at ~525 nm. The test PC 12 cells were cultured in 12-well plate at 1 × 10⁶ cells/well in HBSS, and subsequently incubated in Cal AM and EthD-1 for 5 minute. Using fluorescent microscopy, cell viability was calculated and expressed as a percentage.

The production of reactive oxygen species accelerates as a result of cell membrane damage which in turn drives LPO and toxic aldehyde production. Therefore a comparison of the levels of ROS and LPO following acrolein-exposure and post-treatment with chitosan nanoparticles provides clues for examining the state of cell deterioration in response to oxidative stress. The cultured cells were immersed in 1 mL of PBS with hydroethidium (HE, Invitrogen) at a final concentration of 1 μM for 10 min in the dark. Concentrations of ROS weredetermined by the chemi-luminescence assay based on the conversion of hydroethidine (HE) to ethidium (E⁺) in the presence of intracellular superoxide anion (O₂ ^-). Subsequently, samples were analyzed with epifluorescence on an Olympus BX61 microscope using a standard rhodamine cube (545/590 nm excitation/emission respectively), and quantified using Image J software (NIH) to measure the amount of ethidium bromide uptake. LPO was measured using a lipid hydroperoxide assay (Cayman Chem. Co.). Once cells were exposed to acrolein, the level of hydroperoxides was directly measured utilizing their redox reactions with ferrous ions. Subsequently, the absorbance of each sample was read at 500 nm using a spectrophotometer (SLT spectra plate reader). Lipid peroxidation was calculated and expressed as a percentage of control values.

One-way ANOVA was used to determine the statistical significance between control and experimental groups using InStat software. Results are expressed as mean ± SD. P ≤ 0.05 was considered statistically significant.