Polysaccharide Nanoparticles Bearing HDAC Inhibitor as Nontoxic Nanocarrier for Drug Delivery
Henry Lindemann, Marie Kühne, Christian Grune, Paul Warncke, Susanne Hofmann, Andreas Koschella, Maren Godmann, Dagmar Fischer, Thorsten Heinzel, and Thomas Heinze
Introduction
The systemic administration of pharmaceutically active compounds is often associated with several disadvantages. The compound may be rapidly metabolized, leading to a short serum half-life. Moreover, drugs may cause severe side effects in non-target tissues. To circumvent such drawbacks, the use of bio-based nanoparticles as drug containers is an emerging field in nanomedicine. Nanoparticles may stabilize drugs and achieve targeted delivery. Biocompatibility and biodegradability of the nanoparticles are indispensable prerequisites for materials used as nanocarriers in drug delivery applications.
Polysaccharide-based nanoparticles fulfill these requirements and can be prepared by emulsification, nanoprecipitation, covalent or ionic crosslinking, and polyelectrolyte complexation. Emulsification and nanoprecipitation methods require sufficiently hydrophobic polymers that can be achieved by the covalently bound drug itself (such as ibuprofen or naproxen) or by solubility-controlling substituents.
The concept of pharmaceutically active ingredients covalently linked to polysaccharides has been used for naproxen, mitomycin, and ibuprofen. Direct linking of pharmaceutically active ingredients to the polymer backbone or physical incorporation in nanoparticles may prevent side effects, increase half-life, and circumvent drug insolubility. In many cases, physically incorporated drug molecules can be easily released by diffusion, swelling, and erosion. However, such processes are uncontrolled, and the drug may be released before the carrier has reached the target site. Thus, drugs covalently linked to a polymer are more suitable for drug delivery as the cleavage is adjustable, and a higher loading capacity can be achieved compared with physically immobilized drugs. Contrarily, covalently bound compounds may be cleaved incompletely and at a disadvantageously slow rate.
Polysaccharides possess hydroxyl, carboxylic acid, and/or amino groups that can be used for derivatization reactions. Pharmaceutically active compounds can be attached to the polysaccharide by ester, amide, carbamate, or imine linkages.
An important class of clinically relevant drugs is the group of histone deacetylase inhibitors (HDACi). The balance between the recruitment of histone acetyltransferases (HATs) and HDACs plays a key role in the regulation of gene expression. A deregulation of HDAC activity has been linked to many diseases, including inflammation and various types of cancer. Thus, many different HDACi have been developed as therapeutics. One example is the α-branched fatty acid valproic acid (VPA). Originally, VPA was used as an anti-epileptic drug. The discovery of its HDACi properties opened new possibilities for treatment. VPA was shown to rescue mice from lethal septic shock. However, VPA is known to induce side effects such as hepatic toxicity in a dose-dependent manner. Chemically, VPA is suitable for direct linkage to polysaccharides via its carboxylic acid moiety. Studies of dextran valproate with a low degree of substitution (DS) of around 0.5 showed no hepatic toxicity and ulcerogenicity as the macromolecular prodrug.
The esterification of polysaccharides with carboxylic acids in the presence of activation agents like N,N’-carbonyldiimidazole (CDI), p-toluenesulfonyl chloride (Tos-Cl), or iminium chloride (Im-Cl, formed by the reaction of N,N-dimethylformamide and oxalyl chloride) is a simple modification procedure. CDI offers mild reaction conditions with no polymer degradation or side reactions, making the products suitable for pharmaceutical applications. Tos-Cl is advantageous for upscaling due to easy management of in situ reactions. The reaction of carboxylic acid iminium chlorides with polysaccharides occurs under mild conditions, causing neither degradation nor side reactions.
This work reports on the esterification of cellulose and dextran with the HDACi VPA activated by Tos-Cl, CDI, or Im-Cl. The nanoparticle formation of cellulose- and dextran valproates by emulsification and nanoprecipitation was studied, and the VPA release by lipase treatment was examined in vitro. Cellular uptake was proven in cell culture. Both biocompatibility and nontoxic behavior of the nanoparticles were investigated as well.
Esterification of Cellulose and Dextran with Valproic Acid
Activation agents form reactive carboxylic acid derivatives that efficiently react with the hydroxyl group of the anhydroglucose unit (AGU) of cellulose or dextran. A high degree of substitution (close or equal to 3) is necessary for a high drug loading and renders the polymer hydrophobic, which is a prerequisite for nanoparticle formation.
Cellulose and dextran were dissolved in N,N-dimethylacetamide (DMA)/LiCl. The esterification was performed with VPA in the presence of CDI, Tos-Cl, and Im-Cl for activation. The conversions using CDI and Tos-Cl as activation reagents were carried out for 24 hours at 80 °C, while Im-Cl reactions were performed at 60 °C. The products were isolated by precipitation in ethanol/water mixture (CDI), in aqueous NaHCO3 (Tos-Cl), or in water (Im-Cl). Dextran valproates prepared with CDI or Tos-Cl were precipitated in methanol, and those obtained with Im-Cl in water.
Conversion of the polysaccharides with 3 mol VPA and 3 mol CDI per mol AGU yielded a cellulose valproate with a DS of 0.55 and a dextran valproate with a DS of 0.66. Activation of VPA with Tos-Cl proved more efficient, resulting in cellulose valproate with a DS of 1.62 and dextran valproate with a DS of 1.64 at a 1:3:3 molar ratio. Increasing the molar ratio to 1:6:6 led to cellulose valproate with a DS of 2.09 and dextran valproate with a DS of 2.20. Similar DS values were achieved using Im-Cl at a 1:3:3 ratio, with cellulose valproate at DS 1.69 and dextran valproate at DS 1.75. While the highest DS products were achieved with Im-Cl, Tos-Cl is easier to handle, especially for upscaling.
The DS values were determined by proton NMR spectroscopy after peracetylation of the samples. Samples with high drug loading and remaining hydroxyl groups, which might be used for further modification with dyes or targeting groups, were chosen for additional studies.
All samples are soluble in DMSO and DMA. Dextran valproate with DS 2.20 was not soluble in DMA due to strong hydrophobic character. Cellulose- and dextran valproates with DS above 1.6 are soluble in chloroform, dichloromethane, and ethyl acetate.
The remaining hydroxyl groups of selected samples were labeled with rhodamine B isothiocyanate in pyridine, forming the corresponding thiocarbamate.
NMR Spectroscopy of Cellulose- and Dextran Valproates
The polysaccharide valproates were characterized by proton, carbon, and two-dimensional NMR spectroscopy. In the carbon NMR spectrum of cellulose valproate, the carbonyl carbon signal appeared at 177–175 ppm, with the other substituent carbons observed at 45.3, 34.4–32.8, 20.6, and 13.9 ppm. AGU signals were detected at 101.6, 98.9, 75–69, 62.4, and 60.6 ppm.
For dextran valproate, the carbonyl signal appeared at 175.8–174.6 ppm, with other substituent signals similar to those in cellulose valproate. The characteristic AGU signals corresponded to their respective carbons. Substitution at different positions of the AGU affected chemical shifts.
Acetylation is a known side reaction during activation with Tos-Cl, leading to detectable acetate methyl signals in NMR. The carbonyl carbons for VPA esters were assigned to different AGU positions, revealing preferred esterification at positions 6 and 2. Reactivity follows the order O-6 > O-2 > O-3 without strict preference for cellulose, and O-2 > O-3 > O-4 for dextran derivatives.
Cellulose- and Dextran Valproate Nanoparticles
Cellulose valproate with high DS was shaped into nanoparticles by nanoprecipitation, yielding particles with a hydrodynamic diameter of 207 nm and a polydispersity index of 0.058. Dextran valproate nanoparticles tended to agglomerate, indicating that both polymer backbone and DS impact self-assembly.
The emulsification technique offered an alternative method, requiring a surfactant like polyvinyl alcohol (PVA). This method yielded stable, spherical nanoparticles with diameters between 140 and 150 nm and narrow size distributions. Negative zeta potentials resulted from hydroxy ion adsorption, while PVA’s effect increased the zeta potential through charge shielding.
Rhodamine B-labeled nanoparticles prepared by nanoprecipitation agglomerated quickly, while stable, labeled nanoparticles could be prepared by emulsification in PVA. The formation of nanoparticles by emulsification was independent of backbone polymer and DS value, although solvent, surfactant, and sonication power influenced size and distribution.
Scanning electron microscopy showed self-assembled nanoparticles with spherical morphology and smooth surfaces after nanoprecipitation. Emulsification also produced spherical nanoparticles but revealed the presence of a PVA layer.
Investigation of Valproic Acid Release by Lipase Treatment
Ester bonds may be cleaved under acidic or alkaline conditions, but for intended cellular or in vivo use, VPA release must occur at physiological conditions. Porcine pancreatic lipase was used for in vitro release studies. Nanoparticle suspensions were incubated with lipase at pH 7.4 and 37°C for 24 hours. Dynamic light scattering measurements revealed stable size before and after treatment, with PVA preventing agglomeration.
As VPA is a pan-inhibitor of class I and II HDACs, including HDAC2, released VPA should decrease HDAC2 deacetylase activity. An HDAC2 activity assay showed that only lipase-treated nanoparticles resulted in significant reduction of HDAC2 activity, confirming partial in vitro release of active VPA from nanoparticles.
Cellular Uptake of Polysaccharide Valproate Nanoparticles
Cellular uptake was investigated using rhodamine B-labeled nanoparticles and nanoparticles encapsulating Nile Red. Uptake of labeled cellulose valproate nanoparticles by HeLa cells occurred within seconds and localized mainly in the cytoplasmic compartment, confirmed by z-stack analysis. Dextran valproate and Nile Red-encapsulated nanoparticles showed similar behavior.
Effects of Nanoparticles on Cell Viability
The toxicity of the nanoparticles was evaluated on HEK-293T cells using the MTT assay. Neither free VPA nor cellulose or dextran valproate nanoparticles reduced cell viability compared to untreated controls after 24 hours of exposure, indicating that the nanoparticles are nontoxic to HEK-293T cells.
Hemocompatibility of Dextran and Cellulose Valproate Nanoparticles
Hemocompatibility was studied using isolated sheep erythrocytes. Measurement of hemoglobin release after incubation with nanoparticles classified both valproate nanoparticles and sodium valproate as nonhemolytic up to concentrations of 625 µg mL−1. Evaluation of erythrocyte aggregation showed that neither sodium valproate nor nanoparticles induced aggregation, confirming hemocompatibility.
Toxicity Testing Ex Ovo in the Shell-Less Hen’s Egg Test
Ex ovo testing in a shell-less hen’s egg model was conducted to assess biocompatibility. Dispersions of nanoparticles or sodium valproate solution were injected into the vitelline vein. Negative and solvent controls produced less than 10% hemorrhage associated with microinjection. Positive controls resulted in aggregation and hemorrhage with lethality in all eggs. Both nanoparticle samples showed no toxic effects, with rare hemorrhage comparable to controls, confirming suitability for systemic application.
Conclusions
Cellulose- and dextran valproates with various DS values ranging from 0.55 up to 2.20 were efficiently synthesized using CDI, Tos-Cl, and Im-Cl as activation agents. Tos-Cl and Im-Cl proved most efficient. The structure of the polysaccharide valproates was confirmed by NMR spectroscopy. Cellulose valproate could be shaped into nanoparticles with DH of 207 nm and PDI of 0.064 by nanoprecipitation, while stable nanoparticles could also be formed by emulsification using PVA. Nanoparticles prepared showed long-term stability, rapid cellular uptake, and in vitro release of active VPA upon lipase treatment.
The polysaccharide valproates are biocompatible and nontoxic, constituting potential nanocarriers for drug delivery. These nanocarriers bearing VPA as an HDAC inhibitor offer therapeutic opportunities for inflammatory diseases and selected cancers. Further modification is possible via covalent attachment of targeting groups or dyes and by immobilization of other compounds, including more specific HDAC inhibitors. These nanoparticles provide a versatile and flexible nanocarrier system with wide modification options for diverse treatment strategies.
Experimental Section
Methods and Measurements
NMR spectra were recorded using Bruker spectrometers for both proton and carbon with adequate scans. Scanning electron microscopy (SEM) was conducted on platinum-sputtered samples using a Zeiss microscope. Nanoparticle size and zeta potential were measured by dynamic light scattering. Z-averaged diameters and zeta potential values are averages over several runs and samples.
Degree of substitution (DS) was determined from peracetylated samples using proton NMR. Elemental analysis was performed as required. FT-IR spectra were recorded with dried potassium bromide pellets over the standard infrared range.
Synthesis Protocols
Cellulose valproate was prepared by Tos-Cl activation using typical dry, inert conditions, and stirred at elevated temperatures. After the reaction, products were isolated by precipitation, filtration, and solvent extraction, then dried. The product was characterized by NMR, FT-IR, and standard analytical techniques.
Dextran valproate was synthesized similarly, using CDI activation and standard precipitation and purification procedures before drying and analysis.
Cellulose valproate synthesis with Im-Cl followed addition of oxalyl chloride to DMF at subzero temperatures, further reaction with VPA, followed by stirring with dissolved cellulose and subsequent precipitation, purification, and drying.
Peracetylation and Rhodamine B labeling reactions were performed under controlled conditions, and products purified by precipitation and filtration.
Nanoparticle Preparation
Nanoparticles were prepared either by emulsification, combining a dichloromethane solution of the polymer with an aqueous PVA solution followed by sonication and solvent removal, or by nanoprecipitation, slowly adding water to a DMA solution of the polymer and then dialyzing against water. Encapsulation of Nile Red for imaging used dialysis to remove excess dye.
For drug release studies, nanoparticles were incubated with lipase in buffered saline at physiological conditions.
Assays
HDAC2 activity was assessed with a fluorometric kit, and statistically analyzed. Cell cultures were maintained under standard conditions, and uptake studies used confocal imaging. Cell viability post-exposure was determined by MTT assay, with quantification after solubilization of product and absorbance measurements.
In vitro hemocompatibility was tested using sheep erythrocytes with incubation, aggregation detection by microscopy, and hemolysis measured by spectrophotometry, following established standards and statistical analysis. Ethical protocols for animal procedures were followed.
Ex ovo testing used hen’s egg area vasculosa as a model, with nanoparticle or control solutions injected and toxic responses assessed macroscopically and microscopically at multiple time points after injection. Each sample was tested with adequate replication.