We present the preparation of degradable and biocompatible nanohydrogels from thiol-functional macromers. Linear poly(glycidol) and star-shaped poly(ethylene oxide-stat-propylene oxide) with molecular weights below 15 kDa were functionalized with thiol groups by polymer-analogous reaction and crosslinked in inverse miniemulsion. Particle size was determined by dynamic light scattering, cryo-field emission scanning electron microscopy and scanning force microscopy. The disulfide crosslinked particles readily degrade on addition of 10 mM aqueous glutathione solution what resembles cytosolic conditions. Biocompatibility has been proven by incubation of L929 fibroblasts with nanogels for 24 and 72 h followed by life-dead staining. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]
Hydrogel nanoparticles, so called nanogels, are hydrophilic crosslinked polymeric particles with submicrometer size. They combine characteristics of hydrogels like biocompatibility, high-water content as well as tunable chemical and mechanical properties with the features of nanoparticles such as high-surface area and overall sizes in the range of cellular compartments. These properties make them intriguing candidates for entrapment of hydrophilic bioactive molecules such as DNA or proteins to provide a hydrophilic environment and protect them from degradation.1, 2 So far, little attention has been paid to the size dependence on the cell uptake and pharmacokinetics. A prerequisite for this is, however, synthetic access to well-defined particles.
Several techniques have been used for the preparation of microgels such as precipitation polymerization, photolithography, micromolding, microfluidics, and inverse emulsion techniques.1 Polymerization in inverse miniemulsion is one way that allows preparation of gel particles with controlled size in the submicron range. However, syntheses of nanogels in inverse emulsions are mostly done by radical polymerization techniques3–6 and only little work has been reported on crosslinking of well-defined prepolymers.7, 8 Such a prepolymer condensation concept based on hydrophilic multithiolfunctionalized oligomers and polymers will be particularly suitable for incorporating peptides and proteins in particles that are stable in extracellular physiological conditions but degrade after uptake by cells within the reductive environment of the cytosol where for example glutathione, a tripeptide with a free thiol group, readily cleaves disulfides to thiols.9 In contrast to radical polymerization techniques, oxidative coupling of thio-functional macromers in inverse miniemulsion allows the attachment of cysteine-terminated peptide sequences as the free thiol groups do not interfere with the polymerization. As a best example so far, high-molecular weight, thiol-functionalized hyaluronic acid has been crosslinked in inverse miniemulsion by oxidation.10 However, variation of functionality or molecular architecture of the gel precursor could not be achieved.
Here, we present synthesis and characterization of biocompatible and degradable nanogel carriers composed of thiol-functionalized polymers based on star-shaped poly(ethylene oxide-stat-propylene oxide) (sP(EO-stat-PO)) and linear poly(glycidol) (PG). Nanogel particles are prepared by crosslinking the polymers in inverse miniemulsion via oxidation of thiol groups to disulfide bonds. Synthesis and characterization of the gel precursors, preparation of the nanogels, and their characterization regarding particle size distribution in solution and in the dried state, chemical composition, degradation behavior as well as in vitro biocompatibility studies are presented and discussed. Reduction-sensitive degradation of the nanogels in cytosolic conditions, cytocompatibility, hydrophilicity, as well as a crosslinking chemistry that allows for the introduction of cysteine functional targeting peptides and does not affect the functionality of most biomolecules predetermines these nanogels as carriers for the targeted delivery of proteins, siRNA, and DNA.
Hydroxy terminated, six arm, star-shaped poly(ethylene oxide-stat-propylene oxide) (sP(EO-stat-PO) with a backbone consisting of 80% ethylene oxide and 20% propylene oxide (Mn = 12,000 g/mol, Mw/Mn = 1.12) was obtained from Dow Chemicals (Terneuzen, NL). Before functionalization sP(EO-stat-PO) was purified by precipitation in tetrahydrofuran (THF)/cold diethylether as solvent/nonsolvent system. Linear poly(glycidol) (PG) (Mn = 4500 g/mol, Mw/Mn = 1.17) was synthesized via anionic polymerization in diglyme using potassium tert-butoxide (Aldrich) initiator according to procedure described before.11N,N′-dicyclohexyl-carbodiimide (DCC), (99%, Acros), 4-(dimethylamino)pyridine (DMAP) (>99%, Aldrich), 3,3′-dithiodipropionic acid (99%, DTPA) (Aldrich), tris(2-carboxyethyl) phosphine (TCEP) (99%, Aldrich), dichloromethane (HPLC grade, Aldrich), cysteamine (2-mercaptoethylamine, Biochemica), 5,5′-dithio-bis(2-nitrobenzoic acid) (DTNB) (Merck), anhydrous N,N-dimethylformamide (DMF) (99.8%, Sigma-Aldrich), Span 80 (Sigma), and Tween 80 (Sigma-Aldrich) were used as received. THF was dried over LiAlH4. Dialysis membrane (MWCO = 3500 Da) was purchased from Spectrum Laboratories.
1H NMR spectra were carried out on a Bruker DPX 300 at 300 MHz. DMSO-d6 with TMS as internal standard was used as solvents.
Size exclusion chromatography (SEC) was performed in THF with addition of 250 mg/L 2,6-di-tert-butyl-4-methylphenol. A high-pressure liquid chromatography pump (PL-LC 1120 HPLC) and a refractive index detector (ERC 7515A) were used at 35 °C with a flow rate of 1.0 mL/min. Five columns with MZ gel were applied. The length of the first column was 50 mm, and 300 mm for the other four columns. The diameter of each column was 8 mm, the diameter of the gel particles 5 mm, and the nominal pore widths were 50, 50, 100, 1000, and 10,000 Å, respectively. Calibration was achieved using poly(methyl methacrylate) standards. For PG and SH-PG, SEC analyses were carried out at 80 °C using a high-pressure liquid chromatography pump (Bischoff HPLC 2200) and a refractive index detector (Waters 410). The eluting solvent was N,N-dimethylacetamide (DMAc) with 2.44 g/L LiCl and a flow rate of 0.8 mL/min. Four columns with MZ-DVB gel were applied. The length of each column was 300 mm, the diameter 8 mm, the diameter of the gel particles 5 mm, and the nominal pore widths were 100, 1000, and 10,000 Å. Calibration was achieved using poly(methyl methacrylate) standards.
Raman spectroscopy was recorded on Bruker RFS 100/S spectrometer. The laser used was Nd:YAG at 1064 nm wavelength at a power of 250 mW. On an average, 1000 scans were taken at a resolution of 4 cm−1. For sample holding aluminum pans of 2 mm bore were used. Software used for data processing was OPUS 4.0.
Scanning force microscopy (SFM) studies were conducted using a Nanoscope IIIa from Digital Instruments operating in the tapping mode. Standard silicon cantilevers were used (PPP-NCH from Nanosensors) with a spring constant k ∼ 42 N/m and an oscillation frequency fo ∼ 70 kHz. For the SFM measurements samples were cast onto freshly cleaved mica by spin coating at the rotation speed of 2000 rpm. The samples were dried in vacuum for 2 h before measurements.
Scanning electron microscopy (SEM) was performed with a HITACHI S-4800 instrument in a cryo-mode with secondary electron image resolution of 1.0 nm at 15 kV, 2.0 nm at 1 kV, and 1.4 nm at 1 kV with beam. The material is fixed on a holder and was rapidly frozen with boiling liquid nitrogen. It was then transferred to the high-vacuum cryo-unit, the Balzers BF type freeze etching chamber. The cryo-chamber equipped with a knife can be handled from outside by means of a level to fracture the sample for applications in which imaging of the surface of inner structures is aimed. To remove humidity, the sample is sublimated from 5 to 15 min then the entire material is further inserted into the observation chamber.
Dynamic Light Scattering
The particle sizes were measured by photo correlation spectroscopy using a Malvern Zetasizer Nano ZS at a fixed scattering angle of 173°. Noninvasive back scatter technology takes particle sizing to sensitivity in the 0.6 nm to 6 μm range. Disposable poly(styrene) cuvettes were used for measurement in water. “Expert System” software was used for data interpretation. The presented data is the average value from five measurements. The dynamic light scattering (DLS) measurements give a z-average size (or cumulant mean) value, which is an intensity mean and the polydispersity index (PDI). The cumulant analysis has the following form:
where g(1) is the first order correlation function, is the average decay rate and first cumulant, and μ2 is the second cumulant. The value of μ2/2 is known as PDI.
Synthesis of Thiol Functionalized sP(EO-stat-PO)
Thiol functionalized sP(EO-stat-PO) (SH-sP(EO-stat-PO)) has been prepared in a two step synthesis. In the first step, sP(EO-stat-PO) has been crosslinked with disulfide crosslinker followed by reduction of the disulfide bonds to thiol groups in the second step. Typically, four alcohol groups have been transferred into thiol groups, leaving two unreacted OH groups in the polymers.
For crosslinking, a solution of DTPA (0.5 mmol, 0.105 g) in dried THF (4 mL) was added dropwise to a solution of purified sP(EO-stat-PO) (0.25 mmol, 3 g), DCC (1.1 mmol, 0.227 g), and DMAP (1.1 mmol, 0.134 g) in CH2Cl2 (6 mL) in an ice bath at 0 °C over 10 min under stirring. The resulting mixture was stirred at RT for 5 h to allow for complete reaction and crosslinking. The formed hydrogel was washed three times with CH2Cl2, twice with THF, ethanol, and water followed by evaporation of remaining solvents on the rotary evaporator. Finally, the product was dried in a vacuum oven at 35 °C for 12 h.
For reduction, the hydrogels were immersed in an aqueous solution of TCEP (1.5 eq. with respect to the disulfide units) at room temperature for 4 h. The pH of the solution was adjusted with triethylamine to pH = 6.5. After reduction, the solution was first dialyzed for 2 days at RT against aqueous HCl solution at pH ∼ 3.5, then against pure water under inert gas for 1 day. Finally, the polymer solutions were lyophilized and stored at 4 °C under argon for further use. The free thiol content of the polymer was determined using Ellman's method. SEC (THF): Mn = 12,900 g/mol, Mw/Mn = 1.12.
Synthesis of Thiol Functionalized PG
Thiol-functionalization of PG was done according to the procedure described for sP(EO-stat-PO) with the difference that, because the PG is not soluble in CH2Cl2, DMF was used as a solvent instead. From the 1H NMR, it was determined that ∼16% of the PG units have SH functionality.
Preparation of the Nanogels in Inverse Miniemulsion
For preparation of the miniemulsion, 5 mL of hexane containing 150 mg of a 3:1 weight ratio of Span 80 and Tween 80 were used as organic phase, whereas the aqueous phase consisted of 220 mg polymer dissolved in 0.5 mL 0.04 M PBS buffer (pH = 7.4). The two phases were combined and pre-emulsified by magnetic stirring for 10 min followed by miniemulsification by ultrasonication for 60 s with a Branson sonifier (W450 Digital at duty cycle of 30% and output control of 90%). During the sonication, the reaction vessel was cooled with an ice bath. Subsequently, 20 μL of 0.1 M H2O2 solution was added and the mixture was sonicated again for 60 s. The reaction was allowed to proceed for 15 min at room temperature followed by quenching of the oxidation by addition of 4 mL of acidic water (pH = 3). Separation of the nanogels was achieved by centrifugation at 10,000 rpm for 30 min followed by decantation of the supernatant. Nanogels present in the aqueous layer were carefully washed twice with hexane (10 mL) and four times with THF (20 mL) to remove the surfactants and unreacted polymer. The remaining organic solvents and acid were removed by dialysis. Purified nanogels were stored in millipore water at 4 °C for further use.
Degradation of the Nanogels
Reductive degradation of the nanogels to the corresponding thiol-containing P(EO-stat-PO) and PG fragments was conducted in the presence of 10 mM glutathione in PBS buffer (pH = 7.4) at 37 °C (mass ratio, glutathione/nanogel = 0.4). After 6 h, the solution was dialyzed, freeze-dried followed by analysis with Raman spectroscopy. A part of the solid was dissolved in water, spin casted onto mica and the resulting polymer film was analyzed by SFM.
Cytotoxicity of the Nanogels
Standard indirect cytotoxicity tests were carried out based on the recommendation of the International Standardization Organization (ISO) described in DIN EN ISO 10,993-5. For cytotoxicity testing L929 mouse connective tissue fibroblasts were purchased from DSMZ GmbH (Germany). Before incubating with three different concentrations of nanogel particles, L929 cells were seeded on tissue culture polystyrene and cultivated till subconfluence. Cells were maintained in RPMI supplemented with L-glutamine, 10% fetal calf serum, and 1% penicillin-streptomycin (PAA, Germany) at 37 °C and humidified atmosphere containing 5% CO2. To detect cytotoxic effects, the LIVE/DEAD® Viability/Cytotoxicity Assay Kit from Molecular Probes (MoBiTec GmbH, Germany) was accomplished 24 and 72 h after adding nanogel particles to the cultivated L929 fibroblasts. This test provides a two-color fluorescence cell viability assay that is based on the simultaneous determination of live (green colored) and dead cells (red colored) with two probes that measure recognized parameters of cell viability—intracellular esterase activity and plasma membrane integrity.
RESULTS AND DISCUSSION
We have chosen linear poly(glycidol) (PG) and star-shaped polymers sP(EO-stat-PO) as precursors for the nanogels. Both polymers are hydrophilic and biocompatible, and the molecular weights are well below 30,000 g/mol, which is set as limit that allows renal clearance for linear PEO.12 They have been successfully functionalized with thiol groups through carbodiimide mediated Steglich esterification between the free hydroxyl groups of the polymers and the disulfide containing dicarboxyacid. This reaction yielded hydrogels that were subsequently reduced by cleaving the disulfide crosslinkage to thiol groups (Scheme 1). After reduction, no trimer or dimer content could be detected with SEC, suggesting close to quantitative reduction of the disulfide bonds. Quantification of the thiol groups in the SH-PG polymers by 1H NMR showed that 16% of the SH-PG units bear thiol functionality, which is slightly lower than the maximal theoretical value of 20%. Because of the higher molecular weight and the lower thiol content, quantification of the free thiol groups in SH-sP(EO-stat-PO) was achieved using Ellman's reagent. In several batches, where a functionalization degree of four thiol groups per star molecule was intended, the amount of free thiols was determined as 3.2 ± 0.2 per sP(EO-stat-PO).
It is known that inverse emulsions are generally less stable than regular oil-in-water emulsions due to lack of electrostatic stabilization.7 On the basis of numerous experiments with different surfactants and cosurfactants, we selected a mixture of the FDA approved Span 80 and Tween 80 surfactants for our experiments because they showed most efficient stabilization of the miniemulsion droplets with our experimental setup and procedure. Besides the surfactant mixture, we checked and optimized the oxidative bulk crosslinking kinetics of SH-sP(EO-stat-PO) and SH-PG.
First, preparation of disulfide crosslinked particles from SH-sP(EO-stat-PO) polymers by oxidation in inverse miniemulsion is discussed. Generally, disulfide formation proceeds through the deprotonated form of thiol groups. Thus, to maintain slightly basic reaction conditions, we used PBS buffer (pH = 7.4) as a dispersed phase. Furthermore, the ionic strength of PBS served as osmotic agent in the droplets to stabilize them against Ostwald ripening. Bulk experiments with SH-sP(EO-stat-PO) showed that simple oxygen mediated gelation took >6 h in air at physiological pH = 7.4, thus too long with respect to inverse miniemulsion stability. On addition of H2O2 as an oxidation promoter, the bulk crosslinking reaction could be shortened to 5 min. As a consequence, we found that the presence of H2O2 is indispensable for the disulfide crosslinked nanogel formation in the inverse miniemulsion.
Cryo-scanning electron microscopy (cryo-SEM) and DLS techniques were applied to characterize the nanogel particle size in water, that is, in the swollen state and SFM has been used as a characterization technique for dry particles. Figure 1 shows the cryo-SEM and DLS analysis of SH-sP(EO-stat-PO) nanogels. Particle size analysis with DLS gives the z-average particle diameter of 380 nm with a PDI of 0.24. Cryo-SEM images clearly show that the particles have well-defined spherical shapes, as expected for particles synthesized by miniemulsion techniques, with particle diameters d in the range of 230 nm < d < 350 nm. Taking into account the freezing-step that may lead to partial volume loss of the nanogels during cryo-SEM, these two techniques correlate well to each other with respect to size analysis of the nanogels in the hydrated state. Size determination of such disulfide crosslinked SH-sP(EO-stat-PO) nanogels in the dry state has been achieved by SFM. Samples were prepared by spin-casting from aqueous solution onto mica followed by drying in vacuum. By the spin coating process, a dense monolayer of single particles is initially formed. On drying, the nanogels decrease their volume and a regular pattern of separated single nanogel particles results (Fig. 2).
It is known that PEO-based polymers show a tendency to form aggregates in water that may be easily confused with real crosslinked particles in the DLS analysis.13 Also, the regular pattern formation on mica is not an unequivocal proof for the formation of nanogels because in some cases, dewetting of polymer films may result in similar patterns.14 We have thus optimized the spin-casting parameters and analyzed areas of the sample where the transition between nanogel aggregates and the monolayer occurs. The clear detection of individual nanogels in the larger aggregates that were formed during the spin-coating process undoubtedly proofs the successful crosslinking of the prepolymers to nanogel particles (Fig. 3).
From the SFM topography image showing isolated nanogels (Fig. 2), an average diameter dav = 110 ± 10 nm and an average height hav = 28 ± 2 nm of the nanogels could be determined as adsorbed on mica. The large anisotropy demonstrates that on mica the soft and deformable nanogels adopt a flattened, oblate spheroid shape. This flattening can be contributed to a strong affinity of the hydrophilic particles toward the polar mica surface in air as well as to indentation by the SFM tip. Regardless of the reason for flattening, the volume (V) of the adsorbed particles can be calculated form the formula for spheroid: V = 4/3π abhav (where a = b = average diameter (dav) = 110 ± 10 nm; average height hav = 28 ± 2 nm). With the equation for a perfect sphere V = 4/3πR3, the radius R and thus, also the dav of perfectly spherical nanogels in the dry state could be calculated to be dav = 70 ± 5 nm. Hence, by comparing this value with the size in aqueous environment determined from cryo-SEM and DLS, a strong swelling ability of the nanogels is obvious and the swelling ratio can be estimated in between 3.5 and 5.
To prove degradability, the disulfide crosslinked nanogels were incubated with a glutathione (GSH) solution. Figure 4(a) illustrates the 2000–3500 cm−1 Raman spectra range of SH-sP(EO-stat-PO) nanogels before and after incubation with 10 mM GSH for 6 h. The noteworthy feature is the band at about 2580 cm−1, which corresponds to stretching modes of thiol groups in the reduced particles. This band is not detectable in the crosslinked nanogels and appears on incubation with GSH. In addition, spin-casting the nanogel-solution after reduction onto mica resulted in a dewetted polymer film where no nanogels could be detected [Fig. 4(b)]. These results evidence that the nanogel disulfide crosslinkages can be cleaved with GSH what confirms degradability of the particles in cytosolic conditions.
Generally, particles prepared by oxidation of SH-PG showed very similar behavior regarding dispersion stability and degradation behavior. However, in water, these nanogels possess broader, bimodal size distributions with the z-average particle diameter of 330 nm and PDI = 0.54 as measured by the DLS and particle diameters d in the range of 100 nm < d < 350 nm determined by cryo-SEM (Fig. 5). SFM measurements of nanogels on mica performed in the dry state confirm the DLS measurements and show polydisperse disulfide crosslinked SH-PG particles (Fig. 6).
A live/dead cell viability assay was used to investigate the cytotoxicity of all prepared nanogels. After 24 and 72 h incubation of L929 cells with nanogels (c = 0.5 mg/mL), differential interference contrast microscopy and fluorescence microscopy were used to assess cell morphology and visualize live and dead cells. Live cells with intact membranes are stained with green fluorescence, whereas dead cells were stained with red fluorescence. Figure 7 shows fluorescence images of cells incubated with SH-sP(EO-stat-PO) nanogels after live/dead staining. Careful counting of live and dead cells indicated 94 ± 2% viability of L929 cells in the presence of SH-sP(EO-stat-PO) nanogels, when compared with 96 ± 2% cell viability in the absence of nanogels as a control experiment. Comparable results have been achieved with SH-PG nanogels. This clearly shows the cytocompatibility of nanogels prepared from SH-sP(EO-stat-PO) and SH-PG.
We have demonstrated the formation of nanogels by crosslinking of thiol-functional hydrophilic prepolymers in inverse miniemulsion. Formation of the nanogels was unequivocally proven by a combination of DLS, cryo-SEM, and SFM characterization. Oxidative coupling resulted in particles that are stable in PBS buffer at pH 7.4 but decompose in reductive conditions. Cell viability tests according to ISO 10,993-5 showed that the nanogels are noncytotoxic toward L929 fibroblasts. Biocompatibility, intracellular degradability, fully synthetic polymeric building blocks with molecular weights that allow renal clearance and the possibility to tailor the synthetic precursors regarding structure and functionality make this system a versatile platform for drug delivery.
The authors thank Jochen Salber for performing the cytotoxicity tests. This work was supported by the EU (FP 6, IP NanoBioPharmaceutics; NMP 026723-2).