The copyright line for this article was changed on 10 Mar 2016 after original online publication.
A new class of non-ionic dendronized multiamphiphilic polymers is prepared from a biodegradable (AB)n-type diblock polymer synthesized from 2-azido-1,3-propanediol (azido glycerol) and polyethylene glycol (PEG)-600 diethylester using Novozym-435 (Candida antarctica lipase) as a biocatalyst, following a well-established biocatalytic route. These polymers are functionalized with dendritic polyglycerols (G1 and G2) and octadecyl chains in different functionalization levels via click chemistry to generate dendronized multiamphiphilic polymers. Surface tension measurements and dynamic light scattering studies reveal that all of the multiamphiphilic polymers spontaneously self-assemble in aqueous solution. Cryogenic transmission electron microscopy further proves the formation of multiamphiphiles towards monodisperse spherical micelles of about 7–9 nm in diameter. The evidence from UV–vis and fluorescence spectroscopy suggests the effective solubilization of hydrophobic guests like pyrene and 1-anilinonaphthalene-8-sulfonic acid within the hydrophobic core of the micelles. These results demonstrate the potential of these dendronized multiamphiphilic polymers for the development of prospective drug delivery systems for the solubilization of poorly water soluble drugs.
Several supramolecular architectures on the nanoscale level have become important candidates for potential applications in nanomedicine, particularly for drug, dye, and gene delivery.1–5 Polymeric drugs, polymer–drug conjugates, polymer–protein conjugates, polymeric micelles, and multicomponent polyplexes use specific water-soluble polymers, either as a bioactive component itself or as an inert functional part of a multifaceted construct for improved drug, protein, or gene delivery.1–5 Current approaches towards solubilizing active components involve non-ionic polymeric materials such as Cremophor, Tween, and Pluronics, which have been extensively used to formulate hydrophobic guest molecules.2b, 6 Biocompatiblity and biodegradablity of these nanocarriers are especially important for the clinical application of therapeutic compounds.
The research on the encapsulation of hydrophobic molecules has indicated the need to improve drug-delivery processes with the right transporter architectures with more precise biodistribution and a better pharmacokinetic profile for targeting specific diseases (especially tumors, via the EPR (enhanced permeation and retention) effect).7 Thus, there is a quest to develop larger sized, well-defined, and biocompatible molecular objects as prospective drug-delivery systems. Out of various conceptual extensions for the development of such molecular objects, one approach uses polymers as multifunctional, polydisperse cores to which dendrons are connected via pendant functional groups at every repeat unit and are termed ‘dendronized polymers’ or ‘denpols.’8, 9
Dendronized polymers were originally termed ‘rod-shaped dendrimers,’ and made their first appearance in 1987 in a patent filed by Tomalia et al.10 From their work on dendrimer synthesis, this group set out to synthesize larger highly functionalized structures. Hawker and Fréchet created hybrid architectures of dendrimers and linear polymers by attaching Fréchet-type dendrons to a styrene derivative and copolymerizing it with styrene.11 The predominant role of geometry in the self-assembly of polymers equipped with ‘taper-shaped’ side chains was recognized by Percec et al.12 Extensive work on synthesis and characterization of dendronized polymers was carried out by the Schlüter group who held the world record in shape persistant dendronized polymers.9, 13 They synthesized rod-like polymers with a conjugated backbone and recognized the significance of dendron decoration for the backbone conformation.13c
Dendronized polymers are now being used both as new materials in already established applications, e.g., as organic materials for optoelectronic applications, as well as in intriguing new concepts which may only be accessible with this remarkable class of molecule.8, 14 They have helped to bring different fields of research together.9, 15 Amphiphilic dendronized polymers with nanoscopic dimensions are among the most interesting target structures,8 but have not yet been much explored for drug-delivery applications.
Our interest was to extend the multiamphiphilic PEGs6 with structurally defined glycerol dendrons to synthesize biocompatible dendronized multiamphiphilic polymers. We have recently developed novel biocatalytic approaches for the synthesis of PEG/glycerol/polyglycerol-based polymers and explored their potential as delivery vehicles.16 We have also developed a variety of amphiphilic architectures based on dendritic polyglycerol (dPG) that self-assemble in aqueous solution under physiological conditions into stable aggregates with a nanoscopic size range for the delivery of active agents.6, 17 Recently, it became evident that the introduction of polyglycerol dendrons in monomolecular amphiphiles has a remarkable stabilizing effect, resulting in structurally very defined supramolecular architectures.17e
We assumed that polyglycerol dendrons would also improve the overall structural stability of polymer based supramolecular architectures if covalently linked to a linear polymer backbone. This would combine the biomedical benefits of PEG-type polymers, which have long clearance times, high stability and solubility, no immunogenicity, antigenicity, or toxicity,18, 19 better monodispersity than classical monomolecular amphiphiles,17e and enhanced stability and biocompatibility due to polyglycerol dendrons.2a In contrast to polysoaps where monomolecular amphiphiles are lined up within polymers,20 the layout of such a polymer should allow an independent aggregation of headgroups and tails due to the flexibility of the polyethylene oxide (PEO) chains and thus lead to stable supramolecular nanocarriers.
Therefore, a new class of dendronized multiamphiphilic polymers was synthesized, starting with an (AB)n-type polymer from two biocompatible synthons and then grafting to it hydrophilic polyglycerol dendrons as headgroups as well as hydrophobic alkyl chains as tails. A modular biocatalytic synthetic approach was used to polymerize azido glycerol and PEG-600 diethyl ester with Novozym-435. Click coupling with different generation dendrons and hydrophobic alkyl chains gave well-defined multiamphiphiles that self-assembled at low critical aggregation concentrations. All the new dendronized multiamphiphilic polymers were tested as potential nanocarriers for drug-delivery applications using the hydrophobic fluorescent dye pyrene. Structural parameters influencing the solubilization properties of hydrophobic guests were investigated by varying the hydrophobic/hydrophilic balance and the interaction with photochromic guests was studied using the fluorescent probe pyrene and 1-anilinonaphthalene-8-sulfonic acid (ANS).
2. Results and Discussion
We have designed a biocatalytic approach for the synthesis of biodegradable (AB)n-type diblock polymer (3) from 2-azido-1,3-propanediol (azido glycerol, 1) (Scheme1) and PEG-600 diethylester (2) using Novozym-435 (Candida antarctica lipase) as biocatalyst (Scheme2). A modular approach was developed for synthesis of structurally varied (dendronized) polymers (10a–f) by making use of highly efficient click coupling reactions to attach different generation dendrons and alkyl chains (Scheme 2). This study is focused on the supramolecular self-assembly of these dendronized multiamphiphilic polymers, in particular, the effect of dendron loading onto linear PEG-based polymers and the ability to solubilize hydrophobic guest molecules.
2.1. Synthesis and Characterization
2-Azido-1,3-propanediol (azido glycerol) (4) was synthesized from commercially available glycerol in four steps (Scheme 1). Glycerol was first converted to 2-hydroxypropane-1,3-diyl diacetate (1) using vinyl acetate as the acylating reagent, by following the established procedure.21 The secondary hydroxyl group of 2-hydroxypropane-1,3-diyl diacetate (1) was then mesylated (3) and subsequently an azidation reaction was performed to obtain 2-azidopropane-1,3-diyl diacetate (3). The compound 3 was deacetylated using K2CO3 in anhydrous ethanol to obtain the monomer 2-azido-1,3-propanediol (4).
The structure of compound 4 was confirmed on the basis of its spectral data (IR, 1H and 13C NMR, distortionless enhancement by polarization transfer (DEPT)-135, and high-resolution mass spectra (HRMS). In its 1H NMR spectrum, the methine and the methylene protons recorded upfield shifts from δ 3.89–3.83 and δ 4.22–4.09 in the precursor 2-azidopropane-1,3-diyl diacetate to δ 3.50–3.47 and δ 3.70–3.55, respectively. The characteristic signals for the acetyl moiety at δ 2.08 in 1H NMR spectrum and at δ 20.56 and 170.33 in the 13C NMR spectrum disappeared completely. The peaks for methine carbon showed characteristic downfield shifts from δ 58.38 to δ 66.58 on deacetylation; methine and the methylene carbons were further corroborated with the DEPT-135 NMR spectrum.
Lipases have been employed for their versatile catalytic efficiency for esterification/transesterification reactions and biocatalytic polymerization reactions.22, 23 Drawing upon our previous work on the catalyzed condensation of PEG-600 dimethylester and glycerol under solventless conditions,16a Novozym-435 was further employed for the polymerization of 2-azido-1,3-propanediol (4) and PEG-600 diethylester (5), which was synthesized from its corresponding diacid by a standard esterification protocol16a to afford the polymer poly[(2-azidopropan)-1,3-dioxy-poly(oxyethylene-600)oyl)] (6) at 83% yield (Scheme 2). Use of the diester 5 instead of the diacid not only enhances the degree of polymerization but also introduces an ethyl group that could be used for end-group analysis to determine the average molecular weight of the polymer by 1H-NMR. Polymer 6 was unambiguously characterized from its IR, 1H, 13C, DEPT-135, and 2D-NMR spectral data. In its 1H NMR spectrum, the C-1 and C-3 protons recorded downfield shifts to δ 4.37–4.11 with respect to the monomer 2-azido-1,3-propanediol (4) (δ 3.70–3.55). This downfield shift is attributed to the transesterification reaction during polymerization, and occurs due to the attached primary hydroxyl groups. Protons at C-2 exhibit a downfield shift to δ 3.81–3.43 (compared to δ 3.50–3.47 in 4). Characteristic signals for the PEG chain protons and ethyl end group of PEG-600 diethylester (5) were observed at δ 3.81–3.43 and δ 1.28–1.24, respectively. Assignments of signals were confirmed by 1H-13C heteronuclear multiple quantum coherence (HMQC) correlation spectrum, and DEPT-135 NMR spectrum analysis (see the Supporting Information (SI)). From the DEPT-135 NMR spectrum the end-group carbons of azido glycerol (1) resonating at δ 63.67, 61.72 (C-1 and C-3), and δ 61.48 (C-2) can be distinguished from the main chain carbons, which resonate at δ 63.21 (C-1 and C-3) and 58.45 (C-2). The presence of azide groups can further be confirmed from a signal at 2131.4 cm−1 in the IR spectrum. The integration values for the ethyl ester protons were correlated to the peaks for other protons in 1H-NMR spectrum and the degree of polymerization was observed to be approximately 9. The corresponding molecular weight was calculated to be 6294 g mol−1, while the number-average molecular weight of 6708 g mol−1 was obtained by gel permeation chromatography (GPC) in tetrahydrofuran (THF) (1 mL min−1). The polydispersity index (PDI) was found to be 1.40. The elemental data analysis was also observed to be in accordance (for monomeric unit (C29H53N3O16)n calculated: C, 49.78; H, 7.63; N, 6.01, observed: C, 47.73; H, 7.32; N, 5.33).
Octadecyl-propargyl ether (7)24 and alkyne functionalized [G1.0] and [G2.0] generation polyglycerol dendrons (8 and 9) were synthesized, following well established procedures.17b, 25 The azide functionalized polymer 6 was conjugated with either alkyne functionalized [G1.0] dendrons (8) or octadecyl-propargyl ether (7) in dimethylformamide (DMF) using copper (I) bromide as a catalyst in a one-pot reaction (Scheme 2) to generate the grafted polymers 10a and 10b. The series of dendronized multiamphiphilic polymers 10c–f was synthesized by varying the ratio of the side chains, using mixtures of desired proportions of both the hydrophobic (7) as well as the hydrophilic (8 or 9) alkynes. Progress of the reaction was monitored by the decreasing intensity of the azide signal in the IR spectrum. After stirring for 24 h at room temperature, the product was dialyzed to remove the unreacted starting material.
The functionalized polymers were unambiguously identified on the basis of their spectral data (c.f. SI) as for dendronized multiamphiphilic polymer 10f. In the 1H NMR spectrum of dendronized multiamphiphilic polymer 10f, the C-1, C-3, and C-2 protons recorded a downfield shift and resonated at δ 4.71–4.40 and δ 5.14–5.09 with respect to the precursor polymer, (δ 4.37–4.11: C-1 and C-3 protons and δ 3.94–3.88: C-2 protons). This significant downfield shift in the 1H NMR spectra revealed that the azide functional group of polymer 6 was involved in click coupling with the alkyne functionalized side chains. Assignments of the signals at δ 4.29–4.14 for the α and α' protons of the PEG main chain, and methoxy protons of the side chains in the 1H NMR spectrum were confirmed by an 1H-1H COSY spectrum (SI).
2.2. Supramolecular Self-Assembly in Aqueous Media
The supramolecular self-assembly of the synthesized non-ionic dendronized polymers 10a–f in aqueous media was investigated using surface tension measurements, dynamic light scattering (DLS), and cryogenic transmission electron microscopy (cryo-TEM). Subsequently, the solubilization capacities of the aggregates were evaluated using pyrene as a hydrophobic fluorescent guest and the interactions with the guest molecules were investigated by ANS binding experiments.
2.2.1. Surface Tension Measurements
Surface tension measurements were carried out in a pendant drop apparatus. The data were plotted as logarithmic functions of the surfactant concentrations. The surface tension (γ) of the aqueous solutions was effectively reduced as the polymer concentration increased. A clear break in the curve marks the critical aggregation concentration (CAC; Figure1), for which the values are listed in Table1.
Table 1. Critical aggregation concentration (CAC) of polymers 10a–f in aqueous solution by surface tension measurements at 25 °C.
Composition of R [%]
As was expected from its overall hydrophilic nature, dendronized polymer 10a completely dissolved in water and thus did not show any CAC. Exclusively alkyl-functionalized polymer 10b had a CAC value of 2.00 × 10−5 M (153 mg L−1) and the dendronized multiamphiphilic polymers which differ in percentage of octadecyl chains and in the dendron generations, either [G1.0] (10c and 10d) or [G2.0] (10e and 10f), have CAC values of 2.94 × 10−5 M (262 mg L−1) and 2.50 × 10−5 M (224 mg L−1), and 2.41 × 10−5 M (247 mg L−1) and 2.08 × 10−5 M (203 mg L−1), respectively. Thus, the CAC of the dendronized multiamphiphiles is lowered both by a larger proportion of hydrophobic side chains (10d < 10c and 10f < 10e) as well as a higher generation of the dendrons (10e < 10c and 10f < 10d). Similar correlations were found for monomeric non-ionic dendritic amphiphiles by Haag et al.17e and for amphiphilic block copolymers composed of styrene and sodium acrylate by Eisenberg and co-workers.26 Poly(N-isopropylacrylamide-block-dendronized methacrylate) synthesized by Schlüter et al. showed much lower CAC values, however, ranging from 1.2–6.9 mg L−1.27
2.2.2. Cryogenic Transmission Electron Microscopy and Dynamic Light Scattering
Aggregation behavior of the polymers was studied by cryo-TEM and DLS at a general concentration of 10.0 g L−1, which is well above the CAC of the polymers. The cryo-TEM measurements confirmed that exculsively dendronized polymer 10a, completely dissolves in water without any sign of aggregation as was already anticipated from the absence of a CAC. All other multiamphiphilic polymers (10b–f) formed monodisperse spherical micelles as exemplified in Figure2. It was observed that most of the micelles tended to accumulate into large, densely packed domains, irrespective of compound and preparation method. Isolated particles were monitored only very rarely. Such domains were analyzed by Fourier transform (FT), which gave diffraction rings (insets in Figure 2) indicative of the repetitive packing distance of the micelles. The values can be used for a quite accurate determination of the mean diameter of the micelles as summarized in Table2. Although the diameter value only varies in a narrow margin of 8.1 ± 0.5 nm for the alkylated polymer (10b) and 8.4, 7.5, 8.4, and 7.8 nm (±0.8 nm) for the four dendronized multiamphiphilic polymers (10c–f), it is quite evident that the polymers with the higher alkyl fractions (10d and 10f) have smaller diameters. Errors are estimated from the width of the diffraction rings.
Table 2. PDI and particle size of the micelles 10b–f above the CAC in aqueous solution as measured by DLS (by intensity) and cryo-TEM at 25 °C. These values are compared with those from geometrical calculations. The hydrophobic volume of the micelles was calculated to be 113.1 nm3 for all five polymers.
Calculated Volumes [nm3]
A closer look at the FTs of all samples (cf. SI) unveils the interdependence of size distribution and the content of dendrons. Thus the diffraction rings from the alkylated polymer (10b) are quite sharp, and clearly point to only minor deviations of the micelles mean diameter, while those of the dendronized multiamphiphiles (10c–f) are blurred to some extent, indicating that packing is not so accurate in these cases. This is brought about by a less defined hydrophilic seam of low contrast due to PEG chains, which are barely visible in the images but might cause varying inter-assembly distances. Within the dendronized multiamphiphiles, the diffraction rings from 10e (50% alkyl, 50% [G2.0]) are broadened the most, probably because of the large number (∼127, cf. Experimental Section, Geometrical Calculations of Micellar Dimensions) of the more space-consuming [G2.0] dendrons, which are also not constrained in their conformation.
From our earlier experiments with monomolecular amphiphiles17e we expected a significant contribution of the PG dendrons to the contrast of images for the dendronized multiamphiphilic polymers 10c–f. Anyhow, the observed micelles are very similar to the micelles of the solely alkyl grafted polymer and thus seem to lack this additional high-contrast corona. This indicates less organization of the PG dendrons on the micellar surfaces compared to the monomeric analogues, where intermolecular interactions led to structurally defined micelles.17e
Cryo-TEM measurements were complemented by DLS measurements. The exclusively alkyl grafted polymer 10b provided hydrodynamic diameters of 12.5 nm, which is in keeping with the value obtained from the cryo-TEM measurements (8.1 nm) considering the poor visibility of the hydrated shell. Furthermore, the DLS data display considerably higher hydrodynamic diameters of 15.3 and 14.0 nm for the [G1.0] multiamphiphiles 10c and 10d and 19.3 and 17.2 nm for the [G2.0] multiamphiphiles 10e and 10f. The observed DLS data indicates a general trend that while the more alkyl-grafted polymers form smaller micelles, higher generation dendron [G2.0] grafted polymers form larger ones (cf. Table 2).
Furthermore, the DLS graphs (by intensity) of all polymers (10b–f) display components with hydrodynamic diameters larger than 100 nm, which might be agglomerated aggregates or free (not aggregated) polymer strands. However, the absolute numbers of such particles must be in the ppm range, as can be inferred from the absence of corresponding bands in the volume and number distributions (cf. SI).
It should be emphasised here that the differences between cryo-TEM and DLS data do not necessarily arise from different particles, since both methods display completely different phenomena. TEM pictures depict those parts of individual aggregates where electron density deviates from the surrounding medium (usually water), while strongly hydrated functional groups like PEO are not shown. DLS, on the other hand, measures statistical averages of particle sizes, which, in contrast, not only includes all parts (atoms) of the aggregates along with their hydration shell, but is also influenced by volume-independent properties of the aggregates like shape and mutual interactions. Therefore, hydrodynamic radii often exaggerate proper particle sizes.
A relevant comparsion can be made by the dendronized polymers reported by Chen et al. They studied Fréchet-type benzyl ether dendrons with a carboxyl group (G2, G3) at the apex site attached to poly(4-vinylpyridine) (PVP), forming hydrogen-bonded dendronized polymers. These polymers showed unique self-assembly behavior, forming aggregates with average diameters from 100 to 400 nm. They observed that the size of the vesicles decreased and the thickness of the vascular membrane increased with the increase in the molar ratio of the dendrons to PVP.28
2.2.3. Structural Model of the Micelles
Geometrical calculations of the micelles dimensions (cf. SI) render overall diameters of 8.4 nm (10b), 8.8 and 9.2 nm (10c and 10e) and 8.2 and 8.4 nm (10d and 10f), and point to approximately 127 (10c and 10e) or 60 (10d and 10f) PG groups (whether [G1.0] or [G2.0]), respectively, within their hydrophilic shells. As can be seen from Table 2, the calculated micellar diameters commensurate well with the average values obtained from the Fourier transforms of the cryo-TEM images, within the limits of error (0.5–0.8 nm). These calculations indicate that, from the geometrical point of view, only minor differences in the diameters can be expected for all five aggregating polymers. Graphical models representing the principle differences of the molecular arrangement in the micellar systems are presented in Figure3.
If the micelles are pushed together to form 2D ordered arrays, the FT should yield the inter-assembly distance (center to center), which is indicative of the overall diameter of the micelles in the ideal case. Blurring of the diffraction rings indicates variations in the particle distance, which might be due to variations in the head group corona geometry and packing density. Actually, the surface should be crimped with the dendrons and the overall spherical shape might be distorted by the mutual compression of the micelles. Therefore, it is not surprising that the observed mean diameters (by cryo-TEM) are slightly smaller than the calculated ones, while the hydrodynamic diameters should, in contrast, be slightly larger. The presented DLS data, however, clearly exceed these theoretical values. This points either to an extensive hydration shell, or (depending on the generation of the hydrophilic headgroups) to attractive interactions between micelles, which might enable the formation of bigger clusters of micelles. Such attractive forces can also be concluded from the ordered domains frequently observed on the cryo-TEM micrographs, although the individual number of interacting micelles remains shrouded in the extensive 2D arrangement, which might be also introduced in the sample preparation step.
Discrepancies in the aggregation behavior of the five multiamphiphilic polymers, depending on the type of linked dendrons and the grafting proportions at the polymeric back-bone (Figure 3), became apparent from the DLS measurements. Their nature and origin, however, can only be estimated from a detailed comparison of DLS and cryo-TEM data. The differences as well as the observed clustering in cryo-TEM point to attractive interactions between the micelles, which may be induced (1) by attractions of the [G1.0] or [G2.0] dendrons, or (2) from intermicellar chaining due to the intercalation of the polymer strands, probably due to the nearly 4 nm-long water-soluble PEO spacers (Figure4). Such intermicellar chainings are frequently observed with ABA- or ABC-type copolymers in different media, if the B-block is solvophilic.29–31
2.2.4. Encapsulation Studies
The polymeric multiamphiphiles were evaluated for their guest solubilization properties using pyrene as a hydrophobic guest molecule. To establish the polymer structure with guest solubilization property, the quantity of encapsulated and, hence, solubilized pyrene in the different multiamphiphiles was evaluated. Our focus has been to understand the effect of different hydrophobic (core) and hydrophilic (shell) content on the dendronized multiamphiphilic polymers on its host–guest properties and to evaluate the influence of the hydrophobic/hydrophilic ratio of the multiamphiphiles on the solubilization capacity and solubilization efficiency for lipophilic guests.
Dendronized polymer 10a, which lacks any hydrophobic alkyl chains, did not incorporate the hydrophobic guest pyrene, while polymers 10b–f have considerable guest solubilization properties. Thus, within this group of polymers the presence of an internal hydrophobic core is the essential structural feature for the encapsulation of hydrophobic guest molecules. Figure5a shows UV spectra of pyrene in 5 mg mL−1 aqueous solutions of different multiamphiphiles. The absorption intensity at λmax (329 nm) is directly correlated to their guest solubilization capacities (defined as moles of guest molecule per mole of amphiphile). In the row of the three multiamphiphiles with the same percentage of hydrophobic content (50% alkyl), [G1.0] dendronized multiamphiphile 10c has the lowest solubilization capacity (1 mol pyrene per mol of polymer), followed by non-dendronized multiamphiphile 10b (1.4 mol pyrene per mol of polymer) and the [G2.0] dendronized multiamphiphile 10e (2.1 mol pyrene per mol of polymer) (Table3). In general, multiamphiphiles with [G2.0] dendrons (10e, 10f) provide better guest solubilization capacities (2.1, 2.3 mol pyrene per mol of polymer, respectively) compared to those with [G1.0] dendrons (10c, 10d: 1.0, 1.9 mol pyrene per mol of polymer; Figure 5b). Thus, the generation of the dendrons plays a significant role for the guest solubilization properties of these multiamphiphiles: [G2.0] dendrons stabilize the micelles, as indicated by the higher solubilization capacities, while [G1.0] dendrons seem to destabilize them, if compared to polymer 10b.
Table 3. Guest solubilization capacities and efficiencies of the polymers 10a–f (conc. of 5 g L−1) for solubilizing hydrophobic guest pyrene.
Solubilization capacity [nguest/npolymer]
Solubilization efficiency [mg g−1]
The guest solubilization efficiency (milligrams of guest per gram of amphiphile) increases, however, with an increased percentage of hydrophobic content (Table 3). The dendronized multiamphiphiles 10c and 10e (50% alkyl grafted) encapsulate less pyrene (20.9 mg and 42.4 mg, respectively) than 10d (67.1 mg) and 10f (74.8 mg) (70% alkyl grafted), respectively. Altogether, our new dendronized multiamphiphilic polymers show solubilization efficiencies of up to 7.5 wt% (10f), which is considerably higher if compared with monomolecular non-ionic dendritic amphiphiles (up to 2 wt%),17e and thus, highlights the versatility of the dendronized multiamphiphilic polymeric architectures for prospective drug delivery applications over corresponding monomolecular amphiphiles.
The solubilization site of pyrene in the dendronized multiamphiphilic polymers was also studied based on the well-known sensitivity of pyrene fluorescence on the polarity of its local environment.32, 33 Evidence for the presence of hydrophobic environment of pyrene is obtained from the higher values of polarity index (I3/I1, the ratio of the intensities of the third (384 nm), and the first (373 nm) vibrational peak in the emission spectrum).34 The values of I3/I1 range from 0.555 in water to 1.666 in aliphatic hydrocarbon solvents.34 The maximum values obtained for the dendronized multiamphiphilic polymers 10c–f were 0.790, 0.792, 0.798, and 0.802 respectively, at a polymer concentration of 1.0 mg mL−1, suggesting that pyrene molecules are likely to be located in a hydrophobic pocket. A value of 0.618, was reached for the dendronized polymer 10a, and 0.832 for the non-dendronized multiamphiphilic polymer 10b, which indicates a significantly higher overall polar environment of pyrene in polymer 10a and a significantly higher nonpolar environment of pyrene in polymer 10b, respectively, compared to the dendronized multiamphiphilic polymers 10c–f. These results can be well understood from the structural composition of the polymers, as the dendronized multiamphiphilic polymer 10a lacks a hydrophobic pocket. Thus the I3/I1 value corresponds to an environment of high polarity (most approximates to ethylene glycol).33 Polymers 10b–f, on the other hand, have a hydrophobic pocket composed of alkyl chains to solubilize pyrene which leads to a change in its microenvironment.
2.3. Binding Studies with 1-Anilinonaphthalene-8-sulfonic Acid
The encapsulation experiments with pyrene suggest that these new multiamphiphiles can be employed for encapsulation of active agents/drugs and may find potential biomedical application as drug carriers. However, for drug delivery, it is essential to know the polymer's binding characteristics and association with guest molecules. Quantitative analysis of the polymer molecule binding capacity and affinity can be assessed by encapsulation of small fluorescent dyes.35 Since all relevant multiamphiphiles 10b–f show the similar micellization behaviors, the dendronized multiamphiphile 10f with the most apparent guest solubilization properties was chosen for the ANS binding studies.
The fluorescence of 1-anilinonaphthalene-8-sulfonic acid (ANS) is sensitive to changes in the microenvironment,36 which allows the evaluation of host–guest binding interactions by spectrofluorometric methods. Due to dynamic quenching by water molecules, ANS shows only low fluorescence in the range between 400 and 600 nm with an emission maximum (λmax) at 525 nm in aqueous media. The fluorescence is greatly enhanced and is accompanied by a blue-shift of the band, as the solvents polarity decreases. Upon addition of polymer 10f the ANS fluorescence was significantly increased (Figure6a) and gradually led to a blue shift of λmax to about 477 nm. Since the polymer solution did not show any bands in this range, these changes arose due to the binding of ANS and suggest that the dye was placed in a hydrophobic environment, well shielded from the surrounding water molecules. By further increasing the polymer concentration the fluorescence began to show signs of maximizing as more ANS was bound. From the resulting hyperbolic curve (Figure 6a), the value of Fsp was calculated (cf. Experimental Section). The fluorescence intensity (F) was measured in the second fluorometric titration, where the system was reversed and increasing concentrations of ANS were added to a constant concentration of the multiamphiphile (Figure 6b). From these values the concentrations of bound ANS (CboundANS) and free ANS (CfreeANS) were calculated and the Scatchard's plot (C10h/CboundANS against 1/CfreeANS in Figure 6c) was drawn. The linear regression of the plot (Figure 6c) indicates that ANS penetrates inside the micelles, where only one type of binding center (or hydrophobic pocket) is located. Values for Kb ∼ 1.66 × 105M−1 and n ∼ 0.34 per molecule of the polymer 10f were calculated.
A new class of non-ionic dendronized multiamphiphilic polymers and their self-assembly to micellar nanocarriers has been presented. The polymers are constructed from biocompatible PEG chains and PG dendrons to form a hydrophilic outer shell and alkyl chains moulding for a hydrophobic interior. The polymers backbone was synthesized from linear PEG and 2-azido-1,3-propanediol using a biocatalytic pathway. The versatility of this basic structure originates from its azido groups which facilitate an easy functionalization by click grafting with appropriate alkynes (alkyl groups or dendrons). This opens the way to a new class of multiamphiphilic surfactants, where the hydrophobic/hydrophilic balance as well as functionalities may be easily readjusted as required. The flexibility of the PEG block allows for an independent aggregation of hydrophilic and hydrophobic groups.
As a proof of concept, the polymeric backbone was grafted with polyglycerol dendrons of different generations and octadecyl chains in varying proportion, this led to a series of new multiamphiphilic polymers that self-assemble into well-defined, globular micelles with diameters in the 10 nm range at low critical aggregation concentrations (10−5 M), as was confirmed by DLS, cryo-TEM, and surface tension measurements. The potential ability of the micelles to function as nanocarriers as prospective drug delivery systems was illustrated with the hydrophobic model dye pyrene. The hydrophobic/hydrophilic balance of the multiamphiphiles plays an important role in guest encapsulation and solubilization. Due to the higher ordered structure of larger dendrons, [G2.0]-PG dendron functionalized multiamphiphiles show better guest solubilization efficiencies than [G1.0]-PG dendronized multiamphiphiles and non-dendronized multiamphiphiles. Additional studies with ANS revealed unimolecular binding characteristics within the hydrophobic domain of the micelles that indicates the guest molecules are predominantly located in the internal hydrophobic cores of the aggregates. Alltogether, this new class of dendronized multiamphiphilic polymers demonstrates the formation of notable uniform micelles with higher guest solubilization capacities and efficiencies compared to monomolecular amphiphiles, added biocompatibility due to the use of PEG and PG dendrons, and lower exchange rates due to covalently linked multiamphiphilic architecture, which all ought to enhance the potential use as nanocarriers for biomedical applications. Further work is underway to study the cellular uptake mechanism of these new multiamphiphiles, fine-tune the structure to achieve more stable architectures, and obtain controlled drug release profiles.
We believe that this new concept will facilitate the preparation of tailor-made nanocarriers adapted to specific needs and targeted to specific tissues. Following this concept, multi-functional nanocarriers may in the future be easily constructed by picking up the required side chains from a ‘construction kit’ of suited functional groups and clicking them at a polymeric back-bone of a convenient length.
4. Experimental Section
All the chemicals and solvents used were of analytical grade and purchased from Sigma-Aldrich GmbH (Munich, Germany) and Fisher Scientific GmbH (Schwerte, Germany). Novozym-435 was purchased from Julich Chiral Solutions GmbH (Jülich, Germany). Analytical TLCs were performed on precoated Merck silica-gel 60 F254 plates; silica gel (100–200 mesh) was used for column chromatography. Milli-Q water (Millipore GmbH, Schwalbach/Ts., Germany, resistivity ∼18 MΩ.cm, pH = 5.6 ± 0.2) was used in all experiments and for preparation of all samples. Dialysis was performed using benzoylated dialysis tubing (MW cutoff 2000 Da) from Sigma-Aldrich, changing the solvent three times over a period of 24 h. 1H and 13C NMR spectra were recorded on a Bruker DRX 400, and a Bruker AMX 500 MHz spectrometers (Bruker GmbH, Karlsruhe, Germany). The spectra were calibrated by using the solvent peaks. The chemical shift values are on δ scale and the coupling constant values (J) are in Hz. IR spectra of neat samples were recorded on a Nicolet Avatar 320 FT-IR spectrometer (Thermo Fisher Scientific, Dreieich, Germany). A TSQ 7000 (Finnigan MAT GmbH, Bremen, Germany) instrument was used for ESI measurements, and a JEOL JMS-SX-102A spectrometer (JEOL, Eching, Germany) for high-resolution mass spectra. Elemental analyses were performed on a Perkin-Elmer EA 240. Molecular weight and molecular weight distribution / of polymers were determined using a GPC equipped with Agilent 1100 pump, refractive index detector, and PLgel and Suprema columns. The polymers were eluted with THF at a flow rate of 1.0 mL min−1. Molecular weights were calibrated with pullulan standards. Critical aggregation concentrations of the polymers were determined via surface tension measurements by a commercially available pendant drop tensiometer OCA 20 from Dataphysics, Germany. Aqueous samples in Milli-Q water were prepared 24 h before use and measurements were done at 25 ± 5 °C. The aggregation behavior of the multiamphiphiles in water was studied over the concentration range of 7 × 10−7 to 1 × 10−3 M. The surface tension of each hanging drop was determined two times per minute and the measurement was stopped when the surface tension did not change by more than 0.1 mN m−1 over two minutes. Calculations were done using the Young–Laplace equation.
Light Scattering Measurements: Dynamic light scattering measurements were carried out on a Zetasizer Nano ZS analyzer with integrated 4 mW He–Ne laser, λ = 633 nm (Malvern Instruments Ltd, Worchestershire, UK), using backscattering detection (scattering angle θ = 173°) with an avalanche photodiode detector. The Zetasizer is equipped with a thermostated sample chamber controlled by a thermoelectric peltier. The respective polymer was dissolved in Milli-Q water and vigorously stirred for at least 12 h at room temperature. The mixture was filtrated via 0.45 μm polytetrafluorethylene (PTFE) filters and allowed to equilibrate for 6 h at room temperature. Disposable UV transparent cuvettes (12.5 × 12.5 × 45 mm, Sarstedt AG & Co, Nümbrecht, Germany) were used for all measurements.
Cryogenic TEM: Droplets of the sample solutions were applied to perforated (1 μm hole diameter) carbon film-covered 200 mesh grids (R1/4 batch of Quantifoil MicroTools GmbH, Jena, Germany), which had been hydrophilized before use by 60 s plasma treatment at 8 W in a BALTEC MED 020 device. The supernatant fluid was removed with a filter paper until an ultrathin layer of the sample solution was obtained spanning the holes of the carbon film. The samples were immediately vitrified by propelling the grids into liquid ethane at its freezing point (90 K) with a guillotine-like plunging device. The vitrified samples were subsequently transferred under liquid nitrogen into a Philips CM12 TEM (FEI, Oregon) using a Gatan (Gatan, Inc., California) cryoholder and stage (model 626). Microscopy was carried out at a 94 K sample temperature using the low-dose protocol of the microscope at a primary magnification of 58 300× and an accelerating voltage of 100 kV (LaB6-illumination) with the defocus set to 1.8 μm.
Absorbance Measurements: Absorption spectra were recorded between 220 and 800 nm using a Scinco S-3150 UV–vis spectrophotometer (Scinco Co. Ltd., Seoul, Korea) (range: 187–1193 nm; resolution: 1024 points). All measurements were carried out in a thermostated UV cell (1 cm).
Pyrene Encapsulation Experiment: An excess amount of pyrene was added in solid state to aqueous polymer solutions of fixed concentrations. After stirring for 12 h, the mixture was filtrated twice via 0.45 μm pore size PTFE syringe filters to remove insoluble excess of pyrene. The clear polymer solutions were freeze-dried and re-dissolved in chloroform. The chloroform solutions were then analyzed by means of UV–vis spectroscopy. The molar extinction coefficient (ϵmax) was calculated from a calibration curve of pyrene in chloroform at λ = 329 nm and used for the concentration measurements.
Fluorescence Measurements: Fluorescence emission spectra were taken with a Jasco FP-6500 spectrofluorimeter equipped with a thermostated cell holder, a DC powered 150 W Xenon lamp, a Hamamatsu R928 photomultiplier (Hamamatsu Photonics UK Limited, Essex, UK), and a variable slit system. For pyrene-copolymer solutions emission spectra were recorded between 350 and 500 nm after excitation at 320 nm and for ANS binding experiments emission spectra were recoreded between 400 and 600 nm after excitation at 360 nm. Both excitation and emission slit were set at 5 nm. All data analysis was performed using Microsoft Excel and Sigma Plot 8.0 software.
ANS Binding Experiments: The binding experiments were carried out with dendronized multiamphiphile 10f and ANS as the anionic fluorescent probe. Sample solutions were prepared in phosphate buffered saline (PBS: 20 mmol L−1, pH = 7.6). Binding constant (Kb) and number of binding centers per aggregate (n) were determined by a double fluorometric titration technique and calculated using Scatchard-Klotz analysis.35, 37 For the first fluorometric titration, increasing concentrations of polymers were added to a solution of constant ANS concentration (C1ANS). The maximum intensity of ANS fluorescence (Fmax), indicating complete binding of ANS to the micelles, was then determined. The specific fluorescence intensity (Fsp) for bound ANS was calculated as Fmax/C1ANS. In the second fluorometric titration, the fluorescence intensity (F) in the reversed system (increasing concentrations of ANS and a constant concentration of the polymer (Cpolym)) was measured. The concentration of ANS bound (CboundANS) was calculated as F/Fsp and the concentration of free ANS (CfreeANS) was calculated as (C2ANS–CboundANS). Plots of C10f/CboundANS against 1/CfreeANS and linear regression gave the values for Kb and n according to the following equation:
Supporting Information is available from the Wiley Online Library or from the author.
We gratefully acknowledge the financial assistance by the Center for Supramolecular Interactions (CSI) within the focus area Nanoscale at the Freie Universität Berlin, the Deutsche Forschungsgemeinschaft DFG, and Indo-German Science & Technology Centre (IGSTC) supported by the Bundesministerium für Bildung und Forschung (BMBF) and Departments of Science & Technology (DST), Delhi.