Drug-Induced Morphology Transition of Self-Assembled

Article pubs.acs.org/cm

Cite This: Chem. Mater. 2018, 30, 5227−5236

Drug-Induced Morphology Transition of Self-Assembled Glycopolymers: Insight into the Drug−Polymer Interaction Cheng Cao,†,‡ Jiacheng Zhao,† Fan Chen,† Mingxia Lu,† Yee Yee Khine,† Alexander Macmillan,§ Christopher J. Garvey,*,‡ and Martina H. Stenzel*,† †

Centre for Advanced Macromolecular Design, School of Chemistry, The University of New South Wales, Sydney 2052, Australia Australia Nuclear Science and Technology Organisation, Lucas Heights 2234, Australia § Biomedical Imaging Facility, University of New South Wales, Sydney 2052, Australia Downloaded via REGIS UNIV on October 21, 2018 at 18:35:25 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡

S Supporting Information *

ABSTRACT: It is often assumed that a hydrophobic drug will be entrapped in the hydrophobic environment of a micelle. Little attention is usually drawn to the actual location of the drug and the effect of the drug on properties. In this publication, we show how the chosen drug curcumin is not only unexpectedly located in the shell of the micelle but also that the accumulation in the hydrophilic block can lead to changes in morphology during self-assembly. A block copolymer poly(1-O-methacryloyl-βD-fructopyranose)-b-poly(methyl methacrylate), poly(1-O-MAFru)36-b-PMMA192, was loaded with different amounts of curcumin. The resulting self-assembled nanoparticles were analyzed using transmission electron microscopy, small-angle X-ray scattering (SAXS), and small-angle neutron scattering (SANS). Initial microscopy evidence revealed that the presence of the drug induces morphology changes from cylindrical micelles (no drug) to polymersomes, which decreased in size with increasing amount of drug. SAXS and SANS analysis, supported by fluorescence studies, revealed that the drug is interacting with the glycopolymer block. The drug influenced not only the shape of the drug carrier but also the level of hydration of the shell. Increasing the amount of drug dehydrated the nanoparticle shell, which coincided with a lower nanoparticle uptake by MCF-7 breast cancer cells and noncancerous RAW 264.7 cells. As a result, we showed that the drug can influence the behavior of the nanoparticle in terms of shape and shell hydration, which could influence the performance in a biological setting. Although the depicted scenario may not apply to every drug carrier, it is worth evaluation if the drug will interfere in unexpected ways.

■

INTRODUCTION Polymeric micelles have been applied widely as a drug delivery system due to the high drug loading efficiency, long-circulating properties,1 and the possibility to decorate the surface with a high density of targeting ligands.2 The properties of micelles can be easily adjusted by fine-tuning the block lengths and ratios, which influences size and stability of the micelle. These defining propertiessize, stability, and surface chemistry influence the fate of the micelles in vitro and in vivo.3−7 Although micelles are well-studied in regards to the relationship between physicochemical properties and behavior in biological systems, little attention has been drawn to the effect of the amount and type of drug on the activity. There is currently a surprisingly limited amount of reports available that investigate the effect of drug loading on the activity. The common wisdom is that high amounts of loading is preferred, which is understandable considering that the drug carrier is only the packaging material that needs to be disposed of after usage. Moreover, low drug loading content can have a serious © 2018 American Chemical Society

implication on the amount of nanoparticles required to treat cancer in humans or animals as the required injection volume during administration will exceed feasible amounts.8 Therefore, researchers usually target the highest possible drug loading efficiency to limit the amount of drug carrier required. This can be achieved by reducing the crystallinity of the core-forming block to enhance the drug loading capacity.9 Other approaches to increase loading focused on increasing the compatibility between polymer and drug10,11 or on adjusting the glass transition temperature.12 To push the amount of drug in a micelle even further, stronger polymer−drug interaction is introduced by attaching functional groups such as boronic acid that can display strong donor−acceptor coordination13 or, ultimately, by covalently conjugating the drug to the polymer.14,15 The mantra that the higher drug loading the Received: May 6, 2018 Revised: July 13, 2018 Published: July 16, 2018 5227

DOI: 10.1021/acs.chemmater.8b01882 Chem. Mater. 2018, 30, 5227−5236

Article

Chemistry of Materials

Figure 1. Synthesis and self-assembly of poly(1-O-MAFru)36-b-PMMA192 and various morphologies formed by self-assembly of poly(1-OMAFru)36-b-PMMA192 (2 mg/mL) in existence of different amounts of curcumin: 0, 0.37, and 0.75 mg/mL by TEM. Scale bar = 500 nm.

we are confident, it has been frequently observed in the lab that the presence of drugs can affect the self-assembly process, there are no in-depth studies to understand the underpinning principles that influence the morphology transition. In this paper, we show that the drug can have strong tendency to interact with the shell despite the disparate polarities between drug and polymer shell. We believe that this may be common when polymers such as glycopolymers are employed. Glycopolymers are attractive materials as they display bioactivity, but they can also form strong interactions with other compounds using, for example, hydrogen bonding, which would be absent in more common polymers such as PEO. This interaction can have far-reaching effects as demonstrated here using a block copolymer based on PMMA and a glycopolymer (Figure 1). Curcumin was used as the model drug as it displays antioxidant, anti-inflammatory, and antitumoral properties.24,25 With the help of synchrotron small-angle X-ray scattering (SAXS) and small-angle neutron scattering (SANS), we identified the location of curcumin in the system. The drug position was found to affect the selfassembly process, resulting in morphology transitions. Moreover, understanding the location of the drug and its interaction with the polymer is crucial to appreciate not only the morphology changes but also the effect of the drug location can have on the biological activity.

better is, however, not always valid, as has been shown in a study using docetaxel-loaded nanoparticles. The lower weight loading (9%-NP vs 20%-NP) was observed to have a better pharmacokinetic profile and was more efficient in the fight against murine cancer models.16 At this point, it is crucial to understand the physicochemical parameters to be able to relate the properties of micelles to their performance in media. The question such as the location of the drug in the micelle (shell or core) and how this may affect the mobility of the polymer chains is instrumental to understand why the amount of drug in a carrier can have such large influences. The location of the drugs can be targeted by conjugating drugs either to the shell or the core. Doxorubicin in the shell was more readily cleaved, and therefore, the nanoparticle was more toxic.17 However, even drugs bound to the core-forming block can affect the cellular uptake as their location and the conjugated amount can influence the shell properties such as the ζ-potential18 or the hydrophilicity and mobility of the polymer.19 Moreover, different drugs, despite being bound to the micelle in a similar manner, can influence the physicochemical behavior of the drug carrier and influence the outcomes in vivo.20 Binding the drug onto the polymer allows some control over the location, but a physical encapsulation of the drug into the micelle can lead to surprising results. Although it is assumed that hydrophobic drugs will prefer the hydrophobic environment, the drug can be accumulated within the shell as long as the drug can interact with the respective polymer.21,22 In this example, the size of the core influenced the crowding of the shell which, in turn, affected the strength of binding between drug and shell-forming block.21 The effect of the drug on the properties of the micelles should therefore not be simplified, as the drug type and amount may affect micelle stability23 and the behavior of the shell-forming block.18−20 The drug may also display strong interactions with the shell-forming block.21 However, little is known about how the drug can affect the morphology itself, which ultimately will influence the biological activity. Although

■

RESULTS AND DISCUSSION Self-Assembly in the Presence of Curcumin and Biological Evaluation. The poor bioavailability of curcumin can be addressed by the use of a suitable nanocarrier to enhance the therapeutic efficacy of this promising natural product.26 It was reported that worm-like micelles are promising nanocarriers as they lengthen the circulation time of drugs in the blood.27 We therefore investigated the formation of spherical and cylindrical micelles of different aspect ratios using poly(1-O-MAFru)-b-PMMA. Fructosecontaining polymers were observed to display an enhanced uptake by breast cancer cells and can therefore be used to 5228


Article


Table 1. Size of Self-Assembled Drug-Loaded Poly(1-O-MAFru)36-b-PMMA192 Nanoparticles Using Different Amount of Curcumin nanoparticles

[curcumin] (mg/mL)

drug loading efficiency (%)

drug loading capacity (%)

Dh (nm) (DLS)

PDI

ζ-potential (mV)

sample A (cylinder) sample B (polymersomes) sample C (polymersomes)

0 0.37 0.75

0 36.5 20.8

0 6.8 7.8

167 ± 5 272 ± 3 263 ± 2

0.062 ± 0.012 0.118 ± 0.015 0.124 ± 0.020

−25.7 ± 1.5 −22.6 ± 1.2 −19.5 ± 0.8

Figure 2. Flow cytometry on MCF-7 cell line (a) and RAW 264.7 cell line (b). The polymer concentration was set to 0.2 mg/mL, and the concentration of loaded drug ranges from 0 to 0.0375 mg mL−1 (see Table S2). (c and d) IC50 curves of curcumin-loaded polymersomes in 2D MCF-7 and RAW 264.7 cell culture models. The DLC in panels c and d can be found in Table S1. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

selectively deliver drugs.28,29 While these glycopolymers show a high affinity to breast cancer cells, it is not clear how drug loading will affect the activity of these carriers. A block copolymer, poly(1-O-MAFru)36-b-PMMA192, was synthesized as the chosen block lengths can be processed into cylindrical micelles to create potentially long-circulating drug carriers.29 Typically, block copolymer and curcumin were dissolved in THF, followed by adding MQ water into the solution (Figure 1). After dialysis, the amount of curcumin loaded into the nanoparticle was measured using UV−Vis analysis. From this, the drug loading efficiency (DLE) and the drug loading capacity (DLC) were measured (Table S1). As a control, the block copolymer without added drugs was subjected to the same process. Although the drug-free system led to the expected cylindrical micelles according to transmission electron microscopy (TEM) analysis, polymersomes instead of cylindrical micelles were obtained in the presence of curcumin (Figure 1). The morphologies were stable over an extended period of time. The stabilities of the nanoparticles were confirmed by TEM analysis after the liquid samples were stored for 6 months and reanalyzed (Figure S1). Moreover, dynamic light scattering (DLS) data showed that the size of curcumin-loaded polymersomes is affected by the concentration of curcumin in the solution. The diameter of the

polymersome decreased from 272 to 263 nm when the concentration of curcumin in the solution during assembly increased from 0.37 to 0.75 mg/mL (Table 1). The diameter distributions measured by TEM for samples B and C in Table 1 are shown in Figure S2. The mean diameter is 179 and 134 nm, respectively, which is smaller than the sizes measured by DLS, most likely due to the dehydrated state. The result suggests that the curcumin alters with the self-assembly process in a concentration of the drug fashion. The decline in polymersome size coincides with the increase in surface curvature, which is triggered by increased area demands of the hydrophilic polymer or a decreased volume of the hydrophobic core. Our observations may suggest that curcumin is in the fructose shell, which could have implications on the properties of the shell such as hydration and swelling as well as surface charge. This would ultimately influence the biological activity. It is evident that the presence of curcumin led to a morphology transition, but at this stage, it is unknown if the aggregate changes are permanent or if the morphology can revert back to cylindrical micelles. Initially, the drug release of both drug-loaded carriers (Table 1) was investigated over 72 h in PBS solution (pH 7.4). The drug-loaded nanoparticles were placed in a dialysis bag, and the sample was dialyzed against water. At regular time intervals, a sample was taken from inside 5229


Article


Figure 3. (a) The fluorescence spectra of various concentration of curcumin loaded in the polymersomes and λex = 421 nm. (b) The lifetime plots versus different amounts of curcumin loaded in NPs. (c) The lifetime plots versus different amounts of curcumin in the water. (d) Average lifetime versus different amounts of curcumin loaded in the nanoparticles and different amounts of curcumin in the water.

ments using the concentrations listed in Table S2. The polymers were labeled with conjugated cyanine 5 (Cy5< 1 wt %), which does not interfere with the fluorescence emission spectrum of curcumin. The mean fluorescence intensity (MFI) of polymersomes was normalized after subtraction of the cellular autofluorescence (Table S2 and 3). Surprisingly, polymersomes loaded with a higher amount of drugs showed a lower uptake by MCF-7 and RAW 264.7 cell lines (Figures 2a and b) despite their smaller size. Moreover, the surface charge does not change significantly. If anything, the lower particles with the lower drug loading and the more negative charged surfaces should show preferred uptake, which contrasts actual uptake measurements. The polymersomes with the highest amount of drugs must obviously present cues that may not only lower the cellular uptake but also lower cytotoxicity (Figures 2c and d and Table S4). Therefore, spectroscopic and scattering analyses were carried out to obtain more details on the internal structure of drug-loaded nanoparticles. Physico-Chemical Characterization of CurcuminLoaded Nanoparticles. The local physicochemical environments in self-assembled nanoparticles can be probed by spectroscopic techniques using the solvatochromic shift of the UV−Vis absorbance.31 To understand the effect of the local dielectric environment sensed by the probe, the UV−Vis absorption maxima λmax of curcumin in various solvent as well as 1,4-dioxane−water mixtures were measured (Tables S5 and S6). The measured λmax of curcumin in the NP was equal to a dielectric constant of 39.18 (Figure S5), suggesting that curcumin is not located in the dielectric media of the PMMA shell but rather in a dielectric environment such as a mixture of water and fructose.32

the bag, and the amount of curcumin was measured by UV− Vis spectroscopy. Although typical setup measures the drug on the outside of the dialysis bag, this was not possible here as curcumin accumulated on the surface of the bag. Both carriers displayed similar release rates (Figure S3), albeit the drug carrier with the higher drug loading capacity had a slightly faster rate of release (78 vs 72% of loaded drug released after 72 h). The nanoparticles after the release of curcumin were studied using TEM analysis (Figure S4). It is interesting to note that despite the release, the morphology was retained and does not revert to the cylindrical micelles found when selfassembling the block copolymer without the drug. Once formed, the polymersomes are kinetically stable against rearrangements with changes of drug content as the glass transition temperature of PMMA is well above ambient temperature. Subsequently, the effect of drug loading on the biological activity such as uptake of nanoparticles by cells was evaluated. As cylindrical micelles will naturally have different cellular uptake patterns independent of drug loading, poly(1-OMAFru)36-b-PMMA192 was therefore processed into drug-free polymersomes by applying a stirring rate of 300 r/min. This trapped the block copolymer in the desired spherical shape with a hydrodynamic diameter of 271 nm instead of the usual cylindrical micelle that polymer likes to form. Drug-free polymersomes and drug-loaded polymersomes can then be used for direct comparison in subsequent studies without having to take shape factors into account. At this point, it was expected that the cellular uptake of polymersomes with higher drug loading would be more efficient as the smaller size should display more efficient cell uptake.5,30 The cell uptake using empty and drug-loaded polymersomes was investigated using flow cytometry measure5230


Article

Chemistry of Materials While the measured UV−Vis absorption maxima λmax = 421 nm is independent on the drug loading content (Figure S6), the fluorescence of curcumin is responsive to the environment33 and thus to the changes in the polymer with increasing drug content (Figure 3a). The emission maximum of curcumin undergoes a 10 nm Stokes shift when the curcumin concentration in the carrier increases, indicative of an environment of decreasing polarity. In addition, increasing curcumin concentrations leads to fluorescent self-quenching, which is frequently observed when the dye undergoes aggregation or stacking. The fluorescence lifetime of the drug (fluorophore) is sensitive to the drug’s environment due to the strong effect on the nonradiative decay.34 Figure 3b depicts the lifetime τ decrease with decreasing amount of curcumin loaded in the NPs while the lifetime τ of curcumin in water is less dependent on the concentration (Figure 3c). The fluorescence lifetime of curcumin loaded into nanoparticles leads to a longer decay time, which is increasing with increasing curcumin concentration (Figures 3b and d). Direct comparison of the curcumin fluorescent lifetime in water and in the nanoparticle (Figure 3d, S7 and Table S7) reveals restricted motion of curcumin in the nanoparticle. The lifetime of curcumin in MMA is rather short contrasting the long lifetime in water (Figure 3c). Compared to the actual lifetime of curcumin in the polymersomes, the long lifetime suggests that curcumin is more likely located in the glycopolymer shell. Subsequently, we gained more insight into the location of curcumin in the aggregate by scattering studies. The analysis of small-angle scattering data has been carried out by modeldependent fitting.35 The scattering curves were fitted with a mathematical model containing both spatial and compositional information. The length scales were initially constrained by TEM (shape) and DLS (size and polydispersity) measurements (Figure 1 and Table 1). The scattering length density (SLD) is a description of the atoms in a representative volume: SLD =

Figure 4. (a) SAXS data for the samples A, B, and C and the corresponding fit on the curve of sample A. (b) Core cylinder model and fitting results of sample A. The samples are listed in Table 1.

Figure 5. (a) SLD file for the SANS polymersome Model 2. (b) SANS data for sample [curcumin] = 0 mg/mL (no drug), 0.37 mg/ mL (low drug), and 0.75 mg/mL (high drug) and their respective curve fits.

∑i bci Vm

(1)

where the subscripts refer to the SLD and ⌀ (volume fraction) of the shell, water, curcumin, and fructose. If ⌀curcumin is negligible then, ⌀fructose + ⌀D2O = 1, and the ratio of water to polymer is easily calculated. Initially, the SAXS of NPs without curcumin was fitted with a core−shell cylinder and with curcumin fitted with core−shell, Model 1. There is no difference between NPs with low drug loading and high drug loading for SAXS in the high-q regime (Figure 4a). All the parameters we used are summarized in Table S8. The agreement between the SAXS data for the sample without curcumin with rod fit is obvious. Thus, it can be concluded that the SAXS is sensitive to the shape of the NPs, but not the structure of the shell. The modeling of the SAXS data provided some useful information: it gives an excellent indication of the average shape of the NPs. Due to the poor contrast between the poly(1-O-MAFru), solvent and curcumin (Table S9), the amount of information about the inner and outer shells that can be extracted from the high-q regime is limited. SANS measurements have the possibility of enhancing the contrast between the inner and outer shells and the solvent by using D2O as the solvent. Typical SANS curves from NP polymersomes in D2O for optimal contrast are shown in Figure 5b. We used two different

where bci is the bound coherent scattering density of all i atoms over the molecular or monomer volume, Vm. For SAXS and SANS, the coherent scattering cross sections are from the electronic and nuclear environments, respectively.36 Each sample with different concentrations for SANS and SAXS was measured and normalized by intensity, and they do not depend on the concentrations which can ignore the structure factor. For the fitting of the SAXS and SANS data, a core−shell cylinder and a three-shell around a water core model37 were used. The important parameters used in the models are illustrated in Figures 4b and 5a, respectively. The shell of the cylinder as well as inner and outer shells of the polymersome structures are composed of poly(1-O-MAFru), water, and in some cases curcumin. The middle shell is composed of PMMA with the physical properties of the amorphous material used as a starting point.38 We modeled the SLD and thickness of shells as well as the radius of the nanoparticles. The SLDs of the inner and outer shells allow us to determine the chemical composition: SLDshell = ⌀fructoseSLDfructose + ⌀D2OSLDD2O + ⌀curcuminSLDcurcumin

(2) 5231


Article


According to the variation of the SLD in the outer shell with drug loading (Table S10), we used the eq 2 to calculate the ⌀D2O in the outer shell (Figure 7a and b). This result shows

models to extract structure information from the scattering curves. In the low-q regime, the model can be used to extract the mean core radius of NPs, rcore. The high-q regime is sensitive to the thickness, Tshell, and the SLD of each shell of the NPs. Initially, we kept the SLD of the inner and outer poly fructose/water/curcumin shells equal (Model 1). Although the low-q fit, corresponding to the overall size and shape of the NP converged well, the high-q regime corresponding to the internal structure of the shells was a poor fit (Figure S8, Model 1). A much better fit can be obtained by including slight differences in the SLD’s and thicknesses between inner and outer shells (Figure 5a, Model 2). The results of the fitting are shown in the Table S10. SANS data of NPs with different amounts of curcumin are modeled to obtain the SLD, thickness of the shell. Figure 5b shows the overview of the fitting of SANS curves from samples with three different loadings of curcumin: ([no drug] = 0 mg/ mL; [low drug] = 0.37 mg/mL; and [high drug] = 0.75 mg/ mL). The fitting parameters are given in Table S10. The rcore is insensitive to the drug loading. A parameter which is very sensitive to drug loading is the thickness of PMMA shell; however, the inner and outer shells are poorly sensitive to the drug loading and show opposite trends, the inner shell getting slightly thicker and outer shell getting marginally thinner (Figure 6).

Figure 7. (a) Volume fraction of water in the outer fructose shell with respect to the different amount of curcumin according to the eq 2. (b) Possible morphologies of fructose chains in the outer shell with the different amounts of curcumin.

that the water content of the outer shell decreases with the increasing amount of curcumin. In Figure 2, increasing drug content slows the cellular uptake of NPs. The SANS results indicate that this slower cellular uptake is associated with an outer layer of the NPs which has a lower water content and the more compact layer of poly(1-O-MAFru) and curcumin. SANS and SAXS measurements confirm that the drug is indeed located in the shell of the polymersome, which agrees with the spectroscopic studies. The block copolymer itself has a tendency to form cylindrical micelles; however, with the addition of curcumin, the morphology changed to polymersomes. These polymersomes became smaller when more curcumin was added. Association of hydrophobic curcumin with the hydrophilic glycopolymer instead of the hydrophobic PMMA led to these influences in the self-assembly process. Important here is that the presence of curcumin in the shell does have far-reaching effects on the hydration. The declining levels of hydration may be responsible for the lower cellular uptake when the nanoparticles are loaded with drugs as the fructose polymer is less mobile and less likely recognized by the receptors on MCF-7 cells.19 It needs to be noted here that the morphological transitions were specific to this block length. Polymers with the smaller hydrophobic block, poly(1-OMAFru)35-b-PMMA160, were found to form micelles (Figure S9) independent on the amount of curcumin added, although curcumin will be located in the shell as well. In this scenario,

Figure 6. Variation of core radius and the thickness of inner shell, PMMA shell, and outer shell with drug loading. The lines are only to guide the eye.

The SLD of the PMMA shell 3.55 × 10−6 Å−2 remains the same, close to the initial value for amorphous PMMA, for the three samples. With an increasing amount of curcumin, the SLD of the outer fructose shell decreases from 5.20 × 10−6 to 4.80 × 10−6 Å−2; by contrast, the SLD of the inner shell increases slightly (Table S10). While the SLD should allow us to calculate the chemical composition of the shells (see eq 2), we measured only at one contrast point (100% D2O), and because of the effects of curvature of a thickening PMMA layer on particles which have similar radius (Figure 6), we concentrated on the outer shell where the curvature remains almost constant. An additional assumption is that the density of poly(1-O-MAFru) chains remains constant in the outer shell. While clearly the outer shell is composed of fructose, D2O, and curcumin, from the perspective of SANS, the contribution of curcumin to the overall SLD is negligible because even when it partitions completely in the inner and outer shells, its perturbation to the SLD is very small (Table S11). 5232


Article


methacryloyl-2,3:4,5-di-O-isopropylidene-β-D-fructopyranose (4 g, 12.2 mmol), AIBN (2.6 mg, 1.62 × 10−2 mmol), and CPADB (35 mg, 0.12 mmol) were dissolved in 1,4-dioxane (8.5 mL). Then, the tube was degassed by three freeze−pump−thaw cycles. The polymerization was carried out at 70 °C and stopped at 7 h by cooling the solution in ice water (conversion: 62%). The polymer solution was poured into a large excess of diethyl ether for precipitation. The viscous polymer was dried under vacuum for 24 h. The macro-RAFT agent, poly(1-O-MAFru)36, was subsequently used for chain extension with MMA. The procedure for the synthesis of poly(1-O-MAFru)36-b-PMMA192 was described as follows: macroRAFT agent (200 mg, 2 × 10−2 mmol), MMA (516 mg, 5.16 mmol), and AIBN (0.64 mg, 4 × 10−3 mmol) were dissolved in 1,4-dioxane (3 mL). The tube was degassed by three freeze−pump−thaw cycles. The polymerization was carried out at 70 °C and stopped after 18 h by cooling the solution in ice water. The polymer solution was poured into a large excess of n-hexane for precipitation. The viscous polymer was dried under vacuum for 24 h. The deprotection of the block copolymers was carried out under acidic conditions. Specifically, the polymer (80 mg) was added into 1.59 mL of TFA/H2O (9/1 v/v) in a vial with stirring at room temperature for 30 min. After reaction, the polymer solution was dialyzed against deionized water for 2 days (MWCO 3500). Self-Assembly of Curcumin Encapsulated Nanoparticles. A typical procedure for the encapsulation of curcumin was described as follows: The block copolymer (4 mg) and curcumin (1 mg) were dissolved in THF (0.2 mL). Then, 1.8 mL MQ water was added to the polymer solution using a syringe pump with a rate of 0.2 mL/h. The micelle solution was dialyzed against deionized water to remove THF and excess curcumin. The obtained micellar solutions were used for structural investigations by DLS, TEM, SAXS, and SANS. According to the literature,29 the stirring rate of the solution during the self-assembly process can change the morphologies of NPs; thus, polymersomes composed of poly(1-O-MAFru)36-b-PMMA192 and different concentrations of curcumin (0, 0.1725, and 0.2025 mg/mL) were prepared at a stirring rate of 380 cycles/min) in D2O for SANS measurements. The similarity in size was confirmed by DLS. Drug Release Study. The release study was carried out at 37 °C in the pH 7.4 phosphate buffer (20 mM). Two polymeric nanoparticles (2 mg/mL) carrying different concentrations of curcumin 0.115 and 0.135 mg/mL were transferred into two different 10 mL dialysis bags (MWCO 3500 kDa) in 1 L MQ water at 37 °C. The polymer solution inside the bag was sampled every 2 h. The drug concentration in the nanoparticle was determined using the UV−Vis absorbance at 421 nm. After all curcumin was released, the empty nanoparticles were collected and analyzed using TEM (JEOL1400). Cellular Uptake. Cellular uptake was carried out using flow cytometry. After culturing sufficient amount of cells using Dulbecco’s modified Eagle’s medium (DMEM), including fetal bovine serum (FBS, Bovogen Biologicals) and plasmocin at 37 °C in a humidified atmosphere including 5% CO2, MCF-7 human breast cancer cells RAW 264.7 were seeded in the 6-well plates at a density of 3 × 105 cells per well with DMEM cell culture medium at 37 °C with 5% CO2 for 2 days prior to nanoparticle treatment. For the treatment, micelles (0.6 mg in 3 mL DMEM) were incubated with cells for 1 day. The cells were initially washed 4 times with cold PBS to remove excess nanoparticles before they were detached from the plates using tryspin/EDTA. The cell suspensions were then centrifuged and resuspended in cold Hank’s buffer. The fluorescence intensity of the cells from the suspensions (20 000 cell events) was measured indicative of cellular uptake on a BD FACSCanto TM II Analyzer using an excitation laser wavelength of 640 nm and a band-pass filter of 660/20 nm for emission spectra. Raw data was analyzed using FlowJo software, and the results are shown using median fluorescence intensity (MFI) averaged from 3 individual wells for each sample. Cytotoxicity Test. The cytotoxic effects of free curcumin as well as blank and curcumin-loaded nanoparticles (NPs) were examined on MCF-7 cell lines RAW 264.7 by a standard sulforhodamine B colorimetric proliferation assay (SRB assay). After a sufficient amount of cells were cultured using medium Dulbecco’s modified Eagle’s

the influence of the surface curvature by the drug is less dramatic. We also observed similar results using a related polymer and Paclitaxel as the drug, a suggestion that drugs in general can potentially affect the activity of nanoparticles. This observation will be discussed in a forthcoming paper.

■

CONCLUSION We showed here that the morphologies of poly(1-OMAFru)36-b-PMMA192 self-assembled nanoparticles are affected by the concentration of curcumin, resulting in a transition from cylindrical micelles to polymersomes. Coincidently, the cellular uptake decreased with increasing amount of curcumin drug, suggesting that the drug introduced changes in the surface properties that may have been recognized by cancer cells. Increasing amount of curcumin lead to not only lower uptake but also lower toxicity. We, therefore, used spectroscopic technique and SANS/SAXS studies to learn more about the polymersomes. We found that the drug is indeed located in the shell, which helped to explain the observed morphology changes. More importantly, the location of the drug led to the dehydration of the shell, which may have been the cause for the cellular uptake. In this study, we showed that the interaction between drug and polymer can influence the self-assembly process in unexpected ways. Although it is commonly assumed that hydrophobic drugs will reside in the hydrophobic parts of selfassembled block copolymer aggregate, we showed here that the drug can (a) change the morphology of the nanoparticle and (b) enrich itself in the shell of the self-assembled morphology. The consequence of this behavior is change in not only morphology but also physicochemical properties of the shell, which is now filled with drugs. We studied the encapsulation of curcumin as the drug and a block copolymer based on glycopolymers, but we believe that similar observation can be found when drug and polymer form strong interactions that can overcome disparate polarities. Our observation may well be specific to the glycopolymer system that we used, but the researcher needs to consider that even if the drug is located in the core as expected, the presence of the drug may cause alterations to the property. The take-home message of this manuscript is that the drug cannot be seen as an independent variable, but the drug will interact with the polymer in various ways, which will result in physio-chemical changes of the drug carrier. In some cases, this may lead to change in behavior in a biological environment.

■

EXPERIMENTAL SECTION

Materials. D-Fructose (99%, Aldrich), curcumin (>80%, Aldrich), dichloromethane (DCM; anhydrous, >99.8%, Aldrich), tetrahydrofuran (THF; 99%, Ajax Finechem), fluorescein O-methacrylate (97%, Aldrich) and N,N-dimethylformamide (DMF; 99%, Ajax Finechem) were used as received. 1,4-Dioxane (99%, Ajax Finechem) and pyridine (99%, Ajax Finechem) were purified by reduced-pressure distillation. Methyl methacrylate (MMA, >99%, Aldrich) was passed over basic aluminum oxide to remove the inhibitors. 2,2-Azobis(isobutyronitrile) (AIBN; 98%, Fluka) was recrystallized from methanol for purification. The RAFT agent 4-cyanopentanoic acid dithiobenzoate (CPADB) was synthesized according to a literature procedure.39 Synthesis and Procedures. Synthesis of Poly(1-O-MAFru)36 and Poly(1-O-MAFru)36-b-PMMA192. The synthesis of 1-O-methacryloyl-2,3:4,5-di-O-isopropylidene-β-D-fructopyranose followed the similar procedure as reported previously.40 The polymerization of the glycomonomer was described as follows: in a Schlenk tube, 1-O5233


Article

Chemistry of Materials The drug loading capacity (DLC) (%) was calculated by

medium (DMEM) with fetal bovine serum (FBS, Bovogen Biologicals) and plasmocin at 37 °C in a humidified atmosphere including 5% CO2, cells were seeded in 96-well cell culture plates at a density of 4000 cells per well for 24 h. All the sample solutions were sterilized by UV irradiation for 15 min in the biosafety cabinet, and then the medium was replaced with 100 μL of fresh 2× concentrated medium. Subsequently, the cells were treated with serially diluted (2× dilution) sample solutions: blank nanoparticles, nanoparticles with less curcumin concentration, and nanoparticles with high curcumin concentration, followed by incubation at 37 °C for 3 days. After that, the culture medium was eliminated, followed by the addition of 100 μL of 10% (w/v) trichloroacetic acid (TCA) to each well and incubation of the plates at 4 °C for 30 min. Then, the plates were washed five times with MQ water. Next, the cells were stained with 100 μL of 0.4% (w/v) SRB dissolved in 1% acetic acid, and then the plates were covered with aluminum foil and incubated for 20 min at room temperature. After removal of SRB solution, the plates were washed five times using 1% acetic acid, and the plates were air-dried for 2 h at room temperature. Finally, the dye (SRB) was solubilized using 200 μL of 10 mM Tris buffer. The absorbance of each well was read at 490 nm using a Bio-Rad BenchMark Microplate reader, and the data were analyzed using GraphPad Prism 6.0. In this experiment, nontreated cells were used as controls, and the cell viability was calculated using the following equation (where OD means optical density):

cell viability (%) =

DLC (wt %) =

× 100%

(5)

UV−Vis absorbance of the polymer samples regarding the different concentration of curcumin was used to describe the change of absorbance during the drug release. Fluorescence Spectroscopy. Fluorescence measurements were carried on a CARY Eclipse fluorescence spectrophotometer at room temperature. The excitation of measurements was 421 nm for the experiments. The fluorescent wavelength was collected at 430−750 nm, and the excitation slit was 10 and 20. Fluorescence Lifetime. Fluorescence lifetime was measured by Fluoromax-4 at room temperature. The excitation was 460 nm. All the data were analyzed in custom software written in Labview. The lifetimes were amplitude weighted average lifetimes, and a multiexponential decay law was used to determine the average lifetime (τ). A two-exponential decay is described by τ=

∝1τ12 + ∝2 τ22 ∝1τ1 + ∝2 τ2

(6)

where τ1 and τ2 are the lifetimes of the multiexponential decay model and τ is the average lifetime. α is the amplitude of the individual component. Small-Angle X-ray Scattering and Small-Angle Neutron Scattering. SAXS measurements were made at the Australian Synchrotron’s SAXS/WAXS beamline41 from samples in 1.5 mm quartz capillaries (Charles Supper, Natick, United States). Isotropic scattering patterns were collected on a Pilatus 1 M detector (Dectris, Baden-Daettwil, Switerland) using a wavelength of 1.033 Å (20 keV) and a sample to detector distance of 7.2 m. The measurement geometry was used to convert the counts per detector pixel into the radially averaged intensity versus the scattering vector on an absolute intensity scale after subtraction of background due to solvent filled capillary and normalizing to the scattering of water using the beamline software Scatterbrain. The scattering vector, q, is defined by

OD490, sample − OD490, blank × 100 OD490, control − OD490, blank (3)

Analysis. Size Exclusion Chromatography (SEC). The molecular weight and polydispersity of synthesized polymers were analyzed via SEC. A Shimadzu modular system comprising a SIL-10AD autoinjector, DGU-12A degasser, LC-10AT pump, CTO-10A column oven, and a RID-10A refractive index detector was used. A 5.0-μm bead-size guard column (50 × 7.8 mm) followed by four 300 × 7.8 mm linear columns (500, 103, 104, and 105 Å pore size, 5 μm particle size) were employed for analysis. N,N-Dimethylacetamide [DMAc; HPLC grade, 0.05% w/v 2,6-dibutyl-4-methylphenol (BHT) and 0.03% w/v LiBr] with a flow rate of 1 mL/min at 50 °C was used as mobile phase. Fifty microliters of polymer solution with a concentration of 2 mg/mL in DMAc was used for every injection. The calibration was performed using commercially available narrowpolydispersity PMMA standards (0.5−1000 kDa, Polymer Laboratories). Dynamic Light Scattering (DLS). The hydrodynamic diameter Dh was determined using a Malvern Zetaplus particle size analyzer (laser, angle = 173°) at a copolymer concentration of 1 mg/mL. Samples were prepared in deionized water and sonicated for 30 min prior to the measurements. The polydispersity index (PDI) as a fitting parameter in the cumulants analysis of autocorrelation function shows the DLS data’s quality. ζ-Potential was measured to indicate the surface charge of the nanoparticles. Transmission Electron Microscopy (TEM). The TEM micrographs were obtained using transmission electron microscope (JEOL1400) comprising a dispersive X-ray analyzer and a Gatan CCD, facilitating the acquisition of digital images. The measurement was conducted at an accelerating voltage of 80 kV. The samples were prepared by casting the copolymer micellar solution (1 mg/mL) onto a copper grid. The grids were dried by air and then negatively stained with uranyl acetate. UV−Vis Spectroscopy. UV−Visible spectra were measured between 250−700 nm on the Cary 100 UV−Vis spectrophotometer. The encapsulated curcumin was measured with Cary 100 UV−vis spectrophotometer. The maximum wavelength of curcumin is 421 nm, and the drug loading efficiency (DLE) (%) was calculated by DLE (wt %) =

the amount of curcumin in polymersomes the amount of polymer

q=

4π ij θ yz sinjj zz λ k2{

(7)

and θ/2 is the scattering angle. SANS measurements were made on the fixed wavelength reactorbased pinhole instrument Quokka (ANSTO, Lucas Heights, Australia) .42 For each sample, measurements were made at three different samples to detector distances: 1.3, 10, and 20 m, where the counting times were 30 min, 1 h, and 4 h, respectively. The first two configurations were measured using 5 Å neutrons and the final at 8 Å. In both cases, the Δλ/λ was 10%. The isotropic counts per pixel on the area detector were converted to a continuous scattering curve using radial averaged 2-dimensional data after first correcting for the detector sensitivity, subtracting the dark field counts and the background signal due to the solvent filled empty cell, and finally normalizing to the incident flux of neutrons on the sample.37 Small angle scattering curves were obtained at several different concentrations (SAXS: 1, 2, and 3 mg/mL; SANS: 1, 2, and 3 mg/ mL). When normalized to concentration of particles, both SANS and SAXS curves were independent of concentration. We can therefore assume that all scattering curves were obtained under dilute conditions and that the scattering curve is dominated by the shape and internal structure of the particles.43 The structure model for the NPs described was fitted separately to the reduced SAXS and SANS data by using the program SASview35 and specific macros written for IgorPro.37

the amount of curcumin in polymersomes × 100% the amount of curcumin in feed (4) 5234


Article


■

(10) Liu, J.; Xiao, Y.; Allen, C. Polymer−drug compatibility: A guide to the development of delivery systems for the anticancer agent, ellipticine. J. Pharm. Sci. 2004, 93, 132−143. (11) Kim, B. Y.; Rutka, J. T.; Chan, W. C. Nanomedicine. N. Engl. J. Med. 2010, 363, 2434−2443. (12) Kim, Y.; Liemmawal, E. D.; Pourgholami, M. H.; Morris, D. L.; Stenzel, M. H. Comparison of Shell-Cross-Linked Micelles with Soft and Glassy Cores as a Drug Delivery Vehicle for Albendazole: Is There a Difference in Performance? Macromolecules 2012, 45, 5451− 5462. (13) Lv, S.; Wu, Y.; Cai, K.; He, H.; Li, Y.; Lan, M.; Chen, X.; Cheng, J.; Yin, L. High Drug Loading and Sub-Quantitative Loading Efficiency of Polymeric Micelles Driven by Donor−Receptor Coordination Interactions. J. Am. Chem. Soc. 2018, 140, 1235−1238. (14) Kopeček, J. Polymer−drug conjugates: Origins, progress to date and future directions. Adv. Drug Delivery Rev. 2013, 65, 49−59. (15) Delplace, V.; Couvreur, P.; Nicolas, J. Recent trends in the design of anticancer polymer prodrug nanocarriers. Polym. Chem. 2014, 5, 1529−1544. (16) Chu, K. S.; Schorzman, A. N.; Finniss, M. C.; Bowerman, C. J.; Peng, L.; Luft, J. C.; Madden, A. J.; Wang, A. Z.; Zamboni, W. C.; DeSimone, J. M. Nanoparticle drug loading as a design parameter to improve docetaxel pharmacokinetics and efficacy. Biomaterials 2013, 34, 8424−8429. (17) Zhou, X.; Chang, C.; Zhou, Y.; Sun, L.; Xiang, H.; Zhao, S.; Ma, L.; Zheng, G.; Liu, M.; Wei, H. A comparison study to investigate the effect of the drug-loading site on its delivery efficacy using double hydrophilic block copolymer-based prodrugs. J. Mater. Chem. B 2017, 5, 4443−4454. (18) Zhang, Q.; Vakili, M. R.; Li, X. F.; Lavasanifar, A.; Le, X. C. Terpolymer Micelles for the Delivery of Arsenic to Breast Cancer Cells: The Effect of Chain Sequence on Polymeric Micellar Characteristics and Cancer Cell Uptake. Mol. Pharmaceutics 2016, 13, 4021−4033. (19) Callari, M.; De Souza, P. L.; Rawal, A.; Stenzel, M. H. The Effect of Drug Loading on Micelle Properties: Solid-State NMR as a Tool to Gain Structural Insight. Angew. Chem., Int. Ed. 2017, 56, 8441−8445. (20) Press, A. T.; Ramoji, A.; vd Lühe, M.; Rinkenauer, A. C.; Hoff, J.; Butans, M.; Rössel, C.; Pietsch, C.; Neugebauer, U.; Schacher, F. H.; Bauer, M. Cargo−carrier interactions significantly contribute to micellar conformation and biodistribution. NPG Asia Mater. 2017, 9, e444. (21) Li, Z.; Lenk, T. I.; Yao, L. J.; Bates, F. S.; Lodge, T. P. Maintaining Hydrophobic Drug Supersaturation in a Micelle Corona Reservoir. Macromolecules 2018, 51, 540−551. (22) Li, Z.; Johnson, L. M.; Ricarte, R. G.; Yao, L. J.; Hillmyer, M. A.; Bates, F. S.; Lodge, T. P. Enhanced Performance of Blended Polymer Excipients in Delivering a Hydrophobic Drug through the Synergistic Action of Micelles and HPMCAS. Langmuir 2017, 33, 2837−2848. (23) Chang, T.; Trench, D.; Putnam, J.; Stenzel, M. H.; Lord, M. S. Curcumin-Loading-Dependent Stability of PEGMEMA-Based Micelles Affects Endocytosis and Exocytosis in Colon Carcinoma Cells. Mol. Pharmaceutics 2016, 13, 924−932. (24) Jurenka, J. S. Anti-inflammatory properties of curcumin, a major constituent of Curcuma longa: a review of preclinical and clinical research. Altern Med. Rev. 2009, 14, 141−153. (25) Maheshwari, R. K.; Singh, A. K.; Gaddipati, J.; Srimal, R. C. Multiple biological activities of curcumin: a short review. Life Sci. 2006, 78, 2081−2087. (26) Dutta, A. K.; Ikiki, E. Novel Drug Delivery Systems to Improve Bioavailability of Curcumin. J. Bioequivalence Bioavailability 2014, 06, 1−9. (27) Geng, Y.; Dalhaimer, P.; Cai, S.; Tsai, R.; Tewari, M.; Minko, T.; Discher, D. E. Shape Effects of Filaments Versus Spherical Particles in Flow and Drug Delivery. Nat. Nanotechnol. 2007, 2, 249− 255.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.chemmater.8b01882. Curcumin release, TEM after curcumin release, dielectric constants, UV−Vis spectra, lifetime plots, and SAXS and SANS fitting parameters (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Martina H. Stenzel: 0000-0002-6433-4419 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS M.H.S. and C.J.G. thank the Australian Research Council ARC for funding DP160101172. We also would like to thank the Electron Microscopy Unit of the UNSW Mark Wainwright Analytical Centre for their help. We acknowledge the support of the Australian Centre for Neutron Scattering, Australian NuclearScience and Technology Organization, in providing the neutron research used in this work. This research was also undertaken on the (SAXS/WAXS) beamline at the Australian Synchrotron, part of ANSTO.

■

REFERENCES

(1) Kataoka, K.; Harada, A.; Nagasaki, Y. Block copolymer micelles for drug delivery: design, characterization and biological significance. Adv. Drug Delivery Rev. 2001, 47, 113−131. (2) Miyata, K.; Nishiyama, N.; Kataoka, K. Rational design of smart supramolecular assemblies for gene delivery: chemical challenges in the creation of artificial viruses. Chem. Soc. Rev. 2012, 41, 2562−2574. (3) Cabral, H.; Matsumoto, Y.; Mizuno, K.; Chen, Q.; Murakami, M.; Kimura, M.; Terada, Y.; Kano, M. R.; Miyazono, K.; Uesaka, M.; Nishiyama, N.; Kataoka, K. Accumulation of sub-100 nm polymeric micelles in poorly permeable tumours depends on size. Nat. Nanotechnol. 2011, 6, 815−823. (4) Wang, J.; Mao, W.; Lock, L. L.; Tang, J.; Sui, M.; Sun, W.; Cui, H.; Xu, D.; Shen, Y. The Role of Micelle Size in Tumor Accumulation, Penetration, and Treatment. ACS Nano 2015, 9, 7195−7206. (5) Win, K. Y.; Feng, S. S. Effects of particle size and surface coating on cellular uptake of polymeric nanoparticles for oral delivery of anticancer drugs. Biomaterials 2005, 26, 2713−2722. (6) Skrabania, K.; von Berlepsch, H.; Bottcher, C.; Laschewsky, A. Synthesis of Ternary, Hydrophilic-Lipophilic-Fluorophilic Block Copolymers by Consecutive RAFT Polymerizations and Their SelfAssembly into Multicompartment Micelles. Macromolecules 2010, 43, 271−281. (7) Talelli, M.; Barz, M.; Rijcken, C. J.; Kiessling, F.; Hennink, W. E.; Lammers, T. Core-Crosslinked Polymeric Micelles: Principles, Preparation, Biomedical Applications and Clinical Translation. Nano Today 2015, 10, 93−117. (8) Wilhelm, S.; Tavares, A. J.; Dai, Q.; Ohta, S.; Audet, J.; Dvorak, H. F.; Chan, W. C. W. Analysis of nanoparticle delivery to tumours. Nat. Rev. Mater. 2016, 1, 1−12. (9) Gou, J.; Feng, S.; Xu, H.; Fang, G.; Chao, Y.; Zhang, Y.; Xu, H.; Tang, X. Decreased Core Crystallinity Facilitated Drug Loading in Polymeric Micelles without Affecting Their Biological Performances. Biomacromolecules 2015, 16, 2920−2929. 5235


Article

Chemistry of Materials (28) Zhao, J. C.; Lu, H. X.; Wong, S.; Lu, M. X.; Xiao, P.; Stenzel, M. H. Influence of nanoparticle shapes on cellular uptake of paclitaxel loaded nanoparticles in 2D and 3D cancer models. Polym. Chem. 2017, 8, 3317−3326. (29) Zhao, J.; Lu, H.; Xiao, P.; Stenzel, M. H. Cellular Uptake and Movement in 2D and 3D Multicellular Breast Cancer Models of Fructose-Based Cylindrical Micelles That Is Dependent on the Rod Length. ACS Appl. Mater. Interfaces 2016, 8, 16622−30. (30) Chang, T.; Lord, M. S.; Bergmann, B.; Macmillan, A.; Stenzel, M. H. Size effects of self-assembled block copolymer spherical micelles and vesicles on cellular uptake in human colon carcinoma cells. J. Mater. Chem. B 2014, 2, 2883−2891. (31) Grieser, F.; Drummond, C. J. The physicochemical properties of self-assembled surfactant aggregates as determined by some molecular spectroscopic probe techniques. J. Phys. Chem. 1988, 92, 5580−5593. (32) Patra, D.; Barakat, C. Synchronous fluorescence spectroscopic study of solvatochromic curcumin dye. Spectrochim. Acta, Part A 2011, 79, 1034−1041. (33) Priyadarsini, K. I. Photophysics, photochemistry and photobiology of curcumin: Studies from organic solutions, bio-mimetics and living cells. J. Photochem. Photobiol., C 2009, 10, 81−95. (34) Boreham, A.; Brodwolf, R.; Walker, K.; Haag, R.; Alexiev, U. Time-Resolved Fluorescence Spectroscopy and Fluorescence Lifetime Imaging Microscopy for Characterization of Dendritic Polymer Nanoparticles and Applications in Nanomedicine. Molecules 2017, 22, 17. (35) DANSE SasView for Small Angle Scattering Analysis. http:// www.sasview.org. (36) Kienzle, P. Scattering Length Density Calculator. http://www. ncnr.nist.gov/resources/sldcalc.html. (37) Kline, S. R. Reduction and analysis of SANS and USANS data using IGOR Pro. J. Appl. Crystallogr. 2006, 39, 895−900. (38) White, M. A. Physical Properties of Materials, 2nd ed.; CRC Press: 2011. (39) Mitsukami, Y.; Donovan, M. S.; Lowe, A. B.; McCormick, C. L. Water-Soluble Polymers. 81. Direct Synthesis of Hydrophilic StyrenicBased Homopolymers and Block Copolymers in Aqueous Solution via RAFT. Macromolecules 2001, 34, 2248−2256. (40) Zhao, J.; Babiuch, K.; Lu, H.; Dag, A.; Gottschaldt, M.; Stenzel, M. H. Fructose-coated nanoparticles: a promising drug nanocarrier for triple-negative breast cancer therapy. Chem. Commun. 2014, 50, 15928−15931. (41) Kirby, N. M.; Mudie, S. T.; Hawley, A. M.; Cookson, D. J.; Mertens, H. D. T.; Cowieson, N.; Samardzic-Boban, V. A lowbackground-intensity focusing small-angle X-ray scattering undulator beamline. J. Appl. Crystallogr. 2013, 46, 1670−1680. (42) Gilbert, E. P.; Schulz, J. C.; Noakes, T. J. ’Quokka’ - the smallangle neutron scattering instrument at OPAL. Phys. B 2006, 385−386, 1180−1182. (43) Svergun, D. I.; Feĭgin, L. A.; Taylor, G. W. Structure Analysis by Small-Angle X-ray and Neutron Scattering; Plenum Press: New York, London, 1987.

5236


Drug-Induced Morphology Transition of Self-Assembled

Recommend Documents