Cinnamyl-Modified Polyglycidol/Poly(ε-Caprolactone) Block Copolymer Nanocarriers for Enhanced Encapsulation and Prolonged Release of Cannabidiol

The present study describes the development of novel block copolymer nanocarriers of the phytocannabinoid cannabidiol (CBD), designed to enhance the solubility of the drug in water while achieving high encapsulation efficiency and prolonged drug release. Firstly, a well-defined amphiphilic block copolymer consisting of two outer hydrophilic polyglycidol (PG) blocks and a middle hydrophobic block of poly(ε-caprolactone) bearing pendant cinnamyl moieties (P(CyCL-co-CL)) were synthesized by the click coupling reaction of PG-monoalkyne and P(CyCL-co-CL)-diazide functional macroreagents. A non-modified polyglycidol/poly(ε-caprolactone) amphiphilic block copolymer was obtained as a referent system. Micellar carriers based on the two block copolymers were formed via the solvent evaporation method and loaded with CBD following two different protocols—loading during micelle formation and loading into preformed micelles. The key parameters/characteristics of blank and CBD-loaded micelles such as size, size distribution, zeta potential, molar mass, critical micelle concentration, morphology, and encapsulation efficiency were determined by using dynamic and static multiangle and electrophoretic light scattering, transmission electron microscopy, and atomic force microscopy. Embedding CBD into the micellar carriers affected their hydrodynamic radii to some extent, while the spherical morphology of particles was not changed. The nanoformulation based on the copolymer bearing cinnamyl moieties possessed significantly higher encapsulation efficiency and a slower rate of drug release than the non-modified copolymer. The comparative assessment of the antiproliferative effect of micellar CBD vs. the free drug against the acute myeloid leukemia-derived HL-60 cell line and Sezary Syndrome HUT-78 demonstrated that the newly developed systems have pronounced antitumor activity.


Introduction
Nanoscopic polymer-based drug-delivery systems are rapidly entering the pharmaceutical field since they are beneficial for treating severe human diseases. Such systems enable the introduction of a variety of therapeutic substances into the body and improve their efficacy and safety profiles [1,2]. Despite their vast potential, the key characteristics of the polymer-based nanoformulations still need improvement to achieve superior colloidal stability, protection from enzyme actions, and diminished particle interactions with plasma proteins, which could guarantee a prolonged circulation in the bloodstream, favorable pharmacokinetic profile, and enhanced bioavailability [3,4]. (10 mL) and dialyzed against a methanol/water mixture (10:1 v/v, membrane, MWCO 8 kDa) for 72 h. The methanol was removed using a rotary vacuum evaporator, and the copolymer was recovered by freeze-drying. Yield: 0.258 g (80%); M n = 15,000 g·mol −1 , M w /M n = 1.3.

Preparation of Polymeric Micelles
A copolymer solution (5 mg of copolymer in 5 mL of methanol) was added dropwise to purified water (5 mL) at room temperature under vigorous stirring (1000 rpm). After 30 min, the organic solvent was evaporated under reduced pressure at 37 • C to obtain a slightly opalescent, colorless micellar dispersion with a concentration of 1 mg·mL −1 .

Loading of CBD
The micelles of the novel (PG 50  Protocol A: The selected copolymer (5 mg) and CBD (0.5 mg) were dissolved in methanol (5 mL) at a copolymer/CBD-mass ratio of 10:1. After that, the organic solution was added dropwise to purified water (5 mL) at room temperature under vigorous stirring (1000 rpm). The resulting mixture was stirred additionally for 30 min at the same temperature. Finally, the organic solvent was evaporated by a rotary vacuum evaporator at 37 • C, yielding a slightly opalescent, colorless aqueous micellar dispersion. The dispersion was filtered (Nylon, 0.22 µm) and the filter was rinsed with ethanol. The collected filter fraction was quantified by spectrophotometric measurements (λ = 274 nm) to determine the mass of free, unentrapped CBD.
Protocol B: The selected copolymer (5 mg) was dissolved in methanol (5 mL). After that, the organic solution was added dropwise to purified water (5 mL) at room temperature under vigorous stirring (1000 rpm). The resulting mixture was stirred additionally for 30 min at the same temperature. Next, the organic solvent was evaporated by a rotary vacuum evaporator at 37 • C. The CBD (0.5 mg) (copolymer/CBD-mass ratio 10:1) dissolved in 40 µL methanol was added to the as-prepared micellar dispersion. The resulting mixture was stirred additionally for 30 min at the same temperature. Finally, the traces of organic solvent were evaporated under argon, yielding a slightly opalescent, colorless aqueous micellar dispersion. The dispersion was filtered (Nylon, 0.22 µm) and the filter was rinsed with ethanol. The collected filter fraction was quantified by spectrophotometric measurements (λ = 274 nm) to determine the mass of free, unentrapped CBD. Protocol A: The selected copolymer (5 mg) and CBD (0.5 mg) were dissolved in methanol (5 mL) at a copolymer/CBD-mass ratio of 10:1. After that, the organic solution was added dropwise to purified water (5 mL) at room temperature under vigorous stirring (1000 rpm). The resulting mixture was stirred additionally for 30 min at the same temperature. Finally, the organic solvent was evaporated by a rotary vacuum evaporator at 37 °C, yielding a slightly opalescent, colorless aqueous micellar dispersion. The dispersion was filtered (Nylon, 0.22 µm) and the filter was rinsed with ethanol. The collected filter fraction was quantified by spectrophotometric measurements (λ = 274 nm) to determine the mass of free, unentrapped CBD.
Protocol B: The selected copolymer (5 mg) was dissolved in methanol (5 mL). After that, the organic solution was added dropwise to purified water (5 mL) at room temperature under vigorous stirring (1000 rpm). The resulting mixture was stirred additionally for 30 min at the same temperature. Next, the organic solvent was evaporated by a rotary vacuum evaporator at 37 °C. The CBD (0.5 mg) (copolymer/CBD-mass ratio 10:1) dissolved in 40 µL methanol was added to the as-prepared micellar dispersion. The resulting mixture was stirred additionally for 30 min at the same temperature. Finally, the traces of organic solvent were evaporated under argon, yielding a slightly opalescent, colorless aqueous micellar dispersion. The dispersion was filtered (Nylon, 0.22 µm) and the filter was rinsed with ethanol. The collected filter fraction was quantified by spectrophotometric measurements (λ = 274 nm) to determine the mass of free, unentrapped CBD.

Proton Nuclear Magnetic Resonance ( 1 H-NMR)
Scheme 1. Preparation of CBD-loaded polymeric micelles according to protocol A (a) and protocol B (b).

Proton Nuclear Magnetic Resonance ( 1 H-NMR)
1 H-NMR measurements were conducted on a Bruker Avance II spectrometer operating at 600 MHz using CDCl 3 , DMSO-d 6, or CD 3 OD at 25 • C.

Size Exclusion Chromatography (SEC)
Analyses were performed on Shimadzu Nexera HPLC chromatograph, equipped with a degasser, a pump, an auto-sampler, a RI detector, and three columns: 10 µm PL gel mixed-B, 5 µm PL gel 500 Å, and 50 Å. THF was used as the eluent at a flow rate of 1.0 mL·min −1 and a temperature of 40 • C. The sample concentration was 1 mg·mL −1 and GPC was calibrated with polystyrene standards.

Transmission Electron Microscopy (TEM)
Micrographs were obtained using an HRTEM JEOL JEM-2100 transmission electron microscope operating at 200 kV. The samples were prepared by depositing a drop of the dispersions onto a carbon grid and the subsequent evaporation of the solvent under a vacuum. To visualize the hydrophilic corona with high contrast, a uranyl acetate staining protocol prior to sample preparation was utilized [56].

Atomic Force Microscopy (AFM)
The images were obtained using a Bruker Dimension Icon Instrument operating at a 1.00 Hz scan rate under ambient conditions. Moreover, 2 µL of the copolymer dispersions was placed onto a freshly cleaned glass substrate (1 cm 2 ) and spin-casted at 2000 rpm for a minute. AFM measurements were performed in ScanAsyst (Peak Force Tapping) mode.

Dynamic and Electrophoretic Light Scattering
The preliminary assessment of the particle size and size distribution was carried out by dynamic light scattering (DLS) measurements on a NanoBrook 90 Plus PALS instrument (Brookhaven Instruments Corporation), equipped with a 35 mW red diode laser (λ = 660 nm) at a scattering angle of 90 • . The measurements were taken at 25 • C and 37 • C applying a dust cut-off of 30 micellar dispersions at concentrations of 1 mg·mL −1 and fixed volumes of 1.7 mL.
The apparent hydrodynamic radii (R h 90 ) were determined according to the Stokes-Einstein equation: where k is the Boltzmann constant, η is the solvent viscosity at temperature T in Kelvin and D 90 is the diffusion coefficient measured at an angle of 90 • . Each measurement was performed in triplicate. The electrophoretic light scattering measurements were carried out on the same instrument at a scattering angle of 15 • and 25 • C and 37 • C. The principle of phase analysis light scattering (PALS) was applied for the measurements of electrophoretic mobility. The ζ potentials were calculated using the Smoluchowski equation: where η is the solvent viscosity, υ is the electrophoretic mobility, and ε is the dielectric constant of the solvent.

Multiangle Dynamic and Static Light Scattering
Dynamic light scattering measurements were performed on a Brookhaven BI-200 goniometer with vertically polarized incident light at a wavelength λ = 633 nm supplied by a He-Ne laser operating at 35 mW and equipped with a Brookhaven BI-9000 AT digital autocorrelator. The scattered light was measured for dilute aqueous dispersions of the empty and CBD-loaded micellar dispersions in the concentration range 0.417-1.0 mg·mL −1 at 25 • C. Measurements were made at angles θ in the range of 50-130 • . The autocorrelation functions were analyzed using the constrained regularized algorithm CONTIN [57] to obtain the distributions of the relaxation rates (Γ). The latter provided distributions of the apparent diffusion coefficient (D = Γ/q 2 ), where q is the magnitude of the scattering vector given by q = (4πn/λ)sin(θ/2), n is the refractive index of the medium. The mean hydrodynamic radius was obtained by the Stokes-Einstein equation: where k is the Boltzmann constant, η is the solvent viscosity at temperature T in Kelvin, and D 0 is the diffusion coefficient at infinite dilution. Static light scattering (SLS) measurements were carried out in the interval of angles from 40 to 140 • at 25 • C using the same light scattering set. The SLS data were analyzed using the Zimm plot software provided by Brookhaven Instruments. Information on the weight-average molar mass, M w , the radius of gyration, R g , and the second virial coefficient, A 2 , was obtained from the dependence of the quantity (Kc/R θ ) on the concentration (c) and scattering angle (θ). Here, K is the optical constant given by K = 4π 2 n 0 2 (dn/dc) 2 /N A λ 4 , where n 0 is the refractive index of the solvent, N A is Avogadro's constant, λ is the laser wavelength, and R θ is the Rayleigh ratio at angle θ. dn/dc is the refractive index increment measured in water in separate experiments on an Orange GPC19 DNDC refractometer. The dn/dc values of the investigated systems were in the 0.131-0.133 g·mL −1 range. The DLS and SLS measurements were performed at 25 and 37 • C.

Spectrophotometric Determination of the Critical Micelle Concentration (CMC)
Furthermore, 20 µL of a 0.4 mM solution of 1,6-diphenyl-1,3,5-hexatriene (DPH) in methanol was added to 2 mL micellar dispersions with increasing concentrations in the range 9.765 × 10 −4 -1.0 mg·mL −1 . The samples were incubated in the dark for 24 h at room temperature. UV-vis absorption spectra of DPH in the wavelength interval λ = 300-500 nm at room temperature were recorded on a Beckman Coulter DU 800 UV-vis spectrometer. The intensities of the absorption peak at 356 nm were plotted against the polymer concentration. The CMC value was determined as the break in the absorbance intensity versus the concentration curve.

Drug-Release Study
The release of CBD was evaluated as a function of time using the dialysis method [58]. In short, 2 mL of tested formulations was inserted into a dialysis bag (MWCO 12,000-14,000, Sigma-Aldrich, Steinheim, Germany) and placed in a 50 mL dissolution medium of phosphate-buffered saline pH 7.4 and 10% ethanol to retain the solubility of the released CBD. The acceptor medium was in constant motion (200 rpm) and the circulating water jacket (Huber, Germany) maintained the temperature at 37 ± 0.5 • C during the study. At predetermined time points, 1 mL samples from the released medium were withdrawn and analyzed by UV-vis spectroscopy at 274 nm using a calibration curve with linearity in the concentration range of 0.0025 to 10 µg·mL −1 (correlation coefficient R 2 = 0.995). The aliquots were replaced with equal volumes of fresh medium. The drug-release studies were carried out threefold.

Cytotoxicity Assessment
The cytotoxicity of the copolymers and the micelles prepared thereof was evaluated by standard MTT-dye reduction assay. The assay is based on the aptitude of the mitochondrial succinate dehydrogenase of viable cells to metabolize the yellow tetrazolium MTT dye to a violet formazan. The experiment was performed as described by Mosman [59] with small modifications [60]. Cells in the exponential phase were seeded in 96-well microplates (100 µL/well) at a density of 1 × 10 5 cells·mL −1 and incubated at 37 • C for 24 h. Afterward, they were exposed to various concentrations of empty micelles or free or micellar CBD for a period of 72 h. For each test group, a set of at least eight wells were used. After the exposure time, to each well aliquots of 10 µL of MTT solution (10 mg·mL −1 in PBS) was added. Next, the microplates were incubated for 4 h at 37 • C and the MTT-formazan crystals formed were dissolved by the addition of 100 µL/well of 5% formic acid solution in 2-propanol. The MTT-formazan absorption was recorded using a Beckman-Coulter DTX800 multimode microplate reader at 580 nm. Thereafter the cell survival fractions were calculated as a percentage of the untreated control. In addition, IC 50 values were derived from the concentration-response curves.

Statistical Analysis
The data are presented as the mean standard deviation (SD) of three independent experiments. The correlation coefficients for the linear sections of the curves were in the 0.992-0.999 range.

Copolymer Synthesis and Characterization
The synthetic approach to preparing the novel copolymer is based on a click coupling reaction of appropriately functionalized polymer intermediates. The reaction steps are presented in Scheme 2. Firstly, a PCL-based polymer bearing pendant cinnamyl groups was prepared by a procedure described elsewhere [53] and functionalized with terminal azide groups. The reaction scheme, molar mass characteristics, and composition of this polymer precursor are presented in the ESI. Separately, a monoalkyne functionalized PEEGE was obtained by the ring-opening polymerization of EEGE. Upon completion of the polymerization, a short spacer of four oxypropylene units was introduced to facilitate the subsequent modification of the as-prepared monohydroxy prepolymer with a clickable alkyne end group via esterification with 4-pentynoic acid [48]. A detailed reaction scheme as well as SEC curves and characterization data ( 1 H NMR spectra) are presented in the ESI.
The synthetic approach to preparing the novel copolymer is based on a click coupling reaction of appropriately functionalized polymer intermediates. The reaction steps are presented in Scheme 2. Firstly, a PCL-based polymer bearing pendant cinnamyl groups was prepared by a procedure described elsewhere [53] and functionalized with terminal azide groups. The reaction scheme, molar mass characteristics, and composition of this polymer precursor are presented in the ESI. Separately, a monoalkyne functionalized PEEGE was obtained by the ring-opening polymerization of EEGE. Upon completion of the polymerization, a short spacer of four oxypropylene units was introduced to facilitate the subsequent modification of the as-prepared monohydroxy prepolymer with a clickable alkyne end group via esterification with 4-pentynoic acid [48]. A detailed reaction scheme as well as SEC curves and characterization data ( 1 H NMR spectra) are presented in the ESI. The azide-alkyne click reaction was carried out in THF at 30 °C for 24 h using a CuBr/PMDETA catalytic complex (Scheme 2). SEC and 1 H-NMR analyses were used to characterize the product (Figures 1 and S3). SEC showed a clear shift to lower retention times (corresponding to higher molar mass) of the product in comparison to the PEEGEand PCL-based prepolymers, while maintaining a monomodal and narrow (Mw/Mn = 1.3) molar mass distribution ( Figure 1). Furthermore, in the 1 H-NMR spectrum ( Figure S3a), all proton signals characteristic for the two precursors were evident, whereas the signals assigned to the alkyne protons at 2.01 ppm disappeared and a new signal for the methyne protons of the triazole ring at 8.1 ppm appeared. In the final step, the copolymer was Scheme 2.
Schematic representation of the synthesis of PEEGE 50 -b-PPO 4 -b-[P(CyCL) 4 -co-(CL) 40 ]-b-PPO 4 -b-PEEGE 50 copolymer by copper-catalyzed "click" coupling reaction and subsequent deprotection leading to the amphiphilic PG 50 The azide-alkyne click reaction was carried out in THF at 30 • C for 24 h using a CuBr/PMDETA catalytic complex (Scheme 2). SEC and 1 H-NMR analyses were used to characterize the product ( Figure 1 and Figure S3). SEC showed a clear shift to lower retention times (corresponding to higher molar mass) of the product in comparison to the PEEGE-and PCL-based prepolymers, while maintaining a monomodal and narrow (M w /M n = 1.3) molar mass distribution ( Figure 1). Furthermore, in the 1 H-NMR spectrum ( Figure S3a), all proton signals characteristic for the two precursors were evident, whereas the signals assigned to the alkyne protons at 2.01 ppm disappeared and a new signal for the methyne protons of the triazole ring at 8.1 ppm appeared. In the final step, the copolymer was treated with AlCl 3 .6H 2 O to remove the protective EEGE groups and to convert the flanking PEEGE blocks into blocks of linear polyglycidol (Scheme 2). The 1 H-NMR spectrum of the final product is presented in Figure S3b. The disappearance of the signal for the methyl protons of the protective EEGE groups and the appearance of the new signal assigned to OH groups of PG at 4.46 ppm proved the effective release. The results suggested the high efficiency of both the click reaction for coupling of the prepolymers and the modification reaction to release the protective groups. The composition and molar mass characteristics of the novel and referent copolymers are presented in Table 1 13,200 -a -number-averaged molar mass from 1 H NMR. b -weight-averaged molar mass from SEC; Mw/Mn-molar mass distribution from SEC. Intensity (uRIU) 40

Preparation of Blank and CBD-Loaded Polymeric Micelles
The critical micelle concentration of the modified copolymer was determined by dy solubilization. The method employs specific photophysical properties of the hydrophobi dye 1,6-diphenyl-1,3,5-hexatriene: its UV absorbance in water is minimal, whereas it in creases substantially in a hydrophobic environment so that the appearance of a character istic maximum at 356 nm and its sharp increase upon increasing copolymer concentration indicates an abrupt change in the properties of the system-the formation of hydrophobi domains (presumably cores of the micelles). DPH solubilization has frequently been em ployed for the determination of the CMC of conventional surfactants and amphiphili polymers [48,[61][62][63]. The CMC value was determined from the break of the absorbanc intensity vs. copolymer concentration curve as shown in Figure 2.  50 . THF was used as the eluent at a flow rate of 1.0 mL·min −1 , at a temperature of 40 • C.   45 11,000 -- 50 20,100 15,000 50 13,200 -a -number-averaged molar mass from 1 H NMR. b -weight-averaged molar mass from SEC; M w /M n -molar mass distribution from SEC.

Preparation of Blank and CBD-Loaded Polymeric Micelles
The critical micelle concentration of the modified copolymer was determined by dye solubilization. The method employs specific photophysical properties of the hydrophobic dye 1,6-diphenyl-1,3,5-hexatriene: its UV absorbance in water is minimal, whereas it increases substantially in a hydrophobic environment so that the appearance of a characteristic maximum at 356 nm and its sharp increase upon increasing copolymer concentration indicates an abrupt change in the properties of the system-the formation of hydrophobic domains (presumably cores of the micelles). DPH solubilization has frequently been employed for the determination of the CMC of conventional surfactants and amphiphilic polymers [48,[61][62][63]. The CMC value was determined from the break of the absorbance intensity vs. copolymer concentration curve as shown in Figure 2 The resulting value (0.12 mg·mL −1 ) was slightly larger than that of the reference copolymer PG45-b-PCL35-b-PG45 (0.10 mg·mL −1 ) [48] and probably reflected the effects of the longer hydrophilic polyglycidol blocks.
At concentrations ca. one order of magnitude higher than the CMC, the copolymer micelles were loaded with CBD by applying different protocols, depicted in Scheme 1loading during micelle formation (protocol A) and loading in preformed micelles (protocol B). Based on previous studies [51,53] and preliminary concentration-dependent measurements, the drug/copolymer-mass ratio of 1:10 at a copolymer concentration of 1 mg·mL −1 was determined to be the most effective and optimal, ensuring the highest encapsulation efficiency and drug loading. Preliminary screening of the size and size distributions, performed by DLS at a single angle (90°), revealed monomodal size distributions and apparent hydrodynamic radii, Rh 90 , in the 50-60 nm range ( Table 2). All particles exhibited low ζ potential (very slightly positive or negative, practically neutral), which is in line with the non-ionic nature of the two copolymers and the drug ( Table 2). The simultaneous formation and loading of the micelles (protocol A) yielded higher encapsulation efficiency than protocol B, that is, the loading of the drug into preformed micelles ( Table  2). The aggregates of the novel copolymer invariably exhibited higher encapsulation efficiency than those of the referent copolymer (Table 2), which apparently reflected the introduction of cinnamyl-bearing units intended to increase the loading efficiency. Although exhibiting low ζ potential, the micelles of the two copolymers (blank or CBDloaded) are characterized by enhanced colloidal stability for at least three months. The enhanced colloidal stability is provided by the hydrophilic corona built of polyglycidol chains. Multiangle dynamic and static light scattering was performed to fully characterize the empty micelles of the novel copolymer and the micelles of the novel copolymer loaded The resulting value (0.12 mg·mL −1 ) was slightly larger than that of the reference copolymer PG 45 -b-PCL 35 -b-PG 45 (0.10 mg·mL −1 ) [48] and probably reflected the effects of the longer hydrophilic polyglycidol blocks.
At concentrations ca. one order of magnitude higher than the CMC, the copolymer micelles were loaded with CBD by applying different protocols, depicted in Scheme 1-loading during micelle formation (protocol A) and loading in preformed micelles (protocol B). Based on previous studies [51,53] and preliminary concentration-dependent measurements, the drug/copolymer-mass ratio of 1:10 at a copolymer concentration of 1 mg·mL −1 was determined to be the most effective and optimal, ensuring the highest encapsulation efficiency and drug loading. Preliminary screening of the size and size distributions, performed by DLS at a single angle (90 • ), revealed monomodal size distributions and apparent hydrodynamic radii, R h 90 , in the 50-60 nm range ( Table 2). All particles exhibited low ζ potential (very slightly positive or negative, practically neutral), which is in line with the non-ionic nature of the two copolymers and the drug ( Table 2). The simultaneous formation and loading of the micelles (protocol A) yielded higher encapsulation efficiency than protocol B, that is, the loading of the drug into preformed micelles ( Table 2). The aggregates of the novel copolymer invariably exhibited higher encapsulation efficiency than those of the referent copolymer (Table 2), which apparently reflected the introduction of cinnamyl-bearing units intended to increase the loading efficiency. Although exhibiting low ζ potential, the micelles of the two copolymers (blank or CBD-loaded) are characterized by enhanced colloidal stability for at least three months. The enhanced colloidal stability is provided by the hydrophilic corona built of polyglycidol chains. Multiangle dynamic and static light scattering was performed to fully characterize the empty micelles of the novel copolymer and the micelles of the novel copolymer loaded by protocol A. In the investigated concentration range (0.417-1.0 mg·mL −1 ), which is slightly above the CMC, the relaxation time distributions were predominantly monomodal, indicating the existence of only one population of particles. Additional modes of low amplitude (mainly fast modes of amplitude below 4%) were occasionally and unsystematically observed at certain angles. A representative time distribution is shown in Figure 3a. More relaxation time distributions and converted therefrom particle size distributions are presented in Figure S4 in the ESI. The diffusion coefficients were determined from the angular dependence of the relaxation rate (inversely proportional to relaxation time), measured at different scattering angles and then plotted against concentration to obtain the diffusion coefficient at zero concentration, D 0 , as shown in Figure 3b and c. D 0 was used to calculate the hydrodynamic radii, R h , of the empty and loaded micelles. The R h values of the empty and CBD-loaded micelles are summarized in Table 3.
Pharmaceutics 2023, 15, x FOR PEER REVIEW 12 of 23 by protocol A. In the investigated concentration range (0.417-1.0 mg·mL −1 ), which is slightly above the CMC, the relaxation time distributions were predominantly monomodal, indicating the existence of only one population of particles. Additional modes of low amplitude (mainly fast modes of amplitude below 4%) were occasionally and unsystematically observed at certain angles. A representative time distribution is shown in Figure 3a. More relaxation time distributions and converted therefrom particle size distributions are presented in Figure S4 in the ESI. The diffusion coefficients were determined from the angular dependence of the relaxation rate (inversely proportional to relaxation time), measured at different scattering angles and then plotted against concentration to obtain the diffusion coefficient at zero concentration, D0, as shown in Figure 3b and c. D0 was used to calculate the hydrodynamic radii, Rh, of the empty and loaded micelles. The Rh values of the empty and CBD-loaded micelles are summarized in Table 3.   a -density of the material within the particle. b -expressed as the number of CBD molecules loaded in one copolymer micelle.
Static light scattering was performed to determine the weight-average molar mass (M w ), radius of gyration (R g ), and second virial coefficient (A 2 ) of the empty and loaded micelles. The static parameters were evaluated by the Zimm plot method. Zimm diagrams of the empty and loaded micelles are presented in Figure 4, whereas the derived parameters are collected in Table 3.  a -density of the material within the particle. b -expressed as the number of CBD molecules loaded in one copolymer micelle.
Static light scattering was performed to determine the weight-average molar mass (Mw), radius of gyration (Rg), and second virial coefficient (A2) of the empty and loaded micelles. The static parameters were evaluated by the Zimm plot method. Zimm diagrams of the empty and loaded micelles are presented in Figure 4, whereas the derived parameters are collected in Table 3. Evident from the results in Table 3 is that the CBD-loaded micelles are slightly smaller in size than the empty micelles, whereas the molar masses are comparable within the standard deviation of the method. These findings implied the formation of more compact and dense particles and revealed the effect of the hydrophobic drug molecules as nucleation sites on which copolymer self-assembly occurred, bringing about the enhancement of particle density. A realistic assessment of the different compactness levels of the empty and loaded particles is given by the particle density, ρ (Table 3), calculated from the molar mass and hydrodynamic volume data while assuming the spherical morphology of the particles (see ESI for the calculation of the particle density). Apparently, the ρ value of the CBD-loaded micelles was higher than that of the empty micelles. In addition, Rg/Rh attained values close to unity, which was compatible with the structure of both the empty and loaded micelles-relatively dense and compact particles with "hairy" surfaces [64,65]. A2 values were very small in magnitude (of the order of 10 −6 mL.mol/g 2 ), which is in accordance with the high molar mass of the micelles, and negative. The negative values of A2 normally indicate unfavorable particle-solvent interactions. As the A2 values are very  Table 3 is that the CBD-loaded micelles are slightly smaller in size than the empty micelles, whereas the molar masses are comparable within the standard deviation of the method. These findings implied the formation of more compact and dense particles and revealed the effect of the hydrophobic drug molecules as nucleation sites on which copolymer self-assembly occurred, bringing about the enhancement of particle density. A realistic assessment of the different compactness levels of the empty and loaded particles is given by the particle density, ρ (Table 3), calculated from the molar mass and hydrodynamic volume data while assuming the spherical morphology of the particles (see ESI for the calculation of the particle density). Apparently, the ρ value of the CBDloaded micelles was higher than that of the empty micelles. In addition, R g /R h attained values close to unity, which was compatible with the structure of both the empty and loaded micelles-relatively dense and compact particles with "hairy" surfaces [64,65]. A 2 values were very small in magnitude (of the order of 10 −6 mL·mol/g 2 ), which is in accordance with the high molar mass of the micelles, and negative. The negative values of A 2 normally indicate unfavorable particle-solvent interactions. As the A 2 values are very small, we may speculate here that there are weak attractive interactions between the micelles, previously observed for self-assembled structures of polyglycidol-based copolymers [48,66,67].

Evident from the results in
Another parameter that can be extracted from the light scattering data is the loading capacity expressed as the number of CBD molecules loaded in one micelle of the novel and referent copolymer. The light scattering characterization data of the referent copolymer micelles loaded with CBD as well as the calculation of loading capacity are presented in Figure S5 and Table S1 in the ESI. Evident from the values of this quantity was that the micelles of the novel copolymer bore about 32% more CBD molecules than those of the referent copolymer, which undoubtedly revealed the effect of the copolymer design, namely, the introduction of cinnamyl-bearing units in the PCL block of the copolymer. Measurements were also performed at the physiological temperature of 37 • C. As many of the constituent elements of the copolymer micelles did not exhibit-sensitivity to temperature variations in this temperature range, the characterization parameters were practically the same as those at 25 • C and, therefore, not presented.
Furthermore, data for the enhanced encapsulation efficiency and loading capacity of the modified block copolymer micelles were in excellent agreement with the calculated Flory-Huggins solubilization parameters (χ sp ) for the two core-forming blocks ( Table S2 in the ESI). χ sp for the modified block P(CyCL-co-CL) was 0.0039 vs. 0.1195 for PCL, indicating the significantly greater affinity of the former to CBD [68][69][70].
TEM and AFM were performed to resolve the morphology of the empty and CBDloaded micelles (Figures 5 and 6). The objects were well-separated with dimensions in a dry state that was consistent with the results from the dynamic light scattering. The spherical morphology is dominant; some irregularities in the sphericity were occasionally observed and could be attributed to dehydration-induced artifacts. We suggested that the increased contrast at the periphery of the micelles was due to a selective interaction between PG corona and the staining agent-uranyl acetate. The loading of the micelles with CBD did not affect their morphology. small, we may speculate here that there are weak attractive interactions between the micelles, previously observed for self-assembled structures of polyglycidol-based copolymers [48,66,67].
Another parameter that can be extracted from the light scattering data is the loading capacity expressed as the number of CBD molecules loaded in one micelle of the novel and referent copolymer. The light scattering characterization data of the referent copolymer micelles loaded with CBD as well as the calculation of loading capacity are presented in Figure S5 and Table S1 in the ESI. Evident from the values of this quantity was that the micelles of the novel copolymer bore about 32% more CBD molecules than those of the referent copolymer, which undoubtedly revealed the effect of the copolymer design, namely, the introduction of cinnamyl-bearing units in the PCL block of the copolymer. Measurements were also performed at the physiological temperature of 37 °C. As many of the constituent elements of the copolymer micelles did not exhibit-sensitivity to temperature variations in this temperature range, the characterization parameters were practically the same as those at 25 °C and, therefore, not presented.
Furthermore, data for the enhanced encapsulation efficiency and loading capacity of the modified block copolymer micelles were in excellent agreement with the calculated Flory-Huggins solubilization parameters (χsp) for the two core-forming blocks ( Table S2 in the ESI). χsp for the modified block P(CyCL-co-CL) was 0.0039 vs. 0.1195 for PCL, indicating the significantly greater affinity of the former to CBD [68][69][70].
TEM and AFM were performed to resolve the morphology of the empty and CBDloaded micelles (Figures 5 and 6). The objects were well-separated with dimensions in a dry state that was consistent with the results from the dynamic light scattering. The spherical morphology is dominant; some irregularities in the sphericity were occasionally observed and could be attributed to dehydration-induced artifacts. We suggested that the increased contrast at the periphery of the micelles was due to a selective interaction between PG corona and the staining agent-uranyl acetate. The loading of the micelles with CBD did not affect their morphology.

In Vitro Drug Release and Cytotoxicity Assay
CBD release from prepared micelles was investigated by regular dialysis against PBS at 37 °C and the results are presented in Figure 7. Evident from the presented results is that the elaborated nanocarriers are able to release their cargo in a sustained manner for a prolonged period of time. This effect was more pronounced for micelles based on PG50-b-PPO4-b-[P(CyCL)4-co-(CL)40]-b-PPO4-b-PG50 where less than 45% of the encapsulated CBD was released at the 24th hour, compared to nearly 60% CBD released from the referent PG45-b-PCL35-b-PG45 micelles. This was probably due to the higher affinity of CBD to the cinnamyl-modified core and hence better solubilizing ability and loading efficiency, and this was in line with other reported studies [71]. The method of preparation of the micelles also had an effect on the release profile of CBD. In both types of micelles, those prepared according to protocol B showed a faster release, which can be explained by the entrapment of CBD in the periphery of the PCL core or even in the core-corona interface (Scheme 1a), rather than in the interior of the core. Indeed, PCL is a crystallizable polymer, and micelles based on this polymer are kinetically frozen structures, the hydrophobic domains of which can hardly be deeply penetrated by the external addition of the drug to preformed micelles. In contrast, in micelles prepared according to protocol A, the dominant location of CBD is likely to be well in the interior of the micellar core (Scheme 1b). It is anticipated that such localization would slow down the diffusion and, hence, the release of the drug [72]. In addition, the physicochemical affinity of CBD with the cinnamyl residues may further contribute to the slower drug release.

In Vitro Drug Release and Cytotoxicity Assay
CBD release from prepared micelles was investigated by regular dialysis against PBS at 37 • C and the results are presented in Figure 7. Evident from the presented results is that the elaborated nanocarriers are able to release their cargo in a sustained manner for a prolonged period of time. This effect was more pronounced for micelles based on PG 50 50 where less than 45% of the encapsulated CBD was released at the 24th hour, compared to nearly 60% CBD released from the referent PG 45 -b-PCL 35 -b-PG 45 micelles. This was probably due to the higher affinity of CBD to the cinnamyl-modified core and hence better solubilizing ability and loading efficiency, and this was in line with other reported studies [71]. The method of preparation of the micelles also had an effect on the release profile of CBD. In both types of micelles, those prepared according to protocol B showed a faster release, which can be explained by the entrapment of CBD in the periphery of the PCL core or even in the core-corona interface (Scheme 1a), rather than in the interior of the core. Indeed, PCL is a crystallizable polymer, and micelles based on this polymer are kinetically frozen structures, the hydrophobic domains of which can hardly be deeply penetrated by the external addition of the drug to preformed micelles. In contrast, in micelles prepared according to protocol A, the dominant location of CBD is likely to be well in the interior of the micellar core (Scheme 1b). It is anticipated that such localization would slow down the diffusion and, hence, the release of the drug [72]. In addition, the physicochemical affinity of CBD with the cinnamyl residues may further contribute to the slower drug release.

Micelles PG 50 -b-PPO 4 -b-[P(CyCL) 4 -co-(CL) 40 ]-b-PPO 4 -b-PG 50 :CBD ( Protocol B)
Micelles PG 45  In order to elucidate the release mechanism of CBD from the micelles, the release profiles were fitted by linear regression to several kinetic models: order Higuchi and Korsmeyer-Peppas ( Figures S6 and S7). Data are present For a more accurate interpretation of the data from the dissolution test, we conducted a non-linear regression analysis [73,74] whereby the drug-relea fitted to the Korsmeyer-Peppas model using DDSolver-a freely available software [75]. The results are summarized in Table S3. The obtained correlati data (Table S3) coincided with the corresponding values from the linear anal which proved the accuracy of the model. As can be seen from the data, the be was observed with the Korsmeyer-Peppas kinetic model, with n-values tested formulations below 0.45, indicating that the release followed a mecha cal Fick diffusion [76]. However, taking into account the low values of the k eters (release rate constant and half-release time) calculated for the different 4), it can be assumed that the CBD release mechanism is rather complex an to the diffusion, the probable redistribution of the released drug in the mice fluence due to its strong hydrophobicity. Table 4. Coefficient of determination (R 2 ), release rate constant (K), release half time sion exponent (n), after fitting of release profiles to different drug-release kinetic mo

Kinetic Model
Zero Order First Order Higuchi Korsme  Figure 7. In vitro release of CBD from block copolymer micelles in phosphate buffer (pH = 7.4). The copolymer/CBD weight ratio is 10:1.
In order to elucidate the release mechanism of CBD from the micelles, the data from the release profiles were fitted by linear regression to several kinetic models: zero and first order Higuchi and Korsmeyer-Peppas ( Figures S6 and S7). Data are presented in Table 4. For a more accurate interpretation of the data from the dissolution test, we additionally conducted a non-linear regression analysis [73,74] whereby the drug-release data were fitted to the Korsmeyer-Peppas model using DDSolver-a freely available Excel plug-in software [75]. The results are summarized in Table S3. The obtained correlation coefficient data (Table S3) coincided with the corresponding values from the linear analysis (Table 4), which proved the accuracy of the model. As can be seen from the data, the best correlation was observed with the Korsmeyer-Peppas kinetic model, with n-values found for all tested formulations below 0.45, indicating that the release followed a mechanism of typical Fick diffusion [76]. However, taking into account the low values of the kinetic parameters (release rate constant and half-release time) calculated for the different models (Table 4), it can be assumed that the CBD release mechanism is rather complex and, in addition to the diffusion, the probable redistribution of the released drug in the micelle has an influence due to its strong hydrophobicity.
A comparative evaluation of the antiproliferative effect of micellar CBD vs. free drug (applied as ethanol solution) against the acute myeloid leukemia-derived HL-60 cell line and Sezary Syndrome HUT-78 was performed. The growth-inhibitory concentrationresponse curves are shown in Figure 8 and the derived thereof equieffective IC 50 values are presented in Table 5. As seen from the presented data, the micellar CBD showed slightly lower cytotoxicity as compared to the free drug in both tested tumor lines, thus the concentration-response curves were shifted to the higher concentrations and, respectively, the IC 50 values were higher as compared to those of non-formulated drug, applied as an ethanol solution. Based on the presented results, it is clear that the incorporation of CBD into micelles has a modulatory effect on its range of antitumor activity, possibly due to changes in the release kinetics. These findings paralleled those of the drug-release study, where slower CBD release was reported for both micellar formulations. Table 4. Coefficient of determination (R 2 ), release rate constant (K), release half time (t 1/2 ), and diffusion exponent (n), after fitting of release profiles to different drug-release kinetic models.

Kinetic Model
Zero Order First Order Higuchi Korsmeyer-Peppas centration-response curves were shifted to the higher concentrations and, respectively, the IC50 values were higher as compared to those of non-formulated drug, applied as an ethanol solution. Based on the presented results, it is clear that the incorporation of CBD into micelles has a modulatory effect on its range of antitumor activity, possibly due to changes in the release kinetics. These findings paralleled those of the drug-release study, where slower CBD release was reported for both micellar formulations.   To prove that the observed cytotoxicity of micellar CBD is mainly due to the inherent cytotoxicity of CBD and not the carrier, the antiproliferative effect of unloaded micelles was also investigated in the same concentration range as that of the loaded counterparts ( Figure 9). The obtained results showed that the used polymeric micelles were devoid of cytotoxic potential. No significant differences to the untreated control were measured.

Pure CBD
2.00 7.00 To prove that the observed cytotoxicity of micellar CBD is mainly due to the inherent cytotoxicity of CBD and not the carrier, the antiproliferative effect of unloaded micelles was also investigated in the same concentration range as that of the loaded counterparts ( Figure 9). The obtained results showed that the used polymeric micelles were devoid of cytotoxic potential. No significant differences to the untreated control were measured.

Conclusions
Aiming at increasing the encapsulation efficiency and improving the performance of copolymer micelles as delivery systems of CBD, we designed a novel PEO-free PCL-PG copolymer by introducing a small number (an average of 4 out of 44) of monomer units bearing pendant cinnamyl groups in the middle block of PCL. Azide-alkyne click reactions were employed for the attachment of the pendant groups and for the conjugation of flanking polyether blocks to the middle polyester block. Amphiphilic properties were conferred upon, converting the flanking blocks into blocks of linear polyglycidol by removing the protective EEGE groups. In aqueous solution, the copolymer was found to spontaneously self-associate above a certain critical concentration into well-defined spherical micelles, characterized by moderately large size (Rh = 59.1 nm) and molar mass (Mw = 27.560 × 10 6 g/mol), and slightly negative (−5.90 mV) ζ potential. The micelles of the novel copolymer exhibited higher encapsulation efficiency towards CBD than those of the referent copolymer, independently from the loading protocols applied. Furthermore, the data revealed the enhancement of the loading capacity, expressed as the number of CBD molecules per micelle, by ca. 1/3 as compared to the referent copolymer micelles. Loading CBD during micelle formation (protocol A) resulted in the formation of more dense and com-

Conclusions
Aiming at increasing the encapsulation efficiency and improving the performance of copolymer micelles as delivery systems of CBD, we designed a novel PEO-free PCL-PG copolymer by introducing a small number (an average of 4 out of 44) of monomer units bearing pendant cinnamyl groups in the middle block of PCL. Azide-alkyne click reactions were employed for the attachment of the pendant groups and for the conjugation of flanking polyether blocks to the middle polyester block. Amphiphilic properties were conferred upon, converting the flanking blocks into blocks of linear polyglycidol by removing the protective EEGE groups. In aqueous solution, the copolymer was found to spontaneously self-associate above a certain critical concentration into well-defined spherical micelles, characterized by moderately large size (R h = 59.1 nm) and molar mass (M w = 27.560 × 10 6 g/mol), and slightly negative (−5.90 mV) ζ potential. The micelles of the novel copolymer exhibited higher encapsulation efficiency towards CBD than those of the referent copolymer, independently from the loading protocols applied. Furthermore, the data revealed the enhancement of the loading capacity, expressed as the number of CBD molecules per micelle, by ca. 1/3 as compared to the referent copolymer micelles. Loading CBD during micelle formation (protocol A) resulted in the formation of more dense and compact particles, which together with embedding CBD molecules into the interior of the micellar cores strongly reduced the initial burst effect and prolonged the drug release from the carriers. The copolymers displayed no sign of toxicity, whereas the cytotoxic activity and antiproliferative effect of CBD loaded in the micelles, against acute myeloid leukemiaderived HL-60 cell line and Sezary Syndrome HUT-78, was retained. The results of this demonstrate the effective design of a novel copolymer and the potential of its micelles as delivery vehicles of CBD. With their ability to significantly enhance the solubility of CBD and its loading efficiency, favorable physicochemical characteristics, appropriate release profiles, and excellent biocompatibility, the micelles of the novel copolymer can further enhance the experimental knowledge and therapeutic potential of CBD in neurological diseases and cancer.