Crosslinking Induced Reassembly of Multiblock Polymers: Addressing the Dilemma of Stability and Responsivity.

Abstract Physical or chemical crosslinking of polymeric micelles has emerged as a straightforward approach to overcome the intrinsic instability of assemblies. However, the crosslinking process may compromise the responsivity of nanosystems and result in inefficient release of payloads. To address this dilemma, a crosslinking induced reassembly (CIRA) strategy is reported here to simultaneously increase the kinetic and thermodynamic stability and redox‐responsivity of polymeric micelles. It is found that the click crosslinking of a model multiblock polyurethane at the micellar interface induces microphase separation between the soft and hard segments. The aggregation of hard domains gathers liable disulfide linkages around the interlayer of micelles, which could facilitate the attack of reducing agents and act as an intelligent on‐off switch for high stability and triggered release. As a result, the CIRA approach enables an enhanced tumor targeting, improved biodistribution and excellent therapeutic efficacy in vivo. This work provides a facile and versatile platform for controlled delivery applications.

2D nuclear Overhauser effect spectroscopy (NOESY) 1 H NMR spectra were measured using an AVANCE III HD spectrometer (400 MHz, JEOL) with a sweep width of 4000 Hz into 1024 data points. The relaxation delay was 2 s and the mixing time was 0.3 s. The number of scans was 4. 47200) standards. The sample concentration was 2 mg mL -1 and the flow rate was 1.0 mL min -1 .
The morphology of samples was observed using a transmission electron microscope (TEM), which was acquired on a model H-600-4 (Hitachi, Ltd., Japan) operated at an accelerating voltage of 75 KV. TEM grids were prepared by depositing a diluted suspension of sample onto a copper grid with staining with 1% (w/v) phosphotungstic acid for 3 min, the excessive solution was blotted away and air dried before imaging.
Fluorescence measurement was conducted on an F-4600 FL spectrophotometer (Hitachi, Ltd., Japan). For pyrene fluorescence, the excitation spectra were collected from 206 nm to 406 nm at an emission wavelength (λem) of 372 nm, the emission spectra were collected from 350 nm to 550 nm at an excitation wavelength (λex) of 331 nm. For Forster resonance energy transfer (FRET) measurements, the emission spectra were collected from 500 nm to 800 nm at a λex of 480 nm.
The size and zeta potential of polymer assemblies were obtained on a Zetasizer Nano ZS instrument (Malvern Instruments Ltd., UK) at room temperature at an angle of 90°. The relevant data were presented as mean ± standard deviation (SD) based on triplicate independent experiments.

Synthesis of clickable multiblock polyurethane (MPU)
The multiblock polyurethane was synthesized from PCL, BPEG, LDI and Cys-PA according to our previous reports. [1,2] Briefly, PCL (3.2 g) dissolved in DMAc was first copolymerized with LDI (0.949 g) at 60 °C under a dry nitrogen atmosphere in the presence of stannous octoate catalyst for 1 h. After cooling to room temperature, chain extender Cys-PA (0.628 g) was added and allowed to react with prepolymers for 1 h at room temperature, followed by another 2 h at 60 °C. Finally, BPEG (1.674 g) was added to react for 6 h. The polymer obtained was precipitated in anhydrous ethyl ether and dried under vacuum at 60 °C for 3 d.
The chemical structure of MPU was characterized by 1 H NMR, FTIR and GPC. As shown in Figure S1, all characteristic peaks of PCL, PEG, and LDI can be found. The peaks at 3.97 (-COOCH2-), 2.26 (-CH2COO-),1.52 (-CH2CH2CH2-), and 1.30 ppm (-CH2CH2CH2-) are assigned to the methylene protons of PCL unit. The sharp peak at 3.50 ppm is attributed to the methylene protons of PEG block (-CH2CH2O-). The chemical shifts of methylene (-CH2-OCO-) and methyl protons (-CH3) in the ethoxyl group of LDI are at 4.07 and 1.15 ppm, respectively. The characteristic peak of the imine proton (-HC=N-) is at 8.20 ppm, and the signals at 8.02 and 7.89 ppm are originated from benzene ring of BPEG. In addition, peaks at 2.73 and 3.07−3.16 ppm are ascribed to the alkynyl proton (-C≡CH) and methylene protons next to the disulfide bond (-S−S−CH2-) in Cys-PA, respectively, demonstrating that the multifunctional chain extender has been successfully introduced into the chains of polyurethanes. The FTIR spectra of polyurethanes are shown in Figure S2. The stretching band in the 1600−1800 cm −1 region is overlapped by the absorption of ester carbonyl groups of PCL and free and hydrogen-bonded carbonyl of urethane groups, where a shoulder observed at 1654 cm −1 is ascribed to the hydrogen-bonded carbonyl of urea groups. A broad stretching band around 3340 cm −1 is mainly attributed to the hydrogen-bonded N-H stretching vibration. GPC analysis indicates that the weight average molecular weight of MPU is about 70933 g mol -1 , with monodisperse and quite narrow molecular weight distributions (PDI 1.20, Figure S3). All the above results prove that the polymer has been successfully synthesized.

Synthesis of azidoethylamine (AzEA)
AzEA was synthesized according to Scheme S2. In brief, 2-bromoethylamine hydrobromide (10.25 g) was dissolved in distilled water with mild stirring. Sodium azide (9.75 g) was added carefully and the mixture was kept at 70 °C with refluxing for 12 h. Thereafter, the system was cooled to 0 °C and NaOH (7.00 g) was added. The product was extracted with anhydrous ethyl ether and dried over Na2SO4. After filtration of the solution, the solvent was allowed to evaporate under atmospheric conditions (yield: 50−60%). 1

Synthesis of reduction-cleavable crosslinker (SS-Az)
The reduction-cleavable crosslinker (SS-Az) crosslinker was synthesized from AzEA and DTDPA (Scheme S3). Briefly, DTDPA was completely dissolved in DCM precooled in an ice water bath, then AzEA, DCC and HOBt were added in turn to the solution and stirred for 1 h.
The reaction was kept at room temperature for 24 h. Then the solvent was evaporated under atmospheric conditions. Afterward, 100 mL 0.5 M hydrochloric acid was added, and product was extracted with ethyl acetate. The organic phase was collected and washed with saturated solution of NaHCO3, sodium chloride and distilled water, and dried over anhydrous Na2SO4

Self-assembly of multiblock polyurethane
The assemblies of block copolymers were prepared using a dialysis method. Briefly, a solution of MPU (25 mg) in 2.5 mL of DMAc was added dropwise into 25 mL of deionized water with quickly stirring. Then the solutions were transferred into a dialysis bag (MWCO 3500) and dialyzed against deionized water for 3 d, changing the external water once 3 h.

Computational simulation
To investigate the structure of MPU micelles, computational simulation was carried out using a dissipative particle dynamics (DPD) model. DPD simulation is a particle-based mesoscopic simulation technique originally introduced by Hoogerbrugge and Koelman in 1992, [3,4] and further modified by Español and Warren. [5] It has been established as a powerful tool to investigate the self-assembly of amphiphilic copolymers. [6] In our study, we consider an aqueous solution (W) of MPU. The system comprises clickable multiblock polymers and water in a cubic box of size 20 × 20 × 20 rc 3 with periodic boundary condition.
MPU was divided into six types of beads (E, A, B, L, C, and S). Simple coarse-grained models of these components are shown in Fig. S6. The calculated interaction parameters in polymeric micellar systems at 298 K are given in Table S2. Detailed simulation methods and equations can be found in our previous reports. [7][8][9][10][11] The DPD simulations were conducted using Materials Studio 5.0 software (Accelrys) installed on a DELL PowerEdge SC430 server.
The total beads were 24,000, the spring constant C was chosen as 4.0 and the time step was taken as 0.05. According to Figure S7A, 100,000 DPD steps adopted were sufficient for achieving simulation equilibrium and steady results.
The simulation results were depicted in Figure S7. As shown in the density profiles ( Figure   S7B), front view ( Figure S7C) and cross-sectional view of MPU micelles ( Figure S7D), the multiblock polyurethane self-assembles into a spherical core-shell structure with a hydrophobic core formed by insoluble PCL soft segments (blue) and surrounded by a hydrophilic BPEG corona (green). The hard segments composed by Cys-PA (red) and LDI residues (yellow) are located mainly at the subsurface, with some still distributed in the micellar core due to neighboring hydrophobic soft segments.

Crosslink of multiblock polyurethane assemblies
The obtained crosslinker contains a disulfide linkage and two azide sites, allowing for an efficient crosslinking of MPU micelles via a copper catalyzed alkyne-azide cycloaddition (CuAAC) in aqueous solution. The degree of crosslinking could be controlled by the feed ratio of crosslinker. Briefly, MPU micelles (25 mL) were mixed with SS-Az (0.5 and 10 eq) in the presence of sodium ascorbate (30 mg) and CuSO4· 5H2O (20 mg) and incubated at room temperature under moderate stirring for 24 h. Afterward, the solutions were transferred to dialysis tubes (MWCO 3500) and dialyzed against deionized water for 2 d to remove the unreacted SS-Az and traces of the catalyst. Then the crosslinked assemblies (CMPU) solutions were centrifugalized at 3000 r min −1 for 20 min and passed through a 0.45 μm poresized syringe filter.
The above results verify the successful crosslinking of MPU micelles via click chemistry.

Determination of aggregation number (Nagg)
The weight-average molecular weight of multiblock polyurethane assemblies before and after crosslink were measured by SLS using the Debye plot. The aggregation numbers of MPU and CMPU assemblies were calculated by eq S1: where Mw, aggregate is the weight-average molecular weight of polymeric assemblies estimated by SLS and Mw, block copolymer is the sum of weight-average molecular weight of polymer obtained by GPC analysis.

DOX and Cy5 encapsulation
To load DOX or Cy5, 1 mL of DOX (1 mg mL -1 ) or Cy5 (1 mg mL -1 ) solutions in DCM was added into a bottle, and dried with a flow dry argon. Then, 10 mL of polymer dispersions in water (3mg mL -1 ) were added into the bottle and ultrasonated for 2 h. The solution was transferred into a dialysis bag (MWCO 3500) and dialyzed against water for 24 h, changing the water every 3 h. The fluorescent dye-loaded assemblies (DOX@MPU or Cy5@MPU) were crosslinked as described above to prepare dye-labeled crosslinked micelles (DOX@CMPU or Cy5@CMPU). Finally, all the solutions were centrifugalized for 10 min at 3000 r min -1 and filtered through a 0.45 μm pore-sized syringe filter (Millipore, Carrigtwohill, Co. Cork, Ireland).
To co-encapsulate DOX and Cy5, 1 mL of DOX (1 mg mL -1 ) solutions in DCM was added into a bottle, and dried with a flow dry argon. Then 10 mL of Cy5@MPU solution were added into the bottle and ultrasonated for 2 h. The solution was transferred into a dialysis bag (MWCO 3500) and dialyzed against water for 24 h, changing the water every 3 h. The DOX and Cy5 co-loaded micelles (DOX+Cy5@MPU) were crosslinked as described above to prepare crosslinked FRET micelles (DOX+Cy5@CMPU). Finally, the solutions were centrifugalized for 10 min at 3000 r min -1 and filtered through a 0.45 μm pore-sized syringe filter (Millipore, Carrigtwohill, Co. Cork, Ireland).

DMF dilution test
Aqueous solution of micelles (100 μL) was placed in a vial, followed by the addition of pure DMF (1 mL). The solution was shaken for 1 min to fully mix DMF and water. Then the size and size distribution of micelles were measured by DLS.

Kinetic study of dye-encapsulated micelles
The crosslinking of MPU micelles was proved using fluorescence resonance energy transfer (FRET), which is a facile and straightforward tool to detect the molecular interactions within the range of 10 nm and monitor the process and dynamics of self-assembly in real time. [12,13] As a pair of FRET dyes, DOX (donor) and Cy5 (acceptor) were encapsulated into MPU and CMPU assemblies separately as described above. To determine the kinetic stability of MPU and CMPU micelles, the polymeric assemblies encapsulating DOX or Cy5 were mixed for different times and determined with an F-4600 FL spectrophotometer (Hitachi, Ltd., Japan). The donor (DOX) was excited at 480 nm and the emission spectra were recorded at all wavelengths simultaneously.

Stability of multiblock polyurethane micelles against dilution
To investigate whether crosslinking improves the thermodynamic stability of micelles, the multiblock polyurethane micelles before and after crosslinking were diluted with deionized water and phosphate buffer solution (PBS). The particle sizes under different dilution times were measured with a Zetasizer Nano ZS instrument (Malvern Instruments Ltd., UK) at room temperature at an angle of 90°.
In addition, FRET measurements were also conducted to study the stability of assemblies before and after crosslink. In brief, the fluorescence spectra of DOX and Cy5 co-loaded assemblies before and after crosslink upon dilution was collected. The FRET efficiency was calculated from the intensity ratio IA/ (ID + IA), where IA and ID were the fluorescence intensities at 690 and 550 nm, respectively.

Stability of multiblock polyurethane micelles against surfactant, protein and serum
To investigate the potential stability of multiblock polyurethane micelles under physiological conditions, the MPU and CMPU micelles were incubated with sodium dodecyl sulfate (SDS, 10mg mL -1 ), bovine serum albumin (BSA, 45 mg mL -1 ) or fetal bovine serum (20%) with shaking. The size and size distribution of micelles were monitored over time by Zetasizer Nano ZS instrument (Malvern Instruments Ltd., UK).

Responsiveness of multiblock polyurethane micelles
To evaluate the reduction responsivity of MPU assemblies before and after crosslink, DOX+Cy5@MPU and DOX+Cy5@CMPU micelles were treated with 10 mM of GSH and measured with an F-4600 FL spectrophotometer (Hitachi, Ltd., Japan) at different time points.
The donor (DOX) was excited at 480 nm and the emission spectra were recorded at all wavelengths simultaneously. The ratio of fluorescence intensity at 594 nm to that at 670 nm was normalized and plotted agaist time.

Triggered release of DOX
To verify whether reversible crosslinking of MPU assemblies enables controlled release of payloads in tumor microenvironment, a model drug DOX was encapsulated into the micelles and crosslinked as described above. The release of DOX was evaluated using a dialysis method. Briefly, 3 mL of DOX-loaded assemblies before and after crosslink were added into

Cell internalization
The cellular uptake of DOX and Cy5-coloaded polymeric micelles before and after crosslinking was determined by confocal laser scanning microscope (CLSM) and flow cytometry. For CLSM, MCF-7 breast cancer cells obtained from West China Hospital were seeded in a six-well plate (a coverslip was placed in every well before use) at a density of 1 × 10 5 cells per well and cultured overnight. Then DOX+Cy5@MPU and DOX+Cy5@CMPU assemblies were added separately into the plate with a consistent drug concentration of 10 μg mL -1 and incubated for 4 h. Next, the medium was removed and the plate was washed with PBS for three times. Then the cells were fixed with 4% formaldehyde for 30 min and stained with DAPI for 10 min. At last, the coverslips were mounted with 50% glycerol solution and observed on a CLSM (Olympus FV1000, Japan). For flow cytometry, MCF-7 breast cancer cells were seeded in a six-well plate (a coverslip was placed in every well before use) at a density of 1 × 10 5 cells per well and cultured overnight. The cells were then treated with 10 mM glutathione ethyl ester (GSH-OEt) for 2 h or 0.1 mM L-buthionine-sulfoximine (BSO) for 2 h. Thereafter, DOX+Cy5@MPU and DOX+Cy5@CMPU micelles were added at a consistent drug concentration of 10 μg mL -1 , and the cells were incubated for another 4 h.
After removing the medium, the plate was washed with PBS for three times. Then the cells were digested by trypsin, centrifuged, and re-suspended in 0.5 mL PBS for flow cytometer measurement (Beckman Cytoflex, USA).

Endocytosis mechanism
To evaluate the endocytosis mechanism of multiblock polyurethane micelles before and after crosslinking, MCF-7 cells were seeded in a six-well plate (a coverslip was placed in every well ahead of use) at a density of 1 × 10 5 cells per well and cultured overnight. Then the cells were pre-incubated with different inhibitors: M-β-cyclodextrin (2.5 mM), chloropromazine (10 μg mL -1 ), colchicine (8 μg mL -1 ), genistein (50 μg mL -1 ) for 2 h at 37 °C.
Meanwhile, another two groups of cells were pre-incubated without inhibitor at 4 ℃ and 37 °C for 2 h. Cells without pretreatment were set as control. Then DOX@MPU and DOX@CMPU micelles were added into the plate with a consistent drug concentration of 10 μg mL -1 and incubated at 37 °C or 4 ℃ for 4 h. Finally, the cells were washed, digested, centrifuged and resuspended for flow cytometer measurement.

Construction of tumor model
Five to six-week-old female BALB/c nude mice or KM mice were purchased from Vital River Company in Beijing. All experimental procedures were in accordance with the guidelines for laboratory animals established by the Laboratory Animal Center of Sichuan University. MCF-7 or 4T1 cells were large-scale expanded ex vivo in culture medium and collected in PBS. 100 μL of cell suspensions (2.0×10 7 cells mL -1 ) were injected into the right armpit of BABL/c nude mice or KM mice. The body weight and tumor size were measured every three days. The tumor volume was calculated using the equation V = ab 2 /2, where "a" and "b" represent the length and width of tumors, respectively.

In vivo and ex vivo imaging study
To investigate the targeting property and biodistribution of multiblock polyurethane micelles in vivo, MCF-7 tumor-bearing nude mice or 4T1 tumor-bearing KM mice were randomly divided into three groups. When the tumors had grown to around 100 mm 3 , the mice were intravenously injected with DOX+Cy5@MPU and DOX+Cy5@CMPU micelles via the tail vein, and tracked by an IVIS imaging system (Caliper Life Sciences, USA) at different time points. The excitation filter is 490 nm, and the emission filter are 600 and 700 nm. The animals were sacrificed at 24 h post-administration, and tumor tissues and major organs including heart, liver, spleen, lung, and kidney were collected for ex vivo fluorescence examination using the same imaging system. For the confocal analysis, excised tumors were frozen in optimum cutting temperature (OCT) (Sakura Finetek, USA) at -80 ℃. The corresponding slices (6 μm) were prepared. The red fluorescence emitted from DOX was collected on a CLSM (Olympus FV1000, Japan).
Antitumor treatment MCF-7 tumor-bearing nude mice were divided into four groups (five mice per group): saline, DOX@MPU, DOX@CMPU and DOX. When the tumors had grown to 30-50 mm 3 , the nude mice were treated with DOX@MPU and DOX@CMPU via tail vein every 3 d for 15 d at a DOX dose of 5 mg kg -1 . Mice injected with free DOX and saline were set as positive and negative controls, respectively. The tumor sizes were recorded every three days using a digital caliper. On the day of 15, all the mice were sacrificed and the tumors were excised and weighed, the major organs (heart, liver, spleen, lung and kidney) and tumors were collected for further analysis.
The biological safety in vivo and tumor suppression effect after the treatment were assessed by hematoxylin-eosin (H&E) staining, terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL) assay and nuclear-associated antigen (Ki-67) immunohistochemistry analysis. The tumors and main organs such as liver, kidney, heart, lung, and spleen were collected, embedded with paraffin, and cut into 5-μm-thick sections. The tissues were then stained with H&E, TUNEL, Ki67 and observed with fluorescence microscopy to assess the histopathology alterations.

Statistical analysis
The quantitative data obtained were expressed as means ± standard deviations (SD).
Statistical analysis was carried out using the Statistical Package for the Social Sciences (IBM SPSS Statistics software, Version 19, IBM, New York, USA). Student's t-test or one-way analysis of variance (ANOVA) was performed to determine the statistical significance within the data at 95% confidence levels (P < 0.05).