Polymer Architecture Effects on Poly(N,N-Diethyl Acrylamide)-b-Poly(Ethylene Glycol)-b-Poly(N,N-Diethyl Acrylamide) Thermoreversible Gels and Their Evaluation as a Healthcare Material

Thermoreversible gels which transition between liquid-like and solid-like states when warmed have enabled significant novel healthcare technologies. Poly(N,N-diethyl acrylamide) (PDEA) is a thermoresponsive polymer which can be used as a trigger to form thermoreversible gels, however its use in these materials is limited and crucial design principles are unknown. Herein ABA copolymers with the structure PDEA-b-poly(ethylene glycol) (PEG)-b-PDEA are synthesized to give four block copolymers with varied molecular weight of PDEA and PEG blocks. Rheometry on solutions of the block copolymers reveals that high molecular weight PEG blocks are required to form thermoreversible gels with predominantly solid-like behavior. Furthermore, small-angle X-ray scattering elucidates clear differences in the nanostructure of the copolymer library which can be linked to distinct rheological behaviors. A thermoreversible gel formulation based on PDEA (20 kDa)-b-PEG (10 kDa)-b-PDEA (20 kDa) is designed by optimizing the polymer concentration and ionic strength. It is found that the gel is mucoadhesive, stable, and non-toxic, as well as giving controlled release of a hydrophobic drug. Overall, this study provides insight into the effect of polymer architecture on the nanostructure and rheology of PDEA-b-PEG-b-PDEA and presents the development of a highly functional thermoreversible gel with high promise for healthcare applications.


Introduction
Thermoreversible gels may switch their state from a low viscosity solution to a gel dependent upon temperature. Engineering these materials so that the sol-gel transition occurs upon warming may be achieved by exploiting polymers with lower critical solution temperatures (LCSTs). [1,2] This phenomenon manifests as a solvophilicity when T < LCST but a relative solvophobicity when T > LCST, which in homopolymers is typically accompanied by a coil-to-globule transition and higher order mesoglobular states. [3,4] Thermoreversible gels may be achieved in these systems when the polymer exhibiting an LCST is covalently bonded to a polymer chain exhibiting solvophilicity over all temperatures investigated such that the LCST drives the formation of an amphiphilic state accompanied by self-assembly and/or polymer entanglements to form a percolating gel phase. [1] The triggering of this gel phase by gentle heating presents opportunities in applications such as drug delivery, [2]

Results and Discussion
PDEA-b-PEG-b-PDEA copolymers were synthesized by atom transfer radical polymerization to give a series of samples with molecular weight (in kDa) of each block controlled at two values, namely: A10-B5-A10, A20-B5-A20, A10-B10-A10, and A20-B10-A20. Polymer structure was confirmed by 1 H NMR (Figure 1a), which was also used to determine the number-average molecular weight of the blocks (Table S2, Supporting Information). GPC was used to determine Ð, which was <1.4 in all cases (Figure 1b,c). GPC also confirms the that PEG macroinitiator was not detectable in the product, with a shift of the peak to lower elution times (and thus assumed greater molecular weight). Whilst PEG-b-PDEA copolymers have recently been reported in the literature, [13] to our knowledge this is the first reported synthesis of PDEA-b-PEG-b-PDEA. This is particularly significant where the ABA triblock copolymer architecture gives access to nanostructures unavailable to diblock geometries and is associated with an enhanced ability to form gel phases. [1] It was hypothesized that ABA triblock copolymers of this type should have the ability to act as "loops", "bridges", or "dangling chains" in micellar structures and form percolating gel phases in an analogous manner to telechelic polymers exhibiting solvophilicity in their center and solvophobicity at their termini. [16] The potential of the PDEA-b-PEG-b-PDEA copolymers to form thermoreversible gels in aqueous solution was then evaluated by small-amplitude oscillatory shear rheology as a function of temperature at a fixed concentration of 20% (w/v). In these experiments G′ and G″ are measured as temperature changes at a fixed stress, within the LVR, and a fixed frequency. This ensures that any structures formed within the sample are not altered by high shear. All polymer architectures showed thermothickening from ca. 35-40°C, increasing in viscosity with temperature above this point (Figure 2). The composition of the polymer induced notably different rheological behaviors in the systems. Samples with 5 kDa central PEG blocks increased in viscosity but exhibited predominantly liquid-like behavior above the thermal transition, with G″ never exceeding G′. However, when 10 kDa PEG constitutes the "B" block of the ABA copolymer the materials exhibit a thermoresponsive gelation, with G′ exceeding G″, indicating a dominance of solid-like behavior in the system when heated. The gelation of A10-B10-A10 and A20-B10-A20 occurred at 51.2 ± 1.8 and 45.9 ± 0.5°C, respectively, indicating that the gelation process occurs at temperatures above the reported LCST of DEA homopolymer, ≈33°C. [8] This deviation may be the result of several factors. Conjugation to hydrophilic species has been shown to elevate LCST, [1] and it is known that macromolecular transitions continue to evolve above the LCST. [17,18] For example, the degree of phase separation of PNIPAM chains occurs over a broad transition in PNIPAM-b-PEG-b-PNIPAM copolymers. [17] Thus, nanostructures continue to evolve as heating continues above the LCST, rather than exhibiting a stepwise transition at this temperature. From the materials studied by rheology, A20-B10-A20 was identified as the most promising thermoreversible gel in the library due to its relatively high gel strength of up to 1.5 kPa.
The size of assembled structures of PDEA-b-PEG-b-PDEA copolymers in dilute aqueous solutions (1 mg mL −1 ) was probed by DLS above the transition temperature (Figure 3) in order to extract differences between the constructs, which could give clues into the mechanisms of gelation in more concentrated systems. The constructs with the larger PEG block, A10-B10-A10 and A20-B10-A20, formed well-defined nanoparticles with 68 nm median hydrodynamic diameter and polydispersity indices of 0.22 and 0.08, respectively. The lower molecular weight 5 kDa PEG copolymers assembled into much larger structures than the 10 kDa PEG constructs, with hydrodynamic diameters of 396 and 531 nm for A10-B5-A10 and A20-B5-A20, respectively.
SAXS was employed to further probe the nanostructure of PDEA-b-PEG-b-PDEA copolymers as a function of temperature. Measurements were taken from 5 wt% solutions in H 2 O at 30,37,40,50, and 60°C to span the range of temperatures covered by rheology and elucidate morphology above and below the transition temperature. All copolymers were investigated, however A20-B5-A20 visually showed evidence of phase separation at elevated temperatures-probably due to the large LCST blocks combined with the short hydrophilic PEG block-hence the data were not analyzed. The scattering profiles of all copolymers showed clear structural transitions when warmed above 40°C (Figure 4), in agreement with the rheology presented in Figure 2. The scattering from the A10-B5-A10 system (figure 4a) was fitted with a polydisperse Gaussian coil (PGC) form factor at 30 and 37°C, indicative of the system behaving as a disperse mixture of polymer chains in a theta solvent. [19] At 40°C the system was fitted with a cylinder form factor (radius of 6.2 nm), the length of which was outside the length scale resolved by SAXS and is expected to be greater than ca. 80 nm based on the q range available. At this temperature a power law form factor (exponent 1.61) was added to the model to fit the high q region, which may be a contribution from polymer chains in a good solvent (exponent 1.66), possibly arising from PEG chains in the micellar corona. [20] At 50 and 60°C, the data could be described by a cylinder form factor of radii 9.1 and 8.7 nm, respectively. These elongated structures are in agreement with the DLS data that detected very large aggregates. The construct with a larger PEG central block, A10-B10-A10 (Figure 4b), also revealed the presence of Gaussian coils at 30 and 37°C. At 40°C, a transition  requiring two form factors was again observed accounting for the presence of PGCs in addition to ellipsoidal objects with polar and equatorial radii of 20.5 and 10.2 nm, respectively. At 50 and 60°C, A10-B10-A10 data were best described by a spherical form factor with radii of 12.9 and 13.7 nm, respectively, which also required a contribution from a hard sphere structure factor, accounting for inter-particle repulsions. [21,22] A20-B10-A20 copolymers, which bear larger side chains and the same size of PEG central block, followed a similar trend. At 30°C the scattering data were fitted to a PGC form factor. At 37°C, assembled structures with an elliptical shape, with polar and equatorial radii of 26.2 and 11.3 nm, respectively, were detected, in addition to polymer coils. The same combination of form factors was used at 40°C, with the ellipsoids having polar and equatorial radii of 27.5 and 15.7 nm, respectively. At 50 and 60°C, only the ellipsoid form factor was required to fit the data, giving polar and equatorial radii of 28.5 and 19.1 nm, respectively, at 50°C and of 27.8 and 19.3 nm at 60°C, in broad agreement with the hydrodynamic diameter obtained from DLS. The ellipsoidal objects A20-B10-A20 also required the addition of a hard sphere structure factor to fit the data, which is a reasonable approximation for particle shapes that deviate from spheres. [21] Further detail about the models used and parameters obtained from the SAXS analysis can be found in the SI (Tables S3-S5, Supporting Information).
In summary, SAXS measurements indicate that the A10-B5-A10, A10-B10-A10, and A20-B10-A20 constructs are present as polymer chains in solution at lower temperatures, which is associated with a low viscosity ( Figure 2). When heated, the systems self-assemble into larger structures, driven by the LCST transition and the association of the relatively hydrophobic PDEA chains, leading to an increase in visco-elasticity (Figure 2). Cylindrical structures are present in the A10-B5-A10 systems ( Figure 5) above the transition temperature, which may be rationalized by the classical "molecular packing parameter" theory of surfactants that has been widely applied to self-assembled polymer structures. [23] In brief, the curvature of a  aggregate formed by polymeric amphiphiles is greatly dependent upon the area of the headgroup (in this case PEG) relative to the hydrophobic tail groups (PDEA above the LCST). When the relative area occupied by the headgroup is sufficiently small, the curvature necessary to form a sphere that minimizes the interface cannot be achieved and cylindrical structures are formed, as seen in constructs with a low M n PEG, such as A10-B5-A10, which are expected to be quite long as inferred from the DLS. Instead, the A10-B10-A10 constructs assembled into spheres above the LCST. In this case, the headgroup has a larger size and the packing parameter would allow sufficient curvature to thermodynamically favor a spherical shape. Above the LCST, the A20-B10-A20 copolymers assembled into ellipsoids with an aspect ratio of ≈2:3. In this case, the relative size of the head-to-tail group is lower and again non-spherical structures are formed. Linking these findings to the rheology elucidates an important contribution of particle shape to viscoelasticity. The cylindrical A10-B5-A10 nanostructures form viscous liquids dominated by the dissipative loss modulus (G″ > G′). However, where spherical and ellipsoidal nanostructures are present, the systems form predominantly elastic gels (G′ > G″) at higher concentrations. This may be rationalized by longer PEG chains in these systems allowing the aggregates to be bridged by elastically active unimer chains. Whilst an attractive interaction associated with unimer bridges was not detected via SAXS in these dilute solutions, previous SANS studies of similar systems at high concentrations support their presence. [24] Alternatively, the short chain length of PEG in the A10-B5-A10 and A20-B5-A20 systems may not allow bridge formation between aggregates. In this case, the increased viscosity in the A10-B5-A10 construct may be attributed to entanglements of elongated cylinders in a manner akin to the behavior of wormlike micelles. [25] Interestingly, whilst PEG length is crucial to gelation, no upper limit to this has been identified in the literature, to our knowledge. [14,[26][27][28] Future studies should expand this library to confirm optimal PEG block length at greater degrees of polymerization. The A20-B5-A20 systems additionally showed evidence of phase separation so their rheology is treated with caution.
Cytotoxicity measurements were conducted on human keratinocyte (HaCaT) cells to guide the selection of candidate materials for healthcare applications (Figure 6). There was no evidence of cytotoxicity in either MTS assay, a measurement of mitochondrial activity, or LDH assay, which determines the porosity of the cell membrane when exposed to the A10-B5-A10, A20-B5-A20, A10-B10-A10, or A20-B10-A20 samples. Whilst the dosing interval for these materials was short (2 h), the measurements provide preliminary mitigation of safety risks until optimal materials can be tested under toxicological regimes recognized by regulatory bodies. This testing also provides reassurance that the materials can be used in cell-based applications such as 3D culture or bioprinting.
A20-B10-A20 was selected as the lead thermoreversible gelator for further optimization due to its high gel strength and apparent cytocompatibility. The drawback of this material thus far is the transition temperature, with gelation not observed until 46°C. Initially the lowering of this temperature (T gel ) was attempted by simply increasing the concentration of the material, where competition for solvent is believed to drive the LCST to lower temperatures in similar systems. [14,29,30] Aqueous solutions www.advancedsciencenews.com www.mbs-journal.de  of A20-B10-A20 were prepared between 20 and 50% w/v and evaluated by rheology ( Figure S1, Supporting Information), as previously described, and the thickening temperature (Tthick), T gel and the maximum value of G′ (G′ max ) extracted (Figure 7a). It was observed that T gel decreased from ≈45 to 41°C when the concentration was increased from 20 to 30% (w/v), but plateaued above that concentration. Concurrently, G′ max increased up to ≈5.8 kPa at 30% (w/v), plateauing thereafter. The effect of sodium chloride on thermoreversible gel formation was then investigated to attempt to further reduce the LCST below physiological temperature (37°C) (Figure 7b). A near-linear decrease in T gel with sodium chloride concentration was observed. It is known that sodium chloride can alter the LCST of PNIPAM through the interaction of sodium cations with the oxygen atom in the amide group, which in turn reduces hydration. [31] It is reasonable to assume that this effect is also present in PDEA which is also a poly(N-alkyl acrylamide). The addition of 0.3 m sodium chloride to 30% w/v A20-B10-A20 solutions gave a thermoreversible gel with T gel occurring at 36°C, therefore making it an attractive candidate for the design of in situ gelators intended to transition between room (≈25°C) and body (37°C internally) temperature ( Figure 6c). Hereon, this material is referred to as A20-B10-A20 thermoreversible gel.
A20-B10-A20 thermoreversible gel was subjected to rigorous rheological evaluation to probe its potential as a healthcare material. Oscillatory frequency sweeps (Figure 8a) revealed a low dependence of G′ and G″ on frequency, indicative of a rigid structure which did not undergo structural rearrangements over Figure 7. a) Effect of concentration on the transition temperature and maximum gel strength in aqueous solutions of A20-B10-A20. b) Effect of NaCl on these parameters at a set concentration of A20-B10-A20 (30% (w/v)). c) Image of 30% (w/v) A20-B10-A20 in 0.3 m NaCl at room (≈20°C) and body temperature (after 5 min in a water bath at 37°C).

Figure 8.
Rheological evaluation of the optimized A20-B10-A20 PDEA-b-PEG-b-PDEA copolymer as a 30% (w/v) solution in 0.3 m NaCl. a) Frequency and b) amplitude sweeps at 37°C. c) Reversibility of gelation determined by small-amplitude oscillatory rheology with switching between sol (20°C) and gel (37°C) states (please note there was a 120 s equilibration time at each temperature change during which data was not collected). d) Time taken for gelation to occur determined by holding at 20°C for 60 s, then switching to 37°C. the test. Amplitude sweeps (Figure 8b) determined a LVR with a reduction in G′ indicative of the onset of yield occurring at 862 Pa, over threefold higher than the "gold standard" thermoreversible gel of 20% (w/v) poloxamer 407 which was determined as 256 Pa. [14] This greater yield stress is valuable in applications where the materials are required to retain their shape under shear, such as in topical drug delivery and in maintaining form after bioprinting. The reversibility of the thermal gelation was confirmed by small amplitude oscillatory rheology switching between 25 and 37°C (Figure 7c), confirming that the rheological behavior was preserved after a heating-cooling cycle. Finally, the time required for gel formation was investigated by holding the sample at 25°C then ramping to 37°C and maintaining the sample at that temperature under a small amplitude. The sample took 67 s to form a gel under these conditions which must be considered in its future applications, however thickening occurred immediately.
The stability of A20-B10-A20 was assessed in aqueous solution at 4, 25, or 40°C, reflecting refrigerated storage, storage at room temperature, and an accelerated storage condition, respectively (Figure 9). Accelerated storage at 40°C aimed to predict longerterm storage at room temperature. GPC analysis demonstrated that at 4 and 25°C the reduction in number-average molecular weight of A20-B10-A20 was not statistically significant, while at 40°C there was a significant decrease only by week 12 (p < 0.05 by one-way ANOVA with Tukey post-hoc analysis (GraphPad Prism, USA)). GPC traces of A20-B10-A20 remained monomodal with no shoulder ( Figure S2, Supporting Information), suggesting a reduced likelihood that hydrolysis of the ester moieties between the A and B blocks is a major route of degradation under these conditions. Overall, the A20-B10-A20 copolymer appears stable at and below room temperature in solution for the 12-week study period, supporting its stability during use in pharmaceuticals or cell culture.
The high-performance of the A20-B10-A20 thermoreversible gel makes it an attractive candidate for topical drug delivery to the vagina, where poor retention at this site due to shear forces reduces the efficacy of many semi-solids. [32] Thus, drugs applicable to vaginal drug delivery were incorporated into the thermoreversible gels to evaluate performance in topical drug delivery. Two drugs were investigated for inclusion into the thermoreversible gel, namely progesterone and tenofovir disoproxil fumarate, which have been applied intravaginally for indications including luteal phase support and HIV pre-exposure prophylaxis, respectively. [33,34] Progesterone (clog P = 3.87) is considered to be hydrophobic while tenofovir disoproxil fumarate (clog P = 1.25) is more hydrophilic with saturation solubilities in PBS at 25°C of 9 ± 1 and 7065 ± 412 μg mL −1 , respectively, and in PBS at 37°C of 11 ± 1 and 12 269 ± 393 μg mL −1 , respectively. It was found that the solubility of progesterone was greatly enhanced by inclusion in the A20-B10-A20 thermoreversible gel, particularly above T gel , with progesterone saturation solubilities at 25 and 37°C of 156 ± 6 and 324 ± 8 μg mL −1 , respectively, equivalent to a 17-and 29-fold increase in solubility. This is attributed to the formation of relatively hydrophobic PDEA domains in the material, even below the LCST. The saturation solubility of tenofovir disoproxil fumarate was adversely affected in the thermoreversible gels, having values at 25 and 37°C of 2582 ± 280 and 324 ± 8 μg mL −1 , respectively, equivalent to a 3-and 2-fold decrease, respectively. This decrease in solubility is partly explained by the reduced volume fraction of water but will also be a result of high solute concentration resulting in competition for hydration.
The release of progesterone or tenofovir disoproxil fumarate (50 μg mL −1 ) from the A20-B10-A20 thermoreversible gel across cellulose membrane was investigated at both 25 and 37°C (Figure 10). Progesterone release (Figure 9a) from the A20-B10-A20 thermoreversible gel at 25°C followed Higuchi kinetics and 100% of the drug was released after 32 h. The Higuchi model was applied to this release data as it is suitable for describing the release of both sparingly and abundantly soluble therapeutics from planar dosage forms. [35] This model is based on a linear fit to the fractional drug release with the square-root of time. [36] Increasing the temperature to 37°C retarded the release of progesterone significantly, with 100% drug liberated only after 144 h. This retardation of release at 37°C is attributed to the formation of the gel phase, which provides microphase separated PDEA domains of relative hydrophobicity in which drug solubilization may occur, disfavoring release. Additionally, increased tortuosity (i.e., a pathway having many turns) of the route taken by the drug for liberation within the gel phase will contribute to retardation by ultimately increasing distance over which diffusion must occur. The progesterone release profile at 37°C did not follow the Higuchi model, but its kinetics were linear after a 24 h lag period (R 2 > 0.99), which is highly attractive for controlled drug delivery releasing near-identical doses each day of use. The mechanisms for this zero-order behavior can be rationalized by analogy to biphasic hydrogels containing hydrophilic polymer chains modified to possess hydrophobic moieties, as described by Varelas and co-workers. [37] In these systems where swollen hydrogels contain hydrophobic microdomains, hydrophobic drugs partition preferentially into the microdomains which act as reservoirs. On exposure to excess water of low solute concentration, the drug is released from the gel by diffusion through the bulk hydrogel phase, which is then restored to the bulk from the microdomain reservoirs. If the flux achieved by these two processes are comparable, the concentration of drug in the bulk hydrogel remains steady and the driving force for diffusion is at a pseudo-steady state and zero order release is possible. The release of tenofovir disoproxil fumarate from the thermoreversible gel was more rapid, with ca. 65% drug release occurring after 8 h, and was unaffected by temperature. This has been reported for PNIPAM-b-PEG-b-PNIPAM gels, where it was hypothesized that the effect of increased tortuosity above T gel is counteracted by increased thermal energy favoring diffusion across the membrane. [14] Furthermore, the relatively hydrophilic tenofovir is not expected to partition favorably into the phase separated PDEA domains, counteracting any reservoir effect.
The potential of the A20-B10-A20 thermoreversible gels in vaginal drug delivery was further probed by assessing mucoadhesion using an established flow-through method. [38,39] In brief, fluorescein-doped samples were applied to the surface of ex vivo porcine vaginal mucosa and allowed to equilibrate at 37°C, mimicking internal body temperature. [32] Simulated vaginal fluid was then washed over the surface of the mucosa and the presence of sample determined by fluorescence microscopy (Figure 11a). This retention can be quantified by the measurement of image pixel intensity with time (Figure 11b) to give a percentage of the fluorescence at time zero. The A20-B10-A20 thermoreversible gel was compared against poloxamer 407 (20 % (w/v)) as a widely used thermoreversible gel, [12,40,41] poloxamer 188 (20 % (w/v)) as a non-gel-forming polymer control, and fluorescein solution as a negative non-mucoadhesive control. The A20-B10-A20 thermoreversible gel could match the behavior of poloxamer 407 gel, giving non-significant differences in retention at the experiment end whilst having the benefit of T gel closer to body temperature. The A20-B10-A20 thermoreversible gel outperformed the negative control and the poloxamer 188 control, showing discrimination between mucoadhesive and non-mucoadhesive samples in the experiment. Future tuning of A20-B10-A20 construct could lead to a material that outperforms poloxamer 407, either by further chemical modification of the copolymer to enhance covalent mucosa-polymer interactions (such as by thiol or maleimide chemistry) [42] or by formulation with mucoadhesive polymers. [43]

Conclusions
Triblock copolymers with the structure PDEA-b-PEG-b-PDEA exhibit a temperature-dependent rheological response in aqueous solution, where heating above ≈35°C induced an increase in viscosity. When heated, copolymer solutions with a 5 kDa central PEG blocks formed a viscous phase with dominance of a liquid-like behavior (G″ > G′), whereas constructs with 10 kDa PEG blocks formed gel phases with a predominantly solid-like behavior (G′ > G″). SAXS revealed that A10-B5-A10 copolymers (5 kDa central PEG) formed cylindrical nanostructures above the transition temperature, whereas A10-B10-A10 and A20-B10-A20 (10 kDa central PEG) formed spherical and elliptical aggregates, respectively; it is hypothesized that these aggregates are subsequently connected by long PEG chains forming bridges to form a percolating gel network. Combining rheology with cell culture studies led to the selection of A20-B10-A20 as the most promising thermoreversible gel for healthcare applications. Control of A20-B10-A20 concentration (30% w/v) and ionic strength (0.3 m NaCl) allowed the gelation temperature to be manipulated to below 37°C, enabling in situ gelation upon contact with the body's internal temperature, or warming during cell cul-ture, for instance. This optimized material was highly functional for topical drug delivery, enabling solubilization of poorly watersoluble progesterone and releasing the drug over 6 days, as well as matching the performance of poloxamer 407 in mucoadhesion studies.
Synthesis and Characterization of PDEA-b-PEG-b-PDEA Copolymers: Macroinitiators of 5 and 10 kDa PEG were synthesized and purified as described previously. [14] The reagents and quantities and experimental conditions used for the synthesis of tri-block copolymers by atom transfer radical polymerization (ATRP) are shown in Table S1, Supporting Information. The PEG macro-initiator, ligand (Me 6 TREN) and DEA were placed in a round-bottom flask and dissolved in a 1:1 water:MeOH mixture. Copper(I) bromide was added to a separate flask. Both flasks were then sealed and degassed with nitrogen purge for 30 min. After degassing, the solution containing monomer, macroinitiator and ligand was transferred to the copper(I) bromide flask via a degassed syringe. The flask was then stirred at room temperature for 24 h. It has been reported that removal of copper by neutral alumina alone causes copper-related cytotoxicity, which can be eliminated by passage of sample through neutral alumina and dialysis. [44] Therefore, all copolymers were passed through alumina and dialyzed for purification. Yields were >86% in all cases (Table S1, Supporting Information). These polymers are assigned identifiers based on the molecular weight in kDa of the constituent blocks, namely A10-B5-A10, A20-B5-A20, A10-B10-A10, and A20-B10-A20. 1 H NMR was performed on an Oxford Instrument ECA600 600 MHz NMR spectrometer with Delta 4.3.6 software. All samples were measured in D 2 O. The M n of triblock copolymers was calculated from 1 H NMR. The ratio of the PEG peak integral (3.5 ppm) to that of the DEA proton peaks was used to determine M n , given that the degree of polymerization of PEG is known. [14] An Agilent 12 600 Infinity II GPC equipped with a refractive index (RI) detector was used to characterize triblock copolymers, determining Ð and confirming monomodality. A phenomenex phenogel 10 μm 10E5 Å column was used with DMF with 0.1% LiBr as an eluent, at a flow rate of 0.4 mL min −1 with the column and detector held at 30°C. The GPC was calibrated with Agilent Easivial PMMA standards with peak molecular weight ranging from 370 to 364 000 Da.
Rheological Evaluation of PDEA-b-PEG-b-PDEA solutions: All rheology experiments were performed on a TA AR 1500 ex shear rheometer with a Peltier temperature control unit (±0.1°C) using rheology advantage software and a 40 mm parallel plate. The gap was 650 μm. A solvent trap was used to prevent evaporation. All samples were prepared at the stated composition in aqueous solution and left overnight in the fridge before performing rheometry. All experiments were repeated three times.
Oscillatory shear stress sweep experiments were performed at 20°C between 1 to 100 Pa. The linear viscoelastic region (LVR) was identified as the region before the increase in oscillatory stress caused both storage (G′) and loss (G″) moduli to decrease, which was determined to be the yield stress. Frequency sweeps were conducted between 0.1 and 10 Hz at a stress of 1 Pa.
Temperature ramps were performed at 1 Pa of oscillatory stress and a frequency of 1 Hz, with an increase in temperature from 20 to 70°C at a rate of 2°C per minute. The T gel was determined at the cross-over point between G′ and G″, a criterion widely used in the literature [45] and the gel "strength" was taken as the maximum value of G′ reached across the temperature range. Reversibility of this transition was explored by cycling the sample between 20 at 37°C, holding for 60 s at each temperature with a pre-measurement equilibration time of 120 s whenever the temperature was altered. Time-dependency of gelation was determined by holding the sample at 20°C for 60 s, then transitioning to 37°C and keeping the temperature fixed for 240 s. A stress of 1 Pa and frequency of 1 Hz was used in these experiments.
SAXS Measurements of Copolymer Solutions: Samples of each copolymer were prepared at 5 wt% in deionized water. SAXS measurements were performed on a Nano-inXider SAXS/WAXS instrument (Xenocs, Sassenage, France) at the Materials Characterization Laboratory of the ISIS Neutron and Muon source (STFC, Rutherford Appleton Laboratory, Didcot, UK). The setup is equipped with a micro-focus sealed-tube Cu 30 W/30 μm X-ray source (Cu K-, = 1.54 Å) and two Dectris Pilatus 3 hybrid pixel detectors, covering scattering vector q-ranges (q = 4 sin( /2), where is the wavelength of the incident beam and is the scattering angle) of 0.0045 to 0.37 Å −1 and 0.3 to 4.1 Å −1 for SAXS and WAXS, respectively. Scattering from the samples and water was collected in 1 mm glass thermalized capillaries at temperatures between 30 and 60°C. Data reduction (azimuthal averaging, buffer subtraction, absolute scaling) was carried out using the Foxtrot software. SAXS data was fitted using Sasview 4.2.2 software.
The scattering intensity I(q) can be written as follow: where, A is a scale factor, P(q) A is the form factor of the scattering object, S(q) A is the corresponding structure factor and BKG is the background. If more than one scattering object is present or the object studied has a hierarchical structure that generates scattering at distinct length scales, the expression can be extended to include further terms.
For this work, the polymer constructs, in general, give rise to two scattering components, one derived from their supramolecular structure and another from the polymeric chains. Therefore, the I(q) expression used was extended to: where, Scale, A and B are scale factors, P(q) A is the form factor for the model A, S(q) A is the corresponding structure factor, P(q) PGC is the form factor model of polydisperse polymer coils, and BKG is the background. [19] The SAXS data were fitted as follow: A10-B5-A10, either as polymeric Gaussian coils [19] or with a cylinder model. [46,47] A10-B10-A10, either polymeric Gaussian coils, [19] ellipsoids [48,49] plus polymeric Gaussian coils [19] or sphere [50] model for form factor and hard spheres [51,21] for the structure factor as needed. A20-B5-A20, either polymeric Gaussian coils [19] or spheres [50] models. A20-B10-A20, either polymeric Gaussian coils, [19] ellipsoids [48,49] plus polymeric Gaussian [19] coils or sphere [50] models for form factors and hard spheres structure factor as needed. [21,22] Dynamic Light Scattering: All dynamic light scattering (DLS) experiments were performed on a Malvern Zetasizer Nano ZS with a scattering angle of 173°. 1 mg mL −1 polymer solutions were prepared in deionized water and stored in the fridge overnight before use. Samples were then filtered through a 1 μm syringe filter and size measurements were taken at 50°C as a comparator measurement above the thermal transition in all systems. A 5-min equilibration period was used before each measurement, which was conducted in triplicate.
Investigating the Saturation Solubility of Progesterone and Tenofovir Disoproxil Fumarate in A20-B10-A20 Solution: 30% (w/v) A20-B10-A20 in 0.3 m NaCl was prepared and stored in the fridge overnight. Following this, 1 mL of the polymer solution or water was transferred to a vial and progesterone or tenofovir disoproxil fumarate (≈5 mg) was added. The solutions were then placed in a water bath at 25 or 37°C and allowed to stir for 24 h. If the resultant solution was clear, drug was again added and the sample left for 24 h in the water bath with constant stirring. This process was repeated until the solution remained turbid for 24 h. Excess drug was then removed by centrifugation (10 min at 14 500 rpm) ("high speed microcentrifuge", Four E's) and the clear supernatant was analyzed by HPLC using validated protocols described elsewhere. [14] Each experiment was performed in triplicate.
Investigating the Release of Progesterone and Tenofovir Disoproxil Fumarate in A20-B10-A20 Solution: The release of progesterone and tenofovir disoproxil fumarate from the 30% (w/v) A20-B10-A20 solution in 0.3 m NaCl was investigated using Franz diffusion cells (Soham Scientific, average bore size 174 mm 2 ) equipped with a cellulose membrane (Medicell, MWCO 3.5 kDa) at both 25 and 37°C. The receiver fluid was 10 mL phosphate buffer saline (PBS). The release of 50 μg mL −1 progesterone or tenofovir disoproxil fumarate in the polymer solution was investigated under sink conditions, by ensuring that drug in the receiver fluid could not exceed 10% of the drugs' saturation solubilities. The cells were placed into a water bath for 30 min prior to the experiment to reach the required temperature (25 or 37°C). Cells were dosed with 200 μL of sample and the release of drug measured at regular intervals by sampling receiver fluid (1000 or 200 μL for the progesterone and tenofovir disoproxil fumarate, respectively). The receiver fluid was replaced with an equal volume of pre-warmed PBS. The samples were analyzed by HPLC using established protocols. [14] The experiment was repeated four times.
Culture of HaCaT Cells: The human keratinocyte (HaCaT) cell line was used between passage numbers 2 and 20 from purchase. Cells were maintained in T75 cm 2 cell culture flasks containing 15 mL DMEM with 10% v/v heat-inactivated FBS and supplemented with 100 IU mL −1 penicillin, 100 μg mL −1 streptomycin solution and 2 mm L-glutamine. Cells were cultured in a humidified atmosphere at 37°C with 5% v/v CO 2 . Cells were passaged using 5 mL 0.25% w/v trypsin-EDTA solution once weekly when 90% confluent and subcultured at a ratio of 1:10. For experiments, cells were seeded at a density of 1 × 10 5 cells/well on 96 well tissue culture plates in 100 μL of complete cell culture medium and were cultured for 4 days prior to experimentation.
Cytotoxicity Testing of Triblock Copolymers on HaCat Cells: On the day of cytotoxicity assessment, cell culture medium was removed and replaced with 100 μL cell culture medium containing 10 mg mL −1 copolymer. Controls of untreated cells and cells treated with 0.1% v/v Triton-X 100 were included in each experiment. Cells were incubated at 37°C in 5% v/v CO 2 for 2 h and assessed for viability. Four replicates of the condition were used for each sample.
The CytoTox-ONE Homogenenous Membrane Integrity assay kit was used to quantify the extracellular concentration of lactate dehydrogenase (LDH) which arises due to the presence of pores in the cell membrane, one indicator of cytotoxicity. After 2 h of polymer exposure, 50 μL of cell supernatant was removed from each well and transferred to a black 96 well plate. To this, 50 uL of CytoTox-ONE Reagent (1:1 ratio) was added and cells were incubated in the dark for 10 min at room temperature (≈19°C) before the addition of 25 μL of stop solution. Fluorescence was measured immediately using a Promega Glomax Multi Detection System fluorescence plate reader with excitation wavelength of 560 nm and an emission wavelength of 590 nm. The data was then expressed as a percent cytotoxicity compared to positive control cells dosed with 0.1% v/v Triton-X 100 (Equation × 100 (3) The CellTiter96 Aqueous One Solution Cell Proliferation assay kit was used according to manufacturer's instruction to assess mitochondrial activity of the cells. This is used in combination with LDH assay as