Aqueous Worm Gels Can Be Reconstituted from Freeze-dried Diblock Copolymer Powder †

Worm-like diblock copolymer nanoparticles comprising poly(glycerol monomethacrylate) (PGMA) as a stabilizer block and poly(2-hydroxypropyl methacrylate) (PHPMA) as a core-forming block were readily synthesized at 10% w/w solids via aqueous dispersion polymerization at 70 C using Reversible Addition– Fragmentation chain Transfer (RAFT) chemistry. On cooling to 20 C, soft transparent free-standing gels are formed due to multiple inter-worm interactions. These aqueous PGMA–PHPMA diblock copolymer worms were freeze-dried, then redispersed in water with cooling to 3–5 C before warming up to 20 C; this protocol ensures molecular dissolution of the copolymer chains, which aids formation of a transparent aqueous gel. Rheology, SAXS and TEM studies confirm that such reconstituted gels comprise formed PGMA–PHPMA copolymer worms and they possess essentially the same physical properties determined for the original worm gels prior to freeze-drying. Such worm gel reconstitution is expected to be highly beneficial in the context of various biomedical applications, since it enables worm gels to be readily prepared using a wide range of cell growth media as the continuous aqueous phase.


Introduction
It has been known for more than y years that block copolymers can undergo self-assembly to produce a range of morphologies, including spheres, worms/cylinders or vesicles. [1][2][3][4][5][6] Various commercial applications have been developed for traditional diblock and triblock copolymers prepared using classical anionic polymerization, including thermoplastic elastomers, 7,8 surfactants/dispersants, [9][10][11] biocompatible injectable gels, [12][13][14] and toughening agents for epoxy resins. 15,16 One important parameter that inuences the copolymer morphology is the so-called packing parameter, which is a concept that was originally introduced to explain the selfassembly behavior of small molecule surfactants and later applied to block copolymers. 2,[17][18][19] Of particular relevance to the present study is the worm morphology, which has been recognized for at least three decades. 20,21 Seminal work in this area includes that by Bates and co-workers 4,22,23 and also by Discher's group. 24,25 For example, the latter team demonstrated that poly(ethylene oxide)-based diblock copolymer worms can avoid endocytosis by mammalian cells, which leads to relatively long circulation times and hence new opportunities for drug delivery. 26 More recently, there has also been considerable interest in the generation of diblock copolymer cylinders/rods based on semi-crystalline core blocks. [27][28][29][30][31][32] The development of living radical polymerization techniques such as reversible addition-fragmentation chain transfer (RAFT) polymerization 33,34 has enabled a wide range of novel block copolymers to be prepared directly using various functional monomers without recourse to protecting group chemistry. [35][36][37] As recently reported by our group [38][39][40][41] and others, [42][43][44][45] RAFT dispersion polymerization formulations offer particular advantages for the synthesis of well-dened diblock copolymer nano-objects. This is because polymerization-induced self-assembly (PISA) enables such syntheses to be conducted efficiently at relatively high solids while maintaining relatively low solution viscosity. Moreover, no post-polymerization processing step is required and the generic nature of this PISA approach has been demonstrated by the preparation of spheres, worms or vesicles in either aqueous solution, 38,39 alcoholic media 40,46 or n-alkanes. 41,43,47 In the present study, we have revisited our prototype formulation for RAFT aqueous dispersion polymerization, whereby a poly(glycerol monomethacrylate)-based macromolecular chain transfer agent (PGMA macro-CTA) is chainextended using 2-hydroxypropyl methacrylate (HPMA). 38,39 The latter monomer is water-miscible, but forms a water-insoluble PHPMA block. 38,39,48 This is a pre-requisite for aqueous dispersion polymerization, with the growing PHPMA driving in situ micellar self-assembly. 38 Although the phase space for diblock copolymer worms is relatively narrow, this copolymer morphology can be targeted reproducibly once a detailed phase diagram is established for a given formulation. 39 PGMA-PHPMA worms form so, free-standing gels at ambient temperature, 49 but de-gelation occurs on cooling to 5-10 C due to a worm-to-sphere transition. This reversible thermal transition allows convenient sterilization via cold ultraltration of the spherical nanoparticles, which efficiently removes any micrometer-sized bacteria that may be present. 49 We are currently exploring whether such worm gels are sufficiently biocompatible for long-term cell storage applications. In this particular context, an important question arises: can worm gels be reformed aer freeze-drying with essentially the same physical properties as the original as-synthesized aqueous worm gel? If so, this should enable worm gels to be reconstituted using a wide range of cell growth media instead of pure water. Herein we examine this important technical question using small-angle X-ray scattering (SAXS), dynamic light scattering (DLS), transmission electron microscopy (TEM) and gel rheology measurements. A solution of AIBN (1.23 g, 7.50 mmol) and bis-(2-phenylethanesulfanylthiocarbonyl) disulde (2.13 g, 5.00 mmol) were dissolved in ethyl acetate (50 ml) and purged under N 2 for 30 min. This reaction mixture was immersed in an oil bath at 92 C and reuxed for 16-20 h under N 2 . The organic phase was evaporated and puried by silica column (9 : 1, petroleum ether : ethyl acetate) to isolate the crude product as an orange oil. This oil was washed three times with petroleum to remove traces of ethyl acetate. The puried orange oil obtained at room temperature became a yellow solid when placed in the freezer (yield: 70%). 1  Synthesis of PGMA 49 macro-CTA using PETTC RAFT agent PETTC RAFT agent (0.50 g, 1.47 mmol), ACVA (82.50 mg, 0.29 mmol; PETTC/ACVA molar ratio ¼ 5.0) and GMA monomer (16.51 g, 0.10 mmol) were dissolved in anhydrous ethanol (26.50 ml) in a 100 ml round-bottomed ask and purged under N 2 for 30 min. The sealed ask was immersed into an oil bath set at 70 C for 100 min and quenched by exposure to air. The resulting PGMA macro-CTA was puried by precipitation (twice) into excess CH 2 Cl 2 from methanol. The mean degree of polymerization was calculated to be 49 by 1 H NMR spectroscopy. DMF GPC analysis indicated an M n of 11 100 g mol À1 and an M w /M n of 1.28 (vs. poly(methyl methacrylate) (PMMA) standards).

Experimental
Synthesis of PGMA 57 macro-CTA using PETTCCP RAFT agent The same protocol was used as for the PETTC-mediated synthesis of PGMA 49 macro-CTA described above. PETTCCP RAFT agent (0.

Protocol for redispersion of a freeze-dried copolymer worm gel
A 10% w/w copolymer gel (3.0 g) was freeze-dried overnight, yielding 0.30 g of a yellow copolymer powder. This powder was ground up using a mortar and pestle and deionized water (2.70 g) was added to produce an aqueous slurry. This was allowed to stand for 15 min, and then placed in an ice bath for 7-10 min to cool down to approximately 3-5 C to obtain a yellow liquid with small amount of foam on the top. A so, free-standing, transparent gel was formed aer allowing this copolymer sample to return to room temperature over a period of 10-15 min.

Rheology
Rheology measurements (a TA Instruments AR-G2 rheometer equipped with a Peltier heating/cooling plate) were performed on 10% w/w copolymer gels using cone and plate geometry (aluminum cone with 40 mm diameter and 2 cone). Frequency sweeps were conducted at xed strain (1.0%) and temperature (20 C). In order to study temperature response of the material, temperature sweeps in a range from 4 C to 35 C were performed at an angular frequency of 1.0 rad s À1 , 1.0% strain. Equilibration time of two minutes was used for one centigrade increment during the temperature ramp.

Transmission electron microscopy (TEM)
TEM images were recorded using a Phillips CM100 instrument operating at 100 kV and equipped with a Gatan 1k CCD camera. A 10% w/w PGMA 57 -PHPMA 140 copolymer worm gel was diluted to 0.20% w/w and stirred overnight to ensure complete worm dispersion. Copper/palladium TEM grids (Agar Scientic, UK) were surface-coated in-house to yield a thin amorphous carbon lm. The grids were plasma glow-discharged over 20-30 s to obtain a hydrophilic surface. Each sample (1.0 ml) was placed on the grid using a micropipette and dried under ambient conditions. A dilute aqueous solution of uranyl formate (0.75% w/w) was used as a stain (9 ml of stain, blotted aer 20 s). Each grid was then carefully dried using a vacuum hose. A slightly different protocol was used for the 10% w/w PGMA 49 -PHPMA 130 copolymer worm gel. This worm gel was diluted to 0.20% w/w solids using dilute aqueous HCl to maintain a solution pH of 3.6. Each TEM sample grid was then prepared as described above.
Small-angle X-ray scattering (SAXS) SAXS curves for 10% w/w copolymer worm gels were obtained using a Bruker AXS Nanostar instrument (camera length 1.06 m, CuKa radiation) equipped with a HiStar area detector and a semi-transparent beam-stop. SAXS patterns were recorded over a scattering vector (q) range of 0.1 nm À1 < q < 2.0 nm À1 . A liquid cell with a polytetrauoroethylene spacer (1.0 mm thickness) covered by two mica windows (thickness ¼ 25 mm) on both sides was used as a sample holder. The cell was connected to an ethylene glycol circulator preset to the required temperature. Dry nitrogen gas was used to reduce water condensation on both cell windows at low temperatures.

Visible absorption spectroscopy
The transparency of 10% w/w worm copolymer gels was assessed using a Varian Cary 300Bio UV-visible spectrophotometer equipped with a Peltier temperature-controlled multicell holder. Each gel was equilibrated for 10 min at 20 C before recording its visible absorption spectrum.

Results and discussion
Freeze-drying (or lyophilization) is widely used for the facile isolation of antibiotics, proteins or vaccines, and also for certain foodstuffs such as 'instant' coffee. This technique oen ensures more stable products and longer preservation times. 51,52 Freezedrying is also commonly used in both biology and chemistry to isolate liposomes, proteins, colloids and water-soluble polymers from aqueous solution. For example, Cabane et al. reported that this process can also be used to increase the size of lipid-based vesicles by increasing their concentration. 53 In the context of the present work, Cho and Chung reported that cholesterol-based polymeric vesicles containing [ 14 C] sucrose could be reconstituted aer freeze-drying on addition of water, with 90% of the sucrose payload remaining within the redispersed vesicles. 54 Furthermore, Qi and co-workers reported that freeze-dried poly(ethylene glycol)-stabilized ceria nanoparticles can be readily redispersed in various solvents with good reproducibility. 55 However, to the best of our knowledge, the reconstitution of freeze-dried diblock copolymer worm gels in aqueous solution has not yet been investigated. In order to evaluate such formulations, PGMA-PHPMA diblock copolymer worms were prepared at 10% w/w solids via RAFT aqueous dispersion polymerization at 70 C, as previously reported by Blanazs et al. 49 (Scheme 1). On cooling to 20 C, the worms form so, free-standing transparent gels that undergo de-gelation on cooling because of a reversible worm-to-sphere transition. 49 The two diblock copolymer compositions explored in the present work were PGMA 49 -PHPMA 130 and PGMA 57 -PHPMA 140 , which were prepared using two different trithiocarbonate RAFT CTAs.
Both as-synthesized 10% w/w worm gels were freeze-dried overnight, resulting in pale yellow copolymer powders. In principle, such lyophilization should remove the aqueous continuous phase, but leave the copolymer worm particles intact. 56,57 Grinding using a mortar and pestle degraded the brous nature of the freeze-dried worms and produced a free-owing powder. It was found empirically that this processing step aided subsequent copolymer rehydration: free-owing powder could be redispersed in water much more effectively than the initial brous powder.
An aqueous slurry (Scheme 2, image 3) was formed aer addition of water. TEM analysis (ESI, Fig. S2a †) of the rehydrated copolymer revealed that this initial slurry contained spheres, a minor population of worms and larger colloidal aggregates. In our initial experiments, the freeze-dried PGMA 49 -PHPMA 130 diblock copolymer was dispersed in deionized water at room temperature (20 C) and allowed to stand for several days. In the rst few hours aer rehydration, a turbid suspension was obtained. However, over the next 1-2 days a transparent free-standing gel eventually formed that contained some millimeter-sized trapped air bubbles (denoted PGMA 49 -PHPMA 130 -RT).
Lowering the solution temperature by immersing the initial aqueous copolymer slurry in an ice bath signicantly reduced the time required for gel reconstitution from 1-2 days to approximately 20 min. Thus this cooling stage was used as the standard protocol to redisperse the freeze-dried diblock copolymer worms (denoted PGMA 49 -PHPMA 130 -FD or PGMA 57 -PHPMA 140 -FD). Each step of this gel reconstitution process is summarized in Scheme 2.
TEM studies ( Fig. 1a and b, plus Fig. S3a and b in the ESI †) conrmed that both the original and reconstituted gels Scheme 2 Schematic representation of the standard protocol used to redisperse the freeze-dried PGMA-PHPMA diblock copolymer worms: (1) original worm gel, (2) freeze-drying overnight, followed by grinding the fibrous pale yellow powder using a mortar and pestle, (3) initial dispersion as an aqueous slurry, (4) cooling to 3-5 C using an ice bath, (5) allowing the cold copolymer dispersion to warm up to room temperature (20 C), (6) the reconstituted soft, transparent, free-standing worm gel.  (Table S1 and Fig. S4 in the ESI †) and/or SAXS. Rheology and SAXS studies were conducted in order to ascertain whether the reconstituted worm gels obtained aer freeze-drying had the same physical properties as the original worm gels. Rheological measurements conrmed that almost identical frequency dependence of loss modulus and storage modulus, G 0 and G 00 , respectively, are obtained for the worm gels before and aer freeze-drying (Fig. 2a). Similarly, the temperature response studies of G 0 and G 00 suggest that the critical gelation temperature for the worm-to-sphere remains essentially unchanged (Fig. 2b) and that there is little or no hysteresis (ESI, Fig. S5 †) between the heating and cooling cycles for each of the four copolymer samples investigated. SAXS data obtained for the original and reconstituted worm gels clearly indicate that the nal worm morphology obtained aer freeze-drying is indistinguishable from that of the original worms, since the I(q) vs. q plots almost overlay for both PGMA 49 -PHPMA 130 and PGMA 57 -PHPMA 140 (Fig. 3). This is in agreement with TEM data and conrmed that the reconstituted gel obtained aer freeze-drying is physically identical to the original worm gel ( Fig. 1a and b, and Fig. S3a and b in the ESI †).
As already noted above, lowering the solution temperature enabled signicantly faster reconstitution of worm gels from freeze-dried copolymer powder. Thus this important parameter was investigated in more detail. An initial aqueous slurry of diblock copolymer was placed in a À25 C freezer and became completely frozen (PGMA 49 -PHPMA 130 -FDF or PGMA 57 -PHPMA 140 -FDF). In this case, gel reconstitution appears to differ from the standard protocol: the freeze-dried copolymer dissolved molecularly under these conditions, which allows  more controlled, albeit slower, self-assembly. Thus the individual copolymer chains (or unimers) self-assemble to form rst spheres and then worms (see the SAXS curves shown in Fig. 4a and also DLS data in Fig. 4b). However, the gel strength determined for the PGMA 49 -PHPMA 130 -FDF worm gel is signicantly lower than that of the original PGMA 49 -PHPMA 130 worm gel. This may indicate that shorter worms are formed using this protocol. 48 The series of SAXS curves shown in Fig. 4a reveal the changes in morphology (from worms to spheres to unimers) observed for a 10% w/w aqueous dispersion of PGMA 57 -PHPMA 140 -FD on lowering the temperature from 25 C to À2 C (N.B. due to the Fig. 6 Rheological studies on diblock copolymer worm gels. (a) Frequency dependence of G 0 and G 00 at 20 C and an applied strain of 1.0% for the PGMA 49 -PHPMA 130 -FD worm gel redispersed at four copolymer concentrations (8.5 to 15% w/w). (b) Variation of critical gelation temperature (CGT) and gel strength (G 0 ) with copolymer concentration for four reconstituted PGMA 49 -PHPMA 130 -FD gels prepared at 8.5, 10, 12.5 and 15% w/w (denoted G 49 H 130 -FD). Fig. 4 (a) Small-angle X-ray scattering patterns recorded for a 10% w/ w aqueous dispersion of PGMA 57 -PHPMA 140 obtained after freezedrying and reconstitution using the standard protocol. I(q) vs. q plots were recorded at 25 C, 5 C, 2 C, À1 C, and À2 C; the change in slope of the SAXS curve from a negative gradient (black triangles) to zero gradient (blue inverted triangles) observed for q < 0.2 nm À1 indicates that the worm-to-sphere transition for this copolymer occurs on cooling from 25 C to 5 C. Further cooling from 5 C to À2 C leads to near-molecular dissolution of the copolymer chains. As a comparison, SAXS patterns of a 10% w/w solution of the same diblock copolymer molecularly dissolved in methanol at 20 C are given (red squares). Solid lines represent fittings to SAXS data by using Gaussian coil model (see main text and ESI † for further details). (b) DLS particle size distributions (calculated by volume%; see ESI and also Fig. S4 †) obtained for a 0.20% aqueous dispersion of PGMA 57 -PHPMA 140 -FD at 25, 20, 5 C and also for a 10% methanolic solution of PGMA 57 -PHPMA 140 -FD at 20 C. ***Sample was equilibrated for 1 h at 5 C. copolymer presence freezing of water does not occur at this latter temperature). The morphology of these nano-objects was conrmed by TEM studies (Fig. 1b and d) and is also consistent with DLS data (Fig. 4b and Table S1 and Fig. S4 in the ESI †). A t to the SAXS pattern of the PGMA 57 -PHPMA 140 aqueous solution at À2 C (Fig. 4a, grey diamonds) by using a Gaussian polymer chain model (ESI †) suggests that the sphere micelles in aqueous solution do not completely disintegrate into single molecules upon cooling to À2 C (ESI †). Comparison with the tted SAXS curve obtained for the same diblock copolymer molecularly dissolved as individual copolymer chains in methanol (Fig. 4a, red squares) suggests that near-molecular dissolution is achieved in water at around À2 C (N.B. the aqueous copolymer dispersion becomes frozen below this temperature). This is consistent with previously published data on PHPMA-based diblock copolymers. 49,58 Signicantly lower G 0 values for the reconstituted PGMA 49  In contrast, the slightly higher G 0 observed for the PGMA 49 -PHPMA 130 -RT gel compared to the original PGMA 49 -PHPMA 130 ( Fig. 5 and also Fig. S6a in the ESI †) may indicate longer worms, which would result in more/stronger inter-worm interactions.
Although the PGMA 49 -PHPMA 130 -FDF worm gel is slightly more transparent than the PGMA 49 -PHPMA 130 -FD (Fig. S7a in the ESI †), the visual appearance of these two gels is rather similar. Hence the freezing step at À25 C is not utilized for the standard protocol depicted in Scheme 2. The latter conditions allow the convenient production of a reconstituted worm gel with comparable physical properties to the original worm gel within a relatively short time scale and avoids problems such as trapped air bubbles.
To further probe the mechanism of worm gel reformation, freeze-dried worms (PGMA 57 -PHPMA 140 -CH 3 OH) were molecularly dissolved in methanol, followed by solvent evaporation. Thus this protocol ensures complete destruction of the original worm morphology. Water was then added to the sample vial, which was cooled in the freezer at À25 C prior to warming up to 20 C. The majority of this sample formed a reconstituted gel, as expected. However, this gel was weaker than expected as judged by rheology (Fig. S6b in the ESI †), which was attributed to a minor fraction of non-dispersed copolymer remaining on the walls of the sample vial. TEM analysis of this diluted PGMA 57 -PHPMA 140 -CH 3 OH gel revealed that it consisted of a mixture of worms and spheres, rather than pure worms (Fig. S2b in the ESI †). This observation is consistent with the lower G 0 value indicated by the rheology studies.
Successful gel reconstitution aer copolymer dissolution in methanol (or near-molecular dissolution at À2 C) suggests that the worms formed in aqueous solution is the thermodynamically-favored morphology for this particular copolymer composition, rather than merely a kinetically-trapped morphology.
Gel reconstitution from freeze-dried copolymer worms enables important physical properties such as the gel strength (G 0 ) and critical gelation temperature (CGT) to be tuned by simply varying the copolymer concentration (Fig. 6b). Thus increasing the copolymer content from 8.5 to 15% produces a monotonic reduction in the CGT from 17.5 C to 12 C and a linear increase in gel strength, which is consistent with recently published data. 48 The water content of hydrogels may vary over time because of either syneresis (i.e. the spontaneous expulsion of water from the gel) or evaporative losses. Gel reconstitution from freezedried copolymer powder should minimize this problem and also suggests that these worm gels may be recyclable aer initial use. In this context, it is perhaps also noteworthy that facile (re) sterilization of these aqueous worm gels can be achieved via cold ultraltration, as recently described by Blanazs et al. 49 One potentially decisive advantage offered by the reconstitution of freeze-dried worms is that this protocol enables gels to be conveniently prepared using bespoke aqueous media such as phosphate-buffered saline or various commercial cell growth medium (e.g. standard cell growth media such as DMEM or RPMI, with or without foetal calf serum, or more complex formulations used for stem cell growth). 59,60 This is expected to be benecial for various cell biology applications for these worm gels, which exhibit good biocompatibility. 49 Conclusions PGMA-PHPMA diblock copolymer worm gels can be reconstituted from freeze-dried worms. Gel formation is aided by initial dissolution in water with cooling to 3-5 C, since this leads to near-molecular dissolution of the diblock copolymer chains. Self-assembly to form individual worms, and ultimately so free-standing transparent gels, occurs on warming the cold copolymer aqueous solution/dispersion up to ambient temperature. SAXS, TEM and rheology studies indicate that the particle size/morphology, transparency and gel strength of the reconstituted worm gels are almost indistinguishable from the original worm gels prior to freeze-drying. These ndings are important, because these worm gels have potential applications as cell storage media. In this context, the ability to readily reconstitute worm gels within a range of cell growth media is expected to be an important advantage and also suggests that worm gel recycling should be feasible.