Novel thermoresponsive polymer outperforming Pluronic® F127


 Thermoresponsive polymers featuring the appropriate combination of structural characteristics, i.e. architecture, composition, and molar mass (MM), can form physically crosslinked networks in a solvent upon changes in temperature. This fascinating class of polymers finds utility in various sectors such as formulation science and tissue engineering. Here, we report a novel thermoresponsive triblock terpolymer which out-performs the most commonly used and commercially available thermoresponsive polymer, Poloxamer P407 (also known as Pluronic® F127) in terms of gelation concentration. Specifically, the in-house synthesised polymer forms gels at lower concentrations that is an advantage in biomedical applications. To elucidate the differences in their macroscale gelling behaviour, we investigate their micellization via differential scanning calorimetry, and their nanoscale self-assembly behaviour in detail by means of small-angle neutron scattering by simultaneously recording their rheological properties (Rheo-SANS). Two different gelation mechanisms for the two polymers are revealed and proposed. Ex vivo gelation study upon intracameral injections demonstrated excellent potential for its application to improve drug residence in the eye.


Introduction
Thermoresponsive gels (TRGs) are 3-dimensional (3-D) networks of non-covalently interacting micelles that can reversibly turn to the solution phase as the stimulus (i.e. temperature) is removed. [1][2][3][4][5] For aqueous solutions, the crosslinking points are formed by hydrophobic junctions that are connected via hydrophilic bridges. This fascinating class of polymers has been extensively studied by polymer scientists, engineers, and biologists over the last decades, with the rst systematic study on the subject dating back to 1984. 6 Particularly interesting are the TRGs that are formed upon a temperature increase; these polymers present a lower critical solution temperature (LCST), and their solubility decreases as function of temperature. 1 TRGs with LCST behaviour have attracted much attention in the biomedical sector due to their reversible nature and their inherent ability of physical crosslinking without additional chemicals that might induce toxicity. They have been extensively studied as injectable gels either for tissue engineering (TE), 1,7,8 or to serve as drug delivery systems. 9,10 In this concept, a prerequisite is a sol-gel transition between room temperature and physiological temperature. As a result, the solution phase can be loaded with cells and/or drugs at ambient temperature, in the case of TE and/or drug delivery application, respectively, and easily loaded into a syringe. Upon injection to physiological conditions, a hydrogel matrix is formed that traps the cells and/or drug molecules in place. In the case of drug delivery, the gel matrix provides sustained and topical release of the drug, which minimises undesired side-effects, caused by the systemic release. TRGs are most recently being investigated as 3-D printable (bio-)materials, such as implants. [11][12][13][14][15] The printing can be carried out either at room temperature or body temperature, depending on the desired application, and its feasibility relies on the shear-thinning properties of the hydrogel.
As previously mentioned, the synthesis and the applications of thermoresponsive polymers have been extensively reported. [16][17][18][19][20][21][22][23][24][25][26][27][28] Among these, poloxamer P407 is the most commonly used thermoresponsive polymer, owing to its commercial availability. 21,29,30 This copolymer is a member of the family of Pluronics®, which are ABA triblock copolymers with A and B units being ethylene glycol (EG) and propylene glycol (PG), respectively, i.e. EG m -b-PG n -b-EG m ; where m and n denote the degrees of polymerisation (DP) of EG and PG, respectively. It is known as Pluronic® F127 (registered tradename of BASF), and it consists of EG at 70 mol% and PG at 30 mol%, and its average molar mass (MM) is around 12600 g mol -1 ; its chemical structure is EG 66 -b-PG 99 -b-EG 66 (the given DPs are approximate), and it is presented in Fig. 1a). [31][32][33] Poloxamers are widely used in industry as gelling, emulsifying and defoaming agents. [34][35][36] In the biomedical sector, and more speci cally in the application of injectable gels, to the best of our knowledge, Pluronic® F127 is one of the two polymers that reached clinical trials, along with ReGel® (OncoGel: formulation with paclitaxel, NCT00479765, NCT00573131). [37][38][39] More speci cally, Pluronic® F127 has been evaluated as a Simvastatin gel for the treatment of mucositis (20% gel, NCT03400475), and as a formulation with Pluronic® F68 as a Metronidazole vaginal gel for the treatment of bacterial vaginosis (20% F127/10% F68, NCT02365389). 38 In addition, its mixture with Pluronic® F68 is currently being trialled as a Timolol gel for the treatment of epistaxis (NCT04139018). 38 However, the minimum concentration at which Pluronic® F127 forms hydrogels is 15 w/w%, which might increase the cost of the application. Furthermore, depending on the concentration, Pluronic® F127 forms a gel or a very viscous liquid at room temperature, which makes it di cult to handle for some applications, such as those requiring injection through narrow needles. Therefore, there is a need for new thermoresponsive polymers able to overcome these challenges and that are easy to synthesise, thus rendering them suitable for industrial production.
To meet the current need, we report the design, synthesis and characterisation of a novel thermoresponsive polymer with a gelation concentration (C gel ) which is seven times lower than the one of Pluronic® F127. This copolymer is an ABC triblock terpolymer based on a novel combination of methacrylate units, in which B is the hydrophobic n-butyl methacrylate (BuMA), while A, and C are hydrophilic EG-based methacrylate units of different lengths. More speci cally, A and C are penta-and di-(ethylene glycol) methyl ether methacrylate units, (PEGMA, and DEGMA, respectively); the latter shows cloud point (de ned as the temperature at which the solution turns to cloudy, CP) at around 30 °C, 27,40 while PEGMA presents CP at higher temperatures (~70°C). 40 Notably, this copolymer has been synthesised using Group Transfer Polymerisation (GTP), a polymerisation technique currently used in industry, due to its cost-and time-effectiveness and quantitative yields; 41-43 the latter eliminates the need of intermediate puri cation steps. Therefore, its mass production is certainly feasible. In order to fully harness its potential, an extensive characterisation of both self-assembly and rheological properties in aqueous milieu is necessary. More speci cally, we performed visual tests, dynamic light scattering (DLS), and differential scanning calorimetry (DSC) to investigate its self-assembly and gelling behaviour. Furthermore, we performed rheology in conjunction with small-angle neutron scattering (Rheo-SANS), to scope out the macroscopic rheological behaviour and the self-assembly at the nanoscale at a range of temperatures. We benchmarked our in-house synthesised polymer with Pluronic® F127. To the best of our knowledge, this is the rst Rheo-SANS study on methacrylate ABC triblock terpolymers. The applicability of this novel polymer as an injectable formulation with sodium uorescein was evaluated in comparison with Pluronic F127 in ex vivo experiments on intracameral injections into the bovine eyes.

Results
Design and Thermal Properties in-house Synthesised Polymer In need of new thermoresponsive polymers, we designed and synthesised a triblock terpolymer which is based on a novel combination of methacrylate units. The triblock terpolymer has an ABC linear architecture, which has been shown to provide the best thermogelling properties, i.e. clear sol-gel transition without solubility issues, in the previously studied systems based on PEGMA-BuMA-DMAEMA; 16,19 where DMAEMA stands for 2-(dimethylamino)ethyl methacrylate. This novel polymer consists of two hydrophilic compartments based on PEGMA (A block) and DEGMA (C block), which is also thermoresponsive close to physiological conditions (CP at 30 °C, depending on the MM), and a hydrophobic central block based on BuMA. Thus, its general chemical structure is PEGMA x -b-BuMA y -b-DEGMA z , where x, y, and z are the DPs of PEGMA, BuMA, and DEGMA, respectively, as shown in Fig. 1b), along with the structure of Pluronic® F127. It is noteworthy that even though previous studies investigated polymers consisting of EG-based repeated units, 44,45 this is the rst time that this novel combination of repeated units is reported (patent published 46 ). We believe that the substitution of DMAEMA units with DEGMA units will provide thermoresponsiveness, by avoiding any undesired effect of electrostatic interactions. In addition, the incorporation of BuMA units is bene cial, as it will promote selfassembly, while the incorporation of PEGMA units will balance the hydrophilicity, and thus solubility in aqueous media, while providing well-hydrated bridges between the micelles upon thermoresponse of DEGMA.
To ensure well-de ned structural parameters, i.e. well-de ned MM, and composition, we have implemented the synthesis via GTP, as these properties are crucial for controlling the thermoresponsive properties, i.e. gelation temperature (T gel ) and C gel . 1 The gel permeation chromatography (GPC) traces of the triblock terpolymer ( Supplementary Fig. S1) as well as the proton nuclear magnetic resonance ( 1 H NMR) spectra ( Supplementary Fig. S2) reveal that we have successfully synthesised the triblock terpolymer with narrow MM distribution, indicated by the low dispersity value (Ð<1.2), and controllable MM and composition. The theoretical and experimental structural properties, i.e. MM, Ð, and composition are listed in Table S1.
To investigate the macroscopic differences in aqueous media between PEGMA x -b-BuMA y -b-DEGMA z and Pluronic® F127, we visually inspected the samples across a range of temperatures and concentrations and constructed the relative phase diagrams; the ones in phosphate buffered saline (PBS) are presented in Fig. 1c). The stable gel state, which is de ned visually as the temperatures at which the sample does not ow upon tube inversion, as indicated by the images in Fig. 1c), is of main importance. As extensively reported in the literature, [47][48][49][50] and shown in Fig. 1c) (left panel), the C gel of Pluronic® F127 is 15 w/w%, and it forms gels at room temperature at 20 w/w% and 25 w/w%, which makes handling of the samples challenging, as previously discussed. On the other hand, as visible in Fig. 1c) (right panel), the in-house synthesised polymer, PEGMA x -b-BuMA y -b-DEGMA z , is a runny solution at all concentrations tested at room temperature, which ensures homogeneity and ease in handling and injection. As the temperature is increased, a wide gelation area, shown approximately in black dashed line, is identi ed, with T gel ranging from 32°C to 39°C, depending on the concentration. Most importantly, we report a novel polymer that gels at body temperature (denoted by black dotted line in Fig. 1c) at ve times lower C gel than Pluronic® F127. This is highly advantageous in terms of the cost-effectiveness of the application and low sample viscosity at room temperature upon administration. Interestingly, we observed a controlled decrease in T gel by increasing the concentration from 2 w/w% to 25 w/w%, as opposed to the sudden decrease in T gel for Pluronic® F127. Similar trends are observed in deionised water (see Fig. S3), with the T gel (and C gel for the PEGMA x -b-BuMA y -b-DEGMA z ) being slightly higher compared to the PBS solutions, which is attributed to the absence of ions in solutions. 47 As previously mentioned, the solutions of the PEGMA x -b-BuMA y -b-DEGMA z are runny at room temperature, regardless the concentration, while the solutions of Pluronic® F127 are highly viscous, or gels at room temperature, depending on the concentration. To this end, we evaluated the injectability of their solutions at 15 w/w% in PBS, and it is proven that the injection rate of the polymer solution of PEGMA x -b-BuMA y -b-DEGMA z is at least one order of magnitude higher than to the one of Pluronic® F127, when the same force is applied (see Fig. S4 in Supplementary). Therefore, we demonstrate the ease in administration of the solution of PEGMA x -b-BuMA y -b-DEGMA z , which is highly advantageous especially in the application of injectable gels, as injection of a sample of lower viscosity through narrow needles could be potentially less painful for the patient.
It is well-documented that DSC has been employed to monitor i) the micellisation, i.e. the self-assembly of unimers (free polymer chains) in micelles, the gelation, i.e. the formation of a 3-D network of interconnected micelles, and the cloud point, i.e. the phase separation, of polymer solutions. [50][51][52][53][54][55][56] Therefore, we employed DSC to record the changes in the thermal behaviour of the polymer solutions at 15 w/w% in PBS (C gel for Pluronic® F127), and the results are presented in Fig.2; the repeated thermograms can be found in Fig. S5. As visible, a broad endothermic peak is present on the DSC thermogram of Pluronic® F127 (Fig. 2a)), with an onset temperature (T onset ) at 11.6±0.1 °C, and a maximum temperature (T max ) at 14.5±0.1 °C. This peak is indicative of the micellisation process of Pluronic® F127 caused by the de-hydration of the PG units, and it has been previously reported in the literature. 51,54 The enthalpy of micellisation (ΔH micell. ), calculated by the area of the endothermic peak, is 5.0±0.2 J/g ( equal to 63 kJ/mol). Both the T max and the ΔH micell. for Pluronic® F127 are in a good agreement with previously reported values for similar systems. [50][51][52][53][54] As previously stated, the gelation process is scarcely endothermic (i.e. almost athermal), and thus it has only been observed as a spike on the main peak at concentrations higher than the ones in the present study. 51,54,55 Studies on different systems other than Pluronics® recorded the phase separation at higher concentrations via DSC. 56 On the other hand, the DSC thermogram of PEGMA x -b-BuMA y -b-DEGMA z (Fig. 2 b)) shows no apparent changes up to 45°C, which could be due to the intrinsic micellisation, caused by the incorporation of the BuMA hydrophobic units within the polymer structure.
To con rm this hypothesis, we carried out DLS analysis on solutions of both Pluronic® F127 and PEGMA x -b-BuMA y -b-DEGMA z at 1 w/w% in PBS at 10°C and at room temperature (25°C), and the DLS histograms are shown as insets in Fig. 2. Concerning PEGMA x -b-BuMA y -b-DEGMA z , we observed micelles at both temperatures, as the histograms almost overlap. In contrast, unimers, i.e. free polymer chains, are present in solutions of Pluronic® F127 at 10°C, as this temperature is lower than its micellisation temperature. By increasing the temperature to 25°C, a peak corresponding to micelles appears, the size of which is in agreement with the literature, 53,57,58 in addition to the peak which corresponds to the unimers. DLS at 25°C in deuterated phosphate buffered saline (D 2 O/PBS), which is the solvent used during SANS, as it will be discussed in the following paragraphs, agrees with the results in PBS. The hydrodynamic diameters and the polydispersity (PDI) values, resulted from DLS analysis at 10°C and 25°C are listed in Table S2, and the corresponding DLS histograms by intensity and by number are presented in Figs. S6 and S7, respectively.
Rheological Properties and Self-Assembly behaviour As the two polymers under investigation present clear differences in their macroscopic properties (i.e. T gel and C gel ), in their thermal properties by calorimetry, and in their self-assembly behaviour by DLS, we used a state-of-the-art technique to gain further insights into their nanoscale self-assembly behaviour. Thus, we implemented Rheo-SANS, which is a powerful and non-destructive technique that records the scattering pro les of the polymeric ensembles at the nanoscale and the macroscopic rheological properties simultaneously, at a range of temperatures. In this study, we report and discuss the Rheo-SANS analysis of the polymer solutions at 15 w/w%, as this is the C gel of Pluronic® F127. The solutions were prepared in D 2 O/PBS to achieve good neutron contrast. The temperature ramp pro les of the PEGMA x -b-BuMA y -b-DEGMA z and Pluronic® F127 are presented in Fig.3 a) and b), respectively, while their SANS pro les at selected temperatures are shown in Fig.3 c) and d), respectively. The SANS pro les with the relative t lines at all temperatures investigated can be found in Figs. S8 and S9, while the overlapped SANS pro les at selected temperatures are provided in Fig. S10, for direct comparison.
As corroborated by rheology (Fig. 3 a) and b)), both samples are in the liquid phase at room temperature, while they form gel as the temperature increases; T gel is de ned rheologically as the temperature at which the storage modulus G′ exceeds the loss modulus G′′ (i.e. G′>G′′). 59 Interestingly, both samples are in the gel state at body temperature (T gel, Pluronic® F127 = 28°C, T gel, PEGMAx-b-BuMAy-b-DEGMAz = 32°C), while the gels are destabilised (G′′>G′, T degel ) at 45°C, as expected by the visual tests (see Fig. 1 c) for PBS, and Fig. S11 for D 2 O/PBS). The transition temperatures of PEGMA x -b-BuMA y -b-DEGMA z are in a good agreement with the values determined visually, both in PBS and D 2 O/PBS, while the gelation area of Pluronic® F127 in D 2 O/PBS is wider than in PBS, which is con rmed both visually and rheologically.
As can be seen in Fig. 3 c), and in more detail in Fig. S8 and S9, the in-house synthesised polymer, PEGMA x -b-BuMA y -b-DEGMA z , presents different SANS pro les as the temperature increases, indicating changes in the morphology of its self-assembled structures. We used an elliptical cylinder model with a hardsphere form factor to t the data up to 43°C, while a broad peak model was used to t the data from 45°C; this temperature coincides with the T degel by rheology.
From the SANS data, we inferred that the PEGMA x -b-BuMA y -b-DEGMA z formed micelles shaped as elliptical cylinders, whose best-t values for the radius minor (blue dots) and the length (black dots) are shown in Fig. 4 (top panel, a)) as a function of the temperature. The best-t radius axis ratio and volume fraction as a function of the temperature are shown in Fig. S12. As it is seen, the best-t radius minor decreases from 47 Å to 44 Å, while the best-t axis ratio increases from 1.3 to 1.5, as the temperature is increased from 23°C to 36°C. Interestingly, we observed a clear trend for the length of the cylinder, which increases signi cantly (from 126 Å to 460 Å), within the same temperature range; the tting parameters of the elliptical cylinder above 36°C are not presented in Fig. 4, as the best-t length was outside the limits of the SANS technique (2000 Å).
In order to t the SANS data for PEGMA x -b-BuMA y -b-DEGMA z above the T degel , we used a BroadPeak model, which provided the best-t position of the Bragg peak. 60 This model is a combination of a Lorentzian-peak function and a power law decay and could suggest the presence of a bicontinuous structure 61 above the T degel . As is seen in Fig. 4 (top panel, b)), the obtained Bragg peak position shifted from 0.044 Å -1 at 45°C to 0.059 Å -1 at 55°C. The d-spacing of this peak, calculated as d=2π/Q, is a characteristic distance between the scattering inhomogeneities, 61 and it decreases from d=144.21 Å and d=106.19 Å as the temperature increases from 45°C to 55°C as the sample proceeds from gel syneresis (i.e. disturbance of the gels, attributed to the inhomogeneity in the gel which causes increased internal stress, leading to the exclusion of solvent, rstly reported by Graham in 1864 62, 63 ) to precipitation (i.e. clear separation into two phases -solid phase and liquid phase).
A proposed schematic of the elliptical cylinder structure adopted by PEGMA x -b-BuMA y -b-DEGMA z is also shown in Fig. 4 (top panel, c)), in which the hydrophilic PEGMA blocks, and the hydrophilic and thermoresponsive DEGMA blocks, shown in blue and green respectively, extend from the hydrophobic BuMA core, shown in orange, towards the aqueous environment.
The SANS pro les for Pluronic® F127, presented in Fig. 3 d), S9 and S10, are in a good agreement with the ones previously reported in the literature for concentrated solutions of the same polymer. 50,[64][65][66][67] We used the IGOR software to t the scattering, as it was found to capture the scattering features/peaks best. From the SANS analysis, we observed the presence of globular structures, as indicated by the tted power low value (P~3).
Interestingly, the scattering pro les of the solution of Pluronic® F127 present a series of peaks/shoulders at low Q values, which could be attributed to interparticle interference. 50,65,66 The rst peak, which appears at Q ~ 0.035 Å -1 , is well-distinguishable at all temperatures (see Fig. 3) and it is a characteristic peak of highly concentrated solutions of Pluronic® F127, widely reported in the literature. We observed a shoulder at Q ~ 0.061 Å -1 below 25 °C, at which the sample is in the solution state. At higher temperatures, a clear peak is detected at Q ~ 0.057 Å -1 , as can be seen in Fig. 3 d) middle and right, followed by two shoulders at Q = 0.067 Å -1 and Q = 0.085 Å -1 / 0.094 Å -1 at higher temperatures. These features are also related to the interparticle interference and are associated with the formation of a polymer network. 50,65,66 In addition to the series of peaks/shoulders at low Q values, a broad shoulder is present at Q ~ 0.12 Å -1 , which is due to intraparticle interferences. 65 We obtained a best-t size core for the micellar core of approximately 4.4 nm, similarly to previously reported values. 65 This shoulder remains unchanged over the temperature range tested and it is an indication that the size of the self-assembled structures of Pluronic® F127 is not affected by the temperature, as previously reported. 66 This can also be seen in Fig.   4 (middle panel, a)), which presents the independence of the size of the core as a function of temperature. The suggested globular structure is shown in Fig. 4 (middle panel, b)), in which the wellhydrated PEG corona is shown in grey, and the compact hydrophobic core is illustrated as a black sphere.
It is worth noting that even though the position of the peaks, and thus the size of the globular structure is not affected by temperature, the scattering intensity clearly increases, indicating an increase in the volume fraction of the polymer. Thus, we con rm that the gelation of Pluronic® F127 is caused by an increase in the number of the globular structures, which has also been reported before in the literature. 67 Our extensive Rheo-SANS analysis allowed us to reveal the differences in the nanoscale between PEGMA x -b-BuMA y -b-DEGMA z , and Pluronic® F127. Thus, we conclude that the formation of gel by PEGMA x -b-BuMA y -b-DEGMA z is caused by the growth of the micelles, as indicated by the signi cant increase in their length (Fig. 4). On the other hand, the gelation of Pluronic® F127 is not caused by the change in the micelle size, but by the concentration of the micelles and the close packing, as indicated by the increase in volume fraction. The proposed gelation mechanisms are shown schematically in Fig. 4 (bottom), where the tricomponent system (PEGMA x -b-BuMA y -b-DEGMA z ) is shown in blue, orange, and green (top), while Pluronic® F127, which is a bicomponent system (EG 66 -b-PG 99 -b-EG 66 ), is presented in grey and black (bottom).
In order to evaluate the applicability of novel polymer in drug delivery a series of ex vivo experiments were performed using 15 wt % solutions, containing 1 mg/mL sodium uorescein. In these experiments 15 wt % solution of Pluronic® F127 was used as a positive control (capable of forming gel at physiological temperature) and 15 wt % of Pluronic® F68 was used as a negative control (not capable of forming gel at physiological temperature). Solutions of control polymers were also containing 1 mg/mL sodium uorescein. All these solutions were injected intracamerally into the anterior chamber of freshly excised bovine eyes thermostated at physiological temperature and the spreading of sodium uorescein was monitored visually using video recording (see exemplar videos in Supplementary) and analysed using image analysis (Fig.5). As it was expected, the negative control formulation based on Pluronic® F68 exhibited very quick spreading of sodium uorescein in the anterior chamber due to the absence of gelation. Both PEGMA x -b-BuMA y -b-DEGMA z and Pluronic® F127 formulations exhibited signi cantly slower spreading of sodium uorescein (p<0.05) compared to Pluronic® F68, which is related to their gelation in the anterior chamber. This reduced spreadability of sodium uorescein indicates the applicability of PEGMA x -b-BuMA y -b-DEGMA z for the potential design of injectable delivery systems to the anterior chamber, where longer drug residence will be of great importance. It is interesting to note that the formulation based on PEGMA x -b-BuMA y -b-DEGMA z shows a signi cantly slower spreadability of sodium uorescein compared to Pluronic® F127 (P<0.05). This potentially indicates a further advantage of this polymer compared to Pluronic® F127.
In conclusion, we present a novel thermoresponsive terpolymer, namely PEGMA x -b-BuMA y -b-DEGMA z , that gels at body temperature at a concentration which is ve times lower than the one of the commercially available Pluronic® F127. While PEGMA x -b-BuMA y -b-DEGMA z inherently forms micelles due to the incorporation of permanently hydrophobic block, the micellisation of Pluronic® F127 is temperaturedependent, driven by the thermoresponse of the PG units. We used state-of-the-art characterisation techniques, such as DSC and Rheo-SANS, to probe the self-assembled nanostructures and gain insights into the gelation mechanism of PEGMA x -b-BuMA y -b-DEGMA z . Thus, we conclude that PEGMA x -b-BuMA yb-DEGMA z forms a gel due to the growth of the micelle structures to near cylindrical ensembles. On the other hand, the micelle size and shape adopted by Pluronic® F127 is independent of the temperature, but the volume fraction, and thus the population of globular structures increases as a function of temperature, thus leading to gel formation. The applicability of this novel methacrylate polymer for the use as an injectable intracameral formulation for ocular drug delivery is demonstrated.

Puri cation of Starting Materials
The low MM monomers (BuMA and DEGMA) were puri ed in four steps: i) passing twice through basic alumina to remove the free-radical inhibitor, monomethyl ether hydroquinone, and any acidic components (i.e. methacrylic acid), ii) addition of DPPH to prevent undesired polymerisation, iii) addition of the desiccant CaH 2 and stirring for at least 3h to ensure dry conditions, and iv) vacuum distillation to remove DPPH and CaH 2 /Ca(OH) 2 . The high MM monomer, PEGMA, was puri ed in 2 steps as a solution in THF (50 v/v%) (Steps i) and iii). Direct ltration into the polymerisation ask was performed using PTFE lters to remove the CaH 2 /Ca(OH) 2 . The initiator (MTS) was vacuum distilled prior to use, while the catalyst (TBABB) was previously synthesised following a standard protocol, 68 and puri ed through recrystallisation. The THF used during the polymerisation was puri ed using a solvent puri cation system, which is equipped with an activated alumina column (PureSolvTM Micro 100 Liter solvent puri cation system, purchased from Sigma Aldrich). All the glassware used for the vacuum distillations and polymerisation were dried overnight at 140°C.

Polymer Synthesis and Puri cation
The in-house synthesised triblock terpolymer, PEGMA x -b-BuMA y -b-DEGMA z , was synthesised via sequential GTP. More speci cally, around 10 mg of the catalyst, TBABB, was added in a one-neck round bottom ask, which was then sealed with a septum and purged with argon to ensure complete substitution of the atmosphere by inert argon gas. Subsequently, freshly puri ed THF (59 mL) was syringed into the ask, followed by the addition of the MTS (0.55 mL, 0.5 g, 3 mmol). Three monomer additions were carried out as follows: i) PEGMA solution in THF (17 mL, 8.9g, 30 mmol), ii) BuMA (8.7 mL, 7.8 g, 55 mmol), and iii) DEGMA (5.5 mL, 5.5g, 29 mmol). The temperature changes after each monomer addition was monitored and recorded as follows: i) from 26.9 °C to 36.9 °C, ii) from 30.9 °C to 40.8 °C, and iii) from 32.9 °C to 37.9 °C, respectively. After each polymerisation step was completed, two aliquots (0.1 mL each) were withdrawn for GPC and 1 H NMR analysis. The polymer was puri ed and recovered via precipitation in cold n-hexane. The puri cation was completed by vacuum-drying at room temperature.

Gel Permeation Chromatography (GPC)
The successful sequential GTP was con rmed via GPC technique, which provides the MM values and the distributions of the nal triblock terpolymer and its linear precursors. For this analysis, an Agilent SEC GPC system was used, which was purchased from Agilent technologies UK Ltd., Shropshire, UK. This system is equipped with an Agilent guard column (PL1110-1520, PLgel Mixed, dimensions [mm]: 50 × 7. The experimental MM values (resulted by GPC analysis) were compared to the theoretical ones (MM theoretical ), which were calculated using by summing the product of the MM of each monomer (MM i ) by its degree of polymerisation (DP i ). An additional MM of 100 g mol -1 was added to the result to take into consideration the part of the initiator, MTS, that stays on the polymer chain after the GTP is completed. Simply stated, the theoretical MM values can be calculated using the Equation 1 below:

Nuclear Magnetic Resonance (NMR) Spectroscopy
The experimental compositions of the nal terpolymer and its linear precursors were con rmed by 1 H NMR spectroscopy. The NMR samples were prepared by dissolving the dried polymer samples in deuterated chloroform. The solutions were analysed by using a a 400-MHz Avance Bruker NMR spectrometer from Bruker (Bruker, UK Ltd., Coventry, UK). The data were analysed using a MestReNova software (Version: 11.0.0-17609 ©2016 Mestrelab Research S.L.).
The prove the successful GTP, the experimental weight percentages were calculated by analysing the NMR results and they were compared to the targeted ones. The following peaks were used for the analysis: i) for PEGMA and DEGMA, which are both EG-based methacrylate units, the distinctive peak of their methoxy group (CH 3 O-) at 3.35 ppm was used, and ii) for BuMA, the peak of the methylene group closest to the ester was used (CH 3 CH 2 CH 2 CH 2 O-), which appears at 3.9 ppm.

Dynamic Light Scattering (DLS)
The diluted aqueous solutions (1 w/w%) of both PEGMA x -b-BuMA y -b-DEGMA z and the Pluronic® F127 were analysed using a Zetasizer Nano ZSP from Malvern Instruments Ltd. (Malvern, UK). The polymer solutions were tested under the following solvents: i) PBS, and ii) D 2 O/PBS. The polymer solutions were tested without any further processing, i.e. ltration, to ensure that a direct comparison between the different techniques of determining the size can be made. The DLS experiments were performed at room temperature (25 °C), while in the case of PBS solutions, the samples were also tested at 10 °C. In all the cases, the scattered light was collected at a backscatter angle of 173°. Each sample was analysed three times and the results reported are the mean hydrodynamic diameters (d h ) that corresponds to the maximum of the peak by intensity, and by number. The data were analysed using a Zetasizer software (version 7.11) from Malvern Polyanalytical.
Differential Scanning Calorimetry (DSC) 15 w/w% solutions in PBS of PEGMA x -b-BuMA y -b-DEGMA z and Pluronic® F127 were prepared three times (three polymer solutions for each polymer). Each sample was analysed using DSC three times (nine results per polymer). For the analysis, the samples were placed in T-zero® style hermetic aluminium pans (purchased from Thermal Instruments Ltd., UK). The DSC was performed using the DSC Q2000 (TA Instruments, UK) at a heating rate of 5°C/min between 2 and 60°C under nitrogen atmosphere. The obtained data were analysed using Universal Analysis 2000 software (TA Instruments Version 4.5A).

Rheology-Small Angle Neutron Scattering (Rheo-SANS)
The polymer solutions at 15 w/w% D 2 O/PBS of both PEGMA x -b-BuMA y -b-DEGMA z and Pluronic® F127 were investigated by Rheo-SANS. More speci cally, the self-assembly as well as the rheological properties of the samples were tested at a range of temperatures. This experiment was performed at the time of ight SANS instrument ZOOM at the ISIS pulsed neutron source at the Rutherford Appleton Laboratory (Didcot, UK), using source to sample and sample to detector distance of 4m, and wavelengths 1. 5 -16.5 Å. For this experiment, an Anton Paar rheometer (Physica MCR501) equipped with a special Searle-Couette (rotating stator) measuring geometry, manufactured in Grade-V Titanium. The temperature ramp measurements were performed at constant strain (γ = 1%), and constant angular frequency (ω = 1 rads -1 ), which are within the linear viscoelastic area (see Fig. S13). Data were reduced using MantidPlot, 69 and the resulted SANS curves were tted using the following software: i) SasView software (version 5.0) for the novel polymer and ii) IGOR software for Pluronic® F127. 70 Ex vivo experiments with intracameral injections Bovine eyes were delivered from a local abattoir within three hours. Appropriate randomisation in experiments with intracameral injections was achieved by labelling each eye with a number and placing another set of labels with the same gures in the non-transparent black bag. The randomisation was done by means of pulling labels out of the bag randomly.
15 wt% polymer solutions of PEGMA x -b-BuMA y -b-DEGMA z , Pluronic® F127 and Pluronic® F68 were prepared by dissolving the required amounts of each polymer in 1 mg/mL sodium uorescein in PBS solution and stirred overnight at 4°C to form clear solutions. The PEGMA x -b-BuMA y -b-DEGMA z solution in PBS containing 1 mg/mL sodium uorescein was kept in an incubator at 33 °C for 20 minutes prior to the intracameral injections. The solution of Pluronic® F127 in PBS containing 1 mg/mL sodium uorescein was stored at room temperature for 20 minutes before the injection. Solution of Pluronic® F68 in PBS containing 1mg/mL sodium uorescein was kept in the fridge at 4 °C for 20 minutes prior to the injection.
All nine ex-vivo bovine eyeballs were kept at 37 °C (human body temperature) for at least 40 minutes prior to the injection in the anterior chamber (0.05 mL; 21-gauge needle, 1mL syringe). Three runs for each polymer solution were performed within the incubator at 37 °C. All the experiments were recorded for 3.5 minutes using iPhone XS. All experiments were performed in triplicate and data were analysed using GraphPad Prism 8.0.2 software. where p < 0.05 was used as the statistical signi cance criterion. The signi cance of the calculated mean values ± standard deviations was assessed using one-way analysis of variance (ANOVA) followed by Bonferroni post hoc test, where p < 0.05 was selected as the statistical signi cance criterion.
Then, the extent of uorophore spreading in the anterior chamber was evaluated using ImageJ software (version 1.50i) at different time points (0, 1, 1.5, 2, 2.5 and 3 minutes) starting from the moment of the complete pull of the needle out of the anterior chamber for each run of each polymer solution tested.

Injectability experiments
The injectability of the solutions of PEGMA x -b-BuMA y -b-DEGMA z and Pluronic® F127 at 15 w/w% in PBS was evaluated at room temperature. To accomplish this, a 2 mL gas-tight glass syringe (Cadence Science, Inc.) containing the solution, tted with hypodermic needles, was placed vertically, as shown schematically in Fig. S4 a). The following needles (diameter x length, mm) were used: 23G 1'' (0.6 x 25, ne-ject®), 24G 1'' (0.55 x 25, ne-ject®), 25G 1'' (0.5x 25, ne-ject®), 26G 1'' (0.45 x 25, ne-ject®), and 27G 3/4'' (0.4 x 20 henke-ject®). Two different weights were placed on top of the needle, 200 g and 500 g, with a resulting force applied for the injection equal to 2.0 N and 4.9 N, respectively. The time of injection of 0.5 mL of solution was measured and the injection rate (mL/min) was calculated. Each experiment was repeated six times and the standard error was calculated.