Using Stereochemistry to Control Mechanical Properties in Thiol–Yne Click‐Hydrogels

Abstract The stereochemistry of polymers has a profound impact on their mechanical properties. While this has been observed in thermoplastics, studies on how stereochemistry affects the bulk properties of swollen networks, such as hydrogels, are limited. Typically, changing the stiffness of a hydrogel is achieved at the cost of changing another parameter, that in turn affects the physical properties of the material and ultimately influences the cellular response. Herein, we report that by manipulating the stereochemistry of a double bond, formed in situ during gelation, materials with diverse mechanical properties but comparable physical properties can be obtained. Click‐hydrogels that possess a high % trans content are stiffer than their high % cis analogues by almost a factor of 3. Human mesenchymal stem cells acted as a substrate stiffness cell reporter demonstrating the potential of these platforms to study mechanotransduction without the influence of other external factors.


PEG 42A42S gels preparation and transition from organogels to hydrogels
Thiol-yne organogels were prepared with a solids content of 10 wt% by mixing solutions containing a 1:1 molar ratio of alkyne to thiol polymer precursors at ambient temperature (23 °C) using different solvent mixtures to tune the final stereochemistry of the cross-linked network. Specifically, three solvents (i.e. CHCl3 (non-polar), acetone (moderate polarity), and H2O (polar)) were mixed in different ratios, while triethylamine acted as the catalyst for the thiol-yne click chemistry reaction. The amount of trimethylamine added to each gelation solution, which depended on the solvent polarity, was adjusted so the gelation time for all the gels was ca. 1 minute regardless of the final cis:trans ratio. After mixing the precursors, the organogels cured at ambient temperature (23 °C) for 1 hour to ensure the reaction had gone to completion.
In a typical procedure, stock solutions of the required solvent mixture were prepared with NEt3 (in the range between 0.7 and 34 µL mL -1 ). Then, 42S (20.8 mg) and 42A (19.5 mg) were separately dissolved in 200 µL of the stock solution. Subsequently, the PEG alkyne solution was added to the PEG thiol solution, and the solution was mixed for 5 seconds in a vortex mixer; the solutions were let to gel for 1 h. After such amount of time, the gels were processed for further characterization. Only for FT-IR and Gel Fraction (GF) characterization were the gels used as organogels, thus without solvent exchange. For all the remaining tests, the thiol-yne organogels were converted into hydrogels. To transition from an organogel to a hydrogel, 42A42S gels were immersed in acetone for 5 d with frequent solvent changes. Acetone was chosen since it is miscible with both CHCl3 and H2O, as well as the NEt3 catalyst. Minimal hydrolysis occurred during this washing step. After 5 d, water was introduced gradually into the network over 2 d.

Determination of the cis:trans ratio of 42A42S gels by 1 H NMR spectroscopy
A small molecule study was employed to elucidate the relative stereochemistry of the 42A42S network using 1 H NMR spectroscopy. As a general procedure, the nucleophilic thiol-yne reaction between monofunctionalised alkyne-and thiol-PEG precursors was carried out as follow: a solution of PEG550(SH), (53 mg) in the required solvent system (0.5 mL) was added dropwise to a stirring solution of PEG550(C≡CH) (50 mg in 0.5 mL). The solution was stirred at ambient temperature (23 °C) for 1 h. After such amount of time, the solvent was removed in vacuo and the product was characterized by 1 H NMR spectroscopy.
Model PEG thiol-yne reaction in CHCl3: 1  As stated in the main text, from herein onwards, % cis content of the gels are referred to from the values determined by FT-IR spectroscopy.

Gel fraction (GF)
To determine the gel fraction (GF) of the organogels, samples were lyophilized and their weights (Wg) recorded. The gels were then allowed to swell in deionized water for 3 d, with frequent water changes to extract any unreacted PEG precursors. After that period, the gels were lyophilized, and their weights (Wr) were recorded again. All measurements were repeated in triplicate. GF is expressed as: Gel Fraction (%) = W r W g × 100 % (1)

Equilibrium water content (EWC)
To determine the equilibrium water content (EWC), after the solvent exchange process, hydrogels were allowed to swell in PBS (pH 7.4) for 24 h, and their weight was recorded (Ws). Hydrogels were then lyophilized, and the weights recorded again (Wd). All measurements were repeated in triplicate. EWC is expressed as: (2)

Swelling and degradation studies
Hydrogels (n = 3) were prepared as described above (section 1.5) and then placed in the corresponding solution (i.e. PBS or cell culture medium -MEM-α) at 37 °C in an orbital shakerincubator (Model ES-20, Grant Instruments (Cambridge) Ltd.) with a shaking speed of 60 rpm. The swelling solution was replaced regularly to remove any degradation products and to prevent the build-up of solute concentration. At specific time points, the hydrogels were removed, gently blotted dry, and their weight was recorded. Hence, the swelling was monitored by the percentage of weight of the hydrogel at each time point (Wt) compared to the weight before submersion (W0), which is defined as:

Mesh size determination
Flory-Rehner calculations were used to determine each thiol-yne PEG hydrogel mesh size. [2] All measurements were run in triplicate, and the procedure followed was reported in detail in a previous study. [1a] Fluorescence Recovery after Photobleaching (FRAP) is a very simple technique used to study dynamics, like diffusion rates, at the microscopic level. Here, we exploit this technique to assess the mass transport characteristic of the prepared hydrogels. [3] To that end, FTIC-dextran molecules, with molecular weights of 5 and 10 kDa, were used as macromolecular model compounds. Their hydrodynamic radii, rH, which were calculated according to previous literature, [3a, 4] was determined to be of 1.35 and 1.90 nm, respectively. Hydrogels with a % cis content of 10%, 51%, and 100% were selected for this test (n = 3 for each % cis content) and immersed in aqueous solutions at 0.3 mg mL -1 of the respective FITC-dextran for 36 h, which allowed for FTIC-dextran molecules to diffuse into the hydrogel network.
The FRAP experiments, which were performed on an Olympus Confocal Laser Scanning Microscope (Fluoview FV3000; software FV31S-SW version 2.3.1.163), were based on the procedure described by Brandl et al. [3a] Briefly, all bleaching experiments were performed using the 488 nm-line. After the area of interest was brought into focus (4.0X objective Lens Mag.), a time-series of digital images with a resolution of 1024 × 1024 pixel was recorded. After the acquisition of six pre-bleach images, a uniform disk with a diameter of 32.24 μm was bleached at maximum laser intensity (100% transmission). The bleaching phase happened quickly enough to avoid fluorescence recovery during bleaching. Immediately after bleaching, a stack of at least 30 images (between 30 and 80) was acquired in order to measure the recovery of fluorescence inside the bleached area. Several areas were analysed throughout the hydrogels. The experimental recovery curves were extracted from the images by the software (OLYMPUS cellSens Dimension Desktop 1.18). Indeed, curves that plotted the mean fluorescence intensities inside the bleached region, Ifrap(t), and inside a reference region, Iref(t), versus time were obtained from the stacked images (at least 12 curves for each condition). Next, Ifrap(t) was normalized to the prebleach intensity, Ifrap(pre), and corrected for any bleaching effects that might have occurred during image acquisition: is the normalized fluorescence intensity inside the bleached region, and Iref(pre) is the fluorescence intensity inside the reference region before bleaching. In the following step, f(t) was further normalized to the full scale using: where f(0) is the normalized fluorescence intensity immediately after bleaching, and f(pre) the normalized fluorescence intensity before bleaching. Finally, the characteristic diffusion time τD and the mobile fraction k were determined by a least-squares fit of the following expression to the experimental recovery curve: where I0 and I1 are the modified Bessel functions of the first kind of zero and first order. [4][5] The diffusion coefficient D was then calculated by: D = w 2 /τD, where w is the radius of the bleached spot. The fitting was performed using RStudio (version 1.1.453).

Differential scanning calorimetry (DSC)
The freshly prepared gels were placed in vacuo (ca. 100 mTorr) for 24 h to remove solvent and then sliced into thin discs (ca. 10 mg). The thermal characteristics of the dried gels were determined using differential scanning calorimetry (DSC) (STARe system DSC3, Mettler Toledo) from −100 -100 °C at a heating rate of 10 K min −1 for two heating/cooling cycles unless otherwise specified. The glass transition temperature (Tg), cold crystallization temperature (Tcc), melting temperature (Tm), and enthalpy of melting (ΔHm) were determined from analysis of the second heating cycle to ensure adequate removal of solvent.

Mechanical characterization: uniaxial compressive tests
All uniaxial compressive testing was performed on a M100-1CT Testometric fitted with a load cell of 1 kN. Hydrogel samples, which were prepared as described above (n = 8-10), were tested after the solvent exchange process. A preload force of 0.1 N was set, and each test was carried out at a compression velocity of 5 mm min -1 . Each gel was subjected to 98% strain to determine the ultimate compressive stress and strain. Data was analysed using winTest mechanical analysis software.

Rheological characterization
All rheology was performed on an Anton Parr MCR 302 rheometer fitted with a parallel plate configuration (diameter of 8 mm) and a sandblasted bottom plate to reduce the slippage of the sample. Amplitude sweeps were conducted applying a constant frequency of 10 rad s -1 , while the strain was ramped logarithmically from 0.01% to 10% keeping the normal force constant at 0.04 N. All measurements were repeated in triplicate, and representative charts are shown.

Cryogenic Scanning Electron Microscopy
Cryogenic scanning electron microscopy (cryo-SEM) was performed on ZEISS SUPRA 55-VP equipped with cold stage and sample preparation chamber. Hydrogel samples were carefully placed on a stub and frozen in liquid nitrogen (-195 °C) under vacuum. The stub was then transferred to the cold stage (-125 °C) of the preparation chamber. There, the frozen sample was carefully surface fractured and sublimated at -95 °C for 15 min to reveal the cross-sectional surface. The temperature was then brought down to -125 °C, and the sample was sputter coated with platinum before being transferred under vacuum into the main SEM chamber (kept at -186 °C for imaging). The accelerating voltage was set at 2 kV to avoid burning the sample.

Biocompatibility studies: cytotoxicity of 42A42S hydrogels and degradation products
To confirm that our click-hydrogels and their degradation products were non-cytotoxic, cell viability tests on MC3T3 (murine pre-osteoblasts) and Y201 hTERT-immortalised human clonal mesenchymal stem cell (MSCs) were undertaken. MC3T3 cells were obtained from Public Health England and were cultured in 175 cm 2 tissue culture flasks using MEM alpha medium (Gibco), as advised by the supplier, with addition of 10% FBS and 1% pen/strep, at 37 °C, 5% CO2. Similarly, Y201 MSCs [6] , which were a kind gift from Prof Paul Genever (University of York) were cultured in MEM-α medium supplemented with 10% v/v FCS, 1% penicillin/streptomycin, 10 µM asc-2-phos, and 5 mL Glutamax.
Hydrogels (n = 6-12) were prepared as described above and then placed in a fibronectin solution (25 μg ml -1 in PBS) overnight (Fibronectin Bovine Protein, Plasma -Cat. Number 33010018 -Invitrogen by Thermo Fischer Scientific -Lot 198036). Specifically, hydrogels with a % cis content of 10%, 51%, and 100% were selected for this test, while hydrogels made using PBS as solvent were also included as control. After such amount of time, hydrogels were carefully rinsed with PBS. Cells were seeded on a 2D configuration on the top surface of the click-hydrogels (127k cells/cm 2 ) and left to adhere for 3 h before adding 1 mL of cell culture media. Cells proliferated for 7 d, and culture media was replaced with fresh one every two d during the whole incubation period. Cell viability was measured at specific time points using PrestoBlue® Cell Viability Reagent (Invitrogen TM ) or Alamar Blue® (Invitrogen TM ) for MC3T3 and Y201 cells, respectively, following the supplier's protocol. Cell viability was also assessed using Live/Dead TM Viability/Cytotoxicity Kit (Invitrogen TM ), which includes calcein AM for live cells (λEx. = 495, λEm. = 515) and ethidium homodimer for dead cells (λEx. = 528, λEm. = 617). The staining solution was prepared by dissolving calcein AM (0.5 μL mL -1 ) and ethidium homodimer (2 μL mL -1 ) in PBS and incubated with the samples for 30 min. After washing in PBS, samples were imaged using an Olympus Confocal Laser Scanning Microscope (Fluoview FV3000) and excited using the 488 and 561 nm lasers. Images were processed using CellSens (Olympus) and ImageJ software (1.52i).

Statistical Analysis
One-way analysis of data was used for analyzing the data when appropriate. The data represents the mean and standard deviation. Quantified data were categorized as significantly different when p < 0.05 using Tukey's [7] multiple comparison test. Figure S1. a) PEG precursors synthesized to prepare thiol-yne click-hydrolgels. b) Generic nucleophilic thiol-yne reaction between 42A and 42S and schematic of the resulting hydrogel network.
We determined the cis:trans ratio of the obtained click-networks by means of FT-IR spectroscopy ( Figure S2). For each organogel, at least 10 measurements were performed on different areas of the disk. Additionally, FT-IR spectroscopy gave an insight into the reaction conversion: the alkyne stretching signal (ca. 2100 cm -1 ) disappeared after reacting with the thiol-moiety. The % cis content in the organogel was determined by normalising the area of the bending signal ascribed to the cis alkene (802 cm -1 ) with the area of the stretching signal assigned to C-H bond (2866 cm -1 ). When using H2O as the gelation solvent, the resulting cis content was 94%. Therefore, relative to this value, the cis content for the organogels was determined by dividing the normalised peak area by the area of the material synthesised using aqueous conditions. As a result, the careful optimization of the solvent conditions during gelation yielded a series of thiol-yne PEG organogels with a range of different cis:trans ratio values (i.e. 10, 23, 51 and 82% cis content). Interestingly, even in highly non-polar conditions (i.e. CHCl3), a cis content of 10% was still present. Additionally, a model compound study was used to elucídate the relative stereochemistry of the 42A42S network by 1 H NMR spectroscopy. To that end, a methyl ether PEG (MeO-PEG, 550 g mol -1 ) was monofunctionalised with an alkyne or a thiol end group by Fisher esterification, which resulted in cross-linking precursors with high conversion (> 92%) as determined by 1 H NMR spectroscopy and SEC analysis (Figures S3 and S4). The alkyne-and thiol-functionalised MeO-PEG precursors were dissolved in the appropriate solvent system, and triethylamine was added. The thiol-yne reaction was then sealed and stirred for 1 h before concentrating under vacuum. After the solvent was removed, the product was analysed through 1 H NMR spectroscopy. The signals corresponding to the cis and trans protons, which are distinguishable and well resolved in the 1 H NMR spectrum ( Figure S5 and Figure S6), were assigned using the three bond coupling constants ( 3 JHH). Indeed, these constants differ depending on the stereochemistry of the alkene bond: 3 JHH values for the trans configuration range between 11-19 Hz, whereas 3 JHH values for the cis range between 5-14 Hz. Hence, for the vinyl thioether compound, the two doublets at δ = 6.95 and 5.65 ppm were assigned to the cis isomer (i.e. 3 JHH = 9 Hz), while the signals at δ = 7.42 and 5.55 ppm corresponded to the trans isomer since the coupling constant was greater (i.e. 3 JHH = 15 Hz), and the relative cis:trans ratio was calculated accordingly. In good agreement with previous studies, the increase in polarity of the reaction solvent led to a higher % cis content. [8] Overall, even though the % cis content values determined by FT-IR and 1 H NMR spectroscopy exhibit some discrepancies, probably arising from experimental differences between both techniques, our approach produced 42A42S click-hydrogels with controlled stereochemistry by fine adjustment of the gelation conditions only, thus maintaining the polymer concentration, as well as the cross-linking density, constant throughout the systems. From herein onwards, we will identify the % cis content of the hydrogels referring to the values determined by FT-IR spectroscopy.  Table S1 (CDCl3, 300 MHz, 298 K; (CD3)2CO, 300 MHz, 298 K for H2O system). % cis content values as determined by 1 H NMR spectroscopy.      Normalized ΔH m (J/g) Figure S13. Mesh size determination by FRAP: a) Images of a typical FRAP experiment (100% cis hydrogels, FITC-dextran 10 kDa). Pre-bleach image shows the sample before bleaching (left image); a uniform disk is bleached during the Bleach step (middle image); finally, the laser intensity is reduced, and fluorescence recovery is recorded over time (FRAP step, right image). b-c) Representative recovery curves obtained after the mathematical processing for each % cis content using FITC-dextran with 10 kDa (b) and 5 kDa (c). Table S2. FRAP data determined for FTIC-dextran incoporated into hydrogels. Mesh size is deduced from the mass transport data. To the experimental data shown in Figure S13b and c, a least-squares fit was applied to obtain the mobile fraction (k) and the diffusion coefficient (D). If FTIC-dextran displays a mobile fraction (k) > 55%, it is assumed that the mesh size is similar to the molecule diameter.  Figure S15: Schematic of network formation for cis versus trans bonds. a) Cis bond, functional groups brought into closer proximity, increasing the probability of loop defect formation (i.e. intramolecular reaction), decreasing network formation, resulting in softer click-hydrogels. b) Trans bond, functional groups further apart, increasing the probability that a bond will be formed with another PEG molecule (i.e. intermolecular reaction), increasing network formation, resulting in stiffer click-hydrogels.  Figure S17. Cytocompatibility of of click-hydrogels: a) Cytotoxicity of the degradation products released from hydrogels prepared with 10%, 51%, and 100% cis content, as well as PBS, that had been immersed in cell culture media at 37 °C for 24 h and 72 h. MC3T3 cells were incubated with that media and fresh media (control), and viability (in % relative to control) was determined at time point 96 h. Greek letters on the bars refer to significant differences (p-value <0.05): α vs 10% cis content at 24 h; β vs 51% cis content at 72 h. b) Cytotoxicity of the degradation products released from hydrogels prepared with 10% and 100% cis content, as well as PBS, that had been immersed in cell culture media at 37 °C for 24 h and 48 h. Y201 MCS were incubated with that media and fresh media (control), and viability (in % relative to control) was determined at time point 96 h. Greek letters on the bars refer to significant differences (p-value <0.  Figure S17. Morphometric data of Y201 MSCs seeded on stereocontrolled thiol-yne PEG hydrogels. Greek letters refer to significant differences (p-value <0.05): α vs 10%; β vs 51%; and γ vs 100% cis content.