A bioinspired synthetic soft hydrogel for the treatment of dry eye

Abstract Natural soft hydrogels are unique elastic soft materials utilized by living organisms for protecting delicate tissues. Under a theoretical framework derived from the Blob model, we chemically crosslinked high molecular weight hyaluronic acid at a concentration close to its overlap concentration (c*), and created synthetic soft hydrogels that exhibited unique rheological properties similar to a natural soft hydrogel: being dominantly elastic under low shear stress while being viscous when the stress is above a small threshold. We explored a potential application of the hyaluronic acid‐based soft hydrogel as a long‐acting ocular surface lubricant and evaluated its therapeutic effects for dry eye. The soft hydrogel was found to be biocompatible after topical instillation on experimental animals' and companion dogs' eyes. In a canine clinical study, twice‐a‐day ocular instillation of the soft hydrogel in combination with cyclosporine for 1 month improved the clinical signs in more than 65% of dog patients previously unresponsive to cyclosporine treatment.


| INTRODUCTION
Soft hydrogel is a liquid-like solid 1 commonly found in animals and plants as a soft protectant. These soft hydrogels have extraordinary properties that distinguish them from other materials: the mechanical strength of soft hydrogel is so weak that it could not stand its own weight; the gels are elastic at rest but flowable; and they could retain the elasticity after a history of flow. For example, the mucus in human respiratory track is a soft hydrogel that functions as a protective layer. 2 When the mucus is at rest, it coats and protects the delicate airway surface. During coughing, it easily flows without damaging the underlying tissue. When coughing stops, the protective function is restored. Changes in the mechanical property of this soft hydrogel are associated with lung diseases. [2][3][4] Other examples of the soft hydrogel include the vitreous humor in the eye 5 and egg white, 6 where soft hydrogels provide mechanical support and protection of the most delicate tissues in living organisms including the retina and the embryo.
In rheological terms, soft hydrogel is characterized by an exceptionally low storage modulus (G 0 ) value (on the order of 0.01-1 Pa) at low shear frequency oscillation while its value is still higher than the loss modulus (G 00 ). The gel deforms elastically under a very low stress (on the order of 1 Pa or lower), but deforms viscously when the stress is above a small threshold. Thus, the soft hydrogel is not a viscoelastic polymer solution (or a "structured liquid" 1 ), which has G 0 higher than G 00 only at high shear frequency oscillation, and deforms viscously regardless of the applied stress. Yu Yu and Derek Wai Yee Chow contributed equally to this study.
Since nature has designed soft hydrogel as a protective barrier for various biological tissues, a synthetic soft hydrogel would be of great interest and potential in medicine as a protectant for delicate tissues. One potential is in the treatment of dry eye, which is one of the most common eye diseases affecting 20-30% population globally. 7 Corneal epithelium damage and lid wiper epitheliopathy caused by precorneal friction are one of the core mechanism of dry eye. 8 However, previous effort in making hydrogel for dry eye treatment are not optimum. For example, contact lens has been suggested as a drug delivery device for dry eye treatment 9 ; however, contact lens wear could induce dry eye 10 because of the high modulus of the material. 11 Synthetic soft hydrogels have also been developed. [12][13][14] However, the modulus is on the order of 10-100 Pa, which is too high compares to natural soft hydrogel and not applicable for precorneal application. Thiolated polymers based hydrogels have been used for dry eye treatment; however, the modulus was either too high 12 or the difference in G 0 and G 00 is not significant, 15 which indicated that material is more fluidic than elastic.
Here, we present a theoretical framework and material platform for making synthetic soft hydrogels, and characterize the unique rheological properties of soft hydrogels so prepared. The superior effect of the synthetic soft hydrogel for dry eye treatment was demonstrated to illustrate for its biomedical potential.

| Theoretical consideration
Viscous solution ("structured liquid") that are being widely used in biomedical application are usually composed of polymers dissolved at high concentration (or at the semi-dilute regime, Figure 1). The polymer chains of viscous solution are only weakly associated at the entanglement points, and the chain slips and the bulk material deforms viscously when a stress is applied. In contrast, the natural soft hydrogels are composed of polymers that are chemically or physically crosslinked at an extremely low concentration and crosslinking density. 17 We are inspired by this natural phenomenon and reasoned that a synthetic soft hydrogel can be made when the polymer is crosslinked at a very low crosslinking density ( Figure 1).
According to the Blob model, for hydrogel made by crosslinking of polymers, the maximum density of crosslinks is determined by the density of polymer entanglement points, which is determined by the concentration of a given polymer. 18 For this reason, the lowest possible number of crosslinks appears at the threshold concentration when the polymer chains start to entangle, that is, near the overlap concentration c* for a given polymer solution ( Figure 1). One can roughly relate c* to the properties of the polymer by Equations (1) and (2) 16 : F I G U R E 1 Illustration of the difference between a soft hydrogel, a viscous solution, and a conventional hydrogel. The orange dots represent crosslinks, the threads of different color represent different polymer chains. Center: a polymer solution where the polymer concentration is at the overlap concentration (c*) where the polymer chains are just dense enough to touch each other. The dotted circles represent the average space (Blob size) one polymer occupied. To the right: with increasing polymer concentration above c*, polymer chains become increasingly entangled and the solution turns viscous. 16 Conventional hydrogels are made by crosslinking the polymers are such condition. To the left: the chains of a polymer solution at a concentration about c* are crosslinked to form a hydrogel of the lowest possible crosslinking density for such polymer where M is the molecular weight of the polymer, N A is the Avogadro's number, R g is the radius of gyration of a polymer chain, and [η] is the intrinsic viscosity of the polymer. And since [η] can be measured experimentally by simple techniques and has been tabulated for many polymers, c* can be easily predicted. 18 From this analysis, it is clear that given the same molecular weight, polymers having a more extended configuration will achieve lower c*. For polymer of the same species, lower c* can be achieved by higher molecular weight. A chemical crosslink is preferred over a physical crosslink because the physical crosslink could be unstable under extremely low crosslinking density. In comparison, a conventional hydrogel usually composes of crosslinked polymers at semidilute concentration that is much higher than the overlap concentration (c ) c*, Figure 1).

| Making synthetic soft hydrogel from chemically crosslinked hyaluronic acid
Based on the theoretical consideration, we synthesized soft hydrogels by chemically crosslinking a polymer at a concentration closed to c*.
We identified hyaluronic acid (HA) to be a suitable material for forming soft hydrogel, both as an illustration of the principle and for the application in dry eye treatment. HA is a naturally existing polymer that is found in soft hydrogels in the body 19,20 and on ocular surface. [21][22][23] The configuration of HA in aqueous solution is one of the most extended among polymers that have also been used for biomedical applications (Table S1).
HAs of molecular weight ranging from 120 kDa to 2.6 MDa were modified with vinylsulfone (VS) and thiol (SH) groups with a certain degree of modification (DM) by following our reported modification methods. 18,[24][25][26] The modified HA-VS and HA-SH were tested for the ability to form soft hydrogels at concentrations above and below c*.
As expected, gelation happened only when polymer concentration was above c* ( Table 1). The storage modulus (G 0 ) of the gel is dependent on the concentration of the polymer. Since polymers with higher MW have lower c*, gels formed by HA of 2.6 MDa near c* give the lowest G' ( Table 1). The storage modulus (G 0 ), despite having a very small value on the order of millipascal, is higher than the loss modulus (G 00 ) at low frequency of oscillation, indicating that the solid property is dominant over the liquid property (Table 1 and Figure S4), which is characteristic of a soft hydrogel.

| Elastic (solid) behavior of soft hydrogel
Soft hydrogels with G 0 ranging from 0.05 to 0.5 Pa (prepared using 2.6 M HA as detailed in Table 2) were further investigated for their elastic behavior. All hydrogels had higher G 0 compared to G 00 at the strain and frequency range tested (Figure 2a, Notes: The storage modulus (G 0 ) and loss modulus (G 00 ) were measured by rheometer at 5% strain and 5 rad/s. c* is the overlap concentration of the polymer estimated from [η], c is the concentration of the given polymer precursors. The degree of modification was 28 ± 1% for all polymers. Measurements were triplicated for each formulation (mean ± SD).  (Figure 2c). For gel A1, the strain remained almost constant over time at each stress level, indicating the material responds by deformation rather than flow. Gel A2 behaved similarly to A1 except that a larger strain was found for each stress, and a more significant creep at the larger stress. Gel A3, despite its very low G 0 ($0.05 Pa), behaved as a viscoelastic solid at the lower two stress levels. Another proof of elasticity was shown by the repeats of small magnitude oscillation in strain at the beginning of the test ( Figure S5). This phenomenon, termed inertio-elastic ringing, 27 was a strong indication for elasticity and was observed in all the soft hydrogel tested but not in a viscous solution.

| Viscous (liquid) behavior of soft hydrogel
We next evaluated if the synthetic soft hydrogel could flow (viscous deformation) when the stress was above a threshold level. We first

| Behavior of soft hydrogel after large deformation
The previous sections showed that the synthetic soft hydrogels were able to undergo both elastic and viscous deformation, we went on to evaluate if the elasticity could be retained after the viscous deformation.
The stress response of the soft hydrogel was measured in a continuous shear test at a constant shear rate of 1/s for 100 s (to attain 10,000% strain). The test was repeated 10 times with a 10-s-pause between tests. The total strain applied to the hydrogels after 10 repeats was 100,000%. Surprisingly, for all the three hydrogels tested, the stress response curves were almost identical for all the repeats: there was an initial elastic deformation until about 1500% strain, followed by a viscous deformation at almost identical stress value (Figure 2f and Figure S6).
This experiment showed that the soft hydrogels were able to repeatedly switch from elastic to viscous deformation.

| In vivo study in laboratory animals
The biocompatibility of the hydrogel was evaluated in rats and rabbits for single and repeated instillations ( Figure S7). The eyes were inspected for changes in gross appearance and signs of infection and discomfort, such as swelling, hyperemia, changes in corneal clarity and mucoid discharge. In all the studies, the cornea remained clear and the conjunctiva showed no signs of inflammation. Corneal staining of rabbit eyes was graded 0 immediately and 2 weeks post instillation, suggesting that the soft gels did not cause any damages to the corneal epithelium. Histological examination of the rabbit cornea did not find any gross change in the corneal epithelium ( Figure 3a).
The epithelium thickness and integrity were similar in hydrogel and saline treated eyes. Also, no sign of inflammatory response was observed.
To evaluate whether the soft hydrogel can prolong precorneal residence, the gel was made by crosslinking fluorescent-labled HA-VS with HA-S and instilled on the ocular surface of anesthetized rabbits. The

| Veterinary clinical study
We went on to evaluate if the prolonged retention of the HA soft gel on the ocular surface would be beneficial to the treatment of companion dogs with dry eye.
The biocompatibility of the soft hydrogel (Formulation A3, Table 2) was confirmed in five healthy dogs. We found that the gel was well tolerated by the dogs and well received by the dog owners.
The soft gel can be instilled easily from a unidose eye drop container like conventional eye drops ( Figure S8).
To   The soft hydrogel has very unique physical appearance. For example, when we shake the bottle containing the soft hydrogel, the hydrogel appears to be a nonviscous solution (see Movie S1). However, when we put a fluorescently labeled soft hydrogel in water, it will not be dissolved for days and move along with water when we tilted the tube (Movie S2).
In contrast, a viscous solution of polymer would have a "sticky" appearance and it will be quickly dissolved when mixed with water.
Another surprisingly feature of the soft hydrogel was the ability to retain the elasticity after a history of repeated viscous deformation ( Figure 2f). The study of the mechanism for this behavior is beyond the scope of this study. We hypothesize that the soft hydrogel may consist of a certain percentage of non-covalent entanglements, which may be torn apart more readily and yet quickly recovered afterwards.
The covalent bonds were only minimally disrupted upon shearing, and thus the gel retains the elasticity. An alternative explanation is that although the hydrogel's covalent bonds were broken by the large deformation, the hydrogel was split into a few bulk pieces. Because the gel has low modulus, these pieces may deform to fit tightly into each other and behave as a bulk elastic material. The non-covalent interactions of the polymer chains may further enhance such "healing" effect. Regardless of the mechanism, our result clearly demonstrates that the material can recover from very large deformation.

| Design criteria of soft hydrogel for dry eye and its precorneal retention
Currently, topical application of artificial tears drops is the first line treatment of dry eye. 29 However, artificial tears are cleared rapidly from the corneal surface (typically within minutes), 30,31 and frequent instillations as many as multiple times per hour 32 are needed while the therapeutic effect is still suboptimal. Previous efforts to prolong the precorneal residence time by increasing the viscosity of artificial tear have been ineffective because the solution is quickly diluted and cleared from the lacrimal system. 33,34 The higher viscosity also causes discomfort to the patients and lowers patient compliance.
To design a proper soft hydrogel for dry eye therapy, three aspects should be considered. First, the gel should be able to behave elastically under the stress exerted at the puncta. Precorneal fluid is cleared from flowing through the punta to the lacrimal drainage system. [35][36][37][38] The stress produced by the suction at the puncta is esti-

| Treatment of dry eye disease in dog patients
Dry eye is a multifactorial disease and most of the animal models available are not representative to the human disease. 8  A limitation of the study is that the clinical study is performed as an open label, single-arm study without a regular treatment group.
Although the "switch" design is intended to evaluate therapeutic effect of the conventional and soft gel therapy for the same dog, comparing the soft gel treatment group with a second arm of conventional therapy would be included in a future study.

| Preparation of HA-VS and HA-SH
HAs were modified with pedant VS as described by previously. 24 Briefly, HA was dissolved in double deionized water (DDI water).

| Hydrogel formation
The concentration of HA-VS and HA-SH was first determined from measurement of freeze dried weight or by using a modified cetyltrimethy-lammonium bromide (CTAB) assay. 43 For the CTAB assay, CTAB was dissolved in 2% NaOH solution. HA-VS or HA-SH solution was diluted with 0.1 M phosphate buffer of pH 7.4. The assay was performed in 96 well plate. For each well, 100 μL sample was mixed with 100 μL CTAB solution. After mixing, the plate was immediately transferred to a 37 C oven and incubated for 10 min, and was immediately read in a plate reader for absorbance at 595 nm.
The concentration was determined by comparing the absorbance with unreacted HA processed with the same method.
HA-VS and HA-SH of known concentration was adjusted to pH 7.4 by the addition of 0.5 M PB. The osmolality was then adjusted using 25% NaCl. The polymers were then mixed at various target volume ratio and mass ratio, and adjusted to the target concentration by adding phosphate buffered saline (PBS). The polymers were incubated at 37 C for 24 h for hydrogel formation.
A list of hydrogel formed by 2.6 MDa HA is summarized in Table 2. Formulation A2 were formulation A1 autoclaved at 121 C for 20 min. The higher DM and mass ratio of VS over SH in these formulations was to avoid shrinking of hydrogel during long-term storage. 44  Tissue integrity of corneal epithelium were observed under microscope.

| Mechanical properties measurement of the hydrogel
Gross morphology of the epithelium layer was compared.

| Preparing material for veterinary canine study
The polymers were sterilely filtered through a 0.22 μm sterile filter, mixed to formulate as A3 in Table 2, and filled into a gamma radiated, 330 μL sterile monodose tube strip (Lameplast, Italy). The tube was sealed by a custom-made ultrasonic tube sealer (Looker Machinery Co. Ltd., Guangzhou, China), and the sealed strips were incubated at 37 C for 24 h for gel formation and stored at 4 C. All procedures were performed in a laminar flow clean bench.  Table 3.

ACKNOWLEDGMENTS
The authors acknowledge the Science and Technology Plan of Shenzhen

DATA AVAILABILITY STATEMENT
The raw/processed data required to reproduce the above findings cannot be shared at this time due to time limitations.