Crosslinker-free collagen gelation for corneal regeneration

Development of an artificial cornea can potentially fulfil the demand of donor corneas for transplantation as the number of donors is far less than needed to treat corneal blindness. Collagen-based artificial corneas stand out as a regenerative option, having promising clinical outcomes. Collagen crosslinked with chemical crosslinkers which modify the parent functional groups of collagen. However, crosslinkers are usually cytotoxic, so crosslinkers need to be removed from implants completely before application in humans. In addition, crosslinked products are mechanically weak and susceptible to enzymatic degradation. We developed a crosslinker free supramolecular gelation strategy using pyrene conjugated dipeptide amphiphile (PyKC) consisting of lysine and cysteine; in which collagen molecules are intertwined inside the PyKC network without any functional group modification of the collagen. The newly developed collagen implants (Coll-PyKC) are optically transparent and can effectively block UV light, are mechanically and enzymatically stable, and can be sutured. The Coll-PyKC implants support the growth and function of all corneal cells, trigger anti-inflammatory differentiation while suppressing the pro-inflammatory differentiation of human monocytes. Coll-PyKC implants can restrict human adenovirus propagation. Therefore, this crosslinker-free strategy can be used for the repair, healing, and regeneration of the cornea, and potentially other damaged organs of the body.

Organ transplantation not only saves lives but also improves the quality of life for recipients. However, the demand for solid organs for transplantation is many folds higher than the availability. In the year 2018 (prepandemic condition), 140,964 organ transplantations were carried out worldwide 1 , however the recent data suggests that the number of transplantations fulfilled only 10% of the real need. Organ donation and transplantation became more difficult in the period of COVID-19 2 . Recent advancements in biomaterials development hold tremendous potential to fill the gap between the supply and demand of donated organs by developing tissue-engineered replacements. For engineered organ development, the key challenge is appropriate biomaterial selection and suitable assembly through crosslinking to form apposite scaffolds for the regeneration of the target organ. Chemical crosslinkers can improve the physicochemical properties of artificial biomaterials and decellularized organs 3 . However, the choice of crosslinkers is highly crucial for building and optimizing the properties of the scaffold, as most of these chemicals demonstrate negative effects on the cellular compatibility of the engineered tissue. This is because crosslinkers act on and modify existing functional groups such as free carboxylic acid (-COOH), primary amine (-NH 2 ) within biomaterials, therefore altering the chemical identity and functional group content of the original/parent compound.
Over the past decades, the development of an artificial cornea has emerged as a prime research interest. It is estimated that about 285 million people in the world live with visual impairment, and 39 million of them are blind 4 . Corneal disease is one of the foremost causes of visual impairment 5 . Replacement of a diseased cornea with a donor cornea through transplantation is the most practiced clinical treatment. However, donor corneas are in short supply worldwide, resulting in 10 million untreated patients with 1.5 million additional every year 6 . It has been determined that we currently have only one donor cornea for every 70 persons in need 7 . Additionally, corneal allotransplantation is contraindicated in autoimmune disorders and after severe chemical injury 8 . Donor-derived infection is, furthermore, a significant complication 9  www.nature.com/scientificreports/ Many attempts have been made to find alternatives to corneal allografts, with limited success, but these efforts have contributed to increase our knowledge for the development of artificial corneas. In this context, synthetic polymer-based materials are mechanically robust but lack of cellular biocompatibility, that unlike natural polymer-based biomaterials 10,11 , limit their application. Among the natural polymers used for tissue engineering, collagen-based artificial corneas (CACs) have been extensively studied as a potential alternative for donor corneas for transplantation, as the human cornea is mainly made up of collagen 12 . CACs have been transplanted in humans and post-trial clinical results showed the capacity of the CACs placed as lamellar, partial thickness grafts to restore vision after transplantation in relatively non-severe corneal disease 13 . At 4 years follow-up, the corneas were well integrated into the host tissue. Interpenetrating polymer-based CACs have also been evaluated in a human pilot study 14 , and in a clinical trial 15 and showed the capacity to stabilize corneas and improve vision in aggressive disease condition. Although CACs are highly biocompatible, their mechanical properties do not permit suturing of the implant and require overlying sutures for transplantation 13 . Enzymatic susceptibility of CACs is another weakness concerning their implantation in severely diseased eyes. Hence, an implant is needed that is enzymatically stable and permits the peripheral placement of interrupted or continuous sutures, therefore avoiding crossing the central cornea with overlying sutures with the attendant scarring and altered topography that each reduce visual acuity 16 .
Furthermore, corneal perforations remain an urgent, vision-threatening problem usually requiring immediate sealing of the cornea. The most applied clinical practice involves temporarily sealing the cornea with cyanoacrylate glue in order to maintain globe integrity and save the eye, as initiated in the late 1960s 17,18 . However, cyanoacrylate has low biocompatibility 17 and causes inflammation 19 . Incomplete polymerization leaves behind toxic cyanoacrylate monomers that can undergo hydrolysis to release potentially toxic compounds such as formaldehyde and alkyl cyanoacrylate 20 . These can induce further corneal injury through scarring and neovascularization 20 . Many patients treated with cyanoacrylate eventually require full corneal transplantation. Considering these limitations, patient care would benefit from a nontoxic sealant for perforated corneas.
In this present work, we describe the development of a crosslinker-free collagen scaffold for corneal regeneration. We have previously reported a self-assembly process of a small peptide (PyKC) based gelation strategy to form a water-insoluble supramolecular hydrogel that remained insoluble for more than a year in water and various other solutions 21 . The unique confinement property of this peptide hydrogel can be used to efficiently trap and protect protein molecules from external denaturing agents for a prolonged period 21 . The current study used 1 and 2% PyKC as an assembling units to stitch together the extracellular matrix (ECM) protein collagen (type I) to form CACs. This implant can be transplanted as a substitute for a human donor cornea in patients who would otherwise receive a penetrating keratoplasty. The pyrene modified dipeptide underwent a self-assembly process through hydrophobic interactions between the pyrene rings, as well as via the formation of disulphide bridge (-S-S-) between cysteine residues, which eventually entrapped collagen to form the hydrogel without modifying the collagen molecules as such. This new formulation (Coll x -PyKC y, where x and y are wt% of respective component) can be used to develop a stand-alone artificial cornea or directly on the eye to work as a sealant to repair corneal perforations. In principle, the proposed strategy can further be used to assemble and embed proteins/peptides irrespective of their size, enzymes, and small molecule as a delivery platform for ophthalmic or other applications. The gelation process is not instantaneous; therefore, it allows for adequate homogeneous mixing time. The formulation components are nontoxic, so the homogeneous mixture of PyKC and collagen can be directly applied on organs/wound sites and over time will form a gel. Our Coll x -PyKC y, hydrogels are stable and resistant to collagenase degradation and exhibit dose-dependent protection with increasing PyKC units (y).

Materials and method
Type I porcine atelocollagen was purchased from Nippon Meat Packers Inc. (Tokyo, Japan). PyKC was synthesized according to a previously published report 21 . All reagents were of analytical grade and used as received. All chemicals were procured from Sigma-Aldrich (St. Louis, MO) if not mentioned otherwise.
Fabrication of collagen hydrogel. Control collagen hydrogels were made following a previously published protocol with modifications 22 . In brief, starting solution containing 10% or 15% (w/w) collagen and 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC)-to-collagen primary amine group ratio was kept at 0.7:1.0 molar equivalent. 0.5 mL aliquots of collagen solution was loaded into a syringe mixing system 23 . The syringe mixing system was pre-filled with 0.625 M 2-morpholinoethanesulfonic acid monohydrate (MES) buffer. Excess MES solution was added to the collagen to keep the dilution factor similar for all hydrogels. The collagen solution was then adjusted to pH 5 ± 0.5 with microliter quantities of 2 N aqueous NaOH, followed by thorough mixing. Calculated volumes of aqueous solutions of EDC and its co-reagent N-hydroxysuccinimide (NHS; both at 10% [w/v]) were added to give a 2:1 molar ratio of EDC to NHS and mixed with the collagen solution. The final homogeneous solution was immediately dispensed between two glass slides 500 μm apart. After 24 h, the hydrogels were washed three times with 1 × Phosphate-buffered saline (PBS). Before performing biocompatibility studies, the hydrogels were treated with 3X antibiotic solution consisting of 300 unit/mL penicillin and 300 µg/mL streptomycin as previously described 16 .
For PyKC collagen hydrogels, the above syringe mixing system was used, but the mixing system was prefilled with 20 mM Tris-HCl buffer (pH 8.0). The collagen solution was adjusted to pH 9 ± 0.5 with microliter quantities of 2 N aqueous NaOH, followed by thorough mixing. 100 μL PyKC solution for PyKC 1 (1% PyKC containing hydrogel) or 150 μL of PyKC solution for PyKC 2 (2% PyKC containing hydrogels) was added to the respective collagen solutions in 20 mM Tris-HCl buffer. The homogeneous collagen mix was cast between glass slides spaced 500 µm apart, and follow-up processing was performed as with the control hydrogel. (Hydrogels of 10% and 15% collagen (as controls), 10% collagen and 1% PyKC, 10% collagen and 2% PyKC, 15% collagen Water content measurement. The water content of hydrogels (N = 4) was determined to ensure uniformity using a previously published protocol 24 with minor modifications. In brief, hydrated hydrogels were removed from PBS, the surface was gently blotted and then the samples immediately weighed on a microbalance to record the wet weight (W 0 ) of the hydrogels. Dry weights of the same hydrogels were measured after drying the samples at 50 °C until a constant weight was achieved (W).
The equilibrated water content of the hydrogels (W t ) was calculated according to the following equation: Suture pull-out study. A special setup was used to evaluate the suture pull out resistance 25 . Six mm diameter round hydrogels (N = 3) were placed on a stage and a suture (Nylon suture size 9-0) was placed in the middle of the hydrogel and connected to the bottom part of the machine. The other end of the suture was also attached to the bottom of the machine (Mark-10 ESM 303, Copiague, NY) through the cut located in the middle of the stage. To measure the resistance to suture cutting through the hydrogels, the stage was pulled up and the energy required to cut the hydrogel entirely and remove the suture from each hydrogel was measured.

Fourier-transform infrared spectroscopy (FT-IR).
FT-IR was performed on a Jasco attenuated total reflectance FT-IR 4200 spectrometer (Jasco ATR FT/IR-4200, Japan), averaging 30 scans between 4000 and 600 cm -1 , at a resolution of 2 cm -1 . The measurements were performed on samples in hydrated form, as well as after drying in a vacuum desiccator for 24 h (N = 3). Differential scanning calorimetry (DSC). The thermal properties of the hydrogels (N = 4) were tested using a DSC 1 STAR System (Mettler-Toledo, Columbus, OH). Heating scans were recorded within the range of 10-80 °C at a scan rate of 1 °C/min. Pre-weighed samples of the PBS-equilibrated hydrogels (weights ranging from 8 to 17 mg) were surface-dried with filter paper and hermetically sealed in an aluminium pan to prevent water evaporation during the study. A resulting heat flux versus temperature curve was then used to calculate the denaturing temperature (T m ). T max of the endothermic peak gives the denaturing temperature.

Contact angle measurement.
In vitro biocompatibility. Because the proposed crosslinker-free biomaterials were intended to be used as artificial corneas, human corneal cells were used to evaluate hydrogel biocompatibility. The human cornea is composed of three primary cellular layers, an outermost epithelium, a middle stroma containing fibroblastlike keratocytes and an innermost, single layer of endothelial cells 26 . Each of these corneal cell populations is important for the function of the cornea, and individual cell types can behave differently in response to specific properties of an implanted biomaterial. Hence, we have evaluated the biocompatibility of all three cell types with the respective collagen hydrogels. Cells grown on collagen hydrogel crosslinked with EDC/NHS and on Tissue culture plate (TCP) were used as controls.
Human corneal epithelial cells (HCEC). The biocompatibility of the hydrogels was tested using SV40-immortalized HCEC, as previously reported 24 , and kindly provided by Professor May Griffith. The hydrogels were cut into 6 mm diameter segments and 10,000 HCEC were seeded on top of each for culture with DMEM/Ham's F-12 medium (Corning, Manassas, VA, USA) supplemented with 10% Newborn Calf Serum (NCS) (HyClone, Logan, Utah, USA), 10 ng/mL epidermal growth factor (EGF) (Gibco, California, USA), and 1% penicillin/streptomycin (Gibco, NY, USA), at 37 °C and 5% CO 2 27 . The cell culture medium was changed every other day. Alamar Blue staining was performed on day 1, day 4 and day 6 after cell seeding. At each time point, 4 μL resazurin sodium salt solution was added to the cell culture medium to give a final concentration of resazurin sodium salt at 0.004% w/v, and incubated for 3 h. Afterward, the cell-incubated resazurin sodium salt solution was diluted with fresh culture medium in a new 96 well plate and read on a BioTek plate reader (Synergy 2, BioTek Instruments; Winooski, VT) at 530/25 nm for excitation and 590/35 nm for emission. Relative fluorescence units were obtained by multiplying the measured values with the dilution factors. A cytotoxicity study was performed on the day 1, day 4 and day 6 through live/dead staining (Life Technologies Corporation, Oregon, USA), where cells were double-stained by calceinacetoxymethyl (Calcein AM) and ethidium homodimer-1 (EthD-1). Live cells www.nature.com/scientificreports/ and dead cells stained as green and red, respectively. Images were captured by using a fluorescence microscope (Zeiss Axio Observer Z1, Carl Zeiss Microimaging GmbH, Jena, Germany). For histological evaluation, the hydrogels were fixed in formaldehyde, dehydrated in increasing concentrations of ethanol (70, 96, 100%) for 30 min at each concentration, immersed twice in xylene for 30 min, and then kept in liquid paraffin for 30 min. Paraffin-embedded sections were cut to 6 μm thickness with a microtome and stained for histology with hematoxylin for 2 min and eosin for 1 min (H&E, Hematoxylin Stain, Fisher Chemical, NJ, USA; Eosin Y, Fisher Chemical).
Human corneal fibroblasts (HCF). Immortalized human corneal fibroblasts (HCF), derived from keratocytes and kindly donated by Professor James V Jester, were used for evaluating the biomaterials. 5000 HCF were cultured on the top of each hydrogel and cultured with DMEM/Ham's F-12 media supplemented with 10% fetal bovine serum (FBS) (Life Technologies) for six days at 37 °C in 5% CO 2 . Alamar Blue assay was performed at days 1, 4, and 6 after cell seeding and live/dead staining was performed on the day 1, 4, and 6. 5000 CEC were seeded on the top of each hydrogel and incubated at 37 °C in 5% CO 2 . TCP was coated with cell attachment enhancer coating (FNC Coating Mix, Athena Environmental Sciences, Inc, Baltimore, MD) before seeding the cells. Alamar Blue assay was performed at day 1, 4, and 6 after cell seeding, and live/dead staining was performed on the day 1, 4, and 6.
Immunofluorescence microscopy. The expression of cytokeratin 3 + 12 and the mucin MUC16 by HCEC after culture on different hydrogels was determined by immunofluorescence microscopy, using paraffin embedded tissue sections. The same paraffin embedding protocol was used as described above. Staining was performed according to a previously published protocol 28 . Paraffin was removed from the tissue sections using xylene, and the samples were rehydrated in water through a graded series of alcohols (100, 96, 70, 50%, and water). For antigen retrieval, tissue sections were incubated with 10 mM Sodium citrate buffer, 0.05% Tween 20 (pH 6.0) at 60 °C for overnight. The sections were washed with Tris-buffered saline (TBS) plus 0.025% Triton X-100 followed by blocking any unspecific binding sites using TBS supplemented with 10% FBS and 1% bovine serum albumin (BSA). The sections were then incubated with the primary antibodies overnight at 4 °C in humidifying condition. Corneal epithelial cell specific mouse cytokeratin (anti-cytokeratin 3 + 12, clone AE5, ab68260, Abcam) and mouse monoclonal antibody against corneal mucin (anti-MUC16, clone X75; ab10029, Abcam) were used separately at dilution 1:50 and 1:300, respectively. Incubation with secondary antibodies was carried out for 1 h at room temperature. FITC-conjugated anti-mouse antibody (ab6785, dilution 1:100; Abcam) was used as secondary antibody. Finally, slides were mounted with EthD-1 and finally with VectaShield mounting medium containing DAPI (Vector Laboratories, Inc., CA, USA), and examined by an inverted fluorescent microscope (Zeiss Axio Observer Z1) with a 40 × objective.
To evaluate the expression of ALDH3A1 and alpha smooth muscle actin (α-SMA) marker by primary (p 1 ) human corneal fibroblasts (pHCF) cultured on different hydrogels, we used standard immunocytochemistry assay. pHCF was cultured from deceased donor corneas deemed unsuitable for human transplantation. Briefly, pHCF (10,000 cells per well) were seeded on the hydrogels and cultured for three days, then fixed in paraformaldehyde (4%). For permeabilization and blocking, hydrogels were incubated in PBS containing Triton X-100 (0.25%) for 10 min and PBST (PBS with 0.05% Tween-20) containing 5% FBS for 1 h, respectively. Each hydrogel was then incubated with the mouse monoclonal antibody against ALDH3A1 (clone 1B6; dilution 1:100, GTX84889, GenTex, Irvine, CA); and with mouse monoclonal antibodies against α-SMA (clone 1A4; dilution 1:200, ab7817, Abcam) for overnight at 4 °C. Then, each hydrogel was incubated with FITC-conjugated anti-mouse antibody (dilution 1:1000, ab6785, Abcam) as a secondary anti-body for 1 h at room temperature. Finally, the slides were mounted in VectaShield mounting media containing DAPI and imaged by an inverted fluorescent microscope (Axio Observer Z1, Zeiss) with a 10 × objective.
Standard immunofluorescence assay was also used to evaluate the expression of ZO-1 marker by CEC cultured on the hydrogels 29 . Briefly, CEC (10,000 cells per well) were seeded and on day three, the hydrogels were washed and fixed with paraformaldehyde (4%). Permeabilization and blocking were done as mentioned above. The hydrogels were then incubated with the rabbit polyclonal antibody against ZO-1 (ZO-1 Polyclonal Antibody, dilution: 1:100, 402,200, ThermoFisher Scientific) as a primary antibody overnight at 4 °C in humidifying condition 29 . Afterward, the specimens were incubated with FITC-conjugated anti-Rabbit antibody (dilution: 1:600, ab6717, Abcam) as a secondary antibody for 1 h at room temperature. Samples were then stained with EthD-1. Images were taken by a fluorescent microscope (Axio Observer Z1, Zeiss) with 100 × objective.  . A no-template control and an endogenous control (GAPDH) were measured, and the expression levels were calculated by the 2^-ddCt method and compared with the mock and infected controls. Immunofluorescence assay: Cell culture conditions, and virus infection were the same as described above. At 48 hpi, HCEC cultured on different hydrogels were fixed and ICC was performed. Cells were incubated with anti-adenovirus type 5 antibody (rabbit polyclonal to adenovirus type 5, ab6982, dilution 1:800, Abcam) for overnight, followed by immunostaining with FITC conjugated anti-rabbit antibody (ab6717, dilution 1:1000, Abcam) for 1 h at room temperature. Pictures were taken with a 40 × objective. EthD-1 was added to the hydrogel before adding the DAPI containing mounting media to double stain the cell nuclei.
Hydrogels and their influence on adaptive immunity. Human monocyte THP-1 cells were used to determine the effect of the hydrogels on adaptive immunity. THP-1 cells were cultured on the hydrogels in RPMI (Gibco) media supplemented with 10% FBS, 1% penicillin/streptomycin and 50 μM β-mercaptoethanol (Gibco) (Standard media) and cultured for 7 days at 37 °C in 5% CO 2 . To induced differentiation of the monocytes to antigen presenting dendritic cells (APCs), the standard media was supplemented with 100 ng/mL rhIL-4 (R&D System, Minneapolis, USA) and 100 ng/mL rhGM-CSF (R&D System). THP-1 cells cultured on TCP with or without differentiation ingredients were used as negative controls. For a positive control, another set was performed on TCP where standard media was supplemented with differentiation ingredients together with lipopolysaccharide (1 × LPS, Invitrogen, Carlsbad, CA, USA). Morphological changes of the cells were evaluated by 20 × phase-contrast images taken at days 1 and 7 with Nikon Eclipse (TS100) microscope. The influence of hydrogels on monocyte differentiation with or without differentiation ingredients was evaluated by labeling the cells after culture for 7 days with direct-conjugate antibodies against CD86 (pro-inflammatory activated APCs), and CD206 (inactivated APCs marker) ( Table 1). Data was acquired using a BD LSR II and the flow cytometry results were analyzed using FlowJo™ v10.6.1 Software (BD Life Sciences) 30 .

Statistical analysis.
Data are shown as averages ± standard deviations, and one-way ANOVA with Tukey post hoc test performed to compare groups. A value of p < 0.05 was considered statistically significant. n.s., *, **, *** and **** represent p > 0.05, p < 0.05, p < 0.01, p < 0.001 and p < 0.0001, respectively. GraphPad Prism Software (GraphPad Software, CA, USA) was used to analyze the data.

Results
Hydrogels generated from PyKC and collagen were transparent and could be cast as plain sheets or curvature similar to a human cornea (Fig. 1).
Optical transmission. The light transmittance of all PyKC hydrogels was similar for all samples (Fig. 2a).
Control hydrogels did not block UV light (100-400 nm); however, PyKC containing hydrogels blocked most light in the UV range. The transmittance of the PyKC-containing collagen hydrogel increased with increasing wavelength, similar to the human cornea 31 , and overall results indicated that 10% collagen-containing PyKC gels showed the highest levels of optical transmission (~ 80%) in the visible region (between 400 and 600 nm). This reached ~ 90% at the far visible region (between 600 and 800 nm) for all PyKC-containing hydrogels.
Water content measurement. Water content of all the hydrogels was ~ 90% (Fig. 2b). Increased PyKC concentration alone did not significantly influence the water content of PyKC containing hydrogels; Coll 10 -PyKC 1 vs. Coll 10 -PyKC 2 (p = 0.1959), and Coll 15 -PyKC 1 vs. Coll 15 -PyKC 2 (p = 0.9387). Adding a high concentration of PyKC to highly concentrated collagen decreased the water content compared to control hydrogels; Coll 10 vs. Coll 15 -PyKC 2 (p = 0.0057), and Coll 15 vs. Coll 15 -PyKC 2 (p = 0.0248). Collagen concentration alone had no effect on water content; Coll 10 vs. Coll 15 (p = 0.9788). www.nature.com/scientificreports/ Suture pull-out study. The suture pull-out study was designed to be a useful indicator of biomaterial performance during the implantation of an artificial cornea. Ideally, an implant should withstand high mechanical load during suturing before the suture tears through the material. For the suture pull-out study (Fig. 3a-c), Coll 10 was used as control. The suture resistance of Coll 15 -PyKC 2 (0.49 ± 0.09 N) was the highest among the tested samples and was significantly higher than that of control and other PyKC hydrogels (Coll 10 , Coll 10 -PyKC 1 , and Coll 10 -PyKC 2 ). No difference was observed between Coll 10 -PyKC 1 and Coll 10 -PyKC 2 (p = 0.9972). No statistically significant difference was noticed for 1% PyKC hydrogel; when the collagen concentration was increased from 10 to 15% for the 1% PyKC hydrogel (Coll 10 -PyKC 1 vs. Coll 15 -PyKC 1 , p = 0.1124). Interestingly, high resistance could be achieved when both PyKC and collagen concentrations were enhanced, for example, Coll 10 -PyKC 1 vs. Coll 15 -PyKC 2 (p = 0.0008). The results indicate that the Coll 15 -PyKC 2 can be used as for implantation with interrupted sutures. The difference was also significant when PyKC concentration was increased in high concentration collagen hydrogels, Coll 15 -PyKC 1 vs. Coll 15 -PyKC 2 (p = 0.0449). Based on the suture pull out study we focused on hydrogels with the highest concentration of collagen while varying the concentration of PyKC.

Fourier-transform infrared spectroscopy (FTIR).
The FTIR spectra of the three hydrated gels (Fig. 3d) exhibited amide I bands at 1634 cm -1 , attributed to stretching vibration of C=O bonds coupled to N-H bending vibrations of the polypeptide backbones. In addition, amide II and amide III bands were observed at 1556 cm -1 and 1241 cm -1 respectively, associated with the N-H bending vibrations coupled to C-N stretching vibrations, and the combination peaks between N-H deformation and C-N stretching vibrations, respectively. The potential amide A and amide B bands were obscured by the presence of a broad band centred at 3308 cm -1 characteristic of the O-H stretching vibration mode of water. PyKC containing hydrogels showed similar bands to the collagen control hydrogel.  www.nature.com/scientificreports/ www.nature.com/scientificreports/ The dried hydrogel samples (Fig. 3e) showed similar spectral features to the ones in the hydrated state, with slightly shifted band positions. The FTIR spectra showed amide A bands associated with N-H stretching at 3285 cm -1 , with shoulders at 3081 cm -1 corresponding to sp 2 C-H stretching of aromatic residues. Amide B double bands were observed at 2955 cm -1 and 2872 cm -1 corresponding to the two stretching modes of CH 2 . Stretching of C=O bonds of the polypeptide backbone was indicated by the presence of the amide I band at 1624 cm -1 . Amide II and amide III bands at 1532 and 1237 cm -1 , respectively, were indicative of N-H in plane bending vibrations coupled with C-N and C-H stretching. The remaining signals were assigned as follows: 1456 cm -1 O-H bending coupled with C-H scissoring, 1404 cm -1 carboxyl O-H bending, and broad band at 1033 cm -1 corresponding to C-O stretching. The dried hydrogels induced an additional band at 840 cm -1 corresponding to the pyrene ring, which was missing in the control hydrogel. As expected, the intensity of this band was relatively high in 2% PyKC hydrogel. This pyrene signal was obscured in the hydrated gel by the relatively broader band ascribed to the presence of water.

Contact angle measures.
The contact angle is a measure of the wettability of the sample (Fig. 3f). We found significantly higher contact angle values for the Coll 15 control relative to Coll 15 -PyKC 1 (p < 0.0001) as well as Coll 15 -PyKC 2 (p < 0.0001). The contact angles increased from 21.61 ± 1.08° to 34.59 ± 0.54° as the PyKC amount in the hydrogel was increased from 1 to 2% which indicates the enhanced hydrophobicity of the resultant hydrogel. Thermal properties: differential scanning calorimetry (DSC). Thermal stability was measured through DSC by evaluating the denaturation temperature (Tm; Fig. 3g In vitro implant cellular compatibility. The three major corneal cell types, HCEC, HCF, and CEC, were employed to evaluate the biocompatibility of the newly developed hydrogels. The 15% collagen containing PyKC hydrogels was assessed, and Coll 15 hydrogel and TCP were served as controls. Alamar Blue bioassay was performed to evaluate cell proliferation on the hydrogels, and live/dead staining (Fig. 4a) was used to assess cytotoxicity. Statistical analysis was done on the final day of the Alamar Blue assay and the non-significant differences were presented in the figures (Fig. 4b-d). Cells grown on TCP proliferated the most, for all cell types, and there was no difference in cell growth for any cell type between Coll 15 -PyKC 1 and Coll 15 -PyKC 2 hydrogels.
HCEC growth was prominent on control hydrogels compared to the PyKC hydrogels (Fig. 4b), throughout the study. Alamar Blue assay showed a negligible effect on metabolic activity between Day 4 and Day 6, for all groups. This may be attributed to cell confluence and contact inhibition, which was supported by the live/dead data showing confluence on day 4 of the cultures (Supplementary Fig. 2). H&E images also confirmed a continuous epithelial layer of the top of the hydrogels (Fig. 5a).
HCF growth was similar on Coll 15 and TCP (Fig. 4c). There was a gradual increase in cell proliferation over time for all groups. Although on day 6 Alamar Blue data showed a substantial difference (Coll 15 vs. Coll 15 -PyKC 1 and Coll 15 vs. Coll 15 -PyKC 2 , significance p = 0.0092, p = 0.0006, respectively) in cell proliferation between control and PyKC hydrogels, live/dead analysis showed that all the biomaterials along with TCP led to confluence by day 6 of the culture.
The viability of CEC was greater on TCP compared to the other surfaces; TCP is coated with FNC to facilitate cell growth. Due to low cell seeding number, the changes in cell proliferation between days were prominent. On the day 6, more cells were present on control hydrogel than PyKC hydrogels (Coll 15 vs. Coll 15 -PyKC 1 and Coll 15 vs. Coll 15 -PyKC 2 , significance p = 0.0010 and p < 0.0001, respectively). Although cells were not confluent at day 4 ( Supplementary Fig. 2) on any surfaces, the confluence was apparent on all surfaces by day 6 (Fig. 4d).

Phenotype of corneal cells cultured on hydrogels. Phenotypic evaluation of corneal cells on different
hydrogels was performed by immunofluorescence microscopy. HCEC cultured on different PyKC, and control hydrogels (Fig. 5b) expressed cytokeratin 3 + 12 and MUC16. Autofluorescence from PyKC hydrogels compromised the picture quality. pHCF cultured on all hydrogels expressed α-SMA, whereas very few cells expressed the aldehyde dehydrogenase marker, ALDH3A1 protein (Fig. 5c). PyKC hydrogels showed autofluorescence using the blue laser channel. ZO-1 is a marker of the tight junctions. CEC cultured on the different hydrogels expressed ZO-1 antigen at cell-cell junctions (Fig. 5d).
In vitro evaluation of human adaptive immunity in the presence of the hydrogels. We wanted to determine whether the monocytic cell line THP-1 differentiation to the activated dendritic cells (as implied by CD86 expression) or the inactivated dendritic cells (with CD206 expression) phenotype is impacted by culture on the different hydrogels. Morphological changes of the cells were evaluated on day 1 (Supplementary Fig. 3a) and day 7 (Supplementary Fig. 3b). On day 1, round THP-1 cells were observed for all groups except the cells cultured with LPS. This difference became more prominent at day 7. LPS cultured cells on day 7 were elongated in shape, and tightly attached to the surface. Cells treated with differentiation media appeared larger than cells cultured on hydrogels with standard media. To confirm dendritic cell activation, the expressions of CD86 and CD206 were evaluated after 7 days of cell culture on different hydrogels and TCP. Cells cultured with differentiation media with LPS on TCP showed the highest expression of CD86 and lowest expression of CD206 (Fig. 6a), which was the positive control for cell activation. Cells cultured with differentiation media on Coll 15  www.nature.com/scientificreports/ Coll 15 hydrogels (Fig. 6a). CD86 and CD206 expression was shown in Fig. 6b and   www.nature.com/scientificreports/ sion of CD206 for PyKC hydrogels than with control hydrogels or TCP. The complete statistical analysis between groups is presented in the supplementary Table 1.

Evaluation of preliminary antiviral properties.
To determine any effect of the hydrogels on resistance to viral infection, we grew HCECs on the different hydrogels and infected them with HAdV-D37. Viral DNA replication was evaluated by infecting hydrogel-cell constructs with HAdV-D37 at MOI 1. DNA was isolated from the infected and control constructs at 24 h post infection (hpi) for qRT-PCR. Compared to infected Coll 15 cells, the quantity of E1A DNA was lowest in Coll 15 -PyKC 1 cells, followed by Coll 15 -PyKC 2 cells. The relative fold change of HAdV-D37 DNA is depicted in Fig. 7a. The analysis through immunofluorescence assay also revealed a reduction in HAdV-D37 in the Coll 15 -PyKC 1 group compared to the control Coll 15 . Notably, fluorescence was less in the construct Coll 15 -PyKC 1 , followed by Coll 15 -PyKC 2 at 48 hpi (Fig. 7b). Thus, the measurements of viral DNA were consistent with the immunofluorescence results, which exhibited the same inhibition profile.

Discussion
Cornea transplantation is one of the most common forms of organ transplantation. As the outermost part of the eye, the cornea plays a critical role in vision by focusing and transmitting light to the retina and protecting the eye from external infectious agents. Corneal damage is one of the prime reasons for blindness globally, and the primary treatment is a replacement of the diseased cornea with a donor corneal allograft. However donor corneas remain scarce worldwide, resulting in many patients left untreated 32 . To fill the gap between supply and demand, much research has focused on developing an artificial cornea. Toward this end, CACs have been transplanted into humans 13,15 , in most cases they consisted of collagen crosslinked with EDC and NHS. Carbodiimide-based chemical crosslinkers, such as EDC-based methods are considered a standard strategy with low cytotoxicity, but the mechanical strength of crosslinked collagen is less than satisfactory. EDC crosslinks free primary amino groups of lysine residues with cell-reactive carboxylate anions on glutamate or aspartate residues. This may reduce the accessibility of essential cell attachment points on collagen-based biomaterials 33 . Carbodiimide forms zero-length crosslinking that very quickly produces rigid bonding. Shorter crosslinking times make the implants inhomogeneous, susceptible to microcracks and shearing 13 . In a clinical study with chemically crosslinked collagen based implants, overlying sutures crossing the centre of the cornea were used to avoid implant extrusion 13 . These sutures impeded epithelial growth onto the centre, leaving the implant susceptible to collagenases and generating thinning, haze and astigmatism 13,34 . Other crosslinkers, e.g., glutaraldehyde and hexamethylene diisocyanate, have also been used to produce collagen-based implants; however, modification of important functional groups of the parent biomaterial and low biocompatibility remain limitations. Recently, light-induced crosslinking of ECM components in the presence of photo initiators has become popular 29,35 , although patients with corneal inflammation are highly light-sensitive and it can be difficult for them to tolerate intense visible light application for any length of time without anaesthesia 20 .
Any biomaterials designed for use in an artificial cornea should permit transparency similar to that of the human cornea. All the CACs previously developed were transparent from UV to visible light 13,15 as we observed for the control hydrogel. However, control EDC/NHS crosslinked hydrogels cannot block UV light, but PyKC hydrogels can, while still permissive of transmission of visible light. The human cornea blocks some UV light 31 . Therefore, patients implanted with PyKC hydrogels would not require additional protection from UV light, as is recommended for CAC implants 13 . The blocking of UV light could be ascribed to the presence of pyrene rings in PyKC. Pyrene is the polycyclic aromatic hydrocarbon with high symmetrical properties. Pyrene is one of the most studied organic molecules in terms of its photophysical properties. Pyrene has 16 π electrons, but however, does not follow the Hückel's law of aromatic compounds 36 . This electronic arrangement makes it an interesting compound to evaluate in host-guest systems where two or more molecules combine in special structural relationships. Among the numerous techniques employed to study host-guest chemistry, UV-vis spectroscopy is one of the most frequently used techniques 37 . As pyrene exhibits characteristic absorption spectra in the UV-region, pyrene UV absorption spectra was used to determine the critical micellar concentration in different surfactant solutions 38 . Because of its highly conjugated nature as an aromatic molecule, pyrene has several absorption bands extending from 200 to 360 nm 39 . The equilibrium water content of the hydrogels was measured to ensure uniformity 15 . All the hydrogels contained similar water content of around 90% which is greater than the water content of human corneas (78%) 40 . Collagen-based CACs are vulnerable to enzymatic digestion, and this makes them unsuitable for implantation in patients with inflammatory corneal diseases. Collagenase solution (5 U/ mL) was also used to evaluate the properties of porcine and recombinant human collagen matrices for optically clear tissue engineering applications 41 . This same concentration has been used as a standard in earlier reports on biomaterial development for artificial corneas 16,42 . However, 5 U/mL is considerably more concentrated than in www.nature.com/scientificreports/ some other published studies where biomaterials persisted over many days instead of hours 43,44 . In a study with identical collagenase concentrations 16 , EDC/NHS hydrogels broke down within few hours, as we also observed for our control hydrogels. However, PyKC-containing hydrogels remained stable at the end of the study (28 h) ( Supplementary Fig. 1a). The stability against degradation was attributed to the unavailability of breakable bonds of the PyKC hydrogel and its exceptional compartmentalizing ability through tight packing of the nano-fibres, which create a protective envelope for proteins from denaturation 21 . As EDC/NHS treatment induces zero-length crosslinking (does not take part in linkage), they generate rigid hydrogels with a high compressive modulus, which we also observed for the control hydrogels. Increasing concentrations of both PyKC and collagen in the hydrogel increased the compressive modulus ( Supplementary Fig. 1b). Suturability is also critical to translation of biomaterials to corneal implants, and the lack of suturability contributed to the less than the optimal clinical outcome of previously implanted CACs 13 . Herein, we found that high concentrations of both PyKC and collagen contributed to improved suturability but acknowledge that even PyKC hydrogels cannot yet compare in suturability to the native human cornea, which remains intact even when the pulling force is sufficient to break the suture 45 . FTIR confirmed the presence of pyrene unit in the hydrogel, as pyrene showed characteristic aromatic stretching near 840 cm -1 46 . We also evaluated wettability of the hydrogel  www.nature.com/scientificreports/ surfaces by measuring the contact angles 28 . A high contact angle indicates low wettability or a hydrophobic solid surface and a low contact angle indicates high wettability or a hydrophilic surface 47 . It is generally assumed that hydrophilic biomaterials result in the most optimal biocompatibility 48,49 , however, there are exceptions 47 . Biomaterial surface interactions for extremely hydrophilic surfaces may trap cells by high cell-substrate adhesive forces, and this could reduce cell-cell communication and monolayer confluence 50 . On the other hand, highly hydrophobic surfaces will exhibit less protein interactions, which can prevent or reduce cell adhesion 51 . In parallel, moderately wettable (contact angles between 48 and 62°) surfaces appeared to be optimal for cell adhesion and growth 52 . Our control hydrogel contact angle was close to a moderately wettable surface and was superior for corneal cell proliferation. We observed that the contact angle was reduced by the addition of PyKC, but then increased as the concentration of PyKC was increased. Regardless, for the PyKC hydrogels, wettability did not influence much on the biocompatibility of the hydrogels as shown by cell proliferation and the live/dead assays. It was previously shown that contact angles have a limited impact on biocompatibility when corneal endothelial cells are cultured on hydrogels 53 . We previously showed also that an increasing contact angle did not influence the recellularization of the decellularized porcine cornea 28 . We also measure the T m of the hydrogels to evaluate their thermal stability. The heating of hydrogels leads to the denaturation of the native triple helix structure of collagen and converts the triple helix into a single-chain random coil conformation 16 . The T m depends on the chemical nature and the degree of crosslinking. High T m is desirable as shown for human cornea (65 °C) 54 . Although none of our hydrogels reached that level, PyKC hydrogels showed higher T m (50-51 °C), while the control hydrogels showed similar denaturation temperature to a previously published report (~ 46.8) 16 . Improved thermal stability for PyKC hydrogels suggests that PyKC protects entrapped protein/peptide from the heat 21 .
We did not observe any degradation or contraction of the materials unless we added collagenase to the solution. Cells had no negative effect on the materials. Cytotoxicity studies make clear that PyKC containing hydrogels are not cytotoxic to corneal cells. Collagen-based hydrogels have been previously shown to promote corneal cell growth over a long time, as previously published 16,22 , and adding PyKC did not diminish their superior biocompatibility. In addition, phenotypic evaluation showed that HCEC expressed corneal epithelial specific CK3 + 12 and MUC16; however, autofluorescence of the biomaterials compromised the quality of the images for HCECs as was seen with re-cellularized porcine corneas 28 . Furthermore, HCECs cultured on all hydrogels expressed the HCEC-specific marker zonula occludens-1 (ZO-1) 55 . Primary corneal stromal cells showed expression of α-SMA on all the hydrogels. pHCF are normally quiescent and express the unique markers aldehyde dehydrogenases and keratocan. In serum culture, like our culture condition, they proliferate and become stromal fibroblasts while losing their characteristic phenotype 56 . It was shown previously that cultured primary fibroblasts develop a myofibroblast phenotype even on TCP 57 . It has also been shown that when primary corneal fibroblasts are cultured in a collagen matrix, α-SMA expression increased from day 0 to 7 but then decreased from day 7 to 14 58 . It was found that nearly 100% of fibroblasts seeded in a collagen sponge expressed α-SMA 59 . As expected, given that α-SMA expression is typically inversely proportional to that of the fibroblast specific antigen, ALHD3A1, we observed very low ALHD3A1 expression on hydrogel cultured cells. Regardless, in human studies of implanted The amount of E1A was significantly less with Coll 15-PyKC 1 compared to Coll 15. Data was normalized to GAPDH and are expressed as the mean ± standard deviation. The quantification of E1A was calculated using the 2^-ddCt method, and the results are presented as relative fold change. A value of p < 0.05 was considered statistically significant. *, **, and *** represent p < 0.05, p < 0.01, and p < 0.001, respectively. (b) Immunofluorescence assay. Immunostaining for HAdV-D37 in human corneal epithelial cells (HCECs) cultured on different infected hydrogels. Cells were observed at 48 hpi, green fluorescence indicates the adenovirus load as assessed with anti-Ad5 and secondary antibody conjugated with FITC and red represents nuclei stained with EthD-1. The scale bar is 20 μm. www.nature.com/scientificreports/ CACs 13 , they remained mostly transparent 60 . Once corneal wound remodelling is complete, myofibroblasts are known to disappear from the wound space through apoptosis 61 .
Silencing adaptive immunity considered one of the major factors determining graft success. Antigen presenting cells (such as dendritic cells) play important roles in responding to foreign bodies by activating adaptive immunity as well as maintaining immune tolerance. Due to this unique nature APCs can be a useful tools to evaluate the inflammatory responses of the biomaterials 62 . It was shown previously that that human monocytic (THP-1) cell lines can be differentiated rapidly into dendritic cells 63 . In the absence of inflammatory stimulus, antigen-presenting dendritic cells remain immature; however, inflammation causes activation and functional transformation of inactive dendritic cells into mature dendritic cells 64 . In this work we used THP-1 cell line to generate dendritic cells and LPS was used as positive control of activated dendritic cells. We monitored their expression of CD86 and CD206 to determine their response to different hydrogels. The expression of the CD206 has been regarded a differentiation hallmark of immature dendritic cells whereas monocytes and activated dendritic cells do not express CD206 65 . Though immature dendritic cells also express lower levels of CD86 66 . In contrast, activated dendritic cells expressed prominent CD86 to initiate the adaptive immunity response against the graft 66 . We compared standard media with media containing differentiation inducers, specifically rhIL-4 and rhGM-CSF 63 . We used LPS in the study as a positive control as LPS is known to activate APCs 67 . We found that CD86 expression was down regulated when cells were cultured on PyKC hydrogels as compared to controls, to lower expression levels than even on TCP, suggesting that the PyKC hydrogels did not influence differentiation of the THP-1 cells to activated APCs. To the contrary, we observed heightened expression of CD206 by cells cultured on PyKC hydrogels. It was previously shown that CD206 expression was increased by APCs when cultured on highly biocompatible biomaterials 20 . From our results, we suggest that PyKC hydrogels would perform better if implanted in the corneas of diseased eyes and not contribute to inflammation triggered by adaptive immunity.
Ocular surface infections by human adenovirus are common, and the most frequent cause of viral conjunctivitis 68 . Corneal infection is also an important cause of allograft rejection 69 . PyKC has a polycationic backbone which is usually antimicrobial in nature. Therefore, we evaluated the antiviral behavior of the HCECcovered implants against human adenovirus. Interestingly, PyKC containing implants appeared resistant to infection. This preliminary study supports the idea that PyKC may provide additional protection from viral infection after implantation.

Conclusion
We have shown that collagen-based artificial corneas can be made without using crosslinkers. Enhanced suturing ability and resistance to enzymatic degradation were improved compared to a standard collagen hydrogel, suggesting potential as an artificial cornea without modification of extracellular matrix component and without compromise of its biocompatibility. As the gelation process is not immediate, homogeneous semisolid mixture could be applied directly at the target site to patch defects in potentially any living body organ, in contrast to hydrogels with crosslinkers which solidify immediately and are cytotoxic. Crosslinker-free PyKC-based hydrogels may show utility as a biological glue and sealant for corneal defects and perforation. Our self-assembled PyKC peptide combined with collagen, gelatin, or other ECM, could be applied directly to any wounded area and once gelation has occurred, would protect, and enable regeneration. This approach can also be utilized to embed and encapsulate-for delivery-drugs, growth factors, and antibodies.

Data availability
The datasets used and/or analysed during the current study available from the corresponding author on reasonable request. www.nature.com/scientificreports/