Click chemistry functionalization of self‐assembling peptide hydrogels

Abstract Self‐assembling peptide (SAP) hydrogels provide a fibrous microenvironment to cells while also giving users control of biochemical and mechanical cues. Previously, biochemical cues were introduced by physically mixing them with SAPs prior to hydrogel assembly, or by incorporating them into the SAP sequence during peptide synthesis, which limited flexibility and increased costs. To circumvent these limitations, we developed “Click SAPs,” a novel formulation that can be easily functionalized via click chemistry thiol‐ene reaction. Due to its high cytocompatibility, the thiol‐ene click reaction is currently used to crosslink and functionalize other types of polymeric hydrogels. In this study, we developed a click chemistry compatible SAP platform by addition of a modified lysine (lysine‐alloc) to the SAP sequence, enabling effective coupling of thiol‐containing molecules to the SAP hydrogel network. We demonstrate the flexibility of this approach by incorporating a fluorescent dye, a cellular adhesion peptide, and a matrix metalloproteinase‐sensitive biosensor using the thiol‐ene reaction in 3D Click SAPs. Using atomic force microscopy, we demonstrate that Click SAPs retain the ability to self‐assemble into fibers, similar to previous systems. Additionally, a range of physiologically relevant stiffnesses can be achieved by adjusting SAP concentration. Encapsulated cells maintain high viability in Click SAPs and can interact with adhesion peptides and a matrix metalloproteinase biosensor, demonstrating that incorporated molecules retain their biological activity. The Click SAP platform supports easier functionalization with a wider array of bioactive molecules and enables new investigations with temporal and spatial control of the cellular microenvironment.

biocompatible, relatively easy to use, and provide cells with biochemical cues that promote viability and proliferation. However, naturally derived ECMs often suffer from high variability, limited tunability of individual ECM variables, and ill-defined biochemical compositions. Synthetic hydrogels have also been developed (e.g., polyethylene glycol [PEG]), which enable independent tuning of biochemical and mechanical properties, however, these synthetic systems lack the fibrous architecture of native collagen-rich ECM. While some synthetic hydrogel systems have been designed to contain fibrillar structures, 2-4 they remain overall nanoporous, and cells must degrade the matrix with proteases in order to migrate. Fibrous architecture is a critical regulator of cell-matrix interactions, affecting cell adhesion, migration, proliferation, and drug response in vivo. 5,6 For example, tissues such as tendon and ligament, which are primarily exposed to unidirectional loading, have collagen fibers aligned parallel to the direction of the load, 7 while tissues that are exposed to mechanical loading in multiple directions (e.g., the skin) have less aligned fibers. 8 Additionally, in some cancers, increased ECM alignment is a known negative prognostic indicator. 9,10 In vitro studies show that fiber alignment proceeds and facilitates directional cell migration. [11][12][13] In silico studies suggest that a fibrous microarchitecture greatly enhances force transmission and may allow mechanically mediated signaling between cells. 6,14,15 A hydrogel model system that combines a fibrous architecture with independently tunable properties would enable new, more physiologically relevant studies of fundamental cell biology and drug efficacy.
Self-assembling peptides (SAP) are short amino acid sequences (<40 amino acids) which form hydrogels with a fibrous structure similar to that of collagen I, 16 and enable independent tuning of stiffness and cell-binding site density. 17,18 These models have been used to study many aspects of basic cancer research, 19 including ECM remodeling, 20 metastasis, 21,22 drug sensitivity, [23][24][25][26] tumor-stromal interactions, 20,24 and cell dormancy. 27 SAP have also been used to study stem cell differentiation, 17,28,29 the formation of microvascular networks by endothelial cells, 30,31 and as a tissue engineering scaffold. [32][33][34][35][36] SAP have been developed from a number of different amino acid sequences. We have previously developed a SAP gel system that provides independent control of the biochemical and mechanical properties while presenting cells with a fibrous microarchitecture. This system is based on the previously described peptide sequence KFE (acetyl-FKFEFKFE-CONH 2 ). 30 KFE contains alternating hydrophobic-hydrophilic residues, which allows it to form two antiparallel β-sheet ribbons. These ribbons stack to form a double helical morphology, which then form long fibrils. [37][38][39] Since KFE does not support cell adhesion, 16 it provides a "blank slate" on which other functionalities can be added. 17,30,40 This contrasts with the most widely used SAP, RADA16, known commercially as Puramatrix. Cells are able to adhere to RADA16 without additional adhesion ligands, limiting the ability to control and vary adhesion ligand density and identity. 16 By adding an integrin binding sequence (GRGDSP) to the base KFE, we created a new SAP, "KFE-RGD" (acetyl-GRGDSP-GG-FKFEFKFE-CONH 2 ), that supports cell adhesion. By mixing various amounts of KFE and KFE-RGD prior to gelation, the stiffness and the integrin binding site density of the resulting gel could be independently controlled. 40 Incorporation of biological signals into SAP has previously been achieved through mixing of freely suspended, unbound proteins into the SAP prior to assembly (e.g., laminin 31 ), or by modifying SAP peptides with short amino acid sequences, such as in early work by Semino et al., who demonstrated that cells could interact with SAPs containing cell adhesion sites such as YIGSR, RYVVLPR, AND TAGSCLRKFSTM. 41 Since then, SAP have been synthesized with many other short sequences, such as RGD, 30,40,42 IKVAV, 42,43 PDSGR, 44 MMP-sensitive PVGLIG, 45 and other functional motifs. 46 Incorporation of moieties on the SAP itself presents challenges. First, any moiety being added is constrained by the limitations of solid phase peptide synthesis, namely, it must be a short peptide (<40 amino acids), and it must exhibit the necessary solubility characteristics.
These moieties must already be modified with the correct chemical groups (i.e., protecting groups) to make them compatible with SPPS, greatly increasing their cost. Additionally, the position of the moiety within the SAP and SAP block length may need to be optimized to maintain sufficient gel assembly, which can require significant peptide synthesis time and cost. Therefore, methods that are not constrained by peptide synthesis to covalently incorporate biological signals into SAP could greatly increase the utility of the SAP system and enable incorporation of a wider range of molecules. We reasoned that functionalization through a simple thiol-ene click chemistry process, hereafter referred to as "Click SAP," could provide an alternative method to functionalize SAP. Click chemistry reactions are efficient and cytocompatible, 47 and have been used to functionalize many hydrogel systems. [48][49][50][51][52][53][54][55][56] Thiolene chemistry can be initiated by UV light, which allows for precise spatial and temporal control of covalent bonding between a thiol group and an "ene" group. This feature of the reaction has enabled a new array of models of the cell microenvironment with higher spatiotemporal complexity and flexibility than previously feasible.
SAPs have previously been shown to be compatible with click chemistry reactions. One class of SAPs, known as collagen mimetic peptides, has been successfully conjugated into PEG gels 2-4 and into elastin like peptide nanoparticles. 57 The thiol-ene reaction was also previously used to modify SAP fibers with a biotin-containing peptide, which were then labeled with streptavidin gold nanoparticles as a contrast agent. 58  The photoinitiator lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) was synthesized as previously described. 48 To initiate the thiol-ene reaction, gels or pre-gel solutions containing KFE-alloc, LAP, and either TAMRA, RGD, or MMP biosensor were exposed to 365 nm UV light for 60 seconds (or other durations as indicated) at a measured intensity of 3.87 mW/cm 2 .

| Rheology
A Kinexus Ultra+ rheometer (Malvern Instruments) was used to measure the storage moduli (G') of SAP gels. Gel volumes were 75 μl and were exposed to UV prior to being assembled in PBS unless otherwise described. A biopsy punch was used to cut and peel away the PET membrane from each insert so that the gel could be transferred to the rheometer stage without damage. Gels were measured at room tem-

| Cell viability quantification
Cell viability was measured in 45 μl gels containing 3.34 mM KFE, 0.54 mM KFE-alloc, 2 mM LAP, and 0.5 mM RGD, 4 mM NaOH, with or without 60 seconds of UV exposure. Gels containing encapsulated cells were assembled in culture media, and additional culture media was pipetted on top of gels after 25 min at 37 C and 5% CO 2 . Cells were also encapsulated in Matrigel (Corning, #354234) as a positive viability control, gelled at 37 C and 5% CO 2 for 25 min, and then culture media was added to the well. After 24 h, a standard live/dead assay was performed according to the manufacturer's instructions (Invitrogen #L3224). A Nikon A1R live cell confocal microscope was used to take five random fields of view within each gel at 10x magnification using GFP and Texas Red channels. A custom CellProfiler software routine was written to quantify the number of live and dead cells in each image. 60 Briefly, an adaptive threshold is calculated for each pixel using the "Robust Background" method, and cells at least 10 pixels in diameter are counted.

| Optimization of NaOH neutralization
Cells were encapsulated in 75 μl gels for 24 h as previously described.
A series of SAP solutions were pre-neutralized so that gels would have a final NaOH concentrations ranging from 0 to 8 mM. After 24 h of incubation at 37 C and 5% CO 2 , alamarBlue (Invitrogen #DAL1025) cell viability reagent was added at a 1:10 dilution to the media surrounding each gel, according to manufacturer protocol. After 6 h of incubation, the fluorescent signal was measured using a plate reader at 560 nm excitation/590 nm emission. This extended incubation time was chosen based on previous studies. 59,61

| Clicked-in RGD effect on cell morphology
The cell adhesion peptide CRGDS, referred to here as "RGD," was synthesized by Biomatik. Following gel assembly and UV exposure for

| Fluorogenic biosensor activation by MMP
The fluorogenic MMP biosensor Dabcyl-GGPQG#IWGQK-Fluorescein-AeeaC (where # indicates the protease cleavage site) was synthesized using SPPS and functionalized with a quencher (dabcyl) and fluorophore (fluorescein) as previously described. 50  After 3 h of washing, gels exposed to UV retained approximately five times more TAMRA signal as compared to gels that were not exposed to UV, indicating a covalent click reaction between the TAMRA-C and the alloc-modified SAP. A reduction in the absolute amount of fluorescent signal at time zero between UV exposed gels and non-UV exposed gels indicated possible photobleaching and/or photodamage to the TAMRA dye (Supp. Figure 1). Other durations of UV exposure did not yield significant differences in TAMRA retention (Supp.

| Mechanical characterization of click-modified SAP hydrogels
Since substrate stiffness can affect cell behaviors such as migration, 64 differentiation, 65 and proliferation, 66 Figure 2A,B). Clicking the RGD onto the SAPs before or after gel assembly both resulted in similar stiffnesses compared to gels without UV initiation, suggesting that click chemistry can be done without significantly altering mechanical properties. A non-significant decrease in stiffness on average was observed in Click SAP gels compared to gels with the same total mass density of pure KFE, which agrees with previous rheological studies comparing modified and unmodified SAP gels. 40 Gel stiffness could be tuned by varying the concentration of KFE while holding the concentration of KFE-alloc constant. A range of substrate stiffnesses (~450-3800 Pa) was achieved ( Figure 3B, loss moduli and tanδ in Supp. Figure 2C,D). This showed a strong correlation with a power law relationship (R 2 = 0.95, Equation (1)) as seen previously in KFE SAP gels. 40

| Cell spreading on top of SAP hydrogels with clicked-in RGD
The small peptide RGD, an integrin binding site derived from the ECM protein fibronectin, is often incorporated into biomaterials to promote cell attachment and survival. 69 Here, we tested whether RGD incorporated into SAP hydrogels using thiol-ene chemistry would retain biological activity and promote spreading of HT1080 cells on top of the gels ( Figure 5). HT1080 fibrosarcoma cells are relevant for studying cancer cell migration, ECM remodeling, and drug treatment response. RGD click gels were constructed with 7.11 mM KFE, 0.54 mM KFE-alloc, 2 mM LAP, and 0.5 mM RGD. One set of these gels was exposed to UV light for 60 s prior to washing, while another set was not exposed. KFE-RGD and KFE-RDG scrambled non- F I G U R E 2 TAMRA dye clicked into SAP hydrogels using thiol-ene chemistry. Thiol-ene click chemistry (initiated by UV light) causes an increase in TAMRA retention in alloc-modified SAP hydrogels compared to gels in which the reaction is not initiated. Mean ± SD, **p < .01 versus no UV control. SAP, self-assembling peptide; TAMRA, 5(6)-carboxytetramethylrhodamine amount of spreading ( Figure 5C) compared to Click SAP gels without UV initiation. Importantly, adding the RGD peptide to the KFE-alloc mixture, with no exposure to UV light to initiate the thiol-ene reaction, did not support cell spreading compared to the scrambled control, indicating that the adhesion peptide needed to be clicked into the SAP matrix to support cell adhesion on top of the gels.

| Viability of cells encapsulated in 3D Click SAP gels
To develop Click SAP gels as a model system for physiologically rele-  Figure 4 for NaOH optimization). In initial viability experiments, 75 μl gels were used, but viability was generally at or below~70%. We reasoned that generating thinner gels would decrease the time required for media to diffuse throughout the gel and fully neutralize the cells microenvironment, therefore improving viability. In subsequent experiments, the thinner 45 μl gels consistently had viability at or above~85%. Once optimized, viability was quantified using a live/dead assay and compared to a Matrigel control, an established natural hydrogel for 3D cell culture ( Figure 6). Additionally, the effect of the click reaction on viability was also tested.
Culturing cells in Click SAP hydrogels caused a decrease in viability compared to matrigel, but overall viability remained relatively high.
HT1080 cells were over 85% viable on average, and MSCs were over Finally, to test the ability to temporally control presentation of biochemical cues in alloc-modified SAPs, RGD and LAP were diffused into gels for 1 h after assembly, rather than being mixed into the initial SAP solution. In these gels, subsequent UV initiation of the thiol-ene reaction also significantly increased cell spreading, demonstrating that biochemical cues can be added to the SAPs at desired time points following cell encapsulation. The distribution of MSC circularities in SAP gels with RGD clicked in was similar to that of gels containing KFE-RGD, while those without RGD clicked in resembled that of gels containing scrambled KFE-RDG ( Figure 7C).

| SAP hydrogels functionalized with a fluorogenic MMP biosensor
Hydrogels have previously been functionalized with a FRET-based peptide MMP sensor to measure MMP activity in situ. This has been performed in PEG hydrogels to study cell-derived MMP activity 50,61 and for high-throughput drug screening, 59,71 but has not previously been used to create 3D "smart" reporter SAPs. Here, thiol-ene chemistry was used to bind this biosensor to the Click SAP matrix, and the sensitivity of the biosensor to exogenous collagenase (a mixture of

| DISCUSSION
In an effort to create a simple and flexible system to modify SAP, we used click chemistry to introduce functionality into the blank slate of KFE, and demonstrate that the resulting Click SAP retains the and chosen based on its structure that supports self-assembly. It is reasonable to expect that inclusion of additional amino acids such as the integrin binding sequence could perturb this assembly. We 30,40 and earlier work by others 41 49,54,56 Spatial patterning is not possible when biochemical signals are incorporated prior to gel assembly, but could be important for accurately mimicking aspects of native ECM, which is inherently dynamic, complex, and heterogeneous on many length scales. 75 Additionally, we were able to control presentation of cell adhesion cues in time by clicking in RGD post-self assembly (Figure 7). This ability to temporally control the properties of Click SAPs could be harnessed to model the microenvironment of complex, progressive diseases whose biochemical properties evolve over time. This is not possible in traditional SAPs, where functional moieties are added during the synthesis process or entrapped during hydrogel assembly. Mechanical properties could also be temporally controlled by introducing dithiol crosslinkers to stiffen gels on demand. For example, this temporal stiffening could be used to study how cells respond to a transition from a normal to cancerous physical microenvironment. 76,77 While others have used click chemistry to incorporate assembling peptides into PEG gels, 2-4 the present study is the first to show that biologically active molecules can be clicked into a hydrogel comprised solely of SAP fibers. Previous functionalization of SAP fibers via thiolene chemistry were not verified via cell interaction and were not performed in 3D hydrogels. To demonstrate this, we show that Click SAP with RGD covalently linked to the matrix promotes cell spreading in both 2D and 3D (Figures 5 and 7). RGD is a common integrin binding site and can be used to tune the cellular adhesiveness of a material. 69 We found that cell spreading in 3D occurred even in the presence of a MMP inhibitor GM6001, suggesting that spreading was independent of MMP activity (Figure 7). This highlights a benefit of using ECM-like fibrous biomaterials such as Click SAP to study cell behavior.
In dense, nonfibrous nanoporous gels such as Matrigel or PEG, cells must use a proteolytic migration mode to degrade the surrounding matrix, while in a fibrous material, cells can switch to a proteaseindependent amoeboid mechanism in order to migrate. 5,78 Thus, a fibrous material such as Click SAP could be used to investigate the role of MMPs in cell invasion beyond ECM degradation, such as in growth factor activation, induction of epithelial-mesenchymal transition, or cleavage of cell surface adhesion receptors. 79 The Click SAP platform can be used to generate "smart" SAP biomaterials by covalently incorporating biosensors such as an MMPsensitive fluorogenic peptide (Figure 8). This functionalized hydrogel could be used to study how 3D biophysical and biochemical cues, drugs, and other factors regulate cellular MMP activity in a fibrous tissue. In this study, the clicked-in MMP biosensor was able to measure a range of collagenase concentrations, as well as the MMP activity of encapsulated cells, showing significant sensitivity to as little as 1 Â 10 6 MSCs per ml (45,000 total cells). Since the fluorescein-labeled end of the biosensor remains covalently linked to the matrix, even after peptide degradation by MMPs, future studies could investigate the spatial location of MMP activity in real-time using optical sectioning fluorescence microscopy techniques such as confocal microscopy 17,30 or multiphoton microscopy. 80 Another advantage of Click SAPs over existing biomaterials used for disease modeling is affordability. The total cost of the constituents in the Click SAPs with RGD used to encapsulate cells (e.g., as in Figure 7) is approximately $10/ml, much less than the widely used Matrigel, which is typically $25-30/ml. Additionally, because peptides can be purchased from vendors, no specialized techniques or skills are required, and no special equipment is needed other than a UV lamp (~$500).
In summary, functionalized Click SAPs allow one to conveniently introduce various functionality in the blank slate of the KFE SAP system while retaining a fibrous microarchitecture similar to stromal ECM. The modifications we demonstrated allow one to control both the mechanical (e.g., stiffness) and biochemical cues (e.g., integrin binding cite density) that the 3D ECM biomimetic presents to encapsulated cells, while at the same time using the matrix to monitor cellular activity (e.g., MMP-sensitive biosensor).
Based on the modifications we have already demonstrated, one can envision researchers exploiting these and other modifications to generate ECM biomimetics with specific tunable properties. For example, this platform could be used to build a new system to study how binding site identity and/or density affects the activity of cell-secreted MMPs. Such ECM biomimetics could serve as in vitro models that better replicate the in vivo microenvironment, which could be useful to study physiological and pathological conditions and for drug screening.

| CONCLUSIONS
Addition of a lysine with an alloc group to the previously studied KFE SAP allows multiple types of molecules to be covalently linked to SAP hydrogels through click chemistry thereby circumventing limitations associated with incorporating moieties prior to gel formation via solid phase peptide synthesis. The subsequent clicking in of various functional molecules did not disrupt the fibrous microarchitecture of the SAP or the ability of the user to control the mechanical properties of the resulting hydrogel-both important regulators of cell behavior. The resulting Click SAPs support cell encapsulation with high viability and provide spatiotemporal control of functionalization with biomolecules and biosensors. This platform can be used to generate smart SAP hydrogels that contain an MMP-sensitive fluorogenic peptide, which can be used to study how cellular MMP activity is regulated in fibrous tissue.

ACKNOWLEDGMENTS
Thank you to Dr. Gunjan Agarwal for use of the Atomic Force Microscopy Core, and for assistance in preparing AFM samples. All microscopy was performed using the instruments and services at the Ohio State University Campus Microscopy and Imaging Facility. We would also like to thank Dr. Matthew Reilly for use of the Kinexus rheometer. Thank you to Alondra Montoya for her help with developing TAMRA click chemistry methods. Thank you to Dr. Joshua Zent for synthesizing the TAMRA gel marker.

CONFLICT OF INTEREST
The authors declare no conflicts of interest.

SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.