pH-Responsive Charge-Conversion Progelator Peptides

Abstract

A simple strategy for generating stimuli-responsive peptide-based hydrogels via charge-conversion of a self-assembling peptide (SAP) is described. These materials are formulated as soluble, polyanionic peptides, containing maleic acid, citraconic acid, or dimethylmaleic acid masking groups on each lysine residue, which do not form assemblies, but instead flow easily through high gauge needles and catheters. Acid-induced mask hydrolysis renews the zwitterionic nature of the peptides with concomitant and rapid self-assembly via β-sheet formation into rehealable hydrogels. The use of different masks enables one to tune pH responsiveness and assembly kinetics. In anticipation of their potential for in vivo hydrogel delivery and use, progelators exhibit hemocompatibility in whole human blood, and their peptide components are shown to be noncytotoxic. Finally, demonstration of stimuli-induced self-assembly for dye sequestration suggests a simple, non-covalent strategy for small molecule encapsulation in a degradable scaffold. In summary, this simple, scalable masking strategy allows for preparation of responsive, dynamic self-assembling biomaterials. This work sets the stage for implementing biodegradable therapeutic hydrogels that assemble in response to physiological, disease-relevant states of acidosis.

Graphical Abstract

1. Introduction

Reversible charge-conversion of proteins and polymers with maleamic acid derivatives has been used in seminal work to study protein denaturation and folding pathways,^[1] for structural disassembly of polymer films or colloidal structures,^[2] for polymer film deposition,^[3] for cellular penetration,^[4] and for drug release.^[5] These moieties exhibit acid-triggered amide bond hydrolysis. Furthermore, the pH sensitivity can be tuned for faster responsiveness through increasing bulkiness with aliphatic substitutions onto the alkene backbone.^[6] Although pH-stimulated charge-conversion in vivo has been regularly sought after in biomedicine, efforts have focused almost exclusively on initiating disassembly and/or drug release.^{[4a, 7]} To our knowledge, this strategy has not been applied to induce supramolecular assembly from a soluble precursor. We sought to use charge-conversion chemistry as a facile method for converting inert small molecule peptides into structurally dynamic hydrogels. We anticipate their utility as injectable materials for delivery to sites of tissue acidosis due to injury including myocardial ischemia, rheumatoid arthritis, articular cartilage damage, and epidermal wounds.^[8] For example, during myocardial ischemia, the local pH drops below physiological conditions (~pH 6.0–6.4),^{[8b, 9]} which is caused by a gradual buildup of lactic acid (pKa 3.86) and drop in plasma bicarbonate. Similar drops in pH have been observed in synovial fluid at sites of arthritis or cartilage damage (pH 6.5),^[10] chronic metabolic acidosis due to liver failure (pH 7.1),^[11] and acute (pH 6.5) and chronic (pH 5.4) epidermal wounds.^[12] These forms of tissue acidosis can exacerbate inflammation, leading to further tissue damage.^{[8a, 13]} Injectable hydrogels seeded with cells/growth factors or as acellular scaffolds have been used for the repair of cardiac tissue after myocardial infarction (MI),^[14] articular cartilage damage,^[15] burn and epidermal wounds,^[16] and treatment of arthritic flares from wear and tear.^[17] Introduction of these viscoelastic scaffolds provides lubrication or acts to supplement degraded extracellular matrix (ECM), which provides structural support, a niche for cellular anchoring, and biochemical communications.^[18] Additionally, drug-loaded hydrogels provide a porous network for controlled delivery of encapsulated anti-inflammatories, growth factors, and/or cell therapies.^[19]

Unfortunately, these injuries usually have irregular or gradated boundaries, making localized hydrogel delivery difficult. Furthermore, a balance between material spreading and solidification kinetics is a persistent challenge.^[20] Excess spreading can cause material diffusion away from target tissue, and rapid gelation can preclude noninvasive delivery strategies or prevent appropriate tissue coverage. Conventional strategies with inert injectable hydrogels, such as self-assembling peptides (SAP), rely heavily on varying the peptide concentration and solution ionic strength to control their mechanical properties prior to injection. Although material dilution aids injectability, lack of material can prevent self-assembly. Conversely, removal of salts prior to injection can result in unwanted blood coagulation during systemic delivery. The advent of responsive, structurally dynamic materials presents a useful strategy for controlled assembly in response to endogenous stimuli such as pH, temperature, redox chemistry, metal chelation, enzymes, and mechanical stress.^[21]

As a test of the concept of acidosis-driven self-assembly, we modified the well-known KLD-12 SAP (referred to as KLD control) for tunable pH responsiveness. This peptide has demonstrated utility as a degradable, nonimmunogenic, and nontoxic hydrogel for tissue engineering.^[22] However, practical application in the body following traumatic injury is limited by its high viscosity, precluding the use of catheter delivery.^[21f] Furthermore, lack of stimuli-responsiveness forces this material to assemble at the immediate site of application. We demonstrate that temporary charge-conversion of KLD control, from zwitterionic (Z=0) to polyanionic (Z=−6) is sufficient to yield soluble and low viscosity progelators. These materials are amenable to delivery via an ultrathin catheter, unlike the KLD control. Furthermore, they exhibit pH-tunable angstrom- to millimeter-scale gelation at physiologically relevant levels found during injury-related tissue acidosis, and remain assembled under neutralizing conditions mimetic of healing tissue microenvironments.

2. Results and Discussion

2.1. Synthesis of Charge-conversion Progelators.

We sought a method with synthetic simplicity and reproducibility for scalable, gellable peptide biomaterials.^[23] In addition, we considered that negatively- as opposed to positively-charged biomaterials, generally possess lower cytotoxicity and resist cellular internalization.^[24] We note that without removable anionic masks, pH-activated hydrogels that rely solely on pKa-defined charge would disassemble and flush away upon tissue neutralization, precluding their use for prolonged tissue engineering. Tunable sensitivity and hydrolytic cleavage of simple non-, mono-, and dialkyl-substituted maleamic acids to mildly acidic pH’s has been reported through comprehensive NMR studies in the literature.^[6] As such, we modified KLD control through the addition of either maleic anhydride, citraconic anhydride, or dimethylmaleic anhydride to yield mal-KLD, cit-KLD, or dma-KLD, respectively (Figure 1, Table S1, and Movie S1, Supporting Information). The resulting polyanionic peptides persisted as soluble solutions at high concentrations (up to 90 mM) and physiologically relevant conditions in 1x Dulbecco’s Phosphate Buffered Saline (DPBS, pH 7.4). Peptides were lyophilized for storage and rapidly dissolved upon resuspension. Peptide solutions were stable up to 14 days at pH 8–9 at room temperature.

Figure 1.

Schematic of peptide progelator structures and pH-driven cleavage. The zwitterionic KLD self-assembling peptide is modified with maleic anhydride, citraconic anhydride, or dimethylmaleic anhydride to generate the polyanionic progelators mal-KLD, cit-KLD, and dma-KLD, respectively. Inset images of progelators at 10 mM show nonviscous solutions. Acid-treatment regenerates the starting hydrogel material upon release of maleic acid derivatives.

2.2. Characterization of Progelators with Tunable pH-sensitivity.

Synthesized progelators and KLD control peptide were purified by HPLC and further characterized by liquid chromatography mass spectrometry (LCMS) and analyzed for pH-responsive mask removal (Figure. 2, Figure S1, Supporting Information). The elution time for purified progelators is ~4.2–4.3 min (Figure 2A) and corresponding mass spectra (Figure 2B) show expected [M-2H]^2- species. ¹H NMR spectra of purified progelator in an H₂O/D₂O cosolvent mixture buffered to pH 7.4 reveals the presence of protons unique to the maleamic acid amides on each progelator, which are absent in the KLD control peptide (Figure S2-5, Supporting Information). Both the α- and β-methyl conformers for cit-KLD were identified at a ratio of 0.69:0.31. Splitting of the β-methyl olefin through proximity of the acid amide proton enables conformer identification. We note that reduced signal intensity through proton exchange with solvent, and variable degrees of intramolecular and intermolecular interactions (KLD control > dma-KLD > cit-KLD > mal-KLD) limit peak resolution. Thus, we rely on LCMS to verify purity, as is standard protocol with SAPs.^[25]

Figure 2.

Characterization of progelators by LCMS and pH-responsive unmasking by NMR. (A-B) Liquid chromatography mass spectrometry (LCMS) of mal-KLD, cit-KLD (α- and β-methyl conformers), and dma-KLD. (A) LC spectra show progelator peak (t ~4.5 min) as monitored at 214 nm and (B) corresponding mass spectral patterns with species identities. Spectra show m/z values corresponding to the progelators. See Table S1 in Supporting Information for mass identities. (C) Synthetic scheme and ¹H NMR spectra of pH-induced unmasking reported as signal intensity vs time at 8, 30, 60, 120, 480, and 720 min. Control spectra of starting progelator at pH 7.4 after 720 min is provided at the bottom in black. Peptides were prepared at 2.5 mg/mL (pH 7.4) or 5 mg/mL (pH 3.0, 5.5, 6.5). Proton identities with apostrophes represent the same peak under acidic conditions. Acidic spectra collected in D₂O with 200 mM buffer (d₃-phosphoric acid, d₄-acetate, or d-phosphate) and 400 mM LiCl. Control spectra collected in H₂O/D₂O cosolvent with 200 mM d-phosphate and 100 mM LiCl.

Upon addition of acidic buffer (Figure 2C), a significant upfield shift of β-olefin protons is observed at 8 min for mal-KLD and cit-KLD, which agrees with anhydride spectra at variable pH values (Figure S6, Supporting Information). We note that at pH 3.0, mal-KLD exhibits reduced solubility, whereby the starting material olefin protons disappear from this spectrum. This is not observed at higher pH values with cit-KLD (pH 5.5) and dma-KLD (pH 6.5). High solubility of free vs bound maleic acids and higher LiCl concentrations in acidic experiments improve resolution. Amide bond cleavage was monitored in real time up to 720 min for mal-KLD, cit-KLD, and dma-KLD at pH 3.0, 5.5, and 6.5, respectively. Release of masking moieties is denoted by increasing signal for protons c, f, and i. Olefin peaks for α- and β-methyl conformers (5.56 and 5.85 ppm, respectively) of cit-KLD display differential cleavage kinetics. Notably, the β-methyl conformer engages in rapid cleavage and olefin protons are absent from the spectra after 8 min. In contrast, slow cleavage of the α-methyl conformer reveals olefin peaks (d’) that persist for up to 720 min, which agrees with literature reports.^{[2a, 26]}

Importantly, each cleavage buffer was chosen to match the relative product diacid pKa, and thus their peptide activation pH’s.^[6b] This is most notable for the cleavage of dma-KLD which yields two populations for i (1.78 and 1.84 ppm) resulting from variable protonation states of the product carboxylate. Additionally, proton peaks g/h (1.85 and 1.81 ppm) in dma-KLD merge into two closely spaced populations for masked g’/h’ (1.78 ppm) and unmasked i (1.77 ppm) upon initial incubation in acidic solution (red traces). NMR provides a useful analysis of these unmasking reactions. However, pH effects on the chemical shifts, proton exchange from self-assembly, and overlapping signals from the progelators and resulting cleavage products hamper quantitative analysis.

2.3. Unmasking of Soluble Progelators Initiates Self-assembly.

The synthesis of disperse and fully soluble biomaterials is key for reducing delivery invasiveness. We used circular dichroism (CD) and transmission electron microscopy (TEM) to evaluate secondary structures formed by our peptides in solution before and after accelerated (pH 3.0, 24 hr) pH-responsive charge-conversion (Figure 3 and Figure S7, Supporting Information). All three progelators adopt random coil configurations, as seen by a peak minimum at ~202–205 nm (Figure 3A-D). The most hydrophobic progelator, dma-KLD, also absorbed at longer wavelengths, indicative of weak intra-strand assembly (Figure 3D). Progelators were treated at pH 3.0 for 24 hr, followed by neutralization to pH 7.4 to recapitulate β-sheet assembly (minima ~215–218) observed with the KLD control peptide (minimum 222 nm). Strong absorbance of maleamic acids in our progelators below 200 nm leads to near-saturation of the voltage detector (Figure S8, Supporting Information). Dialysis of the unmasked solutions to remove maleic acid hydrolysis products yielded nearly identical voltage spectra to that of the KLD control.

Figure 3.

Circular dichroism (CD) and transmission electron microscopy (TEM) analysis of self-assembly. (A) Schematic of polyanionic charge-conversion peptides that persist as random coils. Acid-induced lysine unmasking induces self-assembly via ionic crosslinking (depicted) and hydrophobic interactions into β-sheets. (B-D) Circular dichroism of (B) mal-KLD, (C) cit-KLD, and (D) dma-KLD before and after unmasking. KLD control peptide is shown for comparison. Progelators incubated at pH 3.0 for 24 hr, then dialyzed into 50 mM phosphate buffer (pH 7.4) for measurement at 400 μM peptide. (n=3 accumulations) (E-H) Dry state, stained TEM micrographs of progelators and unmasked progelators. (E) mal-KLD, (F) cit-KLD, and (G) dma-KLD progelators with inset chemical structures of maleamic acids before (top) and after (bottom) acid treatment. (H) KLD control with inset chemical structure of unmasked lysine amine. TEM samples (100 μM) were treated at pH 3.0 for 12 hr, then neutralized to pH 7.4 prior to imaging.

The morphology of dilute progelators and pH-induced SAPs was observed with stained dry state TEM (Figure 3E-G and Figure S9, Supporting Information). Both mal-KLD and cit-KLD progelators exhibited no distinct structures (Figure 3E,F), and dma-KLD showed low contrast staining of disordered structures (Figure 3G), which is likely the result of transient oligomers bound by weak hydrophobic interactions, in agreement with CD results. pH-induced unmasking of progelators yields elongated fibrillar networks identical to those of the KLD control peptide (Figure 3H). Interestingly, the TEM structures for each progelator and unmasked adduct with respect to increasing hydrophobicity, are reminiscent of the liquid-liquid phase-separation mediated nucleation-elongation mechanism for self-assembling nanofibrils (Figure S9, Supplementary Information).^[27] Through the different chemistries of our masks, observation of pseudo-intermediate structures for KLD control assembly may be possible.

2.4. Bulk Scale Unmasking Induces Significant Viscoelastic Changes.

The KLD control peptide is a viscoelastic, physical hydrogel with rehealable properties.^[28] We therefore, conducted tests to determine whether unmasking our progelators could recapitulate native gelling behavior. Peptides were incubated at pH 7.4, 6.8, 5.5, and 3.0 for up to 24 hr, then neutralized for rheological measurements (Figure 4). Neutralization enables direct comparison of cleavage kinetics through resulting mechanical analysis independent of pH effects. We note that rheology of KLD control shows no significant dependence at acidosis relevant pH values (~5.4–7.1) (Figure S10). Bulk samples show solid hydrogels, similar to the KLD control following acid-treatment (Figure 4A). Incomplete hydrolysis or potential trapping of the maleic and citraconic acid products is suspected to cause the residual absorbance in the visible spectrum. Acid-triggered self-assembly induces frequency independent viscoelastic properties with increasingly elastic-, as opposed to viscous-, dominant behavior. (Figure S11-13, Supporting Information). Resulting samples for all acid-treated progelators reveal viscoelastic hydrogels with storage moduli (G’) greater than (G”) (Figure 4B). Frequency independence for KLD control and unmasked cit-KLD and dma-KLD persists up to ~50 rad/s (~8 Hz). Conversely, mal-KLD is affected at frequencies exceeding 22 rad/s (3.5 Hz), at which point the three-dimensional network begins to break down. Given that physiologically relevant biological (human pulse 0.6–2 Hz) and activity frequencies (0.3–3.5 Hz)^[29] fall within the frequency independent domains of our hydrogels (Figure 4B), we reason that their relaxation times τ < 0.29 s are sufficient to maintain structural stability as biomaterial implants. Strain sweeps confirm that hydrogel measurements are appropriately measured within the linear viscoelastic region (LVR) (Figure S14, Supporting Information).

Figure 4.

Bulk rheological properties of pH-activated hydrogels and low viscosity progelators for catheter injection. (A) Image of masked progelator solutions, unmasked hydrogels, and KLD control. (B) Frequency sweeps of viscoelastic moduli, G’ and G”, for KLD control and unmasked hydrogels formulated from progelators treated at pH 3.0 for 24 hr, then neutralized to pH 7.4. Angular frequency 100–0.25 rad/s, 0.5% strain, n=3 repeats. Physiologically relevant frequency ranges for human heart rate (red) and normal activity levels (grey) are highlighted. (C) Storage moduli, G’, of peptides treated at pH 7.4, 6.8, 5.5, or 3.0 for 1, 12, or 24 hr. Angular frequency 2.5 rad/s, 0.5% strain. (n=3). Values are mean ± SEM. (D) Representative step-strain oscillations of dma-KLD treated with pH 3.0 for 24 hr, then neutralized to pH 7.4, demonstrate healing capacity. Angular frequency 2.5 rad/s. Measurements conducted with destructive strain at 100% for 3 min, and then regeneration at 0.5% strain for 15 min (n=3 cycles). Progelators persist as low viscosity solutions for catheter injection. (E-F) Viscosity of (E) progelator solutions and (F) acid-treated hydrogels. (G) Progelators solutions flow smoothly through the catheter at 0.6 mL/min. (H) Catheter injection setup with loaded syringe pumping progelators through a catheter submerged in a 37 ˚C water bath. All measurements performed on neutralzed peptide samples in 1x DPBS (pH 7.4) at 10 mM.

The sensitivity of each progelator under concentrated conditions, mimicking that of an injected dose, was tested at pH 7.4, 6.8, 5.5, and 3.0 for 24 hr (Figure 4C). Additionally, each progelator was treated for 1 and 12 hr at corresponding acidities tested in Figure 2C, where significant changes to responsiveness for each maleamic acid have been reported.^[6] One notable difference, is that pH 6.8 instead of 6.5 was used to test increasingly mild conditions for dma-KLD sensitivity. An increase in storage moduli, G’, is observed with increasing acidity and longer treatment times. Furthermore, the stiffest hydrogels were formed by dma-KLD, which possesses the most labile maleamic acid in this study. Indeed, viscoelastic properties for the dma-KLD treated at pH 3.0 for 24 hr resulted in viscoelastic moduli on par with the KLD control, indicative of complete maleamic acid removal. ESI of treated dma-KLD gels reveal partial and completely unmasked products (Figure S15, Supporting Information). Thus, the dma-KLD responded well at physiologically relevant conditions observed in mild tissue inflammation. We note that bulk scale cleavage is likely slowed, due to decreased solvent diffusion as macromolecular self-assembly occurs. We also suspect that in vitro gelation kinetics may differ from those in vivo,^[20] as local salt and pH gradients exist in inflamed tissues. Furthermore, storage moduli of many physical hydrogels are known to be increased to that of native tissue when seeded with cells^[28] or implanted in vivo.^[30] Regardless, these experiments confirm gelation sensitivity can be tuned with simple modification of the charge-conversion moiety.

Finally, it is critical that injectable hydrogels can withstand network disruption. As with many SAPs used for tissue engineering, they are susceptible to hydrogel disruption under high strain but exhibit rehealable behavior.^[31] Step-strain oscillations (Figure 4D and Figure S16, Supporting Information) were performed for each unmasked hydrogel by applying continuous destructive strain (100 %) for 3 min, resulting in liquid-like solutions with G’<G”. Once excess strain was removed, rapid hydrogel regeneration was observed (crossover G’=G” occurs within seconds). Repeated healing of 100% or more was achieved after 900s for both cit-KLD and dma-KLD over several cycles. Furthermore, 80% G’ recovery was achieved within ~10s for each cycle, demonstrating rapid healing kinetics. We note that slow healing kinetics of mal-KLD is likely a result of incomplete progelator unmasking, rendering that progelator an unlikely candidate for clinical application.

2.5. Low Viscosity Progelators are Amenable to Catheter Injection.

Needle-based injections require that materials are shear-thinning. Catheter-based injections further require that the material have low enough viscosity to flow through the catheter during delivery, a process which can take up to 1 hr in a clinical setting.^{[20, 32]} Many physical hydrogels are shear-thinning and rehealable, but aging time can vary from different relaxation and interpenetration properties between fractured domains.^[33] Alternatively, excess viscosity from reassembly in the catheter can stop material delivery altogether. A material that gels after delivery can bypass these issues.

Progelators in this study were over two orders of magnitude lower in complex viscosity than the KLD control gelator, as well as shear-thinning (Figure 4E-F). Unmasking triggered an increase in viscosity. Furthermore, no temperature-dependent changes in viscoelastic properties were observed when samples were slowly heated from 21–37 ˚C (Figure S17, Supporting Information), representative of the temperature variations generally experienced by biomaterials during catheter delivery. Finally, progelators in this study demonstrated no resistance to delivery by hand or catheter failure when pumped at clinically relevant rates (0.6 mL/min) (Figure 4G-H), whereas excess viscosity from the KLD control peptide at 10 mM in 1x DPBS caused catheter failure in vitro (Movie S2, Supporting Information). Thus, we have successfully designed an acid-activatable material for simple formulation and noninvasive delivery. Given the wider inner diameter of clinical over-the-wire infusion catheters (e.g. EmergeTM PTCA, 0.36 mm), we reason that our low viscosity materials are amenable to both transendocardial injection and intracoronary infusion cardiac catheter delivery.

2.6. Progelators are Hemocompatible in Blood Components.

Lack of hemocompatibility can be a limiting factor for injectable biomaterials, which must demonstrate inert activity within the blood during direct (intravenous injection and catheter infusion) or indirect (subcutaneous and intramuscular injection leakage) contact with the bloodstream.^[34] Potential blood interactions may cause thrombosis, induce hemolysis, or alter coagulation kinetics. We incubated our charge-conversion peptides at increasing dosages in whole human blood components up to 1:10 dosing (Figure 5). For reference, generous estimates for clinically relevant doses from transendocardial catheter injection and intracoronary infusion are 1:830 and 1:500, respectively (see Supporting Information). The hemolytic properties of our progelators and the KLD control in isolated red blood cells (RBCs) were assessed for acute toxicity (Figure 5A and Table S2, Supporting Information). Excess doses up to 1:10 for all peptides revealed <5% hemolysis, which is below the limit for consideration as a hemolytic material, according to ASTM F756–17.^[35] Furthermore, all clinically relevant doses are considered nonhemolytic (<2% hemolysis).

Figure. 5.

Hemocompatibility of peptide progelators. (A) Percent hemolysis of human red blood cells (RBCs) after incubation with different concentrations of peptide for 1 hr. Inset line defines the threshold for hemolytic response (> 5%). (n=6) (B) Activated clotting times of whole human blood in the presence of different peptide concentrations. (n=5) (C) Plasma recalcification profiles in platelet poor plasma (PPP) for vehicle standard, and collagen, glass coverslip, and no calcium controls. (n=6) (D) Measurements of coagulation maximal extent, rate, and onset time from plasma recalcification profiles as a function of peptide concentrations. (n=6) All peptide dilutions in biological fluid at 37 ˚C are provided in figures at increasing peptide in blood concentration from 1:10,000, 1:5000, 1:1000, 1:500, 1:100, and 1:10 volume ratios, given an injection concentration of 10 mM stock solution in 1x DPBS. ns (p > 0.05), * (p ≤ 0.05), ** (p ≤ 0.01), *** (p ≤ 0.001), and **** (p ≤ 0.0001). Values are mean ± SEM. Corresponds with Figure S13 and Table S2–6 in Supporting Information.

Activated clotting times (ACT) were used as a general method that encompasses intrinsic and common coagulation pathways to assess thrombogenicity during delivery (Figure 5B and Table S3, Supporting Information). This assay is the preferred test in catheterization labs and cardiac theatres.^[36] Anticoagulative properties were observed at 1:100 and 1:10 blood doses for mal-KLD and cit-KLD, and to a much lesser extent at 1:10 with dma-KLD, indicating a minor effect on the intrinsic coagulation (contact-based) pathway. The KLD control peptide exhibited no significant alteration to normal clotting times over that of the vehicle standard at all doses. In contrast, collagen showed a thrombogenic effect, and chelation of calcium prevented clotting altogether.

Finally, plasma recalcification profiles in platelet poor plasma (PPP) were used to study the intrinsic pathway of clotting, which can reveal adverse blood-biomaterial interactions (Figure 5C-D and Table S4–6, Supporting Information). Collagen, which activates platelet aggregation, showed no early onset of coagulation in the presence of platelet-poor media (Figure 5C). Glass coverslips, which have negatively charged and hydrophilic surfaces, were used as a positive control of contact activation, whereby coagulation onset time was reduced. As a negative control, PPP without Ca²⁺ showed no onset of coagulation. Both mal-KLD and cit-KLD exhibited earlier onset times and marked decrease in extents of coagulation (Figure 5D and Figure S18, Supporting Information). Furthermore, cit-KLD caused a faster rate of coagulation than that of the other peptides. However, dma-KLD and KLD control peptides exhibited minimal deviation from vehicle standard coagulation profiles, with no significant difference between each other at all doses. Given that aberrant effects on hemocompatibility were observed at progelator doses (1:10 and 1:100) that are higher than would theoretically be present in the blood after 1 min of circulation, assuming all material entered into the bloodstream and not the intended tissue, masking has no relevant impact on hemocompatibility.

2.7. Cell Viability of Progelators and Unmasked Products.

Once the peptide has reached its intended acidic tissue target, unmasked KLD control can self-assemble as a local tissue scaffold. This peptide is known to be biocompatible, but the new chemistries presented in this manuscript are not. We assayed cell viability using CellTiter-Blue reagent by incubating progelators (mal-KLD, cit-KLD, and dma-KLD) and free masks (maleic, citraconic, and dimethylmaleic acid) with L-929 cells (mouse, subcutaneous connective tissue) to understand their cytotoxicity (Figure 6). For each concentration tested, the amount of free maleic acid was 3x higher than the progelator concentration to mimic expected release concentrations after cleavage. No changes in buffered media pH were detected. No statistical significance was observed between peptides and the vehicle (1x DPBS) up to 1000 μM after 24 hr incubation (Figure 6A). Similarly, free maleic acids studied at stoichiometrically matched concentrations showed the same result (Figure 6B). In contrast, 10% DMSO exhibited ~19% viability.

Figure 6.

Cell viability of progelator components and drug encapsulation. (A) Cell viability of mal-KLD, cit-KLD, dma-KLD, and KLD control incubated at 0, 7.8 15.6, 31.2, 62.5, 125, 250, 500, and 1000 μM peptide in 1x DPBS (pH 7.4) for 24 hr. (n=3) (B) Cell viability of mal anhydride, cit anhydride, and dma anhydride incubated at 0, 23.4, 46.8, 93.6, 187.5, 375, 750, 1500, and 3000 μM in 1x DPBS (pH 7.4) (n=3). (C) Encapsulation of model drug into fiber network. Normalized fluorescence (F/F_o) of ThT (50 μM) and progelator (500 μM) after incubation at pH 7.4 6.8, 5.5, and 3.0 for 10 min. (n=4). Inset chemical structure of ThT. ns (p > 0.05) and **** (p ≤ 0.0001). Values are mean ± SEM.

A simple strategy for generating structurally dynamic stimuli-responsive peptide-based hydrogels is developed. By utilizing charge-conversional masking groups, zwitterionic gelator peptides can be formulated as soluble, polyanionic progelators for initial catheter or needle-based delivery. Tunable responsiveness to an acidic environment regenerates the initial gelator, providing a scalable strategy for implementing biodegradable therapeutic hydrogels that assemble at sites of disease-related tissue acidosis.

2.8. Self-assembly Induced Encapsulation of a Model Drug.

General strategies for drug delivery using SAPs rely on physical entrapment of small molecules, protein/nucleic acids, and cell therapies into the scaffold prior to injection.^[37] We reason that acid-induced unmasking and simultaneous self-assembly imbues our progelators with the capacity for activatable drug encapsulation. Thioflavin T (ThT) was chosen as a model drug for its stability to acidic pH and fluorogenic capacity to quantitatively detect its physical entrapment .^[38] Serial addition of ThT to KLD control peptide shows a monotonic increase in normalized fluorescence up to 50 μM (Figure S19, Supporting Information). Furthermore, serial dilution of KLD in a fixed concentration of ThT (50 μM) revealed a critical aggregation concentration (CAC) at pH 5.5, 6.8, and 7.4 of 80.0 (± 3.1) μM peptide (Figure S20, Supporting Information). Incubation for 10 min of progelators (500 μM) above their CAC in buffers at associated activation pH values in Figure 4C reveal fluorogenic turn-on as ThT (50 μM) is incorporated into the growing networks (Figure 6C). Finally, coincubation of the solvatochromatic dye, Congo Red, with progelator provides a visual demonstration of small molecule sequestration (Figure S21, Supporting Information). Thus, progelator unmasking recapitulates physical entrapment behavior exhibited by the KLD control. We envision that self-assembly induced encapsulation of therapeutics can provide a facile strategy for localizing treatment through simple co-delivery.

3. Conclusions

We present a straightforward approach to reversible modification of a self-assembling peptide using a one-step synthesis with quantitative yields. The pH-sensitivity of these progelators is readily tuned with substituted maleamic acid moieties, enabling assembly at physiologically relevant tissue acidities (pH ~ranging from 5.4–7.1). Thus, our platform utilizes disparate advantages of both soluble small molecules (injectability and tissue perfusion) and macromolecular hydrogels (stationary support and drug encapsulation). We reason that our materials would have the capacity to spread unhindered until activated by the acidic extracellular microenvironment of inflamed tissue. This structurally dynamic behavior is especially useful in wound-healing applications where the in vivo architectures are tortuous, and in some instances require navigation through narrow pathways prior to self-assembly as a stationary hydrogel. We show that progelator unmasking leads to solidification with increased spreading resistance; the unmasked hydrogels are also able to withstand physiologically relevant activity frequencies and/or reheal following excess strain. We demonstrate that masking with maleic acids has no relevant impact on hemocompatibility at clinical doses. Finally, a general trend presented itself as improved hemocompatibility with increasing progelator hydrophobicity in all clotting and coagulation assays. Additionally, no cytotoxicity was observed with RBCs or L-929 cells.

There is a need for synthetically simple tissue engineering scaffolds that can be delivered non-invasively to injury sites. In many extreme cases, inflammation is not localized (e.g. arthritis, myocardial ischemia, traumatic injury), so single injections of a preformed hydrogel to the site of interest may not be practical. Our straightforward design strategy enables a stimuli-responsive solution that flows freely until activated by inflammation-associated extracellular acidosis. Studies to test tissue accumulation of these materials in injury/inflammation models are underway in our laboratories.