Degradable and Multifunctional PEG-Based Hydrogels Formed by iEDDA Click Chemistry with Stable Click-Induced Supramolecular Interactions

Abstract

The inverse electron demand Diels–Alder (iEDDA) reactions are highly efficient click chemistry increasingly utilized in bioconjugation, live cell labeling, and the synthesis and modification of biomaterials. iEDDA click reactions have also been used to cross-link tetrazine (Tz) and norbornene (NB) modified macromers [e.g., multiarm poly(ethylene glycol) or PEG]. In these hydrogels, Tz-NB adducts exhibit stable supramolecular interactions with a high hydrolytic stability. Toward engineering a new class of PEG-based click hydrogels with highly adaptable properties, we previously reported a new group of NB-derivatized PEG macromers via reacting hydroxyl-terminated PEG with carbic anhydride (CA). In this work, we show that hydrogels cross-linked by PEGNB_CA or its derivatives exhibited fast and tunable hydrolytic degradation. Here, we show that PEGNB_CA (either mono- or octafunctional) and its dopamine or tyramine conjugated derivatives (i.e., PEGNB-D and PEGNB-T) readily cross-link with 4-arm PEG-Tz to form a novel class of multifunctional iEDDA click hydrogels. Through modularly adjusting the macromers with unstable and stable iEDDA click-induced supramolecular interactions (iEDDA-CSI), we achieved highly tunable degradation, with full degradation in less than 2 weeks to over two months. We also show that secondary enzymatic reactions could dynamically stiffen these hydrogels. These hydrogels could also be spatiotemporally photopatterned through visible light-initiated photochemistry. Finally, the iEDDA-CSI hydrogels post ester hydrolysis displayed shear-thinning and self-healing properties, enabling injectable delivery.

Introduction

The inverse electron demand Diels–Alder (iEDDA) reaction is a class of highly efficient click chemistry between an electron-poor diene (e.g., tetrazine or trans-cyclooctyne) and an electron-rich dienophile (i.e., norbornene).^1,2 Owing to their fast reaction kinetics and high biocompatibility, iEDDA click reactions have been utilized for bioconjugation, live cell labeling, and hydrogel cross-linking.³⁻⁶ The Anseth group pioneered the use of iEDDA click reaction to cross-link poly(ethylene glycol) (PEG)-based hydrogels. Specifically, four-arm PEG-tetrazine (PEG4Tz) was modularly cross-linked with bis-norbornene (NB) containing peptide in an aqueous buffer,⁷ yielding hydrogels with spontaneous, mild, and fast gelation kinetics, and high cytocompatibility for in situ encapsulation of human mesenchymal stem cells (hMSCs). Due to NB’s dual reactivity toward free sulfhydryl (i.e., thiol) and Tz groups, the Tz-NB hydrogels were also amenable for postgelation patterning via secondary thiol-NB click reaction. Tz-NB click reaction has also been exploited for the dynamic stiffening of thiol-NB hydrogels.⁸ For example, our lab encapsulated human induced pluripotent stem cells (hiPSCs) into thiol-NB click hydrogels with excess NB groups, enabling dynamic stiffening via simple diffusion of multiarm PEG-Tz.⁹ We showed that the stiffened hydrogels were beneficial for the definitive endoderm differentiation of hiPSCs. The dual reactivity of NB-functionalized macromers was also leveraged for annealing thiol–ene cross-linked microgels to form microporous annealed particle scaffolds (MAPs).^10,11

Typical Tz-NB iEDDA click reaction produces stable cycloaddition products with nitrogen as the only byproduct. Alge and colleagues discovered that, compared with thiol–ene photoclick hydrogel counterparts, the iEDDA click hydrogels exhibited lower swelling ratios, higher storage moduli, and high resistance toward base-induced hydrolytic degradation.¹² They attributed these unique characteristics to the formation of strong supramolecular interactions between the cycloaddition products and validated this hypothesis via experimental and molecular dynamic simulations. Later, the same group leveraged these supramolecular interactions for noncovalent stiffening of a precross-linked thiol–ene click hydrogel network. Briefly, methoxy PEG tetrazine was used to form pendant Tz-NB click adducts, which assembled via supramolecular interactions and increased hydrogel storage moduli (G′ ∼ 2 to 6 kPa).¹³ The stiffening process appeared to be cytocompatible for 3T3 fibroblasts, although no functional demonstration of the stiffening-induced cell fate process was demonstrated.

Conventional PEG-based Tz-NB iEDDA-cross-linked hydrogels display unprecedented resistance against hydrolytic degradation. The Goepferich group observed that iEDDA click hydrogels cross-linked by 8-arm PEGTz and PEGNB did not undergo noticeable hydrolytic degradation until within 100–400 days, depending on the formulations.¹⁴ Even with hydrolytically labile groups (e.g., a phenyl carbonate functional group at the cross-linking site) installed in the polymer backbone, the fastest in vitro degradation time achieved with 8-arm PEGs was approximately 150 days.¹⁵ Further, utilizing lower functionality PEG did not accelerate the hydrolytic degradation time.¹⁶ However, hydrogels cross-linked by the iEDDA click reaction can still be rendered more hydrolytically labile using a degradable polymer backbone. For example, Lueckgen et al. induced hydrolytically labile sites on the alginate backbone through a series of oxidation and reduction reactions.¹⁷ The hydrolytically degradable alginate was then functionalized with NB or Tz groups, and the resulting hydrogel was degraded after 12 days under low norbornene substitution and 5% oxidation. Nonetheless, it will be beneficial to prepare PEG-based iEDDA click hydrogels with highly programmable hydrolytic degradation kinetics.

Recently, our group fabricated a new class of hydrolytically degradable PEG-based hydrogels based on a modified PEGNB, which was synthesized by reacting PEG-hydroxyl with carbic anhydride (CA) via a ring-opening cyclic desymmetrization reaction.¹⁸⁻²⁰ The new macromer, termed PEGNB_CA, underwent fast and efficient gelation with thiol-based cross-linkers via thiol–ene chemistry.¹⁹ Notably, the PEGNB_CA macromer accelerated the hydrolytic degradation of thiol–ene hydrogels, reducing nearly 80% of its initial G′ over 8 days. Surprisingly, when further reacting the carboxylic acid group of the PEGNB_CA with an amino group-containing molecule, the resulting macromer (i.e., PEGNB-X, X = Dopamine, Tyramine, Isopropyl amine, etc.) produced thiol–ene hydrogels with an even faster hydrolytic degradation, with complete dissolution of the gels occurring within hours to a few days.¹⁹ These unique macromers (e.g., PEGNB-D) were subsequently exploited as a sacrificial material to create macroporous hydrogels for facilitating the assembly of solid or hollow cell spheroids.²⁰

The new PEGNB_CA macromer could also be cross-linked with PEG-Tz via iEDDA click reaction to form hydrogels with accelerated hydrolytic degradation.¹⁸ Interestingly, PEGNB_CA-based hydrogels degraded much slower when cross-linked with methyltetrazine (mTz) functionalized PEG (PEG-mTz), with complete degradation occurring after 80 days of aqueous incubation. The degradation kinetics of this new class of (m)Tz-NB hydrogels were governed by the balance between the ester-destabilizing carboxylic acid group on the NB_CA moiety and the hydrophobicity of the methyl group on mTz. In this contribution, we characterized the cross-linking of the PEGNB-X macromers with 4-arm PEGTz. Unlike PEGNB-X hydrogels cross-linked via thiol–ene chemistry that degraded rapidly, the tetrazine-ene PEGNB-X hydrogels remained stable after 24 h of incubation in an aqueous buffer. By combining the unique properties of PEGNB_CA, PEGNB-X, and the supramolecular Tz-NB adducts, we investigated the cross-linking and degradation of various click hydrogels. Furthermore, we exploited PEGNB-X-based iEDDA click-induced supramolecular interactions (iEDDA-CSI) to cross-link hydrogels with tunable stiffness, enzymatic stiffening, and programmable softening (Table 1). Of note, the term “iEDDA-CSI” was used to describe hydrogels initially cross-linked by iEDDA click reaction, but their stability relied on click-induced supramolecular interactions. The hydrogel system also provided inherent reactive handles for photopatterning. We envision this hydrogel system as dynamic hydrogel matrices for 3D cell culture.

Table 1. Summary of the Macromers Used, Crosslinking Mechanisms, and the Rate of Ester Hydrolysis. All Were Crosslinked with PEG4Tz via iEDDA Click Reactions^a.

macromer	cross-linking method	stabilization mechanisms	hydrolysis rate
PEG8NB	iEDDA	covalent + supramolecular interactions	low
PEG8NBCA	iEDDA	covalent + supramolecular interactions	high
PEG8NB-X	iEDDA	supramolecular interactions	highest
mPEGNB	iEDDA-CSI	supramolecular interactions	low
mPEGNBCA	iEDDA-CSI	supramolecular interactions	high
mPEGNB-X	iEDDA-CSI	supramolecular interactions	highest

^aNote: X = Dopamine, Tyramine, Isopropyl amine.

Results and Discussion

Comparison of Hydrogel Cross-Linking Using PEGNB-X Macromers–Thiol–ene vs iEDDA

Despite being synthesized via different routes (Figures 1A and S1–S5), all NB-modified PEGs, including conventional PEGNB, the new PEGNB_CA, and its derivatives PEGNB-X, can all be cross-linked into similar thiol–ene hydrogels with thiol-ether bonds (Figure 1B).¹⁹ High degrees of substitutions were achieved for 8-arm PEG8NB (∼99%), PEG8NB_CA (∼96%), and 4-arm PEG4Tz (∼85%) (Table S1). The conjugation efficiencies of PEG8NB-X were also high (∼96%) but some NB groups were lost during the secondary reaction (Table S1). The overall NB conjugation efficiencies for PEG8NB-D and PEG8NB-T dropped to around 75%. The NB conjugation efficiency for PEG8NB-I was lower at ∼64%. In situ photorheometry demonstrated ultrafast gelation for PEG8NB-D with PEG4SH, with a gel point of around 4 s. When monofunctional PEGNB-D (i.e., mPEGNB-D) was used, no G′/G′’ crossover was observed over the 4 min light exposure period (Figures 1C and S6). As the average functionality of mPEGNB-D and PEG4SH was 2.5, no gelation of this system can be expected according to the Flory–Stockmayer theory, which dictates that gelation occurs only with the average functionality of the macromers is equal to or greater than three.²¹ NB can also spontaneously react with Tz and their derivatives via iEDDA click reaction, forming hydrogels with Tz-NB adducts (Figure 1B)^1,2 that could strengthen the polymer networks.^12,13 The iEDDA click-based cross-linking was considerably slower than that of the thiol–ene counterparts (gel point ∼3 min. Figures 1D and S6). However, unlike in the thiol–ene cross-linking, mPEGNB-D could form hydrogels with PEG4Tz after 5 min (Figures 1D and S6), confirming the existence of noncovalent interactions between the adducts of Tz and NB (or NB-D). The reaction kinetics of mPEGNB-D with PEG4Tz was slightly faster compared with that using PEG8NB-D (Figure S7). The slower gelation time and the lower plateau modulus of mPEGNB-D could be attributed to the lack of covalent cross-links afforded by multiarm macromers (e.g., PEG8NB-D). Our results agree with previous reports where 1-arm PEGTz can be used to cross-link with 4-arm PEGNB and form hydrogels.^12,13 However, to our surprise, hydrogels cross-linked by PEG8NB-D and PEG4Tz remained intact, at least qualitatively, after 24 h (Figure 1E). This finding starkly contrasted our prior results where thiol-norbornene photo-cross-linking of PEG8NB-D and PEG4SH produced rapidly dissolving hydrogels,¹⁹ thus warranted further investigation of the cross-linking, stabilization, and degradation of iEDDA click hydrogels formulated by PEG8NB-D with PEG4Tz.

Figure 1.

Overview of synthesis and cross-linking of PEGNB, PEGNB_CA, and PEGNB-X macromers via thiol-norbornene or tetrazine-norbornene reactions. (A) Schematic of norbornene functionalization onto PEG-hydroxyl through either Steglich esterification with norbornene acid or ring-opening ester conjugation using carbic anhydride, yielding (m)PEG(8)NB and (m)PEG(8)NB_CA, respectively. PEGNB_CA can be further conjugated with different amines, yielding (m)PEG(8)NB-X (i.e., (m)PEG(8)NB-D, (m)PEG(8)NB-T, and (m)PEG(8)NB-I). (B) Cross-linking (m)PEG(8)NB_CA or (m)PEG(8)NB-X with PEG4SH via photoinitiated thiol–ene reaction (top) or with PEG4Tz via spontaneous iEDDA reaction (bottom). (C) In situ photorheometry of PEG8NB-D and mPEGNB-D with PEG4SH in the presence of 2 mM LAP and 365 nm light exposure at 19 mW/cm² intensity. (D) In situ rheometry of PEG8NB-D and mPEGNB-D with PEG4Tz at 37 °C for 30 min. (E) Pictures of hydrogel precursor solution containing 5 wt % PEG8NB-D with PEG4SH and 2 mM LAP or with PEG4Tz at R([Tz]/[NB]) = 1 before cross-linking, posthydrogel cross-linking (i.e., 2 min 365 nm exposure at 3 mW/cm² for thiol–ene or 1 h incubation at 37 °C for tetrazine-ene), and after 24 h of incubation in pH 7.4 PBS at 37 °C. Thiol–ene hydrogels completely degraded, while iEDDA hydrogels were structurally intact after 1 day.

Confirmation of Ester-Hydrolysis and the Presence of Noncovalent Interactions in iEDDA Click Hydrogels

As functional groups attached to the electron-rich alkene have been shown to affect the rates of iEDDA click reaction,²² we conducted a kinetic study to further assess the cross-linking of iEDDA click hydrogels with PEG4Tz and PEG with different NB moieties (i.e., NB, NB_CA, and NB-D) (Figure 2A). Indeed, compared with conventional PEG8NB, the attachment of electron-withdrawing carboxylic acid (on PEG8NB_CA) and dopamine (on PEG8NB-D) significantly reduced the iEDDA reaction rate constant (k). The reaction rate of PEG4Tz with PEG8NB-D was higher than that with PEG8NB_CA, warranting further investigation of hydrogels cross-linked by PEG8NB-D. We next compared the properties of hydrogels cross-linked by PEG8NB-D with PEG4Tz via iEDDA click reaction and with PEG4SH via thiol-NB photoclick reaction. Interestingly, thiol-NB hydrogels exhibited swollen moduli of approximately 19 kPa, a value significantly lower than that of PEG4Tz/PEG8NB-D hydrogels (G′ ∼ 25 kPa) at the same functional group concentrations (Figure 2B). We attributed this to the additional noncovalent interactions between Tz/NB-D adducts.^12,13 We also tracked the reduction of moduli of these hydrogels over time and fitted the results with pseudo-first-order ester hydrolysis kinetics (Figure 2C). We found complete degradation of thiol-NB hydrogels occurred in 7.5 h with a best-fit k_hydrolysis of 0.233 h^–1 (Table S2). In contrast, Tz/NB-D iEDDA hydrogels degraded much more slowly, with a best-fit k_hydrolysis value of 0.093 h^–1 (Table S2). The high R² values for both fittings (0.95 and 0.98 for thiol-NB and Tz-NB reaction, respectively) suggest a high probability that the degradation within this period of time was mediated by hydrolysis of the ester bonds on PEG8NB-D. After 21 h, however, the degrading PEG4Tz/PEG8NB-D iEDDA hydrogels reached a plateau modulus, around 30% of its initial G′ (Figure 2C), indicative of another mechanism that stabilized the hydrogel network following the substantial and rapid degradation. We reasoned that the reduction of iEDDA click hydrogel moduli was due to ester hydrolysis of the PEG8NB-D macromer. To verify this, we first evaluated the gel fractions of iEDDA click cross-linked PEG4Tz/PEG8NB-D hydrogels, with conventional PEG8NB as a control (Figure S8). Regardless of incubation conditions (DI water or PBS), gel fractions of the control PEGNB hydrogels were all close to one, indicating the efficient cross-linking and the lack of hydrogel degradation from the conventional PEG4Tz/PEG8NB hydrogels. PEG4Tz/PEG8NB-D hydrogels incubated in DI water unexpectedly remained stable within 2 days of incubation, as indicated by a near 100% gel fraction. However, when the same hydrogels were swelled in pH 7.4 PBS at 37 °C for 2 days, the gel fraction decreased from nearly 1.0 to ∼0.8. After 14 days of incubation in PBS, the gel fraction of these hydrogels further decreased to ∼0.68, whereas that of the PEG8NB hydrogels remained at ∼1 (Figure S9). Additionally, iEDDA hydrogels cross-linked by PEGNB maintained a swollen mass of approximately 80 mg during a 14 day period with no substantial changes (Figure S10). In contrast, PEG8NB-D hydrogels initially had a swollen mass of approximately 95 mg but gradually decreased to a plateau value of about 80 mg over the same 14 day period. Lastly, the swelling ratio of PEG4Tz/PEG8NB-D iEDDA click cross-linked hydrogels was significantly higher than that of the PEG4Tz/PEG8NB hydrogels (Figure S11). These results substantiated the hydrolytic degradation of the PEG4Tz/PEG8NB-D hydrogels (Figure 2C).

Figure 2.

Cross-linking and degradation of hydrogels formed by variants of PEG-norbornene. (A) Reaction rate constants of PEG8NB, PEG8NB_CA, and PEG8NB-D as determined by photospectrometry. (B) Starting storage moduli (G′) of 5 wt % 8-arm PEG8NB-D hydrogels cross-linked with either PEG4SH (thiol-norbornene reaction) or PEG4Tz (tetrazine-norbornene reaction). Reactive group stoichiometric ratios (R = [Tz]/[NB]) were set at 1. (C) Ester hydrolysis and degradation of iEDDA hydrogels cross-linked by PEG8NB-D with either PEG4SH (thiol-norbornene reaction) or PEG4Tz (tetrazine-norbornene reaction). Changes in G′ were normalized to the starting modulus over a 24 h period. Curve fitting represents best-fit pseudo-first-order kinetics of ester hydrolysis. (D) ¹H NMR spectra of PEG8NB-D, PEG8OH, and leaving degradation products collected through dialysis. (E) ATR-FTIR spectra of PEG8NB-D iEDDA click cross-linked hydrogels before and after hydrolytic degradation.

We hypothesized that the substantial degradation of the PEG4Tz/PEG8NB-D hydrogels was accompanied by the release of the degradation product PEG–OH from the remaining hydrogel network. To test this hypothesis, we isolated and characterized the degradation products via ¹H NMR and ATR-FTIR. As shown in Figure 2D, the polymers collected from the degrading PEG4Tz/PEG8NB-D hydrogels shared identical NMR spectra to that of unmodified PEG–OH. Additionally, ATR-FTIR analysis showed the formation of a new peak in the carbonyl stretch region (∼1700 cm^–1) after the hydrogels had been incubated in PBS,²³ indicative of the emergence of the new carboxylic acid group following ester hydrolysis on the PEG8NB-D macromer (Figure 2E). These results confirm the proposed hydrolysis mechanism of the PEG4Tz/PEG8NB-D iEDDA cross-linked hydrogels (Figure 2C). We reasoned that the stabilization of the degraded iEDDA click hydrogel was due to noncovalent supramolecular interactions formed between the remaining Tz/NB-D adducts. Further investigation is needed to exploit this unique behavior.

Cross-Linking and Degradation of iEDDA Click Cross-Linked Hydrogels

Exploiting the new iEDDA click hydrogels for biomedical applications requires that these hydrogels’ cross-linking and hydrolytic degradation behaviors be comprehensively characterized. As shown in Figure S12, increasing the cross-linking time of PEG4Tz/PEG8NB-D hydrogels from one to four hours led to moderate increases of G′, from ∼23 to ∼29 kPa. Regardless of the initial moduli, however, all hydrogels degraded rapidly and substantially after just 1 day of incubation in PBS, and prolonged incubation for six additional days did not lead to further degradation. To further assess the cross-linking and degradation behaviors of these new iEDDA click hydrogels, we employed PEG4Tz and different norbornene variants (i.e., PEG8NB, PEG8NB_CA, and PEG8NB-D). The total concentration of Tz-NB adducts (equal molarity of Tz and NB variant) was kept at 8 to 32 mM. As shown in Figure S13, increasing [Tz-NB] led to higher G′ for all NB variants (∼3.5 to ∼80 kPa), with PEG8NB-cross-linked iEDDA click hydrogels exhibiting the highest moduli. Except for a low [Tz-NB] of 8 mM, PEG8NB-D hydrogels were all slightly stiffer than PEG8NB_CA hydrogels. After 2 days of incubation, G′ for PEG8NB and PEG8NB_CA hydrogels stiffened slightly (i.e., G′/G′₀ = ∼ 1.2 to ∼1.4) at 8 mM [Tz-NB] and softened slightly (i.e., G′/G′₀ = ∼ 0.6 to ∼0.9) at higher [Tz-NB] concentrations (Figure S14). For PEG8NB-D hydrogels, the storage moduli dropped for all [Tz-NB] concentrations after 2 days, drastically dropping about 75% of their initial moduli (Figure S14). The vastly different degradation behaviors of PEG8NB-D hydrogels compared to the PEG8NB and PEG8NB_CA counterparts prompted us to track the changes of G′ over an extended period (Figure 3A–C). Despite utilizing the same iEDDA click cross-linking mechanism, the identity of NB macromer yielded hydrogels with vastly different stability. Specifically, Tz-NB hydrogels cross-linked by conventional PEG8NB and PEG4Tz remained largely stable over 42 days (Figure 3A), a result consistent with the literature.¹⁴⁻¹⁶ The moderate drops of G′ in gels at higher macromer concentrations were likely due to swelling. In contrast, all iEDDA hydrogels cross-linked by PEG8NB_CA degraded completely in less than 3 weeks (Figure 3B), which was attributed to the ester-destabilizing carboxylic acid on NB_CA.¹⁸ In this case, the previously reported Tz-NB supramolecular interactions did not prevent the disintegration of the hydrogels, suggesting a weakening of the supramolecular interactions between Tz and NB_CA adducts. Finally, all iEDDA hydrogels cross-linked by PEG4Tz and PEG8NB-D degraded rapidly for the first 24 h of incubation but remained exceptionally stable afterward (Figure 3C). Similar trends were observed with the other PEG8NB-X macromers (i.e., PEG8NB-I and PEG8NB-T. Figure S15). These results suggest that the additional amide conjugation aided the supramolecular interactions between Tz/NB-X adducts. The stability of the PEG8NB-D cross-linked iEDDA click hydrogels was attributed to higher binding strength between Tz/NB-D adducts; however, additional testing is needed to confirm this. Remarkably, the stability of the PEG8NB-D cross-linked iEDDA click hydrogels outperformed other previously established PEG8NB ester and amide thiol–ene based hydrogel systems by maintaining the moduli of soft hydrogels (i.e., G′<500 Pa) over a 42 day period.^24,25

Figure 3.

Characterization of iEDDA click hydrogel cross-linking and degradation. (A–C) Changes in G′ of iEDDA hydrogels cross-linked by PEG4Tz and PEG8NB (A), PEG8NB_CA (B), and PEG8NB-D (C). (D) Starting G′ of PEG4Tz/mPEGNB-D hydrogels cross-linked at different concentrations of [Tz-NB-D]. (E) Changes in G′ of mPEGNB-D iEDDA-CSI hydrogels at different [Tz-NB-D]. (F) Changes in G′ of mPEGNB-D hydrogels compared to mPEGNB_CA and mPEGNB. (G) Tuning the degradation behavior through different mPEGNB-D/mPEGNB_CA blends at different total [Tz-NB] concentrations. A higher [NB-D] leads to a higher iEDDA-CSI that stabilizes the hydrogel, while a higher [NB_CA] disrupts the supramolecular interactions between the Tz-NB adducts and weakens the hydrogel.

As the iEDDA cycloaddition products between PEG4Tz and PEG8NB-D exhibited strong iEDDA click-induced supramolecular interactions (iEDDA-CSI) that stabilized the otherwise completely degradable hydrogels (Figure 3C), we sought to investigate whether iEDDA-CSI hydrogels could be formed solely by supramolecular interactions. To this end, we employed monofunctional PEGNB-D (mPEGNB-D), which cannot serve as a covalent cross-linker. Any gelation using this macromer would indicate the presence of strong iEDDA-CSI. As shown in Figure 3D, we successfully synthesized iEDDA-CSI hydrogels of varying moduli using mPEGNB-D and PEG4Tz (G′ ∼ 2 to 8 kPa). However, unlike their multiarm PEGNB-D counterparts, there was no initial drop in G′ for mPEGNB-D cross-linked hydrogels. Instead, all hydrogels slightly stiffened to about 50% of their initial moduli, followed by slight softening back to their initial moduli over the 42 day period (Figure 3E). The initial stiffening effect could be caused by dopamine oxidation and additional cross-linking, while the decrease in storage moduli might be because of the instability of DOPA-quinone linkages over time.²⁶ Alternatively, other studies have shown initial stiffening occurring before degradation within PEG hydrogel systems.^27,28 iEDDA-CSI cross-linked mPEGNB-T hydrogels showed a similar change in storage moduli over time; however, iEDDA-CSI hydrogels formed by mPEGNB-I experienced a reduction of about a third of its initial G′, followed by maintaining the degree of stiffness over the 42 day period (Figure S16).

We next compared iEDDA-CSI hydrogels formed by mPEGNB-D, mPEGNB, and mPEGNB_CA at the same [Tz-NB] concentration. Results show that mPEGNB-D hydrogels initially had a lower G′ compared to mPEGNB_CA and mPEGNB; however, over 2 days, mPEGNB-D hydrogels stiffened, while mPEGNB_CA and mPEGNB gels softened drastically (Figure S17). The storage moduli of mPEGNB-D, mPEGNB_CA, and mPEGNB hydrogels cross-linked with PEG4Tz were then tracked over a 49 day period (Figure 3F). For mPEGNBs cross-linked with multiarm PEG4Tz, mPEGNB_CA hydrogels degraded rapidly and completely in 3 days; mPEGNB hydrogels softened but remained stable over time; finally, the moduli of mPEGNB-D hydrogels remained above its initial modulus over a 49 day period. We suspected that the enhanced degradation of PEG4Tz/PEG8NB_CA iEDDA click cross-linked hydrogels was from the disruption of noncovalent interactions between Tz-NB adducts by the additional carboxylic acid group on NB_CA.¹⁸ On the other hand, the high stability of iEDDA-CSI hydrogels cross-linked by PEG4Tz and mPEGNB-D was a result of enhanced noncovalent supramolecular interactions between the Tz/NB-D adducts. Future studies are needed to elucidate the underlying mechanism of the high stability of Tz/NB-D adducts, including molecular dynamic simulations and isothermal titration calorimetry (ITC).^29,30

Inspired by the drastic variation of iEDDA-CSI hydrogel stability/degradation from using mPEGNB-D or mPEGNB_CA (Figure 3F), we sought to create a noncovalent hydrogel network with highly programmable and predictable degradation patterns. To this end, we formulated iEDDA-CSI hydrogels using PEG4Tz and a mixture of mPEGNB-D and mPEGNB_CA (i.e., [NB-D]/[NB_CA] = 0.5, 1, 2). The blended iEDDA-CSI hydrogels showed storage moduli of approximately 4, 7, and 12 kPa for [Tz-NB] of 12, 16, and 24 mM, respectively (Figure S18 and Table S3). The blended iEDDA-CSI hydrogels with 12 mM [Tz-NB] fully degraded on day 12, day 42, and day 77 for [NB-D/NB_CA] ratios of 0.5, 1, and 2, respectively. The k_hydrolysis for 12 mM [Tz-NB] hydrogels with [NB-D]/[NB_CA] ratios of 0.5, 1, and, 2 were 0.3868, 0.1306, and 0.07676 d^–1, respectively. Similarly, blended hydrogels at a [Tz-NB] of 16 mM had k_hydrolysis constants of 0.2244, 0.07207, and 0.03237 d^–1 for [NB-D]/[NB_CA] ratios of 0.5, 1, and 2 (Table S3). Notably the 0.5 ratio at 16 mM concentration fully degraded on day 23, while hydrogels at the other two ratios remained intact after 77 days. Lastly, at 24 mM [Tz-NB] none of the hydrogel blends fully degraded after 77 days (data not shown); however, significant drops in storage moduli were observed. Specifically, for [NB-D]/[NB_CA] ratio of 0.5, 1, and 2, G′ decreased approximately 98, 88, and 72% from their initial G′, with corresponding k_hydrolysis of 0.08873, 0.03715, and 0.01807 d^–1, respectively (Figure 3G and Table S3). We hypothesized that the difference in hydrolysis kinetics of the iEDDA-CSI blended hydrogels cross-linked with PEG4Tz was from changes in the binding affinity from the presence of different ratios of NB-D (which strengthens supramolecular interactions) or NB_CA (which destabilizes supramolecular interactions). Additional studies are needed to confirm the mechanism of degradation. Interestingly, the degradation follows first-order decay (Tables S2–S3), suggesting that while these hydrogels were cross-linked by noncovalent supramolecular interactions, they could be formulated to exhibit highly predictable degradation akin to typical ester hydrolysis. Nonetheless, this new iEDDA-CSI gelation may benefit light and radical-free hydrogel cross-linking and cell/protein encapsulation. In contrast, the highly tunable and predictable degradation of these hydrogels could be leveraged for autonomous cell/spheroid recovery without using external degradation triggers, such as temperature change or supplying proteases, two common approaches reported in the literature.^31,32 In addition, the new iEDDA-CSI assembled and degradable hydrogels may be leveraged for in vivo drug and cell delivery,^33,34 since the degradation time is ideal for the stability of protein-based therapeutics which range from days to months in an in vivo setting.³⁵

Cytocompatibility of iEDDA Click Hydrogels

To test the cytocompatibility of the new iEDDA click cross-linked hydrogels, we utilized PEG8NB-T, as the additional tyramine motif could be exploited for dynamic tuning of hydrogel properties.³⁶⁻³⁸ The multiarm PEG8NB-T macromer also allows for faster gelation compared to mPEGNB-T and subsequent stabilization of the hydrogel via the noncovalent interactions postester hydrolysis. hiPSCs were chosen as a model cell type for their high value in cell therapy and their known vulnerability under radical-mediated hydrogel cross-linking.³⁹ In the absence of any biological motif, approximately half of the encapsulated hiPSCs were not viable 2 days postencapsulation (Figure 4A), a result consistent with the literature.^9,39,40 For example, the Kloxin group reported viability of 15% using thiol–ene click cross-linked PEG8NB hydrogels without any integrin binding and achieved a higher viability of only 61% when cell-adhesion ligand (RGDs) was immobilized in the gels, in conjugation with the use of antianoikis ROCK inhibitor Y-27632.³⁹ Additionally, the Chaudhuri group reported iPSC viability ranging from less than 10% in slow-stress relaxing alginate hydrogels and up to about 80% in fast-stress relaxing alginate hydrogels containing 1.5 mM RGD concentration.⁴¹ However, even in the absence of any biological motif in our iEDDA-CSI hydrogels, the encapsulated hiPSCs demonstrated significant proliferation, as indicated by progressive increases in spheroid sizes from 25 μm on day 2 to 135 μm on day 7 (Figure 4B). Most cells stained positive for pluripotency markers NANOG, Oct4, and SOX2 (Figure 4C). In addition, a small population of the proliferated hiPSC spheroids showed a lumen structure (Figure 4D). We envision that these hydrogels could be utilized for long-term culture of iPSCs and subsequent maturation into complex organoids which is a lengthy process that can take up to several months.⁴² We reason the viability of iPSCs encapsulated within PEG8NB-T iEDDA click cross-linked hydrogels could be improved by incorporating bioactive components such as cell adhesion ligands or bioactive polymers such as fibronectin and gelatin.⁴³ Norbornene functionalized onto peptides and biopolymers such as gelatin has been demonstrated and could be easily incorporated within the PEG8NB-T iEDDA click cross-linked hydrogel system.^7,44 Chemically defined hydrogel systems have been reported to support iPSC potency and on-target differentiation better than nonchemically defined 3D culture systems such as Matrigel and Geltrex.⁴⁰ Future work will focus on increasing the bioactivity of the iEDDA-CSI hydrogels by implementing cell-adhesion ligands or by rendering the hydrogels more viscoelastic.^41,45

Figure 4.

Cytocompatibility of iEDDA-CSI hydrogels. (A) Confocal images of hiPSCs encapsulated within 3.75 wt % PEG8NB-T cross-linked with 4-arm PEGTz. Cells were stained with calcein AM and ethidium homodimer at different time points. (B) Average spheroid size determined from ImageJ using live/dead confocal images at different time points. (C) Confocal images of hiPSCs stained with NANOG, OCT4, and SOX2. Cells were counterstained with DAPI after 7 days of culture. (D) Confocal image of lumen formation of hiPSC spheroid on day 7. Cells were stained with F-actin and counterstained with DAPI.

Enzymatic Stiffening and Photopatterning of iEDDA Click Cross-Linked (m)PEGNB-T Hydrogels

Inspired by our previous work on the tyrosinase-mediated stiffening of tyrosine-containing polymeric hydrogels, we employed mPEGNB-T to assemble iEDDA-CSI hydrogels. As mPEGNB-T contains a tyrosine-like hydroxyphenyl group, it can be oxidized by tyrosinase to dihydroxyphenylalanine (DOPA), DOPA quinone, and ultimately to DOPA dimer (Figure 5A).^36,37 8-arm PEGNB-T or mPEGNB-T were cross-linked with PEG4Tz to form iEDDA and iEDDA-CSI hydrogels. The initial G′ of iEDDA cross-linked PEG8NB-T hydrogels or iEDDA-CSI cross-linked mPEGNB-T hydrogels was 7.7 and 1.2 kPa, respectively (Figure 5B–S19). The high initial moduli of iEDDA hydrogels were attributed to covalent cross-links, which were not present in iEDDA-CSI hydrogels formed with mPEGNB-T. After 2 days incubation in pH 7.4 PBS, the storage moduli of PEGNB-T iEDDA hydrogels decreased to 2.6 kPa, while the initial moduli of mPEGNB_T iEDDA-CSI hydrogels increased to approximately 2 kPa (Figures 5B and S19). Dityrosine cross-links can be achieved through multiple mechanisms, including various enzymatic and light-initiated reactions.⁴⁶ Both iEDDA and iEDDA-CSI hydrogels were stiffened with mushroom tyrosinase (MT) at concentrations 1.50 and 0.75 kU/mL over an 8 h period (Figures 5C and S20). The G′ of PEG8NB-T iEDDA hydrogels doubled when incubated with 1.50 kU/mL of MT over 6 h, while the same hydrogels incubated with 0.75 kU/mL had an increase of 50% in its initial storage moduli (Figure S20). Similarly, mPEGNB-T iEDDA-CSI hydrogels underwent different degrees of stiffening based on the incubation time and MT concentration (Figure 5C). These new iEDDA-CSI hydrogels should be susceptible to on-demand stiffening using other enzymes such as horse radish peroxidase and light-initiated tyrosine formation using ruthenium/sodium persulfate.^47,48

Figure 5.

Enzymatic stiffening and photopatterning of (m)PEGNB-T iEDDA-CSI click cross-linked hydrogels. (A) Schematics of dityrosine cross-linking catalyzed by mushroom tyrosinase and visible light-mediated ligand conjugation using FMN of FITC-tyramide. (B) Storage moduli of mPEGNB-T iEDDA-CSI hydrogels on days 0 and day 2. (C) Stiffening of mPEGNB-T iEDDA-CSI hydrogels using MT at different concentrations over an 8 hour period with corresponding normalized G′ values. (D) Confocal image of a PEG4Tz/PEG8NB-T iEDDA hydrogel containing photopatterned lines using FITC-tyramide conjugation via FMN and visible light-mediated dityrosine cross-linking. Light intensity across the photopatterns (dashed line) was analyzed by the histogram feature in ImageJ.

The additional tyramine group on PEG8NB-T was also utilized for photopatterning via flavin mononucleotide (FMN)-induced tyrosine conjugation in the presence of visible light (Figure 5A).³⁶ To achieve this, FITC-tyramine along with FMN was diffused into 8-arm PEGNB-T iEDDA click cross-linked hydrogel. A visible light (405 nm) source from a digital light processing bioprinter was projected through a photo pattern consisting of 200 μm thick lines onto the hydrogel to create a patterned area. Confocal microscope images show the presence of three lines approximately 200 μm in width and with small squares of 50 μm, which is the lowest pixel resolution of the DLP projector (Figure 5D). The intensity of the lines was measured, revealing that the photopatterned lines had nearly the exact length as designed, with minimal noise between the lines (Figure 5D). While this proof-of-concept study demonstrated the capability of our new iEDDA click hydrogels in photopatterning, higher resolution or 3D photopatterning can be achieved by using a two-photon laser,⁴⁹ which will enable spatial–temporal conjugation of bioactive ligands to direct cell behavior.⁵⁰

Shear-Thinning and Self-Healing Hydrogels due to Noncovalent Interactions of PEGNB-D iEDDA Click Cross-Linked Hydrogels

As most supramolecular hydrogels exhibit some degree of shear-thinning property and injectability, we asked if our iEDDA and iEDDA-CSI hydrogels support these physical behaviors. To this end, 8-arm PEGNB-D was initially cross-linked by PEGTz at different [Tz-NB] concentrations and swelled for at least 2 days to ensure that ester bonds on PEGNB-D were hydrolyzed entirely and that the hydrogels were stabilized solely by the noncovalent iEDDA-CSI. Subsequently, a viscosity shear rheology test was performed to demonstrate the shear-thinning behavior of noncovalent interactions of PEG8NB-D iEDDA click cross-linked hydrogels (Figure 6A). Results show that the PEG8NB-D cross-linked with PEG4Tz hydrogels display shear thinning behavior with all formulations based upon a decrease in complex viscosity of shear rates from 0.1 to 20 s^–1. Additionally, all PEG8NB-D hydrogels exhibited a more viscous behavior under higher strains (Figure 6B). For 12, 16, and 24 mM conditions, the G′/G′’ crossover occurred at approximately 70, 50, and 20% strain (Figure S21). To assess the self-healing behavior of PEG8NB-D iEDDA click cross-linked hydrogels, a step-strain test was conducted (Figure 6C). The strain rate was alternated between 200 to 1% to show the change of a tangent delta [tan(δ)] greater than 1 and under 1 at high and low strains. PEG8NB-D iEDDA click cross-linked hydrogels across all formulations showed more viscous characteristics under high strain and more elastic properties under low strain. Further, the gels were able to recover for at least three cycles. Lastly, to demonstrate that the noncovalent interactions of PEG8NB-D iEDDA click cross-linked hydrogels were capable of being injected after gelation, three PEG8NB-D hydrogels cross-linked with PEG4Tz at [Tz-NB] concentration of 16 mM postester hydrolysis were loaded into a 1 mL syringe. Subsequently, the hydrogels were extruded through an 18G needle attached to the syringe by hand (Figure 6D). Since the PEG8NB-T iEDDA click cross-linked hydrogels are extrudable, we envision such a hydrogel system to be applied as a bioink for extrusion-based bioprinting without the need for shear-thickeners such as methylcellulose or subsequent cross-linking postprinting.^51,52

Figure 6.

Assessing shear thinning and self-healing properties of iEDDA click cross-linked PEGNB-D hydrogels post ester hydrolysis. (A) Measurement of viscosity at different shear rates of PEG4Tz/PEG8NB-D hydrogels. (B) Gel–sol transition of PEG4Tz/PEG8NB-D hydrogels under increased shear strain. (C) Gel–sol–gel transition based upon alternating between high to low shear strain over time. (D) Pictures of extruding loaded PEG4Tz/PEG8NB-D hydrogels from a syringe through an 18G needle.

Conclusions

The efficient cross-linking of iEDDA click hydrogels often leads to hydrogels with highly stable networks owing to the nondegradability of click reaction products. This is especially true when the click reaction products can self-assemble due to supramolecular interactions. Previously, we developed an approach to create hydrolytically degradable PEG-based hydrogels with tunable degradation. The hydrogels consisted of PEG8NB_CA macromer cross-linked with 4-arm Tz and/or mTz-functionalized PEG with different ratios.¹⁸ Upon secondary conjugation of amines onto the PEG8NB_CA macromer via amide coupling, rapid ester hydrolysis kinetics were observed when cross-linked with thiol-bearing macromers via thiol-norbornene photoclick reaction.¹⁹ In this contribution, we observed PEG8NB-X macromers were capable of cross-linking with PEG4Tz and exhibited initial rapid degradation from ester-hydrolysis and subsequent stabilization via noncovalent interactions between Tz/NB-X adducts. Due to the stability of the noncovalent interactions between the Tz/NB-X adducts, we were capable of cross-linking monofunctional mPEGNB-D with PEG4Tz to form highly stable hydrogels based upon purely the noncovalent interactions. PEG8NB-T iEDDA-CSI hydrogels show good cytocompatibility and maintain the potency of hiPSCs without any bioactive components for at least 7 days. Further, PEG8NB-T iEDDA-CSI hydrogels could be stiffened and patterned via dityrosine cross-links. Blended mPEGNB-D/mPEGNB_CA iEDDA-CSI hydrogels allowed for tunable degradation from less than two months to several months. Hydrogels containing these noncovalent cross-links postester hydrolysis between Tz-NB adducts displayed shear-thinning, self-healing, and injectability properties. We expect this hydrogel system to be highly useful for various applications in the fields of tissue engineering and drug/cell delivery.

Experimental Methods

Materials

Hydroxyl terminated 8-arm PEG (20 kDa) and linear PEG monomethyl ether (4 kDa) were purchased from JenKem Technology and Tokyo Chemical Industry (TCI), respectively. 4-arm PEG-amino succinic acid (PEG-ASA) (10 kDa) and 4 arm PEG thiol (10 kDa) were purchased from Layson Bio Inc. Carbic anhydride, pyridine, and dichloromethane (DCM) were all acquired from Thermo Scientific. 5-Norbornene-2-carboxylic acid, tetrahydrofuran (THF), 4-dimethylaminopyridine (DMAP), Lithium phenyl-2, 4,6-trimethylbenzoylphosphinate (LAP), and N,N-dicyclohexylcarbodiimide (DCC) were obtained from Sigma-Aldrich. Tetrazine-amine was purchased from Click Chemistry Tools. N-hydroxylsuccinimide (NHS) and N,N-diisopropylethylamine (DIPEA) were acquired from TCI. N,N-dimethylformamide (DMF) and 1-[bisdimehtylaminomethylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate (HATU) were purchased from Alfa Aesar and AnaSpec, respectively. Tyramine hydrochloride and N,N′-diisopropylcarbodiimide (DIC) were purchased from Chem-Impex. Dopamine hydrochloride and isopropyl amine were obtained from Alfa Aesar and TCI chemicals, respectively. Hydroxybenzotriazole (HOBt) was purchased from Oakwood Chemical. Riboflavin 5′-monophosphate sodium salt hydrate (FMN) and 5/6-carboxyfluorescein succinimidyl ester mixed isomer were purchased from Thermo Fisher and Millipore Sigma, respectively. Calcein-AM and ethidium homodimer staining were obtained from Biotium. F-actin stain and 4′,6-diamido-2-phenylindole (DAPI) were acquired from Cytoskeleton Inc. and Anaspec, respectively.

Macromer Synthesis and Purification

Synthesis of 8-Arm PEG8NB and mPEGNB

8-arm PEG8NB and monofunctional PEGNB (mPEGNB) was synthesized following previous protocols.^53,54 Briefly, 10 equiv of norbornene acid compared to concentration of hydroxyl functional group on PEG was reacted with 5 equiv of DCC in anhydrous DCM to form norbornene anhydride with dicyclohexylurea as a byproduct. The byproduct was removed through vacuum filtration using filter paper (size 52). Norbornene anhydride was then added dropwise into a flask containing hydroxyl terminated 8-arm PEG (20 kDa) or PEG monomethyl ether (4 kDa) along with 10 equiv of pyridine and 1 equiv DMAP. The reaction vessel was purged with N₂ gas and proceeded overnight. PEGNB and mPEGNB products were precipitated using ice cold diethyl ether and filtered using a fritted glass funnel. PEGNB and mPEGNB were then dried under a desiccator and redissolved in double distilled water (DDH₂O).

Synthesis of 8-Arm PEGNB_CA and mPEGNB_CA

8-arm PEG8NB_CA and monofunctional PEGNB_CA (mPEGNB_CA) were synthesized using previously established protocols with slight modifications.¹⁸⁻²⁰ Briefly, 8-arm hydroxyl terminated PEG and PEG monomethyl ether were dried under a desiccator and reacted with 5 equiv carbic anhydride and 1 equiv DMAP compared to hydroxyl concentration of PEG in anhydrous THF at 60 °C for 16 h. After 16 h, a second portion of 5 equiv carbic anhydride and 1 equiv DMAP was added and reacted for an additional 24 h at 60 °C. For mPEGNB_CA, the protocol was repeated due to the low degree of substitution obtained. PEGNB_CA and mPEGNB_CA was precipitated with diethyl ether and redissolved in DDH₂O.

Synthesis of 8-Arm PEGNB-X and mPEGNB-X

Amide coupling was performed onto the carboxylic acid group on mPEGNB_CA and PEG8NB_CA to form the mPEGNB-X and PEGNB-X macromers, respectively. Coupling of tyramine, dopamine, and isopropyl amine onto the carboxylic acid of PEG8NB_CA was performed in order to obtain PEG8NB-T, PEG8NB-D, and PEG8NB-I, respectively. Similarly, the coupling of tyramine, dopamine, and isopropylamine to mPEGNB_CA resulted in the synthesis of mPEGNB-T, mPEGNB-D, and mPEGNB-I, respectively. The protocol was performed following our previous method⁵⁴, 55]. Briefly, mPEGNB_CA and PEG8NB_CA along with 2.5 equiv compared to norbornene concentration on mPEGNB_CA and PEG8NB_CA of DIC and HOBt were dissolved in anhydrous amine free DMF. The reaction vessel was purged with N₂ gas and was allowed to react for at least 2 h covered from light. After 2 h, 2.5 equiv of tyramine hydrochloride, dopamine hydrochloride, or isopropyl amine along with 2.5 equiv of DIPEA and 1 equiv of DMAP was added into the reaction vessel. The reaction vessel was then purged with N₂ gas and was reacted overnight.

Synthesis of PEG4Tz

[4-(1,2,4,5-Tetrazin-3-yl)phenyl]methanamine (tetrazine amine) was conjugated onto 4-arm PEG-ASA following previously published protocols.^9,18 Briefly, 5 equiv of HATU compared to the concentration of carboxylic acid group on PEG-ASA was dissolved in amine-free anhydrous DMF and allowed to react for 10 min to form an active ester. Subsequently, 1.2 equiv of tetrazine amine was added, and the reaction proceeded for at least 16 h at room temperature.

Macromer Purification and Characterization

PEG8NB, mPEGNB, PEG8NB_CA, and mPEGNB_CA were dialyzed for 3 days in DDH₂O using dialysis tubing with molecular weight cut off (MWCO) of 3.5 kDa. Following 3 days of dialysis, macromers were frozen and then freeze-dried. The same process was repeated for the other PEG macromers with slight variation in the dialysis procedure. PEG4Tz was dialyzed for 1 day in methanol in order to remove DMF quickly then another 2 days in DDH₂O. PEG8NB-X and mPEGNB-X macromers were dialyzed for 1 day in methanol and then 2 days in DDH₂O at 4 °C in order to slow down the rate of ester hydrolysis. Synthesized PEG macromers were characterized using nuclear magnetic resonance (NMR) (deuterium oxide, 500 MHz, Bruker Advance 500, 32 scans). The substitutions of norbornene onto PEG macromers were determined by comparing the integral peaks of the alkene protons of the norbornene functional group to the protons consisting of the PEG backbone. The efficiency of the amide coupling onto PEG8NB_CA and mPEGNB_CA was determined by comparing the integral peaks of the protons on the benzyl ring of the tyramine and dopamine functional groups to the alkene protons on the norbornene functional group. Similarly, the efficiency of isopropyl amine conjugation onto PEGNB_CA or mPEGNB_CA was determined through comparing the integral peaks of the protons on the two methyl groups compared to the alkene protons of the norbornene functional group. Substitution of tetrazine onto PEG-ASA was determined through measuring the absorbance at 523 nm using UV–vis spectroscopy against a standard curve of Tz-amine with concentrations ranging from (1 to 0.015 mg/mL).

Hydrogel Cross-Linking and Mechanical Characterization

(m)PEG(8)NB macromers were cross-linked at stoichiometric R([Tz]/[NB]) ratio equal to 1. The total concentration of tetrazine-norbornene adducts [Tz-NB] was varied between 8 to 32 mM for 8-arm PEG and 12 to 24 mM for monofunctional PEG. Hydrogels were prepared by depositing 45 μL of hydrogel precursor solution between two glass slides coated with hydrophobic windshield coating separated by 1 mm Teflon spacers. The slides containing the hydrogel precursor solution was placed in a sealed container with a 60 mm Petri dish filled with water to prevent dehydration. The sealed container was placed in an incubator set at 37 °C for 1 h. After 1 h, the hydrogels were swelled for 30 min in phosphate buffered saline pH 7.4 at room temperature to obtain the initial elastic shear moduli. Using a Bohlin CVO digital rheometer fitted with an 8 mm diameter parallel geometry plate, strain sweep tests were conducted by measuring the elastic (G′) and viscous (G′‘) moduli of the fabricated hydrogels from starting strain of 0.1 to 5% at 1 Hz oscillating frequency. Gelation time of different hydrogel formulations was obtained through in situ rheometry. Using an Anton-Paar MCR102 rheometer fitted with a 25 mm parallel geometry plate, 200 μL of hydrogel precursor solution was deposited on the rheometer stage. The plate was then lowered to a depth of 0.275 mm. Time sweep test was then conducted at 1% strain with 1 Hz oscillating frequency over 30 min. In situ rheology was performed either at 25 or 37 °C. Gelation point was defined as the time required for the elastic shear moduli (G′) to crossover with the viscous shear moduli (G′‘). When the G′ is higher than the G′‘, a gel, or hydrogel, was formed.

PEGNB macromers cross-linked with PEG4SH was conducted at [NB] of 16 mM with R = 1 with LAP concentration of 2 mM. The hydrogel fabrication process was identical to iEDDA cross-linking except that the glass slides containing the hydrogel precursor solution was placed under a 365 nm UV lamp at light intensity of 3 mW/cm² for 2 min. The hydrogels were then swelled for 30 min at room temperature with pH 7.4 PBS. The procedure for strain sweep test is identical to tetrazine cross-linking. In situ rheology was conducted by using an Anton Paar MCR102 25 mM parallel geometry plate fitted with a UV-curing system. The in situ rheology procedure was identical to tetrazine cross-linking except after 1 min the hydrogel precursor solution was irradiated with UV light produced from OmniCure s2000 containing a 365 nm light filter at light intensity of 19 mW/cm².

Confirmation of Ester Hydrolysis

Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) and NMR was performed to confirm ester hydrolysis of iEDDA click cross-linked hydrogels. For ATR-FTIR, 5 wt % PEGNB-D hydrogels with R = 1 were fabricated at a total volume of 100 μL. Following fabrication, hydrogels were dried in a desiccator for 24 h. ATR-FTIR spectra was obtained using a Nicolet iS50 FTIR spectrometer equipped with a diamond ATR module. The same hydrogels were then swelled using pH 7.4 PBS for 2 days at 37 °C, where ester hydrolysis occurred. Following 2 days, hydrogels were washed in DDH₂O for several hours then dried under a desiccator for 24 h. ATR-FTIR spectra were then obtained postester hydrolysis.

NMR experiment was conducted to confirm the leaving group postester hydrolysis. Three 45 μL hydrogels were fabricated containing 5 wt % PEGNB-D R = 1 and incubated at 37 °C with PBS pH 7.4 for 2 days. After 2 days, the PBS was collected and dialyzed using 3.5 kDa MWCO dialysis tubing in DDH₂O for 1 day. The degradation product was then freeze-dried and dissolved in deuterium oxide where the NMR spectra was obtained.

Programmable Degradation through mPEGNB-D/mPEGNB_CA Blends

To demonstrate programmable degradation, mPEGNB-D/mPEGNB_CA blended hydrogels were cross-linked with PEG4Tz R = 1 over different [NB] values and blend ratios. Specifically, the total [NB] was 12, 16, and 24 mM with blend ratios of [NB-D]/[NB_CA] of 33/67, 50/50, and 67/33. Degradation was tracked by measuring the elastic shear moduli through strain sweep tests over an extended period. Values were then fitted with one-phase decay curves to obtain degradation reaction kinetics.

Photopatterning and Stiffening of PEGNB-T Hydrogels

FITC conjugated tyramine was synthesized by reacting 1.2 equiv of tyramine hydrochloride with 1 equiv of NHS-fluorescein in a 1:3 mixture of dimethyl sulfoxide (DMSO) and 0.1 M bicarbonate buffer. Product was freeze-dried without purification. 400 μL of 3.75 wt % PEG8NB-T cross-linked with PEG4Tz R = 1 was placed in an 8 mL scintillation vial and was centrifuged at 2500 RCF for 60 min during the gelation process to remove bubbles. Subsequently, the fabricated hydrogel was incubated with 0.02 mg/mL FITC-Tyramine and 0.45 mg/mL FMN for overnight. Photopatterning was achieved using a Lumen X digital light processing (DLP) bioprinter. The line constructs were created in Tinkercad having a width of 200 μm and a height of 100 μm and uploaded to the Lumen X bioprinter. The hydrogel was placed in the middle of a polydimethylsiloxane (PDMS) coated plate and was exposed to light from the micro projector for 1.5 min at 80% light intensity (∼24 mW/cm²). The hydrogel was then washed in PBS for 2 days and imaged using confocal microscopy.

mPEGNB-T and PEG8NB-T hydrogels with [NB-T] of 12 mM cross-linked with PEG4Tz at R = 1 were fabricated using the same method as described in this paper. Hydrogels were swelled at 37 °C in PBS pH 7.4 for 2 days before stiffening with mushroom tyrosinase (MT). Hydrogels were then stiffened with 1.5 0.75, or 0 kU/mL of MT dissolved in PBS. The elastic moduli were measured every 2 h using strain sweep tests over an 8 h period.

Assessing Injectable Behavior of PEGNB-D Hydrogels

To assess the reversible behavior of the supramolecular cross-links, additional rheology tests were performed on PEG8NB-D hydrogels with [NB] of 12, 16, and 24 mM cross-linked with PEG4Tz at R = 1. All fabricated hydrogels were incubated for at least 2 days before rheological testing in order to ensure complete ester hydrolysis. All additional rheological testing was performed using an Anton Paar MCR102 rheometer fitted with an 8 mm parallel geometry plate. Viscosity test was performed by measuring the complex viscosity at frequencies between 0.1 to 20 s^–1. To determine the minimum strain of gel–sol transition, the elastic and viscous shear moduli was measured with increasing shear strain from 1 to 200% strain. Self-healing properties were investigated through a step-strain test alternating between 200 and 1% strain. Additionally, injectability of the system was demonstrated by extruding PEG8NB-D with [NB-D-Tz] of 16 mM hydrogels post ester hydrolysis loaded into a 1 mL syringe and extruded through an 18 gauge needle.

Cell Encapsulation and Analysis

Cell Culture

Two different human induced pluripotent cell lines utilized in this study are Cellartis Human iPS cell line 12 (ChiPSC12) from Takara Bio and the mEGFP-tagged SOX2 iPSC cell line developed by the Conklin Laboratory purchased from the Coriell Institute. iPSCs were cultured using vitronectin coated plates containing Essential 8 media. Media changes occurred daily. Cells were passaged every 4 days using TrypLE and cultured with Essential 8 media plus Y27623 at a concentration of 10 μM for 1 day following passaging.

Cell Encapsulation

For hiPSC cell encapsulation, polymer solutions were sterile filtered by using 0.22 μm pore size sterile filters. iPSC cells were detached using TrypLE and dissociated to single cells. The hydrogel system consisted of 3.75 wt % PEG8NB-T along with PEG4Tz at R = 1. The hydrogel precursor solution was allowed to react for 9 min at room temperature before cells were added. The cell laden hydrogel precursor solution was then dispensed at 25 μL on open-tip 1 mL syringes. The reaction then proceeded for an additional 20 min before being placed in E8 media supplemented with 10 μM of Y27623. Media changes then occurred daily using the same media with Y27623.

Live/Dead Staining, F-Actin Staining, and Immunostaining

For iPSCs, live/dead staining occurred on days 2, 4, and day 7. All other staining for iPSCs occurred on day 7. Cell-laden hydrogels were washed with Dulbecco’s PBS (DPBS) for 10 min. Following washing, 0.46 μM of Calcein-AM and 0.30 μM of ethidium homodimer-1 was added and cells were incubated for 45 min at room temperature protected from light. After 45 min, cells were washed three times for 2 min each using DPBS. For F-actin staining and immunostaining, hydrogel samples were washed for 10 min using DPBS and then fixed with 4% paraformaldehyde for 45 min. Following fixing, cells were stained with F-actin or pluripotency markers. For F-actin staining, 70 nM of F-actin (Acti-stain 488 Fluorescent Phalloidin, Cytoskeleton, Inc.) in the presence of 1% (v/v) bovine serum albumin (BSA) and 0.3% (v/v) Triton X100 was added, and cells were incubated at 4 °C overnight. F-actin-stained cells in the hydrogel were then washed three times using 1% (v/v) BSA and 0.3% (v/v) Triton X100 for 20 min each and then counterstained with DAPI overnight at 4 °C. For immunostaining, cell-laden hydrogels were permeabilized for 1 h using 1% (v/v) BSA and 0.3% (v/v) Triton X100. Following permeabilization, cell-laden hydrogels were exposed to Oct4a (Cell Signaling Technology, (C30A3) Rabbit mAb #2840) and NANOG (Cell Signaling Technology, #3580) antibodies at 200x dilution and incubated with BSA/TritonX100 solution for overnight at 4 °C. The next day cell-laden hydrogels were washed using BSA/TritonX100 solution three times for 20 min each then secondary antirabbit antibody with Alexa Fluor 555 (Santa Cruz) attached was added at 200x dilution and was incubated for overnight at 4 °C. The secondary antibody was washed and DAPI was added overnight at 4 °C. All stained hydrogel samples were imaged using confocal microscopy (Olympus Fluoview, FV1000 and Andor BC43). All samples contained three hydrogel samples with at least three random images per gel.

Image Analysis

ImageJ was used to quantify Live/Dead stained images. Cell viability was quantified using the ImageJ plugin called “live dead quantification”. For spheroid diameter measurements, images were uploaded into ImageJ where the diameter of the iPSC spheroids were manually measured using the ROI manager.

Statistics

Data is presented as mean ± standard error of the mean (SEM). For two conditions, unpaired t-test was used to determine statistical significance with p < 0.05. For two or more conditions, one-way ANOVA or two-way ANOVA was performed with Tukey multiple comparison test to determine statistical significance between groups when p < 0.05. For modeling the decrease of G′ of hydrogel systems, one-phase decay fitting was applied using GraphPad Prism 10. From the one phase decay fitting, k_hydrolysis constants along with t_1/2 and R² values were obtained. The equation utilized for one phase decay fitting on the Prism software is Y = (Y₀ – Plateau)*e^–k_hydrolysis^*x + Plateau, where Y is the normalized G′ to day 0, Y₀ was set to 1 to indicate no degradation occurring after normalization, plateau was set to 0 to indicate complete degradation, x is time in days or hours, and k_hydrolysis is the hydrolysis reaction constant in terms of inverse hours (hr^–1) or inverse days (d^–1).

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.macromol.3c01855.

• Description of how gel fraction, swelling ratio, and reaction constants were derived; NMR spectra of synthesized macromers; gelation time of mPEGNB-D and PEG8NB-D cross-linked with either PEG4SH or PEG4Tz; reaction constant between mPEGNB-D and PEG8NB-D with PEG4Tz; gel fraction, swelling ratio, and swollen mass of iEDDA click cross-linked mPEGNB-D or PEG8NB-D hydrogels; changes in G′ overtime of different PEGNBs cross-linked with PEG4Tz; changes in G′ post ester hydrolysis and post enzymatic stiffening of PEG8NB-T iEDDA click cross-linked hydrogels; minimum strain needed to induce G′‘/G′ crossover of PEG8NB-D hydrogels cross-linked with PEG4Tz (PDF)

The authors declare no competing financial interest.