Stabilization of enzyme-immobilized hydrogels for extended hypoxic cell culture

Abstract

In this work, glucose oxidase (GOx)-immobilized hydrogels are developed and optimized as an easy and convenient means for creating solution hypoxia in a regular incubator. Specifically, acrylated GOx co-polymerizes with poly(ethylene glycol) diacrylate (PEGDA) to form PEGDA-GOx hydrogels. Results show that freeze-drying and reaction by-products, hydrogen peroxide, negatively affect oxygen-consuming activity of network-immobilized GOx. However, the negative effects of freeze-drying can be mitigated by addition of trehalose/raffinose in the hydrogel precursor solution, whereas the inhibition of GOx caused by hydrogen peroxide can be prevented via addition of glutathione (GSH) in the buffer/media. The ability to preserve enzyme activity following freeze-drying and during long-term incubation permits facile application of this material to induce long-term solution/media hypoxia in cell culture plasticware placed in a regular CO₂ incubator.

1. Introduction

Creating a hypoxic environment (i.e., [O₂] < 5%) for in vitro cell culture is critically important as hypoxia is implicated in both normal and pathological cell physiology.^1–4 A hypoxic microenvironment can be created by using a cell culture chamber with controlled gas inlet (i.e., hypoxic chamber). However, hypoxic chamber systems can be expensive and require dedicated space and routine maintenance, making them unaffordable and/or not convenient for most biomedical laboratories. In addition, it is challenging to perform real-time imaging or other instrument-based live cell assays within a standard hypoxic chamber system. Furthermore, one hypoxic chamber can only provide one fixed O₂ tension at any given time, which is not ideal for performing experiments under complex hypoxic patterns, such as hypoxia gradients or fluctuating oxygen tensions. In addition to hypoxic chamber systems, alternate methods (e.g., microfluidic systems, devices with diffusion barriers, or enzymatic reactions) have been increasingly developed to create diverse hypoxia patterns for in vitro cell culture and hypoxia-related biomedical applications.^5–11

Enzymatic reactions are particularly attractive for creating hypoxia in biotechnology and therapeutic applications owing to the wide range of enzymes capable of consuming dissolved oxygen (e.g., glucose oxidase, tyrosinase, laccase, bilirubin oxidase).^12,13 Depending on the design, oxygen-consuming enzymes and their respective substrates can be used in soluble or immobilized forms for creating complex hypoxia patterns.^5,7–9 For example, Li et al. immobilized glucose oxidase (GOx) and catalase (CAT) in chitosan-based hydrogels via glutaraldehyde crosslinking.⁶ The enzyme-immobilized materials were coated on to 3D printed solid inserts, which were used to induce solution hypoxia in the range of ~1% to ~12% O₂ for a few hours. This system, however, lacked flexibility (only cell culture in multi-well plates were demonstrated) and long-term enzyme stability. Another disadvantage is that a second enzyme CAT was co-immobilized to reduce cytotoxicity of by-product hydrogen peroxide. Nonetheless, CAT generated additional oxygen that decreased the efficiency of GOx reaction. In addition to using immobilized enzyme, hypoxia can also be induced by immobilized substrates.^8,9,14 In particular, Gerecht and colleagues immobilized ferulic acid (FA) to a variety of polymers, which served as substrates for hydrogel crosslinking and laccase-mediated oxygen consumption.^8,9,14 One shortfall of this approach was that the amount of polymer-immobilized substrate was fixed, which limited the duration of hypoxia. Lewis et al. extended the timespan of laccase-induced hypoxia by imposing longer oxygen diffusion length (e.g., higher volume of media above the hydrogel).¹⁵ The system was successfully used to evaluate the impact of hypoxia on sarcoma cell invasion and migration.¹⁵

To create an enzyme-based flexible hypoxic cell culture system, we have reported GOx-immobilized poly(ethylene glycol)-diacrylate (i.e., PEGDA-GOx) hydrogels for inducing solution hypoxia.¹⁶ We also validated the hypoxia-inducing capability of this system on in vitro culture of anchorage-dependent hepatocellular carcinoma (HCC) cells and anchorage-independent acute myeloid leukemia (AML) cells. This prior work also showed that the PEGDA-GOx hydrogels were capable of inducing sustained hypoxic response in adherent HCC cells and suspending AML cells using multi-well cell culture plates. The major benefit of using enzyme-immobilized hydrogels to create a hypoxic microenvironment for in vitro cell culture is that hypoxia can be induced on-demand and be performed together with the normoxic controls using a single well-plate. It is also possible to create hypoxia gradients through placing the enzyme-immobilized gels into a channel slide. In addition to immobilized GOx, we included CAT in the culture media to mitigate the cytotoxic effects of hydrogen peroxide, a by-product of the GOx reaction. Even though the previous PEGDA-GOX hydrogels were effective in rapidly inducing solution hypoxia (within one hour), the system suffered from the lack of long-term stability. Regardless of gel formulations and immobilized GOx content, solution hypoxia only sustained for roughly 24 hours after placing PEGDA-GOx gels in the media. While duration of 24 hours might be sufficient for some experiments (e.g., migration), it is nonetheless not ideal for long-term cell studies. In addition, all of the GOx-immobilized gels used in our previous report were freshly prepared one day before the experiments. It is not clear whether these hydrogels can be lyophilized without losing the catalytic activity of the immobilized enzyme. Being able to preserve the activity of the lyophilized enzyme-immobilized hydrogels will not only provide great flexibility in experimental design but also present commercialization potential for this hypoxia-inducing material.

Drawing from past work on preserving long-term enzyme activity,^17–24 this study explored methods to increase the flexibility and stability of the PEGDA-GOx hypoxia-inducing hydrogel system. First, attempts were made to provide flexibility of using the hypoxia-inducing materials through lyophilizing the PEGDA-GOx hydrogels. The effect of lyophilization (i.e., freeze-drying) on oxygen consumption capability of immobilized enzyme were evaluated. In addition, cryoprotectants (e.g., trehalose and raffinose) were introduced into the gel formulations for increasing thermo-stability of the enzymes. In particular, unmodified trehalose or raffinose were first used to examine their ability to preserve the activity of reconstituted freeze-dried PEGDA-GOx hydrogels. Next, the influence of glutathione (GSH) on reducing hydrogen peroxide and its effect on sustaining hypoxia were studied. This approach was favorable as opposed to the use of CAT as the reducing activity of GSH would not offset the oxygen consumption ability of GOx. Finally, the cytotoxicity of optimized PEGDA-GOx gels were demonstrated using Molm-14 cells, am AML cell line. The viability and hypoxia-related gene expression in pancreatic cancer cells (COLO-357) cultured in the presence of the improved PEGDA-GOx hydrogels were also evaluated as a means of demonstrating the usefulness of the PEGDA-GOx hydrogels.

2. Materials & methods

2.1. Materials

Linear PEG (MW: 2 kDa) was purchased from Sigma-Aldrich. Glucose oxidase (0243-500KU) and acrylate-PEG-succinimidyl valerate (Acryl-PEG-SVA) were obtained from Laysan Bio and Amresco, respectively. β-D-glucose and glutathione were purchased from Thermo Scientific. D-trehalose was acquired from Acros Organic. Cell culture related reagents, including penicillin-streptomycin, antibiotic-antimycotics, fetal bovine serum (FBS), and Dulbecco’s modified Eagle’s medium (DMEM), were acquired from Life Technologies. HEPES and Dulbecco’s phosphate-buffered saline (DPBS) were purchased from Lonza. Membrane culture plate inserts (PIXP-012-50) were purchased from EMD Millipore. Ellman and AlamarBlue^® reagents were purchased from Fisher Scientific.

2.2. Synthesis of PEGDA-GOx hydrogels

PEG-diacrylate (PEGDA), photoinitiator lithium aryl phosphonate (LAP), and acrylated GOx (i.e., GOxPEGA) were synthesized and purified as described in our previous publication.¹⁶ Prior to preparing PEGDA-GOx hydrogels, all macromer solutions were sterile-filtered through 0.22 μm syringe filters. All hydrogels were photopolymerized from 15 wt% PEGDA, 1 mM LAP, and prescribed concentration of GOxPEGA (0.2–2 mg/mL) and trehalose (0–3 mg/mL). Cylindrical shape gels (30 μL) were crosslinked via UV photopolymerization (365 nm, 5 mW/cm², 2 min exposure). Following gelation, GOx-immobilized hydrogels were either incubated in DPBS for 24 hours at 37°C prior to use or immediately freeze-dried for long-term storage. Freeze-dried gels were reconstituted in DPBS for 24 hours prior to using them in experiments.

2.3. O₂ measurement

All O₂ concentration measurements were performed using a dipping-type oxygen probe connected to Microx4 oxygen sensor (PreSens Precision Sensing GmbH). O₂ probe was immersed in the solution and located at ~2 mm above the surface of the hydrogel.

2.4. Biochemical and cellular assays

GSH and H₂O₂ contents were characterized by Ellman’s assay and Quantichrom Peroxide Assay Kit (BioAssay Systems), respectively. Molm14 cells, an acute myeloid leukemia (AML) cell line, were maintained in RPMI-1640 media supplemented with 10% FBS, 1% antibiotic antimycotics, 25 mM β-D Glucose, and 25 mM HEPES. Molm14 cells were seeded at 400,000 cells/mL in 24 well plate. A PEGDA-GOx hydrogel was placed in the well to induce hypoxia. Cell viability was assessed via trypan blue staining and counted by Countess II automatic cell counter (Life Technologies). COLO-357, a pancreatic cancer cell line, was maintained in high glucose DMEM supplemented with 10% FBS, 1% antibiotic antimycotics, and 25 mM HEPES. Cells were encapsulated in gelatin-norbornene (GelNB)-thiolated hyaluronic acid (THA) hybrid hydrogels via thiol-norbornene photopolymerization as described previously. Cell-laden hydrogel was cultured in the presence of GOx-immobilized hydrogel for 2 weeks with periodical exchange of GOx-immobilized gel to maintain solution hypoxia.

2.5. RNA isolation and gene expression

After the cell culture experiments, cells were collected for RNA isolation using NucleoSpin RNA II kit (TaKaRa Bio). Total RNA concentration and quality were characterized by NanoDrop 2000 (Thermo Scientific). Isolated RNA was stored at −80°C until reverse transcription into complementary DNA (cDNA) using PrimeScript RT reagent kit (TaKaRa Bio). Gene expression was analyzed by real time quantitative PCR (qPCR) using SYBR Premix Ex Taq II Kit (TaKaRa Bio) on a 7500 Fast Real-Time PCR machine (Applied Biosystems). Thermocycling parameters were: 1 cycle at 95°C for 30s, followed by 95°C for 3s, 60°C for 30s, and repeated for 45 cycles. Gene expression results were analyzed using 2^−ΔΔCT method using ribosomal 18S as the housekeeping gene. All groups were normalized to the media control group (expression values were set as one-fold). Table S1 lists all primer sequences used for real-time PCR.

2.6. Statistics

All statistical analyses were performed using GraphPad Prism 5 software. Significance comparison between experimental groups was performed using Two-Way ANOVA with Bonferroni post testing. All experiments were conducted a minimum of three times with data presentation as the mean ± standard error of the mean (SEM). One, two, or three asterisks represent p < 0.05, 0.01, or 0.001, respectively.

3. Results & Discussion

3.1. Effect of freeze-drying on oxygen consumption by GOx-immobilized hydrogels

We have recently reported the development of enzyme-immobilized hydrogels to induce solution hypoxia for in vitro cancer cell culture.¹⁶ In the previous iteration, GOx-immobilized PEGDA hydrogels (i.e., PEGDA-GOx gels) were always prepared fresh (i.e., one-day prior to experiments) to ensure maximum activity of the immobilized enzyme. Although this approach was effective in sustaining solution hypoxia (O₂ < 5%) for 24 hours, the GOxPEGA concentration used was considered high (e.g., 6 mg/mL). In an attempt to improve the flexibility of using PEGDA-GOx hydrogels and to provide long-term storage of the gels, we sought to freeze-dry the hydrogels such that the gels could be stored for longer duration. The PEGDA-GOx gels could be used to induce hypoxia upon rehydration. We first compared the effect of freshly prepared and freeze-dried PEGDA-GOx hydrogels on inducing solution hypoxia. Furthermore, PEGDA-GOx hydrogels were prepared with lower GOxPEGA concentrations (0.2 or 0.4 mg/mL) using photopolymerization of 15 wt% PEGDA_2kDa (Fig. 1A). The enzyme-immobilized hydrogels were either placed in DPBS overnight and used in experiments the next day (i.e., freshly prepared) or freeze-dried immediately after gelation (Fig. 1A). As shown in Fig. 1B, freshly-prepared PEGDA-GOx hydrogels (30 μL) immobilized with 0.2 or 0.4 mg/mL GOxPEGA were capable of maintaining low oxygen content (O₂ < 1 %) in buffer solution containing 25mM glucose for at least 5 hours. At 24 hours after placing the gel into the buffer solution, oxygen contents had increased to ~10% and 6% for gels immobilized with 0.2 and 0.4 mg/mL GOxPEGA, respectively. At 48 hours, solution oxygen contents in both groups increased to above 17%, suggesting that the immobilized GOx had lost its oxygen consumption activity. Further increasing GOxPEGA concentration in the hydrogel to above 1 mg/mL did not improve the duration of hypoxia (data not shown). These results were consistent to our previous work, where hypoxia was maintained roughly for 24 hours.¹⁶

Figure 1. Solution hypoxia induced by freshly prepared or freeze-dried PEGDA-GOx hydrogels.

(A) Schematic of preparing hypoxia-inducing hydrogel by crosslinking PEGDA with GOxPEGA. Gels were either freshly-prepared (B) or freeze-dried (C) before swelling/reconstituting in 1 mL PBS overnight, followed by incubation in buffer solution to induce hypoxia. Gel volume: 30 μl. [PEGDA_2kDa] = 15 wt%. Buffer: DPBS with 25 mM β-d-Glucose and 25 mM HEPES (***p<0.001. Mean ± SEM, n ≥ 3).

The freeze-dried PEGDA-GOx hydrogels were stored at −20°C and rehydrated in DPBS overnight prior to subsequent experiments. Ideally, this approach would not only increase the flexibility of experimental design but also add commercialization potential to this material system. After placing the rehydrated PEGDA-GOx gel in the buffer, solution hypoxia (O₂ ~ 1%) was achieved for both sets of gels (0.2 and 0.4mg/mL GOxPEGA) for only a few hours (Fig. 1C). By 24 hours, solution oxygen levels had risen to above 17%, suggesting that freeze-drying process significantly decreased enzyme stability and activity. This result was not surprising as freeze-drying process has been shown to negatively affect protein structure and enzyme activity.²⁵ Hence, a protein protection strategy was needed to preserve the activity of freeze-dried GOx-immobilized hydrogels.

3.2. Effect of cryoprotectants on catalytic activity of PEGDA-GOx hydrogels

Many oligosaccharides, including trehalose (Fig. 2A) and raffinose (Fig. 2B), have been used as ‘cryo-protectants’ to prevent structural and/or functional damages of protein therapeutics during freeze-drying process.^{18,20,22,26,27} For example, trehalose (Fig. 2A), an alpha-linked disaccharide can stabilize proteins in solution and hydrogels.¹⁷ Trehalose-based glycopolymers have also been synthesized to provide protection against temperature-induced damage.^19,21,23,28 In this study, unmodified D-trehalose or raffinose at 3 mg/mL was added in the prepolymer solution containing PEGDA and GOxPEGA. Note that in this experiment, GOxPEGA was increased to 0.8 mg/mL as this concentration sustained longer hypoxia (at least 24 hours, see No cryoprotectant group in Fig. 2C). Control experiments showed that the incorporation of trehalose/raffinose did not change oxygen consumption profiles of freshly prepared PEGDA-GOx hydrogels (data not shown). However, when the PEGDA-GOx hydrogels were polymerized and freeze-dried in the presence of trehalose or raffinose, oxygen consumption ability of the rehydrated PEGDA-GOx hydrogels was significantly improved (Fig. 2C). At 24 hours, PEGDA-GOx hydrogels with or without cryoprotectant reduced solution oxygen to ~2.5%, or ~3.5%, respectively. At 48 hours, both trehalose and raffinose containing PEGDA-GOx hydrogels maintained oxygen level to about 5%, whereas gels containing no cryoprotectant completely lost its oxygen consumption ability. Increasing oligosaccharide concentration in the prepolymer solution (up to 10 mg/mL) did not further improve oxygen consumption of the freeze-dried and rehydrated PEGDA-GOx hydrogels (data not shown). Interestingly, the protection effect of trehalose/raffinose did not extend to beyond 48 hours, suggesting other factors might exist to hinder long-term activity of immobilized GOx. Results in Fig. 1 and Fig. 2 demonstrated that freeze-drying was detrimental to the oxygen consumption activity of PEGDA-GOx hydrogels and that the inclusion of trehalose or raffinose in the gel formation restored, at least partially, the enzymatic activity of the hydrogels.

Figure 2. Effect of cryoprotectants on solution hypoxia induced by freeze-dried PEGDA-GOx hydrogels.

(A) Chemical structure of trehalose. (B) Chemical structure of raffinose. (C) Oxygen consumption induced by freeze-dried PEGDA-GOx hydrogels in the absence or presence of trehalose or raffinose (3 mg/ml) during gelation. Hydrogel volume: 30 μl. [PEGDA_2kDa] = 15 wt%. LAP: 1mM. [GOxPEGA] = 0.8 mg/ml. Oxygen tensions were measured at room temperature in DPBS with 25 mM β-d-Glucose, 5 mM Glutathione and 25 mM HEPES (**p<0.01. Mean ± SEM, n ≥ 4).

3.3. Attempts to extend solution hypoxia by gel or buffer exchange

While trehalose or raffinose partially restored the enzymatic activity of freeze-dried PEGDA-GOx hydrogels, the gel still lost its oxygen consumption activity sometime between 24–48 hours (Fig. 2C). We evaluated the possibility of extending the duration of hypoxia through replacing the old gel with a new one every 24 hours. Fig. 3A shows the timeline of this experiment, in which the control group (i) was conducted using the same gel throughout (i.e., no gel replacement) and the experimental group (ii) was performed with gel replacement at hour-24 and 48. O₂ measurements were performed at 1, 24, 25 (i.e., 1-hour after gel replacement), 48, 49, and 72 hours. The oxygen content in the control group (i.e., no gel replacement) increased to above 5% before the 48-hour measurement, a result similar to that shown in Fig. 1 and Fig. 2. In the experimental group, oxygen content decreased to ~2% after the first gel replacement (i.e., Gel 2). Within 1-hour of the second gel replacement (i.e., Gel 3), oxygen content decreased rapidly from ~15% to ~2%, indicated that the new gel was active. However, both new PEGDA-GOx gel 2 and 3 failed to maintain hypoxia as oxygen contents in the solution rose back to above 14% prior to the next oxygen measurement (at 48 and 72-hour). Since all gels were identical in crosslinking (i.e., 15 wt% PEGDA, 3mg/mL trehalose, 0.2 mg/mL GOxPEGA) and processing conditions (i.e., freeze-dried immediately after gelation), it could be concluded that the sources contributed to the loss of enzyme activity in PEGDA-GOx gel 2 and 3 were in the buffer solution. In a separate control experiment (data not shown), we utilized gels with 2 mg/mL GOxPEGA (i.e., 10-fold higher GOxPEGA). Surprisingly, gel with higher GOx content performed even worse as hypoxia was only maintained in the first two measurements (1 and 6 hours) and the solution was no longer hypoxic by the 24 hours measurement.

Figure 3. The effects of gel replacement and buffer volume and on oxygen consumption by PEGDA-GOx hydrogels.

(A) Timeline of the gel replacement study. (B) Effect of replacing PEGDA-GOx gel on oxygen consumption. (C) Timeline of the buffer volume study. (D) Effect of buffer volume on oxygen consumption by the same PEGDA-GOx gels. Hydrogel volume: 30 μl. [PEGDA_2kDa] = 15 wt%. LAP: 1mM. [GOxPEGA] = 0.8 mg/ml. Oxygen tensions were measured at room temperature in DPBS with 25 mM β-d-Glucose and 25 mM HEPES (*p<0.05. ***p<0.001. Mean ± SEM, n ≥ 3).

Fig. 3B showed that the gel replacement approach did not yield extended hypoxic effect and that the new gel lost its activity more rapidly than the initial gel with identical formulation and processing conditions. We hypothesized that the accumulation of reaction by-products, particularly H₂O₂, partially inhibited the activity of the immobilized GOx. To test this hypothesis, we setup an experiment where PEGDA-GOx gels were either placed in a low volume buffer (0.8 mL, condition i in Fig. 3C) or in a higher volume buffer (5 mL, condition ii in Fig. 3C) except for when an O₂ measurement was needed. In the latter case, the PEGDA-GOx gels were moved to a well containing 0.8 mL buffer 1 hour prior to O₂ measurements. The purpose of placing PEGDA-GOx gels in a larger volume buffer was to keep the concentrations of reaction by-products at minimum, from which to prevent potential adverse effects exerted on the immobilized enzyme. O₂ measurement results showed that, while the gel kept in the lower volume buffer maintained solution hypoxia for only 24 hours, the ones placed in higher buffer volume unless during O₂ measurement were able to sustain solution hypoxia for at least 48 hours (Fig. 3D).

In a separate experiment, we induced hypoxia with a fresh PEGDA-GOx hydrogel, followed by exogenously adding hydrogen peroxide. To expedite the study, we added concentrated H₂O₂ (i.e., 100 mM and 200 mM) and monitored the solution oxygen levels following the addition. As shown in Fig. 4A, oxygen levels were low before the addition of H₂O₂ (at the 0.5 hour mark) and H₂O₂ was added after the initial oxygen measurement. At 1 hour, solution oxygen contents increased to ~6% and ~13% for buffer added with 100mM and 200mM H₂O₂, respectively. After 2 hours, oxygen contents in both solutions reached near normoxia. Since both freshly-prepared and freeze-dried PEGDA-GOx hydrogels were capable of maintaining solution hypoxia for at least 24 hours when no exogenous H₂O₂ was added (Figs. 1, 2), the rapid loss (within one hour) of oxygen consumption activity of the PEGDA-GOx gels could be attributed, at least in part, to the accumulation of H₂O₂. A similar study performed using concentrated gluconic acid, another GOx reaction by-product, did not cause reduction of immobilized GOx activity (Fig. 4B). These control experiments further verified that the loss of oxygen consumption activity of PEGDA-GOx hydrogel was caused by accumulation of H₂O₂.

Figure 4. Inhibitory effects of GOx reaction by-products on oxygen consumption.

Hydrogen peroxide (A) or gluconic acid (B) was added to the buffer containing PEGDA-GOx hydrogels. Hydrogel volume: 30 μl. [PEGDA_2kDa] = 15 wt%. LAP: 1mM. [GOxPEGA] = 0.8 mg/ml. Oxygen tensions were measured at room temperature in DPBS with 25 mM β-d-Glucose and 0 or 25 mM HEPES for gluconic acid and hydrogen peroxide experiments, respectively. (*p<0.05. ***p<0.001. Mean ± SEM, n ≥ 3).

The negative effects of reaction products on GOx activity have been reported in the literature. For example, Kleppe reported that H₂O₂ inactivated soluble GOx from Aspergillus niger via modification of methionine residues located at or near the active site of the enzyme.²⁹ In another kinetic study of using GOx to produce calcium gluconate, Bao et al. showed that H₂O₂ competitively inhibited catalytic activity of immobilized GOx.³⁰ It is worth noting that while these kinetic studies are highly valuable in the understanding and optimization of enzyme reaction kinetics, the reaction time was relatively short (a few minutes). In the current study, we studied the long-term usage (48–72 hours) of GOx-immobilized PEGDA hydrogels for the purpose of in vitro hypoxic cell culture and the system exhibited similar inhibition as reported in these prior literatures.

3.5. Effect of solution GSH content on sustaining hypoxia

Since replacing gels could not extend the hypoxic environment beyond 24 hours, a new approach was needed for limiting the inhibitory effect of hydrogen peroxide. A common approach to mitigate the adverse effect of the accumulated H₂O₂ is to use catalase (CAT), which reduces H₂O₂ into water and a half mole of oxygen. We have applied this approach in our previous endeavor of using PEGDA-GOx hydrogels to create solution hypoxia for in vitro cell culture. However, the use of CAT hinders overall oxygen consumption as its reaction produces additional oxygen. To this end, we added glutathione (GSH) in the buffer to reduce reactive oxygen species, such as H₂O₂. Fig. 5A shows a potential mechanism by which two GSH molecules form GSSG while reducing H₂O₂ into water.^31,32 We reasoned that the reduction of H₂O₂ into water would prevent its accumulation, which adversely affected GOx activity. We found that oxygen contents in the buffers containing GSH remained at around 2–3% for 24 hours (Fig. 5B). By 48-hour measurements, the oxygen contents in the solutions increased to ~6% and ~3.5% for solution added with 2.5 mM and 5.0 mM of GSH, respectively. By 72-hour measurement, the solution oxygen level remained at around 5% for buffer added with 5 mM GSH (Fig. 5B). Please note that the GOxPEGA concentration used in this study was 0.2 mg/mL (as opposed to 0.8 mg/mL in Fig. 2). A low GOxPEGA concentration was intentionally used here to highlight the effectiveness of adding GSH to mitigate the loss of enzyme activity. We also quantified the concentration of GSH remaining in the solution during the course of the experiment. As shown in Fig. 5C and 5D, GSH was added to the solution at 2.5 mM or 5 mM and the thiol contents were measured hourly for 5 hours using Ellman’s assay. Clearly, GSH contents decreased more rapidly in the presence of PEGDA-GOx hydrogels, presumably due to disulfide bond formation caused on H₂O₂ mediated oxidation (Fig. 5A).

Figure 5.

Effect of GSH on oxygen consumption. (A) Oxidation reaction mechanism of GSH by H2O2. (B) PEGDA-GOx hydrogel induced solution hypoxia was prolonged by the addition of GSH. (C) GSH consumption in the presence or absence of PEGDA-GOx hydrogel. (D) Solution hypoxia induced by PEGDA-GOx hydrogel and exogenously added GSH (5mM, added only at the beginning of experiment or periodically at 6 or 24 hours). Hydrogel (30 μl) were polymerized with 15 wt% PEGDA_2kDa, 0.2 mg/ml GOxPEGA, 3 mg/ml trehalose and 1mM LAP. (***p<0.001. Mean ± SEM, n ≥ 3).

The reduction of H₂O₂ due to GSH likely improves enzyme activity. To further demonstrate this 5mM GSH was either added one time at 0 hour or periodically at 0, 6, and 24-hour. Periodic addition of GSH could improve the functionality of the gel as 5 mM GSH was completely consumed after 5 hours. Oxygen tension was lower with periodic addition GSH as seen in Fig. 5D and oxygen concentration was maintained below 3% after 48 hours, whereas oxygen concentration was around 5% at 48 hours with one time GSH addition.

3.6. Cellular response to the improved hypoxia-inducing PEGDA-GOx hydrogels

In our previous report, soluble CAT was used in conjunction with PEGDA-GOx hydrogels to mitigate the negative effects of hydrogen peroxide on viability of hepatocellular carcinoma cells Huh-7 and acute leukemia cells Molm-14.¹⁶ From the perspective of oxygen consumption, using CAT was not ideal as it generated oxygen that partially offset hypoxia induced by immobilized GOx. Here, we examined the effect of GSH and optimized PEGDA-GOx gels on Molm14 cell viability. As shown in Figure 6A, Molm-14 cell viability was not affected either in normoxic (gel-free) or in hypoxic (with PEGDA-GOx gel) culture for at least 24 hours. Cell viability was significantly impacted under hypoxic culture at 48 hours. To ensure that decreased cell viability was due to low oxygen content and not other factors related to the PEGDA-GOx gels (i.e. reaction by products H₂O₂ and gluconic acid), we compared the relative cell viability when hypoxia was induced with either a hypoxic chamber or with the PEGDA-GOx gel. As shown in Fig. 6B, cell viability was not significantly different between the two methods to induced hypoxia. This result suggests that exposing Molm-14 cells under long-term hypoxia negatively affected cell viability and that the PEGDA-GOx hydrogel was not cytotoxic for long-term cell culture as we stated in our previously publication.¹⁴

Figure 6.

Effect of PEGDA-GOx gel induced hypoxia (A) and gas-controlled chamber induced hypoxia on viability of Mol14 cells over time. 25 mM glucose, 25 mM HEPES, and 5 mM GSH were added to PEGDA-GOx gel containing media. Additional GSH (5mM) was added at 6-hour and 24-hour. PEGDA-GOx gels: 30 μl; 15 wt% PEGDA_2kDa, 0.8 mg/ml GOxPEGA, 3 mg/ml trehalose. (Mean ± SEM, n ≥ 4. n.s.: Not significant. ***p<0.001).

3.7. Effect of hypoxia-inducing PEGDA-GOx gel on pancreatic cancer cells (PCC) cultured in 3D hydrogels.

PEGDA-GOx hydrogels could be used to create long-term hypoxia for 3D cell culture. Specifically, we encapsulated pancreatic cancer cell line COLO-357 in a hyaluronic acid-gelatin hybrid (HAG) hydrogel system previously developed in our laboratory for studying the synergistic effect of matrix stiffness and HA on PCC growth and morphogenesis in 3D.^33,34 Here, we examined the effect of hypoxia (created by PEGDA-GOx hydrogel) on the behaviors of PCCs in 3D over the course of 14 days. We exposed the cell-laden HAG hydrogels to hypoxic media throughout the study by transferring the HAG gels from old media into new media every 48-hour, a time frame coincides with regular media refreshment. To ensure hypoxic environment throughout the study, all fresh media was pre-equilibrated with a new PEGDA-GOx gel for 1 hour, a time sufficient to establish solution hypoxia Fig. 1–3). Although all cell-laden HAG hydrogels were kept in conventional cell culture incubator, only those incubated with PEGDA-GOx gels experienced hypoxic condition, which were measured before and after each gel transfer (<5%, data not shown). Fig. 7A shows that, under normoxic culture, encapsulated PCCs grew into larger spheroids by day 14. On the other hand, the growth of encapsulated cells under hypoxia induced by PEGDA-GOx gels was inhibited, as demonstrated by the formation of smaller spheroids (Fig. 7B). It is worth noting that essentially no dead cells were found in either normoxic or hypoxic culture throughout the 2-week study, suggesting that hypoxia inhibited cell proliferation, rather than cell death. It was not likely that the reduced cell proliferation was caused by the lack of glucose as the media were replenished every other day. Furthermore, since the glucose concentration used in this study (25 mM) was much higher than the maximum oxygen concentration in the solution that the enzyme could consume (~0.27 mM), it was not likely that the reduced cell proliferation was caused by the lack of glucose. In our earlier work,¹⁶ we have shown that bolus addition of glucose did not change oxygen consumption profiles, suggesting that the initial amount of glucose was not a limiting factor in the efficiency of oxygen consumption. We further evaluated mRNA expression of hypoxia-related genes in these cells at day-14, including HIF-1α and sonic hedgehog (SHH). Results showed that cells cultured in 3D and under PEGDA-GOx gel induced hypoxia expressed higher level of these genes. Higher expression of HIF-1α, the master regulator of cellular adaptive response to hypoxia,^35,36 suggested that the cells still experienced hypoxic condition at day-14. Furthermore, the upregulation of SHH in PCCs cultured with PEGDA-GOx hydrogels (i.e., under hypoxia) demonstrated the hypoxic cellular response. This result was in line with a recent finding that the expression and secretion of SHH from PCCs is HIF-1-dependent. Increased SHH secretion potentially is responsible for the formation of stroma-rich microenvironments in pancreatic cancers.³⁷

Figure 7.

Effect of enzyme induced hypoxia on cell fate of COLO 357 cell- laden gels. (A) COLO-357 cell morphology under normoxia (control) or hypoxia. (B) mRNA expression. Housekeeping gene: Ribosomal 18s. (C) Cell size distribution. Hypoxia was induced by 30 μl hydrogels polymerized with 15 wt% PEGDA_2kDa, 0.2 mg/ml GOxPEGA, 3 mg/ml trehalose and 1mM LAP. GOxPEGA gel was placed in the same well as the cell-laden hydrogel (*p<0.05, **p<0.01. Mean ± SEM, n ≥ 3).

4. Conclusion

In summary, we have improved upon the previous GOx-immobilized PEGDA hydrogel system to sustain solution hypoxia for extended in vitro cell culture. The negative effect of freeze-drying on enzyme activity was mitigated with the addition of cryoprotectants (e.g., trehalose and raffinose) in the PEGDA-GOx hydrogel formulation, creating a more convenient hypoxia-inducible hydrogel system. Furthermore, the decrease in enzyme activity due to the accumulation of hydrogen peroxide was alleviated via supplementing reducing agent glutathione in the buffer/medium. The optimized PEGDA-GOx gels were capable of inducing long-term hypoxia for 3D culture of a biomimetic cell-laden hydrogel, demonstrating by the upregulation of hypoxia related genes. Future work will focus on creating oxygen gradients with the PEGDA-GOx hydrogels.