Transient Hypoxia Improves Matrix Properties in Tissue Engineered Cartilage

Abstract

Adult articular cartilage is a hypoxic tissue, with oxygen tension ranging from <10% at the cartilage surface to <1% in the deepest layers. In addition to spatial gradients, cartilage development is also associated with temporal changes in oxygen tension. However, a vast majority of cartilage tissue engineering protocols involves cultivation of chondrocytes or their progenitors under ambient oxygen concentration (21% O₂), that is, significantly above physiological levels in either developing or adult cartilage. Our study was designed to test the hypothesis that transient hypoxia followed by normoxic conditions results in improved quality of engineered cartilaginous ECM. To this end, we systematically compared the effects of normoxia (21% O₂ for 28 days), hypoxia (5% O₂ for 28 days) and transient hypoxia—reoxygenation (5% O₂ for 7 days and 21% O₂ for 21 days) on the matrix composition and expression of the chondrogenic genes in cartilage constructs engineered in vitro. We demonstrated that reoxygenation had the most effect on the expression of cartilaginous genes including COL2A1, ACAN, and SOX9 and increased tissue concentrations of amounts of glycosaminoglycans and type II collagen. The equilibrium Young’s moduli of tissues grown under transient hypoxia (510.01 ± 28.15 kPa) and under normoxic conditions (417.60 ± 68.46 kPa) were significantly higher than those measured under hypoxic conditions (279.61 ± 20.52 kPa). These data suggest that the cultivation protocols utilizing transient hypoxia with reoxygenation have high potential for efficient cartilage tissue engineering, but need further optimization in order to achieve higher mechanical functionality of engineered constructs.

Articular cartilage resides in hypoxic environment with oxygen concentrations ranging from 10% at the surface to <1% in the deep zone.¹ Oxygen and nutrient exchange within cartilage tissue depend on diffusion from the synovial fluid that flows through the tissue during the joint movement.² Chondrocytes, the constitutive cells of articular cartilage, are sparsely distributed in dense extracellular matrix (ECM) that lacks vascular supply. In order to survive in such a harsh environment, chondrocytes must be able to sense oxygen availability and adjust cellular metabolism to consume less oxygen at lower oxygen concentrations.^3,4

Due to its complex biological characteristics, lack of vascular supply and low cell concentration, articular cartilage has poor capability to heal following injury or disease. The need for developing effective modalities for cartilage repair has motivated tissue engineering research toward establishing methods for restoring cell metabolism, tissue architecture, and load bearing capacity. Today, most cartilage constructs are engineered by cultivation of chondrocytes or their progenitors under ambient oxygen concentrations (21%) that is much higher than the oxygen level in native joints.^5–8 In contrast, hypoxic conditions are inherent to many physiological and pathological processes, such as adaptation to high altitudes, wound healing, inflammation, pathology of cancer, and ischemia.⁹

During limb development, the differentiation of mesenchymal cells into chondrocytes and early formation of the tissues in the joints occur at low oxygen levels in which hypoxic inducible factor (HIF) plays an important role in cellular adaptation to hypoxia.¹⁰ HIF is a heteromeric transcription factor that mediates the effects of SOX9, a chondrogenic transcription factor responsible for skeleton formation as it co-localizes in regions where cartilage matrix is being deposited.^11,12 Akiyama et al.¹³ reported that the absence of Sox9 during limb development resulted in malformation of cartilage and bone. They found that Sox9-inactive mesenchymal cells remained at the condensation stage and were unable to undergo chondrogenesis. In addition, Sox9-inactive animals could not produce cartilage ECM.¹³

Overall, the molecular mechanisms of type II collagen and aggrecan synthesis and assembly into mechanically functional cartilage ECM are not fully understood.¹⁴ One putative mechanism involves transcriptional control of ECM synthesis via SOX9 binding to responsive sequences of aggrecan and type II collagen.¹⁵ The orchestrated regulation of cartilage ECM production by the HIF and SOX9 supports the assumption that hypoxia is a favorable condition for maintaining cartilage structure and function. However, not only oxygen tension per se is of interest. Spatial and temporal gradients of oxygen tension may play a crucial role during the formation of functional cartilage tissue as well.¹⁶

This study was designed to examine the effects of transient hypoxia, with an initial exposure to hypoxia (to activate cell proliferation) followed by normoxia (to enhance matrix synthesis), on the composition and mechanical properties of engineered cartilage. Juvenile chondrocytes were selected for these studies in order to better understand the effects of oxygen in fetal and juvenile cartilage development, and thereby improve cartilage regeneration. Chondrocytes were encapsulated in agarose hydrogel, one of the best characterized scaffolds for cartilage tissue engineering, and cultured under three sets of conditions: hypoxic (5% O₂/28 days), normoxic (21% O₂/28 days), and transient hypoxia/reoxygenation (5% O₂/7 days ″ 21% O₂/21 days). Oxygen levels in the culture medium were monitored using in-line oxygen sensors.

METHODS

Cell Isolation

Full-thickness articular cartilage was harvested from fresh bovine carpometacarpal joints obtained from 4- to 6-month-old bovine calves. Cartilage was minced, rinsed in PBS and digested with 390 unit/ml collagenase type IV (Sigma–Aldrich, St. Louis, MO) in Dulbecco’s Modified Essential Medium [hgDMEM supplemented with 10% fetal bovine serum (FBS) and antibiotics] for 10 h at 37°C with stirring.¹⁷ The resulting digest was filtered through a 70 μm pore size cell strainer to isolate individual cells and remove undigested tissue. The cell suspension was centrifuged to form chondrocyte pellets that were rinsed with PBS, resuspended in culture medium and plated at high density (2.5 × 10⁵ cells/cm²) in chondrocyte growth medium (hgDMEM supplemented with 10% FBS, 10 mM HEPES, 100 U/ml penicillin).

Cell Encapsulation

A suspension of cells in culture medium (40 × 10⁶ cells/ml) was mixed with an equal volume of 4% agarose (type VII, Sigma–Aldrich) in PBS at 40°C to yield a final concentration of 20 × 10⁶ chondrocytes/ml in 2% agarose. The cell–agarose mixture was cast between two sterile glass plates separated by a 1 mm spacer to form a rectangular slab (70 mm × 80 mm × 1 mm). Disks were cored out of the slab using a biopsy punch, to obtain cylindrical constructs (4 mm diameter × 1 mm thick), which were transferred into 24-well plates integrated with an oxygen sensor platform (PreSens, Germany). Each construct was cultured in a separate well in 1 ml of chondrogenic medium (high glucose DMEM supplemented with 5 mg/ml proline, 1% ITS+, 100 nM dexamethasone, 50 μg/ml ascorbate-2-phosphate, 10 mM HEPES, 100 U/ml penicillin, and 100 μg/ml streptomycin) with medium change twice a week. For the first 14 days, chondrogenic medium was additionally supplemented with 10 ng/ml TGF-β3 (Invitrogen, Carlsbad, CA) as in previous studies.¹⁸

Experimental Design

All experiments were performed in triplicate, using four joints in each of the three individual studies (n = 12 joints total). Data are represented as mean ± SD for n = 5 constructs engineered using cells from one animal, to minimize batch-to-batch variability, as reported in several previous studies.^19,20 Cartilage constructs were cultured in static culture under three different oxygen supply regimes as (Fig. 1). Normoxic group (21% O₂ for 28 days) was cultured in a chamber (Billups-Rothenberg, Inc., Del Mar, CA) that was maintained in humidified air containing 21% O₂, 5% CO₂ (normal incubator conditions). Hypoxic group (5% O₂ for 28 days) was cultured in an airtight chamber flushed daily with a humidified gas mixture (5% O₂, 5% CO₂, and 90% N₂) to equilibrate culture medium at 5% oxygen. Reoxygenated group was maintained at 5% O₂, 5% CO₂, and 90% N₂ for 7 days and then transferred to 21% O₂, 5% CO₂, and 90% N₂ for additional 21 days. Humidity was maintained by adding 20 ml water into a Petri dish placed in the chamber. To validate the consistency of oxygen levels during cultivation, oxygen levels in culture medium were monitored continuously by oxygen Sensor Dish Reader (PreSens) for 20 h after each medium change (Fig. 1).

Figure 1.

Experimental design. Tissue constructs were cultured in 24-well plates. Each well contained one tissue construct and was fitted with an oxygen sensor that measured oxygen concentration in real time by using a SDR SensorDish^® Reader. Normoxic and hypoxic groups were maintained at 21% and 5% O₂, respectively, for 28 days. Reoxygenation group was maintained at 5% O₂ for 7 days followed by 21% O₂ for 21 days. Medium was changed twice a week (red arrows). Oxygen levels were measured and recorded for 20 h after media replacement (blue arrows).

DNA Content

Constructs (n = 5 per group and time point) were harvested on Days 0, 7, 14, 21, and 28 and digested for 16 h at 56°C with 20 μl/ml papain in 1 mg/ml of proteinase K (Fisher Scientific, Pittsburgh, PA) containing 1 mM iodoacetamide and 10 mg/ml pepstatin-A (Sigma–Aldrich). Total DNA content was quantified using PicoGreen assay (Invitrogen) following the manufacturer’s protocol.

Glycosaminoglycan (GAG) Content

Aliquots of digested tissue samples were analyzed using the 1,9-dimethylmethylene blue dye binding (DMMB) assay²¹ to determine the GAG content.

Collagen Type II ELISA

Constructs harvested after 28 days of culture were homogenized in an ice-cold mortar, resuspended in 0.8 ml of 0.05 M acetic acid containing 0.5 M NaCl, pH 3.0, mixed with 0.1 ml of 10 mg/ml pepsin solution in 0.05 M acetic acid, and stored at 4°C for 48 h. The pH of samples was adjusted to 8.0 using 1 N NaOH. The samples were digested using 0.1 ml of 1 mg/ml pancreatic elastase in 1X TBS (0.1 M Tris, 0.2 M NaCl, 5 mM CaCl₂, pH 8) at 4°C overnight on a rotating rocker and centrifuged at 10,000 rpm for 5 min. This double enzymatic digestion was performed to obtain monomeric collagen, by first digesting collagen fibrils into polymeric collagen by protease (pepsin) and then converting polymeric collagen into monomeric form by elastase digestion. Supernatant was collected and diluted in assay buffer according to the manufacturer’s protocol (M.D. Bioproducts, St Paul, MN). The absorbance at 450 nm was plotted against concentration to obtain a standard curve by a 4-parameter logistic (4-PL) curve fit that was used to determine the amounts of collagen type II.

Mechanical Properties

Compressive properties of constructs were measured in un-confined compression using a custom-made mechanical testing device.²² Constructs were placed in a testing chamber and equilibrated under a creep tare load of 0.5 g for 30 min. Stress-relaxation tests were performed at the ramp velocity of 1 μm/s up to 10% strain. The equilibrium Young’s modulus (E_Y) was determined from the equilibrium stress–strain data.

Real-Time PCR

Constructs were extracted to isolate the total RNA using TRIzol^® Reagent (Invitrogen), treated with DNAse I (Ambion, Austin, TX) and quantified using NanoDrop™ Spectrophotometer (Thermo Scientific, Wilmington, DE). Reverse transcription was performed using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA). Quantitative PCR was carried out using the 7500 fast real-time PCR system. The following TaqMan^® Gene Expression Assays were used for detection of cartilaginous gene expression: COL2A1 (Bt03251861_m1), COL1A1 (Bt03225322_m1), ACAN (Bt03212186_m1), and SOX9 (custom designed forward primer: ACGCCGAGCTCAGCAAGA; reverse primer: CACGAACGGCCGCTTCT; probe: CGTTCA-GAAGTCTCCAGAGCTTGCCCA).²³ Gene expression values were reported in relative levels to GAPDH (Bt03210913_g1) by the 2^−ΔCt method.²⁴ All reactions were performed in triplicates. Representative graphs are shown with error bars indicating standard deviation of four samples for each oxygen condition.

Histology and Immunohistochemistry

Constructs were fixed in 4% paraformaldehyde overnight at 4°C, transferred to 70% ethanol, embedded in paraffin and sectioned at 8 μm. The sections were stained with hematoxylin and eosin for general evaluation, and safranin-O for GAG. Sections for immunohistochemistry staining were hydrated, and antigen retrieval was performed using heated 0.01 M citrate buffer with pH 6.0 for 15 min. Quenching of the endogenous peroxidase was done by immersing the sections in 0.3% H₂O₂/methanol for 10 min at room temperature. The sections were incubated with blocking serum (Vectastain ABC, Burlingame, CA) for 30 min at room temperature, rinsed with PBS, incubated overnight at 4°C with 1:1,000 of type II collagen monoclonal antibody (Millipore, Temecula, CA) and for 30 min with biotinylated secondary antibody (Vectastain ABC). For signal enhancement and detection, Vectastain ABC Kit with peroxidase and DAB Peroxidase Substrate Kit (Vectastain ABC) were added as described in the manufacturer’s protocol.

Statistical Analysis

Statistics were performed with STATISTICA software (Stat-soft, Tulsa, OK). Data were expressed as the average ± SD of n = 4–6 samples per group and time point. The differences in construct properties between the groups were examined by analysis of variance (α = 0.05), with DNA, matrix contents, E_Y or relative level of target gene expression as the dependent variable, followed by Tukey’s Honest Significant Difference Test.

RESULTS

Oxygen Levels in Culture Medium

The level of O₂ in culture medium was measured to validate each of the oxygen regimes (normoxia, hypoxia, reoxygenation). Partial pressures of O₂ were measured in wells containing constructs and reference wells without constructs (Supplementary Fig. 1). Oxygen uptake rate (OUR) was estimated from a steady-state balance of O₂ in medium²⁵:

$$\frac{\mathrm{d}[{\mathrm{O}}_2]}{\mathrm{d}t}=k_{\mathrm{L}}a({[{\mathrm{O}}_2]}_{\text{reference}}-{[{\mathrm{O}}_2]}_{\text{construct}})-\text{OUR}$$

where [O₂] _reference and [O2]_construct are dissolved O2 concentration in wells without constructs and wells containing a cartilage construct, respectively, and k_L a = 0.9/h is the volumetric liquid phase mass transfer coefficient.²⁶ Assuming steady state without change in total O₂ level over a period of 20 h, OUR could be calculated as follows:

$$k_{\mathrm{L}}a({[{\mathrm{O}}_2]}_{\text{reference}}-{[{\mathrm{O}}_2]}_{\text{construct}})=\text{OUR}$$

The cells consumed less oxygen in hypoxic conditions than in normoxic and reoxygenated conditions as indicated by measured values of the oxygen uptake rate (OUR; Supplementary Fig. 2). The calculated values of OUR were in the range of those previously reported (Supplementary Table 1).

Effects of Hypoxia on Cell Proliferation, Proteoglycan Synthesis, and Mechanical Properties of Engineered Cartilage

Chondrocytes encapsulated in agarose hydrogel survived at all oxygen levels, from 21% O₂ (normoxia) to 5% O₂ (hypoxia) continuously or followed by reoxygenation. Live cells were observed in constructs cultured at all oxygen tensions. Cell proliferation under hypoxic conditions increased slightly compared to normoxic conditions by Day 7 (Fig. 2A). Effects of hypoxia on cell proliferation were seen by Day 14 (9.23 ± 0.28 μg DNA in hypoxia versus 7.95 ± 0.46 μg in normoxia, p = 0.016) and Day 21 (12.4 ± 0.32 μg at hypoxia vs 9.99 ± 0.72 μg at normoxia, p = 0.00017).

Figure 2.

Effect of oxygen exposure on cartilage tissue development. Constructs from the three experimental groups described in Figure 1 were analyzed for DNA content (A), GAG production (B), GAG/DNA (C), and compressive Young’s modulus, E_Y (D). Hypoxia-reoxygenation culture best maintained DNA content and promoted GAG synthesis. E: Constructs cultured for 28 days at different oxygen levels were specifically quantified for type II collagen content. There were no differences in type II collagen production between constructs cultured at 21% and 5% O₂, whereas hypoxia-reoxygenation significantly promoted type II collagen synthesis. Error bars denote standard deviation, *p < 0.05 versus 21% O₂, ^Ψp < 0.05 versus reoxy, ^ξp < 0.05 versus previous time point within the same group, n = 5.

Reoxygenated cultures showed similar cell proliferation patterns to hypoxic cultures up to Day 21. The hypoxia and hypoxia-reoxygenation groups demonstrated significant growth initially in comparison with the normoxia group. However, a decrease in DNA content in the hypoxic group (9.48 ± 1.62 μg) was observed by Day 28 while normoxic (11.30 ± 1.41 μg) and reoxygenated groups (12.35 ± 1.5 μg) maintained cell proliferation throughout the study.

Proteoglycan production was initially comparable in normoxic and reoxygenated cultures. However, the GAG content of normoxic group reached a plateau at Day 21 (500.53 ± 15.43 μg on Day 21 and 480.70 ± 16.68 μg on Day 28), whereas that of the reoxygenated group continued to increase (485.71 ± 6.43 μg on Day 21 and 597.35 ± 9.71 μg on Day 28). The GAG content of the reoxygenated group was significantly higher than that in either normoxic or hypoxic groups at Day 28 (Fig. 2B).

GAG accumulation in the hypoxic group on Day 28 was significantly lower in comparison to both the normoxic and reoxygenated groups, consistent with the lower cell numbers resulting from slower cell proliferation under hypoxic conditions. Hypoxic group had low values of GAG/DNA on Day 14, corresponding to the time-point when DNA content was higher and GAG content lower than in the other groups (Fig. 2C). Continuous normoxia maintained the DNA and GAG production over time in culture as indicated by the constant GAG/DNA values in this group.

The reoxygenated group gradually increased GAG/DNA production as the tissue constructs were maturing (Fig. 2C), in accordance with the increase in mechanical properties (Fig. 2D). At the end of the culture period, reoxygenated constructs yielded the highest compressive Young’s modulus (E_Y) of 510.01 ±28.15 kPa as compared to constructs cultured in normoxic (417.60 ± 68.46 kPa) and hypoxic (279.61 ± 20.52 kPa) conditions.

Reoxygenation Promotes Expression of Cartilaginous Genes

Real-time PCR was performed to evaluate cartilage tissue development at the transcriptional level. Total RNA of constructs was used to detect the expression of cartilaginous markers (COL2A1 and ACAN), a key transcription factor of chondrocytes (SOX9), and a key dedifferentiation marker (COL1A1; Fig. 3). During early phases of culture, chondrocytes in all groups showed low expression of COL2A1, the gene encoding for type II collagen. By Day 21, normoxic cultures gradually increased the COL2A1 gene expression and suppressed expression of COL1A1, the gene encoding for type I collagen. The COL2A1 gene expression in the reoxygenated group was upregulated to an even higher degree than in normoxic group, whereas hypoxia downregulated COL2A1 gene expression by half. Reoxygenation temporarily promoted COL1A1 gene expression (on Day 14), followed by suppression of this de-differentiation marker in mature constructs (Days 21 and 28).

Figure 3.

Gene expression of cartilaginous markers in normoxic, hypoxic, and reoxygenated cultures. Cartilaginous gene expression in constructs grown in normoxic, hypoxic, or hypoxia-reoxygenation conditions were determined by real-time PCR and normalized to GAPDH levels. The chondrogenic dedifferentiation marker, COL1A1, decreased with time in all groups. The expression levels of COL2A1, ACAN, and SOX9 genes significantly increased in the hypoxia-reoxygenation group as compared to either normoxic or hypoxic groups. Data are shown as average ± SD (n = 5). *p < 0.05 versus 21% O₂, ^Ψp < 0.05 versus reoxygenated group, ^ξp < 0.05 versus the previous time point within the same group.

The expression of ACAN, the gene encoding for core protein aggrecan, increased over time in all groups and the expression profiles were consistent with type II collagen expression. SOX9 was upregulated in the reoxygenated group by Day 21, and increased further by Day 28. Enhanced expression of SOX9 paralleled with enhanced expressions of COL2A1 and ACAN.

Type II Collagen Synthesis in Engineered Cartilage

Development of functional cartilage, in vitro or in vivo, largely depends on the ability of the cells to synthesize and assemble type II collagen, a trimeric fibrous protein abundant in articular cartilage. To assess the amounts of type II collagen in engineered cartilage constructs, tissue samples were collected and enzymatically digested to obtain monomeric collagen before performing ELISA. Reoxygenated culture resulted in significantly more type II collagen (14.73 ± 1.25 μg/ml) than either normoxic (10.10 ± 1.67 μg/ml) or hypoxic (9.31 ± 1.94 μg/ml) conditions (Fig. 2E).

Histology of Engineered Cartilage

Constructs cultured under normoxic, hypoxic, and transiently hypoxic conditions exhibited similar histomorphologies. Chondrocyte-seeded hydrogels progressively transformed into stiff and opaque tissue constructs over 28 days of culture. Chondrocytes at the construct centers were uniformly distributed in small cell clusters, while chondrocytes at the periphery formed larger clusters (Fig. 4A). Safranin O staining showed homogeneous spatial distributions of GAG (Fig. 4B). Partial GAG loss was observed at the constructs edges, by faint GAG staining, a phenomenon most pronounced under hypoxic conditions. Type II collagen was located in the intercellular spaces and localized around the cells, as shown by immunohistochemistry (Fig. 4C). Notably, reoxygenated constructs showed stronger type II collagen staining than the other two groups.

Figure 4.

Histology and immunohistochemistry of 28-day constructs from the normoxic, hypoxic and hypoxia-reoxygenation groups. A: H&E staining showed that chondrocytes are distributed throughout constructs, with cell clusters located on construct periphery. B: Safranin O stain for glycosaminoglycan (GAG). C: Immunostain for type II collagen. Arrows indicate strong type II collagen staining areas in the reoxygenated group.

DISCUSSION

Cultivations of engineered cartilage at reduced levels of oxygen tension have been investigated with varying degrees of success. In general, cultures were subjected to a constant level of hypoxia, for a period of up to 4 weeks,^27–30 without transferring cultures low and high oxygen environments. The effects reported from these studies were controversial. The implementation of hypoxia in cartilage tissue engineering resulted in either adverse effects of low oxygen on cell growth and ECM assembly, or no significant effects.^31–33 The studies by Yang et al.³² examining juvenile chondrocytes seeded in PGA scaffolds showed no difference in cell numbers and GAG contents in static cultures maintained at 5% and 21% O₂. They also found that total collagen content significantly decreased in static culture under hypoxia compared to normoxia. Hypoxia thus plays a crucial roles in the chondrogenic differentiation of stem cells.^34–36 Adipose-derived adult stromal cells (ADASs) that were expanded at 2% O₂ for two passages (8 days) and subjected to chondrogenic differentiation at 21% O₂ (6 days) increased their potential for differentiation into chondrogenic linages and attenuated expression of the osteogenic transcription factor Runx-2.³⁴

Engineered cartilage constructs were prepared in small dimensions (1 mm thick × 4 mm in diameter) in order to minimize the distance of diffusion to the centers of constructs. In this way we ensured that the low oxygen tension throughout the construct was the result of the applied 5% O₂ culture conditions and not of limited diffusion from the construct edges to the center. One limitation of the present study is that the chondrocytes were obtained from young animals (4–6 months old bovine calves). We have chosen to use juvenile chondrocytes to mimic some aspects of cartilage development,³⁷ such as the avascular and hypoxic environment in epiphyseal cartilage that is associated with the production of angiogenic inhibitors.³⁸ As a result, the center of the epiphysis is hypoxic, and survival of the juvenile chondrocytes is dependent on anaerobic glycolysis regulated by HIF-1α.³⁹ Juvenile chondrocytes express HIF-1α and their response to hypoxia resembles some aspects of the native development.

We demonstrated that engineered cartilage can survive and maintain matrix production in a 5% O₂ environment, in accordance with some published studies.^29,40 Although chondrocytes can adapt to low oxygen tension, long-term hypoxia had negative effects on GAG production, DNA content, and type II collagen synthesis (Figs. 2–4). Reoxygenation maintained cell proliferation, and enhanced proteoglycan and collagen type II along with upregulation of COL2A1 throughout the duration of culture (Fig. 3) significantly higher GAG and collagen contents as compared to either normoxic or hypoxic group (Fig. 2B,E).

We also observed that the Young’s modulus (E_Y) in the reoxygenation group and normoxic group was significantly higher than in the hypoxia group (Fig. 2D). Comparable values of E_Y in groups that exhibited differences in gene expression and ECM composition confirm previous findings that total GAG and collagen contents are not the only predictors of mechanical properties of tissue-engineered constructs.^19,20 Mechanical properties of engineered cartilage are also influenced by crosslinking molecules (such as cartilage oligomeric matrix protein-COMP, type IX collagen, type XI collagen) and ECM organization.⁴¹

ECM network formation is a multistep process that involves fibrous protein synthesis, secretion, and assembly.^42–45 Ultrastructural organization including the alignment of collagen fibers to form collagen bundles and the attachment of GAGs such as N-acetylgalactosamine and N-glucuronic acid on the core proteins^46,47 occurs during distinct time periods.¹⁶ In the current study, we observed a trend of improved mechanical properties in the reoxygenated constructs within 28 days, that was significant compared to the hypoxic group but not the normoxic group. The prolongation of the culture time up to 42 days^17,19 would likely provide sufficient time for complete matrix organization and result in a significant increase in E_Y modulus. Another factor that may delay ECM elaboration is the low cell seeding density. Mauck et al.⁴⁸ showed that constructs with seeding density of 60 × 10⁶ cells/ml in agarose hydrogels had higher GAG content and mechanical stiffness than constructs seeded at 20 × 10⁶ cells/ml.

Constructs cultured in hypoxic conditions always showed the lowest GAG/DNA ratio compared to the other groups (Fig. 2C). We hypothesize that under extended hypoxia, chondrocytes may preferentially maintain cellular metabolism and viability via increased glucose uptake,⁴⁹ maintenance of pH homeostasis,^50,51 and reduction of oxygen consumption rate⁵² rather than undergoing the heavy metabolic demands of ECM synthesis. Mobasheri et al.³ presented the dual model of oxygen and glucose sensing, which proposed that transcription of hypoxia-responsive glucose transporter is mediated by HIF-1α during oxygen deprivation in combination with low intracellular ATP and low extracellular glucose. This finding suggested that hypoxia plays a role not only in activation of ECM synthesis but is also in maintenance of the energy status when cells are exposed to low oxygen tension. Furthermore, the application of acute hypoxia stimulated vesicular ATP release, which in turn influenced local purinergic signaling and affected cell metabolism.⁵³ More studies are needed to assess the relationship of hypoxia-induced purinergic signaling with cell metabolism and ECM synthesis.

We found that priming cartilage constructs in low oxygen conditions promoted cell proliferation and activated cartilaginous gene expression. Therefore, introducing low oxygen tension could be an alternative approach, especially if used in combination with other physiological stimuli such as deformational loading¹⁹ and hydrostatic pressure⁵⁴ to trigger cartilage tissue development in vitro. It is possible that utilization of hypoxia and mechanical stimuli in tissue culture may activate different tissue development pathways to coax cells into producing cartilage tissues.

Several studies have shown that utilizing two or more physiological environments during in vitro culture results in improvement of mechanical properties and GAG production, including mechanical loading plus growth factors,¹⁹ low oxygen tension and mechanical loading,^55,56 and growth factors and temporally graded hypoxia in the current study. However, introducing physiological stimuli in cartilage tissue engineering requires optimization due to contradictory outcomes when low oxygen tension and mechanical loading were applied to hMSC-seeded agarose hydrogels.⁵⁷

The necessity of dynamic loading in hypoxic culture may in fact depend on the size of constructs. Cylindrical cartilage construct sized 10 mm × 2 mm³² and 8 mm × 4 mm,⁵⁵ which are up to 15 times larger than the constructs used in our study (4 mm × 1 mm), showed significant increase in type II collagen synthesis and GAG content when mechanical stimulation was applied under 5% O₂. This result suggested that mechanical stimulation is required to enhance oxygen delivery to the center of constructs to avoid anoxia (≪1% O₂), which is linked to nitric oxide-induced damage to collagen fibrils and is associated with pathological conditions such as osteoarthritis.^1,58

This study raises the question of whether the in vitro engineered cartilage grown under hypoxic conditions has energy levels insufficient for ECM production, in contrast to the native cartilage, which can maintain cellular functions and matrix biosynthesis under chronic hypoxia.

Based on the collected data, we propose that the timing of exposure of hydrogel-encapsulated chondrocytes to hypoxia and normoxia can be optimized to enhance the expression of cartilaginous genes, synthesis of cartilage proteins, and the assembly of mechanically functional cartilage matrix. We assessed the effects of transient hypoxia with reoxygenation on the expression of cartilaginous genes, the composition of ECM (GAG and collagen type II contents) and mechanical properties of engineered cartilage. To the best of our knowledge, this is the only study exploring this important question. Other studies have focused on gene arrays and the production of proinflammatory mediators and reactive oxygen species.^59–61 We show that transient hypoxia can lead to greater increases in mechanical properties, levels of ECM components, and expression of chondrogenic genes (COL2A1, ACAN, and SOX9) compared to continuous hypoxia. In comparison to normoxic conditions, transient hypoxia also induced greater production of ECM components and chondrogenic gene expression, but the functional mechanical properties remained comparable. We hypothesize that short-term hypoxia increases cartilaginous gene expression, and that transfer to energy-rich normoxic conditions enhances the production and functional assembly of cartilaginous ECM. However, the precise duration of transient hypoxia intervals remains to be determined in order to achieve more functional mechanical properties of engineered cartilage.

In summary, both the level of oxygen and the timing of exposure to hypoxia and normoxia play important roles in the in vitro formation of engineered cartilage by chondrocytes encapsulated in agarose hydrogel. An initial exposure to hypoxia (to activate cell proliferation) followed by normoxia (to enhance matrix synthesis) resulted in best ECM composition and structure, with significant increases in cartilaginous gene expression (COL2A1, ACAN, and SOX9). Further optimization of the culture period and duration of transient hypoxia is needed in order to achieve higher mechanical functionality of engineered constructs. Other cell types, such as pluripotent stem cells not yet committed to chondrogenic linages, may require different regimes of exposure to hypoxia and normoxia for optimal outcomes.