Engineering zonal cartilage through bioprinting collagen type II hydrogel constructs with biomimetic chondrocyte density gradient

Chondrocytes were isolated from the knee joints of juvenile New Zealand white rabbits via enzymatic digestion. The animals were provided by the laboratory animal center and were sacrificed with the approval of the Animal Care and Use Committee at the Third Military Medical University. The cartilage was sequentially digested with 0.1 % trypsin in PBS at 37 °C and 5 % CO₂ for 30 min, followed by treatment with 0.025 % collagenase II (Sigma, USA) in high-glucose Dulbecco’s modified Eagle’s medium (DMEM; HyClone, China) supplemented with 10 % fetal bovine serum (FBS; HyClone, China) and 1 % penicillin and streptomycin (P/S; Beyotime, China) at 37 °C and 5 % CO₂ overnight. The isolated chondrocytes were expanded in 225 cm² culture flasks in high-glucose DMEM supplemented with 10 % FBS.

Collagen type II was extracted from swine knees via pepsin treatment and salt precipitation, as previously described [19]. The swine knee joints were obtained from the nearby slaughter market and used with the approval of the suppliers. The collagen type II precipitant was dissolved in 0.15 M hydrochloride acid to obtain pre-gel at a 10 % (wt/vol) final concentration, which was then stored at 4 °C.

A bioprinter developed by our laboratory based on open-resource systems consists of robotic arms with three-axis motion, a stationary print platform, and a changeable print head. The print head on the X-axis robotic arm consists of piston-driven microscale depositional equipment. The bioprinter is controlled by software written by our laboratory.

Passage 3 chondrocytes were detached from culture flasks using 0.25 % trypsin (HyClone, China). Cell viability was assessed using the trypan blue exclusion test. The cells were counted with a hemocytometer and aliquoted according to different seeding densities. The collagen type II pre-gel was adjusted to pH 7.0 through a dropwise titration of 10 M sodium hydrate. The chondrocytes and collagen type II pre-gel were gently mixed in Eppendorf tubes at temperature below 20 °C.

Our earlier study found that cell distribution patterns could not be well maintained in a printed hydrogel construct as thin as articular cartilage tissue due to zonal fusion caused by the hydrogel’s low mechanical strength. In the present study, a computer-aided design (CAD) model of a cylinder with a 6 mm height and a 4 mm radius was created, as shown in Fig. 1(a). The constructs were printed layer by layer, from the bottom to the top, on a sterile platform on a super-clean bench at a temperature below 20 °C, as shown in Fig. 1(b). A 25-gauge needle (0.25 mm inner nozzle diameter) was used. The deposition speed was set to 20 mm/s, and the layer thickness was set at 0.4 mm. Every construct had 15 layers. Three groups were created according to the total cell seeding density in the collagen type II pre-gel: Group A, 2 × 10⁷ cells/mL; Group B, 1 × 10⁷ cells/mL; and Group C, 0.5 × 10⁷ cells/mL. Based on Hunziker’s study [14], the total cell density within the articular cartilage of the human medial femoral condyle is 1 × 10⁷ cells/mL. Group B was designed to have a biomimetic total cell density, while Group A was designed to have a higher total cell density, and Group C was designed to have a lower total cell density. Each group included two types of construct: one with a cell density gradient and the other with a homogeneous cell distribution. According to previous research [14], the ratio of chondrocyte densities in the superficial zone (top 10 %) to the transitional zone (middle 10 %) to the deep zone (remaining 80 %) of normal adult human articular cartilage is approximately 3:2:1. As depicted in Fig. 2, the first 11 layers (deep 74 %) of the printed construct were designed to be the deep zone, the middle 2 layers (middle 13 %) were designed to be the middle zone, and the top 2 layers (top 13 %) were designed to be the superficial zone. The construct with the homogeneous cell distribution had a single cell density, which was equal to the total cell density. For the condition of the construct with the cell density gradient, the ratio of three zonal cell densities was 3:2:1, and the ratio of the three zonal volumes was 13 %:13 %:74 %. As the total cell densities were known for each group, the zonal cell densities shown in Table 1 could be obtained by solving a three-variable linear equation.

The fabricated constructs were cross-linked in a humidified incubator at 37 °C for 30 min [20] and were then moved into a 24-well plate and cultured in high-glucose DMEM supplemented with 10 % FBS, 1 % P/S, and 2.5 μg/mL amphotericin B (Beyotime, China) at 1.3 mL per well at 37 °C and 5 % CO₂, as shown in Fig. 1(c). The culture medium was changed every other day. The constructs were harvested at 0, 1, 2 and 3 weeks, as shown in Fig. 1(d).

Cell viability was assessed using the Calcein AM (KeyGen Biotechnology Nanjing, China) and propidium iodide (PI, KeyGen Biotechnology Nanjing, China) molecular probes. Live cells were stained with Calcein AM (green), and dead cells were stained with PI (red). The constructs were sectioned and incubated in PBS containing 0.2 μM Calcein AM and 0.6 μM PI at room temperature for 1 h and were then rinsed in PBS. A full-thickness plane was imaged by fluorescent microscopy (Olympus IX71 microscope). The total cell number in each field of view could be quantified by separately counting green and red fluorescent dots. The percentage of live cells was calculated by dividing the number of live cells by the total cell number in 3 random images at 20× magnification.

Total RNA was extracted from the constructs using the Cartilage RNAout Kit (TIANDZ, China). The RNA concentrations and purity were measured using a NanoDrop 2000 (Thermo Scientific, USA). Reverse transcription was performed with the QuantiTect Reverse Transcription Kit (Qiagen, USA). Real-time PCR was performed with the QuantiFast SYBR Green Kit (Qiagen, USA). All procedures followed the manufacturers’ protocols. The expression levels of the collagen type I (COL1A1), collagen type II (COL2A1) and aggrecan (ACAN) genes were normalized to GAPDH, which was used as an endogenous control gene, according to the 2-△△CT method. The target gene primers were designed as follows: COL1A1 forward, 5-TTGGCAAAGAGGGCGGCAAAGGC-3, and reverse, 5-GAGCACCAGCAGGTCCGTCAGCA-3; COL2A1 forward, 5-ACGCTCAAGTCCCTCAACAACCAG-3, and reverse, 5-GTCTATCCAGTCACCGCTCTTCCA-3; ACAN forward, 5-ACTTCCGCTGGTCAGATGGA-3, and reverse, 5-TCTCGTGCCAGATCATCACC -3; and GAPDH forward, 5-GAAGGTGAAGGTCGGAGTC-3, and reverse, 5-GAAGATGGTGATGGATTTC-3.

After harvesting, the constructs were weighed and then digested for 24 h in 1 mL of 0.1 M PBS containing 2.5 U of papain (Beyotime, China), 5 mM cysteine-HCl and 5 mM Na2EDTA at 60 °C. The cell number was assessed based on the DNA content using Hoechst 33258 dye (KeyGen Biotechnology Nanjing, China), as previously described [21]. Hoechst 33258 is a fluorescent nucleic acid stain used to quantitate double-stranded DNA in solution. The fluorescence of the samples was converted to cell numbers using a cell-number standard curve created previously. The glycosaminoglycan (GAG) content was determined using the 1, 9-dimethylmethylene blue (DMMB; Jianglaibio, China) dye-binding assay and a shark cartilage chondroitin sulfate reference sample (C4348; Sigma, USA). The GAG content was normalized to the cell number to assess the single-cell biosynthetic activity.

The constructs were fixed in 4 % paraformaldehyde, dehydrated in a sequential ethanol series and embedded in paraffin. Sections (5 μm) were cut, deparaffinized and stained with hematoxylin-eosin (H&E) for cell distribution assessment and with Alcian blue for sulfated GAG assessment. For immunohistochemistry, rabbit polyclonal antibody Anti-Collagen I (bs-0578R; Bioss, China), Anti-Collagen II (bs-11929R; Bioss, China) and Anti-PRG4 (bs-1175R; Bioss, China) primary antibodies were used to detect collagen type I, collagen type II, and PRG4, respectively, synthesized by the rabbit chondrocytes. The primary antibodies were recognized using the Biotin-Streptavidin HRP Detection System (SP-9000; ZSGB-BIO, China) and a DAB kit (ZLI-9017; ZSGB-BIO, China). All procedures followed the manufacturers’ protocols.

The constructs were successfully fabricated and had uniform geometrical dimensions, with an average volume of 0.3 ± 0.05 mL. The physiochemical properties of the 10 % (wt/vol) collagen type II pre-gel adequately matched the requirements for hydrogel biofabrication [22]. As shown in Fig. 3, the results of the H&E staining showed that both the homogeneous and the gradient cell distribution patterns were effectively established and were maintained throughout the whole culture period.

The trypan blue exclusion test showed that 98 ± 1 % of the chondrocytes that had detached from the culture flasks were alive. To assess the damaging effect of printing on cell viability, cell viability tests were performed on the first day after fabrication, as shown in Fig. 4(a). The average live cell percentage was 93 ± 3 %. No significant difference was observed between the two types of construct or among the different groups. As shown in Fig. 4(b), a slight decrease in the total cell number was observed in all groups during culture, but the difference was not statistically significant.

To evaluate the phenotypic alterations of the chondrocytes in the collagen type II hydrogel constructs, the relative expression levels of COL1A1, COL2A1 and ACAN compared to GAPDH were determined by real-time PCR, as shown in Fig. 5. During the 3-week in vitro culture, COL1A1 expression remained low, and it was down-regulated throughout the culture period. However, the expression levels of both COL2A1 and ACAN were significantly increased (p < 0.05).

The total GAG contents in the constructs are shown in Fig. 6(a). During the first week of culture, the GAG content was positively correlated with the total cell density. Group A had the highest GAG content, and Group C had the lowest. During the second and third week, only Group C had a lower GAG content compared to Group A or B (p < 0.05). The differences between Group A and B were not significant.

To assess the average single-cell GAG production, the total GAG content was normalized to the cell number, as shown in Fig. 6(b). In the first week, no significant differences were observed among the three groups. In the last 2 weeks, the average single-cell GAG production in Group A was the lowest among the three groups. Because there was no significant decrease in the cell number, several of the cells in Group A might have entered the static state due to the limited nutrient supply. Group B had the highest average single-cell GAG production. Interestingly, there were significant differences between the construct with the cell density gradient in Group B and the constructs in Group A and C, although there was no significant difference within Group B. This result might be due to a synergistic effect of the total cell density and the cell distribution pattern. At 3 weeks, a significant difference was observed in the constructs with the homogeneous cell distribution between Group B and Group C, which might have been due to an effect of the total cell density.

To assess the ECM distribution in the constructs, Alcian blue staining for GAG and immunohistochemical analysis of collagen types I and II and PRG4 were performed. As shown in Fig. 7, the positively stained cells were brown, the negatively stained cells exhibited blue nuclei, and GAG in the matrix was also stained blue. Nearly all of the chondrocytes in the constructs stained positively for collagen type II, many stained positively for PRG4, and a few stained positively for collagen type I. In the constructs with the cell density gradient, the collagen type II, PRG4 and GAG contents were concentrated in the superficial zone and decreased with depth.