Polylactic acid (PLA)-polyglycolic acid (PGA) co-polymer (PLGA) scaffolds were obtained from Jinan Daigang Biomaterial Co., Ltd. (Jinan, China) and included low-temperature, high-pressure PLA (PLA/PGA ratio, 75:25; porosity, >90%; aperture, 200–250 µm; molecular weight, 300,000 kDa). Transforming growth factor (TGF)-β1 was obtained from Peprotech (Rocky Hill, NJ, USA). BM-MSCs with fluorescent markers (third generation, Odcell HUxMA late 1101080925A01) were purchased from Guangzhou Cyagen Biotechnology Co., Ltd. (Guangzhou, China). Streptavidin-peroxidase immunohistochemical staining kits (SP-0022) were obtained from Beijing Biosynthesis Biotechnology Co., Ltd (Beijing, China).

Once informed consent was obtained, a total of 3 voluntary donations of aborted human fetuses (~35 weeks of gestation) were used, which was approved by the Ethics Committee of the Affiliated Hospital of Qingdao University Medical College. NPs from fetal tissue were cut away, ground into a paste and then digested with 0.2% collagenase II (MP Biomedicals, LLC; Santa Ana, CA, USA). This produced a clear liquid (5 ml) that was centrifuged with 5 ml low-glucose Dulbecco's modified Eagle's medium (L-DMEM; GE Healthcare Life Sciences, Chalfont, UK) containing 10% fetal calf serum (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA). This mixture was placed in an incubator at 37°C containing 5% CO₂ in air and cells were grown to 90% confluence.

Samples were placed in a warm water bath to thaw and then transferred to a centrifuge tube. L-DMEM was added to the tubes and samples were mixed with gentle pipetting and then centrifuged 3 times. Cells were re-suspended in 4 ml cell culture medium and seeded in culture bottles.

PLGA scaffolds were cut into 4×2-mm blocks, rinsed with saline, dried and sterilized with ethylene oxide. Scaffolds were immersed in L-DMEM, followed by fixing in glutaraldehyde, dehydration with a graded ethanol series and drying. An ion sputtering apparatus was used to electroplate samples in gold for subsequent electron microscopy.

BM-MSCs and NPs were detached by digestion in 0.25% and mixed at a l:3 ratio. The resulting cell suspension (lx10⁵ cells/ml) was seeded onto the PLGA scaffold and cultured for 1 h. The scaffold was inverted and allowed to incubate for 5 h in L-DMEM containing 5 µg/l TGF-βl.

Male New Zealand white rabbits (n=20; age, 6 months; weight, 2.0–2.5 kg) were obtained from the Qingdao Institute for Drug Control (Qingdao, China). The rabbits were maintained separately at room temperature, exposed to a 16:8 h light: Dark cycle and were fed with rabbit feed (Qingdao Kangda Foodstuffs Co., Ltd., Qingdao, China). The rabbits had free access to water. The rabbits were anesthetized with Su Mian Xin II (0.15 ml/kg intramuscularly; Kangda Medical Products Co., Ltd., Shanghai, China) and atropine sulfate (0.05 mg/kg intramuscularly; Jiangsu Lianshui Pharmaceutical Co., Ltd., Lianshui, China). Animals were subjected to lumbar magnetic resonance imaging (MRI) and then randomly divided into 2 groups. Each animal received an anterolateral abdomen oblique incision parallel to the iliac crest through an extraperitoneal operative route to expose the L4-5 intervertebral space. Pre-vertebral annuli of control animals were destroyed and the NP was removed. Subsequently, a BM-MSC and NP suspension (1×105 cells/ml; ratio=1:3) was implanted into the intervertebral space (50 µl). Animals in the experimental group were treated equally, but PLGA scaffolds (4×2-mm blocks) with an adherent equivalent of BM-MSCs and NPs were implanted into the intervertebral space. After 12 weeks, another lumbar MRI was performed. Intervertebral signal intensity on MRI was evaluated using a 5-stage Thompson grading scale for describing the gross morphology of the intervertebral discs at 12 weeks after the second MRI was obtained (8). A Thompson grade of 1 indicated normal signal expression in intervertebral discs. For grade 2, the signal was mildly impaired and had a narrow signal range. A Thompson grade of 3 referred to weakened signals and grade-4 discs showed impaired signaling. Subsequent to MRI, animals were sacrificed. The L4-5 intervertebral discs were removed, frozen, sectioned (6 µm) and histochemically assessed for differences between control and experimental animals.

At 12 weeks after the operation, the expression of aggrecan in the intervertebral discs was detected by benzene-3-phenol spectrophotometric detection used to represent protein polysaccharide (9). For detection of collagen II, 10% paraformaldehyde-fixed intervertebral disc tissue were embedded in paraffin. Paraffin sections were baked in a 60°C oven (Fuzhou Maixin Biotechnology Co., Ltd., Fuzhou, China) for 2 h, deparaffinized with xylene, rehydrated in descending alcohol and then rinsed with sterile distilled water for 3–5 min. Antigen retrieval in tissue sections was performed by using a modified method of microwave antigen retrieval. Tissue sections were cooled to room temperature and washed by using phosphate buffer saline (PBS) three times for 5 min each. After PBS flushing, the fixed intervertebral disc tissues were incubated for 2 h at 37°C with anti-human collagen type II antibody (1:100 dilution; cat. no. ab3092; Abcam Ltd., Cambridge, MA, USA). Samples were then incubated for 1 h at room temperature with goat anti-mouse horseradish peroxidase-conjugated secondary antibodies (1:1,000 dilution; cat. no. CW0102S; CWBio, Beijing, China), and then rinsed with PBS. Subsequent to visualization with diaminobenzidine and counterstaining with hematoxylin, staining was observed under a microscope.

All values were expressed as the mean ± standard error of the mean. An independent-samples t-test and a Mann-Whitney rank sum test were applied to assess differences between groups. P<0.05 was considered to indicate a statistically significant difference. Statistical analyses were performed using SPSS software for Windows (version 19; International Business Machines, Corp., Armonk, NY, USA).

NPs attached to culture bottles at 24 h after inoculation and adhered well after 10 days. Under an inverted microscope, chondrocytes or cartilage-like cells were evident with large round or oval nuclei with 2–3 nucleoli, as well as spindle- or multi-angular fiber-like cells with round or ovoid nuclei and acytoplasm extending as pseudopodia (Fig. 1).

Approximately 95% of BM-MSCs were round or ovoid and were adherent at 24 h after recovery. BM-MSCs gradually became spindle or multi-angular shaped over time, and most of them had slim and extended pseudopodia that fluoresced under an inverted fluorescence microscope (Fig. 2). BM-MSCs and NPs adhered and grew well with a vigorous metabolism, and the cell cytoplasm was filled with numerous secretory granules. Each of the two cell types had a spindle- or multi-angular fibro-like phenotype in co-culture and was difficult to distinguish from the other based on morphology. Mixed BM-MSCs and NPs grew well and fluorescence was strong (Fig. 3).

Electron microscopy showed that BM-MSCs and NPs grew well with normal morphology when attached to PLGA scaffolds, which suggested that the PLGA scaffolds had an affinity for BM-MSCs and NPs (Fig. 4). A white jelly-like NP (3 mm diameter) was removed via an extraperitoneal approach and BM-MSCs and NPs in suspension were implanted into intervertebral discs of control animals and PLGA scaffolds with adherent BM-MSCs and NPs were implanted in intervertebral discs of the experimental animals. MRI was performed and disk morphology after surgery and 12 weeks' post-surgery was evaluated by Thompson grading: All of rabbits in control and experimental groups (n=20) were Thompson grade 1 and 2 immediately after surgery. After 12 weeks, 3 animals in the control group were Thompson grade 3 and 7 were grade 4. Furthermore, 2 animals in the experimental group were grade 2, 6 were grade 3 and 2 were grade 4. Pulposus signal intensity in each group gradually weakened, and the intervertebral space narrowed and degenerated over time. Intervertebral signal intensity in the experimental group was significantly higher than that in controls (P=0.015). Thus, the recovery of the intervertebral discs in the experimental animals at 12 weeks was superior than that in controls according to Thompson grading (Fig. 5).

Green fluorescence was observed in intervertebral disc slices of the two groups at 12 weeks after implantation, which indicated the presence of live cells within the intervertebral discs (Fig. 6). The proteoglycan content in the experimental group was significantly higher than that in the control group (3.93±0.31 vs. 3.52±0.26 mg/100 mg; P=0.014; Fig. 7). Type II collagen was measured immunohistochemically. Collagen II-expressing cell cytoplasts in the experimental and control groups stained brown or pale yellow (Fig. 8). The expression was scored according to standard methods. The collagen II content was significantly higher in the experimental group compared with the control group (P=0.012; Fig. 9).