A total of 15 goats (female, 24 months, 48.3 $\pm$ 3.6 kg), were used in this study. All goats were healthy and received X-ray and Magnetic Resonance Imaging (MRI) scans to assure the absence of IVDD-related diseases before the study. All goats were fasted for 24 h before surgery. Five goats were randomly selected, and the BMA was aspirated from the iliac crest for enrichment to determine the effect of the BONE GROWTH PROMOTER (BGP, FUWOSI, Chongqing, China), as a SCR technology using negative pressure to facilitate the permeation of the biomaterials by bone marrow under conditions of aspiration [25, 34, 35, 36]. The remaining 10 goats were used for IVD defect repair. BMA aspirated from randomly one side of iliac crest was cultured to obtain autologous BMSCs, and then the autologous BMSCs combined with gelatin sponge matrix were implanted into the IVD defect; the autologous BMA aspirated from the other side of the iliac crest was enriched by BGP and implanted into IVD after discectomy. The goats were euthanized 24 weeks after operation to compare the repair effect of the two methods (Fig. 1).

Fig. 1.

Schematic illustration of the study design of in vivo repair IVD defect after discectomy at goat model.

Five goats were anesthetized by intramuscular injection of atropine (0.01 mL/kg) and intravenous injection of 3% pentobarbital sodium (1 mL/kg). The BMA was aspirated from iliac crest of the goat using a 16 G bone marrow puncture needle, which was rinsed with 1000 U/mL heparin to prevent coagulation (Fig. 2A). Two puncture holes were made at one side iliac crest, and 10 mL BMA was aspirated from each hole [37]. For preparation of the enriched BMA-matrix grafts, two pieces of gelatin sponge (60 mm $\times$ 20 mm $\times$ 5 mm, XiangEn Medical Technology Development Co. Ltd, Nanchang, China) were chopped into small cubes (5 mm $\times$ 5 mm $\times$ 2.5 mm), and were loaded into the BGP (Fig. 2A). Then, a total of 20 mL BMA aspirated from one side iliac crest was injected into the BGP for enrichment (Supplementary Fig. 2A–D). Slowly press the handle of the BGP to the end to send the BMA into the conical cup filled with gelatin sponge, release the handle to automatically rise, and form a negative pressure at the bottom of the conical cup to draw the BMA back into the tube, and repeat the above operation for 5 times. Five cycles were processed until MSCs and effective constituent in BMA were fully filtered and seeded into the gelatin sponge cubes.

Fig. 2.

Surgical approach. (A) The BMA was aspirated from the ilium crest of goat. (B) Exposure of the IVD. (C) Construction of IVD defect model after discectomy (arrow). (D) Graft implantation. (E) Sutured annulus fibrosus. (F) All animals recovered from the procedure without complication and retained full lumbar spine function.

The enriched BMA-matrix was taken out and fixed in 4% paraformaldehyde for 12 h. After gradient ethanol dehydration and critical point drying, the sample and gelatin sponge was cut into 1 mm slices and pasted on the carrier table with conductive tape, respectively. The surfaces were sprayed with gold and observed by scanning electron microscope (SEM, Quanta 200, Fei, Hillsboro, OR, USA).

To assess the enrichment effect, the volume of BMA was measured before and after enrichment process, and the samples were taken for flow cytometry (FACS Aria III, BD Corporation, Franklin Lakes, NJ, USA) to detect the number of nucleated cells and the number of the target cells with negative expression of CD34 and CD45 and positive expression of CD90. The number and concentration of cells in the BMA before and after BMA enrichment using BGP were assessed by flow cytometry. The steps are as follows: (1) Take three samples of BMA before and after enrichment, isotype and corresponding isotype control, and put 300 $\mu$L of each sample into a tuber. (2) Add CD34, CD45, CD90 and CD34-FITC, CD45-PEcy${}^{\text{TM}}$7, CD90-PE (BD Corporation, Franklin Lakes, NJ, USA) to the corresponding tuber according to the antibody instructions and flow cytometry requirements. Incubate in a 37 ${}^{\circ }$C incubator protected from light for 20 min. (3) Add 1.5 mL erythrocyte lysate to each tube for 10 min in the dark. Centrifuge at 400 g for 10 min and discard the red supernatant. If there are still many red blood cells, repeat the above operation. (4) Wash twice with PBS (phosphate buffered saline) and resuspend with 500 $\mu$L PBS for flow cytometry detection. Then the concentration and adhesion multiple of the retention cells in the gelatin sponge matrix were calculated according to the calculation formula as follows: retention cells concentration = retention cells number/volume of retention fluid; cells adhesion multiple = the retention cells concentration/original BMA cells concentration [25].

Furthermore, 100 $\mu$L original BMA and a small amount of enriched matrix were cultured in a culture dish respectively with 10 mL of the DMEM/F12 culture medium (containing L-Glutamine and 10% FBS) for 11 days. In our study, there is all a penicillin-streptomycin solution in the culture medium, the concentration of penicillin is 100 U/mL, and the concentration of streptomycin is 0.1 mg/mL, and the culture conditions are all in an incubator with 20% oxygen and 5% carbon dioxide. Then, the cells adhered to the culture plate surface were identified by crystal violet stanning to assess the enrichment effect, followed by observation under microscope (Leica, DM3000, Wetzlar, Germany) [37].

For isolation of BMSCs, 20 mL of BMA was collected from the iliac crest and mixed with 2 mL heparin (1000 U/mL). Bone marrow mononuclear cells were isolated by density-gradient centrifugation using lymphoprep [38]. The mononuclear cells were cultured for 4 days in a 100 cm${}^2$ culture flask with DMEM/F12 culture medium containing 20% FBS. Adherent cells were considered to be BMSCs. When BMSCs were cultured to reach 80–90% confluence, they were passaged and seeded in DMEM/F12 with 10% FBS. BMSCs were identified by flow cytometry to analyze the stem cell surface markers of CD90, CD34, and CD45 and analyzed the multilineage capability of adipogenic, chondrogenic, and osteogenic differentiation.

To facilitate analysis of the tracking of transplanted BMSCs in IVD, the BMSCs were infected with lentivirus-GFP (catalog number: GV492, Shanghai Genechem Co., Ltd, China), a lentivirus vector expressing the green fluorescent protein (GFP) gene. BMSCs were cultured and washed with PBS, and infected with lentivirus-GFP at optimal multiple of infection (MOI = 1:10). After 3 days, GFP expression was observed using a fluorescence microscope (Leica, DMI4000, Wetzlar, Germany). 100 $\mu$L GFP-labeled autologous BMSCs of 1 $\times$ 10${}^6$ passage 3 was seeded into gelatin sponge cubes to obtain cultured BMSCs-matrix grafts, which is prepared for intraoperative implantation of disc defects of three goats.

The goats were general anesthetized by intramuscular injection of atropine (0.01 mL/kg) and xylazine hydrochloride (0.02 mL/kg). All animals receive perioperative intravenous antibiotic (ceftiofur sodium, 25 mg/kg). Once anaesthetized, the goat is placed on the operating table in the lateral position, and the operation area was shaved and disinfected routinely with iodophor. A longitudinal 10 cm incision parallel and 1 cm anterior to the transverse processes is made according to the desired discs level using a left lateral retroperitoneal approach [39]. The subcutaneous layers were dissected using monopolar cautery. With the thoracolumbar fascia divided, the retroperitoneal space can be entered. The psoas and quadratus lumborum muscles are retracted posterolaterally, by the assistant, using a retractor, and blunt separation of the muscles attached to the disc, further exposing the target discs (Fig. 2B). A small incision was made at the left-anterior surfaces of the L2/3, L3/4 and L4/5 discs with a sharp knife, and approximately 50 mg NP tissues was removed from a single disc with small NP forceps to construct IVD defect model (Fig. 2C). The L2/3 disc was subjected to discectomy only (DO group), the L3/4 disc was implanted with enriched BMA-matrix (CE group), and the disc L4/5 was implanted cultured BMSCs-matrix (CC group) (Fig. 2D). And the intact L1/2 disc served as a non-injured control (NC group). The AF incision was closed using an absorbable suture (Fig. 2E). Then hemostasis is achieved, and muscle, fascia and skin were separately sutured layer by layer. Strict sterile precautions are maintained at all times. The goat was awakened by benzoxazole hydrochloride injection (0.01 mL/kg IM). After operation, all goats received intramuscular injection of flunixin meglumine (0.2 mL/kg) for analgesia for three days. Dressing change regularly until healing of the incision site was complete. The animals were observed at least once daily for general health and appearance by the veterinary staff and carefully monitored for signs of pain, discomfort, gait or posture. The animals were followed up for 24 weeks after the operation.

Lumbar X-ray imaging of goats was performed under general anesthesia at pre-operation and 4, 24 weeks after operation. The change in IVD height was evaluated by the disc height index (DHI) [40, 41]. Disc height and the adjacent vertebral body heights were measured on the midline and 25% of the disc’s width from the midline on either side. The DHI was expressed as the mean of the 3 measurements from midline to the boundary of the central 50% of disc width divided by the mean of the 2 adjacent vertebral body heights. Changes in DHI were expressed as %DHI and were standardized to the preoperative DHI as follows: %DHI = ((postoperative DHI/preoperative DHI) $\times$ 100).

MR scans were obtained in T2-weighted images in all groups at pre-operation and 4, 24 weeks after operation in the sagittal and axial planes using a 1.5 T scanner (GE, Boston, MA, USA). T2-weighted sections in the sagittal plane were obtained under the following settings: fast spin echo sequence with time to repetition of 2500 seconds and time to echo of 85 seconds; 384 (h) $\times$ 224 (v) matrix; field of view of 260; and four excitations. The section thickness was 3 mm with a 0.3 mm gap. The grayscale of the NP and the cerebrospinal fluid on T2-weighted sagittal MRI was determined by Image J version 1.51j8 (National Institutes of Health, MD, USA). The disc relative gray index (RGI) was calculated by dividing the gray value of the NP by the gray value of the cerebrospinal fluid [31]. The Pfirrmann classification scheme was applied to grade IVD degeneration [42]. All image assessments were performed by three independent observers who were blinded to the samples, and the mean of the three evaluations was recorded.

Goats were sacrificed by intravenous injection of overdosed pentobarbital at 24 weeks after surgery, and the ten lumbar spines were harvested. Biomechanical testing was performed on the spinal motion segment with ligamentous attachments and muscles removed. The denture base resin embedding sample was vertically installed on the biomechanical machine (MTS 858 Bionix II, Maumee, IN, USA) for biomechanical testing. Every vertebra of L1-5 was marked by markers in the same position (the specific position of every vertebra), and the motion of the markers was recorded by 3D Dynamic Capture System Camera (NDI Corporation, Canada) from different direction (Fig. 3A). The fresh spinal specimens were kept wet during the process of testing. Preconditioning was repeated three times (30 seconds/time) with 1 N preloading. The multidirectional bending moments (flexion and extension ($\pm$3 N m, x axis), left and right bending ($\pm$3 N m, z axis), and left and right torsion ($\pm$3 N m, y axis) were applied to the head end of the sample, while the tail end of the sample remained fixed. The multidirectional range of motion (ROM) of individual motion segments of every IVD was measured three times respectively according to the motions of markers.

Fig. 3.

Biomechanical analysis. (A) Spine specimens were tested in biomechanical tester. (B) The ROM of flexion-extension and left-right bending in the CE and CC groups were larger than that of the DO group at 24 weeks after operation. However, there was no significant difference in ROM between the CE and CC groups. ROM of left-right rotation was not statistically different in the four groups. All the data represent mean $\pm$ SD, n = 10. ‘#’ refers to NC group compared to DO group. ‘*’ refers to CE and CC groups compared to DO group respectively; #, *, = p 0.05; ns, = p $>$ 0.05.

After the biomechanical testing, the L1/2, L2/3, L3/4, and L4/5 discs of nine goats selected randomly were cut transversally at the center of the NP for macroscopic evaluation. A previously published histological grading system was applied to assess disc degeneration [43, 44]. Two histologists who were blinded to the samples performed evaluation of these sections, with their intra-observer error being minimum. One half of every disc was used for histological studies and identify the tracking transplanted cells in the NP tissues, and the other half was stored in liquid nitrogen (–196 ℃) for realtime PCR analysis of gene expression.

The remaining discs of one goat were isolated consisting of the total IVD and 0.5 cm of adjacent vertebral body and fixed in 4% paraformaldehyde for 5 days. Then the samples were decalcified in 15% EDTA for 6 months and subsequently embedded in paraffin. To visualize intact disc, the coronal sections of entire disc were cut using microtome (Leitzl516, Leica, Wetzlar, Germany) and stained with hematoxylin and eosin (H&E). To visualize histologic changes in NP tissue, NP tissues were fixed in 4% paraformaldehyde for 24 h and processed for paraffin embedding and sectioning into transversal sections (6 $\mu$m). The transversal sections were stained with H&E, Safranin O and Sirius Red for evaluation.

Immunohistochemistry was performed for Collagen II (II-II6B3, DSHB, Iowa City, IA, USA) and Aggrecan (ABT1373, Millipore, Burlington, AL, USA) analysis. Rehydrated sections were serially incubated at room temperature in proteinase K (Sigma, Hesse, Germany) for 5 min, 3% hydrogen peroxide for 10 min, blocking reagent for 5 min (BD5001, Bioworld Technology, Nanjing, China), and incubated with Collagen II (4 $\mu$L/mL) and Aggrecan (1:250 dilution) primary antibodies. After incubation at 4 ºC overnight, sections were incubated with antibody amplifier (BD5001, Bioworld Technology, Nanjing, China) for 10 min and horseradish peroxidase polymer for 10 min. Positive signal was visualized by diamino-benzidine (DAB). Finally, the sections were counterstained with hematoxylin, dehydrated and sealed. Then, the stained sections were observed under stereo-microscope (Leica, DM3000, Wetzlar, Germany).

To identify the existence of implanted cells in the NP tissues, the NP tissues were embedded in optimal cutting temperature (OCT) compound (Fisher Scientific, Pittsburgh, PA, USA) and sectioned at 6 $\mu$m. The GFP-positive cells were observed under fluorescence microscope (Leica, DMI4000, Wetzlar, Germany).

Total RNA was extracted from the NP using the TRIspin [45]. After the cDNA had been obtained by reverse transcription using AMV reverse transcriptase (Takara, Beijing, China), relative gene expressions of Collagen I, Collagen II, Aggrecan and SOX-9 were determined by Real-time PCR (RT-PCR). GAPDH were applied as normalization control. The obtained cDNA was amplified by qPCR using SYBR Premix Ex Taq${}^{\text{TM}}$ (Takara, Beijing, China). A cycle threshold (Ct) value was obtained for each sample, and triplicate sample values were averaged. The 2${}^{-\mathrm{\Delta }\mathrm{\Delta }Ct}$ value was then used to calculate relative expression of each target gene [46]. The primers being used were listed in Table 1.

The statistical analyses were performed using SPSS version 24.0 software (IBM, Armonk, NY, USA). Data was assessed for normality using the Shapiro Wilks test. The radiograph measurements, biomechanical analyses, and RT-PCR analyses data were all fitted with normal distribution, and statistical analysis was performed using analysis of variance (ANOVA) and the Fisher’s least significant difference post-hoc test. The MRI Pfirrmann grading and histological grading data did not conform to a normal distribution, and Kruskal Wallis H test was used for statistical analysis. Statistical significance was set at p 0.05.

BMA was aspirated from 10 iliac crests of 5 goats for enrichment. After the BMA was enriched by the BGP, the adhesion multiple of nucleated cells and target cells was 6.40 $\pm$ 0.93 and 4.20 $\pm$ 0.65 respectively, which were measured respectively by flow cytometry (Fig. 4J).

Fig. 4.

Bone marrow enrichment. (A) Gelatin sponge cubes (5 mm $\times$ 5 mm $\times$ 2.5 mm) in the BGP. (B) SEM images shows the porous structure and their interconnection of gelatin sponge. (C) Multiple round cells adhered to the inner wall of the enriched matrix. The red circles indicate the nucleated cells. (D) After the BMA culturing for 11 days, the fibroblastic colony formation could be observed. (G) After the enriched BMA-matrix culturing for 11 days, a lot of cells were visible in and around the gelatin sponge. (E,H) Many colony-forming units stained by crystal violet in the culture dish of the BMA and the enriched matrix respectively. (F,I) Microscopic observation showed that a number of fusiform or polygonal cells stained by crystal violet were adherent to the bottom of culture dish from the BMA and the enriched matrix. respectively ($\times$100). (J) The number of nucleated cells and target cells in the sample were detected by flow cytometry to calculate the adhesion multiple of the BGP. Calculation formula: adhesion multiple $\alpha$ = cells concentration of retentate/cells concentration of BMA. After the BMA enriched by the BGP, the adhesion times of the retained nucleated cells and the target cells were 6.40 $\pm$ 0.93 and 4.20 $\pm$ 0.65 respectively.

BMSCs differentiation assay demonstrated the autologous BMSCs possess the capability of osteogenic, chondrogenic and adipogenic differentiation. (Supplementary Fig. 1A–C) Meanwhile, flow cytometry analysis showed the CD90 was positively expressed, and CD34 and CD45 were negatively expressed on the surface of the target cells, which was consistent with the BMSCs surface markers, indicating that BMSCs was successfully isolated and identified (Supplementary Fig. 1D).

SEM images showed that the gelatin sponge was macroporous (Fig. 4B), and there were multiple round cells adhered to the porous wall of the enriched BMA matrix (Fig. 4C). After the BMA and the enriched BMA matrix were cultured for 11 days, crystal violet staining showed that there were a large number of colonies forming unit-fibroblast (CFU-F) at the bottom of the culture dish, and fusiform or polygonal cells were adherent to the wall. No bacterial growth was observed in all sample cultures (Fig. 4D,G). The results of flow cytometry detection and crystal violet staining showed that the MSCs in BMA could be effectively retained in gelatin sponge cubes by the BGP, and the cells viability were good.

The mean weight of the NP removed from the DO, CE and CC group was 50.18 $\pm$ 3.95, 51.12 $\pm$ 5.38 and 51.26 $\pm$ 4.34 mg (p $>$ 0.05), which indicated that the quantity of NP removed from each disc was consistent (Supplementary Fig. 2E). The average intraoperative blood loss was about 60 mL. The incisions of all goats were free of infection and healed smoothly. All goats successfully survived to the defined 24-weeks endpoint and showed no signs of neurological deficits or pain as determined by veterinary care staff.

Radiographic analysis at 4 weeks after operation showed that the disc spaces in DO, CE and CC groups narrowed significantly, and the DO group continued to decrease at 24 weeks, while rate of decrease in the CE and CC groups slowed down after 4 weeks. With a mean DHI for NC group at beginning of the study expressed as 100%, the mean DHI of the DO, CE and CC group was 89.3 $\pm$ 3.3%, 89.7 $\pm$ 2.7% and 90.2 $\pm$ 2.5% at 4 weeks after surgery respectively (p $>$ 0.05). The mean DHI of the CE and CC group was 82.0 $\pm$ 3.8% and 83.3 $\pm$ 3.2% at 24 weeks after surgery (p = 0.503), and all were higher than 68.3 $\pm$ 3.4% of the DO group (p = 0.004, 0.00) (Fig. 5A,B).

Fig. 5.

Radiographic and MRI assessment. (A,C) Representative radiographic and MRI images at 0 week before, 4, and 24 weeks after operation in goats. (B,D) The change of disc height index (DHI) and relative grey index (RGI) at 4 and 24 weeks after operation. All the data represent mean $\pm$ SD, n = 10. ‘&’ refers to NC group compared to DO group. ‘#’, ‘*’ refers to CE and CC groups compared to DO group respectively; ‘^’ refers to CE group compared to CC group. #, *, ^, = p $>$ 0.05; &, ##, **, = p 0.01; ^^, = p 0.05. (E,F) Pfirrmann grading of IVD degeneration in T2-weighted images. (E) Discs showed no degeneration in all groups before the operation. (F) There was significantly lower grading of MRI in the CE and CC groups compared with DO group at 24 weeks after operation (p 0.05).

According to the MRI T2-weighted results, the mean signal intensity of NC group had barely changed throughout the study (Fig. 5C,D). Signal intensity in the DO group decreased significantly shortly and continued to develop after the operation, while those of transplantation groups were significantly delayed. The change of MRI T2-weighted signal intensity was evaluated by RGI. For the DO, CE and CC groups the mean RGI were 0.56 $\pm$ 0.05, 0.58 $\pm$ 0.03 and 0.58 $\pm$ 0.04 at 4 weeks after surgery respectively (p $>$ 0.05). At 4 weeks, discs in CE and CC groups showed stronger signal intensities than that in DO group. At 24 weeks, the mean RGI of CE and CC groups were 0.41 $\pm$ 0.03 and 0.47 $\pm$ 0.04 respectively (p = 0.044), and these two groups showed higher RGI compared to DO group (0.26 $\pm$ 0.03, p = 0.006, 0.003).

The Pfirrmann classification analysis showed that no obvious disc degeneration in all groups before the operation (Fig. 5E). There was no significant difference between the Pfirrmann classification of CE group and CC group (p = 1.00), which was significantly lower than that of DO group (p = 0.021, 0.01), but all of them were significantly higher compared with the NC group at 24 weeks after operation (p 0.05) (Fig. 5F). Together, the analysis of imaging results shows that enriched BMA matrix and cultured BMSCs matrix transplantation can effectively maintain the IVD height and delay degeneration of the discs.

We further compared the biomechanical effect from the three operated discs (DO, CE, and CC groups) and intact disc (NC group) at 24 weeks after operation. As shown in Fig. 3, the mean ROM of flexion-extension and left-right bending in the DO, CE and CC groups were 13.3 $\pm$ 1.2, 16.2 $\pm$ 1.8, 16.8 $\pm$ 1.5 degree and 7.6 $\pm$ 0.9, 11.8 $\pm$ 1.7, 12.5 $\pm$ 1.3 degree respectively, which of them significantly decreased compared with those of the NC group (18.2 $\pm$ 1.8, 14.4 $\pm$ 1.4 degree) (p = 0.013, 0.035, 0.042 and p = 0.00, 0.040, 0.037). The ROM of flexion-extension and left-right bending in the CE or CC group showed significantly less reduction compared with the DO group (p = 0.038, 0.041 and p = 0.015, 0.010). No significant differences were seen between CE and CC groups. The ROM of left-right rotation was 4.0 $\pm$ 0.7, 3.7 $\pm$ 0.5, 3.7 $\pm$ 0.6 and 3.8 $\pm$ 0.6 degree in the NC, DO, CE and CC groups respectively, however, there was not statistically different among them (p $>$ 0.05). The result showed that the enriched BMA matrix and cultured BMSCs matrix transplanted into the discs maintained the stability of the degenerative IVD.

Gross observation of the lumbar spine showed that the fibrous connective tissue at the incision of AF in the CE and CC groups was rich, tough, and well repaired. However, the DO group had less connective tissue at the incision of the AF, and it was depressed and poorly repaired. The condition of NP and AF could be observed from the macroscopic view of freshly dissected IVD samples (Fig. 6A). In the DO group, in DO group, the NP and part of the AF tissue were black and yellow with severe degeneration. The CE and CC groups showed mild changes in NP tissue compared to the intact NC group. There are no NP protrusions in the discs of the CE and CC groups. Nine goats were assessed, showing that the histological grades of the discs in the CE and CC groups were mainly 2–3 grades, and that in DO group were mainly 4–5 grades. There was no significant difference in the grade of the discs between the CE and CC groups (p = 1.00), but both were lower than those of the DO group (p = 0.043, 0.024). Discs in NC group maintained grade 0 throughout the study (Fig. 6C).

Fig. 6.

Histological grading for assessment of IVD repair. (A) Representative macroscopic view of discs in four groups at 24 weeks. (B) The histological grading system for disc degeneration that focuses on morphological change in the inner annulus fibrosus structure was used in this study. (C) Histological grades were used to evaluate discs in the four groups at 24 weeks. There was no significant difference in histological grade between CE and CC groups (p = 1.0), but they were significantly lower than DO groups (p 0.05).

In this study, GFP-labeled BMSCs combined with gelatin sponge matrix were transplanted into the IVD in the CC group. The discs of three goats were used to detect the tracking of transplanted BMSCs by fluorescence microscopy. As shown in Supplementary Fig. 3D, the fluorescence microscopy confirmed the tracking of the GFP-labeled BMSCs in the discs of the CC group.

No GFP-positive cells were observed in the discs of the DO or CE group (Supplementary Fig. 3C). The results showed that the GFP-labeled BMSCs was successfully implanted into the IVD and grew in the CC group.

H&E, Safranin O and Sirius red staining revealed that compared with the DO group, the structure of NP tissue in CE and the CC group was more complete, the cell proliferation was more active, and the extracellular matrix was more abundant (Fig. 7A). HE staining of the mid-coronal sections showed that disc height was well maintained in NC, CE and CC groups, whereas disc narrowing was obvious in DO group (Fig. 7B). The immunohistochemical staining of NP tissue in transversal sections indicated that the staining intensity of NP in the CE and CC groups was strong positive in aggrecan and type II collagen, while the staining intensity in the DO group was weak or intermediate positive (Fig. 7C).

Fig. 7.

Histological and real-time PCR analysis. (A) H&E, Safranin O and Sirius red staining of NP in the transversal plane. (B) the intact IVD of mid-coronal sections were stained with HE staining. (C) The immunohistochemical staining for aggrecan and type II collagen. (D) Real-time PCR analysis. The results showed that aggrecan, type II collagen, and SOX-9 mRNA expression markedly increased in discs from the CE and CC group compared with the DO groups. Bar plots are shown as means $\pm$ SD, n = 9. ‘#’ refers to NC group compared to DO group. ‘*’ refers to the comparison among CE, CC and DO groups; #, *, = p 0.05.

For quantitative analysis the NP matrix, RT-PCR was used to measure the levels of aggrecan, type II collagen, type I collagen, and SOX-9 in the NC, DO, CE, and CC groups. As shown in Fig. 7D, the levels of aggrecan, type II collagen, and SOX-9 mRNAs were decreased in the DO, CE, and CC groups as compared with the NC group, while the type I collagen expression was significantly increased in the DO, CE, and CC groups (p 0.05). Importantly, we found that the levels of aggrecan, type II collagen, and SOX-9 mRNAs were significantly higher in the CE and CC groups than those in the DO group, and the type I collagen expression was significantly lower in the CE and CC groups than that in the DO group (p 0.05). Together, the results show that compared with the DO group, transplantation groups (CE group and CC group) increased the expression of aggrecan, type II collagen, and SOX-9, especially in CC group.