The present study was approved by the ethics committee of the Sun Yat-Sen Memorial Hospital of Sun Yat-Sen University, Guangzhou, China. In addition, informed consent was obtained from all patients and all healthy donors (HDs). Fifty-one ASp (eight women and 43 men) with an average age of 29.4 years (17 to 45 years) and 49 HDs (eight women and 41 men) with an average age of 27.3 years (18 to 39 years) were included in the study. All of the AS patients were diagnosed according to the New York modified criteria 34 and were HLA-B27-positive; conversely, 37 healthy donors were HLA-B27-negative (HD1) and 12 healthy donors were HLA-B27-positive (HD2). Sixteen patients were diagnosed for the first time, and the research samples from all ASp were taken at the active stage (all Bath Ankylosing Spondylitis Disease Activity Index ≥4) and without taking any medicine for at least 2 weeks.

After being informed regarding the scientific contributions, possible risks and complications and the corresponding prevention and treating measures for bone marrow aspirations, all of the healthy controls and ASp expressed approval and signed the informed consent. The bone marrow aspirations were all performed by skilled allied health professionals strictly according to the international standardized procedure for bone marrow aspirations. The bone marrow samples of AS patients and HDs were diluted with DMEM (low-glucose DMEM) containing 10% FBS. The mononuclear cells were prepared by gradient centrifugation at 900 × g for 30 minutes on Percoll (Pharmacia Biotech, Uppsala, Sweden) of density 1.073 g/ml. The cells were washed, counted, seeded at 2 × 10⁶ cells/cm² in 25-cm² flasks containing low-glucose DMEM supplemented with 10% FBS and cultured at 37°C, 5% carbon dioxide. Medium was replaced and the cells in suspension were removed at 48 hours and every 3 or 4 days thereafter. BMSCs were recovered using 0.25% Trypsin-ethylenediamine tetraacetic acid and replated at a density of 5 × 10³ to 6 × 10³ cells/cm² surface area as passage 1 cells when the culture reached 90% confluency. BMSCs after the third subculture were used for described experiments. PBMCs were obtained by the Ficoll-Hypaque (Pharmacia Biotech, Uppsala, Sweden) gradient separation of the buffy coat of ASp and HDs.

BMSCs were seeded in 96-well plates at a concentration of 1 × 10⁴/ml, in a final volume of 100 μl fresh medium (10% FBS + low-glucose DMEM), and three wells of each sample were digested using 0.25% Trypsin-ethylenediamine tetraacetic acid for cell counting per day up to 12 days. The BMSC growth curves were made using the data for cell proliferation obtained above. Using MTT (5 mg/ml; Sigma-Aldrich Co., St. Louis, MO, USA), dimethyl sulphoxide (Sigma) and an EL800 microplate reader (BioTek Instruments, Winooski, VT, USA) that was to determine absorbance at 490 nm, the cell viability curves for BMSCs were acquired in the same way according to the day and the absorbance. The BMSC proliferation ability was also examined by ³H-TdR assay. Fresh medium was used as a negative control.

To induce osteogenic differentiation, BMSCs were initially seeded in six-well plates at a concentration of 10⁴/cm². After preculturing for 24 hours, the BMSCs were allowed to grow in osteogenic medium (high-glucose DMEM supplemented with 10% FBS, 50 mg/l ascorbic acid, 10 mM β-glycerolphosphate and 10 nM dexamethasone; all these inducing reagents from Sigma). The BMSCs were then incubated in 5% carbon dioxide at 37°C, according to the experimental requirements for up to 14 days, and the medium was replaced every 3 days before harvest. The alkaline phosphatase (ALP) and mineralization of BMSCs were assayed using Cell Alkaline Phosphatase Staining assay (Sigma) and Alizarin Red staining (AR-S, 40 mmol/l, pH 4.2; Sigma) on the 14th day, respectively.

To induce adipogenic differentiation, the BMSCs were seeded in six-well plates at a concentration of 10⁴/cm². After preculturing for 24 hours, the BMSCs were shifted to adipogenic medium (low-glucose DMEM supplemented with 10% FBS, 1 μM dexamethasone, 10 μg/ml insulin, 0.5 mM 3-isobutyl-1-methylxanthine and 0.2 mM indomethacin; all these inducing reagents from Sigma). The BMSCs were then incubated in 5% carbon dioxide at 37°C, and the medium was replaced every 3 days before harvest. The intracellular lipid accumulation as an indicator was visualized on the 14th day by Oil Red O staining after fixed with 4% cold paraformaldehyde in PBS (pH 7.4) and washed with distilled water.

To induce chondrogenic differentiation, aliquots of 2.5 × 10⁵ BMSCs were centrifuged at 1,000 rpm for 5 minutes in 15-ml polypropylene conical tubes to form pellets, which were then cultured in high-glucose DMEM supplemented with 1% ITS-Premix (Becton-Dickinson, Mountain View, CA, USA), 50 mg/ml ascorbic acid (Sigma), 10^-3 M sodium pyruvate (Sigma), 10^-7 M dexametazone (Sigma), and 10 ng/ml transforming growth factor-β3 (R&D Systems, Minneapolis, MN, USA) for 28 days. The pellets were then fixed with 4% paraformaldehyde, embedded in paraffin, and subjected to Alcian blue staining to confirm chondrogenic differentiation.

The BMSCs in fresh medium (high-glucose DMEM supplemented with 10% FBS) without these differentiation-inducing factors were used as the experimental control, and fresh medium without any cells was used as a negative control. All measurements were performed in triplicate. The images were visualized using an inverted phase contrast microscope (Nikon Eclipse Ti-S, Nikon Corporation, Tokyo Prefecture, Japan).

On the 14th day the osteogenic medium was removed, and then 1.0 ml Triton X-100 (Sigma) was added to each well. A cell scraper was used to remove the BMSCs from the well bottom, and then the 1.0 ml cell lysate were placed in a 1.5-ml centrifuge tube. The samples were then processed through two freeze-thaw cycles (-70°C and room temperature, 45 minutes each) to rupture the cell membrane and extract the proteins and DNA from the cells. A p-nitrophenyl phosphate liquid substrate system (Stanbio, Boerne, TX, USA) was used to analyze the ALP concentration from the cells of each group. Then 10 μl each cell lysate solution was added to 190 μl p-nitrophenyl phosphate substrate and incubated in the dark at room temperature for 1 minute. The absorbance was read using a plate reader (M5 SpectraMax; Molecular Devices, Sunnyvale, CA, USA) at 405 nm and normalized to the PicoGreen assay 35 . DNA was quantified using the Quant-iT PicoGreen Kit (Invitrogen, Carlsbad, CA, USA) following standard protocols. Briefly, 100 μl each cell lysate solution was added to 100 μl PicoGreen reagent and incubated in the dark at room temperature for 5 minutes. The absorbance was read at an excitation/emission of 480 to 520 nm on the plate reader.

The inhibitory effects of BMSCs on mixed PBMC reaction (MLR) and PBMC proliferation stimulated by phytohemagglutinin (PHA) (4 μg/ml; Roche, Mannheim, Germany) were measured using the MTT assay 36 and the ³H-TdR assay 10 as described previously. Briefly, BMSCs were seeded in V-bottomed, 96-well culture plates for 4 hours for adherence, and then irradiated (30 Gy) with Co⁶⁰ before being cultured with the mixed PBMCs or the PBMCs stimulated by PHA.

BMSCs were trypsinized and then irradiated (30 Gy) with Co⁶⁰ before being co-cultured with PBMCs from a healthy volunteer in the presence of PHA (4 μg/ml; Roche) in 24-well plates (Nunclon, Roskilde, Denmark) and plated at a ratio of 1:10 in a total volume of 2 ml/well in triplicate for 72 hours. The cell density was 5 × 10⁴/cm² BMSCs and 5 × 10⁵/cm² PBMCs in a mix. Phorbol myristate acetate (50 ng/ml; Sigma, St Louis, MO, USA) and calcium ionomycin (1 μg/ml; Sigma) were added 6 hours prior to the end of the 72-hour co-culture. All of the PBMCs were then collected to be assayed by flow cytometry for the CCR4⁺CCR6⁺ Th and Treg cells. PBMCs were also grown alone in BMSC-free medium and used as control.

To detect the surface markers 37 of BMSCs and the frequency of CCR4⁺CCR6⁺ Th and Treg cells in PBMCs, the antibodies (Table 2) - including CD105(FITC), CD73(FITC), CD90(FITC), CD34(FITC), CD45(FITC), CD14(PE) and HLA-DR(FITC) for BMSCs; CCR4(PE-Cy7), CD196(CCR6)(PE) and CD4(FITC) for CCR4⁺CCR6⁺ Th cells 29 ; and CD4(FITC), CD25(APC) and Fox-P3(PE) antibodies for Treg cells - were used according to the manufacturers' recommendations. BMSCs and PBMCs marked with appropriate antibodies were measured with a FACScan laser flow cytometry system (Becton Dickinson) immediately. In each experiment, control staining with the appropriate isotype monoclonal antibodies was included. Results were expressed as the mean (frequency, %) ± SD.

Data are expressed as the mean ± SD, and the significance of the results was determined using the unpaired Student's t test. The product-moment correlation coefficient was used to test the correlations between the suppression ratios of BMSCs and the ratio of CCR4⁺CCR6⁺ Th cells to Treg cells in peripheral blood. Statistical analysis was performed using the SPSS computer program (SPSS Inc., Chicago, IL, USA). P < 0.05 was considered statistically significant.

To evaluate the biological properties of AS-BMSCs, compared with those of HD-BMSCs, the studies for growth characteristics, cell viability and multiple differentiation potentials in vitro were performed. The AS-BMSC growth curves have the same tendency as those for HD-BMSCs. The BMSC proliferation data of these two groups at each day (1 to 12 days) were tested by unpaired Student's t test, and the statistical result indicates that there was no statistically significant difference in BMSC growth characteristics between ASp and HDs (HD1 and HD2) (P > 0.05, Student's t test for independent samples). Established cultures (12 days) of BMSCs exhibited close, even equivalent, cell viability at each point of time from 1 to 12 days, as determined by cellular viability assays, and the difference of OD at 490 nm between ASp and HDs (HD1 and HD2) at each day (1 to 12 days) was not also statistically significant (P > 0.05, Student's t test for independent samples). The cultures have similar purities: (Q_L [the lower point of interquartile range], Q_U [the upper point of interquartile range]) = (95%, 99%) for AS-BMSCs, (Q_L, Q_U) = (96%, 98%) for HD1-BMSCs and (Q_L, Q_U) = (96%, 99%) for HD2-BMSCs.

To explore whether the multiple differentiation potentials of BMSCs in AS were abnormal, we investigate the osteogenic, adipogenic and chondrogenic differentiation potentials of AS-BMSCs and HD-BMSCs in the present study. Obvious differentiated osteocytes and adipocytes were detected as early as the 7th day after being induced for osteogenic and adipogenic differentiation, and obvious differentiated chondrocytes were seen at about 14 days since induction (Figure 1A to 1C).

There appeared to be two stages in the BMSC differentiation process for both ASp and HDs. In the early stage, only a few osteocytes, adipocytes and chondrocytes were found within the undifferentiated BMSCs. Gradually, these three kinds of cells increased; simultaneously, the cell's body got bigger and cytoplasm became more abundant because, for example, osteocytes made closer contact, fat vacuoles of adipocytes multiplied and grew bigger, and chondrocytes began to gain many collagen fibers. In the later stage, these three kinds of cells increased rapidly and nearly predominated. For the adipocytes, osteocytes and chondrocytes derived from BMSCs, the purities were (Q_L, Q_U) = (90%, 97%), (Q_L, Q_U) = (91%, 96%) and (Q_L, Q_U) = (88%, 95%) for ASp, (Q_L, Q_U) = (88%, 98%), (Q_L, Q_U) = (90%, 97%) and (Q_L, Q_U) = (90%, 96%) for HD1, and (Q_L, Q_U) = (86%, 95%), (Q_L, Q_U) = (89%, 98%) and (Q_L, Q_U) = (92%, 97%) for HD2, respectively.

The calcium nodules were stained to present a red color (Figure 1A1 to 1D1), after Alizarin Red staining for calcium deposits of osteocytes was performed to determine the mineralization of BMSCs. For the adipogenic differentiation, the mass fat vacuoles of adipocytes were also stained to present a red color by Oil Red O staining (Figure 1A3 to 1C3). The well-differentiated chondrocytes were Alcian Blue-positive, and presented a bright blue color after staining (Figure 1A4 to 1C4).

The ALP activity, normalized to DNA concentration, is plotted in Figure 1E. The ALP activity (mean ± SD) was 644 ± 45 (mM p-nitrophenyl phosphate/minute per mg DNA) for AS-BMSCs (n = 51), which is lower than the 655 ± 49 for HD1-BMSCs (n = 37) (P > 0.05) and the 646 ± 51 for HD2-BMSCs (n = 12) (P > 0.05). All three values were much higher than those of the baseline ALP for BMSCs of ASp, HD1 and HD2 (85 ± 40, 88 ± 48 and 82 ± 13, respectively) (P < 0.001) in control medium without the osteogenic factors. ALP staining was performed on the 14th day to investigate the maturity degree of osteocytes in the groups of Asp, HD1 and HD2 (Figure 1A2 to 1C2).

The AS-BMSCs and HD-BMSCs were then examined for typical MSC phenotypic surface markers. Flow cytometric analysis showed that the AS-BMSCs and HD-BMSCs (HD1-BMSCs and HD2-BMSCs) have the same phenotypic surface markers, just as the typical MSCs did. The samples all express high levels of the surface markers CD105, CD73 and CD90, and lack expression of CD45, CD34, CD14 and HLA-DR surface molecules (Figure 2).

Under the condition that the proliferation characteristics, cell viability, multiple-differentiation potentials and surface markers of AS-BMSCs were normal, compared with HD-BMSCs, the immunomodulation potential of AS-BMSCs was evaluated in the present study. The effects of BMSCs from ASp (n = 51), HD1 (n = 37) and HD2 (n = 12) on two-way MLR or PBMC proliferation in the presence of PHA were evaluated by mixing BMSCs and mixed PBMCs for two-way MLR, or by PBMCs from a third healthy volunteer in the presence of PHA for PBMC proliferation assay at five BMSC:PBMC ratios of 1:20, 1:10, 1:5, 1:2 and 1:1, respectively.

Recent studies have independently revealed enhanced Th17 response and weakened Treg response in some autoimmune diseases 38 39 , so we also examined the frequencies of CCR4⁺CCR6⁺ Th and Treg cells in PBMCs of ASp and HDs (Figure 4). The PBMCs from ASp and HDs were examined for the subset populations using flow cytometry, defined as the percentages of CCR4⁺CCR6⁺ Th cells (CCR4/CCR6 double-positive) 32 and Treg cells (CD4/CD25/Fox-P3 triple-positive) accounting for the total CD4-positive Th cells (CCR4⁺CCR6⁺ Th/Th, Treg/Th), CD3-positive T cells (CCR4⁺CCR6⁺ Th/T, Treg/T), lymphocytes (CCR4⁺CCR6⁺ Th/L, Treg/L), and peripheral blood mononuclear cells (CCR4⁺CCR6⁺ Th/PBMCs, Treg/PBMCs) respectively. The proportions of Fox-P3-positive occupying CD4/CD25 double-positive cells (Fox-P3⁺/CD4⁺CD25⁺) and PBMCs (Fox-P3⁺/PBMCs) were also tested. Compared with healthy donors (HD1 and HD2), the CCR4⁺CCR6⁺ Th population of ASp was significantly increased (P < 0.001, Table 3 and Figure 5), whereas Treg cells and Fox-P3-positive cells were found to be significantly decreased (P < 0.001, Student's t test for independent data, Table 3 and Figure 5). There were no significant differences between HD1 and HD2.

We performed the direct contact co-culture of BMSCs and PBMCs to explore whether the reduced immunomodulation potential of AS-BMSCs altered the balance of CCR4⁺CCR6⁺ Th/Treg cells. The PBMCs were collected to be assayed by flow cytometry for the CCR4⁺CCR6⁺ Th and Treg cells after co-culture with BMSCs of ASp and HDs (HD1 and HD2) for 3 days. The percentages of Treg cells (0.63 ± 0.23%) and Fox-P3-positive cells (0.74 ± 0.11%) in PBMCs after co-culture with AS-BMSCs for 3 days reduced significantly, whereas the percentages of CCR4⁺CCR6⁺ Th cells (1.87 ± 0.29%) in PBMCs after co-culture with AS-BMSCs for 3 days increased significantly, compared with these values of groups, including 3-day HD1, 3-day HD2, 3-day control, and 0 days (P < 0.001, Figure 6). Impressively, the ratio of CCR4⁺CCR6⁺ Th cells to Treg cells (CCR4⁺CCR6⁺ Th/Treg) in PBMCs after co-culture with AS-BMSCs for 3 days increased significantly (P < 0.001, Figure 6).

When examining data from all subjects tested, we observed positive correlations between the percentage inhibition of BMSCs (MLR) and the percentage inhibition of BMSCs (PHA) for all of the 51 ASp, 37 HD1 and 12 HD2. Interestingly, however, for all of the ASp (Figure 7A), HD1 (Figure 7B) and HD2 (Figure 7C) there were significantly negative correlations between the ratios of CCR4⁺CCR6⁺ Th cells to Treg cells in the peripheral blood and either percentage inhibition of BMSCs (MLR) or percentage inhibition of BMSCs (PHA) at all five ratios (P < 0.01, respectively).