Patients, sample collection, and ethics approval

A group of 11 patients (6 female, 5 male) who fulfilled the Assessment of SpondyloArthritis International Society (ASAS) criteria for AS [17] were included in the study. Demographic and clinical characteristics of the patients are presented in Table I.

This study meets all criteria contained in the Declaration of Helsinki and was approved by the Ethics Committee of the National Institute of Geriatrics, Rheumatology, and Rehabilitation, Warsaw, Poland (approval protocol no: KBT-8/4/20016). All patients gave their written informed consent prior to enrolment.

Adipose tissue-derived mesenchymal stem cell isolation and culture

Specimens of subcutaneous abdominal fat were taken from the patients by 18 G needle biopsy. Tissue processing, ASC isolation and culture were performed as described previously [18]. Five human adipose-derived mesenchymal cell lines (Lonza Group, Lonza Walkersville Inc., MD, USA) were used as a control.

All experiments were performed using ASCs at 3–5 passages. Adipose tissue-derived mesenchymal stem cells were cultured in complete culture medium composed of DMEM/F12 (PAN Biotech United Kingdom Ltd., Wimborne, UK), 10% fetal calf serum (FCS) (Biochrom, Berlin, Germany), 200 U/ml penicillin, 200 µg/ml streptomycin (Polfa Tarchomin S.A., Warsaw, Poland) and 5 µg/ml Plasmocin (InvivoGen, San Diego, CA, USA).

Both untreated and cytokine pre-stimulated ASCs were applied. For pre-stimulation, ASCs were cultured for 24 hours with human recombinant tumour necrosis factor (TNF) and interferon gamma (IFN-γ) (both from R&D Systems, Minneapolis, MN, USA; each applied at 10 ng/ml).

Co-culture of adipose tissue-derived mesenchymal stem cells with purified allogeneic CD4+ T-cells

Peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats obtained from healthy male honorary blood donors (< 60 years old), according to the routinely applied procedure using Ficoll-Paque (GE Healthcare, Uppsala, Sweden). The CD3+CD4+ cells were isolated from PBMCs using the EasySep Human CD4+ T-Cell Isolation Kit (Stemcell Technologies, Vancouver, Canada).

After isolation, CD4+ T-cells (1.2 × 10⁶/well/2 ml of medium) were seeded either directly (contacting cultures) or on 0.4 mm pore size Transwell filters (MD24 with carrier for inserts 0.4 MY, Thermo Fisher Scientific, Massachusetts, MA, USA) (non-contacting co-culture) into 24-well plates with adherent ASCs (5 × 10⁴/well/2 ml of medium), then T-cells were activated with Dynabeads Human T-Activator CD3/CD28 (Thermo Fisher Scientific, Massachusetts, MA, USA).

After 5 days of co-culture, CD4+ T-cells and culture supernatants were harvested for further analysis, i.e. flow cytometry to identify proliferating and non-proliferating cells, and the measurement of concentrations of kynurenines, PGE-2, IL-10, and IL-1Ra, respectively. The CD4+ T-cells cultured separately were used as a control.

Proliferation assay

For proliferation assay, purified CD3+CD4+ T lymphocytes were stained with carboxyfluorescein diacetate succinimidyl ester (CFSE) or cell trace violet (CTV) (Thermo Fisher Scientific, Massachusetts, MA, USA), then co-cultured with ASCs as described above.

Cells harvested from cultures were analysed by flow cytometry to identify proliferating and non-proliferating cells. To characterize the cellular proliferation response, the percentage of proliferating cells, proliferation index (PI – number of division per proliferating cell), and replication index (RI – fold expansion) were calculated as described elsewhere [19].

Measurement of soluble factor concentrations in culture supernatants

The concentration of PGE-2 was measured using the Parameter Kit (R&D Systems, Minneapolis, MN). Kynurenine concentration was measured spectrophotometrically as described previously [16].

To this aim, culture supernatants were mixed with 30% trichloroacetic acid at a 2 : 1 ratio and incubated for 30 min at 5°C, then centrifuged at 10 000 × γ for 5 min and finally diluted at a 1 : 1 ratio in Ehrlich’s reagent (100 mg of p-dimethyl benzaldehyde and 5 ml of glacial acetic acid; Sigma-Aldrich, St. Louis, MO, USA).

The optical density of the samples was measured at a wavelength of 490 nm. L-kynurenine (Sigma-Aldrich, St. Louis, MO, USA) diluted in culture medium was used to prepare the standard curve. The concentrations of cytokines were measured by specific enzyme-linked immunosorbent assays (ELISAs), i.e. IL-1Ra using ELISA DuoSet Kits from R&D Systems, and IL-10 using human IL-10 ELISA (cat. no. 88-7104-88) from Invitrogen (Vienna, Austria). All measurements were taken in duplicate.

Data analyses

Data were analysed using GraphPad Prism software version 7. The Shapiro-Wilk test was used as a normality test. The Wilcoxon test was applied to compare the secretion of soluble factors by resting and activated CD4+ T-cells.

One-way analysis of variance (ANOVA) with repeated measures and the post-hoc Tukey test were used to assess the effect of untreated and TNF/IFN-γ (TI)-treated ASCs (ASCs_TI) on target cells, as well as to compare contacting versus (vs) non-contacting co-cultures.

The Mann-Whitney test was applied to analyse differences between effects exerted by ASCs of healthy donors versus ASCs of ankylosing spondylitis patients. Parametric (Pearson’s linear) and non-parametric (Spearman’s rank) correlation tests were used to assess associations between analysed parameters. Probability values less than 0.05 were considered significant.

Patients

The patient cohort was heterogenous with respect to demographic and clinical data (Table I). All patients were HLA-B27 positive, 50% of them presented ocular symptoms (iritis), and none of them had peripheral arthritis. They were mostly treated with non-steroid anti-inflammatory drugs (NSAIDs), while application of non-biologic disease-modifying anti-rheumatic drugs (DMARDs) and glucocorticosteroids was less frequent.

Inhibition of T-cell proliferation by adipose tissue-derived mesenchymal stem cells

In control, separately cultured, purified CD4+ T-cells activated via the CD3/CD28 pathway the majority of cells proliferated (mean ±SEM = 63.9 ±3.4%).

In the presence of both untreated and TI-stimulated ASCs, the number of proliferating T-cells decreased significantly. The number of proliferating CD4+ T-cells in T-ASCs co-cultures was inhibited in the presence of HD/ASCs and HD/ASCs_TI by 50.7 ±5.9% and 63.8 ±5.1%, respectively, while AS/ASCs and AS/ASCs_TI reduced it by 42.2 ±6.8% and 49.2 ±7.6%, respectively (mean ±SEM; data not shown).

However, as shown in Figure 1A–C, there were no significant differences between inhibitory effects exerted by HD/ASCs and AS/ASCs on T-cell proliferation response, i.e. the percentage of proliferating cells, the proliferation, and replication indices.

Fig. 1

Inhibition of T-cell proliferation by adipose-derived stem cells of healthy donors and ankylosing spondylitis patients. T helper (CD3+CD4+) lymphocytes, isolated from peripheral blood of 13 (A–C) or 3 (C–F) healthy donors, were stimulated with anti-CD3/anti-CD28 antibodies and cultured alone (T) or co-cultured for 5 days with either untreated or TNF+IFN-γ (TI)-stimulated ASCs from 5 HD (HD/ASCs) or 10 AS patients (AS/ASCs). The cell co-cultures were performed in conditions allowing direct cell contact (A–C) and in a transwell system (D–F, as indicated). Proliferation of CD4+ T-cells was analysed by flow cytometry. Data are the results of the indicated number of experiments (n), performed using random combination of ASCs and T-cells. Upper panels (A–C) – lines between points identify cultures containing the same combination of ASCs and T-cells. Lower panels (D–F) – results are expressed as the median (horizontal line) with interquartile range (IQR, box), lower and upper whiskers (data within 3/2 x IQR) and outliers (points) (Tukey’s box). Values: *p = 0.05–0.01; **p = 0.01–0.001, ***p = 0.001–0.0001 for intra-group comparisons. The inter-group (HD vs. AS and contact vs. transwell co-cultures) differences were statistically insignificant.

Moreover, untreated and TI-treated AS/ASCs reduced T-cell proliferation to a similar extent, while TI-treated HD/ASCs were more potent than untreated HD/ASCs in diminishing the number of proliferating cells (Fig. 1A), but the reduction of PI and RI values by these cells was comparable (Figs. 1B and 1C).

Contribution of cell-to-cell contact and soluble factors to anti-proliferative effect of adipose tissue-derived mesenchymal stem cells

Results of co-culture of untreated and TI-treated HD/ASCs with CD4+ T lymphocytes in the conditions allowing or preventing (transwell) direct cell-to-cell contact showed that inhibition of T-cell proliferation did not differ significantly between these culture systems (Figs. 1D–F), although in transwell cultures more scattered data were obtained.

These results point to secretory factors as critical mediators of ASC-triggered anti-proliferative effects. Therefore, we further investigated the production of several soluble factors in ASC/CD4+ T-cell co-cultures.

Production of soluble factors in co-cultures of CD4⁺ T-cells with adipose tissue-derived mesenchymal stem cells

Untreated and activated CD4+ T-cells produced small quantities of kynurenines (mean ±SEM = 0.004 ±0.001 vs. 0.05 ±0.02 mmol/ml, n = 7, p = 0.02 for untreated vs. anti-CD3/CD28-stimulated T-cells) and PGE-2 (0 vs. 815 ±531 pg/ml, n = 15, p = 0.09 for untreated vs. anti-CD3/CD28-stimulated T-cells) (data not shown). In the co-cultures of activated CD4+ T-cells with HD/ASCs and AS/ASCs significant increases of kynurenines and PGE-2 production were observed and untreated and TI-treated ASCs exerted similar enhancing effects (Figs. 2A and 2B). Importantly, both untreated and TI-treated AS/ASCs induced higher secretion of these factors, especially PGE-2, than HD/ASCs (Figs. 2A and 2B).

Fig. 2

Concentrations of soluble factors in the co-cultures of T-cells with adipose tissue-derived mesenchymal stem cells. Cells were prepared and co-cultured as described in Figure 1. CD4+ T-cells were isolated from peripheral blood of 10–17 healthy blood donors. Five HD/ASCs lines and AS/ASCs obtained from 6–8 patients were used. The concentrations of indicated soluble factors in culture supernatants were measured as described in the material and methods section. Data are the results of the indicated number of experiments (n), performed using random combination of ASCs and T-cells, and are shown as Tukey’s boxes. Values: ***p = 0.001–0.0001 for intra-group (T vs. T+ASCs) comparisons; #p = 0.05–0.01, ##p = 0.01–0.001 for inter-group (HD vs. AS) comparison.

In separate cultures, CD4+ T-cells activated via the CD3/CD28 pathway produced significantly more IL-10 and IL-1Ra than resting cells. The concentrations of these cytokines in the culture supernatants of resting versus activated T-cells (mean ±SEM, n = 7) were as follows: 0 vs. 1906 ±165 pg/ml, p = 0.02, for IL-10 and 70 ±33 vs. 216 ±33 pg/ml, p = 0.02, for IL-1Ra. In the co-cultures of activated CD4+ T-cells with untreated and TI-treated HD/ASCs and AS/ASCs no significant changes were observed in the concentrations of IL-10 (Fig. 2C), while a significant and comparable increase of IL-1Ra levels was found in these conditions (Fig. 2D).

Dependence of kynurenines, PGE-2, IL-10, and IL-1Ra production on cell-cell contact and soluble factors

In the cell-to-cell contact and transwell co-cultures of CD4+ T-cells with HD/ASCs the concentrations of PGE-2 were similar (Fig. 3B), whereas kynurenines levels were higher in transwell than cell contacting co-cultures, especially in those containing TI-treated HD/ASCs (Fig. 3A). By contrast, the concentrations of IL-10 (Fig. 3C), and to a lesser extent also IL-1Ra, (Fig. 3D) were significantly lower in the transwell system than in cell contacting co-cultures.

Fig. 3

Comparison of soluble factor production in contact and transwell co-cultures. Cells were prepared and co-cultured as described in Figure 1. Five HD/ASCs lines and CD4+ T-cells from peripheral blood of 5–7 healthy blood donors were used. The concentrations of indicated soluble factors in culture supernatants were measured as described in the material and methods section. Data are the results of the indicated number of experiments (n), performed using random combination of ASCs and T-cells, and are shown as Tukey’s boxes. Values: *p = 0.05–0.01, **p = 0.01–0.001, ***p = 0.001–0.0001 for intra-group (T vs. T+ASCs) comparison; #p = 0.05–0.01, ##p = 0.01–0.001, ###p = 0.001–0.0001 for comparison of contact vs. transwell cultures.

Thus, these results indicate that in the co-cultures of purified CD4+ T-cells with HD/ASCs up-regulation of kynurenines and PGE-2 release is dependent mostly on soluble factors, while cell-to-cell contact is important to raise IL-1Ra and to maintain IL-10 levels, respectively.

Association of soluble factor secretion with anti-proliferative effect of adipose tissue-derived mesenchymal stem cells

As shown in Table II, in the co-cultures containing untreated HD/ASCs or AS/ASCs the proliferative response of T-cells inversely correlated with the concentration of PGE-2 or kynurenines and IL-10, respectively. The association was more evident in the presence of TI-treated HD/ASC and AS/ASCs (Table II, Fig. 4).

Fig. 4

Correlation between concentrations of soluble factors and T-cell proliferative response. Cell co-culture, the evaluation of T-cell proliferation, and measurement of indicated soluble factor concentrations were performed as described in the materials and methods section. Pearson’s (R) or Spearman’s rank (Rs) correlation coefficient and p-values are shown. Other explanations as in Figure 1.

There was a moderate to strong negative correlation between the replication index on the one hand, and the concentrations of kynurenines, PGE-2, and IL-10, but not IL-1Ra, on the other hand (Fig. 4).

However, in the presence of HD/ASCs_TI, the association between RI values and kynurenines levels did not reach statistical significance (R = –0.652, p = 0.079), and PGE-2 concentration inversely correlated with the percentage of proliferating T-cells but not with RI (Fig. 4, Table II).

In the case of AS/ASCs_TI-containing co-cultures an inverse correlation was also found between the proliferation index and the levels of soluble factors, such as kynurenines, PGE-2, and IL-10 (Table II).