Human samples were obtained after informed consent as approved by the Spanish Ethics Committee of reference for the site of tissue procurement (Clínica de la Luz Hospital, Madrid, Spain). Human adipose tissue aspirates from healthy donors were washed twice with phosphate-buffered saline and digested with 0.075% collagenase (Type I, Invitrogen, Carlsbad, CA, USA). The digested sample was washed with 10% fetal bovine serum (FBS), treated with 160 mM NH₄Cl to eliminate remaining erythrocytes and suspended in culture medium (Dulbecco’s Modified Eagle Medium, with 10% FBS). Cells were seeded in tissue culture flasks and expanded (37°C, 5% CO₂) with change of culture medium every 3–4 days. Cells were transferred to a new flask when they reached 90% confluence. Cells were expanded up to duplication 12–14 and frozen. Experiments were performed with a pool of cells from three male and three female adult donors at population doublings 12–14. eASCs were thawed, seeded in tissue culture flasks, and trypsinized before administration to mice.

eASCs were defined according to the criteria of the International Society for Cellular Therapy (35), being positive for CD73 (AD2) and CD90 (5E10) from Becton Dickinson (Franklin Lakes, NJ, USA) and CD105 (43A3) from Biolegend (San Diego, CA, USA) and negative for CD14 (RM052) from Immunotech (Monrovia, CA, USA), CD19 (4G7), HLA-DR (L243), and CD34 (8G12) from Becton Dickinson and CD45 (J33) from Beckman Coulter (Brea, CA, USA) (Figure S1A in Supplementary Material).

For immunosuppression assays, peripheral blood mononuclear cells (PBMCs) were isolated by density centrifugation gradient using Ficoll-Paque Plus (GE Healthcare Biosciences AB, Uppsala, Sweden) from buffy coats provided by the National Transfusion Centre of the Comunidad Autonoma of Madrid, and splenocytes were obtained from C57/BL6 male mice.

For carboxyfluorescein diacetate N-succinimidyl ester (CFSE) labeling, PBMCs or splenocytes were washed extensively to remove FBS, resuspended in a 10 µM CFSE (Sigma-Aldrich, St Louis, MO, USA) solution (10⁷ PBMC or splenocytes per 200 µl of solution), and incubated under constant shaking at 37°C for 10 min. Reaction was stopped by adding ice-cold medium (RPMI + 10% FBS), and cells were washed three times with ice-cold phosphate buffer saline. Cells were then cultured overnight, and one aliquot was used to set up and control the FL-1 voltage for CFSE. After resting overnight, CFSE-labeled PBMCs were activated with the Pan T Cell Activation Kit (microbeads coated with anti-CD3, anti-CD2, and anti-CD28; Miltenyi Biotec, Auburn, CA, USA) following the manufacturer’s instructions. CFSE-labeled splenocytes were activated with anti-CD3 (Becton Dickinson) and IL-2 (Novartis, Basel, Switzerland). PBMCs or splenocytes (10⁶ cells/well) were cultured in 24-well plates alone or with eASCs (4 × 10⁴ cells/well; ratio 1:25 of eASC:PBMC or eASC:splenocytes) in a total volume of 2 ml of RPMI + 10% FBS. The 1:25 ratio was chosen because it provided a high inhibitory effect, on the basis of previous studies (36, 37). After 5 days for PBMCs and 3 days for splenocytes, cells were harvested, labeled with 7-AAD and anti-CD3 antibody and cell proliferation of the CD3⁺/7-AAD⁻ population (viable CD3 T lymphocytes) was determined by flow cytometry, according to loss of CFSE signal. Data were analyzed with the use of FCSExpress 4 (De Novo Software, Glendale, CA, USA) and BD CellQuest™ Pro analysis (Becton Dickinson) softwares. CaliBRITE beads (BD Bioscience, Erembodegem-Aalst, Belgium) were used to calibrate the acquisition events in the cytometer. As shown in Figure S1B in Supplementary Material, human eASCs strongly inhibited the proliferation of both human T lymphocytes and mouse splenocytes, demonstrating their immunomodulatory properties both in allogeneic and xenogeneic conditions.

DBA/1 male mice, 8-week-old, were obtained from Janvier SAS, France.

Each mouse was injected subcutaneously in the tail with a first dose of an emulsion of Chicken Collagen Type II (1 mg/ml final concentration, Col II, Chondrex, Redmond, WA, USA) in Complete Freund’s Adjuvant (Mycobacterium tuberculosis) in a volume of 0.1 ml/animal. Twenty-one days after the first immunization with collagen II, a second injection was administered to each animal, again subcutaneously in the tail. In this occasion, the collagen suspension was made using Incomplete Freund’s Adjuvant (no M. tuberculosis). Clinical signs of arthritis were evaluated daily after the second immunization to determine clinical evidence of arthritis of the limb joints by macroscopic examination. The Arthritis Index Score was conducted until the end of the study. The severity of the arthritis was scored in both, front and hind, paws according to the following arthritis index scoring system: 0, no signs of arthritis; 1, swelling and/or reddening of the paw or 1 digit; 2, two groups of joints inflamed, swelling and/or reddening; 3, more than two groups of swelling and/or reddening joints; and 4, inflammation of the whole paw (38).

Paw edema was assessed daily as the volume of both, hind and front, paws by the use of a plethysmometer.

In mice, the most suitable lymph nodes for injection are the inguinal ones due to their relatively easy access. Mice were anesthetized and, after depilation and disinfection of the inguinal region with 70% ethanol, a 6–8 mm incision was made. Injections were made with a Hamilton 30 gauge syringe. Three doses of 8 µl containing 5 × 10⁴ eASCs resuspended in Ringer’s Lactate (Grifols, Barcelona, Spain) were injected in both the right and left inguinal lymph nodes (total dose 1 × 10⁵/mouse) at day 1, day 8, and day 15. As a control, three doses of 8 µl of Ringer’s Lactate were injected in both the right and left inguinal lymph nodes at day 1, day 8, and day 15. Three doses of 200 µl of 2 × 10⁵ eASCs in Ringer’s Lactate were intravenously injected by the tail vein at day 1, day 8, and day 15. Mice were treated with eASCs when arthritis score was 2–4 according to Szabo et al. (38).

At day 22, mice were culled and cells were isolated from spleen and draining lymph nodes (inguinal and popliteal). Spleen and lymph nodes from each mouse were processed independently and disaggregated mechanically using a gentleMACS Dissociator AP^® (Miltenyi Biotec, Bergisch Gladbach, Germany). Cell counts were determined by Nucleocounter (Chemometec, Copenhagen, Denmark). Cells were surface stained with antibodies directed against with anti-CD4-FITC (GK1.5), anti-CD25-PE (REA568), and anti-CD3-PECy5 (REA641) (Miltenyi Biotec Bergisch Gladbach, Germany) and stained intracellularly with anti-Foxp3 (FKJ-16a, eBioscience, San Diego, CA, USA) using the Foxp3/Transcription factor staining buffer set (eBioscience) according to the manufacturer’s instructions. Appropriate isotype matched controls were included for the surface and intracellular staining.

For intracellular analysis of IL-10 cytokine expression, cells were stimulated in vitro with 5 ng/ml phorbol myristate acetate (Sigma-Aldrich, USA) and 500 ng/ml ionomycin (Sigma-Aldrich) for 4 h. GolgiPlug (BD Pharmingen) was added after 1 h. Cells were fixed and stained according to manufacturer’s instructions (Cytokine BD kit, BD Pharmingen). For intracellular staining, anti-IL-10 antibody (JES5-16E3, Becton Dickinson, BD Pharmingen) was used. Cells were acquired in a BD FACS CantoTM (Becton Dickinson). At least 10,000 events were acquired. Data were analyzed by BD FACSDiva™ (Becton Dickinson) and Flow Jo (FlowJo) softwares.

Plasma concentration of autoantibodies against chicken collagen II was determined by ELISA (MD Bioproducts, Zurich, Switzerland) following manufacturer’s instructions.

Plasma levels of IL-6, TNF-α, IL-17, and IL-23 were determined by Luminex© multiplex technology (Life Technologies, Madrid, Alcobendas) following manufacturer’s instructions (Affymetrix-eBioscience, Santa Clara, CA, USA).

For in vitro detection of IL-10, IL-4, TNF-α, IL-17, and IFN-γ, splenocytes and lymph node cells were seeded in 96-well U-plates in 10% FBS supplemented RPMI (5 × 10⁶ and 1 × 10⁶ cells/well, respectively) and stimulated with chicken collagen II (25 µg/ml), and then supernatants were collected after 72 h of culture and cytokines determined as indicated above.

At day 50 after treatment, mice were sacrificed and hind paws were removed and fixed in 10% neutral buffered formalin. Before micro-CT scanning, paws were washed with running water for 15 min. The imaging of three-dimensional micro-CT was performed with GE eXplore Locus micro-CT scanner (GE Healthcare, Spain). Data were acquired at 27 µm isotropic voxel size with 720 projections by 360° scan, integration time of 2,000 ms with three frames, photon energy of 80 KeV, and current of 450 µA. The duration of imaging time was 100 min per scan and followed by 1 h of projection correction and volume reconstruction of three-dimensional representation. Three-dimensional render images of hind paws were generated through original volumetric reconstructed images by MicroView software (GE Healthcare, Spain). Comparable regions of interest of four metatarsal joints per hind paw (right and left, total eight joints per mouse) were determined. Bone mineral density (BMD = BM/BV mg/cm³) and trabecular mineral density (TMD) were quantified from micro-CT scans using GE MicroView software v2.2. Data are expressed as the BMD and TMD variation in untreated and eASC-treated CIA mice compared to healthy animals.

Data are expressed as mean plus the SEM. The mean data were analyzed with unpaired U-test (Mann–Whitney test) in the comparison between two groups or one-way analysis of the variance (Kruskal–Wallis test) with Dunn’s multiple comparison post-test where more than two groups were compared. A value of p < 0.05 was regarded as statistically significant. Analysis was performed using Graphpad software (Graphpad, La Jolla, CA, USA).

We and others have previously demonstrated the immunomodulatory capacity of human eASCs administered intravenously in experimental CIA mice (10, 11, 39, 40). In the present study, we investigated the feasibility and efficacy of the eASC administration into the lymph nodes in CIA mice and compared it with the IV administration.

To first demonstrate the feasibility and efficacy of the IL administration of eASCs, CIA mice were injected at day 1, day 8, and day 15 in the right and left inguinal lymph nodes with either eASCs (5 × 10⁴ cells per lymph node; total dose per time point: 1 × 10⁵) or Ringer’s lactate (the vehicle of the eASCs), and the clinical score was monitored for 50 days after the treatment. In order to better reproduce the clinical situation, treatment was initiated when arthritis was already established (arthritis score 2–4) (day 1). As shown in Figure 1A, IL administration of eASCs significantly attenuated the severity of arthritis from day 8 to day 50. In contrast, the progression of the arthritis in CIA mice treated with vehicle was similar to untreated CIA mice. These results demonstrated that IL administration of eASCs, but not vehicle, had immunomodulatory effects and reduced the progression of arthritis in CIA mice.

We next investigated the efficacy of eASCs administered through the IL route compared to the IV route in CIA mice. In these studies, we used the lowest eASC dose per route of administration that, in our hands, efficiently modulated the progression of arthritis (minimal therapeutic dose per route, in our experimental settings). Thus, CIA mice were either untreated or treated at day 1, day 8, and day 15 with 2 × 10⁵ eASCs by the IV route or with 1 × 10⁵ eASCs by the IL route. As shown in Figure 1B, IV administration of eASCs significantly attenuated the severity of CIA from day 15 to day 29. Interestingly, the immunomodulatory effect of the IL administration showed a prolonged therapeutic effect that lasted for up to 50 days. The mean arthritis score for the whole study was 11.7 ± 0.3 for untreated, 7.8 ± 0.3 for IL eASC-treated, and 9.9 ± 0.3 for IV eASC-treated CIA mice, showing a higher efficacy of IL route compared to IV administration (Figure 1C).

Taken together, these results demonstrate that the IL administration of eASCs is feasible and could modulate the progression of arthritis in CIA mice potentially more efficiently than the IV administration.

In order to further confirm the therapeutic effect of the IL treatment with eASCs, joint damage on the hind paws of untreated and eASC-treated CIA mice was measured by means of micro-CT at day 50. As shown in Figure 2, the loss of TMD and BMD was significantly lower in eASC-treated CIA mice than in untreated CIA mice. The micro-CT images indicated that the administration of eASCs into the lymph nodes in CIA mice significantly reduced clinical symptoms, joint pathology, and local bone destruction in CIA mice.

We next determined the effect of eASCs on the generation of autoantibodies and cytokine secretion profile in peripheral blood. Mice were sacrificed 1 week after last infusion of eASCs (day 22), when the highest reduction in the arthritis score (Figure 1) and paw swelling (Figure 3A) was observed. As expected, anti-chicken collagen II antibodies, TNF-α, IL-6, IL-23, and IL-17 levels were increased significantly in CIA mice with respect to healthy mice (Figure 3). While IL eASC treatment reduced the level of anti-chicken collagen II IgG compared to untreated CIA mice, no significant effects were observed in the levels of circulating cytokines. This was similar to previous reports (41). However, a tendency to reduce the levels of TNF-α and IL-23 and some increment in IL-6 levels were observed in the eASC-treated CIA group.

These results suggest that the systemic inflammation in arthritic mice is not clearly affected upon IL treatment with eASCs and that their beneficial effects are mainly due to their effect in the lymphoid tissues and in the joints.

Several studies, in different experimental animal models of inflammation, have indicated that the therapeutic effect of eASCs is in part mediated by increasing immune T cells with a regulatory phenotype, such as regulatory T cells (CD25⁺Foxp3⁺ CD4⁺) and Tr1 cells (IL10⁺ CD4⁺) (10, 11, 42, 43). Therefore, we investigated the effects of the eASC administration into the lymph nodes on the levels of these regulatory T cell populations in spleen and lymph nodes (inguinal/popliteal) at day 22. A higher cellularity in the spleen and lymph nodes was found in CIA mice compared to healthy mice. In the eASC-treated CIA mice, total cell numbers tended to be reduced, although not significantly, with respect to untreated CIA mice (Figure 4).

As expected, CIA mice showed a significant reduction in IL10⁺CD4⁺ T cells and CD25⁺Foxp3⁺CD4⁺ T cells in spleen (Figure 4A; Figure S2A in Supplementary Material) and lymph nodes (Figure 4B; Figure S2B in Supplementary Material), compared to healthy mice. Interestingly, administration of eASCs into the lymph nodes increased the levels of IL10⁺CD4⁺ and CD25⁺Foxp3⁺CD4⁺ T cells in both tissues. Levels of IL10⁺CD4⁺ T cells in spleen and lymph nodes were similar to healthy mice suggesting that the eASCs tend to restore the homeostasis in the tissues.

Taken together, these results suggest that the inflammatory environment in CIA mice is accompanied by a broad reduction of regulatory T cells in lymphoid tissues and that the administration of eASCs in lymph nodes has a positive major impact on these two populations in spleen and draining lymph nodes.

The secretion of IL-10, IL-4, IFNγ, TNFα, and IL-17 cytokines in the supernatant of in vitro cultured splenocytes and draining lymph node cells in response to collagen II were determined (Figure 5). Increased levels of IL-10, IL-4, and IFNγ were found to be increased in eASC-treated CIA mice compared to CIA and healthy mice (Figure 5). Additionally, no clear effects were observed in the secretion of TNFα and IL-17 in both tissues.

Interestingly, the inflammatory/regulatory balance in spleen and lymph nodes, as shown by the TNFα/IL-10 and IL17/IL-10 ratios, was reduced in eASC-treated CIA mice with respect to the untreated CIA mice (Figure 6). The inflammatory/regulatory balance measured by IFNγ/IL-10 ratio was significantly increased in the spleen in eASC-treated CIA mice with respect to untreated CIA mice. This was in contrast to the lymph nodes where IFNγ/IL-10 ratio was decreased in eASC-treated CIA mice with respect to untreated CIA mice. These results are in agreement with previous data obtained by us and others (10, 44).

These results further confirm that the administration of eASCs into the lymph nodes reduces the inflammatory/regulatory balance mainly in the draining lymph nodes of CIA mice where the eASCs were delivered.