Adipose tissue-derived mesenchymal stromal cells (ADSCs) are an ex vivo expanded heterogenous cell population that includes progenitor and stem cells endowed with multipotential differentiation capacity in vitro and angiogenic and immunomodulatory properties in vivo (1). Due to the relatively easy access to subcutaneous adipose tissue (SAT) and several preclinical results showing the positive role of ADSCs in the resolution of inflammation during tissue repair processes, these cells became the preferred choice for application in regenerative medicine protocols envisioning the treatment of inflammatory conditions (2). However, since stem and progenitor cells are directly affected by their tissue microenvironment (3) and obesity continues to increase worldwide (4), concerns have been raised about the ability of ADSCs isolated from obese and former obese subjects (Ex-Ob ADSCs) to modulate inflammatory responses (5).

During the weight gain process, progressive changes occur in the SAT, affecting its components and functional homeostasis: vascularization is impaired, increasing tissue hypoxia; the secretion of inflammatory mediators increases (6); and the number of metabolic activated macrophages (MMe), which simultaneously express M1 (pro-inflam-matory) and M2 (anti-inflammatory) markers, increases (7). Accordingly, studies have shown that ADSCs from obese animals and humans have reduced proliferation in vitro, primed differentiation to adipocytes, and impaired angiogenic capacity compared to cells from non-obese subjects (8-10). Furthermore, ADSCs of obese individuals were shown to have increased expression of pro-inflam-matory cytokines, such as IL-1β, IL-6, TNF-α (Tumor Necrosis Factor alpha), MCP-1 (Monocyte Chemoattra-ctant Protein) and activated NLRP3 (NOD- and Pyrin Domain-Containing Like Receptor 3) inflammasome signaling (11, 12).

Following bariatric surgery and significant weight loss, Ex-Ob ADSCs were shown to have persistent impairments in their replicative lifespan and differentiation potential (10, 13, 14), but the effects on their immunomodulatory potential remained unclear. Because the main goal of the application of ADSCs in cell therapy is to explore their ability to down-modulate inflammatory responses, we asked whether Ex-Ob ADSCs would recover their anti-inflam-matory activity. Therefore, here we investigated how Ex-Ob ADSCs would influence macrophage differentiation and function in vitro, considering that macrophages have a preponderant role in the induction and resolution of inflammation during tissue repair. Our observations corroborated previous findings (15) that SAT of Ex-morbidly obese patients still presents low-grade inflammation, even after up to six years of weight loss. Despite it, Ex-Ob ADSCs instructed macrophage differentiation toward an M2-like profile, just as ADSCs isolated from overweight subjects and, to some extent, as bone marrow mesenchymal stem/stromal cells (BMSCs), the two main alternative cell candidates available for application in cell therapy. A deeper understanding of the effects of obesity and weight loss on ADSCs function will be fundamental in determining the effectiveness of its application in different eligible cell therapies.

Samples and donors

All sample collections and procedures were approved by the Ethics Committee for Human Research (Nº 06023913. 6.0000.5285) of the Clementino Fraga Filho University Hospital (HUCFF) of the Federal University of Rio de Janeiro (UFRJ). Donors were informed about the study and provided their written informed consent for tissue donation. Subcutaneous adipose tissue and/or lipoaspira-tes were obtained from 26 individuals undergoing plastic surgery at the HUCFF. Of these, nine were overweight (OW) and 17 were ex-morbidly obese (Ex-Ob), that is, former class III obese patients (BMI≥40 kg/m²) that had undergone bariatric surgery at least 18 months before plastic surgery (time of tissue harvest) and had stabilized weight during the last six months. Bone marrow samples were obtained by collecting leftover iliac crest aspirate from discarded collection bags from the Bone Marrow Transplant Unit of the HUCFF. Because the bone marrow collections were for transplantation purposes, all came from healthy donors. As a source of monocytes, buffy coats of peripheral blood were obtained from the HUCFF Hemotherapy Service Blood Bank, also from healthy volunteers. Donor data are summarized in Table 1.

Table 1 . Clinical and general characteristics of the patients

	Overweight (n=9)	Ex-Ob (n=17)	p-value
Age (yr)	43 (31∼59)	47 (32∼64)	0.6044
Sex (female/male)	8/0	14/3	-
BMI before bariatric surgery (kg/m2)	-	55.17 (39.51∼63.65)	<0.001*
BMI at plastic surgery (kg/m2)	25.8 (25.4∼26.89)	32.15 (22.07∼45.67)	<0.05
Weight loss (kg)	-	60.2 (28∼84)	-
Time between surgeries (mo)	-	35 (17∼78)	-
Glycemia before plastic surgery	-	80.5 (74∼96)	-
Seric triglycerides before plastic surgery	-	83 (30∼140)	-

Median values and variation are shown.

Ex-Ob: ex-morbidly obese, BMI: body mass index.

*Significantly different compared to Ex-Ob BMI values at plastic surgery.

Histology and immunohistochemistry

The fragments of the subcutaneous adipose tissue samples were fixed in 10% buffered formalin, processed for routine paraffin histology, cut into 5 μm sections, and stained with H&E. Heat-induced epitope retrieval (HIER) for CD68 immunostaining was performed by placing sections in citrate buffer pH 6.0 in a steamer at 96℃ for 30 min. The sections were then treated with 0.1% trypsin solution with 0.1% CaCl₂ in Tris-HCl pH 7.2 (Sigma-Aldrich, Merck Millipore, Burlington, MA, USA) for 5 min. For the immunostaining of FXIII-A and CD206, HIER was performed by incubating sections for 5 min in citrate buffer pH 6.0 (Zymed Laboratories, Invitrogen Immunodetec-tion, South San Francisco, USA) in a microwave oven. After inhibition of endogenous peroxidase with a 0.3% H₂O₂ solution in methanol (Sigma-Aldrich) for 15 min, sections were incubated with phosphate buffer saline (PBS) containing 5% goat serum (Dako Agilent, Santa Clara, CA, USA) and 10% bovine serum albumin (Sigma-Aldrich) for 1 h. Incubation with mouse anti-human CD68 (clone Kpi, Dako Agilent, 1：50), rabbit anti-human FXIII-A (CalbiochemTM, Merck Millipore, 1：300) and rabbit anti-human d-CRM (CD206, produced and gently provided by Dr. A. Régnier-Vigouroux, University of Mainz, Mainz, Germany, 1：300) was performed for 16 h at 4℃ in a humidified chamber. The sections were washed and incubated with the EnVisionTM+Dual Link System HRp (Dako, Agilent). Antibody staining was revealed using 3-3-diaminobenzidine (Liquid DAB, Dako, Agilent) and counterstaining was performed with hematoxylin.

Immunofluorescence

Paraffin sections were dewaxed, rehydrated, washed with PBS, and incubated with 5% bovine serum albumin in PBS for 1 h to block autofluorescence. The sections were then incubated with 50 mM ammonium chloride (Sigma-Aldrich) in PBS pH 8.0 for 15 min, washed in PBS, and treated with 0.1% trypsin solution with 0.1% CaCl₂ in Tris-HCl pH 7.2 (Sigma-Aldrich) for 5 min. After incubation with primary antibodies against CD68 and FXIII-A, as described above, sections were washed with PBS containing 0.25% Tween-20 (Sigma-Aldrich) and incubated with the following secondary conjugated antibo-dies: anti-mouse IgG Alexa 488 and anti-rabbit IgG Alexa 546 (both from Molecular Probes, Thermo Fisher Scienti-fic, Waltham, MA, USA) for 1 h at room temperature. The nuclei were stained with 10 μg/ml 4’-6-Diamidino-2-Phenylindole (DAPI, Sigma-Aldrich) and sections were mounted with Vectashield (Vector Laboratories, Burlin-game, CA, USA).

Imaging and macrophage quantification

The images were obtained with a light microscope (Eclipse E400, Tokyo, Japan) coupled to a digital camera (Evolution Media Cybernetics Inc., Bethesda, MD, USA) and a computer equipped with the Q Capture 295.0 2.0.5 software (Silicon Graphic Inc, Milpitas, CA, USA). The quantification of macrophages was done as previously described (16). Briefly, CD68⁺, CD206⁺, and FXIII-A⁺ cells and adipocytes were counted in 20∼30 microscopical fields (40× objective) from three sections of each paraffin block, obtained in 100 μm intervals. Data were expressed as the number of macrophages per 100 adipocytes.

Isolation of ADSCs and BMSCs

Lipoaspirates and/or subcutaneous adipose tissue fragments were incubated with 1 mg/ml collagenase IA (Sigma-Aldrich) for 1 h at 37℃, under agitation. The suspension was then mechanically dissociated by pipetting and centrifuged at 400× g for 20 min. The supernatant was discarded and the cell pellet was resuspended in Dulbecco’s low glucose medium (DMEM, LGC, São Paulo, SP, Brazil) supplemented with 10% fetal bovine serum (FBS, Cultilab, Campinas, SP, Brazil) and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin, both Sigma-Aldrich) and passed through a 70 μm nylon mesh. Cells were seeded at a density of 1,0∼2,0×10⁴ cells/cm² and incubated at 37℃ and 5% CO₂. The following day, nonadherent cells were removed by washing with PBS and the medium was changed. Subsequently, the adherent ADSCs were maintained as above until 70% confluence, when cells were harvested with 0.125% trypsin and 0.78 mM EDTA (Sigma-Aldrich) for further expansion (1, 13). BMSCs were isolated from bone marrow aspirates obtained from healthy volunteers at the HUCFF Bone Marrow Transplant Unit, as previously described (17, 18). Briefly, bone marrow suspensions were diluted 6：1 in Hespan^Ⓡ (Hydroxyethylstarchhaline, American Hospital Supply Corp., McGaw Park, IL, USA) and incubated for 30 min for hemosedimentation. Cell suspensions were collected, washed with PBS and seeded at 4,0∼6,0×10⁴ cells/cm² in DMEM low glucose supplemented with 10% FBS and antibiotics. After 3∼4 days, nonadherent cells were removed and BMSCs were cultured for up to 21 days. All isolated ADSCs and BMSCs were expanded until passage 3 for use in subsequent experiments.

Monocyte isolation

Mononuclear cell fractions were isolated from peripheral blood buffy coats by Ficoll HypaquePlus^TM gradient density centrifugation (GE Healthcare, Chicago, IL, USA), washed two times with 1 mM PBS-EDTA, resuspended in serum-free Iscove’s Modified Dulbecco’s Medium (IMDM, LGC) and loaded in a 1：1 (v/v) ratio on top of a 46% Percoll solution (Sigma-Aldrich). After 30 min of centrifugation at 550× g, the monocyte-containing cellular fraction was collected and washed two times with 1 mM PBS-EDTA. The cells were then resuspended in IMDM supplemented with 2% FBS, seeded at a density of 0,5∼1,0×10⁵ cells/cm², and allowed to adhere for 90 min at 37℃ and 5% CO₂. To remove nonadherent cells, cultures were washed three times with PBS, and fresh IMDM supplemented with 10% FBS, 4 mM L-glutamine, 1 mM sodium pyruvate, MEM nonessential amino acids (all from Gibco, Thermo Fisher Scientific), and 4 μg/ml insulin (Humu-lin, Lilly, São Paulo, SP, Brazil) was added (19). Cultures were maintained for 7 days, with medium exchange every 3 days. Alternatively, CD45⁺ CD14⁺ CD16⁻ monocytes were isolated from mononuclear cell fractions by magnetic and FACS sorting. In this case, mononuclear cells were incubated with purified mouse anti-human CD3 primary antibody (clone S4.1; Caltag, Thermo Fisher Scientific) for 30 min at 4℃, washed with PBS and incubated with Dynabeads^Ⓡ Pan Mouse IgG (Dynal Biotech, Thermo Fisher), according to the manufacturer’s instructions. After magnetic separation, the negative cell fractions were collected, washed with PBS, and incubated with primary fluorochrome conjugated mouse anti-human CD45 (Clone H130, BD Biosciences, Franklin Lakes, NJ, USA), CD14 (Clone TUK4, Caltag, Thermo Fisher Scientific) and CD16 (Clone 3G8, Caltag, Thermo Fisher Scientific) antibodies for 30 min at 4℃. After washing with FACS buffer (PBS with 3% FBS), cell sorting was performed on a Moflow^TM cell sorter (Beckman Coulter, Brea, CA, USA) equipped with Summit software (Dako Cytomation, Fort Collins, CA).

Monocyte coculture with ADSCs and BMSCs

On day 7 of total monocyte culture, OW-ADSCs, Ex-Ob ADSCs, and BMSCs were seeded over the monocytes at a density of 1,0×10⁴ cells/cm² and the cultures were maintained for 4 additional days (19). For the CD45⁺ CD14⁺ CD16⁻ monocyte sorted population, 0,5∼1,0×10⁵ cells/cm² were instead seeded over semi-confluent monolayers of OW-ADSCs, Ex-Ob ADSCs, and BMSCs. These cocultures were kept in modified Dexter medium (20), that is, DMEM supplemented with 12.5% FBS, 12.5% horse serum (LGC), 10^-3M hydrocortisone (Blau Farmaceutica SA, Cotia, SP, Brazil) and antibiotics for 48 hours.

ADSC flow cytometry

The ADSCs in third passage were harvested enzymatically with 0.125% trypsin and 0.78 mM EDTA (Sigma-Aldrich), and washed with staining buffer (PBS with 3% FBS and 0.1% sodium azide, Sigma-Aldrich). Cells were then incubated for 30 minutes on ice with the following mouse anti-human antibodies conjugated to fluorochro-mes: anti-CD14 (clone TÜK4; Caltag, Thermo Fisher Scientific), anti-CD31 (clone WM59, Biolegend, San Diego, CA, EUA), anti-CD34 (clone HPCA-2, 8G12; BD-Bioscie-nces), anti-CD44 (clone IM7; eBiosciences, Thermo Fisher Scientific), anti-CD45 (clone HI-30; BD-Biosciences), anti-CD105 (clone 43A4; Biolegend), and anti-CD90 (Thy-1; clone 5E10; e Bioscience).

Flow cytometry of cocultured monocytes with stromal cells

After 4 days of total monocyte cocultures with OW ADSCs, Ex-Ob ADSCs, and BMSCs, cells were incubated overnight with 1 μg/ml of bacterial lipopolysaccharide endotoxin (LPS, Sigma-Aldrich) and with the Brefeldin A protein transporter inhibitor (GolgiStop^TM, BD Biosciences Brasil, São Paulo, SP, Brazil), according to the manufacturer’s instructions. Unsorted LPS-stimulated and sorted monocytes cocultures were enzymatically harvested, washed with staining buffer, and incubated with 10% mouse serum in PBS for 10 min at 4°C, to block immunoglobulin Fc receptors. The cells were then incubated for 30 min at 4℃ with the following fluorochrome conjugated mouse anti-human monoclonal antibodies: anti-CD14 (clone TÜK4), anti-CD16 (3G8) and anti-CD86 (clone BU65), all Caltag, Thermo Fisher Scientific; anti-CD206 (clone 19.2) and anti-CD45 (clone H130), BD Biosciences; and anti-HLA-DR (clone 9∼49), Immunotech, Beckman Coulter. For intracellular staining, after surface labeling, cells were fixed and permeabilized with Cytofix/Cytoperm (BD Biosciences) or Fixation/Permeabilization buffer (Biolegend, San Diego, CA, USA), according to the manufacturer’s instructions, and incubated with anti-TNFα (clone Mab11, BD Biosciences). Acquisition was performed on an FACSCalibur or FACSCanto II instrument, respectively, equipped with CellQuest Pro and FACSDiva software (BD Biosciences). Analysis was performed using FACSDiva or FlowJo (Tree Star, Inc., Ashland, OR).

ELISA

After 4 days of total monocyte cocultures with OW-ADSCs, Ex-Ob ADSCs, and BMSCs, 1 μg/ml of bacterial lipopolysaccharide endotoxin (LPS, Sigma-Aldrich) was added to the cultures as described above. The supernatants were collected, centrifuged for 15 min at 300× g, and stored at −20℃ until analysis. The concentration of TNF-α, IL-10, the CC-chemokine ligand 3/macrophage inflammatory protein-1α (CCL3/MIP-1α), and the CC-chemokine ligand 4/macrophage inflammatory protein-1β (CCL4/MIP-1β) was quantified with commercially available kits (all from R&D Systems, MN, USA), following the manufacturer’s instructions. Measurements were made in duplicate. The cytokine concentration in each supernatant was calculated using internal standard curves generated with known concentration ranges of recombinant proteins.

Statistical analysis

Data are shown as mean±standard deviation (mean± SD). The difference between groups was evaluated using the two-tailed paired t-test or One-Way Anova, as appro-priate. Correlation analysis was performed with Spearman’s test. Values of p<0.05 were considered statistically significant.

Adipose tissue from Ex-Ob patients has an increased number of macrophages

The overweight (OW) and ex-morbidly obese (Ex-Ob) groups were composed mainly of females and had a similar median age (Table 1). Body mass index (BMI) analysis showed that even after significant weight loss (median BMI of 55.17 before bariatric surgery vs. 32.15 at plastic surgery; p<0.001), most Ex-Ob subjects were still classified as class I obesity, that is, with BMI≥30 and<35 (Table 1).

Histological analysis of subcutaneous adipose tissue (SAT) from both groups showed that the degree of hypertrophy (Fig. 1A-1K) and the number of adipocytes per field (6.92±0.61 in OW vs. 7.43±0.73 in Ex-Ob, Fig. 1C) were similar. CD68⁺ macrophages were observed in OW and Ex-Ob SAT, but with a different distribution pattern. While in OW, CD68⁺ macrophages were mostly isolated in the interstitium (Fig. 1D), in Ex-Ob, these cells were preferentially clustered around the adipocytes, in a crown-like shape (Fig. 1E), corroborating previous observations (21). Since areas of fibrosis with an increased number of blood vessels were described in the SAT of Ex-Ob patients (13), sections without fibrosis were selected for the analysis of macrophage infiltration. The quantification of CD68⁺ macrophages showed an approximately two-fold increase in their number within the SAT of Ex-Ob patients (23.50±3.75 in OW vs. 44.45±8.57 in Ex-Ob, Fig. 1F). However, no correlation was observed between the number of CD68⁺ macrophages and the BMI (Fig. 1K), which was still significantly higher in Ex-Ob patients compared to OW (Table 1).

Figure 1. Characterization and quantification of macrophages in the subcutaneous white adipose tissue of overweight (OW) and ex-morbidly obese (Ex-Ob) subjects. Hematoxylin-Eosin staining of OW (A) and Ex-Ob (B) adipose tissue. Immunohistochemistry for CD68 (D, E), FXIII-A (G, H), and CD206 (I, J) in OW and Ex-Ob. (C) Number of adipocytes per microscopic field of view. Bars express mean±SEM (p=0,0846); (F) Percentage of CD68⁺ (n=9 for OW and n=15 for Ex-Ob, *p<0.0001) and FXIII-A⁺ (n=9 for OW and n=17 for Ex-Ob, p=0,3738) macrophages in relation to the number of adipocytes. (K) Correlation analysis between the number of CD68⁺ macrophages and the BMI of Ex-Ob patients, immediately before the plastic surgery (Spearman’s correlation=0.4545 and linear correlation=0.1336). Immunofluo-rescence for CD68 (green, L, P, O, S) and FXIII-A (red, M, Q, O, S) in OW (L∼O) and Ex-Ob (P∼S) samples. Dapi (blue, N, R, O, S) stains nuclei. Merged sections (O, S). Scale bars=25 μm.

To further characterize this exceeding macrophage population, we first evaluated the expression of FXIII-A, a member of the transglutaminase enzyme family, expressed by tissue-resident M2 macrophages (22). Immunohisto-chemistry staining of SAT sections showed FXIII-A⁺ macrophages in the OW and Ex-Ob groups (Fig. 1G and 1H), in a similar percentage (24.23±2.96 in OW vs. 23.70±5.49 in Ex-Ob, Fig. 1F). Next, we evaluated the expression of CD206, a scavenger receptor expressed by murine M2 macrophages and by mixed M1/M2 populations in human adipose tissue (23). CD206⁺ macrophages were evenly dispersed among adipocytes (Fig. 1I and 1J), and quantification of three Ex-Ob and two OW-derived samples showed that their numbers did not differ from FXIII-A⁺ macro-phages.

Then we evaluated whether the excess pool of CD68⁺ macrophages infiltrating the Ex-Ob SAT was otherwise polarized towards the pro-inflammatory M1 phenotype. Co-immunofluorescence staining of SAT sections with anti-CD68 and anti-FXIII-A antibodies showed that in OW SAT, CD68⁺ macrophages were all FXIII-A⁺ (Fig. 1L-1O), while in Ex-Ob SAT both CD68⁺ FXIII-A⁺ and CD68⁺ FXIII-A⁻ cells were present (Fig. 1P-1S). Collectively, these findings suggested that tissue remodeling and lipolysis were still active in Ex-Ob SAT, contributing to the maintenance of metabolic activated macrophages (MMe) and a somewhat inflammatory environment (24).

Ex-Ob ADSCs modulate macrophage differentiation and activation in vitro similarly to OW ADSCs

Therefore, we sought to investigate the immunomodu-latory potential of ADSCs isolated from Ex-Ob SAT. Specifically, we asked whether Ex-Ob ADSCs would still induce macrophage polarization toward the anti-infla-mmatory M2 spectrum, as described for ADSCs from non-obese subjects and other types of mesenchymal stem/stromal cells (25-28). To this end, we first isolated ADSCs in vitro and checked their immunophenotype (Fig. 2 and Supplementary Fig. S1), as well as the absence of contaminating hematopoietic CD45⁺ and CD14⁺ cells between them (Fig. 2A, 2B). We found that both Ex-Ob and OW-ADSCs were negative for CD34 (Fig. 2C, 2D) and CD31 (Fig. 2E, 2F) and homogeneously expressed CD105 (Fig. 2G, 2H), CD44 (Fig. 2I, 2J), and CD90 (Fig. 2K, 2L).

Figure 2. Immunophenotype of ADSCs isolated from Ex-Ob and OW patients. FACS analysis of the hematopoietic markers CD45 (A) and CD14 (B) in cultured ADSCs from Ex-Ob patients (solid red lines). (C∼L) Histograms showing the expression of CD34 (C, D), CD31 (E, F), CD105 (G, H), CD44 (I, J) and CD90 (K, L) antigens in cultured ADSCs from Ex-Ob (solid red lines) and OW (solid blue lines) individuals. Solid gray lines in histograms show the negative control. Data are representative of n=3 Ex-Ob and n=3 OW subjects.

Next, because it was shown that unstimulated classical peripheral monocytes (CD14^Hi CD16⁻) maintain high expression of CD14 and acquire CD16, adopting a M2-like macrophage profile when in contact with BMSCs (26), we added sorted CD14^Hi CD16⁻ monocytes (>96% pure, Fig. 3A-3D) to OW and Ex-Ob ADSCs cultures. BMSCs were also used, for comparison (Fig. 3E). After 48 h of coculture with the three types of stromal cells, immuno-phenotyping analysis showed that the majority population of CD14^Hi monocytes differentiated into CD16⁺ macrophages (Fig. 3F-3H). On the other hand, when monocytes were cultured solely on plastic, less than half of the population became CD16⁺ macrophages (Fig. 3I). While the percentage of CD14^Hi cells that acquired the expression of CD16⁺ were significantly (p<0.05) different in cells cultured onto plastic, no differences were observed comparing cells cultured onto OW-ADSC, Ex Ob-ADSC, and BMSC (Fig. 3J) , indicating that Ex-Ob ADSCs behaved similarly to OW-ADSCs and BMSCs in monocyte differentiation instruction.

Figure 3. Expression of CD16 in peripheral blood CD14⁺⁺ monocytes following coculture with stromal cells. (A, B) Sorting strategy of peripheral blood monocytes showing selected region of FSC versus SSC. (A) combined with the region of CD14^Hi, CD16⁻ events (B). (C, D) post-sort analysis showing 96.4% pure CD14^Hi, CD16⁻ cells. (E) Experimental design of CD14^Hi, CD16⁻ selected peripheral blood monocytes co-cultured with BMSC and ADSCs. (F∼I) FACS analysis of CD14 and CD16 expressions by sorted monocytes after 48 h of culture in contact with OW-ADSCs (F), Ex-Ob-ADSCs (G), BMSCs (H), and plastic (control; No MSC) (I). Numbers represent the percentage of positive cells inside the region. (J) Percentage of CD16⁺ cells within the CD14⁺ cell population. Bars express mean±SEM (*p<0,05). Data of n=3 independent experiments.

Afterward, we investigated the influence of Ex-Ob ADSCs on the function of LPS-stimulated macrophages. Untou-ched monocytes (whole fraction, unsorted) were co-cultured for 4 days in the presence or absence of ADSCs and BMSCs and challenged with LPS during the last 18 hours (Fig. 4A). Regarding the expression of immunophenotypic markers, we found that macrophages derived from ADSCs and BMSCs cocultures were CD14^Hi and HLA-DR⁺, while those maintained in the absence of stromal cells were CD14⁺ and HLA-DR^Hi (Fig. 4B, 4C). The mean fluorescence intensities (MFI) of CD14 and HLA-DR were significantly different between macrophages cultured in contact with stromal cells and those kept solely on plastic (Fig. 4F, 4G). The expressions of CD206 and CD86 were unchanged by interaction with the stromal cells (Fig. 4D, 4E, 4H, 4I). It should be noted that the modulation of all these macrophage markers occurred at equal levels in all three stromal cell cocultures, thus indicating that Ex-Ob ADSCs were able to mirror the effects of OW-ADSCs and BMSCs on macrophage phenotypic modulation (Fig. 4F-4I).

Figure 4. Expression of macrophage markers by unsorted peripheral blood monocytes cocultured with stromal cells and challenged with LPS. (A) Experimental design: Whole fraction, unsorted monocytes obtained from peripheral blood were maintained in culture for seven days, then co-cultured with OW-ADSCs, Ex-Ob ADSCs, and BMSCs for another four days, and finally treated with LPS for 18 hours. At the end of the experiment, cells were harvested for flow cytometry (B∼E) and the supernatants were harvested for ELISA (F∼I). (B∼E) Histograms showing the expression of CD14 (B), HLA-DR (C), CD206 (D), and CD86 (E) by macrophages in co-culture with Ex-Ob ADSCs (solid red line), OW ADSCs (solid black line), BMSCs (solid blue line), and alone (solid green line). Isotype controls are shown as solid gray line. Data are representative of n=4 independent experiments. (F∼I) Mean fluorescence intensity (MFI) of CD14 (F), HLA-DR (G), CD206 (H), and CD86 (I) expressions. Bars show mean±SEM of n=4 independent experiments. *p<0,05. MC: macrophage-only cultures.

To further investigate the influence of Ex-Ob ADSCs on macrophage function under challenging inflammatory conditions (experimental design shown in Fig. 4A), we evaluated the cytoplasmic expression of the cytokine TNF-α in macrophages treated with LPS, by flow cytometry (Fig. 5A-5C), and the levels of TNF-α, IL-10, MIP-1α, and MIP-1β secretion in the coculture supernatants, by ELISA (Fig. 5D-5G). In the intracellular analysis, a similar percentage of CD14⁺ TNF-α⁺ macrophages was observed under all coculture conditions (Fig. 5A, 5B). Additionally, no significant differences were observed in the level of TNF-α expression, reflected by the mean fluorescence intensities (MFI), in macrophages derived from all conditions (Fig. 5C).

Figure 5. Secretion of cytokines by LPS-activated macrophages cultured for 4 days in the absence (MC, macrophage-only) or presence of stromal cells and stimulated with LPS. (A) Representative histogram of the FACS analysis of the intracellular TNF-α expression in CD14⁺ macrophages cultured in the absence (solid green line) or presence of Ex-Ob ADSCs (solid red line), OW ADSCs (solid black line), and BMSCs (solid blue line). Isotype control is represented by the solid gray line. Percentage of CD14⁺ TNF-α⁺ cells (B) and the MFI for TNF-α (C). Data show mean±SEM of n=3 independent experiments. (D∼G) Concentration of TNF-α (D), IL-10 (E), MIP-1α (F), and MIP-1β (G) in the supernatants of macrophages cultured alone (MC) or in the presence of Ex-Ob-(MC+Ex-Ob ADSC) and OW-(MC+OW ADSC) derived ADSCs and BMSCs (MC+BMSC). Ex-Ob ADSC, OW ADSC, and BMSC represent the conditions where stromal cells were cultured alone. Data show mean±SEM of n=3 independent experiments. *p<0.05.

In the supernatants, an increased secretion of TNF-α and IL-10 was detected, after LPS treatment, under the condition in which macrophages were cultured alone (Fig. 5D, 5E). Interestingly, this increase in TNF-α secretion by LPS-challenged macrophages was not significantly inhibited by any ADSC type, as did BMSCs (Fig. 5D). On the other hand, IL-10 secretion was equally inhibited by all three types of stromal cells (Fig. 5E). Regarding MIP-1α and MIP-1β, the levels of their secretions by LPS-challenged macrophages were not affected by the presence of any type of stromal cell (Fig. 5F, 5G). It is noteworthy that we did not detect TNF-α, MIP-1α, and MIP-1β in the conditioned media of ADSCs and BMSCs only cultures, treated or not with LPS (Fig. 5D, 5F, 5G), which confirms that these cytokines were indeed secreted only by macrophages. IL-10 was the only exception, being also detected in the conditioned media of stromal cells (Fig. 5E). However, no increases above the basal level of IL-10 secretion were observed after stimulation of the stromal cells with LPS, indicating that LPS did not influence the levels of IL-10 secretion by these cells (Fig. 5E). Taken together, the results presented herein indicate that Ex-Ob ADSCs behave similarly to OW-ADSCs in the education of macrophage profile and function, that is, they favor the adoption of an M2-like immunophenotype and a mixed (M1/M2) secretory response.

The differentiation potential, together with the proangiogenic and immunomodulatory properties of ADSCs, opened promising therapeutic avenues for degenerative and inflammatory diseases (1, 2). However, the increasing incidence of obesity (29), a chronic low-grade systemic inflammatory disorder that significantly modifies the microenvironment of the adipose tissue and consequently the functional properties of ADSCs runs against this hope (30). Although the transcriptomic profile, replicative lifespan, and differentiation potential of ADSCs are known to not fully recover after weight loss (10, 31), it was not yet clear whether modifications in the immunomodulatory properties of ADSCs could be reversed. Here, we showed that despite the persistent higher number of M1 macrophages in the SAT of ex-morbidly obese individuals, isolated Ex-Ob ADSCs were able to stimulate a pro-repair response in macrophages, similarly as ADSCs of overweight individuals and BMSCs.

The persistence of CD68⁺ macrophages within the SAT has previously been documented, in studies with distinct follow-up (15, 32). First, Cancello et al. (33) showed that although the number of CD68⁺ macrophages infiltrating the adipose tissue of Ex-Ob patients decreased after three months of bariatric surgery, their percentage remained significantly higher compared to lean subjects after two years of sustained weight loss (15). Then Ara et al. (32) reported that CD68⁺ macrophage infiltration can persist for up to 10 years after weight loss. Our data corroborated these findings, and since we found that exceeding CD68⁺ macrophages within the crown-like structures of Ex-Ob SAT did not express FXIII-A, a marker of tissue-resident M2 macrophages (22), we reasoned that this pool was composed of M1 and/or the so-called metabolic active macrophages (MMe), a subtype that accumulates in adipose tissue depots during obesity (24). Therefore, our study strengthens the notion that macrophage infiltration takes a long time or may not even completely disappear after weight loss, despite the overall systemic improvement of the inflammatory and metabolic status of Ex-Obese patients (34, 35).

However, our analysis of the immunomodulatory properties of Ex-Ob ADSCs showed that their persistent contact with the infiltrating M1 or MMe macrophages within the SAT microenvironment did not affect their ability to stimulate macrophage polarization toward the M2 spectrum, as reported for other tissue-derived mesenchymal stromal cells (26, 28). Immunophenotypic analysis of macrophages co-cultured with Ex-Ob ADSCs revealed that these cells, similarly to OW-ADSCs and BMSCs, were able to induce CD16 expression by macrophages. Chiossone et al. (26) also reported this observation in cocultures of murine macrophages with BMSCs, which occurred together with the up-regulation of class II MHC, CD11b, CD209, CD163, and CD206, a typical signature of M2 macrophages. Furthermore, when our macrophages were treated with LPS, the presence of Ex-Ob ADSCs inhibited the down-modulation of CD14 and the up-regulation of HLA-DR, which occurs when macrophages orient their polarization toward the M1 spectrum (26).

Although the analysis of the functional secretory profile of macrophages generated in coculture with stromal cells provided seemingly contradictory results with respect to previous literature reports (26), i.e., the lack of inhibition of TNF-α secretion by LPS-treated macrophages in coculture with both ADSCs types and the suppression of IL-10 secretion by these macrophages in all cocultures of stromal cells, we attribute these asymmetries to the heterogeneity of macrophage functional responses and/or to distinctions in the methods of macrophage isolation and activation. Most studies obtain macrophages by positive selection from a population of peripheral blood monocytes, such as CD14 staining, followed by FACS and the addition of stimulating cytokines, such as M-CSF (26). As CD14 is an essential part of the LPS receptor complex, macrophage isolation using this marker can, by itself, prone macrophages to the M1 phenotype, allowing them to display more intense inflammatory responses when activated. In this study, we isolated macrophages by plastic adhesion followed by lymphocyte depletion. Furthermore, we did not provide any cytokines until ADSCs and BMSCs were added, thus favoring the maintenance of macrophages in the resting M0 state, in which, depending on the type and strength of the activation signaling provided, macrophages can acquire intermediate characteristics between M1 and M2, instead of a full M1 or M2 profile (36). This fully goes along and might explain our observations.

Despite the subtle differences observed between the effects of ADSCs and BMSCs on macrophage education, which can be attributable to their different tissue origins, and the limitations of our study in fully mapping the expression and secretory profile of macrophages co-cultured with stromal cells, we can reasonably conclude that Ex-Ob ADSCs mirrored the properties of OW-ADSCs and, to some extent, of BMSCs, inducing a common M2-like immunophenotype and a functional mixed (M1/M2) secretory response on macrophages. Therefore, beyond underscoring further knowledge about the contribution of ADSCs to macrophage instruction within the SAT, we provide evidence that Ex-Ob ADSCs behave similarly to ADSCs isolated from overweight subjects in the matter of immuno-modulation. Although OW-ADSCs and BMSCs are not true controls, these are the two main other progenitor/stromal cell types that are largely characterized, experi-mentally tested in several disease models, and therefore conceivable for application in regenerative medicine strategies, thus justifying their choice as comparison parameters in this study.

In conclusion, our results indicate that the adipose tissue of ex-obese individuals might still be a useful source of ADSCs for the treatment of inflammatory diseases, such as Crohn’s disease and spinal cord injury, two conditions for which our group has already investigated the therapeutic application of ADSCs, with promising results (37, 38). Further in vivo studies will be fundamental to validate the in vitro observations reported here and certify the immunomodulatory roles of Ex-Ob ADSCs.

We thank Dr. Wagner Baetas da Cruz for providing the scientific protocol for CD206 immmunohistochemistry; Dr. Vívian Samoto for technical assistance during immunohistochemistry experiments; Dr. Alex Duarte for providing the protocol for intracellular cytokine analysis; and the Multiuser FACS Facility of the Health Sciences Center of the Federal University of Rio de Janeiro for assistance during cell sorting experiments. This work was supported by grants from the Brazilian government agencies FAPERJ, CAPES, and CNPq. The sponsors had no role in study design, data collection and analysis, decision to publish or manuscript preparation.

The authors have no conflicting financial interests.

Supplementary data including one figure can be found with this article online at https://doi.org/10.15283/ijsc22172