Impaired immunomodulatory capacity in adipose tissue‐derived mesenchymal stem/stromal cells isolated from obese patients

Abstract Immune‐modulatory properties of adipose tissue‐derived mesenchymal stem/stromal cells (MSCs) might be susceptible to metabolic disturbances. We hypothesized that the immune‐modulatory function of MSCs might be blunted in obese human subjects. MSCs were collected from abdominal subcutaneous fat of obese and lean subjects during bariatric or kidney donation surgeries, respectively. MSCs were co‐cultured in vitro for 24 h with M1 macrophages, which were determined as M1or M2 phenotypes by flow cytometry, and cytokines measured in conditioned media. In vivo, lean or obese MSCs (5 × 105), or PBS, were injected into mice two weeks after unilateral renal artery stenosis (RAS) or sham surgeries (n = 6 each). Fourteen days later, kidneys were harvested and stained with M1 or M2 markers. Lean MSCs decreased macrophages M1 marker intensity, which remained elevated in macrophages co‐cultured with obese MSCs. TNF‐α levels were four‐fold higher in conditioned media collected from obese than from lean MSCs. RAS mouse kidneys were shrunk and showed increased M1 macrophage numbers and inflammatory cytokine expression compared with normal kidneys. Lean MSCs decreased M1 macrophages, M1/M2 ratio and inflammation in RAS kidneys, whereas obese MSCs did not. MSCs isolated from lean human subjects decrease inflammatory M1 macrophages both in vivo and in vitro, an immune‐modulatory function which is blunted in MSCs isolated from obese subjects.

class-II molecules, 6 allowing them to evade immune recognition.
Thus, exogenously delivered MSCs can modulate the host immune response. The mechanisms involved in their immunomodulatory effects have not been fully elucidated and may depend on both cell contact-dependent mechanisms and paracrine effects through production of cytokines and various soluble factors that regulate immune cell functions. [7][8][9] Previous studies 10,11 demonstrated in vitro that healthy MSCs inhibit adhesion and invasion of inflammatory cells and promote polarization of macrophages from an inflammatory (M1) to anti-inflammatory (M2) phenotype. Furthermore, adipose tissue-derived MSCs exhibit advantageous capabilities in polarization of M1 macrophages and anti-inflammatory effects compared to bone marrow-derived MSCs. 12 Obesity induces metabolic dysregulation including hyperglycaemia, hyperinsulinemia and dyslipidaemia, which are major risk factors for diabetes and cardiovascular complications. Shree et al. 13 showed that injection of healthy human adipose tissue-derived MSCs exerts beneficial effects in high-fat diet-fed mice. Moreover, obesity can trigger an early senescence programme in human adipose tissue-derived MSCs, which exhibit lower proliferative capacities than non-obese MSCs, suggesting impaired function. 14 However, whether obesity affects the immunomodulatory capability of MSCs is unknown.
This study was therefore designed to test the hypothesis that MSCs from obese human subjects have impaired capability to reverse inflammatory macrophage phenotypes. To this end, we exposed macrophages to 'lean' or 'obese' human MSCs both in vitro and in a RAS mouse model in vivo.

| ME THODS
To evaluate the effects of obesity on their immunomodulation capacity, MSCs were obtained from human subjects with obesity and from lean controls, injected into mice with unilateral RAS, and renal macrophage phenotypes and spleen monocyte reservoir then determined 2 weeks after injection. Furthermore, to compare their direct impact on macrophages, obese and lean MSCs were co-cultured with activated macrophages, and their phenotypes and cytokine release were determined.

| Human subject recruitment and MSCs isolation
All human study procedures were approved by the Mayo Clinic Institutional Review Board. Eligible subjects were 18-80 years of age with a BMI < 30 kg/m 2 (lean) or BMI > 35 kg/m 2 (obese), who underwent kidney donor nephrectomy or weight reduction gastric surgery, respectively (n = 6 each), at Mayo Clinic in Rochester, Minnesota, between October 2017 and March 2019. All subjects gave written informed consent. Blood samples and body weights were collected prior to surgery. At the time of their surgical procedures, adipose tissue (0.5-2.0 g) samples were harvested and MSCs isolated following standard protocols. [2][3][4]14 Briefly, adipose tissue was minced, incubated with collagenase-H at 37°C for 45 min, and after stopping, digestion was filtered through a 100μm cell strainer. The cellular suspension was centrifuged for 10 min at 233 g to pull down cells, and the cellular pellet re-suspended in Advanced Minimum Essential Medium supplemented with 5% platelet lysate (PLTmax, Mill Creek Life Sciences). MSCs were then expanded in culture for three passages to prepare for experimentation.

| In vivo study
To investigate the effects of obesity on the ability of MSCs to modulate macrophage phenotype in vivo, MSCs obtained from either lean or obese subjects (n = 6 each) were injected in mice with unilateral RAS. Kidney M1 and M2 macrophages were counted ex vivo in kidneys harvested 2 weeks after injection. Male 129-S1 mice (Jackson Furthermore, kidney gene expression of the inflammatory cytokines TNFα, IL-6, IL-1a and MCP-1 was evaluated using qPCR.

| In vitro macrophages study
To investigate the direct effects of obesity on the ability of MSCs to modulate macrophage phenotype in vitro, MSCs obtained from either lean or obese subjects were co-cultured with M1 macrophages, and the M1/M2 phenotypes were subsequently determined. To generate M1 macrophages, human monocytes (U937 cells, ATCC, CRL-1593.2™) were cultured in RPMI-1640 supplemented with 10% FBS, 1% penicillin and streptomycin at 37°C in 5% were stained to detect M1, and CD163 and CD200R for M2 (all from BioLegend, 5 ul per test/10 6 cells). These antibodies were selected based on their suitability for flow cytometry, as previously reported. 16 At least 10000 events were counted, and the percentages of doublepositive CD64/CD86 (M1) and CD163/CD200R (M2) cells were determined. In addition, macrophages (M0, M1, M1+Lean MSCs and M1+obese MSCs) and their culture media were collected. Arginase-1 expression in macrophages was then evaluated using Western blotting, and levels of IL-1ra and TNFα in culture media using ELISA. 16

| Statistical analysis
Results of normally distributed data are expressed as mean ± SEM and non-normally distributed as median(Range). Comparisons within groups were performed using the paired Student's t-test or Wilcoxon/Kruskal-Wallis test, depending on normality of distribution, and among groups using ANOVA and unpaired t-test. Chi-square or Fisher's exact tests were used for categorical variables analysis. A statistical difference was considered significant for p ≤ 0.05.

| RE SULTS
Lean and obese subjects were of a similar age, but there were more women in the obese than in the lean group, BMI was higher (p < 0.001), and a history of hyperlipidaemia and diabetes was significantly more common in the obese group. As for medications, statins were primarily prescribed for hyperlipidaemia and metformin or glipizide for diabetes. The history of hypertension or the use of angiotensin-converting enzyme inhibitors or calcium blockers was not different between lean and obese subjects. Furthermore, obese subjects had increased triglyceride and haemoglobin A1c levels, reflecting metabolic alterations in these subjects, although glucose levels were unchanged (Table 1).

| MSC characterization
Flow cytometry analysis showed that MSCs from both lean and obese subjects express CD73, CD90 and CD105 (purity 96.7%-99%) and were negative to CD45 and CD34 ( Figure 1A), confirming their MSC nature.

| In vivo macrophage polarization
Mesenchymal stem cells retention rate in the kidney was not different between RAS+ lean and RAS+ obese (1.1 ± 0.6 vs. 0.7 ± 0.6 per field, p > 0.05) MSCs. No co-localization of MSCs and CD3 + lymphocytes was observed, suggesting no immune rejection of human MSCs ( Figure 1B). The weight ratio of stenotic/contralateral kidneys in RAS was significantly decreased in RAS+ vehicle, confirming renal TA B L E 1 The demographics of lean and obese subjects (n = 6 each) atrophy. This ratio was slightly improved in RAS+ lean MSCs, but not in RAS+ obese MSCs ( Figure 1C).

Renal artery stenosis kidneys showed an increase in both M1
(

| In vitro macrophage polarization
Flow cytometry demonstrated that in vitro, lean MSCs more effectively decreased M1 phenotype and M1/M2 ratio compared with obese MSCs ( Figure 3B). Macrophages stimulated with IFNγ, and LPS showed increased CD68/iNOS intensity, indicating M1 induction.
Co-culture with lean MSCs decreased M1 intensity, which remained elevated in macrophages co-cultured with obese MSCs (Figure 4).
The downregulated arginase-1 expression and increased release of TNFα levels by M1 macrophages were also consistent with phenotype switching, and MSCs reversed the switching. Importantly, lean MSCs were more effective in this respect ( Figure 5A). Furthermore, the levels of IL-1ra in cultured media decreased in M1, indicating decreased anti-inflammatory potential, which were restored after MSC co-culture, with lean MSCs more effective than obese MSCs. Levels of TNFα were higher in conditioned medium collected from obese MSCs than that from lean MSCs, whereas other cytokines tended to be elevated (p = 0.09), but did not reach statistical significance levels. Interestingly, levels of the immunosuppressive factor IDO-1 were increased, whereas PGE2 decreased in obese MSCs ( Figure 5B).

| DISCUSS ION
This study reveals that obesity impairs the capability of MSCs to  to suppress inflammation and may contribute to impaired wound healing in these subjects.

ACK N OWLED G EM ENTS
This research was partly supported by National Institutes of Health grants DK122734, DK120292, AG062104, and DK102325.

CO N FLI C T S O F I NTE R E S T
Dr. Lerman is an advisor to AstraZeneca and Janssen Pharmaceuticals.
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Original data are available per request.