Mesenchymal stem cells protect renal tubular cells via TSG-6 regulating macrophage function and phenotype switching

Abstract

Tumor necrosis factor (TNF)-α-induced gene/protein (TSG)-6 regulates the immunomodulatory properties of mesenchymal stem cells (MSCs), but its ability to protect the ischemic kidney is unknown. In a swine model of renal artery stenosis (RAS) and metabolic syndrome (MetS), we assessed the contribution of TSG-6 produced by MSCs to their immunomodulatory properties. Pigs were studied after 16 wk of diet-induced MetS and unilateral RAS and were either untreated or treated 4 wk earlier with intrarenal autologous adipose tissue-derived MSCs (n = 6 each). Lean, MetS, and RAS sham animals served as controls. We studied renal function in vivo (using computed tomography) and kidney histopathology and macrophage phenotype ex vivo. In vitro, TSG-6 levels were also measured in conditioned media of human MSCs incubated with TNF-α and levels of the tubular injury marker lactate dehydrogenase in conditioned media after coculturing macrophages with injured human kidney 2 (HK-2) cells with or without TSG-6. The effects of TSG-6 on macrophage phenotype (M1/M2), adhesion, and migration were also determined. MetS + RAS showed increased M1 macrophages and renal vein TNF-α levels. After MSC delivery, renal vein TSG-6 increased and TNF-α decreased, the M1-to-M2 ratio decreased, renal function improved, and fibrosis was alleviated. In vitro, TNF-α increased TSG-6 secretion by human MSCs. TSG-6 decreased lactate dehydrogenase release from injured HK-2 cells, increased expression of macrophage M2 markers, and reduced M1 macrophage adhesion and migration. Therefore, TSG-6 released from MSCs may decrease renal tubular cell injury, which is associated with regulating macrophage function and phenotype. These observations suggest that TSG-6 is endowed with renoprotective properties.

NEW & NOTEWORTHY Tumor necrosis factor-α-induced gene/protein (TSG)-6 regulates the immunomodulatory properties of MSCs, but its ability to protect the ischemic kidney is unknown. In pigs with renal artery stenosis, we show that MSC delivery increased renal vein TSG-6, decreased kidney inflammatory macrophages, and improved renal function. In vitro, TSG-6 decreased inflammatory macrophages and tubular cell injury. Therefore, TSG-6 released from MSCs may decrease renal tubular cell injury, which is associated with regulating macrophage function and phenotype.

INTRODUCTION

Metabolic syndrome (MetS) is a cluster of cardiovascular disease-related risk factors that is frequently associated with chronic kidney disease (CKD) and increases its progression toward end-stage renal disease (1). Renal artery stenosis (RAS) is a major cause for secondary hypertension and may produce chronic underperfusion of the renal parenchyma, leading to progressive loss of renal mass and function and the development of kidney ischemia (2). Coexisting MetS and RAS are linked to poorer outcomes after revascularization, underscoring the need for targeted interventions capable of preserving the poststenotic kidney in subjects with MetS.

Mesenchymal stem cells (MSCs) have a number of potential advantages for kidney repair. MSCs are undifferentiated nonembryonic stem cells that can be isolated from a variety of tissues and possess the ability to differentiate into a broad spectrum of cell lineages. MSCs are immunoprivileged and confer immunomodulatory effects by paracrine mechanisms (3). Previous studies have shown in several kidney injury models that MSCs promote tubular regeneration and reduce inflammation and vascular injury. In kidneys of atherosclerotic RAS pigs, we observed that MSCs restored renal blood flow (RBF) and increased vascularity (4). In addition, several studies have demonstrated that MSCs increase levels of anti-inflammatory cytokines and attenuate renal fibrosis. However, the molecular mechanism underlying that effect of MSCs on the MetS + RAS kidney has not been fully elucidated.

As a secretory cytokine, tumor necrosis factor (TNF)-α-induced gene/protein-6 (TSG-6) is associated with immunomodulation and is a key factor in endowing anti-inflammatory responses of MSCs in inflammatory diseases such as peritonitis (5), pancreatitis (6), inflammatory bowel disease (7), myocardial infarction (8), and skin wounds (9). Indeed, TSG-6 has been shown to restore renal structure and function in experimental rodent models of acute or chronic renal disease (10–12). However, whether TSG-6 is involved in the ability of adipose tissue-derived MSCs to rescue kidney function in chronic ischemic injury in a large preclinical animal model is unknown. Moreover, the mechanism associated with such a renoprotective effect remains to be defined.

The present study took advantage of a novel porcine model of coexisting MetS and RAS (MetS + RAS), comorbidities that accentuate renal inflammation, associated with renal tissue damage. We tested the hypothesis that intrarenal delivery of MSCs is associated with increased TSG-6 levels and attenuated injury in the MetS + RAS kidney. Furthermore, we tested whether the renoprotective properties of TSG-6 might be determined by regulating macrophage function and phenotype.

METHODS

Animal Experiments

All procedures in animals were approved by the Mayo Clinic Institutional Animal Care and Use Committee. Thirty domestic female pigs were studied during 16 wk of observation. Pigs were randomly divided into the following five groups: lean, RAS, MetS (high-cholesterol/high-carbohydrate diet), MetS + RAS, and MetS + RAS + pig MSCs (p-MSCs) (n = 6 each).

At baseline, 18 pigs started the high-cholesterol/high-carbohydrate diet [MetS, 5B4L, protein 16.1%, ether extract fat 43.0%, and carbohydrates 40.8%, Purina Test Diet, Richmond, IN (13)], whereas 12 others were fed standard pig chow (lean). Six weeks later (Fig. 1A), 18 pigs were anesthetized, and RAS was induced by placing a local irritant coil in the main renal artery (14). Sham renal angiography was performed in the remaining six lean and six MetS pigs. Abdominal adipose tissue was collected in all pigs at that time and used to harvest autologous p-MSCs.

Figure 1.

Pig mesenchymal stem cell (p-MSC) characterization and tracking. A: schematic of the experimental protocol. B: images of CM-DiI-labeled (red) p-MSCs in frozen sections of the poststenotic kidney of pigs from the metabolic syndrome (MetS) + renal artery stenosis (RAS) group 4 wk after cell delivery. C: in frozen sections of MetS + RAS kidneys that received MSCs, immunofluorescent staining with cytokeratin staining identified MSCs within tubular cells. D: lactate dehydrogenase (LDH) levels in urine. Results are means ± SD (t test). *P < 0.05 vs. the lean group (n = 5 or 6 pigs/group).

Six weeks later, after the stenosis was assessed by angiography, p-MSCs (10⁷) (4) were slowly injected into the stenotic kidney of six MetS + RAS pigs through the renal artery. The others underwent sham angiography.

Four weeks later, pigs were anesthetized, and systemic blood samples were collected under fluoroscopic guidance. Urine and blood samples were taken for the measurement of biochemical parameters, and an arterial catheter was used to measure mean arterial pressure (MAP). Glomerular filtration rate (GFR) and RBF were then measured using MDCT (Multidetector CT, Somatom Sensation-128, Siemens Medical Solution, Forchheim, Germany) from images acquired following a central venous injection of iopamidol (0.5 mL/kg) and calculated from time-density curves as previously described (1).

One week later, pigs were euthanized, and the kidneys were collected for histological and molecular assays (Fig. 1A).

p-MSC Isolation, Characterization, Delivery, and Tracking

p-MSCs were cultured from swine abdominal adipose fat (10 g) that was digested in collagenase-H for 45 min, filtered, and cultured in advanced MEM (Gibco/Invitrogen) supplemented with 5% platelet lysate (Mill Creek Life Sciences, Rochester, MN) at 37°C with 5% CO₂ and were kept in cell recovery medium at 80°C. p-MSCs were characterized by the expression of common MSC markers, and their potential to differentiate into adipocytes, chondrocytes, and osteocytes was assessed as previously described (15). p-MSCs were then labeled with a fluorescent membrane dye (CM-DiI), suspended in 10 mL of PBS (10⁶ cells/mL), and injected slowly through a balloon catheter placed in the renal artery proximal to the stenosis (15). After the completion of in vivo experiments, the kidneys were dissected, and frozen tissue was cut into 5-μm sections and mounted with DAPI. Prelabeled p-MSCs were detected by red labeling and identified in each slide.

Kidney Tissue Experiments

Kidney sections were stained with Masson’s trichrome staining and then examined by light microscopy (×40) in a blinded manner. Twenty randomly selected nonoverlapping slides from each kidney were analyzed using ImageJ software (National Institutes of Health). Periodic acid-Schiff-stained paraffin sections were analyzed using a blinded scoring method using a grid. For each square, the presence of tubule injury, including intratubular proteinaceous casts or tubular atrophy, was documented. The final score was presented as the percentage of positive squares.

For immunofluorescence, the tissue was stained for CD68 (MA5-13324, ThermoFisher), arginase-1 (Arg-1, sc-20150, Santa Cruz Biotechnology), and inducible nitric oxide synthase (iNOS; PA1-036, ThermoFisher). After nuclei had been stained with DAPI, double immunostainings for CD68 and iNOS, CD68, and ARG-1 were visualized with a fluorescence microscope (×400). To explore whether macrophages uptake MSCs, CD68 staining was performed in frozen kidney sections.

Furthermore, to explore potential downstream mechanisms involved in the TSG-6 pathway, kidney protein expression was measured using Western blot analysis. We assessed TSG-6 (ARP63662-P050, 1:1,000, Avivasysbio), transforming growth factor (TGF)-β1 (sc-52892, 1:500, Santa Cruz Biotechnology), TNF-α (ab6671, 1:200, Abcam), collagen type I (ab34710, 1:1,000, Abcam), α-smooth muscle actin (α-SMA; ab6671, 1:131.58, Abcam), Toll-like receptor (TLR)-4 (nb100-56566, 1:1,000, Novusbio), myeloid differentiation primary response protein (MyD)88 (ORB583123, 1:1,000, Biorbyt), and NF-κB (cs-6956, 1:1,000, Cell Signaling).

Urinary levels of the tubular injury marker lactate dehydrogenase (LDH; ab102526, Abcam) and plasma levels of neutrophil gelatinase-associated lipocalin (NGAL) and lipids were measured by standard methods.

Activation of Human MSCs with TNF-α to Secrete TSG-6

For clinical translation, we used human MSCs (h-MSCs) collected from healthy kidney donors at the time of donation. h-MSCs were cultured in advanced MEM (Gibco/Invitrogen) supplemented with 10% FBS (Sciencell) with 5% platelet lysate, 1% glutamine (GlutaMax), 1% penicillin-streptomycin, and 0.2% Heparin and incubated at 37°C in 5% CO₂. When MSCs reached ∼70% confluence, they were stimulated with 10 ng/mL TNF-α (R&D Systems) for 18 h (5). Expression of TSG-6 in cells and levels in supernatants were measured by Western blot analysis and ELISA (Sigma), respectively. Written informed consent was obtained from all study participants, and study procedures were approved by the Mayo Clinic Institutional Review Board (No. 18–005076).

Macrophage Migration and Adhesion

Human monocytes (U937 cells, RCB0435, RIKEN) were cultured in RPMI-1640 supplemented with 10% FBS and 1% penicillin-streptomycin at 37°C in 5% CO₂. First, macrophages were induced by the addition of 100 ng/mL propylene glycol methyl ether acetate (PMA) (16) and after 12 h were washed with PBS (pH 7.2). Classical models of activated M1 macrophages were then established in vitro by treating some of the cells with 100 U/mL interferon (IFN)-γ + 5 ng/mL LPS (Sigma). Cells were randomly divided into the following three groups: M0, M1 (M0 + IFN-γ + LPS), and M1 + TSG-6 (100 ng/mL, R&D), and macrophage phenotype, migration, and adhesion were studied by assays.

For Transwell migration assays (Costar polycarbonate filters, 5-μm pores), macrophages (0.5 × 10⁵ cells/well) were added to the upper chamber without serum, and complete medium (10% serum) was added to the lower chamber. Cells migrating to the lower chamber after 8 h of incubation were washed, stained with Hoechst 33342 (ThermoFisher Scientific), and counted with a Celigo Image Cytometer (Nexcelom Bioscience).

For adhesion, fibronectin (Sigma) was used to coat 96-well plates overnight at 4°C. Macrophages (0.1 × 10⁵ cells/plate) were added to the fibronectin-coated plates and incubated for 4 h at 37°C. Adhered cells were then washed, fixed in formaldehyde, stained with Hoechst 33342, and counted with the Celigo Image Cytometer.

Human Kidney 2 Cells Coculture with Macrophages

Cell viability was measured using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Promega), a time-dependent assay. Human kidney 2 (HK-2) cells (∼ 2 × 10⁴) were seeded into 96-well plates overnight. After 12, 24, 48, and 72 h of 10 ng/mL TNF-α (R&D) and 10 ng/mL antimycin-A (AMA; Sigma) treatment (17), 20 μL of MTT was added into each well, and cells were incubated for 4 h at 37°C. The optical absorbance of each sample was measured at 490 nm using a PowerWave XS machine (BioTek).

In the Transwell coculture system, HK-2 cells (4 × 10⁵) plated on six-well plates (lower chamber) were cultured for 24 h with or without TNF-α and AMA, which induce tubular injury. Macrophages (M0 or M1, 0.5 × 10⁵ cells/well) were added to the upper chamber for 24 h in the absence or presence of TSG-6 (R&D Systems). LDH levels in supernatants of the lower chamber released by HK-2 cells were measured by ELISA.

Statistical Analysis

Data were analyzed using GraphPad Prism 5 (GraphPad Software, San Diego, CA), and results are expressed as means ± SD for normally distributed variables and as medians (ranges) for non-Gaussian distributed data. Comparisons within groups were performed using a paired Student’s t test and among groups using ANOVA and an unpaired t test with Bonferroni correction. A statistical difference was considered significant at P ≤ 0.05.

RESULTS

p-MSCs Ameliorate Renal Damage and Regulate M1/M2 Macrophage Phenotype In Vivo in MetS + RAS Pigs

All p-MSCs were characterized by the expression of traditional surface markers (CD90, CD105, and CD73) and transdifferentiation into adipocytes, chondrocytes, and osteocytes, as we have previously shown (2, 18). Four weeks after intrarenal administration, red fluorescent signal was detected within the stenotic kidney of treated but not untreated animals (Fig. 1B). MSCs mainly integrated into interstitial and tubular compartments (Fig. 1C).

Renal function was reduced in RAS pigs, whereas body weight, renal volume, blood pressure, GFR, and lipid levels were comparably elevated in MetS and MetS + RAS groups compared with the lean group (P < 0.05; Table 1), due to hyperfiltration. However, compared with MetS controls, MetS + RAS pigs exhibited increased serum creatinine and decreased GFR and RBF, which improved 4 wk after p-MSC treatment (P < 0.05). In addition, the LDH level in the urine was significantly increased in MetS, RAS, and MetS + RAS pigs but not in MetS + RAS + p-MSC pigs (Fig. 1D). Compared with lean pigs, plasma levels of the acute kidney injury marker NGAL were unchanged in RAS and MetS + RAS + p-MSC pigs (Table 1), given the chronic nature of their disease. Interestingly, NGAL levels were decreased in both MetS and MetS + RAS pigs, possibly because high-fat, high-fructose diets might affect hepatic NGAL production (19).

Table 1.

Systemic characteristics and single kidney function in study groups at 16 wk

	Lean Group	MetS Group	RAS Group	MetS + RAS Group	MetS + RAS + p-MSC Group
Body weight, kg	55.1 ± 5.6	90.9 ± 3.7*	46.3 ± 5.9*‡	86.2 ± 9.0*§	79.3 ± 19.2*§
Renal volume, mL	127.4 ± 15.4	200.8 ± 31.1*	93.1 ± 25.2*‡	158.1 ± 33.3‡§	153.2 ± 16.4‡§
Serum creatinine, μmol/L	1.5 ± 0.2	1.6 ± 0.2	1.7 ± 0.2	2.1 ± 0.2†*	1.6 ± 0.1*†
Mean arterial pressure, mmHg	96.1 ± 7.9	123.8 ± 14.3*	123.2 ± 24.5*	128.4 ± 14.8*	118.2 ± 18.0*
Degree of stenosis, %	NA	NA	84.8 ± 11.0*‡	89.2 ± 25.1*	77.3 ± 23.3*
Glomerular filtration rate, mL/min	83.1 ± 10.7	136.2 ± 8.1*	65.5 ± 21.1*‡	101.7 ± 16.2*‡§	119.9 ± 14.6*†‡§
Renal blood flow, mL/min	603.6 ± 84.9	877.5 ± 87.5*	420.7 ± 46.7*‡	509.4 ± 94.7*‡	739.4 ± 78.7†§
Plasma renin activity, ng/mL/h	0.13 ± 0.17	0.13 ± 0.1	0.14 ± 0.07	0.23 ± 0.10	0.15 ± 0.13
Total cholesterol	82.4 ± 6.8	361.7 ± 60.3*	94.9 ± 18.4‡	467.5 ± 199.9*§	455.6 ± 240.3*§
High-density lipoprotein-cholesterol	49.7 ± 17.2	125.4 ± 45.3*	50.7 ± 13.4‡	98.2 ± 40.7*§	97.8 ± 17.1*§
Low-density lipoprotein-cholesterol	37.7 ± 16.6	410.0 ± 105.5*	41.6 ± 5.4‡	436.0 ± 134.1*§	395.0 ± 222.5*§
Triglycerides	8.7 ± 2.9	14.8 ± 9.6*	12.9 ± 2.0*	13.2 ± 11.5*	14.8 ± 8.8*
Neutrophil gelatinase-associated lipocalin, ng/mL	134.5 ± 6.6	85.2 ± 5.2*	147.6 ± 22.7†	110.3 ± 8.5*	152.8 ± 13.9†‡

Values are means ± SE. Pigs were divided into the following groups: lean, metabolic syndrome (MetS), renal artery stenosis (RAS), MetS + RAS, and MetS + RAS + pig mesenchymal stem cell (p-MSC) treatment. *P < 0.05 vs. the lean group; †P < 0.05 vs. the RAS + MetS group; ‡P < 0.05 vs. the MetS group; §P < 0.05 vs. the RAS group.

Histology showed that cortical and medullary fibrosis (trichrome staining) increased in MetS + RAS compared with lean, MetS, and RAS pigs but decreased in MetS + RAS + p-MSC pigs, as did renal tubular score (Fig. 2, A–F). Congruently, kidney protein expressions of collagen type I, α-SMA, and TGF-β increased in MetS or MetS + RAS pigs but decreased in MetS + RAS + p-MSC pigs (Fig. 1), supporting the antifibrotic effects of MSC.

Figure 2.

Pig mesenchymal stem cells (p-MSCs) induce a shift in renal pathology and macrophage phenotype. At the end of the study, kidneys were harvested, paraffin sections were used for Masson’s trichrome staining (A and B), and semiquantitative analysis of the renal cortex (D) and medulla (E) demonstrated renal fibrosis in the metabolic syndrome (MetS) + renal artery stenosis (RAS) group that decreased after p-MSC treatment. Periodic acid-Schiff (PAS) stains showed a similar trend for renal tubular injury (C and F). G–J: representative merged images (×40) of immunofluorescence staining for stenotic kidney M1 macrophages [CD68 (red)/inducible nitric oxide synthase (green)] and M2 macrophages [CD68/arginase-1 (green)] (double staining yellow). Quantification of stenotic kidney M1 and M2 macrophages and the M1-to-M2 ratio is shown. Results are means ± SD (t test). *P < 0.05 vs. the lean group; #P < 0.05 vs. the MetS + RAS group; †P < 0.05 vs. the MetS group; &P < 0.05 vs. the RAS group (n = 5 or 6 pigs/group).

Furthermore, the number of infiltrating inflammatory M1 macrophages (CD68⁺/iNOS⁺), which was higher in MetS and RAS kidneys compared with lean kidneys, further increased in MetS + RAS but decreased in MetS + RAS + p-MSC kidneys. In contrast, fewer reparative M2 macrophages (CD68⁺/Arg-1⁺) were observed in MetS, RAS, and MetS + RAS kidneys compared with lean kidneys, but their number was normalized in p-MSC-treated pigs (Fig. 2, G–I). Consequently, the M1-to-M2 ratio was markedly elevated in the MetS + RAS group but blunted in the MetS + RAS + p-MSC group (Fig. 2J).

TNF-α and TSG-6 Expression in MetS + RAS

TNF-α levels increased in renal veins from the MetS + RAS group (P = 0.038 vs. the lean group) but decreased after p-MSC treatment (P = 0.043 vs. the MetS + RAS group; Fig. 3A), whereas TSG-6 conversely increased only in the MetS + RAS + p-MSC group compared with the lean group (P = 0.041; Fig. 3B). Their levels did not correlate with each other (r = 0.147, P = 0.46). However, TNF-α showed a direct correlation with the number of M1 macrophages (r = 0.547, P = 0.02; Fig. 3C) and an inverse correlation with the number of M2 macrophages (r = −0.4935, P = 0.015; Fig. 3D). Given the unchanged levels of TSG-6 in most experimental groups, its levels were not significantly correlated with macrophage phenotypes (M1: r = 0.098, P = 0.320; M2: r = 0.191, P = 0.181). However, within the MSC-treated and untreated MetS + RAS groups, TSG-6 levels directly correlated with the number of M2 macrophages (r = 0.543, P = 0.034; Fig. 3F) but not M1 macrophages (r = 0.383, P = 0.109; Fig. 3E).

Figure 3.

Link between tumor necrosis factor (TNF)-α and the macrophage M1/M2 phenotype. Renal vein plasma samples were collected using a catheter at the end of the study, and levels of TNF-α (A) and TNF-α-induced gene/protein (TSG-6; B) were determined. Renal vein TNF-α correlated directly with M1 (C) and inversely with M2 (D) macrophage numbers. Renal vein TSG-6 levels showed no correlation with M1 macrophages (E) but correlated directly with M2 macrophages (F). G: in frozen sections of metabolic syndrome (MetS) + renal artery stenosis (RAS) kidneys treated with pig mesenchymal stem cells (p-MSCs), p-MSCs (red) incorporated within kidney macrophages (green). Results are means ± SD (t test). *P < 0.05 vs. the lean group; #P < 0.05 vs. the MetS + RAS group; †P < 0.05 vs. the MetS group; &P < 0.05 vs. the RAS group (n = 5 or 6 pigs/group, except for E and F, where n = 6 pigs each in the MetS + RAS and MetS + RAS + p-MSC groups).

In kidney tissue, protein expression of TNF-α slightly increased in all experimental groups compared with the lean group but achieved statistical significance only in the RAS group (Supplemental Fig. S1; see https://doi.org/10.6084/m9.figshare.13557539.v2). Kidney TSG-6 expression was not different among the groups.

M1 Macrophages Pretreated with TSG-6 Reduce Renal Tubular Cell Injury In Vitro

In vitro, TSG-6 protein expression increased in h-MSCs cocultured with TNF-α and its levels in their supernatant (Fig. 4, A and B).

Figure 4.

Tumor necrosis factor (TNF)-α increased TNF-α-induced gene/protein (TSG-6) expression, which reduced renal tubular cell injury. A: representative Western blot of TSG-6 protein expression in TNF-α-stimulated human mesenchymal stem cells (h-MSCs) harvested from healthy human subjects. GAPDH was used as an internal control. B: supernatant TSG-6 levels in TNF-α-induced h-MSCs. Human tubular cell line HK-2 cell vitality fell time dependently by TNF-α and antimycin-A (AMA) treatment (C), whereas lactate dehydrogenase (LDH) levels rose (D). E and F: supernatant TNF-α (E) and LDH (F) levels after TSG-6 stimulation of M1 macrophages. Results are means ± SE for experiments performed in triplicate (t test). *P < 0.05 vs. the normal control (NC); #P < 0.05 vs. TNF-α + AMA.

To explore the direct effect of TSG-6 on injured tubular cells, TNF-α and AMA were used to induce damage in vitro in HK-2 cells, which were then treated with TSG-6 with or without M1 macrophages. First, to select the most appropriate intervention time that induced moderate injury, HK-2 cells were stimulated with TNF-α and AMA for different durations (Fig. 4C). We observed a decrease in vitality and increase in LDH levels by 24 h, which was selected for subsequent experiments (Fig. 4D).

Preinjured HK-2 cells and macrophages were then cocultured in the presence or absence of TSG-6 for 24 h. LDH in the supernatant was tested to examine the effect of TSG-6-stimulated M1 macrophages on damaged HK-2 cells. TNF-α (P = 0.045 vs. control) and LDH (P = 0.022 vs. control) decreased with TSG-6-pretreated M1 macrophages, whereas no changes were noted when HK-2 cells were treated with M1 macrophages or TSG-6 separately (P = not significant vs. TNF-α + AMA), although TSG-6-treated cells were not different from controls either (Fig. 4, E and F).

TSG-6 Promotes Macrophage Phenotype Switching and Decreases M1 Migration and Adhesion In Vitro

To further explore the mechanism by which TSG-6-pretreated macrophages protected damaged cells, the direct effects of TSG-6 on M1/M2 macrophage phenotype were studied. We found that in preestablished M1 macrophages, TSG-6 upregulated M2 but not M1 markers (Fig. 5A), suggesting the initiation of phenotype switching.

Figure 5.

Tumor necrosis factor-α-induced gene/protein (TSG-6) attenuates M1/M2 macrophage phenotype, adhesion, and migration. Human monocytes (U937 cells) were stimulated and polarized toward M1 macrophages and then treated with TSG-6 in vitro. A: TSG-6 increased M2 marker mannose receptor (MR) expression in macrophages (M0) (GAPDH loading control). B: TSG-6 decreased M1 macrophage adhesion and migration. iNOS, inducible nitric oxide synthase. Results are means ± SE for experiments performed in triplicate (t test). *P < 0.05 vs M0; #P < 0.05 vs M1.

We then performed functional assays in M1 macrophages cultured with or without TSG-6. M1 macrophages showed augmented adhesion to fibronectin and migration compared with control macrophages (P < 0.05). Cells incubated with TSG-6 decreased adhesion by nearly 20% and migration by 50% (Fig. 5B). Importantly, MSCs also incorporated within macrophages in MetS + RAS + p-MSC kidneys (Fig. 3G), supporting the notion that MSC-derived TSG-6 might directly regulate macrophage function and phenotype in vivo.

The TLR4-MyD88-NF-κB Axis

Exploration of downstream mechanisms involved in the TSG-6 pathway showed that MyD88 expression rose significantly in MetS and RAS groups, strongly tended to rise in the MetS + RAS group (P = 0.06 vs. the lean group), and significantly decreased after p-MSC treatment (P = 0.001 vs. the MetS + RAS group). MyD88 expression inversely correlated with M2 numbers (r = −0.546, P = 0.002). TLR4 expression was not increased in the experimental groups and only decreased in the MSC-treated MetS + RAS group, whereas NF-κB expression was unchanged among the groups. TSG-6 showed an inverse correlation with TLR4 (r = −0.782, P = 0.004) and MyD88 (r = −0.626, P = 0.026; Supplemental Fig. S2; see https://doi.org/10.6084/m9.figshare.13557542.v1).

DISCUSSION

Our study demonstrates that a rise in TSG-6 levels is associated with improvement of renal structure and function in experimental MetS + RAS pigs treated with adipose tissue-derived p-MSCs, suggesting that TSG-6 might play a role in this mechanism. Improvement of stenotic kidney GFR and RBF, as well as serum creatinine, was associated with decreased fibrosis and M1 macrophage numbers and with an increase in TSG-6 in the stenotic kidney vein. Furthermore, in vitro TSG-6 attenuated tubular injury by increasing M2 phenotype markers in M1 macrophages and reducing M1 macrophage adhesion and migration. These observations may implicate TSG-6 in the beneficial effects of MSCs on stenotic pig kidneys.

MSCs have been proposed as a novel therapeutic approach for acute or chronic kidney disease thanks to their proangiogenic, anti-inflammatory, and antifibrotic activity (20, 21). In recent years, studies have identified beneficial effects of MSCs on regeneration of the stenotic kidney parenchyma to blunt vascular rarefaction and fibrosis. MSC delivery prevented the progressive increase in systolic arterial pressure, reduced sympathetic hyperactivity, improved renal morphology, induced neovascularization, and reduced fibrosis in stenotic rat kidneys (22, 23). We have previously found that MSCs restored renal hemodynamics and function and decreased tissue perturbations in stenotic pig (2, 4, 15) and human (24) kidneys. The present study extends our previous studies by showing that MSCs also restore renal function and alleviate fibrosis in kidneys of a novel MetS + RAS pig model, which are markedly damaged (1, 18). However, the mechanism of these protective effects has not been fully elucidated.

MSCs exert renal protection mainly through paracrine mechanisms (20, 25). TNF-α and other cytokines in inflamed tissues can activate MSCs to secrete a variety of anti-inflammatory factors. TSG-6 is a key regulator in maintaining stemness and biological properties of MSCs (26, 27) and mediates their ability to attenuate peritoneal inflammation and fibrosis (28). Consistent with those studies, the present study shows that TSG-6 increased and TNF-α decreased after MSC injection into the stenotic renal artery of MetS + RAS pigs. Furthermore, we also found a decrease in urinary levels of LDH, a marker of tubular injury. These findings indicate that a decrease in renal inflammation and tubular injury by MSCs might be linked to TSG-6. However, further studies are needed to identify the exact biological role of TSG-6 on renal tubular protection. Moreover, given that TNF-α and TSG-6 levels in vivo did not correlate, additional mechanisms likely regulate TSG-6 levels and warrant additional studies. Indeed, we observed that TSG-6 levels correlated inversely with expression of inflammatory TLR-4 and its downstream mediator MyD88 (29), although NF-κB may be less involved in this interaction.

Interestingly, in vitro TSG-6 alone was incapable of improving TNF-α- and AMA-induced tubular injury, whereas TSG-6-pretreated M1 macrophages effectively alleviated HK-2 injury. These data underscore the mandatory role of macrophages in the renoprotection bestowed by TSG-6. MSC incorporation within kidney macrophages further supports the notion that MSC-derived TSG-6 might directly regulate renal macrophage function and phenotype and play an indirect role in renal tubular protection.

Kidney injury in the MetS + RAS group reflects complex interactions among vascular rarefication, oxidative injury, and recruitment of inflammatory cellular elements that ultimately produce glomerular and tubular atrophy and local interstitial fibrosis. Macrophage accumulation characterizes inflammatory diseases, with their numbers and phenotypes strongly associated with the severity of renal injury and progressive chronic renal failure. After adhesion and migration to injury sites, monocytes transform to different macrophage phenotypes, depending on the microenvironment. M1 macrophages contribute to tissue inflammation and damage, whereas M2 macrophages display anti-inflammatory and tissue repair properties (30–32). We have recently shown that reparative macrophage populations in stenotic mouse kidneys are endowed with proangiogenic and prohealing properties (33). Thus, regulation of macrophage function and phenotype may be useful for the prevention and treatment of ischemic kidney injury. For example, Song et al. (7) found that TSG-6 released from MSCs ameliorated inflammatory bowel disease by inducing macrophage phenotypic switching from M1 to M2. Our in vitro study also shows that TSG-6 can promote M1 macrophage switching to M2 macrophages and decrease M1 macrophage adhesion and migration. Furthermore, TSG-6 reduced inflammation and protected renal tubules or cells, possibly by regulating macrophage function and phenotype. Although the precise mechanism remains unclear, the TLR4-MyD88-NF-κB axis (34) and TSG-6/CD-44 receptor signaling pathways (35) might be involved in mediating the effects of TSG-6 driving the macrophage phenotype switching to an M2 type. In addition, we previously found that active vitamin D regulated macrophage M1/M2 phenotypes via the STAT1-triggering receptor expressed on myeloid cells (TREM)-1 pathway in diabetic nephropathy (36), and similar mechanisms might be activated by TSG-6 as well.

Limitations in our study include the use of young animals and short duration of the disease. Given their selective increase, renal venous TSG-6 levels correlated with other factors mostly in MSC-treated or untreated MetS + RAS pigs. Kidney TSG-6 expression was not different among the groups. Notably, we have previously reported a similar discrepancy between renal vein levels and kidney tissue expression of TGF-β in both pigs (37) and patients (38) with renovascular disease. TSG-6 is secreted to the plasma from stimulated blood vessels (39), so that its renal venous plasma levels might be higher than in the kidney parenchyma. Clearly, regulation of TSG-6 in kidney disease is complex. Extension of our findings to male pigs may also require further studies. Nevertheless, the strengths of the present investigation include many human-like features of the pig model, with similar renal injury and dysfunction compared with humans. Additional studies are needed to explore the direct effects of TSG-6 on the kidney in pig models and in human subjects.

In conclusion, our data suggest that MSCs and TSG-6 act in concert to repair the ischemic kidney. TSG-6 released from MSCs may orchestrate this process by regulating macrophage function and phenotype and, in turn, decreasing renal tubular cell injury (Fig. 6).

Figure 6.

Illustration of the renoprotective properties of mesenchymal stem cells (MSCs) potentially mediated through tumor necrosis factor-α-induced gene/protein (TSG-6) regulation of macrophages. TNF-α-induced TSG-6 release from MSCs may decrease renal tubular cells injury, which is associated with, and may be partly mediated by, inducing a macrophage phenotypic switch from M1 to M2 macrophages and reducing M1 macrophage adhesion and migration.

Perspectives and Significance

MSC-derived TSG6 regulates renal macrophage function and phenotype and might thereby indirectly confer renal tubular protection. Its mechanisms might involve TLR and MyD88 signaling and result in blunted fibrosis. Hence, TSG-6 might be a therapeutic target in regenerative medicine. Further studies are needed to leverage the protective effects of MSCs and TSG-6 on renal tubular cells.

GRANTS

This work was partly supported by National Institutes of Health Grants DK122734, DK104273, DK120292, AG062104, and DK102325.

DISCLOSURES

L. O. Lerman is a consultant for AstraZeneca and Janssen Pharmaceuticals. No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

L.O.L. conceived and designed research; Y.Z., T.S., L.Z., A.E., S.C., H.T., I.S., and K.J. performed experiments; Y.Z. analyzed data; L.O.L. interpreted results of experiments; Y.Z., X.-Y.Z., A.L., and L.O.L. drafted manuscript; Y.Z., X.-Y.Z., A.L., and L.O.L. edited and revised manuscript; Y.Z., X.-Y.Z., T.S., L.Z., A.E., S.C., H.T., I.S., K.J., A.L., and L.O.L. approved final version of manuscript.