Adipose Tissue-Derived Mesenchymal Stem Cells Improve Revascularization Outcomes to Restore Renal Function in Swine Atherosclerotic Renal Artery Stenosis§


  • Author contributions: A.E.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; X.Z.: collection and/or assembly of data, data analysis and interpretation, and manuscript writing; J.D.K., H.T., and K.L.J.: collection and/or assembly of data and data analysis and interpretation; J.P.G.: final approval of manuscript; A.L. and S.C.T.: manuscript writing and final approval of manuscript; L.O.L.: conception and design, financial support, collection and/or assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript.

  • Disclosure of potential conflicts of interest is found at the end of this article.

  • §

    First published online in STEM CELLSEXPRESS January 30, 2012.


Reno-protective strategies are needed to improve renal outcomes in patients with atherosclerotic renal artery stenosis (ARAS). Adipose tissue-derived mesenchymal stem cells (MSCs) can promote renal regeneration, but their potential for attenuating cellular injury and restoring kidney repair in ARAS has not been explored. We hypothesized that replenishment of MSC as an adjunct to percutaneous transluminal renal angioplasty (PTRA) would restore renal cellular integrity and improve renal function in ARAS pigs. Four groups of pigs (n = 7 each) were studied after 16 weeks of ARAS, ARAS 4 weeks after PTRA and stenting with or without adjunct intrarenal delivery of MSC (10 × 106 cells), and controls. Stenotic kidney blood flow (renal blood flow [RBF]) and glomerular filtration rate (GFR) were measured using multidetector computer tomography (CT). Renal microvascular architecture (micro-CT), fibrosis, inflammation, and oxidative stress were evaluated ex vivo. Four weeks after successful PTRA, mean arterial pressure fell to a similar level in all revascularized groups. Stenotic kidney GFR and RBF remained decreased in ARAS (p = .01 and p = .02) and ARAS + PTRA (p = .02 and p = .03) compared with normal but rose to normal levels in ARAS + PTRA + MSC (p = .34 and p = .46 vs. normal). Interstitial fibrosis, inflammation, microvascular rarefaction, and oxidative stress were attenuated only in PTRA + MSC-treated pigs. A single intrarenal delivery of MSC in conjunction with renal revascularization restored renal hemodynamics and function and decreased inflammation, apoptosis, oxidative stress, microvascular loss, and fibrosis. This study suggests a unique and novel therapeutic potential for MSC in restoring renal function when combined with PTRA in chronic experimental renovascular disease. STEM CELLS 2012;30:1030–1041


Renal artery stenosis (RAS) is one of the reversible mechanisms for hypertension. Atherosclerosis is the most common cause of RAS, accounting for 90% of the cases [1]. Based upon community-based screening, atherosclerotic RAS (ARAS) exceeding 60% lumen occlusion averages 6.8% in the elderly population [2]. ARAS can accelerate hypertension and lead to loss of kidney function, which are known to increase cardiovascular morbidity and mortality [3].

Renal revascularization using endovascular percutaneous transluminal renal angioplasty (PTRA) and stenting has been a common treatment strategy in patients with ARAS both to reduce blood pressure and to improve renal function. To date, however, randomized, prospective trials fail to identify major benefits from restoring blood flow for preservation of renal function [4, 5] compared with medical therapy alone. This might be due to lingering kidney tissue damage that is not reversed by restoring blood flow with PTRA alone. In line with these clinical observations, we have previously shown in a swine model of non-ARAS that PTRA partially restores the renal microvascular network and improves renal function, but vascular wall remodeling and fibrosis are incompletely reversed [6].

The presence of an atherosclerotic environment compounds these effects. Renal revascularization in a swine model of ARAS normalizes blood pressure levels but fails to improve tubulointerstitial injury, microvascular rarefaction, and renal function in the stenotic kidney [7]. This dissociation between the effects of revascularization on blood pressure and renal function underscores the need to identify more effective strategies to restore the structures within the stenotic kidney in ARAS in addition to PTRA.

Our previous studies demonstrated that intrarenal delivery of autologous hematopoietic endothelial progenitor cells (EPCs) can increase neovascularization and mitigate renal injury in non-atherosclerotic RAS [8]. However, the capacity of this cell-based therapy to reverse the more profound damage observed in the ARAS kidney was more limited in that EPC only partially improved microvascular density and failed to fully restore renal blood flow (RBF) and glomerular filtration rate (GFR) [9]. We speculated that both securing renal arterial patency and at the same time improving the regenerative capacity of the poststenotic kidney using cell-based therapy might be a more effective strategy to preserve the stenotic kidney. As a practical matter, autologous EPCs are difficult to isolate and expand. Mesenchymal stem cells (MSCs) are undifferentiated nonembryonic stem cells present in adult tissues, which have the ability to differentiate into a broad spectrum of cell lineages [10]. Moreover, MSC can be isolated from a variety of tissues, including adipose tissue and bone marrow, and possess immunomodulatory properties that decrease inflammation and immune responses [11].

Previous studies showed that MSCs restore renal structure and function in experimental rodent models of acute renal failure [12]. Whether MSC might augment renal function and structure improvement in response to PTRA in a large animal model remains unknown. Thus, we hypothesized that intrarenal infusion of allogeneic MSC at the time of revascularization would restore renal cellular integrity and repair mechanisms in experimental ARAS.


Six weeks after induction of RAS and before PTRA, all ARAS pigs demonstrated hemodynamically significant stenosis (79.4% ± 2.7%, p = .26 analysis of variance) [13], and mean arterial pressure (MAP) was elevated compared with normal pigs (p < .01 in all). The systemic characteristics in all pigs 4 weeks after PTRA or sham are summarized in Table 1. Total cholesterol, high-density lipoprotein (HDL), and low-density lipoprotein (LDL) levels were elevated in all ARAS groups compared with normal. As common in chronic ARAS [14, 15], plasma renin activity (PRA) levels were similar among the groups.

Table 1. Systemic characteristics (mean ± SEM) in normal, ARAS, ARAS + PTRA, and ARAS + PTRA + MSC pigs (n = 7 each) 4 weeks after PTRA or sham
inline image

PTRA Successfully Reduced Blood Pressure

There was no residual stenosis at 16 weeks in PTRA-treated pigs (Fig. 1A). Continuously measured MAP decreased immediately after PTRA and persisted at normal levels until the end of the study (p < .05 vs. ARAS, p > .05 vs. normal) (Table 1, Fig. 1B).

Figure 1.

PTRA + MSC restored blood pressure, renal hemodynamics, and function. (A): Renal angiography in a pig with ARAS before (left) and 4 weeks after (right) revascularization with PTRA. (B): Mean arterial pressure measured using telemetry decreased after PTRA. (C): Red fluorescence of chloromethylbenzamido-DiI (arrows, ×40) MSC cytokeratin (green)-stained stenotic kidney 4 weeks after administration. Blue: 4′,6-diamidino-2-phenylindole nuclear stain. (D): Single-kidney RBF and GFR were restored in MSC-treated pigs, although RBF response to Ach remained blunted. *, p < .05 versus normal, , p < .05 versus ARAS + PTRA + MSC, and , p < .05 versus baseline. Abbreviations: Ach, acetylcholine; ARAS, atherosclerotic renal artery stenosis; GFR, glomerular filtration rate; MSC, mesenchymal stem cell; PTRA, percutaneous transluminal renal angioplasty; RBF, renal blood flow.

MSC Characterization and Culture

MSC displayed a fibroblast-like, spindle-shaped morphology (Supporting Information Fig. 2As), expressed CD44, CD90, and CD105 markers (Supporting Information Fig. 2Cs), and secreted vascular endothelial growth factor (VEGF) and tumor necrosis factor (TNF)-α in the culture media (Supporting Information Fig. 2Bs). Furthermore, successful transdifferentiation of MSC into osteocytes, chondrocytes, and adipocytes in vitro supported their mesenchymal origin (Supporting Information Fig. 3As).

MSC Home to the Stenotic Kidney

The MSC retention rate in the kidney 4 weeks after intra-arterial administration was 13.1% ± 2.2% (1–2 cells per field (×40) or 800–1,000 per an entire slide). Chloromethylbenzamido-DiI-labeled MSCs were mostly detected at the renal cortical interstitium (cortical-medullary engraftment ratio = 5:1) 4 weeks after injection, although cytokeratin staining showed that some MSCs were incorporated into renal tubules (Fig. 1C). In addition, few MSC costained with the endothelial marker CD31 and the proliferating cell nuclear antigen (PCNA) marker (Supporting Information Fig. 4s). In contrast to the stenotic kidney, very few (1–2 per an entire slide) cells were detected in the contralateral kidney (Supporting Information Fig. 3sB) and heart (Supporting Information Fig. 3sC), as previously shown [8, 16]. Histological analysis showed no evidence of cellular rejection (e.g., CD3 clusters), microinfarcts, or tumors in tissue sections from ARAS + PTRA + MSC pigs.

MSCs Restore Renal Hemodynamics and Function

Basal stenotic kidney GFR and RBF were similarly attenuated in ARAS and ARAS + PTRA (Fig. 1D, p < .05 vs. normal) but not different than normal levels in ARAS + PTRA + MSC, although the large variability of RBF might have contributed to its apparent increase. GFR responses to acetylcholine (Ach) were normalized in MSC-treated pigs, while RBF responses remained blunted (p = .32 vs. baseline). Serum creatinine levels were higher in ARAS compared with normal and remained elevated after PTRA (p = .04 vs. normal; p = .77 vs. ARAS). Treatment with MSC led to a fall in serum creatinine to normal levels (p = .14 vs. normal, Table 1).

Microvascular Architecture Is Improved in MSC-Treated Pigs

Spatial density of cortical microvessels was similarly diminished in ARAS and ARAS + PTRA but improved after MSC treatment to levels not different from normal pigs (Fig. 2A, 2B). In particular, the number of small vessels (<40 μm) was significantly reduced in ARAS and ARAS + PTRA compared with normal but normalized after MSC administration (Fig. 2C, p = .27 vs. normal). The fraction of glomeruli and tubulointerstitial area stained with the endothelial markers CD31 and Von Willebrand factor (vWF) was similarly decreased in ARAS and ARAS + PTRA compared with normal and returned to normal levels in ARAS + PTRA + MSC pigs (p > .05 vs. normal, Supporting Information Fig. 5sB). Vessel diameter was similarly increased in ARAS and ARAS + PTRA compared with normal but decreased to normal levels in MSC-treated pigs (Fig. 2D, p = .12 vs. normal). Tortuosity (a measure of angiogenesis and vascular remodeling) was increased in MSC-treated pigs compared with normal but was lower than in ARAS (p = .03) and ARAS + PTRA (Fig. 2E, p = .05) pigs.

Figure 2.

Treatment with MSC improved angiogenesis and microvascular architecture. (A): Microcomputer tomography three-dimensional images of the kidney showing improved microvascular architecture in ARAS + PTRA + MSC. Spatial density (B) and its classification by vessel size (C), average vessel diameter (D), and tortuosity (E) of renal cortical microvessels. (F): Renal protein expression of VEGF and VEGFR-2, eNOS, and bFGF in normal, ARAS, ARAS + PTRA, and ARAS + PTRA + MSC pigs. *, p < .05 versus normal and , p < .05 versus ARAS + PTRA + MSC. Abbreviations: ARAS, atherosclerotic renal artery stenosis; bFGF, basic fibroblast factor; eNOS, endothelial nitric oxide synthase; MSC, mesenchymal stem cell; PTRA, percutaneous transluminal renal angioplasty; VEGF, vascular endothelial growth factor; VEGFR, VEGF receptor.

Angiogenic Factors Were Upregulated in Animals Treated with MSC

Expression of VEGF was reduced in ARAS and ARAS + PTRA (Fig. 2F, p < .05 vs. normal) but restored to normal levels in MSC-treated pigs (p < .05 vs. ARAS and ARAS + PTRA, p = .61 vs. normal). In addition, expression of the proangiogenic factors, endothelial nitric oxide synthase (eNOS), and basic fibroblast growth factor (bFGF) was downregulated in ARAS and ARAS + PTRA animals, but treatment with MSC restored their expression to levels not different from normal (p > .05 vs. normal). Contrarily, renal expression of VEGF receptor (VEGFR)-2 that was attenuated in ARAS was similarly normalized in both PTRA-treated groups (Fig. 2F, p > .05 vs. normal both).

MSC Reduced Oxidative Stress

Circulating levels of 8-isoprostanes were significantly higher in sham-treated and PTRA-treated ARAS compared with normal (p = .04 and p = .03, respectively) but were restored to normal levels after renal administration of MSC (Table 1, p > .05 vs. normal). Moreover, in situ production of superoxide anion was similarly increased in ARAS and ARAS + PTRA (p = .01 vs. normal and p = .99 vs. ARAS) and decreased to levels not different from normal in ARAS + PTRA + MSC (Fig. 3A, p = .10 vs. normal). Also, the increased protein expression of the NAD(P)H-oxidase subunit p47phox observed in ARAS and ARAS + PTRA was normalized after MSC treatment, suggesting a decreased potential for superoxide generation. Furthermore, protein expression of nitrotyrosine (NT), which was similarly and significantly elevated in ARAS and ARAS + PTRA kidneys compared with normal (p < .05), was substantially reduced in MSC-treated pigs (Fig. 3B, p < .05 vs. ARAS and ARAS+PTRA, p = .29 vs. normal), implying decreased production of peroxynitrite.

Figure 3.

Oxidative stress declined after PTRA + MSC. (A): Renal production of superoxide anion (A), detected by DHE (×40), and its quantification (B). (C): Representative immunoblots and renal protein expression of NT and p47 in normal, ARAS, ARAS + PTRA, and ARAS + PTRA + MSC. (D): Renal protein expression of TNF-α, MCP-1, IF-γ, and NFκβ in normal, ARAS, ARAS + PTRA, and ARAS + PTRA + MSC pigs. *, p < .05 versus normal and , p < .05 versus ARAS + PTRA + MSC. Abbreviations: ARAS, atherosclerotic renal artery stenosis; DHE, dihydroethidium; IF-γ, interferon γ; MCP, monocyte chemoattractant protein; MSC, mesenchymal stem cell; NFκB, nuclear factor kappa B; NT, nitrotyrosine; PTRA, percutaneous transluminal renal angioplasty; TNF-α, tumor necrosis factor α.

MSC Decreased Inflammation in the Stenotic Kidney

The number of CD45+, CD3+, and CD163+ cells infiltrating the kidney was similarly increased in ARAS and ARAS + PTRA pigs compared with normal (p < .05 for both) but decreased to normal levels after treatment with MSC (Fig. 4A–4D). Although MSC failed to fully restore renal expression of the proinflammatory cytokines TNF-α and interferon (IF)-γ (p < .05 vs. normal), they tended to decrease their levels compared with ARAS and ARAS + PTRA pigs (Fig. 3D). Furthermore, renal expression of monocyte chemoattractant protein (MCP)-1 and nuclear factor kappa B (NFκB) was attenuated only in ARAS + PTRA + MSC pigs. Moreover, systemic levels of interleukin (IL)-1β, which were significantly increased in ARAS (p = .02 vs. normal) and ARAS + PTRA (p = .04) pigs, were restored to normal levels in animals treated with MSC (p = .1 vs. normal) (Table 1).

Figure 4.

PTRA + MSC decreased renal inflammation. Representative immunostaining (×40) of B-T lymphocytes (CD45+), T lymphocytes (CD3+), and macrophages (CD163+) (A) and their quantification in tubular and glomerular compartments (B). *, p < .05 versus normal and , p < .05 versus ARAS + PTRA + MSC. Abbreviations: ARAS, atherosclerotic renal artery stenosis; MSC, mesenchymal stem cell; PTRA, percutaneous transluminal renal angioplasty.

MSC Decreased Apoptosis and Increased Proliferation of Renal Cells

The number of apoptotic cells was similarly increased in ARAS and ARAS + PTRA compared with normal (p < .05 both) but decreased to normal levels in pigs treated with MSC (Fig. 5D, p > .05 vs. normal). The number of cells stained with PCNA was significantly elevated in ARAS + PTRA + MSC pigs compared with the other three groups (Supporting Information Fig. 5sA).

Figure 5.

Renal fibrosis and apoptosis were reduced in MSC-treated pigs. (A): Representative renal trichrome staining (×40) in normal, ARAS, ARAS + PTRA, and ARAS + PTRA + MSC pigs. Periglomerular and tubulointerstitial fibrosis (B) and glomerular score (C, % of sclerotic glomeruli) decreased after PTRA + MSC. *, p < .05 versus normal and , p < .05 versus ARAS + PTRA + MSC. (D): TUNEL staining showing increased number of apoptotic cells (green) in ARAS and ARAS + PTRA, which decreased in MSC-treated animals. Abbreviations: ARAS, atherosclerotic renal artery stenosis; MSC, mesenchymal stem cell; PTRA, percutaneous transluminal renal angioplasty; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling.

Renal Scarring Was Reduced in ARAS + PTRA + MSC

Tubulointerstitial fibrosis observed in ARAS and ARAS + PTRA pigs was attenuated in ARAS + PTRA + MSC (Fig. 5A, 5B, p = .02 vs. ARAS, p = .05 vs. ARAS + PTRA, and p = .12 vs. normal). Furthermore, glomerular score, which did not improve after revascularization alone, declined in MSC-treated pigs (Fig. 5C, p = .002 vs. ARAS, p < .04 vs. ARAS + PTRA, and p = .29 vs. normal). Collagen content, assessed by Sirius red, was similarly elevated in ARAS and ARAS + PTRA compared with normal (p < .05 for both) but decreased in PTRA + MSC-treated animals (Fig. 6A, p = .05 vs. ARAS and ARAS + PTRA, p = .07 vs. normal). PTRA alone also failed to restore the expression of the fibrogenic factor transforming growth factor (TGF)-β1 (Fig. 6B, p < .05 vs. normal), but its levels were significantly reduced in ARAS + PTRA + MSC (p = .04 vs. ARAS, p = .04 vs. ARAS + PTRA, and p = .73 vs. normal). No significant differences were found in the expression of podocin and nephrin among the groups (data not shown) or urinary albumin excretion (Table 1).

Figure 6.

Collagen formation diminished after PTRA+ MSC. (A) Sirius red staining, viewed under polarized light (top) and its quantification (bottom) decreased by MSC, as did TGF-β1 expression (B). *p<0.05 vs. normal, p<0.05 vs. ARAS+PTRA+MSC. Abbreviations: ARAS, atherosclerotic renal artery stenosis; MSC, mesenchymal stem cell; PTRA, percutaneous transluminal renal angioplasty; TGF, transforming growth factor.


This is, to the best of our knowledge, the first study demonstrating that adjunctive cell-based therapy in conjunction with restoration of vascular patency can improve renal structural and functional responses in experimental ARAS. Intrarenal administration of adipose tissue-derived MSC improved renal function and structure 4 weeks after revascularization and reduced oxidative stress, apoptosis, fibrosis, inflammation, and microvascular remodeling in the stenotic ARAS kidney. This may thus provide a novel therapeutic approach to preserve the stenotic kidney.

Although renal stenting is commonly used in clinical ARAS, data supporting PTRA and stenting versus medical therapy are controversial [17]. Restoring vessel patency alone has been shown to confer limited benefit as regarding recovery of kidney function [18, 19]. Consequently, the use of this procedure has declined considerably over the past few years. One result of this trend has been the alarming increase in the appearance of medically treated subjects with more severe kidney injury beyond the stenotic lesion.

We have recently shown in a swine model of non-atherosclerotic RAS that PTRA successfully restored blood pressure, GFR, and renal endothelial function but not RBF or fibrosis in the stenotic kidney [6]. We have also shown in a swine model of ARAS that PTRA improved blood pressure but failed to restore renal function beyond the stenotic lesion [7]. This study extends these observations and underscores the ability of PTRA to reverse renovascular hypertension but not necessarily dysfunction in the stenotic ARAS kidney. Thus, whereas blood pressure may be controlled successfully with antihypertensive medication and/or revascularization, renal dysfunction requires additional strategies, targeted to repair the kidney parenchyma.

Several previous studies demonstrated that MSC can protect the kidney from ischemia/reperfusion injury and stimulate renal parenchymal regeneration [12, 20–23]. Ninichuk et al. also demonstrated that weekly injections with MSC prevent loss of peritubular capillaries and reduce interstitial fibrosis in collagen4a3-deficient mice [24]. Our study now shows in a large animal model that intra-arterially delivered MSC restored basal hemodynamics and function in the kidneys beyond a stenotic lesion. These observations were accompanied by normalized GFR (although not RBF) responses to endothelium-dependent challenge with Ach, suggesting improvement of endothelial function.

We have previously shown that impaired renal function in ARAS is associated with oxidative stress and decreased expression of angiogenic factors, leading to microvascular loss and renal dysfunction in the stenotic kidney [25]. In this study, both renal oxidative stress, assessed by the in situ production of superoxide anion, and systemic oxidative stress, as reflected by isoprostane levels, were decreased in MSC-treated pigs. This antioxidant effect was also illustrated by the decreased expression of NAD(P)H-oxidase (p47phox) and peroxynitrite (NT) formation observed in ARAS + PTRA + MSC pigs. These findings extend previous studies showing an antioxidant effect of MSC through modulation of pathways associated with the activation of antioxidant pathways such as superoxide dismutase and glutathione peroxidase [26, 27].

A selective loss of small microvessels, which led to increased average vessel diameter in ARAS and ARAS + PTRA, was substantially improved in ARAS + PTRA + MSC pigs. These findings were accompanied by restoration of endothelial cell density, reflected by the increased glomeruli and tubulointerstitial staining with CD31 and vWF, respectively. Similarly, MSC attenuated microvascular remodeling in the stenotic kidney, reflected by the decrease in vessel tortuosity. Tortuous newly formed microvessels are frequently unstable and hyperpermeable [28], and the decreased tortuosity associated with the increase in vascular density might thus represent increase vascular maturity. Because reactive oxygen species in high concentrations inhibit angiogenesis, the decreased oxidative stress in PTRA + MSC-treated pigs may have improved microvascular architecture reflected by their increased spatial density. Alternatively, the improved microcirculation might have resulted from increased renal expression of the proangiogenic factors, VEGF, VEGFR-2, eNOS, and bFGF, possibly via paracrine production or an indirect effect of MSC. Previous studies in an acutely ischemic rat model suggested that VEGF is an important mediator of the renoprotective effect of MSC [29]. Likewise, the potent angiogenic properties of MSC were supported in our study by the active secretion of VEGF in cell culture. Nevertheless, baseline stenotic kidney RBF was only slightly improved in ARAS + PTRA + MSC, and RBF responses to Ach did not differ among the three ARAS groups. The blunted RBF reactivity suggested residual endothelial dysfunction in ARAS + PTRA + MSC, and that restoration of the microcirculation was at least functionally incomplete. Importantly, GFR was contrarily more prominently improved by MSC, implying a role for this adjunct therapy in improving renal function.

This study also showed that treatment with MSC reduced renal inflammation, as evidenced by decreased infiltration of CD163+ macrophage and CD45+ and CD3+ lymphocytes in the stenotic kidneys of ARAS + PTRA + MSC animals. Furthermore, systemic IL-1β levels and renal expression of TNF-α, MCP-1, IF-γ, and NFκB decreased (although not necessarily normalized) in MSC-treated pigs, suggesting that MSC decreased local production of inflammatory mediators. These results agree with a previous study that showed reduced levels of MCP-1, a proinflammatory cytokine that plays a critical role in the pathogenesis of atherosclerosis [30], after infusion of bone marrow-derived MSC in hypercholesterolemic mice [31]. This anti-inflammatory effect of MSC might be related to their immunomodulatory properties, as well as paracrine release of anti-inflammatory cytokines, which may influence surrounding parenchymal cells in the stenotic kidney. Indeed, MSCs have the ability to inhibit maturation of dendritic cells and suppress the function of B cells, T cells, and natural killer cells, by decreasing the surface expression of class I and II major histocompatibility molecules and the expression of costimulatory molecules [32]. In contrast, a recent study in a porcine model of ischemia reperfusion injury showed limited immune-modulating activity of MSC with no beneficial for kidney function and histology [33]. However, the different study design (bilateral acute kidney injury model) and infusion of bone marrow-derived MSC into the suprarenal aorta might partly account for the disparity from our study.

The recruitment of exogenous MSC to the injured renal tissue may be mediated by their expression of the multifunctional receptor CD44 [34]. In addition, we have shown that ARAS kidney releases a plethora of injury signals and homing factors that attract circulating progenitor cells [9], which likely also promote adhesion and retention of exogenously delivered MSC. Their subsequent protective effects might be attributed partly to their capacity to engraft in the damaged kidney. Four weeks after their infusion, MSCs were detected in renal tissue sections, mostly at the insterstitium. This observation may partly explain their ability to decrease inflammation, as T cells and macrophages tend to infiltrate the interstitium [35]. Remarkably, treatment with MSC decreased apoptosis and increased proliferation (PCNA staining) of renal glomerular and tubular cells, underscoring for the ability of MSC to increase cellular survival. However, fewer injected MSCs were incorporated into renal tubules and blood vessels (cytokeratin and CD31 costaining, respectively). This relatively minor engraftment of MSC in renal structures, and likely primarily secretion of paracrine factors, may account for the substantial improvement in renal function in the postischemic kidney. In turn, by decreasing rarefaction of interstitial vessels, an important determinant of GFR [36], MSC might have restored renal function beyond the stenotic lesion.

Although ARAS can sometimes lead to proteinuria [37], no differences in urinary protein excretion were found among the groups, and renal expressions of podocin and nephrin were similar, arguing against structural alterations in the glomerular basement membrane in this early stage of the disease. On the other hand, delivery of MSC in conjunction with PTRA decreased the number of sclerotic glomeruli in the stenotic kidney. Downregulation of the fibrogenic factor TGF-β1 might have contributed to the reduction in tubulointerstitial fibrosis and glomerulosclerosis in the stenotic kidney of MSC-treated pigs. Our results are underscored by previous observations showing a renoprotective effect of arterially delivered MSC decreasing renal fibrosis in rats with unilateral ureteral obstruction [38]. Finally, PTRA + MSC might have attenuated fibrosis by decreasing the number of renal apoptotic cells, in agreement with previous observations in experimental models of acute kidney injury [39].

Notably, in our study, allogeneic MSCs were isolated from healthy animals. Autologous cells derived from patients might be susceptible to potential deleterious effects of atherosclerosis, as a recent study showed impaired multilineage differentiation potential of adipose tissue-derived MSC isolated from obese compared with lean subjects [40]. Furthermore, allogeneic MSC may incur a transition from an immunoprivileged to an immunogenic state after differentiation [41], which may limit their beneficial effects. However, although the use of autologous cells remains preferable, no evidence of cellular rejection was found in our allogeneic MSC-treated pigs. Pertinently, even in an atherosclerotic environment, MSCs have retained their capacity to decrease atherosclerotic changes by stimulating endothelial repair, decreasing proliferation of smooth muscle cells, and decreasing thrombosis [42–44].


Our study is limited by the short duration of ARAS, lack of additional comorbid conditions, and use of relatively young animals. However, similar to observations in humans, PTRA alone failed to improve within 4 weeks renal functional compromise, intrarenal inflammation, and fibrosis in the RAS kidney distal to the stenosis. While blood pressure does not consistently decline in patients after revascularization as it did in our experimental model, human ARAS might be superimposed on essential hypertension or pre-existing kidney disease. We also did not examine the independent effects of MSC on the stenotic ARAS kidney without PTRA, which warrant independent and thorough future studies. In addition, a potential bias in the calculation of MSC retention rate is the fact that their in vitro migration capabilities are largely influenced by the systemic and local inflammatory state [45]. The possible proliferation of MSC in the kidney, suggested by some costaining with PCNA, may raise concerns regarding formation of ectopic masses, yet none were observed in this study or in another preliminary study in pigs followed for 3 months after MSC delivery (unpublished data). Future studies are needed to establish the lack of adverse effects and the persistence of the beneficial effects of MSC on the kidney and its response to PTRA over longer periods of time and in humans.


Taken together, our observations have identified a synergistic beneficial role of adipose tissue-derived MSC and PTRA in reversing renal injury induced by ARAS. The beneficial effect of MSC delivered after PTRA appears to be mediated by improvement of microvascular remodeling, reduction of oxidative stress, and inflammation in the stenotic kidney via paracrine mechanisms. Hence, our data indicate that intrarenal delivery of MSC constitutes an effective approach to improve the response to revascularization in swine ARAS. Additional studies are needed to provide further insight into the role of MSC in improving renal function and response to revascularization in this increasingly prevalent disease.


Twenty-eight domestic female pigs (50–60 kg) were studied during 16 weeks of observation (Supporting Information Fig. 1s) after approval by the Institutional Animal Care and Use Committee. At baseline, pigs were randomized into two groups, which were fed either a 2% high-cholesterol diet (n = 21), as a surrogate for early atherosclerosis, or normal pig chow (n = 7). Six weeks later, all animals were anesthetized with 0.5 g of intramuscular ketamine and xylazine, and anesthesia was then maintained with i.v. ketamine (0.2 mg/kg per minute) and xylazine (0.03 mg/kg per minute). RAS was induced in high-cholesterol animals by placing a local irritant coil in the main renal artery, leading to a gradual development of unilateral RAS, as previously described [46]. The rationale for initiating diet before RAS was to reproduce the clinical situation in ARAS in which early atherosclerosis precedes the stenosis [47]. A sham procedure, which involved cannulating the renal artery (without placement of irritant coil), was performed in normal animals. In addition, a telemetry system was implanted in the left femoral artery to continuously measure MAP for 10 additional weeks.

Six weeks after induction of RAS (Fig. 1), animals were similarly anesthetized, and the degree of stenosis was determined by angiography. Fourteen ARAS pigs were treated with PTRA, while the others underwent a sham procedure. Immediately after PTRA, labeled allogeneic adipose tissue-derived MSCs isolated from normal swine were delivered into the stenotic kidney of seven ARAS + PTRA pigs. A vehicle (saline) was administered in other seven ARAS + PTRA pigs. Intrarenal delivery of MSC is supported by low engraftment of injected cells in the kidney after i.v. infusion [48].

Four weeks after PTRA or sham, the pigs were again similarly anesthetized, and the degree of stenosis was determined by angiography. Blood samples were collected from the inferior vena cava for PRA, total cholesterol, triglycerides, HDL, LDL, IL-1β, and creatinine measurements. In addition, urine samples were collected and albumin concentration was quantified by enzyme-linked immunosorbent assay (ELISA) (Bethyl Laboratories, Montgomery, TX, Renal hemodynamics and function in each kidney were then assessed using multidetector computer tomography (MDCT).

Two days after completion of all studies, pigs were euthanized with a lethal i.v. dose of 100 mg/kg of sodium pentobarbital (Sleepaway, Fort Dodge Laboratories, Inc., Fort Dodge, IA, [49]. The kidneys and heart were removed using a retroperitoneal incision and immediately dissected, and sections were frozen in liquid nitrogen (and maintained at −80°C) or preserved in formalin [15] for in vitro studies.

In Vivo Studies

Renal Hemodynamics and Function

Basal regional perfusion, RBF, and GFR in each kidney were noninvasively assessed using MDCT, an ultrafast scanner that provides accurate and noninvasive quantifications of single-kidney volume, hemodynamics, and function [14, 15, 50]. In brief, 160 consecutive scans were performed following a central venous injection of iopamidol (0.5 ml/kg per 2 seconds). The same procedure was repeated after 15 minutes toward the end of a 10-minute suprarenal infusion of Ach (5 mg/kg per minute) to test endothelium-dependent microvascular reactivity [14]. Images were reconstructed and displayed with the Analyze software package (Biomedical Imaging Resource, Mayo Clinic, Rochester, MN, Tissue attenuation curves obtained from cross-sectional images from the aorta, renal cortex, and medulla were fitted by curve-fitting algorithms to obtain measures of renal function [6, 7, 14]. Cortical and medullary volumes were calculated with planimetry.

In Vitro Studies

Renal Morphology and Fibrosis

MSCs were tracked in frozen kidney sections stained with 4′,6-diamidino-2-phenylindole as well as cytokeratin, the endothelial marker CD31, or PCNA. The number of retained MSC (retention rate) was calculated and their distribution was evaluated (Supporting Information). In addition, midhilar 5-μm cross-sections of each kidney (one per animal) were stained with Masson's trichrome and glomerular and peritubular regions were examined using an image analysis program (MetaMorph, Meta Imaging Series 6.3.2, Allentown, PA, In each slide, staining intensity or positivity was semiautomatically quantified in 15–20 fields, expressed as fraction of kidney surface area or number of positive cells, and the results from all fields were averaged [51]. Glomerular score (% of sclerotic out of 100 glomeruli) was also assessed [14, 15]. Glomerular and tubulointerstitial collagen content was evaluated by Sirius red staining as previously described [52]. Slides were visualized under polarized light microscope, and pictures of the entire slice were taken with identical exposure settings for all sections and analyzed using MetaMorph. Results were quantified as percent area staining. In addition, Western blotting protocols were followed in each kidney sample using specific polyclonal antibodies against TGF-β1 (Santa Cruz, Biotechnology (Santa Cruz, CA, 1:200) [14]. All quantifications were performed in a blinded manner.

Oxidative Stress

Systemic levels of isoprostanes were assessed using an enzyme immunoassay kit [53]. Renal redox status was evaluated by the in situ production of superoxide anion, detected by fluorescence microscopy using dihydroethidium [51] and by the expression of the NADPH-oxidase subunit p47 (Santa Cruz, 1:200) and NT (Cayman Chemical (Ann Arbor, MI, 1:200), determined by Western blotting [9, 53].


Renal inflammation was evaluated by standard immunostaining with antibodies against macrophages (antimacrophage CD163), B-T lymphocytes (CD45), and CD3 T lymphocytes. Positive cells were manually counted under ×40 in random glomerular or cortical fields and averaged from 20 fields in each sample in a blinded manner. In addition, renal expression of TNF-α (Santa Cruz, 1:200), MCP-1 (MyBioSource, San Diego, CA, 1:7,500), IF-γ, and NFκB was quantified by Western blot [47, 54].

Systemic inflammation was assessed by IL-1β levels, quantified by ELISA for porcine IL-1β developed using Nunc MaxSorp plate (ThermoFisher Scientific, Waltham, MA,; 437796) coated with 100 μl of IL-1β capture antibody 2.0 μg/ml (R&D Systems, (Minneapolis, MN,; DY681) and 100 μl of pig plasma samples. The plate was read by the SynergyMx plate reader (BioTek, Seattle, WA,, set at standard luminescence reading mode, for 1 hour at 10-minute intervals. The maximum readout was then further analyzed.


The number of apoptotic cells was quantified in renal sections by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) method (Promega, Madison, WI,, as we have previously shown [52].

Glomerular Damage

The degree of glomerulosclerosis was assessed by glomerular score and podocyte injury by renal expression of the podocyte proteins podocin and nephrin (Abcam, Boston, MA,; 1:200).

Microvascular Architecture

Stenotic kidneys were perfused with microfil MV122 (an intravascular contrast agent) under physiological pressure using a saline-filled cannula ligated in a segmental artery. Samples were prepared and scanned, and images were analyzed as previously described [6, 52]. Spatial density, average diameter, and tortuosity of renal cortical microvessels (diameters of 20–500 μm) were calculated [52] using Analyze. In addition, renal expression of VEGF (Santa Cruz, 1:200), VEGFR-2, eNOS, and bFGF (Millipore, (Billerica, MA, 1:1,000) was quantified by Western blot. Finally, glomeruli and tubulointerstitial endothelial cell density were assessed in renal tissue sections stained with the endothelial markers CD31 and vWF (Abcam, cat# ab6994), respectively [55].

Statistical Methods

Statistical analysis was performed using JMP software package version 8.0 (SAS Institute Inc., Cary, NC, The Shapiro–Wilk test was used to test for any deviation from normality. Results were expressed as mean ± SEM for normally distributed variables. Comparisons within groups were performed using the paired Student's t test and among groups using ANOVA followed by the unpaired t test with Bonferroni correction. Data that did not show a Gaussian distribution were expressed as median (range) and comparisons within and among the groups were performed using nonparametric tests (Wilcoxon and Kruskal-Wallis H test, respectively). Statistical significance for all tests was accepted for p ≤ .05.


This study was partly supported by NIH grant numbers DK73608, DK77013, HL77131, HL085307, and UL1-RR024150 and by the American Heart Association.


The authors indicate no potential conflicts of interest.