, Volume 16, Issue 3, pp 609–624

In acute kidney injury, indoxyl sulfate impairs human endothelial progenitor cells: modulation by statin


  • Vin-Cent Wu
    • Department of Internal MedicineNational Taiwan University Hospital
  • Guang-Huar Young
    • Department of SurgeryNational Taiwan University Hospital
  • Po-Hsun Huang
    • Division of Cardiology, Department of Internal MedicineTaipei Veterans General Hospital
  • Shyh-Chyi Lo
    • Department of Laboratory MedicineNational Taiwan University Hospital
  • Kuo-Chuan Wang
    • Department of NeurologyNational Taiwan University Hospital
  • Chiao-Yin Sun
    • Division of NephrologyChang Gung Memorial Hospital
  • Chan-Jung Liang
    • Department of Anatomy and Cell Biology, College of MedicineNational Taiwan University Hospital
  • Tao-Ming Huang
    • Yun-Lin BranchNational Taiwan University Hospital
  • Jou-Han Chen
    • Department of Internal MedicineNational Taiwan University Hospital
  • Fan-Chi Chang
    • Chu-Tung BranchNational Taiwan University Hospital
  • Yuh-Lien Chen
    • Department of Anatomy and Cell Biology, College of MedicineNational Taiwan University Hospital
  • Yih-Shing Kuo
    • Bei-Hu branchNational Taiwan University Hospital
  • Jin-Bor Chen
    • Division of NephrologyChang Gung Memorial Hospital
  • Jaw-Wen Chen
    • Division of Cardiology, Department of Internal MedicineTaipei Veterans General Hospital
    • Department of Internal MedicineNational Taiwan University Hospital
  • Wen-Jo Ko
    • Department of SurgeryNational Taiwan University Hospital
  • Kwan-Dun Wu
    • Department of Internal MedicineNational Taiwan University Hospital
  • The NSARF group
Original Paper

DOI: 10.1007/s10456-013-9339-8

Cite this article as:
Wu, V., Young, G., Huang, P. et al. Angiogenesis (2013) 16: 609. doi:10.1007/s10456-013-9339-8


Renal ischemia rapidly mobilizes endothelial progenitor cells (EPCs), which provides renoprotection in acute kidney injury (AKI). Indoxyl sulfate (IS) is a protein-binding uremic toxin with a potential role in endothelial injury. In this study, we examined the effects and mechanisms of action of IS on EPCs in AKI. Forty-one consecutive patients (26 male; age, 70.1 ± 14.1 years) diagnosed with AKI according to the AKIN criteria were enrolled. The AKI patients had higher serum IS levels than patients with normal kidney function (1.35 ± 0.94 × 10−4M vs. 0.02 ± 0.02 × 10−4M, P < 0.01). IS levels were negatively correlated to the number of double-labeled (CD34+KDR+) circulating EPCs (P < 0.001). After IS stimulation, the cells displayed decreased expression of phosphorylated endothelial nitric oxide (NO) synthase, vascular cell adhesion molecule-1, increased reactive oxygen species, decreased proliferative capacity, increased senescence and autophagy, as well as decreased migration and angiogenesis. These effects of IS on EPCs were reversed by atorvastatin. Further, exogenous administration of IS significantly reduced EPC number in Tie2-GFP transgenic mice and attenuated NO signaling in aortic and kidney arteriolar endothelium after kidney ischemia–reperfusion injury in mice, and these effects were restored by atorvastatin. Our results are the first to demonstrate that circulating IS is elevated in AKI and has direct effects on EPCs via NO-dependent mechanisms both in vitro and in vivo. Targeting the IS-mediated pathways by NO-releasing statins such as atorvastatin may preempt disordered vascular wall pathology, and represent a novel EPC-rescued approach to impaired neovascularization after AKI.


Indoxyl sulfateStatinEndothelial nitric oxide synthaseEndothelial progenitor cellsAutophagy



Acute kidney injury






Chronic kidney disease


Activated caspase 3




DNA fluorochrome 4′-6-diamidine-2-phenyl indole


Dichlorofluorescin diacetate


Endothelial nitric oxide synthase


Endothelial progenitor cells


End stage renal disease


Fluorescein isothiocyanate


Generalized additive model


Intercellular adhesion molecule 1


Indoxyl sulfate


Light Chain 3


3-(4,5-cimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide




Nitric oxide


Reactive nitrogen species


National Taiwan University Hospital Study Group on Acute Renal Failure


Organic anion transporter


Reactive oxygen species


Sodium nitroprusside




Transforming growth factor


Terminal deoxyribonucleotidyl transferase (TDT)-mediated dUTP-digoxigenin nick end labeling


Vascular cell adhesion molecule


Von Willebrand factor


Acute kidney injury (AKI) is a serious complication in the hospital, resulting in a prolonged hospital stay and heightened mortality [1]. Renal recovery from AKI, depending on the severity and duration, is often not achieved due to concomitant nonrenal organ injuries. Further, the retained nitrogenous and non-nitrogenous waste products may cause serious metabolic disturbances and impair the recovery during the repair phase [2].

Indoxyl sulfate (IS) is synthesized in the liver from indole, a metabolite of tryptophan produced by the intestinal flora [3]. Recent studies have shown that IS was associated with endothelial dysfunction [4, 5], induction of oxidative stress by modification of the balance between pro- and antioxidant mechanisms [6] and associated with increased cardiovascular events [7, 8], in term of cardiorenal syndrome [9]. IS-increased endothelial injury through increased oxidative stress [6], impaired nitric oxide (NO) metabolism or availability [4], and then resulted in blunting of the endothelial healing ability [5]. IS has been identified as a potential vascular toxin in AKI mice and has been reported to increase 15 times more than sham control [10].

Hypoperfusion of the kidney affects the function and structure of tubular epithelial cells and peritubular capillary endothelial cells, which inhibits the process of post-ischemic renal reperfusion (I/R), thereby prolonging kidney dysfunction [11, 12]. In contrast to the proximal tubules, the renal vasculature lacks efficient regenerative capacity and mesenchymal transition, leading to a persistent 30–50 % reduction in vascular density following I/R [13, 14]. The loss of renal microvessels following AKI and the resultant impairment of vascular remodeling processes, if progressive irreversible, will lead to the development of progressive chronic kidney disease [14].

The extent of endothelial injury, which predicts cardiovascular events, represents a balance between the magnitude of injury and the capacity for repair [15, 16]. There is growing evidence that endothelial progenitor cells (EPCs) may improve vascular regeneration in ischemic organs. Recent data suggest that after AKI, EPCs are mobilized and recruited to ischemic kidney areas to ameliorate AKI through restoration of injured microvasculature [17]. In ischemic AKI, which is characterized by hypoperfusion of peritubular capillaries, renal function can be preserved by systemic administration of EPCs [18]. It has been reported that EPCs act by incorporating into sites of neo-vascularization and endothelial denudation in murine models of ischemic AKI [19, 20].

In patients with AKI, the retained uremic toxins are shown to impair EPC mobilization and the resultant accumulation of EPCs is considered renoprotective [21]. It is not known, however, if EPCs are functionally competent under uremic milieu of AKI. In the present study, we hypothesize that sudden retention of IS in AKI is associated with low EPC numbers and impairs EPC function, which can be attenuated by antioxidative treatment and the NO–releasing-moiety-containing statins.



All patients were registered in the National Taiwan University Hospital Study Group on Acute Renal Failure (NSARF) between Feb 2009 and Jan 2011. The database was constructed for quality assurance in 1 medical center (National Taiwan University Hospital in Taipei, Taiwan) and its 3 branch hospitals in different cities [2225].

AKI patients in the present study were consecutively enrolled when AKI was diagnosed. The definition of AKI was based on the criteria established by Acute Kidney Injury Network (AKIN) [26]. Briefly, AKI was defined as serum creatinine elevated ≥1.5 times the baseline value, absolute elevation of ≥0.3 mg/dL, or urine output of <0.5 mL/kg per hour for ≥6 h in the first 48 h post-operation. In the NSARF database, serum creatinine and urine output were recorded daily after surgery [22]. We excluded patients who had received acute dialysis or had a baseline kidney function less than 45 mL/min/1.73 m2, defined as chronic kidney disease [27], were excluded. Ten patients with end-stage renal disease (age, 69.4 ± 13.9 years, dialysis period, 4.5 ± 2.5 years, and average KT/V, 1.5 ± 0.18) were enrolled, and 10 volunteers with kidney function >60 mL/min/1.73 m2 were enrolled as controls.

This study was approved by the Institutional Review Board of the National Taiwan University Hospital, Taipei, Taiwan (ClinicalTrials, NCT00451373). All participants provided written informed consent before inclusion in the study.

Human EPC isolation and cultivation

Peripheral blood samples (40 mL) were obtained from patients and volunteers as our previously description [2830], and total mononuclear cells (MNCs) were isolated by density-gradient centrifugation with Ficoll-Paque™ PLUS (GE Healthcare). After a 4–day culture, non-adherent cells were removed, and the attached, elongated, spindle-shaped cells were identified as early EPCs [28, 29, 31]. The medium was changed every 2 days. A certain number of MNCs were allowed to grow continuously into colonies of late EPCs, exhibiting a cobblestone morphology and monolayer growth pattern typical of mature endothelial cells at confluence [31]. The reagents used were sodium nitroprusside (SNP; Sigma-Aldrich); ROS scavenger; N-acetylcsyteine (NAC; Merck) and NADPH oxidase inhibitor; and Apocynin (Apo; Sigma-Aldrich). Atorvastatin (ATO) was a gift from Pfizer Inc. (New York, NY, USA).

Flow cytometry analysis

A volume of 1,000 μL peripheral blood was incubated for 30 min in the dark with 0.1 μg fluorescein isothiocyanate (FITC)-conjugated monoclonal antibodies against human CD34 monoclonal antibodies (R&D, Minneapolis, MN, USA) and 0.1 μg allophycocyanin (APC)-labeled monoclonal antibodies against human VEGF R2/KDR (R&D, Minneapolis, MN, USA). Each analysis included 1 × 105 events according to our protocol using a FACS Calibur™ flow cytometer (BD Bioscience) [28]. As shown in supplementary Figure 1, the numbers of circulating EPCs were gated among monocytes and defined as CD34+KDR+ and expressed as percentage.

To assess the reproducibility of EPC measurements, circulating EPCs were measured from 2 separate blood samples in 10 subjects, and there was a strong correlation between the two measurements (r = 0.82, P < 0.001) [31]. The coefficient of variation (CV) % of 20 times repeated measurement were 6.2 %.

For the cell cycle analysis, the cells were stained with 50 μg/ml propidium iodide and 40 μg/ml RNase A for 30 min in the dark. Cell cycle distribution was determined by flow cytometric analysis (Becton–Dickinson FACScan) and Cell Quest software (Becton–Dickinson).

Immunohistochemical analysis

For early EPCs, we grew cells on fibronectin-coated 8-chamber slides and performed double-positive staining with 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate-acetylated low-density lipoprotein (Dil-ac-LDL; Molecular Probe) and FITC-labeled lectin from Ulex europaeus agglutinin (UEA-1; Sigma-Aldrich) [2830]. The late EPC-derived outgrowth of endothelial cells was characterized by the expression of Von Willebrand factor (vWF; AbD, SeroTec), VE-cadherin (Cell Signaling), and KDR/VEGF receptor 2 (Cell Signaling) [30]. Secondary antibodies were conjugated with DyLight 488 or DyLight 594-labeled (Jackson Laboratories). After counterstaining with DAPI (Sigma-Aldrich), cell images were captured with a confocal laser-scanning microscope (LSM510, Zeiss).

Proliferation assay

The viability of EPCs was determined by counting the positive cells in 8 random high-power microscopic fields (100×), and by the 3-(4,5-dimethylthiazol-2-yl)- 2,5,diphenyltetrazolium bromide (MTT) assay[28, 30]. Cumulative population doubling at each passage was calculated from the cell count with the following equation: 24 NH/N1 = 2× or [log10(NH) − log 10 (N1)]/log10(2) = X where N1 is the inoculum number, NH is the cell harvest number, and X is population doublings. The population doublings were added to yield cumulative population doubling [32].

Apoptosis and autophagy assay

In situ EPC apoptosis was determined by a TUNEL kit (fluorescein-TUNEL; Roche Applied Science) for apoptotic nuclei. The antibodies against apoptosis were caspase 3 (Imgenex) and BCL2 (Epitomic); autophagy was light chain 3b (LC3b, Cell Signaling Technology).

Directed in vivo angiogenesis angioreactor

The in vivo EPCs function was determined by directed angiogenesis angioreactor [33],incorporating with EPCs. The angioreactor was quantitated by intravenous injection of 100 μL of Lectin-FITC (1 mg/ml in 10 m M HEPES (4-(2- hydroxyethyl)-1- piperazineethanesulfonic acid), 150 m M NaCl, pH7.5; Sigma, St Louis, MO, USA) 5 min before killing of the mice. The angioreactors were trimmed away, dispased, digested, and images captured by fluorescence spectroscopy.

Migration assays

For migration assays, a uniform cell-free area was created by scratching confluent monolayers with a plastic micropipette tip. The wound area was inspected at different time intervals (0–16 h) to determine the area percentage traversed by the cells. The larger the area, the faster the cell migrated. At each time point, 4 photographs were taken and the widths of the wound space were evaluated by multiple measurements [31].

Angiogenic growth factors assays

NOx production was measured by nitrite and nitrate ELISA kit (NOx; R&D Systems) and vascular endothelial growth factor ELISA kits (sVCAM-1; R&D Systems) according to the manufacturer’s instructions.

Reactive oxygen species assays

To evaluate cellular levels of ROS, EPCs in 96-well plates were pretreated with IS (10−3 M) for 3 h. After the removal of IS from wells, cells were incubated with peroxide-sensitive probe 2′7′-dichlorofluorescein diacetate (DCFDA; Invitrogen) according to the manufacturer’s instructions. The fluorescence intensity (relative fluorescence units) was measured at 485-nm excitation and 530-nm emission by using a fluorescence microplate reader. Cellular oxidative stress was detected using the cell-permeable fluorogenic probe CellROX (Molecular Probes), and the image was captured with a confocal laser-scanning microscope (LSM510, Zeiss).

RNA analysis

EPCs were pooled into RNAprotect Cell Reagent (QIAGEN) and total RNA was extracted by the RNeasy Mini Kit (QIAGEN). Reverse transcription was performed using 2 μg of total RNA, and real-time RT-PCR was performed with a FastStart Universal SYBR Green Master mix (Roche) by using an ABI 7900 system and SDS2.3 software (Applied Biosystems). All primer sets are listed in supplementary data. Relative transcript levels were obtained with normalization to GAPDH transcript levels.

Western blot analysis

EPCs were lysed in RIPA buffer (Genestar). Equal amounts of total proteins were separated by SDS-PAGE and transferred onto a PVDF membrane (Millipore).

AKI mice model

IS infusing AKI mice

C57BL/6 mice (20 ± 2 g body weight or 8–9 weeks of age) were briefly anesthetized with anesthesia cocktails (80 mg/kg ketamine plus 100 mg/kg xylazine) i.p. injection, and unilateral AKI was induced by complete ligature of the left renal artery) for 40 min with a microaneurysm clamp. The body temperature of the mice was maintained at 37 °C by a heating pad.

Mice were randomly assigned to 4 groups as follows. (1) Control group (n = 10): sham-operated mice were subjected to the same surgical procedure except that the left renal pedicles were not clamped. (2) AKI group (n = 10): unilateral renal I/S injury was induced in mice. (3) Indoxyl sulfate group (n = 10): after unilateral AKI surgery, mice received intraperitoneal injection with IS (Sigma) at a dosage of 100 mg/kg/day for 3 days as previously report [34]. (4) Indoxyl sulfate with ATO group (n = 20): intraperitoneal injection with IS, as in group 3, with addition of oral gavages with ATO at a dosage of 10 mg/kg for 3 days. I/R model had been reported to show the elevation of serum IS at least until 48 h after reperfusion without IS injection; [10] therefore IS was expected to accelerated the pathophysiological change and it’s serum level was relative high by daily injection in our model [34]. This model might lessen confounding factors from endogenous toxins and allow us to focus more on the effects of exogenous IS effect on EPCs.

The numbers of circulating EPCs form bone marrow, spleen, and peripheral blood were measured. The thoracic aorta was dissected out, immersion-fixed in 4 % buffered paraformaldehyde, paraffin-embedded, and cross-sectioned for immunohistochemistry. Two serial sections were examined by immunostaining for TGF-β1 (Genetex) or eNOs (Upstate).

Donor transgenic Tie2-GFP mice

Bone marrow (BM) chimeric mice were generated as previously described [35]. In brief, ten million BM cells from Tie2-GFP male donors in 200 μl of PBS were injected into the lateral tail vein of lethally irradiated (1,000 Rads over 30 min) isogenic female recipients. Chimerism was confirmed after 6 weeks by quantitative duplex polymerase chain reaction using specific detection probes for Y-chromosome Y6 and autosomal GAPDH primer amplicons [36].

All experimental procedures and protocols involving animals were in accordance with the local institutional guidelines for animal care, and approved by the institutional animal care committee of National Taiwan University (Taipei, Taiwan), and complied with the Guide for the Care and Use of Laboratory Animals (NIH publication No. 86-23, revised 1985).

Statistical analysis

Differences between 2 or more groups were analyzed using a 2-tailed Student’s t test. P < 0.05 was considered statistically significant. Statistical analyses were performed with R software, version 2.8.1 (Free Software Foundation, Inc., Boston, MA, USA) and SAS software, Version 9.1.3 (SAS Institute Inc., Cary, NC, USA). A normal distribution was obtained by appropriate transformations of skewed variables, such as the percentage of circulating EPCs. To display the implications of EPCs for individual patients, a generalized additive model (GAM) (with spline) incorporating subject-specific (longitudinal) random effects was plotted, adjusted for the sex and age of the patient, SOFA score, and serum potassium at the diagnosis of AKI. We used a transformation of the β coefficients [100 * (eβ − 1)] to obtain the percentage increase of circulating EPCs per SD increase in IS. This approach permitted adjustments for possible nonlinear effects of the continuous variables [37, 38].


Decreased circulating EPC number correlated to IS serum level in AKI patients

Forty-one consecutive patients (26 male; age, 70.1 ± 14.1 years) were enrolled at diagnosis of postcardiac surgery AKI. There were 17 patients in stage 1 (mean creatinine level, 1.7 ± 0.2 mg/dL), 12 patients in stage 2 (mean creatinine level, 2.3 ± 0.17 mg/dL), and 12 patients in stage 3 (mean creatinine level, 4.1 ± 0.9 mg/dL) according to AKIN criteria. IS was measured by ultra performance liquid chromatography (UPLC) system (supplementary methods). The average disease severity, as determined by the SOFA score at enrollment, was 9.5 ± 4. When AKI patients (1.35 ± 0.94 × 10−4 M) were compared with 10 volunteers with normal kidney function and 10 end-stage kidney disease patients, IS was the highest in ESRD (1.87 ± 0.51 × 10−4 M) and the lowest in patients with normal kidney function (0.02 ± 0.02 × 10−4 M, P < 0.01; Fig. 1a). At AKI, 5 patients received statin treatment. There were a significantly higher percentage of CD34+KDR+ EPCs in AKI patients with statin therapy than those not taking statin (−3.13 ± 0.58 vs. −4.05 ± 0.57 %, P = 0.014).
Fig. 1

Elevated indoxyl sulfated (IS) and negative correlated with circulating endothelial progenitor cells (EPCs) in AKI patients. a Comparison the serum IS level of AKI patients (n = 41) (IS, 1.35 ± 0.94 × 10−4 M) with 10 volunteers with normal kidney function (IS, 0.02 ± 0.02 × 10−4 M) and 10 end-stage kidney disease patients (IS, 1.87 ± 0.51 × 10−4 M). Serum IS level was the highest in ESRD and the lowest in patients with normal kidney function (**P < 0.001 Data are expressed as mean ± SEM). b Generalized additive model (GAM) with spline was plotted to predict the association of IS and EPCs [(Log [(CD34+KDR+) cells (%))] in patients with AKI. The plot was incorporated with the subject-specific (longitudinal) random effects, including age, sex, SOFA score, and creatinine level to predict the association. The curve was centered to have an average of zero over the range of the data. The dashed lines indicate an approximated point-wise 95 % confidence interval. Note that after the cut-off point of 2.4 × 10−4 M, IS was nearly linearly negatively correlated to EPC number

To explore the effect of IS on the number of circulating EPCs in AKI, multivariate GAM regression was plotted after adjusting the co-variables, SOFA score, age, sex, and creatinine level when AKI was diagnosed. The IS serum level was inversely correlated to the number of circulating EPCs, as defined by the number of double-labeled cells (CD34+KDR+; P = 0.006; Fig. 1b) in multivariate model. The apparent turning point of IS concentration negatively related to EPC number was 2.4 × 10−4 M.

Effects of IS on EPC proliferation and senescence

To elucidate the potential mechanisms whereby IS acts to decrease EPC number, early EPCs as characterized by positive double staining with Dil-ac-LDL and UEA-1 were incubated with different concentrations of IS for 3 days in vitro. The results showed that the numbers of early EPCs were significantly reduced in a dose-dependent manner (Fig. 2a). The organic anion transporters (OATs) 1 and 3 were not present in the EPCs, as confirmed by PCR (supplementary figure). Therefore, there was no OAT-mediated response.
Fig. 2

EPC proliferation inhibited by IS and rescued by ATO. a Dose-dependent effect of indoxyl sulfate (IS) on early EPC number. Early EPCs, double positive for Dil-acetylated LDL and FITC, were incubated in media with and without different concentrations of IS for 5 days. IS affected the number of EPC identified by fluorescence microscopy (*P < 0.001 vs. untreated control). Scale bars 50 μm. b Early (left panel) and late EPC were incubated with various concentration of IS for 3 days, and cell viability was detected by MTT assay. c Late EPCs were pre-incubated with agents for 30 min and then IS was added (5 × 10−4, 1 × 10−3, and 5 × 10−3.) for 3 days. The viability of late EPCs attenuated by IS improved after the addition of ATO, and NAC. However, the effect of probenecid, inhibitor of OAT, on EPCs viability was not apparent. Cell viability was determined by MTT assay, and data are presented as percentage of viability (*P < 0.05; #P = not significant vs. untreated control). Data are expressed as mean ± SD, n = 6 independent experiments

EPC proliferation analyzed by the MTT assay also showed a dose–response type decrease in response to IS stimulation, both in early and late EPCs (P < 0.001 vs. control; Fig. 2b and supplementary figure). The quantitative results of cell cycle analysis confirmed the findings. The cell cycle analysis showed that in late EPCs, IS treatment resulted in significant blockade of the G0/G1 transition after 1-day infusion, and the effect was inhibited by NAC (vs. control; P < 0.01; supplementary figure). This indicates that IS may slow the growth of EPCs by artificially imposing a cell cycle checkpoint. The effect of IS on EPCs appears to be dose-dependent. However, the viability of late EPCs attenuated by IS was blunted by adding ATO (10−7 M), but not oxidative stress scavengers or probenecid, an inhibitor of OAT (Fig. 2c).

Effects of IS on EPC senescence

EPCs cultured in IS 10−3 M medium for 4 days showed dose-dependent increases in the percentage of senescence-associated β-galactosidase-positive EPCs (P < 0.001; Fig. 3a). Cumulative population doubling of EPCs decreased after adding IS, and administration of ATO attenuated the IS-induced senescence.
Fig. 3

Amelioration of IS induced EPC senescence and autophagy by ATO. a Dose-dependent increase in senescence-associated β-galactosidase-positive EPCs after culture in IS medium for 4 days (left panel). ATO (5 × 10−7 M) ameliorated IS induced senescence (***P < 0.001; **P < 0.01 vs. untreated control). Cumulative population doubling decreased after adding IS and restored by ATO. b TUNEL+ apoptotic cells (green fluorescence) in late EPCs did not increase after IS (10−3 M) infusion, like when there was no IS. EPCs treated with DNase were used as positive control. Nuclei were stained with DAPI. c Activated caspase 3 (CASP3) was not detected after IS infusion even after more than 5 days. The expression autophagy marker Light chain 3 (LC3) and antiapoptosis marker BCL 2 were dose-dependently increased after IS stimulation. The upregulated LC3 and BCL 2 were ameliorated by ATO. d The distribution of LC3B was determined by confocal fluorescent microscopy. IS treatment induced a characteristic pattern of LC3B, indicating increased autophagy, and produced an increase in fluorescence intensity in vWF (+) progenitor cells as compared to the control cells. DAPI was the nuclear counterstain. The intensity dose- dependently increased but was attenuated by ATO pretreatment. e Aggregated accumulation and progression of autophagosomes after IS 10−3 M stimulation. After IS simulation for 24 h, there were many more vacuoles, often filled with fibrillar electron-dense material. These areas sometimes contained membrane-rimmed vacuoles on electron micrographs. After coincubation in IS with ATO or NAC, autophagosomes decreased

Effects of IS on EPC apoptosis and autophagy

IS, in either a low or high dose, did not affect the rate of apoptosis in EPCs, and in situ detection of fragmented DNA (TUNEL assay) failed to discriminate for apoptosis (Fig. 3b). The presence of activated caspase 3 did not change after IS infusion for more than 5 days (Fig. 3c). However, as seen by immunoblots in Fig. 3d, the standard autophagy marker LC3b (P = 0.035) and the antiapoptosis marker BCL 2 (P = 0.041) were dose-dependently increased after IS stimulation. The accumulation of fluorescent LC3b dots was increased by IS stimulation and attenuated by ATO treatment (Fig. 3d). These events occurred in association with upregulation of LC3b messenger RNA quantified by real-time PCR (Supplemental figure). This was consistent with electron microscopic analysis indicating that the formation of autophagosomes increased after IS stimulation but decreased by pretreatment with ATO or NAC (Fig. 3e).

IS induced oxidative stress in EPC

IS was added to EPCs and resulted in significant increase in intercellular oxidative stress at 1–3 h expressed by fluorescent intensity (P < 0.01; Fig. 4a). ROS were detected by the incorporation of a fluorogenic probe (ROX) in live EPCs exposed to IS. Activation of the ROX probe, which results in bright red fluorescence, was dramatically increased after stimulation with IS but blunted by ATO and NAC (Fig. 4b).
Fig. 4

Effect of IS on oxidative stress in EPCs. a High IS markedly increased ROS production determined by relative DCF-DA fluorescence intensity. Cells were stimulated with IS and increased ROS in time-dependent (left panel) and dose-dependent manners. (*P < 0.05, **P < 0.01; ***P < 0.001 vs. untreated control). b Intracellular ROS production was dramatically increased after IS (5 × 10−4 M) infusion but blunted by ATO (10−7 M). ROS were detected by a cell-permeable fluorogenic probe (5 μmol/L) that emitted red fluorescence upon oxidation by the ROS (Molecular Probes CellROX™ Deep Red reagent), and nuclei were counter- stained with DAPI. Scale bar 50 μm

IS attenuated eNOs activation and NO production

The expression of eNOs and p-eNOs at Ser1177 was significantly downregulated in EPCs after incubation with 10−3 M IS, (decreased p-eNOs, 68 %, Supplemental figure). Upregulation of p-eNOs was noted after administration of ATO (Fig. 5a), APO and SNP at 24 h (Fig. 5b). The marker of reactive nitrogen species (NOx) assayed from conditional medium showed downregulation after IS simulation (11 % enhancement compared with the control group, P < 0.001). (Supplemental figure) We further found in our I/R model that the reduced serum NOx caused by IS injection could be restored by prefeeding the mice with ATO (Fig. 5c).
Fig. 5

IS decreased eNOs and ameliorated by ATO. a ATO (10−7 M) significantly ameliorated p-eNOs expression after IS (10−3 M) stimulating late EPCs for 24 h. The blots show representative immunoblots of the phosphorylation of eNOs (p-eNOs), and the graphs show the quantification of the bands by densitometry. b Restoration of p-eNOs by APO, SNP and ATO after IS stimulation. c Derived production of NOx was assayed from conditioned medium of late EPCs by ELISA assay (Nox; R&D Systems) and showed a dose-responsive downregulation after IS stimulation and was restored by ATO. d Serum NOx was assayed by ELISA assay from I/R mice receiving IS injection and pre-fed with ATO or not. mRNA Data were normalized with GADPH so that the value of the control group was regarded as 1 and expressed as mean ± SEM (n = 4). *P < 0.01 (vs. untreated control) #P = not significant

Effects of IS on EPC vasculogenesis and wound healing

In vivo matrigel-filled implantable angioreactors (directed in vivo angiogenesis assay), incorporating with late EPCs was used to investigate the effect of IS on EPC. Control angioreactors, which contained only growth factor-depleted matrigel, revealed no capillary ingrowth. The EPCs-containing angioreactors showed upregulated angiogenesis; however, this was blunted by IS infusion and restored by ATO feeding within 9 days of implantation (Fig. 6a).
Fig. 6

IS induced EPC dysfunction. a In vivo directed angiogenesis assay, incorporating with late EPCs showed the effect of IS on EPC neovascularizaiton. Accompanied with the onset of angiogenesis, vascular endothelial cells proceed to grow and form vessels in the angioreactor. Angioreactors were prepared as described and recovered after 9 days [33]. The florescence units, counting by Lectin positive cells, increased after EPC incorporated however decreased by IS injection and then restored by feeding with ATO. b Effects of IS (10−3 M) and interaction with ATO, SNP, NAC, and APO on healing scratch wounds in cultured late EPCs. The wounded cultures were allowed to re-epithelialize for 16 h at 37 °C in the presence of VEGF. Results are presented as extent of healing (total covered as 100 %). c Real-time PCR was performed to quantify the RNA transcripts of cell adhesion molecules. After IS simulation, the mRNA expression of integrin α4, β4, and VCAM-1 decreased in EPCs. Cell function data are expressed as mean ± SEM, n = 4, *P < 0.05 vs. control, #P = not significant

An in vitro angiogenesis assay also confirmed this result. The functional capacity for tube formation in late EPCs was significantly reduced after incubation with IS (10−3 M). However, the capacity for tube formation was improved by adding ATO (10−6 M) and SNP (5 × 10−4 M). Moreover, the effect of the NO supply from ATO was functionally confirmed by adding the NO inhibitor L-NAME (10−3 M) to blunt the angiogenesis effect (Supplemental method and figure).

A similar result was obtained by the scratch wound-healing assay. Only 81.2 % of the scratch wound was covered after infusion with IS when compared to the control after 6 h (P < 0.01). In the presence of ATO, SNP, and APO, the wound was almost completely healed (significant increase compared with IS infusion; all P < 0.01; Fig. 6b). The integrin α4β1–VCAM-dependent cell–cell attachment is reported to promote the function of endothelial cells during angiogenesis [39] and to mediate EPC recruitment [40]. Our in vitro findings of real-time PCR indicate that after IS simulation, the expression of integrin α4, β1, and VCAM mRNA levels decreased in EPCs (Fig. 6c).

Effects of IS on EPC mobilization in vivo

To investigate EPC-like cell mobilization after I/R and high IS stimulation in mice, levels of Sca-1+/Flk-1+ cells were determined by flow cytometry and the model of high IS and AKI mice was used. The uninephrectomized B-6 mice without contralateral nephrectomy received the uremic toxins IS had significant higher total and free IS than control as our previously report [34]. After reperfusion, the number of EPC-like cells from the peripheral blood (PB) and bone marrow (BM) of I/R mice at 3 days increased significantly when compared to the sham-operated controls (all n = 25; P = 0.047; P = 0.038, respectively). However, the upregulation of EPC-like cells was significantly blunted in I/R mice receiving IS injection in both PB (P = 0.162 vs. control mice) and BM (P = 0.556 vs. control mice) (Fig. 7a).When the IS-treated AKI mice were fed with ATO (10 mg/Kg) for 3 days, the number of EPCs was restored in both PB (IS vs. ATO, P = 0.037) and BM (IS vs. ATO, P = 0.031; Fig. 7a).
Fig. 7

EPC mobilization impaired by IS was restored by statin in vivo. a Effects of IS on EPC mobilization in I/R mouse model. After reperfusion, the EPC-like cells from the peripheral blood and bone marrow at 3 days increased significantly when compared to the cells from sham-operated animals with or without feeding with ATO. However, the increase in the number of EPC-like cells was significantly attenuated in AKI mice receiving IS injection. When the IS-treated AKI mice were fed with ATO (10 mg/kg) for 3 days, the reduced EPCs were restored both in peripheral blood and in bone marrow (*P < 0.05, vs. Sham control). b The mRNA expression of LC3B increased while that of eNOs and VCAM-1 decreased after IS infusion. This was ameliorated by ATO pretreatment. The serum level of VCAM-1 significantly decreased after IS infusion and was restored after ATO pretreatment by ELISA (*P < 0.05 vs. AKI mice) c Histochemical staining shows that IS blunted the upregulation of eNOs in the aortic root and kidney arterioles after I/R. The blunted eNOs expression in the vessels was restored after ATO pretreatment. In contrast, the IS-augmented TGF-β in vessel wall was reversed by ATO pretreatment. The lumen is uppermost in all sections. Bar 25 μm (*P < 0.05 vs. AKI mice). The lower panel of immunoblot shows that the IS-blunted upregulation of renal p-eNOs was restored by ATO pretreatment. Data are expressed as mean ± SEM, n = 6 independent experiments. d Effect of IS and statin on the recruitment of bone marrow-derived cells into vascular lesions after I/R. Representative cross sections through lesions at the kidney arteriole of Tie2-GFP transgenic mice. Double immunofluorescence analysis confirmed the co-localization of GFP+ (green signal) and CD34+ (red signal), and DAPI-positive (blue) cells. The GFP+ CD34+ endothelial cells in transgenic mice were increased after AKI, decreased after IS infusion and ameliorated by feeding mice with ATO. Magnification, ×50 uM. (Color figure online)

eNOs and VCAM are the critical mediators for EPC migration into ischemic tissue [4, 41]. In IS-treated AKI mouse kidneys, eNOs and VCAM-1 mRNAs were decreased, while LC3b mRNA was increased, as shown by real-time PCR. These IS-induced changes of gene expression were reversed by ATO pretreatment (Fig. 7b). Similarly, the elevated serum creatinine and decreased VCAM-1 levels after IS infusion were restored by ATO pretreatment (Fig. 7b). To determine the effect of IS on vascular endothelial eNOs expression during AKI in vivo, the mouse thoracic aorta and renal arterioles were stained after treatment. As shown in Fig. 7c, IS blunted the upregulation of eNOs and p-eNOs in the vasculature after I/R, and this phenomenon could be restored by ATO pretreatment. In contrast, IS-augmented TGF-β expression in vessel wall, which was reversed by ATO pretreatment.

In sections examined by confocal microscopy, the bone marrow derived GFP+CD31+ cells in donor transgenic Tie2-GFP mice were increased after AKI, decreased after IS infusion and restored by feeding mice with statin. (Fig. 7d) The GFP- and CD31-positive cells were found in the kidney arteriolar endothelium. Importantly, very few endothelial cells expressing GFP could be seen in the contralateral kidneys.


One of the novel findings in this study is that the level of the protein-binding uremic toxin IS was elevated almost 50-fold during AKI. The influence of IS was confirmed by the observation that high serum IS inhibits endothelial proliferation and wound repair in uremia [5]. Uremic toxins have been implicated as the missing link between AKI, tubular and endothelial injury, and nonrenal organ failure [42]. Our results showed that IS was nearly inversely related to the EPC number above 2.4 × 10−4 M, a concentration that is nearly 70 times higher than the reported normal range [42]. This effect of IS was further supported in Tie2-GFP chimeric mice after AKI, where infusion of IS appreciably attenuated the recruitment of EPCs to the kidney arterioles during recovery phase. We found that suppression of EPCs by IS in AKI results from an imbalance in favor of increased injury, decreased cell number, and decreased repair capacity, thus altering the vascular competence of AKI patients.

IS-induced LC3b protein and autophagosomes could be restored by ATO, an NO supplier, or an oxidative scavenger. These findings demonstrate that statins or NAC can inhibit autophagy, suggesting that the NO system and oxidative stress may potentially contain therapeutic targets for recovery from AKI.

EPCs showed decreased expression of NO and VCAM, decreased proliferative capacity, decreased angiogenesis, increased ROS and cellular aging, and defects in migration after IS stimulation. In this study, we found that hyper IS was able to quench NO and increase the generation of oxidant species. We have previously shown that NO is a key molecule for EPC differentiation, survival, and function [30]. Further, NO is known to be a crucial regulator of endothelial cell homeostasis and angiogenesis [43, 44] and is also implicated in the cardio-protective action of statins in the setting of ischemia [45]. eNOs produces NO within the differentiated endothelial cells and governs EPC migration, proliferation, and sprouting or tube-forming [28, 46].

We showed that in I/R mice, IS attenuated the eNOs expression in the endothelium of arteries and the ischemic kidney in terms of decreased EPC mobilization and that it could contribute to impaired reendothelialization. Additionally, this attenuating effect of IS was ameliorated by statin treatment. These findings thus establish an additional mechanism by which NO may specifically preempt disordered vascular wall pathology and augment angiogenesis during AKI recovery.

The production of ROS in endothelial cells, which is stimulated by IS, may also contribute to this inflammatory reaction during AKI [6]. In this study, the total oxidative stress from EPC increased after IS stimulation in vitro. IS could inhibit EPC function through ROS, and the NADPH oxidase inhibitor, apocynin, could reverse IS-attenuated eNOs activity. These findings are reminiscent of the notion that the uncoupling eNOS-derived superoxide amplifies the oxidative stress initiated by NADPH oxidase activation and leads to a futile cycle of ROS production and the creation of an inflammatory and prothrombogenic environment in the microvasculature [47]. The NO donors, ATO and SNP, could further up-regulate the phosphorylation of eNOS after IS stimulation and this result could be seemly to endothelial cells activated by SNP via the NO–AMPK–eNOs–NO pathway [48]. Therefore, oxidative stress was enhanced under uremic milieu, which induced EPC impairment at least partly through the activation of NADPH oxidase.

Statins have been shown to aid progenitor cell functions in vivo and in vitro [49]. These agents, which bear an NO-releasing moiety, have come into the focus for their enhanced anti-inflammatory/anti-thrombotic properties [50]. Some have suggested that statins can potentially decrease endothelial injury through lowering of oxidative stress and inflammation as well as improvement of endothelial repair via recruitment and survival of EPCs [51]. The present study found that when IS-infused AKI mice are treated with statin, the downregulated expression of vascular eNOs and renal VCAM-1 are enhanced, and the adhesion and transendothelial migration of EPCs are amplified.These data add new evidence to indicate that statins help endogenous vascular repair during AKI recovery by aiding the function of EPCs through the activation of the eNOs pathway [52].

The expression of the adhesion molecules VCAM and integrin α4β1 is also attenuated by IS in EPCs. Vascular adhesion molecules play an important role in the recruitment of circulating leukocytes to the endothelium [53]. EPCs are recruited to ischemia-damaged muscle to reconstitute a functional microcirculation, and integrin α4β1 and VCAM-1 facilitate a critical cell–cell adhesion event required for the survival of endothelial cells during neovascularization [39]. Decreased expression of β1 integrin has been implicated in decreased competence of EPC renoprotective mobilization in the setting of AKI [2].

It’s worthy of mention the limitations in our study. First, the concentration of IS in the cultured cells was higher (ex. 10−3M) than that in plasma of patients with ESRD. Nonetheless, it mimics the additive effect of IS and other uremic toxin during kidney failure in vivo as previously reports [34, 54]. Likewise, the IS injection in our unilateral I/R seems not physiologic. As our previously report [34], we used the unilateral I/R model because the relatively normal, contralateral kidney without I/R could eliminate the various toxins accumulated during AKI. Thus, this model might lessen confounding factors from endogenous toxins and allow us to address the effects of exogenous IS effect on EPCs.


In patients with AKI, the serum concentration of total IS increased and was negatively correlated with the EPC number. In vitro, IS suppressed the number of EPCs and compromised their homing capacity in angiogenesis, proliferation, and senescence. The increased serum IS level also significantly attenuated EPC function by decreasing NO bioavailability and increasing oxidative stress. These findings shed more light on the mechanisms underlying the detrimental effects of IS on re-endothelialization and neovasculization during AKI. Targeting IS and/or its downstream pathways with agents such as NO-releasing statins and ROS scavengers may represent a novel EPC-rescue approach to the management of endothelial injury in AKI (Fig. 8).
Fig. 8

Schematic representation of IS-induced EPC dysfunction during AKI restored by ATO via NO- and ROS-de pendent mechanisms. IS impaired the proliferation and re-endothelialization of EPCs. The resultant decrease in VCAM and eNOs expression attenuated EPC mobilization to AKI sites. These impairments caused by IS can be reversed by ATO via restoration of NO- and ROS-mediated pathways


The authors would like to thank the staff of the Second Core Lab of the Department of Medical Research in the National Taiwan University Hospital for technical assistance. This study was supported by The Ta-Tung Kidney Foundation, Taiwan National Science Council (Grant NSC 101-2314-B-002-132-MY3, Grant NSC 101-2314-B-002-085-MY3, and Grant NSC 100-2314-B-002-119-), 100-N1776 from NTUH, 101-M1953, 102-S2097 and NTUH-TVGH Joint Research Program (VN9803, VN9906 and VN10009).

Conflict of interest

The authors have nothing to disclose.

Supplementary material

10456_2013_9339_MOESM1_ESM.docx (2.7 mb)
Supplementary material 1 (DOCX 2727 kb)

Copyright information

© Springer Science+Business Media Dordrecht 2013