Enhanced progenitor cell recruitment and endothelial repair after selective endothelial injury of the mouse kidney

Abstract

Primary and/or secondary injury of the renal microvascular endothelium is a common finding in various renal diseases. Besides well-known endothelial repair mechanisms, including endothelial cell (EC) proliferation and migration, homing of extrinsic cells such as endothelial progenitor cells (EPC) and hematopoietic stem cells (HSC) has been shown in various organs and may contribute to microvascular repair. However, these mechanisms have so far not been studied after selective microvascular injury in the kidney. The present study investigated the time course of EPC and HSC stimulation and homing following induction of selective EC injury in the mouse kidney along with various angiogenic factors potentially involved in EC repair and progenitor cell stimulation. Erythropoietin was used to stimulate progenitor cells in a therapeutic approach. We found that selective EC injury leads to a marked stimulation of EPCs, HSCs, and various angiogenic factors to orchestrate microvascular repair. Angiogenic factors started to increase as early as 30 min after disease induction. Progenitor cells could be first detected in the circulation and the spleen before they selectively homed to the diseased kidney. Injection of a high dose of erythropoietin 2 h after disease induction markedly attenuated vascular injury through nonhemodynamic mechanisms, possibly involving vascular endothelial growth factor release.

renal microvascular endothelium is the primary or secondary site of injury in many renal diseases, and progressive capillary loss is associated with glomerulosclerosis and fibrosis during chronic kidney disease (5, 26, 27, 31, 36). Because of the relevance of an intact microvascular endothelium to kidney function, immediate repair is essential, and the success of repair may be decisive in the subsequent decline or preservation of organ function.

Although the migration and proliferation of mature endothelial cells represent a well-known mechanism of repair, sex chromosome-linked studies after bone marrow transplantation or kidney transplantation have shown the presence of engrafted cells in the kidney (14, 35). The bone marrow has frequently been considered as the origin of such engrafted cells, which could be derived from hematopoietic stem cells (HSC). In addition, more differentiated and endothelial-specific progenitor cells have been proposed as endothelial progenitor cells (EPC). Even though controversy about the exact definition of the EPC exists (22), studies clearly suggest a role of such progenitor cells for tissue repair (2, 38).

Despite the high prevalence of endothelial injury during kidney disease, to date, only few studies directly investigated the role of extrinsic cells for endothelial repair in the kidney.

We have recently described a model of selective, unilateral endothelial injury of the mouse kidney (19). With the use of this renal endothelial injury model, the present study was performed to address the question whether HSC and EPC are selectively recruited to the diseased kidney after induction of endothelial injury and whether the known stem cell-activating capability of erythropoietin (EPO) can modulate progenitor cell recruitment and thereby contribute to improved endothelial repair.

MATERIALS AND METHODS

Experimental design.

All animal experiments were done according to American Physiological Society guidelines and duly approved by local government authorities. For the present study, 103 C57Bl/6J mice obtained from Jackson Laboratories (Bar Harbor, ME) and 6 Tie-2/green fluorescent protein (GFP) mice expressing GFP driven by the endothelial-specific and selective promoter for Tie-2 receptor, thereby resulting in specific fluorescence of endothelial cells (32), were used.

Site-specific endothelial injury was induced as previously described (19) by direct perfusion of the left kidney with the lectin concanavalin A (con A) followed by a rabbit anti-con A serum. For generation of the rabbit anti-con A serum, New Zealand White rabbits were immunized on day 0 using 0.3 mg con A with the same volume of complete Freund's adjuvant and on days 28, 42, and 56 with incomplete Freund's adjuvant, killed under general anesthesia, and bled by heart puncture. After centrifugation of the blood at 2,000 g for 15 min and removal of clotted blood cells, complement inactivation was done at 56°C for 30 min.

In the first experiment, mice were killed 30 min, 1 and 3 h, 1 day, 3 days, and 5 days after disease induction. Five mice served as healthy controls for evaluation of baseline values of progenitor cells and cytokine levels. Five mice were perfused with PBS in lieu of con A and served as controls (30-min time point) for cytokine measurements.

In a second experiment, 30 mice were subjected to unilateral renal injury. Twelve mice (6 in each group) were killed at day 1; 16 mice (8/group) were killed on day 5. At each time point, one group of mice received EPO (kind gift from CJ, Seoul, Korea) at a dose of 1,000 U/kg body wt ip in PBS, whereas the other group received an equal amount of PBS only. To ensure successful antibody binding and therefore initiation of disease, therapy was started 2 h after disease induction and repetitively given after 24, 36, and 48 h for animals killed at day 5.

Mice were kept in metabolic cages 24 h before death on day 5. On the day of death, anesthesia was induced by intraperitoneal injection of ketamine/xylazine. Blood was drawn from the inferior caval vein. The spleen was removed, and mice were perfused with 10 ml of PBS via the heart to remove blood components from both kidneys. Both kidneys were then removed, and a small slice of kidney tissue was processed for histological evaluation, whereas the majority of both kidney tissues were used for cell isolation and subsequent fluorescence-activated cell sorting (FACS) analysis.

FACS analysis.

To quantify peripheral circulating and splenic EPC and HSC by FACS, mononuclear cells were isolated from either 500 μl of peripheral blood or from tissue homogenates by density gradient centrifugation using Histopaque-1077 solution (Sigma-Aldrich, St. Louis, MO). For preparing splenic tissue homogenates, the whole organ was placed in 2 ml of DMEM-F-12 medium (Cambrex, Walkersville, MD) at 4°C. The tissue was minced and immediately homogenized. Cells were incubated for 45 min on ice with fluorescein isothiocyanate (FITC)-conjugated anti-mouse CD34 (RAM34) (eBioscience, San Diego, CA) and phycoerythrin (PE)-conjugated anti-mouse Flk-1 (Avas12α1) (BD Biosciences, Rockville, MD) or with FITC-conjugated anti-mouse CD117 (c-Kit) (BD Biosciences) and PE-conjugated anti-mouse CD150 (eBioscience). After incubation, cells were washed with PBS with 1% BSA (Sigma-Aldrich) and fixed in 4% paraformaldehyde. Data were acquired using a FACScan cytometer equipped with a 488-nm argon laser and analyzed using CellQuest software (Becton Dickinson, San Jose, CA). To quantify EPCs, the number of CD34/Flk-1 double-positive cells within the monocytic cell population was counted. For HSCs, the number of CD117/CD150 double-positive cells within the monocytic cell population was counted.

To quantify the percentage of EPCs and HSCs in renal tissue, FACS was also applied to this experiment. Before analysis, renal tissue was placed in culture dishes and minced to small pieces (1–3 mm) immediately after washing in Hank's balanced salt solution (Mediatech, Herndon, VA). Collagenase type II (1 mg/ml; Invitrogen) was added and incubated at 37°C in 5% CO₂ for 45 min. After digestion, the suspensions were filtered through 41 μm nylon mesh (Millipore). After being washed with PBS with 1% BSA, cells were incubated for 45 min on ice with FITC-conjugated anti-mouse CD34 (RAM34) (eBioscience) and PE-conjugated anti-mouse Flk-1 (Avas12α1) (BD Biosciences) or with FITC-conjugated anti-mouse CD117 (c-Kit) (BD Biosciences) and PE-conjugated anti-mouse CD150 (eBioscience). After incubation, cells were washed with PBS with 1% BSA (Sigma-Aldrich) again and fixed in 4% paraformaldehyde. Data were acquired using a FACScan cytometer equipped with a 488-nm argon laser and analyzed using CellQuest software (Becton Dickinson).

Laser-Doppler flowmetry.

For laser-Doppler measurements, a total number of 11 C57Bl/6J mice were used. Mice were anesthetized by intraperitoneal injection of a mixture of 6% ketamine/3% xylazine in sterile PBS adjusted to body weight. Mice were either pretreated with EPO or PBS 1 h before induction of endothelial injury. Laser-Doppler measurements were performed before and 30 min after disease induction. Therefore, both the left and the right kidneys were prepared through a flank incision as described for the animal model under direct vision through a stereomicroscope after 30 min; before disease induction, laser-Doppler flowmetry was performed for left kidneys only. The kidney capsule was removed. After surgical preparation, the kidneys were placed in a Plexiglas cup and immobilized without exerting tension on the renal vessels. The kidney surface was gently flushed using 0.9% saline solution. After placing the probe in a perpendicular position to the surface of the cortex, cortical blood flow on both sides was measured simultaneously by a dual-channel laser-Doppler flowmeter (Periflux 5000; Perimed) using two calibrated needle probes with a fiber separation of 0.25 mm (model 403; Perimed).

Red blood cell count.

The number of red blood cells was counted using a hematocytometer. Three samples were analyzed for each mouse, and the cell count was given as means ± SD.

Intravital videomicroscopy.

Intravital microscopy was performed using six Tie-2 GFP transgenic mice. A charge-coupled device videomicroscope (Nikon MM-40; Nikon) with an attached Photometric CoolSnap HQ digital camera was used. The field size is 0.0432 mm² at ×400 magnification and 0.171 mm² at ×200 magnification, resulting in resolutions of 0.35 and 0.68 μm, permitting the visualization of single erythrocytes.

For single point measurements, mice were anesthetized by intraperitoneal injection of a 2% solution of α-chloralose (3.75 μl/g body wt), and 10% urethane (11.25 μl/g body wt) for repeated intravital microscopy anesthesia was used as described for laser-Doppler measurements.

Mice were evaluated 24 h after disease induction. At least 10 fields of vision under a ×200 magnification were evaluated under fluorescent light and brightfield view. Video sequences of at least 5 s (∼100 frames) were recorded for each area and later evaluated using ImageJ (ImageJ).

During these experimental procedures, mice were placed on a heating plate with a constant temperature of 38°C. All animals were fed standard mice chow and tap water ad libitum.

Tissue processing and immunohistochemical staining.

Tissue for light microscopy was fixed in methyl Carnoy's solution, embedded in paraffin, and cut into 3-μm sections for indirect immunoperoxidase staining as described elsewhere (10, 20).

To perform immunoperoxidase staining, tissue sections were incubated with the following primary and secondary antibodies as indicated: 19A2, a murine IgM monoclonal antibody (mAb) against the proliferating cell nuclear antigen (PCNA) (20) to detect actively proliferating cells (Chemicon, Temecula, CA); mouse endothelial cell antigen 32 (MECA-32), a murine IgG1 mAb specific for detecting endothelial cells (16); and biotinylated lectin from Lycopersicon esculentum (tomato) for staining of glomerular capillaries (Sigma-Aldrich, Munich, Germany) (46). Negative controls for immunostaining included either deletion of the primary antibody or substitution of the primary antibody with equivalent concentrations of an irrelevant murine or rat mAb. All tissue sections were incubated with primary antibodies overnight at 4°C. Afterward, specific biotinylated (all by Zymed, San Francisco, CA) or Alexa-Fluor fluorescent secondary antibodies (Invitrogen, Karlsruhe, Germany) were applied followed by peroxides conjugated avidin D (Vector Laboratories, Burlingame, CA) and color development with diaminobenzidine for biotinylated antibodies.

Quantitative analysis of immunostaining and capillary rarefaction.

Digital images from fluorescence staining were analyzed using ImageJ. Therefore, ImageJ generated a grid with fields of a size of 10,000 μm², resulting in 88 fields/image. Evaluation was performed in at least 20 cortical images sparing glomeruli. Each square that contained no MECA-32 positive capillary was counted. This scoring system inversely reflects peritubular endothelial injury/rarefaction whereby low values represent intact capillaries and higher values indicate loss of capillaries (maximum is 100). PCNA positive cells were counted in 20 cortical fields of vision for assessment of the tubulointerstitium under a 400-fold magnification.

Multiplexed biomarker immunoassays.

Plasma was stored at −80°C until analysis. The multiplex mouse cardiovascular diseases biomarkers panel-1 and the mouse cytokines/chemokines panel-1 were used (MCVD1–77AK-4Plex, MCYTO-70K-15plex; Linco, part of Millipore) for the simultaneous quantification of the following analytes: soluble intercellular adhesion molecule (ICAM)-1, soluble vascular cell adhesion molecule (VCAM), soluble E-selectin, and matrix metalloprotease-9 (MMP-9) and cytokines/chemokines interleukin (IL)-1α, IL-1β, IL-10, IL-12, IL-2, IL-6, interferon-inducuble protein-10 (IP-10), keratinocyte-derived chemokine (KC), monocyte chemotactic protein (MCP)-1, macrophage inflammatory protein (MIP)-1, rapid upon activation, normal T cell expressed, and secreted (RANTES), tumor necrosis factor (TNF)-α, granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage colony-stimulating factor (GM-CSF), interferon (IFN)-γ, and vascular endothelial growth factor (VEGF). All examined analytes were tested individually and in combination to ensure that there were no cross-reactions. All measurements were performed in duplicate.

Briefly, the multibiomarkers and cytokine standards were resuspended in provided assay buffer and then differently serially diluted. Reconstituted standard (25 μl), Quality Controls, or sample (MCVD1–77AK-4Plex required a 1:50 serum dilution) were added to each well of a 96-well plate with 25 μl of the bead suspension. The plate was sealed, covered with aluminum foil, and incubated overnight (16–18 h) with agitation on a plate shaker at 4°C. Next, the plate was washed two times with 200 μl/well of wash buffer, removing buffer by vacuum filtration (<100 mmHg) between each wash. This was followed by addition of 25 μl of a detection antibody cocktail in each well and incubation at room temperature for 60 min. A streptavidin-PE solution (25 μl) was added to each well and incubated at room temperature for 30 min. The plate was then analyzed on the LuminexIS100 analyzer (Luminex, Austin, TX). The data were saved and evaluated as median fluorescence intensity using appropriate curve-fitting software (Luminex 100IS software version 2.3). A five-parameter logistic method with weighting was used.

Statistical analysis.

All values are expressed as means ± SD. Statistical analysis was performed using one-way ANOVA (Kruskal-Wallis test) for nonparametric data and two-way ANOVA with Bonferroni posttesting for multifactoral analysis. Posttesting was performed using Dunn's multiple-comparison test. Single comparisons were done using the nonparametric Mann-Whitney test. Statistical significance was defined as a P of <0.05 (GraphPad Prism Version 5.00; GraphPad Software, San Diego, CA).

RESULTS

Induction of endothelial injury in mice.

Tissue sections from all mice were evaluated for successful disease induction using periodic acid-Schiff (PAS) and acid fuchsin orange G (AFOG) staining. All left kidneys perfused with the antibody demonstrated injury analogous to that previously described (19). Starting 3 h [Fig. 1, A and B (PAS staining)] after disease induction, kidneys already showed detectable signs of thrombotic microangiopathy (arrows) that reached maximum after 3 days. Glomeruli showed fibrin thrombi [Fig. 1C (arrow, bright orange by AFOG staining)] and also significant tubulointerstitial injury [Fig. 1D (PAS)]. Contralateral right kidneys showed normal histology at all times tested [Fig. 1E (3 days ×20) and Fig. 1F (3 h ×40)].

Fig. 1.

Successful induction of endothelial injury in mice. Starting 3 h after disease induction, signs of thrombotic microangiopathy could be detected in glomerular capillaries (A) and arterioles (B) by periodic acid-Schiff (PAS) staining. Glomeruli presented with fibrin thrombi (C) in acid fuchsin orange G (AFOG) stainings and also significant tubulointerstitial injury as depicted in D (PAS staining). Arrows indicate thrombotic material. E and F: contralateral (right, not diseased) kidney of diseased mice after 3 days (E; PAS, ×20 magnification) and 3 h (F; PAS, ×40 magnification).

Mobilization of EPC and HSC after selective renal endothelial injury.

After induction of endothelial injury in the left kidney, all mice demonstrated a uniform sequence of progenitor/stem cell mobilization. We detected a rise of EPC and HSC in the peripheral blood starting 3 h after disease induction and persisting up to day 3 (Fig. 2, A and B). This was followed by an increased number of splenic EPC and HSC (for HSC significant on day 3; Fig. 2, C and D) before the surge in EPC and HSC could finally be detected in diseased kidneys (Fig. 2, E and F). Interestingly, the homing of both EPC and HSC during the surge was restricted to the injured left kidneys, whereas the contralateral right kidneys of the same mice showed no increase in EPC or HSC, indicating specific homing factors (Fig. 2 I and J).

Fig. 2.

Modulation of endothelial progenitor cells (EPC) and hematopoietic stem cells (HSC) recruitment after induction of endothelial injury in the kidney. Fluorescence-activated cell sorting (FACS) analysis of EPC (A) and HSC (B) from blood samples, spleen [EPC (C) and HSC (D)], left kidneys [EPC (E) and HSC (F)], and right kidneys [EPC (G) and HSC (H)] was performed after induction of unilateral, left-sided endothelial injury in the mouse kidney. I and J: direct comparison of left and right kidney for EPC (I) and for HSC on day 5 (J). *P < 0.05, **P < 0.01, and ***P < 0.001 for 1-way ANOVA. #Individual differences.

Response of angiogenic factors following renal microvascular injury.

A large variety of chemokines and angiogenic factors have been shown to be involved in the response to endothelial injury (42, 44) on the one hand and in the recruitment and homing of EPC and HSC (13, 18) on the other. Hence, we decided to profile the cytokines in parallel with the analysis of EPC and HSC. To exclude chemokine stimulation induced by the injection itself, an additional control sham injection group was examined. The first and most significant increase occurred in MMP-9 levels, which peaked after 30 min and demonstrated values up to 10 times above baseline (Fig. 3A). KC, the murine analog to IL-8 increased after 30 min and remained increased up to 3 h (Fig. 3B). The proinflammatory IL-6 (Fig. 3C) and the anti-inflammatory IL-10 (Fig. 3D) peaked 3 h after disease induction. TNF-α (Fig. 3E) followed a similar pattern, whereas MCP-1 was increased after 3 h and 5 days (Fig. 3F), and MIP-1α levels were increased after 3 h only (Fig. 3G). GM-CSF (Fig. 3H) and G-CSF (Fig. 3I) peaked later on days 3 and 5, whereas IP-10 exhibited a peak 5 days after disease induction (Fig. 3J). These observations demonstrate that a large variety of modulatory angiogenic factors are released upon endothelial injury of a single kidney, thus potentially contributing to either the inflammatory response or early induction of protective and/or repair mechanisms, including progenitor and stem cell activation.

Fig. 3.

Modulation of angiogenic factors after induction of endothelial injury in the kidney. A Luminex multiplex analysis was performed to assess chemokine levels of matrix metalloproteinase (MMP)-9 (A), keratinocyte-derived chemokine (KC, B), interleukin (IL)-6 (C), IL-10 (D), tumor necrosis factor (TNF)-α (E), monocyte chemotactic protein (MCP-1) (F), macrophage inflammatory protein (MIP)-1α (G), granulocyte macrophage colony-stimulating factor (GM-CSF) (H), granulocyte colony-stimulating factor (G-CSF) (I), and interferon-inducible protein-10 (IP-10) (J) during the time course of disease. *P < 0.05, **P < 0.01, and ***P < 0.001 for 1-way ANOVA. #Individual differences.

EPO therapy prevents endothelial injury.

Previous experimental studies have shown that EPO application in high doses (1,000 U/kg body wt) can stimulate the activation of EPC (17).

Therefore, we decided to use EPO administration to enhance endothelial repair by progenitor cell activation. Two groups of mice were investigated 24 h and 5 days after disease induction. Treatment was started 2 h after induction of endothelial injury.

Unexpectedly, we detected the most significant effect of EPO already at the early time point after 24 h. As depicted in Fig. 4A,controls had a higher degree of endothelial injury/loss as evaluated with MECA-32 staining compared with EPO-treated mice (Fig. 4B). EPO treatment significantly prevented endothelial injury (Fig. 4C) at day 1, whereas no differences were apparent on day 5. Because endothelial cell proliferation is one of the most important repair mechanisms, we evaluated whether EPO exerts its effect via endothelial cell proliferation. Indeed, there was a much higher number of proliferating endothelial cells on day 1 but not at 5 days after disease induction (Fig. 4D). This might be related to the reduced loss of endothelial cells due to EPO treatment or to effects on endothelial cell proliferation.

Fig. 4.

Reduced endothelial injury and enhanced endothelial cell proliferation due to erythropoietin (EPO) therapy. Microvascular endothelium was specifically labeled using fluorescent staining for mouse endothelial cell antigen 32 (MECA-32) (red). As depicted in A, endothelial injury was much more severe in control mice compared with sections from EPO-treated mice on day 1 (B). C shows the reduction of EC injury in EPO-treated mice on day 1, whereas there are no differences on day 5. Endothelial cell proliferation (D) is also significantly enhanced on day 1. **P < 0.01 by 2-way ANOVA. #Significant difference by posttest.

To verify our findings in vivo, we performed intravital videomicroscopy in a small number of EPO-treated and control mice. Disease was induced in Tie-2 GFP transgenic mice (n = 6), and animals received either EPO or PBS as control. On day 1, we evaluated microvascular perfusion of these mice, which was indeed significantly improved after EPO treatment (videos 1 and 2, EPO-treated mouse after 24 h, GFP and brightfield view; videos 3 and 4, control mouse after 24 h, GFP and brightfield view) (Supplemental data for this article may be found on the American Journal of Physiology: Renal Physiology website.).

Because the recruitment of EPC and HSC to the kidney 5 days after endothelial injury was documented, we next evaluated both cell types in EPO-treated mice.

EPO therapy has modulatory effects on EPC activation.

The number of circulating EPC (Fig. 5A) and HSC (Fig. 5B) was reduced (significant only for HSC) in EPO-treated mice, most probably as a result of reduced endothelial injury. We detected enhanced EPC sequestration by the spleen in EPO-treated mice (Fig. 5C), although we could not detect any differences for HSC (Fig. 5D). Interestingly and in line with the other findings, EPC were also reduced in the left kidney of EPO-treated mice (Fig. 5E), whereas we detected no differences for HSC. There were no differences for right kidneys of EPO-treated and control mice. It should be mentioned that EPC and HSC levels in diseased kidneys and controls reached exactly the same values as in the first set of experiments depicted in Fig. 2.

Fig. 5.

Reduced EPC and stem cell recruitment to the kidney, but enhanced EPC stimulation due to EPO therapy. Circulating levels of EPC (A) and HSC (B) as well as splenic EPC (C) and HSC (D) were measured using FACS analysis. EPO significantly reduced renal EPC recruitment (E), and HSC were unchanged (F). No differences occurred for EPC (G) and HSC (H) of the right kidneys. *P < 0.05, **P < 0.01, and ***P < 0.001 by Mann-Whitney test.

EPO treatment modulates release of angiogenic factors and leads to increased VEGF.

We also sought to examine if EPO treatment modulates the cytokine release after endothelial injury. Commensurate with histological findings, MMP-9 was significantly reduced on day 1 in EPO-treated mice (Fig. 6A), whereas reduction of the proinflammatory IL-6 was only nearly significant (F = 0.069 by 2-way ANOVA; Fig. 6C). KC, a stem cell recruiting cytokine, was increased in the EPO group on day 5 (Fig. 6C).

Fig. 6.

EPO significantly altered chemokine responses. Levels of MMP-9 (A),IL-6 (B), KC (C), and VEGF (D) were altered due to EPO therapy. IL-6 levels failed to demonstrate significance by 2-way ANOVA (F = 0.069). *P < 0.05 and **P < 0.01 by 2-way ANOVA. #Significant difference by posttest. §One data point out of range.

Searching for a potential mediator of the remarkable endothelial protective effects of EPO treatment, we next measured VEGF serum levels in EPO-treated and control mice. We detected significant differences between groups with higher VEGF levels in EPO-treated mice on day 1 (Fig. 6D).

These data, together with the tempered proinflammatory cytokine profile of EPO-treated mice, might explain their endothelial protection against con A/anti-con A thrombotic microangiopathy.

Effects of EPO therapy do not relate to changes in microvascular hemodynamics or red blood cell count.

EPO treatment may cause additional effects on the microvasculature, and also changes in red blood cell count (RBCC) could interfere with the disease process. Several studies suggest direct effects of EPO treatment on vascular blood flow and tone (8, 25). To exclude such immediate effects, we decided to monitor microvascular circulation in mice with and without EPO treatment. Mice were pretreated 1 h before disease induction with EPO or PBS, and laser-Doppler measurements were performed 30 min after disease induction (by renal arterial perfusion). Parallel recordings of left (perfused) and right kidneys are depicted in Fig. 7A, demonstrating no difference.

Fig. 7.

EPO did not influence microvascular hemodynamics and red blood cell count (RBCC). A: microvascular hemodynamics were assessed using laser-Doppler flowmetry in diseased and control kidneys. B: RBCC was assessed using a hematocytometer. No differences occurred due to EPO therapy.

In addition, previous studies have reported a significant influence of EPO administration on RBCC a few days after administration (17). However, in our study, RBCC were unchanged in EPO-treated compared with control mice 5 days after EPO administration (Fig. 7B). Therefore, the above mechanisms are unlikely to cause the observed protective effects.

DISCUSSION

Injury to the microvascular endothelium is very common in renal disease, and rapid and adequate repair of endothelial injury is critical for preservation of organ function, while persistent endothelial damage is associated with disease progression (27).

Experimental studies suggest that bone marrow-derived HSC and more specific EPC might contribute to endothelial repair in the kidney (15, 33, 40, 47). Studies in kidney allografts also emphasize the chimerism of the donor endothelial cells by invading cells from the recipient (39).

Having established an inducible selective and reproducible endothelial injury model in the mouse kidney (19), we sought to evaluate the potential role of mobilized endogenous HSC and EPC after endothelial injury and repair in progression of kidney disease. Because endothelial injury in our model is restricted to the left kidney, we were able to use the contralateral kidney as internal control. In a first step, we described the chronology of HSC and EPC mobilization and engraftment upon endothelial injury of the kidney, which surge after 3 h when they can be detected in the circulation. Subsequently, HSC and EPC demonstrated a uniform splenic sequestration (maximum on day 3 for HSC) before they could be detected in the affected kidney (day 5). This recruitment to the kidney was selective for the affected side and not observed in the nondiseased kidney of the same mouse, emphasizing the specificity of this process. The chronology and magnitude of progenitor cell homing after endothelial injury in this study is clearly more prominent than those described after ischemia-reperfusion injury (33). Data provided herein offer a much more detailed picture than that drawn from histological studies mainly performed in anti-Thy 1 nephritis of the rat after bone marrow transplantation (21, 40), where endothelial injury is a secondary phenomenon. We believe that the specific impact on endothelial cells in this experimental endothelial cell injury model accounts for the rapid stimulation of progenitor cells. Data from other selective endothelial injury models in the kidney is lacking, but the relevance of bone marrow-derived cells has been indicated in a case study of a 28-yr-old woman with thrombotic microangiopathy (41). Still, it is not clear whether recruited cells participate in endothelial repair directly or through paracrine mechanisms in the kidney (1, 2, 24, 45).

Potential cytokine signaling to mobilize stem cells was considered and tested using a multiplex approach. Whereas we did not detect significant modulation of cytokines such as VCAM-1, ICAM-1, and selectins, many other angiogenic factors and chemokines demonstrated changes during the experimental time frame. The most prominent increase could be detected for MMP-9, where levels increased up to 10-fold above normal within 30 min after disease induction. MMP-9 is known to be relevant for Kit ligand processing and has been shown to originate from ischemic tissue and endothelial cells. It plays a role in platelet aggregation, typically induced after endothelial injury due to platelet activation (6, 18, 37), and appears to be a strong candidate for the early progenitor cell recruitment. Various cytokines appearing in peripheral blood at later times, including KC, the putative murine analog of IL-8 (13, 42), GM-CSF and G-CSF, as well as cytokines responsible for inflammatory responses such as TNF-α, IL-6, IL-10, and MCP-1 were increased. Especially, the induction of KC/IL-8 has two very important aspects in this setting. Experimental studies have shown that IL-8 plays an important role in the repair response after endothelial injury, promotes endothelial cell survival, and stimulates MMP production, which may in turn participate in the regulation of angiogenesis (44). On the other hand, IL-8 has also been directly linked with the recruitment of EPC (13, 28). Both G-CSF (23) and GM-CSF (45, 48) have been involved in stem cell activation, both are in clinical use in stem cell transplantation, and G-CSF also participates in endothelial repair and neovascularization (49). In contrast, IL-6 and -10 as well as MCP-1 are much more prominent mediators of the inflammatory response upon endothelial injury (42, 43) but have also been linked to some extent with the regulation of progenitor cell activity (29) and endothelial repair (11). In contrast, TNF-α is considered a potent proinflammatory mediator that promotes endothelial injury/dysfunction and inhibits stem cells (50). MIP-1α, found to be enhanced in parallel with the above chemokines, has to be considered as a negative regulator of stem cell proliferation and a protective chemokine (9), whereas IP-10 (also known as CXCL10) seems to be involved in the suppression of endothelial cell proliferation (30).

In summary, there are significant responses of angiogenic factors that have been linked with the endothelial repair response by previous studies and might influence the endothelial repair response in this model in part also through the induction of progenitor cell activation.

On the basis of these findings, we then aimed to mitigate injury and improve the repair process through stimulation of stem/progenitor cells. Several angiogenic factors mentioned above, including VEGF, have been described as potent stimulators of stem and progenitor cells, and many studies have used the potency of EPO to stimulate these cells.

To evaluate the effect of enhanced progenitor cell stimulation, we therefore decided to use EPO, a well-known stimulator of EPC activation, for therapeutic intervention (3, 17). We induced endothelial injury in mice and started EPO treatment not earlier than 2 h after disease induction to assure the prior onset of endothelial injury. Dosage was adjusted to previous studies that have shown relevant effects using 1,000 U/kg body wt of EPO (17). Application was repeated every 24 h for a period of 3 days for the long-term group. FACS analysis was performed in day 5, since the first experiment showed major HSC and EPC recruitment at that time point.

Unexpectedly, we found that, already on day 1, 22 h after EPO application, EPO treatment markedly affected the course of the disease. Endothelial injury as evaluated using a MECA-32 stain was significantly reduced, and endothelial cell proliferation in turn markedly enhanced. Because of these protective effects, the recruitment of EPC to the injured kidney was reduced on day 5 compared with controls, whereas there was still an enhanced splenic sequestration of EPC at that time point, clearly indicating the potency of EPO (17). This was also reflected by the cytokine response, with reduced production of MMP-9 and enhanced amounts of KC on day 5. In contrast to a study investigating regulatory effects of EPO on EPC (3), we also detected significantly elevated levels of circulating VEGF in EPO-treated mice. VEGF is the most potent mitogen and a survival factor for endothelial cells (12). Although it has the ability to stimulate stem cell mobilization (37) and promote transendothelial migration of EPC (44), this mechanism does not appear to be dominant, since endothelial cytoprotective effects outweigh it. Therefore, the rapid release of VEGF is a potential mediator of endothelial cell protection, whereas EPO-mediated effects on renal microvascular hemodynamics (8, 25) and enhanced RBC production could be excluded (17).

EPO has previously been recognized as a potent cytoprotective agent for endothelial cells in ischemic and vascular injury (7, 34) as well as in chronic kidney disease (3, 4, 15). Signaling through the Janus kinase-2 pathway and protein kinase B/AKT-1 have been proposed as antiapoptotic signaling pathways (7), whereas, in contrast to the present study, VEGF has not been considered as a potential mediator.

In summary, the present study illustrates the time course of stem and EPC homing after renal endothelial injury together with angiogenic factors that might influence the repair response. In a therapeutic approach, EPO proved its potency to prevent endothelial injury in this experimental injury model and to stimulate EPC activation. Enhanced circulating VEGF could be detected as a potential mediator of protective effects. Further studies will be necessary to elucidate whether EPC and HSC directly, indirectly, or in both ways participate in endothelial repair during this disease model and whether the apparent changes of angiogenic factors directly relate to the repair response. The EPO effect is confined to early endothelial cytoprotection, results in the early improvement in microvascular injury and perfusion, and therefore bypasses the need for EPC-mediated regeneration. The latter property may be especially useful in cases of EPC dysfunction. EPO might also represent a treatment option to protect the renal microvascular endothelium from acute injury.

GRANTS

This work was supported by a stipend of the “Stiftung Familie Bouhon” to B. Hohenstein and grants to B. Hohenstein (IZKF Erlangen, TP B14) and M. S. Goligorsky (National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-052783, DK-45462, and DK-054602).

DISCLOSURES

K.-U. Eckardt received lecture and consulting fees from Affymax, Amgen, Johnson & Johnson, Kirin, Roche, and Sandoz Hexal; B. Hohenstein received lecture and consulting fees from Bristol-Myers Squibb and Nycomed. C. P. M. Hugo has received honoraria by Astellas, Bristol-Myers Squibb, Wyeth, Roche, and Novartis and grant support by Wyeth, Roche, and Novartis.