Renal Physiology

Soluble thrombomodulin reduces inflammation and prevents microalbuminuria induced by chronic endothelial activation in transgenic mice

Gangaraju Rajashekhar, Akanksha Gupta, Abby Marin, Jessica Friedrich, Antje Willuweit, David T. Berg, Martin S. Cramer, George E. Sandusky, Timothy A. Sutton, David P. Basile, Brian W. Grinnell, Matthias Clauss


Chronic kidney disease pathogenesis involves both tubular and vascular injuries. Despite abundant investigations to identify the risk factors, the involvement of chronic endothelial dysfunction in developing nephropathies is insufficiently explored. Previously, soluble thrombomodulin (sTM), a cofactor in the activation of protein C, has been shown to protect endothelial function in models of acute kidney injury. In this study, the role for sTM in treating chronic kidney disease was explored by employing a mouse model of chronic vascular activation using endothelial-specific TNF-α-expressing (tie2-TNF) mice. Analysis of kidneys from these mice after 3 mo showed no apparent phenotype, whereas 6-mo-old mice demonstrated infiltration of CD45-positive leukocytes accompanied by upregulated gene expression of inflammatory chemokines, markers of kidney injury, and albuminuria. Intervention with murine sTM with biweekly subcutaneous injections during this window of disease development between months 3 and 6 prevented the development of kidney pathology. To better understand the mechanisms of these findings, we determined whether sTM could also prevent chronic endothelial cell activation in vitro. Indeed, treatment with sTM normalized increased chemokines, adhesion molecule expression, and reduced transmigration of monocytes in continuously activated TNF-expressing endothelial cells. Our results suggest that vascular inflammation associated with vulnerable endothelium can contribute to loss in renal function as suggested by the tie2-TNF mice, a unique model for studying the role of vascular activation and inflammation in chronic kidney disease. Furthermore, the ability to restore the endothelial balance by exogenous administration of sTM via downregulation of specific adhesion molecules and chemokines suggests a potential for therapeutic intervention in kidney disease associated with chronic inflammation.

  • chronic inflammation
  • CKD
  • ICAM1
  • tmTNF

in diseases associated with chronic inflammation, endothelial activation plays a role in both initiation and exacerbation of the pathology. Such chronic endothelial activation is accompanied by endothelial dysfunction, characterized by sustained expression of leukocyte adhesion molecules, chemokine production, and localized or disseminated tissue dysfunction and damage (30, 46, 49). Comparable to the development of atherosclerotic lesions, these changes are usually subtle, develop gradually, and are not comparable to the immediate and severe damage seen in acute injuries such as by ischemia-reperfusion (21). An example of slow chronic disease development is diabetic nephropathy, which is the most common type of chronic kidney disease (CKD) progressing to terminal renal failure (20). Despite abundant investigations to identify additional risk factors for CKD, the involvement of endothelial dysfunction in developing nephropathies and appropriate treatment modalities are insufficiently explored.

The vascular endothelium, which is a target of many proinflammatory cytokines, can also evoke anti-inflammatory activity in response to activation. One possible mechanism by which the endothelium is protected is through thrombomodulin, which we and others recently described as an anti-inflammatory agent. Indeed, soluble thrombomodulin (sTM) can reduce acute kidney injury and experimental glomerulonephritis (17, 29, 41). Surprisingly, our earlier study showed that a point mutation in thrombomodulin, F376L, that is incapable of acting as a cofactor for activating protein C, was equally effective in protecting against acute kidney injury (AKI). This suggested that sTM's effects in protecting endothelial cells are independent of activated protein C (APC) or thrombin activation, which was previously described as a novel mechanism for thrombomodulin in vitro and in vivo (8).

Although sTM has been shown to play an important protective role in acute inflammatory injury, there are little data to suggest a protective role in situations associated with chronic activation of the vasculature. Therefore, we sought to examine the potential anti-inflammatory effect of sTM in a model of chronic endothelial activation and have employed our endothelial TNF-overexpressing transgenic mouse model (48). In this model we have used the endothelial promoter tie2 to drive overexpression of transmembrane TNF, which by mutation of its TNF-α-converting enzyme (TACE) cleavage site remains bound to endothelium (48). Starting from 3 mo of age, these mice develop proinflammatory exudates, which are most prominent in the kidney and liver but also were seen in other organs such as the heart and lung. Thus, starting from month 3 on, we applied sTM for an extended time period of 3 mo and analyzed the effect of sTM to reduce TNF-induced chronic endothelial cell activation, inflammation, and kidney dysfunction.


tie2-TNF transgenic animals and sTM treatment in vivo.

Animal studies were carried out according to the guidelines of the Institutional Animal Care and Use Committee Review Board, IU School of Medicine. The construction of transgene and generation of tie-2-TNF transgenic animals was described previously (48), in which the cDNA of the uncleavable murine tmTNF-α mutant [mTNFα Δ1–9, K(11)E] cloned between the endothelial-specific tie2 promoter and the tie2 first intron to localize TNF-α specific to the endothelium. Mice used for this study had been back crossed for more than eight generations in C57BL/6 animals. To evaluate the effect of pretreatment with sTM on nephropathy, 3-mo-old female mice heterozygous for the transgene (n = 16) and nontransgenic littermates (n = 16) were divided into two groups. While the control group received 0.9% normal saline, the treatment group received 2.5 mg/kg of murine sTM (Lilly Laboratories, Indianapolis, IN) twice weekly subcutaneously for 3 mo. Mice were sacked at 6 mo of age, and renal function was assessed using serum albumin/creatinine measurements. Urine was collected from bladder and albumin and creatinine were determined with specific ELISA kits as described previously (41).

Tissue culture.

Murine endothelial cells transfected with the noncleavable transmembrane mutant form of murine TNF (TNF) and control endothelial cells (transfected with the empty vector) as described previously (34) were cultured in DMEM supplemented with 10% FBS, 2 mM glutamine and 1× penicillin/streptomycin. The cells were maintained as monolayers in Nunclon 75-cm2 cell culture flasks (Nalge Nunc, Naperville, IL) at 37°C in a humidified atmosphere of 5% CO2 in air. The THP-1 (human acute monocytic leukemia cell line, ATCC, Rockville, MD) cells were cultured in RPMI supplemented with 10% FCS, 2% glutamine, and 1% penicillin/streptomycin and maintained as suspension culture at 37°C in a humidified atmosphere of 5% CO2 in air.

Immunohistochemistry and microscopy.

To assess the extravasation of leukocytes, the right kidneys were fixed in paraformaldehyde and processed for immunohistology for CD45 as per the standard procedures of the ABC method (36). Briefly, paraffin sections were blocked with 1× universal blocking serum (Dako North America, Carpinteria, CA) and incubated with rat anti-mouse CD45 primary antibody or control isotype-matched IgG antibodies (both BD Biosciences, Franklin Lakes, NJ) at 4°C overnight. Bound antibody was detected with a biotinylated secondary antibody and avidin-biotinylated peroxidase complex as per the manufacturer (Vector Laboratories, Burlingame, CA). Finally, the immune complex was detected with AEC chromogen (Sigma, St. Louis, MO) and counterstained with hematoxylin. Microscopy was performed on a Nikon Eclipse 80i upright system. Images were captured in a blinded fashion, and quantitative intensity (expression) data were obtained by MetaMorph Imaging software (Molecular Devices, Downingtown, PA).

To determine fibrotic lesions in renal sections, Masson's trichrome staining was used. The percentage of fibrotic area relative to overall fibrosis in the section was evaluated under high-power magnification, and the amount of collagen deposit (stained in blue) was examined by a board-certified veterinary pathologist unaware of the clinical or experimental findings as described previously (11).

Immunofluorescent studies were carried out in kidney tissues as described previously (2). Briefly, perfused kidneys were fixed in 100% ice-cold methanol and sectioned at 100 μm using a vibratome. For double-staining of endothelial-specific platelet endothelial cell adhesion molecule (PECAM) with ICAM-1, labeling of primary antibodies was carried out by incubating tissues with anti-PECAM (SEW31, Dr. Peter Newman, Blood Center of Southeast Wisconsin, 1:200,000) and anti-ICAM (1:200,000, eBioscience, San Diego, CA) antibodies followed by incubation with goat anti-rabbit-Alexa Fluor 488 and donkey anti-mouse-Cy3 (both 1:10,000, Invitrogen). Following development using a tyramide Cy3 signal amplification kit (PerkinElmer), images were obtained using a Zeiss LSM NLO confocal microscope equipped with Ar and HeNe lasers.

Modified Miles assay.

Mice were injected intravenously with 2.5 mg FITC-labeled BSA (Sigma) in 100 μl PBS. After 10 min, mice were anesthetized and thrombin (100 U in 20 μl saline, Sigma) or saline only was injected into the left or right ear, respectively. Seven minutes later, the animals were euthanized, and the ear was removed, minced, and digested with protease K overnight. The extravasated fluorescent dye was measured using a fluorescence photometer, and thrombin-induced vascular permeability was calculated as the ratio between thrombin- and saline-injected ears.

mRNA quantification using QuantiGene Plex 2.0 branched DNA assay.

For mRNA quantification, the perfused left kidneys were collected in RNAlater RNA stabilization solution (Ambion, Foster City, CA) at 4°C overnight followed by storage at −80°C until use. The tissue was homogenized using a QuantiGene 2.0 Sample Processing Kit (Panomics, Fremont, CA). After tissue homogenization, the QuantiGene Plex 2.0 Reagent bDNA (51) assay (Panomics) was used according to the manufacturer's instructions to evaluate the expression of a panel of 36 selected biomarkers that have been implicated to play a role in kidney injury. These genes included proinflammatory markers [RANTES, ITAC, Ccl3 (MIP-1α), VCAM-1, IL-18], markers of kidney injury [cystatin C, osteopontin 1, transforming growth factor (TGF)-β receptor 2 (TGFβ1), FGL-2 (fibroleukin), and TGFβR1], and endothelial dysfunction [Cybb (gp91phox) regulatory subunit of the NADPH oxidase, inducible nitric oxide synthase (iNOS), ICAM1, and serpine 2].

Transmigration assay.

The transmigration assay was performed as described before (39). Briefly, control and TNF-expressing endothelial cells were cultured on laminin-coated Transwell filter inserts (Corning Costar, Bodenheim, Germany) for 2 days. Both control and TNF-expressing endothelial cells were incubated with 20 μg/ml of monoclonal antibodies [25ZC7 (anti-ICAM-1) (44), YNI1.7 (anti-ICAM-1) (31), 6C7, and 4B12 (anti-VCAM-1) (38) were a kind gift of D. Vestweber, Münster, Germany] for 20 min at 37°C before the assay or incubated with hirudin (Sigma), a thrombin inhibitor throughout the assay. To test the effect of sTM on monocyte migration across intact endothelial monolayer, 2.5 × 105 THP-1 cells were preincubated with 3 μg/ml (50 nM) murine sTM (Lilly Laboratories) for up to 24 h at 37°C in a humidified atmosphere in migration medium (DMEM+ supplemented with 5% FCS, 2% glutamine, and 25 mM HEPES). For pertussis toxin (PTX) treatment, cells were incubated with 200 ng/ml PTX (Sigma) for 2 h at 37°C and washed four times before the assay. The number of transmigrated leukocytes was measured using a cell counter (Beckman Coulter Counter), and each value was performed in triplicate. The integrity of the endothelial monolayer on each filter was checked by staining the cells with a solution of 0.5% crystal violet in 20% methanol and light microscopic examination.

Real-time quantitative RT-PCR.

To assess the differential expression of proinflammatory genes without and with sTM treatment, quantitative real-time PCR was performed. Both control and TNF-expressing endothelial cells without and with sTM (3 μg/ml, 24 h) treatment were processed for total RNA using a Qiagen RNeasy kit (Qiagen, Valencia CA). Quantitative real-time PCR was performed using an MJ Research PTC-200 Chromo4 sequence detector. An iScript one-step qRT-PCR kit with SYBR Green (Bio-Rad, Hercules, CA) was used for cDNA synthesis and PCR amplification as per the manufacturer's instructions using gene-specific primer pairs. The amount of target gene transcript normalized to the endogenous elongation factor 1-α, housekeeping gene transcript was computed based on a comparative Ct method as described by us previously (34). The results were expressed as fold-change in mRNA levels relative to control untreated cells.

Western blot analysis.

To perform immunoblot analysis, the whole cell extracts from control, control treated with sTM, TNF, and TNF treated with sTM endothelial cell populations were probed with anti-RANTES antibodies (R&D Systems, Minneapolis, MN) as per the standard procedure described earlier (25, 36).

Statistical analysis.

For QuantiGene Plex 2.0 analyses, two-way comparisons of control and experimental groups were performed using a Welch's t-test. The results were filtered for present/absent call [generated by the MAS 5 software with numeric values of absent (A = 0), marginal (M = 0.5), or present (P = 1)]. All samples with an average “Fraction Present” value of <0.5 for each group were excluded from the analysis. Additionally, the results were filtered for statistical significance based on a P value of <0.05. For quantitative real-time PCR, the data are presented as means ± SD for each group, performed in duplicate, and repeated an additional three times. For in vitro transendothelial migration assay, data are expressed as means ± SD for each group performed in triplicate. For immunohistological analysis, the data are presented as means ± SE from all animals/group performed two times individually and counted by an investigator who was blinded to the treatment regimen. Statistical significance was determined by ANOVA using the Microsoft Excel Statistical Package. A probability value of P < 0.05 was considered statistically significant.


Continuous activation of endothelium in tie2-TNF transgenic mice induces markers of nephropathy.

To assess the relevance of our transgenic endothelial activation model for kidney disease, we analyzed kidneys of age-matched female wild-type and tie2-TNF transgenic mice. Whereas at 3 mo tie2-TNF mice did not show any changes in the albumin/creatinine ratio or in proinflammatory exudates (data not shown), analysis of 6-mo-old mice demonstrated a significant increase in the albumin/creatinine ratio in the tie2-TNF group compared with age-matched wild-type mice (Fig. 1A). This decrease in renal function was accompanied by increased fibrosis in the glomerulus as noted by trichrome staining for collagen deposition (Fig. 1B). Quantification of this interstitial glomerular fibrosis by the scoring classification most commonly used in pathology(11) revealed a significant (P < 0.02) but moderate increase in the tie2-TNF group compared with the wild-type control group (Fig. 1, B and C). A regional comparison did not show any differential fibrosis index in deep cortical and surface glomeruli, but there was a trend toward increased fibrosis in deep cortical glomeruli. We found no or undetectable levels of fibrosis in the tubulointerstitial compartments.

Fig. 1.

Continuous activation of endothelium in tie2-TNF transgenic mice induces markers of nephropathy. Both albumin normalized to creatinine levels from urine samples (A) and interstitial glomerular fibrosis from trichrome-stained kidney sections (B) were assessed in wild-type (Wt) or tie2-TNF (TNF) mice. Quantification of glomerular thickness scores are shown in C. Data shown are an average from a group size of n = 8. Scale bar = 100 μm.

To gain further molecular support for the observed kidney pathology, we analyzed the expression of selected kidney disease genes using a quantitative gene expression method (51) from kidneys of 6-mo-old tie2-TNF-overexpressing mice compared with age-matched wild-type littermate kidneys. mRNA expression changes showed a significant increase in genes associated with tubular damage, fibrosis, and inflammation in tie2-TNF mice (P < 0.01) compared with wild-type controls (Table 1).

View this table:
Table 1.

Selective gene induction in kidneys of tie2-TNF mice and TNF-expressing endothelial cells

Increased endothelial expression of proinflammatory proteins and leukocyte infiltration and sensitivity to thrombin in tie2-TNF transgenic mice.

Many of the genes upregulated in the kidneys of tie2-TNF transgenic mice have been previously identified to be upregulated in TNF-expressing endothelial vs. control endothelial cells (34), including the chemokines MCP-1, the adhesion molecule VCAM-1, the NADPH oxidase gp91phox subunit, and TFGβR2 (see Table 1). Because ICAM-1 is also expressed in nonendothelial cells, we confirmed endothelial expression of this intracellular adhesion protein in kidney sections by using costaining with the endothelial cell marker PECAM (CD31) and confocal immunofluorescence analysis. As shown in Fig. 2, ICAM-1 (green) was strongly upregulated in kidney section of 6-mo-old tie2-TNF mice (Fig. 2B) compared with age-matched wild-type mice (Fig. 2A). The overlay with endothelial cell marker protein CD31 (red) revealed that only endothelial cells displayed increased ICAM-1 levels (Fig. 2, E and F). Again, inflammatory exudates were observed in kidney sections of 6-mo-old tie2-TNF mice but not in 3-mo-old tie2-TNF mice or in wild-type mice of either age group (Fig. 3, A and B, and data not shown). Finally, based on our previous finding that TNF is permissive for growth factor-induced vascular permeability in the skin (7), we tested whether these tie2-TNF mice also displayed increased vascular permeability in response to thrombin. When the effect of thrombin in a modified Miles assay for vascular permeability in the ear was tested, the permeability-inducing effect of thrombin was doubled in transgenic tie2-TNF mice (Fig. 4). However, using FITC-dextran injections and intravital confocal microscopy in the kidney, we did not see baseline differences between tie2-TNF and wild-type mice (data not shown). Although we could not recapitulate our ear permeability finding in the kidney due to adverse effects of thrombin when injected systemically, when all data are taken together these mice show chronic features of endothelial cell activation, increased response to vascular permeability factors, and inflammation.

Fig. 2.

Endothelial location of ICAM-1 in kidneys from tie2-TNF mice. Kidney sections from perfusion-fixed animals of tie2-TNF or wild-type background were incubated with monoclonal antibody anti-platelet endothelial cell adhesion molecule (PECAM; red) and for colocalization with monoclonal anti-ICAM-1 antibody (green) and observed by confocal microscopy. A: wild-type ICAM-1. B: tie2-TNF ICAM-1. C: tie2-TNF PECAM; D: tie2-TNF IgG background. E: tie2-TNF ICAM-1 and PECAM overlay. F: enlargement of inset in E. Scale bars = 50 μm except 100 μm in F. Data shown are from a representative experiment performed with a group size of n = 8 with similar results.

Fig. 3.

Amelioration of leukocyte extravasation in tie2-TNF mice after soluble thrombomodulin (sTM) treatment. Kidney sections from wild-type (A and C) or tie2-TNF (B and D) mice without (A and B) or with sTM treatment (C and D) were prepared and stained with a rat anti-mouse CD45 antibody using the AEC chromogen method. CD45-positive cells were quantified in a double-blinded fashion from at least 125 paraffin sections (E) representing n = 8/group. Scale bars = 20 μm.

Fig. 4.

Increased reactivity to thrombin in tie2-TNF mice. tie-2 TNF mice (age 3 mo) were tested for vascular permeability by injecting FITC-BSA into the tail vein followed by interstitial local injection of thrombin (100 U) into one ear and saline into the other. Dye was isolated from ears, and the fluorescence intensity ratio between thrombin- and saline-treated ears was employed to measure fold-induction of vascular permeability. Data shown are an average from a group size of n = 8.

sTM reduced chronic inflammation in kidneys of tie2-TNF transgenic mice.

To test the hypothesis that the anti-inflammatory protection by thrombomodulin (41) could be rescued by administration of recombinant thrombomodulin, we treated animals long term with recombinant murine sTM. Mice were euthanized, and the kidney sections were analyzed for CD45-positive leukocytes. Leukocyte infiltration was remarkably reduced after treatment with sTM in tie2-TNF transgenic mice (Fig. 3D, TNF+sTM) compared with saline-treated control tie2-TNF mice (Fig. 3B, TNF+Ctr). Of note, no infiltrates were visible in wild-type mice with or without sTM treatment (Fig. 3, A and C).

sTM reduced markers of nephropathy in tie2-TNF mice.

Having shown that sTM can reduce chronic inflammation in tie2-TNF mice, we next tested whether murine sTM treatment could also normalize markers of nephropathy. In fact, 3-mo-long treatment with sTM normalized renal function as evidenced by a significant reduction in the urinary albumin/creatinine ratio (Fig. 5A), decreased fibrosis (Fig. 5B), as well as a significant reduction in interstitial glomerular thickening (Fig. 5C) compared with saline-treated control tie2-TNF mice, as shown in Fig. 1. Furthermore, mRNA analysis from kidney extracts demonstrated that treatment with sTM suppressed gene transcripts shown to be elevated in tie2-TNF mice, including inflammatory chemokines CCL5, CXCL11, IL-18, and VCAM-1 as well as markers of kidney injury (cystatin C, osteopontin and TGFβR2; see heat map of regulated genes and representative individual regulation in Fig. 6).

Fig. 5.

Amelioration of albuminuria and fibrosis in sTM-treated tie2-TNF mice. Both albumin normalized to creatinine levels from urine samples (A) and interstitial glomerular fibrosis from trichrome-stained kidney sections (B) were assessed in tie2-TNF (TNF) or Wt mice without and with sTM treatment. ACR, albumin/creatinine ratio. Glomerular thickness scores (11) are shown in C. Data shown are an average from a group size of n = 8. Scale bars = 100 μm.

Fig. 6.

Profile of nephropathy and proinflammatory genes in kidney sections of sTM-treated tie2-TNF mice. A: heat map of kidney disease- and inflammation-relevant genes altered in tie2-TNF mice vs. wild-type mice without or with sTM treatment. B: downregulation of representative genes implicated in inflammation (VCAM-1, IL-18) and kidney remodeling [transforming growth factor-β receptor 2 (TGFβR2)] are shown.

sTM downregulated chemokine and cell adhesion molecule expression in tmTNF-expressing endothelial cells.

Based on the evidence that the sTM has anti-inflammatory properties (8) and downregulates chemokine and vascular adhesion molecules in vivo (Figs. 5 and 6), we tested whether sTM could reduce the expression of the vascular adhesion molecules and chemokines in cultivated endothelial cells expressing cellular TNF. Indeed, in vitro the expression of the same proinflammatory cell adhesion and chemokine molecules identified in the kidney of tie2-TNF animals was significantly reduced in TNF-expressing endothelial cells upon pretreatment with murine sTM (Fig. 7A). These genes included the vascular adhesion molecules VCAM-1 and ICAM-1 as well as the chemokines RANTES and MCP-3. In addition to this, we performed Western blot analysis of total cell lysates from treated and untreated TNF-expressing endothelial cells. As shown in Fig. 7B, pretreatment with sTM for 24 h significantly reduced the release of RANTES in the supernatants of TNF- expressing endothelial cells.

Fig. 7.

Normalization proinflammatory molecules in sTM-treated TNF endothelial cells. A: quantitative real-time PCR analysis of inflammatory gene transcripts in control (Ctrl-EC) and activated endothelial cells (TNF-EC) without and with sTM treatment was determined. Data are shown as normalized mRNA fold-change to control cells. B: Western blot analysis for RANTES. Immunoblot analysis was performed as described in materials and methods with anti-RANTES antibody, stripped, and reprobed with GAPDH antibodies. Quantification of fold-change in protein expression normalized to GAPDH from 3 different experiments was analyzed by National Institutes of Health Image J densitometry analysis software. Data shown are from a representative experiment and repeated at least 3 additional times with similar results.

sTM reduced enhanced transendothelial migration of monocytes through TNF-expressing endothelial cells.

To obtain more mechanistic insights underlying the finding that sTM can protect against chronic inflammation and nephropathy, we used transmembrane TNF-expressing endothelial cells, which are deficient in thrombomodulin. Because both chemokines and cell adhesion molecules ICAM-1 and/or VCAM-1 are a prerequisite for the adhesion and subsequent transendothelial migration of mononuclear cells at the sites of inflammation (43), we tested their involvement in transendothelial migration of monocytes across TNF-expressing endothelial cells and control cells in an in vitro transmigration assay. The human monocytic leukemia cell line THP-1 exhibited only a low spontaneous migration through an endothelial monolayer of control cells, which is dramatically increased in TNF-expressing endothelial cells (Fig. 8, A and B). Next, we tested the general involvement of chemokines in this enhanced migration through TNF-expressing endothelial cells by employing PTX, an inhibitor of chemokine receptor function. Indeed, pretreatment of monocytes and lymphocytes with PTX blocked transmigration across TNF-expressing endothelial cells, indicating participation of chemokines in this process (Fig. 8A). Furthermore, transendothelial migration of THP-1 cells across TNF-expressing cells was reduced to ∼50% with neutralizing antibodies to VCAM-1 and by ∼90% inhibition with antibodies to ICAM-1 (Fig. 8B).

Fig. 8.

Amelioration of increased migration of leukocytes across continuously activated endothelial cells dependent on cooperation of chemokine signaling and vascular adhesion molecules by sTM. A: increased migration of THP-1 monocytic cells across TNF-expressing endothelial cells can be abrogated by pretreatment of monocytes with pertussis toxin (PTX). B: increased migration of THP-1 monocytic cells across TNF-expressing endothelial cells can be blocked by pretreatment with neutralizing antibodies to ICAM-1 and VCAM-1. C: amelioration of increased migration of monocytes across activated endothelium without and with sTM treatment is shown. D: transmigration of monocytes across activated endothelium is independent of hirudin. The data presented are from a representative experiment performed in triplicate, repeated 4 times independently with similar results.

Next, we tested the involvement of sTM in reducing the expression of VCAM-1, ICAM-1, and chemokines in TNF-expressing endothelial cells, which may explain the reduced transmigration of monocytes across monolayers of these cells. In fact, pretreatment of endothelial cells with sTM dramatically reduced monocyte migration in TNF-expressing endothelial cells (Fig. 8C). This effect of sTM did not appear to be due to its ability to modulate thrombin, as the addition of the potent thrombin-inhibitor hirudin was unable to affect monocyte transmigration in TNF-expressing endothelial cells (Fig. 8D).


Our results show that continuous endothelial activation is sufficient to cause chronic inflammation in the kidney and increase markers of nephropathy in mice. These results are in line with the hypothesis that chronic endothelial activation is involved in CKD. Importantly, this proinflammatory endothelial activation and nephropathy can be abolished with prolonged treatment with sTM. Our findings contribute to understanding the role of the endothelium in CKD and suggest that pathological processes initiated in the endothelium may be sufficient to cause a cascade of proinflammatory events leading to renal injury and dysfunction. Thus novel therapeutic agents targeting endothelial activation, including sTM, may be important for treating CKD.

Our interpretation that sTM may ameliorate CKD development was based on its ability to normalize inflammation, fibrosis, albuminuria, and markers of kidney diseases in our model of endothelial-only expression of TNF. In addition to markers of kidney dysfunction (cystatin C) and tissue remodeling (osteopontin and TGFβR2), markers of inflammation (RANTES, ITAC, VCAM-1, IL-18) were most significantly upregulated with chronic inflammation in tie2-TNF mice and downregulated with sTM therapy. This is in line with the finding that inflammatory markers such as IL-18 and TNF (which we used in our model) are increased in the serum of patients with diabetic nephropathy, the most common type of CKD progressing to terminal renal failure (18). This occurs at a very early stage of disease and correlates with the degree of albuminuria (14). Although the pathogenesis of kidney diseases, and in particular of diabetic nephropathy, is multifactorial, local inflammatory stress may contribute and pharmacological interventions such as angiotensin-converting enzyme inhibitors or angiotensin receptor blockers (42), aldosterone antagonists (14), and PKC inhibitors that display anti-inflammatory properties in addition to normalizing vascular and endothelial barrier integrity (4, 14) have been shown to be beneficial. Finally, the endothelial cell-specific growth and survival factor angiopoietin-1 can abrogate endotoxin-induced acute kidney injury (22). Thus with its ability, angiopoietin-1 can activate the tie-2 receptor, compete for binding with the tie2 antagonist angiopoietin-2 to reduce endothelial cell reactivity to TNF (13). In line with these findings, a positive correlation of angiopoietin-2 and a reciprocal correlation of angiopoietin-1 with kidney diseases have been demonstrated (9, 10, 19, 23, 50).

Endothelial dysfunction has been demonstrated to be a starting point in vascular changes leading to other chronic diseases, in particular cardiovascular diseases such as atherosclerosis and artery calcification (20, 33). The vascular endothelium senses a variety of stimuli, and, in response, the endothelial cells produce a range of compounds which modulate vascular tone, coagulation, cell proliferation, and inflammation (6). Oxidative stress by reactive oxygen species (ROS) reportedly is increased in vessels associated with chronic inflammatory diseases, leading to nitric oxide (NO) quenching and subsequent endothelial dysfunction. Since physiological flow is essential for NO production, supplemental l-citrulline has been suggested as a therapeutic adjunct in disease states associated with l-arginine deficiencies based on its ability to restore NO signaling in endothelial cells (40). In excess, these ROS are believed to induce cytokines and profibrotic growth factors implicated in the pathogenesis of chronic diseases including TGF-β, TNF, IL-18, and cell adhesion molecules (12). In this context, we have previously demonstrated that ROS influence intracellular signaling and gene expression, including TNF expression in human endothelial cells, which leads to a positive feed-forward loop involving TNF-induced NADPH oxidase-dependent ROS production (34, 35, 37). Increased cell adhesion molecules have been shown to be required for leukocyte emigration and subsequent damage of the local tissues in inflammatory diseases. Our studies using activated endothelial cells demonstrating increased emigration of leukocytes that is dependent on VCAM-1/ICAM-1 and subsequent decrease with sTM (Fig. 8) in vitro and a further confirmation of amelioration of CD45-positive leukocyte extravasation in tie2-TNF kidneys in vivo with sTM suggest a therapeutic role for sTM in CKD.

Endothelial injury was suggested to link kidney injury and the development of nephropathy (1). Early alterations in peritubular capillary blood flow during reperfusion has been documented and associated with loss of normal endothelial cell function, which can be replaced pharmacologically or with cell replacement interventions. Modified urinary fluid shear stress induced by variations of urinary fluid flow and composition is observed in early phases of most kidney diseases. Interestingly, renal tubular fluid shear stress promotes endothelial TNF production and cell activation, which supports our rationale for choosing endothelial TNF expression as a model for chronic inflammation in the kidney (26).

Previously, we have shown that TNF sensitizes endothelial cells to respond to stimuli by increasing their vascular permeability (7). In this study, we also found that tie2-TNF mice displayed increased sensitivity to vascular permeability in the ear when induced by thrombin. Of note, these transgenic mice did not display significantly increased vascular leakage in the absence of stimulation. It is likely that endothelial activation and vascular dysfunction increase local thrombin generation, which together with increased endothelial responsiveness can contribute to increased vascular permeability and interstitial changes in the kidney. Although we could not test the activity of thrombin in the kidney due to its procoagulant adverse activity, it is possible that at least some of the normalizing effects of sTM in kidney pathophysiology of the tie2-TNF mice were caused by direct inactivation of thrombin. This indicates that systemic therapy with sTM may have wide protective effects, spanning from reducing vascular permeability to blocking inflammation.

The finding that sTM can reverse the kidney phenotype observed in an endothelial activation-induced model of chronic inflammation is in accordance with previous demonstrations that sTM administration can prevent acute kidney injury caused by ischemia (41) and that thrombomodulin and a lectin binding domain of sTM have anti-inflammatory activities via direct action on endothelial cells (8). During inflammation, circulating sTM has been shown to increase in human plasma, cleaved by soluble TNF-α and normalized by the anti-inflammatory compound pentoxifylline (3, 24, 28). Although we observed a decreased thrombomodulin staining in the kidneys of tie2-TNF mice, we could not detect increased levels of sTM in plasma (data not shown). This could be explained by different transcriptional vs. posttranslational regulation (24) in chronic inflammation. While sTM has been well studied as a marker of endothelial activation, its pathophysiological role is unknown (16, 24, 28, 45). Our results of chronic administration of sTM primarily suggest that therapeutic intervention with circulating sTM may play a protective role against endothelial cell-dependent inflammation. However whether such regulation of endogenous thrombomodulin is involved in chronic inflammatory processes would need to be addressed by back crossing tie2-TNF mice into mice with thrombomodulin deficiency (47). This is expected to display an earlier and more pronounced onset of proinflammatory activity and tissue damage and may provide further evidence for the role of thrombomodulin.

Our tie2-TNF model used a background (C57BL/6), which is known to be highly resistant to the induction of kidney disease (15, 32). As such, we were not surprised that the albuminuria seen was relatively mild compared with a more direct renal injury model. Pathological alterations of the kidney generally precede the clinical development of microalbuminuria in patients with type 1 diabetes, and these alterations are even more significant by the relatively late appearance of overt macroalbuminuria (5). In addition, the mild kidney phenotype of the tie2-TNF model resembles features of the streptozotocin-induced rodent models of diabetes, in which albuminuria is also seen only at moderate levels (18, 27).

The exact mechanism by which sTM exerts its protective effect in the tie2-TNF mouse is unknown. In addition to APC generation, which depends on procoagulant thrombin formation, sTM also appears to directly act on endothelial cells. Indeed, our findings that sTM reduced adhesion molecule, chemokine gene expression, and monocyte transmigration in continuously activated, but not in control endothelial cells, support the hypothesis that the observed effects of thrombomodulin in vivo are due to a direct effect on activated endothelium. Although further studies identifying the receptor(s) mediating these activities in endothelial cells are needed, this is the first demonstration of sTM downregulation of CKD characteristics in a murine model of chronic endothelial activation via specific adhesion molecules and chemokines. Thus sTM emerges as a novel nonsteroidal- and nonarachidonic acid-targeting anti-inflammatory agent. Given the side effects of previously described anti-inflammatory pharmacological treatment modalities, sTM may offer a promising alterative for suppression of endothelial cell activation.


This research was supported by an unrestricted research grant from Eli Lilly, Co. to M. Clauss, National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grant DK-63114 to D. P. Basile, Altana Pharma (now Nycomed) support to M. Clauss, the Cryptic Masons to the Indiana Center for Vascular Biology and Medicine to G. Rajashekhar, the NIDDK (79312) George M. O'Brien Award to the Indiana Center for Biological Microscopy, and NIDDK Grants 77124 and 79312 to T. A. Sutton.


No conflicts of interest, financial or otherwise, are declared by the authors.


Author contributions: G.R., A.G., A.M., J.L.F., A.W., D.T.B., M.S.C., G.E.S., T.A.S., and D.P.B. performed experiments; G.R., A.G., A.W., G.E.S., T.A.S., D.P.B., and M.C. analyzed data; G.R., A.G., G.E.S., T.A.S., D.P.B., and M.C. interpreted results of experiments; G.R. and G.E.S. prepared figures; G.R. and M.C. drafted manuscript; G.R., D.P.B., B.G., and M.C. edited and revised manuscript; G.R., B.G., and M.C. approved final version of manuscript and provided conception and design of research.


We are thankful to Dietmar Vestweber (Münster, Germany) for providing neutralizing anti-ICAM-1 and anti-VCAM-1 antibodies. The technical assistance of Nagesh Gollahalli is gratefully acknowledged.

Present address of A. Willuweit: Forschungszentrum Juelich, Institute of Neuroscience and Medicine, Juelich, Germany.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
View Abstract