Skip Navigation


NDT Advance Access originally published online on November 28, 2007
Nephrology Dialysis Transplantation 2008 23(4):1144-1156; doi:10.1093/ndt/gfm774
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
23/4/1144    most recent
gfm774v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Hohenstein, B.
Right arrow Articles by Hugo, C. P. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hohenstein, B.
Right arrow Articles by Hugo, C. P. M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author [2007]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org


A murine model of site-specific renal microvascular endothelial injury and thrombotic microangiopathy

Bernd Hohenstein1, Andrea Braun1, Kerstin U. Amann2, Richard J. Johnson3 and Christian P. M. Hugo1

1 Department of Nephrology and Hypertension, University Erlangen-Nuremberg, Erlangen, Germany 2 Division of Pathology, University Erlangen-Nuremberg, Erlangen, Germany 3 Division of Nephrology, Hypertension, and Renal Transplantation, University of Florida, Gainesville, FL, USA

Christian Hugo, Department of Nephrology and Hypertension, University Erlangen-Nuremberg, Loschgestrasse 8, 91054 Erlangen, Germany. Tel: +49-9131-8539002; Fax: +49-9131-8539209; E-mail: christian.hugo{at}rzmail.uni-erlangen.de



   Abstract
Top
 Abstract
Introduction
 Material and methods
 Results
 Discussion
 References
 
Despite the importance of endothelial injury and healing for primary and secondary renal disease and the availability of genetically engineered mouse models, to date no generally applicable murine disease model with site-specific renal endothelial injury has been established. We induced specific microvascular renal injury via selective renal arterial perfusion of the lectin concanavalin A (Con A) followed by sheep anti-concanavalin A and harvested tissues after 4 h, 24 h, days 3 and 7. Compared to control kidneys, histological evaluation demonstrated endothelial cell injury with subsequent complement, and platelet activation and thrombosis by light and electron microscopy. Mouse kidneys showed histologic evidence of severe glomerular and peritubular microvascular thrombosis with acute tubular necrosis, proteinuria, increased BUN and presence of schistocytes. Initial cell death of intrinsic renal cells resulted in a decrease of the glomerular cell count by 50% after 4 h followed by a proliferative response of glomerular (day 3, P < 0.05), interstitial (day 3, P < 0.05) and tubular cells leading to increased total glomerular cell count by day 7. This study establishes the Con A anti-Con A model as specific endothelial injury model in the mouse kidney providing a novel tool for investigating endothelial injury and repair mechanisms as well as elucidating the role of platelets in genetically engineered mice.

Keywords: Con A; platelets; renal endothelial injury; thrombotic microangiopathy



   Introduction
Top
 Abstract
 Introduction
Material and methods
 Results
 Discussion
 References
 
The renal endothelium represents the major nutritional barrier and functional surface between the blood and other renal cell types and is the predominant site of injury in diseases such as HUS/TTP-syndrome, eclampsia, systemic sclerosis, malignant hypertension, vasculitis and vascular rejection after renal transplantation causing thrombotic microangiopathy in the kidneys. Mechanisms of development of thrombotic microangiopathy (factors H and I, ADAMTS 13) [1,2] are increasingly considered in human disease. Nevertheless, the pathophysiology and time course of progression versus healing of microangiopathic renal lesions once disease is established is largely unknown, since very few disease models in general and especially no murine model of thrombotic microangiopathy is available to date. The importance of endothelial injury and repair is even greater, since almost any experimental kidney disease model showed also secondary endothelial injury. Hereby, these experimental studies demonstrated a close link between the progressive loss of peritubular capillaries and progression of renal disease as well as renal scarring and fibrosis [3–5].

In addition, platelets, as blood cells highly interacting with the endothelial surface, have been investigated in several experimental studies and are regarded as an important inflammatory cell type during glomerulonephritis [6–10]. Nevertheless, the role of platelets and platelet-endothelial cell interaction in renal thrombotic microangiopathy has not been systematically investigated. Furthermore, therapeutic options to modulate platelet as well as endothelial cell properties after injury as interventional strategies during primary and secondary renal microvascular lesions are not existent. This most likely relates to the fact that studies in human disease are problematic due to the relatively low frequency and variability of each disease entity. While existing rat endothelial disease models suggest a major contribution of platelet activation and platelet–endothelial interaction to the time course of renal thrombotic microangiopathies, the spectrum of research regarding novel therapeutic targets and principles is much more limited when compared to murine models due to less developed manipulatory genetic techniques. In mice, many gene-deficient and transgenic models have been developed and many molecules influencing platelet and endothelial cell properties have been modified [11–14], but to date no mouse model of site-specific renal microvascular injury has been described.

The present study introduces the Con A anti-Con A model as specific renal endothelial injury model in the mouse resulting in severe thrombotic microangiopathy. Hereby, it offers the possibility to systematically investigate the pathophysiological role of endothelial and also platelet functional molecules and potentially identify new therapeutic targets for renal microvascular injury.



   Material and methods
Top
 Abstract
 Introduction
 Material and methods
Results
 Discussion
 References
 
Description of the animal model
This disease model is adopted and modified from the immune complex models described by Golbus and Wilson [15] and Johnson et al. [7] in the rat kidney. Endothelial injury results from the binding of the lectin concanavalin A (Con A; Sigma-Aldrich, Taufkirchen, Germany), a protein extract from the jack bean Concanavalia ensiformis, to endothelial glycoproteins and the subsequent perfusion of the kidney with an anti-Con A antibody. Studies using ferritin-labelled Con A have shown that this lectin specifically binds to the endothelium [16]. The subsequent infusion of anti-Con A antibody results in immune complex formation and complement activation on the endothelial surface.

Renal arterial perfusion in mice
Male C57/bl6 mice (Charles River WIGA, Sulzfeld, Germany) at the age of 12–14 weeks were used. Mice were anaesthetized using isoflurane. A paravertebral incision at the left flank was made, the left kidney was mobilized and surrounding fat was carefully removed. After cautious preparation of the left renal artery and the abdominal aorta the left renal artery was cannulated via the abdominal aorta using a 31-gauge micro-cannula (FST, Heidelberg, Germany). This micro-cannula has a tip length of 1 mm and fits the inner diameter of the ~4-mm-long and 0.4-mm-wide renal artery. The micro-cannula has a luer lock and thereby was connected to a short piece of an infusion line which itself was adapted to a 20-gauge intravenous cannula (both from Braun, Melsungen, Germany). The complete device has a dead space of 150 µl and enables the consecutive injection of PBS, Con A and anti-Con A by an assisting person without significant movement of the micro-cannula after insertion into the renal artery. The complete procedure was performed under view with a Zeiss OPMI 1-FC operation microscope. Since the micro-cannula filled more than 90% of the murine renal artery, the perfusion procedure was performed without clamping of the arterial blood flow. Perfusion was initiated by application of 150 µl of PBS to remove blood cells from the kidney. The kidney was then perfused with 200 µg of Con A followed by 150 µl PBS. Next, 18 mg of a sheep anti-Con A antibody was applied followed by 150 µl PBS. After the perfusion procedure, the abdominal aorta was clamped for vessel closure for <3 min. Afterwards, the blood flow was restored by unclamping the aorta and kidney perfusion as well as the absence of bleeding was ensured visually. The wound was then closed with a continuous suture and mice were placed under a heat lamp for 1 h.

Experimental design
For the present study, a sheep anti-Con A IgG antibody was used after preparation as previously described [7]. Two different, independent experiments were performed. An initial low-dose model was induced in 18 C57Bl/6 mice by the perfusion of 100 µg Con A followed by 4.5 mg anti-Con A. An additional 18 mice underwent renal arterial perfusion with 100 µg Con A followed by PBS to serve as controls. In addition, the right non-perfused kidneys of the Con A anti-Con A mice also served as controls. Since this model system only induced some transient but specific endothelial activation but not severe injury in the perfused Con A anti-Con A kidneys, a second high-dose experiment was performed in 20 mice and 25 controls: Con A anti-Con A disease was induced in 20 mice, while 25 mice served as controls. In diseased mice, renal arterial perfusion of 200 µg Con A was followed by 18 mg anti-Con A. PBS was administered in lieu of anti-Con A in 20 mice serving as controls. Five control mice received anti-Con A antibody alone. Kidneys were harvested at 1, 4 and 24 h in the low-dose and 4h, 24 h, day 3 and day 7 in the high-dose model. All animals were fed standard mice chow (Altromin 1324, Spezialfutterwerke GmbH, Lage, Germany) and tap water ad libidum.

The day before sacrifice, a 24-h urine collection to assess proteinuria was performed. At sacrifice, mice were anaesthetized by isoflurane and blood was collected via puncture of the inferior caval vein and a blood smear was performed immediately. Mice were perfused via the heart with 0.9% (w/v) NaCl solution and kidneys were harvested for immunohistochemical analysis and electron microscopy. Serum and urine samples were stored at –70°C until analysis.

Tissue processing and immunohistochemical staining
Tissue for light microscopy was fixed in methyl Carnoy`s solution or 3% paraformaldehyde, embedded in paraffin, and cut into 3 µm sections for indirect immunoperoxidase staining as described elsewhere [17,18].

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 [18] to detect actively proliferating cells (PCNA; Chemicon, Temecula, CA, USA); biotinylated lectin from Lycopersicon esculentum (tomato) for staining of glomerular and peritubular capillaries (Sigma–Aldrich, Munich, Germany) [19]; F4/80, a murine IgG1 mAb to a cytoplasmic antigen present in monocytes, macrophages and dendritic cells (Caltag Lab., Burlingame, CA, USA); A-20, a rabbit mAb specific for the detection of all VEGF isoforms (Santa Cruz Biot., CA, USA) [20] and a rat anti-mouse glycoprotein Ib antibody for specific staining of platelets (gift from B. Nieswandt, Virchow-Institute, Wuerzburg, Germany) [21]. 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. Afterwards specific biotinylated secondary antibodies (all by Zymed, San Francisco, CA, USA) were applied followed by peroxidase conjugated Avidin D (Vector Lab., Burlingame, CA, USA) and colour development with diaminobenzidine.

Tissues for immunofluorescent staining were embedded in OCT (Lab-Tek Products, Naperville, IL, USA) and snap frozen in liquid nitrogen. Tissue sections were incubated overnight with the following primary and secondary antibodies: a specific Alexa Flour 555 donkey anti-sheep antibody (Invitrogen, Karlsruhe, Germany) for evaluation of the binding of the disease inducing antibody; a FITC-conjugated antibody against complement C3 (Cappel, ICN Biomedicals, Eschwege, Germany) and a monoclonal rat anti-mouse CD31 ab (Chemicon, Temecula, CA, USA) followed by an Alexa-Flour 488 donkey anti-rat ab (Invitrogen, Karlsruhe, Germany). A rabbit anti-Con A serum was generated in a New Zealand white rabbit to detect Con A bound to the renal endothelium. Therefore, the rabbit was 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, sacrificed under general anaesthesia and bled by heart puncture. After centrifugation of the blood at 2000 g for 15 min and removal of clotted blood cells, the serum complement inactivated at 56°C for 30 min.

Quantitative analysis of immunostaining and capillary rarefaction
Glomerular expression/deposition of C3 was quantified using computer-assisted image analysis software (MetaVue, Visitron Systems, Puchheim, Germany). At least 50 glomeruli per section were analysed using a 400-fold magnification. C3 positivity is given as percent positive area per glomerular cross-section. Rarefaction of CD31-positive capillaries was assessed semiquantitatively after fluorescence staining for PECAM/CD31 on frozen sections using a score from 0 to 4. A score of 0 was assigned if a homogeneous positivity of all peritubular capillaries was present (normal finding). Scoring was 1, if single CD31-negative segments could be detected; scoring was 2, if CD31 was negative in up to 25% of the tissue section, 3 if CD31 was negative in 25–50% of the tissue section and 4 if CD31 staining was completely absent. A higher score, therefore, reflects a higher degree of capillary loss. In addition, peritubular and glomerular capillary loss were examined by staining for lectin from L. esculentum (tomato) as previously described [22,23]. Peritubular capillary rarefaction was evaluated through a 10 x 10 eyepiece grid using a x40 objective in at least 20 cortical fields. The grid covers an area of 0.0625 mm2 at this magnification. Each square that contained no lectin-positive capillary was counted. This scoring system inversely reflects peritubular capillary rarefaction, whereby low values represent intact capillarization and higher values indicate loss of capillaries (maximum is 100). Glomerular capillary rarefaction was determined using a semiquantitative scoring system from 0 to 3 where 0 means 0–25%, 1 means 25–50%, 2 means 50–75% and 3 means 75–100% loss of glomerular capillaries. At least 50 glomeruli were evaluated. PCNA-positive cells were counted in 50 consecutive glomeruli and in 20 cortical fields of vision for assessment of the tubulointerstitium under a 400-fold magnification. Glomerular cell number was assessed in PAS-stained sections in 50 consecutive glomeruli under a 400-fold magnification. Glomerular and peritubular platelet infiltration were assessed after staining for glycoprotein Ib using a semiquantitative scoring system from 0 to 4, where 0 means absence of platelets, 1 refers to the presence of glomerular platelets in <10% of all glomeruli, 2 for the presence of platelets in up to 50% of glomeruli, 3 when >50% of glomeruli contained platelet thrombi, often with peritubular thrombi, and 4 for severe (sub)total glomerular and peritubular thrombosis [23]. Each biopsy was completely evaluated reflecting 12 cortical fields of vision at a 200-fold magnification. Fibrin deposits were identified on tissue sections stained with acid fuchsin orange G (AFOG), where fibrin is indicated by an intense orange-red colour [24]. Glomerular fibrin deposition was evaluated in at least 50 randomly selected glomeruli under a x400 magnification. Therefore, we counted the number of glomeruli containing fibrin deposits and calculated the percentage of positive glomeruli. Cortical VEGF, sheep-IgG and Con A positivity were assessed in at least 20 cortical fields of vision at a 400-fold magnification using computer-assisted image analysis (Metavue). Data are given as percent positive area per cortical field vision. F4/80-positive infiltrating monocytes/macrophages were counted under a 400-fold magnification in 50 consecutive glomeruli and in 20 cortical fields to assess the tubulointerstitium. Invading neutrophils were counted after PAS staining in 50 consecutive glomeruli under a 600-fold magnification.

TUNEL assay
Apoptotic cells were detected by the terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-biotin nick end labelling assay as described previously [25]. The number of TUNEL-positive apoptotic cells were either counted in 50 sequentially selected glomeruli and given as the mean number per glomerular cross-section, or per mm2 after counting 20 peritubular areas of each section.

Staining and evaluation of blood smears
Blood smears were done immediately after the blood was drawn, air dried and fixed in 100% methanol. After fixation, blood smears were stained according to the Pappenheim method and evaluated using a 600-fold magnification.

Electron microscopy
For electron microscopy, renal tissues were fixed in a buffer containing 48 mM Na2HPO4 and 14 mM KH2PO4 containing 3% glutaraldehyde. Prior to electron microscopy semithin sections (0.5 µm) were made to depict morphological findings. In several randomly selected animals per group, ultrathin sections of the renal cortex were qualitatively investigated using a Zeiss EM 902 (Zeiss Co., Oberkochen, Germany) at various magnifications.

Statistical analysis
All values are expressed as mean ± SD. Statistical analysis was performed using a two-sided non-parametric Mann–Whitney U-test and statistical significance was defined as P <0.05 (* SPSS 12.0, SPSS Software, Munich, Germany).



   Results
Top
 Abstract
 Introduction
 Material and methods
 Results
Discussion
 References
 
Kidneys perfused in the low-dose Con A anti-Con A model demonstrated mild and short-lived endothelial activation as indicated by transient platelet infiltration limited to the first few hours after disease induction compared to contralateral kidneys and perfused kidneys of control animals (not shown). In the low-dose model, no severe signs of endothelial injury or renal thrombotic microangiopathy comparable to the high-dose model could be detected. Therefore, all data demonstrated below relate to experiments with the high-dose Con A anti-Con A model.

Homogeneous antibody binding after selective renal arterial perfusion in mice
Renal arterial perfusion could be performed with a low mortality. Death of mice usually resulted from a problematic anatomy, when branches of the renal artery left the common renal artery after a short distance. A minimum length of 1 mm of the common renal artery was needed to successfully place the micro-cannula and to perfuse kidneys. After injection of PBS, the complete kidney was bloodless and pale, whereas all other visible abdominal organs including bowel, liver and spleen demonstrated regular blood perfusion. In addition, injection of methylene blue as colour indicator demonstrated the effective perfusion of the left kidney without visible accumulation in other organs (not shown).

In the first step, animals were perfused with Con A alone. In these mice, Con A binding to the renal endothelium did not lead to endothelial damage as indicated by a lack of TUNEL positivity, complement activation or loss of CD31 positivity, but apparently led to some platelet activation as indicated by some increased influx of platelets (not shown).

In the second step, mice were perfused with anti-Con A antibody following Con A binding and were investigated regarding disease induction. Therefore, kidneys were stained using an anti-sheep antibody since after the induction sequence of Con A followed by sheep anti-Con A, the endothelium should be loaded with the sheep antibody. After 4 h, perfused (left) kidneys were positive in glomeruli and peritubular capillaries (Figure 1A) compared to the contralateral non-perfused kidneys and Con A alone perfused controls demonstrating specificity (Figure 1B: contralateral kidney). Kidneys perfused with anti-Con A antibody alone also demonstrated mild positivity of glomerular and peritubular capillaries most likely due to non-specific antibody deposition. Compared to controls, diseased kidneys demonstrated enhanced positivity for sheep-IgG up to day 3 (Table 1). Staining for Con A using a rabbit anti-Con A antibody demonstrates that the glomerular and peritubular endothelium is loaded with a high amount of Con A with the same staining pattern depicted in Figure 1A for sheep-IgG. In parallel to sheep-IgG (and thereby deposition of anti-con A), con A binding was increased up to day 3 compared to controls (Table 1). Enhanced positivity for sheep-IgG and Con A on day 3 related to an increased number of positively stained injured tubules.


Figure 1
View larger version (20K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1 Complete antibody binding and autologous immune response following renal arterial perfusion in mice. Staining for sheep-IgG was performed in frozen tissue sections of perfused (A) and control kidneys (B) after 4 h. Perfused left kidneys were not only positive in glomerular, but also peritubular capillaries after 4 h (A) compared to the control kidneys (B). The autologous immune response was evaluated using an antibody against complement C3. Glomerular and peritubular capillaries of Con A/anti-Con A-perfused kidneys were positive for complement C3 (C) whereas all controls revealed some background around Bowman's capsule and few interstitial areas but were clearly negative in glomeruli (D).

 

View this table:
[in this window]
[in a new window]

  		Table 1 Injury, repair and inflammatory response

 
Subsequently, local complement activation was assessed using an antibody against complement C3. Kidneys perfused with Con A followed by anti-Con A antibody were highly positive for C3 in glomerular as well as peritubular capillaries (Figure 1C). Control kidneys (contralateral, Con A and anti-Con A alone) demonstrated some background staining around Bowman's capsule and few interstitial areas but were negative in glomeruli for complement C3 by immunostaining (Figure 1D). C3 positivity in Con A anti-Con A-perfused kidneys was 37 ± 14% (positive area per glomerular cross-section) compared to negativity in controls (contralateral 0.4 ± 0.23, Con A 0.6 ± 0.4, anti-Con A 0 ± 0% positive area per glomerular cross-section; all P < 0.01). The increase of C3 does not necessarily prove the subsequent generation of C5b-9 membrane attack complexes, but their formation and relevance has been shown in the rat Con A anti-Con A model [26].

Selective endothelial cell activation and injury after disease induction leads to severe thrombotic microangiopathy
The endothelial cell marker CD31 is constitutively expressed on the endothelial surface with maximal expression at endothelial cell–cell contacts [11]. Staining for CD31 demonstrated a dimorphic pattern, with both focal areas of increased staining and areas of absent staining at 4 h (Figure 2A, left half with increased CD31 intensity, right half with absent CD31 signal). In contrast, controls demonstrated a very homogeneous endothelial staining pattern in glomerular and peritubular capillaries (Figure 2B). In parallel, glomerular and peritubular capillary loss was evaluated after staining for tomato lectin. Arrows in Figure 2C depict glomerular capillary rarefaction in a diseased kidney. Figure 2D visualizes the assessment of peritubular capillary rarefaction using a 10 x 10 eyepiece grid where every square without a capillary section was counted (score here 79). Quantitation of CD31-negative areas (2E) as well as evaluation of peritubular (2F) and glomerular (2G) capillary rarefaction after lectin staining documented the presence of significant endothelial injury at all time points compared to contralateral kidneys.


Figure 2
View larger version (54K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2 Disease induction results in endothelial injury. Staining for platelet endothelial cell adhesion molecule 1 (PECAM-1/CD31) was performed on frozen kidney sections to evaluate endothelial injury and capillary rarefaction. Diseased kidneys demonstrated absence of CD31 staining in many areas of the kidney (A, right half of the image) indicating the absence of CD31-positive endothelium next to areas with increased CD31 intensity (A, left half of image). Control kidneys demonstrated a very homogeneous and scattered staining pattern without any differences in the expression or distribution (B). Staining of glomerular (C) and peritubular capillaries (D) in paraffin-embedded tissues using tomato lectin verified the findings of the CD31 staining. Arrows in C indicate areas with glomerular capillary loss. (D) Visualizes the scoring using a 10 x 10 eyepiece grid (score here 79). Evaluation of endothelial injury score was assessed using a semiquantitative scoring system reflecting the loss of CD31 positivity of capillaries (E). In parallel, rarefaction of peritubular (F) and glomerular (G) capillaries was assessed in lectin stained tissue sections (P < 0.05: *versus both controls).

 
Diseased kidneys demonstrated a pronounced platelet influx [glycoprotein Ib (GP Ib) positivity] in almost all glomeruli and in many peritubular capillaries (Figure 3A, B zoomed) after 4 h. Contralateral controls demonstrated no platelet influx (Figure 3C), whereas kidneys perfused with Con A alone demonstrated some glomerular, but never peritubular platelet accumulation. Accumulation of glomerular and peritubular platelets was observed up to 72 h but markedly decreased on day 7 (Figure 3D). Fibrin thrombi (identified by AFOG staining as an orange colour, Figure 3E) could also be demonstrated consistent with the histology of severe thrombotic microangiopathy (Figure 3F).


Figure 3
View larger version (112K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3 Increased platelet activation after endothelial injury leads to glomerular and peritubular capillary thrombosis. Staining for glycoprotein Ib was performed on frozen tissue sections. After 4 h, disease induction led to severe glomerular and peritubular platelet accumulation following endothelial injury (A, B zoomed). Control kidneys had no platelet positivity (C). The score for platelet accumulation was assessed as described in the ‘Material and methods’ section and reflects the amount of glomerular and peritubular platelet thrombus formation (D). Glomerular fibrin deposition was assessed using an acid fuchsin orange G (AFOG) stain and was detected by its bright orange colour (E). Evaluation was performed using a semiquantitative score (F) (P < 0.05:* versus both controls, +P < 0.05 versus right kidney).

 
Macro- and microscopy demonstrate glomerular and peritubular thrombotic microangiopathy
In contrast to control kidneys, all diseased kidneys perfused with the Con A anti-Con A sequence were enlarged at sacrifice without any necrotic areas (Figure 4A).


Figure 4
View larger version (115K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4 Macro- and microscopy demonstrate severe endothelial injury and platelet activation. After 4 h and up to 3 days all high-dose model kidneys demonstrated a characteristic macroscopic finding (A). The microscopic findings are depicted on semithin sections (0.5 µm) for diseased kidneys (B). Arrows in (B) indicate glomerular (arrow 1) and peritubular (arrow 2) thrombi. PAS-stained sections demonstrate the severity of this model after 4 h (C) compared to controls (D). After 3 days (E) and 7 days (F) kidneys still demonstrated severe glomerular and tubulointerstitial injury. Con A-perfused control kidneys (G) and right kidneys (H) demonstrated normal histology on semithin sections.

 
By light microscopy, both glomerular (Figure 4B, arrow 1) and peritubular capillaries displayed thrombi (Figure 4B, arrow 2), interstitial oedema and tubular necrosis 4 h after disease induction. Glomeruli demonstrated collapsing capillary loops and a large number of vacuolizations (Figure 4C) 4 h after disease induction compared to controls (Figure 4D). As depicted in Figure 4E, kidney sections on day 3 were still characterized by PAS-positive material in capillary loops and by signs of acute tubular necrosis despite ongoing repair. Areas with dilated tubules and residues of tubular necrosis could be detected up to day 7 in model kidneys (Figure 4F). Contralateral (Figure 4G) as well as Con A-perfused controls (Figure 4H) demonstrated normal histology.

Electron microscopy demonstrates severe endothelial injury shortly after disease induction
By EM, kidneys demonstrated endothelial cell swelling (Figure 5A–D) with the presence of cytoplasmic vacuoles and condensated nuclear chromatin (Figure 5B; x7000). The endothelial layer was widely destroyed and basement membranes demonstrated thickening and lamellation (Figure 5C; x20 000). In some capillaries swollen endothelial cells led to subtotal occlusion (Figure 5D). Control kidneys showed normal glomerular ultrastructure with widely open capillary spaces and the endothelial layer did not show any signs of injury (Figure 5E; x4400 and Figure 5F; x7000). In parallel, evaluation of peritubular capillaries demonstrated endothelial swelling and attachment of leukocytes as well as red blood cells (Figure 5G; x3000), whereas control kidneys demonstrated a regular ultrastructure (Figure 5H; x3000).


Figure 5
View larger version (125K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 5 Electron microscopy demonstrates the severe endothelial injury shortly after disease induction. Electron microscopy of model kidneys and controls was performed at various magnification. Starting 4 h after perfusion, kidneys of the high-dose model demonstrated severe endothelial injury (A; x3000). At higher magnifications endothelial cells appeared enlarged with cell swelling, cytoplasmic vacuoles and condensated nuclear chromatin (B; x7000). The endothelial layer was widely destructed and basement membranes demonstrated thickening and lamellation (C; x20 000). Swollen endothelial cells lead to subtotal occlusion of capillaries (D; 4400). Podocytes also demonstrated damage, i.e. cytoplasmic swelling with increased number of cytoplasmic vacuoles and foot process effacement (C; x20 000). Podocyte damage was most likely a secondary phenomenon. By contrast, perfused controls and right kidneys showed normal glomerular ultrastructure with widely open capillary spaces (E; x4400). The endothelial layer did not show any signs of injury (F; x7000). Endothelial injury was also detected in peritubular capillaries 4 h after disease induction as depicted in (G; x3000) with swollen endothelial cells and attached leukocytes and red blood cells. In contrast, control kidneys demonstrated a normal endothelial layer by electron microscopy (H; x3000).

 
Endothelial injury leads to glomerular and tubulointerstitial cell death and subsequent proliferation
Increased glomerular cell death could be detected after 4 and 24 h in diseased kidneys (4 h: 0.1 ± 0.07; 24 h: 0.58 ± 0.38) compared to contralateral (4 h: 0 ± 0; 24 h: 0.02 ± 0.04; P < 0.05) and Con A-perfused (4 h: 0 ± 0; 24 h: 0 ± 0; P < 0.05) control kidneys (Table 1) as indicated by TUNEL staining. TUNEL positivity was significantly increased in the interstitium of diseased kidneys at 4, 24 h and day 3 compared to the contralateral kidney (Table 1; all P < 0.05) and Con A-perfused control kidneys. Tubular injury with death of tubular cells also occurred in most diseased kidneys (1.3 ± 1.3 TUNEL + cells versus 0 ± 0 cells in contralateral kidneys and 0; 0 ± 0 cells in Con A alone perfused kidneys, 24 h, P < 0.05) (Table 1).

Since cell injury subsequently leads to the increase of cell repair, we further evaluated glomerular and tubulointerstitial cell proliferation as one major repair mechanism. Glomerular proliferation was decreased after 4 h in diseased kidneys (0.003 ± 0.008) compared to controls (contralateral: 0.04 ± 0.03; Con A: 0.05 ± 0.04 PCNA-positive cells per glomerular cross-section; both P < 0.05) and increased after 3 days compared to contralateral controls (0.7 ± 0.6 versus 0.03 ± 0.04; P < 0.05) This was also true for proliferating interstitial cells on day 3 (23.2 ± 21.9 versus 2.1 ± 1.4 PCNA-positive cells per mm2 of the contralateral kidney; P < 0.05; Table 1).

Parallel to the initial tubular injury, the number of proliferating tubular cells was reduced 4 h after disease induction compared to contralateral kidneys (0.6 ± 0.7 versus 2.5 ± 1.8 cells per mm2; P < 0.05; Table 1). The persistent presence and variability of tubular injury on days 3 and 7 might explain why the proliferation rate of tubular cells was not increased enough to lead to significant differences during tubular regeneration compared to control mice (Table 1).

To further evaluate the balance of glomerular cell death and regeneration, the glomerular cell count was assessed in diseased and Con A-perfused control kidneys after 4 h and 7 days. The glomerular cell number rapidly decreased by about 50% 4 h after disease induction (20.35 ± 2 versus 11.4 ± 3.1; P < 0.05; Table 1). On day 7, glomerular regeneration resulted in an increase of glomerular cells in diseased kidneys (25.9 ± 2.3) compared to Con A-perfused controls (18.5 ± 1; P < 0.05; Table 1).

VEGF expression is increased during the repair process
VEGF is induced in response to hypoxia and represents the major proangiogenic cytokine. We, therefore, assessed VEGF distribution by immunostaining during this disease model. While controls demonstrated VEGF mainly restricted to podocytes and tubules (4 h Con A control, Figure 6A), VEGF was upregulated on day 3 in animals with endothelial injury (Figure 6B, C). VEGF was lost in many severely injured tubules and glomeruli (podocytes), but in parallel strongly upregulated in less severely injured tubules and glomeruli (Figure 6B, C). This pattern could be detected up to day 3, when the VEGF increase was significant and mainly regenerating parts of the kidney demonstrated increased tubular and also tubulointerstitial VEGF expression (Figure 6D). This suggests a role of VEGF during regeneration of this disease model. In contrast, on day 3 many injured glomeruli still were negative for VEGF.


Figure 6
View larger version (101K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 6 VEGF is increased after induction of specific endothelial injury. Cortical VEGF expression was assessed after staining with an antibody detecting all VEGF isoforms. Controls demonstrated VEGF mainly restricted to podocytes and single tubules (A; x400). During disease up to day 3, VEGF was heterogeneously distributed within the cortex, whereas it was increased in areas of intermediate injury but decreased in areas of very severe tubular and glomerular injury with apparent cell death (B; x400). Regenerating tubules were highly positive for VEGF (C; x200). Cortical VEGF expression was quantitated during disease using computer-assisted image analysis (D) (P < 0.05: *versus both controls).

 
Influx of inflammatory cells after endothelial injury
Evaluation of infiltrating monocytes and macrophages in response to injury was done by staining for the F4/80 antigen. An increased number of F4/80-positive cells in glomeruli starting 4 h after disease induction and lasting up to day 3 (all P < 0.05) were observed compared to contralateral control kidneys (Table 1). In contrast, F4/80-positive cells in the tubulointerstitium were increased only on day 7 (12.1 ± 9.6) compared to contralateral (0.3 ± 0.2; P < 0.05) and Con A-perfused control kidneys (2.1 ± 2.5 cells per mm2; P < 0.05; Table 1). In parallel, an increased number of invading neutrophils could be detected in glomeruli at 24 h (0.18 ± 0.08 versus contralateral: 0.01 ± 0.02 and versus Con A: 0.04 ± 0.04; both P < 0.05) and 3 days (0.21 ± 0.28 versus contralateral: 0 ± 0 and versus Con A: 0.01 ± 0.01; both P < 0.05; Table 1) after disease induction.

Endothelial injury is accompanied by altered renal function, proteinuria and presence of schistocytes in the peripheral blood
Since renal endothelial injury is characterized by the presence of schistocytes, proteinuria and impaired renal function, these parameters were investigated in our murine model. Schistocytes were present in blood smears of diseased mice between days 1 and 3, but were absent in controls. While serum creatinine values were not significantly different (Table 2), measurement of blood urea nitrogen (BUN) on day 3 demonstrated increased values in diseased mice compared to Con A-perfused controls (77.6 ± 19.9 versus 51.3 ± 10.1 mg/dl; P = 0.05, Table 2) indicating some functional consequences of this renal injury model despite unilateral induction. In addition, proteinuria expressed as urinary protein per mg urinary creatinine was significantly increased 24 h after model induction (85.7 ± 38.9 versus 30.1 ± 14.9 µg protein per mg creatinine; P = 0.05; Table 2) compared to controls, but decreased thereafter.


View this table:
[in this window]
[in a new window]

  		Table 2 Functional parameters

 


   Discussion
Top
 Abstract
 Introduction
 Material and methods
 Results
 Discussion
References
 
We present a murine model of site-specific endothelial injury associated with platelet activation and fibrin deposition. The model was induced by unilateral renal artery perfusion of the endothelial-specific lectin, Con A [7,16], followed by anti-Con A antibody. The in situ formation of Con A anti-Con A complexes on the endothelial cell surface resulted in complement activation with endothelial injury manifested by endothelial swelling and injury as detected by electron microscopy. This was accompanied by a loss of CD31- and lectin-positive endothelial cells and the rapid influx and activation of platelets leading to glomerular and peritubular thrombus formation as indicated by the large number of fibrin-positive thrombi. The severe impact on the renal microvascular endothelium was reflected by an increase of TUNEL-positive cells in glomeruli and the interstitium during the first 3 days and the marked decrease in glomerular cell numbers. These histological changes were associated with the development of proteinuria, a rise in BUN, and the presence of schistocytes on peripheral blood smears. Thus, this model has features of thrombotic microangiopathy, although it lacks certain characteristics (arteriolar lesions) as well has other characteristics (peritubular capillary injury) that are not commonly present in human thrombotic microangiopathy.

Interestingly, whereas CD31 staining demonstrated a very homogeneous PECAM distribution at EC–cell contacts in normal kidneys, capillaries of diseased kidneys either lost CD31 positivity or focally (in less severely injured areas) demonstrated a remarkable increase of CD31. Since CD31 is constitutively expressed on the EC surface, its loss most likely reflects capillary loss [11]. In contrast, the upregulation of CD31 might also indicate a role of CD31 as an important signalling molecule with respect to EC survival and injury/death [11,27] or as a consequence of platelet––endothelial cell interactions. Glomerular and peritubular evaluation of capillary loss after staining for the endothelial binding lectin from L. esculentum (tomato) verified our findings of capillary loss detected by evaluation of CD31. The model was also associated with tubular injury. Interestingly, the TUNEL assay did not reliably and persistently detect significantly increased number of apoptotic cells in the tubular compartment even though diseased animals demonstrated severe tubular injury. This possibly may reflect the relatively short time of TUNEL positivity during the death of these cells on the one, and the presence of acute tubular necrosis rather than apoptosis on the other hand. The decreased proliferation rate of tubular cells 4 h after disease induction also indicates the rapid and severe tubular injury in this model.

Renal cellular repair as assessed by the proliferative response was significantly increased on day 3 in both glomeruli and the interstitium, while the tubular repair process seemed to be maximal on day 3 but did not result in significant differences. The lack of significant differences regarding tubular cell proliferation as well as the relative low numbers of proliferating cell types for all renal compartments may indicate additional alternative repair mechanisms, such as replacement by stem cells and/or relate to the variability of the time course of disease. Nevertheless, we also could demonstrate that VEGF as one major endothelial cell mitogen and survival factor is upregulated in injured kidneys and might play an important role during renal regeneration in this disease model.

A significant increase of F4/80-positive monocytes/macrophages was also demonstrated starting after 4 h in glomeruli and on day 7 in the tubulointerstitium. In parallel, influx of neutrophils was also increased in glomeruli after disease induction. Chemotactic cytokines released at the site of EC injury have been shown to induce the influx of inflammatory cells, such as monocytes/macrophages and neutrophils. The secretion of monocyte chemoattractant protein-1 (MCP-1) by endothelial cells can be induced by assembly and activation of the C5b-9 membrane attack complex [28].

Although we detected similarities in proteinuria, renal function and influx of inflammatory cells early on in this murine model compared to the rat Con A anti-Con A model, there also exist clear differences. Endothelial injury in the murine model is more severe and involves peritubular capillaries. This peritubular endothelial injury most likely is the cause for subsequent acute tubular necrosis and the remarkable macroscopic changes we found. The severe impact on the kidney might also explain the slower repair process compared to the rat model with incomplete restitution after 7 days. Even though the aetiology of human lesions might be different, we believe that this model with site-specific endothelial injury and subsequent glomerular and peritubular thrombotic microangiopathy is a promising relevant research tool to investigate endothelial injury and repair [29–36]. This model should be helpful to elucidate the role of platelet–endothelial interactions and platelet activation during renal disease, where capillary injury/protection and regeneration is critical [4,37].



   Acknowledgments
 
We thank Professor B. Nieswandt, Virchow-Institute, Wuerzburg for kindly providing antibodies against glycoprotein Ib. The technical assistance of S. Weber and S. Cabric is gratefully acknowledged. This work was supported by grants to C.H. (IZKF TP A12) and B.H. (IZKF TP B14).

Conflict of interest statement. None declared.



   References
Top
 Abstract
 Introduction
 Material and methods
 Results
 Discussion
 References
 

  1. Zipfel PF, Heinen S, Jozsi M, et al. Complement and diseases: defective alternative pathway control results in kidney and eye diseases. Mol Immunol (2006) 43:97–106.[CrossRef][Web of Science][Medline]
  2. Dong JF, Moake JL, Nolasco L, et al. ADAMTS-13 rapidly cleaves newly secreted ultralarge von willebrand factor multimers on the endothelial surface under flowing conditions. Blood (2002) 100:4033–4039.[Abstract/Free Full Text]
  3. Kang DH, Hughes J, Mazzali M, et al. Impaired angiogenesis in the remnant kidney model: II. Vascular endothelial growth factor administration reduces renal fibrosis and stabilizes renal function. J Am Soc Nephrol (2001) 12:1448–1457.[Abstract/Free Full Text]
  4. Kang DH, Kanellis J, Hugo C, et al. Role of the microvascular endothelium in progressive renal disease. J Am Soc Nephrol (2002) 13:806–816.[Abstract/Free Full Text]
  5. Masuda Y, Shimizu A, Mori T, et al. Vascular endothelial growth factor enhances glomerular capillary repair and accelerates resolution of experimentally induced glomerulonephritis. Am J Pathol (2001) 159:599–608.[Abstract/Free Full Text]
  6. Johnson RJ. Platelets in inflammatory glomerular injury. Semin Nephrol (1991) 11:276–284.[Web of Science][Medline]
  7. Johnson RJ, Alpers CE, Pritzl P, et al. Platelets mediate neutrophil-dependent immune complex nephritis in the rat. J Clin Invest (1988) 82:1225–1235.[Web of Science][Medline]
  8. Johnson RJ, Garcia RL, Pritzl P, et al. Platelets mediate glomerular cell proliferation in immune complex nephritis induced by anti-mesangial cell antibodies in the rat. Am J Pathol (1990) 136:369–374.[Abstract]
  9. Barnes JL, Hevey KA. Glomerular mesangial cell migration. Response to platelet secretory products. Am J Pathol (1991) 138:859–866.[Abstract]
  10. Barnes JL, Venkatachalam MA. The role of platelets and polycationic mediators in glomerular vascular injury. Semin Nephrol (1985) 5:57–68.[Web of Science][Medline]
  11. Jackson DE. The unfolding tale of PECAM-1. FEBS Lett (2003) 540:7–14.[CrossRef][Web of Science][Medline]
  12. Leon C, Hechler B, Freund M, et al. Defective platelet aggregation and increased resistance to thrombosis in purinergic P2Y(1) receptor-null mice. J Clin Invest (1999) 104:1731–1737.[Web of Science][Medline]
  13. Andre P, Delaney SM, LaRocca T, et al. P2Y12 regulates platelet adhesion/activation, thrombus growth, and thrombus stability in injured arteries. J Clin Invest (2003) 112:398–406.[CrossRef][Web of Science][Medline]
  14. Hechler B, Zhang Y, Eckly A, et al. Lineage-specific overexpression of the P2Y1 receptor induces platelet hyper-reactivity in transgenic mice. J Thromb Haemost (2003) 1:155–163.[CrossRef][Web of Science][Medline]
  15. Golbus SM, Wilson CB. Experimental glomerulonephritis induced by in situ formation of immune complexes in glomerular capillary wall. Kidney Int (1979) 16:148–157.[CrossRef][Web of Science][Medline]
  16. Fries JW, Mendrick DL, Rennke HG. Determinants of immune complex-mediated glomerulonephritis. Kidney Int (1988) 34:333–345.[Web of Science][Medline]
  17. el Nahas AM, Bassett AH, Cope GH, et al. Role of growth hormone in the development of experimental renal scarring. Kidney Int (1991) 40:29–34.[Web of Science][Medline]
  18. Hugo C, Pichler R, Meek R, et al. Thrombospondin 1 is expressed by proliferating mesangial cells and is up-regulated by PDGF and bFGF in vivo. Kidney Int (1995) 48:1846–1856.[Web of Science][Medline]
  19. Thurston G, Murphy TJ, Baluk P, et al. Angiogenesis in mice with chronic airway inflammation: strain-dependent differences. Am J Pathol (1998) 153:1099–1112.[Abstract/Free Full Text]
  20. Hohenstein B, Hausknecht B, Boehmer K, et al. Local VEGF activity but not VEGF expression is tightly regulated during diabetic nephropathy in man. Kidney Int (2006) 69:1654–1661.[CrossRef][Web of Science][Medline]
  21. Bergmeier W, Rackebrandt K, Schroder W, et al. Structural and functional characterization of the mouse von willebrand factor receptor GPIb-IX with novel monoclonal antibodies. Blood (2000) 95:886–893.[Abstract/Free Full Text]
  22. Hohenstein B, Kasperek L, Kobelt DJ, et al. Vasodilator-stimulated phosphoprotein-deficient mice demonstrate increased platelet activation but improved renal endothelial preservation and regeneration in passive nephrotoxic nephritis. J Am Soc Nephrol (2005) 16:986–996.[Abstract/Free Full Text]
  23. Hohenstein B, Renk S, Lang K, et al. P2Y1 gene deficiency protects from renal disease progression and capillary rarefaction during passive crescentic glomerulonephritis. J Am Soc Nephrol (2007) 18:494–505.[Abstract/Free Full Text]
  24. Mihatsch MJ, Bremer J. Acid fuchsin orange G-stain (AFOG) for glomerular protein deposits. Clin Nephrol (1978) 9:259.[Medline]
  25. Hughes J, Brown P, Shankland SJ. Cyclin kinase inhibitor p21CIP1/WAF1 limits interstitial cell proliferation following ureteric obstruction. Am J Physiol (1999) 277:F948–F956.[Web of Science][Medline]
  26. Hughes J, Nangaku M, Alpers CE, et al. C5b-9 membrane attack complex mediates endothelial cell apoptosis in experimental glomerulonephritis. Am J Physiol Renal Physiol (2000) 278:F747–F757.[Abstract/Free Full Text]
  27. Newman PJ, Newman DK. Signal transduction pathways mediated by PECAM-1: new roles for an old molecule in platelet and vascular cell biology. Arterioscler Thromb Vasc Biol (2003) 23:953–964.[Abstract/Free Full Text]
  28. Kilgore KS, Flory CM, Miller BF, et al. The membrane attack complex of complement induces interleukin-8 and monocyte chemoattractant protein-1 secretion from human umbilical vein endothelial cells. Am J Pathol (1996) 149:953–961.[Abstract]
  29. Amigo MC, Garcia-Torres R, Robles M, et al. Renal involvement in primary antiphospholipid syndrome. J Rheumatol (1992) 19:1181–1185.[Web of Science][Medline]
  30. Andreoli SP. Renal manifestations of systemic diseases. Semin Nephrol (1998) 18:270–279.[Web of Science][Medline]
  31. D’Agati V, Kunis C, Williams G, et al. Anti-cardiolipin antibody and renal disease: a report three cases. J Am Soc Nephrol (1990) 1:777–784.[Abstract]
  32. Hughson MD, Nadasdy T, McCarty GA, et al. Renal thrombotic microangiopathy in patients with systemic lupus erythematosus and the antiphospholipid syndrome. Am J Kidney Dis (1992) 20:150–158.[Web of Science][Medline]
  33. Kincaid-Smith P, Fairley KF, Kloss M. Lupus anticoagulant associated with renal thrombotic microangiopathy and pregnancy-related renal failure. Q J Med (1988) 68:795–815.[Medline]
  34. Leibovich SJ. Vasculopathy and renal injury in lupus erythematosus: does shedding of the endothelial protein C receptor play a role? Kidney Int (2005) 68:407–408.[Web of Science][Medline]
  35. Magil AB, McFadden D, Rae A. Lupus glomerulonephritis with thrombotic microangiopathy. Hum Pathol (1986) 17:192–194.[CrossRef][Web of Science][Medline]
  36. Ruggenenti P, Remuzzi G. Malignant vascular disease of the kidney: nature of the lesions, mediators of disease progression, and the case for bilateral nephrectomy. Am J Kidney Dis (1996) 27:459–475.[Web of Science][Medline]
  37. Johnson RJ. The glomerular response to injury: progression or resolution? Kidney Int (1994) 45:1769–1782.[Web of Science][Medline]
Received for publication: 21. 5.07
Accepted in revised form: 4.10.07


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 CirculationHome page
J. Hirahashi, K. Hishikawa, S. Kaname, N. Tsuboi, Y. Wang, D. I. Simon, G. Stavrakis, T. Shimosawa, L. Xiao, Y. Nagahama, et al.
Mac-1 (CD11b/CD18) Links Inflammation and Thrombosis After Glomerular Injury
Circulation, September 29, 2009; 120(13): 1255 - 1265.
[Abstract] [Full Text] [PDF]

Home page
 BloodHome page
S. V. Seshan, C.-W. Franzke, P. Redecha, M. Monestier, N. Mackman, and G. Girardi
Role of tissue factor in a mouse model of thrombotic microangiopathy induced by antiphospholipid antibodies
Blood, August 20, 2009; 114(8): 1675 - 1683.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Pathol.Home page
T. Kosugi, M. Heinig, T. Nakayama, T. Connor, Y. Yuzawa, Q. Li, W. W. Hauswirth, M. B. Grant, B. P. Croker, M. Campbell-Thompson, et al.
Lowering Blood Pressure Blocks Mesangiolysis and Mesangial Nodules, but Not Tubulointerstitial Injury, in Diabetic eNOS Knockout Mice
Am. J. Pathol., April 1, 2009; 174(4): 1221 - 1229.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
23/4/1144    most recent
gfm774v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Hohenstein, B.
Right arrow Articles by Hugo, C. P. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hohenstein, B.
Right arrow Articles by Hugo, C. P. M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?