Plasma leakage through glomerular basement membrane ruptures triggers the proliferation of parietal epithelial cells and crescent formation in non-inflammatory glomerular injury

ORIGINAL PAPERJournal of PathologyJ Pathol (2012)Published online in Wiley Online Library(wileyonlinelibrary.com) DOI: 10.1002/path.4046

Plasma leakage through glomerular basement membrane rupturestriggers the proliferation of parietal epithelial cells and crescentformation in non-inflammatory glomerular injuryMi Ryu,1 Adriana Migliorini,1 Nicolai Miosge,2 Oliver Gross,3 Stuart Shankland,4 Paul T Brinkkoetter,5Henning Hagmann,5 Paola Romagnani,6 Helen Liapis7 and Hans-Joachim Anders1*

1 Nephrologisches Zentrum, Medizinische Klinik und Poliklinik IV, Klinikum der Universitat Munchen, Germany2 Tissue Regeneration Work Group, Georg August University, Gottingen, Germany3 Department of Nephrology and Rheumatology, University Medicine Goettingen, Germany4 Division of Nephrology, University of Washington Medical Center, Seattle, WA, USA5 Renal Division, Department of Medicine and Center for Molecular Medicine, University of Cologne, Cologne, Germany6 Excellence Centre for Research, Transfer and High Education for the Development of De Novo Therapies (DENOTHE), University of Florence,Florence, Italy7 Department of Pathology and Immunology, Washington University, School of Medicine, St Louis, MO, USA

*Correspondence to: Hans-Joachim Anders, MD, Medizinische Klinik und Poliklinik IV der LMU, Pettenkoferstrasse 8a, 80336 Munchen, Germany.e-mail: [email protected]

AbstractGlomerular crescents are most common in rapidly progressive glomerulonephritis but also occur in non-inflammatory chronic glomerulopathies; thus, factors other than inflammation should trigger crescent formation,eg vascular damage and plasma leakage. Here we report that Alport nephropathy in Col4A3-deficient Sv129mice is complicated by diffuse and global crescent formation in which proliferating parietal epithelial cells arethe predominant cell type. Laminin staining and transmission and acellular scanning electron microscopy ofacellular glomeruli documented disruptions and progressive disintegration of the glomerular basement membranein Col4A3-deficient mice. FITC-dextran perfusion further revealed vascular leakage from glomerular capillariesinto Bowman’s space, further documented by fibrin deposits in the segmental crescents. Its pathogenic rolewas validated by showing that the fibrinolytic activity of recombinant urokinase partially prevented crescentformation. In addition, in vitro studies confirmed an additional mitogenic potential of serum on murine andhuman parietal epithelial cells. Furthermore, loss of parietal cell polarity and unpolarized secretion of extracellularmatrix components were evident within fibrocellular crescents. Among 665 human Alport nephropathy biopsies,crescent formation was noted in 0.4%. We conclude that glomerular vascular injury and GBM breaks cause plasmaleakage which triggers a wound healing programme involving the proliferation of parietal cells and their lossof polarity. This process can trigger cellular and fibrocellular crescent formation even in the absence of cellularinflammation and rupture of the Bowman’s capsule.Copyright © 2012 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Keywords: crescent; Alport; glomerulosclerosis; chronic kidney disease; parietal epithelial cells; collagen

Received 13 February 2012; Revised 19 April 2012; Accepted 23 April 2012

No conflicts of interest were declared.

Introduction

Crescent formation has been reported in manyglomerulopathies, but is a defining feature of rapidlyprogressive glomerulonephritis (RPGN) due to immunecomplex GN, pauci-immune vasculitis, and anti-glomerular basement membrane (GBM) GN [1,2].In RPGN, glomerular crescents consist of a mixtureof inflammatory cells, extracellular matrix, fibrinoidnecrosis, fibroblasts, and leukocyte infiltrates, but cres-cents also occur in the late stages of many chronicglomerulopathies [3]. Epithelial cellular crescents indi-cate acute damage, whereas older crescents becomefibrocellular and reflect a poor prognostic indicator for

progression to end-stage renal disease [3–5]. Eleganttransgenic mouse studies using lineage tracing of eitherparietal epithelial cells (PECs) or visceral podocytesdemonstrated recently that cellular crescents originatelargely from parietal epithelial cells in several mod-els of glomerular injury [6]. However, other studiessuggested that podocytes can also participate in thelesion [7,8]. However, the potential for podocytes tore-enter the cell cycle and to contribute to multi-layercell growth is unlikely, although this remains underdebate [9,10]. Recent studies show that crescents thatstain positive with antibodies directed against maturepodocyte-related proteins may rather represent acti-vated PECs [5,9–12].

Copyright © 2012 Pathological Society of Great Britain and Ireland. J Pathol (2012)Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk www.thejournalofpathology.com

M Ryu et al

Most studies address inflammatory cell mediatorsor distinct leukocyte populations as potential media-tors of crescent formation [3]. Thus, one view is thatpro-inflammatory elements trigger glomerular epithe-lial proliferation, similar to synovial hyperplasia inthe arthritic joint (synovitis). The therapeutic effectsof immunosuppressants in RPGN (as well as inarthritis) further support this concept [13]. However,glomerular inflammation may not primarily drive cres-cent formation in late-stage chronic nephropathies.An unanswered question is what factors initiate anddrive glomerular crescent formation in the absence ofglomerular inflammation. For example, we ask whetherpodocyte proliferation is a mechanism of homeosta-sis (as in intestinal epithelia) or a process of repairafter injury (as in wound healing) [14]. Woundingat first activates coagulation to prevent blood lossand only later tissue inflammation joins in to pro-tect from pathogen entry and infection [15]. However,injury initiates regenerative epithelial proliferation veryearly. Similarly, in acute kidney injury, epithelialrepair is the first line of response [16]. In searchfor non-inflammatory triggers of epithelial prolifera-tion, we hypothesized that vascular injury, bleeding,and plasma leakage provide a trigger for glomerularcrescent formation. To test this hypothesis, we usedcollagen (Col)4A3-deficient 129X1/SvJ ‘Alport’ mice,which develop a progressive glomerulopathy devoid ofglomerular inflammation [17].

Materials and methods

Animal studiesCol4A3-deficient and wild-type littermate mice in iden-tical 129X1/SvJ genetic backgrounds were bred underspecific pathogen-free housing conditions and geno-typed as previously described [18]. Col4A3-deficientmice and controls were sacrificed at 6 and 9 weeks ofage to assess kidney pathology. We injected anothercohort subcutaneously with 5000 U of urokinase(Medac, Hamburg, Germany) daily for 3 weeks andsacrificed them at the end of week 9. Interventionalstudies from our laboratories were retrospectively anal-ysed for crescent formation. These treatments includedanti-CCL2/MCP-1 Spiegelmers [19], the chemokinereceptor CCR1 antagonist BX471 [20], mesenchymalstem cells [21], the TNFα blocker etanercept [22],the angiotensin-converting enzyme inhibitor ramipril[23], and the vasopeptidase inhibitor AVE7688 [24].All experimental procedures were performed accord-ing to German animal care and ethics legislation andwere approved by local government authorities.

Histopathology and scanning electron microscopyKidney sections were fixed in 10% formalin in PBSand embedded in paraffin. Two-micrometre sectionswere stained with periodic acid–Schiff reagent (PAS)and silver [25]. Glomerular crescents were evaluated

as follows: number of (a) glomeruli with no cres-cents; (b) glomeruli with segmental crescents (< 50%Bowman’s space); and (c) glomeruli with global cres-cents (> 50% of Bowman’s space). Fifty glomeruli persection were analysed. We reviewed the electronic filesof the Department of Pathology at Washington Uni-versity in St Louis and the Mayo Clinic, Rochester,MN, covering 66 500 renal biopsies from 1991 to 2011.Key words in the search were ‘Alport’ and ‘crescents’in the text and diagnosis. We retrieved and reviewedthe light microscopy slides and electron microscopy ofthe Alport renal biopsies with crescents plus slides of∼10% of the total Alport biopsies (n = 60) withoutcrescents found in the electronic database. No addi-tional cases with crescents were found. For transmis-sion electron microscoy, kidney tissues from 9-week-old Col4A3-deficient and wild-type mice were fixedwith 0.4% paraformaldehyde and 2% glutaraldehyde in0.1M PBS buffer (pH 7.4). Fresh Epon resin polymer-ization was carried out at 60 ◦C for 24 h as previouslydescribed [26]. Preparation of acellular glomeruli andscanning electron microscopy (SEM) were performedas previously described [27].

ImmunohistochemistryAll immunohistological studies were performed onparaffin-embedded sections as described using thefollowing primary antibodies [18]: rat anti-Wilms’tumour (WT)-1 (Santa Cruz Biotechnology, SantaCruz, CA, USA; 1 : 200); guinea pig anti-nephrin(Acris Antibodies, Herford, Germany; 1 : 100); rabbitanti-mouse claudin-1, a marker for parietal epithelialcells (Bioworld, St Louis Park, MN, USA; 1 : 50); Ki-67, a marker for proliferating cells (Dianova, Hamburg,Germany; 1 : 50); and rabbit anti-mouse laminin (1 : 50)and rabbit anti-mouse fibrinogen (both Abcam, Cam-bridge, UK; 1 : 500).

Vascular permeability studiesNine-week-old Col4A3-deficient and wild-type micewere intravenously injected with 10 mg per mouseof FITC-dextran (MW 500 000; Sigma-Aldrich, Stein-heim, Germany) and sacrificed 30 min later. Kidneytissue was fixed in 10% formalin and paraffin-embedded. Two-micrometre sections were stained withlaminin and viewed under a fluorescence microscope.

Real-time quantitative RT-PCRTotal RNA was isolated from kidney or cells usingan RNA extraction kit (QIAGEN, Hilden, Germany)according to the manufacturer’s instructions, and RNAquality was assessed using agarose gels [28]. Afterisolation of RNA, cDNA was generated using reversetranscriptase (Superscript II; Invitrogen, Carlsbad, CA,USA). The SYBR Green Dye detection system wasused for quantitative real-time PCR on a LightCycler 480 (Roche, Mannheim, Germany) using 18srRNA as a housekeeper gene. Gene-specific primers


Crescents without glomerular inflammation

(300 nM; Metabion, Martinsried, Germany) were usedas follows: 18s: forward, GCAATTATTCCCCATG-AACG; reverse, AGGGCCTCACTAAACCATCC; col-lagen Iα1: forward, ACATGTTCAGCTTTGTGGACC;reverse, TAGGCCATTGTGTATGCAGC; fibronectin:forward, GGAGTGGCACTGTCAACCTC; reverse,ACTGGATGGGGTGGGAAT; vimentin: forward,AGAGAGAGGAAGCCGAAAGC; reverse, TCCACTTTCCGTTCAAGGTC.

In vitro studiesHuman renal PECs using CD133+/CD24+ as selec-tion markers were cultured in EGM-MV 5% FBS(Hyclone, Logan, UT, USA) as previously described[29]. The cells were seeded in 96-well plates andstimulated with different serum concentrations for72 h. Murine primary epithelial cells (mPECs) werecultured at 33 ◦C in 10 cm plates (BD Biosciences,Heidelberg, Germany) coated with type I collagenin RPMI 1640 containing 5% FBS and 50 U/mlIFN-γ (Roche Molecular Biochemicals) as previouslydescribed [30]. To induce differentiation, cells werecultured under growth-restrictive conditions at 37 ◦Cand in the absence of IFN-γ for 14 days [31]. Undergrowth-restrictive conditions, mPECs stop replicatingand display a polygonal shape and expressed claudin-1 as previously described [30]. Cells exhibiting thischaracteristic were used in the experiments. The pro-liferation was determined using the microtetrazoliumassay (Promega, Madison, WI, USA) [29].

Statistical analysisPaired Student’s t-test was used for the comparison ofsingle groups. A value of p < 0.05 was considered toindicate statistical significance.

Results

Glomerular pathology in Col4A3-deficient miceCol4A3-deficient mice in a mixed genetic back-ground develop GBM multi-lamination and splitting,podocyte damage, and glomerulosclerosis [32]. Wenoted that during the later stages of disease, Col4A3-deficient mice on a 129X1/SvJ genetic backgroundhave additional glomerular abnormalities, includingcrescent formation (Supporting information, Supple-mentary Figure 1 and Figure 1). This occurs much laterin Col4A3-deficient C57BL/6 mice [33]. PAS stain-ing demonstrated segmental crescents in 32 ± 1%, andglobal crescents in 43 ± 4% of cortical glomeruli at9 weeks of age (Figure 1). Global crescents showeda multi-layered cell growth filling the extracapillaryspace over collapsed outer glomerular capillaries andsclerotic mesangium (Figure 1). At earlier time points(6 weeks), the percentage of segmental and global cres-cents was significantly smaller compared with latertime points (Figure 1G, p < 0.01 and p < 0.001),suggesting that crescents develop gradually mainly

0Wild-type9 weeks 6 weeks 9 weeks

Col4A3-/-

**

*

* ****

No crescentsSegmental crescentsGlobal crescents

*20

40

60

80

% o

f glo

m. a

ffect

ed b

y 100

Figure 1. Crescent formation in late-stage Alport mice. Renalsections from 9-week-old wild-type (A, B) and Col4A3-deficientmice (C–F) Note the normal cortex (A) and glomerulus (B) in9-week-old wild-type mice. Age-matched Col4A3-deficient miceshow diffuse crescentic glomerulonephritis with cellular casts intubular lumens (C). Representative images illustrate the differentstages of crescent formation. (D) Segmental crescent with multi-layer growth of extra capillary cells surrounded by PAS-positivehyalin matrix. (E) This glomerulus illustrates retracted capillaryloops encircled by an epithelial crescent in multiple layers.Note the intratubular cellular cast in the upper left corner asa marker of proteinuria. (F) This glomerulus displays extensiveproliferation of cells completely obstructing Bowman’s space.Original magnification: PAS × 100 (A, C) and × 600 (B, D–F).(G) Graphic representation of the percentage of crescents in 6-and 9-week-old Col/4A3−/− and wild-type controls (∗∗p < 0.01;∗∗∗p < 0.001).

throughout the late phase of the glomerular diseaseprocess. To exclude the hypothesis that crescents devel-oped by GBM autoantibodies which is similar toAlport patients transplanted with a normal allograftkidney and subsequent anti-GBM glomerulonephritisdue to immunization against the neoantigen [34], weused anti-mouse IgG immunostaining and found thatin Col4A3-deficient mice, IgG stains were negative(not shown). This excludes the possibility of autolo-gous triggers for crescent formation in these mice. Insummary, late-stage Alport nephropathy in 129X1/SvJCol4A3-deficient mice is associated with glomerular


M Ryu et al

crescent formation, a process not involving anti-GBMautoantibodies.

Glomerular crescents in Col4A3-deficient miceinvolve parietal epithelial cellsWe used the nuclear podocyte marker WT-1 to deter-mine the nature of cells within crescents in late-stageCol4A3-deficient Alport mice. However, WT-1 stain-ing was detected not only in visceral podocytes, butalso in PECs (Figure 2B). This was not observedin cells at the vascular pole of age-matched wild-type mice (Figure 2A). De novo WT-1 expressionindicated parietal epithelial cell activation and didnot distinguish PECs from visceral podocytes in seg-mental and global crescents of Col4A3-deficient mice(Figures 2C and 2D). We therefore used WT-1/nephrindouble immunostaining to distingush differentiated vis-ceral podocytes (WT-1-posive and nephrin-positive)from activated PECs (WT-1-positive and nephrin-negative) (Figure 2E). Segmental and global crescentsshowed a continuity between WT-1+/nephrin− pari-etal epithelial cells along Bowman’s capsule and thecells in crescents, in contrast to WT-1+/nephrin+podocytes in the uninvolved areas of the glomeruli(Figures 2F and 2G). Although it is theoreticallypossible that the WT-1+/nephrin− cells representdedifferentiated podocytes, we consider this possi-bility unlikely because intratubular nephrin + cellswere detected, indicating that these are detachedvisceral podocytes (Figure 2H). In addition, doubleimmunopositivity for WT-1 and the PEC markerclaudin-1 identified parietal epithelial cells in cres-cents but there was only WT-1 positivity in intratubularpodocytes (insets in Figures 2E, 2G, and 2H). Thus,in Col4A3-deficient mice, crescent formation largelyinvolves activated PECs.

Parietal cell proliferation in Col4A3-deficient miceTo determine if the excess of parietal epithelial cellsfilling the glomerular urinary space was due to pro-liferation, immunostaining was performed for Ki-67, amarker of cell cycle entry. Ki-67 staining cells werebarely detected in the kidneys of 9-week-old wild-type mice (Figure 3A). In contrast, Ki-67 stainingwas marked in the kidneys of age-matched Col4A3-deficient mice. In glomeruli with crescents, Ki-67 pos-itivity was mostly in the parietal epithelial cell layeralong the Bowman’s capsule (Figures 3C and 3D),although Ki-67 positivity was also noted in all areasof the crescent. Ki-67 staining in the glomerularmesangium was infrequent. We conclude that parietalepithelial cells proliferate in crescentic glomeruli ofCol4A3-deficient mice.

Therapeutic interventions and crescent formationin Col4A3-deficient miceNext, we carefully re-evaluated all kidney sectionsof 8- to 9-week-old Col4A3-deficient mice on a

Figure 2. WT-1 and nephrin expression in late-stage Alportnephropathy. Kidney sections from 9-week-old wild-type (A, E) andCol4A3-deficient mice (B–D, F–H) were stained with WT-1. Notethat WT-1 positivity by single immunostaining is largely limited tovisceral podocytes, although transitional cells at the vascular poleand occasionally along the circumference of Bowman’s capsulealso show focal WT-1 positivity (white arrows). (B) Non-crescentic(normal-appearing) glomeruli in age-matched Col4A3-deficientmice display WT-1 positivity in most of the parietal epithelialcells. (C, D) Podocytes and WT-1-positive parietal epithelial cellsbecome indistinguishable in focal points of contact where thesemerge (transition zone) (C, white arrows) or in glomeruli containingglobal crescents (D). WT-1/nephrin double staining distinguisheddifferentiated podocytes from other WT-1-positive cells suchas activated PECs. (E) Note that in 9-week-old wild-type mice,nuclear WT-1 positivity (red) is strong in visceral podocytes, whichalso show linear nephrin positivity along the capillary wall. Incontrast, nuclear WT-1 positivity is weak in parietal epithelialcells, which stain negative for nephrin (white arrow). We also usedclaudin-1(red) and WT-1 (green) double staining to differentiatebetween PECs and visceral podocytes in crescents and foundthat parietal epithelial cells, but not visceral podocytes, stainedwith claudin-1 (red) in wild-type mice (inset in E). (F) Nine-week-old Col4A3-deficient mouse glomeruli with crescents showWT-1/nephrin positivity in the lower part of the glomerular tuft,while cells within the crescent are WT1+/nephrin− in the upperpart of the glomerulus. (G) Many glomeruli are affected by markedproliferation of WT-1-positive/nephrin-negative cells. Note thesame staining pattern in parietal epithelial cells (white arrows).The inset shows WT-1 (green) and claudin-1 (red) double stainingin a segmental crescent. (H) Occasionally, intratubular casts stainpositive for nephrin and WT-1, indicating that detached podocytesare incorporated into cellular casts (white arrows). The inset showsWT-1 (green) positivity but lack of claudin positivity (red) as indirectevidence that these casts include podocytes and not PECs. Originalmagnification: × 600. Figure 2 E,G, and H were replicated withpermission from Ryu M, et al. J Pathol 2012; 226:120-131.



0

InterventionSurvival benefit

n.a.n.a.

n.a.αCCL2 αCCR1 MSC αTNF αVPαACE

79d14d14d 0d0d 101d-

--

-+

-+ +

+++

++GS improved

IF improved

** * * *

No crescentsSegmental crescentsGlobal crescents

20

40

60

80

% o

f glo

m. a

ffect

ed b

y

100

Figure 3. Proliferation of glomerular cells in Alport nephropathy.Nine-week-old wild-type (A) and Col4A3-deficient mouse kidneysections (B–D) were stained with Ki-67. Note that Ki-67+ cellsare rare in glomeruli of healthy wild-type mice (A). In Col4A3-deficient mice, Ki-67+ cells were either predominately in thebasal cell layer of parietal epithelial cells (B) or throughoutthe entire crescent (C, D). Ki-67 proliferating cells were rarein the glomerular tuft (dotted area). Original magnification:× 600. (E) Segmental and global glomerular crescents werequantified in PAS-stained renal sections of 9-week-old Col4A3-deficient mice undergoing therapeutic interventions as indicated.αCCL2 = CCL2/MCP-1 blockade with anti-CCL2 Spiegelmers;αCCR1 = chemokine receptor CCR1 blockade with BX471; MSC =therapeutic injections with mesenchymal stem cells; αTNF = TNFblockade with etanercept; αACE = angiotensin-converting enzymeinhibition with ramipril; αVP = vasopeptidase inhibition withAVE7688. Data are mean percentages ± SEM. p < 0.05 versusno treatment.

129X1/SvJ genetic background with experimental drugtherapies previously performed in our laboratories. Aquantitative analysis revealed a comparable amountof glomeruli with no crescents, segmental or globalcrescents in the untreated or vehicle-treated Col4A3-deficient control groups of all interventional stud-ies (not shown) compared with the data reportedin this paper. Chemokine receptor CCR1 block-ade or injecting mesenchymal stem cells signif-icantly improved tubulointerstitial disease withoutaffecting glomerulosclerosis [20,21]. In both studies,the numbers of segmental and global crescents weresimilar to those in age-matched Col4A3-deficient con-trol mice (Figure 3E). The same was found uponblockade of CCL2/MCP-1 (Figure 3E), which didnot affect glomerular or tubulointerstitial damage in

Col4A3-deficient mice [19]. In contrast, interventionssuch as TNF blockade with etanercept, ramipril or avasopeptidase inhibitor significantly prolonged the lifespan of Col4A3-deficient mice by delaying the progres-sion of glomerular pathology and were now found tobe associated with significantly fewer glomeruli withglobal crescents (Figure 3E) [22–24].

GBM defects and vascular leakagein Col4A3-deficient miceCrescents are mostly limited to severe forms ofglomerulonephritis where focal necrosis or enzymesreleased by infiltrating leukocytes promote GBM rup-ture [27]. In the kidneys of 9-week-old Col4A3-deficient mice, laminin staining provided a first hintof thinning and focal discontinuations of the GBM inareas adjacent to crescents (Figures 4A and 4B). GBM

Figure 4. GBM breaks in late-stage Alport nephropathy. Renalsections from 9-week-old wild-type (A) and Col4A3-deficient mice(B–D) were stained with laminin. Note the normal structure ofthe GBM, Bowman’s capsule, and the tubular BM in healthywild-type mice (A). In Col4A3-deficient mice, glomeruli withsegmental crescents reveal discontinuations of the GBM (B–D)which occasionally display erythrocytes passing from the capillarylumen into the urinary space (marked with white arrows in Cand D). Original magnification: × 400 (A); × 600 (B); × 1000(C, D). Transmission electron microscopy from 9-week-old Col4A3-deficient mice. (E) Note two small GBM gaps (marked by whitearrows). A larger GBM gap of about 4 μm is flanked by whitearrows (F). Original magnification: × 10 000.


M Ryu et al

Figure 5. Scanning electron microscopy of acellular glomeruli. Nine-week-old wild-type and Col4A3-deficient mice show that all glomerulifrom wild-type mice have smooth GBM surfaces (A, B). In contrast, glomeruli of Col4A3-deficient mice display marked GBM abnormalitieswith large openings, discontinuities, and profound disintegration (C, D).

gaps were easily detected when associated with redblood cells escaping through GBM gaps (Figures 4Cand 4D). GBM discontinuities became evident ultra-structurally with transmission electron microscopy(Figures 4E and 4F). To further confirm these observa-tions, we performed scanning electron microscopy onacellular glomeruli, a technique that removes glomeru-lar cells, allowing us to image the GBM [27]. Nine-week-old wild-type mice displayed a uniform andsmooth GBM surface (Figures 5A and 5B), whilethe glomeruli of 9-week-old Col4A3-deficient miceshowed a spectrum of defects depending on the stageof glomerular damage. The GBM appeared shrivelled,with numerous perforations ranging from small holesto complete GBM disintegration (Figures 5C and 5D).Such severe GBM damage may have caused vascu-lar leakage in Alport nephropathy, as documented byhaematuria. Fibrin deposits and tissue necrosis werelong thought to trigger glomerular crescent formation,a notion that was recently confirmed in fibrinogen-deficient mice [35,36]. In order to document vascularleakage in Col4A3-deficient mice, FITC-dextran injec-tion studies were performed. When given to 9-week-oldwild-type mice, FITC was limited to inside glomeru-lar capillaries (Figure 6A). In contrast, in null micewith glomerular disease, FITC was widespread in theextracapillary space, especially in crescentic glomeruli(Figure 6B). Thus, late-stage Alport nephropathy inCol4A3-deficient mice is associated with podocyte

injury and GBM disruption, which results in vascularleakage.

Fibrinolysis with urokinase prevents crescentformation in murine Alport nephropathyBecause serum leaks across damaged podocytes andGBM, we next asked which serum factors might drivethe proliferation of the parietal epithelial cells exposedto the serum. Fibrinogen activation and fibrin depo-sition trigger crescents in antiserum-induced glomeru-lar necrosis and immune complex glomerulonephritis[35,36]. We speculated that this mechanism contributesto crescent formation in murine Alport nephropathybecause vascular damage and disintegration shouldactivate coagulation factors independent of a majorinflammatory response. Our data showed that fibrinimmunostaining of kidneys from 9-week-old Col4A3-deficient mice is markedly positive within crescents(Figure 7A). To show that filtered fibrin is a mechanismunderlying crescent formation, mice were injected withthe fibrinolytic agent urokinase. Urokinase significantlyreduced the number of global crescents in 9-week-old Col4A3-deficient mice (Figure 7B, p < 0.05). Onlyfocal crescents were detected and this comprised sub-stantially fewer glomeruli than in control diseased mice(Figure 7B). We conclude that vascular damage pro-motes fibrinogen activation in Bowman’s space whichappears to stimulate PEC proliferation in murine Alportnephropathy.



Figure 6. Vascular leakage in late-stage Alport nephropathy. Nine-week-old wild-type (A) and Col4A3-deficient mice (B) perfusedwith FITC-dextran. (A) In wild-type mice, FITC positivity (green)was detected only inside the glomerular capillaries highlightedby laminin (red) co-staining. (B) In Col4A3-deficient mice,most glomeruli revealed additional diffuse FITC positivity ofextracapillary cells and matrix as well as in downstream tubules,suggesting dextran leakage from glomerular capillaries (originalmagnification × 400).

Serum stimulates the proliferation of culturedparietal epithelial cells

We used cultured human primary and immortalizedmouse parietal epithelial cells to investigate whetherGBM rupture and plasma leakage into the urinaryspace promote crescent formation in Alport nephropa-thy. To determine whether factors other than fibrino-gen, thrombin, and fibrin underlie this effect, culturedcells were exposed to serum largely devoid of the

2.0

1.5

1.0

0.5

0.0

1.00

0.75

0.50

MT

T a

bsor

banc

e(O

D 5

70nm

)M

TT

abs

orba

nce

(OD

570

nm)

0.25

0.00

0% 0.5% 1% 5% 10% 20% 40% 80%Serum

0% 0.5% 1% 5% 10% 20% 40% 80%Serum

Human renal CD133+ CD24+ cells

Murine parietal epithelial cells

***

***

***

***

***

***

*************

Figure 8. In vitro studies with parietal epithelial cells. Humanrenal CD133+CD24+ cells (A) and mouse parietal epithelialcells (B) were cultured with increasing serum concentrations.Cell viability was assessed by MTT assay after 72 h. Data aremeans ± SD of three independent experiments, respectively.Concentration-dependent mitogenic effect by serum occurs inboth cell lines. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 versus 0%serum concentration.

above coagulation factors [29,30,37]. Serum induced aconcentration-dependent mitogenic effect on both celltypes (Figure 8). We conclude that parietal epithelialcells, a cell type that resides in the glomerular ultrafil-trate and usually is not exposed to serum, proliferateupon serum exposure.

Crescent-related parietal epithelial cells losepolarity and extracellular matrix

Physiologically, PECs grow in single layers. However,in crescents, PECs form multi-layers, which implicatesa significant phenotype switch. In addition, epithelial

100

80

% o

f glo

m. a

ffect

ed b

y

60

40

20

0Wild-type

Nil Saline UrokinaseCol4A3-/-

No crescents

Segmental crescents

Global crescents

Figure 7. Fibrin and crescent formation in late-stage Alport nephropathy. (A) Nine-week-old Col4A3-deficient mice were stained withantibody to fibrin. Note fibrin positivity in the areas of segmental crescents. Original magnification: PAS × 400. (B) Wild-type orCol4A3-deficient mice treated with either saline or urokinase. Urokinase-treated mice show significantly reduced crescent formationcompared with saline-treated mice (∗p < 0.05).


M Ryu et al

cells are defined by a polar orientation. As one aspectof their polarity, epithelial cells release matrix proteinsmainly to their basal and not to their apical/luminalaspect. In contrast, mesenchymal cells are unpolar-ized and secrete matrix components in all directions.To assess whether proliferating parietal epithelial cellsmaintain their polar shape in crescents, we stainedextracellular matrix with silver stain. Silver stains illus-trated mesangial sclerosis and collapse of the capillaryloops in Alport mice, but also the argyrophilic extracel-lular matrix produced by proliferating PECs (Figure 9).It is of note that cells within crescents were surroundedby silver positive matrix, documenting their switch tounpolarized cells (Figure 9). Matrix production wasalso evident by scanning electron microscopy of acellu-lar glomeruli. Glomeruli displayed dense matrix fibrescoupling the capillary loop surface, giving them a cob-bled appearance (Figures 10A–10C). Real-time PCRfor collagen Iα, vimentin, and fibronectin showed anincrease in these elements in the kidney (Figure 10D).Thus, activated PECs acquire an unpolarized phenotypeand produce matrix in all directions like mesenchy-mal cells. This transition process is similar to what hasbeen widely discussed in the context of tubulointersti-tial fibrosis [38,39] or for glomerular epithelial cells[40–42].

Crescents in human Alport nephropathyHow frequent is crescent formation in human Alportnephropathy? We reviewed the files in the Departmentof Pathology at Washington University in St Louis andthe Mayo Clinic, Rochester, MN, covering 66 500 renalbiopsies from 1991 to 2011. Six hundred and sixty-five biopsies with a final diagnosis of Alport wereidentified. Fibrous or epithelial crescents were foundin three cases (0.4%), but in each case only one outof 9–36 glomeruli available for analysis containedcrescents (Table 1 and Figure 11).

Discussion

The pathophysiology of glomerular crescents, definedas layers of cells lining Bowman’s capsule and invad-ing the urinary space, remains controversial [2,4]. Pre-vious studies have mostly focused on inflammatorymediators to underlie crescent formation in RPGNbecause crescents are a defining feature [2,3]. How-ever, crescents have been reported in the late stagesof almost every form of glomerular disease [42–44].In order to better define and understand if and howa primary lesion in the GBM might cause glomeru-lar crescents, we studied 129X1/SvJ Col4A3-deficientmice of varying age. The major finding was that therewas an increase in the number of crescents with ageand that they also increased in severity, judging by theextent to which they filled the glomerulus. This hasnot been previously recognized. Our work was trig-gered by our observations in late-stage murine Alport

Figure 9. Silver stain of late-stage Alport nephropathy. Nine-week-old wild-type (A) and Col4A3-deficient mice (B–H). Normal-appearing glomeruli (A). (B) Age-matched Col4A3-deficient miceshow altered glomerular pathology with areas of maintainedGBM coverage by podocytes (thin arrow), while other areasshow podocyte detachment, erythrocytes in Bowman’s space, andcytoplasmic adhesions extending to activated parietal epithelialcells (large arrows). (C) Few glomeruli display synechiae betweendenuded glomerular capillaries and parietal epithelial cells (arrows).(D) Glomeruli affected by segmental lesions illustrate collapse ofglomerular capillaries (thin arrows). Note the extracellellular matrixproduced by proliferating parietal epithelial cells (white arrows) andred blood cells (RBCs) in intratubular lumens. (E) Many glomerulishow global cellular crescents with admixed RBCs (white arrow).(F) Parietal epithelial cell proliferation starts at the Bowman’scapsule (thin arrows). These cells grow in an unpolarized mannerand produce extracellular matrix all around them (white arrows),giving the capsule a multi-layer appearance. (G) Note that at laterstages, the matrix produced by unpolarized parietal epithelial cells(white arrows) exceeds the thickness of the GBM, which can clearlybe identified by intracapillary erythrocytes in the left part of thisglomerulus. (H) Late-stage Alport glomerulopathy is characterizedby fibrous crescents where matrix-producing parietal epithelialcells dominate and only a few glomerular capillaries are left (asindicated by luminal RBCs).

nephropathy using 129X1/SvJ Col4A3-deficient mice.We were intrigued by the frequent crescent forma-tion in these mice and the paucity of related litera-ture [32,33,45]. We reviewed the kidney sections fromour own previously published interventional studies of129X1/SvJ Col4A3-deficient mice and by reviewingthe images of published reports on Col4A3-deficient



Col1a1

Vimentin

2.0x10-4

1.5x10-4

1.0x10-4

5.0x10-5

04 wk

mR

NA

/18s

rR

NA

7.5x10-4

6.0x10-4

4.5x10-4

3.0x10-4

4.0x10-4

3.0x10-4

2.0x10-4

1.0x10-4

1.5x10-4

0

0

mR

NA

/18s

rR

NA

mR

NA

/18s

rR

NA

6 wk 9 wk

WT

Col4a3-/-

4 wk 6 wk 9 wk

WT

Col4a3-/-

Fibronectin

4 wk 6 wk 9 wk

WT

Col4a3-/-

Figure 10. Fibrocellular crescents in late-stage Alport nephropathy. (A–C) Acellular glomeruli of 9-week-old wild-type (A) and Col4A3-deficient mice (B, C) viewed with scanning electron microscopy. Note that wild-type glomeruli have smooth GBM surfaces without matrixin the extracapillary space. (B, C) In contrast, glomeruli of Col4A3-deficient mice display a rough GBM surface (∗) forming comb-likescaffolds and thick Bowman’s capsule (black arrowhead) characterized by multiple layers of extracellular matrix (white arrowheads inB). (C) Extracellular matrix forms dense plaques covering the entire surface of the glomerulus. (D) RNA was prepared from 4-, 6-, and9-week-old wild-type and Col4A3-deficient mice and real-time PCR was performed for collagen Iα, vimentin, and fibronectin.

Table 1. Clinical and pathological findings of Alport patients with crescents on renal biopsySerum Alport family

Case Age/sex Proteinuria Haematuria creatinine history GLS SGS GBM on EM Crescents

1 15/F X X 0.4 mg/dl Yes 0/36 Thin and split 1/362 31/M No X 2.5 mg/dl Yes GLS:10/19SGS:4/19 Thin and split 1/193 13/F No X ‘Normal’ Yes GLS:3/9 SGS:2/9 Thin with ‘crumbs’ 1/9

Age (years). GLS = global glomerulosclerosis; SGS = segmental glomerulosclerosis; GBM = glomerular basement membrane; EM = electron microscopy.


M Ryu et al

Figure 11. Representative pathology from Alport patients with crescents on renal biopsy as described in Table 1. (A, B) Patient 1. Focalsegmental epithelial crescent (silver stain; original magnification: × 200). (B) Electron microscopy from the same biopsy shows thin andsplit GBM (EM; original magnification: × 6000). (C) Patient 2. Fibrocellular crescent from a man with advanced disease (50% globallysclerosed glomeruli; Trichrome; original magnification: × 400). (D) Patient 3. A focal segmental fibrocellular crescent is adjacent to aglobally sclerosed glomerulus (upper left corner) and tubulointerstitial atrophy and chronic inflammation (Trichrome; original magnification:× 100).

mice, we confirmed that crescents are a feature inAlport mice that has not yet been carefully evaluated.Following a review of our own renal biopsies of humanAlport cases with crescents, we found that crescentshad occasionally been reported previously in the Alportliterature [46–48], albeit this phenomenon seems tooccur at low frequency. This discrepancy between miceand humans may relate entirely to the specific geneticbackground of 129X1/SvJ mice [33,45]. However, inhumans, a selection bias cannot be excluded becauserenal biopsy is usually performed at earlier stages ofAlport; therefore, biopsy cohorts may not adequatelymirror the glomerular changes of late-stage Alportglomerulopathy [49]. Nevertheless and apart from thefrequency of crescents in human Alport nephropa-thy, we considered 129X1/SvJ Col4A3-deficient micea suitable model to study crescent formation in non-inflammatory forms of chronic kidney disease.

Fibrinogen deposition contributes to crescent for-mation in anti-GBM disease, a highly inflammatorydisorder associated with a massive induction of pro-inflammatory cytokines, ruptures of Bowman’s cap-sule, and an influx of neutrophils, macrophages, andT cells into the glomerulus [2,3]. However, fibrino-gen activation rather represents an element of clot-ting upon traumatic vascular injury [14]. Therefore,we propose that glomerular fibrinogen activation isa secondary event following vascular injury withinthe glomerular tuft, rather than secondary to cellularinflammation or related to Bowman’s capsule ruptures.Col4A3-deficient mice allow us to dissect vascular

from inflammatory glomerular injury because glomeru-lar inflammation and ruptures of Bowman’s capsuleare not apparent even at later stages of the disease pro-cess [19]. We documented vascular damage by demon-strating GBM discontinuities, a structural prerequisiteof haematuria and a hallmark of Alport nephropathyalso in humans [49]. In the GBM of Col4A3-deficientmice, the physiological collagen IVA3/4/5 chains arereplaced by collagen IVA1/2 chains with fewer disul-phide bonds, reducing the mechanical strength of theGBM [50]. Cosgrove et al showed that the abnormalcollagen assembly in the GBM of Col4A3-deficientmice is associated with proteolytic degradation of theGBM by matrix metalloproteinases such as MMP12[51,52]. This process usually begins at the capillaryloops while the mesangium undergoes progressive scle-rosis [32]. As such, the characteristic thinning andsplitting of the GBM in Alport nephropathy ultimatelyleads to GBM perforation and diffuse disruption, asillustrated by us using transmission and scanning elec-tron microscopy of acellular glomeruli. We observedGBM perforations similar to those that Steve Bonsibhad described in human crescentic glomerulonephritisusing the same imaging technology [27]. Here, the genedefect-driven vascular injury allows red blood cells andprotein to pass through the GBM; that is why Alport’ssyndrome was initially described as ‘hereditary nephri-tis’ [53]. The functional significance of GBM rupturesfor renal outcome is also illustrated by the fact thatACE inhibitors, which reduce filtration pressure acrossthe glomerular filtration barrier, improve glomerular



pathology and significantly prolong the (renal) life spanof mice and humans with Alport nephropathy [23,54].

By injecting macromolecular FITC-dextran intoCol4A3-deficient or wild-type mice, we demonstratedthat serum components leak into the urinary space,causing proliferation of PECs. Under physiologicalconditions, PECs are exposed to serum-free glomeru-lar ultrafiltrate only. GBM ruptures may expose murinePECs to a markedly increased serum concentrationin vivo, similar to the concentration-dependent mito-genic effect on murine and human PECs that weobserved in vitro. In line with other published reports,fibrinogen inactivation prevents crescent formation inCol4A3-deficient mice, indicating that activation of thecoagulation cascade during vascular damage serves asa mitogenic stimulus on epithelial cells [35,36]. Ourfinding that PECs and not visceral podocytes consti-tute the majority of the cells within the crescents ofCol4A3-deficient mice is in agreement with a numberof older and more recent studies that have character-ized the composition of glomerular crescents in humanand experimental glomerulonephritis [55,56].

The proliferation of epithelial cells upon vascularinjury represents a mechanism of epithelial repair [14].This process remains enigmatic because repairing thepodocyte sheet would require considerable structuraldifferentiation and redistribution of the local podocyteprogenitors onto the glomerular tuft, which may occur,albeit only in certain settings [9,37,44,57,58]. How-ever, sufficient podocyte repair may be impossibleat the stage of a largely destroyed GBM. In addi-tion, de novo podocytes would create an abnormalGBM in Col4A3-deficient mice. The persistent mito-genic stimuli therefore seem to drive the migration andproliferation of PECs inside Bowman’s space like itwas recently documented for epidermal growth factorreceptor signalling [59] or after subtotal PEC ablation[60]. During this process, the parietal epithelial cellsundergo a phenotype switch involving loss of polar-ity and de novo expression of mesenchymal markersas observed in human crescentic glomerulonephri-tis [6,12,40,55]. This phenomenon, named epithe-lial–mesenchymal transition (EMT) by some authors[40–42], is proposed to occur in tubular epithelial cells,triggered by TGF-β [55,61]. Furthermore, the polarityof epithelia implies that extracellular matrix compo-nents are secreted only at their basal aspect. However,our silver stains and scanning electron microscopyimages clearly demonstrate that in Col4A3-deficientmice, epithelial cells produce matrix components allaround their circumference. This loss of epithelialpolarity is consistent with a transition to a mesenchy-mal phenotype.

A limitation of our study is the use of a single ani-mal model with doubtful clinical significance. From aclinical point, late stages of diabetic nephropathy andhypertension are much more frequent causes of over-shooting PEC repair but we had to choose an exper-imental system which is less complex and that lacksthe bias of metabolic or oxidative stress. We propose

that the mechanism demonstrated here contributes tothe late stage of many glomerular diseases wheneverdirect or indirect vascular injury adds to the more orless complex but progressive disease process.

Altogether, our data support the concept thatglomerular vascular damage causes plasma leakage,activation of fibrinogen and possibly of other serumfactors as well. These factors, which may physiolog-ically promote wound healing, appear to trigger PECproliferation but often no repair of the glomerular fil-tration barrier. Proliferating PECs pile up and blockBowman’s space. This process involves a phenotypictransition of PECs, characterized by loss of polarityand increased production of extracellular matrix com-ponents, resulting in fibrocellular crescents. These dataprovide a unifying concept for crescent formation innon-inflammatory glomerular injury.

AcknowledgmentThe work was supported by the Association pourl’Information et la Recherche sur les Maladies RenaleGenetiques, France. We also thank Iana Mandelbaumand Dan Draganovic for their technical assistance,as well as the O’Brian Kidney Center at Washing-ton University in St Louis for support to HL. Adri-ana Migliorini is a graduate student of the DeutscheForschungsgemeinschaft (GRK 1202).

Author contribution statement

HJA and MR designed the study. MR, AM, OG,HH, and HL conducted experiments. MR, AM, NM,OG, PTB, PR, HL, and HJA analysed data. SS andPR provided cells. HJA, HL, SS, and PR wrote themanuscript.

References

1. Whitworth JA, Morel-Maroger L, Mignon F, et al . The signifi-cance of extracapillary proliferation. Clinicopathological review of60 patients. Nephron 1976; 16: 1–19.

2. Tipping PG, Kitching PR, Holdsworth SR. The formation of theglomerular crescent. In Immunologic Renal Diseases (2nd edn),Neilson EG, Couser WG (eds). Lippincott Williams & Wilkins:Philadelphia, 2001.

3. Atkins RC, Nikolic-Paterson DJ, Song Q, et al . Modulators ofcrescentic glomerulonephritis. J Am Soc Nephrol 1996; 7:2271–2278.

4. Singh SK, Jeansson M, Quaggin SE. New insights into the patho-genesis of cellular crescents. Curr Opin Nephrol Hypertens 2011;20: 258–262.

5. Shirato I, Asanuma K, Takeda Y, et al . Protein gene product 9.5is selectively localized in parietal epithelial cells of Bowman’scapsule in the rat kidney. J Am Soc Nephrol 2000; 11: 2381–2386.

6. Smeets B, Uhlig S, Fuss A, et al . Tracing the origin of glomerularextracapillary lesions from parietal epithelial cells. J Am SocNephrol 2009; 20: 2604–2615.

7. Moeller MJ, Soofi A, Hartmann I, et al . Podocytes populate cellu-lar crescents in a murine model of inflammatory glomerulonephri-tis. J Am Soc Nephrol 2004; 15: 61–67.


M Ryu et al

8. Ding M, Cui S, Li C, et al . Loss of the tumor suppressor Vhlh

leads to upregulation of Cxcr4 and rapidly progressive glomeru-

lonephritis in mice. Nature Med 2006; 12: 1081–1087.

9. Lasagni L, Ballerini L, Angelotti ML, et al . Notch activation dif-

ferentially regulates renal progenitors proliferation and differenti-

ation toward the podocyte lineage in glomerular disorders. Stem

Cells 2010; 28: 1674–1685.

10. Shkreli M, Sarin KY, Pech MF, et al . Reversible cell-cycle entry

in adult kidney podocytes through regulated control of telomerase

and Wnt signaling. Nature Med 2011; 18: 111–119.

11. Albaqumi M, Barisoni L. Current views on collapsing glomeru-

lopathy. J Am Soc Nephrol 2008; 19: 1276–1281.

12. Ohse T, Vaughan MR, Kopp JB, et al . De novo expression of

podocyte proteins in parietal epithelial cells during experimental

glomerular disease. Am J Physiol Renal Physiol 2010; 298:

F702–F711.

13. Wilde B, van Paassen P, Witzke O, et al . New pathophysiological

insights and treatment of ANCA-associated vasculitis. Kidney Int

2011; 79: 599–612.

14. Gurtner GC, Werner S, Barrandon Y, et al . Wound repair and

regeneration. Nature 2008; 453: 314–321.

15. Martin P, Leibovich SJ. Inflammatory cells during wound repair:

the good, the bad and the ugly. Trends Cell Biol 2005; 15:

599–607.

16. Maeshima A, Yamashita S, Nojima Y. Identification of renal

progenitor-like tubular cells that participate in the regeneration pro-

cesses of the kidney. J Am Soc Nephrol 2003; 14: 3138–3146.

17. Jedlicka J, Soleiman A, Draganovici D, et al . Interstitial

inflammation in Alport syndrome. Hum Pathol 2010; 41:

582–593.

18. Ryu M, Kulkarni OP, Radomska E, et al . Bacterial CpG-DNA

accelerates Alport glomerulosclerosis by inducing an M1

macrophage phenotype and tumor necrosis factor-alpha-mediated

podocyte loss. Kidney Int 2011; 79: 189–198.

19. Clauss S, Gross O, Kulkarni O, et al . Ccl2/Mcp-1 blockade

reduces glomerular and interstitial macrophages but does not ame-

liorate renal pathology in collagen4A3-deficient mice with autoso-

mal recessive Alport nephropathy. J Pathol 2009; 218: 40–47.

20. Ninichuk V, Gross O, Reichel C, et al . Delayed chemokine recep-

tor 1 blockade prolongs survival in collagen 4A3-deficient mice

with Alport disease. J Am Soc Nephrol 2005; 16: 977–985.

21. Ninichuk V, Gross O, Segerer S, et al . Multipotent mesenchymal

stem cells reduce interstitial fibrosis but do not delay progression

of chronic kidney disease in collagen4A3-deficient mice. Kidney

Int 2006; 70: 121–129.

22. Ryu M, Mulay SR, Miosge N, et al . Tumour necrosis factor-alpha

drives Alport glomerulosclerosis in mice by promoting podocyte

apoptosis. J Pathol 2012; 226: 120–131.

23. Gross O, Beirowski B, Koepke ML, et al . Preemptive ramipril

therapy delays renal failure and reduces renal fibrosis in COL4A3-

knockout mice with Alport syndrome. Kidney Int 2003; 63:

438–446.

24. Gross O, Koepke ML, Beirowski B, et al . Nephroprotection by

antifibrotic and anti-inflammatory effects of the vasopeptidase

inhibitor AVE7688. Kidney Int 2005; 68: 456–463.

25. Ninichuk V, Khandoga AG, Segerer S, et al . The role of intersti-

tial macrophages in nephropathy of type 2 diabetic db/db mice. Am

J Pathol 2007; 170: 1267–1276.

26. Miosge N, Klenczar C, Herken R, et al . Organization of the

myotendinous junction is dependent on the presence of alpha7beta1

integrin. Lab Invest 1999; 79: 1591–1599.

27. Bonsib SM. Glomerular basement membrane discontinuities. Scan-

ning electron microscopic study of acellular glomeruli. Am J Pathol

1985; 119: 357–360.

28. Ninichuk V, Clauss S, Kulkarni O, et al . Late onset of Ccl2 block-ade with the Spiegelmer mNOX-E36-3’PEG prevents glomeru-losclerosis and improves glomerular filtration rate in db/db mice.Am J Pathol 2008; 172: 628–637.

29. Sagrinati C, Netti GS, Mazzinghi B, et al . Isolation and char-acterization of multipotent progenitor cells from the Bowman’scapsule of adult human kidneys. J Am Soc Nephrol 2006; 17:2443–2456.

30. Ohse T, Pippin JW, Vaughan MR, et al . Establishment of condi-tionally immortalized mouse glomerular parietal epithelial cells inculture. J Am Soc Nephrol 2008; 19: 1879–1890.

31. Teixeira Vde P, Blattner SM, Li M, et al . Functional consequencesof integrin-linked kinase activation in podocyte damage. Kidney Int

2005; 67: 514–523.32. Cosgrove D, Meehan DT, Grunkemeyer JA, et al . Collagen

COL4A3 knockout: a mouse model for autosomal Alport syn-drome. Genes Dev 1996; 10: 2981–2992.

33. Andrews KL, Mudd JL, Li C, et al . Quantitative trait loci influ-ence renal disease progression in a mouse model of Alport syn-drome. Am J Pathol 2002; 160: 721–730.

34. Browne G, Brown PA, Tomson CR, et al . Retransplantation inAlport post-transplant anti-GBM disease. Kidney Int 2004; 65:675–681.

35. Drew AF, Tucker HL, Liu H, et al . Crescentic glomerulonephritisis diminished in fibrinogen-deficient mice. Am J Physiol Renal

Physiol 2001; 281: F1157–F1163.36. Tipping PG, Thomson NM, Holdsworth SR. A comparison of

fibrinolytic and defibrinating agents in established experimentalglomerulonephritis. Br J Exp Pathol 1986; 67: 481–491.

37. Ronconi E, Sagrinati C, Angelotti ML, et al . Regeneration ofglomerular podocytes by human renal progenitors. J Am Soc

Nephrol 2009; 20: 322–332.38. Zeisberg M, Duffield JS. Resolved: EMT produces fibroblasts in

the kidney. J Am Soc Nephrol 2010; 21: 1247–1253.39. Kriz W, Kaissling B, Le Hir M. Epithelial–mesenchymal transi-

tion (EMT) in kidney fibrosis: fact or fantasy? J Clin Invest 2011;121: 468–474.

40. Bariety J, Hill GS, Mandet C, et al . Glomerular epithelial–mesenchymal transdifferentiation in pauci-immune crescen-tic glomerulonephritis. Nephrol Dial Transplant 2003; 18:1777–1784.

41. Fujigaki Y, Sun DF, Fujimoto T, et al . Mechanisms and kineticsof Bowman’s epithelial–myofibroblast transdifferentiation in theformation of glomerular crescents. Nephron 2002; 92: 203–212.

42. Ng YY, Fan JM, Mu W, et al . Glomerular epithelial–myofibroblast transdifferentiation in the evolution of glomerularcrescent formation. Nephrol Dial Transplant 1999; 14: 2860–2872.

43. Stejskal J, Pirani CL, Okada M, et al . Discontinuities (gaps) ofthe glomerular capillary wall and basement membrane in renaldiseases. Lab Invest 1973; 28: 149–169.

44. Benigni A, Morigi M, Rizzo P, et al . Inhibiting angiotensin-converting enzyme promotes renal repair by limiting progenitor cellproliferation and restoring the glomerular architecture. Am J Pathol

2011; 179: 628–638.45. Cosgrove D, Kalluri R, Miner JH, et al . Choosing a mouse model

to study the molecular pathobiology of Alport glomerulonephritis.Kidney Int 2007; 71: 615–618.

46. Afonso A, Valente I, Macedo L, et al . Alport syndrome—a rarehistological presentation. Port J Nephrol Hypert 2010; 24: 51–55.

47. Spear GS, Slusser RJ. Alport’s syndrome. Emphasizing electronmicroscopic studies of the glomerulus. Am J Pathol 1972; 69:213–224.

48. Gubler M, Levy M, Broyer M, et al . Alport’s syndrome. A reportof 58 cases and a review of the literature. Am J Med 1981; 70:493–505.



49. Heidet L, Gubler MC. The renal lesions of Alport syndrome. J Am

Soc Nephrol 2009; 20: 1210–1215.50. Hudson BG, Tryggvason K, Sundaramoorthy M, et al . Alport’s

syndrome, Goodpasture’s syndrome, and type IV collagen. N Engl

J Med 2003; 348: 2543–2556.51. Rao VH, Meehan DT, Delimont D, et al . Role for macrophage

metalloelastase in glomerular basement membrane damage asso-ciated with Alport syndrome. Am J Pathol 2006; 169: 32–46.

52. Zeisberg M, Khurana M, Rao VH, et al . Stage-specific action ofmatrix metalloproteinases influences progressive hereditary kidneydisease. PLoS Med 2006; 3: e100.

53. Alport AC. Hereditary familial congenital haemorrhagic nephritis.Br Med J 1927; 1: 504–506.

54. Gross O, Licht C, Anders HJ, et al . Early angiotensin-convertingenzyme inhibition in Alport syndrome delays renal failure andimproves life expectancy. Kidney Int 2011; 81: 494–501.

55. Smeets B, Angelotti ML, Rizzo P, et al . Renal progenitor cellscontribute to hyperplastic lesions of podocytopathies and cres-centic glomerulonephritis. J Am Soc Nephrol 2009; 20:2593–2603.

SUPPORTING INFORMATION ON THE INTERNET

The following supporting information may be found in the online version of this article.

Figure S1. Representative images from 6- and 9-week-old Col4A3-deficient or wild-type mice at low and high magnification.

56. Romagnani P. Parietal epithelial cells: their role in health anddisease. Contrib Nephrol 2011; 169: 23–36.

57. Lasagni L, Romagnani P. Glomerular epithelial stem cells: thegood, the bad, and the ugly. J Am Soc Nephrol 2010; 21:1612–1619.

58. Sayyed SG, Hagele H, Kulkarni OP, et al . Podocytes producehomeostatic chemokine stromal cell-derived factor-1/CXCL12,which contributes to glomerulosclerosis, podocyte loss and albu-minuria in a mouse model of type 2 diabetes. Diabetologia 2009;52: 2445–2454.

59. Bollee G, Flamant M, Schordan S, et al . Epidermal growth factorreceptor promotes glomerular injury and renal failure in rapidlyprogressive crescentic glomerulonephritis. Nature Med 2011; 17:1521.

60. Sicking EM, Fuss A, Uhlig S, et al . Subtotal ablation of parietalepithelial cells induces crescent formation. J Am Soc Nephrol 2012;23: 629–640.

61. Zeisberg M, Neilson EG. Biomarkers for epithelial–mesenchymaltransitions. J Clin Invest 2009; 119: 1429–1437.


Plasma leakage through glomerular basement membrane ruptures triggers the proliferation of parietal epithelial cells and crescent formation in non-inflammatory glomerular injury

Documents