Mast Cells in Rapidly Progressive Glomerulonephritis

TIBOR TOTH,* RIA TOTH-JAKATICS,* SHIRO JIMI,* MAKOTO IHARA,†

HIDENORI URATA,† and SHIGEO TAKEBAYASHI**Second Department of Pathology and†Department of Internal Medicine, Fukuoka University, School ofMedicine, Fukuoka, Japan.

Abstract. The role of mast cells (MC) in tubulointerstitialdamage in glomerulonephritis (GN) is not fully understood.The distribution of MC was compared in renal biopsies from50 patients with different stages of rapidly progressive GN(RPGN) and in 20 control samples. The immunoreactivity ofrenal MC with anti-tryptase and anti-chymase antibodies wasstudied. Interstitial myofibroblasts were stained with anti-a-smooth muscle actin (a-SMA) antibody, and inflammatorycells were identified by anti-CD3, -CD20, and -CD68 mono-clonal antibodies. Positively stained cells were counted, andthe relative interstitial and fractional areas of anti-a-SMA-stained cells were measured. MC were rarely found in controlsamples. In contrast, samples showing crescentic GN containednumerous tryptase-positive MC (MCT) (43.7 6 4.65 versus7.14 6 1.3/mm2) and fewer tryptase- and chymase-positiveMC (MCTC) (13.8 6 1.86 versus1.89 6 0.86/mm2) in therenal interstitium but never in the glomerulus. Double immu-

nostaining demonstrated the presence of both phenotypes ofMC. Accumulation of MC was significantly correlated with thenumbers of T lymphocytes (MCT, r 5 0.67) and interstitialmacrophages (MCT, r 5 0.455). There was also a significantcorrelation between the number of MCT and the relative inter-stitial area. The number of MCTC was well correlated with thefractional area ofa-SMA-positive interstitium (r 5 0.749) andthe percentage of the interstitial fibrotic area (r 5 0.598). Therewas also a significant negative correlation between interstitialMCTC accumulation and creatinine clearance (r 5 0.661). Thedensity of MCTC was higher (1.4-fold) in advanced forms ofGN associated with fibrocellular crescents and interstitial fi-brosis. These results show the potential involvement of MC inthe fibroproliferative process in the renal interstitium of pa-tients with RPGN. The results indicate that these cells consti-tute part of the overall inflammatory cell accumulation inRPGN.

Tubulointerstitial (TI) lesions are considered prognostic fea-tures of various renal diseases. Despite numerous studies inhuman subjects and experimental animal models, the exactpathogenic factors involved in TI damage remain unclear. Inparticular, the role of mast cells (MC) in glomerulonephritis(GN) with TI lesions has not yet been fully analyzed. MC playan important role in allergic inflammatory conditions such asasthma, and their number is increased in chronic inflammatoryconditions. Their presence in tissue is often accompanied byfibrosis, such as liver cirrhosis, pulmonary fibrosis, and woundhealing, suggesting that they might be involved in these patho-logic conditions (1–3).

Immunohistochemical studies have demonstrated thepresence of two MC phenotypes, which are distinguished bytheir neutral serine protease contents. The tryptase-positiveMC (MCT) phenotype contains only tryptase, whereas thetryptase- and chymase-positive MC (MCTC) phenotype con-tains both tryptase and chymase (4). The MCT phenotype

seems to play a role in host defense, whereas the MCTC

phenotype seems to represent “non-immune system-related”cells involved in fibrosis (5). MC have been demonstratedby histochemical or immunohistochemical methods in var-ious renal diseases, including nephrosclerosis, pyelonephri-tis, chronic GN, amyloidosis (6,7), renal vasculitis (8), acutecellular rejection of renal allografts (9) and IgA nephritis(10). Furthermore, both MC phenotypes have been detectedin diabetic nephropathy (11). There is general agreementthat MC increase in number during chronicity and progres-sion of various renal lesions (6,9,11)

Activated MC synthesize, store, and release several bioac-tive mediators. For example, histamine and heparin are mito-genic for fibroblasts and induce collagen synthesis, as well asactivating collagenase (12–17). MC are also known to secretebasic fibroblast growth factor in IgA nephritis (10). Using thein situ hybridization technique, Rugeret al. (11) demonstratedthat human MC can produce type VIII collagen bothin vitroand in vivo in diabetic nephropathy. However, whether MCmodulate the inflammatory and fibrotic processes in GN re-mains to be elucidated.

In this study, we evaluated the potential role of MC in thepathogenesis of TI lesions in rapidly progressive GN (RPGN),which is characterized by glomerular crescent formation,marked inflammatory processes, and TI fibrosis. Specifically,we examined the role of MC in the fibrotic process in TIlesions.

Received September 22, 1998. Accepted December 26, 1998.Correspondence to Dr. Tibor To´th, The Second Department of Pathology,Fukuoka University, School of Medicine, Nanakuma 7-45-1, Jonan-ku,Fukuoka 814-0180, Japan. Phone:181-92-801-1011; Fax:181-92-863-8383;E-mail: mm038492@Msat.fukuoka-u.ac.jp

1046-6673/1007-1498Journal of the American Society of NephrologyCopyright © 1999 by the American Society of Nephrology

J Am Soc Nephrol 10: 1498–1505, 1999

Materials and MethodsDefinition

RPGN was defined as glomerular disease with extensive glomer-ular crescent formation (.50%) associated with rapid loss of renalfunction (usually 50% reduction of GFR within 3 mo).

Tissue SamplesKidney tissue samples were obtained from 50 randomly selected

patients (23 men and 27 women; mean age, 53.36 15.7 yr) withcrescentic GN at the time of renal biopsy. Only samples containing atleast 10 glomeruli in the paraffin blocks were included in this study.All samples showed extensive crescent formation in.50% of theglomeruli. Histologic studies using light, immunofluorescence, andelectron microscopy demonstrated typical extracapillary cell prolifer-ation with crescent formation. Patients with RPGN were divided intothree subgroups on the basis of the etiologic classification criteria ofCouser (18) (Table 1). Control samples consisted of 20 randomlyselected renal biopsy samples from patients with thin glomerularmembrane disease, with normal extraglomerular renal structure andwithout immune deposition, glomerulosclerosis, or inflammation. Allpatients in both groups were adults (.16 yr of age). Clinical andlaboratory data were recorded at the time of renal biopsy. The studyprotocol was approved by the Human Ethics Review Committee ofFukuoka University, and signed consent forms were obtained.

Patients with RPGN were divided into two groups according thetype of glomerular crescents. Stage I tissue samples (n 5 25) con-tained fresh large cellular crescents around the glomerular capillarytufts, representing early stages of the disease process. Stage II tissuesamples (n 5 25) contained fibrocellular crescents in more thanone-half of the crescentic glomeruli, representing a progressive glo-merular process with or without interstitial fibrosis.

Tissue PreparationTissue sections were stained with hematoxylin and eosin, Masson’s

trichrome, periodic acid-Schiff, and periodic acid-methenamine-silverstains. To identify MC, sections were stained with toluidine blue.

Electron MicroscopyA portion of each biopsy sample was fixed in 1.4% phosphate-

buffered glutaraldehyde, post-fixed in OsO4, and embedded in Eponresin. Ultrathin sections were cut and stained with uranyl acetate andlead citrate. All samples were examined by electron microscopy.

Immunohistochemical AnalysesAnti-human chymase rabbit IgG (protein-origin antibody) was ob-

tained using the methods described previously (19). Peptide-originmonoclonal antibody for chymase was a kind gift from Drs. Y. Kisoand N. Ishihara of Suntory Biomedical Research Center (Osaka,Japan). Anti-human tryptase monoclonal antibody was purchasedfrom Chemicon International (Temecula, CA). Anti-CD68, anti-CD20, anti-CD3, and anti-a-smooth muscle actin (a-SMA) monoclo-nal antibodies were purchased from Dako (Glostrup, Denmark). Forimmunohistochemical staining of the tissue samples, one part of eachbiopsy sample was fixed for light microscopy in 10% buffered for-malin and was embedded in paraffin. Three-micrometer-thick, depar-affinized, serial sections were washed in 0.05 M Tris-HCl buffercontaining 0.145 M NaCl, pH 7.5 (Tris-buffered saline [TBS]). Pro-tease pretreatment was performed for staining for chymase and CD68,whereas autoclave pretreatment (121°C, 10 min) was performed forstaining fora-SMA. Sections were initially treated with 1% skim milk(DIFCO Laboratories, Detroit, MI) and incubated for 1 h at20°C withthe first antibody (anti-chymase antibody, 50 mg/ml; anti-tryptaseantibody, 0.32 mg/ml; each dissolved in 3% bovine serum albumin).For negative control staining, we used vehicle alone or nonimmunizedanimal Ig. After incubation, sections were washed with TBS and thenincubated for 30 min with alkaline phosphatase (ALP)-conjugatedsecond antibody against rabbit or mouse Ig (Dako). After washing,sections were incubated for 30 min with ALP-conjugated third anti-body against ALP (Dako). The sections were then stained with asolution of 0.01% new fuchsin (Merck, Darmstadt, Germany), 0.01%NaN, 10 mg of naphthol AS-BI phosphate (Sigma Chemical Co., St.Louis, MO), and 0.1 ml ofN,N-dimethylformamide (Wako, Osaka,Japan) in 40 ml of 0.2 M Tris-HCl buffer, pH 8.2. After being washedwith TBS, sections were post-fixed in 2% buffered glutaraldehyde,counterstained with hematoxylin, and used for immunohistochemicaldetection.

In a preliminary study, we immunohistochemically compared twodifferent antibodies for human chymase,i.e., a peptide-origin anti-body and a protein-origin antibody (Chemicon International). Theresults showed that the two antibodies yielded similar precisions;therefore, the protein-origin antibody was used in the following ex-periments. Tissue for positive controls consisted of biopsy samplesfrom the small intestine for chymase and tryptase and the renalvasculature fora-SMA.

Double ImmunostainingAfter deparaffinized sections were treated with 0.005% protease

(Dako-Japan, Osaka, Japan) for 10 min, they were incubated for 60min at 20°C with the first mouse antibody against human chymase(Chemicon). After being washed with TBS, sections were incubatedfor 60 min at 20°C with tetrarhodamine isothiocyanate-conjugatedsecond rabbit antibody against mouse Ig (Dako). Sections were thentreated with 1% skim milk for 30 min to minimize the backgroundlevels. Sections were again incubated with the first rabbit antibodyagainst human tryptase (BioPur Co., Switzerland), for 60 min at 20°C.After being washed with TBS, sections were incubated for 60 min at20°C with FITC-conjugated swine antibody against rabbit Ig (Dako).In the next step, nuclei were stained with 49,6-diamidino-2-phenylin-

Table 1. Classification and distribution of cases of RPGNa

Type No. ofCases

Anti-GBM antibody nephritisnephritis with lung hemorrhage

(Goodpasture’s syndrome)7

crescentic nephritis without lung hemorrhage 8complicating membranous nephropathy 1

Immune complex nephritisIgA nephritis 6lupus nephritis 2postinfectious glomerulonephritis 2idiopathic nephritis 4

Nonimmune deposit nephritispauci-immune (idiopathic) nephritis 18Wegener’s granulomatous nephritis 2

a RPGN, rapidly progressive glomerulonephritis; GBM,glomerular basement membrane.

J Am Soc Nephrol 10: 1498–1505, 1999 Mast Cells in RPGN 1499

dole (Sigma). Sections were then mounted and examined under afluorescence microscope (Axioplan; Carl Zeiss, Jena, Germany), witha fluorescence imaging system (Isis; MetaSystems, Altlussheim, Ger-many).

Morphometric AnalysesSeveral indices of the extent of the pathologic processes were

derived using a previously published method (20). (1) A standardpoint-counting method was used to quantify the relative interstitialarea (Aint) of the renal cortex. In this method, sections stained by theperiodic acid-methenamine-silver method were examined under highmagnification (3400), using a 121-point (100-square) eyepiece mi-crometer of 1 mm2. A total of 10 consecutive nonoverlapping corticalfields (area, 0.625 mm2) were analyzed in each section of the biopsy.Points overlying the tubular basement membrane and interstitial spacewere counted, whereas those falling on either Bowman’s capsule orthe peritubular capillaries were not. Points falling on glomerularstructures or on larger vessels were excluded from the total count. Theresults were expressed according to the following formula:Aint 5[(Number of grid intersections on the cortical interstitium/Total num-ber of grid intersections)3 100]. (2) Interstitial immunoperoxidasestaining for a-SMA was quantified by the point-counting methoddescribed above. The fractional area was also calculated, using theformula given above. (3) The fractional area stained by Masson’strichrome stain was quantified by the point-counting method de-scribed above, and the results were expressed as the fractional areausing the formula given above. (4) The numbers of interstitial lym-phocytes (CD201 and CD31) and macrophages (CD681 and Ki-11)present in the cortical interstitial area were counted, under highmagnification (3400), in 10 adjacent nonoverlapping cortical fields(total area, 0.625 mm2/biopsy specimen). Only cells with a clearlyidentifiable nucleus were counted. Finally, the number of countedcells was expressed as cells per unit area (square millimeter).

Statistical AnalysesData were expressed as mean6 SEM. Differences between groups

were examined for statistical significance using thet test and one-way

ANOVA. Association of categorical variables was examined using thex2 test. Correlation between variables was assessed by linear regres-sion analysis. AP value of ,5% denoted a statistically significantdifference.

ResultsImmunohistochemical Observations

A representative tissue sample containing MCT is shown inFigure 1. Double staining for both proteases showed MCpositive for tryptase only, whereas others were stained for bothtryptase and chymase (Figure 2). Chymase was not observed incells lacking tryptase. MC were mostly round in shape, with acentral nucleus and abundant cytoplasm containing numeroustryptase- or chymase-positive granules (Figure 3, inset). Aproportion of MC appeared elongated, with cytoplasmic pro-cesses, and contained fewer granules. Electron-microscopicexamination clearly demonstrated the presence of numerousintracytoplasmic dense granules (Figure 3). MC were localizedin the peritubular or periglomerular region, in close proximityto interstitial fibroblasts (Figure 3), and were surrounded bycell processes of fibroblasts. Intraglomerular MC were neverdetected in our tissue samples, although a few intratubularMCT and/or MCTC were found. MC were mostly located on theperiphery, rather than within areas of lymphocytic accumula-tion. Furthermore, MC were observed around interstitialCD681 macrophages anda-SMA-positive interstitial areas(Figure 4).

The densities of both MC phenotypes were significantlyhigher in the RPGN group than in control samples (MCT,43.76 4.6 versus7.1 6 1.3/mm2; MCTC, 13.86 1.86versus1.9 6 0.86/mm2). Specifically, the MCT phenotype was thepredominant MC type, whereas fewer MCTC were found in thestudy group. Furthermore, MC constituted the third most abun-dant interstitial inflammatory cell population in RPGN (Figure

Figure 1.Photomicrograph showing diffusely distributed tryptase-positive mast cells (MCT) in the renal interstitium. Magnification,3200.

1500 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 1498–1505, 1999

5). Figure 5 shows the distribution of both MC types in relationto the development of RPGN. The numbers of MC in differenttypes of RPGN were not significantly different and were lessthan the numbers of T cells.

We also analyzed the distribution of different MC subsetsaccording to the histologic stages of RPGN. No major differ-ence was observed in the number of MCT (41.86 6.5 versus45.8 6 6.7/mm2). However, the mean number of MCTC washigher (1.4-fold) in stage II than in stage I (11.46 1.9 versus16.2 6 3.1/mm2, P . 0.05), although the difference was notstatistically significant. In stage II, representing a progressiveform of RPGN, inflammatory cells (T cells, B cells, andmacrophages) were more abundant than in stage I, but thedifference was not significant.

Morphometric CorrelationWe also analyzed the correlation between the MC phenotype

and the pathologic findings. There was a significant correlationbetween the percentage of glomerular crescents and the num-ber of MCT (r 5 0.497,P , 0.001). Furthermore, there was asignificant correlation between the number of interstitial Tlymphocytes and the number of MCT (r 5 0.67,P , 0.0001),as well as that of MCTC (r 5 0.719,P , 0.0001). However,there was no relationship between the density of B cells and thenumbers of MCT and MCTC (r 5 0.358 and 0.238, respec-tively; P . 0.05). There was a significant correlation betweenthe number of interstitial CD681 macrophages and the num-bers of MCT (r 5 0.455,P , 0.001) and MCTC (r 5 0.646,P , 0.0001). A significant relationship was detected betweenthe relative interstitial volume (reflecting all changes associ-ated with interstitial expansion, such as edematous expansion,

focal fibrosis, and inflammatory cell accumulation) and thenumber of MCT (r 5 0.45,P , 0.0014).

A close relationship was observed between the fractionalarea ofa-SMA staining and the number of MCTC (r 5 0.749,P , 0.0001) (Figure 6), but this relationship was weaker withMCT (r , 0.525,P , 0.0001). The fractional area of fibrosiswas significantly higher in stage II RPGN than in the earlystage (stage I) of the disease (10.86 9.4versus23.76 14.9%,P 5 0.0006). In addition, there was a significant correlationbetween the fractional area of fibrosis and the number of MCTC

but not that of MCT (r 5 0.598,P , 0.0001, compared withr 5 0.269,P . 0.05).

Clinicopathologic CorrelationThe number of MC was not influenced by age or gender but

was significantly correlated with serum creatinine clearance,particularly for MCTC (r 5 0.661,P , 0.0001). Furthermore,significantly greater numbers of MCTC were detected in pa-tients with hypertension (.165/95 mmHg,n 5 16), includingMCTC (22.46 15.3versus9.7 6 9.8/mm2, P , 0.0001).

DiscussionThe major finding of this study was the significant accumu-

lation of both MC phenotypes in RPGN (sixfold), comparedwith control cases. Although MC have been demonstratedpreviously in renal diseases (6–11), to our knowledge this isthe first demonstration, using the double immunostainingmethod, of the simultaneous presence of both MCT and MCTC

phenotypes in RPGN. Accumulation of MC in the interstitiumwas significantly correlated with the loss of renal function, as

Figure 2. Photomicrograph showing both phenotypes of mast cells (MC). The MCT phenotype is indicated by rhodamine-labeled tryptaseparticles (red). The tryptase- and chymase-positive (MCTC) phenotype is shown as yellow, mixing the rhodamine-labeled tryptase (red) andFITC-labeled chymase (green). Note the absence of only chymase-positive MC. Double immunostaining; magnification,3200.

J Am Soc Nephrol 10: 1498–1505, 1999 Mast Cells in RPGN 1501

reported for other renal diseases (9,10), and was well correlatedwith TI damage.

In the early stages, RPGN is histologically characterized byextracapillary glomerular epithelial cell proliferation associ-ated with intra- and extraglomerular inflammation. However,with the progression of the disease process, histologic exami-nation usually shows fibrocellular and then fibrous transfor-mation of the glomerular crescents, together with TI damage,including interstitial fibrosis. Our data suggest the presence ofMC in both stages of renal lesions, but the MCTC phenotype ismore frequently (1.4-fold) present in the late stage than in theearly stage. Despite the report in 1960 of the detection of MCin “subacute GN” (7), the mechanism by which MC proliferatein the interstitium in RPGN is still not known.

Consistent with the results of recent studies (6–11), ourfindings showed that MC were localized in the renal intersti-tium but not in the glomerular tufts or spaces. Although ourresults showed a good correlation between the number of MCand the frequency of glomerular crescents, the distribution ofthese cells indicated that MC accumulation was correlatedmore strongly with TI injury than with glomerular lesions.

More importantly, we also demonstrated a strong correlationbetween the number of MCTC and TI damage of RPGN,including inflammation, myofibroblast proliferation, and fibro-sis.

The results depicted in Figure 5 showed no differences in theaccumulation of MC and that of other types of inflammatorycells in different types of RPGN. These results suggest thatinfiltration of MC is an integral part of the inflammatoryprocess, rather than a specific process in these renal diseases.In this regard, previous studies have convincingly demon-strated that the pathologic processes in these nephropathies aredependent to a large extent on inflammatory cells, especially Tcells and macrophages (21,22). Although no correlation wasobserved between the accumulation of MC and that of Blymphocytes in previous studies (9–11), histologic examina-tion and morphometric analysis in this study revealed a closerelationship between the accumulation of MC and that of Tlymphocytes. Previousin vivo and in vitro studies establishedthat MC maturation and activation is regulated by products ofT lymphocytes (e.g., interleukin-3 and interleukin-4) (23,24).Therefore, it is possible that infiltration of T cells in the

Figure 3.Photomicrograph showing peritubular (T) MC. An oval-shaped MC completely filled by electron-dense granules is in the interstitium(INT). The MC is partially surrounded by cytoplasmic processes (arrows) of a peritubular interstitial fibroblast (F). Electron microscopy;magnification,36000. Inset, granulated MCT. Immunostaining; magnification,31000.

1502 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 1498–1505, 1999

interstitium in RPGN might directly or indirectly increase MCingress from the circulation, as well as MC maturation andactivation. Furthermore, it has been demonstrated that tryptasecan also increase microvascular permeability (25), enhanceinflammatory cell transmigration (26), and stimulate matura-tion as well as cytokine release (5,27). Additional studies are

necessary to determine whether MC or lymphocytes can initi-ate the induction of interstitial inflammation and whether ac-tivated T cells within the diseased kidney synthesize moleculesthat participate in MC proliferation and maturation orviceversa. Although our analysis demonstrated a close structuraland statistical relationship between monocyte/macrophagesand MC, similar to previous data on human lung MC activationby macrophage inflammatory protein-1a (28), the functional

Figure 4. Two serial 3-mm sections of a rapidly progressive glomerulonephritis (RPGN) renal biopsy, immunostained for chymase (a) anda-smooth muscle actin (a-SMA) (b). Note that MCTC accumulation is colocalized with thea-SMA-positive myofibroblast proliferation area.Magnification,3400.

Figure 5.Distribution of renal interstitial inflammatory cells in casesof RPGN, according to etiologic differences. Mo/Ma, monocytes/macrophages.

Figure 6.Correlation between the number of MCTC and the fractionalarea ofa-SMA-positive cell proliferation (FAa-SMA).

J Am Soc Nephrol 10: 1498–1505, 1999 Mast Cells in RPGN 1503

relationship between these cells in renal diseases is still notclear.

In this study, we found colocalization of accumulated MCanda-SMA-positive myofibroblasts, indicating that MC maybe involved in the fibrotic process in RPGN. Previous studiesdescribed the presence of MC in fibrotic areas of variousorgans (1–3,29,30). For example, Liet al. (30) found thatcardiofibrosis was more severe when a large proportion of MCwere present in heart transplants. Furthermore, Lajoieet al. (9)demonstrated a strong relationship between the number ofMCT and interstitial edema, as well as fibrosis, in renal allo-grafts. However, in another study, despite the significant (10-fold) accumulation of MC, Colvinet al. (6) did not detect anycorrelation between interstitial damage and the accumulationof MC in chronic rejection in renal allografts stained withtoluidine blue. The discrepancy between these results might beattributable to different methods used for the identification ofMC. Despite thein vitro effects of MC on the development offibrosis (31,32), the role of MC in fibrogenesis in GN has notbeen demonstrated. In this analysis, there was a significantcorrelation between the accumulation of both phenotypes ofMC (particularly the MCTC phenotype) and the degree ofinterstitial fibrosis. The density of MCTC in the late stage ofRPGN was higher than that in the early stage. Whether selec-tive proliferation or activation of a MCTC subpopulation occursin RPGN remains to be elucidated.

Consistent with previous observations (10), we also found aclose structural relationship between MC and interstitial fibro-blasts (Figure 3, inset). The latter cells are known to take upMC granules (33), followed by phenotypic modulation of MCby cytokines (34). Furthermore, MC products (e.g., tryptase,chymase, and histamine) have mitogenic activity for fibro-blasts and enhance collagen synthesis (12–14,16,17,32,35).Recently, mitogenic basic fibroblast growth factor was de-tected in renal MC cytoplasm in biopsy specimens from pa-tients with IgA nephritis (10), suggesting the potential fibro-genic role of MC in IgA nephritis. On the other hand,in vitroand in vivo analyses demonstrated that MC products such asheparin and tumor necrosis factor-a influencea-SMA expres-sion (32,35). Rugeret al. (11) observed periglomerular accu-mulation of MC, type VIII collagens, anda-SMA-positivecells in diabetic nephropathy, suggesting that MC may upregu-late renal fibroblasts expressinga-SMA. These activated fi-broblasts (myofibroblasts) are well known as the cells respon-sible for extracellular matrix production (36,37) in renalfibrosis. Our results demonstrated a significant correlation be-tween the accumulation of both phenotypes of MC and thefractional area ofa-SMA staining, suggesting that the role ofMC is not only in the early inflammatory process but also inthe late fibrotic renal process in RPGN.

In summary, this study demonstrated that MC accumulationis a feature of a more aggressive form of RPGN and that suchaccumulation may be an important mechanism for amplifyingMC-mediated renal injury. Additional studies are necessary toclarify the exact role of MC in renal inflammatory and fibro-proliferative processes. Understanding the mechanisms for MCinterstitial inflammation and fibrosis may facilitate the design

of new therapeutic strategies for the treatment and preventionof GN progression.

AcknowledgmentsWe are grateful to Drs. Y. Kiso and N. Ishihara of the Suntory

Biomedical Research Center (Osaka, Japan) for providing anti-chy-mase antibody. We also thank Noriko Kawamoto and Masako Ish-iguro for skilled technical assistance and Dr. F. G. Issa (Word-Medex,Sydney, Australia) for careful reading and editing of the manuscript.

References1. Claman HN: Mast cells, T cells and abnormal fibrosis.Immunol

Today6: 192–195, 19852. Metcalfe DD, Costa JJ, Burd PR: Mast cells and basophils. In:

Inflammation: Basic Principles and Clinical Correlates, 2nd Ed.,edited by Gallin JI, Goldstein IM, Snyderman R, New York,Raven Press, 1992, pp 709–725

3. Ambrust T, Batusic D, Ringe B, Ramadori G: Mast cell distri-bution in human liver disease and experimental rat liver fibrosis:Indications for mast cell participation in development of liverfibrosis.J Hepatol26: 1042–1054, 1997

4. Irani AM, Schechter NM, Craig SS, DeBlois G, Schwartz LB:Two types of human mast cells that have distinct neutral proteasecompositions.Proc Natl Acad Sci USA83: 4464–4468, 1986

5. Church MK, Levi-Schaffer F: The human mast cell: Updates oncells and cytokines.J Allergy Clin Immunol99: 155–160, 1997

6. Colvin RB, Dvorak AM, Dvorak HF: Mast cells in the corticaltubular epithelium and interstitium in human renal disease.HumPathol 5: 315–326, 1974

7. Pavone-Macalusco M: Tissue mast cells in renal diseases.ActaPathol 50: 337–346, 1960

8. Harrison DJ, Neary C, Wathen CG: Distribution of IgE-bearingcells in renal biopsies in Wegener’s granuloma.Nephrol DialTransplant3: 24–27, 1988

9. Lajoie G, Nadasdy T, Laszik Z, Blick K, Silva FG: Mast cells inacute cellular rejection of human renal allografts.Mod Pathol9:1118–1125, 1996

10. Ehara T, Shigematsu H: Contribution of mast cells to the tubu-lointerstitial lesions in IgA nephritis.Kidney Int54: 1675–1683,1998

11. Ruger BM, Hasan Q, Greenhill NS, Savis PF, Dunbar PR, NealeTJ: Mast cells and type VIII collagen in human diabetic nephrop-athy.Diabetologia39: 1215–1222, 1996

12. Russel JD, Russel SB, Trupin KM: The effect of histamine on thegrowth of cultured fibroblasts isolated from normal and keloidtissue.J Cell Physiol93: 389–394, 1977

13. Norrby K: Effect of heparin, histamine and serotonin on thedensity-dependent inhibition of replication in two fibroblasticcell lines.Virchows Arch B Cell Pathol15: 75–93, 1973

14. Hatamochi A, Fujiwara K, Ueki H: Effects of histamine oncollagen synthesis by cultured fibroblasts derived from guineapig skin.Arch Dermatol Res277: 60–64, 1985

15. Folkman J, Kagsbrun M: Angiogenic factors.Science235: 442–447, 1987

16. Young JD-E, Liu CC, Butler G, Cohn ZA, Galli SJ: Identifica-tion, purification, and characterization of a mast cell-associatedcytolytic factor related to tumor necrosis factor.Proc Natl AcadSci USA84: 9175–9179, 1987

17. Sugarman BJ, Aggarwal BB, Hass PE, Figari IS, Palladino MAJr, Shepard HM: Recombinant human tumor necrosis factor-a:

1504 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 1498–1505, 1999

Effects on proliferation of normal and transformed cellsin vitro.Science230: 943–945, 1985

18. Couser WG: Rapidly progressive glomerulonephritis: Classifica-tion, pathogenetic mechanisms, and therapy.Am J Kidney Dis11: 449–464, 1988

19. Urata H, Kinoshita A, Misono KS, Bumpus FM, Hussain A:Identification of a highly specific chymase as the major angio-tensin II-forming enzyme in the human heart.J Biol Chem265:22348–22357, 1990

20. Toth T, Takebayashi S: Glomerular hypertrophy as a prognosticmarker in childhood IgA nephropathy.Nephron80: 285–291,1998

21. Hooke DH, Gee DC, Atkins RC: Leukocyte analysis usingmonoclonal antibodies in human glomerulonephritis.Kidney Int31: 964–972, 1987

22. Lan HY, Nikolic-Paterson DJ, Mu W, Atkinson RC: Localmacrophage proliferation in the progression of glomerular andtubulointerstitial injury in rat anti-GBM glomerulonephritis.Kid-ney Int48: 753–760, 1995

23. Nabel G, Galli SJ, Dvorak AM, Dvorak HF, Cantor H: InducerT lymphocytes synthesize a factor that stimulates proliferation ofcloned mast cells.Nature (Lond)291: 332–334, 1981

24. Bhattacharyya SP, Drucker I, Reshef T, Kirshenbaum AS, Met-calfe DD, Mekori YA: Activated T lymphocytes induce degran-ulation and cytokine production by human mast cells followingcell-to-cell contact.J Leukocyte Biol63: 337–341, 1998

25. He S, Walls AF: Mast cell tryptase: A stimulus of microvascularleakage and mast cell activation.Eur J Pharmacol328: 89–97,1997

26. He S, Peng Q, Walls AF: Potent induction of neutrophil andeosinophil-rich infiltratein vivo by human mast cell tryptase:Selective enhancement of eosinophil recruitment by histamine.J Immunol159: 6215–6225, 1997

27. Cairns JA, Walls AF: Mast cell tryptase is a mitogen for epithe-lial cells: Stimulation of IL-8 production and intercellular adhe-sion molecule-1 expression.J Immunol56: 275–283, 1996

28. Baghestanian M, Hofbauer R, Kiener HP, Bankl HC, Wimazal F,Willheim M, Scheiner O, Fureder W, Muller MR, Bevec D,Lechner K, Valent P: The c-kit ligand stem cell factor andanti-IgE promote expression of monocyte chemoattractant pro-tein-1 in human lung mast cells.Blood 90: 4438–4449, 1997

29. Inoue Y, King TE Jr, Tinkle SS, Dockstader K, Newman LS:Human mast cell basic fibroblast growth factor in pulmonaryfibrotic disorders.Am J Pathol149: 2037–2054, 1996

30. Li Q-Y, Raza-Ahmad A, MacAulay MA, Lalonde LD, RowdenG, Trethewey E, Dean SH: The relationship of mast cells andtheir secreted products to the volume of fibrosis in posttransplanthearts.Transplantation53: 1047–1051, 1992

31. Rubinchik E, Levi-Schaffer F: Mast cells and fibroblasts: Twointeracting cells.Int J Clin Lab Res24: 139–142, 1994

32. Cairns JA, Walls AF: Mast cell stimulates the synthesis of typeI collagen in human lung fibroblasts.J Clin Invest99: 1313–1321, 1997

33. Subba Rao PV, Friedman MM, Atkins FM, Metcalfe DD: Phago-cytosis of mast cell granules by cultured fibroblasts.J Immunol130: 341–349, 1983

34. Davidson S, Mansour A, Gallily R, Smolarski M, Rofolovitch M,Ginsburg H: Mast cell differentiation depends on T cells andgranule synthesis on fibroblasts.Immunology48: 439–452, 1983

35. Desmouliere A, Rubbia-Brandt L, Grau G, Baggiani G: Heparininducesa-smooth muscle actin expression in cultured fibroblastsand in granulation tissue myofibroblasts.Lab Invest67: 716–726, 1992

36. Alpers CE, Hudkins KL, Floege J, Hohnson RJ: Human renalcortical interstitial cells with some features of smooth musclecells participate in tubulointerstitial and crescentic glomerularinjury. J Am Soc Nephrol5: 201–210, 1994

37. Goumenos D, Tsomi K, Iatrou C, Oldroyd S, Sungur A, Papa-ioannides D, Moustakas G, Ziroyannis P, Mountokalakis T, ElNahas AM: Myofibroblasts and the progression of crescenticglomerulonephritis.Nephrol Dial Transplant13: 1652–1661,1998

This article can be accessed in its entirety on the Internet athttp://www.lww.com/JASN along with related UpToDate topics.

J Am Soc Nephrol 10: 1498–1505, 1999 Mast Cells in RPGN 1505