Distribution of basement membrane antigens in glomeruli of mice with autoimmune glomerulonephritis.

(1)

Distribution

of

Basement

Membrane

Antigens

in

Glomeruli

of

Mice

With Autoimmune

Glomerulonephritis

SEIICHI MAINUO, MD, JAN R. BRENTJENS, MD, GIUSEPPE ANDRES, MD,

JEAN-MICHELLE FOIDART, MD,

GEORGE R. MARTIN, PhD, and

ANTONIO MARTINEZ-HERNANDEZ, MD

Glomerulonephritiswasinducedinmice by the repeated

injection ofhuman glomeruliorpurifiedglomerular base-mentmembrane. The glomerularbasementmembranes ofnephritic animalswere observed to develop subepi-thelial extensions, "spikes?' Although normallyType IV

collagen isfoundthroughout the fullthickness of

base-THEBIOCHEMICAL composition of the

glomeru-larbasement membrane (GBM) has been studied

in-tensively, and several components havebeenidentified, including laminin, Type IVcollagen, heparan sulfate

proteoglycan,entactin, nidogen andamyloidP.`-9 The

distribution of basement membrane components has

been established in normal glomeruli,"2,4,6'7,9"10-14 in

several types of human'522 and of experimentally

induced823glomerular diseasesusingspecific antibod-ies. Thesestudies, combined with data obtainedfrom

glomerularcellsgrown in vitro, provideafoundation forexploringtherole ofindividual GBM components in pathologic conditions.

Patients with Stage 2and 3 membranous

glomeru-lonephritisshowas acharacteristicabnormalityofthe

glomerular capillary wall, the presence of basement membrane "spikes" between

subepithelial

immune

deposits.24The

pathogenesis

ofthisdisease is

usually

ascribedto

antigen-antibody complexes

formed in situ or in thecirculation.25 Recently, however,Bolton and

co-workers26 observed that mice immunized with

hu-man GBMpreparationsdeveloped

glomerular

lesions characterized by diffuse and

unusually large spikes

not

separatedbyimmune

deposits.

This

finding

is consis-tentwith thenotionthatthis

glomerular

disease is due

totheinteraction ofcirculatingantibodieswith

"struc-turalantigensoftheGBM."26Thepurpose ofour

study

was to establish the distribution oftwo well-defined

From theDepartmentsof Microbiology, Pathology, andMedicine,

State UniversityofNew YorkatBuffalo, Buffalo, New York;

the Laboratory ofDevelopmental Biology and Anomalies,

NationalInstituteofDental Research, National Institutesof Health,

Bethesda, Maryland; and theDepartmentof Pathology,

Hahnemann University, Philadelphia, Pennsylvania

mentmembranes, the"spikes" reacted withantibodies

tolaminin butnotwith antibodiesto Type IVcollagen.

Itisproposed thatinmurineautoimmune glomerulone-phritis, the visceral epithelialcellsproduceanexcess of

laminin. (AmJ Pathol 1986, 122:36-49)

basement membrane components,TypeIVcollagenand

laminin,intheglomeruliofmicewiththis form of

anti-GBMantibody-mediated glomerulonephritis. The

re-sultsindicate thatlan-tinin, butnotType IVcollagen, is amajor component ofthe spikes.

Materials and Methods

Animals

Female Swiss-ICR mice (20-25 g) were purchased

fromWestSenecaLaboratory (West Seneca, NY)and allowed free access to food and water.

Antigen Preparation

Normal humankidneyswereobtained from

autop-siesperformedwithin 12hoursonaccident victims and storedat -60 Cuntil used. Glomeruliwereisolated

ac-cordingtothemethod of Krakower and

Greenspon.27

Afterisolation,glomeruliweredialyzed againstdistilled

Supportedin part byNIH GrantsAI-10334, AM-25254,

andAM-28488.

Accepted for publication July31, 1985.

AddressreprintrequeststoG.Andres,Pathology

Depart-ment, 208 Farber Hall, State University of New York at

Buffalo, Buffalo, NY 14214.

(2)

BASEMENT MEMBRANE ANTIGENS 37

water andlyophilized (glomerularfraction).Partofthis

fraction was further processed for separation of the

GBM from cellular components, with amodification

ofthe detergent method describedbyMeezan etal.28 Glomeruliweretreated with407odeoxycholate

contain-ing0.lIlo sodium azide for 3.5 hours at room tempera-ture with constant stirring. After centrifugation at

10,000 rpm for 10 minutes, the supernatant was dia-lyzed against distilled water andlyophilized (cellular fraction). The sediment waswashed three times with

distilled water andlyophilized (GBMfraction).The pro-teincontentof each fraction (weightpercent)was

de-termined by the modification ofLowry's method.29

Immunization

Sixteen mice wereimmunized with the glomerular

fraction, 5micewith theGBMfraction, and5mice with thecellularfraction. The initial immunizationconsisted

of subcutaneous injection at multiple sites ofatotal of0.75mg ofprotein,in0.075 mlphosphate-buffered

saline(PBS) mixed with 0.075 ml of Freund'scomplete

adjuvant (FCA), (DifcoLaboratories, Detroit, Mich).

Furtherimmunizations were carried out by subcutane-ous injections of the same antigenic mixture every 2 weeks until themice died or were sacrificed (24 weeks afterimmunization). Sevencontrolmiceweregiven

in-jections, following the schedule described above, ofa

mixture of 0.075 ml of PBS and FCA. Antibodies

Forimmunofluorescence microscopy, rabbit antibod-iesdirected against mouse Type IVcollagen, laminin andfibronectinwereused.Forimmunoelectron micros-copy, the Fabfragment of monospecific rabbit antibod-iesdirected against mouse Type IV collagen or fibronec-tinoragainstratlaminin were used. Thepurification, characterization, and specificity of these antibodies has been described previously.

3.4'612'30'31

Fluorescein

isothiocyanate (FITC)-conjugated goat antibodies to rabbit IgG, mouse IgG, or mouse IgM and mouse C3

werepurchased from Cappel Laboratories

(Cochran-ville, Pa).FITC-conjugatedgoatanti-rabbit IgG did not

react with mouse IgG when tested by direct im-munofluorescence on kidney sections of mice with nephritis. Goat anti-rabbit Fab and rabbit Fab-PAP were prepared as previously described.12,32.33

Urinary Protein Excretion and Blood Sampling

Mice from each group were placed every 2 weeks in metabolism cages, and 24-hour urine specimens were collected.Theprotein excretion wasdetermined by the

biuret method. Blood samples were taken from the

retroorbitalplexus, beforeimmunization, 8 weeks

af-terinitial immunization, andatthe time of sacrifice. Thesera weretested forthepresenceofbasement

mem-braneantibodiesbyindirectimmunofluorescence

mi-croscopy with frozen sections of normal Swiss-ICR

mouse kidneyas substratum.

Elution ofImmunoglobulins From Kidney Tissue

Elution was carried out by treatment of renal

homogenates with citrate bufferatpH 3.2orwith

gly-cine buffer at pH 2.2, as previously described.34

Light and ElectronMicroscopy

Forlight microscopy, kidney tissuewasfixed in 10Wo

buffered formalinovernight and embedded inparaffin.

Sections were stained with hematoxylin and eosin

(H&E) or with periodic acid-Schiff (PAS) reagent.

Some blockswereembeddedinmethacrylate, and l-,u sections were reacted with H&E and Jones' silver methanamine. For electronmicroscopysmallpieces of

cortex werefixed inamixture ofparaformaldehyde and glutaraldehyde,35 postfixed in1%Wo osmium tetroxide, and

embedded in Epon 812-Araldite. Thin sections were

stained with uranylacetateand lead citrate. For silver staining, tissue blocks were fixed in osmium

tetrox-ide andembeddedinEpon 812-Araldite.Thinsections wereoxidized byperiodic acid, reacted withamixture ofsilvernitrate,hexamethylene tetramine, andsodium borate, apd fixed with sodium thiosulfate.36

Immunofluorescence Microscopy

Sections of frozenorparaffin-embedded kidney

tis-sue were used for the study. Frozen sections were

processed as described in previous publications.37

Paraffin sections were deparaffinized with xylene and ethanol, washedin distilled waterfor severalminutes, incubated in 0.10%0 trypsin (Cat. No.610-5095, GIBCO

Laboratories, Grand Island, NY) and 0.05% CaCl2 in

PBS at 37 C for 15, 30, 60, 120, 180, and 240 minutes. Afterwashing in distilled water and in PBS the sections were stained for mouse IgG, Type IV collagen, lami-nin, and fibronectin. All the sections were studied with

aLeitz Ortholuxmicroscope equipped with

epifluores-cence optics and appropriate filters. Immunoelectron Microscopy

Thekidneys of anesthesized animals were perfused with 4% formaldehyde via the abdominal aorta for 10 minutes accordingtopreviously describedmethods.46'12

(3)

(4)

Figure 1-Glomeruli from mice immunizedwith theglomerular fraction. Arrowspointtosomeargyrophilic spikes. a-Methacrylate-embedded sec-tionreactedwithJones'silver methanamine.TheGBM ofeverycapillaryloopcontainsmultiplespikes. (x 750) b-Epon-embeddedsectionreacted

with silver.Thespikes arelocatedattheepithelial side of the GBM.(x3000) c-Standard electronmicrograph of theglomerularcapillarywall.Note the absence ofsubepithelialforeignelectron-densedepositsbetweenthespikes. (x 12,000) Abbreviations used in allfigures:B,Bowman'scapsule;

Cl, capillary lumen;GBM, glomerular basementmembrane; En,endothelialcell; Mc, mesangialcell; Mm, mesangialmatrix; N,nucleus; Pe,parietal

epithelialcell;RER,roughendoplasmic reticulum; Tc,tubularcell; TBM,tubular basementmembrane;Us, urinaryspace;Ve,visceralepithelialcell.

Afterperfusion, thekidneyswereremoved, minced into

1 x 1 x 0.5cmblocks, andpostfixed in 407 formalde-hyde at4Cwith continuousstirring for3hours. After fixation, the blocks were rinsedinmultiple changes of PBS with 4% sucrose at4 Covernight. After afinal 1-hour wash in PBS,with 4%sucroseand7%glycerol, the tissuesweresnap-frozen inmethylbutaneprecooled

toliquid nitrogen temperature. Six-micron sectionswere cutinacryostat and allowedtodryonalbumin-coated slides.Sectionsweresequentially reacted with sodium

borohydride,normal goat serum, normal rabbit serum,

orprimary antibody(Fab). The localizationwas

com-pleted with goat anti-rabbit Fab, Fab-PAP, 2.5%7o

glutaraldehyde, glycine, and diaminobenzidine, accord-ing to previously described methods.12.33Sections were reacted withlo OS04, dehydrated in ethanol, and em-bedded inplastic via propylene oxide. Ultrathin sections

were examined and photographed without further

staining.

Results

Proteinuria

About20weeks afterimmunization the mice receiv-ing either glomerular or GBM fractions developed significant proteinuria (>20 mg/24 hours). Two of the

5mice immunized with cellular fraction and one of the

7 control mice had transient proteinuria during the

course of the experiment.

Circulating Antibodies

Sera frommice immunized with the glomerularor

GBMfraction bound inalinear patternto mouse

glo-merular and tubular basement membrane. Somesera

ofmice treated with theglomerular fraction alsoreacted

with the brush border of proximal tubules. The

anti-body titers were low (1:1-1:4). Sera from mice

im-munized with the cellular fraction reacted with the

brush border ofmouse proximal tubules and,

possi-bly, with some cellular components of mouse glomer-ularcapillary wall. Sera from mice given injectionsof

PBS and FCAalone did not react with mouse kidney.

Histologic Findings

Inimmunized mice the most characteristic light- and electron-microscopic lesion of the glomerular capillary wall was thediffuse presence of GBM spikes toward the epithelium (Figure 1). This change was more

pro-nounced in mice immunized with the glomerular or

GBMfraction thaninthose immunized with the

cellu-larfraction. In addition, fibrocellular crescents were ob-servedin afewof themice giventheglonierular or GBM

(5)

Figure2-Immunofluorescence microscopyofglomerulifromanimalsgiventheglomerular fraction,24weeks after immunization. a-MouseIgG.

Lineardeposits ofIgGaredemonstrablealongtheGBM;thespikescontain IgGdeposits. (x 1600) b-Laminin. Laminin is present inBowman's capsule,mesangial matrix,andglomerularand tubular basement membrane. Inaddition,laminin isprominentinallspikes. (x 1000) c-TypeIV

(6)

collagen.Type IVcollagenisdemonstrableinBowman's capsule, mesangialmatrix, andglomerularandtubularbasement membrane.Notice thatthe

spikes do not containdemonstrabletype IVcollagen(comparewithb, laminin). (x1000) d-Fibronectin. Fibronectinispresentinmesangialmatrix. The GBM andthespikes have onlyoccasional,irregular, weakreactivity.(x 1000)

(7)

42 MATSUO ET AL

%?Us

Ve

Us

.4x

v0.

...s

Cl

v

'.,

.

3b Figure 3-Electronimmunohistochemicallocalization ofTypeIVcollageninglomeruliof animalsgiventheglomerular fraction,17weeks after

immuni-zation. a-TypeIVcollagenisdemonstrableinmesangialmatrix andalongtheGBM,inalinearpattern.Thespikes (arrows)arenegative (compare

with Figure 4a).(x 8000) b-Detail ofcapillary loop. TypeIVcollagenispresent inthe lamina rara interna andlaminadensaandonly weaklyin the lamina raraexterna.Thespikes(arrowheads)lackingdemonstrableTypeIVcollagenaresurroundedbyepithelial cytoplasm. (x 22,000)

(8)

Vol.122 * No. 1

Figure 4-Electron immunohistochemicallocalization of Type IV collagen inglomeruli ofmiceimmunizedwith theglomerular fraction, 17 weeks after immunization. a-The tubularbasement membrane andBowman'scapsule arehomogeneously positive. In the GBM only the lamina rara interna and densa react withtheantibody.Aparietal epithelial cell containsTypeIV collagen antigens in the cytoplasm. (x 8500) b-Detail of aparietal epithelialcell. The cisternaeof theroughendoplasmicreticulum containType IVcollagenantigens.(x17,000)

(9)

44 MATSUO ET AL

fraction. The kidneys of mice given- FCA alone were consistently normal.

Immunofluorescence Findings

The results obtained with sections of frozen or

paraffin-embedded kidney tissue were comparable, al-thoughsuperior resolution was obtained with paraffin sections. The glomeruli of mice given FCA alone had

minimal deposits of mouse IgM in the mesangium.

Laminin and TypeIVcollagenwereseenin a linear pat-terninthe GBM.Fibronectin was present in the mesan-gium and only in trace amounts in the GBM.

Mice immunized with theglomerular or GBM

frac-tion had diffuse deposits of mouse IgG in the GBM,

in the spikes (Figure 2a), and in the mesangium.

Deposits of mouse IgM were restricted to the mesan-gium. Deposits of mouse C3 were seen only in a few mice and had a distribution similar to that of IgG. All themice hadabundant laminin in the GBM and in the spikes, withapatternsimilartothat of mouse IgG

(Fig-ure2b). This localization of laminin was incontrast tothat ofcollagen Type IV, whichwasdetectable, with

a sharp linear distribution, in the GBM and in the

mesangium, butnotin thespikes (Figure 2c). Fibronec-tinwaspresentin themesangium,andonlytracescould befound in theGBM(Figure 2d).Theglomeruliof mice immunized with thecellular fraction had lineardeposits

ofmouseIgG in the GBM.Laminin, collagen Type IV,

IgM, and C3 fibronectin hadadistribution similarto

that ofcontrol mice.

Antibodies Eluted From Nephritic Kidneys

Theeluates, containing IgG,weretestedbyindirect

immunofluorescence microscopy, using Swiss-ICR

mousekidneyassubstratum. TheIgGreacted inasharp

linear pattern with glomerularandtubular basement membrane.

Immunoelectron Microscopy

TypeIVcollagenandlaminin had different

localiza-tions in mice immunized with the glomerular or the

GBMfraction. Antibodies

against

TypeIV

collagen

had adiscretelocalizationintheGBM. Thereaction

prod-uct was prominent in the laminarara interna and in

thelamina densa and less intense or absent from the

laminarara externa(Figures3and4). The

spikes

were almostinvariably

negative (Figures

3 and4).The

mesan-gial matrixwas always

homogeneously positive

(Fig-ure3).Theparietalcells ofBowman's

capsule

reacted with the antibody(Figure 4a), and it was obvious at highermagnificationthatthereaction

product

was

pres-ent in the cisternae of the rough endoplasmic reticu-lum (Figure 4b). The staining of tubular basement mem-brane (Figure 4a) and Bowman's capsule (Figure 4) was identical to that seen in control animals.

Antibodies against laminin had a localization radi-cally different from that of antibodies to Type IV col-lagen. The GBM reacted with anti-laminin antibodies in a diffuse, homogeneous manner (Figure 5). The spikes (negative withanti-Type IV collagen antibodies) were invariably positive with anti-laminin antibodies (Fig-ure 5).The mesangial matrix was homogeneously posi-tive (Figure 5a). In addition, epithelial cells of Bow-man's parietal layer frequently had demonstrable laminin antigens in the cisternae of the rough endoplas-mic reticulum (data not shown) in a pattern similar to that observed for collagen Type IV. Tubular basement membrane stained in a pattern identical to that seen in control animals.

Antibodies against fibronectin had a localization similar to that in normal animals. Fibronectin was found in the mesangial matrix, and only traces were presentin the GBM (Figure 6). The spikes were consis-tently negative (Figure 6).

Antibodies eluted from diseased kidneys reacted with

basementmembranes of normal mouse kidney. In the

GBMthelocalization was preferentially subendothelial, inthe lamina rara interna and lamina densa, whereas the laminararaexterna wasconsistently negative

(Fig-ure7a). Intubular basement membranes the eluted an-tibodies bound in a homogeneous manner, with a some-what preferential localization in the lamina densa (Figure 7b).

Discussion

The glomerular basement membrane is a complex

structurecomposed ofavarietyof macromolecules. The presence of

laminin,6'7

TypeIVcollagen,89heparan

sul-fateproteoglycan,2 8and entactin6inbasement

mem-branes has beenconvincinglydemonstrated.Nidogen38

and amyloid P

component1

have also beenproposed

asbasement membrane components. The fibronectin

tracesoften found in GBM andoriginallydescribedas

anotherbasement membranecomponent10'39 probably

representplasmafibronectin"trapped"in the

glomer-ularfilter.'2'32'40'4' TypeIVcollagen, forminganopen network, is assumedtobe themajorstructural element of the basementmembrane.Theheparansulfatechains of theproteoglycanarearrangedalongthesurface and

create anionic barriertothepassage ofplasma

macro-molecules. Laminin isalargeglycoproteinthat hasbeen proposedtobindtoTypeIVcollagenandheparan sul-fate proteoglycan and also to cell surface receptors. Nidogen and entactin, possibly related proteins, are

(10)

45

Vol. 122 * No. 1 BASEMENT MEMBRANE ANTIGENS

ak IV.

i't~~~~~~~:4

Ve

J

0.:.

Us'

V_

GBM

et-Pr I

Ci

5b

Figure5-Electronimmunohistochemicallocalizationoflaminin inglomeruliof mice immunized with theglomerularfraction,17weeksafter

immuniza-tion. a-Lamininantigensarepresentinmesangial matrix,capillaryloops,basementmembrane,and in allspikes(arrows)(comparewith3a).The breaks in theglomerularcapillarywall(asterisk)are anartifact offreezing.(x5000) b-Detail ofacapillary loop.The full thicknessofthe GBMreacts

with theantibody.Thespikesareintenselypositiveforlaminin,but notashomogeneouslyastheGBM. (x 15,000) 4

(11)

46 MATSUO ET AL _AJP

*January 1986

r

~

Figure 6-Electronimmunohistochemicallocalization of fibronectin in glomeruli of mice immunized with glomerular fraction, 17 weeks after immuniza-tion.Some fibronectin ispresent inthe mesangialmatrix.TheGBM and thespikes (arrowheads) are negative. (x15,000)

found in basement membranes but arerelatively

mi-nor constituents.

Changesinsomebasement membranecomponents

have beenreportedinhumanandexperimental diseases.

Forinstance, inearlyand moderate diabetic

nephrop-athy,along withvisiblythickened basementmembranes,

anincrease in immunofluorescence reactivityofType

IVcollagen,laminin,and fibronectin inGBM,

mesan-gium, and tubular basement membrane has been

reported.'6

Bycontrast,the immunofluorescence

reac-tivityof thesecomponents inthe kidneys ofpatients

with late and severe diabetic nephropathy was

de-creased.'6 Another study reportsthat fibronectin,

al-though absent from normalglomeruli, waspresent in

theglomerular capillaryloops ofsome patients with mesangiocapillary, diffuseproliferative, and membra-nousglomerulonephritis.42 An increased

immunoflu-orescencereactivityoflaminin,TypeIV

collagen,

and fibronectinhas beenreportedinglomeruli ofpatients

withpreeclampsiaandhypertensive syndromes of

preg-nancy.'I TypeIVcollagenand lamininwerepresent in themesangium ofpatients with focal

glomeruloscle-rosis.'8The glomerulardistribution of Goodpasture's

antigen,2' of the antigenrecognized in human kidneys

by rabbit anti-human GBM sera,20and of amyloid-P

component'"wasaltered in severaltypesof human

glo-merulonephritis. Heparan sulfate proteoglycan was

decreased in glomeruli of patients with congenital

nephroticsyndrome,22 in diabetes mellitus,19 inratswith aminonucleoside induced nephroticsyndrome,8 and in

the extracellular matrix secreted by the EHS tumor

grown indiabeticmice.43Inpatients with advanced

idi-opathicmembranousglomerulonephritis, the

glomer-ularbasementmembranespikes havebeenreportedto

containfibronectin, laminin, and Type IV collagen.'8

These studies shouldbeinterpreted with caution. Itis difficulttoknowwhichchangesrepresentthecause,and which theconsequenceofthedisease. Forinstance, it islikely that increases in

fibronectin'1742

maysimply be

the consequenceofaleaky glomerular filter.Decrease

ofonecomponent mayresultinincreased compensa-torydeposition of others. Suchamechanism hasbeen

proposedtoexplainthedecrease of heparan sulfate

pro-teoglycan seen in diabetic mice and the subsequent

(12)

CI

Us

7a

Figure 7-Electron immunohistochemical localizE

neyofantibodieseluted fromkidneysof mice imn larfraction. a-Inthe GBM the eluted antib pattern.They localizepreferentiallyinthelamin;

densa. Thelaminararaexterna isconsistently

bodies.(x10,000) b-Inthe tubulesthe elu allbasement membranes. Thereactivityismore

sa.(x 7500)

thickening of the basement membi

the mechanisms involved inbasem

sembly areunknown. Complexme

ist to regulate the assembly of at molecules into the structure that

membrane; however, the exact arm

ecules asdetermined bytheirrelatiP

constants, and covalent cross-lini known.5 Interpretationof the chan, ishampered bythelack ofthisesse

It is worthnoting that in oursti

obviousmorphologic abnormality in

dothelialcells, whereasthevisceral dergo significantchanges,ofwhicht

istheeffacement offootprocesses.

changemaybeindicative ofanalteredfunction ofthe podocytes. Many of the findings in this studycan be interpreted in this light. In the normal murineGBM,

Type IV collagen is foundthroughoutthe full

thick-nessof theGBM, withapreferentialdistribution inthe laminadensa.8'10 In themice with autoimmune

glomer-ulonephritis, Type IV collagen could not be

demon-strated in the laminarara externa. Thisredistribution indicates thatduring thecourseof thedisease, either

TypeIVcollagenis selectivelyremoved from the

lam-inararaexterna

(but

notfromthe laminararainterna and densa), or, more likely, it is not being deposited

inthe lamina rara externa. We know that the mature GBMis the result of thefusion,atthelevel of the

lam-..;

, * minadensa,of two basement membranes:theepithelial

andendothelial basement

membranes.5

Itwould

ap-pearthat the mostlikely explanation for the change inTypeIVcollagen distribution isthat the visceral epi-thelial cells, affected bythe immune-mediated

glomer-ularinjury,decrease theirsyntheticrateofTypeIV

col-lagen, the resultbeing analmostcompleteabsence of

TypeIV collagen in thelamina rara externa and inthe spikes. Ifthis is thecase, it appearsthat the visceral

epithelial cells respond tothe injury, notonlywith a

decreasedsecretion ofTypeIVcollagen,but also with anincreased secretionof laminin. The result of this

in-creased secretion oflamininis the formation ofspikes. Obviouslywe cannotruleouttheredistributionof other

At:9]_

s basement membrane components, such as entactin,

nidogen,orheparan sulfateproteoglycan.Inaddition,

thepossiblerelationshipbetweenchangesin structural ation in normalmousekid-

composition

ofthe GBM and

proteinuria

wasnot

ad-nunized with theglomeru- dressed in the present study.

odies localize inalinear In

nephritic

miceTypeIV

collagen

and lamininwere

a rarainternaandlamina

negativewith these anti- observed within the roughendoplasmic reticulum of ited antibodies localizein parietal

epithelial

cells

(Figure 4).

Inthenormaladult

itneinthe lamina den-(Fgr4)Inteomaadl

murineglomerulus,laminin and TypeIVcollagenhave notbeenlocalized

intracellularly,'79

afact that

proba-blyreflects the lowsynthetic rate of these molecules.

rane.43 At present, Inautoimmuneglomerulonephritis in mice, the

parie-ent membrane as- talepithelial cells must receive some signal that trig-chanismsmustex- gersan increasedsynthesis and secretion of laminin and least five macro- TypeIV collagen. Thesynthetic rate must be increased

we call basement to such an extent that the antigensbecome detectable

angement of mol- byimmunohistochemistry. The fate of these basement

veaffinity, binding membranecomponents is not clear: Bowman's capsule

king remains un- does not appear thickened;

therefore,

these antigens gesseenindisease mustbe either lost in theglomerular filtrate or

incor-ntial information. porated into the GBM.

adies we found no Inthe present study, mice wereimmunized not only

theglomerularen- with whole glomeruli or isolated GBM but also with

epithelialcellsun- acell-derived glomerular preparation. This was done hemostprominent totest thehypothesis that the pathogenic

antigen-anti-Thismorphologic bodyinteraction in murine autoimmune

(13)

48 MATSUO ET AL AJP * January1986

phritis occurs at the level of the plasmamembrane of

the epithelial and/or endothelial cells, cells that

pro-duce the GBM. Our failure to induce significant dis-ease inmice immunizedwith cellular fraction does not

completely ruleoutthispossibility, because basement

membrane antigens expressed on thesurface of

glomer-ular capillary cells may have been lost during the

preparative process.

Itwould be important to know, of all the basement

membrane antigens present in human GBM, which

one(s) elicits an antibody response, and which one(s) is responsible for the immunologic injury. No

defini-tive answeris available, but it is clear that the ultra-structurallocalization ofthe IgG eluted from the

dis-eased mouse kidneys is different from that of antibodies directed against laminin, heparan sulfate proteoglycan, andTypeIVcollagen. The eluted antibodieslocalized

in normal mouse GBM in the lamina rara interna and to some extent in the lamina densa. Type IV collagen

is present in the three laminae, but preferentially in the lamina densa.8'9 Laminin is also found in the three layers, but preferentially in thelaminae rarae.' Hepa-ransulfate proteoglycan is predominantly localized in discrete clusters in both laminaerarae.28The localiza-tion ofthe eluted antibodies closely resembles that of

entactin. Thisglycoproteinis foundpreferentiallyin the

lamina rara interna and lamina densa.6 The localiza-tion of theeluted antibodies in tubular basement mem-brane is also similar to that of entactin.6 Obviously, similar ultrastructural localization isnotdefinitive evi-dence, butit seems toruleoutthree components and

suggests that entactin, or other GBM components

with alocalizationsimilar to that of entactin, may play

a crucial role in this murine model of anti-GBM

antibody-mediated glomerulonephritis.

References

1. Dyck RF, LockwoodCM,KershawM,McHugh N, Du-anceVC, BaltzML,PepysMB:AmyloidP-component

isaconstituent of normal humanglomerularbasement membrane. J Exp Med 1980, 152:1162

2. KanwarYS,FarquharMG:Presenceofheparansulfate

in theglomerularbasement membrane.ProcNatl Acad Sci USA 1979,76:1303

3. Laurie GW, LeblondCP, InoueS, MartinGR,Chung

A:Finestructureof theglomerularbasementmembrane and immunolocalization of five basement membrane

componentstothelamina densa(basal lamina)and its extensions in bothglomeruliandtubules of therat

kid-ney. Am J Anat 1984, 169:463

4. MadriJA, Roll FJ,Furthmayr H,Foidart JM. Ultrastruc-turallocalizationof fibronectin and lamininin basement

membrane of the murinekidney.JCell Biol1980,86:682 5. Martinez-Hernandez A,AmentaPS:The basement

mem-branein pathology. Lab Invest 1983,48:656

6. Martinez-HernandezA,ChungA:Theultrastructural lo-calization oftwo basement membrane components

-entactin and laminin-in rat tissues. J Histochem Cytochem 1984, 32:289

7. Martinez-Hernandez A, Miller EJ, Damjanov I, Gay S: Laminin secreting yolk sac carcinoma of the rat: Biochem-ical and electronimmunohistochemical studies. Lab In-vest 1982, 47:247

8. Mynderse LA, Hassele J, Kleinman HK, Martin GR, Martinez-Hernandez A: Loss of heparan sulfate pro-teoglycan from glomerular basement membrane of nephrotic rats. Lab Invest 1983, 48:292

9. RollFJ, Madri JA, Albert J, Furthmayr H: Codistribu-tionofcollagens type IV and AB2 in basement membrane andmesangium of the kidney-An immunoferritin study ofultrathin frozen sections. J Cell Biol 1980, 85:597 10. Courtoy PJ, Timpl R, Farquhar MG:Comparative

dis-tribution of laminin, type IV collagen and fibronectin in theratglomerulus. J Histochem Cytochem 1982, 30:874 11. Fish AJ, Carmody KM, Michael AF: Spatial orientation anddistribution of antigens within human glomerular basement membrane. J Lab Clin Med 1979, 94:447 12. Martinez-Hernandez A, Marsh CA, Clark CC, Macarak

EJ,Brownell AG: Fibronectin: Its relationship to base-mentmembranes. II. Ultrastructural studies in rat kid-ney. Collagen Rel Res 1981, 5:405

13. Scheinman JI,Foidart JM, Michael AF: The immuno-histology of glomerular antigens: V. Thecollagenous an-tigens of the glomerulus. Lab Invest 1980, 43:3759 14. Scheinman JI, Foidart JM, Robey PG, Fish AJ, Michael

AF: Theimmunohistology of glomerular antigens: IV. Laminin, a defined non-collagenous basement membrane glycoprotein. Clin Immunol Immunopathol 1980, 15:175 15. DyckRF, Evans DJ, Lockwood CM, Rees AJ,TurnerD,

Pepys MB:Amyloid P-component in human glomerular

basementmembrane-abnormal patterns of

immunofluo-rescentstaininginglomerulardiseases. Lancet 1980, 2:607 16. Falk RJ, Scheinman JI, Mauer SM, Michael AF: Poly-antigenic expansion of basement membrane constituents indiabetic nephropathy. Diabetes 1983,32:(Suppl2):82 17. Foidart JM, Nochy D, Nusgens B,Foidart JB, Mahieu PR,Lapiere CM, Lambotte R,Bariety J: Accumulation ofseveral basement membraneproteinsinglomeruliof

patientswithpreeclampsiaand otherhypertensive syn-dromes of pregnancy: Possible role of renal

prostaglan-dins and fibronectin. Lab Invest 1983, 49:250

18. Killen PD, Melcion C, Bonadio JF, Morel-Maroger L, Striker GE: Glomerular responsetoimmunological

in-jury. Studiesonprogression. SpringerSemin

Immuno-pathol 1982, 5:297

19. Parthasarthy N, Spiro RG: Effect of diabetes on the glycosaminoglycan component of human glomerular

basement membrane. Diabetes 1982, 31:738

20. SchifferMS, Michael AF, KimY, FishAJ:Distribution ofglomerularbasement membraneantigensin diseased human kidneys. Lab Invest 1981, 44:234

21. Sisson S, DysartNK, Jr, FishAJ,VernierRL:

Localiza-tion of theGoodpastureantigen byimmunoelectron

mi-croscopy. Clin ImmunolImmunopathol 1982, 23:414

22. Vernier RL,KleinDJ,SissonSP,MahanJD,OegemaTR,

Brown DM:Heparansulfate-rich anionic sites in the

hu-manglomerular basementmembrane-Decreased

concen-trationincongenitalnephroticsyndrome.NEnglJ Med

1983, 309:1001

23. AbrahamsonDR,HeinA,CaulfieldJP:Laminin in

glo-merular basement membranes of aminonucleoside

ne-phroticrats:Increasedproteinuriainducedby

antilami-ninimmunoglobulin G. Lab Invest 1983,49:38 24. EhrenreichT, Churg J: Pathologyof membranous

ne-phropathy. PatholAnnu 1968, 3:145

25. WilsonCB, Dixon FJ: Therenal responseto immuno-logicalinjury,TheKidney.EditedbyBMBrenner,FC

Rec-tor,Jr.Philadelphia,W. B.SaundersCompany, 1981,p

(14)

Vol. 122 * No. 1 BASEMENT MEMBRANE ANTIGENS 49

26. Bolton WK, Benton FR, Sturgill BC: Autoimmune glomerulotubular nephropathy in mice. Clin Exp

Im-munol 1978, 33:463

27. Krakower CA,GreensponSA:Localization of the nephro-toxicantigen within the isolated renalglomerulus.AMA Arch Pathol 1951, 51:629

28. Meezan E,Hjelle JT, Brendel K: A simple, versatile,

non-disruptive method for the isolation of morphologically andchemicallypurebasement membranes fromseveral

tissues. Life Sci 1976, 17:1721

29. PetersonGL:Asimplification of the proteinassaymethod ofLowry etal. which ismoregenerally applicable.Anal

Biochem 1977, 83:346

30. Foidart JM, Yaar M, Figueroa A, Wilk A, Brown KS,

LiottaLA:Abortion in mice induced by intravenous

in-jections ofantibodiesto type IVcollagenorlaminin:Am JPathol 1983, 110:346

31. Laurie GW,Leblond CP, Martin GR: Localization oftype IVcollagen,laminin, heparan sulfate proteoglycan,and

fibronectintothe basal lamina of basementmembranes. JCell Biol 1982, 95:340

32. Boselli JM, Macarak EJ, Clark CC, Brownell AG, Martinez-Hernandez A: Fibronectin: Itsrelationshipto

basement membranes.I.Lightmicroscopic studies. Col-lagen Rel Res 1981, 5:391

33. Martinez-HernandezA:Thehepatic extracellular matrix:

I. Electronimmunohistochemical studiesinnormalrat

liver. Lab Invest 1984, 51:57

34. Brown CA,Carey K,Colvin RB: Inhibition of

autoim-munetubulointerstitial nephritis in guinea pigs by heter-ologousantisera containinganti-idiotype antibodies.J

Immunol 1979, 123:2102

35. Karnovsky MJ:Formaldehyde-glutaraldehyde fixative of

highosmolality for use in electron microscopy. J Cell Biol 1965, 27:137A

36. Churg J, Mautner W,Grishman E: Silver impregnation for electron microscopy.JBiophys Biochem Cytol 1958, 4:841

37. Andres GA,Acinni L,Beiser SM, Christian CL, Cinotti GA,Erlanger BF, Hsu KC, Seegal BC: Localization of fluorescein labeledantinucleoside antibodies in glomeruli of patients with active systemic lupus erythematosus nephritis. J Clin Invest 1970, 49:2106

38. Timpl R, Dziadek M, Fujiwara S, Nowak H, Wick G: Nidogen: A newself-aggregating basement membrane protein. Eur J Biochem 1983, 137:455

39. Courtoy PJ, Kanwar YS, Hynes RO, Farquhar MG: Fibronectin localization in theratglomerulus. J Cell Biol

1980, 87:691

40. Amenta PS, Clark CC,Martinez-Hernandez A: Deposi-tion offibronectinandlaminin in the basement

mem-brane of theratparietalyolksac:Immunohistochemical

and biosynthetic studies. J Cell Biol 1983, 96:104 41. Courtoy PJ,Boyles J: Fibronectin in the

microvascula-ture:Localization in thepericyte-endothelialinterstitium. J Ultrastruct Res 1983, 83:258

42. Weis MA, Ooi BS, Ooi YM, Engvall E, Ruoslahti E: Im-munofluorescent localization of fibronectin in the human kidney. Lab Invest 1979, 41:340

43. Rohrbach DH, HassellJR,Kleinman HK, Martin GR: Alterations in basement membrane (heparan sulfate) pro-teoglycan in diabetic mice. Diabetes 1982, 31:185

44. SariolaH,Timpl R, vonder Mark K, Mayne R, Fitch JM,Linsenmayer TF, Ekblom P: Dual origin of glomer-ular basement membrane. Dev Biol 1984, 101:86