Journal of Clinical InvestigationVol. 45, No. 2, 1966

Acute Poststreptococcal Glomerulonephritis: ImmuneDeposit Disease *

ALFREDF. MICHAEL, JR.,t KEITH N. DRUMMOND,tROBERTA. GOOD,§ANDROBERTL. VERNIER || WITH THETECHNICALASSISTANCEOF

AGNESM. OPSTADANDJOYCEE. LOUNBERG(From the Pediatric Research Laboratories of the Variety Club Heart Hospital and the

Department of Pediatrics, University of Minnesota, Minneapolis, Minn.)

The possible role of immunologic mechanisms inacute poststreptococcal glomerulonephritis wassuggested in 1908 by Schick (2), who comparedthe delay in appearance of serum sickness afterinjection of heterologous serum to the latent pe-riod between scarlet fever and onset of acute glo-merulonephritis. Evidence in support of this con-cept is the depression of serum complement duringthe early stages of the disease (3) and glomerularlocalization of immunoglobulin. Immunofluores-cent studies have revealed either no glomerulardeposition of a-globulin (4) or a diffuse involve-ment of the capillary wall (5-9). Seegal, Andres,Hsu, and Zabriskie (10) demonstrated the pres-ence of 7 S y-globulin, /31c-globulin, and strepto-coccal antigen in the glomeruli of most patientswith this disease. By studies employing ferritin-labeled antibodies, these proteins could be demon-strated in the mesangium, between the endothelialcells, and within and adjacent to the basementmembrane. Electron microscopic observations of

* Submitted for publication May 3, 1965; accepted No-vember 4, 1965.

Aided by grants from the U. S. Public Health Service(AI-02168 and HE-05662), the American Heart Associ-ation, and the Cardiovascular Clinical Research Center(HE-06314).

Presented in part at the American Society for ClinicalInvestigation, Fifty-sixth Annual Meeting, May 4, 1964,Atlantic City, N. J. (1).

t Established Investigator, American Heart Associa-tion.

t Present address: Dept. of Pediatrics, McGill Uni-versity, Montreal Children's Hospital, Montreal, Canada.

§ American Legion Memorial Heart Research Profes-sor of Pediatrics and Microbiology.

Address requests for reprints to Dr. Robert A. Good,Dept. of Pediatrics, University of Minnesota MedicalSchool, Minneapolis, Minn. 55455.

11 Established Investigator, American Heart Association.Present address: Dept. of Pediatrics, Center for Health

Sciences, University of California, Los Angeles, Calif.

the kidney in acute glomerulonephritis have alsorevealed the presence of discrete electron densemasses adjacent to the epithelial surface of theglomerular basement membrane (11-18).

The purpose of this paper is to describe immuno-fluorescent and electron microscopic observationsof the kidney in 16 children with acute poststrepto-coccal glomerulonephritis. This study demon-strates 1) the presence of discrete deposits of yG-and fl3c-globulins along the glomerular basementmembrane and its epithelial surface that are similarin size and location to the dense masses seen byboth thin section microscopy and electron micros-copy; 2) the characteristic and unique specificity ofthis lesion in acute poststreptococcal glomerulo-nephritis, its difference from other glomerular dis-eases, and the similarity to experimental antigen-antibody complex nephritis; and 3) the resolutionof these deposits during recovery from the diseaseand their absence from the glomeruli 10 years af-ter epidemic acute poststreptococcal glomerulo-nephritis.

Methods

Patients. The major portion of the study concernedsixteen children with acute poststreptococcal glomerulo-nephritis having the characteristic clinical course and re-nal histology as well as prior P-hemolytic streptococcalinfections and elevated or rising antistreptococcal anti-body titers. A total of 22 renal biopsies were performedon these patients by the method previously described(19) and at the times indicated in Table I. As part ofa separate follow-up study (20) 10 years after well-documented acute poststreptococcal glomerulonephritis(Red Lake, Minnesota, epidemic of 1953), additional re-nal biopsies were performed on 14 children and studiedby fluorescent microscopy.

Electron and fluorescent microscopic observations ofthe kidney were also made in one patient with lupusnephritis and one with Goodpasture's disease.

Histologic methods. Part of each biopsy was placedin 10% buffered formalin (pH 7.35) for standard paraffin

237

MICHAEL, DRUMMOND,GOOD, AND VERNIER

embedding and sectioning and subsequently stained withhematoxylin-eosin, azocarmine, and periodic-acid Schiff.Light microscopy was carried out on 21 of the 22 biopsies.A total of 12 to 30 glomeruli were evaluated in 17 of the21 biopsies; in each of the remaining specimens, five tosix glomeruli were present.

Another portion of the biopsy specimen was immedi-ately cut into cubes (0.5 mme) under a drop of buffered(1%o) osmic acid solution, dehydrated by standard tech-niques, and embedded in Vestopal-W1 polyester resin.Sections 0.5 to 1.0 j in thickness cut from these blockswere stained by either the Wright-Giemsa (21) or basicfuchsin method (22) and examined under oil immersionlight microscopy. Very thin sections for electron micros-copy were stained with either uranyl acetate (23) or leadcitrate (24) and examined with either an RCAEMU-3Dor a Phillips EMU-200 electron microscope at originalmagnifications of 2,500 to 75,000 times and enlargedphotographically as desired.

Immunofluorescent methods (25-27) were carried outon renal biopsy tissue that had been frozen in isopentanein liquid nitrogen, fixed to a small cellulose sponge,2 andsectioned in a Lipshaw cryostat. Fluorescent staining ofunfixed tissue was carried out according to the methodof Ortega and Mellors (28). In some instances an indi-rect technique was also employed. Inhibition of fluores-cence was carried out by pretreating the section once ortwice with unlabeled antiserum.

The sections were viewed in a Zeiss microscope withan HB-200 light source or a mercury lamp (GE-AHG)mounted in a Scopicon water-cooled unit. The ultra-violet activating and heat absorption filters were UG-1and KG-1, respectively. The barrier filters consisted ofOG-4 and a 2A Wratten (Kodak). Photographs weretaken with an Exacta camera using Kodak Tri-X orhigh speed Ektachrome film. The intensity and amountof fluorescence were arbitrarily graded as negative, trace,1+, 2 +, and 3 +. Six to 18 glomeruli were availablefor evaluation from each of the 22 biopsies.

Preparation of antisera. Antisera were prepared againstthe following antigens: 1) Human -yG-globulin was iso-lated by DEAE Sephadex A-50 chromatography; 3 im-munoelectrophoresis of this antigen against antihumanserum revealed a single precipitin line. 2) Human Pic-globulin, a part of the third component of complement,was prepared from 400 ml of fresh human serum by thecolumn chromatographic method of Muller-Eberhard,Nilsson, and Aronsson (29); the 8,c-globulin isolatedformed one precipitin line with both antihuman serumand with known rabbit anti-Pic-globulin serum.4 3) Hu-

1 Martin Jaeger Co., Geneva, Switzerland.2 Onkosponge no. 1, obtained from Histomed, Paterson,

N. J.8 Some of the -yG-globulin used was obtained from Im-

munology, Inc., Lombard, Ill.4 During the initial part of this investigation the rab-

bit anti-Pie serum was kindly supplied by Dr. H. Muller-Eberhard, Scripps Clinic and Research Foundation,La Jolla, Calif.

man fibrinogen was prepared from human Fraction I 5

according to the method of Laki (30, 31). 4) Humanfibrin was prepared from human fibrinogen; 25.0 mg offibrinogen was diluted in isotonic saline to a final con-centration of 5.0 mg per ml, and then 10.0 ml of 0.075 Msodium chloride in phosphate buffer (pH 6.1) and 0.5 mlof bovine thrombin were added. After standing for 1hour, the clots were washed with distilled water, lyophi-lized, and suspended in particulate form in 0.15 M saline.Antisera were also prepared against 5) human albumin(Cohn Fraction V),6 6) rabbit y-globulin (Cohn Frac-tion II) , and 7) heat-killed and ground-cell suspensionof nephritogenic, type 1, group A, fi-hemolytic strepto-coccus. 8) Extracellular products of the same strepto-coccus were prepared according to the method of Wan-namaker (32) and concentrated by ammonium sulfateprecipitation or by pervaporation.

Preparation of antibody. Antibody to these human anti-gens was prepared in rabbits. Immunizations were car-ried out with alum-precipitated human 'yG-globulin andalbumin, fibrin and Pic-globulin incorporated in completeFreund's adjuvant,8 and a saline solution of fibrinogen.Alum-precipitated rabbit y-globulin was injected intogoats. The specificity of the antisera was shown by im-munoelectrophoresis of the immunizing antigen and wholeserum developed against the antisera. Antisera showingmore than one major precipitin band were not used.A horse anti-'yM-globulin serum,9 employed in thesestudies, formed one major line on immunoelectrophoresisagainst normal serum; an additional faint line was alsoseen in the a2-globulin .region.

Antisera to the various streptococcal preparations wereobtained by immunizing rabbits three times weekly fora total of 2 months with 0.5 ml of antigen in solution or

suspension. Antigen preparations were injected intra-venously during the first month and subcutaneously dur-ing the second month of immunization. The developmentof antibodies to multiple antigens in these crude vaccinematerials was demonstrated by the appearance of a num-ber of lines of precipitation in a double-diffusion agarsystem.

The ammonium sulfate-precipitated globulins were di-alyzed and then conjugated with fluorescein-isothiocya-nate 10 (33, 34) with 0.015 g per g of protein. The la-beled antiserum was then passed through a column ofSephadex G-25 to remove the free fluorescein (35, 36).Antisera were stored as small portions at - 220 C untilused. Immediately before use, the antisera were ab-

5 Provided by the American Red Cross and Dr. Rich-ard von Korff, Department of Biochemistry, Universityof Minnesota, Minneapolis, Minn.

6 Cutter's normal human serum albumin, Cutter Labora-tories, Berkeley, Calif.

7 Rabbit y-globulin (Fraction II), Nutritional Bio-chemical Corp., Cleveland, Ohio.

8 Difco Laboratories, Detroit, Mich.9 Hyland Laboratories, Los Angeles, Calif.10 Baltimore Biological Laboratories, Baltimore, Md.

238

ACUTEPOSTSTREPTOCOCCALGLOMERULONEPHRITIS:IMMUNEDEPOSIT DISEASE

be(U.5!9

3: tU)

co

0 0

rxs>3 bo twf2==

U) )

Cd Cd Cbe tO be

Vf

Cd co+ b+ w

. . .Z

U) U)

Cd Cd ci

U) U)

z z z

V

+ + + 40

U

..v v ev) U)

kZ Z

+ +N' -

U) U) U) U) cu~~.

00 bO 0a be be4 4 4

+ ++

.mm

v -

+ + + "m VI +

C') eC'iC' Z C

^ vD - - Q _4 > = = 0 =

++ @+++ I +(4 m Z C ) - H "i

+ ++

+ + ++ ++C' - C4 - ( '

(U-'

+ + OUbO+ 0U + X

l 40 -- 00n

_ In oco 0 0o

0o s

(U

SC G4i

0

(Ud CaCO*_ 4.& e

0

(40 M

O

0'.0.':g .2

Cd _*C.rU (d(U'.°

OU. *.

U)(U(U

0440._

U0

IV Id wo

*'r.0 .X

Ck a.o(U.0

0 E.

&doY 2

0 be

,/4 0CdvSL

0w >cd o

o >, 0

44 .)C.(U%)

co0~a

-.U)0 .0(U rU vU( U

0CU) (U(

(Ur.o* R44U)-4

cU'

(U( c

239

bO(.

0 3 3c 0o cU~t~t ol (U(I.

4i V

m 8.

30

U)

.0.2

44

Q 0

C

W0

YCU

-(

.0

O(U

0 .

0

U)

(U..U 300a

0

.44

- U,)

m)

.4

C.o

.-

O75.00

44.Q044 -

+ + + + + w

-- v- V.= Sc > = >

U- -,- C

*

Cd + + be +bO

.-

.00

..acdbe

m z

.-(U c

1.fU U

At 0 0 ) 0l I'd, O (4~ .-_ 0 (4 N N " --U) t-

m -4-_ _-_4

-U))'(400

OU

vo- C

._

(U-E,

-1

_4 C_4 _.

MICHAEL, DRUMMOND,GOOD, ANDVERNIER

With the basic fuchsin technique the nodules werepale red or pink, whereas the basement membraneproper was intensely red.

II) Fluorescent microscopy

A) Acute glomerulonephritis. The glomerulardeposition of yG- and fl-globulins is summarized

FIG. 1. THIN SECTION (.5 ,) OF RENAL TISSUE OB-TAINED FROMPATIENT 6, 19 DAYS AFTER ONSET OF ACUTEGLOMERULONEPHRITIS, SHOWINGA SEGMENTOF THE GLO-MERULUS. Discrete deposits are visible along the epi-thelial surface of the basement membrane (arrows) orappear to lie in the epithelium (dotted arrow). (Wright-Giemsa x 900.)

sorbed twice with mouse liver powder.'0 Labeled anti-sera having low fluorescein to protein ratios were usedas recommended by Goldstein, Slizys, and Chase (37)and Frommhagen and Martins (38).

Results

I) Light microscopyLight microscopy of standard 4-,u sections in all

patients showed findings characteristic of acutepoststreptococcal glomerulonephritis: glomerularendothelial proliferation, decreased erythrocytesin the glomeruli, polymorphonuclear leukocyteinfiltration, segmentation, and in some casescrescents.

Thin sections (Patients 1 to 4, 6, 10, 12). Al-though nodular masses seen along the basementmembrane of glomerular capillaries of fluorescentand electron microscopy (see below) were notvisible in routine 4-,u sections because of thicknessof the tissue, these masses could be resolved bylight microscopy with thinner sections. Examina-tion of sections of 0.5 to 1.0 pt, cut from plasticblocks embedded for electron microscopy, disclosednodular masses along the basement membranes(Figure 1). Variable numbers of these masseswere seen in seven of seven cases (1 to 4, 6, 10,12) studied by this technique. The masses stainedsomewhat differently from the membrane itself.With Wright-Giemsa stain, the masses were adeeper blue than the basement membrane proper.

FIG. 2. RENAL BIOPSY OBTAINED FROM PATIENT 11, 23DAYS AFTER ONSET. A) Discrete linear deposits of yG-globulin are present throughout the glomerulus. A simi-lar morphologic picture is seen when the tissue isstained with fluorescein-conjugated rabbit anti-Pic-globu-lin. (Fluorescent microscopy X 350.) B) Small seg-ment of a glomerular capillary loop showing immunemasses (m) of 'yG-globulin along the epithelial surfaceof the basement membrane (BM). The basement mem-brane stains weakly. The deposits in this photograph are2 to 4 g in diameter. (Fluorescent microscopy X 2,200.)

240

ACUTEPOSTSTREPTOCOCCALGLOMERULONEPHRITIS:IMMUNEDEPOSIT DISEASE

in Table I. In 12 of 16 patients, yG-globulin wasdemonstrated in the glomeruli of -the initial biopsy;f31c-globulin was detected in glomeruli of all pa-tients. Typically, the fluorescence was discrete andnodular, scattered diffusely throughout the glo-meruli (Figure 2A); linear areas of basementmembrane and mesangial fluorescence were alsoseen, especially in patients with more severe dis-ease (Patients 10, 12, 14) or patients biopsied laterin convalescence, but this-morphologic distributionwas less frequent. In glomeruli from some pa-tients the fluorescent masses approached the sizeof erythrocytes and in some areas seemed to befused with other masses in a local distributionsparing parts of the glomerulus. In tissue fromother patients the nodules were smaller and gave aspeckled appearance to the glomerulus at thesemagnifications. The deposits appear to follow thecourse of the glomerular capillary walls at highermagnification, and in glomeruli from some patientsthe immune desposits clearly were present on theepithelial surface of the glomerular basement mem-brane (Figure 2B). There was no evidence thatthe fluorescence represented nuclear, cytoplasmic,or cellular staining. GammaG-globulin was notpresent on the glomeruli in Patients 4, 9, 13, and

FIG. 3. RENAL TISSUE FROMPATIENT 3, 8 DAYS AFTER

ONSET, STAINED WITH FLUORESCEIN-LABELED RABBIT ANTI-

HUMANFIBRIN SERUM. The exact localization in the glo-merulus cannot be determined although the fibrin is prob-ably present within Bowman's space or along the base-ment membrane; it is not deposited in the nodular formseen with -yG- or 8,c-globulins (cf. Figure 2). (Fluo-rescent microscopy X 350.)

FIG. 4. TISSUE SHOWINGA PORTION OF A GLOMERULUSOBTAINED FROMNPATIENT 3, 8 DAYS AFTER ONSETOF ACUTEGLOMERULONEPHRITIS, STAINED WITH FLUTORESCEIN-CON-JUGATEDRABBIT ANTIBODY AGAINST A GROUND-CELLPREPA-RATION OF A NEPHRITOGENIC TYPE 1, GROUPA, S-HEMO-LYTIC STREPTOCOCCUS. The discrete fluorescence cannotbe anatomically localized with respect to the basementmembrane, but appears to be present between cells orwithin the cytoplasm. (Fluorescent microscopy X 350.)

15. In these patients considerable amounts of 8,c-globulin were observed as discrete nodules.

Beta1c-globulin was also demonstrated by anindirect technique in nine patients so studied, butwhen this method was used, the fluorescence wasmore linear in appearance than when the directtechnique was used.

In Patient 10, the amounts of yG- and /1c-glob-ulins were decreased at 112 days. At 272 days,only trace amounts of yG-globulin were noted onthe glomeruli in focal distribution, but no flic-globulin could be detected. The tissues from Pa-tients 5 and 6, 193 and 184 days after onset, dem-onstrated neither Plic- nor yG-globulin. Slightlocal and focal deposition of both yG and f13c wasobserved in tissue obtained from Patient 12, 198days after onset.

The specificity of the fluorescence was confirmedby inhibition of fluorescence with unlabeled anti-serum to -yG-globulin in 11 of 11 instances, by in-hibition of fluorescence with antiserum to /,3c-globulin, and by negative fluorescence for humanalbumin in eight of eight preparations. No yM-

241

MICHAEL, DRUMMOND,GOOD, AND VERNIER

FIG. 5. ELECTRONMICROGRAPHOF A SMALL PORTION OF A GLOMERULUSFROMPATIENT 4, ILLUSTRAT-ING THE FREQUENCYAND VARIABILITY OF FORMOF DENSE IMMUNE DEPOSITS. ml (m = electron densemasses) appears to be attached to the basement membrane, which is more dense than normal in thearea of attachment. A polymorphonuclear leukocyte (PMN) is adjacent to the endothelium at thissite. m2 and m, appear to lie entirely within the epithelial cell although this is not proven. Notethe dense halo of cytoplasmic material around M2. EP = epithelial cell; END= endothelial cell.(x 16,500.)

globulin could be shown in the glomeruli of sixpatients so studied.

Fibrin was observed in the glomeruli of initialbiopsy specimens from five of eight patients; how-ever, the distribution of the fluorescence differedfrom the nodular distribution observed with yG-or /l3c-globulins (Figure 3). The pattern of fibrindeposition was often ribbon-like as though it werepresent within Bowman's space, along the base-ment membrane, or within the capillary lumen.

In ten patients (No. 2 to 6, 10 to 12, 13. 15)the tissue was stained with fluorescein-conju-gated globulin against two cellular preparations ofa nephritogenic type 1, group A, /8-hemolyticstreptococcus: a heat-killed and a ground-cell sus-pension." Definite discrete glomerular fluores-

11 These studies were done in collaboration with Dr.Lewis W. Wannamaker under the sponsorship of theCommission on Streptococcal and Staphylococcal Diseases,

cence was present in Patients 2 and 3 when eitherantiserum was used (Figure 4); this was inhibitedby the prior application of unlabeled antisera. Atrace of fluorescence was also seen in the glomeruliof Patient 4, but none was seen in the other pa-tients. It was not possible to localize exactly thefluorescence with respect to the glomerular base-ment membrane, although no discrete nodules wereseen. Fluorescein-conjugated globulin against theextracellular streptococcal products failed to showlocalization in Patients 2, 3, 4, 5, and 10; the re-mainder of the patients were not evaluated withthis antiserum because of lack of tissue. Fluores-cent antibody against the ground-cell preparation

Armed Forces Epidemiological Board. The support ofthe offices of the Surgeon General, Department of theArmy, Washington, D. C., and of the American HeartAssociation (Dr. Wannamaker, Career Investigator) inthis collaborative effort is hereby acknowledged.

242

ACUTEPOSTSTREPTOCOCCALGLOMERULONEPHRITIS:IMMUNEDEPOSIT DISEASE

of type 1 streptococcus showed no glomerularstaining in kidney tissue obtained from patientswith the nephrotic syndrome (2), chronic glo-merulonephritis (2), subacute glomerulonephritis(1), and diabetes mellitus (1), or in a normalindividual (1).

B) Ten years after acute glomerulonephritis.Renal biopsies from 14 patients who had had acutepoststreptococcal glomerulonephritis 10 yearspreviously were examined for the presence of yG-or 81c-globulins. All had normal renal function.The Addis counts in these patients were slightlyabnormal, but the frequency of this abnormalitywas not significantly different from that of a con-trol Indian population with no history of glomeru-lonephritis (20).

In 11 biopsies no glomerular deposition of yG-or,/1c-globulins was found; nine of these biopsiesshowed normal histology, but two showed mildfocal and local glomerular abnormalities. In threebiopsies, there was focal and local deposition of-yG- and fl3c-globulins along the basement mem-brane of approximately one-half of the glomeruli;the deposition was restricted to small loci and rep-resented a very small area of the glomerulus. His-tologically, two of these biopsies also showed mildfocal and local glomerular abnormalities, and onewas normal. No immune complex type of deposi-tion resembling that found during acute glomerulo-nephritis was seen.

III) Electron microscopy

Two to four glomeruli in each of 11 renal biopsyspecimens from 9 of the 16 patients included in thisreport were studied by electron microscopy.

The ultrastructural changes within glomeruliof the patients fell roughly into two patterns, eachrelated to the interval between onset of clinicaldisease and renal biopsy.

A) Electron microscopic findings in six renalbiopsy specimens obtained 7 to 26 days after onset(Patients 1 to 4, 6, 12). The prominent abnor-malities of ultrastructure in these biopsies included(Figures 5 and 6) 1) an increase in volume oflow density cytoplasmic material of luminal endo-thelial cells (endothelial swelling) ; 2) an increasednumber of cells identified as either luminal endo-thelial cells (endothelial proliferation) or centro-lobular cells (mesangial cells); 3) a striking in-

FIG. 6. TYPICAL EARLY CHANGESIN ENDOTHELIALCELLSAND BASEMENTMEMBRANEIN PATIENT 3. The lumen ofthis capillary is obliterated by blebs of endothelium.Densely stained material (arrows) lies within the base-ment membrane and bulges outward toward the epi-thelium in one area. (X 25,700.)

crease in the number of polymorphonuclear leuko-cytes within the glomerular capillary lumina usu-ally adjacent to the dense deposits within the base-ment membrane or opposite nodules on the epi-thelial side of the basement membrane; 4) irregu-lar circumscribed and linear areas of increaseddensity within the basement membrane with oc-casional zones of thinning of the lamina densa orirregular scalloped thickening (these were espe-cially prominent in Patients 2, 3, and 12); and 5)numerous dense, homogeneous, spherical depositsthat projected from the epithelial side of the base-ment membrane. The smaller masses often in-dented the lamina densa of the basement mem-brane (Figure 5) and appeared to lie wholly withinthe basement membrane. These ovoid bodies werealways more densely stained with lead citrate oruranyl acetate than the lamina densa. The largermasses were pedunculated and attached to the base-ment membrane by a narrow base that fused withthe lamina densa. Other larger spheres often ap-peared to lie entirely within the epithelial cell al-though this is uncertain since serial sections wouldbe needed to establish this point. Rarely, densegranular material could be observed within endo-thelial cytoplasmic vacuoles.

243

MICHAEL, DRUMMOND,GOOD, AND VERNIER

B) Electron microscopic findings in five kidneybiopsy specimens obtained 1 month or more afteronset (biopsy 2 from Patients 5, 6, and 14 andbiopsies 2 and 3 from Patient 10). The electronmicroscopic findings in the group were similar tothose described above except that the amount of

FIG. 7. A) TISSUE OBTAINEDFROMA 15-YEAR-OLD MALE

DURING THE COURSEOF SEVERE LUPUS NEPHRITIS. Fluo-rescent microscopy of a segment of a glomerulus revealsthe presence of 7yG-globulin along the glomerular base-ment membrane. Deposits are primarily along the epi-

thelial side of the basement membrane (arrows), but are

also present within it. Betaic-globulin was present in a

similar distribution. Compare with electron microscopicappearance in Figure 8.

B) TISSUE OBTAINED FROMA 26-YEAR-OLD MAN WITH

GOODPASTURE'SDISEASE. Gamma G-globulin is presentwithin the glomerular basement membrane giving a dif-fuse linear pattern. No isolated or discrete extramem-branous deposits are seen. Betaic-globulin was presentin a similar distribution. Compare with electron micro-scopic appearance in Figure 8. (Fluorescent microscopyx 875.)

dense material was decreased, and there werefewer deposits within the basement membrane.Nodular masses, where present, appeared to bewithin enlarged epithelial cells that contained nu-merous dilated cisternae and a decreased numberof mitochondria and lysosomes. A dense granularlayer of material approximately 200 A in thicknesssurrounded the nodular deposits and appeared toblend into the epithelial cytoplasm. The base-ment membrane from later biopsy specimensshowed focal areas of thickening alternating withareas of normal thickness and density. Glomerulifrom two biopsy specimens (Patients 5 and 6),obtained 6 months after onset of the disease, dem-onstrated only an occasional nodular mass thatappeared to be entirely within the epithelial cell;there was also an increase in membranous (stalk)material.

IV) Comparison of the immunofluorescent andultramicroscopic morphology of acute glomeru-lonephritis with other renal diseases

A nodular deposition of -yG- and flc-globulinshas also been seen in tissue from certain patientswith chronic glomerulonephritis and systemic lu-pus erythematosus (SLE). In the latter diseaselarge amounts of immune-complex deposits maybe present within or beneath the glomerular base-ment membrane or projecting from its epithelialsurface (Figure 7A). Electron microscopy of asimilarly involved glomerulus from the same pa-tient reveals numerous irregular dense masseswithin and projecting from the epithelial side ofthe basement membrane (Figure 8A). Thesemasses are very different from the nodules de-scribed in acute glomerulonephritis in that the le-sions observed in SLE are less compact, arerarely pedunculated, and have not been observedas nodules entirely within the visceral epithelium.

In other diffuse glomerular diseases such ascertain types of chronic glomerulonephritis andGoodpasture's disease, no nodules or masses areseen by light or fluorescent microscopy. Figure7B illustrates the fluorescent microscopic appear-ance of the glomerulus in a patient with Good-pasture's disease. The fine linear basement mem-brane fluorescence is due to y-globulin depositionin the basement membrane, which is uniformlythickened and without nodules when viewed un-

244

ACUTE POSTSTREPTOCOCCALGLOMERULONEPHRITIS:IMMUNEDEPOSIT DISEASE

41~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i

FIG. 8. SECTIONS OF TYPICAL GLOMERULARCAPILLARIES FROMA) A PATIENT WITH SYSTEMIC LUPUSERYTHEMATOSUS(SLE) AND B) A PATIENT WITH PULMONARYHEMORRHAGEAND GLOMERULONEPHRITIS(GOODPASTURE'S SYNDROME). Mottled density within the basement membrane and loosely packed nodularmasses (arrows) on the epithelial side of the basement membrane are features of the ultrastructural pa-thology in SLE. Compare with Figure 7. In Goodpasture's syndrome the basement membrane is dif-fusely and rather uniformly infiltrated with electron dense material (,y-globulin). Compare with Figure7. L = capillary lumen. (A, X 26,000; B, X 18,000.)

der the electron microscope (Figure 8B). Thispicture stands in sharp contrast to the morpho-logic appearance of the glomerulus in acute post-streptococcal glomerulonephritis.

DiscussionStrict criteria for the diagnosis of acute post-

streptococcal glomerulonephritis in 16 of our pa-tients included a characteristic clinical picture,

evidence of preceding streptococcal infection, andcharacteristic renal morphology. The salient find-ing was the presence of discrete nodular depositsof either 8,c- or yG-globulin, or both, in the glo-meruli of all 16 patients with acute disease andthe correlation of this type of deposition with den-sities, similar in conformation and location, in elec-tron microscopy and thin section microscopy oftissue from the same biopsies. These deposits are

245

MICHAEL, DRUMMOND,GOOD, AND VERNIER

characteristic of the disease. In a few patientswith severe disease, a prominent diffuse linearbasement membrane and mesangial fluorescencewas also observed. During convalescence, theimmune-complex deposits were either not present

at all or in minimal quantity by fluorescence mi-croscopy and seemed to be resolving in the fewpatients studied by electron microscopy. Inthese late specimens the densities seemed to beentirely within the epithelial cells and showedchanges interpretable of ongoing lysis.

This lesion is morphologically distinct from thediffuse linear distribution of y-globulin and com-

plement observed by fluorescent microscopywithin glomerular basement membrane in Good-pasture's disease and in some patients with chronicglomerulonephritis, in which no discrete depositsare present. Nodular basement membrane fluo-rescence has also been observed in disseminatedlupus erythematosus and in some cases of chronicglomerulonephritis (39); however, the yG andPic are not as discretely deposited.

Beta1c-globulin was demonstrated more consis-tently in the glomeruli of patients than was yG-globulin, which was not demonstrable even intrace amounts in four children. This may beanalogous to the non-y-globulin Coombs' test incertain hemolytic anemias where complement or itscomponents, and not y-globulin, may be found on

the red cell surface (40, 41). Leddy, Hill,Swisher, and Vaughan (41) have suggested thatthe y-globulin is present, but sterically hinderedfrom reacting with its homologous antiserum, or

that y-globulin has in fact dissociated from thecell leaving complement behind. It may also bethat the location of the deposit in some way deter-mines detectability of the y-globulin.

The closest experimental analogue to the lesionsin acute poststreptococcal glomerulonephritis isantigen-antibody complex nephritis, as observedby Dixon, Feldman, and co-workers (42-45).Our own studies (3) and those of Germuth andMcKinnon (46, 47) support the concept that thelesions of serum sickness reflect the toxicity ofthese complexes, which are also complement fix-ing. Light microscopy of the kidney in rabbitswith experimental serum sickness shows acuteproliferative glomerulonephritis and vasculitis, andelectron microscopy reveals swelling and prolifera-tion of the endothelial cells and deposition of elec-

tron-dense material adjacent to and within the glo-merular basement membrane (42-44). Ferritin-labeled studies have shown that these depositscontain bovine serum albumin (antigen) and rab-bit y-globulin (antibody) (48). The discretenodular fluorescent and electron microscopic ap-pearance of these nodular masses on the epithelialside of the basement membrane in acute serumsickness in rabbits is indistinguishable from therenal lesion of acute poststreptococcal glomerulo-nephritis in children (49).

Although the exact mechanism of production ofacute glomerulonephritis is speculative, the se-quence of events may be similar to that occurringin experimental antigen-antibody complex disease.The presence of streptococcal antigen in the glo-meruli of some patients with this disease tends tosupport but does not prove the antigen-antibodycomplex hypothesis.

Although the deposits of acute glomerulo-nephritis bear similarities to those of lupus nephri-tis and some forms of chronic glomerulonephritis,there are differences. In acute nephritis the de-posits tend to be discrete and appear to be presentwholly within the epithelial cell; in the other dis-eases, the nodular deposits may be located entirelywithin the basement membrane, as in some casesof chronic glomerulonephritis (39) ; the electron-dense material or yG- and fl1c-globulins may alsoappear as a linear density within the basementmembrane without extramembranous deposits asin Goodpasture's disease. The appearance of theglomeruli on immunofluorescent staining of thebiopsies in the latter disease is identical to thatpreviously shown to be characteristic of nephro-toxic serum nephritis (3). In acute nephritisthere is little distortion or disruption of the base-ment membrane, and the masses are not usuallytrapped or incorporated within the basement mem-brane; in the other diseases the basement mem-brane is extensively involved and becomes thick-ened and distorted.

Summary

1. A coordinate study of the light, immunofluor-escent, and electron microscopy of the kidney wascarried out in 16 children with acute poststrepto-coccal glomerulonephritis.

2. Discrete deposits of fic-globulin or yG-globu-lin or both were present within and along the epi-

246

ACUTE POSTSTREPTOCOCCALGLOMERULONEPHRITIS:IMMUNEDEPOSIT DISEASE

thelial surface of the glomerular basement mem-brane during the acute phase of the disease. Byelectron and thin section microscopy, these im-mune deposits were shown as dense masses lyingon the epithelial surface of the basement mem-brane, and rarely within the endothelial cell.This discrete deposition of yG- and /l3c-globulinsis morphologically distinct from other glomerulardiseases. Correlative experimental evidence indi-cates that these deposits represent immune aggre-gates, probably antigen-antibody complexes. Inthree of ten patients streptococcal cell antigen wasalso present in the ilomerulus.

3. The amount of yG- and /31c-globulin in theglomeruli decreased markedly or disappeared infour patients studied up to 9 months after onset.

4. Immunofluorescent studies on kidney tissueobtained from 14 patients 10 years after epidemicacute poststreptococcal glomerulonephritis revealsno deposition in 11 and minimal abnormalities in 3.

Acknowledgments

We appreciate the thoughtful editorial assistance ofMiss Ann Gabrielsen.

References1. Michael, A. M., K. N. Drummond, R. A. Good, and

R. L. Vernier. Immunoglobulins: clarification oftheir significance in renal disease and demonstra-tion of response to immunosuppressive therapy(abstract). J. clin. Invest. 1964, 43, 1291.

2. Schick, B. Die Nachkrankheiten des Scharlach. Jb.Kinderheilk. 1907, 65, 132.

3. Michael, A. F., K. N. Drummond, R. L. Vernier, andR. A. Good. Immunologic basis of renal disease.Pediat. Clin. N. Amer. 1964, 11, 685.

4. Taft, L. I., J. K. Dineen, and I. R. Mackay. Thelocalization and binding of serum proteins in theglomeruli of kidney biopsies in disseminated lupuserythematosus and glomerulonephritis. Aust. Ann.Med. 1958, 7, 5.

5. Freedman, P., J. H. Peters, and R. M. Kark. Lo-calization of gamma-globulin in the diseased kid-ney. Arch. intern. Med. 1960, 105, 524.

6. Freedman, P., and A. S. Markowitz. Gammaglobu-lin and complement in the diseased kidney. J. dlin.Invest. 1962, 41, 328.

7. Seegal, B. Fluorescent antibody studies of humanglomerulonephritis in Hereditary, Developmental,and Immunologic Aspects of Kidney Disease, J.Metcoff, Ed. Evanston, Northwestern UniversityPress, 1962, p. 148.

8. Lachmann, P. J., H. J. Miller-Eberhard, H. G.Kunkel, and F. Paronetto. The localization of

in vivo bound complement in tissue sections. J.exp. Med. 1962, 115, 63.

9. Koffler, D., and F. Paronetto. Immunofluorescentlocalization of immunoglobulin, complement, andfibrinogen in human disease. II. Acute, subacute,and chronic glomerulonephritis. J. dlin. Invest.1965, 44, 1665.

10. Seegal, B. C., G. A. Andres, K. C. Hsu, and J. B.Zabriskie. Studies on the pathogenesis of acuteand progressive glomerulonephritis in man by im-munofluorescein and immunoferritin techniques.Fed. Proc. 1965, 24, 100.

11. Steiner, J. W., and R. Volpe. Autoimmunization-apossible mechanism of tissue injury. III. Clinicalaspects of autoimmunity. Canad. med. Ass. J.1961, 84, 1297.

12. Movat, H. Z., J. W. Steiner, and D. Huhn. Thefine structure of the glomerulus in acute glomeru-lonephritis. Lab. Invest. 1962, 11, 117.

13. Kimmelstiel, P., 0. J. Kim, and J. Beres. Studieson renal biopsy specimens, with the aid of theelectron microscope. II. Glomerulonephritis andglomerulonephrosis. Amer. J. dlin. Path. 1962,38, 280.

14. Trump, B. F., and E. P. Benditt. Electron micro-scopic studies of human renal disease. Observa-tions of normal visceral glomerular epitheliumand its modification in disease. Lab. Invest. 1962,11, 753.

15. Churg, J., W. Mautner, E. Grishman, and G. M.Eisner. Structure of glomerular capillaries inproteinuria. Arch. intern. Med. 1962, 109, 97.

16. Movat, H. Z., J. W. Steiner, and R. J. Slater. Thefine structure of the glomerulus in Bright's dis-ease: a clinico-pathological study in Ciba Founda-tion Symposium on Renal Biopsy, G. E. W. Wol-stenholme and M. P. Cameron, Eds. London, J.and A. Churchill, 1961, p. 103.

17. Strunk, S. W., W. S. Hammond, and E. P. Benditt.The resolution of acute glomerulonephritis. Anelectron microscopic study of four sequential biop-sies. Lab. Invest. 1964, 13, 401.

18. Farquhar, M. G., R. L. Vernier, and R. A. Good.An electron microscope study of the glomerulusin nephrosis, glomerulonephritis, and lupus erythe-matosus. J. exp. Med. 1957, 106, 649.

19. Vernier, R. L. Kidney biopsy in the study of renaldisease. Pediat. Clin. N. Amer. 1960, 7, 353.

20. Perlman, L. V., R. C. Herdman, H. Kleinman, andR. L. Vernier. Poststreptococcal glomerulonephri-tis: a ten year follow-up of an epidemic. J. Amer.med. Ass. 1965, 194, 63.

21. Condie, R. M., R. Y. Tousignant, and R. A. Good.Rapid staining of osmic acid-fixed methacrylate-embedded tissue sections for light microscopy.Anat. Rec. 1958, 130, 399.

22. Chandra, S., and F. R. Skelton. Staining juxtaglo-merular cell granules with toluidine blue or withbasic fuchsin for light microscopy after Epon em-bedding. Stain Technol. 1964, 39, 107.

247

MICHAEL, DRUMMOND,GOOD, AND VERNIER

23. Watson, M. L. Staining of tissue sections for elec-tron microscopy with heavy metals. J. biophys.biochem. Cytol. 1958, 4, 475.

24. Reynolds, E. S. The use of lead citrate of high pHas an electron-opaque stain in electron microscopy.J. Cell Biol. 1963, 17, 208.

25. Coons, A. H., and M. H. Kaplan. Localization ofantigen in tissue cells. II. Improvements in amethod for the detection of antigen by means offluorescent antibody. J. exp. Med. 1950, 91, 1.

26. Coons, A. H. Fluorescent antibody methods in Gen-eral Cytochemical Methods, J. F. Danielli, Ed.New York, Academic Press, 1958, vol. 1, p. 399.

27. Mellors, R. C. Fluorescent antibody-method in Ana-lytical Cytology, R. C. Mellors, Ed. New York,McGraw-Hill, 1959, p. 1.

28. Ortega, L. G., and R. C. Mellors. Analytical pa-thology. IV. The role of localized antibodies inthe pathogenesis of nephrotoxic nephritis in therat. J. exp. Med. 1956, 104, 151.

29. Muller-Eberhard, H. J., U. Nilsson, and T. Aronsson.Isolation and characterization of two 6,-glycopro-teins of human serum. J. exp. Med. 1960, 111, 201.

30. Laki, K. The polymerization of proteins: the actionof thrombin on fibrinogen. Arch. Biochem. 1951,32, 317.

31. Mills, D. A., R. Coyne, B. Pollara, and R. W. vonKorff. The action of plasmin on fibrinogen andfibrin. I. Changes in the N-terminal residues.Biochim. biophys. Acta (Amst.) 1964, 86, 527.

32. Wannamaker, L. W. Electrophoretic studies of theextracellular products of group A streptococci.J. exp. Med. 1958, 107, 783.

33. Marshall, J. D., W. C. Eveland, and C. W. Smith.Superiority of fluorescein isothiocyanate (Riggs)for fluorescent-antibody technic with a modifica-tion of its application. Proc. Soc. exp. Biol.(N. Y.) 1958, 98, 898.

34. Riggs, J. L., R. J. Seiwald, J. H. Burckhalter, C. M.Downs, and T. G. Metcalf. Isothiocyanate com-pounds as fluorescent labeling agents for immuneserum. Amer. J. Path. 1958, 34, 1081.

35. Fothergill, J. E., and R. C. Nairn. Purification offluorescent protein conjugates: comparison ofcharcoal and "Sephadex." Nature (Lond.) 1961,192, 1073.

36. George, W., and K. W. Walton. Purification andconcentration of dye-protein conjugates by gel fil-tration. Nature (Lond.) 1961, 192, 1188.

37. Goldstein, G., I. S. Slizys, and M. W. Chase. Stud-ies on fluorescent antibody staining. I. Non-specific fluorescence with fluorescein-coupled sheepantirabbit globulins. J. exp. Med. 1961, 114, 89.

38. Frommhagen, L. H., and M. J. Martins. A com-parison of fluorescein-labeled 'y-globulins purifiedby rivanol and DEAE chromatography. J. Im-munol. 1963, 90, 116.

39. Michael, A. F. Unpublished observations.40. Muiller-Eberhard, H. J. Identification of complement

antigens coating red cells in acquired hemolyticanemia in Immunopathology, Third InternationalSymposium, P. Grabar and P. A. Miescher, Eds.Basel, Schwabe, 1963, p. 332.

41. Leddy, J. P., R. W. Hill, S. N. Swisher, and J. H.Vaughan. Observations on the immunochemicalnature of red cell autosensitization in Immuno-pathology, Third International Symposium, P.Grabar and P. A. Miescher, Eds. Basel, Schwabe,1963, p. 318.

42. Dixon, F. J., J. D. Feldman, and J. J. Vazquez. Ex-perimental glomerulonephritis. The pathogenesisof a laboratory model resembling the spectrum ofhuman glomerulonephritis. J. exp. Med. 1961, 113,899.

43. Feldman, J. D. Electron microscopy of serum sick-ness nephritis. J. exp. Med. 1958, 108, 957.

44. Feldman, J. D. Pathogenesis of ultrastructuralglomerular changes induced by immunologic meansin Immunopathology, Third International Sym-posium, P. Grabar and P. Miescher, Eds. Basel,Schwabe, 1963, p. 263.

45. Dixon, F. J., J. J. Vazquez, W. 0. Weigle, and C. G.Cochrane. Pathogenesis of serum sickness. Arch.Path. 1958, 65, 18.

46. Germuth, F. G., Jr. A comparative histologic andimmunologic study in rabbits of induced hyper-sensitivity of the serum sickness type. J. exp.Med. 1953, 97, 257.

47. Germuth, F. G., Jr., and G. E. McKinnon. Studieson the biological properties of antigen-antibodycomplexes. I. Anaphylactic shock induced bysoluble antigen-antibody complexes in unsensitizednormal guinea pigs. Bull. Johns Hopk. Hosp.1957, 101, 13.

48. Andres, G. A., B. C. Seegal, K. C. Hsu, M. S.Rothenberg, and M. L. Chapeau. Electron micro-scopic studies of experimental nephritis with ferritin-conjugated antibody. J. exp. Med. 1963, 117, 691.

49. Fish, A., A. F. Michael, R. L. Vernier, and R. A.Good. Correlation of experimental serum sick-ness nephritis with acute post-streptococcal glo-merulonephritis in man, Seventeenth Annual Con-ference on the Kidney, J. Metcoff, Ed. Warrenton,Va., 1965, in press.

248