Immune Complex Glomerulonephritis - NCBI

American Journal ofPathology, Vol. 136, No. 3, Marcb 1990 Copyright © American Association ofPathologists

Mechanism of Formation of Subepithelial Electron-dense Deposits in Active In Situ Immune Complex Glomerulonephritis Shoji Kagami, Koichiro Kawakami, Kaname Okada, Yasuhiro Kuroda, Tetsuo Morioka, Fujio Shimizu, and Takashi Oite From the Department ofPediatrics, Tokushima University School ofMedicine, Tokushima, and the Department of Immunology, Institute ofNephrology, Niigata University School ofMedicine, Niigata, Japan

The influences of the epitope density on cationic antigens on the fate of immune reactants and theformation of subepithelial electron dense deposits (EDD) were studied in a model of active in situ immune complex glomerulonephritis (ICGN), using

ahapten-carriersystem. Threeweeksafterimmunization with trinitrophenol conjugated bovine serum albumin (TNPJ7.3-BSA), the left kidneys of rats were perfused with 500 ,ug of TNP6.2-cationized human immunoglobulin G (C-HIgG) or TNP3 1.3- C-HIgG. The renal tissues were then examined at intervals by light, immunofluorescence, and electron microscopies. The perfused kidneys of rats given high-valency antigens (TNP31.3) showed marked subepithelial EDDs with foot process retraction associated with proteinuria. In contrast, those of rats given low-valency antigens (TNP6.2) showed only small subepithelial EDDs beneath the slit membrane, which consisted of apparently normal epithelial cells, and did not develop proteinuria. Kinetic studies on immunofluorescence showed thatglomerular depositions of immune reactants (TNP-carrier conjugate, rat IgG, and C3) were present longer in rats treated with high-valency antigens than in those treated with low- valency antigens. We conclude that the epitope density on cationic antigens strongly influences the retention ofimmune reactants and the formation of subepithelial EDDs, as well as development of glomerular injury. (Am JPathol 1990, 136:631-639)

Most immune renal diseases are caused by deposition of circulating immune complexes or by local (in situ) formation of immune complex deposits; either of these are the

result of the combination of antibodies with endogenous or planted antigens.1 Recent progress has shown that the type of glomerular lesion produced depends largely on the site at which deposit formation occurs, which in turn determines what mediators of tissue injury are activated.2 Subepithelial immune deposits are believed to be mainly due to local formation of immune complexes and to be associated with marked increase in glomerular permeability. Activation of the complement system may have a direct effect in this process and result in development of the nephrotic syndrome.3 Little is known, however, about the molecular mechanism underlying the induction of glomerular injury or the accumulation of subepithelial immune deposits leading to formation of electron dense deposits (EDD). Agodoa et a14 reported that formation of subepithelial EDDs requires a precipitating antigen-antibody system in a passive model of in situ immune complex glomerulnephritis (ICGN). Similarly, Allegri et a15 showed that highly cross-linked immune complexes formed by polyvalent antibodies on the epithelial cell-surface antigen 330-kd glycoprotein (GP 330) are necessary for the induction of subepithelial EDDs in passive Haymann's nephritis. According to findings in these two models, EDDs seem to be due to subepithelial formation of large aggregates by interaction of antibodies with nephritogenic antigens. But the nature of these immune complexes that cause tissue damage associated with complement activation, and the factors determining the size of the EDDs are not fully understood. Recently we established a model of active in situ ICGN using a hapten-cationic carrier system and demonstrated that glomerular injury and inflammation were affected by the valence of the cationic antigen.6 In this work, we studied the effect of epitope density on cationic antigens on the formation of subepithelial EDDs and the fate of the immune reactants that were the main constituents of Supported by the Ministry of Education of Japan Grants 59570145 and 60570153. Accepted for publication November 1, 1989. Address reprint requests to Shoji Kagami, Department of Pediatrics, School of Medicine, Tokushima University, Kuramoto-cho 3-chome, Tokushima 770, Japan.

631

632

Kagami et al

AJPMarch 1990, Vol. 136, No. 3

these EDDs. The results showed that the epitope density on the cationic antigen influenced the morphologic features and biologic properties of subepithelial EDDs associated with glomerular injury.

Epon. Thin sections were cut on an LKB ultramicrotome and stained for 2 hours with uranyl acetate and for 5 minutes with lead citrate. The sections were examined in a Hitachi (H-S 12) electron microscope. At least eight glomeruli in each specimen were examined.

Materials and Methods Preparation of Antigens and Antibodies

Experimental Design

The antigens used in these studies were trinitrophenol (TNP) 6.2-C-HIgG (pl > 8.5), TNP31.3-C-HIgG (pl > 8.5), C-HIgG (pl > 9.0), TNP1 7.3-BSA (pi = 4.6), and TNP49.7KLH (pl, not determined), which have been described pre-

Four groups of male Wistar rats weighing 160 to 180 g were used. Three to five rats per group were examined at various times during the experiment.

viously.6 For use in examination of the localization of haptenic groups in renal tissue specimens, FITC-labeled anti-TNP antibody was obtained as described previously.6

Measurements of Serum Antibodies and Proteinuria Anti-TNP and anti-C-HIgG antibodies were determined by ELISA as described in detail elsewhere.6 The protein content of 24-hour urine samples was determined by the biuret method.7 The mean amount of urinary protein excreted in the age-matched, normal rats (without treatment) was 8.5 mg per 24 hours.

Histologic Examination For light microscopy (LM), renal tissue was fixed in buffered formalin, embedded in paraffin, and 4-um sections were stained with hematoxylin and eosin (H&E), periodic acidSchiff (PAS), or periodic acid silver hematoxylin (PAM). For immunofluorescent (IF) study, kidney specimens were snap frozen in liquid hexane at -70 C. Frozen sections were stained with FITC-labeled anti-human IgG (Medical and Biological Laboratories Co., Nagoya, Japan), anti-rat IgG, and anti-rat C3 (Nordic Immunochemicals, Holland, The Netherlands). Antibodies to human IgG and rat IgG were absorbed with rat IgG and human IgG, respectively, to remove cross-reactive materials. The staining intensity of immune reactants deposited along glomerular capillary walls and the mesangial region was semiquantitatively graded by two independent observers from 0 to 4+ as described previously.8 At least 40 glomeruli in each specimen were examined to determine the localizations of immune reactants. For electron microscopy (EM), 1-mm3 blocks of tissue were fixed in 2.5% glutaraldehyde in 0.1 mol phosphated buffer (pH 7.4), postfixed in 1% OS04, and embedded in

Group A (High-valency Group) Rats were immunized by injection of 1 mg of TNP1 7.3BSA in 0.5 ml of Freund's complete adjuvant (FCA) into several subcutaneous and intramuscular sites. Twentyone days later the left kidney was perfused with 500 ,ug of TNP31.3-C-HIgG (high-valency antigen) via the left renal artery. Rats were killed for histologic examination on days 1, 3, and 7, and then at weekly intervals until week 8.

Group B (Low-valency Group) Rats were immunized in the same way as those in group A and 21 days later the left kidney was perfused with 500 ,ug of TNP6.2-C-HIgG (low-valency antigen). Rats were killed at the same times as those in group A.

Group C (Control Group-1) Rats were immunized like those in group A and B and 21 days later the left kidney was perfused with 500 ,g of C-HIgG (carrier molecule). Rats were killed on days 1, 3, and 7, and in weeks 4 and 8.

Group D (Control Group-2) Rats were given only FCA and 21 days later and the left kidney was perfused with 500 Mug of high-valency antigen, as in group A. The rats were killed at the same times as those in group C.

Statistical Analysis Student's t-test was used.

Subepithelial EDDs in Active In Situ ICGN

633


(a) Anti-TNP Ab titers

(b) Anti-C-HIgG Ab titers

0.8 r OD492

0.8 -OD492

0.7

0.71-

0.6

[

0.61-

0.5 I

O.5s

0.4

0.4 _

Group Group Group Group

0.3 1

0.2 F 0. 1

A B C D

- *-.* o-o - a * *-e

0.3

0.2 0.1

N.R. *

Group A Group B Group C Group D

-N. R .

I

Before ID

Figure

1.

3D

1W

2W

3W

4W

6W

8W

Before ID

3D

1W

2W

3W

4W

6W

8W

Serum anti-TNP and anti-C-HIgG antibody level. Points are means- lSDfor values at least 10 rats (*normal rats).

Results

in groups A, B, and C at any time. No anti-C-HigG antibody actvity was detected in any group throughout the experiment, as shown in Figure 1 b.

Serum Antibody Levels

Urinary Protein Excretion As shown in Figure 1a, the anti-TNP antibody titers were significantly higher in groups A, B, and C than were those in group D and in rats that received no treatment from the time of renal perfusion until week 8 (P < 0.01). The antiTNP antibody levels gradually decreased but statistically, however, there was no significant difference in the levels Table 1.

In group A, proteinuria (>20 mg/24 hours) was detectable on day 1 (average 35 mg/24 hours) and reached a mean maximum of 58 mg/24 hours between weeks 2 and 4. No proteinuria was detected in other groups during the

experiment.

Presence of immune reactants in glomeruli

A

Group

B

Time

TNP.

HlgG.

RatUgG.

C3.

TNP.

HIgG.

ID

4+a

4+8a

4a

4+a

4+a

4+a

RatlgG. 4+a

C3. tr

3D

3+

4+

4+

3+

1+

4+

4+

0

1W

2+

3+

4+

2+

0

3+

4+

0

1+

2+

0

1 +a

0 0

2W

1+

2+

4+

1+

0

3W

°

2+a

4+a

0

0

4W

0

1+

3+

0

0

0

0

6W

0

0

2+

0

0

0

0

0

8W

0

0

2+

0

0

0

0

0

Immunofluorescent localization of immune reactants in glomeruli was graded from 0 to 4+ on the basis of their intensity of staining and distribution tr, trace amount. At least four rats were examined at each time. aExamples of these observations are shown in Figures 2 and 3.

634 Kagami et al AJPMarch 1990, Vol. 136, No. 3

Figure 2. Immunofluorescence staining of immune reactants in glomeruli of rats ofgroup A (a to d) and B (e to h) on day 1. Similar deposits of TNP(a, e), carrier C-HIgG (b, f) and rat IgG (c, g) were present on capillary walls, whereas the patterns and intensities of stainingfor rat C3 in the two groups (d, h) were quite different (X380).

Histologic Findings In group A, LM showed mild exudative changes from day 1 and moderate endocapillary proliferation on day 3, as

described previously.6 These changes then gradually subsided, but glomeruli showing slight hypercellularity with slight expansion of mesangial matrix, as well as those showing partial or total collapse, were observed until the


635

AJPMarch 1990, VoL 136, No. 3

11 I

l1 Figure 3. Localization of immune reactants in glomerulifrom groupA (a, b) andB(c, d) in week 3. GroupA showed granular deposits of rat IgG (a) andfine granular deposits of C-HIgG(b) on capillary walls. Group B showed a segmental loop pattern of deposits of rat IgG(c). Stainingfor C-HIgG(d) was negative (X360).

end of the experiment. LM in other groups appeared essentially normal throughout experiment. An IF study was performed to determine the glomerular localization of immune reactants and findings are summarized in Table 1, in which an average of the semiquantitative scores within each group was taken according to the gradings. On day 1, the glomeruli of groups A and B, given high- or low-valency antigens, respectively, showed intense depositions of the TNP haptenic group (4+), carrier molecule (4+), and rat antibodies (4+) on capillary walls (Figure 2). The carrier molecule and antibodies stained equally in the two groups after 3 days and 1 week (HIgG, 4+ after 3 days and 3+ after 1 week; rat IgG, 4+ after 3 days and 1 week) and decreased in parallel thereafter, but their staining decreased more slowly in group A than in group B (Figure 3). The haptenic group as antigenic determinant disappeared within 3 weeks in group A and within in 1 week in group B. The deposition of rat C3 differed greatly in groups A and B (Figure 2d and h). In group A, deposits were graded 4+ on day 1, and decreased to 3+ on day 3 and 1 + after 1 week, whereas in group B only a trace amount was detected on day 1, and no deposit was detected on day 3. Groups C and D showed mesangial staining (1 +) for each antigenic molecule on day 1. No other immune reactants were seen in either group on day 1 or thereafter. Ultrastructural examination of glomeruli of group A on day 1 revealed extensive accumulation of dense deposits

in the subepithelial region and beneath slit pores associated with marked foot process effacement (Figure 4a). A few polymorphonuclear and mononuclear leukocytes were seen in glomerular capillary loops. Small EDDs were observed in the mesangial area near the glomerular hilus. In sharp contrast, the glomeruli in group B examined on day 1 exhibited multiple small subepithelial deposits located regularly under slit pores between foot processes that had a normal appearance (Figure 4b). After 3 days and 1 week, the subepithelial deposits in groups A and B, respectively, were similar to those seen on day 1. In the next two or three weeks, the subepithelial deposits in group A became smaller and were gradually incorporated into the glomerular basement membrane (GBM) (Figure 4c). After 6 to 8 weeks, the GBM was irregularly thickened and had small translucent areas, indicative of reabsorbed deposits. In group B, the small subepithelial deposits seen initially rapidly disappeared, no EDDs were detected after 3 weeks (Figure 4d), and no abnormalities were observed thereafter. Control groups (C and D) showed no morphologic changes at any time examined.

Discussion From the present study using our model of in situ ICGN, we conclude that the persistence of immune reactants

(antigen, antibody, and complement) and the formation

,A.

Us

W!.:

:H:e 1:

a1.... ..

-~~~~~~~~~~~. . -i_ i. ik'' "Id m

I .

*!

t:

ii*

'4

...

Cr C

;,... 4

.W;

I\

;.q?'

64

.4

p

'P 4$

i

CL

/

so g

::

*

it 1. ...A c:

R,

I

S~#'

_

P7s / a C a

a

,4

w;

o ;9S

:'3 W"

B.=_ 0x

Figure 4. Electron micrographs ofglomerular capillary walls ofgroupA (a, c) and B(b, d) on day 1 (a, b) and in week 3 (c, d). a: Large subepithelial deposits (arrows) and loss of discrete foot processes are seen (X36, 000). b: Numerous small deposits are seen beneath slit pores (arrow-heads). Epithelial cells show normal interdigitating foot processes (X 72,000). c: Small subepithelial deposits are seen, associated with localfoot process retraction. Afew deposits are covered with newlyformed GBM-like material (arrow-heads) (X27,000). d: No detectable abnormalities are seen (X30,000). Abbreviations: US, urinary space; Ep, glomerular epithelial cells; End, glomerular endothelial cells; CL, capillary lumen.

iwr (s.L w , *.

F .


'#

-dim,

4r

.:s.. s; 1.;

.t

637

AJPMarch 1990, Vol 136, No. 3

-.,4 I.

.'i.0

+-.;

':

s

I

b

.1.

pw r-~

i.,

,.

'I

P4

.CL

N

-~ l-

.......I

-PA

:;

44.

:.J"i

-Y

I

O.

.4 Figure 4 (continued)

4"

638

Kagami et al


of EDDs in the subepithelial area are affected by the number of haptenic groups conjugated to the carrier molecule (epitope density). This conclusion is drawn from the observation that the glomerular deposition of immune reactants detected by IF was more persistent in rats treated with high-valency antigens (TNP31.3) than in those treated with low-valency antigens (TNP6.2). In addition, an EM study showed that large subepithelial deposits were formed in glomeruli of rats given high-valency antigens and were then retained in this area for a long time, whereas the deposits formed in rats given low-valency antigens were small and soon disappeared. There was no significant difference in the serum antiTNP antibody titers of the groups treated with high- and low-valency antigens (groups A and B) during the experiment. Similar levels of glomerular deposits of carrier-HIgG and rat IgG molecules were detected semiquantitatively in the two groups early in the experiment (from day 1 to week 1). Furthermore, using radiolabeled antigens we recently detected no significant differences in the amounts of antigens bound to the left kidney 1 hour after renal perfusion in groups A and B.6 Together these results suggest that the morphologic differences between the two groups resulting from subepithelial accumulation of immune complexes was not simply due to the amount of immune complex deposits in this area, but was probably caused by differences in the nature of immune complexes, particularly, the valence of the antigens in the immune complexes. In view of the role of lattice formation in both biologic activity and deposition of immune complexes in glomeruli, it is of considerable interest that complement activation was marked in group A but was minimal in group B. In general, the lattices of immune complexes are influenced not only by the degree of antigen excess, but also by the antigen valence, the affinity of the antigens and antibodies, and the absolute concentration of immunoreactants.9 For example, antigens with high valence tend to form large latticed complexes with the corresponding antibodies, whereas antigens with low valence tend to form small latticed complexes.4 The requirement for a sufficient lattice structure to achieve complement activation has been demonstrated by studies on circulating immune complexes.'° Thus the lattices in immune deposits composed of TNP31.3-C-HIgG in subepithelial areas may be large enough to activate the complement system, but lattices in immune deposits with TNP6.2-C-HIgG may not. The significant difference in the sizes of subepithelial EDDs in groups A and B is consistent with this concept and seems to substantiate findings on the influence of the lattice structure in formation of EDDs. Mannik et a11 also reported that only immune complexes in which the antigen and antibody can rearrange and condense into larger lat-

ticed immune deposits persist in glomeruli and become EDDs. Recently we investigated the process of antigen localization and the interaction of immune deposits with anionic sites of the GBM in an active model of in situ ICGN using cationized ferritin as detectable antigen.12 We found that initially the immune complexes were mainly formed on the subendothelial side of the GBM, corresponding to anionic sites in the lamina rara interna. These complexes may then dissociate into individual reactants or small immune complexes that penetrate the lamina densa and become rearranged into larger immune complex aggregates in the subepithelial space.13 After 24 hours, subepithelial ferritin aggregates did not correspond in position to the anionic sites on the lamina rara externa but were preferentially situated beneath the slit membrane and were sometimes in contact with the cell membranes of the foot processes. From these findings and the observed sequence of development of EDDs composed of cationic antigens with different epitope densities and the fate of immune reactants, we speculate that the following events occur in formation of EDDs in the subepithelial region: First, local immune complex formation mediated by charge-charge interaction occurs on the subendothelial side. Then the immune complexes dissociate into subunits that are transferred across the lamina densa and recombine on the subepithelial side. The immune complexes containing low-valency antigens reformed in this area became concentrated locally and condensed beneath the slit-pore membrane to form small latticed immune deposits, due to their low valence, although they are of sufficient size to become EDDs. On the other hand, reformed immune complexes containing high-valency antigens form larger latticed immune deposits resulting in massive EDDs, occupying not only slit pores but also other region of the subepithelial space. These EDDs cause complement activation resulting in a glomerular permeability defect associated with epithelial cell damage.14 This proposal is consistent with the association between the intensities of capillary deposits of IgG, always combined with C3, the degree of subepithelial EDDs, and the level of proteinuria observed clinically in patients with poststreptococcal GN

(PSGN).15 The mechanism-of local increase in concentration of immune complexes to form large aggregates at slit pores is not known, but seems likely to affect the function of slit membranes as a filtration barrier for macromolecules.16 During the course of disease, the immune deposits formed are not static, but are slowly removed. It is noteworthy that IF studies showed that antigens disappear faster than antibodies from immune deposits in both groups A and B. Because similar differences in the rates of disappearance of antigens and antibodies have been observed in a passive in situ ICGN model (cationic anti-


639


gen) and acute serum sickness model (anionic antigen), the antigen charge has no effect on elimination of antigens from immune deposits.17'18 This phenomenon must be closely associated with the fate of immune reactants in naturally occurring acute glomuerulonephritis (GN), particularly in GN in which the initiating antigen is not replicated. In fact, there is evidence that suggests that streptococcal antigens can be detected only in biopsy specimens from patients in the early stage of PSGN.19,20 In this work, we demonstrated the importance of the epitope density on cationic molecules in subepithelial EDD formation leading to glomerular capillary wall injury. Our results also increase our understanding of formation and removal of glomerular immune deposits observed in many renal diseases.

References 1. Couser WG, Salant DJ: In-situ immune complex formation and glomerular injury (editorial review). Kidney Int 1980,17:

1-13 2. Salant DJ, Adler S, Darby C, Capparell NJ, Groggel GC, Feintzeig ID, Rennke HG, Dittmer JE: Influence of antigen distribution on the mediation of immunological glomerular injury. Kidney Int 1985, 27:938-950 3. Couser WG: Mechanism of glomerular injury in immune complex disease. Kidney Int 1985,28:569-583 4. Agodoa LYC, Gauthier VJ, Mannik M: Precipitating antigenantibody systems are required for the formation of subepithelial electron-dense immune deposits in rat glomeruli. J Exp Med 1983,158:1259-1271 5. Allegri L, Brianti E, Chatelet F, Manara C, Ronco P, Verroust P: Polyvalent antigen- antibody interactions are required for the formation of electron-dense immune deposits in passive Haymann's nephritis. Am J Pathol 1986,126:1-6 6. Kagami S, Miyao M, Shimizu F, Oite T: Active in situ immune complex glomerulonephritis using the hapten-carrier system: Role of epitope density in cationic antigens. Clin Exp Immunol 1988, 74:121 -125 7. Weichselbaum TM: An accurate and rapid method for the determination of proteins in small amounts of blood serum and plasma. Am J Clin Pathol 1946,10:40-46

8. Gauthier VJ, Mannik M: Only the initial binding of cationic immune complexes to glomerular anionic sites is mediated by charge-charge interactions. J Immunol 1986, 136:32663271 9. Mannik M: Pathophysiology of circulating immune complexes. Arthritis Rheum 1982, 25:783-787 10. Mannik M, Arend WP, Hall AP, Gilliland BC: Studies on antigen-antibody complexes 1. Elimination of soluble complexes from rabbit circulation. J Exp Med 1971,133:713-739 11. Mannik M, Agodoa LYC, David KA: Rearrangement of immune complexes in glomeruli leads to persistence and development of electron-dense deposits. J Exp Med 1983, 157:1516-1527 12. Oite T, Shimizu F, Suzuki Y, Vogt A: Ultramicroscopic localization of cationized antigen in the glomerular basement membrane in the course of active, in situ immune complex glomerulonephritis. Virchows Arch (Cell Pathol) 1985, 48: 107-118 13. Vogt A, Rohrbach R, Shimizu F, Takamiya H, Batsford S: Interaction of cationized antigen with rat glomerular basement membrane: In situ immune complex formation. Kidney Int 1982, 22:27-35 14. Couser WG, Baker PJ, Adler S: Complement and the direct mediation of immune glomerular injury: A new perspective (editorial review). Kidney Int 1985,28:879-890 15. Sorger K, Gessler U, Huebner FK, Koehler H, Schulz W, Stuehlinger W, Thoenes GH, Thoenes W: Subtypes of acute postinfectious glomerulonephritis. Synopsis of clinical and pathological features. Clin Nephrol 1982,17:114-128 16. Kelley VE, Cotran RS: Mesangial and subepithelial localization of ferritin immune complexes in mouse glomerulus. Lab Invest 1972, 27:144-150 17. Oite T, Batsford SR, Mihatsch MJ, Takamiya H, Vogt A: Quantitative studies of in situ immune complex glomerulonephritis in the rat induced by planted cationized antigen. J Exp Med 1982,155:460-474 18. Wilson CB, Dixon FJ: Antigen quantitation in experimental immune complex glomerulonephritis. I. Acute serum sickness. J Immunol 1970,105:279-290 19. Lange K, Ahmed U, Kleingerger H, Treser G: A hitherto unknown streptococcal antigen and its probable relation to acute poststreptococcal glomerulonephritis. Clin Nephrol 1976, 5:207-215 20. Vogt A, Batsford S, Rodriguez-lturbe B, Garcia R: Cationic antigens in poststreptococcal glomerulonephritis. Clin Nephrol 1983, 20:271-279