A new role for complement in experimental membranous nephropathy in rats

Urinary excretion of C5b-9 reflects disease activity in passive Heymann nephritis

Passive Heymann nephritis (PHN) is a model of membranous nephropathy in rats in which glomerular injury is mediated by the terminal C5b-9 membrane attack complex of complement. This model has been shown to be associated with markedly elevated urinary excretion of C5b-9, compared to other experimental models of glomerulonephritis To determine if urinary C5b-9 excretion could serve as an index of disease activity by correlating with the formation and quantity of glomerular subepithelial immune deposits in PHN, we measured urinary excretion of C5b-9 in PHN under several experimental conditions. In the heterologous phase a direct correlation was demonstrated between levels of urinary C5b-9 excretion and the amount of anti-Fx1A IgG deposited in glomeruli (r = 0.85). In the autologous phase, C5b-9 excretion correlated with the amount of deposit forming antibody present in the serum and resolved when antibody disappeared, despite persistence of glomerular deposits of antigen, antibody, C5b-9 and heavy proteinuria. Glomerular C3 deposits paralleled urinary C5b-9 excretion. Re-initiation of active deposit formation by a second injection of anti-Fx1A produced new C3 deposits and a marked rise in C5b-9 excretion. Finally, complete abrogation of deposit formation by transplanting PHN kidneys into normal recipients also halted C5b-9 excretion. Our findings demonstrate that urinary excretion of C5b-9 is a sensitive index of on-going immunologic disease activity in the PHN model of membranous nephropathy.

Depletion of C6 prevents development of proteinuria in experimental membranous nephropathy in rats

To study the possible role of the complement membrane attack complex, C5b-9, in an experimental rat model that is morphologically indistinguishable from membranous nephropathy in man (passive Heymann nephritis [PHN]), an antibody to rat C6 was used to deplete C6 levels to less than 5% of pretreatment values (C6D) during disease development. C3, C7, C8, and C9 levels were not different in C6D and control rats. After injection of nephritogenic quantities of 125I-anti-Fx1A antibody, the kinetics of disappearance of labeled IgG from the blood were identical in the complement deficient and sufficient groups, and glomerular deposition of 125I-antibody was the same in both groups at 5 days. Glomerular deposits of sheep IgG and C3 were also similar in C6D and controls, but glomerular deposits of C6 and C5b-9 neoantigens were markedly reduced or absent in C6 depleted rats. However, despite equivalent antibody deposits, proteinuria was abolished in C6D rats compared with normocomplementemic controls. Similar results were obtained when F(ab')2 anti-rat C6 IgG was used to deplete C6 during development of PHN. These results demonstrate that C6 is required for the development of the increased glomerular permeability that occurs in PHN, presumably because C6 is required for formation of C5b-9. We conclude that glomerular injury in the PHN model of membranous nephropathy in the rat is mediated by C5b-9.

Preparative polyacrylamide gel electrophoresis in the isolation of the nephritogenic proteins of passive Heymann nephritis

To study kidney antigens involved in the formation of glomerular subepithelial immune deposits in passive Heymann nephritis polypeptides of 500, 130 and 105 kDa were isolated from rat kidney brush border (BB) membrane fraction using preparative polyacrylamide gel electrophoresis. Polyclonal antibodies raised against these proteins were specific for their respective antigens in immunoblotting. All three antisera bound to proximal tubular BB of kidney and to apical surfaces of several other epithelia as shown by indirect immunofluorescence on frozen sections of normal rat tissues. The anti-500 kDa and anti-105 kDa, but not the anti-130 kDa, antibodies also stained glomeruli and the anti-105 kDa antibodies also endothelial cells. After injection into rats the anti-500 kDa IgG bound to kidney glomeruli forming diffuse, granular deposits of rabbit IgG along the glomerular capillary walls, as shown by direct immunofluorescence. In electron microscopy the immune deposits were subepithelial and electron dense. The deposits remained in glomeruli for at least 60 days and increased with time. Deposits of C3 were not detected and proteinuria did not develop. The anti-130 kDa and the anti-105 kDa IgGs did not form glomerular deposits after in vivo injections. The results suggest that the 500 kDa and the 105 kDa proteins or related antigens are present in glomeruli and the 500 kDa protein is located on the epithelial side of the glomerular basement membrane. Circulating antibodies can bind to the 500 kDa protein forming immune complexes which rearrange and form electron dense deposits. The results further demonstrate that preparative gel electrophoresis is a useful technique for the isolation of kidney proteins of immunopathologic interest.

Studies with antibodies to cultured rat glomerular epithelial cells. Subepithelial immune deposit formation after in vivo injection

To investigate the role of glomerular epithelial cell (GEC) membrane proteins in the in situ formation of subepithelial immune deposits, the authors raised a rabbit antiserum against GEC that had been grown in culture (anti-GEC). By indirect immunofluorescence (IF) on normal rat kidney, anti-GEC stained proximal tubular brush border (BB). After intravenous injection into animals, granular glomerular capillary wall staining for IgG was present by IE and subepithelial immune deposits were identified by standard transmission and immunoelectron microscopy. Using the latter technique, injected anti-GEC IgG was identified beneath slit diaphragms and in endocytic-coated pits and intracellular vesicles of podocytes. Anti-GEC immunoprecipitated gp330 and two other proteins from radiolabeled BB. These proteins also were identified by sheep anti-rat Fx1A, the antiserum responsible for passive Heymann nephritis. Anti-GEC and anti-Fx1A also immunoprecipitated five identical proteins from surface-labeled GEC. Biosynthetically-labeled but not surface-labeled GEC contained immunoprecipitable gp330. Thus, injection into rats of antibodies raised against cultured GEC can produce subepithelial immune deposits, a disease process classically induced by antibodies to BB or its purified components. In addition to gp330, GEC and BB share other antigenic determinants that may contribute to the formation of these immune deposits.

Leprosy and glomerulonephritis: case report and review of the literature

Membranoproliferative glomerulonephritis (MPGN) is a common form of glomerulonephritis and frequently is associated with chronic infections. Leprosy, one of the most common infections worldwide, was found in conjunction with MPGN, type I, in a patient. Serological abnormalities typical of MPGN, improvement in renal function with therapy of acute complications of leprosy, and long-term renal improvement with antileprosy therapy all occurred in this patient. Others have found that MPGN is found in 11% to 43% of leprosy patients undergoing renal biopsy. Serological abnormalities typical of MPGN frequently are found in patients with lepromatous leprosy. The associations of MPGN and leprosy, and the susceptibility of the glomerulonephritis to therapy, should be emphasized.

Increased urinary excretion of C5b-9 distinguishes passive Heymann nephritis in the rat

Increased urinary excretion of C5b-9 distinguishes passive Heymann nephritis from other forms of experimental glomerulonephritis in the rat. In the passive Heymann nephritis (PHN) model of membranous nephropathy (MN) subepithelial deposits form from anti-Fx1A antibody reacting with antigen expressed on the glomerular epithelial cell membrane followed by membrane patching and shedding of immune complexes. Immune complex deposits are accompanied by deposits of C5b-9 which is required for the mediation of proteinuria. We tested the hypothesis that C5b-9 assembly on the epithelial cell membrane might result in C5b-9 excretion in the urine, which would distinguish this autoimmune mechanism of MN from other processes that result in subepithelial immune complex deposits. Using monoclonal antibodies developed to rat C6 and a rat C5b-9 neoantigen, in a sensitive ELISA assay, elevated urinary excretion of rat C5b-9 was documented in PHN associated with on-going glomerular immune deposit formation. No urinary C5b-9 was detectable in MN induced by an exogenous antigen (cationized IgG) despite equivalent glomerular C5b-9 deposits, or in models of nephrotoxic nephritis, subendothelial immune complex nephritis, anti-mesangial cell membrane antibody-induced nephritis or two non-immune nephropathies. Infusion of preformed C5b-9 in proteinuric animals excluded glomerular filtration of C5b-9 as a contributing mechanism to urinary C5b-9 excretion. We conclude that in the rat, increased urinary excretion of C5b-9 is a marker of MN induced by antibody to a glomerular epithelial cell antigen. Urine C5b-9 excretion reflects active glomerular immune deposit formation and distinguishes MN induced by this mechanism from other forms of MN as well as from other glomerular diseases with equivalent glomerular C5b-9 deposits.

Activation of complement in IgA nephropathy

Considerable evidence supports a role for the complement system in the pathogenesis of IgA nephropathy (IgAN). The alternative pathway components C3 and properdin (P) and the membrane attack complex (C5b-9) are generally found in the mesangial deposits in IgAN, while the classical pathway components C1q and C4 are usually absent. This pattern of immunofluorescence staining for complement components suggests activation of the alternative and terminal pathways in most patients. Despite normal serum concentrations of C3 and other complement proteins, fragments generated by activation of C3, including iC3b, C3d, and iC3b-C3d neoantigen, and sometimes C4, are often detected in plasma. We found that the severity of the histologic changes in the renal biopsy specimens correlated with plasma iC3b-C3d neoantigen concentrations as measured by an enzyme-linked immunosorbent assay. However, no other clinical feature correlated with the plasma concentrations of this neoantigen.

Determinants of immune complex-mediated glomerulonephritis

We have studied the influence of steric factors on the clinico-pathologic expression of immune complex-mediated glomerular diseases, utilizing ferritin as an exogenous antigen. The tracer was planted in the left kidney either in the subepithelial layer of the glomerular capillary wall or on the endothelium and lamina rara interna. Subepithelial immune complex formation resulted in non-inflammatory injury with heterologous and autologous proteinuric phases (115 +/- 16 mg/24 hrs on day 2; 183 +/- 16 mg/24 hrs on day 9) lasting four to five weeks. The glomerular filtration rate of the experimental left kidney was reduced by 19% at day 3, and was increased by 20% at day 12 over right kidney values. Immune complexes persisted for more than seven weeks in the lamina rara externa. In contrast, immune complex deposition on the endothelium and in the lamina rara interna led to acute transient anuria, with a 38% drop in glomerular filtration rate at one hour, massive platelet accumulation, followed by a strong inflammatory response. Proteinuria did not develop. Functional and structural integrity was restored within 24 hours, with complete clearing of immune deposits. We conclude that the distribution of exogenous antigens within the capillary wall determines the structural and functional expression of immune-mediated glomerular diseases.

Nephritogenicity of proteoglycans. II. A model of immune complex nephritis

Antibodies to glomerular basement membrane, heparan sulfate-proteoglycans are nephrotoxic but possess a weak nephritogenic potential. In order to enhance the nephritogenic potential, the antibodies were intravenously administered into rats presensitized with heterologous rabbit IgG. This resulted in the integration of heterologous and autologous phases, the two phases characteristic of the traditional model of nephrotoxic serum nephritis. The presensitization caused a dramatic shift in the binding characteristics of the heterologous antibodies between the kidney and lymphoid tissues. A proliferative form of immune complex glomerulonephritis associated with a remarkable proteinuric response was observed. In addition, a moderate degree of hematuria was noted as well. The proteinuria was largely complement-dependent and may possibly be cell-mediated as well. The proteinuria became severe with increasing production of host IgG antibodies and with their subsequent sequestration in the glomeruli. The predominant glomerular lesions were in the form of epimembranous/subepithelial immune deposits, which became more frequent with timely increasing titer of host autologous IgG antibodies. These findings indicate that antibodies to heparan sulfate-proteoglycan, an authentic component of the basement membrane, are capable of mediating a glomerular injury with acquisition of nephritogenic potential in an appropriate environment of the host. At present, it seems that this is the sole constituent of the basement membrane whose antibodies are capable of inducing an immune complex nephritis.

Anti-Fx1A produces complement-dependent cytotoxicity of glomerular epithelial cells

Glomerular injury in passive Heymann nephritis (PHN) in rats is mediated by the C5b-9 membrane attack complex (MAC) and is associated with morphologic changes in glomerular visceral epithelial cells (GEC). We determined if the nephritogenic antibody of PHN (gamma 1 sheep anti-Fx1A IgG) directs insertion of the MAC into GEC plasma membranes with consequent cytotoxicity. Antibody-sensitized GEC were exposed to various sera serving as sources of complement. Loss of cell viability was determined by trypan blue uptake and/or by release of cellular lactate dehydrogenase (LDH). Incubation of antibody-sensitized primary and passaged GEC in fresh human serum (FHS) resulted in sigmoidal relationships between cytotoxicity and complement dose (r = 0.97 and 0.94, respectively) such that cytolysis approached 100% with FHS (10% vol/vol). Cytotoxicity was not evident if C8-deficient (C8D) plasma was substituted for FHS, but was restored in a dose-dependent manner by reconstitution with purified rat C8. Sublytic injury was demonstrated by wide separation between simultaneous release curves of cell-incorporated biscarboxyethyl carboxyfluorescein (BCECF; mol wt approximately equal to 520) and LDH at limiting doses of complement (at 2% FHS, BCECF release was 51.1 +/- 0.6% of maximum vs. 3.2 +/- 1.3% for LDH; N = 3) and by blebbing of the plasma membrane on electron microscopy. Thus, the pathogenic antibody of PHN produces complement-mediated sublytic as well as lytic cytotoxicity of GEC.

Renal cell culture

Methods for the establishment and growth of renal cell types in culture are reviewed, with emphasis on current trends. General techniques available for the isolation and culture of glomerular cells have progressed from explant to enzyme dissociation and cloning techniques. The growth characteristics and properties of cultured glomerular endothelial, epithelial, mesangial, and bone-marrow-derived cells are discussed. Studies are described in which cultures of contractile mesangial cells have led to an elucidation of their role both in normally functioning glomeruli and in disease states. Renal tubule culture techniques also have progressed from mixed tissue explants and cell isolates to fractionation of enriched tubule populations and growth of specific, individually microdissected proximal convoluted, proximal straight, thick ascending limb of Henle's loop, and collecting tubules. The differentiated tubule epithelial-specific properties of such primary cultures are discussed in relation to those of permanently growing cell lines such as MDCK and LLC-PK1. Renal tubule cultures will be invaluable for the study of the role of hormones and extracellular matrix in epithelial growth and polarity of normal structure and function. In addition, in vitro models of cultured renal tubules have been established to study the effects of age, nephrotoxins, and anoxic injury.

Epitope specific induction of proteinuria by monoclonal antibodies

A monoclonal antibody, K9/9, directed against a novel epithelial cell surface sialo-glycoprotein, SGP-115/107, present in the rat glomerulus, has been shown to induce glomerular epithelial cell effacement and retraction, and an increase in protein excretion rate upon in vivo administration. Such damage is not seen upon administration of two additional monoclonal antibodies that recognize this epithelial cell antigen, but with different epitope specificities. To further clarify the mechanism of the epithelial cell abnormality, in vitro studies were performed on glomerular epithelial cells established in primary culture. None of these antibodies alone appeared to induce alterations in the cultured cells. However, an antibody of the IgG2a isotype induced complement-dependent cell damage in vitro, although failed to be pathogenic when administered in the intact animal. The pathogenic potential of K9/9 cannot be attributed to its isotype or rates of association or dissociation from the antigen. Studies suggest that all three monoclonal antibodies recognize different, though spatially close epitopes on SGP-115/107. These results demonstrate, for the first time, a complement- and leukocyte-independent mechanism of tissue injury that results from an epitope-specific interaction between a monoclonal antibody and its specific, epithelial cell surface-antigen. Results obtained in other cell systems suggest that abnormalities of epithelial cell structure and function can result from the interaction between specific cell surface components, particularly growth factor receptors, and monoclonal antibodies that mimic the actions of the specific agonist.

Effect of diet, age and sex on the renal response to immune injury in the rat

We investigated the effect of three factors, namely dietary protein intake, age and sex, on the susceptibility of the renal glomerulus to the binding of antiglomerular basement membrane antibody (anti-GBM) in the early (heterologous) phase of anti-GBM nephritis, and the consequent reduction in glomerular filtration rate (GFR) as measured by inulin clearance (CIn). The effect of diet was examined in approximately equal to 8 week-old female Munich-Wistar rats fed a 40% high (HP) or a 6% low (LP) protein diet, and that of sex and age in male and female rats, 6 week or 10 month old. Following an intravenous dose (3 to 20 micrograms/g body wt) of radiolabeled nephritogenic anti-GBM, assessment of glomerular function was followed by quantitation of anti-GBM binding (values corrected for GBM surface area) in isolated glomeruli. At a given plasma level of antibody, the degree of binding of anti-GBM was slightly but significantly higher in HP than LP-fed rats; the decrease in GFR was significantly more pronounced in HP than LP-fed animals. The amount of anti-GBM binding was significantly greater in adult than young animals; however, the consequent decrease in GFR was more pronounced in the young than adult animals. Sex dependency was not discernible in anti-GBM binding or reduction in GFR. In all of the above experimental groups, the degree of anti-GBM binding was closely correlated with the plasma level of anti-GBM, but not with effective renal plasma flow rate, measured by PAH clearance. Separate groups of rats were subjected to experimental manipulation of single nephron GFR, glomerular capillary hydraulic pressure and glomerular plasma flow rate, by partial aortic constriction and saralasin administration. This set of experiments, using a tracer amount of non-nephritogenic anti-GBM, revealed that glomerular anti-GBM binding is independent of any of the above parameters. The studies indicate that dietary protein intake and age, but not sex, are among the factors determining the susceptibility of the glomerulus to acute immune injury. Since the binding of anti-GBM is determined by the affinity property of the glomerulus per se, and not by the prevailing hemodynamic pattern, the observed dependence of susceptibility to functional impairment on age and protein intake appears to also reflect a property of the glomerulus, which is influenced by age and the degree of dietary protein intake.

Anti-Fx1A induces association of Heymann nephritis antigens with microfilaments of cultured glomerular visceral epithelial cells

Binding of anti-Fx1A to Heymann nephritis antigens (HA) on rat glomerular epithelial cells (GECs) in culture leads to capping and disappearance of antigens from the cell surface. This process may contribute to the formation of glomerular subepithelial immune deposits in vivo. The authors differentially extracted GECs to determine whether HA redistribution is mediated by cytoskeletal components. Observations were made by phase-contrast and immunofluorescence microscopy on primary and passaged GECs in monolayer culture and by spectrofluorimetry on GECs in suspension. GEC-bound sheep anti-Fx1A IgG was detected by fluoresceinated anti-sheep IgG. Microfilaments were identified by rhodamine-phalloidin staining of F-actin. After cross-linking HA on GECs by anti-Fx1A IgG at 0 C, GECs remained polygonal in shape and had diffuse granular IgG staining of their plasma membranes. Treatment of GECs at 0 C with hypotonic buffer containing 0.5% Triton X-100 produced microfilament-rich cytoskeletons that retained the shape of unextracted GECs. Further incubation with DNase I at 37 C removed microfilaments (mean fluorescence declined by 90%) and resulted in the rounding of cytoskeletons. After Triton X-100 treatment, 85% of initial GEC-bound anti-Fx1A IgG remained, but only 29% remained after DNase I. In contrast to intact IgG, detergent-extraction resulted in the complete loss of GEC-bound anti-Fx1A Fab'. Anti-Fx1A IgG did not bind to GECs pretreated with Triton X-100. Thus, cross-linking of HA by anti-Fx1A converts HA from a detergent-soluble, membrane-associated form to an insoluble, cytoskeleton-bound form. Attachment of cross-linked HA to the cytoskeleton is mediated by microfilaments.

Terminal complement complex in plasma from patients with systemic lupus erythematosus and other glomerular diseases

To evaluate terminal complement pathway activation in plasma from patients with systemic lupus erythematosus (SLE) and primary glomerular diseases, we developed an enzyme-linked immunosorbent assay (ELISA) for measuring the terminal complement complexes (TCC). The method is based on a sandwich technique using rabbit antibodies against native human C5, C7 and C9. To avoid interference by native components, we equilibrated plasma specimens with 5% polyethylene glycol buffer. The precipitates were measured by ELISA. TCC was detectable in all 14 normal controls (0.48 +/- 0.06 AU/ml; mean +/- s.e.m.). TCC levels were elevated in 18 of 54 patients with SLE (0.89 +/- 0.07 AU/ml; P less than 0.01) and in eight of 11 patients with membranoproliferative glomerulonephritis (MPGN) (3.15 +/- 0.62 AU/ml; P less than 0.01). However, only one of six patients with membranous nephropathy and none of 13 with mesangial proliferative glomerulonephritis showed high values. In SLE, TCC was correlated with circulating immune complexes and inversely with CH50, C3, C4, C5 and alternative complement pathway activity (AH50), and showed significantly high values even in normal CH50 cases (n = 34; P less than 0.01). In MPGN, TCC was inversely correlated with CH50, AH50, C3, C5 and C9. These results suggest that the classical pathway plays an important role for TCC generation in SLE and that the alternative pathway does in MPGN.

Role of abnormally high transmural pressure in the permselectivity defect of glomerular capillary wall: a study in early passive Heymann nephritis

To explore the mechanism of glomerular permselectivity defect in passive Heymann nephritis, an experimental model of human membranous glomerulopathy, Munich-Wistar rats were subjected to both micropuncture assessment of glomerular hemodynamics and whole kidney clearance measurements of graded size dextrans 10 days after injection of sheep anti-rat tubular antigen (anti-Fx1A). Compared with normal control rats, anti-Fx1A-treated animals were characterized by marked proteinuria (65 +/- 8 micrograms/min versus 6 +/- 1, p less than 0.001), markedly and significantly higher glomerular transcapillary hydraulic pressure difference (40 +/- 1 mm Hg versus 33 +/- 1, p less than 0.001), depressed ultrafiltration coefficient and impaired glomerular size-selective function as determined by fractional clearance of dextrans. Calculation of membrane parameters based on a recently defined heteroporous membrane model revealed abnormally high availability of non-size selective, large pore pathways in the glomerular capillary wall of the rats with passive Heymann nephritis. To ascertain the role of the altered hemodynamic pattern in the observed defect in the size-selective function of the glomerular capillary wall, glomerular transcapillary hydraulic pressure difference was manipulated experimentally in these proteinuric rats by intra-aortic infusion of acetylcholine or angiotensin II. These agents respectively suppressed and augmented glomerular transcapillary hydraulic pressure difference and brought about a decline of and a further rise in fractional clearance of larger dextrans along with parallel changes in both urine protein excretion rate and availability of nonselective channels. These results indicate that the permselectivity defect in passive Heymann nephritis is attributable, at least in part, to impaired size selectivity of the glomerular capillary wall caused by a prevailing abnormally high transcapillary hydraulic pressure difference.

Participation of the myeloperoxidase-H2O2-halide system in immune complex nephritis

Neutrophils (PMNs) mediate injury in experimental glomerulonephritis (GN) in part via the release of reactive oxygen species, particularly hydrogen peroxide (H2O2). Recent kidney perfusion studies demonstrate that H2O2 can cause glomerular injury by reaction with halides in the presence of the PMN cationic enzyme myeloperoxidase (MPO) to form oxidants which can oxidize and halogenate tissue. We sought evidence for participation of the MPO system in a model of PMN-mediated immune complex (IC) GN. A PMN-dependent model of GN was developed in rats by perfusing the renal artery with concanavalin A followed by anticoncanavalin A antibody. PMN depletion abolished glomerular PMN infiltration and significantly reduced proteinuria (35 +/- 7 mg/day vs. 113 +/- 10, P less than 0.001). Rats that received Na125I (5.0 microCi) three and six hours following disease induction had more 125I incorporation in glomeruli and GBM at 48 hours than similarly treated rats that were PMN depleted (1200 cpm vs. 88 cpm, P less than 0.01). Glomerular iodination could not be demonstrated in a PMN-independent model of nephrotoxic nephritis induced with noncomplement fixing anti-GBM antibody. These data indicate that this model of PMN-mediated IC GN is associated with activation of the MPO-H2O2-halide system, which may participate in mediating glomerular injury.

Mechanism of prostaglandin E2 inhibition of acute changes in vascular permeability

In order to determine the mechanism of antiinflammatory activity, prostaglandin E2 (PGE2) or diluent was administered to rats 2 h prior to intradermal injections of various mediators of inflammatory vascular permeability changes. Vascular permeability was measured as the accumulation of [125I]rat serum albumin at the site of mediator injunction. PGE2 at 500 micrograms significantly inhibited protein leakage produced by histamine, platelet activating factor, zymosan, and zymosan-activated plasma. Pretreatment with PGE2 had no effect on protein leakage induced by injection of lysosomal enzymes, glucose oxidase, or xanthine oxidase. The accumulation of polymorphonuclear leukocytes (PMNs) at the site of injection of zymosan or zymosan-activated plasma was not altered by PGE2 administration. In separate experiments, the ability of PGE2 to alter phagocytosis and oxygen radical production by PMN was examined. PGE2 significantly inhibited phagocytosis at 2 h, but this returned to normal by 6 h. Production of hydrogen peroxide by PMN was not affected by PGE2. These results suggest that PGE2 prevents acute changes in vascular protein leakage by preventing endothelial cell contraction and by inhibiting specific PMN functions.

A role for thromboxane in complement-mediated glomerular injury

The membrane attack complex (MAC) of complement (C) has been shown to stimulate prostaglandin (PG) and thromboxane (Tx) synthesis in nucleated cells. Because glomerular epithelial cell injury and altered permeability in rat membranous nephropathy are mediated by the MAC, the authors examined whether MAC-induced proteinuria is linked to glomerular prostanoid synthesis. In kidneys containing non-nephritogenic, non-C-fixing gamma 2 sheep anti-Fx1A (planted antigen) that were perfused in vitro with C-fixing guinea pig anti-sheep IgG and a C source (fresh human plasma, 50% vol/vol in buffered bovine albumin), heavy proteinuria developed, reaching 4.27 +/- 1.20 mg/min/g at 100-120 minutes (n = 8). Cyclooxygenase blockade with 10(-4) M indomethacin (n = 6) inhibited urinary PGE2 excretion (569 +/- 47 to 124 +/- 18 pg/min/g, P less than 0.001) and lowered proteinuria (1.06 +/- 0.42 mg/min/g, P less than 0.001). Reduced protein excretion (0.88 +/- 0.12 mg/min/g, n = 6, P less than 0.001) also occurred with inhibition of Tx synthetase by OKY-046, 10(-4) M, a dose that was shown in separate perfusions to inhibit urinary TxB2 excretion by greater than 85%. Control kidneys, without planted antigen and perfused with anti-sheep IgG and plasma, excreted 0.30 +/- 0.05 mg protein/min/g (n = 6). Because inulin clearance was reduced by indomethacin, renal hemodynamic factors may have contributed to the reduction in proteinuria observed with this drug. However, insulin clearance was not significantly affected by OKY-046, implying that inhibition of Tx synthetase reduced proteinuria independently of changes in renal hemodynamics. Thus, proteinuria in rat membranous nephropathy is due to MAC-dependent glomerular epithelial injury and is mediated, in part, by Tx.

Initial events in the formation of immune deposits in passive Heymann nephritis. gp330-anti-gp330 immune complexes form in epithelial coated pits and rapidly become attached to the glomerular basement membrane

The nephritogenic antigen of Heymann's nephritis (HN), gp330, was previously demonstrated (4-9) to be a resident glycoprotein of coated pits in the glomerular and proximal tubule epithelium of rats, and anti-gp330 IgG given intravenously was found to form IDs in glomeruli (passive HN). The purpose of this study was to investigate the detailed events that occur in the formation of IDs in passive HN. HN was induced by the injection of either 125I-labeled or unlabeled anti-gp330 IgG. At various times after injection (15 min to 8 d) the kidneys of some of the injected rats were fixed by perfusion, and the distribution of the rabbit IgG was determined by immunofluorescence and by immunoelectron microscopy. Glomeruli were isolated from the kidneys of injected rats and used for isolation of GBM fractions or for elution of the bound IgG. At 15 min to 1 h after injection, the rabbit IgG was localized by immunocytochemistry exclusively in coated pits along the podocyte plasmalemma facing the GBM. By 1-8 d, anti-gp330 IgG was detected in larger electron-dense IDs often located under the slit diaphragms. Serial sectioning revealed that each of the IDs maintained contact with a coated pit at some level. When GBMs isolated from rats given radiolabeled anti-gp330 IgG were examined by electron microscopy, the IDs were found to remain attached to the GBMs as early as 15 min after injection and coisolated with them at all time points. By double-immunolabeling of the isolated GBMs with two sizes of gold particles, both the antigen (gp330) and the anti-gp330 IgG could be demonstrated in IDs at all time points. When the amount of radiolabeled anti-gp330 bound to GBM fractions was compared with that of isolated glomeruli, it was found that 20% of the radiolabel remained bound to the purified GBMs at 15 min after injection, and 90% at 3 d. The bound IgG was released only by treatments that disrupt antibody-antigen complexes (high and low pH), but not by the other treatments we tried (detergent, high salt, heparinase, or collagenase digestion). When the IgG bound to glomeruli was eluted with acid citrate buffer 3 d after injection, it was found to specifically immunoprecipitate only gp330 from detergent-solubilized 125I-labeled kidney microvillar vesicles. By isoelectric focusing the eluate was found to be enriched in IgGs with acidic isoelectric points.(ABSTRACT TRUNCATED AT 400 WORDS)

An evaluation of the development of experimental membranous nephropathy

Heymann nephritis is a rat model of glomerulonephritis with morphologic manifestations of human membranous nephropathy. This model is generated by immunizing rats with Fx1A antigen. Passive Heymann's nephritis (PHN) can be produced by the administration of anti-Fx1A antibody (anti-Fx1A Ab) (with abnormal proteinuria appearing in 5 days). Studies were designed to examine the evolution of temporal changes in protein excretion, the glomerular ultrafiltration coefficient (LpA) and morphology of glomerular capillary three and five days after induction of PHN. Glomerular hemodynamic evaluation by micropuncture in euvolemic rats with PHN revealed normal values for nephron filtration rate (SNGFR), LpA and the glomerular hydrostatic pressure gradient (delta P) at day three, but by day five the whole kidney GFR and SNGFR were decreased, delta P increased and LpA significantly reduced. Glomerular binding of anti-Fx1A Ab increased from 38 micrograms/7.6 X 10(4) glomeruli on day three to 52 micrograms on day five. Immune complex deposits evaluated by immunofluorescence and electron microscopy appeared larger and were better defined on day five than on day three. Epithelial foot process fusion was more extensive on day five than day three. The onset of increased proteinuria correlated temporally with a reduction in LpA on day five, which in turn correlated with increased antibody binding, immune deposit accumulation and fusion of epithelial cell foot processes.

Enhanced glomerular prostaglandin formation in experimental membranous nephropathy

To determine whether the induction of immune-mediated glomerular injury influences the formation of cyclooxygenase products by glomerular cells, we determined prostaglandin E2 (PGE2) and thromboxane B2 (TXB2) (as the stable metabolite of TXA2) formation in isolated glomeruli of rats with passive Heymann nephritis (PHN). PHN is a model of membranous nephropathy mediated by antibody and complement independent of inflammatory cells. Five days following induction of PHN by injection of heterologous antibody to rat proximal tubular brush border antigen (Fx1A) rats developed proteinuria 36.5 +/- 34 (controls 3.8 +/- 1 mg/day). Treatment with cobra venom factor, which depleted complement C3 levels to less than 10% of baseline, prevented the development of proteinuria (6.9 +/- 2 mg/day). The development of subepithelial, glomerular immune-complex deposits and proteinuria was associated with a significant stimulation of glomerular PGE2 (87%) and TXB2 (183%) formation. This increment in glomerular prostanoid biosynthesis was significantly inhibited (PGE2 increased 22%, TXB2 increased 75%) in animals that were complement depleted with cobra venom factor. Cobra venom factor had no effect on glomerular prostanoid formation in normal rats. In additional experiments we tested the hypothesis that TXA2 may contribute to mediation of proteinuria in PHN. We utilized a thromboxane synthetase inhibitor UK38485. UK38485 reduced glomerular TXB2 formation by 80% without influencing glomerular deposition of 125I-labeled antibody, and did not alter levels of urine protein excretion in rats with PHN (control 42 +/- 21, UK 38485, 39 +/- 24 mg/day, P greater than 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)

Proteolytic enzyme treatment reduces glomerular immune deposits and proteinuria in passive Heymann nephritis

We investigated the effect of proteolytic enzyme treatment on the course of passive Heymann nephritis (PHN). PHN was induced by intravenous injection of Heymann antibody into Sprague Dawley rats. Protease-treated rats received intraperitoneal chymopapain and subtilisin. In rats given subnephritogenic doses of Heymann antibody (5 or 10 mg, insufficient to cause proteinuria), glomerular immune deposits were assessed by immunofluorescence and electron microscopy. In rats given 5 mg Heymann antibody and treated with protease in the heterologous phase of the disease (days 1-7), fewer animals were positive for rabbit IgG and rat IgG, as determined by immunofluorescence on day 12, compared with controls (p less than 0.01). Rats given 10 mg Heymann antibody and treated on days 1-5 were less frequently positive for rabbit IgG on day 5 than controls (p less than 0.05). When treatment was given on days 6-12 (autologous phase), fewer rats had glomerular rabbit and rat IgG compared with controls (p less than 0.025). Protease treatment of rats given nephritogenic doses of Heymann antibody (greater than or equal to 40 mg, causing proteinuria) did not result in significant differences in immunofluorescence deposits. However, protease treatment significantly reduced the number of electron dense deposits at all doses of antibody (p less than 0.01). Furthermore, rats given 60 mg Heymann antibody followed by enzyme treatment in the heterologous phase (days 1-7) or throughout the autologous phase (days 6-18) had significantly reduced protein excretion during the autologous phase compared with control rats (p less than 0.05). After onset of significant proteinuria on day 15 in rats given 40 mg Heymann antibody and treated from day 15 until day 25, there was significantly less (p less than 0.05) proteinuria on days 21-22 and 24-25 than in control rats; thus, enzymes could reverse proteinuria. In normal rats, administration of proteases did not have significant effects on urinary protein excretion, serum creatinine, or renal morphology, nor did protease affect anti-rabbit IgG antibody production in rats injected with Heymann antibody. The overall results indicate that proteolytic enzyme treatment can prevent or remove glomerular immune deposits and can prevent or reverse proteinuria.

Immune complex induced glomerular lesions in C5 sufficient and deficient mice

The role in the pathogenesis of immune complex-mediated glomerulonephritis of C5 or some terminal complement component dependent upon C5 for activation was explored in a congenic strain of C5 sufficient (NSN) and C5 deficient (OSN) mice. When these mice were given daily injections of heterologous protein, horse apoferritin (HAF), there were profound differences between the strains in the development of glomerulonephritis and renal dysfunction. When NSN and OSN mice produced low levels of anti-HAF, NSN mice developed extensive glomerular deposits of HAF and immune reactants and a mild proliferative glomerulonephritis. In contrast, comparable OSN mice developed only trace mesangial localization of HAF and no glomerular lesions by light microscopy. When NSN and OSN mice produced high levels of anti-HAF, both strains had equivalent glomerular immune deposits; however, NSN mice developed a severe necrotizing and crescentic glomerulonephritis, while OSN mice had much less glomerular injury. Compared to OSN mice, these NSN mice also had much more severe tubulointerstitial injury, and significantly higher serum creatinine levels. Thus, in this experimental model, the absence of C5 resulted in reduced glomerular immune complex localization when there were small amounts of circulating immune reactants; and in markedly reduced glomerular leukocyte influx, necrosis and crescent formation, when large amounts of immune reactants have localized in glomeruli. These effects could be mediated by C5 (such as C5a) or by some terminal complement component(s) dependent upon C5 for activation.

Effects of the anaphylatoxin, C5a, on renal and glomerular hemodynamics in the rat

The effects of intrarenal infusion of the complement-derived anaphylatoxin, C5a, upon glomerular hemodynamics were examined in the Munich-Wistar rat, a strain with glomeruli on the kidney surface. Human C5a (1.5 micrograms/min) or vehicle was infused into the left renal artery for 12 min, and glomerular capillary (PG) and Bowman's space pressures, nephron plasma flow (SNPF) afferent and efferent arteriolar protein concentrations, nephron filtration rate (SNGFR) and the glomerular ultrafiltration coefficient (LpA) determined. Human C5a infusion resulted in a reduction in SNPF due to increased efferent arteriolar resistance, and PG increased which maintained SNGFR constant. LpA was numerically lower but not significantly decreased. Infusion of porcine C5ades Arg decreased glomerular filtration rate and renal blood flow. No polymorphonuclear leukocytes were observed within glomerular capillaries of C5a infused rats, and rat leukocytes did not exhibit receptors for human C5a infused. Renal artery infusion of either human C5a or porcine C5a resulted in renal hemodynamic alterations and, as documented for human C5a, effects of C5a upon renal vascular resistance can be added to the known effect of C5a on the polymorphonuclear leukocyte.

Complement-induced glomerular epithelial cell injury. Role of the membrane attack complex in rat membranous nephropathy

In passive Heymann nephritis (PHN) in rats, antibody (anti-Fx1A) reacts in situ with a glomerular epithelial antigen and induces complement (C)-mediated cell-independent proteinuria. To assess the role of the membrane attack complex (MAC), we determined the need for C8 in the pathogenesis of proteinuria in an autologous-phase model of PHN. Isolated rat kidneys, containing nonnephritogenic, non-C-fixing gamma 2 sheep anti-Fx1A (planted antigen), when perfused in vitro with C-fixing guinea pig anti-sheep IgG and a source of C (fresh human plasma 50% vol/vol in buffer containing bovine serum albumin), developed marked proteinuria after 20 min (0.58 +/- 0.08 mg/min X g, n = 8) that increased further to 3.20 +/- 0.93 mg/min X g after 80 min. In contrast, identical kidneys perfused with antibody and heat-inactivated or C8-deficient human plasma and normal kidneys perfused with antibody and fresh plasma excreted only 0.27 +/- 0.03 (n = 6), 0.27 +/- 0.04 (n = 5), and 0.40 +/- 0.05 mg/min X g (n = 6) after 20 min, and 0.13 +/- 0.02, 0.22 +/- 0.03, and 0.32 +/- 0.05 mg/min X g after 80 min, respectively. When C8-deficient plasma was reconstituted with sources of C8 (n = 3), proteinuria was restored to the level observed with fresh normal plasma. Differences in protein excretion could not be explained by quantitative differences in glomerular antigen or antibody content. Extensive ultrastructural damage to glomerular visceral epithelial cells was exclusively seen in antigen-containing kidneys perfused with antibody and C8-replete plasma. Thus, glomerular injury in this model results from an antigen-specific, antibody-directed, C8-dependent reaction involving assembly of the MAC. The ultrastructural findings argue in favor of MAC-induced cytotoxicity of the glomerular visceral epithelial cells.

Charge selective properties of the glomerular capillary wall influence antibody binding in rat membranous nephropathy

IgG antibodies, eluted from kidneys of rats with Heymann nephritis were separated into cationic and anionic fractions, labeled with 125I and 131I, respectively, mixed in equal amounts, and then injected in incremental doses into 10 rats. Glomerular antibody binding was highly correlated with blood concentration of antibody at 24 hr, however, significantly more cationic antibody bound to glomeruli than did anionic antibody at all blood levels studied. The differences were not due to greater antibody content and/or avidity of the cationic preparation, as measured by binding to isolated glomeruli in vitro. These studies demonstrate the influence of glomerular permselectivity and antibody charge on subepithelial antibody deposition.

Complement membrane attack complex stimulates production of reactive oxygen metabolites by cultured rat mesangial cells

To explore possible mechanisms by which complement membrane attack complexes (MAC) that are deposited in the glomerular mesangium might be pathogenic, we stimulated rat glomerular mesangial cells grown in vitro with nascent MACs formed from the purified human complement components C5b6 and normal human serum and measured production of superoxide ion (O2-) and hydrogen peroxide (H2O2). Mesangial cells incubated with C5b6 + serum, which results in cell membrane interaction with the MAC, produce 0.9 +/- 0.15 nmol O2-/10(5) cells per 30 min, which was significantly greater than the amount produced by cells incubated with C5b6 alone, serum alone, or decayed MACs that can no longer interact with the cell membrane (0.3 +/- 0.2, 0.4 +/- 0.1, 0.3 +/- 0.2 nmol O2-/10(5) cells per 30 min, respectively; P less than 0.02). Production of O2- after stimulation with MACs increased during the first 20 min of incubation but then plateaued. Cells exposed to decayed MACs produced small amounts of O2-, which did not increase from 20 to 60 min. Production of H2O2 was also observed after stimulation with MACs, and continued to increase during 60 min of incubation (1.22 +/- 0.16 nmol H2O2/10(5) cells per 60 min), whereas H2O2 production could not be detected after exposure to decayed MACs. Cell viability was not adversely affected by exposure to nascent MACs as determined by trypan blue exclusion or chromium-51 release. These results demonstrate that glomerular mesangial cell membrane interaction with the MAC stimulates the production of the toxic oxygen metabolites O- and H2O2. Activation of the terminal complement pathway by mesangial immune deposits in vivo might lead to tissue injury by stimulation of local production of toxic oxygen-free radicals.

Antibody, antigen, and glomerular capillary wall charge interactions: influence of antigen location on in situ immune complex formation

These studies examined the charge interactions between the glomerular capillary wall, antibody and antigen at different sites in the glomerulus. Sheep IgG was separated into differently charged subclasses and immunologically placed in one of two glomerular locations (subepithelial or subendothelial) to serve as planted antigen. Single kidneys with planted antigen were transplanted into uninephrectomized recipients that received affinity-purified, cationic and anionic rat anti-sheep IgG labelled with 125I and 131I, respectively. Glomerular bound antibody was determined and corrected for antibody delivery. Specificity of antibody binding was confirmed by comparison of kidneys with or without planted antigen. The results indicate that the influence of charge on glomerular antibody binding depends on the site of the antigen. When antigen was planted in the subepithelial space, significantly more (15 to 25%) cationic than anionic antibody bound despite the fact that the antigen was cationic. Conversely, when the antigen was planted subendothelially, significantly more anionic (13 to 22%) antibody bound when the antigen was cationic, and significantly more cationic (7 to 16%) antibody bound when the antigen was anionic. Thus, the negatively-charged glomerular filtration barrier retards the permeation of anionic antibodies that complex with antigens located in the subepithelial space, but antigen-antibody charge interactions appear to predominate when the antigen is more proximally located.

Tubular antigen-associated renal disease in New Zealand white rabbits

Rabbits immunized with autologous renal tubular antigen (Fx1A) developed tubulointerstitial nephritis whereas sheep anti-Fx1A antibody administered i.v. produced glomerulonephritis (GN). This lesion showed heavy granular glomerular deposition of immunoglobulin and subepithelial electron dense deposits, early proteinuria, leucocyte independence and a temporal pattern of quantitated glomerular antibody binding distinct from that reported to occur in passive Heymann's nephritis in rats. Isoelectric focusing followed by immunoblotting of deoxycholate-soluble Fx1A antigens with the heterologous and autologous antibodies, indicated species differences in epitope recognition which could account for dissociation between the two models.

Superimposed nephritis: a separate entity among glomerular diseases?

The concomitant occurrence of two glomerular diseases in the same patient was diagnosed in seven out of 105 patients undergoing renal biopsy for suspected glomerulopathy. The most frequently associated disease was a membranous type glomerulopathy. The follow-up was characterized by a rapid deterioration of renal function and two patients were required to start a chronic hemodialysis program soon after the diagnosis. It is suggested that the observed coexisting patterns of glomerular injury do not occur on the basis of chance alone and should be considered as a separate entity in glomerular pathology. In all cases, clinical and pathologic findings were strongly suggestive for two consecutive distinct pathologic processes, thus justifying the use of the term superimposed nephritis. It is reasonable to assume that the mechanisms responsible for glomerular damage and for the evolution of the disease in superimposed nephritis are different from those regulating the corresponding glomerulonephritis when occurring alone. The high prevalence of membranous pattern in superimposed nephritis indicates that pre-existing glomerular alterations might favor an immune reaction in the subepithelial space.

Complement membrane attack complex stimulates production of reactive oxygen metabolites by cultured rat mesangial cells

To explore possible mechanisms by which complement membrane attack complexes (MAC) that are deposited in the glomerular mesangium might be pathogenic, we stimulated rat glomerular mesangial cells grown in vitro with nascent MACs formed from the purified human complement components C5b6 and normal human serum and measured production of superoxide ion (O2-) and hydrogen peroxide (H2O2). Mesangial cells incubated with C5b6 + serum, which results in cell membrane interaction with the MAC, produce 0.9 +/- 0.15 nmol O2-/10(5) cells per 30 min, which was significantly greater than the amount produced by cells incubated with C5b6 alone, serum alone, or decayed MACs that can no longer interact with the cell membrane (0.3 +/- 0.2, 0.4 +/- 0.1, 0.3 +/- 0.2 nmol O2-/10(5) cells per 30 min, respectively; P less than 0.02). Production of O2- after stimulation with MACs increased during the first 20 min of incubation but then plateaued. Cells exposed to decayed MACs produced small amounts of O2-, which did not increase from 20 to 60 min. Production of H2O2 was also observed after stimulation with MACs, and continued to increase during 60 min of incubation (1.22 +/- 0.16 nmol H2O2/10(5) cells per 60 min), whereas H2O2 production could not be detected after exposure to decayed MACs. Cell viability was not adversely affected by exposure to nascent MACs as determined by trypan blue exclusion or chromium-51 release. These results demonstrate that glomerular mesangial cell membrane interaction with the MAC stimulates the production of the toxic oxygen metabolites O- and H2O2. Activation of the terminal complement pathway by mesangial immune deposits in vivo might lead to tissue injury by stimulation of local production of toxic oxygen-free radicals.

Nature of the glomerular capillary injury in human membranous glomerulopathy

A differential solute clearance technique was used to evaluate glomerular capillary wall function in 20 patients with membranous glomerulopathy and massive proteinuria. The clearance of inulin, the filtration fraction, and the fractional clearance of uncharged dextrans of a radius of 28-48 A were depressed significantly below control values in 20 healthy volunteers (P less than 0.01). In contrast, the fractional clearance of dextrans of radius greater than 50 A was elevated markedly. A theoretical model of solute transport that depicts the major portion of the glomerular capillary wall as an isoporous membrane and the minor portion as a nondiscriminatory shunt pathway revealed the calculated glomerular ultrafiltration coefficient to be five times lower and mean pore radius of the major membrane component to be 4 A smaller than control values. However, the fraction of filtrate volume permeating the shunt pathway was three- to fourfold above control values and correlated strongly in individual patients with the fractional clearance of albumin (r = 0.76) and of IgG (r = 0.80). Lowering renal plasma flow by 24% during indomethacin therapy in seven patients resulted in a 74% reduction in proteinuria accompanied by a corresponding diminution of filtrate formed through the shunt pathway. Morphometric analysis of glomerular ultrastructure revealed the magnitude of depression of the glomerular filtration rate and of urinary protein leakage to be related strongly to changes in the epithelial layer of the glomerular capillary wall, but not to the density of subepithelial immune deposits. We conclude that glomerular capillaries in membranous glomerulopathy are characterized by a loss of ultrafiltration capacity and of barrier size-selectivity, and that subepithelial immune deposits do not provide a structural basis for these functional alterations.

Ultrastructural localization by monoclonal antibodies of brush border antigens expressed by glomeruli. I. Renal distribution

The authors have previously reported the production of monoclonal anti-brush border antibodies defining two glycoprotein (gp) antigens of 90 and 330 kd which are also expressed on glomerular cells and may thus be of significance for in situ formation of immune complexes. The 330-kd antigen is involved in Heymann's nephritis, whereas the 90-kd antigen induces transient glomerular immune deposits. In this report, the authors describe the renal ultrastructural localization of the two antigens, which are both detectable on brush border and glomerular epithelial cells but assume contrasting patterns: gp 90 is diffusely present on the cell membranes, whereas gp 330 is concentrated in the intermicrovillar region and within the coated pits. Their localization within proximal tubular cells, on the membrane of intracytoplasmic vesicles, suggests that they are involved in endocytosis. In addition, gp 90 is expressed on endothelial cells lining glomerular capillaries. These observations may help us to understand the formation and different kinetics of immune glomerular deposits.

An arthritogenic lymphokine in the rat

A type II collagen-specific arthritogenic lymphokine has been identified in the rat. Arthritogenic factor (AF) is a 65 kD protein generated in vitro by T cells from rats with collagen arthritis, and it induces an erosive, proliferative synovitis when injected into the knee joint of syngeneic naive recipients. Complement does not appear to be required. These data identify a potential T cell-mediated effector mechanism in this model, and suggest that AF may function in other inflammatory synovial diseases.

Pathophysiology of chronic tubulo-interstitial disease in rats. Interactions of dietary acid load, ammonia, and complement component C3

The human end-stage kidney and its experimental analogue, the remnant kidney in the rat, exhibit widespread tubulo-interstitial disease. We investigated whether the pathogenesis of such tubulo-interstitial injury is dependent upon adaptive changes in tubular function and, in particular, in ammonia production when renal mass is reduced. Dietary acid load was reduced in 1 3/4-nephrectomized rats by dietary supplementation with sodium bicarbonate (NaHCO3), while control rats, paired for serum creatinine after 1 3/4 nephrectomy, were supplemented with equimolar sodium chloride. After 4-6 wk, NaHCO3-supplemented rats demonstrated less impairment of tubular function as measured by urinary excretory rates for total protein and low molecular weight protein and higher transport maximum for para-aminohippurate per unit glomerular filtration rate, less histologic evidence of tubulo-interstitial damage, less deposition of complement components C3 and C5b-9, and a lower renal vein total ammonia concentration. Such differences in tubular function could not be accounted for simply on the basis of systemic alkalinization, and differences in tubular injury could not be ascribed to differences in glomerular function. Because nitrogen nucleophiles such as ammonia react with C3 to form a convertase for the alternative complement pathway, and because increased tissue levels of ammonia are associated with increased tubulo-interstitial injury, we propose that augmented intrarenal levels of ammonia are injurious because of activation of the alternative complement pathway. Chemotactic and cytolytic complement components are thereby generated, leading to tubulo-interstitial inflammation. Thus, alkali supplementation reduces chronic tubulo-interstitial disease in the remnant kidney of the rat, and we propose that this results, at least in part, from reduction in cortical ammonia and its interaction with the alternative complement pathway.