Hepatitis C virus (HCV) in lymphocyte subsets and in B lymphocytes expressing rheumatoid factor cross-reacting idiotype in type II mixed cryoglobulinaemia.

The IgMk rheumatoid factors (RF) of type II mixed cryoglobulinaemia (MC) react, in 95% of cases, with MoAbs against the cross-reactive idiotypes (CRI) Cc1 or Lc1 (corresponding to the products of the VH1 and VH4 genes). MC is closely associated with HCV infection, a virus which infects lymphocytes and may replicate in B cells. It has been suggested that HCV may induce clonal selection of B cells producing monoclonal IgMk RF in type II MC. To verify whether HCV is enriched in B cells, and in the subsets expressing Cc1 and Lc1 CRI, we studied peripheral blood lymphocytes from eight patients with MC and HCV RNA-positive sera. Seven patients had RF reacting with anti-Cc1, the other with anti-Lc1 CRI. Total lymphocytes, T cells, B cells, and Cc1+ or Lc1+, Cc1- or Lc1- B cells were purified using MoAb-coated magnetic beads. Lymphocyte subsets were then diluted to give a range of 1 x 106-1 x 103 cells and tested for HCV RNA by reverse transcriptase-polymerase chain reaction. HCV was found exclusively in B cells in seven out of eight patients. In three patients HCV was enriched in the Cc1+ cells. In one of these patients, HCV was found exclusively in Cc1+ cells, with Cc1- cells being HCV-. The data indicate that B cells from type II MC patients are almost constantly infected by HCV. In selected cases, B cell subsets expressing IgMk RF CRI are the prevalent cell type infected by HCV. Our data suggest HCV involvement in B cell dysregulation leading to cryoprecipitable IgMk RF production.

Renal manifestations associated with hepatitis C virus.

Among the several types of chronic glomerulonephritis (GN) described in association with hepatitis C virus (HCV) infection, cryoglobulinemic glomerulonephritis is by far the most frequent. It is usually associated with type II cryoglobulinemia with IgM k rheumatoid factor. It is a membranoproliferative GN, which shows some distinctive histologic features (intraglomerular monocyte infiltration, intraluminal thrombi due to massive precipitation of cryoglobulins, renal vasculitis), has a chronic course with acute recurrent episodes that can be controlled by corticosteroids more than by antiviral therapy (interferon alpha). More controversial is the association with type I non-cryoglobulinemic membranoproliferative GN, which has been found in some series from the USA and Japan but not in others. The demonstration of HCV antibodies and/or HCV-RNA in other types of chronic glomerulonephritis is usually reported in a small minority of cases suggesting the possibility of a coincidental finding more than an etiologic factor.

Electron microscopy study of genesis and dynamics of immunodeposition in IgMk-IgG cryoglobulin-induced glomerulonephritis in mice.

Cryoglobulinemic glomerulonephritis is particularly frequent in type II mixed IgMk-IgG cryoglobulinemia. The typical form is a membranoproliferative glomerulonephritis with a particular monocyte infiltration. In the most severe cases, there is occlusion of the capillary lumina by the same immunoglobulin constituents of the cryoprecipitate. While it is generally accepted that the "hyaline thrombi" are endoluminal aggregates of IgG-IgM immune complexes, probably favored by high endocapillary concentration of cryoglobulins, the modality of generation has not been studied. To study the dynamic formation of such "thrombi," we reproduced an experimental model of cryoglobulinemic glomerulonephritis in mice by injecting them twice a day for 3 days with 4 mg human IgMk-IgG cryoglobulins previously solubilized at 37 degrees C. The dynamic formation of immunodeposits was studied by immunofluorescence and electron microscopy. After 1 day, only mesangial deposits were found; after 3 days, in addition to mesangial deposition, all the capillary lumina were occluded by huge electron-dense bodies. To look for and quantify the contacts between such "thrombi" and mesangial or subendothelial deposits, we obtained serial, ultrathin, 0.5-microm sections that allowed us to reconstruct the whole glomerular tuft. Within each serial section, there was continuity between hyaline thrombi and mesangial or subendothelial deposits ranging from 80% to 85% of the capillary loops. The percentage was 100% for two adjacent serial sections. In conclusion, our data demonstrate directly for the first time that hyaline thrombi follow mesangial deposits. The high percentage of contacts between thrombi and mesangial or subendothelial deposits suggests that they result from in situ build-up of true huge endoluminal immunodeposits after saturation of the clearance capacity of the mesangium.

Characterization of nephropathy induced by immunization with high molecular weight dextran.

BACKGROUND: Injection of DEAE dextran into Lewis rats can produce proteinuria and has been reported as a model of IgA nephropathy. METHODS: Cationic diethyl aminoethyl (DEAE) dextran of molecular weight 500 kDa was injected into male Lewis rats. After a pre-immunization period of 3 weeks, the animals were divided into two groups: group 1 (n = 14) received daily i.v, injections of 3.5 mg of antigen, group 2 (n = 14) was injected with 1.5 mg three times per week for a total period of 6 weeks. I.v. treatment was initiated with gradually increasing doses of DEAE dextran in both groups for 1 week, after which the maintenance dose was reached. RESULTS: We observed the appearance of proteinuria in a nephrotic range after 5 weeks of i.v. injections in group 1 (urinary excretion: 332 +/- 83 mg/24 h, controls: 53 +/- 14 mg/24 h). In group 2, the proteinuria was almost equal to protein excretion of healthy rats of the same weight (67 +/- 20 mg/24 h). The serum and urine creatinine were normal. By light microscopy of kidney biopsies, the presence of focal and segmental proliferation of mesangial cells after 6 weeks of i.v. injections was identified. Immunohistochemistry revealed no deposition of IgA, IgM, IgG, or C3. Using anti-ED1 antibodies, there was no evidence of interstitial infiltration of monocytes/macrophages after 6 weeks of i.v. injections. Staining for proliferating cell nuclear antigen (PCNA) did not show the presence of proliferating cells either in glomeruli or in the interstitium. Staining with FITC-WGA lectin revealed focal and segmental loss of the negative charge in the capillary wall. By electron microscopy there was deposition of dextran in the basal membrane and segmental and focal damage of the podocyte foot processes. As the chemokine RANTES may be involved in glomerular injury, we examined the kidneys of proteinuric and non-proteinuric rats for the presence of RANTES. By indirect immunofluorescence only the proteinuric rats showed RANTES deposition in the mesangium. CONCLUSIONS: Injection of rats with DEAE dextran leads to dose-dependent proteinuria without deposition of immune complexes but with podocyte damage. This is associated with local expression of the chemokine RANTES which may play a role in proteinuria of glomerular disease.

Structural and functional characterization of three human immunoglobulin kappa light chains with different pathological implications.

The structural properties of three immunoglobulins light chains: kappa SCI, responsible for light chain deposition disease (Bellotti, V., Stoppini, M., Merlini, G., Zapponi, M.C., Meloni, M.L., Banfi, G. and Ferri, G. (1991) Biochim. Biophys. Acta 1097, 177-182), k INC responsible for light chain amyloidosis (Ferri, G., Stoppini, M., Iadarola, P., Bellotti, V. and Merlini, G. (1989) Biochim. Biophys. Acta 995, 103-108) and the non-pathogenic kappa MOS were analyzed by fluorescence spectroscopy and circular dichroism. Comparative evaluation of the data shows that SCI and MOS have similar stability under different conditions, while the amyloid k INC behaves as a very unstable protein. As calculated from the GdnHCl curves, the midpoint of unfolding transition was 1.35 M for SCI, 1.20 M for MOS and 0.1 M for INC. Analysis of CD spectra evidences that the three proteins conserve their conformation in the range of pH 4-8. Change in temperature at pH 4.0 produces the premature transition of INC (Tm 40 degrees C) with respect to SCI and MOS (Tm 50 degrees C). At this pH both the pathological SCI and INC light chains aggregate at a temperature of 20 degrees C lower than the normal counterpart. The specific kidney deposition of kappa SCI has been evidenced after injection of the 125I labelled light chain into mice. No deposition was detectable in the case of INC and MOS.

High binding of immunoglobulin M kappa rheumatoid factor from type II cryoglobulins to cellular fibronectin: a mechanism for induction of in situ immune complex glomerulonephritis?

In our previous experimental work we suggested that the frequent nephritogenicity of type II cryoglobulins could depend on a particular affinity of the immunoglobulin (Ig) M kappa rheumatoid factor (RF) component for mesangial matrix. Since cellular fibronectin (cFN) in the human kidney is mainly represented in glomerular mesangium, we studied the binding capacity to cFN of IgM kappa RFs from type II cryoglobulins compared with other different monoclonal and polyclonal IgM and IgM RFs. We purified 13 IGM kappa from human IgM kappa/IgG cryoglobulins, eight monoclonal IgM from patients with Waldenström's macroglobulinemia, nine polyclonal IgM from normal donors, and eight polyclonal IgM RFs from patients with rheumatoid arthritis. Purified IgM were used at the same concentration in enzyme-linked immunosorbent assay (ELISA) on cFN-coated plates. All the cryoglobulin IgM showed high specific binding to cFN while IgM from Waldenström's macroglobulinemia, normal IgM, and polyclonal IgM RFs had low or absent binding. These data were confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis of cFN followed by Western blot analysis with purified IgM. The IgM kappa binding to cFN persisted using IgM kappa monomers, and was inhibited by cFN but not by plasma FN in a specific inhibition test. Further enzyme-linked immunosorbent assay studies showed that cryoglobulin IgM kappa RFs are still able to bind IgG in a dose-dependent manner once linked to solid-phase cFN. The data suggest that the affinity of cryoglobulin IgM kappa RFs for immobilized cFN could be involved in the particular high nephritogenicity of type II cryoglobulins and might lead to in situ immune complex formation.

Type II mixed cryoglobulinaemia, hepatitis C virus infection, and glomerulonephritis.

Mixed cryoglobulins (MC) are immunoglobulins which precipitate reversibly in the cold. Type II mixed cryoglobulins are composed of a monoclonal component (usually IgMk) with rheumatoid factor (RF) activity against polyclonal IgG. In type III MC, all the components are polyclonal. The majority of MC are secondary to connective tissue diseases, infectious or lymphoproliferative disorders, hepatobiliary diseases, or immunologically mediated glomerular diseases. The aetiology of MC is not clear and cryoglobulinaemia was considered 'essential' until an association between hepatitis C virus (HCV) infection and MC was recognized. The renal pattern includes typical glomerular lesions characterized by a particular glomerular monocyte infiltration, double-contoured appearance of the glomerular basement membrane (GBM) and by the presence of intraluminal 'hyaline thrombi' due to deposition of circulating cryoglobulins. The progression of renal disease is variable: in one-third of patients remission of renal symptoms occurs, 20% of patients experienced nephritic or nephrotic flare-ups during the course of the disease. Uraemia is observed in only 10% of patients 10 years after renal disease onset, but 50% of patients had already died from cardiovascular disease, infectious liver failure, or neoplasia during those 10 years. This review analyses the pathogenic mechanisms of MC and associated GN, with particular attention to the role of HCV infection. HCV RNA is detected in most patients with MC. HCV, by infecting B cells, could trigger abnormal production of polyclonal RF in type III MC and, together with other factors, a clone selection of B cells to produce monoclonal IgMk RF in type II MC. The presence of IgMk in serum appears essential for glomerular damage to occur. Cryoglobulinaemic GN might be initiated by IgG antibody-HCV complexes binding to IgMk RF, either in situ or in circulation, nephrotoxicity being due to a particular affinity of the IgMk RF for cellular fibronectin present in the mesangial matrix. Glomerular damage can be perpetuated by the reduced effectiveness of monocytes to remove cryoglobulins.

Hepatitis C virus genotype in patients with essential mixed cryoglobulinaemia.

We studied 54 patients with essential mixed cryoglobulinaemia (EMC), (23 males, 31 females) mean age 61 years (range 28-77). Forty-one (76%) had type II cryoglobulinaemia and 13 (24%) type III. Antibodies to HCV were detectable by second-generation ELISA in 49 patients (91%) with confirmed or indeterminate RIBA results. HCV RNA was detected by RT PCR using 5' UTR nested primers; HCV genotypes 1a, 1b, 2 and 3a were identified by genotype-specific core-region nested primers. All patients (49) with antibodies to HCV in their serum were HCV-RNA positive; 27 (55.1%) had HCV subtype 1b and 21 (42.8%) type 2. In one patient the HCV genotype could not be determined. The genotype distribution was not different from that found in patients with chronic hepatitis C without cryoglobulinaemia. However, the presence of HCV subtype 1b correlated significantly with signs of chronic hepatitis and presence of peripheral neuropathy. Severity of disease tended to be worse in patients infected with HCV subtype 1b, but this was mainly due to liver disease. HCV genotypes may influence the clinical expression and, in particular, the severity of liver involvement in patients with EMC. Extent and severity of EMC disease in general may also be affected by the different HCV genotypes. These findings may have therapeutical implications, since the different HCV genotypes respond differently to interferon treatment.

Significance of hepatitis C virus genotypes and viral load in patients with "essential" mixed cryoglobulinemia.

Several studies have established a strong association between hepatitis C virus (HCV) infection and essential mixed cryoglobulinemia (EMC). However, the mechanisms by which HCV infection may result in cryoglobulinemia in some patients but not in others remain unknown. In this paper we shall summarize some of the work done in our laboratories on certain aspects of HCV in patients with EMC.

Cross-reactive idiotypes on cryoprecipitating monoclonal IgMk rheumatoid factors.

Shared idiotype specificities (CRIs) among monoclonal IgM rheumatoid factors from patients with essential mixed cryoglobulinemia have been demonstrated with the use of polyclonal antisera and have been confirmed, more recently, using monoclonal antibodies. In this paper we will summarize some of the work that has been performed in our laboratory using anti-idiotypic monoclonal antibodies. Such reagents allowed us to detect CRIs on cryoprecipitable rheumatoid factors, to detect idiotype-positive cells in bone marrow and peripheral blood, and to identify glomerular immune deposits.

Anti-mesangial and anti-endothelial cell antibodies in IgA mesangial nephropathy.

In the present study we verified by solid phase ELISA the presence of antibodies against mesangial and endothelial cell constituents in patients with IgA-GN and Schoenlein-Henoch syndrome (SH). An antigen extract was prepared by sonication of human mesangial cell (MC) monolayers between third and fifth subculture and coated at 20 micrograms/ml on microtiter plates where sera were tested by incubation for 2 h at 37 degrees C and addition of peroxidase-conjugated anti-human IgG or IgA. In comparison to 86 normal controls, increased levels of IgG anti-MC antibodies were found in 15/84 patients with IgA-GN and 4/11 with SH. IgA antibodies were always negative. Furthermore anti-endothelial cell antibodies (AECA) were sought in the same patients and controls by ELISA as previously described. Increased levels of IgG and IgA AECA were found in 25/62 and 24/46 patients respectively. A cross-inhibition test showed that preadsorbment of positive sera for both IgG anti-MC and IgG AECA on endothelial cells in culture resulted in an inhibited binding of IgG to MC. HPLC-ELISA and Western blot analysis of the MC extract showed a significant binding of IgG from ELISA-positive sera to a protein band of 25-50 kD. Similar results were obtained by Western blot analysis of an endothelial cell extract. These results suggest the identity of the antigens recognized by IgG antibodies on endothelial cells and MC in patients with IgA-GN.

Cryoglobulinemic glomerulonephritis: a membranoproliferative glomerulonephritis induced by hepatitis C virus.

Mixed cryoglobulins (MCs) are proteins that precipitate from cooled serum, and are composed of a polyclonal immunoglobulin G (IgG) bound to another immunoglobulin that acts as an anti-IgG rheumatoid factor (RF). In type II mixed cryoglobulinemia, the antiglobulin component, usually of the IgM class, is monoclonal; it is polyclonal in type III mixed cryoglobulinemia. The majority of MCs are found in patients with connective tissue diseases, infectious or lymphoproliferative disorders, hepatobiliary diseases, or immunologically mediated glomerular diseases (secondary MCs). The etiology is not clear for 30% of all MCs, and this type of cryoglobulinemia is called "essential." There is a common clinical syndrome in types II and III essential mixed cryoglobulinemia (EMC) characterized by purpura, weakness, and arthralgia. In type II EMC only, in which an IgMk is the monoclonal RF, a membranoproliferative glomerulonephritis (MPGN) occurs with some peculiar morphologic and clinical features; this is termed "cryoglobulinemic GN." Glomerulonephritis can be differentiated from idiopathic MPGN, especially in the acute stage, which is characterized by an acute nephritic syndrome, by the following findings: (1) the presence of large deposits filling the capillary lumen that sometimes are shown to have a characteristic fibrillar or crystalloid structure by electron microscopy; (2) the extent of the exudative component consequent to the frequently massive infiltration of monocytes; (3) a more diffuse and evident thickening of the glomerular basement membrane, which has a double-contoured appearance that is mainly due to the peripheral interposition of monocytes, with less evident mesangial expansion; and (4) possibly some vasculitis in small and medium-sized renal arteries without concomitant features of segmental necrotizing GN or crescentic GN.(ABSTRACT TRUNCATED AT 250 WORDS)

Glomerulonephritis induced by human IgMK-IgG cryoglobulins in mice.

BACKGROUND: Human cryoglobulinemia is sometimes associated with glomerulonephritis (GN) due to deposition of cryoglobulins (cryos). To see whether human cryos can induce GN in mice and to study time-related changes of glomerular lesions and possible factors of cryos' nephritogenicity, we developed an experimental passive model of cryoglobulinemic GN. EXPERIMENTAL DESIGN: Two cryos IgMk-IgG from 2 patients with active GN (OLD and SOR), 2 cryos IgMk-IgG (TAC and GRO) and 1 IgM lambda (CHI) from 3 patients without GN were purified, solubilized at 37 degrees C and injected intravenously into BALB/c mice, 4 mg, twice a day. To study the possible factors of cryo nephritogenicity, we analyzed: (a) the presence, amount, and size of complexed IgMk-IgG at 37 degrees C by fast flow liquid chromatography; (b) the Cc1 or Lc1 subclass of rheumatoid factors; (c) the isoelectric points of the IgMks; (d) The proportion of IgG subclasses in cryos. RESULTS: On day 1 from the beginning of intravenous injections, cryos OLD had induced mesangial deposits of human IgM, human IgG, mouse C3 and mesangial hypercellularity. On day 2, phagocytizing cells were found along with massive endoluminal and subendothelial deposits of IgM, IgG, and C3. On day 6, perivascular infiltrates of mononuclear cells were also seen. Cryos SOR induced a similar but milder form of GN. After administration of purified OLD IgMk, OLD IgG, GRO IgMk or GRO IgG, only OLD IgMk was deposited in the mesangium. Analysis of all the cryos revealed that: the amount of complexed IgMk-IgG at 37 degrees C was always less than 1% of cryos; Cc1 and Lc1 idiotypes were not related to the nephritogenicity of cryos, the isoelectric points of IgMks were 4.5 to 5.5 and IgG1 was the prevalent subclass. CONCLUSIONS: Data demonstrate that human cryos from patients with GN can induce GN in mice that resembles the corresponding human pathology. The affinity of IgMk for glomeruli and the unexpectedly small amounts of IgMk-IgG complexes at 37 degrees C suggest that there is a role of in situ binding in nephritogenicity which is independent of the isoelectric point, rheumatoid factor idiotype, or IgG subclass.

Induction of experimental SLE in naive mice by immunization with human polyclonal anti-DNA antibody carrying the 16/6 idiotype.

Recently, the induction of SLE in naive mice employing monoclonal anti-DNA antibodies (anti-DNA Ab) carrying the pathogenic idiotype 16/6 (16/6 Id) has been reported. In the current study we report on the induction of experimental SLE by polyclonal IgG anti-DNA Ab derived from a patient with active SLE and carrying the 16/6 Id. Two different experiments were conducted in which BALB/c mice were immunized in the footpads with 1 microgram/ml or 5 micrograms/ml of anti-DNA Ab. The first experiment showed the appearance in the immunized mice of high titre anti-DNA Ab together with antinuclear antibodies, alopecia and proteinuria. In the second experiment we compared, as immunizing agents, 16/6 positive anti-DNA Ab and 16/6 negative anti-tetanus toxoid antibodies (anti-TT Ab) obtained from the serum of the same patients. Our results show that only mice immunized with 16/6 positive antibodies produced anti-DNA Ab, while mice immunized with anti-TT Ab did not show any DNA-binding activity but, surprisingly, developed high titre anti-cardiolipin antibodies.

The clinical significance of IgA rheumatoid factor in idiopathic IgA mesangial nephropathy (Berger's disease).

IgA rheumatoid factor, IgA and IgG immune complexes were measured in 119 patients with IgA nephropathy. IgA rheumatoid factor was detected in 62/119 (52%) patients and in 92/265 (35%) serum samples. There was a good correlation (p less than 0.001) between the presence of IgA rheumatoid factor and the presence as well as levels of IgG immune complexes, but not between levels of IgA rheumatoid factor and other clinical or immunological parameters. However, higher levels of serum IgA were found in the subgroup of patients with constantly positive IgA rheumatoid factor. Using aggregated human IgG, we could not demonstrate antiglobulin activity in renal biopsy specimens from 36 patients. These results suggest that IgA rheumatoid factor does not play a primary role in renal damage in IgA nephropathy, but could simply reflect a response to IgG immune complexes in a disorder characterized by abnormalities of IgA production. Nevertheless, the presence of circulating IgA rheumatoid factor in a substantial proportion of patients, especially in those with features of polyclonal IgA activation, provides additional evidence for a general perturbation of IgA metabolism in this disease and could represent an antigen-specific system with which to study regulation of IgA synthesis.

Polymeric IgA and IgA rheumatoid factor decrease the capacity of serum to solubilize circulating immune complexes in patients with primary IgA nephropathy.

Primary IgA nephropathy (IgAN) is characterized by the presence of immune complexes (IC), high levels of polymeric IgA (pIgA), and IgA rheumatoid factor (RF) in the blood. The impaired capacity of serum to solubilize IC in the presence of normal values of C hemolytic activity as well as high serum levels of C3, C4, and properdin factor B have led us to analyze whether pIgA and IgA RF from patients with IgAN where capable of inhibiting the capacity of normal human serum to solubilize immune precipitates (BSA-anti-BSA) preformed at equivalence. The results showed a significant reduced mean capacity of serum from patients with IgAN to solubilize "in vitro" immune precipitates (p less than 0.001) and significant high mean levels of pIgA (p less than 0.001) and IgA RF (p less than 0.005) in the blood. Increasing amounts of pIgA inhibited solubilization of IC in the fluid phase, and inhibitory activity was also shown by the IgA RF. There were inverse correlations between pIgA and the capacity of serum to solubilize IC (r = -0.36; p less than 0.05), and between IgA RF and the complement-mediated solubilization (r = -0.57; p less than 0.001). It is suggested that pIgA and IgA RF may be responsible for the impaired complement-mediated solubilization of serum and the persistence of insoluble nephritogenic IC in the blood of patients with primary IgAN.

Identification of glomerular immune deposits in cryoglobulinemia glomerulonephritis.

To provide further evidence of the nature of intraglomerular immune deposits in essential mixed cryoglobulinemia (EMC), we used two mouse monoclonal antibodies against cross-reactive idiotypes present on monoclonal rheumatoid factors (MoRFs) from patients with type II-EMC. MoAb Cc1 reacted with 9 of 16 circulating IgMk MoRFs tested, and MOAb Lc1 with four of the remaining. Using indirect immunofluorescence and immunoperoxidase techniques, we could identify the same cross-reactive idiotype of the serum MoRF in the renal biopsy specimens from 11 of 13 patients with EMC glomerulonephritis. Kidney specimens from the three patients, whose MoRF was not recognized by MoAbs Cc1 and Lc1, were negative. Two out of 30 control renal biopsies from patients with other forms of glomerulonephritis were shown to contain idiotype (Cc1 and Lc1) positive material. Both patients had serum polyclonal RF which could account for this finding. In conclusion, our results provide direct evidence that serum cryo-MoRF participate in the formation of glomerular immune deposits and, presumably, in the pathogenesis of renal damage in EMC glomerulonephritis.

Food antigens, IgA-immune complexes and IgA mesangial nephropathy.

To investigate whether patients with IgA nephropathy have an exaggerated serum IgA response to ubiquitous food antigens we measured serum IgA antibodies to gliadin, ovalbumin, bovine serum albumin (BSA), beta-lactoglobulin and casein in 120 patients and 53 normal controls, using ELISA. No significant differences were observed between patients and controls in serum IgA antibodies against each of the antigens tested. Moreover, no correlation was found between serum IgA antibodies and IgA-immune complexes (IgA CIC). However, nine patients but no controls had an association of two or more IgA antibodies to dietary antigens. Sixty-six per cent of these patients (vs 24% in the remaining population) had IgA CIC, suggesting a possible involvement of these antibodies in the constitution of IgA CIC. Analysis of sera by HPLC revealed that both monomeric and higher molecular forms of IgA antibodies were present, the latter being coincident with the peak of IgA CIC. Preincubation of sera with serial concentrations of the specific antigen decreased significantly IgA CIC, suggesting that in this subgroup of patients IgA antibodies to food antigens (mainly BSA) are involved in the formation of IgA CIC. BSA-containing IgA CIC were in fact demonstrated by ELISA using rabbit IgG anti-BSA coated plates and peroxidase-conjugated anti-human IgA. The role of these CIC in the pathogenesis of IgA nephropathy needs to be further elucidated.