While the use of monoclonal antibodies has added a new dimension to the detection and characterization of glomerular immune deposits in human renal diseases, it carries at least one serious limitation: that non-detection of an epitope may not be equivalent to the absence of a particular molecule. This problem is highlighted in the present study in which antibodies specific for different regions of the IgA1 molecule were used to evaluate tissue from patients with systemic lupus erythematosus (SLE), IgA nephropathy (IgAN), idiopathic membranous nephropathy (IMN), and anti—glomerular basement membrane (GBM) nephritis. One of the IgA1 epitopes was masked in subepithelial and linear BM deposits but was detected in mesangial/subendothelial deposits. This epitope was also hidden following binding of IgA1 antinuclear antibodies (ANA) to nuclear substrates, in vitro. The presence of IgA1 in all loci was confirmed by reactivity of two other monoclonal anti-IgA1 antibodies. Methods. Patients with the following diseases were evaluated: SLE, ten patients, IgAN, ten; IMN, two; and anti-GBM nephritis, one. The diagnosis of SLE was made according to criteria established by the American Rheumatism Association [1], and morphologic categories were defined on the basis of light and electron microscopy as previously described [2]: Two biopsies demonstrated predominantly mesangial proliferation, four with membranous nephropathy, and four with diffuse proliferation. Patients with IgA nephropathy had mesangial deposits exclusively and no evidence of systemic disease. The two patients with IMN showed only subepithelial deposits. A patient with familial nephritis who developed anti-GBM nephritis in a transplanted kidney manifested by linear deposition of IgG and IgA along the GBM also was evaluated. Goat anti-human IgA (a chain specific) (Cappel Laboratories, Cochranville, Pennsylvania, USA) was used to evaluate IgA in renal tissue. The IgA localization was confirmed using a monoclonal antibody (murine y1K designated 1-29) against whole IgA as previously described [3, 4]. Those tissues positive for IgA were evaluated by indirect immunofluorescence for the presence of IgA1 and IgA2. Three monoclonal antibodies against IgA1 were used. 1-155-1 (y3X) has been described previously [3, 4]. DLD B4 and DLD A7 (y1K antibodies) were produced by immunizing Balb/c mice (previously made tolerant to an IgA2 myeloma) with an Fc fragment of an IgA1 myeloma protein. In a competitive radioimmunoassay DLD B4 and DLD A7 bind all intact IgA1 myeloma proteins, isolated IgA1 heavy chains, and Fc fragments of IgA1, but do not bind an IgA1 myeloma that lacks the CH3 domain (designated WAL) (Delacroix, manuscript in preparation). These findings indicate that DLD B4 and DLD A7 are directed against determinants on the CH3 or carboxyterminal portion of the IgA1 molecule. In contrast, 1-155-1 binds to the CH3 deficient IgA1 myeloma, WAL, but does not bind to either the Fc or Fab portion of IgA1 myeloma proteins. These findings suggest that 1-155-1 binds to a carbohydrate antigen found in the hinge region of IgA1. Two monoclonal antibodies against IgA2 were used. 14-3-26 (y2BK) has been described previously [3, 41. The other antibody, DLD D2 (y1K) was produced by immunizing a Balb/c mouse (previously made tolerant to an IgA1 myeloma protein) with an intact IgA2 myeloma protein. The epitopes recognized by the anti-IgA2 monoclonal antibodies have not been determined because of resistance of IgA2 to bacterial proteases and the lack of well characterized CH domain deficient IgA2 myeloma proteins. All antibodies were used as clonal supernatants. The secondary antibody was fluorescein isothiocyanate (FITC) conjugated goat anti-mouse immunoglobulin (FITC-GAMIg) (Cappel Laboratories). Controls were carried out as previously described [5]. In addition, myeloma proteins (courtesy of Dr. Michael Potter, National Institutes of Health, Bethesda, Maryland, USA) of the same subclasses as the monoclonal antibodies, were applied to tissue sections followed by FITC-GAMIg with no reactivity detected. These included: MOPC21-yiK, Y5606-y3X, and MOPCI41-y2BK. Kidney biopsy specimens were obtained percutaneously for clinical indications. The tissue samples were frozen in isopentane prechilled in liquid nitrogen, and 4-nm sections were prepared for direct and indirect immunofluorescence as previously described [3, 5]. All kidney sections were evaluated with respect to localization, pattern, and distribution as well as intensity graded on the following scale: 0, no staining; 1+, mild to moderate staining; and 2+, intense staining. Sections were evaluated independently by three observers and observations were uniformly concordant. The location of immune deposits with relationship to the GBM was determined by the use of simultaneous