Using antibody directed against cyclosporine (CsA-Ab) in an avidin-biotin-complex immunoperoxidase technique on routine formalin-fixed tissue specimens, 46 renal biopsies from CsA-treated renal allograft recipients and 23 biopsies from non-CsA-treated patients were examined to identify staining patterns potentially associated with CsA nephrotoxicity (CsA-NT). All specimens were examined independently in a masked fashion by two pathologists for the intensity of (CsA-Ab) specific staining for each of the four distinct patterns identified: diffuse interstitial staining (DIS, 0-3+); fine granular staining of tubular epithelium (FGS, 0-3+); coarse granular staining of tubular epithelium (CGS, 0-3+); and dark cellular staining of mononuclear inflammatory cells (DCS, 0-3+). Based on standard clinical criteria all CsA-treated patients were categorized according to the degree of CsA nephrotoxicity (grades 1-4). Significant differences in the intensity of CsA-Ab labeling were found for three of the four immunohistologic grading patterns (DIS, FGS, DCS). For the DIS and FGS patterns, cases with clinical evidence of moderate-severe CsA-NT (grades 3 and 4) had significantly higher average staining than either the slight-mild CsA-NT cases (grade 1 and 2) or the control cases (grade 0). The DCS pattern demonstrated significantly more staining in the moderate-severe CsA-NT cases than in the controls. The sum of individual scores for the DIS, FGS, and DCS immunohistologic pattern (CsA index) was also significantly higher in the moderate-severe CsA-NT cases (3.57 +/- 0.44) than in either the controls (1.73 +/- 0.23, P less than 0.05) or the slight-mild CsA-NT cases (1.98 +/- 0.27, P less than 0.05), demonstrating a specificity of 78% for identifying moderate-severe CsA nephrotoxicity. The extent of infiltration by Leu 2 (CD8; T cytotoxic-suppressor phenotype) positive cells was not significantly different between the moderate-severe versus the slight-mild CsA-NT outcome groups. However, when CsA index values were used in combination with the intensity of Leu-2-positive cell infiltrates to predict CsA-NT (CsA-treated patients with high CsA-Ab staining as well as low levels of Leu-2-positive infiltrate) the specificity for identifying moderate-severe CsA-NT was 96%. These results suggest that CsA labeling of renal allograft biopsies may be useful for identifying CsA-NT, especially when considered with the nature of inflammatory cell infiltration.