Determination of accuracy of quantitative immunohistochemical results with an antigen-immobilized filter model system

To determine accuracy of quantitative immunohistochemical results, serially diluted liver cell lysates from methylcholanthrene (MC)-treated rats containing cytochrome P-450 (P-450) 1A were immobilized on nitrocellulose (NC) filters and stained by the indirect immunoperoxidase method under saturation conditions. The stained filters were processed for image analysis and the relationship between the resulting immunostaining intensity due to P-450 1A and the antigen amount immobilized on the filters was examined. The relationship examined in the filters fitted with an exponential curve. Thus, the intensity due to P-450 1A is not simply proportional to the antigen amount. Subsequently, immunostaining intensity due to P-450 1A was measured in sections from control and methylcholanthrene (MC)-treated rats by image analysis with an image processor, and the antigen content in cell lysates from control and MC-treated animals was measured immunochemically by quantitative single radial immunodiffusion. Although immunochemically measured P-450 1A content in the lysates increased markedly (26 times), average staining intensity in sections increased slightly (2-3 times) after MC injection. However, when the exponential curve obtained from the filter binding assay was used as a standard curve to convert staining intensity in sections to molar content of P-450 1A, the resulting content was compatible with the immunochemical content. The divergence between the increase in immunochemically measured P-450 1A content and that in immunostaining intensity in sections after MC injection is therefore attributed primarily to the curvilinear relationship between the intensity and the content.