Abstract
The influence of the primary antibody, the fixative, and the antigen unmasking technique on the method sensitivity of immunohistochemistry as a method for the identification of viral hemorrhagic septicemia (VHS) virus in paraffin-embedded specimens of naturally infected rainbow trout (Oncorhynchus mykiss) was examined. Fish (200-300 g) were collected during an outbreak of VHS. Parallel specimens from liver, spleen, kidney, and brain were fixed by immersion in 10% phosphate-buffered formalin, periodate-lysine-paraformaldehyde (PLP), Bouin's fluid, or absolute ethanol. Virus cultivation was also performed on parallel specimens, and the virus titer (TCID50/ml) was determined. Purified nucleocapsid protein (N-protein) of the virus was incorporated in an artificial antigen substrate (polymerized bovine serum albumin), fixed as described above, and embedded in paraffin wax. Microwave unmasking was performed on formalin-, PLP-, and Bouin's fluidfixed specimens. The presence of virus peptides in situ or N-protein in the artificial antigen substrates was visualized using an immunohistochemical method based on alkaline phosphatase or peroxidase and one polyclonal and five monoclonal polypeptide-specific antibodies. VHS virus was identified in situ in specimens with high virus titers (107–8 TCID50/ml) regardless of the fixative and without the need of an unmasking procedure. A pronounced masking effect was observed for the cross-linking formalin and PLP fixatives. Regardless of the primary antibodies used, there was a significantly higher epidemiologic sensitivity (the proportion of virus positive samples that tested positive by immunohistochemistry) using ethanol and Bouin's fluid compared with formalin and PLP (P < 0.05). At 105 TCID50/ml, the average sensitivity reached 0.5, and at ≥106 TCID50/ml, sensitivity was 0.9. Unmasking procedures showed a moderate effect and did not result in significantly higher epidemiologic sensitivity (P = 0.17). There was great variation for the different monoclonal antibodies/antigens and fixatives. Sensitivity studies on antigen substrates were in accordance with results of in situ studies that showed the highest sensitivity for ethanol and Bouin's fluid. Virus cultivation was more sensitive than immunohistochemistry. This study showed that the fixative and the primary antibody both influence method sensitivity and that VHS virus antigens concealed during fixation are difficult to reexpose. Immunostaining for VHS virus should be performed with monoclonal antibodies specific for the N-protein, and tissue samples should be fixed in either ethanol or Bouin's fluid. Immunohistochemistry is specific but is less sensitive than virus cultivation. Immunostaining for VHS virus can be a valuable supplement to virus cultivation during acute outbreaks of disease.