Procedure for immunocytochemical detection of P16INK4A antigen in thin-layer, liquid-based specimens

Abstract

Objective: To develop a procedure for the immunocytochemical detection of P16INK4A in ThinPrep specimens.

Study design: Archived ThinPrep, liquid-based cervical/endocervical cytology specimens (Cytyc Corp., Boxborough, Massachusetts, U.S.A.) diagnosed as LSIL, HSIL and WNL were resampled and fixed in 95% ethanol for at least three days. Rehydration and endogenous peroxidase blocking of both ThinPreps and formalin-fixed, paraffin-embedded tissues were accomplished on a Leica Autostainer (Leica, Deerfield, Illinois, U.S.A.). Microwave antigen retrieval with CitraPlus (Biogenex, San Ramon, California, U.S.A.) was performed using a Panasonic microwave oven (Matsushita Cooking Appliances, Franklin Park, Illinois, U.S.A.) on the high setting twice for five minutes each. After cooling for 20 minutes and undergoing a buffer rinse, the slides were placed in a Dako autostainer (Dako-USA, Carpinteria, California, U.S.A.). The P16INK4A primary antibody, clone E6H4 (MTM Laboratories, Heidelberg, Germany) was diluted 1:200 in antibody diluent buffer. Detection was accomplished with a mouse non-avidin-biotin EnVision+ polymer (Dako). The expression of P16INK4A in ThinPreps and corresponding biopsies were scored by two pathologists. A ThinPrep case was scored as positive if it contained > 10 abnormal cells with nuclear and cytoplasmic immunocytochemical staining. Corresponding biopsies were scored as exhibiting negative, sporadic, focal or diffuse staining, as described by Klaes et al, Overexpression of P16INK4A as specific marker for dysplastic and neoplastic epithelial cells of the cervix uteri (Int J Cancer 2001;92:276-284).

Results: The P16INK4A antibody assay was positive in 14 of 19 (73.68%) LSIL ThinPrep cases and in 25 of 26 (96.15%) HSIL ThinPrep cases. Thirty-eight of the 39 (97.44%) biopsies corresponding to the positively stained ThinPreps also were positive, with a staining score of at least focal positivity in the dysplastic regions. The P16INK4A antibody assay was negative in 5 of 19 (26.32%) LSIL ThinPrep cases and negative in 1 of 26 (3.85%) HSIL ThinPrep cases. The six biopsies corresponding to the negative ThinPreps were similarly negative. The two cytologic specimens diagnosed as WNL were negative for P16INK4A, as were two tissue control cases with benign diagnoses. Nondysplastic squamous epithelium, identified in 17 biopsy cases, did not stain, nor did nondysplastic squamous cells identified in ThinPrep cases. Sporadic staining of bacteria, inflammatory cells and occasional endocervical glandular cells was identified.

Conclusion: P16INK4A expression in ThinPrep specimens correlates with tissue expression of P16INK4A, as implemented in the above protocol. P16INK4A may thus serve as a surrogate marker in gynecologic cytology for high-risk HPV infection and for the development of cervical neoplasia.