Comparison of an immunoperoxidase "sandwich" staining method and western blot detection of P-glycoprotein in human cell lines and sarcomas.

The applicability of a multilayer immunoperoxidase "sandwich" method (IpS) developed by Chan14 for the amplified detection of P-glycoprotein (Pgp) was investigated. The authors examined 15 formalin-fixed cell lines, as well as formalin-fixed, paraffin-embedded sections from single biopsies of 46 sarcomas. The cell lines included sensitive and multidrug resistant sublines (KB, A2780, MCF-7, HeLa) with various relative degrees of resistance to doxorubicin (Dox). The sarcoma biopsy specimens were selected on the basis of the results obtained in Western blot (WB) detection of Pgp (22 positive and 24 negative by WB) using C219 and C494 monoclonal antibodies to Pgp. The IpS method employed C219. The least resistant cell line in which Pgp could be detected by IpS was fivefold resistant to doxorubicin, whereas Pgp was detected by WB in cells greater than twofold resistant. Cell lines having greater than fivefold resistance to Dox were positive by both IpS and WB analyses. The less resistant cell lines contained more nonreactive cells whereas the highly resistant cell lines showed more homogeneous strong membrane reactions. Among the six cell lines determined to be Pgp negative by WB analysis, no false positive immunostaining by IpS existed. One of 22 WB positive and 7 of 24 WB-negative sarcoma biopsy specimens were positive by IpS methods. Reaction varied and was always focal (a minimum of 3-5 cells, ranging up to 3-4 high power fields) indicating pronounced heterogeneous distribution of Pgp. Thus, WB can detect low average (overall) levels of Pgp in tumor samples but such low concentrations of PgP at the single cell are not detectable by IpS methods. However, IpS can discern among many Pgp-negative cells small subpopulations of immunoreactive cells, which are not detected by WB analysis due to Pgp dilution by the membrane protein of numerous Pgp negative cells. IpS and WB used together as complementary methods can provide more complete information about Pgp distribution and content, particularly in the case of heterogeneous human tumors. The IpS method is more suitable for less drastically treated (not embedded) cell line specimens than for paraffin-embedded (routine) sections. Some modification of the present IpS protocol seems necessary to increase its sensitivity and reduce the disparity with WB results.