Human tumor xenografts in nude mice are widely utilized model system for the transplantation of human surgical specimens and human established cell lines. Gene expression studies are often carried out in these model systems. With an increasing use of PCR based analyses, the extreme sensitivity of this technique poses a serious challenge with regards to the extent of contaminating host mouse cells in the human tumor xenografts. These xenografts are never free of host cell contamination. We detected mouse estrogen receptor expression in several human tumor xenografts using RT-PCR demonstrating that precaution is necessary when utilizing PCR based analyses in human tumor xenografts. A cytologically based methodology which distinguishes human versus mouse cells will be more suitable for ER expression studies using human xenograft models. Both (1) in situ hybridization using human probe and (2) immunocytochemistry using a monoclonal antibody directed against human cytokeratin have been used successfully to distinguish human cells versus host mouse cells in human xenografts in nude mice. Immunostaining of ER can then be utilized to determine the expression pattern of ER in the transplanted human cells.