For detecting human tumor cells in mouse tissue, we usually stain with a anti-human nuclei antibody. In most of the cases this is a better way to detect human cells than trying to search for a cell line specific marker.

Thank you doctor but I will be injecting the mice with human HCT116 cells and human NK cells so how I will differentiate between them if I used anti human nuclei antibody ?

I was thinking to use specific antibodies against receptors/ proteins expressed by each type of these cells..

That's of course a problem. Then fluorescent labelling of the tumor cells with e.g. GFP might be an option to detect the micrometastases by either the fluorescent signal directly, or using an anti-GFP antibody. Another markers we used for HCT-116 detection, although as an alternative CSC marker, was epithelial specific antigen (ESA), but i am not aware if it might be by any chance expressed by NK cells, even if i doubt it.