The tumors arising in response to a different protocol, but using the same mouse strain and carcinogen, were used as a useful cohort for direct functional and molecular comparison. Thirteen mice were administered DMBA by intraperitoneal injection at 12–14 weeks old (0.01 µmol/gbw, 10× less than the gavage protocol); 10 mice developed mammary gland tumors within 14 months. This protocol induced many types of tumors (principally lung and liver, see [27] ), but more interestingly for this study, induced mammary tumors with a distinct histopathology (not microacinar, and no expression of basal cell associated cytokeratin5).
For serial transplantation experiments, tumors were dissociated, and 5000–10000 single cells were isografted into cleared fat pads as described [28] (link). To test the contribution of host cells to tumor outgrowth, BALB/c tumor cells were transplanted into F1 C57Bl6(EGFP): BALB/c recipients, to ensure immuno-compatibility. Tumor initiating cell frequency was calculated using limdil software (