These are very old and well established mammary tumor models (most of the preclinical work with tamoxifen was done in the DMBA model - we likely wouldnt have one of the most useful drugs in breast cancer without the work in this model). If you are interested in the back story, Craig Jordan has written a nice description of the history of these type of models (J Lab Clin Med. 1987 Mar;109(3):267-77).
Most of the data in breast cancer is with DMBA or NMU administration to adult rats. In mice, priming with medroxyprogesterone acetate before giving DMBA works pretty well.
Both NMU (or MNU; same compound) and DMBA induce mostly ER+ mammary adencarcinomas, although other tumors also arise and the study design may alter their distribution (so histopathological analysis is required). The window of admin for optimum tumor incidence is about the same for both (at least in adult animals).
DMBA requires metabolic activation, whereas NMU is a direct acting carcinogen. If there is a co-treatment with the DMBA that could alter its pk/pd, it will likely also alter tumor (and type) incidence.
Most NMU-induced tumors (in rats) have a v-HA-Ras mutation; prevalence is much lower for DMBA-induced tumors and closer to the proportion of ras-overexpression seen in human breast tumors (ras mutation is rare in humans breast cancer but overexpression occurs in perhaps one-third or a little more - depending on what study you read). Depending on what you want to do, DMBA may be more useful if you want to avoid tumors potentially dominated by mutated ras signaling.
So, like much else, your choice may depend on the study goals. I've probably missed a few points but maybe something here will be helpful...Bob
From my experience NMU is more site-specific for mammary carcinogenesis induction. DMBA appplication can occasionally induce gastrointestinal tumors too.
Yes, for sure there are non mammary gland tumors with the chemical carcinogens.
Also, not all mammary tumors are adenocarcinomas, e.g., some less common mammary tumor histologies are seen with both NMU and DMBA, e.g., papillomas. Not all tumors are cancerous, e.g., benign fibroadenomas also arise.
Whichever carcinogen you choose, you should get histopathologic confirmation of those you choose to include in your estimates on your experimental endpoints, (e.g., latency, multiplicity, growth, e3tc).