*2.4. Chemical and Radiation-Induced Models*

These animal models are valuable tools to study mechanisms underlying human carcinogenesis. From that perspective, these murine models play an essential role in the interpretation of epidemiological observations: to assess the natural history, mechanism, and modifying factors of cancer development. Furthermore, these models allow evaluating the growth and histologic characteristics of the primary tumor and metastases. Nonetheless, they show high variability in terms of the time of onset and prevalence, as well as a long latency in metastatization; thus, influencing the number of animals required for a study [18,19].

Chemical agents or physical agonists and radiation, as well as the administration of known carcinogenic substances, have been used to induce thyroid cancer in mice. Mainly these are goitrogens that decrease thyroid hormone levels, resulting in alterations that then induce thyroid tumorigenesis. Furthermore, it is known that ionizing radiation is a determinant risk factor for thyroid cancer, but it is difficult to obtain a systematic and rapid model [63,64].

Radiation-induced mouse models and chemical carcinogenesis have been developed for the study of breast cancer. Through these models, the natural development of breast cancer linked to exposure to carcinogens has been evaluated. Radiation-induced models are relevant since radiation exposure is one of the epidemiologically-proven etiological factors of human breast cancer [65]. Moreover, chemical-induced breast cancer can be induced in some strains of rats by administering dimethylbenzanthracene and *N*-methyl-*N*-nitrosourea (MNU) [19,66].

Prostate cancer rodent models are rarely used. In particular, the most common is the chemically-induced, following the administration of methyl nitrosourea and testosterone in rats; these models infrequently recapitulate metastatization, and only will to the lymph nodes and lungs [66].

Chemically-induced colorectal cancer models were developed in mice following exposure to carcinogens, such as dimethylhydrazine, *N*-methyl-*N*-nitro-*N*-nitrosoguanidine, and MNU. The incidence depends on the dosage, duration, and frequency of administration; the route of administration; i.e., oral, subcutaneous, or intrarectal; and the timing of administration. These models are useful for studying the influence of diet on tumor development. However, the occurrence of cancer is low, making the use of a large number of animals necessary, and the metastatic phase is prolonged and infrequent [67].

#### *2.5. Metastatic Models*

The development of animal models of metastases is useful for the evaluation of new therapies that can prevent the metastatic phase or arrest the growth of metastasis already formed, and to identify new molecular targets. There are numerous approaches for the development of such models, including experimental metastases, spontaneous metastases, and the GEMMs' metastatic approach models [60,68]. Experimental metastasis models are established by direct injection of the selected cell line into the bloodstream through the tail vein or via intracardiac injection. Such an approach allows the diffusion of metastases to various organs, including the lungs, bones, brain, liver, and lymph nodes. The time-course is generally short, and the biology of metastases results is reproducible, but the injection site influences the target organ, and above all, this approach bypasses the early phases of the metastatic cascade [68]. The spontaneous metastasis model was derived from the orthotopic model, and more closely resembles the human pathology, providing the cascade of metastatic processes with adequate reproducibility [68,69]. The GEMMs' metastasis models show different degrees of penetrance and latency, and therefore, allow evaluating the metastatic process in a heterogeneous genetic background. Indeed, the activation or the loss of a gene may not replicate completely natural metastasis. Besides, the penetrance for the metastatic process is often low. Finally, long latency times and difficult recognition of metastases have been reported [68].

In thyroid cancer research, all the models above of metastatization are used. The orthotopic model is essential for evaluating the molecular mechanisms preceding metastatization. Often, the local invasion of neck's structures by the primary tumor, i.e., large vessels, the esophagus, and the trachea with consequent respiratory insufficiency or inability to feed, do not allow enough time to study metastases. On the other hand, most transgenic mouse models develop a limited number of spontaneous lung metastases, which are only detected post-mortem at the end of the experiment. Tail vein and intracardiac injections of thyroid cancer cell lines are used, as described in [32,70,71].

The orthotopic injection of cells into the mammary glands offers a comprehensive model of the metastatic process. A robust breast cancer metastasis model is obtained using immunocompromised NSG mice that facilitate understanding the mechanism underlying breast cancer metastatization [60,72,73]. Usually, this model causes a low incidence of bone metastasis, frequent in human beings. In any case, metastatization in mice models of breast cancer is strictly linked to the cell line used. Some transgenic models of breast cancer spontaneously metastasize to the lungs, but there is a very low incidence of bone metastases, probably due to the rapid progression of primary tumors [60].

Human prostate cancer cell lines have been orthotopically implanted into mice to study the different stages of cancer progression up to metastatization. This model also reproduces bone metastases, which, however, show a low incidence. Furthermore, the so-called mouse prostate reconstitution model is considered a valid approach for the development of bone micro-metastases. This model combines human neoplastic and non-neoplastic prostate cells with the murine urogenital sinus mesenchyme, and thus, the cells are implanted under the renal capsule of an immunodeficient mouse. This method allows for exploring the genetic pathogenesis of prostate cancer, but technical complexity often limits its use [60]. Most transgenic models of prostate cancer do not develop metastases or show a very low incidence, especially of bone metastases, often due to the rapid progressions of the primary tumors [60,66].

The metastatic colon cancer model is developed by orthotopic microinjection of human colon cancer cells directly in the subserosa of the cecum, via surgical celiotomy, or the injection into the portal vein. These traditional methods reproduce metastatic colonization of the liver and lungs and facilitate the study of the metastatic cascade that follows both qualitatively and quantitatively, the natural course. A noninvasive trans-anal model has been developed, with a trans-anal rectal injection of colon cancer cells after disruption of the mucosa with irritant agents administered via enema. This model allows assessing essential aspects of the development of metastases, without the negative influence of the surgical intervention. Transgenic metastasis models of colorectal cancer offer reproducible information on tumor onset and in the early phases of the metastatic process but have a low metastatic rate and limited dissemination to the target organs [74–78].
