**4. Animal models of tuberculosis**

A wide variety of animal models have been used to test new vaccines and drugs [15]. Mice can harbor high numbers of *M. tb* within lung tissue without showing clinical signs [16]. Mice do not cough nor form cavitary lesions, making them a poor model for transmission studies [17]. Fibrous capsules are not observed histologically, which can affect the validity of antibiotic studies, as *M. tb* would be more easily accessed by drugs in the mouse lung. In addition, because of their short life span, mice are poor models for the study of latent infection. Rat TB also showed similar pathophysiology to murine TB [31]. Guinea pigs develop robust DTH response to mycobacterial antigens and, after infection with *M. tb*, reproduce many of the aspects of human infection, such as caseous and mineralized granulomas, primary and hemato-genous pulmonary lesions, fibrous capsule formation, and dissemination [19], however, pulmonary lesions in guinea pig contain a high proportion of granulocytes, partic‐ ularly eosmophils, which are not common features of human disease [20]. The rabbit is the only common laboratory animal in which the disease closely resembles the typical chronic cavitary type found in the majority of human beings [21, 22]. Rabbits infected with *M. tb* mount a moderate DTH response and form caseous granulomas and cavitary lesions [23-25]. Rabbits, including currently available inbred strains, are relatively resistant to *M. tb*, however, requiring the inhalation of 500 to 3000 bacilli to form one grossly visible tubercle at 5 weeks postinfection [23]. Most rabbits will also overcome disease completely, with few culturable bacilli [24]. This model is useful in the study of latent or paucibacillary TB states, however, without the use of antibiotics as in the Cornell model. Rabbits do need to be experimentally immunosuppressed as they will not spontaneously reativate disease [26]. There are minimal immune reagents, however, for this model, and the larger size of rabbits makes them more costly to use. There are inbred strains of rabbits, such as the Lurie and Thorbecke rabbits, which are more suscep‐ tible to *M. tb* infection. This susceptibility has been linked to suppressed macrophage antimy‐ cobacterial activity, decreased MHC Class 2 expression, and impaired development of type 4 hypersensitivity [27]. Other animal models, such as nonhuman primates, which are susceptible to *M. tb* and full spectrum of granuloma types can be observed [28], have not been widely used. Using mycobacterial inoculation into trachea, at necropsy, all unvaccinated monkeys (*Macaca fascicularis* and *Macaca mulatta*) exhibited extensive bilateral lung pathology characterized by the presence of multiple granulomas. These granulomas exhibited conglomeration to larger caseous areas, especially in the hilar region [55].
