Why doesn't Mycobacterium tuberculosis spread in animals?

In their recent Spotlight article, entitled ‘Why doesn’t the bovine tubercle bacillus spread in humans?’ [1], Berg and Smith discussed work by Gonzalo-Asensio and colleagues who investigated mutations specific to the bovine tubercle bacillus, Mycobacterium bovis [2]. It has been known for about a decade that, of the two major forms of the tubercle bacillus, Mycobacterium tuberculosis has the more ancestral genome structure, with the M. bovis genome marked by several mutations and deletions [3,4]. In their recent work, Gonzalo-Asensio et al. used elegant molecular studies to define functional consequences of these mutations in M. bovis by recapitulating evolutionary steps between M. tuberculosis and M. bovis [2]. Given that M. tuberculosis is pandemic in humans, while M. bovis causes sporadic disease that rarely transmits among humans, there is a natural urge to link these molecular differences to a loss-of-virulence phenotype. However, this putative link overlooks three natural features of M. bovis that argue, if anything, towards a greater pathogenicity. First, setting aside humans, whose health is of the most interest to authors and readers of this journal, M. bovis is the more successful in nature, maintaining and transmitting successfully among cattle, goats, East-Asian water buffalos, deer, bison, brush-tailed possums, and last, but not least, the badger. Second, while the mutations engineered by Gonzalo-Asensio and colleagues resulted in diminished virulence as assayed in experimental infections, it must be remembered that the intermediate forms of the M. tuberculosis complex that separate M. tuberculosis from M. bovis are successfully spreading from antelope to antelope (in the case of Mycobacterium orygis [5]) and from seal to seal (in the case of Mycobacterium pinnipedii), the latter causing zoonotic infections in New World humans before M. tuberculosis came to the Americas [6]. Third, when M. tuberculosis and M. bovis have been contrasted directly in a neutral host, such as mice, guinea pigs, and rabbits, it has been M. bovis that is the more virulent of the two [7]. Indeed, long before these pathogens could be distinguished based on the genomic deletions that brand their lineages, the first-generation assay for differentiation of human and bovine bacilli was lethality in rabbits, with M. tuberculosis producing the nonlethal