Mycobacterium tuberculosis infections claim more than a million lives each year, and better treatments or vaccines are required. A crucial pathogenicity factor is translocation from phagolysosomes to the cytosol upon phagocytosis by macrophages. Translocation from the phagolysosome to the cytosol is an ESX-1-dependent process, as previously shown in vitro Here, we show that in vivo, mycobacteria also translocate to the cytosol but mainly when host immunity is compromised. We observed only low numbers of cytosolic bacilli in mice, armadillos, zebrafish, and patient material infected with M. tuberculosis, M. marinum, or M. leprae In contrast, when innate or adaptive immunity was compromised, as in severe combined immunodeficiency (SCID) or interleukin-1 receptor 1 (IL-1R1)-deficient mice, significant numbers of cytosolic M. tuberculosis bacilli were detected in the lungs of infected mice. Taken together, in vivo, translocation to the cytosol of M. tuberculosis is controlled by adaptive immune responses as well as IL-1R1-mediated signals.IMPORTANCE For decades, Mycobacterium tuberculosis has been one of the deadliest pathogens known. Despite infecting approximately one-third of the human population, no effective treatment or vaccine is available. A crucial pathogenicity factor is subcellular localization, as M. tuberculosis can translocate from phagolysosome to the cytosol in macrophages. The situation in vivo is more complicated. In this study, we establish that high-level cytosolic escape of mycobacteria can indeed occur in vivo but mainly when host resistance is compromised. The IL-1 pathway is crucial for the control of the number of cytosolic mycobacteria. The establishment that immune signals result in the clearance of cells containing cytosolic mycobacteria connects two important fields, cell biology and immunology, which is vital for the understanding of the pathology of M. tuberculosis.