Novel Screening System of Virulent Strains for the Establishment of a Mycobacterium avium Complex Lung Disease Mouse Model Using Whole-Genome Sequencing

The establishment of animal models reflecting human Mycobacterium avium complex (MAC) lung disease (LD) pathology has the potential to expand our understanding of the disease pathophysiology. However, inducing sustained infection in immunocompetent mice is difficult since MAC generally shows less virulence and higher genetic variability than M. tuberculosis. To overcome this hurdle, we developed a screening system for identifying virulent MAC strains using whole-genome sequencing (WGS). We obtained nine clinical strains from Mycobacterium avium complex lung disease (MAC-LD) patients and divided them into two groups to make the mixed strain inocula for infection. Intranasal infection with the strain mixture of both groups in BALB/c mice resulted in progressive infection and extensive granuloma formation in the lungs, suggesting the existence of highly pathogenic strains in each group. We hypothesized that the change in the abundance of strain-specific single-nucleotide variants (SNVs) reflects the change in bacterial number of each strain in infected lungs. Based on this hypothesis, we quantified individual strain-specific SNVs in bacterial DNA from infected lungs. Specific SNVs for four strains were detected, suggesting the pathogenicity of these four strains. Consistent with these results, individual infection with these four strains induced a high lung bacterial burden, forming extensive peribronchial granuloma, while the other strains showed a decreased lung bacterial burden. The current method combining mixed infection and WGS accurately identified virulent strains that induced sustained infection in mice. This method will contribute to the establishment of mouse models that reflect human MAC-LD and lead to antimycobacterial drug testing. IMPORTANCE To promote research on Mycobacterium avium complex (MAC) pathogenicity, animal models reflecting human progressive MAC lung disease (MAC-LD) are needed. Because there is high genetic and virulence diversity among clinical MAC strains, choosing a suitable strain is an important process for developing a mouse model. In this study, we developed a screening system for virulent strains in mice by combining mixed infection and whole-genome sequencing analysis. This approach is designed on the hypothesis that in vivo virulence of MAC strains can be examined simultaneously by comparing changes in the abundance of strain-specific single-nucleotide variants in the mouse lungs after infection with mixed strains. The identified strains were shown to induce high bacterial burdens and cause extensive peribronchial granuloma resembling the pulmonary pathology of human MAC-LD. The current method will help researchers develop mouse models that reflect human MAC-LD and will lead to further investigation of MAC pathogenicity.