Mycobacterium africanum: a new opportunistic pathogen in HIV infection?

Abstract

Mycobacterium africanum, a member of the Mycobacterium tuberculosis complex, is endemic to West Africa, where it causes almost half of pulmonary tuberculosis [1]. Mycobacterium africanum is similar in sequence to M. tuberculosis[2], except that it lacks certain genomic regions, which may account for the lower virulence reported in early animal studies. However, although M. tuberculosis strains with fewer deletions have been associated with reduced rates of pulmonary cavitation [3], there are no reports of differences in clinical presentation or disease course between M. tuberculosis and M. africanum. Although cases of M. africanum have sporadically been identified outside West Africa [4], no chains of transmission have been reported. In Cameroon, M. africanum was identified as causing 56% of pulmonary tuberculosis three decades ago, dropping to 9% in 2001; a new strain of M. tuberculosis accounted for 40% of tuberculosis cases [1]. We hypothesized that M. africanum is a more opportunistic pathogen than M. tuberculosis in HIV-infected subjects. As part of a tuberculosis case contact study in The Gambia, a country with an HIV prevalence of approximately 2%, the mycobacterial isolates from sputum smear-positive tuberculosis index cases were classified by spoligotyping. Isolates were further tested by polymerase chain reaction for RD9 and RD702, two regions of difference that characterize M. africanum subtype I [2]. Patients were tested for HIV-1 and HIV-2 and those who tested positive were referred for HIV clinical care. Of 317 cases, 300 (95%) had an isolate available for analysis, and 286 (90%) had an interpretable spoligotype pattern; 108 (38%) of the isolates were M. africanum and 178 M. tuberculosis. Of the respective index cases 247 consented for HIV testing and 19 (7.7%) tested positive. Cases infected with M. africanum were more likely to be HIV co-infected than those with M. tuberculosis (14 versus 4%, respectively; P = 0.008); conversely 63% of HIV-infected tuberculosis patients were infected with M. africanum, versus 33% of HIV-negative tuberculosis patients. In a logistic regression model, adjusting for age, the odds ratio of HIV infection was 3.2 (95% confidence interval, 1.2–8.6; P = 0.021) when infected with M. africanum compared with M. tuberculosis. A minority of the co-infected patients turned up at the HIV clinic, which precluded analysis by CD4 cell count. Similarly, the numbers of HIV-1, HIV-2 or dually infected patients were too few to seek a specific association by HIV type. We report that M. africanum-infected patients are more likely to be HIV positive than M. tuberculosis-infected patients. The data suggest that M. africanum is more of an opportunist than M. tuberculosis. Within the phylogenetic tree of the M. tuberculosis complex, M. africanum is positioned between M. tuberculosis sensu stricto and Mycobacterium bovis, which was also found to be associated with HIV infection [5]. Mycobacterium bovis has a wider host range and is mainly transmitted from cows through contaminated dairy. It causes up to 7% of human pulmonary tuberculosis in areas where bovine tuberculosis control is poor [6]. Outbreaks of tuberculosis with multi-drug-resistant M. bovis have occurred among HIV-infected patients in nosocomial settings [7], but human-to-human transmission of this organism is an exceptional event in the absence of immune compromise [8]. In contrast, we have found that skin test positivity rates in household contacts of M. africanum cases are similar to those in contacts of M. tuberculosis cases (B. de Jong, unpublished data). Where rates of HIV are low, as was likely to be the case in Cameroon in the 1970s and 1980s, one might expect M. africanum to be out-competed by M. tuberculosis, while increasing HIV prevalence would be associated with an increase in M. tuberculosis strain diversity, including lower virulence strains such as M. africanum. These results highlight new complexities that should be considered when interpreting the changing pattern of important infectious diseases in the presence of an expanding HIV epidemic [9]. If M. africanum is taking advantage of HIV to increase its prevalence then it may well start to appear in other parts of Africa where HIV prevalence is greatest. Further studies typing M. tuberculosis complex strains in HIV cohorts, including studies of M. africanum incidence by CD4 cell count, may delineate its role as an opportunistic infection with respect to HIV in the region and elsewhere in sub-Saharan Africa. Acknowledgement Sponsorship: This work was supported by grant TW06083-01 and by MRC funding.