Impact of Methoxymycolic Acid Production by Mycobacterium bovis BCG Vaccines

Impact of Methoxymycolic Acid Production by Mycobacterium bovis BCG Vaccines
May 2004
Received 9 September 2003/ Returned for modification 4 December 2003/ Accepted 31 January 2004
Adam Belley,1 David Alexander,2 Tania Di Pietrantonio,1 Manon Girard,1 Joses Jones,2 Erwin Schurr,1 Jun Liu,2 David R. Sherman,3 and Marcel A. Behr1*
Infection and Immunity, May 2004, p. 2803-2809, Vol. 72, No. 5
Centre for the Study of Host Resistance, McGill University, Montreal H3G 1A4,1 Department of Medical Genetics and Microbiology, University of Toronto, Ontario M5S 1A8, Canada,2 Department of Pathobiology, University of Washington, Seattle, Washington 981953
ABSTRACT
BCG vaccines are a family of closely related daughter strains of an attenuated isolate of Mycobacterium bovis derived by in vitro passage from 1908 to 1921. During subsequent laboratory propagation of the vaccine strain until its lyophilization in 1961, BCG Pasteur underwent at least seven further genomic mutations. The impact of these mutations on the properties of the vaccine is currently unknown. One mutation, a glycine-to-aspartic acid substitution in the mmaA3 gene, occurred between 1927 and 1931 and impairs methoxymycolic acid synthesis in BCG strains obtained from the Pasteur Institute after this period. Mycolic acids of the cell wall are classified into three functional groups (alpha-, methoxy-, and ketomycolic acids), and together these lipids form a highly specialized permeability barrier around the bacterium. To explore the impact of methoxymycolic acid production by BCG strains, we complemented the functional gene of mmaA3 into BCG Denmark and tested a number of in vitro and in vivo phenotypes. Surprisingly, restoration of methoxymycolic acids alone had no effect on cell wall permeability, resistance to antibiotics, or growth in cultured macrophages and C57BL/6 mice. Our results demonstrate that the loss of methoxymycolic acid production did not apparently affect the virulence of BCG strains.
INTRODUCTION
Bacille Calmette-Gu?rin (BCG) vaccines are live attenuated strains of Mycobacterium bovis derived by in vitro passage from 1908 to 1921. BCG vaccines are given to millions of infants each year as antituberculosis vaccines, although their capacity to prevent tuberculosis in clinical trials has ranged from 80% protection to no detectable benefit (11). Several hypotheses have been proposed to explain this variable protection, including exposure to environmental mycobacteria (26) and differences between BCG vaccine strains (8).
From genomic analyses, it is now known that during in vitro passage, M. bovis lost a genomic region called RD1 (19), which has been shown to contribute to the observed attenuation of virulence of M. bovis BCG strains (16, 27). However, complementation of RD1 in the Pasteur strain of BCG did not completely restore pathogenicity in immunocompetent mice (27), suggesting that further mutations contribute to the observed phenotype of BCG strains. Because BCG stocks were propagated for another 40 to 50 years in various vaccine production laboratories, it has been hypothesized that ongoing evolution of BCG in vitro may have resulted in additional attenuation to the detriment of protective efficacy (5). Early reports on BCG suggest a second phase of attenuation in the late 1920s (24), with a decrease in BCG virulence in animal models (15) and reduced persistence of BCG in the mesenteric lymph nodes of vaccinated children (35). These observations are consistent with recent genomic analysis of existing BCG strains that revealed numerous mutations occurring after 1921 (6, 23).
Further evidence of evolution of BCG strains followed from studies demonstrating the loss of methoxymycolic acid production in BCG Pasteur and other strains (1, 20, 22). Prompted by these observations, it was determined that the production of methoxymycolates by BCG strains corresponded to their pattern of distribution from the Pasteur Institute, as strains obtained prior to 1927 (Birkhaug, Brazil, Japan, Russia, and Sweden) produce methoxymycolates, whereas strains distributed after this time (Connaught, Denmark, Frappier, Glaxo, Pasteur, Phipps, Prague, and Tice) do not (4).
A single-nucleotide nonsynonymous point mutation in mmaA3 causes a glycine-to-aspartic acid substitution at position 293 that impairs methoxymycolic acid production. Mycolic acids are long-chain -alkyl, ?-hydroxy fatty acids that are characteristic of the mycobacterial cell wall and are classified according to their functional group; the Mycobacterium tuberculosis complex consists of alpha-, keto-, and methoxymycolic acids (3, 21). The role of subclasses of mycolates has been explored in previous work, in which disruption of the hma gene in M. tuberculosis impaired both methoxy- and ketomycolic acid production and decreased cell wall permeability to the small molecules chenodeoxycholic acid and glycerol (10). The hma mutant also manifested reduced growth in a mouse model. Another study showed that heterologous promoter-driven overexpression of mmaA3 in BCG strains increased the production of methoxymycolates but, interestingly, impaired production of ketomycolates (34). Paradoxically, resistance to hydrophobic antibiotics increased in this mutant, whereas the uptake of the permeability marker chenodeoxycholate was unaffected. Overexpression of mmaA3 also appeared to reduce virulence, in this case assessed by growth inhibition in the human monocytic THP-1 cell line.
To further explore the importance of methoxymycolic acid production by BCG strains, we expressed the functional mmaA3 gene from its native promoter in a late strain of BCG and tested the impact on a number of in vitro and in vivo phenotypes.
Please visit the Web site to view the report in its entirety.