Our results clearly demonstrate that isocitrate lyase is essential for growth only when C2
-generating carbon sources are present. This confirms a previous study in which specific inhibition of isocitrate lyase by 3-nitropropionate inhibited germination on an acetate medium (9
). However, during cometabolism with other carbon sources, which do not require the glyoxylate bypass, isocitrate lyase seems to play only a minor role.
Most interestingly, we were able to show that isocitrate lyase is not important for the development of invasive aspergillosis in a murine infection model. This result was rather unexpected, because it was generally believed that lipids and fatty acids serve as one of the major carbon sources during the infection process and that the glyoxylate bypass is needed to avoid a shortage of oxaloacetate. This assumption was supported by the results obtained with isocitrate lyase mutants of various microorganisms, including both human and plant pathogens. Deletion of the two isocitrate lyases present in M. tuberculosis
, which can substitute for each other, led to the rapid elimination of bacteria from the lungs and to the impairment of intracellular replication (19
). Deletion of the single isocitrate lyase from C. albicans
led to a strong attenuation in virulence in a murine infection model (15
), and furthermore, the plant-pathogenic fungi Leptosphaeria maculans
) and Magnaporthe grisea
) displayed significantly reduced virulence levels when the isocitrate lyase gene was deleted.
In our study, we used a corticosteroid-based immunosuppression model in order to test the A. fumigatus
isocitrate lyase mutant for its virulence. Although this system does not cause neutropenia, it is suitable as a model for invasive aspergillosis. In our cortisone acetate immunosuppression system, mice infected with wild-type, isocitrate lyase mutant, or isocitrate lyase-complemented strains died from invasive aspergillosis. Although a severe inflammatory response accompanied by a massive recruitment of neutrophils to the lung tissue occurs under cortisone acetate immunosuppression, a dramatic decrease in the production of tumor necrosis factor alpha (TNF-α) has been observed (8
). TNF-α is supposedly essential for the induction of protective immunity against A. fumigatus
). However, in a neutropenic model, although large amounts of TNF-α have been found in the bronchoalveolar lavage fluids of infected mice, this cytokine was not efficient to prevent the development of invasive aspergillosis (1
). Therefore, in both corticosteroid-based and neutropenic immunsuppression models, extensive pulmonary invasion by growing hyphae occurs, and this results in the death of mice, caused by an acute cardiorespiratory insufficiency. Despite the differences in immune response, both models seem to be suitable to test the attenuation in virulence of Aspergillus
Due to the strong hyphal growth of our isocitrate lyase mutant, which was comparable to that of the wild type and the complemented mutant, we conclude that isocitrate lyase is not required for invasive growth of A. fumigatus
. In agreement with this finding is the finding that isocitrate lyase from another pathogenic fungus, Cryptococcus neoformans
, was shown not to be required for pathogenesis (21
). In C. neoformans
, isocitrate lyase was first thought to represent a putative target for antifungal drugs, because the gene was found to be up-regulated when reisolated from a rabbit infection model 7 days postinfection. This was explained by the fact that immune effector cells, like macrophages, at that time acquire their maximum activation state for fungicidal activity, which leads to phagocytosis of yeast cells and subsequent activation of isocitrate lyase because of lipids which are present in macrophages. Disruption of the isocitrate lyase coding region, however, revealed that this had no effect on virulence when the mutant strain was tested in a murine inhalation model for cryptococcosis.
Therefore, we can conclude that, although lipid-specific Nile red staining of the macrophages confirmed the presence of large lipid loads (9
), isocitrate lyase is also not required for development of invasive aspergillosis. This could be explained by the fact that the macrophages may contain not only lipids but also proteins or carbohydrates. Furthermore, germinating conidia grow for only a short time within the macrophages, because the elongating hyphae destroy them (27
). This makes available alternative carbon sources like proteins, which are released from the surrounding tissue during invasive growth, and leads to the assumption that isocitrate lyase may, in the macrophages, be required only for long-term persistence. This is also supported by investigations of the pathogenic bacterium Salmonella enterica
serovar Typhimurium (10
). An isocitrate lyase mutant of this bacterium shows no reduction in acute virulence. However, when long-term persistence in the macrophages was examined, the wild-type strain was able to persist in the macrophages, whereas the isocitrate lyase mutant was progressively cleared. Whether the persistence of A. fumigatus
conidia also plays a role in the infection of humans has not yet been investigated. It therefore remains unclear whether all cases of invasive aspergillosis derive from a “de novo infection” with conidia taken up from the environment or from persisting conidia which start to germinate when the immune system becomes suppressed.
We showed that during cometabolism of proteins and lipids, isocitrate lyase is hardly induced and, therefore, may not be required for anaplerosis of oxaloacetate. Additionally, immunostaining of isocitrate lyase in the tissues of patients suffering from an invasive fungal infection revealed only negative results (Frank Ebel, Max von Pettenkofer Institute, Munich, Germany, personal communication). Theoretically, the possibility that the antibodies were unable to detect isocitrate lyase in the tissue sections after they were fixed and embedded in paraffin cannot be excluded. However, the fact that the antibodies worked well on infected and formaldehyde-fixed macrophages, and that they clearly showed the presence of isocitrate lyase (9
), seems to exclude this possibility.
We conclude that isocitrate lyase, and hence the glyoxylate cycle, is not required for the virulence of A. fumigatus in the murine infection model. Therefore, from our results, isocitrate lyase does not represent an antifungal drug target. Furthermore, results show that even among different fungal species infecting human tissues, diverse metabolic pathways are required. Although lipids may be consumed during invasive growth of A. fumigatus, they do not provide the major carbon source. Hence, it seems very likely that proteins released from the host tissue represent at least one additional carbon source supporting the growth of A. fumigatus during invasive aspergillosis. This hypothesis is currently being tested.