Relative to C. neoformans
, C. gattii
more frequently causes lung infections, and C. gattii
VGII strains especially are less frequently reported to be neurotropic in the human host (4
). This clinical observation was confirmed by our animal experiments. Mice infected via the pulmonary route with H99 appeared to have been killed by brain infection, while those infected with R265 were killed by lung infection, in both BALB/c and C57BL/6 strains. Interestingly, while inferior to H99 in extrapulmonary dissemination, R265 was superior to H99 in intrapulmonary growth. This may be explained by stronger pulmonary immunosuppression with R265, as fewer inflammatory cells were present in lungs infected with R265 than in lungs infected with H99. Cheng et al. (18) reported similar findings with C. gattii
strains, including the R265 strain, which was isolated from Vancouver outbreak cases. All the C. gattii
strains induced a less protective inflammatory response in C57BL/6 mice by suppressing neutrophil migration to the infection sites. Cheng et al. also found that C. gattii
strains failed to elicit the production of protective cytokines, such as tumor necrosis factor, in comparison with H99 (18
). However, the difference between C. gattii
and C. neoformans
in terms of their fate during infection has not been elucidated. We expanded on the pathological differences between the two species by monitoring the fate of the yeast cells inoculated via two different routes. Our results support the previous report (18
), but details of the immunosuppressive mechanisms that C. gattii
employs have yet to be deciphered.
The adaptive immune response, both antibody mediated and cell meditated, has been proven to be essential against cryptococcal infection (30
). C. gattii
cannot cause fulminant meningoencephalitis upon pulmonary inoculation, although the yeast cells must have been released into the bloodstream, since they can cross the blood-brain barrier (BBB). We speculate, therefore, that the host may manifest stronger protective immunity against R265. This speculation is supported by the fact that mice infected intravenously with R265 survived longer than those infected with H99 when they had been preinfected via the pulmonary route 2 weeks prior to the intravenous challenge. On the other hand, when mice were preinfected with R265 via the pulmonary route and challenged by intravenous inoculation after 5 days, their survival was not prolonged (data not shown). These findings suggest that an adaptive systemic immune response likely affects the dissemination to the brain and delays the establishment of severe meningoencephalitis in mice infected with R265 but not those infected with H99.
Since the protective effect of dual infections was not robust enough to prolong survival in R265 infected mice, additional mechanisms may prevent R265 from causing fulminating brain infection. In fact, even with a single high-dose i.v. inoculum, the fungal load of R265 in the brains was lower than that of H99 (). Furthermore, while most mice infected with a very low i.v. inoculum of R265 cleared the infection, the majority of mice infected with the same number of H99 cells did not (). In light of the consistent failure to recover R265 from the blood of mice inoculated intrapharyngeally (i.p.) and the poor yield of R265 in blood or serum cultures in vitro, it was not surprising to find limited numbers of CFU in extrapulmonary organs. Although the detailed immunological mechanism has yet to be investigated, our results suggest differences in the host innate immune responses to the two species. Therefore, comparative studies on the detailed host immunological response to C. neoformans versus C. gattii are warranted.
The mechanism by which C. neoformans
cells cross the blood-brain barrier (BBB) has been studied extensively, and both direct transcytosis (31
) and a Trojan horse mechanism (32
) have been proposed. A recent study using intravital microscopy reported that crossing the BBB by C. neoformans
is urease dependent, requires viability, and involves cellular deformation, which suggest direct invasion (33
). To date, however, neither mechanism has been confirmed directly within the host. It is possible that C. neoformans
may use both mechanisms to cross the BBB (34
). Whether and how C. gattii
crosses the BBB has never been studied. When mice were inoculated via the i.p. route, C. gattii
disseminated to the brain far less effectively than C. neoformans
and failed to cause meningoencephalitis despite its higher growth rate in the lungs. At the time of death, only 0 to 200 and 0 to 20 CFU were found in the brains of R265-infected C57BL/6 and BALB/c mice, respectively, indicating that death was not caused by a brain disease. This was also true even with i.v. inoculation when the inoculum size was 50 cells or less. Since growth of C. gattii
in blood or serum in vitro
was significantly slower than that of C. neoformans
, there may be a factor(s) in mouse blood or serum that inhibits C. gattii
cells from multiplying and eventually reaching the brain. However, R265 caused severe meningoencephalitis when the i.v. inoculum was large (50,000 cells/mouse). Even with the large inoculum, the brain fungal load of R265 trailed that of H99. These data suggest that R265 can cross the BBB and invade the brain but less efficiently than H99. The pattern of CNS infection is also different between the two species, since prominent cryptococcomas with hemorrhage were commonly observed in mice infected intravenously with R265 but were rarely seen in those infected with H99 (see Fig. S5
in the supplemental material; also data not shown).
In conclusion, our study validates previous reports on clinical manifestations of the two agents of cryptococcosis (4
): mice infected with C. gattii
R265 strain via the pulmonary route died due to overwhelming intrapulmonary growth, while those infected with H99 died due to fulminating brain disease. Despite the differences in their clinicopathology, both H99 and R265 crossed the blood-brain barrier effectively and established fatal brain infection upon intravenous inoculation. Although the growth rate of R265 in the blood was lower than that of H99 in vitro
, this does not explain why mice infected with R265 via the pulmonary route failed to exhibit fatal meningoencephalitis. For instance, mice infected via inhalation with 50 R265 cells survived for more than 90 days without noticeable brain lesions, while an i.v. inoculum of five R265 cells established fatal brain infection in 10% of mice in less than 30 days. It is possible that some immunological protection resulting during lung infection provided an additional mechanism(s) for reducing R265 dissemination to the extrapulmonary organs. Thus, it can be assumed that both innate and adaptive immune responses are responsible for such containment of C. gattii
. However, it is paradoxical, since R265 induces a relatively poor inflammatory response compared to H99 and yet can produce more adaptive immunity. The observation that growth of C. gattii
in the blood or serum is clearly slower than that of C. neoformans
raises the question of whether there are factors in mouse blood that actively suppress R265 but not H99 and might be further induced during infection. This would offer some protection against blood exposure to R265 but not H99. It is important to find anti-C. gattii
compounds in serum, since the agent in question does not appear to be complement. We believe our results have laid the foundation for further in-depth studies of the host immune response to the two species.