Caenorhabditis elegans as an Alternative Model Host for Legionella pneumophila, and Protective Effects of Bifidobacterium infantis

As a standard strain of C. elegans, Bristol N2 was maintained and propagated on nematode growth medium (NGM) according to standard techniques using OP (25). Tryptone soya agar (Oxoid, Basingstoke, United Kingdom) was used to culture OP at 37°C. The virulent and attenuated L. pneumophila strains used in this study are listed in Table 1. All attenuated strains were produced by transposon insertion into the Icm/Dot (intracellular multiplication/defect in organelle trafficking) type IV secretion system genes, which have been shown previously to be essential for virulence (5, 22). Legionella strains were cultured using BCYE agar plates (Oxoid) at 37°C for 2 days. Bifidobacterium infantis ATCC 15697 was also used to feed the worms, being one of five lactic acid bacteria that we previously found to have a longevity effect on nematodes (13). Transoligosaccharide (TOS) propionate agar (Eiken Chemical Co., Tochigi, Japan) was used to grow B. infantis anaerobically at 37°C (32). Bacteria were recovered from the agar plates, and each 10 mg (wet weight) of the bacteria suspended in 50 μl of M9 buffer (25) was spread onto peptone-free modified NGM (mNGM) in 5.0-cm-diameter petri dishes for feeding or infecting of C. elegans.

Worms were generated from eggs released after exposure of adult hermaphrodites to a sodium hypochlorite-sodium hydroxide solution as described previously (28). The fertilized egg suspension was incubated overnight at 25°C to allow hatching, and the suspension of larval stage 1 (L1 stage) worms was centrifuged at 156 × g for 1 min. The supernatant was removed, and the remaining larvae were transferred onto fresh mNGM plates covered with OP and incubated at 25°C. Because the reproductive system regulates aging in C. elegans (12), in order not to disturb natural pubescence, worms were fed on OP until the start of Legionella infection.

Nematocidal assays were begun with adult worms, which were allocated at 30 each onto mNGM plates covered with suspensions of each Legionella strain. Once the feeding bacteria were switched from OP to Legionella, worms were given the pathogen during the entire assay time until death. The plates were incubated at 25°C, and live and dead worms were counted at least every 24 h. Worms produce progeny that develop into adults in 3 days, rendering it difficult to identify the original worms. Therefore, the original worms were transferred onto fresh plates at approximately 2-day intervals. A worm was considered dead when it failed to respond to a gentle touch with a worm picker. Worms that died as a result of getting stuck to the wall of the plate were excluded from the analysis. All nematode survival studies were conducted by a researcher with no prior knowledge of the bacterial strains fed to the worms. Nematode survival was calculated by the Kaplan-Meier method, and survival differences were tested for significance using the log rank test.

No significant differences in survival between 3-day-old worms fed OP and 3-day-old worms infected with virulent Legionella strains were seen (Fig. 1 A). However, when the worms were infected from 7.5 days after hatching, the virulent Legionella strains were obviously nematocidal (Fig. 1B).

The present study is the first to show that L. pneumophila is virulent to elderly worms. Our previous study showed that Salmonella enterica is clearly virulent to both older and younger worms, although more so to elderly worms (13). These findings appear to be consistent with the epidemiological characteristics of both pathogens in humans: Legionella tends to infect older people in an opportunistic manner, while Salmonella can cause enteritis irrespective of host age.

To examine whether the virulence of Legionella correlates with its ability to accumulate in the nematodes, the number of organisms in the nematodes was determined according to the method of Garsin et al. (9) with some modifications. Each nematode was mechanically disrupted using a microtube pestle (Scientific Specialties Inc., Lodi, CA). The number of Legionella cells was determined using BCYE agar plates supplemented with polymyxin B at 80 IU/ml. The number of Legionella cells recovered from worms infected from 3 days after hatching was smaller than the number recovered from worms infected from 8 days after hatching (Fig. 1C); this logarithmic difference cannot be explained by the difference in body size (the area of the worm's projection) between elderly and young worms since the bodies of the 8-day-old worms were 20% larger than those of the 3-day-old worms.

Although there are no data to explain the vulnerability of the older worms, the ability of the older worms to adapt to microbes might have deteriorated, a possibility consistent with the finding that the level of heat shock protein 16 in 16-day-old worms responding to heat stress had declined to about 50% of that in 4-day-old young adults (34). Since nematodes have to switch distinct genes on and off to defend against and digest different species of ingested pathogens (35), it is likely that the worms that first encountered the Legionella cells when they were young were able to respond effectively to Legionella and that the 8-day-old worms were not.

On the basis of these findings, 7.5- to 8-day-old nematodes were used as the hosts to examine the virulence of the Legionella strains. Although the parent strain JR32 showed significant nematocidal activity, all the genetically attenuated mutants, having been demonstrated to be less virulent in the lungs of guinea pigs (19) or in human macrophages (22), were considered nonpathogenic, similar to OP (Table 2). Interestingly, the attenuated mutant LELA1718, which is reportedly cytolethal in contrast to the other attenuated mutants in a cytotoxicity assay with HL-60-derived human macrophages (22), showed modest virulence in the nematodes compared to other avirulent mutants.

The pathogenicity of L. pneumophila in C. elegans seems to correlate well with that in macrophages, and the nematode could serve as a unique host of Legionella spp. Indeed, several genes of human pathogens have been recently discovered to be involved in virulence in C. elegans, and some of these are also required for virulence in mammalian systems (4, 16, 26, 29). However, the pathogenesis of infections with Legionella spp. in nematodes remains to be elucidated. The numbers of organisms recovered from the worms were not significantly different among the three Legionella strains (JR32, LELA1718, and LELA4432) that showed different levels of pathogenicity. Since the nematodes actively take in bacteria, the differences in virulence would not result in a difference in bacterial number as long as the levels of bacterial resistance to antimicrobial peptides in the worms did not differ among the bacterial strains. Functional defects in the Icm/Dot type IV secretion system could have decreased the secretion of toxic effectors from the attenuated mutants and consequently increased worm survival. Programmed cell death in C. elegans appears to be an important defense response against pathogen attack and stress (1). Since Icm/Dot plays a role to inhibit programmed cell death (6), the attenuated mutants appeared not to reduce worm survival.

Three-day-old worms were assigned to either a control group that continued to be fed OP or to a group that was fed bifidobacteria for 5 days. The 8-day-old worms were then transferred onto mNGM covered with Legionella. Nematodes fed bifidobacteria were clearly resistant to subsequent Legionella infection, in contrast to nematodes fed OP before the Legionella infection, and nematodes fed bifidobacteria survived longer than the control worms fed OP throughout the experiment (Fig. 1D); however, the numbers of Legionella organisms recovered from the worms showed no significant difference between groups fed with bifidobacteria or OP. These results are similar to our previous findings that lactic acid bacteria are effective in increasing both the longevity of C. elegans and the host defense against Salmonella (13). Our ongoing experiments using mutated nematodes suggest that bifidobacteria enhance the worms' host defense via the PMK-1/p38 mitogen-activated protein kinase signaling pathway (unpublished results), which plays an important role in the innate immunity of C. elegans (27).

In conclusion, we found that L. pneumophila significantly reduced the life span of C. elegans via icm/dot-associated pathogenicity when the pathogen infected older worms. Since the worms died over the course of several days, the pathogen is considered to have caused what is known as slow killing (30). C. elegans can be an additional unique model to study opportunistic infection with Legionella spp.

We are grateful to Howard A. Shuman for kindly providing the L. pneumophila AM511, JR32, and LELA strains. The nematodes used in this work were kindly provided by T. Stiernagle at the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources (NCRR).

This study was supported in part by a grant-in-aid from the Graduate School of Human Life Science, Osaka City University.