In a comprehensive review of the literature, Kristensen et al. (33
) found a small number of publications describing randomized controlled trials of probiotics that included the characterization of fecal microbiota. The surprising conclusion of that survey was that no publications reported a significant change in the microbiota based on OTU richness, evenness, or diversity analysis. As no uninfected controls were included in our experiments, the impact of the probiotics on the microbiota could be observed only in the initial phase of the experiment, as described above for days 1 and 2 after initiation of the probiotic treatment. The increase in facultative aerobes later in the infection likely represents the effect of C. parvum
multiplication in the intestinal epithelium, rather than a direct impact of probiotics.
Because we did not observe a significant increase in probiotic bacteria in the feces of treated mice, we postulate that some of the bacterial or prebiotic ingredients present in the probiotic product induced changes in the mouse intestinal environment, favoring the proliferation of C. parvum
. Proliferation of the parasite then led to extensive secondary modifications of the microbiota, as shown in Fig. 4
. An impact of the prebiotics present in the product, i.e., acacia gum, larch gum, oligosaccharides, and l
-glutamine, on the microbiota cannot be excluded. Elucidation of the mechanism by which probiotic administration promotes proliferation of C. parvum
will require testing of individual probiotic species or defined combinations of species and/or prebiotics (34
) and metabolomic analysis to identify mediators of the probiotic effect. This research is of primary importance to enable targeted manipulations of the microbiota aimed at limiting the proliferation of Cryptosporidium
parasites. Zhu et al. (21
) described methods to “edit” the gut microbiota, in that case by inhibiting the multiplication of facultative anaerobes. An analogous approach could be used to investigate the causal link between parasite proliferation and dysbiosis. Although mice infected with C. parvum
do not develop diarrhea, the fecal microbiota from heavily infected animals in our experiments resembles the fecal microbiota of humans suffering from cholera diarrhea (29–31
) or diarrhea of other etiologies (31
). This observation is significant, because it indicates that neither the actual pathogen nor diarrhea is important for inducing dysbiosis. A characteristic of many intestinal pathologies of infectious or other causes is an increase in the proportion of Gammaproteobacteria
). Although exceptions to this trend have been reported (36
), a shift toward facultative anaerobes, reflecting increased permeability of the gut epithelium, is a hallmark of infectious (29
), inflammatory (39–41
), and other (42
) intestinal pathologies. The abundance of Gammaproteobacteria
in the distal gut of mice heavily infected with C. parvum
indicates a shift in the luminal oxygen gradient (43
), likely a consequence of epithelial erosion and villus atrophy (44–46
). These observations raise the question of whether Cryptosporidium
proliferation responds to the oxygen concentration in the gut lumen. Selective inhibition or promotion of oxygen-consuming bacteria (21
), to increase or to deplete luminal O2
), could potentially be investigated to assess the response of C. parvum
and to explore dietary interventions to mitigate the severity and duration of cryptosporidiosis.
Few studies have reported on the effects of diet on C. parvum
infection. Liu et al. (47
) found that protein deficiency increased the concentration of C. parvum
DNA in feces; in that study, however, the difference between normal and protein-deficient animals was reported at 20 h p.i. Since C. parvum
is not known to complete its life cycle in less than 72 h (48
), the results are difficult to interpret. Another study found a positive effect of pomegranate extract on cryptosporidiosis in calves (49
). The authors reported that calves fed milk supplemented with extract excreted fewer oocysts. The very limited range of the literature on the effects of diet on cryptosporidiosis illustrates the need for additional research, particularly basic research on mechanistic aspects of parasite-microbiota interactions.
Although the results across experiments were consistent, differences were also noticed. Most notably, the average oocyst output in experiment 2 was higher, indicating more severe infection. Corroborating the model discussed above, more severe infection was associated with greater relative abundance of Gammaproteobacteria
). The reason for the difference in the severity of infection is difficult to determine, but the different route of dexamethasone administration in experiment 2 might have contributed to this outcome. Differences in fecal oocyst outputs between cohoused mice were also observed in experiments 1 and 2 (Fig. 1
). As described in Materials and Methods, in those experiments, animals from the same cage were housed individually for 16 h three times a week for collection of feces but otherwise were housed together in 2 cages per treatment (4 cages for each experiment). The differences in oocyst outputs and microbiota profiles between cage mates are difficult to explain, given the close contact between animals. This phenomenon justifies mice being sampled individually, rather than being sampled by group, as is common practice, and is the reason why experiment 3 samples were not subjected to a full set of analyses as for samples from experiments 1 and 2.
In the absence of effective drugs to control cryptosporidiosis, a search for alternative treatments is warranted. In most cases, however, we do not know how perturbation of the microbiota, whether induced by diet, probiotics, antibiotics, or prebiotics, affects enteric pathogens, particularly Cryptosporidium
parasites. Identifying specific mechanisms affecting pathogen virulence in response to diet may enable the development of targeted microbiota-editing measures to mitigate the severity of cryptosporidiosis. Methods designed to detect changes in the metabolome (50
) will be needed to supplement taxonomic analyses based on 16S amplicon sequencing. Lastly, enhancing the value of the rodent cryptosporidiosis model, the observed shift toward facultative anaerobes in the infected gut indicates common pathogenic changes in the human and rodent intestines in response to enteric infections.