Introduction

A variety of microbial species inhabit the human gut, and every person has a unique gut microbiota consisting of about 170 microbial species, mainly reflecting different species of bacteria [1, 2]. Billions of bacteria colonize the adult gastrointestinal tract (GIT), and colonization potentially begins already at the fetal stage and develops throughout our lives [3, 4]. Sequencing of bacterial DNA to determine the intestinal microbiome has identified Firmicutes, Bacteroidetes, and Actinobacteria as the dominant bacterial phyla in the adult human intestine [5]. Firmicutes comprise butyrate-producing bacterial groups such as Clostridial cluster XIV and IV, while Bacteroidetes primarily comprise Bacteroides, Prevotella, and Porphyromonas [6, 7].

Arumugam et al. demonstrated that humans could be stratified into enterotypes, the three most common ones being defined by a high abundance of Bacteroides, Prevotella, and Ruminococcus, respectively [8]. Until now, most comprehensive studies of human gut microbiota, i.e., gut microbiome and metagenomics studies, have focused entirely on bacteria; hence, little is known about other types of intestinal organisms such as parasites, fungi and viruses in a microbiome context [9, 10].

Blastocystis spp. and D. fragilis represent two common, single-celled intestinal parasites. While the role of these parasites in human health and disease remains unsettled, recent data suggest that these parasites are more common in healthy individuals than in patients with functional and inflammatory bowel diseases [11–13]. Differences in interactions between parasites and bacteria may reflect differences in the clinical and public health significance of common intestinal parasites. The impact of bacterial communities on the pathogenicity and growth of parasites has been documented to some extent; for instance, the presence of Lactobacillus may reduce the growth of Giardia [14]. Additionally, a recent study analyzing metagenomics data showed a negative association between Blastocystis spp. and the Bacteroides enterotype in human fecal samples [15], suggesting that parasite colonization might be a biomarker of gut microbiota structure. However, the original material used to generate the metagenomics data was not available for studies that could corroborate these finding by other methods.

The aim of the present study was to validate previous data suggesting a clear negative association between Blastocystis spp. and the Bacteroides enterotype, and to identify associations between common intestinal parasites with intestinal bacterial groups using qPCR. This is the first time that bacterial DNA from clinical stool samples is investigated by targeted qPCR to reveal associations between two common intestinal parasitic protists, Blastocystis spp. and D. fragilis, and selected groups of intestinal bacteria with potential relevance for human health.

Materials and methods

Sample inclusion/exclusion criteria

Since the purpose of this study was to test for associations between common intestinal parasites and bacteria, 750 clinical fecal samples submitted by unique patients (in the period from 28th of March to 25th of April 2013) were considered as candidates. DNA from stool samples was extracted using the NucliSENS® easyMag® protocol (bioMérieux, Denmark) according to the manufacturer’s recommendations (Protocol B, 200 mg stool, 60 μL silica). All 750 samples were screened for Blastocystis spp., Dientamoeba fragilis, Giardia intestinalis, Entamoeba dispar, Entamoeba histolytica, and Cryptosporidium spp. by PCR using methods previously described [16, 17]. Exclusion criteria were as follows: (1) Samples positive for G. intestinalis, E. histolytica, E. dispar, and Cryptosporidium spp. as determined by qPCR, (2) patients who had been travelling abroad, and (3) patients less than 6 years of age since the intestinal microbiota is not yet completely stable in this age group [18]. All patients were referred from hospitals, specialists, or general practitioners. No information was available regarding potential consumption of antibiotics. A total of 84 samples were selected from eligible samples with a view to obtaining four groups with similar age and gender distribution and evenly distributed in terms of being Blastocystis spp.- and D. fragilis-positive/-negative, with 21 patients in each group, as follows: Blastocystis spp.- and D. fragilis-positive patients (B+ D+ group; age range, 9–70 years; mean age, 31 years [SD, 19.1]); Blastocystis spp.-positive and D. fragilis-negative patients (B+ D− group; age range, 6–69 years; mean age, 34 years [SD, 17.2]); Blastocystis spp.-negative and D. fragilis-positive patients (B– D+ group; age range, 9–69 years; mean age, 29 years [SD, 18.4]); and Blastocystis spp.- and D. fragilis-negative patients (B– D– group; age range, 17–69 years; mean age, 38 years [SD, 15.9]).

Detection of bacteria

The abundance of Bacteroides, Prevotella, Clostridial cluster IV, Clostridial cluster XIVa, Akkermansia muciniphilia, and Bifidobacterium was determined in quadruplets by qPCR on a LightCycler® 480 II System (Roche applied Systems, Penzberg, Germany). The relative abundance of each taxon was calculated as the ratio of the mean abundance of the specific taxon and the mean abundance of the total bacteria targeted by the primers HDA1 and HDA2 [19]. The LinRegPCR software program (The Heart Failure Research Center [HFRC], Amsterdam, Netherlands) was used to calculate abundance data. Amplification by qPCR, applied primers, and the calculation of relative abundance was performed as previously described [20]. One sample in the B+ D− group failed in qPCR, and this group therefore contained only 20 samples.

Data analysis

For statistical testing of differences between the four individual groups, a one-way ANOVA test was applied with the post-hoc Tukey’s range test [21]. To test the 0-hypothesis that there was no statistical difference in mean values between parasite-positive samples (B+ D+, B+ D− and B− D+ combined) and parasite-negative samples (B– D–), the Wilcoxon rank-sum test was applied. This variant was chosen since the number of samples differed in the two groups (N = 62 and N = 21 for parasite-positive and parasite-negative samples, respectively) [22].

Results

Analyzing for differences between the four individual groups, the relative abundance of Bacteroides was found to be significantly higher in the B− D− group compared with the B+ D+ and B+ D− groups (P < 0.01 and P < 0.05, respectively) (Fig. 1). No statistically significant difference was observed between the B-D+ group and any of the other groups with respect to the abundance of Bacteroides.

Fig. 1
figure 1

The log-normalized relative abundance of Bacteroides, Clostridial cluster IV, and Clostridial cluster XIVa in each of the four groups (B+ D+, B− D+, B+ D−, and B− D−). The middle bar indicates the mean value, and whiskers indicate the standard deviation. * P < 0.05, *** P < 0.001. (+/−) denotes whether the group is positive and/or negative for Blastocystis spp. and Dientamoeba fragilis (B/D)

Full size image

The relative abundance of Clostridial cluster IV was significantly lower in the B+ D− group compared with the B− D+ group (P < 0.05; Fig. 1). Furthermore, the relative abundance of Clostridial cluster XIVa was found in higher levels in the B– D– group compared with the B+ D− group (P < 0.05; Fig. 1). The abundance of Akkermansia muciniphilia and Bifidobacterium did not differ among the four groups.

Combining the two Blastocystis spp.-positive groups and the two Blastocystis spp.-negative groups, the relative abundance of Bacteroides was significantly lower in Blastocystis spp.-positive patients (P = 0.006). A comparison of all parasite-positive samples and all parasite-negative samples showed an even more pronounced difference in Bacteroides abundance (P < 0.001), and a borderline significant difference in the relative abundance of Clostridial cluster XIVa (Fig. 2). When grouping individuals based on age and disregarding parasite presence, the relative abundance of Clostridial cluster XIVa was higher in individuals aged 41–60 years than in those aged 61–80 years (P < 0.05).

Fig. 2
figure 2

Comparison of the log-normalized relative abundance of Bacteroides and Clostridial cluster XIVa in the combined parasite-positive groups (B+ D+, B− D+, B+ D−) versus the parasite-negative group (B− D−). *** P < 0.001. The middle bar indicates the mean value, and whiskers indicate the standard deviation

Full size image

The relative abundance of Prevotella did not differ significantly between the four groups. A bimodal distribution of the Prevotella-to-Bacteroides ratio was found (Fig. 3). Of note, most individuals in the B– D– group were found to have a low P/B ratio (Fig. 3), which was due to higher levels of Bacteroides in the B– D– group (Fig. 1). The other three groups were more similar in terms of the distribution of low and high P/B ratios.

Fig. 3
figure 3

Bimodal distribution of individuals according to Prevotella to Bacteroides (P/B) ratio. (+/−) denotes whether the group is positive or negative for Blastocystis spp. and/or Dientamoeba fragilis (B/D)

Full size image

Discussion

A negative association between Blastocystis spp. and Bacteroides in stool samples has previously been reported in healthy subjects [15]. In concordance with this, we found a higher relative abundance of Bacteroides in B– D– individuals than in Blastocystis-positive individuals, suggesting a negative association between Bacteroides and Blastocystis. The previous study [15] described a negative association between the Bacteroides enterotype and the presence of Blastocystis, but not in a quantitative manner; meanwhile, the present study sought to focus on the quantitative relationships between the studied taxa. When taking the relative abundance of Bacteroides of every single study individual into account—not only that of the Bacteroides-dominated individuals—a quantitative relationship between Bacteroides and Blastocystis spp. was revealed. Our data also revealed that the lower abundance of Bacteroides in Blastocystis spp.-positive individuals is not only observed in healthy subjects [15], but also in patients with intestinal symptoms. This suggests that the association between Bacteroides and Blastocystis is independent of health status.

The B+ D− group showed a significantly lower abundance of Bacteroides than the B– D– group (P < 0.05), and the two Blastocystis spp.-positive groups exhibited an even higher negative association to Bacteroides compared with the two Blastocystis spp.-negative groups (P = 0.006). Eventually, when combining all parasite-positive individuals, a strong negative association between Bacteroides and parasites was observed (P < 0.001). Our data indicate that Blastocystis spp. drive the difference in Bacteroides abundance observed across study individuals but also appear to indicate that Blastocystis spp. and D. fragilis in conjunction have an even more pronounced influence on the abundance of Bacteroides. It could be speculated that the presence of another organism such as D. fragilis could increase this association; however, further analysis of healthy individuals should also be performed in order to confirm this. Additionally, it would be interesting to explore whether the negative correlation existing between these common intestinal parasites and Bacteroides also parallels an increase in bacterial richness in the presence of these parasites, as indicated in a previous study [15]. If high bacterial richness reflects a healthy gut ecology, the influence of these common parasites on bacterial richness warrants further investigation. If colonization by Blastocystis is linked to high gut microbiota diversity, the use of in vitro and in vivo models may help clarify whether Blastocystis selects for high gut microbiota diversity or whether it is the other way around. For instance, experiments involving the introduction of Blastocystis in a Blastocystis-naïve bacterial ecosystem could focus on potential changes in bacterial structure over time.

The abundance of Clostridial cluster XIVa was lower in parasite-positive individuals. As there was no statistical difference in age distribution between the four groups, the establishment of cluster XIVa also appears to be influenced by co-colonization with parasites and not just determined by the individual’s age as described by Claesson et al. [23].

The bimodal distribution of Prevotella/Bacteroides ratio as detected by qPCR has previously been reported in children [24, 25]. Our data indicate that the absence of Blastocystis spp. and D. fragilis is associated with a low Prevotella/Bacteroides ratio, which may correspond to the Bacteroides-driven enterotype as defined by Arumugam et al. [8]. However, the ecology underlying this association remains to be elucidated.

In the present study, no information was available regarding potential consumption of antibiotics. Of course, previous antibiotic use in some of the patients could have resulted in perturbation of the microbiota in certain directions, including clearance of Blastocystis and selection for a certain microbiota structure. However, if any drug would be used to treat patients suspected of intestinal parasitosis, this would be metronidazole or mebendazole, both of which are quite inefficient in terms of eradicating Blastocystis [26]. Meanwhile, metronidazole is a broad-spectrum antibiotic potentially resulting in shifts in gut microbiota; for example, extensive metronidazole use may lead to Clostridium difficile infections, which reflect gut microbiota perturbation. Therefore, any metronidazole use leading to gut microbiota perturbation may indirectly result in eradication of Blastocystis. This situation, however, does not affect the hypothesis that Blastocystis is a biomarker of gut microbiota diversity and/or intestinal homeostasis.

Conclusion

Our data corroborate the theory that common intestinal parasitic micro-eukaryotes can be viewed as markers of gut microbiota structure. Further studies are needed to elaborate on the associations between common intestinal parasitic micro-eukaryotes and the composition of the intestinal microbiota in relation to health and disease.