Background

Blastocystis is a common gut eukaryote detected in human and many animal hosts [1, 2]. It is classified under the group Stramenopiles which mostly comprises unicellular flagellated or ciliated free-living organisms [2, 3]. Blastocystis, however, is an obligately anaerobic and parasitic protist and is transmitted via the fecal-oral route [1]. Estimates put the number of individuals infected by this parasite to more than 1 billion worldwide [4]. Although it is more common in developing countries, surveys in developed countries often indicate prevalence rates of more than 5% in the general population [5]. The role of Blastocystis in disease has been the subject of many investigations. There are studies associating it with symptoms of a gastrointestinal disease [6,7,8] while others could not find the basis for defining it as pathogenic [9, 10]. More recently, infection with Blastocystis has been linked with irritable bowel syndrome (IBS) [11] and inflammatory bowel disease (IBD) [12]. There are however conflicting reports on whether Blastocystis was really the sole causative agent in these cases [12,13,14,15]. In another perspective, IBS [16,17,18] and IBD [19, 20] were also linked to the disruption of the gut microbiota or dysbiosis. Blastocystis’ role in IBS or IBD may thus be mediated by altering gut microbiota composition. However, the few microbiome studies on Blastocystis generally identified it as a common commensal in the human gut. These analyses associated the presence of Blastocystis with higher diversity of gut microbiota [21,22,23,24]. However, one study indicated that it caused a decrease in beneficial bacteria particularly Bifidobacterium and Lactobacillus spp. [25]. These discrepancies may be due to the complex nature of Blastocystis wherein several genetically distinct subtypes (ST) exist. Different Blastocystis STs could exhibit different growth rates, drug susceptibilities, host ranges, and other biological features [1, 26]. These differences could therefore influence the protist’s influence on the gut microbiota. Indeed, it has been suggested that microbiota composition in relation to Blastocystis may be dependent on the organism’s subtype identity [27].

With these in mind, the current study explored the interactions between a particular ST of Blastocystis, ST7, and prokaryotic representatives of the gut microbiota. ST7 isolates possess pathogenic properties not observed in other STs. For example, drug susceptibility assays indicated that ST7 isolates are resistant to metronidazole, the usual drug of choice to clear protistan parasites [28]. In vitro culture assays also revealed that ST7, but not ST4 isolates, could compromise the intestinal epithelial barrier [29]. In vivo experiments also revealed that isolates from this ST could cause tissue damage in the mouse intestines [30]. This ST appears rarely in surveys but has been reported to be strongly associated with gastrointestinal symptoms [31]. We used co-culture experiments to determine the effect of Blastocystis ST7 isolates on the viability of select gut bacteria representatives. Biological assays as well as gene expression analyses were used to investigate a possible mechanism on how Blastocystis affect these bacterial populations. Lastly, we conducted in vivo experiments involving infection of mice with Blastocystis and subsequent analyses of bacterial content in the fecal samples. The results of this study indicated that Blastocystis can disrupt gut microbiota populations particularly decreasing the content of Bifidobacteria and Lactobacillus but increasing Escherichia coli. Possible explanations of these occurrences point to oxidative stress caused by Blastocystis as well as host factors induced by the parasite. Our data indicates that while Blastocystis spp. may be a member of healthy gut microbiota, specific isolates or rare ST may disrupt homeostasis leading to pathological states in the host.

Methods

Blastocystis cultures

Human Blastocystis isolates were acquired from patients at the Singapore General Hospital in the early 1990s before the Institutional Review Board was established at the National University of Singapore (NUS). Blastocystis ST7 isolates B and H are maintained at a microbial collection at the Department of Microbiology and Immunology of the NUS. Both isolates ST7-B and ST7-H were axenized previously [32] and maintained in 8 ml pre-reduced Iscove’s modified Dulbecco’s medium (IMDM) (Gibco) supplemented with heat-inactivated 10% horse serum (Gibco). These were incubated in anaerobic jars (Oxoid) with Anaerogen gas packs (Oxoid) at 37 °C and subcultured every 3–4 days. Blastocystis cell counts were done manually using hemocytometer (Kova International).

Bacterial cultures

Escherichia coli ATCC 11775, Enterococcus faecalis ATCC 29212, Bacillus subtilis ATCC 6633, and Bacteroides fragilis ATCC 25285 were cultured and maintained in Luria-Bertani (LB) broth and agar (Sigma). Bifidobacterium longum ATCC 15707 and Lactobacillus brevis ATCC 14869 were cultured and maintained in Bifidus selective medium (BSM) (Sigma) and deMan, Rogosa, Sharpe (MRS) medium (Sigma), respectively, in broth and agar forms. B. fragilis and B. longum were maintained in anaerobic condition inside anaerobic jars (Oxoid) with Anaerogen gas packs (Oxoid). All cultures were incubated at 37 °C. Absorbance readings of bacterial broth cultures prior to experiments were done using Tecan Infinite F200 microplate reader.

Co-culture experiments

Blastocystis cells and bacterial cells were washed twice in phosphate-buffered saline (PBS) at 1000×g for 10 min. A concentration of 1 × 107 cells/ml of Blastocystis ST7-B or ST7-H and 1 × 109 CFU/ml of bacteria (E. coli, E. faecalis, B. longum, L. brevis, B. subtilis, or B. fragilis) was incubated for 24 h at 37 °C in pre-reduced PBS. Controls with only 1 × 107 cells/ml Blastocystis and only 1 × 109 CFU/ml bacteria, both re-suspended in pre-reduced PBS, were also incubated for 24 h at 37 °C. After 24 h, Blastocystis cells were counted using a hemocytometer (Kova International) after a 50-fold dilution of the neat cultures. A drop plate method was utilized for the enumeration of bacterial colony-forming units (CFUs) [33]. Bacterial colony-forming unit per milliliter was determined when the colonies appeared on the agar plates.

B. longum ROS staining and flow cytometry

To determine cellular reactive oxygen species (ROS) content in B. longum cells, the stain 2′,7′–dichlorofluorescein diacetate (DCFDA) (Sigma) was used at a concentration of 20 μM for 30 min at 37 °C. Before the co-culture experiment, B. longum cells were stained with Baclight Red (Thermofisher) at a concentration of 1 μM for 15 min at room temperature to be able to gate for these cells in flow cytometry. The cells were run in Attune Nxt Flow Cytometer (Life Technologies) using blue (488 nm) and yellow (561 nm) lasers.

B. longum oxidoreductases genes expression analysis

mRNA from B. longum cells were extracted using RNAzol RT (Sigma-Aldrich) following the manufacturer’s instructions. cDNA was synthesized using iScript cDNA kit (Bio-Rad). The gene expression of Bifidobacterium oxidoreductases were determined in a qPCR assay using primers reported previously [34]. SsoAdvanced™ Universal SYBR Green Supermix (Bio-Rad) was used and amplifications performed in an iCycler thermocycler with iQ5 attachment (Bio-Rad).

HT-29 monolayer

Cells were maintained in T-75 flasks (Corning) in a humidified incubator with 5% CO2 at 37 °C. Culture medium consisted of 10% heat-inactivated FBS (Gibco) and 1% each of sodium pyruvate (Gibco), non-essential amino acids (Gibco), and penicillin-streptomycin in Dulbecco’s modified Eagle’s medium (DMEM) (Thermo Scientific). HT-29 cells were then used for co-culture experiments with Blastocystis and B. longum at 1 × 106 cells/ml and 1 × 109 CFU/ml, respectively.

Epithelial permeability measurement

HT-29 Cells were seeded with complete medium onto Millicell hanging cell culture insert with 0.4-μm-sized pores (Merck) placed on 6-well plates (Greiner). After reaching confluence, the monolayers were stimulated for 48 h with 3 mM sodium butyrate (Sigma-Aldrich) in serum-free medium. Conditioning of differentiated HT-29 monolayers by Blastocystis was performed for 24 h at 37 °C in anaerobic condition. Incubation of the HT-29 monolayers with B. longum was performed for 6 h for viability determination. Transepithelial electrical resistance across the monolayers was measured using Millipore-ERS-2 volt-ohm-meter. Flux assay was performed using fluorescein isothiocyanate-conjugated Dextran 4000 (FITC-Dextran) (Sigma). The assay included washing of the monolayer on the inserts twice with Hank’s balanced salt solution (HBSS) (Thermofisher). FITC-Dextran at a concentration of 100 μg/ml in HBSS was added on the apical compartment, and the plate was incubated at 37 °C. The buffer at the basolateral compartment was collected after 1 h and transferred to a black 96-well plate (Nunc). Fluorescence was measured using Tecan Infinite F200 microplate reader at excitation and emission wavelengths of 492 nm and 518 nm, respectively.

Acute infection of Blastocystis in a mouse model

The animal experiments were performed according to the Singapore National Advisory Committee for Laboratory Animal Research guidelines. The protocol (R13-5890) was approved by the NUS Institutional Animal Care and Use Committee. The infection of Blastocystis into mice was carried out according to a previous protocol [30]. C57BL/6 male mice, aged 5 to 6 weeks, were given 2% DSS in drinking water for 4 days followed by a recovery period of 5 days. After the recovery period, they were injected with 5 × 107 live Blastocystis cells intracecally. The mice were subjected to anesthesia (ketamine 75 mg/kg + medetomidine 1 mg/kg via intraperitoneal (IP) injection) then a vertical incision was made on the abdomen. The cecum was exteriorized, and 50 μl Blastocystis suspended in PBS was injected into the caecum using a 27G needle. Sham surgical controls were injected with 50 μl PBS intracecally. The incision was then closed with two layers of sutures. Subsequently, anesthesia was reversed (Atipamezole 1 mg/kg via subcutaneous (SC) injection), and antibiotics (Enrofloxacin10 mg/kg SC) and analgesic (Carprofen 5 mg/kg SC) were given. Fecal samples were collected at various time-points—before surgery at day 0, day 1 post-infection, day 2 post-infection, and day 3 post-infection. A total of 24 mice were included, with 8 mice in each of the 3 groups—control, ST7-B-infected and ST7-H-infected. The rest of the mice were euthanized on day 3 post-infection. The colon and cecum were extracted for histology.

Determination of bacterial abundance in mice fecal samples

DNA from 42 mg each of mouse fecal samples were extracted using QIAamp Fast DNA Stool Mini Kit (Qiagen) following the manufacturer’s instructions. The relative abundance of selected bacterial groups was determined in a qPCR assay using the DNA primers listed (Table 1). SsoAdvanced™ Universal SYBR Green Supermix (Bio-Rad) was used, and amplifications were carried out in an iCycler thermocycler with iQ5 attachment (Bio-Rad).

Table 1 DNA primers used for qPCR assay to determine the relative abundance of bacterial group in mice fecal samples
Full size table

Statistical analysis

Comparisons of two groups were done using Student’s t test for paired samples. Comparisons of more than two groups were done using analysis of variance (ANOVA). Analyses and generation of graphs were done using Prism GraphPad version 5.

Results

Gut bacteria exerted positive effects on Blastocystis cell count in vitro

To determine whether the presence of gut commensal bacteria affects Blastocystis cell count in vitro, ST7-H and ST7-B were individually co-incubated with representative bacteria of the gut microbiota—E. coli, E. faecalis, B. longum, L. brevis, B. fragilis, and B. subtilis. The reduced PBS condition used for co-incubation ensured a low oxygen environment necessary for Blastocystis viability while the simple PBS formulation minimizes potential exogenous growth factors that would otherwise complicate the assay, resulting in bacterial overgrowth. Generally, both ST7-B and ST7-H displayed higher cell counts when co-incubated with gut commensal bacteria, with differential effects observed depending on the species of bacteria (Fig. 1a). More specifically, significant positive effects were observed when ST7-B was co-incubated with E. coli, E. faecalis, B. longum, and B. fragilis, and when ST7-H was co-incubated with E. faecalis, B. longum, and B. fragilis. The highest observed positive effect was observed between ST7-B and B. longum.

Fig. 1
figure 1

Interactions of Blastocystis with representatives of gut bacteria. Blastocystis ST7 isolates B and H were incubated for 24 h at 37 °C in PBS with each of the following bacterial cultures: E. coli, E. faecalis, B. longum, L. brevis, B. fragilis, and B. subtilis. There were higher counts of Blastocystis when incubated with bacteria (a). The highest increase in Blastocystis count was observed between isolate ST7-B and B. longum. The representative gut bacteria also had higher colony-forming unit per milliliter as observed (b) and counted (c) from agar plates, except for B. longum. B. longum’s colony-forming unit per milliliter was not significantly different in the presence of Blastocystis (c). *p < 0.05; **p < 0.01; ***p < 0.001

Full size image

Blastocystis exerted positive effects on some gut bacteria in vitro

The CFU counts of E. coli, E. faecalis, L. brevis, B. longum, B. subtilis, and B. fragilis were also examined when they were co-incubated with Blastocystis ST7-B and ST7-H. Gut commensal bacteria colony-forming unit per milliliter values were generally higher when co-incubated with Blastocystis cells. Representative images of the bacterial colonies on agar plates for the co-incubation assay are shown (Fig. 1b). E. coli, E. faecalis, B. fragilis, and B. subtilis had significantly higher CFU count when co-incubated with both ST7-B and ST7-H (Fig. 1c). L. brevis also had higher CFU count after co-incubation, but this was only significant for ST7-B. Interestingly, B. longum displayed lower CFU count when co-incubated with ST7-H. Average CFU count was moderately higher when co-incubated with ST7-B, but the differences did not reach statistical significance. An overall greater growth effect on CFU count was observed when ST7-B was co-incubated with gut commensal bacteria compared to ST7-H.

Blastocystis positively affects E. coli and negatively affects B. longum in a three-way co-culture setup

To investigate further if the effect of Blastocystis on gut bacteria is selective, a co-culture arrangement involving Blastocystis, E. coli, and B. longum was prepared. After incubation, E. coli had significantly higher colony-forming unit per milliliter compared to controls when incubated with B. longum. The presence of Blastocystis further increased the CFU of E. coli (Fig. 2a). On the other hand, B. longum displayed significantly lower colony-forming unit per milliliter compared to controls when incubated with E. coli, which was further reduced by both Blastocystis ST7-B and ST7-H (Fig. 2a).

Fig. 2
figure 2

Blastocystis inhibition of B. longum is linked to an increase in cellular ROS. B. longum exhibited lower colony-forming unit per milliliter when incubated with E. coli for 24 h at 37 °C in PBS. The count is even lower in the presence of Blastocystis (a). There is an increase in some oxidoreductase genes in B. longum when it is incubated with E. coli and Blastocystis indicating that the bacterium is under oxidative stress (b). Flow cytometry analysis shows B. longum cells’ shift to the right indicating more cells have cellular ROS content when co-incubated with E. coli, Blastocystis, or both (c). Blastocystis caused a greater increase in ROS content compared to E. coli (d). *p < 0.05; **p < 0.01; ***p < 0.001

Full size image

E. coli and Blastocystis caused oxidative stress to B. longum

As oxidative stress is a known contributor to dysbiosis [37], we explored whether Blastocystis and E. coli was impacting the viability of B. longum via such a mechanism. After incubation with Blastocystis and E. coli, the oxidoreductase gene expression of B. longum was analyzed. Results showed that two of the oxidoreductase genes, ferredoxin and ferridoxin, were upregulated when Blastocystis and E. coli were present (Fig. 2b). This suggests that the bacterium is undergoing oxidative stress in the presence of the organisms mentioned. Flow cytometry analysis of ROS content demonstrated that Blastocystis, E. coli, or both caused more B. longum cells to convert DCFDA stain indicating the presence of cellular ROS (Fig. 2c). Interestingly, the presence of Blastocystis alone caused significantly more production of ROS in B. longum than in combination with E. coli (Fig. 2d).

B. longum protects intestinal epithelial barrier against Blastocystis-induced damage

To determine the significance of B. longum for the host, HT-29 monolayers were grown and incubated with Blastocystis and B. longum. TEER measurements showed that B. longum can help maintain the epithelial barrier as observed from higher TEER compared to controls (Fig. 3a). Flux assays using FITC-dextran also showed that less number of reporter molecules can pass through the barrier when B. longum is present (Fig. 3b). Both assays showed that B. longum is beneficial to the host in maintaining the barrier and even negating the damage caused by Blastocystis. However, host factors induced by Blastocystis could inhibit B. longum growth. This is shown in co-culture assays whereby HT-29 monolayers were previously conditioned by Blastocystis. In these test wells, the colony-forming unit per milliliter count of B. longum was significantly lower than the control (Fig. 3c).

Fig. 3
figure 3

Host responses to Blastocystis-B. longum interaction. B. longum had a protective effect on the integrity of the epithelial barrier as observed on HT-29 monolayers incubated with Blastocystis. Monolayers had higher transepithelial electrical resistance values when B. longum is present (a) even in the presence of Blastocystis. Likewise, flux assay showed less FITC-Dextran molecules pass through the layer when B. longum is present (b). However, Blastocystis-induced host responses have a negative effect on B. longum as shown by the lower bacterial counts when this bacterium was incubated with Blastocystis-primed HT-29 monolayer (c). *p < 0.05; **p < 0.01; ***p < 0.001

Full size image

Blastocystis-infected mice had lower Bifidobacterium sp. and Lactobacillus sp. but higher E. coli abundance in the fecal samples

To determine whether Blastocystis infection alters the gut microbiota after acute infection in mice, fecal pellets were collected from various time-points before and after Blastocystis infection. The fecal pellets were subjected to qPCR to quantify the relative abundance of total bacteria, Bacteroides, Lactobacillus, Bifidobacterium, and E. coli. Bacterial relative abundance in the mouse fecal samples at equal weight was compared between the Blastocystis-infected mice with the control mice. The values were first normalized to the relative abundance found at their respective conditions before surgery. There was little to no difference in total bacteria between controls and Blastocystis-infected mice (Fig. 4). However, significant reduction in Bifidobacterium sp. was observed on day 3 after infection with Blastocystis ST7-B or ST7-H. Significant reduction in Lactobacillus sp. was also observed but only in fecal samples of ST7-H-infected mice at days 1 and 3 post-infection. E. coli had significantly higher abundance in ST7-B-infected mice on both days 1 and 2 post-infection. These observations suggest that, in general, Blastocystis can selectively influence gut microbiota populations, and in this case, could negatively affect beneficial bacterial populations.

Fig. 4
figure 4

Relative abundance of representatives of gut bacteria in mouse fecal samples. The mice were surgically infected with Blastocystis ST7-B and ST7-H isolates, and fecal samples were collected before surgery and 1-, 2-, and 3-day post infection. The DNA was extracted from 42 mg of samples and gut bacteria 16S rDNA genes were detected in qPCR. Analyses showed an increase in E. coli in mice infected with Blastocystis ST7-B at 1- and 2-day post infection. On the other hand, there was a decrease in Lactobacillus and Bifidobacterium in Blastocystis-infected mice. Blastocystis ST7-H isolated caused a decrease in Lactobacillus in mice after 1 and 3 days of infection. Blastocystis ST7-B and ST7-H caused a decrease in Bifidobacterium after 3 days of infection. *p < 0.05; **p < 0.01

Full size image

Histopathology examination revealed tissue damage in the colon of Blastocystis-infected mice

To further determine the effect of Blastocystis in mouse intestinal tissue, a histopathological examination and scoring was done on mouse colon and cecum tissue. The results showed damage and ulceration in the colon from Blastocystis-infected mice (Fig. 5a), with significantly higher pathological scores in the colon tissue of ST7-H-infected mice compared to control mice (Fig. 5b). The cecum did not appear to be damaged by Blastocystis.

Fig. 5
figure 5

Histological examination of mouse tissue after Blastocystis infection. a Histology images show the tissues with the highest histological scores within each treatment group. Epithelial lesions/ulcers and loss of crypt architecture were observed for the colon of mice infected with ST7-B (black arrows) and ST7-H (boxed area). Scale bar = 100 μm. b Dot plots show the histological scores for individual mice. Histological scores for ST7-H were significantly higher than that of ST7-B and the control in the colon. *p < 0.05

Full size image

Discussion

Although previous studies have reported associations between Blastocystis and gastrointestinal disorders, the protist’s pathogenic potential and clinical significance still remains to be established [38]. To address the issue of Blastocystis’ pathogenesis, it could be useful to determine whether Blastocystis colonization is associated with gut dysbiosis, which is known to affect intestinal health [4]. Various epidemiological studies have been executed in the past to investigate the associations between Blastocystis and dysbiosis, with conflicting results obtained [21,22,23,24,25, 39, 40]. One important limiting factor of some of these previous studies was that the subtype of Blastocystis, which has variations in terms of pathogenic potential, was not controlled for or identified [41, 42]. This study was therefore conducted with the aim of using a specific subtype (ST7) to study Blastocystis-gut bacteria interactions and to determine whether Blastocystis infection could disrupt the gut microbiota in vitro and in vivo.

Two Blastocystis ST7 isolates, ST7-B and ST7-H, were used in this study since previous reports indicate their pathogenic potential. In vitro assays on ST7 revealed that the isolate caused disruptions in the gut epithelial barrier by disrupting tight junction proteins such as occludin and zonula occludens-1 (ZO-1), and also have greater adhesiveness than ST4 isolates to intestinal epithelial cells [29, 43]. Furthermore, ST7 was shown to have significantly greater cysteine protease activity compared to ST4 [29]. Blastocystis ST7 have been shown to be more resistant to anti-parasitic drugs [28, 44] and against the host innate immune response [45] compared to ST1 and ST4 isolates. E. coli, E. faecalis, B. longum, L. brevis, B. fragilis, and B. subtilis were chosen for the co-incubation assay as representative species of the gut microbiota [46,47,48,49]. Among these, L. brevis and B. longum are well-known probiotic species which contribute protective benefits for the gut [49, 50]. More importantly, they have been found to improve intestinal conditions through the mitigation of IBD and IBS [51]. Although the other species (E. coli, E. faecalis, B. fragilis, B. subtilis) are not widely considered as probiotic, they are still important commensal gut bacteria, playing roles in carbohydrate metabolism, production of important metabolites, and the exclusion of potential pathogens [52]. Changes in the viability of these bacterial species (represented by CFU count in the assay) in the presence of Blastocystis may lead to potential disruptions in the microbiota. This may have important implications in relation to IBD and IBS.

The in vitro co-incubation assay demonstrated that Blastocystis cell count was higher in the presence of gut bacteria (Fig. 1a). It is possible that the bacteria secretory products or dead cells in the suspension act as a nutrient source for Blastocystis, allowing it to survive better when incubated in PBS. It is still unclear how Blastocystis obtains nutrients, but the mechanisms involved can be speculated. Although Blastocystis does not have true mitochondria, it has been found to possess mitochondrion-derived double-membrane-bound organelles called mitochondrion-like organelles (MLOs) [53]. These MLOs are likely hydrogenosomes, which are found in “amitochondriate protists,” which can play roles in carbohydrate and amino acid metabolism. Within the MLOs, two key enzymes involved in anaerobic energy metabolism have been identified, namely, pyruvate:ferredoxin oxidoreductase (PFO) and [FeFe] hydrogenase [54]. PFOs function in carbohydrate metabolism, catalyzing the conversion of pyruvate to acetyl-CoA and CO2 [55]. [FeFe] hydrogenases function in hydrogen metabolism [56]. These two enzymes may be activated in the presence of bacterial products, which may be utilized as nutrient sources for Blastocystis. In comparison, Blastocystis cells in the control may have been starved of nutrients when they are incubated in PBS. The co-incubation assay also demonstrated higher bacterial CFU count when gut commensal bacteria were co-incubated with Blastocystis (Fig. 1b, c). Higher bacterial CFU count may be a result of bacteria breaking down dead cells (from both Blastocystis and existing bacteria cells), in order to obtain a nutrient source in the PBS suspension. It was also observed that there were differential effects across bacterial species, with E. coli, E. faecalis, B. subtilis, and B. fragilis displaying more prominent positive effects compared to L. brevis and B. longum after co-incubation with Blastocystis. Blastocystis and gut commensal bacteria generally exhibit a mutualistic relationship when co-incubated in vitro, evidenced by higher parasite numbers and bacterial CFU counts after co-incubation. E. coli, E. faecalis, B. fragilis, and B. subtilis appeared to have more significant positive effects than L. brevis and B. longum. This observation suggests that these bacteria received less beneficial effects and may have a weaker mutualistic relationship with ST7-B and ST7-H in vitro. Overall, the co-incubation assays show that Blastocystis can interact with gut-commensal bacteria, which may lead to changes in the gut microbiota. Furthermore, there may be differential interactions depending on the species of bacteria present. Hence, this served as a basis for the subsequent in vivo assays.

A three-way setup involving Blastocystis, E. coli, and B. longum was used to further investigate if Blastocystis had a selective influence on specific groups of gut bacteria. Results showed that Blastocystis could boost the growth of E. coli while inhibiting B. longum (Fig. 2a). E. coli is a facultative anaerobe while B. longum is an obligately anaerobe. These interactions suggest oxidative stress as a factor in the experimental outcome [57]. Indeed, gene expression analysis of B. longum showed that some of the bacterium’s oxidoreductase genes are upregulated, suggesting that it is undergoing oxidative stress in the presence of Blastocystis and E. coli (Fig. 2b). In addition, a greater percentage of B. longum cells exhibited cellular ROS content when these were incubated with Blastocystis and E. coli (Fig. 2c, d). Our results, however, do not exclude other mechanisms of Blastocystis- and E. coli-mediated killing of B. longum. Possible implications of redox-mediated killing of obligate anaerobes would be decreased diversity in gut bacteria ultimately leading to dysbiosis [58].

The significance of B. longum in the context Blastocystis infection was explored using intestinal epithelial monolayer assays. Specifically, the role of B. longum as well as the effect of Blastocystis on the epithelial barrier integrity was investigated using TEER measurements and flux assay. These in vitro assays showed that B. longum helps to maintain the intestinal epithelial barrier (Fig. 3a, b), even in the presence of Blastocystis. This supports the notion that B. longum and similar groups of gut bacteria are essential for the health of the gut [51]. Past studies reported that the presence of Bifidobacterium attenuated the decrease in transepithelial electrical resistance and increase in paracellular permeability in Caco-2 cells treated with LPS. Bifidobacterium was also found to upregulate the expression of tight junction proteins occludin, claudin-3, and ZO-1 as well as aid the localization of these proteins to the epithelial tight junctions [58]. Aside from maintaining the epithelial barrier, Bifidobacterium can also exert anti-inflammatory properties as it can reduce the production of pro-inflammatory cytokines IL-6 and TNF-α [58]. In contrast, Blastocystis ST7 disrupts tight junction proteins such as occludin and ZO-1 [43, 59] as well as increases the levels of pro-inflammatory cytokines to trigger an inflammatory response [60, 61]. These show that Bifidobacterium can potentially negate the cytopathic effects of Blastocystis on the hosts. However, host-secreted factors that result from Blastocystis infection could be limiting to Bifidobacteria, as shown from the co-culture assays involving HT-29 cells previously conditioned by Blastocystis (Fig. 3c). Blastocystis therefore may not only affect Bifidobacteria directly but could also limit the bacterium through the host. These host factors may include elements of the innate immunity such as antimicrobial peptides. We have previously shown that Blastocystis can induce intestinal epithelial cells to secrete LL-37, a fragment of cathelicidin with antimicrobial properties [45]. These factors, however, have broad effects that do not only affect invading pathogens but could also impact local microbial populations when overly secreted. This added pressure could therefore result in lower diversity of microbial populations in the gut. Overall, the Blastocystis-Bifidobacterium-host epithelial cell interactions are complex and could involve numerous signaling and effector molecules. Our study reveals that ROS and host factors may play roles in limiting B. longum viability, providing new clues on how Blastocystis influences specific gut microbiota populations (Fig. 6).

Fig. 6
figure 6

Interactions of Blastocystis with gut bacteria and the effect on the host. Blastocystis could disrupt gut microbiota selectively. In this study, Blastocystis caused reduction of B. longum but an increase in E. coli. This could happen by several mechanisms. There is a direct effect of Blastocystis through oxidative stress, limiting the viability of obligately anaerobic bacteria. Host immune responses as induced by Blastocystis could also limit Bifidobacterium. This bacterium is important to protect the epithelial barrier from Blastocystis-mediated damage. Red and blue arrows signify negative and positive interactions respectively

Full size image

In this study, an acute infection of Blastocystis ST7-B and ST7-H on mice was performed to assess Blastocystis-induced changes in the gut microbiota using qPCR (Fig. 4). This study utilized a DSS colitis mouse model, which improves Blastocystis colonization rates, as previously demonstrated for ST7-B- and ST7-H-infected C57BL/6 mice treated with low concentrations of DSS [27]. Total bacteria levels and the relative abundance of Bacteroides, Lactobacillus, Bifidobacterium, and E. coli populations were quantified using qPCR after Blastocystis infection. The reduction was observed in Bifidobacterium in mice infected by ST7-B and ST7-H. There was also lower abundance observed in Lactobacillus in ST7-H-infected mice. Interestingly, there was a higher abundance of E. coli in ST7-B-infected mice. These results are in concordance with what has been obtained using in vitro assays; Bifidobacterium was reduced, and E. coli’s abundance increased. In the case of Lactobacillus, its reduction was only observed in vivo. It is possible that host factors, which are not present in the in vitro assays, come into play in the observed reduction of Lactobacillus. Histological examination of mouse tissues (Fig. 5) corroborated with our previous study [30]. The pathology scoring also identified ST7-H as more able to cause tissue damage than ST7-B. However, in this study, ST7-B appeared to be a better driver of dysbiosis than ST7-H. These observations point to differences between the isolates’ mechanism of pathogenesis, with ST7-H causing more direct damage to host cells and ST7-B causing harm through dysbiosis.

Together with Bifidobacterium spp. discussed above, Lactobacillus is another group of probiotic bacteria in the gut. Like Bifidobacterium, members of this genus also have similar anti-inflammatory properties [62]. An in vitro study showed that L. casei could reduce T cell response by dendritic cells in healthy and ulcerative colitis patients, thus decreasing the inflammation. This is achieved through increased production of IL-4 and decreased secretion in IL-22 and IFN-γ [63, 64]. Lactobacillus has also been found to significantly increase IgA levels [65]. Several in vivo studies using the DSS colitis mouse model showed that administration of both Lactobacillus and Bifidobacterium improved clinical symptoms of colitis and enhanced mucus production [66, 67]. Hence, a reduction in both gut bacteria would remove an element of protection from the gut epithelium, aiding the pathogenesis of Blastocystis. This could explain the intestinal tissue damage seen in the histology results of the current study. Epidemiological studies have also shown that reductions in these two bacteria could increase the susceptibility to gastrointestinal disorders. UC and CD patients were found to possess lower levels of Lactobacillus and Bifidobacterium populations, respectively [68, 69]. Similarly, Lactobacillus and Bifidobacterium levels are lower in IBS patients than that in healthy controls [70, 71]. Hence, the presence of Blastocystis ST7 could cause disease not only just directly but also through reduction of beneficial bacteria.

Various epidemiological studies have been conducted to investigate the links between Blastocystis and dysbiosis. Previous surveys have observed certain characteristics in the microbiota of Blastocystis-positive subjects, leading to the association of Blastocystis with a healthy human gut [21,22,23,24, 27]. A study found that Blastocystis-positive individuals free from IBD had higher fecal bacterial diversity, higher abundance of Clostridia, and lower abundance of Enterobacteriaceae [22]. A recent study done across 12 metagenomic datasets found a strong association between Blastocystis and the enrichment of Firmicutes and Clostridiales, as well as the reduction in Bacteroides [21]. Additionally, another group showed that Blastocystis is linked to a healthy gut, based on the high F. prausnitzii–E. coli ratio in Blastocystis-positive subjects [23]. These and other studies formed the basis for asserting that Blastocystis is a member of the normal, healthy gut microbiota [72, 73]. It is important to note that two of the mentioned studies did not identify the subtype of Blastocystis present in its subjects [22, 23]. For the study which did identify the subtypes, a whole array was found in the subjects, including ST1 and ST3 which are associated with asymptomatic infections [21]. The subjects from these studies could be predominantly colonized with subtypes associated with lower pathogenic potential and may be associated with healthy gut microbiota. However, this present study used a specific subtype, ST7, which exhibits high pathogenic potential. Therefore, the conflicting results may be due to the differential effects of various subtypes on the host and gut microbiota. It is therefore important for future studies to control or stratify for Blastocystis subtypes to avoid self-limiting results. The results of this study are in line with one epidemiological study, which concluded that Blastocystis is linked to dysbiosis [25]. In that study, Blastocystis-positive patients with IBS-C had a significant decrease in Bifidobacterium and Lactobacillus populations. It was also noted that the subtype of Blastocystis involved in the study by Nourisson et al. was predominantly ST4. Compared to other STs except ST7, ST4 has a moderate pathogenic potential [29], and this could be the reason that Blastocystis was associated with dysbiosis, unlike the other studies. Overall, the findings from Nourisson et al. corroborate with the results of this study, both suggesting that virulent subtypes of Blastocystis are more likely to be associated with dysbiosis, and its pathological outcomes, including IBD and IBS.

Conclusion

Overall, this study investigated the interactions of pathogenic isolates of Blastocystis ST7 with known members of the gut microbiota. To our knowledge, this is the first time wherein in vitro setups complemented by an in vivo system were utilized to investigate the interactions of Blastocystis with the gut microbiota. In addition, this study also focused on a specific ST of Blastocystis. While most reports on Blastocystis label it as a commensal and a member of healthy gut microbiota, the findings in this study indicate that different ST of Blastocystis, represented by two pathogenic isolates, may modulate gut microbiota differently from more common STs (e.g., ST1–3). Future work should include other Blastocystis STs with lesser pathogenic potential as well as involving more representatives of gut bacteria. This should provide a clearer picture on where Blastocystis and its STs really stand on gut health and disease.