Background
The protozoan parasite,
Giardia lamblia (synonymous with
G. duodenalis, G. intestinalis), is the most commonly identified agent causing waterborne diarrheal disease worldwide.
G. lamblia is a strictly lumen-dwelling pathogen characterized by trophozoites that adhere to and colonize the small intestine, resulting in symptoms of diarrhea, abdominal pain, and weight loss in humans. In some cases, giardiasis can present as a self-limiting acute infection. Recent evidence suggests that acute
Giardia infection may lead to the development of chronic functional gastrointestinal disorders, such as post-infectious irritable bowel syndrome (IBS) and functional dyspepsia, by unknown mechanisms [
1,
2]. Follow-up studies of a waterborne outbreak of giardiasis in Norway have reported that 40–80% of patients experienced abdominal symptoms consistent with IBS after the eradication of parasites [
3‐
6]. The percentage of patients who developed post-infective gut disorders following giardiasis is higher than those following bacterial or viral gastroenteritis (3–30%) [
1,
7].
Intestinal pathologies in giardiasis include villous atrophy, brush border shortening, and epithelial barrier disruption in infected humans and animals [
8‐
10]. Tight junctional damage has been demonstrated in epithelial cell cultures following exposure to live
Giardia or trophozoite lysates [
10‐
12]. Besides direct parasitic effects, host immune factors responding to
Giardia infection are responsible, in part, for increased intestinal permeability [
8,
13‐
16]. It remains unclear whether these gut barrier defects are corrected after parasitic expulsion.
The gut lumen is the largest reservoir of commensal bacteria in the human body. It harbors an estimated 100 trillion bacteria that gradually increase along proximal to distal segments, reaching their highest number in the colon [
17‐
19]. Bacterial overgrowth in the small intestine was documented in an early clinical report of giardiasis in which 8 of 17 patients showed increased bacterial counts in jejunal aspirates [
20]. Whether colonic bacterial numbers are affected by parasitic infection of the small intestine remains unknown. Although
Giardia parasites do not disseminate beyond the epithelial layer during colonization, we speculate that infection with
Giardia may trigger long-lasting changes in the spatial distribution of commensal microorganisms and induce microbial penetration and inflammatory responses in the gut mucosa during the post-clearance phase.
Discussion
Accumulating evidence indicates that Giardia infection may lead to chronic, post-infective gastrointestinal dysfunction for unknown reasons. The present study detected persistent gut barrier damage and commensal bacterial influx, accompanied by long-term mucosal inflammation, during the post-clearance phase of Giardia infection. These novel findings suggest that the host’s unresolved immune response toward its own microbiota, due to an impaired epithelial barrier, may partly contribute to the development of post-infective gut disorders.
In our mouse model of
G. lamblia infection, peak colonization was detected during the first week, with parasite clearance occurring 2–3 weeks after inoculation. Consistent with other reports, the infection course of
G. lamblia is much shorter than that of
G. muris in mice, which lasts for more than 6 weeks [
16,
21,
22]. Instead of murine-specific
G. muris, a human isolate of
G. lamblia was used in our model as a zoonotic pathogen to mimic the intestinal pathology in humans. Murine models of
G. lamblia infection have been utilized to investigate the pathogenesis of villus blunting and epithelial brush border shortening, which are mechanisms responsible for diarrhea, malabsorption, and malnutrition in human patients [
23‐
26].
Based on our observed time course of
G. lamblia colonization in mice, subsequent experiments were conducted on PI day 7 (colonization period) and PI day 35 (post-clearance phase) to assess gut barrier changes. Consistent with previous studies that demonstrated tight junctional disruptions in giardiasis [
9,
10,
12], we detected occludin cleavage in the jejunum of
G. lamblia-infected mice during peak parasite colonization. Tight junctional damage in the small intestine persisted during the post-clearance phase, when parasites were no longer detected. Interestingly, colonic tissues, which did not show increased macromolecular flux during peak parasite colonization, displayed heightened epithelial permeabilities after the elimination of parasites. This finding suggests that post-infective host factors may be involved in the mechanism of barrier dysfunction.
During
Giardia colonization, increased gut-associated bacterial counts were accompanied by enhanced bacterial adherence to and invasion of the small and large intestines, suggesting that parasite infection of the proximal small intestine may lead to bacterial overgrowth and internalization throughout the intestinal tract. Our data indicate that colonization with non-invasive
Giardia triggered an influx of commensal bacteria, possibly via impaired tight junctions. Recent cell culture studies have suggested that commensal bacteria can pass through paracellular spaces or be engulfed into enterocytes in the presence of other pathogenic invasive bacteria (
e.g. Campylobacteri jejuni or
Salmonella enterica) [
27,
28]. Taken together, colonization by enteropathogens, whether invasive or non-invasive, may affect the luminal confinement of commensals during infection.
The present findings also demonstrate that the phenomena of tight junctional damage and mucosal bacterial invasion persist into the post-clearance phase (at least 5 weeks after the eradication of parasites). It is noteworthy that gut-associated bacterial counts, which mainly include bacteria in the mucus layer, were not higher among post-infective mice, suggesting that increased bacterial penetration may not necessarily correlate with bacterial overgrowth. We identified several bacterial strains (e.g. Bacillus, Staphylococcus, Lactobacillus, and Phenylobacterium) in the mucosal endocytosed colonies. Whether these invasive bacteria reflect a compositional change in commensals requires further investigation.
A key role of microbiota in the onset or perpetuation of intestinal inflammation has been implicated in models of enterocolitis using gnotobiotic animals monoassociated with particular bacterial strains or with germ-free versus conventionally raised animals [
29,
30]. These studies have indicated that chronic gut inflammation may be driven by inappropriate and excessive mucosal immune responses against the host’s own commensal microflora. Aberrant microbiota compositions have been documented in patients with inflammatory bowel disease (IBD) and IBS [
31‐
33]. The emergence of mucosal adherent and invasive bacteria has also been reported in cases of IBD, surgical stress, and necrotizing enterocolitis [
34‐
36]. Whether the composition of commensal bacteria is altered in
Giardia-infected mouse models remains unknown. Short- and long-term alterations in microfloral diversity associated with giardiasis warrant further investigation.
The activation of immune factors, such as T cells, neutrophils, mast cells, secretory IgA, proinflammatory cytokines, and nitric oxide, have been reported in
Giardia-infected humans and mice [
14,
15,
22,
23,
37‐
39]. A recent report described long-term CD4
+ T cell proliferative responses in patients 5 years after acute giardiasis [
40], suggesting unresolved immune activation during the post-clearance phase. We examined the mucosal inflammatory response in post-infective mouse intestines and identified increased phagocytic infiltration and activation as well as IFNγ and IL-8 production during infection, which persisted into the post-clearance phase. TNFα and IL-1β levels were only upregulated after parasite elimination, whereas no change was detected in IL-6 and IL-12 levels. Previous studies have suggested that the attenuated inflammatory response during
Giardia infection, which was linked to the activity of parasitic proteases, was a mechanism for
Giardia to evade host immune responses [
41]. We speculate that the post-infective proinflammatory cytokine increase may be caused by secondary bacterial influx in the absence of parasites. The differential patterns of proinflammatory cytokines throughout the course of infection and their relationships with commensal bacterial penetration require further investigation.
Unresolved inflammation also may trigger bacterial influx through either transcellular or paracellular routes. Previous reports have documented tight junctional disassembly and increased paracellular permeability by neutrophil migration [
42], or by synergistic effects of IFNγ and TNFα [
43]. Recent studies have shown that stimulation with low-dose IFNγ or TNFα was able to induce epithelial internalization and endocytosis of nonpathogenic bacteria [
44,
45]. Taken together, abnormal bacterial entry and mucosal inflammation may be two events that aggravate each other in chronic post-infective disorders.
Materials and Methods
Animals
Specific pathogen free BALB/c mice (4–6 weeks old) were obtained from the Animal Center of National Taiwan University. Animals were raised in a temperature-controlled room (20 ± 2°C) with 12/12-h light/dark cycles and were fed regular mice chow and water ad libitum. All experimental procedures were approved by the Animal Care and Use Committee of National Taiwan University.
Giardia infection protocol
Axenic
G. lamblia GS/M trophozoites (ATCC 50581) were cultured
in vitro and harvested at log-phase as previously described [
14,
22]. Mice received 10
7G. lamblia trophozoites suspended in 0.2 ml of sterile phosphate-buffered saline (PBS), or the same volume of PBS alone, by orogastric gavage. To verify the infection time course, mice were sacrificed at various time points (PI day 0, 4, 7, 14, 21, 28, 35, and 49) for the enumeration of trophozoites in the small intestine [
8,
10]. Briefly, the entire length of the small intestines were opened longitudinally and placed in 10 ml of ice-cold saline for 3 min. Trophozoites were detached from the tissues by vigorous shaking and were counted using a hemocytometer. The detection limit was ~2 × 10
3 trophozoites per small intestine. Samples of intestinal tissues, liver and spleen were collected for additional analyses on PI days 7 and 35.
Histological examination
Intestinal segments were fixed in 4% paraformaldehyde and were carefully embedded in paraffin wax to ensure proper orientation of crypt to villus axis. Sections of 4 μm thickness were stained with hematoxylin and eosin (H&E) and observed under a light microscope (Zeiss, Göttingen, Germany). Photoimages were captured using AxioVision Rel. 4.6 software (Zeiss) [
46,
47].
Western blotting for occludin cleavage
Scraped intestinal mucosal samples were homogenized in ice-cold complete radio-immunoprecipitation assay (RIPA) buffer, and lysates were sonicated and centrifuged. Supernatant protein concentrations were determined using a protein assay (Bio-Rad, Hercules, CA), adjusted to 5 mg/ml, and diluted 1:1 (v/v) in 2× electrophoresis sample buffer containing 2% (w/v) sodium dodecyl sulfate (SDS), 100 mM dithiothreitol (DTT), and 62.5 mM Tris/HCl (pH 6.8). Samples then were heated to 95°C in a heat block for 5 min and were stored at −20°C until immunoblotting. The extracted proteins were resolved by SDS-polyacrylamide gel electrophoresis, and electrotransferred onto membranes. After blocking with 5% non-fat milk in Tris buffer saline (TBS), membranes were incubated with monoclonal mouse anti-occludin (1:1000; Invitrogen, Carlsbad, CA) at 4°C overnight. Monoclonal mouse anti-β-actin (1:10,000; Sigma, St. Louis, MO) was used to control for equal loading of each sample. Membranes were washed with 0.1% Tween 20 in TBS and were incubated with horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (1:1000; Cell Signaling, Danvers, MA). Antigens were revealed and band densities were quantified by photoimage analysis [
48,
49].
Intestinal permeability assay
Intestinal tissues were opened along the mesenteric border and mounted in Ussing chambers (World Precision Instruments, Sarasota, FL) [
46‐
48]. The opening of each chamber (0.7 cm
2) exposed the tissue to 5 ml of circulating oxygenated Krebs buffer. The serosal buffer contained 10 mmol/l of glucose that was osmotically balanced with 10 mmol/l of mannitol in the mucosal buffer. A circulating water bath was used to maintain the temperature of the buffer at 37°C. Intestinal permeability was determined as the level of mucosal-to-serosal flux of dextran conjugated to fluorescein isothiocyanate (dextran-FITC, MW = 4 kDa; Sigma). The dextran probe was added to the mucosal buffer at a final concentration of 500 μM. Samples (250 μl) of serosal buffer were collected at 0, 30, 60, and 90 min after addition of the dextran probe and were replaced with Krebs buffer/glucose. The fluorescence units of dextran-FITC in serosal buffer were determined at excitation/emission = 490/530 nm using a multimode plate reader (Paradigm Detection Platform; Beckman Coulter, Indianapolis IN), and dextran-FITC concentrations (nM) were calculated according to a standard curve.
Measurement of intestinal bacteria numbers
Gut-associated, superficial, and mucosal endocytosed bacterial counts were determined in mice. To evaluate gut-associated bacterial numbers, 1-cm segments of the mid-jejunum and colon were excised and cut longitudinally, and luminal contents were rinsed off with sterile saline. Tissues were weighed aseptically and transferred to PBS at a ratio of 1 mg to10 μl for homogenization and sonication. Tissue lysates were plated onto fresh blood agars (Scientific Biotech, Taipei, Taiwan) for aerobic and anaerobic bacterial culturing overnight at 37°C. Colony-forming units (CFU) were normalized to grams of intestinal tissues (CFU/g) to represent gut-associated bacterial counts [
48].
To examine superficial adherent bacterial numbers, 5-cm intestinal segments were excised longitudinally, and luminal contents were rinsed off with sterile saline prior to weighing aseptically. Intestinal tissues were cut into 1-cm pieces and were incubated in 30 ml of sterile PBS with 0.016% DTT (Sigma) at room temperature for 15 min to remove mucus. Tissues were transferred to sterile saline (1:100 w/v) and vortexed for 30 s. Tissues were washed three times with PBS, and the last solution was plated onto fresh blood agars (Scientific Biotech) for aerobic bacterial culturing. CFUs were normalized to grams of intestinal tissues (CFU/g) to represent superficial bacterial counts [
50].
To examine mucosal endocytosed bacterial counts, 10-cm-long segments of mid-jejunum (distal of ligament of Triez) and 5-cm segments of colon were cut longitudinally and rinsed in sterile PBS. Tissues were cut into 1-cm portions, incubated in 10 ml of sterile PBS with 1 mM ethylenediaminetetraacetic acid (EDTA) and 5.55 mM glucose (Sigma) at 37°C for 10 min, and then were shaken vigorously. Tissues in solution were passed through nylon mesh (BD Biosciences, San Jose, CA) with a pore size of 80 μm, and single-celled eluates were collected by centrifugation for hemocytometer counting. Our flow cytometry pilot study indicated that 64.0 ± 2.8% of isolated cells were BerEP4
+ (epithelial marker), 22.0 ± 0.6% were CD68
+ (macrophage marker), and 38.7 ± 2.0% were CD3
+ (T cell marker). The amount of intracellular bacteria was determined using a gentamicin protection assay, as described previously [
51]. Briefly, single cells were incubated with 300 μg/ml of gentamicin (Invitrogen) for 1 h with gentle shaking to kill extracellular bacteria. Cells then were washed twice with sterile PBS and were incubated with 1% Triton X-100 in PBS for 10 min on ice at a concentration of 5×10
6 cells/ml. Lysates (200 μl) were plated onto fresh blood agar (Scientific Biotech) and were incubated overnight at 37°C. Bacterial CFUs were normalized to CFU/10
6 cells to represent mucosal endocytosed bacterial counts.
Analysis of bacterial translocation to extra-intestinal organs
Livers and spleens were removed from animals using sterile instruments and weighed. Tissues were homogenized and sonicated in sterile PBS at a ratio of 1 mg to 10 μl. Undiluted lysates (200 μl) were inoculated onto Luria-Bertani (LB) agar plates with or without 50 μg/ml ampicillin for aerobic bacterial culturing. Following incubation at 37°C for 24 h, bacterial colonies were counted and normalized to CFU/g [
46,
48].
Gram staining
Mucosal endocytosed bacterial colonies grown on agar plates were rapidly classied by Gram staining. Crystal violet was used for peptidoglycan staining, followed by rapid decolorization with alcohol and counterstaining with safranin. All reagents were purchased from Sigma.
Bacterial 16S rDNA sequencing
Mucosal endocytosed bacterial colonies were picked for DNA extraction using a QIAamp DNA Stool Mini kit (Qiagen, Hilden, Germany). Extracted DNA samples were amplified by polymerase chain reaction (PCR) using a T3000 Thermocycler (Biometra, Göttingen, Germany) with universal bacterial primers detecting the 16S rDNA gene (341F and 534R of
Escherichia coli) [
52]. After amplification, PCR products were sent for sequencing (Genomics BioSci & Tech, Taipei, Taiwan), and sequences were classified using the naïve Bayesian classifier provided by the Ribosomal Database Project (RDP;
http://rdp.cme.msu.edu/classifier/classifier.jsp) [
53]. Sequences from 16S gene libraries were submitted to the DNA Data Bank of Japan (DDBJ) and were cataloged under the DDBJ accession numbers AB777576–AB777579.
Fluorescence in situ hybridization (FISH)
Microbiological analysis of intestinal tissues was conducted by FISH, with some modifications to previously described protocols [
54]. Briefly, tissues were fixed in Carnoy’s solution (Ricca Chemical Company, Arlington, TX), embedded in paraffin, and sectioned at 5 μm. Sections were dewaxed, placed in 1% Triton X-100 for 90 s, washed three times in PBS, and incubated in 5 mg/ml lysozyme at 37°C for 20 min. Sections were washed in PBS three times and were incubated in prewarmed hybridization solution (12% formamide, 20 mM Tris–HCl, 0.9 M NaCl, 0.01% SDS; pH 7.4) containing 0.02–0.1 μM oligonucleotide probes targeting bacterial 16S rDNA at 46°C overnight. The probes included a FITC-labeled universal bacterial probe (EUB-338; 5'-GCTGCCTCCGTAGGAGT-3') and a negative-control probe (Non-338; 5'-CGACGGAGGGCATCCTCA-3') (Genomics BioSci & Tech). Sections then were rinsed in washing buffer (20 mM Tris–HCl, 0.9 M NaCl, 0.01% SDS; pH 7.4) and air-dried prior to staining with Hoechst dye. Slides were mounted and viewed under a fluorescent microscope (Axio Imager A1; Zeiss) equipped with a CCD camera.
Neutrophil-specific esterase staining of intestinal mucosa
Neutrophils in the intestinal mucosa were detected in paraffin-embedded tissue sections using a naphthol AS-D chloroacetate (neutrophil-specific esterase) staining kit (Sigma) [
47]. The substrate, naphthol AS-D chloroacetate, is hydrolyzed by a specific esterase in neutrophils, liberating a free naphthol that couples with a diazonium salt and forms dark purple deposits at sites of enzymatic activity. Tissue sections were stained according to the manufacturer’s instructions, and photoimages were taken using AxioVision Rel. 4.6 software.
Myeloperoxidase (MPO) activity assay
Intestinal segments were opened longitudinally, rinsed, and weighed. The tissues then were suspended in 0.5% hexadecyltrimethylammonium bromide (Sigma) in 50 mM potassium phosphate buffer (PPB; pH 6) at a ratio of 1 g tissue to 10 ml buffer. Samples then were homogenized and sonicated on ice. Lysates were centrifuged at 12,000 ×
g for 20 min at 10°C, and the resultant supernatants were collected. Supernatants (7 μl) were diluted with 200 μl of reactive buffer (PPB solution containing 0.167 mg/ml of
o-dianisidine dihydrochloride and 0.0005% of H
2O
2) in 96-well plates. Enzyme concentrations were determined as the absorbances at 460 nm measured every 50 s during a 5-min period. One unit of MPO activity was defined as the quantity of enzyme degrading 1 μmol of H
2O
2/min at 25°C, expressed as unit/mg tissue [
47].
Total and differential leukocyte counts
Blood was drawn by cardiac puncture with a heparinized syringe into EDTA-containing, pyrogen-free vacutainer tubes (BD Biosciences). Total leukocyte numbers and hematocrit levels in whole blood were determined using a hematology analyzer (Medonic CA-620, Sweden). For differential leukocyte counts, 10-μl blood samples were spread in thin layers on micro-slide glass and stained with Liu’s stain (ASK, Taipei, Taiwan). Differential counts of neutrophils, monocytes, and lymphocytes were determined based on cell morphology under a light microscope.
Enzyme-linked immunosorbent assay (ELISA)
Jejunal segments (1 cm in length) were weighed and homogenized at a ratio of 1 g tissue to 10 ml PBS which was dissolved with one Complete Mini protease inhibitor cocktail tablet (Roche, Mannheim, Germany). Samples then were sonicated on ice. Lysates were centrifuged at 12,000 × g at 4°C for 30 min, and supernatants were collected. Supernatant protein concentrations were measured using a protein assay (Bio-Rad). Cytokine levels in mucosal samples were measured with ELISA development kits according to the manufacturer’s instructions (PeproTech, Rocky Hill, NJ). Briefly, microplates were coated overnight with capture antibodies. Plates were blocked with PBS containing 1% BSA for 1 h and then washed. Sample and standard solutions were added, and plates were incubated for 2 h. Biotinylated antigen-affinity detection antibodies were added, and plates were incubated for an additional 2 h. After washing, an avidin-HRP conjugate was added for 30 min followed by incubation with 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) liquid substrate for color development. Absorbance was measured at 405 nm with correction set at 650 nm. Cytokine levels in intestinal mucosa were expressed as pg cytokine/mg total protein.
Statistical analysis
All values, except for bacterial CFUs, were expressed as means ± SEM, and means were compared by one-way analysis of variance followed by a Student Newman-Keuls test. For bacterial count data, pairwise rankings of median CFUs were performed using the nonparametric Mann–Whitney U test. Differences were considered statistically significant at P < 0.05.
Acknowledgements
We thank the staff of the imaging facility at the First Core Lab, and the AM2 Pharmacological Studies of Drugs in Animal Models for Human Diseases, Disease Animal Research Center, National Taiwan University College of Medicine, for their technical assistance.
Sources of support
This study was supported by grants from the National Research Program for Biopharmaceuticals (NRPB), National Science Council (NSC100-2325-B-002-035, 101-2325-B-002-035-031, 102-2325-B-002-035-030).
Competing interest
The authors declare that there are no conflicts of interest.
Authors' contributions
Guarantor of integrity of entire study, LCY; study concepts and design: LCY and AGB; data acquisition: TLC, HC, HWW, YZL, and LLW; data analysis/interpretation: TLC, HC, HWW, and TCL; statistical analysis: TLC, HWW; material and technical support: YHN, CHS, WHY; obtained funding: LCY; manuscript drafting or revision for important intellectual content, literature research, manuscript editing, and manuscript final version approval: all authors.