ABSTRACT
Giardia duodenalis (syn. G. intestinalis, G. lamblia) infections are a leading cause of waterborne diarrheal disease that can also result in the development of postinfectious functional gastrointestinal disorders via mechanisms that remain unclear. Parasite numbers exceed 106 trophozoites per centimeter of gut at the height of an infection. Yet the intestinal mucosa of G. duodenalis-infected individuals is devoid of signs of overt inflammation. G. duodenalis infections can also occur concurrently with infections with other proinflammatory gastrointestinal pathogens. Little is known of whether and how this parasite can attenuate host inflammatory responses induced by other proinflammatory stimuli, such as a gastrointestinal pathogen. Identifying hitherto-unrecognized parasitic immunomodulatory pathways, the present studies demonstrated that G. duodenalis trophozoites attenuate secretion of the potent neutrophil chemoattractant interleukin-8 (CXCL8); these effects were observed in human small intestinal mucosal tissues and from intestinal epithelial monolayers, activated through administration of proinflammatory interleukin-1β or Salmonella enterica serovar Typhimurium. This attenuation is caused by the secretion of G. duodenalis cathepsin B cysteine proteases that degrade CXCL8 posttranscriptionally. Furthermore, the degradation of CXCL8 via G. duodenalis cathepsin B cysteine proteases attenuates CXCL8-induced chemotaxis of human neutrophils. Taken together, these data demonstrate for the first time that G. duodenalis trophozoite cathepsins are capable of attenuating a component of their host's proinflammatory response induced by a separate proinflammatory stimulus.
INTRODUCTION
Giardia duodenalis (syn. G. intestinalis, G. lamblia) is a noninvasive protozoan parasite of the upper small intestine of many animals, including humans, which is now included in the WHO's Neglected Disease Initiative (1). Giardiasis is a leading cause of waterborne diarrheal disease worldwide, and the infection is known to result in the development of postinfectious functional gastrointestinal disorders, as well as in extraintestinal complications (2, 3). G. duodenalis is currently subdivided into eight distinct genetic assemblages designated A through H, with assemblages A and B isolates being infective to humans (4, 5). There is an ongoing discussion as to whether at least some G. duodenalis assemblages may represent distinct Giardia species (6, 7). At the height of Giardia infections, millions of trophozoites closely associate with the apical surface of the intestinal epithelium and induce pathophysiological responses that can culminate in malabsorptive diarrheal disease (reviewed in reference 8). With the exception of the small but significant increase in numbers of intraepithelial lymphocytes (9, 10), acute infection with G. duodenalis is not associated with the infiltration of inflammatory cells, for reasons that remain obscure. These observations represent a counterintuitive observation not only in view of the direct presence of large numbers of parasites but also because G. duodenalis breaks the epithelial barrier via direct effects on tight junctional proteins (11–14) and therefore likely facilitates the translocation of potent proinflammatory luminal antigens. In addition, G. duodenalis infections can occur concurrently with other proinflammatory gastrointestinal pathogens, such as Cryptosporidium parvum (15), Helicobacter pylori (16), rotavirus (17, 18), and Salmonella (19).
To date, little research has focused on the ability of G. duodenalis to modulate host intestinal proinflammatory responses and the recruitment of proinflammatory immune cells induced by coinfecting proinflammatory gastrointestinal pathogens. Previous studies have demonstrated that mast cell hyperplasia occurs in the late stages of a Giardia infection or following parasite clearance (20, 21) and that eosinophil accumulation may occur in vivo in an isolate-dependent manner (22). Furthermore, G. duodenalis parasite products have been shown to modulate dendritic cell responses to lipopolysaccharide (23, 24), while separate studies have demonstrated that intestinal epithelial cells (IECs) exposed to Giardia trophozoites produce a unique chemokine profile (25). The need to investigate G. duodenalis's ability to modulate its host's proinflammatory responses is now apparent due to data collected from in vivo animal and human studies. Microarray analysis of jejunal tissues collected from assemblage E Giardia-infected calves revealed decreased mRNA expression of several proinflammatory mediators and increased expression of anti-inflammatory transcription factors (26). Several human studies have suggested that Giardia infections in children may reduce the incidence or severity of diarrheal disease (17, 27, 28). One study demonstrated that Tanzanian children infected with Giardia had a reduced likelihood of developing fever and had lower levels of serum C-reactive protein, a classic marker of inflammation, than did their noninfected counterparts (27). A separate study suggested that children coinfected with rotavirus and Giardia displayed a marked reduction in the severity of diarrheal disease compared to children infected with only rotavirus (17). However, findings from the latter study directly conflict with findings obtained by other authors who were unable to find a reduction in the severity of diarrheal disease in children coinfected with rotavirus and Giardia (18). These contrasting studies may suggest that Giardia-mediated protection from diarrheal disease may be isolate specific, reliant on host genetics, or due to other, yet-to-be-determined factors.
As G. duodenalis trophozoites are noninvasive, the intestinal epithelium represents the primary point of contact between parasite and host; this structure is comprised of a single, polarized layer of intestinal epithelial cells that function to separate the external environment of the intestinal lumen from underlying host tissues (reviewed in reference 29). The intestinal epithelium is also involved in the induction of acute inflammatory responses within the intestinal mucosa, with an important function being the secretion of proinflammatory chemokines (reviewed in references 30 and 31). In response to a variety of proinflammatory stimuli, including direct exposure to translocated bacterial antigens, intestinal epithelial cells secrete different classes of chemokines, including the potent neutrophil/polymorphonuclear leukocyte (PMN) chemoattractant interleukin-8 (CXCL8) (32–34); this intermediary chemokine recruits extravasated PMNs to the basolateral membrane of the intestinal epithelium so that subsequent signals can, if necessary, drive their migration across to the apical surface of the intestinal epithelium (35–37). During an acute intestinal inflammatory response, PMNs are typically the first leukocyte to exit the vasculature and be recruited to the inflammatory site (reviewed in reference 38).
Cathepsin cysteine proteases are defined by a catalytic dyad comprised of active-site cysteine (Cys) and histidine (His) residues and classified as clan CA cysteine proteases; these are further subdivided into superfamilies, including cathepsin L (catL)-like or cathepsin B (catB)-like proteases (reviewed in reference 39). catB proteases contain an additional 20-amino-acid insertion, referred to as the occluding loop, containing two characteristic His residues; this insertion enables their function as an endo- or exopeptidase (40). The G. duodenalis genome contains genes for numerous cathepsin cysteine proteases, the majority of which have no described function (41, 42). However, several G. duodenalis cathepsin proteases are upregulated upon exposure to IECs (43). While some parasites may use cathepsin cysteine proteases to evade or modulate their host's immune responses (reviewed in references 44 and 45), previous research has demonstrated that Entamoeba histolytica cysteine protease 2 (EhCP2) cleaves CXCL8 into a more potent isoform that enhances PMN chemotaxis (46). We hypothesized that Giardia cathepsin cysteine proteases may be involved in modulating acute proinflammatory responses within the intestinal epithelium. Specifically, we hypothesized that G. duodenalis cathepsin proteases are capable of attenuating IEC-induced CXCL8 secretion and the resulting PMN chemotaxis, which explains, at least in part, the lack of overt inflammation in the intestine during infection. Our study is the first to describe a role for G. duodenalis catB proteases in degrading intestinal epithelial proinflammatory secretion of CXCL8 following exposure to host- or pathogen-derived proinflammatory stimuli. Furthermore, degradation of CXCL8 by G. duodenalis catB proteases resulted in attenuation of CXCL8-induced PMN chemotaxis.
MATERIALS AND METHODS
Ethics statement.All studies involving human small intestinal mucosal biopsy tissues were approved by the Conjoint Health Research Ethics Board (CHREB) at the University of Calgary and the Calgary Health Region. In accordance with CHREB guidelines, adult subjects used in this study provided informed, written consent and a parent or guardian of any child participant provided informed, written consent on their behalf.
Reagents.Recombinant human interleukin-1β (IL-1β), CXCL8, and all corresponding enzyme-linked immunosorbent assays (ELISAs) were purchased from R&D Systems. The broad-spectrum, clan CA membrane-permeative cysteine protease inhibitor (2S,3S)-trans-epoxysuccinyl-l-leucylamido-3-methylbutane ethyl ester (E-64d) (47) was purchased from Sigma-Aldrich. The membrane-permeative, catB-specific inhibitor l-3-trans-(propylcarbamoyl)oxirane-2-carbonyl)-l-isoleucyl-l-proline methyl ester (Ca-074Me) (40), the catB/L fluorgenic substrate benzyloxycarbonyl-l-phenylalanyl-l-arginine 4-methyl-coumaryl-7-amide (ZFR-AMC), and the catB fluorogenic substrate benzyloxycarbonyl-l-arginine-l-arginine 4-methyl-coumaryl-7-amide (ZRR-AMC) were purchased from Peptides International (48, 49). QIAzol, RNeasy RNA extraction kits, QuantiTect reverse transcription kit, QuantiFast SYBR green PCR kits, and human CXCL8 (GenBank accession no. NM_000584; Qiagen catalog number PPH00568A) and β-2 microglobulin (β2M) (GenBank accession no. NM_004048; Qiagen catalog number PPH01094E) qPCR primers were purchased from Qiagen. The validity of these primers has been validated previously (50, 51).
Human biopsy tissues and cell lines.Adapting a previous protocol (52), small intestinal mucosal biopsy tissues were obtained from the terminal ileum of patients with Crohn's disease (CD) in remission or, in separate experiments, areas of active inflammation. Samples were washed three times in Dulbecco's phosphate-buffered saline (PBS; Sigma-Aldrich) containing 0.016% 1,4-dithioerythritol (Sigma-Aldrich) to remove loosely adherent mucus and bacteria, followed by one wash in PBS. Washed biopsy tissues were placed into 96-well plates and incubated in 300 μl of Opti-MEM (Life Technologies) at 37°C, 5% CO2, and 96% humidity. Biopsy tissue weights were, on average, approximately 30 mg. The human adenocarcinoma Caco-2 cell line (ATCC HTB-37) was grown in ATCC complete growth medium comprised of minimum essential medium Eagle (MEME) (Sigma-Aldrich; M5650) supplemented with 100 g/ml of streptomycin, 100 U/ml of penicillin, 200 mM l-glutamine, 5 mM sodium pyruvate, and 20% heat-inactivated fetal bovine serum (FBS; VWR). Cells were passaged at 80% confluence with 2× trypsin-EDTA and seeded onto 6-well plates, 12-well plates with 0.4- or 3.0-μm Transwells (Corning), or small petri dishes pretreated with poly-l-ornithine (Sigma-Aldrich). Cells were maintained at 37°C, 5% CO2, and 96% humidity and media replaced every 2 to 3 days. Cells were used between passages 22 and 34.
Transepithelial resistance.Caco-2 cells were grown to confluence (>500 Ω) on 3.0-μm Transwell filter units and treated according to experimental design. Transepithelial electrical resistance was recorded with an electrovoltometer (World Precision Instruments).
Parasites.All Giardia duodenalis trophozoite isolates used in this study were previously obtained and isolated. Giardia duodenalis NF trophozoites were obtained from a water sample during an outbreak of giardiasis in Newfoundland, Canada (53), WB trophozoites (ATCC 30957) were obtained from a symptomatic patient with chronic giardiasis (54), and GS/M clone H7 trophozoites (ATCC 50581) were isolated during a previous study (55). Trophozoites were grown axenically at 37°C in Keister's modified TYI-S-33 medium (56, 57) supplemented with piperacillin (Sigma-Aldrich) in 15-ml polystyrene tubes (Becton, Dickinson Falcon) and used at peak density culture. For sonication, trophozoites were resuspended in 1 ml of PBS and sonicated three times on ice with three bursts of 30 s each (550 sonic dismembrator; Fisher Scientific).
Salmonella Typhimurium.Salmonella Typhimurium ATCC 14028 was a gift from Kenneth Sanderson, University of Calgary. A nonagitated microaerophilic culture of log-phase S. Typhimurium was generated by inoculation of 10 μl of an overnight stationary-phase culture into 10 ml of Luria broth and incubation at 37°C until log phase was attained. The number of CFU per ml was determined by measuring the optical density at 600 nm (OD600). Cultures were subsequently centrifuged at 1,000 × g and resuspended in a volume of Caco-2 growth media without antibiotics. As previously described (58), Caco-2 cells were infected at a multiplicity of infection (MOI) of 100:1 for 5 or 7 h to induce CXCL8 secretion. Therefore, 100 CFU of S. Typhimurium were added per Caco-2 cell. To determine the number of S. Typhimurium cells associated with Caco-2 monolayers following 5 h of incubation, Caco-2 supernatants were decanted and monolayers were lysed in sterile radioimmunoprecipitation assay (RIPA) buffer, serial diluted, and spot plated onto LB agar plates.
Giardia duodenalis trophozoite infection.Confluent tubes of G. duodenalis trophozoites were harvested by cold shock on ice for 30 min and subsequently pooled into 50-ml polypropylene tubes (Falcon) and centrifuged at 500 × g for 10 min. Resulting pellets were collectively resuspended in 10 ml of ice-cold PBS (Sigma-Aldrich) and centrifuged at 500 × g for 10 min. The pellet was resuspended in 3 ml of fresh PBS, and trophozoites were enumerated with a hemocytometer and adjusted to the appropriate concentration. For ex vivo human biopsy experiments, G. duodenalis trophozoites were adjusted to a concentration of 5.0 × 106 trophozoites/well, while trophozoites were coincubated with in vitro Caco-2 monolayers at an MOI of 1:1, 10:1, or 50:1. For all experiments involving coincubation of ex vivo human small intestinal biopsy tissues or Caco-2 monolayers in vitro with G. duodenalis trophozoites, cells were maintained at 37°C, 5% CO2, and 96% humidity for the experimental duration.
Giardia modulation of intestinal epithelial CXCL8 secretion.In accordance with a previously established protocol (52), ex vivo human small intestinal mucosal biopsy tissues were coincubated with G. duodenalis NF trophozoites in Opti-MEM (Life Technologies) for 2 h and subsequently administered 1.0 ng/ml of proinflammatory IL-1β or a vehicle (0.05% bovine serum albumin in PBS) control for 4 h. In separate experiments, ex vivo inflamed human small intestinal biopsy tissues collected from the terminal ileum of patients with active Crohn's disease were coincubated with G. duodenalis NF trophozoites in Opti-MEM for 6 h. Similarly, G. duodenalis trophozoites (NF, WB, or GS/M) were coincubated in the presence or absence of Caco-2 monolayers in ATCC Caco-2 complete growth media for 2 h and subsequently administered IL-1β (1.0 ng/ml) or CXCL8 (1.0 ng/ml) for 4 h or S. Typhimurium (MOI, 100:1) for 5 or 7 h. Positive control groups were administered IL-1β (1.0 ng/ml), CXCL8 (1.0 ng/ml), or S. Typhimurium (MOI, 100:1) for appropriate incubation times. In all experiments, supernatants were collected and centrifuged at 500 × g for 10 min at 4°C. Resulting supernatants were decanted and stored at −70°C. Biopsy tissues were homogenized in 1 ml of RIPA buffer (1% IGEPAL, 0.1% SDS, and 0.5% sodium deoxycholate in PBS) containing a protease inhibitor tablet (Roche; 04693159001), centrifuged at 10,000 × g, aliquoted, and stored at −70°C. CXCL8 and IL-1β levels in supernatants and/or biopsy tissues were determined using commercially available human CXCL8 and IL-1β ELISAs (R&D Systems) with samples assayed in triplicate.
CXCL8 mRNA expression.Caco-2 monolayers were coincubated with G. duodenalis NF or GS/M trophozoites at an MOI of 10:1 for 2 h in Caco-2 growth media and subsequently administered 1.0 ng/ml of IL-1β. At 1 and 4 h after IL-1β administration, cell monolayers were washed with PBS, lysed in 1 ml of QIAzol, and stored at −70°C in RNase-free tubes. mRNA was isolated using a modified RNeasy protocol from Qiagen. Briefly, 0.2 ml of chloroform was added to tubes and samples were shaken for 15 s. After 3 min of incubation at room temperature, samples were centrifuged at 12,000 × g for 15 min at 4°C. The top mRNA-containing layer was mixed with an equal volume of 70% ethanol and applied to RNeasy spin columns; at this point, the Qiagen RNeasy mRNA isolation protocol was followed. Samples were assessed with a NanoDrop to determine mRNA concentrations. Only mRNA samples with 260 nm/280 nm ratios greater than 1.8 were transcribed to cDNA. Following this, 1 μg of mRNA from each sample was reverse transcribed into cDNA via a QuantiTect reverse transcription kit and a PCR thermal cycler (Bio-Rad). Quantitative PCR (qPCR) was run on cDNA samples using a QuantiFast SYBR green PCR kit and a RotorGeneQ qPCR machine (Qiagen). A positive-control sample was used to generate a relative standard curve that consisted of six 1-in-10 dilutions. All samples were diluted 1 in 10 to fit within the standard curve. Levels of CXCL8 were normalized against loading control β2M. All primers were predesigned and purchased from Qiagen.
Cathepsin cysteine protease inhibition.Previous reports have suggested that 1 μM E-64d is the maximal inhibitory dose that does not affect G. duodenalis trophozoite viability (59). The MIC of Ca-074Me was determined by performing experiments with increasing concentrations of Ca-074Me. Therefore, Caco-2 supernatants in ATCC growth medium were pretreated with 1 μM E-64d, increasing concentrations of Ca-074Me (1, 10, or 50 μM), or a vehicle control (dimethyl sulfoxide [DMSO]) for ∼10 to 15 min. After this incubation period, G. duodenalis trophozoites were added to supernatants and incubated for 2 h, and IL-1β was subsequently added for 4 h. Cell supernatants were collected and assayed for cathepsin cysteine protease activity (see below). In separate experiments, confluent tubes of G. duodenalis trophozoites were pretreated with 1 μM E-64d, increasing concentrations of Ca-074Me (1, 10, or 50 μM), or a vehicle control (DMSO) for 30 min. Following this, trophozoites were harvested via cold shock on ice for 30 min. G. duodenalis trophozoites were then coincubated with Caco-2 monolayers for 2 h and administered recombinant CXCL8 for 4 h. In addition, G. duodenalis trophozoites were collected and sonicated (see above), and 1-in-10 dilutions of sonicates were assayed for cathepsin cysteine protease activity (see below).
G. duodenalis viability assays.In our laboratory, cold shock is not effective at detaching G. duodenalis trophozoites from the surface of Caco-2 monolayers after a 6-h coincubation. Therefore, motility of supernatant G. duodenalis trophozoites was used as an indicator of viability (60). After a 6-h coincubation with Caco-2 monolayers, supernatants were collected and vortexed, and 10 μl was plated onto a hemocytometer. The ratio of the number of motile trophozoites to total trophozoites was determined and expressed as a percentage of the control.
Detection of cathepsin cysteine protease activity.Cathepsin cysteine protease activity was determined by measuring the liberation of 7-aminomethylcoumarin (AMC) from fluorogenic substrates. Cathepsin protease activity corresponds to a change in reflective light units (RFUs) over time (48, 49). In vitro cell supernatants and trophozoite sonicates (diluted 1:10) were plated on a 96-well microplate along with cathepsin assay buffer (100 mM sodium acetate, 10 mM dithiothreitol [DTT], 0.1% Triton X-100, 1 mM EDTA, 0.5% DMSO, and 200 μM ZFR-AMC or ZRR-AMC) at a 1:2 ratio of supernatant to assay buffer. Assay buffer was adjusted to pH 7.2 to mimic the luminal pH of the upper small intestine. Microplates were incubated at 37°C for 5 min and subsequently measured kinetically using a microplate reader (SpectraMax M2e; Molecular Devices, Sunnyvale, CA) at 37°C with excitation and emission wavelengths of 354 nm and 445 nm, respectively. Measurements were recorded every 30 s for 5 min.
Visualization of Giardia duodenalis cathepsin proteases.A previously described protocol (61) was adapted to visualize G. duodenalis intratrophozoite cathepsin cysteine protease activity. Isolated G. duodenalis NF or GS/M trophozoites were adjusted to a concentration of 1.0 × 108 trophozoites/ml in cathepsin assay buffer (100 mM sodium acetate, 10 mM DTT, 0.1% Triton X-100, 1 mM EDTA, 0.5% DMSO [pH 7.2]) and sonicated (see above). Samples were subsequently centrifuged at 10,000 × g for 10 min at 4°C. Following centrifugation, resulting supernatants were decanted and incubated with an equal volume of nondenaturing electrophoresis buffer for 15 min at room temperature. At 4°C and protected from light, 50 μl of this solution was then added to 7% SDS-PAGE gels containing 200 μM Z-Phe-Arg-AMC and run at 100 V. To remove SDS, gels were protected from light and washed 4 times for 15 min in a cold 2.5% Triton X-100 distilled water solution, followed by the same number of washes in cold distilled water. Gels were subsequently incubated in cathepsin assay buffer overnight at 37°C. Bands were visualized using a ChemiDoc XRS System (Bio-Rad).
Giardia modulation of CXCL8-induced neutrophil chemotaxis. (i) Chemotactic supernatant generation.Caco-2 monolayers grown to confluence on poly-l-ornithine (Sigma-Aldrich)-pretreated small petri dishes were coincubated with G. duodenalis NF, WB, or GS/M trophozoites in Opti-MEM for 2 h and subsequently administered CXCL8 (100 ng/ml) for 4 h. Experiments were performed in serum-reduced Opti-MEM to prevent potential serum confounding factors when analyzing PMN chemotaxis. Cell supernatants were collected and centrifuged at 500 × g for 10 min, and the resulting supernatant was decanted and stored at −70°C until further use. Positive-control supernatants were administered only CXCL8 (100 ng/ml).
(ii) PMN isolation and chemotaxis assay.Human blood was obtained from healthy volunteers according to standard techniques in accordance with protocols approved by the University of Calgary. Human neutrophils were isolated by dextran sedimentation and Percoll density gradient separation, as described previously (62, 63). In short, blood was collected in acid-citrate-dextrose Vacutainers (Becton, Dickinson), pooled, and centrifuged at 350 × g for 20 min. The resulting platelet-rich plasma was discarded, and leukocytes were separated from erythrocytes via 0.6% dextran (Sigma-Aldrich) sedimentation. The leukocyte-rich upper layer was fractionated using isotonic Percoll (Sigma-Aldrich). PMNs were collected from the 70% to 81% interface and subsequently washed twice in Hanks balanced salt solution (Sigma-Aldrich) without Ca2+ or Mg2+. PMNs were enumerated using a hemocytometer, and cell viability was assessed via trypan blue exclusion. Chemotactic supernatants, collected from the experiments described in the previous section, were thawed, and 600 μl was added to the lower chambers of 8.0-μm Transwells (Costar), while 5.0 × 105 PMNs in Opti-MEM were added in 150 μl of Opti-MEM to the upper chambers. Plates were allowed to incubate for 1 h at 37°C, 5% CO2, and 96% humidity before upper and lower supernatants were collected and assessed via myeloperoxidase (MPO) assay to quantify PMN chemotaxis.
Myeloperoxidase assay.PMN chemotaxis was assessed via MPO activity assays using the O-dianisidine method (64). Briefly, PMNs in the upper and lower chambers of 8.0-μm Transwells were lysed in 150 μl of a 1:1 ratio of 1 M sodium citrate and 10% Triton X-100 (Sigma-Aldrich); upper chambers were adjusted to a final volume of 600 μl prior to lysis. Following addition of lysis buffer, samples were incubated for 15 min at 4°C with shaking. After 15 min of incubation, 100 μl of supernatant from PMN chemotaxis assays was added to a 96-well microplate along with 150 μl of the reaction mixture (comprised of 0.005 g of O-dianisidine [Sigma-Aldrich], 30 ml of distilled H2O, 3.33 ml of potassium phosphate buffer, and 17 μl of 1% H2O2). Using a microplate scanner (SpectraMax M2e; Molecular Devices, Sunnyvale, CA), three absorbance readings at 450 nm were recorded every 30 s. PMN chemotaxis was determined by taking the bottom-to-top ratio of myeloperoxidase activity. PMNs applied to supernatants administered only CXCL8 (100 ng/ml) were used as positive controls.
Statistics.Values are represented as means ± standard errors of the means (SEM). Statistical analysis was performed using GraphPad Prism 6 software; the normality of the data was assessed prior to analysis. All comparisons were made using one-way analysis of variance (ANOVA) with Tukey's post hoc analysis. Statistical significance was established at a P value of <0.05.
RESULTS
Giardia duodenalis trophozoites attenuate CXCL8 secretion from human intestinal tissues and intestinal epithelial cells.Initial experiments were performed to determine whether G. duodenalis NF trophozoites were capable of attenuating IL-1β-induced CXCL8 secretion from human small intestinal mucosal biopsy tissues. This isolate has previously been used in our laboratory to characterize the pathophysiological effects of G. duodenalis infections (11, 65). Adapting from previous studies (52), we developed a novel method of modeling G. duodenalis interactions with human small intestinal tissues, whereby human small intestinal mucosal biopsy tissues collected from the terminal ileum of patients with CD in remission were coincubated with live G. duodenalis trophozoites. Ileal biopsy tissues from CD patients exhibit cytokine profiles that differ from those from individuals without CD (66, 67) and therefore may have abnormal responses to proinflammatory cytokine stimulation. As a result, initial experiments sought to determine whether administration of IL-1β to these tissues resulted in enhanced secretion of CXCL8 ex vivo. As determined via ELISA, IL-1β administration to uninflamed CD biopsy tissues significantly increased the levels of CXCL8 detected within supernatants (Fig. 1A). These results indicate that mucosal biopsy tissue collected from patients with CD in remission were capable of increasing secretion of CXCL8 in response to proinflammatory IL-1β ex vivo and therefore represent an experimental model to determine if G. duodenalis trophozoites attenuate CXCL8 secretion. CXCL8 supernatant levels in uninflamed CD biopsy tissues coincubated with G. duodenalis NF trophozoites and subsequently administered IL-1β were not significantly different from uninflamed CD biopsy tissues not administered IL-1β (Fig. 1A). These results also demonstrated that G. duodenalis trophozoites fail to induce the release and/or accumulation of supernatant CXCL8 from ex vivo CD small intestinal mucosal biopsy tissues, as levels were not significantly different from those in groups not administered IL-1β (Fig. 1A). Similar results were observed in the biopsy tissue homogenates: CXCL8 concentrations in groups containing G. duodenalis NF trophozoites in the presence or absence of IL-1β were significantly lower than in CD biopsy tissues administered only IL-1β (Fig. 1B). These results demonstrate that G. duodenalis NF trophozoites attenuate CXCL8 levels in ex vivo CD small intestinal mucosal biopsy tissues administered IL-1β. To determine if the baseline inflammatory state of collected biopsy tissues affected our observations, similar experiments were performed with biopsy tissues collected from areas of active inflammation in the terminal ileum of patients with CD. Supernatant CXCL8 levels were significantly reduced in inflamed biopsy tissues coincubated with G. duodenalis NF trophozoites for 6 h compared to inflamed biopsy tissues not incubated with G. duodenalis NF trophozoites (Fig. 1C). These results demonstrate that G. duodenalis trophozoites are capable of attenuating levels of CXCL8 in both inflamed and uninflamed small intestinal mucosal biopsy tissues.
Giardia duodenalis NF trophozoites attenuate IL-1β induced CXCL8 from small intestinal mucosal biopsy tissues. Human small intestinal biopsy tissues obtained from the terminal ileum of patients with Crohn's disease in remission were incubated with 5.0 × 106 G. duodenalis NF trophozoites for 2 h and subsequently administered 1.0 ng/ml of recombinant IL-1β for 4 h. Biopsy tissues administered IL-1β in the absence of G. duodenalis trophozoites was used as a positive control for supernatant CXCL8 levels. CXCL8 levels in supernatants (A) and biopsy tissues (B) were determined via ELISA. (C) Human small intestinal biopsy tissues obtained from areas of active inflammation in the terminal ileum of patients with active Crohn's disease were incubated with 5.0 × 106 G. duodenalis NF trophozoites for 6 h. Supernatant CXCL8 levels were determined by ELISA. All data are representative of at least three independent experiments (n = 3 to 9/group) and represented as means ± SEM. n.s., not significant. *, P < 0.05.
The intestinal epithelium represents the primary point of contact between G. duodenalis trophozoites and the host. Previous reports from our laboratory and others have demonstrated that G. duodenalis trophozoites induce pathophysiological events within intestinal epithelial cells (11, 53, 65, 68, 69). As these epithelial cells also participate in the induction of acute intestinal inflammatory responses by secreting chemokines such as CXCL8 (32–34), we investigated whether the attenuation of IL-1β-induced CXCL8 secretion from CD small intestinal mucosal biopsy tissues implicated a G. duodenalis-mediated modulation of human intestinal epithelial CXCL8. Experiments similar to those described above were performed in vitro using the human Caco-2 intestinal epithelial cell line; this cell line has previously been used to delineate pathophysiological events induced by G. duodenalis and other enteropathogens in human enterocytes (65, 70, 71) and has well-defined CXCL8 signaling pathways in response to host- and pathogen-induced proinflammatory stimuli such as IL-1β and S. Typhimurium (34, 72). Our results demonstrate that CXCL8 supernatant levels from Caco-2 monolayers coincubated with several MOIs (1:1, 10:1, and 50:1) of G. duodenalis NF trophozoites administered IL-1β were significantly reduced compared to those from control monolayers administered only IL-1β (Fig. 2A). Levels of CXCL8 detected in IL-1β-stimulated Caco-2 supernatants coincubated with G. duodenalis NF trophozoites at an MOI of 50:1 were also significantly reduced from those of IL-1β-stimulated Caco-2 monolayers incubated with trophozoites at MOIs of 1:1 and 10:1 (Fig. 2A). Therefore, the attenuation of supernatant CXCL8 in Caco-2 monolayers administered IL-1β occurs in a G. duodenalis trophozoite dose-dependent manner. Our results demonstrated attenuation of CXCL8 in ex vivo CD small intestinal mucosal biopsy tissues and in vitro Caco-2 monolayers coincubated with G. duodenalis trophozoites. Therefore, G. duodenalis trophozoites are capable of reducing supernatant CXCL8 levels in several experimental models. This attenuation of supernatant IL-1β-induced CXCL8 is, at least partially, mediated via attenuation of IEC CXCL8.
Giardia duodenalis trophozoites attenuate CXCL8 secretion from in vitro intestinal epithelial Caco-2 monolayers. (A) Caco-2 monolayers were coincubated with G. duodenalis NF trophozoites at an MOI of 1:1, 10:1, or 50:1 for 2 h and incubated with IL-1β for 4 h. (B) Caco-2 monolayers were coincubated with G. duodenalis NF, WB, or GS/M trophozoites (MOI, 10:1) for 2 h and incubated with IL-1β for 4 h. (C) Caco-2 monolayers were coincubated with G. duodenalis NF, WB, or GS/M trophozoites (MOI, 10:1) for 24 h and incubated with IL-1β for another 4 h. (D) Caco-2 monolayers were coincubated with G. duodenalis NF trophozoites (MOI, 10:1) in direct contact with monolayers (Giardia NF) or separated by 0.4-μm Transwells [Giardia NF (T)] for 24 h and incubated with IL-1β for another 4 h. (E) Caco-2 monolayers were coincubated with G. duodenalis NF trophozoites (MOI, 10:1) for 2 h and subsequently administered S. Typhimurium (MOI, 100:1) for 5 h. (F) Caco-2 monolayers were lysed in sterile RIPA buffer and spot plated onto LB agar to determine the number of S. Typhimurium-associated CFU. Caco-2 monolayers administered IL-1β or S. Typhimurium, where appropriate, in the absence of G. duodenalis trophozoites were used as positive controls for supernatant CXCL8 protein levels. CXCL8 levels were determined via ELISA. All data are representative of at least two independent experiments (n = 2 or 3/group) and represented as means ± SEM. n.s., not significant. *, P < 0.05.
Experiments were performed to determine whether different G. duodenalis trophozoite isolates were capable of attenuating supernatant CXCL8 levels, induced via administration of IL-1β, from in vitro Caco-2 monolayers. Caco-2 monolayers were coincubated with G. duodenalis GS/M, NF, or WB trophozoites at an MOI of 10:1 for 2 h and then incubated with IL-1β for an additional 4 h. After the 6-h incubation period, CXCL8 supernatant levels collected from Caco-2 monolayers coincubated with G. duodenalis NF or WB trophozoites and administered IL-1β were significantly reduced compared to those from monolayers administered IL-1β alone (Fig. 2B). Interestingly, CXCL8 supernatant levels collected from Caco-2 monolayers coincubated with G. duodenalis GS/M trophozoites and administered IL-1β were not significantly different from those from groups given IL-1β alone (Fig. 2B). Information available through the Giardia genome indicates that coincubation of G. duodenalis trophozoites with Caco-2 monolayers in vitro results in changes in cathepsin cysteine protease parasite mRNA levels over time (42). Moreover, isolate-dependent differences in the proteolytic activity of G. duodenalis trophozoites have been reported (73). Therefore, our IL-1β experiments were repeated after a longer initial incubation period. Caco-2 monolayers were coincubated with G. duodenalis GS/M, NF, or WB trophozoites for 24 h and then incubated with IL-1β for an additional 4 h. CXCL8 ELISA analysis demonstrated that supernatant levels were significantly reduced in supernatants collected from Caco-2 monolayers coincubated with G. duodenalis GS/M, NF, and WB trophozoites compared to those from monolayers administered only IL-1β and that these levels were not significantly different from each other (Fig. 2C). Therefore, multiple G. duodenalis isolates are capable of attenuating supernatant levels of CXCL8 released by Caco-2 monolayers following administration of proinflammatory IL-1β. Follow-up experiments sought to determine whether direct contact between G. duodenalis trophozoites and in vitro Caco-2 monolayers was required for attenuation of supernatant CXCL8 to occur. As a result, Caco-2 monolayers were coincubated with G. duodenalis NF trophozoites in direct contact with monolayers or separated via 0.4-μm Transwells for 24 h and exposed to IL-1β for 4 h. ELISA analysis demonstrated that CXCL8 supernatant levels were significantly reduced compared to those from control monolayers administered only IL-1β when G. duodenalis NF trophozoites were directly contacting Caco-2 monolayers or when separated by 0.4-μm Transwells and subsequently administered IL-1β, and these groups were not significantly different from each other (Fig. 2D). Therefore, G. duodenalis-mediated attenuation of IL-1β-induced supernatant CXCL8 levels does not require contact between Caco-2 monolayers G. duodenalis trophozoites.
CXCL8 secretion also occurs in response to a variety of pathogens and pathogen-derived proinflammatory stimuli. Previous reports have demonstrated that administration of S. Typhimurium ATCC 14028 to intestinal epithelial monolayers results in CXCL8 secretion (35). As our results demonstrated that all G. duodenalis isolates used in this study were capable of attenuating IL-1β-induced CXLC8 secretion, follow-up experiments investigated whether an initial 2-h coincubation between G. duodenalis NF trophozoites and in vitro Caco-2 monolayers resulted in attenuated CXCL8 supernatants after 5 h of exposure to the enteropathogen S. Typhimurium. As determined via ELISA, G. duodenalis NF trophozoites significantly inhibited the production of CXCL8 induced by S. Typhimurium (Fig. 2E). The number of S. Typhimurium CFU associated with Caco-2 monolayers was not significantly different whether experiments were performed in the presence or absence of G. duodenalis NF trophozoites, excluding the possibility that the observations resulted from a reduced microbial load (Fig. 2F). Therefore, G. duodenalis NF trophozoites directly attenuate CXCL8 secretion from IECs in response to the enteropathogen S. Typhimurium, and this does not occur via G. duodenalis antimicrobial effects.
Giardia duodenalis-mediated attenuation of CXCL8 secretion occurs via a parasite-mediated degradation of CXCL8.Additional experiments sought to elucidate whether G. duodenalis trophozoites prevent or reduce transcription of CXCL8 mRNA. Caco-2 monolayers were coincubated with G. duodenalis NF or GS/M trophozoites at an MOI of 10:1 for 2 h and then administered 1.0 ng/ml of IL-1β. CXCL8 mRNA levels relative to loading control β-2 microglobulin (β2M) were determined via real-time reverse transcriptase PCR 1 and 4 h after IL-1β administration to ascertain whether G. duodenalis NF or GS/M trophozoites modulated CXCL8 mRNA transcription and/or stability. After 1 h of incubation with IL-1β, CXCL8 mRNA levels from Caco-2 monolayers given IL-1β alone were not significantly different from levels found in monolayers that had been initially coincubated with G. duodenalis NF or GS/M trophozoites (Fig. 3A). After 4 h of incubation with IL-1β, CXCL8 mRNA levels in monolayers coincubated with G. duodenalis NF trophozoites were significantly greater than in control monolayers administered IL-1β and monolayers initially coincubated with G. duodenalis GS/M trophozoites prior to IL-1β administration; moreover, CXCL8 mRNA levels in Caco-2 monolayers coincubated with G. duodenalis GS/M trophozoites were not significantly different from those in control monolayers administered IL-1β (Fig. 3B). CXCL8 mRNA levels in Caco-2 monolayers coincubated with G. duodenalis NF and GS/M trophozoites in the absence of IL-1β were not significantly different from those of controls, indicating that the parasite does not induce transcription of CXCL8 mRNA at the time points assessed (Fig. 3). Based on these results, we concluded that isolate-dependent attenuation of IL-1β-induced CXCL8 secretion by G. duodenalis trophozoites occurred via a posttranscriptional mechanism.
Giardia duodenalis trophozoites attenuate IL-1β-induced CXCL8 secretion via a posttranscriptional mechanism. Caco-2 monolayers were coincubated with G. duodenalis NF or GS/M trophozoites (MOI, 10:1) for 2 h and IL-1β was subsequently administered to supernatants and incubated for 1 h (A) or 4 h (B) to induce CXCL8 secretion. Caco-2 monolayers administered IL-1β in the absence of G. duodenalis trophozoites were used as positive controls for CXCL8 mRNA. Levels of CXCL8 mRNA relative to loading control β-2 microglobulin (β2M) were determined. All data are representative of at least two independent experiments (n = 2 or 3/group) and represented as means ± SEM. n.s., not significant. *, P < 0.05.
Giardia duodenalis trophozoites secrete factors that degrade interleukin-8.Additional experiments were performed to establish whether G. duodenalis trophozoites directly degraded CXCL8 protein. G. duodenalis NF, WB, or GS/M trophozoites were coincubated in the presence of Caco-2 monolayers at an MOI of 10:1 for 2 h, and then the coculture was given 1.0 ng/ml of CXCL8 and incubated for an additional 4 h. ELISA analysis determined that the remaining levels of CXCL8 detected in supernatants of G. duodenalis NF and WB trophozoites coincubated with Caco-2 monolayers were significantly reduced compared to those of control CXCL8 groups (Fig. 4A). Moreover, supernatant levels of CXCL8 from the coincubation of G. duodenalis GS/M trophozoites with Caco-2 monolayers were not significantly different from those of groups spiked with CXCL8 (Fig. 4A). These results corroborate previous data from our study demonstrating that G. duodenalis NF or WB isolates were capable of attenuating IL-1β-induced CXCL8 secretion from Caco-2 monolayers at the time point assessed (Fig. 2B). To determine whether these effects required the presence of host cells or whether G. duodenalis trophozoites were directly responsible for degrading CXCL8, G. duodenalis NF or WB trophozoites at an MOI of 10:1 were incubated in Caco-2 growth media in the absence of Caco-2 monolayers for 2 h and subsequently given CXCL8 for 4 h. As determined by ELISA, G. duodenalis NF and WB trophozoites significantly decreased CXCL8 levels in the absence of Caco-2 monolayers (Fig. 4B). Based on these results, experiments were performed to uncover whether G. duodenalis trophozoites promoted degradation of other proinflammatory mediators, such as IL-1β. Therefore, the above-described experiments were repeated, but CXCL8 was replaced with IL-1β. As determined via ELISA, Giardia NF or GS/M trophozoites did not alter the levels of IL-1β (Fig. 4C). Therefore, G. duodenalis trophozoites are capable of releasing factors into cell supernatants that degrade CXCL8.
Giardia duodenalis trophozoites degrade CXCL8. (A) G. duodenalis NF, WB, or GS/M trophozoites (MOI, 10:1) were incubated in the presence of Caco-2 monolayers for 2 h. CXCL8 (1.0 ng/ml) was administered to supernatants and incubated for 4 h. Supernatant levels of CXCL8 were determined via ELISA. (B) G. duodenalis NF or WB trophozoites (MOI, 10:1) were incubated in ATCC complete Caco-2 growth media without Caco-2 monolayers for 2 h, and CXCL8 (1.0 ng/ml) was administered to supernatants and incubated for 4 h. CXCL8 administered to supernatants lacking G. duodenalis trophozoites were used as a positive control. Supernatant levels of CXCL8 were determined via ELISA. (C) G. duodenalis NF or GS/M trophozoites (MOI, 10:1) were incubated in the presence of Caco-2 monolayers for 2 h. IL-1β (1.0 ng/ml) was administered to supernatants and incubated for 4 h. Administration of IL-1β to supernatants in the absence of G. duodenalis trophozoites was used as a positive control. Supernatant levels of IL-1β were determined via ELISA. All data are representative of at least two independent experiments (n = 2 or 3/group) and represented as means ± SEM. n.s., not significant. *, P < 0.05.
Giardia duodenalis trophozoite cathepsin-like cysteine proteases degrade interleukin-8.To determine whether G. duodenalis cathepsin cysteine proteases were involved in the degradation of CXCL8, initial experiments sought to elucidate whether the observed increase in supernatant proteolytic activity was sensitive to inhibition with the broad-spectrum cysteine protease inhibitor E-64d (47) or the catB-specific inhibitor Ca-074Me (74). Our experiments confirmed that pretreatment of Caco-2 supernatants with 1 μM E-64d significantly decreased the hydrolysis of ZFR-AMC when Caco-2 monolayers were coincubated with G. duodenalis NF trophozoites at an MOI of 10:1; this finding was corroborated by slopes not significantly different from those of control supernatants (Fig. 5A). Supernatants pretreated with increasing concentrations of Ca-074Me (1, 10, and 50 μM) did not affect G. duodenalis NF trophozoite viability after the 6-h incubation period. Indeed, the number of trophozoites detected within cell supernatants (Fig. 5D) or the ratio of motile to total trophozoites (Fig. 5E) was unchanged compared to that of vehicle control (DMSO)-pretreated supernatants. Slope values calculated from the supernatant hydrolysis of ZFR-AMC (Fig. 5B) and ZRR-AMC (Fig. 5C) were significantly decreased when G. duodenalis NF trophozoites were coincubated with Caco-2 monolayers pretreated with 10 or 50 μM Ca-074Me. As a result, follow-up experiments examining supernatant inhibition of catB were performed with Caco-2 supernatants pretreated with 10 μM Ca-074Me. As isolate-dependent differences in the ability of G. duodenalis trophozoites to attenuate supernatant CXCL8 levels induced by IL-1β were observed in vitro (Fig. 2B and C), experiments were performed to determine if differences in cathepsin cysteine protease activity within G. duodenalis trophozoites could be observed. G. duodenalis trophozoite sonicates were run under nondenaturing conditions through SDS-PAGE gels copolymerized with the catB/L fluorogenic substrate Z-Phe-Arg-AMC. Different banding patterns were observed when G. duodenalis NF or GS/M sonicates were incubated in SDS-PAGE gels copolymerized with ZFR-AMC (Fig. 5F). These results suggest that cathepsin cysteine proteases can differ between G. duodenalis trophozoite isolates. When Caco-2 monolayers were coincubated with G. duodenalis NF trophozoites at an MOI of 10:1 in the presence of 1 μM E-64d and administered IL-1β, CXCL8 levels detected within supernatants were not significantly different from those of respective IL-1β-induced controls (Fig. 6A). Therefore, G. duodenalis-mediated attenuation of IL-1β-induced CXCL8 secretion in Caco-2 monolayers occurs via the secretion of a cysteine protease sensitive to inhibition with E-64d. Similarly, levels of recombinant CXCL8 administered to supernatants were significantly reduced when Caco-2 monolayers were initially coincubated with G. duodenalis NF trophozoites at an MOI of 10:1; these effects were also reversed when experiments were performed in the presence of 1 μM E-64d or 10 μM Ca-074Me (Fig. 6B). Based on these observations, we concluded that G. duodenalis trophozoites secrete catB-like proteases that degrade CXCL8.
Giardia duodenalis trophozoites secrete cysteine proteases that are inhibited by E-64d and Ca-074Me. Caco-2 monolayers were pretreated with 1 μM E-64d (A) or increasing concentrations of Ca-074Me (1, 10, or 50 μM) (B to E) and subsequently coincubated with G. duodenalis NF trophozoites (MOI, 10:1) for 6 h. Supernatants were incubated with the cathepsin B and L fluorogenic substrate ZFR-AMC (A and B) or the catB fluorogenic substrate ZRR-AMC (C) (200 μM, 5 min, 37°C, and pH 7.2). Proteolytic activity was determined by graphing the change in reflective light units over time. Caco-2 monolayers not containing G. duodenalis trophozoites were used as controls. Supernatants were collected, and the log total of supernatant trophozoites (D) and the ratio of motile to nonmotile trophozoites (E) were determined via enumeration on a hemocytometer. (F) Giardia duodenalis NF and GS/M trophozoite sonicates were run through an SDS-PAGE gel copolymerized with Z-Phe-Arg-AMC under nondenaturing conditions and visualized with a ChemiDoc XRS system. All data are representative of at least two independent experiments (n = 2 or 3/group) and represented as means ± SEM. n.s., not significant. *, P < 0.05.
Giardia duodenalis cathepsin B proteases degrade interleukin-8. (A) Caco-2 monolayers were pretreated with 1 μM E-64d and subsequently coincubated with G. duodenalis NF trophozoites (MOI, 10:1) for 2 h. After this, IL-1β (1.0 ng/ml) was administered to supernatants and incubated for 4 h. (B) Caco-2 monolayers were pretreated with 1 μM E-64d or 10 μM Ca-074Me and subsequently coincubated with G. duodenalis NF trophozoites (MOI, 10:1) for 2 h. After this, CXCL8 (1.0 ng/ml) was administered to supernatants and incubated for 4 h. Caco-2 monolayers administered IL-1β or CXCL8, where appropriate, not containing G. duodenalis trophozoites were used as positive controls. CXCL8 supernatant levels were determined via ELISA. All data are representative of at least three independent experiments (n = 3 or 4/group) and represented as means ± SEM. n.s., not significant. *, P < 0.05.
Basolateral intestinal epithelial CXCL8 recruits PMNs to the basolateral membrane of the epithelium so other signals can, if necessary, promote PMN transepithelial migration (35–37). Experiments sought to determine whether apically produced G. duodenalis cathepsin cysteine proteases may translocate to the basolateral side of the epithelium. Results indicate that concurrently with exposure to S. Typhimurium, G. duodenalis proteases may translocate to the basolateral side in a time- and dose-dependent manner (Fig. 7). A significant increase in both apical (Fig. 7A) and basolateral (Fig. 7B) cysteine protease activities was observed in supernatants collected from the coincubation of polarized Caco-2 monolayers and G. duodenalis NF trophozoites at an MOI of 50:1 administered S. Typhimurium for 7 h. In the same groups, we observed attenuation of apical (Fig. 7C) and basolateral (Fig. 7D) supernatant CXCL8. These results suggest that basolateral translocation of G. duodenalis cathepsin cysteine proteases occurs following exposure to S. Typhimurium and results in attenuated basolateral supernatant CXCL8.
Basolateral translocation of cathepsin cysteine proteases and attenuation of basolateral CXCL8 by G. duodenalis are facilitated by Salmonella Typhimurium. Caco-2 monolayers were coincubated with G. duodenalis NF trophozoites at an MOI of 10:1 or 50:1 for 2 h and subsequently administered S. Typhimurium for 7 h. (A and B) Apical (A) and basolateral (B) supernatants were incubated with the cathepsin B and L fluorogenic substrate ZFR-AMC (200 μM, 5 min, 37°C, and pH 7.2), and proteolytic activity was determined by graphing the change in reflective light units over time. Caco-2 monolayers not containing G. duodenalis trophozoites were used as controls. (C and D) Apical (C) and basolateral (D) CXCL8 supernatant levels were determined via ELISA. Caco-2 monolayers administered S. Typhimurium in the absence of G. duodenalis trophozoites were used as positive controls. All data are representative of at least two independent experiments (n = 2 or 3/group) and represented as means ± SEM. n.s., not significant. *, P < 0.05.
Giardia duodenalis trophozoites secrete cathepsin B proteases and attenuate interleukin-8-induced neutrophil chemotaxis.As inhibition of catB-like proteases prevents G. duodenalis-mediated degradation of CXCL8, experiments were performed to determine whether this resulted in the functional impairment of CXCL8-induced PMN chemotaxis. PMN chemotaxis was significantly reduced when supernatants collected from the coincubation of Caco-2 monolayers, G. duodenalis NF trophozoites at an MOI of 50:1, and CXCL8 were used (Fig. 8A). Supernatants collected from the coincubation of Caco-2 monolayers with G. duodenalis NF trophozoites at an MOI of 10:1 and CXCL8 had a slight but nonsignificant decrease in CXCL8-induced PMN chemotaxis compared to that of the control (Fig. 8A). The results indicate that G. duodenalis NF trophozoites attenuate CXCL8-induced PMN chemotaxis in a dose-dependent manner. Having established that supernatants collected from the coincubation of 100 ng/ml of CXCL8 with Caco-2 monolayers and G. duodenalis NF trophozoites at an MOI of 50:1 were less effective at inducing PMN chemotaxis, experiments were repeated using G. duodenalis GS/M and WB trophozoite isolates. CXCL8-induced PMN chemotaxis was again significantly reduced when supernatants were used from the coincubation containing Caco-2 monolayers, G. duodenalis WB trophozoites, and CXCL8 compared to those of positive-control supernatants (Fig. 8B); conversely, PMN chemotaxis was not altered by G. duodenalis GS/M trophozoites (Fig. 8B). These results are consistent with the above-described observations that only G. duodenalis NF and WB isolates were capable of degrading CXCL8 after 6 h of incubation (Fig. 4A and B). Together, the data indicate that G. duodenalis trophozoites are capable of degrading CXCL8 and preventing CXCL8-induced PMN chemotaxis in an isolate-dependent manner.
Giardia duodenalis trophozoite secretions attenuate interleukin-8-induced neutrophil chemotaxis. Caco-2 monolayers were coincubated with G. duodenalis NF trophozoites at an MOI of 10:1 or 50:1 (A) or G. duodenalis WB or GS/M trophozoites at an MOI of 50:1 for 2 h (B). After 2 h of incubation, CXCL8 (100 ng/ml) was administered to supernatants and incubated for an additional 4 h. Supernatants were collected and applied to the bottom chambers of 8-μm Transwells, and human PMNs were applied to the top chambers and incubated for 1 h. Supernatants administered CXCL8 in the absence of G. duodenalis trophozoites were used as positive controls. PMN chemotaxis was quantified by determining the bottom-to-top myeloperoxidase ratio. All data are representative of at least three independent experiments (n = 3 or 4/group) and represented as means ± SEM. n.s., not significant. *, P < 0.05.
In order to implicate G. duodenalis catB-like activity in the attenuation of CXCL8-induced PMN chemotaxis, experiments were modified to specifically inhibit catB proteases in G. duodenalis trophozoites. This was essential, as preliminary experiments indicated that the presence of E-64d or Ca-074Me in PMN chemotactic supernatants negatively affected CXCL8-induced PMN chemotaxis (data not shown). Initial experiments sought to ascertain whether administration of E-64d or Ca-074Me to confluent tubes of G. duodenalis trophozoites would result in inhibition of cysteine protease activity. Cysteine protease activity assays and corresponding slopes demonstrated that hydrolysis of ZFR-AMC was significantly reduced in groups where G. duodenalis trophozoites were pretreated with Ca-074Me at concentrations of 10 and 50 μM and 1 μM E-64d and resulted in decreased slope values (Fig. 9A). Similarly, hydrolysis of ZRR-AMC and corresponding slope values were significantly reduced in groups in which G. duodenalis trophozoites were pretreated with 10 and 50 μM Ca-074Me; interestingly, E-64d pretreatment did not affect hydrolysis of ZRR-AMC (Fig. 9B). Moreover, the number of trophozoites detected within cell supernatants (Fig. 9C) and the ratio of motile to total trophozoites (Fig. 9D) were not significantly different from those of vehicle control (DMSO)-pretreated trophozoites. Based on these results, it was determined that G. duodenalis trophozoites would be pretreated with 10 μM Ca-074Me to study its effects on PMN chemotaxis, as it resulted in inhibition of the hydrolysis of ZFR-AMC and ZRR-AMC and did not affect trophozoite viability.
Pretreatment of Giardia duodenalis NF trophozoites with Ca-074Me inhibits cathepsin activity. Confluent tubes of G. duodenalis NF trophozoites were treated with increasing concentrations of Ca-074Me (1, 10, or 50 μM) or 1 μM E-64d for 30 min. After 30 min, trophozoites were isolated and sonicated. Trophozoite sonciates were incubated with the cathepsin B and L fluorogenic substrate ZFR-AMC (A) or the catB fluorogenic substrate ZRR-AMC (B) (200 μM, 5 min, 37°C, and pH 7.2). Proteolytic activity was determined by graphing the change in reflective light units over time. Supernatants were collected, and the log total of supernatant trophozoites (C) and the ratio of motile to nonmotile trophozoites (D) were determined via enumeration on a hemocytometer. Vehicle (DMSO)-treated G. duodenalis trophozoites were used as positive controls. All data are representative of at least two independent experiments (n = 2 or 3/group) and represented as means ± SEM. n.s., not significant. *, P < 0.05.
As described above, confluent tubes of G. duodenalis trophozoites were pretreated with a vehicle control (DMSO) or Ca-074Me 30 min prior to isolation and subsequently coincubated with Caco-2 monolayers and CXCL8. Compared to supernatants containing vehicle control-treated G. duodenalis NF trophozoites, hydrolysis of ZFR-AMC and corresponding slopes indicated a significant reduction in cathepsin activity within supernatants when trophozoites were pretreated with Ca-074Me (Fig. 10A). These results suggest that Ca-074Me pretreatment of G. duodenalis NF trophozoites significantly reduced their ability to secrete active cathepsin cysteine proteases into supernatants. For PMN chemotaxis assays, supernatants collected from the coincubation of vehicle control-treated (DMSO) G. duodenalis NF trophozoites, Caco-2 monolayers, and CXCL8 displayed significantly less chemotactic potential for PMNs than did control supernatants (Fig. 10B). More importantly, the inhibitory effect of Giardia was abolished by Ca-074Me (Fig. 10B). These results demonstrate that G. duodenalis catB-like cysteine protease activity attenuates CXCL8-induced PMN chemotaxis in an isolate-dependent manner.
Inhibition of cathepsin B activity in Giardia duodenalis NF trophozoites prevents attenuation of interleukin-8-induced neutrophil chemotaxis. Confluent tubes of G. duodenalis NF trophozoites were treated with 10 μM Ca-074Me for 30 min. After 30 min of incubation, trophozoites were coincubated with Caco-2 monolayers at an MOI of 50:1 for 2 h, administered CXCL8 (100 ng/ml), and then incubated for an additional 4 h. (A) Supernatants were collected and incubated with the cathepsin B and L fluorogenic substrate ZFR-AMC (200 μM, 5 min, 37°C, and pH 7.2). Proteolytic activity was determined by graphing the change in reflective light units over time. (B) Supernatants were collected and applied to the bottom chambers of 8-μm Transwells, and human PMNs were applied to the top chambers and incubated for 1 h. Supernatants administered CXCL8 in the absence of G. duodenalis trophozoites were used as positive controls. PMN chemotaxis was quantified by determining the bottom-to-top myeloperoxidase ratio. All data are representative of at least three independent experiments (n = 3 or 4/group) and expressed as means ± SEM. n.s., not significant. *, P < 0.05.
DISCUSSION
Results from the present study reveal a hitherto-unrecognized immunomodulatory effect by a parasite, in which G. duodenalis trophozoites attenuate supernatant levels of CXCL8, a potent PMN chemokine, that have been released from inflamed ex vivo CD small intestinal mucosal biopsy tissues or uninflamed tissues administered proinflammatory IL-1β and in vitro Caco-2 monolayers administered IL-1β or S. Typhimurium. Interestingly, attenuation of supernatant CXCL8 by G. duodenalis GS/M trophozoites occurred with different kinetics than for other isolates. This attenuation of CXCL8 by G. duodenalis trophozoites was the result of secreted G. duodenalis catB proteases that degraded CXCL8. Degradation of CXCL8 by G. duodenalis trophozoites resulted in an end product less effective at promoting PMN chemotaxis. The inhibition of G. duodenalis catB proteases prevented the parasite-mediated degradation of CXCL8 and subsequent attenuation of CXCL8-induced PMN chemotaxis. Therefore, G. duodenalis trophozoites secrete catB proteases capable of degrading CXCL8 and attenuating CXCL8-induced chemotaxis. These results demonstrate that G. duodenalis trophozoites possess immunomodulatory factors that may inhibit proinflammatory intestinal responses in G. duodenalis-infected individuals.
To investigate the immunomodulatory properties of G. duodenalis, we developed a novel model to investigate G. duodenalis-human intestinal mucosal interactions using human small intestinal biopsy tissues ex vivo, given proinflammatory IL-1β to mimic an inflamed intestinal tissue milieu. The findings were supported by studies with epithelial Caco-2 monolayers in vitro, which permitted the identification of the mechanisms involved. Finally, using a well-known PMN chemotaxis assay (75) and freshly isolated human neutrophils, we demonstrated that G. duodenalis catB proteases degraded CXCL8 into a product less effective at promoting CXCL8-induced PMN chemotaxis. Results from our study confirm previous observations that G. duodenalis trophozoites do not induce CXCL8 secretion from IECs (32, 71), and this lack of induction of IEC CXCL8 mRNA or protein supports the notion that human G. duodenalis infections occur in the absence of overt intestinal inflammation (9). The results presented in this report demonstrate for the first time that G. duodenalis trophozoites secrete catB proteases capable of attenuating intestinal epithelial proinflammatory responses that are associated with the recruitment of proinflammatory PMNs. Additional research is needed to determine whether these observations may help explain previous reports suggesting that giardiasis in children living in areas in which a variety of enteric infections are endemic have decreased incidence rates of diarrheal disease (17, 27, 28). G. duodenalis infections have also been reported to modulate the severity of diarrhea during rotavirus coinfection (17, 18). Future research will assess whether G. duodenalis trophozoites possess other immunomodulatory molecules also capable of attenuating or modulating other aspects of their host's inflammatory response. Similarly, research is needed to clarify whether both assemblage A and B trophozoites are capable of attenuating aspects of a host's proinflammatory response via different mechanisms. Ongoing research is now assessing the effects of Giardia on inflammation using experimentally infected animal models in vivo.
Accumulation of neutrophils is a hallmark of acute intestinal inflammatory responses. CXCL8 and the lipid mediator leukotriene B4 (LTB4) are intermediate host-derived neutrophil chemoattractants, whose signals are overridden by host-derived end-target chemoattractants such as anaphylatoxins, including C5a (reviewed in reference 76). CXCL8 secretion by epithelial cells rapidly occurs during acute intestinal inflammatory responses and recruits extravasated PMNs to the basolateral surface of the intestinal epithelium (35, 37). The present study demonstrates that G. duodenalis catB proteases degrade CXCL8 produced by human enterocytes and reduce its chemotactic potential toward neutrophils, thereby suggesting a role for these parasitic proteases in preventing neutrophil recruitment and accumulation (77, 78). Receptors for CXCL8 are also expressed on a variety of proinflammatory cells, including mast cells and eosinophils (77, 78). In addition, the apical administration of CXCL8 has been shown to induce gene transcription within IECs in vitro via CXCR1 (79). Therefore, attenuation of CXCL8 by G. duodenalis trophozoites and its roles in affecting other proinflammatory cells, such as mast cells and eosinophils, and in IEC gene transcription require further study.
The intestinal epithelial barrier functions to restrict access to underlying host tissues, and its impairment can be associated with the migration of various lumenal products to underlying host tissues (reviewed in reference 80). A variety of infectious and inflammatory intestinal disorders disrupt the epithelial barrier (81–83), as illustrated in the present study using coinfection with S. Typhimurium and G. duodenalis. In these instances, Giardia cathepsins are able to reach the basolateral side of the epithelium. Several studies have demonstrated that G. duodenalis infections can occur concurrently with infections with proinflammatory gastrointestinal pathogens (15–18). As previous studies have suggested that neutrophils can induce intestinal epithelial anion hypersecretion and, consequently, water loss and diarrhea via several mechanisms (84, 85), our results may help to explain the lower inflammatory scores and reduced incidence of diarrhea in children infected with G. duodenalis (17, 27, 28). Indeed, diarrhea caused by experimental infection with enterohemorrhagic Escherichia coli was found to be directly induced by neutrophilic infiltration, independently of toxin production by the bacteria (86). The present findings also revealed that G. duodenalis catB failed to have a similar proteolytic effect on IL-1-β, but it remains to be seen whether G. duodenalis is capable of degrading or modulating other PMN chemokines. As a previous report has demonstrated that E. histolytica cysteine proteases degrade the anaphylatoxins C3a and C5a (87), research examining the potential of G. duodenalis catB proteases to degrade these factors seems warranted. Furthermore, microarray analysis of G. duodenalis assemblage E-infected cattle demonstrated a decrease in genes associated with lipid metabolism (26). Therefore, additional research is required to determine whether G. duodenalis trophozoites may also modulate the expression of proinflammatory lipid mediators.
Cathepsin cysteine proteases of single-celled parasites have been well described, but their function remains incompletely understood (reviewed in references 44 and 45). Parasitic cysteine proteases have been shown to prevent activation of proinflammatory transcription factors (88), degrade host effector molecules (89), promote specific host immune responses (90, 91), and cleave host chemokines (46, 87). Prior to this study, reports had described roles for a G. duodenalis catB protease in trophozoite encystation and excystation (41), but the role of these proteases in disease pathogenesis and host immunity has remained largely speculative (43, 92, 93). The results presented herein demonstrated for the first time that catB proteases secreted by G. duodenalis trophozoites might be associated with the attenuation of aspects of their host's proinflammatory immune response. Our study further demonstrates isolate-dependent differences in G. duodenalis catB/L cysteine proteases. As the G. duodenalis genome contains genes for several cathepsin cysteine proteases, including 9 catB proteases (42), it remains to be seen whether G. duodenalis cathepsin cysteine proteases may modulate various aspects of their host's immune response and target a variety of cells within the intestinal mucosa. Similarly, the role of these cathepsin cysteine proteases in pathogenesis warrants further investigation. Analysis of the coding region of assemblage A and B G. duodenalis genomes has indicated an amino acid identity of only 78% (6). As a result, research needs to assess whether cathepsin cysteine proteases in assemblage A and B G. duodenalis isolates play different roles in modulating or attenuating their host's immune response. Moreover, the kinetics of cathepsin cysteine protease secretion by assemblage A and B G. duodenalis trophozoites requires further study. This research may explain results in this study demonstrating that G. duodenalis GS/M trophozoites attenuated IL-1β-induced supernatant CXCL8 only when incubated with Caco-2 monolayers for 24 h prior to addition of IL-1β.
This study demonstrated that treatment with 10 μM Ca-074Me was sufficient to inhibit degradation of CXCL8 by G. duodenalis NF trophozoites while also having a more significant impact on limiting hydrolysis of the catB/L substrate ZFR-AMC than processing of the catB substrate ZRR-AMC by trophozoite sonicates. Therefore, certain G. duodenalis catB proteases appear to possess unique structural features that prevent them from processing ZRR-AMC while still being capable of inhibition by Ca-074Me. Similar studies have shown that Leishmania major catB is unable to hydrolyze ZRR-AMC but is susceptible to inhibition with Ca-074 (94). Furthermore, as the G. duodenalis genome contains genes for numerous cathepsin proteases, it is possible that additional cathepsins or other proteases are capable of cleaving CXCL8. Indeed, a basic local alignment search tool (BLAST) (95) search of the protein responsible for cleaving CXCL8 in E. histolytica (46) against the G. duodenalis genome has indicated structural similarity to several catL proteases, suggesting that multiple G. duodenalis proteases may be capable of degrading or cleaving CXCL8. Additionally, G. duodenalis cathepsin cysteine proteases may require activation via separate proteases. For example, Toxoplasma gondii catB proteases require proteolytic activation by the parasite's catL cysteine proteases (96). Additional research is required to examine the structure and substrate specificities of G. duodenalis catB proteases to determine whether these factors exhibit unique substrate specificities and, furthermore, if other proteases participate in the degradation of CXCL8. Regardless, our study is the first to demonstrate an anti-inflammatory role for G. duodenalis trophozoites while specifically identifying the host cells targeted and the parasite factors involved.
In conclusion, our data reveal a novel role for G. duodenalis catB cysteine proteases. G. duodenalis catB proteases can degrade CXCL8 secreted by human intestinal epithelial cells in response to host- and pathogen-derived proinflammatory stimuli, such as IL-1β and S. Typhimurium. The degradation of CXCL8 by G. duodenalis catB cysteine proteases directly attenuated CXCL8-induced neutrophil chemotaxis. Our findings are the first to demonstrate that G. duodenalis trophozoites are capable of attenuating acute inflammatory responses within the intestinal mucosa that promote the recruitment of proinflammatory PMNs and assign a role for G. duodenalis catB cysteine proteases in this process.
ACKNOWLEDGMENTS
We are grateful to Kenneth Sanderson for providing us with S. Typhimurium ATCC 14028.
A. G. Buret's laboratory is funded in part by a training grant from the Natural Sciences and Engineering Research Council of Canada (NSERC) CREATE program and an NSERC Discovery grant. J. A. Cotton is a recipient of an NSERC graduate student scholarship and a joint IBD studentship from Alberta Innovates Health Solutions (AIHS) and the Crohn's and Colitis Foundation of Canada (CCFC).
FOOTNOTES
- Received 21 March 2014.
- Returned for modification 6 April 2014.
- Accepted 8 April 2014.
- Accepted manuscript posted online 14 April 2014.
- Copyright © 2014, American Society for Microbiology. All Rights Reserved.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
