Introduction

The intestinal protozoan parasite Giardia duodenalis has a wide vertebrate host range with high prevalences found both in production and companion animals. In humans, infections reach up to 280 million each year1. Giardia has a simple and direct life cycle, presenting itself in 2 morphological stages, i.e. the cyst and the trophozoite stage. Cysts are able to withstand the harsh environmental conditions outside the host, whereas trophozoites are adapted to the specific conditions in the intestine of the host. After the ingestion of infective cysts, which are present in the environment, such as in contaminated water or food, excystation occurs inside the host. This leads to the release of two flagellated trophozoites from each cyst. These trophozoites reside in the small intestine, where they attach to the mucosa and the surface of the epithelium without invading the tissue2. An infection with Giardia can pass without any clear symptoms, but can also lead to the clinical condition of giardiasis, which is characterized by gastro-intestinal complaints such as diarrhoea, abdominal pain, nausea and weight loss. While in most hosts these symptoms rapidly disappear, in others a chronic situation develops, which can last for several months3.

In recent years, several studies have highlighted the importance of IL-17A in orchestrating the protective immune response against Giardia. First, Solaymani-Mohamaddi and Singer showed in vitro production of IL-17 by spleen cells after stimulation with Giardia extract4. Furthermore, an in vivo study in calves demonstrated a strong proliferation of IL-17-producing CD4+ αβ T-cells starting from 5 weeks post infection5. Similarly, Saghaug et al. showed that a G. duodenalis infection in humans also evoked an IL-17A-producing memory CD4+ T cell response6. Finally, infections in mice with their natural parasite G. muris 7, 8 or with G. duodenalis 8 provoked a strong upregulation of IL-17A starting from week 1 post infection onwards. Importantly, mice deficient in either IL-17A or IL-17 receptor A (IL-17RA) were unable to clear a Giardia infection in comparison to wild type mice7, 8.

The functional role of IL-17A in the intestine can be very broad and has been described as being dual with both protective and pathological effects. IL-17A can influence many different cell types, such as T cells, B cells, macrophages, neutrophils and epithelial cells to produce various factors such as cytokines, chemokines, antimicrobial peptides, mucins and tight junction proteins9. Pathological, pro-inflammatory roles are mainly seen in autoimmune and inflammatory disorders, yet it is increasingly clear that IL-17A plays an important role in intestinal immune homeostasis and host defense mechanisms. However, insights in the precise protective effector mechanism(s) triggered by IL-17A following a Giardia infection are still scarce. Dann et al.8 previously reported on the outcome of a microarray analysis performed to identify genes differentially expressed between C57BL/6 WT and IL-17RA KO mice 2 weeks after a G. muris infection. This resulted in the identification of 6 genes that were preferentially induced in the C57BL/6 WT mice and not in the IL-17RA KO mice, including resistin-like molecule β (Retnlb), β-defensin 1 (Defb1), serum amyloid A1 and A2 (Saa1 and Saa2), a putative TNF-resistance related protein (U90926) and a phospholipase A2 (Pla2g4c)8. The authors also reported a decreased mucosal IgA secretion in IL-17A−/− mice. In order to further extend our knowledge on the IL-17A-induced anti-Giardia immune response, in the current study RNA deep sequencing technology was applied as an unbiased way to analyse the intestinal response following a G. muris infection in both C57BL/6 WT and C57BL/6 IL-17RA KO mice. The outcome of this analysis indicated that a Giardia infection triggered an IL-17RA-dependent transcriptional upregulation of a wide array of antimicrobial proteins and complement factors, which, in combination with an intestinal IgA response seems to be important to confer protection against this parasite.

Materials and Methods

Infection experiments and tissue sample collection

All animal experiments were conducted in accordance with the E.U. Animal Welfare Directives and VICH Guidelines for Good Clinical Practice. Ethical approval to conduct the studies was obtained from the Ethical Committee of the Faculty of Veterinary Medicine, Ghent University (ethical committee approval numbers EC 2012/027 and EC 2012/176).

In all infection experiments, mice were orally infected with 103  G. muris cysts suspended in 0.2 mL phospahte-buffered saline (PBS) and uninfected control groups were included.

Infection experiment 1 was previously described by Dreesen et al.7. In summary, this study consisted of 2 groups of 10 female C57BL/6 IL-17RA KO and 10 female C57BL/6 WT mice of 6 weeks old at the start of the experiment. Five mice of each group were orally infected with G. muris cysts. The five remaining mice of each group served as uninfected negative controls. Cyst counts were performed every 2 days to monitor the course of infection. All mice were sacrificed 21 days post infection (p.i.) and intestinal tissue samples were collected and processed as previously described7. The samples collected during this experiment were subjected to RNA sequencing.

For additional qRT-PCR analyses, small intestinal samples from another experiment described by Dreesen et al.7 were used. Briefly, 6 week-old female C57BL/6 mice were infected with G. muris cysts, and uninfected control groups of mice were included. Mice were sacrificed at day 7, day 14 and day 21 p.i., and small intestinal tissue was collected. The number of mice was 5 in each group for each timepoint7.

For further analysis of circadian clock-associated genes, a third infection experiment was performed in which 5 female C57BL/6 WT mice were likewise infected with G. muris cysts and 5 female C57BL/6 WT mice served as uninfected controls. The mice were sacrificed 21 days p.i., at 1 pm, and intestinal tissue samples were collected and processed as previously described7.

Infection experiment 4 involved a G. muris infection in B6.129S4-Mbl1tm1Kata Mbl2tm1Kata/J mice on a C57BL/6 background (C57BL/6 Mbl KO) (Jackson Laboratories) and their respective C57BL/6 WT controls. C57BL/6 Mbl KO and C57BL/6 WT mice were housed together in the same cage. Ten female C57BL/6 WT and 10 C57BL/6 Mbl KO mice were infected with 103 G. muris cysts. Cyst counts were monitored daily, starting from day 1 p.i., as previously described7. Five animals of each group were sacrificed at day 7 p.i. and the remaining animals at day 21 p.i. Intestinal tissue was collected and the trophozoites present in the intestine counted. Hereto, the small intestine was removed, starting from 2 cm distal to the gastrointestinal junction. The segment was placed in PBS and incubated on ice for 20 minutes, before counting the trophozoites by using a hemacytometer. The counts are expressed as the absolute number of trophozoites present in the small intestine. In addition, a duodenal tissue sample of 2 cm long, distal to the gastrointestinal junction, was taken from each animal. The samples were snap-frozen in liquid nitrogen and subsequently used for the extraction of total RNA, as previously described7.

RNA sequencing and analysis

Total RNA from experiment 1 (RNA integrity number (RIN) >8.5) was processed using an Illumina TruSeq RNA sample prep kit following the manufacturer’s instructions (Illumina, San Diego, CA, USA). Individually barcoded libraries were then pooled at an equal molar ratio and sequenced at 2 × 50 bp/read using an Illumina HiSeq. 2000 sequencer, as described previously10, 11. Approximately 31 million paired-end sequence reads per sample (mean ± SD = 31.3507 ± 5.3253 million; N = 20) were generated. The quality of the samples was assessed via FastQC software. After inspection 2 pre-processing steps were performed using custom python scripts: (i) end trimming: 3 bp’s from 5′ site and 1 bp from 3′ site (lower quality), (ii) elimination of reads with mean quality score <31 and total number of non-identified bases (N’s) >2. The obtained reads, varying between 50 mio and 80 mio reads per sample, were then mapped to the mouse mm10 genome assembly using Tophat (v2.0.11). Count tables were created with htseq-count (v 0.6.0) software using UCSC mm10 annotation and were used as an input for the analysis using the DESeq. 2 R package. Differential expression analysis was performed, based on the following 4 major comparisons: C57BL/6 WT infected versus C57BL/6 WT uninfected control, C57BL/6 IL-17RA KO infected versus C57BL/6 IL-17RA KO uninfected control, C57BL/6 WT uninfected control versus C57BL/6 IL-17RA KO uninfected control and C57BL/6 WT infected versus C57BL/6 IL-17RA KO infected. The filters that were used to identify differential expression were an adjusted p-value lower than 0.05 and an absolute log2 fold change greater than 1. The four comparisons where then further subdivided into 15 exhaustive categories and a Venn diagram was created (Venny 2.1 software). Ingenuity Pathway Analysis software was used to link genes in the dataset to particular biological functions and pathways. Heatmaps were drawn using the gplots R package and show scaled expression values clustered based on Euclidean distance.

Real-time quantitative PCR

A quantitative Real-Time PCR (qRT-PCR) approach was used to measure the relative mRNA expression levels of several genes in the small intestine. Samples obtained at day 7, day 14 and day 21 p.i. from C57BL/6G. muris infected mice and from C57BL/6 uninfected control mice, as well as at day 21 p.i. from Mbl KO G. muris infected mice and Mbl KO uninfected control mice, were used. The sequences of all the primers that were used can be found in supplementary Table 1. QRT-PCR and normalization of the data, which was based on the housekeeping genes Hprt1 and Tbp, was essentially performed as previously described7. Gene transcription levels were evaluated based on fold between the different groups. Statistical analysis was carried out using GraphPad Prism software. A one-way ANOVA followed by a Dunn’s multiple comparison test was used to determine differences between the different groups. A P-value ≤ 0.05 was considered significant.

Histology

The presence of angiogenin 4 (Ang4) protein was visualised in intestinal tissue. Hereto, small intestinal tissue obtained from C57BL/6 WT G. muris infected mice was stained with a polyclonal sheep anti-mouse-Ang4 antibody, kindly provided by Dr. R. Forman, using a general staining protocol for paraffin sections. Briefly, 4 μm thin sections of formalin-fixed, paraffin-embedded tissues were cut and placed onto 3-aminopropyltriethoxysilane coated glass slides. The sections were deparaffinised and subsequently blocked with a 4% bovine serum albumin solution in PBS, followed by incubation with the primary anti-Ang4 antibody (12 μg/ml). Next, an incubation step with a biotin labeled secondary donkey anti-sheep IgG antibody (SantaCruz) was performed, followed by DAB staining by means of a Vectastain Elite ABC kit (Vector Laboratories). A negative antibody control was included, in which sections were incubated with the secondary donkey anti-sheep IgG antibody after blocking and DAB stained.

In vitro incubation of G. duodenalis trophozoites with angiogenin 4

Plasmid containing the murine Ang4 coding sequence (pET3a-Ang4) (kindly provided by Prof. Simon Carding and Dr. Isabelle Hautefort), was introduced into a pPICZalfaB expression vector, followed by transformation into electrocompetent Top10F-cells. Plasmid DNA was purified from pPICZalfaB Top10F-cells with QIAfilter Plasmid Purification Kit (Qiagen) and subsequently linearised. Linear plasmid DNA was transformed into KM071H-cells, which were grown by fermentation in BMGY- and BMMY-medium. Expressed protein was isolated by collecting, filtrating and concentrating culture supernatant and purified by cation exchange on a Resource S column (Sigma Aldrich). G. duodenalis trophozoites, human assemblage B isolate GS, were kept in axenic culture in modified TYI-S-33 medium12. G. duodenalis trophozoites were incubated with the Ang4 recombinant protein in 96-well plates. Trophozoite numbers ranged from 1 × 104 to 5 × 105 per well. They were incubated with Ang4 protein for 2 h, 8 h, 16 h and 24 h. The concentration of Ang4 ranged from 0,5 to 100 μM in 300 of μL of Giardia TYI-S-33 medium. Fenbendazole 0,5 μM (Panacur, MSD) served as a positive control for the killing of trophozoites. After the respective time points, TYI-S-33 medium was removed and replaced by PBS. Finally, 2 μg of resazurin (Sigma-Aldrich) was added per well, as described by Bénéré et al.13. This blue dye is converted into fluorescent pink by metabolically active trophozoites. After 8 h, fluorescence was measured with a Tecan SpectraFluor fluorescence reader (λex 492 nm–λem 595 nm).

Western blot

The presence of Mbl2 protein in the small intestinal tissue of C57BL/6 WT G. muris infected mice and C57BL/6 WT uninfected control mice, sampled at 21 days p.i., was measured by means of Western blotting. Hereto, the tissue was ground into powder in liquid nitrogen. A water-soluble extract was prepared by dissolving the powder in 150 mM PBS. After centrifugation at 16 000 rpm for 15 minutes, the supernatant was frozen at −80 °C until further use. The pellet was subsequently solubilised in PBS with a 1% protease inhibitor cocktail (Sigma-Aldrich) and with 1% Triton X-100, centrifuged at 16 000 rpm for 1 hour and the supernatant collected and frozen at −80 °C. Protein content of the supernatant was measured with the Pierce BCA protein assay (Thermo Scientific). Two 15% SDS-PAGE gels were run with 30 μg of protein and 1 gel was stained with SimplyBlue SafeStain (Thermo Scientific), to confirm equal loading and concentration. The second gel was blotted onto a nitrocellulose membrane. Subsequently, the blot was washed and incubated with blocking solution (PBS with 0,5% Tween80), followed by an incubation for 24 h at 4 °C with primary rat anti-Mbl2 antibody (Abcam), diluted 1/50 in blocking solution. Following washing, the blot was incubated for 1 h with secondary goat anti-rat-HRP (SantaCruz), diluted 1/1000 in blocking solution. Mbl2 protein was visualised by incubation with DAB dissolved in Tris-buffered saline pH 7.6 (Sigma-Aldrich).

ELISA

Fecal IgA antibody levels were measured by enzyme-linked immunosorbent assays (ELISA). Total IgA levels were measured in fecal samples from WT and Mbl KO uninfected control mice. Giardia-specific IgA levels were measured in fecal samples from WT infected and uninfected control mice. Fecal samples were dissolved in PBS with 0,01% sodium azide and 1% protease inhibitor cocktail (Sigma-Aldrich). Proteins were extracted by vortexing and centrifuging at 16 000 g for 10 min. Supernatants were stored at −80 °C until further use. For measurement of total IgA, Maxisorp 96-well plates (Nunc) were coated with 2 μg/ml sheep anti-mouse IgA (Sigma-Aldrich), for 16 h at 4 °C. For measurement of Giardia-specific IgA levels, plates were coated with G. duodenalis trophozoites. Hereto, trophozoites that were kept in axenic culture in modified TYI-S-33 medium, were centrifuged for 5 min at 800 g, resuspended and washed 3 times in PBS. Next, they were resuspended in 8% formaldehyde solution. 2 mL of formaldehyde solution was added to trophozoites derived from 10 mL of culture. 100 μl of this solution was added to each well of the plates, which were pre-incubated with poly-L-lysine (Sigma-Aldrich). Trophozoites were allowed to attach to the plates for 48 hours and the plates were subsequently dried at 37 °C for 24 hours. After washing with PBS-Tween 20 (0,05%) solution (PBST), all plates were blocked with 2% bovine serum albumin (BSA) in PBST. After blocking, the plates were incubated for 1 hour at room temperature with the fecal extracts in a 1/5 dilution. For measurement of total and Giardia-specific IgA levels, the plates were washed and incubated for 1 hour at room temperature with goat anti-mouse IgA-HRP (Sigma-Aldrich) in a 1/1000 dilution in PBST.

Binding of recombinant Mbl2 to Giardia trophozoites was also measured by ELISA. Hereto, plates were coated with trophozoites as described above. In a first step, recombinant Mbl2 (R&D Systems) was added to the plates in different concentrations, ranging from 0.02–4 μg/ml and incubated for 2 hours at room temperature. Plates were washed 3 times and incubated for 1 hour with rat anti-Mbl2-antibody (2 μg/ml) (Abcam). After washing, plates were finally incubated for 1 hour with goat anti-rat-HRP (SantaCruz Biotechnologies) in a 1/500 dilution in PBST. In a second step, trophozoite-coated plates were incubated for 1 hour at room temperature with fecal samples of WT infected and uninfected control mice, in a 1/5 dilution. After washing, 2 μg/mL of recombinant Mbl2 (R&D systems) was added to each well and the binding of Mbl2 to the trophozoites was subsequently measured as described above.

All reactions were visualised with ABTS in ABTS buffer (Roche). Optical density was measured at 405 nm with a Tecan plate reader, diminished by conjugate control levels measured at 492 nm.

Results and Discussion

Transcriptional changes in C57BL/6 wild type mice following a G. muris infection

Differential expression analysis performed on the sequence reads obtained from C57BL/6 WT infected versus C57BL/6 WT uninfected control mice resulted in the identification of 844 differentially expressed (DE) genes. Of these, 483 genes were upregulated and 361 genes downregulated in the intestinal tissue of infected mice in comparison to uninfected controls (Table S2). In order to validate the RNA-seq data, qRT-PCRs assays were performed for 6 DE genes, including angiogenin 4 (Ang4), dual oxidase 2 (Duox2), mannose binding lectin 2 (Mbl2), nitric oxide synthase 2 (Nos2), 2′-5′-oligoadenylate synthetase 2 (Oas2) and regenerating islet-derived 3 beta (Reg3B). Consistent with the RNA-seq analysis, an upregulation could be observed for Mbl2, Duox2 and Nos2 and a downregulation for Oas2 and Reg3B (Fig. 1). Ingenuity Pathway Analysis (IPA), performed on the 844 DE genes, revealed an increase in the activity of functions mainly related to the development and repair of tissue (Table S3). The genes involved in most of the top 5 impacted and increased functions were IL-17A, Nos2, T-cell acute lymphocytic leukemia 1 (Tal1), the transcription factor GATA binding protein 2 (Gata2) and tumor necrosis factor superfamily member 15 (Tnfsf15).

Figure 1
figure 1

qRT-PCR analysis for genes identified by RNA-seq as differentially expressed between G. muris infected and uninfected C57BL/6 WT mice. Genes that are significantly differentially regulated between the 2 groups, as measured by qRT-PCR, are marked with an asterisk (*p < 0.05).

Full size image

Interestingly, among the 15 highest upregulated genes in C57BL/6 WT infected mice (Table 1), 3 genes were found that are associated with the circadian or day-night rhythm, i.e. Per3, Per2, and Ciart. An additional three genes involved in this pathway were also identified as being upregulated following infection, i.e. Per1, Cry2 and Bhlhe41. All these genes encode for inhibitors of the transcription factors Bmal1 and Clock, which are the central regulators of the circadian clock. CLOCK and BMAL1 will heterodimerize in the cytoplasm to form a complex that, following translocation to the nucleus, initiates transcription of a range of target genes, including their own inhibitors Per1, Per2, Per3, Cry2, Bhlhe41 and Ciart, creating a negative feedback loop14. In accordance with the transcriptional upregulation of these inhibitors following a Giardia infection, the RNAseq data also indicated that the transcription of Bmal1 was downregulated in C57BL/6 WT infected mice compared to C57BL/6 WT uninfected controls (Fig. 2A). Because the transcription levels of genes associated with circadian rhythm vary substantially depending on the time of the day, an additional infection study was performed in which C57BL/6 WT infected mice and C57BL/6 WT uninfected control mice were sacrificed at the exact same time of day, i.e. 1 pm. Consistent with the first study, a significant upregulation was again observed for Per1, Per2 and Per3 at 21 days post infection in infected mice compared to uninfected control animals (Fig. 2B). The functional implications of the transcriptional changes of clock genes following Giardia infection are still unclear. Infection experiments in mice deficient for genes associated with the circadian rhythm, such as Clock, Per2, Bmal1 and Cry, have shown an effect of these genes on the response to bacterial challenge, as reviewed by Tsoumtsa et al.15. The time of day during which organisms are infected with certain microbials also seems to influence the extensiveness of the infection and the associated antimicrobial response, as is the case for a Salmonella Typhimurium infection in mice16. In addition, it has been shown that expression of clock genes in the intestinal tissue controls gastrointestinal functions such as motility, cell proliferation and migration17. It is therefore possible that this response is part of a defence mechanism in an attempt to remove the parasite from the intestinal lumen.

Table 1 Top 15 most upregulated genes as identified by RNA-seq in C57BL/6 WT infected mice versus C57BL/6 WT uninfected control mice.
Full size table
Figure 2
figure 2

G. muris infection in C57BL/6 WT mice has an impact on the circadian clock. (A) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway for the circadian clock (©Kanehisa Laboratories)50. Red and green respectively indicates transcriptional up- or down-regulation in C57BL/6 WT infected mice compared to C57BL/6 WT uninfected control mice, as identified by RNA-seq. (B) Relative mRNA expression levels of the circadian clock associated genes Ciart, Per1, Per2 and Per3 in G. muris infected C57BL/6 WT mice as measured by qRT-PCR at day 21 post infection compared to uninfected control mice (*p < 0.05, **p < 0.01).

Full size image

Another interesting gene amongst the top 15 highest upregulated genes in C57BL/6 WT mice was angiogenin 4 (Ang4). Although Ang4 was originally identified as a factor that induces neovascularization, it is also known as an important antimicrobial protein produced by Paneth cells18. Experimental infections of mice with the gastrointestinal nematode Trichuris muris revealed a positive correlation between Ang4 expression and expulsion of the parasite19, 20, suggesting a potential anti-parasitic function. Interestingly, Ang4 was also identified in the microarray study of Tako et al.21 as being 4-fold upregulated in the intestine of mice 10 days after a G. duodenalis infection. In order to further investigate the role of Ang4 following a G. muris infection, the transcriptional response of this gene was analysed at different time points following infection by qRT-PCR. The outcome of this analysis indicated that a significant transcriptional upregulation was already observed from week 1 post infection onwards (Fig. 3A). Furthermore, histological analysis revealed the presence of Ang4 in the small intestinal paneth cells of G. muris infected mice (Fig. 3B). Finally, in order to assess whether Ang4 had a direct effect on the viability of Giardia, trophozoites of G. duodenalis were incubated in vitro with different concentrations of a recombinantly produced Ang4. As shown in Fig. 3C the presence of Ang4 in G. duodenalis trophozoite cultures, up to a concentration of 100 μM, appeared to have no direct effect on their viability, this in contrast to fenbendazole which was used as a positive control. The reasons for the lack of a direct effect of Ang4 on trophozoites are currently unclear. In bacteria, Ang4 is capable of impairing membrane integrity to exert a direct effect22. However, Hooper et al. have also demonstrated that although Ang4 has microbicidal activity against certain bacteria, other bacteria appeared to be resistant. This shows that species-specific features can influence the susceptibility to Ang418.

Figure 3
figure 3

Ang4 is transcriptionally upregulated in paneth cells in the small intestine of mice following an infection with G. muris but has no direct effect on trophozoite viability. (A) Relative mRNA expression levels of Ang4 in the small intestine of C57BL/6 WT G. muris infected mice versus C57BL/6 WT uninfected controls, as measured by qRT-PCR. Expression levels were measured at day 7, day 14 and day 21 post infection (*p < 0.05). (B) Histological analysis with an anti-Ang4 antibody revealed the presence of Ang4 protein in paneth cells (indicated by black arrow) in the small intestine of C57BL/6 WT infected mice. (C). In vitro incubation of G. muris trophozoites with angiogenin 4 recombinant protein (100 μM) and fenbendazole (0.5 μM) as a positive control. Metabolic activity of trophozoites was measured as fluorescence intensity of the reduced colorimetric agent resazurin (***p < 0.001).

Full size image

Next to Ang4, a large number of additional antimicrobial proteins were also transcriptionally upregulated following infection, including α- and β-defensins, lectins and phospholipases (summarized in Table 2), with β-defensin 1 (Defb1) being the most upregulated. The upregulation of Defb1 and 3 α-defensins (Defa3, Defa5 and Defa22) at day 21 after G. muris infection was confirmed by qPCR analysis. This analysis also revealed a small upregulation at day 7 and day 14 p.i. (Fig. 4). Although Defb1 was also identified as being upregulated in the microarray study of Dann et al.8, its role in the anti-Giardia immune response is still unclear. It is also noteworthy that matrix metalloprotease 7 (Mmp7), which is responsible for the activation of α-defensins through the proteolytic removal of the inhibitory pro-peptide23, was also significantly upregulated in infected mice. Mmp7 was also one of the induced transcripts in the microarray study of Tako et al. 10 days post a G. duodenalis infection in mice. They further showed that mice homozygous for a deletion of the Mmp7 gene were unable to fully eliminate G. duodenalis 21. On the other hand, Eckmann reported that Mmp7-deficient mice were fully able to control and eliminate a G. muris infection24. Further research is therefore necessary to fully understand the role of Mmp7, and by extension the α-defensins, in the protective immune response against Giardia.

Table 2 List of upregulated antimicrobial genes as identified by RNA-seq in C57BL/6 WT infected mice versus C57BL/6 WT uninfected control mice.
Full size table
Figure 4
figure 4

Alpha- and beta-defensins are transcriptionally upregulated in the intestine of C57BL/6 WT mice following G. muris infection. Relative mRNA expression levels of Defb1, Defa3, Defa5 and Defa22 in C57BL/6 WT G. muris infected mice were measured by qRT-PCR at day 7, 14 and 21 post infection, compared to C57BL/6 WT uninfected controls (*p < 0.05).

Full size image

Another group of genes that was identified in the study of Tako et al.21 as being highly upregulated following G. duodenalis infection, were genes linked to mast cell responses. Mast cells have previously been shown to play a role in the immune response against G. duodenalis 25, 26 and G. muris 27 in several studies. In contrast, in our current dataset genes linked to mast cell responses were not impacted. This could be due to the different timepoints of analysis, e.g. 10 days p.i. in the study of Tako et al.21 versus 21 days p.i. in the current study. It is possible that these genes were also upregulated earlier on following G. muris infection, or that this response is more specific for G. duodenalis infection.

Furthermore, some downregulated transcripts that were identified in the study of Tako et al.21 and that were linked to enterocyte function and nutrient absorption, such as Amy2, Akp3, Try4 and Ctrb1, were not found in the current study. This could be due to similar reasons as the ones outlined for the discrepancy in mast cell-linked transcripts.

Identification of the IL-17RA dependent transcriptional changes following a G. muris infection

As previously published, IL-17RA KO G. muris infected mice showed higher cumulative cyst counts than C57BL/6 WT G. muris infected mice. While C57BL/6 WT mice cleared the infection within a timeframe of 3 weeks, cyst excretion was still detectable in IL-17RA KO mice at 21 days p.i.7.

In order to identify the IL-17RA dependent intestinal responses following a G. muris infection, the transcriptome datasets obtained in C57BL/6 WT were subsequently compared to the datasets from C57BL/6 IL-17RA KO mice. In a first step, the datasets of the C57BL/6 IL-17RA KO infected mice and C57BL/6 IL-17RA KO uninfected control mice were analysed and compared to each other. Remarkably, while in C57BL/6 WT infected mice 844 genes were impacted by the infection, in C57BL/6 IL-17RA KO infected mice only 153 genes were differentially expressed compared to the C57BL/6 IL-17RA KO uninfected control mice, with 62 upregulated genes and 91 downregulated genes (Table S4). Pathway analysis on the complete set of DE genes revealed that functions related to apoptosis and damage were predicted to be increased, with important roles for IL-10 and Nos2 (Table S5). This effect is likely due to the prolonged presence of the parasite in the intestine, as it has been shown both in vitro 28, 29, as in human patients with chronic giardiasis30, that certain Giardia strains can induce apoptosis in enterocytes. The increased expression of the immune-regulatory cytokine IL-10 is likely important in the prevention of further tissue damage31. It is also interesting to note that Nos2 was transcriptionally upregulated following infection in the susceptible C57BL/6 IL-17RA KO mice. Although some studies have indicated that Nos2 was important in controlling a Giardia infection32,33,34, the data presented here supports the observations of Maloney et al.35 and Tako et al.21 that Nos2 did not play a crucial role in the protective immune response against Giardia.

By crosschecking the lists of DE genes for the different comparisons, graphically represented in Fig. 5A, 124 genes were identified (highlighted with an * in Fig. 5A and listed in Table S6), that responded differently following a Giardia infection in C57BL/6 WT mice compared to C57BL/6 IL-17RA KO mice, suggesting that their expression is regulated through IL-17RA. A heatmap showing the relative expression patterns of these genes in the different groups is shown in Fig. 5B and the 15 most upregulated genes in C57BL/6 WT mice compared to C57BL/6 IL-17RA KO mice are listed in Table 3.

Figure 5
figure 5

Differential expression analysis on the RNA-seq dataset was performed for 4 different comparisons between C57BL/6 WT and C57BL/6 IL-17RA KO infected and uninfected control mice. Venn diagrams and heatmaps that represent the different comparisons and subcategories were created. (A) Venn diagram representing differentially expressed genes in the RNA-seq dataset of C57BL/6 WT and C57BL/6 IL-17RA KO mice. Four different comparisons between WT and IL-17RA KO mice were further subdivided into 15 exhaustive subcategories. Subcategories marked with an *, contain genes that differentially respond to infection in WT infected mice versus IL-17RA KO infected mice, and in WT infected mice versus WT uninfected controls. The subcategory marked with **, contains genes that were differentially expressed between WT uninfected control versus IL-17RA KO uninfected control mice and between WT infected mice versus IL-17RA KO infected mice. (B). Heatmap of genes that differentially respond to infection in WT infected mice versus IL-17RA KO infected mice, and in WT infected mice versus WT uninfected controls (* in Fig. 5A). The color scale shows the relative expression patterns. (C) Heatmap of genes that were differentially expressed between WT uninfected control versus IL-17RA KO uninfected control mice and between WT infected mice versus IL-17RA KO infected mice (** in Fig. 5A).

Full size image
Table 3 Top 15 most upregulated genes as identified by RNA-seq that differentially respond to infection in C57BL/6 WT infected mice versus C57BL/6 IL-17RA-KO infected mice (A) and also in WT infected mice versus WT uninfected controls (B).
Full size table

Amongst the differentially expressed genes are several genes encoding antimicrobial proteins, such as defensin alpha pseudogene 1 (Defa-ps1), gasdermin C and C2 (Gsdmc/2), regenerating islet-derived family member 4 (Reg4) and resistin-like gamma (Retnlg), supporting previous observations that IL-17A can enhance the production of a range of antimicrobial proteins36. In addition, one of the most differentially expressed genes was mannose-binding lectin 2 (Mbl2) with almost 30-fold higher transcript levels in C57BL/6 WT mice compared to C57BL/6 IL-17RA KO mice. Mbl2 is a circulating C-type lectin that can activate the complement system upon binding carbohydrates with its carbohydrate recognition domains. Mbl2 has been shown to play a critical role in the innate immune defence against different pathogens37, 38. In vitro testing also revealed binding of Mbl2 to the surface of G. duodenalis trophozoites and proved its requirement in the complement-mediated killing of G. duodenalis trophozoites39. Furthermore, in a recent study, Li et al.40 showed that Mbl KO mice exhibited a delayed clearance of G. duodenalis. The data presented in the current study shows for the first time that the expression of Mbl2 is regulated through IL-17A and IL-17RA. Apart from the differential expression between the infected C57BL/6 WT and C57BL/6 IL-17RA KO mice, also the C57BL/6 WT uninfected control mice had higher Mbl2 transcript levels compared to C57BL/6 IL-17RA KO uninfected control mice (Table S7). Furthermore, the transcription profile of Mbl2 following a G. muris infection showed the same kinetics as IL-17A, with a significant upregulation 3 weeks post infection (Fig. 6A). This increased mRNA expression profile also coincided with higher levels of Mbl2 protein in the intestinal tissue of infected mice, as measured by Western blotting (Fig. 6B). A full-length uncropped blot is presented in Supplementary Figure 1. Once the lectin pathway is activated, it proceeds through the action of C2 and C4 to produce activated complement proteins further down the cascade, ultimately leading to the formation of the membrane attack complex and the chemotaxis and activation of inflammatory cells41. Importantly, in addition to Mbl2, several other components of the complement cascade, such as C7, C4bp, C1qtnf3 and C1qtnf9, as well as of mannose-binding lectin-associated serine protease 1 (Masp1), were also transcriptionally upregulated following a G. muris infection in C57BL/6 WT mice (Table S1). Furthermore, additional qRT-PCR analyses showed that also the core complement components C2 and C3 were significantly upregulated at day 21 post infection (Fig. 6C).

Figure 6
figure 6

Mbl2 mRNA levels and protein levels and complement factors are upregulated following G. muris infection in C57BL/6 WT mice. Deficiency of Mbl has an impact on the course of a G. muris infection and on IL-17A mRNA levels. (A) Relative mRNA expression profile of Mbl2 in the small intestine of WT infected mice versus WT uninfected controls, as identified by qRT-PCR. mRNA levels in WT infected mice and WT uninfected control mice were estimated at day 7, day 14 and day 21 post infection (**p < 0.01). (B) Western blot with an anti-Mbl2 antibody on a water-insoluble extract of small intestinal tissue of WT infected (21 days p.i.) and WT uninfected control mice. Numbers 1 to 5 at the left represent 5 replicates of G. muris infected mice and at the right of uninfected control mice. A full-length uncropped blot is presented in Supplementary Figure 1. (C) Relative mRNA expression levels of the complement factors c2 and c3 in C57BL/6 WT infected mice were estimated by qRT-PCR at day 21 post infection, compared to C57BL/6 WT uninfected controls (*p < 0.05). (D) Total number of trophozoites present in the small intestine of infected WT and Mbl KO mice at day 7 and day 21 post infection (*p < 0.05). Values are represented on a logaritimic scale. (E) Relative mRNA expression levels of IL-17A in infected WT and Mbl KO mice were estimated by qRT-PCR at day 21 post infection and in WT and Mbl KO uninfected control mice (*p < 0.05, **p < 0.01).

Full size image

It has previously been shown that the activation of the complement cascade can induce the intestinal IL-17A axis through both C3aR42 as well as the complement anaphylatoxin C5a43, 44. In addition, Mbl2 can also modulate the release of both IL-1β and IL-6 production from monocytes, which could further trigger an IL-17 response42, 43, 45. It is therefore unclear whether the impaired ability of Mbl KO mice to clear a Giardia infection, as previously described (33), is due to the absence of a direct effect of Mbl2 on the trophozoites, or rather caused by a reduced IL-17A response in these mice. In order to investigate this, a G. muris infection study was performed in C57BL/6 Mbl KO mice to compare the IL-17A response to that of C57BL/6 WT mice. Similar as previously described40, trophozoite counts were significantly higher in C57BL/6 Mbl KO mice both at days 7 and 21 post infection compared to C57BL/6 WT mice (Fig. 6D). Analysis of the IL-17A expression levels by qRT-PCR showed that IL-17A was significantly upregulated in both the C57BL/6 WT and C57BL/6 Mbl KO mice at day 21 post infection. Interestingly, IL-17A levels were even higher in the C57BL/6 Mbl KO mice compared to C57BL/6 WT mice (Fig. 6E). This effect has previously also been reported in C57BL/6 Mbl KO mice that were subjected to intestinal inflammation by the administration of dextran sodium sulphate45. Based on this, the authors hypothesized that Mbl could act as an inhibitor to prevent an excessive inflammatory response. As an alternative explanation in the case of a Giardia infection, however, it is also possible that the higher number of trophozoites still present in the intestine of C57BL/6 Mbl KO mice, further triggers the IL-17A response. In contrast, Li et al. demonstrated that splenocytes collected from c3aR KO mice following G. duodenalis infection, produced reduced amounts of IL-17 in vitro in comparison to splenocytes from WT BALB/c mice. The authors suggest that these conflicting results could be attributed to the different species of mice that were used or the different timepoints that were analyzed40.

In addition to Mbl2, IgA also has a protective role against Giardia infections, since mice deficient for IgA can not eradicate a G. muris infection46. Interestingly, previous research has shown that Mbl can actually bind to IgA and consequently activate the lectin pathway of the complement system47. In order to test this possibility in the context of a Giardia infection, first, we investigated the level of Mbl2 binding to trophozoites in the absence of parasite specific IgA’s. Plates were coated with G. duodenalis trophozoites and incubated with different concentrations of recombinant Mbl2. Mbl2 was shown to bind the trophozoites in a concentration-dependent manner. A concentration of at least 2 μg/ml of recombinant Mbl2 resulted in a saturated binding of the trophozoites (Fig. 7B). In a second phase, the effect of parasite specific IgA’s on Mbl2 binding was analysed. For this, the level of Giardia-specific IgA present in fecal samples from WT G. muris infected mice and WT uninfected control mice first needed to be determined. Hereto, 96-well plates were coated with Giardia trophozoites and incubated with fecal extracts. Samples from WT G. muris infected mice contained higher levels of Giardia-specific IgA that bound to the trophozoites, in comparison to WT uninfected control mice (Fig. 7A). Finally, plates coated with Giardia trophozoites were incubated with fecal samples from WT G. muris infected mice, containing Giardia-specific IgA, and from WT uninfected control mice and recombinant Mbl2 was subsequently added at 2 μg/ml. The presence or absence of bound IgA, originating from the fecal extracts, did not alter the amount of recombinant Mbl2 that bound to the trophozoites (Fig. 7C), suggesting that there was no binding between Mbl2 and IgA. This is in line with Terai et al. who reported that Mbl was not able to bind native IgA but only to aberrantly formed IgA that had abnormal glycosylation or denaturation48. In addition to the direct interaction between Mbl2 and IgA, we also investigated whether Mbl2 would affect IgA production and secretion. This was done by measuring total fecal IgA levels in WT and Mbl KO mice by ELISA. The results showed that there was no significant difference in total IgA levels between WT and Mbl KO mice (Fig. 7D), which is consistent with results shown by Li et al.40 in c3aR KO mice. This suggests that the diminished ability of Mbl KO mice to clear a Giardia infection is not due to reduced IgA levels.

Figure 7
figure 7

Fecal total IgA levels, Giardia-specific IgA levels, and Mbl2 levels, as measured by ELISA. (A) Giardia-specific IgA levels were measured in fecal samples from C57BL/6 WT uninfected control and G. muris infected mice. Optical density is shown. (**p < 0.01). (B) ELISA plates coated with trophozoites were incubated with recombinant Mbl2 in different concentrations (0.02–4 μg/mL). The amount of bound Mbl2 was measured through optical density. (C) ELISA plates coated with trophozoites were incubated with fecal samples from either C57BL/6 WT uninfected control mice or WT G. muris infected mice. Recombinant Mbl2 (2 μg/mL) was added and the amount of bound Mbl2 was measured. (D) Total fecal IgA levels were measured in C57BL/6 uninfected WT mice and Mbl KO uninfected mice.

Full size image

In addition to the abovementioned cluster of genes that are transcriptionally regulated through IL-17RA following infection, there was a second cluster of 31 genes of which the expression was also IL-17RA dependent but independent on whether the animals were infected or not. These genes (highlighted with ** in Fig. 5A) are listed in Table 4 and their relative expression patterns are shown in Fig. 5C. A gene in this list that is extremely relevant in the context of a Giardia infection is the polymeric immunoglobulin receptor (pIgR). This receptor is responsible for the transport of IgA and IgM across epithelial tissues. Transport of IgA appears to be essential for the clearing of G. muris, since mice deficient for pIgR were unable to clear an infection with the parasite49. The transcriptome data presented here showed a 2-fold decrease in pIgR transcript levels in C57BL/6 IL-17RA KO mice compared to C57BL/6 WT mice. Quantitative RT-PCR analysis indicated that pIgR transcript levels were not affected by a Giardia infection (results not shown). It is also noteworthy that the gene encoding the J chain, which is a protein component of IgA and IgM that is required for their secretion into the mucosa, was also significantly upregulated following infection in C57BL/6 WT mice.

Table 4 List of upregulated genes as identified by RNA-seq in C57BL/6 WT uninfected control mice versus C57BL/6 IL-17RA KO uninfected control mice (A) and that also differentially respond upon infection in C57BL/6 WT infected mice versus C57BL/6 IL-17RA KO infected mice (B).
Full size table

Conclusions

The outcome of this study provides further insights into the central role of the IL-17A/IL-17RA axis in orchestrating the protective immune response against Giardia. The upregulation of IL-17A and subsequent activation of IL-17RA does not only induce the production of an array of antimicrobial peptides and complement factors but also regulates the secretion of IgA into the intestinal lumen. The data presented here further indicate that the combination of complement activation through Mbl2 and parasite-specific IgA’s is crucial to combat a Giardia infection.

Data availability statement

The RNAseq data is available upon request.