The role of serine protease HtrA in acute ulcerative enterocolitis and extra-intestinal immune responses during Campylobacter jejuni infection of gnotobiotic IL-10 deficient mice

Campylobacter jejuni infections have a high prevalence worldwide and represent a significant socioeconomic burden. C. jejuni can cross the intestinal epithelial barrier as visualized in biopsies derived from human patients and animal models, however, the underlying molecular mechanisms and associated immunopathology are still not well understood. We have recently shown that the secreted serine protease HtrA (high temperature requirement A) plays a key role in C. jejuni cellular invasion and transmigration across polarized epithelial cells in vitro. In the present in vivo study we investigated the role of HtrA during C. jejuni infection of mice. We used the gnotobiotic IL-10−/− mouse model to study campylobacteriosis following peroral infection with the C. jejuni wild-type (WT) strain NCTC11168 and the isogenic, non-polar NCTC11168ΔhtrA deletion mutant. Six days post infection (p.i.) with either strain mice harbored comparable intestinal C. jejuni loads, whereas ulcerative enterocolitis was less pronounced in mice infected with the ΔhtrA mutant strain. Moreover, ΔhtrA mutant infected mice displayed lower apoptotic cell numbers in the large intestinal mucosa, less colonic accumulation of neutrophils, macrophages and monocytes, lower large intestinal nitric oxide, IFN-γ, and IL-6 as well as lower TNF-α and IL-6 serum concentrations as compared to WT strain infected mice at day 6 p.i. Notably, immunopathological responses were not restricted to the intestinal tract given that liver and kidneys exhibited mild histopathological changes 6 days p.i. with either C. jejuni strain. We also found that hepatic and renal nitric oxide levels or renal TNF-α concentrations were lower in the ΔhtrA mutant as compared to WT strain infected mice. In conclusion, we show here that the C. jejuni HtrA protein plays a pivotal role in inducing host cell apoptosis and immunopathology during murine campylobacteriosis in the gut in vivo.


INTRODUCTION
Campylobacter species are classical zoonotic pathogens, living predominantly as commensals in the gastrointestinal tract of a wide range of birds and mammals, including agriculturally important animals (Young et al., 2007;van Putten et al., 2009;Dasti et al., 2010;Gaynor and Szymanski, 2012). Thus, contaminated animal food products serve as a major source of Campylobacter infections in humans (Alter et al., 2011;Oyarzabal and Backert, 2011). The most prevalent Campylobacter species in human disease is C. jejuni, which represents the leading cause of bacterial infections in the gut and acute diarrheal disease worldwide (Friedman et al., 2000;Young et al., 2007;Mukhopadhya et al., 2011). Disease outcome in humans varies from mild, non-inflammatory, self-limiting diarrhea to severe, inflammatory, bloody diarrhea lasting for several weeks (Young et al., 2007;Oyarzabal and Backert, 2011). In addition, in a minority of infected persons, C. jejuni can be associated with the development of reactive arthritis and peripheral neuropathies, the Miller-Fisher and Guillain-Barrè syndromes, respectively (Nachamkin et al., 2008;Szymanski and Gaynor, 2012). In vivo and in vitro research studies performed in the last two decades revealed that C. jejuni exhibits various remarkable properties during infection. An important feature of C. jejuni is their ability to bind to and enter human gut epithelial cells causing intestinal tissue damage (Ó'Cróinín and . C. jejuni adherence to epithelial cells has been shown to involve a variety of proposed and confirmed outer membrane adhesins, including JlpA, PEB1, CadF, FlpA among others (Pei et al., 1998;Konkel et al., 2005;Poly and Guerry, 2008;Novik et al., 2010;Eucker and Konkel, 2012). C. jejuni invasion of cultured INT-407 and other cell lines has been observed to induce rearrangements of the host cytoskeleton by small Rho GTPases, Rac1, and Cdc42, which are directly linked to bacterial uptake (Krause-Gruszczynska et al., 2007aBoehm et al., 2011;Eucker and Konkel, 2012). C. jejuni can also cross the intestinal epithelial barrier as visualized in biopsies derived from human patients . Efforts with rodent and chicken infection model systems have been made to study pathogenicity mechanisms of C. jejuni in vivo, but each animal system has diverse limitations. Currently, disease manifestation such as gastroenteritis mimicking human camplylobacteriosis can be achieved in distinct infection models such as conventionally colonized infant wild-type (WT) and gnotobiotic IL-10 −/− mice Haag et al., 2012a,b). Notably, when infecting with a C. jejuni B2 strain (which is well known for its effective colonization properties) immediately after weaning, approximately 90% of conventionally colonized 3weeks-old infant mice developed self-limiting enterocolitis within 6-8 days resolving within 2 weeks post infection (p.i.) (Haag et al., 2012b). However, when infecting with other C. jejuni strains than B2 such as strains 81-176 or 11168, a huge variability in colonization and disease development could be observed in the infant mouse model. Irrespective of the C. jejuni strain, however, gnotobiotic IL-10 −/− mice get readily colonized by the pathogen at high loads following peroral infection. Very similar to immunocompromised patients, infected mice develop non-self limiting wasting ulcerative enterocolitis within 1 week p.i. (Haag et al., 2012a;Heimesaat et al., 2014). Importantly, the intestinal inflammation induced by C. jejuni in mice and humans is aggravated by C. jeuni lipooligosaccharide (LOS) via Toll-like-receptor-4 (TLR-4), as we could show previously (Haag et al., 2012a).
We and others have recently reported that the HtrA (high temperature requirement A) protein of C. jejuni is a novel virulence factor (Brøndsted et al., 2005;Baek et al., 2011a,b;Boehm et al., 2012;Hoy et al., 2012). Bacterial HtrA proteins represent a class of conserved heat shock induced serine proteases with additional chaperone activity, which were shown to have a significant impact on the virulence capabilities of various bacterial pathogens (Ingmer and Brøndsted, 2009;Frees et al., 2013;Skorko-Glonek et al., 2013). For example, it was demonstrated that growth of the htrA mutant was severely impaired at 44 • C as compared to WT C. jejuni and tolerance of the mutant bacteria against oxygen stress is strongly reduced (Brøndsted et al., 2005). In many different bacterial species, HtrA proteins are localized in the periplasm, where they form proteolytically active multimers with crucial function in the intracellular protein quality control machinery (Clausen et al., 2002(Clausen et al., , 2011. The class of HtrA proteins typically consists of a signal peptide, a trypsinlike serine protease domain and one or two PDZ [post synaptic density protein (PSD95), Drosophila disc large tumor suppressor (Dlg1), and zonula occludens-1 protein (ZO-1)] domains for protein-protein interactions (Kim and Kim, 2005). For long time it was assumed that HtrA family members are strictly acting intracellularly within the bacteria. However, very recently we have discovered a remarkably new feature of HtrA during infection. In various ε-proteobacteria such as H. pylori and C. jejuni, HtrA is actively secreted into the extracellular environment, where it cleaves the cell surface adhesion protein and tumor-suppressor E-cadherin (Hoy et al., 2010Boehm et al., 2012Boehm et al., , 2013. Infection experiments with C. jejuni in vitro indicated that HtrA can open the cell-to-cell junctions in cell monolayers by cleavingoff the ∼90-kDa extracellular NTF domain of E-cadherin Hoy et al., 2012). Deletion of the htrA gene leads to a defect in E-cadherin shedding and transmigration of C. jejuni across monolayers of polarized human MKN-28 epithelial cells in vitro . However, the potential relevance of the htrA gene for the interaction of C. jejuni with the host immune system has not been investigated so far.
To address this important question, we applied in the present study the C. jejuni infection model system of gnotobiotic IL-10 −/− mice. Here we investigated the role of C. jejuni HtrA in (i) colonization capacity, (ii) translocation, (iii) clinical outcome, (iv) intestinal inflammation, and (v) extra-intestinal sequelae including systemic immune responses following infection of gnotobiotic IL-10 −/− mice with the C. jejuni WT strain NCTC11168 and the isogenic knockout mutant strain NCTC11168 htrA.

ETHICS STATEMENT
All animal experiments were conducted according to the European Guidelines for animal welfare (2010/63/EU) with approval of the commission for animal experiments headed by the "Landesamt für Gesundheit und Soziales" (LaGeSo, Berlin, Germany; registration numbers G0123/12). Animal welfare was monitored twice daily by assessment of clinical conditions.

C. JEJUNI STRAINS AND GENETIC COMPLEMENTATION OF HTRA
For genetic complementation of htrA, we used the pCam-148 chromosomal C. jejuni complementation vector kindly provided by Dr. Dennis Linton (University Manchester, UK). Briefly, pCam-148 contains a 2,178 bp fragment of C. jejuni NCTC11168 genomic DNA (position 205,297 to 207,475) cloned into the SmaI restriction site of plasmid pUC18. pCam-148 contains a singular SpeI restriction site in the C. jejuni sequence within the pseudogene downstream of Cj0208. We used this SpeI site to introduce three additional restriction sites (NotI, MluI, and NruI) using the primer annealing approach. Subsequently, we amplified by PCR a 1,694 bp fragment of the htrA gene of C. jejuni NCTC11168, including 200-bp upstream and 75-bp downstream sequences using the primers HtrA-1 5 GTTATATTTCCTTAAAAATTTTTAC and HtrA-2 5 AGTTTTCCTTTATTTTAAACTTAAT. The resulting PCR product was cloned into the pSB-249 vector containing flanking NotI and MluI sites, respectively. The htrA gene was then further subcloned into the NotI and MluI sites of pCam-148. As a selection marker, we used a 795-bp kanamycin-resistant Aph cassette with its own promoter from plasmid pRYSK12 (kindly provided by Dr. Sabine Kienesberger, University Graz, Austria). This Aph cassette was cloned into the MluI and NruI restriction sites of pCam-148 next to the NCTC11168 htrA gene. The resulting htrA complementation vector was called pSB-250. pSB-250 was then transformed into the C. jejuni NCTC11168 htrA deletion mutant  and called NCTC11168 htrA/htrA. Correct integration of htrA in the C. jejuni chromosome was confirmed by PCR and standard sequencing. Expression of HtrA proteins was verified by Western blotting.

CASEIN ZYMOGRAPHY
Bacterial lysates, culture supernatants or recombinant HtrA were separated under non-reducing conditions in gels containing casein. Subsequently, gels were renatured in 2.5% Triton-X-100 and equilibrated in developing buffer Hoy et al., 2012). Caseinolytic activity was visualized by staining with 0.5% Coomassie Blue R250.

C. JEJUNI INFECTION OF MICE
Mice were infected with 10 9 viable colony forming units (CFU) of the C. jejuni parental strain NCTC11168 WT strain or the isogenic mutant strain NCTC11168 htrA by gavage in a total volume of 0.3 mL PBS on 2 consecutive days (day 0 and 1) as described (Haag et al., 2012a,b).

QUANTITATIVE ANALYSIS OF C. JEJUNI AND BACTERIAL TRANSLOCATION INTO OTHER ORGANS
Live C. jejuni were detected at time of necropsy (day 6 p.i.) in luminal samples taken from the duodenum, ileum, or colon diluted in sterile PBS by culture as described earlier . To quantify bacterial translocation into different organs, MLNs, spleen, liver, and kidney were homogenized in sterile PBS and analyzed by cultivating in dilution series on karmali agar (Oxoid, Wesel, Germany) in a microaerobic atmosphere at 37 • C for at least 48 h (Heimesaat et al., 2013). In addition, 0.5 mL of cardiac blood was streaked out immediately on karmali agar. The respective weights of luminal fecal or tissue samples were determined by the difference of the sample weights before and after asservation.

CYTOKINE DETECTION IN SERUM SAMPLES AND CULTURE SUPERNATANTS OF EX VIVO BIOPSIES TAKEN FROM COLON, MESENTERIC LYMPH NODES, SPLEEN, LIVER, AND KIDNEY
Colonic biopsies and kidneys were cut longitudinally and the former washed in PBS. MLNs or strips of approximately 1 cm 2 colon and liver tissue and additionally half of spleen and kidney were placed in 24-flat-bottom well culture plates (Nunc, Wiesbaden, Germany) containing 500 μL serum-free RPMI 1640 medium supplemented with penicillin (100 U/mL) and streptomycin (100 μg/mL; PAA Laboratories). After 18 h of incubation at 37 • C, culture supernatants as well as serum samples were analyzed for IFN-γ, TNF-α, and IL-6 by the Mouse Inflammation Cytometric Bead Assay (CBA; BD Biosciences) on a BD FACS Canto II flow cytometer (BD Biosciences). Nitric oxide (NO) was determined by Griess reaction as described earlier (Heimesaat et al., 2006).

QUANTITATIVE REAL-TIME PCR (QRT-PCR)
RNA was isolated from colonic tissues using the RNeasy Mini Kit (Qiagen). mRNA was reversed transcribed and analyzed in triplicate assays by TaqMan PCR using a sequence detection system (ABI Prism 7700; Applied Biosystems) as described previously (Wolk et al., 2002;Munoz et al., 2009). For detection of murine MUC-2 assays including double-fluorescent probes in combination with assays for the mouse housekeeping gene hypoxanthine phosphoribosyltransferase (HPRT) were purchased from Applied Biosystems. Expression levels were calculated relative to the HPRT expression.

ANTIBODIES AND WESTERN BLOTTING
C. jejuni cell pellets were lysed and proteins were separated by SDS-PAGE (Krause-Gruszczynska et al., 2007b;Wiedemann et al., 2012). The polyclonal rabbit α-HtrA antibody was raised against a conserved peptide corresponding to amino acid (aa) residues 288-301: C-QGDTKKAYKNQEGA. The α-CiaB antibody was generated against the epitope 597-610 (C-EIDNSGEFERYKKK) and the α-MOMP antibody against aa residues 400-413 (C-NLDQGVNTNESADH) in the corresponding proteins, respectively. All three peptides were conjugated to Limulus polyphemus haemocyanin carrier protein, and two rabbits each were immunized by Biogenes GmbH (Berlin, Germany) using standard protocols . The resulting antiserum was affinity-purified and the specificity against the proteins in C. jejuni was confirmed by Western blotting Backert and Hofreuter, 2013). Horseradish peroxidase-conjugated anti-rabbit polyvalent sheep immunoglobulin was used as secondary antibody (DAKO Denmark A/S, DK-2600 Glostrup, Denmark). Blots were developed with ECL Plus Western blot reagents (GE Healthcare, UK limited Amersham Place, UK) as described (Conradi et al., 2012;Hirsch et al., 2012).

STATISTICAL ANALYSIS
Mean values, medians, and levels of significance were determined using Mann-Whitney-U-test. Two-sided probability (P) values ≤ 0.05 were considered significant. All experiments were repeated various times as indicated in the corresponding figure legends.

HTRA DOES NOT AFFECT THE COLONIZATION CAPACITY OF C. JEJUNI IN GNOTOBIOTIC IL-10 −/− MICE
In order to eradicate the chronic colitogenic stimulus derived from the commensal intestinal microbiota, IL-10 −/− mice were pre-treated with a quintuple antibiotic regimen for approximately 4 months starting immediately after weaning (Haag et al., 2012a). The resulting gnotobiotic IL-10 −/− mice were then perorally infected with 10 9 CFU of either C. jejuni NCTC11168 WT or isogenic htrA mutant (NCTC11168 htrA) strain, each grown to stationary phase on 2 consecutive days (day 0 and 1). Control blots demonstrate that equal amounts of C. jejuni were infected per sample and HtrA is not expressed in the htrA mutant as expected ( Figure 1A). We also confirmed that our htrA mutant is non-polar because genetic complementation experiments with the corresponding WT gene restored (i) expression of proteolytically active HtrA multimers (Figures S1A,B), (ii) C. jejuni growth at high temperature (44 • C) (Figures S2A,B) and (iii) growth under high oxygen stress conditions (Figures S2A,C). Six days following peroral challenge all mice harbored comparable pathogen loads with either strain in the duodenum, ileum and colon, which is indicative for an uncompromised colonization capacity of the htrA mutant strain in vivo ( Figure 1B). After 6 days 12.5% of IL-10 −/− mice infected with either the htrA mutant or the C. jejuni WT strain contained viable pathogens also in MLNs as shown by culture. Notably, bacterial presence in extra-intestinal compartments such as spleen, liver, kidney, or blood could not be observed with either strain based on CFU determination (data not shown).

IMPACT OF HTRA ON ACUTE ENTEROCOLITIS IN C. JEJUNI INFECTED GNOTOBIOTIC IL-10 −/− MICE
Six days p.i. with the C. jejuni NCTC11168 WT strain, gnotobiotic IL-10 −/− mice were decisively compromised by acute enterocolitis as indicated by wasting clinical symptoms, diarrhea and occurrence of blood in liquid feces in up to 90% of cases (Figures 2A,B). In contrast, mice infected with the htrA mutant displayed significantly less severe clinical symptoms (p < 0.0005;  Figure 2A) and lower frequency of bloody diarrhea as compared to controls infected with the WT strain (50.0 vs. 89.7%, respectively; p < 0.001, Figure 2B). Given that intestinal inflammation results in a significant shortening of the intestines (Heimesaat et al., 2006;Bereswill et al., 2011;Haag et al., 2012a), we further assessed the colonic lengths upon C. jejuni infection. At day 6 p.i., IL-10 −/− mice infected with the htrA mutant displayed longer colons as compared to WT strain infected control animals (p < 0.05; Figure 2C). These results provide first evidence that HtrA aggravates the inflammatory outcome of C. jejuni infection. This was further confirmed by histopathological analysis of paraffin embedded colonic sections. Microscopical investigations of H&E-stained tissues revealed that htrA mutant infected mice displayed significantly lower histopathological scores as compared to mice infected with the C. jejuni WT strain at day 6 p.i. (p < 0.0001, Figure 2D). The htrA mutant induced rather mild inflammatory changes whereas C. jejuni WT strain infected mice exhibited acute enterocolitis characterized by ulcerations of and bleeding into the colonic mucosa as well as by diffuse mucosal and submucosal leukocytic infiltrates, loss of goblet cells and crypt drop-outs ( Figure 2D).

HTRA TRIGGERS C. JEJUNI-MEDIATED INTESTINAL IMMUNE RESPONSES
The role of C. jejuni HtrA in intestinal inflammation was next assessed by microscopical quantification of apoptotic and proliferating cells as well as infiltrating immune cells. This was achieved by specific immunohistochemical stainings of colonic paraffin sections. Six days following C. jejuni WT strain infection, gnotobiotic IL-10 −/− mice displayed a multifold increase of apoptotic cells, neutrophils, macrophages and monocytes, B and T lymphocytes as well as regulatory T cells (Treg) in the colonic mucosa and lamina propria (p < 0.001 − 0.0001 as compared to naive control animals; Figure 3). The induction of apoptosis and the increase in inflammatory immune cells, however, were significantly less pronounced in animals infected with the C. jejuni htrA mutant except for the T cell and Treg populations (p < 0.05 − 0.0001 as compared to animals infected with the C. jejuni WT strain; Figure 3). Given that Ki67 comprises a nuclear protein that is associated with and necessary for cellular proliferation (Scholzen and Gerdes, 2000), we stained colonic paraffin sections against Ki67 to detect proliferative measures of the colonic epithelium counteracting cellular destruction during the inflammatory process. Interestingly, the htrA mutant induced significantly higher numbers of Ki67 + proliferating cells as compared to the WT strain (p < 0.05; Figure 3B). The C. jejuni induced colonic immune cell responses were accompanied by increased expression of pro-inflammatory cytokines in the large intestine. Until day 6 p.i., levels, of IFN-γ and IL-6 and, in addition, NO were multifold increased in colonic ex vivo biopsies upon C. jejuni infection. Cytokines increased to a significantly lower extent in gnotobiotic IL-10 −/− mice infected with the htrA mutant as compared to WT strain infected mice (p < 0.05 − 0.01; Figures 4A-C). In addition, C. jejuni infected mice displayed increased NO secretion into draining MLNs at day 6 p.i. as compared to uninfected controls. This increase, however, was significantly lower in mice infected with the htrA mutant vs. the parental WT strain infected animals (p < 0.05; Figure 4D). Given that a proper mucus layer is a pivotal barrier protecting the intestinal epithelium from intestinal pathogens, we next determined the mRNA expression levels of mucin-2 (MUC-2), which is mainly secreted from goblet cells in the epithelial lining of the large intestine (Allen et al., 1998;Naughton et al., 2014). Six days following C. jejuni infection of IL-10 −/− mice, colonic mucin-2 mRNA was significantly downregulated by C. jejuni WT strain, but to a lesser extent by the htrA knockout mutant (p < 0.05; Figure 4E). Taken together, the less severe clinical, histopathological and inflammatory outcome of enterocolitis in mice infected with the C. jejuni htrA mutant strain was paralleled by a higher mucin-2 expression level as compared to control animals infected with the parental strain.

HTRA IS ALSO INVOLVED IN THE INDUCTION OF EXTRA-INTESTINAL PRO-INFLAMMATORY IMMUNE RESPONSES IN C. JEJUNI INFECTED GNOTOBIOTIC IL-10 −/− MICE
We next investigated potential systemic pro-inflammatory immune responses upon C. jejuni infection. At day 6 p.i. with either strain, TNF-α and IL-6 serum levels were increased as compared to naïve controls, but less distinctly in htrA mutant strain infected mice (p < 0.005 and p < 0.05, respectively; Figure 5). Hence, C. jejuni htrA knockout mutation not only reduces intestinal but also systemic inflammatory responses upon infection. Unexpectedly, 6 days following infection with the htrA gene mutant but not the WT strain, IFN-γ, TNF-α, and NO levels were up-regulated in splenic ex vivo biopsies (p < 0.05 − 0.0005; Figures 6A-C). Furthermore, secretion of IL-6 into spleens were higher following htrA mutant infection as compared to the parental strain at day 6 p.i. (p < 0.0001, Figure 6D).
We have recently shown that following long-term C. jejuni infection (i.e., more than 100 days p.i.) conventionally colonized infant mice exhibited pro-inflammatory immune responses at extra-intestinal locations such as liver and kidneys (Haag et al., 2012b;Heimesaat et al., 2013). We were therefore interested to investigate whether extra-intestinal sequelae might arise even after relatively short-term C. jejuni infection of gnotobiotic IL-10 −/− mice suffering from acute enterocolitis. To address this idea, we assessed inflammatory changes in H&E-stained paraffin sections of liver and kidneys. Six days following C. jejuni infection with either strain mice displayed only minimal to mild lobular  and/or portal inflammatory infiltrates in livers and mild focal hypercellularity in kidneys (Figure 7).
We next determined pro-inflammatory cytokine responses in ex vivo biopsies from liver and kidneys upon C. jejuni infection of gnotobiotic IL-10 −/− mice. Six days following C. jejuni infection hepatic IFN-γ, TNF-α, and IL-6 concentrations increased (p < 0.05 − 0.0001; Figures 8A-C), but to even higher levels upon htrA knockout mutant as compared to parental C. jejuni WT strain infection (p < 0.05; Figures 8A-C). Notably, the standard deviation in the cohort infected with the htrA gene knockout mutant was rather high. Interestingly, hepatic NO levels increased 6 days following WT infection, but not during htrA mutant strain infection (p < 0.05; Figure 8D), which helds also true for NO secretion in ex vivo biopsies taken from kidneys (p < 0.05; Figure 8E). Furthermore, TNF-α levels increased in kidneys of C. jejuni infected IL-10 −/− mice (p < 0.0001; Figure 8F), but less distinctly in the htrA knockout mutant infected group (p < 0.05; Figure 8F). Taken together, even as early as 6 days following C. jejuni infection pro-inflammatory immune responses could be detected in extra-intestinal compartments such as liver and kidneys that were even more pronounced in livers, but less distinct in kidneys of htrA mutant as compared to parental WT strain infected IL-10 −/− mice suffering from acute enterocolitis.

DISCUSSION
The mucosa in the intestine of mammals forms a tight barrier, which protects against commensals and other microbes  present in the intestinal lumen. Previous infection studies in mice have revealed that mutation of the serine protease HtrA in a number of pathogenic bacteria results in attenuated virulence properties (Li et al., 1996;Humphreys et al., 1999;Wilson et al., 2006;Yuan et al., 2008;Lewis et al., 2009;De Stoppelaar et al., 2013). Bacterial HtrA may be involved in degradation of damaged bacterial proteins that accumulate during the hostile conditions in the macrophages and accordingly, htrA mutants of the facultative intracellular Salmonella typhimurium, Listeria monocytogenes, and Yersinia enterocolitica are all sensitive to oxidative agents and show reduced survival rates in macrophages (Bäumler et al., 1994;Yamamoto et al., 1996;Wilson et al., 2006). The individual contribution of chaperone and protease activity of HtrA to bacterial survival and multiplication in the host is difficult to dissect. However, for S. typhimurium, both activities appear to be required for proper systemic infection in mice (Lewis et al., 2009). Interestingly, after oral infection with an htrA mutant, the pathogen cannot translocate accurately from the Peyer's patches to other organs. In contrast, intravenous infection of mice resulted in full colonization of livers and spleens by the S. typhimurium htrA mutant, suggesting that HtrA is required to overcome the intestinal barrier (Humphreys et al., 1999). Expression of HtrA is also important for cell-to-cell spread of Shigella flexneri (Purdy et al., 2002(Purdy et al., , 2007. Previous in vitro infection studies with C. jejuni have indicated that inactivation of htrA reduced the bacterial adherence to cultured epithelial cells (Brøndsted et al., 2005). In addition, C. jejuni htrA mutants exhibited reduced invasion rates (Novik et al., 2010;Baek et al., 2011a). Interestingly, HtrA chaperone activity appears to be important for efficient binding of C. jejuni to epithelial cells, while the HtrA protease activity seems required for maximal host cell entry after the bacteria have adhered to them (Baek et al., 2011a). Under laboratory conditions in vitro, HtrA chaperone activity is necessary for growth of C. jejuni at high temperatures or under oxidative stress conditions, whereas HtrA protease activity is only essential during severe stress (Baek et al., 2011b).
We have recently shown in a series of in vitro experiments that HtrA secreted by C. jejuni constitutes a novel bacterial virulence determinant, which opens cell-to-cell junctions through cleavage of E-cadherin and probably other host factors Hoy et al., 2012). In the present study, we investigated for the first time the impact of the htrA gene in C. jejuni-induced immunopathology in vivo and applied the acute C. jejuni infection model of gnotobiotic IL-10 −/− mice. To prevent conventionally colonized IL-10 −/− mice from chronic colitis exerted by antigenic stimuli through the conventional intestinal microbiota, mice were pre-treated for 4 months with a quintuple antibiotic regimen starting immediately after weaning (Haag et al., 2012a). Six days following peroral infection with C. jejuni, mice harbored high intestinal loads of the non-polar knockout mutant strain NCTC11168 htrA, which were comparable to those detected in mice upon infection with the parental WT strain NCTC11168. Hence, inactivation of the htrA gene did not down-regulate the overall high colonization capacity of C. jejuni in gnotobiotic mice. At first sight, these data seem to contradict a previous report where HtrA chaperone activity was shown to be required for efficient binding of C. jejuni to cultured INT-407 epithelial cells (Baek et al., 2011a). However, INT-407 cells do not form polarized cell layers, while polarized epithelial cell layers are found in the intestine of live mice. Thus, the receptor availability in both systems is certainly different, which could be one reason to explain these findings.
Importantly, gnotobiotic IL-10 −/− mice infected with the parental WT strain were severely compromised and developed ulcerative enterocolitis with bloody diarrhea and wasting symptoms, hence mimicking severe campylobacteriosis in immunocompromised patients (Haag et al., 2012a). Remarkably, at day 6 p.i. ulcerative enterocolitis was less distinct in mice infected with the htrA mutant, which displayed significantly less severe immunopathology in the intestinal tract as compared to mice infected with the C. jejuni WT strain. Interestingly, htrA mutant infected mice exhibited higher Ki67 + proliferating cell numbers in the colonic mucosa as compared to WT strain infected controls. This might be indicative for up-regulated compensatory properties in order to counteract cell destruction during immunopathology. Furthermore, not only local, but even systemic immune responses were less pronounced upon htrA mutant infection as indicated by significant lower serum levels of pro-inflammatory cytokines such as TNF-α and IL-6 in htrA mutant as compared to WT strain infected mice. The amelioration of devastating enteric and systemic disease, as in this hyper-acute model system, underlines the biological relevance of individual C. jejuni virulence factors in vivo. Furthermore, in the present and previous infection studies with conventional infant, germfree or with human microbiota re-associated germfree WT mice Haag et al., 2012b), translocation of viable C. jejuni of either strain from the intestine to the MLNs could be observed in a subset of animals, whereas no bacterial CFUs could be cultured from extra-intestinal locations such as spleen, liver, kidney, or cardiac blood. This is good agreement with the observation of very rare cases of extra-intestinal C. jejuni-associated disease affecting the liver, lung, heart, or spleen that have been reported in immunocompromised patients suffering from C. jejuni bacteremia (Pigrau et al., 1997;Tee and Mijch, 1998;Crushell et al., 2004) and previous infection studies of isolator-raised germfree mice (Fauchère et al., 1985). Whereas viable C. jejuni could be cultured from MLNs of infected mice more than 3 weeks p.i., the pathogen was cleared from the circulation and extra-intestinal organs such as liver and spleen within 24 h. The authors proposed that non-specific bactericidal factors such as complement or phagocytic cells might have been responsible for rapid pathogenic clearance. However, the histopathological sequelae within the respective organs were not reported (Fauchère et al., 1985). Surprisingly, in the present study histopathological analysis revealed rather mild inflammatory changes in extra-intestinal organs such as liver and kidneys that did not differ 6 days p.i. with either C. jejuni strain. In gnotobiotic WT mice, however, which can be readily colonized by C. jejuni, we did not observe any overt extra-intestinal histopathological changes in respective H&E stained paraffin sections by day 42 p.i. (unpublished observations). It is therefore most likely that a significant intestinal pro-inflammatory scenario (as seen in the gnotobiotic IL-10 −/− mouse model) is a prerequisite for subsequently induced extra-intestinal inflammatory responses as shown in the present study. In line with this, in another infection model, 3 weeks old infant mice (harboring a conventional microbiota) were infected with C. jejuni B2 strain immediately after weaning and exhibited histopathological sequelae in liver, kidneys, lungs, and cardiac muscle more than 100 days following C. jejuni infection (Haag et al., 2012b). Interestingly, mice were asymptomatic C. jejuni carriers until day 103 p.i. and the extra-intestinal organs were exclusively C. jejuni culture negative (Haag et al., 2012b). Furthermore, the vast majority of inflammatory cells that had accumulated in the respective extraintestinal organs comprised CD3 + T lymphocytes (Heimesaat et al., 2013). It is highly likely that pro-inflammatory immune cells have been attracted to the site of infection and cleared the pathogen early in the course of infection and reside further in the respective organ. This would explain the sterile inflammatory responses in liver and kidneys, both in less than 1 week (this study) and more than 3 months p.i. (Haag et al., 2012b;Heimesaat et al., 2013). Hence, the influx of pro-inflammatory immune cells upon even short-term C. jejuni infection might explain the increased levels of pro-inflammatory cytokines such as IFN-γ, IL-6, TNF-α, and NO detected in livers and the latter two in kidneys at day 6 p.i. Like in the colon, renal TNF-α and NO as well as hepatic NO levels were significantly lower in htrA knockout mutant as compared to parental strain infected gnotobiotic IL-10 −/− mice at day 6 p.i. Unexpectedly, IFN-γ, IL-6, and TNF-α concentrations in spleen and liver were even higher 6 days following htrA knockout mutant as compared to parental WT strain infection. Notably, the standard deviations in the cohorts infected with the htrA gene knockout mutant were rather high, which held true for individual experiments as well as for pooled data sets. Hence, the individual variabilities upon infection needs to take into account when judging for the biological relevance of the observed effects in the absence of htrA. In line with these unexpected data derived from livers of infected mice, splenic IFN-γ, IL-6, and TNF-α levels increased only upon htrA knockout mutant, but not parental WT strain infection when compared to naive controls. It is therefore tempting to speculate whether in the course of the observed systemic immune responses more immune cells (such as dendritic cells and/or lymphocytes) might have been activated/imprinted in the spleen by the C. jejuni htrA mutant present in the intestine as compared to parental WT strain infection or stimulated by distinct (so far unknown) circulating soluble bacterial factors derived from the pathogen which in turn could counteract intestinal disease. Further studies need to unravel this fascinating infection phenomenon.
Taken together, C. jejuni is one of the most important zoonotic pathogens causing food-borne gastroenteritis and potentially more severe diseases. Crossing the intestinal epithelial barrier and host cell invasion by C. jejuni are considered to constitute primary reasons of gut tissue damage in humans. However, the molecular mechanisms as well as major bacterial and host cell factors involved in these activities are poorly understood. Using the IL-10 −/− knockout mouse infection model system, the results presented in this study demonstrate for the first time that the C. jejuni HtrA serine protease is a novel virulence factor which aggravates enterocolitis in situ by causing a substantial amount of cell damage, aggravation of intestinal apoptosis and inflammation upon C. jejuni infection in vivo accompanied by significant systemic pro-inflammatory immune responses. These observations are in line with our earlier in vitro studies showing that HtrA of C. jejuni targets epithelial E-cadherin-based cell-to-cell junctions Hoy et al., 2012). Future work should address important questions such as how C. jejuni can trigger HtrA secretion into the extracellular space, to identify E-cadherin cleavage sites by HtrA and to search for novel host cell targets involved in the above discovered activities in vivo. We further demonstrated non-polarity of the htrA knockout mutant strain by complementation of phenotypes in vitro; however, this does not warranty non-polarity of the mutation in vivo. The investigation of epithelium-bound or internalized C. jejuni in ongoing studies will further complete our understanding of the complex interactions of HtrA with the epithelium.

FINANCIAL DISCLOSURE, GRANT SUPPORT
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.