**3. Discussion**

Infections caused by the bacteria of the genus *Brucella* were mainly transmitted orally to human and animals. *Brucella* has a rapid capacity for infectivity in the oral infection model (by gavage or inoculation in the oral cavity), where after one hour of infection, bacteria were already found in the lumen and in the epithelium of the duodenum [42]. Few virulence factors of *Brucella* that are important for the establishment of infection through the oral route have been described, such as urease [43], which confers resistance to gastric acidity, cholylglycine hydrolase (CGH) [44] which induces resistance to bile salts, and the *Brucella* protease inhibitor Omp19 [42] which induces resistance to the action of proteases.

When gastrointestinal tract defense cells fail to capture microorganisms, they are drained mainly through the portal vein into the liver [45]. Previous studies have shown that in bacterial infections, higher concentrations of LPS are detected in the portal vein when compared to other hepatic or peripheral veins and, interestingly, bacteria can be cultivated even from healthy liver explants [46,47]. The phenotype of resistance exhibited by ST2-receptor deficient mice during oral infection was lost when intraperitoneal infection was performed. Considering that the portal vein might be the main route for systemic dissemination of *Brucella*, we first speculated that the liver from ST2-deficient mice might be mounting its own immune response and consequently, decreasing the number of viable bacteria and their ability to spread systemically. However, when we measured IFN-γ and TNFα production by liver cells, ST2 knockout and WT mice produced similar levels of these cytokines (Figure S1 found in the Supplementary Materials). Therefore, we sugges<sup>t</sup> that other mechanisms might be involved in reduced bacterial counts observed in ST2−/− livers.

The gut-associated lymphoid tissue (GALT), such as the Peyer's patches (PPs) along with the intestinal mucosal epithelium, act as a sentinel for recognition and initiation of immune responses against pathogenic bacteria [48]. The process of invasion of *Brucella* into the gastrointestinal tract occurs through its ability to translocate via M cells, which occurs by interaction with the prionic protein PrPc that are highly expressed on the apical surface of these cells [49]; however, this process does not lead to the rupture of the cell–cell junctions [50]. Another mechanism related to the invasion process is through the intestinal epithelial cells [51], but this mechanism has not ye<sup>t</sup> been fully clarified. Tight junctions (TJs) play an important role in intestinal function. TJs in intestinal epithelial cells are composed of di fferent junctional molecules, such as claudins, zonula occludens (ZO-1, -2, and -3), and occluding, among others. In this study, we determined the role of ST2 in *ZO-1, -2*, and *-3* and *claudin-1* expression in intestinal tissue. Herein, we showed that animals lacking ST2 had reduced expression levels of *ZO-1* and to a less extent that of *ZO-2* and *ZO-3*, when compared to WT mice. Regarding *claudin-1* mRNA transcripts, the levels of this TJ remained similar between both mouse groups. The reduced expression of zonula occludens (ZO) molecules might not have a direct relationship to intestinal permeability in this model since ST2−/− mice had reduced intestinal permeability, compared to WT animals, as measured by the FITC-dextran method. Rather, diminished expression of these tight junction gene products might correlate with enhanced IFN-γ production observed in ST2−/− animals, as recently demonstrated in the *Salmonella enteritis* infection model [52]. Breaking the epithelial barrier after oral infection resulted in increased intestinal permeability observed in WT mice and could be one important mechanism that facilitates the entry and spread of this pathogen. Studies using a model of ex vivo infection in the ileal bowel loop showed that the migration of *Brucella* through the intestinal epithelium occurs via endocytosis by the follicle-associated epithelium (FAE) in Peyer's patches or by its uptake by the penetrating dendritic cells of the FAE [49,53]. Additionally, Rosseti and collaborators (2013) [54] observed through microarray analysis that two pathways related to the intestinal epithelial barrier were repressed during the initial phase of *Brucella* infection, suggesting the subversion of the barrier function and facilitating transepithelial migration. Thus, a variety of pathogens use molecules involved in cell adhesion and invasion, such as the *Helicobacter pylori*, whose type IV secretion system injects one of its e ffectors (CagA) into the host cell, modifying several processes and culminating in the rupture of the epithelial barrier and invasion of the bacteria [55]. Although, we did not explore the infection of *Brucella abortus* in intestinal epithelial cells in vitro, the breaking of the epithelial barrier in vivo might be associated with the presence of the ST2 receptor, since no increase in intestinal permeability was observed in ST2 knockout mice after infection, when compared to WT animals.

Another mechanism related to the process of maintaining the epithelial barrier, involves amphiregulin (AREG), which plays a role in intestinal epithelial regeneration after injury [56] and in cellular proliferation [57]. The level of expression of this molecule after infection was similar in both animals analyzed, suggesting that the change in intestinal permeability observed in WT mice is not mediated by the participation of ST2 in transcriptional regulation of *AREG*. The intestinal mucus is one of the main components of defense against invasion of pathogens and protects the epithelium from physical damage. Muc2 mucin is produced and secreted by intestinal goblet cells. We believe that the increased expression of MUC2 in WT mice might be linked to augmented intestinal permeability. Recently, a higher expression level of the mucin glycoprotein Muc2 in enteroids following *Shigella flexneri* infections was reported [58]. These findings sugges<sup>t</sup> that mucus production might not be an important factor involved in the phenotype of decreased intestinal permeability observed in ST2 knockout-infected mice, and that other mechanisms are involved in the intestinal barrier of ST2−/− animals.

During infections, depending on the organ involved, IL-33/ST2 signaling might induce the necessary immune response to control the infectious foci, which might be a Th1 or Th2 type of response [38]. The increase of IFN-γ and TNFα after the infection observed in the knockout mice might be associated with a greater defense of the intestine against the invasion of *B. abortus*, which might be contributing to the phenotype of resistance observed in these animals. Previous studies have demonstrated the requirement of Th1-type cytokine profile to induce protection against *Brucella* infection [40,41]. IL-1β plays an important role in the defense against pathogens and also in the maintenance of the intestinal homeostatic balance and in the regeneration of the epithelium [59,60]. High levels of this cytokine are found in the intestinal mucosa in a normal state (steady-state), implying its importance in maintaining the mucosal barrier and in immune monitoring. The decrease in IL-1β levels observed in WT-infected mice corroborates the data of altered intestinal permeability through infection, suggesting that ST2 might have a regulatory role of this cytokine and consequently a function in the maintenance of intestinal permeability. Several studies have proposed that cell injury or death are the dominant mechanisms through which the IL-33 reaches the extracellular environment. Therefore, in a steady state, the IL-33 is not actively secreted by cells [35,61], being an important tool for the immune system, when there is a violation in the integrity of the mucosa, secondary to damage to the epithelial cells [61]. In this study, we observed that production of this cytokine in the small intestine was naturally higher in WT mice, compared to the knockout animals, suggesting that the change in intestinal permeability induced by oral infection was not through tissue damage, but via other mechanisms that need to be investigated.

The increase in MPO and EPO as an indirect measurement of neutrophils and eosinophils in the intestine of WT mice might contribute to the establishment of infection. Since *Brucella abortus* is an intracellular pathogen, and it is already described in the literature that neutrophils infected with *Brucella* are readily phagocytized by macrophages and replicate extensively within these cells, neutrophils then end up serving as "Trojan horse" vehicles for e fficient bacterial dispersion, intracellular replication, and establishment of chronic infections [62]. In ST2 knockout mice, MPO and EPO were decreased after infection when compared to WT animals, which might contribute to the initial resistance profile exhibited by these mice, since there are less infected granulocytes that can carry the pathogen to spread into other host cells and organs. The absence of ST2 increased the bactericidal activity of neutrophils and macrophages against *Staphylococcus aureus* in a sepsis model [63], by increasing the production of nitric oxide of these cells. Thus, we sought to investigate whether, in an in vitro scenario, macrophages would show higher production of NO against *Brucella* infection. We observed that nitric oxide production rate is similarly influenced in WT and ST2−/− macrophages, either through stimulation with *B. abortus* or LPS, in the presence or absence of IFN-γ, suggesting that bone marrow-derived macrophages have the same microbicidal potential, and that ST2 in the context of *B. abortus* infection is not involved in the regulation of NO production by these cells.

In summary, we observed that lack of ST2 is important in the model of *Brucella* oral infection but not when the animals are infected by the intraperitoneal route. In this study, we revealed that the oral infection by *Brucella abortus* alters the intestinal homeostasis in favor of its invasion and establishment of systemic infection, and the mechanisms involved in this process were partially dependent on the ST2 receptor. The ST2 receptor proved to be important in maintaining the epithelial barrier and in the negative regulation of the inflammatory immune response to oral infection through *B. abortus*.

### **4. Materials and Methods**
