**3. Discussion**

*Brucella* infections have been associated with abortion in both humans and animals. Although the pathophysiology of this complication has not been completely elucidated, the inflammatory phenomena observed in the affected placenta [3,4] sugges<sup>t</sup> that, as with other pathogens causing abortion, placental inflammation may have a role in *Brucella*-induced abortion. As *Brucella* can reach the placenta by the hematogenous route, the maternal decidua is probably the initial site of placental colonization [14,15]. Given the known ability of decidual cells to respond to microbial PAMPs with an enhanced production of proinflammatory cytokines, and the known deleterious effect of placental inflammation on gestation, we decided to assess the ability of *Brucella* spp. to colonize decidualized stromal endometrial cells (T-HESC) and to induce the production of proinflammatory cytokines.

As shown here, *B. abortus* was able to infect both decidualized and non-decidualized T-HESC cells, although the initial number of intracellular bacteria was slightly higher for non-decidualized cells. This may relate to the fact that decidualized cells form an organized layer thus exposing less membrane surface to the environment. In addition, the pathogen was able to survive and replicate inside these cells. These findings are in line with the reported ability of *B. abortus* for intracellular replication in several phagocytic and non-phagocytic cells, including macrophages, epithelial cells and trophoblasts [13,25,37]. It has been widely demonstrated that this ability for intracellular survival in different cell types depends on the expression of a type IV secretion system encoded by the *virB* operon, which allows *Brucella* to modulate phagosome-lysosome fusion [33,34]. In line with this, we found that a *B. abortus* mutant lacking the *virB10* gene was unable to survive and replicate inside decidualized and non-decidualized T-HESC despite a similar ability of invasion compared to the wild type strain. In contrast, a mutant lacking the genes for the BtpA and BtpB proteins that interfere with TLR signaling exhibited invasion and replication abilities similar to the wild type strain. Importantly, *B. abortus* infection did not induce cytotoxicity, nor did it affect the decidualization status of cells, suggesting that the decidua might sustain the infection in affected individuals.

The mechanisms for *Brucella* invasion of non-phagocytic cells may vary according to the cell type considered. Whereas actin polymerization and microtubules have been involved in many cells [37], internalization in Vero cells does not depend on microtubules but depends on clathrin-mediated endocytosis [38]. The requirements for invasion of endometrial cells have not been reported. We found that *B. abortus* internalization was inhibited by cytochalasin D and to a lower extent by colchicine, which inhibit actin polymerization and microtubule formation, respectively. In contrast, internalization was not affected by MDC, an inhibitor of clathrin-mediated endocytosis.

As mentioned previously, placental *Brucella* infection is accompanied by the infiltration of inflammatory cells [3,4], which suggests that placental inflammation may have a role in *Brucella*-induced abortion as it does in abortion triggered by other pathogens. Our results show that *B. abortus* infection elicits the secretion of IL-8 and MCP-1 in either decidualized or non-decidualized T-HESC cells. IL-8 levels were higher in non-decidualized cells than in decidualized ones, whereas no significant difference was found for MCP-1. The higher production of IL-8 in non-decidualized cells may relate to the higher number of intracellular bacteria found in this condition as compared to decidualized cells, a downmodulating effect of decidualization on IL-8 production [39], or both. Nonetheless, these results sugges<sup>t</sup> that, although the decidualization status may influence the levels of some proinflammatory mediators, decidualized endometrial cells are capable of mediating a proinflammatory response to *B. abortus*. A few previous studies have shown that *Brucella* BtpA and BtpB proteins, which contain TIR motifs and can thus modulate TLR signaling, can reduce cytokine production in dendritic cells in vitro (IL-12, TNF-α) and in lung tissues in vivo (IL-12, CXCL-1, MCP-1) [35,36]. However, the potential modulating role of these proteins in *Brucella*-infected endometrial cells was unknown. At variance with those previous studies, we did not detect significant differences in the levels of the two chemokines here evaluated (IL-8 and MCP-1) between T-HESC infected with the wild type *B. abortus* strain or the *btpAbtpB* mutant strain. These results agree with those reported for the same chemokines in *Brucella*-infected human trophoblasts [12], and add support to the hypothesis that the immune responses of professional phagocytes are more influenced by the action of Btp proteins than those of non-phagocytic cells.

The secretion of both cytokines was reduced to basal levels by all the inhibitors tested, suggesting that all the signaling pathways (p38, Jnk1/2, and NF-kB) are involved. In line with these findings, previous studies have shown that several signaling pathways are involved in cytokine production by different cell types in response to *B. abortus*. For example, CCL20 secretion by human bronchial epithelial cells depends on p38, Jnk1/2, Erk1/2, and NF-kB [40], whereas in murine astrocytes TNF-α secretion depends on p38 and Erk1/2 signaling pathways [41].

Previous studies in several non-phagocytic cells have shown that not only live *B. abortus* but also some of its antigens can elicit the production of proinflammatory cytokines [42–44]. In line with these reports, we found that HKBA and OMVs from *B. abortus* elicit IL-8 and/or MCP-1 secretion in T-HESC cells. Obviously, these findings imply that the induction of chemokines in these cells does not depend on *Brucella* viability. At variance with HKBA and OMVs, *B. abortus* LPS did not elicit the production of the chemokines analyzed. This result is in line with previous studies in other cell types, which demonstrated that *B. abortus* LPS is a poor inducer of proinflammatory responses [42–45]. In contrast, most inflammatory responses are triggered by outer membrane lipoproteins, which induce TLR2 signaling [45].

As shown in this study, decidualized T-HESC produce proinflammatory mediators, including MCP-1, in response to infection with *B. abortus* or stimulation with its antigens. However, the decidua contains not only DSC but also a significant proportion of macrophages [19], with which DSC can establish reciprocal interactions [22,23]. In addition, the number of decidual macrophages could eventually augmen<sup>t</sup> in the context of locally increased MCP-1 levels induced by an infectious process. Therefore, it can be speculated that, during *B. abortus* infection in the pregnan<sup>t</sup> uterus, endometrial cells may respond not only to the stimulus of bacterial antigens but also to stimulation by factors produced by adjacent *Brucella*-infected macrophages. In support of this hypothesis, we found that the stimulation of decidualized T-HESC with CM from *B. abortus*-infected macrophages induced the production of IL-6, MCP-1, and IL-8 in a dose-dependent manner, a phenomenon not produced by stimulation with CM from non-infected monocytes. Additional studies using specific blocking agents revealed that IL-6 induction by CM is mediated by TNF-α and IL-1β, whereas the induction of MCP-1 and IL-8 is mediated by TNF-<sup>α</sup>. These findings are similar to those reported for the interaction between *Brucella*-infected macrophages and human trophoblasts [12].

Overall, these results sugges<sup>t</sup> a possible scenario in which DSC produce IL-6 and chemoattractants for monocytes/macrophages in response to *B. abortus* infection and/or in response to cytokines produced by *Brucella*-infected placental macrophages. Reciprocal stimulations between DSC and phagocytes may amplify these phenomena. These interactions may be long-lasting due to the ability of *Brucella* to survive and replicate within macrophages and DSC. Altogether, these proinflammatory responses may contribute to the gestational complications of brucellosis.

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