**2. Results**

### *2.1. Brucella abortus Infects and Replicates in Both Decidualized and Non-Decidualized T-HESC Cells*

Both decidualized and non-decidualized T-HESC endometrial cells were infected with *B. abortus* at a multiplicity of infection (MOI) of 250 bacteria/cell, and colony-forming units (CFU) of intracellular bacteria were determined at different times post-infection (p.i.). As shown in Figure 1, *B. abortus* was able to infect T-HESC cells in both conditions, although the initial number of intracellular bacteria (2 h p.i.) was slightly higher for non-decidualized cells (1125 ± 250 vs. 345 ± 32 CFU/well, mean ± SD). Besides wild type *B. abortus*, two additional strains carrying mutations in genes relevant for virulence were also tested for their capacity to infect and survive in T-HESC cells. These included a mutant lacking the *virB10* gene, widely reported as essential for the intracellular survival and replication of *Brucella* [33,34], and a double mutant lacking *btpA* and *btpB* genes which encode proteins able to interfere with TLR signaling [35,36]. As shown in Figure 1A, both mutant strains were able to infect decidualized and non-decidualized T-HESC at levels similar to the wild type strain. However, the ability to survive and replicate intracellularly differed between the *virB10* mutant and the other two strains. While CFU of intracellular bacteria increased along time for wild type *B. abortus* and the *btpAbtpB* mutant, showing intracellular replication, the CFU of the *virB10* mutant declined at the same time and no viable bacteria were detected in either condition at 48 h p.i. This later result confirmed in endometrial cells the essential role of *virB10* for the intracellular survival of *Brucella*.

Infection experiments were also carried out in the presence of specific inhibitors to examine whether *B. abortus* internalization by T-HESC cells depends on actin polymerization (cytochalasin D), microtubules (colchicine), or clathrin-mediated endocytosis (monodansylcadaverine, MDC). As shown in Figure 1B, *B. abortus* internalization was highly inhibited by cytochalasin D and to a lower extent by colchicine, but was not affected by MDC.

**Figure 1.** *Brucella abortus* invades and replicates in T-HESC cells. (**A**) Non-decidualized (T-HESC) and decidualized (T-HESCd) endometrial cells were infected with wild type *B. abortus* and two isogenic mutants (*virB10* and *btpAbtpB*), and colony forming unit (CFU) numbers of intracellular bacteria were determined at different times post-infection (p.i.). (**B**). Decidualized T-HESC were pretreated for 1 h with different doses of Colchicine (10, 5, 2.5 μM), Monodansylcadaverine (MDC; 200, 100, 5 μM), Cytochalasin D (2, 1, 0.5 μg/mL), or DMSO (vehicle) before infection with wild type *B. abortus*. Intracellular CFU were determined at 1 h p.i. Results are expressed as mean ± SD from three independent experiments run in duplicates. \*\*\* *p* < 0.001 versus control.

To assess whether the infection affected the viability of T-HESC cells or their decidualization status, the levels of lactate dehydrogenase (LDH) and prolactin were measured in culture supernatants of infected cells at 24 and 48 h p.i. and also in non-infected cells cultured in parallel. As shown in Figure 2, the infection with either wild type *B. abortus* or the *btpAbtpB* mutant did not modify the levels of LDH or prolactin as compared to non-infected cells at any time point, showing that it does not induce cytotoxicity or affect the decidualization of cells.

**Figure 2.** *B. abortus* infection does not induce cytotoxicity or alterations in decidualization in T-HESC cells. Decidualized T-HESC cells were infected or not (NI) with *B. abortus* wild type or its isogenic *btpAbtpB* mutant, and culture supernatants were harvested at 24 and 48 h p.i. to measure the levels of prolactin by ELISA (**a**) and the activity of lactate dehydrogenase (LDH) using a commercial non-radioactive cytotoxicity assay (**b**). In the latter assay, a control of 100% cell lysis (Max) was obtained by hypotonic lysis of the same number of non-infected cells. Results are expressed as mean ± SD from three independent experiments run in duplicates. ns: non-significant versus NI.

### *2.2. B. abortus Infection Induces the Secretion of Proinflammatory Chemokines in T-HESC Cells*

As mentioned above, DSC and ESC cell lines express several TLRs and NLRs, and respond to microbial PAMPs with an enhanced production of proinflammatory cytokines, including MCP-1, IL-6, IL-8, IL-1β, and RANTES [17]. To assess the ability of *B. abortus* to induce a proinflammatory response in T-HESC, these cells were infected with the wild type strain and the *btpAbtpB* mutant, and the levels of IL-8 and MCP-1 were measured in culture supernatants. The studies were performed on decidualized and non-decidualized cells to determine whether the proinflammatory response depends on the decidualization status. As shown in Figure 3, the infection with any of the *B. abortus* strains elicited the secretion of both chemokines in either decidualized or non-decidualized T-HESC, and this effect was mostly evident at 48 h p.i. At this time point, IL-8 levels were higher in non-decidualized cells than in decidualized ones (mean, 8539 vs. 4948 pg/mL), whereas no significant difference was found for MCP-1 (mean, 3197 vs. 3621 pg/mL).

**Figure 3.** *B. abortus* infection elicits chemokine secretion in T-HESC cells. Decidualized (T-HESCd) (**<sup>a</sup>**,**<sup>c</sup>**) and non-decidualized (T-HESC) (**b**,**d**) endometrial cells were infected or not (NI) with wild type *B. abortus* and the *btpAbtpB* mutant, and the levels of IL-8 (**<sup>a</sup>**,**b**) and MCP-1 (**<sup>c</sup>**,**d**) were measured by ELISA in culture supernatants harvested at 24 or 48 h p.i. Results are expressed as mean ± SD from three independent experiments run in duplicates. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001 versus NI.

To determine which signaling pathways may be involved in the induction of chemokine secretion, decidualized T-HESC cells were treated with SB203580 (p38 MAPK inhibitor), SP600125 (Jnk1/2 inhibitor), BAY 11-7082 (NF-κB inhibitor), or vehicle (dimethyl sulfoxide, DMSO) before and during the infection with *B. abortus*, and IL-8 and MCP-1 were measured as above. As shown in Figure 4, the secretion of both cytokines was not affected significantly by DMSO but was reduced to basal levels by all the inhibitors tested, suggesting that all the signaling pathways (p38, Jnk1/2, and NF-kB) are involved.

**Figure 4.** Several signaling pathways are involved in chemokine secretion in *Brucella*-infected T-HESC cells. Decidualized T-HESC cells were treated or not (NT) with SB203580 (SB, p38 MAPK inhibitor), SP600125 (SP, Jnk1/2 inhibitor), BAY 11-7082 (BAY, NF-κB inhibitor), or vehicle (DMSO) for 1 h, and were infected with wild type *B. abortus*. The inhibitors were kept throughout the experiment. At 48 h p.i. culture supernatants were harvested for measuring IL-8 (**a**) and MCP-1 (**b**) by ELISA. Non-treated non-infected cells (NI) served as controls. Results are expressed as mean ± SD from three independent experiments run in duplicates. Asterisks over bars indicate \*\*\* *p* < 0.001 or \*\*\*\* *p* < 0.0001 versus NI. Asterisks over lines indicate \*\* *p* < 0.01, \*\*\* *p* < 0.001 or \*\*\*\* *p* < 0.0001 versus NT. ns: non-significant.

Given the ability of *B. abortus* infection to induce the secretion of IL-8 and MCP-1 in T-HESC cells, experiments were carried out to determine whether such responses can be also elicited by stimulation with *B. abortus* antigens or, conversely, depend on bacterial viability. For this purpose, cells were stimulated with either heat-killed *B. abortus* (HKBA), or LPS or outer membrane vesicles (OMVs) from this bacterium, and chemokine levels were measured at 48 h p.i. As shown in Figure 5, HKBA (at 10<sup>9</sup> CFU/mL) elicited IL-8 and MCP-1 secretion by T-HESC cells, albeit at lower levels than those attained by the infection. In addition, IL-8 secretion was significantly induced by *B. abortus* OMVs. These results show that the induction of chemokines in these cells does not depend on *Brucella* viability.

**Figure 5.** The chemokine response of T-HESC to *B. abortus* does not require bacterial viability. Decidualized T-HESC cells were stimulated or not (NS) with two doses (10<sup>8</sup> and 10<sup>9</sup> CFU/mL) of heat-killed *B. abortus* (HKBA), or with lipopolysaccharide (LPS) or outer membrane vesicles (OMVs) from this bacterium, and IL-8 (**a**) and MCP-1 (**b**) levels were measured in culture supernatants at 48 h post-stimulation. Results are expressed as mean ± SD from three independent experiments run in duplicates. \* *p* < 0.05, \*\*\* *p* < 0.001 versus NS.

### *2.3. Factors Produced by Brucella-Infected Macrophages Stimulate Proinflammatory Responses in Decidualized T-HESC Cells*

The results shown above demonstrate that decidualized T-HESC produce proinflammatory mediators in response to infection with *B. abortus* or stimulation with its antigens. In the context of infection in the pregnan<sup>t</sup> uterus, however, endometrial cells may also receive stimulation by factors produced by adjacent infected macrophages [22,23]. To model this scenario in vitro, decidualized T-HESC cells were stimulated with conditioned media (CM) from *B. abortus*-infected macrophages and the levels of proinflammatory cytokines were measured in culture supernatants 24 h later. The preexisting levels of these cytokines in the CM were subtracted in order to calculate the secretion specifically induced by the stimulation. As shown in Figure 6, stimulation with CM from *Brucella*-infected macrophages induced the production of IL-6, MCP-1, and IL-8 in a dose-dependent manner (higher secretion for stimulation with CM diluted at 1/2). No significant secretion of any of these cytokines was induced by stimulation with CM from non-infected monocytes. Previous similar studies on the stimulation of other non-phagocytic cells have shown that IL-1β and TNFα are involved in the inducing effect of CM from *Brucella*-infected macrophages. To test whether these cytokines are also involved in the stimulation of IL-6, IL-8, and MCP-1 in decidualized T-HESC cells, experiments were performed in which CM were preincubated with a TNF-neutralizing antibody or T-HESC were preincubated with the natural antagonist of the IL-1 receptor (IL-1Ra). As shown in Figure 6, the stimulating effect of the CM on the secretion of IL-6 was significantly reduced by both pretreatments, implying that both TNFα and IL-1β are involved. For MCP-1 and IL-8, in contrast, only the preincubation with the anti-TNF antibody produced a significant reduction. Although the isotype control also produced a significant reduction of MCP-1 levels, the reducing effect of the specific anti-TNF antibody was much more pronounced. In summary, TNFα and/or IL-1β are involved in the ability of CM from *Brucella*-infected macrophages to stimulate the production of proinflammatory cytokines by decidualized T-HESC.

**Figure 6.** *Cont*.

**Figure 6.** Factors produced by *B. abortus*–infected macrophages stimulate cytokine production by endometrial cells. Decidualized T-HESC cells were stimulated or not (NS) with conditioned media from *B. abortus*-infected macrophages (CM Inf) or from uninfected macrophages (CM NI) at different dilutions (1/2, 1/5 or 1/10), and 24 h later culture supernatants were harvested to measure IL-6 (**a**), MCP-1 (**b**) and IL-8 (**c**) levels. In parallel experiments, CM Inf was preincubated with a TNF-α neutralizing antibody or an isotype control before addition to cells, or T-HESC were preincubated with the natural antagonist of the IL-1 receptor (IL-1Ra) before stimulation with CM Inf, and cytokine levels were measured as described. Results are expressed as mean ± SD from three independent experiments run in duplicates. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001, ns: non-significant. Asterisks over bars indicate differences versus NS.
