*2.7. Statistical Analysis*

Statistical analysis was carried out with PrismTM 7 (GraphPadTM, San Diego, CA, USA). The normality of the data obtained was evaluated using the Kolmogorov–Smirnov test. Accordingly, Kruskal–Wallis and Sidak's multiple comparison tests were applied and data were depicted as means of all independent experiments. Differences among groups were considered significant when *P* < 0.05.

#### **3. Results and Discussion**

Candidemia has been increasing in the last decades, especially among individuals under chemotherapy programs, as well as in those who are HIV-positive, hospitalized, or catheterized [2,53]. *C. albicans* is still the most frequent isolated yeast, but *C. glabrata* has become one of the most threatening non-*Candida albicans Candida* (NCAC) spp., mostly due to its high antifungal resistance [2,54]. Though human clinical data demonstrate that immunosuppression is a risk factor for *C. glabrata* infections, it is not an absolute prerequisite for *C. glabrata* candidiasis [55]. Hence, increasing the data on the host immune response to *C. glabrata* and revising the efficacy of chemotherapeutic approaches to treat infections caused by this fungus are of major value. The murine model is a suitable one to address both issues, alone or combined [56].

#### *3.1. Fungal Burden Progression Differs Substantially between Liver and Kidneys*

The fungal burden of CD1 mice infected intravenously with *C. glabrata* biofilm cell suspensions and subsequently treated with echinocandins was assessed in the liver and kidneys 72 h post-infection. No differences were observed among the different infected groups.

In contrast to *C. albicans*, which can heavily infect the kidneys [57], a tropism of *C. glabrata* to the liver was clearly noticed. High CFU counts were detected on this organ (Figure 1), in contrast to the low or non-detected CFU counts in the kidneys ( ≤3 × 10<sup>4</sup> CFU/g kidney). The low colonization of this organ, as compared to the liver or brain in immunocompetent mice systemically infected with *C. glabrata*, was also reported by other authors [1,31,58–60]. Nevertheless, Kaur et al. [59], Srikantha et al. [60], and Brieland et al. [58] stated that *C. glabrata* could be recovered after several days in the kidneys, liver, spleen, hearts, lungs, brains, and lungs. Moreover, Atkinson et al. [61] described that fungal burdens were 10<sup>4</sup> to 10<sup>8</sup> in immunocompromised mice in the spleen and kidneys. Nonetheless, it is important to stress that the differences in mouse strains and immunocompetence status, *C. glabrata* strains, animal age and gender, or even the concentration of the inoculum used do not allow a direct comparison of published data [31]. In addition, past *in vitro* reports have shown that susceptible *C. glabrata* strains can become resistant in less than four days of continuous culture with low doses of drugs, such as fluconazole [1,16] and echinocandins [62–65]. Thus, it is plausible that a fast increase of resistance could have been observed in vivo. Moreover, the inoculum exclusively contained biofilm cells, known to be more resistant than their planktonic counterparts [66–72].

**Figure 1.** Liver fungal burden of CD1 mice 72 h after intravenously challenged with 1 × 10<sup>8</sup> biofilm cells plus two cycles of treatment with PBS, caspofungin (Csf), or micafungin (Mcf). Data are representative of two independent experiments. Each symbol represents an individual mouse, and horizontal bars are means of colony forming unit (CFU) numbers for each group. The obtained results are displayed as CFU/liver. Controls (naïve; PBS + Csf; PBS + Mcf), *n* = 2; Cg + Csf, *n* = 8; Cg + Mcf, *n* = 8. No statistical differences were observed among infected groups (evaluated by Kruskal–Wallis (Overall ANOVA *P* < 0.05) and post hoc Sidak's multiple comparison tests). Cg—*Candida glabrata* ATCC2001.

#### *3.2. Host Immune Response to Hematogenously Disseminated Candidiasis*

In contrast to the considerable work that has been described on the host immune response to *C. albicans*, the immune mechanisms elicited in the course of *C. glabrata* infections are far less explored.

Neutrophils and macrophages are in the first line of host immune defence against *Candida* spp. cells infecting the bloodstream or the endothelia [73–75]. Clinical observations and experimental studies have demonstrated the main role of polymorphonuclear leukocytes in mediating host protection against systemic *C. albicans* infections [76–78]. In mice, neutrophils have a Gr-1high surface phenotype and macrophages typically express the F4/80 cell surface marker. Previous reports have shown that, in *C. albicans* infections*,* Gr-1+ splenocytes may have immunosuppressive function and F4/80+ cells may play a pro-inflammatory role [79,80]. The expression of these two surface markers was analyzed using flow cytometry in the spleen of CD1 mice 72 h after i.v. infection with 1 × 10<sup>8</sup> *C. glabrata* biofilm cells. Myeloid cells (CD11b+) displaying the phenotypes F4/80high Gr-1neg, F4/80high Gr-1high, and F4/80neg/low Gr-1high were respectively considered macrophages, inflammatory monocytes, and neutrophils [81]. The gating strategies employed in this study are shown in Figure 2. As shown in Figure 3A, a significant increase in the numbers of inflammatory monocytes was observed in the spleen of infected mice, while those of neutrophils and macrophages remained within control values. No significant differences, however, were observed among treated groups. These results are in accordance with previous reports [31,82,83]. Unlike *C. albicans* infections, for which high neutrophil infiltration is a commonly observed feature, *C. glabrata* infections are not associated with massive neutrophil infiltration. Indeed, *C. glabrata* infection has mainly been associated with mononuclear cell infiltration and is far less inflammatory. One of the reasons given to explain this

disparate outcome is that *C. albicans* hyphae cause significant host cell damage, which results in the extensive recruitment of myeloid cells and the production of pro-inflammatory cytokines [31,82,83].

**Figure 2.** Gating strategy applied for the flow cytometry data analysis. Following leukocyte selection based on Forward Scatter Area (FSA) and Side Scatter Area (SSA), doublets were excluded based on FSA and Forward Scatter Height (FSH) parameters, and dead cells were further excluded by fixable viability dye (FVD) incorporation. Dendritic cells were gated as CD11chigh MHC class II+ cells. Myeloid cells were defined as CD11b+ cells that were further divided into macrophages (CD11b+ F4/80high MHC class IIlow) and Gr-1+ cells. Within the latter, neutrophils were defined as CD11b+ Gr-1+MHC class II− and inflammatory monocytes were gated as CD11b+ Gr-1+MHC class II+ cells.

Additionally, other reports have shown that *C. glabrata* is recognized and phagocytized by macrophages at a much higher rate than *C. albicans* [84]. After recognizing pathogens, macrophages release cytokines that help coordinate the immune response. However, when *C. glabrata* is internalized by macrophages, it interferes with the phagosome maturation process [85], surviving through autophagy and replicating inside the phagosome until the eventual bursting of the phagocyte [59,85,86]. Here, no elevated numbers of macrophages were detected in the spleen of infected mice as compared to noninfected controls (Figure 3C), which indicates that the recruitment or local proliferation of these cells does not occur in response to *C. glabrata*.

**Figure 3.** CD1 mice were challenged intravenously with 1 × 10<sup>8</sup> biofilm cells and then treated with PBS, caspofungin (Csf), or micafungin (Mcf). The obtained results are displayed as the total number of cells of indicated populations: (**A**) inflammatory monocytes, (**B**) neutrophils, and (**C**) macrophages. The numbers of animals used were as follows: controls (naïve; PBS + Csf; PBS + Mcf), *n* = 2; Cg + Csf, *n* = 8; Cg + Mcf, *n* = 8. Statistical differences were evaluated using Kruskal–Wallis and post hoc Sidak's multiple comparison tests (Overall ANOVA *P* < 0.05). Cg—*Candida glabrata* ATCC2001. \* *P* < 0.05; \*\* *P* < 0.001.

In addition to macrophages, dendritic cells (DC) play a major role in the induction of the T cell-mediated immune response to *Candida* spp. infections [86,87] and may determine the infection outcome [88,89]. DC are able to modulate adaptive responses, depending on the *Candida* spp. morphotype encountered [73,74,90]. DC can initiate and shape the antimicrobial immunity and, since candidiasis appears frequently in immunocompromised patients, these cells may hold the key to new antifungal strategies [91]. Accordingly, the numbers of splenic conventional DC, defined as CD11chigh cells, and surface maturation markers were evaluated upon *C. glabrata* systemic infection (Figure 4). A slight increase in splenic DC as compared to noninfected controls was observed in the infected mice, indicating that *C. glabrata* promoted the mobilization of these cells to the spleen or promoted their local proliferation. DC surface expression of the costimulatory molecule CD86, as evaluated by the mean fluorescence intensities (MFIs) due to antibody staining (Figure 4A,B), was elevated in infected mice, showing that *C. glabrata* induced the maturation of these cells. However, the stimulatory effect was not different among the treated and nontreated groups.

In contrast, the expression of MHC class II molecules on the surface of splenic DC of mice infected with this yeas<sup>t</sup> was found to be below control levels, an effect that reached statistical difference in mice treated with caspofungin. As CD86 expression in infected mice was found to be elevated, it is unlikely that this could represent a suppressive mechanism and could just be subsequent to a previous stimulatory effect. A kinetic study would be necessary to elucidate this point.

**Figure 4.** CD1 mice were challenged intravenously with 1 × 10<sup>8</sup> biofilm cells and then treated with PBS, caspofungin (Csf), or micafungin (Mcf). The obtained results are displayed as (**A**) total number of dendritic cells or mean fluorescence intensities (MFI) due to antibody staining against (**B**) CD80, (**C**) CD86, and (**D**) MHC class II on the surface of dendritic cells. The numbers of animals used were: controls (naïve; PBS + Csf; PBS + Mcf), *n* = 2; Cg + Csf, *n* = 8; Cg + Mcf, *n* = 8. Statistical differences among infected groups were evaluated using Kruskal–Wallis (overall ANOVA *P* < 0.05), post hoc Sidak's, and Dunn's multiple comparisons tests (\* *P* > 0.05). Cg—*Candida glabrata* ATCC2001. *\* P* < 0.05.

The expression of CD80, CD86, and MHC class II molecules on the surface of inflammatory monocytes was observed to be similar or slightly lower in the infected groups as compared to noninfected controls (Figure 5A–C). Likewise, and as observed on DC, no differences were observed among infected mouse groups, indicating that the treatment did not affect the expression of these activation markers on these innate immune cell populations. Finally, liver and kidney histopathologies were analyzed in infected mice, as these organs are preferred targets in i.v. *Candida* spp. infections [31,92]. As could be expected, no yeasts were found in the non-challenged control groups, and their organs presented no significant histological alterations.

Challenged mice showed inflammatory infiltrates in the liver. They were also shown, albeit less markedly, in the kidneys (nontreated and treated groups). The presence of polymorphonuclear cells was observed, but in general the infiltration remained mostly mononuclear. Yeasts were found in the liver, but not in the kidneys of treated and nontreated challenged groups. This fact corroborated the low CFU counts found in the kidneys.

**Figure 5.** CD1 mice were challenged intravenously with 1 × 10<sup>8</sup> biofilm cells and then treated with PBS, caspofungin (Csf), or micafungin (Mcf). The obtained results are displayed as mean fluorescence intensities (MFI) due to antibody staining against (**A**) CD80, (**B**) CD86, and (**C**) MHC II on inflammatory monocytes. The numbers of animals per group were: controls (naïve; PBS + Csf; PBS + Mcf), *n* = 2; Cg + Csf, *n* = 8; Cg + Mcf, *n* = 8. Statistical differences among infected groups were evaluated using Kruskal–Wallis (Overall ANOVA *P* < 0.05) and post hoc Sidak's multiple comparison tests. Cg—*Candida glabrata* ATCC2001.

Together, these observations confirmed *C. glabrata* as a low inflammatory species and indicated that two-dose treatment with caspofungin and micafungin does not have a significant impact on liver and kidney fungal burden or recruited inflammatory infiltrate when mice are i.v. infected with *C. glabrata* biofilm-grown cells. These results confirm the biofilm in vitro outcome our group previously reported [93,94].

Finally, liver and kidney histopathologies were analyzed in infected mice, as these organs are preferred targets in i.v. *Candida* spp. infections [45,86]. As could be expected, no yeasts were found in the non-challenged control groups, and their organs presented no significant histological alterations (Figure 6). Challenged mice showed inflammatory infiltrates in the liver, and less markedly in the kidneys (nontreated and treated groups, data not shown). The presence of polymorphonuclear cells was observed, but in general the infiltration remained mostly mononuclear. Yeasts were found in the liver (Figure 6), but not in the kidneys (data not shown) of treated and nontreated challenged groups. This fact corroborated the low CFU counts found in the kidneys.

**Figure 6.** Analysis of liver histology in CD1 mice. (**A**) Representative hematoxylin-eosin and (**B**) periodic acid–Schiff (PAS)-stained examples of liver tissue from the indicated mouse groups. Black arrows denote inflammatory infiltrates that were mostly of the mononuclear type. Insets correspond to higher magnification micrographs. White arrows indicate PAS-stained *Candida glabrata* ATCC2001 cells. Scale bars are shown and apply to similar sized micrographs (100 μm) or insets (20 μm), as indicated.
