Evaluation of Candida peritonitis with underlying peritoneal fibrosis and efficacy of micafungin in murine models of intra-abdominal candidiasis

Candida peritonitis is a crucial disease, however the optimal antifungal therapy regimen has not been clearly defined. Peritoneal fibrosis (PF) can be caused by abdominal surgery, intra-abdominal infection, and malignant diseases, and is also widely recognized as a crucial complication of long-term peritoneal dialysis. However, the influence of PF on Candida peritonitis prognosis remains unknown. Here, we evaluated the severity of Candida peritonitis within the context of PF and the efficacy of micafungin using mice. A PF mouse model was generated by intraperitoneally administering chlorhexidine gluconate. Candida peritonitis, induced by intraperitoneal inoculation of Candida albicans, was treated with a 7-day consecutive subcutaneous administration of micafungin. Candida infection caused a higher mortality rate in the PF mice compared with the control mice on day 7. Proliferative Candida invasion into the peritoneum and intra-abdominal organs was confirmed pathologically only in the PF mice. However, all mice in both groups treated with micafungin survived until day 20. Micafungin treatment tends to suppress inflammatory cytokines in the plasma 12 h after infection in both groups. Our results suggest that PF enhances early mortality in Candida peritonitis. Prompt initiation and sufficient doses of micafungin had good efficacy for Candida peritonitis, irrespective of the underlying PF.

(2019) 9:9331 | https://doi.org/10.1038/s41598-019-45776-x www.nature.com/scientificreports www.nature.com/scientificreports/ For fungal peritonitis, the International Society for Peritoneal Dialysis (ISPD) guideline updated in 2016, advocates for immediate catheter removal and antifungal therapy; however, the appropriate antifungal agent, for the initial therapy, has not been clearly defined 12 . Intra-abdominal candidiasis, such as peritonitis or abscesses, is considered much more common than recognized 13 . The causes of intra-abdominal candidiasis are not only from PD but also from intra-abdominal surgery, anastomotic leakage, and pancreatitis 14 . The Infectious Disease Society of America (IDSA) guideline on the management of candidiasis, updated in 2016, recommends echinocandin as an initial therapy for intra-abdominal candidiasis irrespective of the presence or absence of PF 14 . Micafungin is one of the antifungal agents belonging to the echinocandins class; and has been shown to have good antifungal activity even against non-albicans Candida species 15,16 . Non-albicans Candida species have recently increased as causative organisms in candidiasis patients [17][18][19] . However, C. albicans remains the predominant species causing Candida peritonitis. Furthermore, echinocandins are easy to use, with minimal adverse effects [20][21][22] . For these reasons, the clinical use of micafungin has recently increased.
PF and Candida peritonitis, two complications in PD patients, have been evaluated thus far; and severe or prolonged infective peritonitis is known to induce PF 23,24 . Yet, the influence of PF on peritoneal infection has not been adequately investigated. Therefore, evaluation of the efficacy of antifungal therapy for Candida peritonitis developed in the context of PF is essential among PD patients who suffer from Candida peritonitis. Furthermore, it could be useful for treating patients with Candida peritonitis in the context of abdominal surgery, intra-abdominal infection, and malignant diseases who may develop PF [25][26][27] . In this study, we evaluated whether PF has any influence on the prognosis of Candida peritonitis and on the efficacy of micafungin treatment, using PF mouse models.

Results
Evaluation of PF. PF was experimentally induced in mice by repeated intraperitoneal administration of chlorhexidine gluconate (CG) (Fig. 1A). In the control group, the mice peritoneal tissue consisted of a peritoneal mesothelial monolayer, with sparse connective tissues below the layer. Compared to the control group, the peritoneal tissues of the mice in the PF group showed significant thickening of the submesothelial compact zone (p < 0.0001, unpaired t test), and the presence of numerous inflammatory cells and fibroblasts as previously reported 4 (Fig. 1B).

Evaluation of survival.
We evaluated survival curves of the control and PF mice with or without micafungin treatment ( Fig. 2A). When untreated with micafungin, the PF mice showed a higher mortality rate than the control mice on day 7 (p = 0.013, Kaplan-Meier log-rank test) but there was no statistically significant difference in survival rate on day 20 between these two groups (p = 0.29, Kaplan-Meier log-rank test). In both PF and control groups, all mice treated with micafungin survived until the end of the experiment (day 20).
Evaluation of the fungal burden in the liver, spleen, kidneys, and blood. The fungal burden in the liver, spleen, kidneys, and blood were evaluated at 6 h, 12 h, and 7 days after the infection (Fig. 2B). The fungal burden of the PF mice without micafungin treatment on day 7 was not assessed because majority of this group died as shown in Fig. 2A. At early timepoints, micafungin treatment significantly reduced fungal burden in all the organs examined in both control and PF mice at 12 h, but only in the control mice at 6 h. Blood stream infection (positive blood culture) was detected in some PF mice at 6 and 12 h but successfully treated with micafungin. Micafungin treatment also reduced fungal burden of spleen and kidneys in control mice on day 7 compared with no-treatment mice. In both control and PF groups with micafungin treatment, fungal burden was significantly decreased in all the organs on day 7 compared with the data at 12 h (except for spleen of PF group).

Histopathological evaluation of Candida cells in the peritoneum and organs. Grocott staining
of the peritoneum showed abscess formation in a markedly thickened submesothelial zone with the presence of many Candida cells in most PF mice without micafungin treatment at 12 h after the infection (Fig. 3A,B). hematoxylin and eosin (H & E) staining confirms numerous inflammatory cells in the same sections (Fig. 3C). However, Candida proliferation was hardly observed in PF group with micafungin treatment (Fig. 3D). Candida cells were also hardly confirmed in both control groups with or without micafungin treatment probably due to the early timepoint of evaluation.
Grocott staining of the liver (Fig. 3E,F) and spleen (Fig. 3H,I) showed proliferative Candida infiltration into the subcapsular structure inside the organ in most PF mice without micafungin treatment at 12 h after the infection. H & E staining confirms numerous inflammatory cells in both organs (Fig. 3G,J). Adhesion of Candida cells to the surface of the organ was observed in some mice in both control groups (with or without micafungin treatment) and PF group (with micafungin treatment), although infiltration inside the organ was not detected in any mice in these groups. No Candida cell was detected in the kidneys in all the groups (photographs not shown).
Evaluation of the cytokines in the blood. We quantified the plasma cytokine concentrations to evaluate an inflammatory response in the Candida peritonitis mouse models. At 12 h after the infection, interferon-gamma, interleukin-10, interleukin-17A, interleukin-1 beta, interleukin-6, and tumor necrosis factor alpha (TNF-α) tended to be suppressed with micafungin treatment in both PF and control mouse models (Fig. 4A). We also used lipopolysaccharide (LPS) instead of C. albicans to evaluate if micafungin exerts an anti-inflammatory effect irrespective of the Candida infection. The plasma concentration of TNF-α was increased by intraperitoneal LPS injection but not suppressed with micafungin administration in both PF and control groups (Fig. 4B).

Discussion
Long-term PD causes histopathological changes, such as PF, in the peritoneum 24 , associated with mesothelial loss, severe thickening of the submesothelial compact zone, and vascular alterations 3 . Severe or prolonged infective peritonitis also leads to PF, eventually leading to membrane 23,24 and organ failures 28 . However, the influence of PF in the prognosis of peritonitis remains to be evaluated. In the present study, we evaluated the severity of C. albicans peritonitis in the context of PF using mouse models induced by CG exposure. Although, some differences exist between CG-induced PF and human PF with PD therapy, most of the pathological changes between them are similar. These include, expression of collagen and alpha-smooth muscle actin, macrophage infiltration, and neovascularization in the peritoneum 29 . Therefore, CG mouse models have been used to investigate the pathogenesis and therapy of PF 4,30-32 . In our study, the presence of PF enhanced early mortality in mice with Candida peritonitis during the first 7 days. In a histopathological evaluation, we hereby showed the proliferation of Candida cells, invading into the liver and spleen tissues, only in the PF group without micafungin treatment; whereas, with micafungin treatment, it was inhibited. Interestingly, no infiltration was found in all the control mice regardless of micafungin treatment. That no Candida cell invasion was detected histopathologically in the kidneys might have resulted from the location of these organs in the retroperitoneal space, which is separated from the intraperitoneal space. In addition, the absolute number of Candida cells in kidneys were approximately 100 times less than those in other two organs. The proliferation of Candida cells in the peritoneum was also detected only in the PF mice without micafungin treatment, histopathologically. Considering all these findings, the damage to the peritoneal membrane and a remarkable intraperitoneal Candida proliferation followed by candidemia, which was confirmed In the control group, the monolayer of mesothelial cells covered the entire surface of the peritoneum. In the PF group, the submesothelial compact zone was markedly thickened. Bar graph shows the thickness of the submesothelial compact zones in the control and PF groups. The PF group showed significant thickening compared to the control group. An asterisk indicates p < 0.0001 (unpaired t test). Both control and PF groups in this assay were not inoculated with Candida. These results were confirmed on two separate occasions and representative data are shown. PF, peritoneal fibrosis; CG, chlorhexidine gluconate; SS, sterile saline; i.p., intraperitoneal injection; and s.c., subcutaneous injection. ; control with micafungin treatment (filled circles); and PF with micafungin treatment (filled triangles). Without micafungin treatment, the PF mice had significantly higher mortality than the control mice on day 7 (p = 0.013, Kaplan-Meier log-rank test), although both groups showed no significant difference on day 20 (p = 0.29). All the PF and control mice treated with micafungin survived until the end of the experiment (day 20) (vs. untreated group, p < 0.01 each). Similar results were obtained in three different experiments and representative data are shown. (B) Fungal burden in the liver, spleen, bilateral kidneys, and blood was evaluated 6 h, 12 h, and 7 days after the infection (n = 10 for each group). In the control mice, fungal burden in the three target organs at 6 and 12 h, and that in spleen and kidneys on day 7 was significantly reduced by micafungin treatment. In the PF mice, there was significant difference in the fungal burden between micafungin-treated and -untreated groups at 12 h, but not at 6 h. Fungal burdens of all the three organs in the control mice and two organs (liver and kidneys) in the PF mice were significantly reduced by micafungin treatment on day 7 compared with the results at 12 h. Asterisks indicate p < 0.0083 (Mann-Whitney U test adjusted with Bonferroni correction). PF, peritoneal fibrosis. www.nature.com/scientificreports www.nature.com/scientificreports/ by the significant increase of fungal burden in the bloodstream at 12 h, are thought to have caused the early high mortality phase in the PF group.
In our study, we confirmed all mice in both PF and control groups were successfully treated with early initiation of micafungin 2 h after C. albicans inoculation based on the results of survival, fungal burden, and histopathological findings. Drug penetration into the site of infection to achieve microbe-eliminating concentrations is a key requirement for effective antimicrobial treatment. It should be noted that micafungin concentration at the site of infection may be above MICs for micafungin susceptible Candida species but below mutant prevention concentrations 33 . This caution is important because intra-abdominal candidiasis is a hidden reservoir for emergence of echinocandin resistance particularly in cases of Candida glabrata infection 34 . A limitation of the current study is that we were unable to evaluate effects of PF on drug penetration into the peritonea and peritoneal cavity. www.nature.com/scientificreports www.nature.com/scientificreports/ Concerning the good efficacy of micafungin demonstrated in this study, the regulation of cytokines might also be playing an important role. During acute inflammation induced by infection in the peritoneal cavity, proinflammatory cytokines are activated during the early period, and the neutrophils recruited are subsequently replaced by monocytes 35 . While proinflammatory and anti-inflammatory cytokines are critical to the elimination of the infection, excessive production can cause tissue and organ damages 36 . It was reported that micafungin suppresses LPS-induced TNF-α production and may have immunomodulative effects 37 . In the present study, we could not confirm the suppression of plasma TNF-α concentration by micafungin administration in LPS-exposed mouse models, in both PF and control groups. Therefore, the tendency for the suppression of excessive inflammatory cytokines in Candida peritonitis mice by micafungin is considered to be due mainly to the secondary effect of its direct fungicidal activity against Candida, resulting in an improved survival rate. We considered that anti-inflammatory cytokine IL-10 was induced secondarily to inflammatory cytokine production and reduced at 12 h as a result of the suppression of inflammatory cytokines by micafungin treatment.
In conclusion, the present study demonstrated that Candida peritonitis itself showed a high mortality even in subjects without PF, and PF enhanced early mortality in Candida peritonitis. Micafungin showed a good efficacy The plasma cytokine concentrations of IFN-γ, IL-10, IL-17A, IL-1β, IL-6, and TNF-α were quantified by enzyme-linked immunosorbent assay in the following groups: control group (control) and PF group (PF) just before infection (0 h); and infected with or without micafungin administration, in both PF and control groups at 12 h after the infection (four mice each). All the six cytokines at 12 h after the intraperitoneal inoculation of C. albicans tended to be suppressed with micafungin treatment in both PF and control groups. (B) The plasma TNF-α concentrations of LPS-exposed mice instead of C. albicans inoculation were quantified by enzyme-linked immunosorbent assay, and no significant difference was observed between micafungin-treated and -untreated mice in both control and PF groups. Similar results were obtained in two different experiments and representative data are shown. Asterisks indicate p < 0.05 (Mann-Whitney U test). IFN-γ, interferon-gamma; IL-10, interleukin-10; IL-17A, interleukin-17A; IL-1β, interleukin-1 beta; IL-6, interleukin-6; TNF-α, tumor necrosis factor alpha; PF, peritoneal fibrosis; and LPS, lipopolysaccharide. (2019) 9:9331 | https://doi.org/10.1038/s41598-019-45776-x www.nature.com/scientificreports www.nature.com/scientificreports/ for Candida peritonitis even in the context of PF. This result may also apply to patients who develop PF due to intra-abdominal inflammation caused by other reasons besides PD, such as abdominal surgery, intra-abdominal infection, and malignant diseases [25][26][27] . Considering the high mortality result in the PF groups during the early period, early initiation of micafungin is considered important. The Candida peritonitis mouse model with underlying PF developed in the present study could also be useful in future studies to evaluate the pathogenicity of other Candida strains and the efficacies of other antifungal agents in similar conditions.

Ethics. Our animal experiment protocol was pertinently reviewed and approved by the Institutional Animal
Care and Use Committee of Nagasaki University (approval number 1407281164). All animal experiments were performed at the Nagasaki University Laboratory Animal Center for Biomedical Research in accordance with the Guide for the Care and Use of Laboratory Animals (National Research Council, National Academy Press, Washington DC, 2011) and the institutional regulations and guidelines for animal experimentation.
Murine model of PF. The animals used in this study were specific-pathogen-free male ICR mice (10 weeks of age, weighing approximately 40 g; CLEA Japan, Inc., Japan). They were housed in the Biomedical Research Center, Life Science Support Center, Nagasaki University.
We induced PF by intraperitoneal administration of CG as described previously 4,[30][31][32] . Mice received injections of 0.1% CG in 15% ethanol, or 15% ethanol alone, at a dose of 0.2 mL/mouse into the peritoneal cavity on alternate days for 3 weeks, 9 times overall.
Evaluation of the peritoneum thickness. For the evaluation of PF, the PF and control mice were sacrificed 4 days after the last CG injection (n = 6 for each group). In order to prevent bias, peritoneal membrane was dissected in four locations (upper right, lower right, upper left, and lower left), and evaluated with Masson's trichrome staining. Thus, four sections were analyzed for each mouse. The thickness of the submesothelial compact zone for each section was measured at 10 positions and compared between the PF and control groups, using the unpaired t test.
Murine Candida peritonitis and antifungal treatment. Candida peritonitis was induced by intraperitoneal inoculation of the C. albicans wild-type strain SC5314. C. albicans cells grown in yeast extract-peptone-dextrose (YPD) broth overnight were washed and resuspended in sterile saline and adjusted to 5.0 × 10 7 CFU/mL. Mice were infected intraperitoneally with 1000 µL of the Candida suspension 4 days after the last CG injection. Non-infected groups were intraperitoneally injected with 1000 µL of sterile saline instead of C. albicans cell suspension.
All the mice in the treatment groups were treated with 5 mg/kg of micafungin (500 µg/mL) (Astellas Pharma Inc., Tokyo, Japan) injection subcutaneously in the neck, once a day for seven days, beginning 2 h after the intraperitoneal injection of C. albicans cell suspension; based on the same administration schedule against candidiasis reported previously 38 .
The MIC of micafungin was ≤0.015 mg/L, determined by dry plate antifungal susceptibility testing of yeasts; Eiken (Eiken Chemical Co., Ltd., Tokyo, Japan). We used the subcutaneous route of administration because repeated injection into the tail vein of mice is difficult and it is reported that there is only a marginal difference in the efficacy of this drug between subcutaneous and intravenous administrations 38 . The dosage of 5 mg/kg corresponds to 125 mg/day in humans 39 consistent with serum concentration-time curve data from 0 to 24 h (AUC 0-24 ) between humans and mice 40,41 . The mice in the non-treatment groups were administered with sterile saline alone during the same period.
Evaluation of survival, fungal organ and bloodstream burdens and histopathological examination in mice. The survival of the mice was recorded daily after the intraperitoneal injection of C. albicans until day 20 (n = 8 for each group). Statistical analysis was performed using the Kaplan-Meier method.
To assess the viable cell count of Candida in the organs; the liver, spleen, and bilateral kidneys were excised and placed in sterile 0.9% saline at 4 °C immediately after sacrifice, at 6 h, 12 h, and 7 days after the intraperitoneal injection of C. albicans (n = 10 for each group). The homogenate was then serially diluted 1:10, and aliquots were plated on YPD agar. Blood samples were also collected by cardiac puncture and plated on YPD agar without dilution. Viable fungal colony counts were determined after 24-48 h incubation at 30 °C. The lower limit of detection was 10 CFU/organ. The results were expressed as log 10 CFU/organ for the three organs and CFU/mL for blood, and analyzed with the Mann-Whitney U test followed by Bonferroni correction.
For the histopathological evaluation, the peritoneum, liver, spleen, and kidney, obtained 12 h after the intraperitoneal injection of C. albicans were stained with H & E and Grocott.
Evaluation of the cytokines in the murine blood. The plasma cytokine concentrations of interferon-gamma, interleukin-10, interleukin-17A, interleukin-1 beta, interleukin-6, and TNF-α were quantified using a Bio-Plex Pro TM Mouse Cytokine Th17 Panel A 6-Plex (#m6000007ny, Bio-Rad Laboratories, Hercules, CA) for the following groups. These were, the control and PF groups just before infection; and infected with or without micafungin administration, in both PF and control groups at 12 h after the infection (four mice each).
The plasma TNF-α concentrations in mice exposed to LPS were also quantified as follows: 20 mg/kg of LPS were administered intraperitoneally, after 2 h, 5 mg/kg of micafungin (500 µg/mL) was injected subcutaneously to the treatment group. Sterile saline was subcutaneously injected to the non-treatment groups. Plasma TNF-α concentrations, obtained 6 and 12 h after LPS administration were measured using Mouse TNF alpha Uncoated ELISA (88-7324, Thermo Fisher Scientific Inc., Waltham, MA). Each of the groups consisted of four mice.
www.nature.com/scientificreports www.nature.com/scientificreports/ Statistical analysis. The unpaired t test was used to assess differences of the PF thickness. The differences of fungal burden in the target organs and the cytokines in the plasma were evaluated by using the Mann-Whitney U test. Multiple comparisons were adjusted with the Bonferroni method. For all statistical analyses, a value of p < 0.05 was considered significant. All statistical analyses were performed using Prism 5.0 (GraphPad Software, Inc., La Jolla, CA).