Activation of the inflammasome drives peritoneal deterioration in a mouse model of peritoneal fibrosis

During peritoneal dialysis (PD), the peritoneum is exposed to a bioincompatible dialysate, deteriorating the tissue and limiting the long‐term effectiveness of PD. Peritoneal fibrosis is triggered by chronic inflammation induced by a variety of stimuli, including peritonitis. Exposure to PD fluid alters peritoneal macrophages phenotype. Inflammasome activation triggers chronic inflammation. First, it was determined whether inflammasome activation causes peritoneal deterioration. In the in vivo experiments, the increased expression of the inflammasome components, caspase‐1 activity, and concomitant overproduction of IL‐1β and IL‐18 were observed in a mouse model of peritoneal fibrosis. ASC‐positive and F4/80‐positive cells colocalized in the subperitoneal mesothelial cell layer. These macrophages expressed high CD44 levels indicating that the CD44‐positive macrophages contribute to developing peritoneal deterioration. Furthermore, intravital imaging of the peritoneal microvasculature demonstrated that the circulating CD44‐positive leukocytes may contribute to peritoneal fibrosis. Bone marrow transplantation in ASC‐deficient mice suppressed inflammasome activation, thereby attenuating peritoneal fibrosis in a high glucose‐based PD solution‐injected mouse model. Our results suggest inflammasome activation in CD44‐positive macrophages may be involved in developing peritoneal fibrosis. The inflammasome‐derived pro‐inflammatory cytokines might therefore serve as new biomarkers for developing encapsulating peritoneal sclerosis.


| INTRODUCTION
Peritoneal dialysis (PD) is a viable option of renal replacement therapy for treating patients with end-stage renal disease. 1 PD is a home-based treatment suitable for ameliorating the patient's quality of life. 2 However, prolonged exposure to the dialysis solutions often deteriorates the peritoneum. 3 Peritoneal deterioration is a comprehensive concept based on the decreased peritoneal function and morphological changes in the peritoneum. 4 Prolonged exposure to PD fluid triggers chronic inflammation and may be one of the factors contributing to peritoneal deterioration. 5 Therefore, it is important to elucidate the main locus of chronic inflammation pathogenesis for preventing peritoneal deterioration.
Inflammasomes are intracellular multimeric complex molecules recognizing either the pathogen-associated molecular patterns (PAMPs) or danger-associated molecular patterns (DAMPs). 6 The NLRP3 inflammasome can trigger chronic low-grade inflammation 7 and the pathogenesis of multiple complex diseases such as chronic kidney disease, 8,9 atherosclerosis, 10 Type 2 diabetes, 11 and Alzheimer's disease. 12 The activation of the NLRP3 inflammasome and its downstream pathway particularly enhances the pro-caspase-1 activation and caspase-1-mediated interleukine-1β (IL-1β) maturation in the macrophages, which contribute to the progression of inflammation. 13 Alterations in the tissue-resident macrophages have been reported to render the long-term PD patients sensitive to developing peritonitis and consequently fibrosis/sclerosis. 14 Moreover, the NLRP3 activation and IL-1β release are critical for solute transport defects and tissue remodeling in PD-related peritonitis. 15 However, the involvement of inflammasome activation in chronic inflammation after the onset of peritonitis, as well as its role in peritoneal deterioration has not been clearly elucidated.
Therefore, the present study aimed to elucidate the mechanism by which peritoneal deterioration is caused by inflammasome activation using in vivo imaging techniques, and genetically engineered animals. Inflammasome activation in macrophages thus plays a pivotal role in peritoneal deterioration.

| Animals
The study protocol was approved in advance by the Ethics Review Committee for Animal Experiments of the Kawasaki Medical School (approval no. . At the beginning of the study, 8-week-old male C57BL/6J mice weighing 20-30 g were designated wild-type (WT) mice. The ASC homozygous knockout (ASC −/− ) mice were kindly provided by M. Takahashi (Jichi Medical University, Shimotsuke, Japan). 16,17

| Establishment of mouse peritoneal fibrosis models and drug administration
Two mouse peritoneal fibrosis models were used. In the first model, peritoneal fibrosis was induced in mice by intraperitoneal injection of 0.3 mL of 0.1% chlorhexidine gluconate (CG) in 15% ethanol and 85% PBS thrice a week for 4 weeks, as previously reported. 18 The control mice received an intraperitoneal injection of PBS. In the second model, a vascular access port (Access Technologies, Skokie, IL, USA) was implanted following which the catheter end was advanced into the peritoneal cavity. 19 Seven days after the procedure, the mice with the implanted peritoneal access port were infused with 0.2 mL saline with 1 IU/mL heparin to facilitate wound healing. Thereafter, over the next 5 weeks, the five mice were administered with 1.5 mL saline daily (saline) and another five mice were administered with 1.5 mL standard PD fluid (PDF) buffered with lactate and containing 4.25% glucose (Dianeal; Baxter Limited, Tokyo, Japan). The mice were deeply anesthetized using sevoflurane inhalation and sacrificed, and the peritoneal tissues were collected for histopathological examination. We used a high glucose-based PD solution to induce peritoneal fibrosis during bone marrow transplantation. In addition, we used liposome-encapsulated clodronate (LC) to induce macrophage apoptosis. LC or PBS-liposomes were purchased from Katayama Chemical Industries Co., Ltd. (Osaka, Japan). WT mice were intravenously injected with LC (5 mg/kg) or PBS-liposomes (control) once weekly for 3 weeks. Moreover, we orally administered VX-765 (100 mg/kg/day), a caspase-1 specific inhibitor, for 3 weeks after 1 week of CG injection, with the vehicle group used as the control.
1. Experimental 1. Analysis of whether inflammasome activation is involved in peritoneal deterioration in WT and ASC −/− mice with or without CG injection (Figures 1-3).
2. Experimental 2. Examination of whether CD44positive cells accumulate in peritoneal blood vessels in WT mice with or without CG injection using intravital imaging ( Figure 4). 3. Experimental 3. Analysis of whether inflammasome activation of hematopoietic cells contributes to peritoneal deterioration in WT and ASC −/− mice with or without high glucose-based PD solution injection ( Figure 5).

| Bone marrow transplantation
The bone marrow transplantation (BMT) was performed according to a previously described standard protocol. 13 Here, the adult (8-week-old) WT and ASC −/− mice received 9 Gy of total body X-ray irradiation at two doses separated by 3 h to minimize the gastrointestinal toxicity.

| Intravital serial multiphoton microscopy (MPM)
The animals were anesthetized using sevoflurane and placed on the stage of an inverted microscope, as described previously. [20][21][22] The rectal temperature was monitored and maintained using a homeothermic blanket system at 37°C. The peritoneum was exposed by incising the skin and placed on the stage of a microscope. The images were acquired using a Nikon A1R-MP multiphoton confocal microscope (Tokyo, Japan) equipped with an inverted imaging system and an Apo LWD25X 1.10 W DIC N2 objective lens. The mice were injected with Alexa Fluor 594-conjugated bovine serum albumin (BSA; Invitrogen) intravenously for labeling plasma flow. 20,23 The Alexa Fluor 488-labeled anti-CD44 antibody (BioLegend) was used for detecting the CD44-positive leukocytes and to directly and quantitatively visualize their peritoneal intravascular homing and rolling in mice using MPM. The Alexa Fluor 594-conjugated BSA was detected through 525/50 nm bandpass filters and FITC-conjugated BSA through 595/50 nm bandpass filters employing 800 nm laser excitation using multiphoton imaging. The potential toxicity of the laser excitation and fluorescence to the cells was minimized using low laser power and high scan speeds to minimize the total laser exposure as minimal as possible.

| Histology, immunohistochemistry, immunofluorescence, and immunocytochemistry
The peritoneal membrane sections (4 μm thick) were obtained from the paraffin-embedded tissue blocks and stained using hematoxylin-eosin and Masson's trichrome. The thickness of the peritoneal membrane in the parietal peritoneum was evaluated by measuring the distance from the surface of the mesothelium to the upper limit of the muscular tissue using the measurement module in the BZ-H1M software (Keyence Co., Osaka, Japan). The peritoneal thickness was measured at 10 random points, and the mean thickness of each tissue sample was compared. The measurements were performed as previously reported. 19,24 The immunohistochemistry was performed using serial sections of the paraffin-embedded specimens rehydrated in the phosphate-buffered saline and subjected to antigen retrieval in a microwave or treated with proteinase K ( antibodies. The blue-stained scarred areas of Masson's trichrome and areas positively stained for Collagen IV, α-SMA, and F4/80 were quantified using a color image analyzer (Keyence). Immunocytochemistry of the peritoneal macrophages was performed using the FITC anti-mouse CD44 antibody (BioLegend). The antibodies were incubated for 6 h at room temperature. DAPI (4,6-Diamidino-2-phenylindole, Sigma-Aldrich) was used as the nuclear counterstain.

| Western immunoblotting
The peritoneal tissue samples were obtained as described previously. 19 Immunoblotting analysis was performed

| Real-time reverse transcription-quantitative PCR
The total mRNA extraction and real-time reverse transcription-quantitative PCR (RT-qPCR) were performed as described previously. 19 The total RNA was extracted from the anterior abdominal wall using the TRIzol reagent (Life Technologies, Grand Island, NY, USA), followed by DNase digestion (Sigma-Aldrich, Japan). The primer and probe sequences are listed below. The first-strand complementary DNA was synthesized from the total RNA (1 μg) using the Moloney murine leukemia virus reverse transcriptase (Life Technologies, Grand Island, NY, USA) with an oligo(dT) [12][13][14][15][16][17][18] primer. The primers and probes for TaqMan analysis were designed using the GenBank sequence data and Primer 3 online software (http://frodo.wi.mit.edu/ prime r3/). The real-time quantitative PCR was performed using the ABI Prism 7700 sequence detection system (Applied Biosystems, Foster City, CA, USA). The expression levels in each sample were relative to the 18S rRNA levels.

| Flow cytometric analysis
The peritoneal macrophages were isolated from the peritoneal cavities of WT Vehi and WT CG mice. The peritoneal cavity was lavaged twice with 10 mL of ice-cold, RPMI 1640 medium supplemented with 10% FBS. The recovered medium was centrifuged at 250× g for 10 min. The cell pellet was resuspended in 6 mL RPMI 1640 medium. The cells were plated in 6-well plates (1 × 10 6 cells/ well), incubated at 37°C overnight to enable cell adherence, and subsequently washed with fresh medium to remove any unattached cells. The cells were dissolved in the FACS buffer for FACS analysis using FACS CantoII™ (BD Bioscience, Tokyo, Japan) and FlowJo software (Tree Star, Ashland, OR, USA). Allophycocyanin (APC)-labeled anti-mouse F4/80 antibody (B202565; BioLegend, San Diego, CA, USA) and phycoerythrin (PE)-labeled anti-mouse/human CD11b antibody (B194776; BioLegend) were used. 25

| Statistical analysis
Data are expressed as mean ± standard error of the mean (SEM). Statistical analyses were performed using the GraphPad Prism6 software (GraphPad Software). Comparisons between two groups were performed by using the unpaired or paired two-tailed Student's t-test or Mann-Whitney U test. Comparisons between multiple groups with a normal distribution were performed by using one-way ANOVA followed by the Tukey's multiple comparison test. Comparisons between multiple groups with non-normal distributions were performed by using the Kruskal-Wallis tests followed by Dunn's multiple comparisons test. Statistical significance was set at p < .05.

| Inflammasome activation exacerbates peritoneal fibrosis
First, the role of inflammasome activation was evaluated in the mouse model of peritoneal fibrosis. The maximum thickness of the peritoneal membrane was drastically increased in the WT-CG mice compared to that in the WT-Vehi mice ( Figure 1A,B). Peritoneal fibrosis was significantly higher in the WT-CG mice than that in the WT-Vehi mice ( Figure 1A,C,D). After CG injection, the mRNA expression of the profibrotic genes, TGF-β and Col1a1 in the WT mice was enhanced compared to that in the vehicle-injected mice ( Figure S1A,B). The role of myofibroblasts was also investigated in the peritoneal membrane as the cells play a central role in tissue fibrosis. There was a higher number of α-SMA positive cells in the WT-CG mice than in the WT-Vehi mice ( Figures 1A,E and S1C,D). There was a decrease in the protein levels of E-cadherin, a transmembrane glycoprotein in the epithelial cell adherens junction in the WT-CG mice compared to the WT-Vehi mice ( Figure S1C,D). These changes were significantly ameliorated in the ASC −/− -CG mice. Next, the possible involvement of inflammasome activation was assessed. The levels of mRNAs encoding the inflammasome proteins, such as Pycard and Nlrp3, were markedly higher in the WT-CG mice than in the WT-Vehi mice (Figure 2A,B). The protein expression of ASC and active caspase-1 (p20) was significantly higher in the WT-CG mice than in the WT-Vehi mice ( Figure 2C,D). The production of the mature IL-18, as a marker of inflammasome activation, was higher in the WT-CG than in the WT-Vehi mice ( Figure 2C,D). These changes in the WT-CG were significantly ameliorated in the ASC −/− -CG mice.

| Inflammasomes are activated in the infiltrating macrophages
The macrophage infiltration was increased in the submesothelial layers of the peritoneum. A large number of F4/80-positive macrophages were observed in the WT-CG mice, which was significantly higher than those in the WT-Vehi mice ( Figure 3A,B). The levels of Adgre1, Itgax, and Chemokine (C-C motif) ligand 2 (CCL2), regulating the migration and infiltration of the macrophages, were found to increase in the WT-CG compared to the WT-Vehi mice ( Figure 3C-E). All of these changes were attenuated in the ASC −/− -CG mice. To determine whether the inflammasomes were activated in the infiltrating macrophages, the ASC and F4/80 cells were double-stained. The ASC-positive cells expressing F4/80 were detected in the submesothelial layers of the peritoneum ( Figure 3F). The merged image indicated that the ASC-positive macrophages were present in the submesothelial layers of the peritoneum. Moreover, some of the F4/80 positive cells were expressed CD44 in the submesothelial layers of the peritoneum ( Figure 3G). Next, we assessed the phenotype of peritoneal macrophages. The peritoneal macrophages were collected from the WT-Vehi and WT-CG mice and analyzed using fluorescence-activated cell sorting (Figure S2A,B). The number of CD11b high F4/80 low cells were higher in the WT-CG mice than in the WT-Vehi mice (Figure 3H,I). On the contrary, the CD11b high F4/80 high cells were decreased in the WT-CG mice compared to the WT-Vehi mice ( Figure 3H,J).

| Visualization of the peritoneal microvasculature, and the effect of circulating CD44-positive leukocytes on peritoneal deterioration using MPM imaging
Peritoneal macrophages from WT-CG mice expressed high levels of CD44, indicating that the CD44-positive cells may contribute to peritoneal fibrosis ( Figure S2C-E). Therefore, we used MPM imaging techniques to evaluate whether CD44-positive cells accumulated in the microvasculature of the peritoneum. The peritoneal microvessels in living mice were successfully visualized using intravital MPM ( Figure 4A and Video S1). The CD44-positive leukocytes were increased in the WT-CG compared to the WT-Vehi mice and were home to the peritoneal microvessels of the WT-CG mice ( Figure 4B-D). In addition, the rolling of the CD44-positive leukocytes was observed in the endothelial cells (Video S2A,B). These data suggested that CD44-positive leukocytes may be associated with peritoneal pathology.

| Activation of the inflammasomes in hematopoietic cells is required for peritoneal fibrosis
To assess the role of the inflammasomes of hematopoietic cells in peritoneal fibrosis under high glucose peritoneal dialysis fluid injection, the WT chimeric mice for ASC expression in hematopoietic cells were created using BMT. The peritoneal thickness and fibrotic lesions were found to be drastically attenuated in the BMT (ASC −/− → WT) mice compared to the BMT (WT → WT) mice ( Figure 5A-C). The expression of Collagen IV and α-SMA was attenuated in the BMT (ASC −/− → WT) mice compared to that in the BMT (WT → WT) mice ( Figure 5D,E). The mRNA levels of Nlrp3 and IL-18 were markedly lower in the BMT (ASC −/− → WT) mice than in the BMT (WT → WT) mice ( Figure S3A,B). Moreover, the protein expression of the mature IL-1β and IL-18 to significantly decreases in BMT (ASC −/− → WT) mice compared to that in the BMT (WT → WT) mice ( Figure 5F,G).

| Liposome-encapsulated clodronate induced macrophage depletion attenuates peritoneal fibrosis
We also assessed the effects of liposome-encapsulated clodronate (LC), which induces macrophage apoptosis, on CG-induced peritoneal fibrosis. There was a significant decrease in maximal peritoneal thickness and fibrotic lesions in WT-CG/ LC mice compared to those in WT-CG mice ( Figure 6A-C). The expression level of Adgre1 mRNAs was lower in WT-CG/LC mice than in WT-CG mice ( Figure 6D). The mRNA expression levels of TGFβ and Col1a1 in WT-CG/LC mice were lower than those in WT-CG mice ( Figure 6E,F). The number of CD44-positive leukocytes were decreased in the WT-CG/LC mice compared with those in the WT-CG mice ( Figure 7A-C). The protein expression of ASC, mature IL-18, and IL-β was lower in the WT-CG/LC mice than in the WT-CG mice ( Figure 7D,E).

| Therapeutic interventions for peritoneal fibrosis using caspase-1 inhibitor
Finally, the therapeutic interventions for inhibiting peritoneal deterioration were assessed by treating with the caspase-1 inhibitor VX-765 (referred to WT-CG/VX-765).
There was a drastic decrease in the maximal peritoneal thickness and fibrotic lesions in the WT-CG/VX-765 mice compared to WT-CG mice ( Figure 8A-C). The expression levels of TGFβ and Col1a1 mRNAs in the WT-CG/ VX-765 mice were lower than those in the WT-CG mice ( Figure S4A,B). The α-SMA expression was also lower in the WT-CG/VX-765 mice than in WT-CG mice (Figure S4C,D). The production of mature IL-1β and IL-18 was lower in the WT-CG/VX-765 mice than in the WT-CG mice ( Figure 8D,E).

| DISCUSSION
To our knowledge, this is the first study assessing inflammasome activation by infiltrated CD44-positive macrophages in progressing the peritoneal deterioration. The potential molecular mechanism of peritoneal deterioration via inflammasome activation was proven in the mouse model of peritoneal fibrosis. Interestingly, intravital imaging of the peritoneum demonstrated the circulating CD44-positive leukocytes to possibly contribute to peritoneal fibrosis. Bone marrow transplantation in the ASC −/− mice indicated inflammasome activation in the macrophage-mediated PD solution-induced peritoneal fibrosis. Furthermore, the caspase-1 inhibitor preserved the peritoneum and prevented the development of fibrosis. Our data indicated for the first time that sustained activation of inflammasomes in macrophages is an important factor in the deterioration of the peritoneum during PD treatment (Figure 9). Peritoneal deterioration is inevitable in long-term PD. Therefore, it is essential for treating patients with PD while minimizing the peritoneal burden as much as possible. Peritonitis is the most important complication in patients undergoing PD. Peritonitis is one of the most important risk factors for developing encapsulated peritoneal sclerosis (EPS). In Japan, EPS decreases the spread of pHneutral PD solutions. 26 Neutral-pH peritoneal dialysates have been suggested to reduce the damage of the peritoneal membrane. 27 However, peritonitis has been reported to be a strong factor in patients with EPS treated with pHneutral PD solutions. 28 Why does peritonitis cause EPS despite using biocompatible PD fluid? NLRP3 activation and IL-1β release have been previously reported to have critical roles in solute transport defects and tissue remodeling during PD-related acute peritonitis. 15 We thought that chronic inflammation might persist after the onset of peritonitis and that inflammasome-medicated chronic inflammation might be involved in EPS. Therefore, evaluating the cytokine expression through long-term observation after remission of peritonitis is a subject for future investigation.
Understanding the cellular and molecular mechanisms underlying the fibrosis of the peritoneal membrane has both basic and translational relevance, as it might be useful for developing therapies aimed at counteracting the deterioration and restoring the homeostasis of the peritoneal membrane. There are three reports of inflammasome activation in peritoneal fibrosis. Two studies have reported the importance of inflammasome activation in the peritoneal mesothelial cells. 29,30 One study reported the importance of inflammasome activation in the peritoneal endothelial cells. 31 We strongly advocate for the importance of inflammasome activation in the infiltrating macrophages. The first two reports were mainly based on the experiments with the cultured cells and did not provide the in vivo data on peritoneal degradation. Hishida et al. reported the role of NLRP3 inflammasome in the murine model of the PD-related peritoneal fibrosis induced by methylglyoxal (MGO). They reported the myeloid cell-specific ASC deficiency to fail in inhibiting the MGO-induced peritoneal fibrosis and there are possibly two reasons for the discrepancy between our results and theirs. First, the bone marrow transplantation experiment demonstrated the importance of inflammasome activation in the macrophages. Second, different models were used to induce peritoneal fibrosis. NLRP3 inflammasome-induced PD-related acute peritonitis has been suggested to occur predominantly in the resident peritoneal monocytes and macrophages. 15 Moreover, the macrophages have been reported to play a pivotal role in the progression of peritoneal deterioration. 14,32,33 Macrophages are likely to remain in peritoneal tissue after the onset of PD-related acute peritonitis. Notably, it has been reported that significance changes of the prevalence of macrophage/monocyte populations varies widely in PD patients, depending on the history of peritonitis. 34 A recent clinical study demonstrated that PD patients with a history of peritonitis had higher levels of dialysate MCP-1 compared to the patients without a history of peritonitis. 35 Accordingly, the macrophage accumulation in peritoneal tissue may reflect chronic peritoneal inflammation after PD-related acute peritonitis. This establishes that the inflammasome activation in the macrophages is critical for peritoneal degradation.
How do macrophages infiltrate the peritoneum? To test this question, a state-of-the-art technique was applied in the unique intravital imaging technology where a technique was established in evaluating the kidneys of the living mice 20,22,23 and also the peritoneal microcirculation. First, the peritoneal microcirculation was successfully visualized. Interestingly, massive homing of the CD44positive leukocytes was observed in the peritoneal microvessels in the mouse model of peritoneal fibrosis. CD44 is a family of cell surface glycoproteins that function as cell adhesion molecules in leukocyte extravasation, lymphocyte homing, and binding to the extracellular matrix. 36 Furthermore, its ligand hyaluronan (HA) has emerged as an important adhesion molecule for cell trafficking in multiple organs and contributes to the pathogenesis of a variety of inflammatory diseases. 37 HA in the peritoneal dialysate might be useful as a marker to assess functional and morphological changes in patients undergoing longterm PD. 38 CD44 and subsequent HA catabolism have been reported to trigger inflammasome activation. 39 The present study confirmed some F4/80 positive cells in the submesothelial layer to express CD44. Peritoneal macrophages from the WT-CG mice were found to express high levels of CD44 indicating that the CD44-positive macrophages contribute to developing peritoneal fibrosis. Additionally, HA in peritoneal tissues may trigger inflammasome activation in macrophages. HA accumulation and several CD44-positive macrophages may regulate the transition from acute to chronic inflammation. This is the first report of the CD44-positive leukocytes in the circulating blood contributing to peritoneal fibrosis using MPM imaging.
In summary, the findings of this study strongly suggested the sustained activation of inflammasomes in the macrophages to be an important factor in peritoneal deterioration during PD treatment. Thus, inflammasome activation in the macrophages might serve as a new therapeutic target for chronic inflammation-induced peritoneal deterioration. Furthermore, the inflammasome-derived pro-inflammatory cytokines might serve as new biomarkers for peritoneal deterioration.