Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Directly recruited GATA6 + peritoneal cavity macrophages contribute to the repair of intestinal serosal injury

Directly recruited GATA6 + peritoneal cavity macrophages contribute to the repair of intestinal... ARTICLE https://doi.org/10.1038/s41467-021-27614-9 OPEN Directly recruited GATA6 + peritoneal cavity macrophages contribute to the repair of intestinal serosal injury 1,3 1,3 1,2 2 1 Masaki Honda , Masashi Kadohisa , Daiki Yoshii , Yoshihiro Komohara & Taizo Hibi Recruitment of bone marrow derived monocytes via bloodstream and their subsequent conversion to CX3CR1 macrophages in response to intestinal injury is dependent on CCR2, Nr4a1, and the microbiome. This process is critical for proper tissue repair; however, GATA6 peritoneal cavity macrophages might represent an alternative, more readily avail- able source of mature and functional myeloid cells at the damaged intestinal locations. Here hi + we show, using spinning-disk confocal microscopy, that large F4/80 GATA6 peritoneal cavity macrophages promptly accumulate at damaged intestinal sites upon intestinal thermal injury and upon dextran sodium sulfate induced colitis in mice via a direct route from the peritoneal cavity. In contrast to bloodstream derived monocytes/macrophages, cavity mac- rophages do not depend on CCR2, Nr4a1 or the microbiome for recruitment, but rather on the ATP-release and exposed hyaluronan at the site of injury. They participate in the removal of necrotic cells, revascularization and collagen deposition and thus resolution of tissue damage. In summary, peritoneal cavity macrophages represent a rapid alternative route of intestinal tissue repair to traditional monocyte-derived macrophages. 1 2 Department of Transplantation and Pediatric Surgery, Kumamoto University, 1-1-1 Honjo, Chuo-ku, Kumamoto 860-8556, Japan. Department of Cell Pathology, Kumamoto University, 1-1-1 Honjo, Chuo-ku, Kumamoto 860-8556, Japan. These authors contributed equally: Masaki Honda, Masashi Kadohisa. email: honda.masaki@kuh.kumamoto-u.ac.jp NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 1 1234567890():,; ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 issue-resident macrophages are major players in maintain- monocytes/macrophages depends on CCR2, Nr4a1, and the ing local homeostasis in response to environmental changes microbiome . This process contributes to the resolution of Tsuch as infection and inflammation .Unlike mostother inflammation and tissue repair but requires at least a few days for organs, the intestinal mucosal lamina propria is constantly conversion from a classical pro-inflammatory monocyte to an replenished with macrophages from bone marrow progenitor alternative phenotype. However, because the intestine is a portal to 2–5 hi monocytes .Ininflammation, rapid accumulation of Ly6C the external environment, we speculated that repair is initiated monocytes into the injured area and conversion to CX3CR1 immediately via an alternative route during severe intestinal injury. 2 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE hi Fig. 1 Large F4/80 macrophages rapidly accumulate in response to intestinal thermal injury. a Representative stitch images of colonic LP (left) in GFP/+ + Cx3cr1 mice. Representative still (middle) and three-dimensional (3D) image (right) of CX3CR1 macrophages (green) in colonic LP. Scale bars, 50 μm. + + GFP/+ RFP/+ b Representative images of colonic LP CX3CR1 monocytes/macrophages and CCR2 cells (red) 6 h after focal intestinal injury in Cx3cr1 Ccr2 + + mouse. Scale bars, 50 μm. c Migration paths, d crawling velocities of CCR2 cells and CX3CR1 cells in response to intestinal injury. n= 30/group (obtained hi by three independent images). e Quantification of large F4/80 macrophages within intestinal injury at indicated time points. n= 3 (2 h) and 5 (6, 24, 48 h). hi f Representative image of large F4/80 macrophages in injury area at 24 h post-injury. Higher magnification of the indicated area is shown in right. Scale bars, 100 μm(left)and 50 μm (right). g H&E and h immunofluorescence staining for F4/80 (brown) and GATA6 (green) of injured colon 24 h post-injury (cross section). Arrowheads indicate GATA6 peritoneal macrophages. Scale bars, 10 μm. Data were representative of three independent experiments. Data represent box and whiskers (Fig. 1d) and mean ± SEM (Fig. 1e). Source data are provided as a Source Data file. Large GATA6 peritoneal cavity macrophages rapidly invade were adjacent to the injury site remained sessile and did not move the subcapsular region of the liver following sterile injury/ from their original position towards the damaged site (Fig. 1b–d inflammation where they contribute to tissue repair . The zinc and Supplementary Movie 1). Despite their sessile nature, when finger transcription factor, GATA6, controls proliferation, survi- topical F4/80 antibody was applied to the injury site, a very sig- 8–10 hi val, and metabolism . Wang et al. revealed that peritoneal nificant population of large F4/80 macrophages accumulated macrophages expand rapidly at liver injury sites and up-regulate within 2 h post-injury in C57BL/6 mice (Fig. 1e and Supple- hi CD273, CD206, and arginase-1, markers of alternatively activated mentary Fig. 1a). The accumulation of these large F4/80 cells 7,11,12 macrophages . The intestine is also surrounded by a peri- peaked at 24 h after injury and persisted for at least 48 h (Fig. 1e toneal cavity; however, whether peritoneal macrophages can (1) and Supplementary Fig. 1a). To confirm that they were not GFP/+ RFP/+ detect luminal injury within the peritoneum, (2) invade into the derived from monocytes, we imaged Cx3cr1 Ccr2 mice intestine across a formidable mesothelial barrier under patholo- 6 h after injury with topical administration of F4/80 antibody. RFP GFP gical conditions, and (3) affect tissue repair, have not been This revealed that neither CCR2 nor CX3CR1 co-localized hi explored. The molecular mechanisms involved in this process are with large F4/80 cells (Supplementary Fig. 1b). The neutrophil also unknown. The mammalian intestinal tract is populated with marker, Ly6G, also did not show any co-localization with F4/80 over 100 trillion microbes, the majority of which reside in the (Supplementary Fig. 1b). At 24 h after injury, CCR2 monocytes 13,14 colon so they could be a potent recruitment cue for peritoneal formed a ring surrounding the injury site and their accumulation hi macrophages. was via blood vessels, whereas the large F4/80 cells could not be Here, we examine the recruitment and function of peritoneal seen in blood vessels and were positioned within the center of the cavity macrophages in intestinal injury. Using spinning-disc dual- injury as a large aggregate (Fig. 1f). There was a striking differ- + + hi laser intravital microscopy and a sterile burn injury model of the ence in size between CCR2 CX3CR1 cells and large F4/80 hi intestine, we detect large F4/80 peritoneal macrophages accu- cells, the latter being at least twice the size of the former (Sup- hi mulating at the intestinal injury site. They are recruited to the plementary Fig. 1c, d). Importantly, these large F4/80 cells in the injured intestine via a unique peritoneal route in response to ATP intestinal injury site expressed GATA6, a transcription factor released by necrotic cells. Interaction between hyaluronan and specific for large peritoneal cavity macrophages, but not intestinal CD44 is indispensable for the recruitment, while CCR2, Nr4a1, F4/80 macrophages (Fig. 1g, h and Supplementary Fig. 2a–d). and the microbiome are not involved. Accumulated large peri- Using flow cytometry, we confirmed the intravital microscopy toneal macrophages contribute to the removal of necrotic cells data showing that there was a population of GATA6 hi hi within the intestinal injury site and help with revascularization CD11b F4/80 macrophages in the injured colon (Supplemen- and collagen deposition. We also focus on the most common tary Fig. 3a–c). The other populations of macrophages did not form of intestinal disease, inflammatory bowel disease (IBD), and express GATA6 (Supplementary Fig. 3c). observe more severe dextran sodium sulfate (DSS)-induced colitis activity in the absence of peritoneal macrophages. Overall, these Peritoneal macrophages from the peritoneal cavity accumulate data demonstrate the importance of the immune player at the site of intestinal injury independent of CCR2 or Nr4a1. “GATA6 peritoneal cavity macrophages” in intestinal injury. Luminex assays revealed induction of MCP-1 (CCL2), the key These findings prompt further studies to explore the mechanisms + chemokine that attracts CCR2 monocytes, and neutrophil che- and/or therapeutic strategies for disorders of the intestinal tract mokine, KC, in the colon at 24 h post-injury (Fig. 2a). However, with a particular focus on IBD. the lack of the CCL2 receptor, CCR2, did not affect the recruit- ment of GATA6 macrophages (Fig. 2b, c) indicating that (1) Results CCL2 was not responsible for peritoneal macrophage recruitment hi Large F4/80 macrophages accumulate promptly in response and (2) CCR2-positive monocytes were not important for peri- hi to sterile intestinal injury. Intestinal macrophages express a toneal macrophage recruitment. The recruitment of large F4/80 specific cell surface marker, CX3CR1. Using intravital imaging cells within damaged intestine at 24 and 48 h after injury was not −/− lo and mice with a fluorescent reporter for CX3CR1, we imaged different in Nr4a1 mice, which lack Ly6C monocytes at the steady state intestinal lamina propria macrophages and found injury site, compared with wild-type mice (Fig. 2b, c). The that these cells form an interdigitated physical chain surrounding CX3CR1 ligand is involved in the recruitment of macrophages in the blood vessels (Fig. 1a). We examined the recruitment of some tissues; however, CX3CR1-deficient mice accumulated hi monocytes and macrophages in response to sterile thermal injury equivalent numbers of large F4/80 macrophages to wild-type GFP/+ RFP/+ of the intestine using Cx3cr1 Ccr2 mice. A 500 μm focal mice after intestinal injury (Supplementary Fig. 4a, b). necrotic lesion extending into the lamina propria was created To confirm that the peritoneum was the source of the large F4/ hi + from the serosa side of the colon using a thermal probe. This 80 GATA6 macrophages, we depleted peritoneal macrophages model allowed us to image recruitment of immune cells in an area by intraperitoneal administration of clodronate liposome (CLL), + 7 eradicated of resident cells. Imaging showed that CCR2 as described previously . Intraperitoneal CLL treatment did not + + monocytes but not CX3CR1 monocytes infiltrated into the affect the distribution of intestinal CX3CR1 macrophages + + injury site within 6 h. Meanwhile, CX3CR1 macrophages that (Fig. 2d, e) or the accumulation of CCR2 monocytes within NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 the intestinal injury site (Supplementary Fig. 5a, b) but incapable of entering the injury site when transferred intrave- hi significantly suppressed the number of large F4/80 macrophages nously (Fig. 2h–j) further indicating that this was not the route of in the intestinal injury site (Fig. 2f, g). Additionally, peritoneal infiltration. When peritoneal cells from LysM-eGFP mice with cells transferred from LysM-eGFP mice (in which > 85% of the depleted GATA6 macrophages by intraperitoneal CLL admin- + + 7 GFP cells in the peritoneal cavity are GATA6 macrophages ) istration were transferred into C57BL/6 mice intraperitoneally, into the peritoneum of C57BL/6 mice, resulted in accumulation GFP cells were not found within the intestinal injury site of these cells at the intestinal injury site. GFP macrophages were (Fig. 2k, i). Next, peritoneal cells from LysM-eGFP mice were 4 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE Fig. 2 Peritoneal Macrophages accumulate into the intestinal injury site directly via the peritoneal route regardless of CCR2 or Nr4a1. a Luminex assays of chemokines in colon tissue samples at steady state and 24 h after thermal injury. n = 5 (control group) and 8 (injury group). b Representative hi RFP/RFP −/− images of large F4/80 cells accumulated in injury site at 48 h after focal intestinal injury in WT, Ccr2 , and Nr4a1 mouse. Scale bars, 50 μm. hi + c The number of large F4/80 cells at indicated time points are quantified (n = 5/group). d Representative images and e quantification of CX3CR1 cells per field of view (FOV) in lamina propria of the colon in control and 2, 7, 14 days after the intraperitoneal administration of CLL. Data were collected by GFP/+ hi imaging of CX3CR1 mice. n = 4 (CLL 2, 7, 14 d) and 6 (control). Scale bars, 50 μm. f Representative images of large F4/80 cells accumulated in hi intestinal injury site at 48 h post-injury. Mice were treated by PBS-L or CLL 4 days before the injury. Scale bars, 50 μm. g The number of large F4/80 cells are quantified. n = 3(PBS-L group) and 5(CLL group). h Schematic protocol for peritoneal cell (PC) transfer experiments (from LysM-eGFP mice to C57BL/ 6 mice). i Representative images and j the number of LysM-eGFP cells accumulated in C57BL/6 mice (n = 3/group). Scale bars, 100 μm. k Representative images and l quantification of LysM cells within the injured colon (6 h) of C57BL/6 mice that were transferred PCs intraperitoneally from LysM-eGFP mice with i.p. administration of PBS-L or CLL. Scale bars, 50 μm. n = 3/group. Data represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, NS not significant. P values were calculated with a two-tailed unpaired Student t-test (a, g, j, l) and one-way ANOVA followed by Tukey’s post hoc test (c, e). Source data are provided as a Source Data file. hi stained using a PE-ICAM2 (CD102) antibody. In total, 78.6% of the peritoneal cavity nor did it affect the amount of large F4/80 + + GFP peritoneal cells were also positive for CD102, which GATA6 macrophages that accumulated at the intestinal injury demonstrated that the double-positive cells are large peritoneal site (Fig. 4c–f), indicating that the phenotype and recruitment of macrophages (Supplementary Fig. 6a–c). We next transferred large peritoneal macrophages in intestinal injury is independent peritoneal cells from LysM-eGFP mice to the peritoneal cavity of of the microbiome. CD44, the predominant adhesion molecule, C57BL/6 mice after staining CD102. Imaging of the intestine at was also not significantly affected by the diminution of the gut + + 6 h after injury showed that LysM CD102 cells had accumu- microbiota (Fig. 4c, d). lated in the injured area but not in the uninjured area (Supplementary Fig. 6d–f). Peritoneal macrophages promote intestinal repair. We pre- RFP/RFP viously reported that healing in Ccr2 mice at 48 h after Recruitment of peritoneal cavity macrophages to the intestinal injury and later was delayed. Time-lapse imaging of the colon at hi injury site is dependent on ATP and hyaluronan-CD44 inter- 24 h after thermal injury showed that large F4/80 macrophages action. ATP is an important damage-associated molecular pat- were already at the site disassembling the nearby SYTOX tern (DAMP) in sterile injury that recruits both neutrophils and necrotic cells (Fig. 5a). Next, we imaged the SYTOX green- macrophages to the injury site . Pretreatment with apyrase, an positive cells within the intestinal injury site after depletion of ATPase, or an ATP receptor antagonist inhibited the accumula- peritoneal macrophages by intraperitoneal CLL administration. hi tion of large F4/80 macrophages at the intestinal injury site At 48 h post-injury, the clearance of necrotic cells was delayed in (Fig. 3a, b). To further explore the mechanism of peritoneal CLL-treated mice compared with PBS-liposome (PBS-L)-treated macrophage dynamics in intestinal injury, we blocked CD44, a mice (Fig. 5b, c). Furthermore, peritoneal macrophage depletion leukocyte adhesion molecule highly expressed on GATA6 was associated with significantly impaired revascularization and peritoneal macrophages, and imaged intestinal injury sites. Pre- collagen deposition within the intestinal injury site (Fig. 5d–g). treatment with anti-CD44 antibody prevented the recruitment of The lack of clearance of SYTOX-positive cells was even more hi RFP/RFP large F4/80 macrophages to the intestinal injury site (Fig. 3c, d). pronounced in peritoneal macrophage-depleted Ccr2 mice Importantly, immunofluorescent staining showed exposed hya- (Fig. 5b, c), indicating that both CCR2 monocytes, which + + luronan, which is the ligand for CD44, within the intestinal injury became CX3CR1 monocytes/macrophages, and large GATA6 site but not on the serosal surface (Fig. 3e). Administration of peritoneal macrophages contribute to necrotic cell clearance, hyaluronidase, an enzyme that breaks down hyaluronan, also perhaps in separate layers of the intestine. Indeed, imaging of the prevented the recruitment of large peritoneal macrophages injured colon showed that CCR2 monocytes accumulated hi (Fig. 3f). Intriguingly, injury from the mucosal side induced some mainly in the lamina propria, whereas large F4/80 macrophages hi accumulation of large F4/80 macrophages from the serosal accumulated mainly in the muscularis (Supplementary Fig. 7a, b). surface, indicating that these accumulation mechanisms are also The pathological analysis confirmed that the majority of F4/ hi + partially functional in severe mucosal injury (Fig. 3g, h). 80 GATA6 macrophages were infiltrating the muscularis in the injured colon (Supplementary Fig. 5c, d). It is also worth noting that the muscularis has a lower vascular density than the lamina The gut microbiota is not critical for the recruitment of propria (Supplementary Fig. 6a, b), further emphasizing the GATA6 peritoneal macrophages in response to intestinal importance of blood flow-independent peritoneal macrophage injury. Recruitment of CCR2 monocytes to the intestine and accumulation when the injury reaches the muscularis. conversion to CX3CR1 monocytes/macrophages is microbiome- dependent in both the emergency repair as well as in steady state 4,6,15 turn over . Because the burn injury extended from the serosa Peritoneal macrophages accumulate in the colon in response to all the way through to the lamina propria, we predicted that gut DSS-induced colitis. To evaluate whether peritoneal macro- microbiota would be critical for the recruitment of GATA6 phages can respond to inflammation that starts in the lamina peritoneal cavity macrophages. We used broad-spectrum anti- propria and develops towards the serosa, mice were orally biotics from before birth until adulthood (Fig. 4a), which deci- administered 4% DSS-containing water for 5 days. Intravital mated the intestinal microbiota as assessed by SYTOX green imaging of the colon in LysM-eGFP mice with F4/80 antibody + hi staining (Fig. 4b). Moreover, sequencing analyses of feces showed topically applied to the serosa showed that LysM F4/80 large the remarkable differences of microbial communities between macrophages were infiltrating the muscularis (Fig. 6a, b). When control and Abx-treated mice (Fig. 4c and Supplementary peritoneal cells from LysM-eGFP mice were intraperitoneally Fig. 7a–d). However, these changes had no impact on the number transferred to C57BL/6 mice, a similar accumulation of peritoneal hi hi + + hi of CD11b F4/80 GATA6 peritoneal macrophages found in LysM F4/80 macrophages was found in the DSS-induced colitis NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 GFP/+ RFP/+ model (Fig. 6c–e). We also subjected Cx3cr1 Ccr2 mice in the colon of DSS-induced colitis compared with the control to DSS-induced colitis and found that no CCR2 and/or (Fig. 6g, h). As described in the thermal injury model, Abx + hi hi CX3CR1 signal co-localized with the large F4/80 cells (Fig. 6f) treatment did not influence on the recruitment of F4/80 large indicating that these are distinct cell lineages. These findings peritoneal macrophages to the colon in DSS-induced colitis indicate that these large invading macrophages are derived from (Fig. 6i, j). the peritoneum, not the vasculature. For that matter, immuno- To further analyze the role of peritoneal macrophages in DSS- fluorescence staining showed increased expression of hyaluronan induced colitis, we depleted peritoneal macrophages by 6 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE Fig. 3 Recruitment of peritoneal macrophages to the site of intestinal injury depends on ATP and hyaluronan-CD44 interaction. a Representative hi images and b quantification of the number of F4/80 macrophages within intestinal injury site 24 h post-injury in mice that were pretreated with apyrase or p2rx7 antagonist. Scale bars, 50 μm. n = 4 (Apyrase), 5 (control), and 7 (P2X7 antagonist). c Representative images and d quantification of the number hi of F4/80 macrophages within intestinal injury site 24 h post-injury in mice that were pretreated with isotype control or anti-CD44 antibody. Scale bars, 50 μm. n = 4 (isotype) and 5 (anti-CD44). e Immunofluorescence staining of the colon harvested from control or 2 h after injury with hyaluronic acid hi binding protein (HABP). The dashed line indicates the injury border. f Quantification of the number of F4/80 macrophages within intestinal injury site 6 h post-injury in mice that were treated with hyaluronidase. Scale bars, 50 μm. n= 4 (hyaluronidase) and 5 (control). g Representative images and hi h quantification of the number of F4/80 macrophages in the colonic muscularis with no injury or 24 h post-injury from mucosa or serosa. Scale bars, 50 μm. n = 5 (no injury, injury from serosa) and 6 (injury from mucosa). Data represent mean ± SEM. *p < 0.05, **p < 0.01, ****p < 0.0001. P values were calculated with two-tailed unpaired Student t-test (d, f) and one-way ANOVA followed by Tukey’s post hoc test (b, h). Source data are provided as a Source Data file. intraperitoneal administration of CLL and compared body weight (Fig. 8). However, somewhat surprising was the observation that change, disease activity index, colon length, and pathological these cells could also accumulate in the intestinal tract when the findings with those of PBS-L-administered DSS-treated mice. We injury was induced via DSS colitis, an injury that starts primarily found that CLL-treated mice had increased levels of injury and in the lamina propria. inflammation, including more severe weight loss and increased Pleural, pericardial, and peritoneal cavity macrophages have disease activity score (Fig. 7a, b and Supplementary Table 1). In been studied extensively, primarily in vitro . The homeostatic the CLL-treated mice, colon length, which is a marker of and transcriptomic signatures of these tissue-resident macro- inflammatory injury, was often shorter on the 5th day of DSS- phages are maintained by Wt1 mesothelial and fibroblastic induced colitis, but this difference was not significant compared stromal cells via the generation of retinoic acid . Cavity mac- with that of the PBS-L-treated mice (Fig. 7c). Moreover, colon rophages express CD11b, F4/80, and CSF1R and can be divided + hi - lo tissue sections obtained 5 days after the start of DSS adminis- into large GATA6 F4/80 and small GATA6 F4/80 8–10,18–20 tration were carefully evaluated. CLL-treated mice showed more macrophages . Small macrophages play a pivotal role in 18,21 severe histomorphological scores, including marked inflamma- inflammatory responses and infections . Kim et al. showed tory cell infiltration and disrupted intestinal architecture, that they are continuously replenished by blood monocytes compared with PBS-L mice (Fig. 7d, e). The degrees of infiltration dependent on the transcription factor, IRF4, and signals from the by neutrophils, macrophages, and lymphocytes were compared microbiome . By contrast, the large macrophages are derived between PBS-L and CLL groups, but no significant difference was from the yolk sac and have multiple functions regarding recog- observed in the lamina propria or muscularis, respectively nition and phagocytosis of pathogens, antigen presentation, and (Supplementary Fig. 11a, b). Although the number of neutrophils resolution of inflammation . They are also known to disappear 21,23,24 in the muscularis of the DSS-induced colitis group treated with in the early stage of infection or inflammation although CLL was approximately twice that of the group treated with PBS- this may be caused by their aggregatory properties that lead to L, it did not reach statistical significance. Indeed, chemokine clump formation rather than actual disappearance . These large expression in the colon tissue was not affected by CLL treatment GATA6 peritoneal macrophages express phagocytic receptors (Supplementary Fig. 11c). In the final series of experiments, such as Tim4, MerTK, CD36, and genes for cell adhesion and 19,26,27 peritoneal cell transfer experiments were performed for DSS- angiogenesis, suggesting a repair phenotype . Using intra- induced colitis to evaluate its effect on inflammation and vital imaging, we visualized peritoneal GATA6 macrophages symptoms (Fig. 7f). In comparison with the PBS-treated control recruited to the injured intestine. These cells exerted an in vivo group, the peritoneal cell transfer group had a lower disease reparative function to a focal injury that started in the serosa and activity index, while the other parameters did not show to DSS which causes inflammation primarily starting in the statistically significant differences (Fig. 7g–k). lamina propria and moving towards the serosa. The interaction between gut microbiota and the immune system in tissue repair is an important issue. Gut microbiota can shape Discussion intestinal mucosal T cells, including Foxp3 regulatory T cells and The intestinal tract is a barrier to trillions of bacteria; therefore, 28–30 mucosal-associated invariant T cells . Additionally, Furusawa tissue repair is vital. Resident macrophages play a pivotal role in et al. provided the mechanistic insight that microbe-derived buty- intestinal repair helping remove debris and also helping to induce rate regulates the differentiation of regulatory T cells .Using angiogenesis. However, tissue-resident CX3CR1 macrophages intravital imaging, we recently revealed that prenatal antibiotic do not move during intestinal injury and depend on classical + + administration and germ-free conditions altered the localization of CCR2 monocytes to be recruited and to convert to CX3CR1 intestinal mucosal CX3CR1 macrophages and impeded the monocytes/macrophages. This conversion process takes more development of a perivascular anatomical barrier .Bycontrast, the than 2 days and is dependent on the transcription factor, Nr4a1 , hi microbiome did not regulate the function of GATA6 peritoneal indicating that early healing must be dependent on the Ly6C cavity macrophages because their numbers were not altered in mice monocytes. However, the timing and inherent inflammatory nature of these cells make it difficult to explain how healing could given antibiotics from birth and they still accumulated the injured intestine. Despite significant alterations in many different immune begin so rapidly. In this study, we identified a third macrophage type that contributed to tissue repair, namely GATA6 peritoneal cells in the dysbiotic intestine, none of the many differences affected peritoneal macrophage accumulation to the injured intestine indi- cavity macrophages. They rapidly accumulated via the peritoneal route at the injured site in response to intestinal injury within the cating that these cells are not affected by alterations in the micro- serosa. Both the alarmin ATP and CD44-hyaluronan interactions biome or intestinal immunobiome. were key for recruitment from the peritoneal cavity to the This study also has clinical implications because humans have hi hi damaged intestinal serosa, which led to tissue repair independent an intraperitoneal mature CD14 CD16 macrophage subset that + + 32 of CCR2 monocyte-derived CX3CR1 monocytes/macrophages shows intracellular GATA6 expression . As a typical digestive NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 tract disease, IBDs, including ulcerative colitis and Crohn’s dis- GATA6 macrophage function is also worth considering as a ease, are a worldwide burden with high incidence in developed form of therapy. Indeed, adoptive transfer of peritoneal macro- countries and increasing incidence in developing countries . phages with or without IL33, which can induce the proliferation Currently, new therapeutic strategies targeting the recruitment and alternative activation of peritoneal macrophages, ameliorated 35,36 and activation of immune cells from the vasculature are being inflammation in a preclinical colitis model . Whether these tested . Our results raise the intriguing possibility that enhancing peritoneal cavity macrophages also contribute to colorectal cancer 8 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE Fig. 4 Gut microbiota does not affect the phenotype and dynamics of peritoneal macrophages in intestinal injury. a Schematic protocol for the generation of dysbiotic mice. b Representative stitch images (from 12 different fields of view) of SYTOX green particles in intestinal contents of SPF control and Abx-treated mice. Scale bars, 50 μm. c Relative abundance of bacterial amplicon sequence variants in feces obtained from control or Abx- treated mice. The bar plot is displayed at the phylum level. n = 4/group. d Flow cytometry analysis of GATA6 and CD44 expression in peritoneal hi hi + CD11b F4/80 macrophages obtained from control or Abx-treated mice. Cells were pregated on size, viability, and CD45 . Data were representative of hi hi three independent experiments. e Quantification of the proportion of peritoneal CD11b F4/80 macrophages (left). n = 6 (control) and 5 (Abx). An hi hi hi absolute number of peritoneal CD11b F4/80 macrophages (right). n = 4 (control) and 3 (Abx). f Representative images of the large F4/80 hi macrophages within intestinal injury in control and Abx mice. Scale bars, 50 μm. g Quantification of the number of large F4/80 macrophages in injury site at 6 and 24 h post-injury. n = 3 (Abx at 6 h) and 5 (control at 6 h, control and Abx at 24 h). Data represent mean ± SEM. NS not significant. P values were calculated with a two-tailed unpaired Student t-test (e, g). Source data are provided as a Source Data file. and other intestinal diseases is an intriguing but as yet untested maintain a body temperature of 37 °C throughout imaging. Exposed abdominal tissues were covered with saline-soaked gauze to prevent dehydration. hypothesis. Further studies are needed to address this question. In summary, this study demonstrates GATA6 peritoneal Spinning-disc confocal intravital microscopy (SD-IVM) and image analysis.A macrophages to be mobile immune players with a specific repair multichannel spinning-disk confocal microscope was used to image the mouse function in intestinal injury and inflammation. Our data indicate intestine. Image acquisition of the intestine was performed using an Olympus IX81 that enhancing their repair functions is a potential therapeutic inverted microscope (Olympus, Center Valley, PA), equipped with an Olympus approach for numerous intestinal diseases that include IBD. focus drive and a motorized stage (Applied Scientific Instrumentation, Eugene, OR) and fitted with a motorized objective turret equipped with 4×/0.16 UPLAN- SAPO, and 10×/0.40 UPLANSAPO objective lenses and coupled to a confocal light Methods path (WaveFx; Quorum Technologies) based on a modified CSU-10 head GFP/GFP Mice. C57BL/6 mice (#000664), LysM-eGFP mice, Cx3cr1 (CX3CR1-defi- (Yokogawa Electric Corporation, Tokyo, Japan). Cells of interest were visualized −/− cient) mice (#005582), and Nr4a1 mice (#006187) were obtained from The using fluorescently labeled antibodies, and fluorescent reporter mice. In some RFP/RFP Jackson Laboratory. Generation of Ccr2 (CCR2-deficient) mice have been experiments, necrotic cells were labeled by superfusion of the intestinal surface 37 GFP/+ RFP/+ previously described . Cx3cr1 Ccr2 mice were generated by crossing with 1 μM SYTOX Green. Laser excitation wavelengths of 491, 561, 642, and GFP/GFP RFP/RFP 37 Cx3cr1 Ccr2 mice with C57BL/6 mice . All mice were on a C57BL/6 730 nm (Cobolt, AB, Solna, Sweeden) were used in rapid succession together with background. Animals were maintained in a specific pathogen-free environment the appropriate band-pass filters (Semrock Inc., Rochester, NY). A back-thinned with ad libitum access to food and water. Mice were housed under standardized electron-multiplying charge-coupled device 512 × 512-pixel camera (Hamamatsu conditions of temperature (21–22 °C) and illumination (12/12 h light/dark cycle). Photonics) was used for fluorescence detection. Volocity software 6.1 (Perki- Mice of 8–12 weeks of age were used for experiments. Mice were gender-matched nElmer) was used to drive the confocal microscope and analysis of images. for experiments and experimental/control mice were bred separately. Mice were Acquired images were analyzed or exported as TIF images using Volocity euthanized by cervical dislocation after imaging or for tissue sampling. All software. The minimum threshold values were adjusted for each of the fluorescence experiments were approved by the Kumamoto University Ethics Review Com- channels to reduce the background. Exported images were imported to the Image J mittee for Animal Experimentation and were performed according to guidelines of 38 software package (NIH) for analysis . For quantification in the number of large the Institutional Animal Committee of Kumamoto University. hi F4/80 macrophages in the intestine, images were acquired for each mouse using a 10× objective and was evaluated using the “analyze and measure” command in hi Image J software. F4/80 cell whose length is larger than 15 μm was defined as Antibodies and reagents. Antibodies against CD11b (#17-0112-82; M1/70) (1:100 “large” macrophage. Quantification of SYTOX Green-positive dead cells in the dilution), CD31 (#12-0311-82; PECAM-1, 390) (1:100 dilution), CD45 (#45-0451- 7,38 intestinal injury area was performed using Image J as previously described . 82; 30-F11) (1:150 dilution), F4/80 (#12-4801-82; BM8) (1:100 dilution) were Autofluorescence induced by debris was excluded from the analysis. obtained from eBioscience. Antibodies against Ly6G (#127612; 1A8) (1:100 dilu- tion) were obtained from Biolegend. Antibody against CD44 (#553134; IM7) (1:150 dilution) was obtained from BD Biosciences. Antibody against GATA6 (#26452; Depletion of gut commensal bacteria. Gut commensal bacteria were depleted D61E4) (1:50 dilution) was obtained from Cell Signaling Technology. SYTOX using a method modified a protocol as previously described . Mice were provided Green (S7020) was obtained from Thermo Fisher Scientific. For intravital imaging, with ampicillin (1 g/L), vancomycin (0.5 g/L), neomycin (1 g/L), metronidazole we used a very low concentration at 2 μg per mouse of each antibody. Each anti- (1 g/L), and ciprofloxacin (0.2 g/L) in drinking water. All antibiotics were obtained body was diluted in PBS as appropriate. from Sigma-Aldrich. Antibiotics-mixed water was started from E14.5 and con- tinued until the experiments. In vivo treatment. Peritoneal macrophages were depleted by intraperitoneal admin- istration of 100 μL/mouse (0.69 mol/L) clodronate liposome (#CP-020-020; clo- Sterile inflammation induced by focal necrotic injury. Mice were anesthetized dronateliposomes.org, Vrije Universiteit, Netherlands) 4 days prior to the experiment. with isoflurane and a small midline laparotomy was made to exteriorize the colon. For The same dose of PBS-liposome (#CP-020-020) was used for the control experiment. imaging, a single focal injury was induced on the serosal surface of the colon to a depth For adhesion molecule blocking experiments, mice were administered either 50 μganti- of 80 μm using the tip of a heated 30-gauge needle mounted on an electrocautery CD44 (#CL8944AP; KM81, Cedarlane) monoclonal antibody or isotype control anti- device. Injury from the mucosal side was performed by puncturing the needle from the body 1 h prior to the injury. Apyrase treatment was performed by intraperitoneal opposite side, and the puncture hole was sutured closed. After the induction of thermal administration of 25 U apyrase (#A2230; Sigma-Aldrich). For blocking ATP receptor, injury, the abdominal incision was sutured closed and mice were allowed to recover for mice received 10 μg P2X7 antagonist (#A438079; R&D) intraperitoneally 1 h prior to imaging of later time points (2, 6, 24, 48 h). For sham experiments, mice underwent the theinjury. Hyaluronidasetreatment wasperformed by intraperitoneal administration of same surgical procedure, but no injury was induced. 100 U hyaluronidase (#H3506; Sigma-Aldrich) just after the intestinal injury. Each reagent was diluted in PBS as appropriate. Immunocytochemistry. Peritoneal macrophages were obtained from peritoneal exudates of mice, and erythrolysation was performed. Peritoneal macrophages were Preparation of the mouse intestine for intravital imaging. Mice were anesthe- cultured in low-glucose DMEM supplemented with 2% FBS and 1% penicillin/strep- tized by s.c. injection of 200 mg/kg ketamine (Bayer Animal Health) and 10 mg/kg tomycin. Cytospin smears were prepared by placing 1 mL (1.0 × 10 cells/ml) cultured xylazine (Bimeda-MTC). After anesthesia, the right jugular vein was cannulated to fluid in the cytospin funnel with filter paper being placed between the funnel and the administer fluorescent dyes and additional anesthetic. For intestinal imaging, a slide, followed by centrifugation at 800 rpm for 5 min resulting in the formation of a midline incision followed by a left lateral incision along the costal margin to the monolayered sheet of cells within a small circumference. After drying the sections, midaxillary line was performed to expose the intestine. To image the intestinal wall sheeted peritoneal macrophages were fixed with 1% paraformaldehyde for 5 min. from the serosal site, the oral side of the intestine was ligated with 5-0 silk string. Subsequently, the sections were incubated with blocking solution (1.0% bovine serum PBS (200 μL/20 mm of the colon) was introduced into the intestinal lumen using a albumin in TBS) for 20 min at room temperature (RT), then reacted with primary syringe with a 30-gauge needle, and the anal side of the intestine was ligated with antibodies. The sections probed with primary antibodies were incubated with secondary 5-0 silk string. Mice was placed in a left lateral position and the intestine was placed antibodies. Antibodies used in the immunofluorescence (IF) method are listed in on a glass coverslip and imaged if blood flow was normal. We defined “muscularis” Supplementary Table 2. IF sections were mounted using a mounting medium con- as 10–30 μm of depth, “submucosa” as 30–50 μm of depth, and “lamina propria” as taining 4′,6-diamidine-2′-phenylindole dihydrochloride (DAPI) (#SCR-38448; Dianova 50–80 μm of depth by imaging from serosa. Mice were placed on a heating pad to GmbH, Hamburg, Germany). NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 hi + Fig. 5 Peritoneal cavity macrophages facilitate intestinal repair. a Time-lapse images showing F4/80 large macrophage pulling off SYTOX green particles within the intestinal injury site. Elapsed time is shown. Scale bar, 20 μm. b Representative images of SYTOX green cells within intestinal injury +/+ RFP/RFP + site at 48 h post-injury in PBS-L- or CLL-treated CCR2 and CCR2 mice. Scale bars, 100 μm. c Quantification of SYTOX green area within injury in each group. n = 6/group. d Representative images of intestinal injury site 48 h after injury in PBS-L- or CLL-treated mice. Mice were administered an anti- CD31 (red) antibody intravenously to visualize vasculature. The white dashed line highlights the original injury border. Scale bars, 100 μm. e Quantification of revascularization (CD31 area within injury). n = 6/group. f Representative images of collagen (purple) within intestinal injury site 72 h post-injury in PBS-L- or CLL-treated mice. Scale bars, 50 μm. g Quantification of collagen deposition. n = 4/group. Data represent mean ± SEM. *p < 0.05, **p < 0.01, ****p < 0.0001. P values were calculated with two-tailed unpaired Student t-test (e, g) and one-way ANOVA followed by Tukey’s post hoc test (c). Source data are provided as a Source Data file. 10 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE Immunohistochemistry. Specimens were fixed at RT for from 24 to 48 h in 10% for antigen retrieval. After cooling down to RT, the sections were treated with 0.3% formalin (#134-10047; FUJIFILM Wako Pure Chemical, Corp., Osaka, Japan) and H O in methanol (30 min at RT) and were subsequently incubated with the 2 2 embedded in paraffin (#7810; Sakura Finetek Japan Co., Ltd., Tokyo, Japan). To blocking solution for 20 min at RT, following which the sections were reacted with prepare the paraffin-embedded cell block specimens, obtained peritoneal macro- primary antibodies. The sections were then incubated with horseradish peroxidase phages were fixed in 10% neutral buffered formalin. Then cells were suspended in (HRP)-labeled secondary antibodies. Antibodies used in the immunoenzyme 1% sodium alginate and solidified by the addition of 1 M calcium chloride. Finally, method are listed in Supplementary Table 2. Immunoreactions were visualized gelatinous specimens containing macrophages were embedded in paraffinin a using the diaminobenzidine (DAB) substrate kit (#425011, Nichirei Biosciences). routine manner. Specimens were then sectioned at 3 µm. Sections were pretreated For double-immunostaining, sections visualized with DAB in the first NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 11 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 Fig. 6 Peritoneal macrophages accumulate to the colon in DSS-induced colitis. a Representative stitch images of the colon in LysM-eGFP mice 5 days after the start of 4% DSS-containing water. Higher magnification of the indicated area was shown in right. Scale bars, 50 μm. b Quantification of the + hi number of large LysM F4/80 cells per FOV in LysM-eGFP mice that were treated with normal water or DSS-containing water. n = 4/group. c Schematic protocol for peritoneal cell transfer experiments (from LysM-eGFP mice to C57BL/6 mice). d Representative images of the colon in C57BL/6 mice that was administered peritoneal cells obtained from LysM-eGFP mice intraperitoneally 24 h before the imaging. Images were taken 5 days after the start + hi of 4% DSS-containing water. Anti-F4/80 antibody (red) was applied topically to the serosa. Arrows indicate LysM F4/80 cells. Scale bars, 50 μm. + hi e Quantification of the number of large LysM F4/80 cells per FOV in WT mice that were treated with normal water or DSS-containing water. Mice were pretreated by intraperitoneal administration of peritoneal cells from LysM-eGFP mice 24 h before the imaging. n = 4/group. f Representative images of the GFP/+ RFP/+ colon in Cx3cr1 Ccr2 mice 5 days after the start of 4% DSS-containing water. Anti-F4/80 antibody (purple) was applied topically. Scale bars, 50 μm. Data were representative of three independent experiments. g Immunofluorescence staining of the colon harvested from control or 5 days after the start of 4% DSS-containing water with hyaluronic acid binding protein (HABP). Scale bars, 500 μm. h Quantification of the HABP-positive area (%). n = 5 (normal water) and 10 (DSS water). i Representative images of the colonic muscularis in control and Abx mice 5 days after the start of 4% DSS-containing hi water. Anti-F4/80 antibody (red) was applied topically to the serosa. Scale bars, 50 μm. j Quantification of the number of large F4/80 cells per FOV. n = 5/group. Data represent mean ± SEM. *p < 0.05, ****p < 0.0001. P values were calculated with a two-tailed unpaired Student t-test (b, e, h, j). Source data are provided as a Source Data file. immunostaining were then used to perform the second immunostaining, with the run using a flow cytometer (FACSCanto; BD Biosciences) and analyzed using reaction visualized using HistoGreen (#E109; LINARIS Biologische Produkte FlowJo software (Tree Star). GmbH, Dossenheim, Germany). Sections were mounted using malinol (Muto Pure Chemicals Co., Ltd, Tokyo, Japan). Measurement of colon tissue chemokine concentration. Quantification of concentrations of chemokines was performed using the validated Luminex bead- Image processing, cell counting, and area measurement for immunostained based assay from R&D Systems (Minneapolis, MN) according to the manu- facturer’s instructions. Briefly, a 20-mm-long colon tissue sample was placed into sections. Immunostained sections were photographed using a microscope (BX51, 1 mL PBS, and they were mechanically disrupted. The mixture was then cen- Olympus Corporation, Tokyo, Japan). To quantify immunostaining of Gr-1 by cell trifuged at 20,124 × g for 5 min three times and the supernatant was transferred to counting, two pathologists (D.Y. and Y.K.), blinded to mouse information eval- uated sections. To quantify immunostaining by calculating areas of immunostained a 0.22 μm PVDF DuraPore centrifugal filter (EMD Millipore, Billerica, MA) to remove any particles from the solution. Filtered samples were then incubated with sections, we used BZ-X800 (Keyence corp., Osaka, Japan). Immunostained sections were photographed and then entire fields were reconstructed into one picture. a capture bead cocktail on a 96-well plate in the dark for 2 h at room temperature. After incubation, wells were washed with wash buffer, incubated with a biotin IBA1 and CD3 area proportions of the colonic mucosa and muscularis for each antibody cocktail for 1 h next before another round washing, and incubated for entire field were measured, respectively. HABP area proportion for the entire field another 30 min with Streptavidin-PE. Following washes, the plate was read using a was also measured. Luminex 200 apparatus (Applied Cytometry Systems, UK) and analyzed with StarStation V.2.3 (Applied Cytometry Systems, UK). Total protein concentration of Peritoneal cell transfers. Peritoneal lavage was performed as described before filtered samples was measured using Bio-Rad Protein Assay (Bio-Rad Labora- using sterile PBS . Cells were washed with cold PBS twice and resuspended in tories). Results were normalized against the amount of total protein extracted from 100 μL PBS. 5 × 10 cells isolated from LysM-eGFP mice were directly transferred the colon tissues or the weight of the colon tissue. into naïve mice. Both intraperitoneal and intravenous transfers were performed 1 h prior to injury induction. In the 4% DSS-induced colitis model, peritoneal cells Bacterial DNA extraction and amplification. DNA was extracted from fecal pellet isolated from LysM-eGFP mice were transferred 4 days after the start of DSS- using a QIAamp Fast DNA Stool Mini Kit (QIAGEN) according to manufacturer’s containing water and the colon was imaged 24 h later. instructions and 16 S rRNA genes were amplified and sequenced using an Illumina MiSeq (Illumina, San Diego, CA, U.S.A.) as described previously . The 16 S rRNA SYTOX green staining of intestinal content. Bacterial load of intestinal contents operational taxonomic units (OTUs) were selected from the combined reads using was measured as previously described . Briefly, a fecal pellet was collected from a de novo OTU picking protocol clustered at 97% identity using the Quantitative control or an Abx-treated mouse and each sample was added 500 μL 4% paraf- Insights Into Microbial Ecology (QIIME) pipeline software. To investigate the ormaldehyde and mixed. After the incubation for 30 min at room temperature, bacterial diversity of each sample, the number of OTUs, Chao 1, and Shannon were 100 μLof fixed content was resuspended in 800 μL sterile PBS and 1 μL of SYTOX calculated, and a rarefaction curve was generated using QIIME. Differences of green (100 μg/mL) was added. After the incubation for 30 min in the dark, samples microbial communities were evaluated using phylogeny-based unweighted or were centrifuged and 50 μL of supernatant was spread on a slide. The image was weighted UniFrac distance matrices. Principal coordinate analysis (PCoA) graph recorded by microscopy in a green fluorescence channel using a 10× objective. and Unweighted Pair Group Method with Arithmetic mean (UPGMA) tree graph were depicted using QIIME. Detection of intestinal collagen by multi-photon microscopy. Intestinal fibrillar collagen was imaged using second-harmonic generation (SHG). In brief, focal DSS colitis model. Mice were given 4.0% DSS (molecular mass, 36,000–50,000, injured colons were removed, maintained in cold PBS, and imaged using a BX61WI MP Biomedicals, LLC) in the drinking water for continuous 5 days. Body weight upright microscope and FV1000MPE (Olympus, Tokyo, Japan) laser-scanning and disease activity score was observed daily in PBS-L treated control and CLL- microscope system equipped with a MaiTai HP Deep See femtosecond-pulsed laser treated mice. Disease activity index was evaluated according to the previous (Spectra-Physics, Santa Clara, CA, USA). Colonic tissue was visualized using a report and was described as follows; (a) general appearance: normal = 0; combination of multi-photon fluorescence and SHG using 890 nm excitation . piloerection = 1; lethargy and piloerection = 2; motionless, sickly = 4; (b) weight The signal was detected by the external non-descanned detectors (505 nm mirror loss: no change = 0; < 5% = 1; 6–10% = 2; 11–20% = 3; > 20% = 4; (c) feces and band-pass emission filters at 465–485 nm for SHG). Z-stacks were recorded at consistency: normal = 0; pasty, semi-formed = 2; liquid, sticky, or unable to 0.5 μm intervals with an Olympus XLPLN 25X WMP (water immersion; numerical defecate after 5 min = 4; and (d) rectal bleeding: no blood = 0; visible blood in aperture, 1.05; working distance 2.0 mm) objective lens. rectum = 2; visible blood on fur = 4. At 5 days after the start of DSS-containing water, colon length was measured, and the pathological findings were evaluated in a blinded fashion using a previously reported scoring system: (a) inflammatory Flow cytometry. After mice were anesthetized, colon or peritoneal lavage were cell infiltrate (severity, extent): mild, mucosa = 1; moderate, mucosa and sub- collected and placed in PBS on ice. Colonic cells were obtained from intestinal mucosa = 2; marked, transmural = 3; (b) intestinal architecture (epithelial changes, biopsies of the uninjured or injured area. After the homogenization, single-cell mucosal architecture): focal erosions = 1; erosions ± focal ulcerations = 2; extended suspensions were generated by mechanical disruption through a 40-µm nylon ulcerations ± granulation tissue ± pseudopolyps = 3. mesh (BD Bioscience). Residual red blood cells were lysed using ACK lysing buffer (Invitrogen). The cells were blocked using anti-CD16/32 antibody (#BE0307; 2.4G2 clone; Bio X Cell) (1:100 dilution) for 30 min. Then, cells were stained for 30 min Statistical analysis. Data were expressed as mean ± SEM. Unpaired Student t-test or with antibodies for specified markers. Nonviable cells were identified using viability Mann–Whitney U-test was used to compare between two groups as appropriate. One- TM dye efluor 780 (#65-0865-14; eBioscience) (1:1000 dilution) or Ghost Dye Red way ANOVA was used to compare more than two groups, followed by Tukey’spost 710 (#13-0871-T100; TONBO biosciences) (1:1000 dilution). A Foxp3 staining hoc test. Percent body weight change or disease activity index in PBS-L- or CLL-treated buffer set (eBioscience) was used for intracellular GATA6 staining. Samples were mice with 4% DSS colitis were compared by two-way repeated-measures ANOVA. 12 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE Fig. 7 Peritoneal macrophages exert a tissue repair function in DSS-induced colitis. a Body weight change of PBS-L treated control and CLL-treated mice in 4% DSS-induced colitis. n = 5 (normal water group treated by PBS-L or CLL) and 20 (DSS water group treated by PBS-L or CLL). Data were pooled from three independent experiments. b Disease activity index of PBS-L treated control and CLL-treated mice in 4% DSS-induced colitis. n = 5 (normal water group treated by PBS-L or CLL) and 10 (DSS water group treated by PBS-L or CLL). c Macroscopic findings of the colon 5 days after the start of 4% DSS- containing water (left) and quantification of colon length (right). n = 5 (normal water group treated by PBS-L or CLL) and 10 (DSS water group treated by PBS-L or CLL). d Representative H&E staining of the colon 5 days after the start of 4% DSS-containing water and e quantification of histomorphological scores. n = 5 (normal water group treated by PBS-L or CLL) and 10 (DSS water group treated by PBS-L or CLL). Scale bars, 100 μm. f Schematic protocol for peritoneal cell (PC) transfer experiments in DSS-induced colitis. g Body weight change and h disease activity index of PBS-treated control and PC-treated mice in 4% DSS-induced colitis. n = 5/group. i Macroscopic findings of the colon 5 days after the start of 4% DSS-containing water (left) and quantification of colon length (right). n = 5/group. j Representative H&E staining of the colon 5 days after the start of 4% DSS-containing water and k quantification of histomorphological scores. n = 5/group. Scale bars, 100 μm. Data represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. P values were calculated with two-tailed unpaired Student t-test (c, i), Mann–Whitney U-test (e, k), and two-way repeated-measures ANOVA (a, b, g, h). Source data are provided as a Source Data file. NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 13 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 Fig. 8 GATA6 peritoneal cavity macrophages infiltrate an injured intestine via a direct peritoneal route and contribute tissue repair. A scheme + + + showing the mechanism of GATA6 peritoneal macrophage recruitment and bloodstream-derived CCR2 monocyte to CX3CR1 monocyte/macrophage conversion in intestinal injury/inflammation. Experimental findings were reproduced at least twice to ensure consistency. A p value 9. Okabe, Y. & Medzhitov, R. Tissue-specific signals control reversible program <0.05 was considered statistically significant. All tests were two-tailed. All statistical of localization and functional polarization of macrophages. Cell 157, 832–844 analyses were performed using GraphPad Prism v8.0 software (GraphPad Software Inc., (2014). La Jolla, CA). 10. Rosas, M. et al. The transcription factor Gata6 links tissue macrophage phenotype and proliferative renewal. Science 344, 645–648 (2014). 11. Huber, S., Hoffmann, R., Muskens, F. & Voehringer, D. Alternatively activated Reporting summary. Further information on research design is available in the Nature macrophages inhibit T-cell proliferation by Stat6-dependent expression of Research Reporting Summary linked to this article. PD-L2. Blood 116, 3311–3320 (2010). 12. Martinez, F. O. & Gordon, S. The M1 and M2 paradigm of macrophage Data availability activation: time for reassessment. F1000Prime Rep. 6, 13 (2014). The sequence data generated in this study have been deposited in the NCBI repository 13. Hooper, L. V. & Macpherson, A. J. Immune adaptations that maintain GenBank Nucleotide under accession numbers OK665944-OK666378. The first is homeostasis with the intestinal microbiota. Nat. Rev. Immunol. 10, 159–169 accessible and the others can be accessed by editing the accession number in the (2010). hyperlink. Further data that support the findings of this study are available from the 14. Brown, E. M., Sadarangani, M. & Finlay, B. B. The role of the immune system corresponding author upon reasonable request. Source data are provided with this paper. in governing host-microbe interactions in the intestine. Nat. Immunol. 14, 660–667 (2013). 15. Niess, J. H. & Adler, G. Enteric flora expands gut lamina propria CX3CR1+ Received: 28 November 2020; Accepted: 29 November 2021; dendritic cells supporting inflammatory immune responses under normal and inflammatory conditions. J. Immunol. 184, 2026–2037 (2010). 16. Bain, C. C. & Jenkins, S. J. The biology of serous cavity macrophages. Cell Immunol. 330, 126–135 (2018). 17. Buechler, M. B. et al. A stromal niche defined by expression of the References transcription factor WT1 mediates programming and homeostasis of cavity- 1. Mowat, A. M., Scott, C. L. & Bain, C. C. Barrier-tissue macrophages: resident macrophages. Immunity 51, 119–130 e115 (2019). functional adaptation to environmental challenges. Nat. Med. 23, 1258–1270 18. Ghosn, E. E. et al. Two physically, functionally, and developmentally distinct (2017). peritoneal macrophage subsets. Proc. Natl Acad. Sci. USA 107, 2568–2573 (2010). 2. Cerovic, V., Bain, C. C., Mowat, A. M. & Milling, S. W. Intestinal macrophages 19. Gautier, E. L. et al. Gene-expression profiles and transcriptional regulatory and dendritic cells: what’s the difference? Trends Immunol. 35, 270–277 pathways that underlie the identity and diversity of mouse tissue (2014). macrophages. Nat. Immunol. 13, 1118–1128 (2012). 3. Zigmond, E. & Jung, S. Intestinal macrophages: well educated exceptions from 20. Louis, C. et al. Specific contributions of CSF-1 and GM-CSF to the dynamics the rule. Trends Immunol. 34, 162–168 (2013). of the mononuclear phagocyte system. J. Immunol. 195, 134–144 (2015). 4. Bain, C. C. et al. Constant replenishment from circulating monocytes 21. Cassado Ados, A., D’Imperio Lima, M. R. & Bortoluci, K. R. Revisiting mouse maintains the macrophage pool in the intestine of adult mice. Nat. Immunol. peritoneal macrophages: heterogeneity, development, and function. Front. 15, 929–937 (2014). Immunol. 6, 225 (2015). 5. Bain, C. C. et al. Resident and pro-inflammatory macrophages in the colon 22. Kim, K. W. et al. MHC II+ resident peritoneal and pleural macrophages rely represent alternative context-dependent fates of the same Ly6Chi monocyte on IRF4 for development from circulating monocytes. J. Exp. Med. 213, precursors. Mucosal Immunol. 6, 498–510 (2013). 1951–1959 (2016). 6. Honda, M. et al. Perivascular localization of macrophages in the intestinal 23. Barth, M. W., Hendrzak, J. A., Melnicoff, M. J. & Morahan, P. S. Review of the mucosa is regulated by Nr4a1 and the microbiome. Nat. Commun. 11, 1329 macrophage disappearance reaction. J. Leukoc. Biol. 57, 361–367 (1995). (2020). 24. Stengel, S. et al. Peritoneal level of CD206 associates with mortality and an 7. Wang, J. & Kubes, P. A reservoir of mature cavity macrophages that can inflammatory macrophage phenotype in patients with decompensated rapidly invade visceral organs to affect tissue repair. Cell 165, 668–678 (2016). cirrhosis and spontaneous bacterial peritonitis. Gastroenterology 158, 8. Gautier, E. L. et al. Gata6 regulates aspartoacylase expression in resident 1745–1761 (2020). peritoneal macrophages and controls their survival. J. Exp. Med. 211, 25. Zhang, N. et al. Expression of factor V by resident macrophages boosts host 1525–1531 (2014). defense in the peritoneal cavity. J. Exp. Med. 216, 1291–1300 (2019). 14 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE 26. Deniset, J. F. et al. Gata6(+) pericardial cavity macrophages relocate to the licensed under a Creative Commons Attribution 3.0 Unported License (https:// injured heart and prevent cardiac fibrosis. Immunity 51, 131–140 e135 (2019). creativecommons.org/licenses/by/3.0/). The source of the mouse icons used in Figs. 2h, 27. Gautier, E. L., Ivanov, S., Lesnik, P. & Randolph, G. J. Local apoptosis 4a, 6c, 7f, and Supplementary Figs. 6a, 6d, 10a are Free vector graphics on Pixabay. mediates clearance of macrophages from resolving inflammation in mice. Content on Pixabay is licensed under Pixabay License (https://pixabay.com/service/ Blood 122, 2714–2722 (2013). terms/#license). We also thank Jeremy Allen, Ph.D., from Edanz Group (https://en- 28. Atarashi, K. et al. Treg induction by a rationally selected mixture of Clostridia author-services.edanzgroup.com/ac) for editing a draft of this manuscript. M.H. is strains from the human microbiota. Nature 500, 232–236 (2013). supported by the Research Fellowship of the Uehara Memorial Foundation and grants 29. Atarashi, K. et al. Induction of colonic regulatory T cells by indigenous from the Ministry of Education, Culture, Sports, Sciences and Technology of Japan Clostridium species. Science 331, 337–341 (2011). (KAKENHI 19H03716). 30. Constantinides, M. G. et al. MAIT cells are imprinted by the microbiota in early life and promote tissue repair. Science 366, eaax6624 (2019). Author contributions 31. Furusawa, Y. et al. Commensal microbe-derived butyrate induces the M.H. designed and performed experiments, analyzed data, and wrote the manuscript. differentiation of colonic regulatory T cells. Nature 504, 446–450 (2013). M.K. performed experiments, analyzed data, and wrote the manuscript. D.Y. performed 32. Ruiz-Alcaraz, A. J. et al. Characterization of human peritoneal monocyte/ experiments and analyzed pathological specimens. Y.K. and T.H. provided material macrophage subsets in homeostasis: phenotype, GATA6, phagocytic/oxidative support and reviewed the manuscript. activities and cytokines expression. Sci. Rep. 8, 12794 (2018). 33. Ng, S. C. et al. Worldwide incidence and prevalence of inflammatory bowel disease in the 21st century: a systematic review of population-based studies. Competing interests Lancet 390, 2769–2778 (2018). The authors declare no competing interests. 34. Honda, M. & Kubes, P. Neutrophils and neutrophil extracellular traps in the liver and gastrointestinal system. Nat. Rev. Gastroenterol. Hepatol. 15, Additional information 206–221 (2018). Supplementary information The online version contains supplementary material 35. Seo, D. H. et al. Interleukin-33 regulates intestinal inflammation by modulating available at https://doi.org/10.1038/s41467-021-27614-9. macrophages in inflammatory bowel disease. Sci. Rep. 7, 851 (2017). 36. Liu, T. et al. Treatment of dextran sodium sulfate-induced experimental colitis Correspondence and requests for materials should be addressed to Masaki Honda. by adoptive transfer of peritoneal cells. Sci. Rep. 5, 16760 (2015). 37. Dal-Secco, D. et al. A dynamic spectrum of monocytes arising from the in situ Peer review information Nature Communications thanks the anonymous reviewer(s) for reprogramming of CCR2+ monocytes at a site of sterile injury. J. Exp. Med. their contribution to the peer review of this work. Peer reviewer reports are available. 212, 447–456 (2015). 38. Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH image to imageJ: 25 Reprints and permission information is available at http://www.nature.com/reprints years of image analysis. Nat. Methods 9, 671–675 (2012). 39. McDonald, B. et al. Intravascular danger signals guide neutrophils to sites of Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in sterile inflammation. Science 330, 362–366 (2010). published maps and institutional affiliations. 40. Macpherson, A. J., Geuking, M. B., Kirundi, J., Collins, S. & McCoy, K. D. Gnotiobiotic and Axenic Animals. In Encyclopedia of Microbiology (ed. Moselio, S.) 237–246 (Elsevier, 2009). Open Access This article is licensed under a Creative Commons 41. Gailhouste, L. et al. Fibrillar collagen scoring by second harmonic microscopy: Attribution 4.0 International License, which permits use, sharing, a new tool in the assessment of liver fibrosis. J. Hepatol. 52, 398–406 (2010). adaptation, distribution and reproduction in any medium or format, as long as you give 42. Omori, M. et al. Fecal microbiome in dogs with inflammatory bowel disease and intestinal lymphoma. J. Vet. Med. Sci. 79, 1840–1847 (2017). appropriate credit to the original author(s) and the source, provide a link to the Creative 43. Taylor, B. C. et al. TSLP regulates intestinal immunity and inflammation in Commons license, and indicate if changes were made. The images or other third party mouse models of helminth infection and colitis. J. Exp. Med. 206, 655–667 material in this article are included in the article’s Creative Commons license, unless (2009). indicated otherwise in a credit line to the material. If material is not included in the 44. Erben, U. et al. A guide to histomorphological evaluation of intestinal article’s Creative Commons license and your intended use is not permitted by statutory inflammation in mouse models. Int J. Clin. Exp. Pathol. 7, 4557–4576 (2014). regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. Acknowledgements We would like to thank Servier for providing Servier Medical Art, which was used for the © The Author(s) 2021 creation of figures (Fig. 8 and Supplementary Fig. 7a). Servier Medical Art by Servier is NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 15 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals

Directly recruited GATA6 + peritoneal cavity macrophages contribute to the repair of intestinal serosal injury

Loading next page...
 
/lp/springer-journals/directly-recruited-gata6-peritoneal-cavity-macrophages-contribute-to-n7JM10KkPb

References (50)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
eISSN
2041-1723
DOI
10.1038/s41467-021-27614-9
Publisher site
See Article on Publisher Site

Abstract

ARTICLE https://doi.org/10.1038/s41467-021-27614-9 OPEN Directly recruited GATA6 + peritoneal cavity macrophages contribute to the repair of intestinal serosal injury 1,3 1,3 1,2 2 1 Masaki Honda , Masashi Kadohisa , Daiki Yoshii , Yoshihiro Komohara & Taizo Hibi Recruitment of bone marrow derived monocytes via bloodstream and their subsequent conversion to CX3CR1 macrophages in response to intestinal injury is dependent on CCR2, Nr4a1, and the microbiome. This process is critical for proper tissue repair; however, GATA6 peritoneal cavity macrophages might represent an alternative, more readily avail- able source of mature and functional myeloid cells at the damaged intestinal locations. Here hi + we show, using spinning-disk confocal microscopy, that large F4/80 GATA6 peritoneal cavity macrophages promptly accumulate at damaged intestinal sites upon intestinal thermal injury and upon dextran sodium sulfate induced colitis in mice via a direct route from the peritoneal cavity. In contrast to bloodstream derived monocytes/macrophages, cavity mac- rophages do not depend on CCR2, Nr4a1 or the microbiome for recruitment, but rather on the ATP-release and exposed hyaluronan at the site of injury. They participate in the removal of necrotic cells, revascularization and collagen deposition and thus resolution of tissue damage. In summary, peritoneal cavity macrophages represent a rapid alternative route of intestinal tissue repair to traditional monocyte-derived macrophages. 1 2 Department of Transplantation and Pediatric Surgery, Kumamoto University, 1-1-1 Honjo, Chuo-ku, Kumamoto 860-8556, Japan. Department of Cell Pathology, Kumamoto University, 1-1-1 Honjo, Chuo-ku, Kumamoto 860-8556, Japan. These authors contributed equally: Masaki Honda, Masashi Kadohisa. email: honda.masaki@kuh.kumamoto-u.ac.jp NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 1 1234567890():,; ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 issue-resident macrophages are major players in maintain- monocytes/macrophages depends on CCR2, Nr4a1, and the ing local homeostasis in response to environmental changes microbiome . This process contributes to the resolution of Tsuch as infection and inflammation .Unlike mostother inflammation and tissue repair but requires at least a few days for organs, the intestinal mucosal lamina propria is constantly conversion from a classical pro-inflammatory monocyte to an replenished with macrophages from bone marrow progenitor alternative phenotype. However, because the intestine is a portal to 2–5 hi monocytes .Ininflammation, rapid accumulation of Ly6C the external environment, we speculated that repair is initiated monocytes into the injured area and conversion to CX3CR1 immediately via an alternative route during severe intestinal injury. 2 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE hi Fig. 1 Large F4/80 macrophages rapidly accumulate in response to intestinal thermal injury. a Representative stitch images of colonic LP (left) in GFP/+ + Cx3cr1 mice. Representative still (middle) and three-dimensional (3D) image (right) of CX3CR1 macrophages (green) in colonic LP. Scale bars, 50 μm. + + GFP/+ RFP/+ b Representative images of colonic LP CX3CR1 monocytes/macrophages and CCR2 cells (red) 6 h after focal intestinal injury in Cx3cr1 Ccr2 + + mouse. Scale bars, 50 μm. c Migration paths, d crawling velocities of CCR2 cells and CX3CR1 cells in response to intestinal injury. n= 30/group (obtained hi by three independent images). e Quantification of large F4/80 macrophages within intestinal injury at indicated time points. n= 3 (2 h) and 5 (6, 24, 48 h). hi f Representative image of large F4/80 macrophages in injury area at 24 h post-injury. Higher magnification of the indicated area is shown in right. Scale bars, 100 μm(left)and 50 μm (right). g H&E and h immunofluorescence staining for F4/80 (brown) and GATA6 (green) of injured colon 24 h post-injury (cross section). Arrowheads indicate GATA6 peritoneal macrophages. Scale bars, 10 μm. Data were representative of three independent experiments. Data represent box and whiskers (Fig. 1d) and mean ± SEM (Fig. 1e). Source data are provided as a Source Data file. Large GATA6 peritoneal cavity macrophages rapidly invade were adjacent to the injury site remained sessile and did not move the subcapsular region of the liver following sterile injury/ from their original position towards the damaged site (Fig. 1b–d inflammation where they contribute to tissue repair . The zinc and Supplementary Movie 1). Despite their sessile nature, when finger transcription factor, GATA6, controls proliferation, survi- topical F4/80 antibody was applied to the injury site, a very sig- 8–10 hi val, and metabolism . Wang et al. revealed that peritoneal nificant population of large F4/80 macrophages accumulated macrophages expand rapidly at liver injury sites and up-regulate within 2 h post-injury in C57BL/6 mice (Fig. 1e and Supple- hi CD273, CD206, and arginase-1, markers of alternatively activated mentary Fig. 1a). The accumulation of these large F4/80 cells 7,11,12 macrophages . The intestine is also surrounded by a peri- peaked at 24 h after injury and persisted for at least 48 h (Fig. 1e toneal cavity; however, whether peritoneal macrophages can (1) and Supplementary Fig. 1a). To confirm that they were not GFP/+ RFP/+ detect luminal injury within the peritoneum, (2) invade into the derived from monocytes, we imaged Cx3cr1 Ccr2 mice intestine across a formidable mesothelial barrier under patholo- 6 h after injury with topical administration of F4/80 antibody. RFP GFP gical conditions, and (3) affect tissue repair, have not been This revealed that neither CCR2 nor CX3CR1 co-localized hi explored. The molecular mechanisms involved in this process are with large F4/80 cells (Supplementary Fig. 1b). The neutrophil also unknown. The mammalian intestinal tract is populated with marker, Ly6G, also did not show any co-localization with F4/80 over 100 trillion microbes, the majority of which reside in the (Supplementary Fig. 1b). At 24 h after injury, CCR2 monocytes 13,14 colon so they could be a potent recruitment cue for peritoneal formed a ring surrounding the injury site and their accumulation hi macrophages. was via blood vessels, whereas the large F4/80 cells could not be Here, we examine the recruitment and function of peritoneal seen in blood vessels and were positioned within the center of the cavity macrophages in intestinal injury. Using spinning-disc dual- injury as a large aggregate (Fig. 1f). There was a striking differ- + + hi laser intravital microscopy and a sterile burn injury model of the ence in size between CCR2 CX3CR1 cells and large F4/80 hi intestine, we detect large F4/80 peritoneal macrophages accu- cells, the latter being at least twice the size of the former (Sup- hi mulating at the intestinal injury site. They are recruited to the plementary Fig. 1c, d). Importantly, these large F4/80 cells in the injured intestine via a unique peritoneal route in response to ATP intestinal injury site expressed GATA6, a transcription factor released by necrotic cells. Interaction between hyaluronan and specific for large peritoneal cavity macrophages, but not intestinal CD44 is indispensable for the recruitment, while CCR2, Nr4a1, F4/80 macrophages (Fig. 1g, h and Supplementary Fig. 2a–d). and the microbiome are not involved. Accumulated large peri- Using flow cytometry, we confirmed the intravital microscopy toneal macrophages contribute to the removal of necrotic cells data showing that there was a population of GATA6 hi hi within the intestinal injury site and help with revascularization CD11b F4/80 macrophages in the injured colon (Supplemen- and collagen deposition. We also focus on the most common tary Fig. 3a–c). The other populations of macrophages did not form of intestinal disease, inflammatory bowel disease (IBD), and express GATA6 (Supplementary Fig. 3c). observe more severe dextran sodium sulfate (DSS)-induced colitis activity in the absence of peritoneal macrophages. Overall, these Peritoneal macrophages from the peritoneal cavity accumulate data demonstrate the importance of the immune player at the site of intestinal injury independent of CCR2 or Nr4a1. “GATA6 peritoneal cavity macrophages” in intestinal injury. Luminex assays revealed induction of MCP-1 (CCL2), the key These findings prompt further studies to explore the mechanisms + chemokine that attracts CCR2 monocytes, and neutrophil che- and/or therapeutic strategies for disorders of the intestinal tract mokine, KC, in the colon at 24 h post-injury (Fig. 2a). However, with a particular focus on IBD. the lack of the CCL2 receptor, CCR2, did not affect the recruit- ment of GATA6 macrophages (Fig. 2b, c) indicating that (1) Results CCL2 was not responsible for peritoneal macrophage recruitment hi Large F4/80 macrophages accumulate promptly in response and (2) CCR2-positive monocytes were not important for peri- hi to sterile intestinal injury. Intestinal macrophages express a toneal macrophage recruitment. The recruitment of large F4/80 specific cell surface marker, CX3CR1. Using intravital imaging cells within damaged intestine at 24 and 48 h after injury was not −/− lo and mice with a fluorescent reporter for CX3CR1, we imaged different in Nr4a1 mice, which lack Ly6C monocytes at the steady state intestinal lamina propria macrophages and found injury site, compared with wild-type mice (Fig. 2b, c). The that these cells form an interdigitated physical chain surrounding CX3CR1 ligand is involved in the recruitment of macrophages in the blood vessels (Fig. 1a). We examined the recruitment of some tissues; however, CX3CR1-deficient mice accumulated hi monocytes and macrophages in response to sterile thermal injury equivalent numbers of large F4/80 macrophages to wild-type GFP/+ RFP/+ of the intestine using Cx3cr1 Ccr2 mice. A 500 μm focal mice after intestinal injury (Supplementary Fig. 4a, b). necrotic lesion extending into the lamina propria was created To confirm that the peritoneum was the source of the large F4/ hi + from the serosa side of the colon using a thermal probe. This 80 GATA6 macrophages, we depleted peritoneal macrophages model allowed us to image recruitment of immune cells in an area by intraperitoneal administration of clodronate liposome (CLL), + 7 eradicated of resident cells. Imaging showed that CCR2 as described previously . Intraperitoneal CLL treatment did not + + monocytes but not CX3CR1 monocytes infiltrated into the affect the distribution of intestinal CX3CR1 macrophages + + injury site within 6 h. Meanwhile, CX3CR1 macrophages that (Fig. 2d, e) or the accumulation of CCR2 monocytes within NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 the intestinal injury site (Supplementary Fig. 5a, b) but incapable of entering the injury site when transferred intrave- hi significantly suppressed the number of large F4/80 macrophages nously (Fig. 2h–j) further indicating that this was not the route of in the intestinal injury site (Fig. 2f, g). Additionally, peritoneal infiltration. When peritoneal cells from LysM-eGFP mice with cells transferred from LysM-eGFP mice (in which > 85% of the depleted GATA6 macrophages by intraperitoneal CLL admin- + + 7 GFP cells in the peritoneal cavity are GATA6 macrophages ) istration were transferred into C57BL/6 mice intraperitoneally, into the peritoneum of C57BL/6 mice, resulted in accumulation GFP cells were not found within the intestinal injury site of these cells at the intestinal injury site. GFP macrophages were (Fig. 2k, i). Next, peritoneal cells from LysM-eGFP mice were 4 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE Fig. 2 Peritoneal Macrophages accumulate into the intestinal injury site directly via the peritoneal route regardless of CCR2 or Nr4a1. a Luminex assays of chemokines in colon tissue samples at steady state and 24 h after thermal injury. n = 5 (control group) and 8 (injury group). b Representative hi RFP/RFP −/− images of large F4/80 cells accumulated in injury site at 48 h after focal intestinal injury in WT, Ccr2 , and Nr4a1 mouse. Scale bars, 50 μm. hi + c The number of large F4/80 cells at indicated time points are quantified (n = 5/group). d Representative images and e quantification of CX3CR1 cells per field of view (FOV) in lamina propria of the colon in control and 2, 7, 14 days after the intraperitoneal administration of CLL. Data were collected by GFP/+ hi imaging of CX3CR1 mice. n = 4 (CLL 2, 7, 14 d) and 6 (control). Scale bars, 50 μm. f Representative images of large F4/80 cells accumulated in hi intestinal injury site at 48 h post-injury. Mice were treated by PBS-L or CLL 4 days before the injury. Scale bars, 50 μm. g The number of large F4/80 cells are quantified. n = 3(PBS-L group) and 5(CLL group). h Schematic protocol for peritoneal cell (PC) transfer experiments (from LysM-eGFP mice to C57BL/ 6 mice). i Representative images and j the number of LysM-eGFP cells accumulated in C57BL/6 mice (n = 3/group). Scale bars, 100 μm. k Representative images and l quantification of LysM cells within the injured colon (6 h) of C57BL/6 mice that were transferred PCs intraperitoneally from LysM-eGFP mice with i.p. administration of PBS-L or CLL. Scale bars, 50 μm. n = 3/group. Data represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, NS not significant. P values were calculated with a two-tailed unpaired Student t-test (a, g, j, l) and one-way ANOVA followed by Tukey’s post hoc test (c, e). Source data are provided as a Source Data file. hi stained using a PE-ICAM2 (CD102) antibody. In total, 78.6% of the peritoneal cavity nor did it affect the amount of large F4/80 + + GFP peritoneal cells were also positive for CD102, which GATA6 macrophages that accumulated at the intestinal injury demonstrated that the double-positive cells are large peritoneal site (Fig. 4c–f), indicating that the phenotype and recruitment of macrophages (Supplementary Fig. 6a–c). We next transferred large peritoneal macrophages in intestinal injury is independent peritoneal cells from LysM-eGFP mice to the peritoneal cavity of of the microbiome. CD44, the predominant adhesion molecule, C57BL/6 mice after staining CD102. Imaging of the intestine at was also not significantly affected by the diminution of the gut + + 6 h after injury showed that LysM CD102 cells had accumu- microbiota (Fig. 4c, d). lated in the injured area but not in the uninjured area (Supplementary Fig. 6d–f). Peritoneal macrophages promote intestinal repair. We pre- RFP/RFP viously reported that healing in Ccr2 mice at 48 h after Recruitment of peritoneal cavity macrophages to the intestinal injury and later was delayed. Time-lapse imaging of the colon at hi injury site is dependent on ATP and hyaluronan-CD44 inter- 24 h after thermal injury showed that large F4/80 macrophages action. ATP is an important damage-associated molecular pat- were already at the site disassembling the nearby SYTOX tern (DAMP) in sterile injury that recruits both neutrophils and necrotic cells (Fig. 5a). Next, we imaged the SYTOX green- macrophages to the injury site . Pretreatment with apyrase, an positive cells within the intestinal injury site after depletion of ATPase, or an ATP receptor antagonist inhibited the accumula- peritoneal macrophages by intraperitoneal CLL administration. hi tion of large F4/80 macrophages at the intestinal injury site At 48 h post-injury, the clearance of necrotic cells was delayed in (Fig. 3a, b). To further explore the mechanism of peritoneal CLL-treated mice compared with PBS-liposome (PBS-L)-treated macrophage dynamics in intestinal injury, we blocked CD44, a mice (Fig. 5b, c). Furthermore, peritoneal macrophage depletion leukocyte adhesion molecule highly expressed on GATA6 was associated with significantly impaired revascularization and peritoneal macrophages, and imaged intestinal injury sites. Pre- collagen deposition within the intestinal injury site (Fig. 5d–g). treatment with anti-CD44 antibody prevented the recruitment of The lack of clearance of SYTOX-positive cells was even more hi RFP/RFP large F4/80 macrophages to the intestinal injury site (Fig. 3c, d). pronounced in peritoneal macrophage-depleted Ccr2 mice Importantly, immunofluorescent staining showed exposed hya- (Fig. 5b, c), indicating that both CCR2 monocytes, which + + luronan, which is the ligand for CD44, within the intestinal injury became CX3CR1 monocytes/macrophages, and large GATA6 site but not on the serosal surface (Fig. 3e). Administration of peritoneal macrophages contribute to necrotic cell clearance, hyaluronidase, an enzyme that breaks down hyaluronan, also perhaps in separate layers of the intestine. Indeed, imaging of the prevented the recruitment of large peritoneal macrophages injured colon showed that CCR2 monocytes accumulated hi (Fig. 3f). Intriguingly, injury from the mucosal side induced some mainly in the lamina propria, whereas large F4/80 macrophages hi accumulation of large F4/80 macrophages from the serosal accumulated mainly in the muscularis (Supplementary Fig. 7a, b). surface, indicating that these accumulation mechanisms are also The pathological analysis confirmed that the majority of F4/ hi + partially functional in severe mucosal injury (Fig. 3g, h). 80 GATA6 macrophages were infiltrating the muscularis in the injured colon (Supplementary Fig. 5c, d). It is also worth noting that the muscularis has a lower vascular density than the lamina The gut microbiota is not critical for the recruitment of propria (Supplementary Fig. 6a, b), further emphasizing the GATA6 peritoneal macrophages in response to intestinal importance of blood flow-independent peritoneal macrophage injury. Recruitment of CCR2 monocytes to the intestine and accumulation when the injury reaches the muscularis. conversion to CX3CR1 monocytes/macrophages is microbiome- dependent in both the emergency repair as well as in steady state 4,6,15 turn over . Because the burn injury extended from the serosa Peritoneal macrophages accumulate in the colon in response to all the way through to the lamina propria, we predicted that gut DSS-induced colitis. To evaluate whether peritoneal macro- microbiota would be critical for the recruitment of GATA6 phages can respond to inflammation that starts in the lamina peritoneal cavity macrophages. We used broad-spectrum anti- propria and develops towards the serosa, mice were orally biotics from before birth until adulthood (Fig. 4a), which deci- administered 4% DSS-containing water for 5 days. Intravital mated the intestinal microbiota as assessed by SYTOX green imaging of the colon in LysM-eGFP mice with F4/80 antibody + hi staining (Fig. 4b). Moreover, sequencing analyses of feces showed topically applied to the serosa showed that LysM F4/80 large the remarkable differences of microbial communities between macrophages were infiltrating the muscularis (Fig. 6a, b). When control and Abx-treated mice (Fig. 4c and Supplementary peritoneal cells from LysM-eGFP mice were intraperitoneally Fig. 7a–d). However, these changes had no impact on the number transferred to C57BL/6 mice, a similar accumulation of peritoneal hi hi + + hi of CD11b F4/80 GATA6 peritoneal macrophages found in LysM F4/80 macrophages was found in the DSS-induced colitis NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 GFP/+ RFP/+ model (Fig. 6c–e). We also subjected Cx3cr1 Ccr2 mice in the colon of DSS-induced colitis compared with the control to DSS-induced colitis and found that no CCR2 and/or (Fig. 6g, h). As described in the thermal injury model, Abx + hi hi CX3CR1 signal co-localized with the large F4/80 cells (Fig. 6f) treatment did not influence on the recruitment of F4/80 large indicating that these are distinct cell lineages. These findings peritoneal macrophages to the colon in DSS-induced colitis indicate that these large invading macrophages are derived from (Fig. 6i, j). the peritoneum, not the vasculature. For that matter, immuno- To further analyze the role of peritoneal macrophages in DSS- fluorescence staining showed increased expression of hyaluronan induced colitis, we depleted peritoneal macrophages by 6 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE Fig. 3 Recruitment of peritoneal macrophages to the site of intestinal injury depends on ATP and hyaluronan-CD44 interaction. a Representative hi images and b quantification of the number of F4/80 macrophages within intestinal injury site 24 h post-injury in mice that were pretreated with apyrase or p2rx7 antagonist. Scale bars, 50 μm. n = 4 (Apyrase), 5 (control), and 7 (P2X7 antagonist). c Representative images and d quantification of the number hi of F4/80 macrophages within intestinal injury site 24 h post-injury in mice that were pretreated with isotype control or anti-CD44 antibody. Scale bars, 50 μm. n = 4 (isotype) and 5 (anti-CD44). e Immunofluorescence staining of the colon harvested from control or 2 h after injury with hyaluronic acid hi binding protein (HABP). The dashed line indicates the injury border. f Quantification of the number of F4/80 macrophages within intestinal injury site 6 h post-injury in mice that were treated with hyaluronidase. Scale bars, 50 μm. n= 4 (hyaluronidase) and 5 (control). g Representative images and hi h quantification of the number of F4/80 macrophages in the colonic muscularis with no injury or 24 h post-injury from mucosa or serosa. Scale bars, 50 μm. n = 5 (no injury, injury from serosa) and 6 (injury from mucosa). Data represent mean ± SEM. *p < 0.05, **p < 0.01, ****p < 0.0001. P values were calculated with two-tailed unpaired Student t-test (d, f) and one-way ANOVA followed by Tukey’s post hoc test (b, h). Source data are provided as a Source Data file. intraperitoneal administration of CLL and compared body weight (Fig. 8). However, somewhat surprising was the observation that change, disease activity index, colon length, and pathological these cells could also accumulate in the intestinal tract when the findings with those of PBS-L-administered DSS-treated mice. We injury was induced via DSS colitis, an injury that starts primarily found that CLL-treated mice had increased levels of injury and in the lamina propria. inflammation, including more severe weight loss and increased Pleural, pericardial, and peritoneal cavity macrophages have disease activity score (Fig. 7a, b and Supplementary Table 1). In been studied extensively, primarily in vitro . The homeostatic the CLL-treated mice, colon length, which is a marker of and transcriptomic signatures of these tissue-resident macro- inflammatory injury, was often shorter on the 5th day of DSS- phages are maintained by Wt1 mesothelial and fibroblastic induced colitis, but this difference was not significant compared stromal cells via the generation of retinoic acid . Cavity mac- with that of the PBS-L-treated mice (Fig. 7c). Moreover, colon rophages express CD11b, F4/80, and CSF1R and can be divided + hi - lo tissue sections obtained 5 days after the start of DSS adminis- into large GATA6 F4/80 and small GATA6 F4/80 8–10,18–20 tration were carefully evaluated. CLL-treated mice showed more macrophages . Small macrophages play a pivotal role in 18,21 severe histomorphological scores, including marked inflamma- inflammatory responses and infections . Kim et al. showed tory cell infiltration and disrupted intestinal architecture, that they are continuously replenished by blood monocytes compared with PBS-L mice (Fig. 7d, e). The degrees of infiltration dependent on the transcription factor, IRF4, and signals from the by neutrophils, macrophages, and lymphocytes were compared microbiome . By contrast, the large macrophages are derived between PBS-L and CLL groups, but no significant difference was from the yolk sac and have multiple functions regarding recog- observed in the lamina propria or muscularis, respectively nition and phagocytosis of pathogens, antigen presentation, and (Supplementary Fig. 11a, b). Although the number of neutrophils resolution of inflammation . They are also known to disappear 21,23,24 in the muscularis of the DSS-induced colitis group treated with in the early stage of infection or inflammation although CLL was approximately twice that of the group treated with PBS- this may be caused by their aggregatory properties that lead to L, it did not reach statistical significance. Indeed, chemokine clump formation rather than actual disappearance . These large expression in the colon tissue was not affected by CLL treatment GATA6 peritoneal macrophages express phagocytic receptors (Supplementary Fig. 11c). In the final series of experiments, such as Tim4, MerTK, CD36, and genes for cell adhesion and 19,26,27 peritoneal cell transfer experiments were performed for DSS- angiogenesis, suggesting a repair phenotype . Using intra- induced colitis to evaluate its effect on inflammation and vital imaging, we visualized peritoneal GATA6 macrophages symptoms (Fig. 7f). In comparison with the PBS-treated control recruited to the injured intestine. These cells exerted an in vivo group, the peritoneal cell transfer group had a lower disease reparative function to a focal injury that started in the serosa and activity index, while the other parameters did not show to DSS which causes inflammation primarily starting in the statistically significant differences (Fig. 7g–k). lamina propria and moving towards the serosa. The interaction between gut microbiota and the immune system in tissue repair is an important issue. Gut microbiota can shape Discussion intestinal mucosal T cells, including Foxp3 regulatory T cells and The intestinal tract is a barrier to trillions of bacteria; therefore, 28–30 mucosal-associated invariant T cells . Additionally, Furusawa tissue repair is vital. Resident macrophages play a pivotal role in et al. provided the mechanistic insight that microbe-derived buty- intestinal repair helping remove debris and also helping to induce rate regulates the differentiation of regulatory T cells .Using angiogenesis. However, tissue-resident CX3CR1 macrophages intravital imaging, we recently revealed that prenatal antibiotic do not move during intestinal injury and depend on classical + + administration and germ-free conditions altered the localization of CCR2 monocytes to be recruited and to convert to CX3CR1 intestinal mucosal CX3CR1 macrophages and impeded the monocytes/macrophages. This conversion process takes more development of a perivascular anatomical barrier .Bycontrast, the than 2 days and is dependent on the transcription factor, Nr4a1 , hi microbiome did not regulate the function of GATA6 peritoneal indicating that early healing must be dependent on the Ly6C cavity macrophages because their numbers were not altered in mice monocytes. However, the timing and inherent inflammatory nature of these cells make it difficult to explain how healing could given antibiotics from birth and they still accumulated the injured intestine. Despite significant alterations in many different immune begin so rapidly. In this study, we identified a third macrophage type that contributed to tissue repair, namely GATA6 peritoneal cells in the dysbiotic intestine, none of the many differences affected peritoneal macrophage accumulation to the injured intestine indi- cavity macrophages. They rapidly accumulated via the peritoneal route at the injured site in response to intestinal injury within the cating that these cells are not affected by alterations in the micro- serosa. Both the alarmin ATP and CD44-hyaluronan interactions biome or intestinal immunobiome. were key for recruitment from the peritoneal cavity to the This study also has clinical implications because humans have hi hi damaged intestinal serosa, which led to tissue repair independent an intraperitoneal mature CD14 CD16 macrophage subset that + + 32 of CCR2 monocyte-derived CX3CR1 monocytes/macrophages shows intracellular GATA6 expression . As a typical digestive NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 tract disease, IBDs, including ulcerative colitis and Crohn’s dis- GATA6 macrophage function is also worth considering as a ease, are a worldwide burden with high incidence in developed form of therapy. Indeed, adoptive transfer of peritoneal macro- countries and increasing incidence in developing countries . phages with or without IL33, which can induce the proliferation Currently, new therapeutic strategies targeting the recruitment and alternative activation of peritoneal macrophages, ameliorated 35,36 and activation of immune cells from the vasculature are being inflammation in a preclinical colitis model . Whether these tested . Our results raise the intriguing possibility that enhancing peritoneal cavity macrophages also contribute to colorectal cancer 8 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE Fig. 4 Gut microbiota does not affect the phenotype and dynamics of peritoneal macrophages in intestinal injury. a Schematic protocol for the generation of dysbiotic mice. b Representative stitch images (from 12 different fields of view) of SYTOX green particles in intestinal contents of SPF control and Abx-treated mice. Scale bars, 50 μm. c Relative abundance of bacterial amplicon sequence variants in feces obtained from control or Abx- treated mice. The bar plot is displayed at the phylum level. n = 4/group. d Flow cytometry analysis of GATA6 and CD44 expression in peritoneal hi hi + CD11b F4/80 macrophages obtained from control or Abx-treated mice. Cells were pregated on size, viability, and CD45 . Data were representative of hi hi three independent experiments. e Quantification of the proportion of peritoneal CD11b F4/80 macrophages (left). n = 6 (control) and 5 (Abx). An hi hi hi absolute number of peritoneal CD11b F4/80 macrophages (right). n = 4 (control) and 3 (Abx). f Representative images of the large F4/80 hi macrophages within intestinal injury in control and Abx mice. Scale bars, 50 μm. g Quantification of the number of large F4/80 macrophages in injury site at 6 and 24 h post-injury. n = 3 (Abx at 6 h) and 5 (control at 6 h, control and Abx at 24 h). Data represent mean ± SEM. NS not significant. P values were calculated with a two-tailed unpaired Student t-test (e, g). Source data are provided as a Source Data file. and other intestinal diseases is an intriguing but as yet untested maintain a body temperature of 37 °C throughout imaging. Exposed abdominal tissues were covered with saline-soaked gauze to prevent dehydration. hypothesis. Further studies are needed to address this question. In summary, this study demonstrates GATA6 peritoneal Spinning-disc confocal intravital microscopy (SD-IVM) and image analysis.A macrophages to be mobile immune players with a specific repair multichannel spinning-disk confocal microscope was used to image the mouse function in intestinal injury and inflammation. Our data indicate intestine. Image acquisition of the intestine was performed using an Olympus IX81 that enhancing their repair functions is a potential therapeutic inverted microscope (Olympus, Center Valley, PA), equipped with an Olympus approach for numerous intestinal diseases that include IBD. focus drive and a motorized stage (Applied Scientific Instrumentation, Eugene, OR) and fitted with a motorized objective turret equipped with 4×/0.16 UPLAN- SAPO, and 10×/0.40 UPLANSAPO objective lenses and coupled to a confocal light Methods path (WaveFx; Quorum Technologies) based on a modified CSU-10 head GFP/GFP Mice. C57BL/6 mice (#000664), LysM-eGFP mice, Cx3cr1 (CX3CR1-defi- (Yokogawa Electric Corporation, Tokyo, Japan). Cells of interest were visualized −/− cient) mice (#005582), and Nr4a1 mice (#006187) were obtained from The using fluorescently labeled antibodies, and fluorescent reporter mice. In some RFP/RFP Jackson Laboratory. Generation of Ccr2 (CCR2-deficient) mice have been experiments, necrotic cells were labeled by superfusion of the intestinal surface 37 GFP/+ RFP/+ previously described . Cx3cr1 Ccr2 mice were generated by crossing with 1 μM SYTOX Green. Laser excitation wavelengths of 491, 561, 642, and GFP/GFP RFP/RFP 37 Cx3cr1 Ccr2 mice with C57BL/6 mice . All mice were on a C57BL/6 730 nm (Cobolt, AB, Solna, Sweeden) were used in rapid succession together with background. Animals were maintained in a specific pathogen-free environment the appropriate band-pass filters (Semrock Inc., Rochester, NY). A back-thinned with ad libitum access to food and water. Mice were housed under standardized electron-multiplying charge-coupled device 512 × 512-pixel camera (Hamamatsu conditions of temperature (21–22 °C) and illumination (12/12 h light/dark cycle). Photonics) was used for fluorescence detection. Volocity software 6.1 (Perki- Mice of 8–12 weeks of age were used for experiments. Mice were gender-matched nElmer) was used to drive the confocal microscope and analysis of images. for experiments and experimental/control mice were bred separately. Mice were Acquired images were analyzed or exported as TIF images using Volocity euthanized by cervical dislocation after imaging or for tissue sampling. All software. The minimum threshold values were adjusted for each of the fluorescence experiments were approved by the Kumamoto University Ethics Review Com- channels to reduce the background. Exported images were imported to the Image J mittee for Animal Experimentation and were performed according to guidelines of 38 software package (NIH) for analysis . For quantification in the number of large the Institutional Animal Committee of Kumamoto University. hi F4/80 macrophages in the intestine, images were acquired for each mouse using a 10× objective and was evaluated using the “analyze and measure” command in hi Image J software. F4/80 cell whose length is larger than 15 μm was defined as Antibodies and reagents. Antibodies against CD11b (#17-0112-82; M1/70) (1:100 “large” macrophage. Quantification of SYTOX Green-positive dead cells in the dilution), CD31 (#12-0311-82; PECAM-1, 390) (1:100 dilution), CD45 (#45-0451- 7,38 intestinal injury area was performed using Image J as previously described . 82; 30-F11) (1:150 dilution), F4/80 (#12-4801-82; BM8) (1:100 dilution) were Autofluorescence induced by debris was excluded from the analysis. obtained from eBioscience. Antibodies against Ly6G (#127612; 1A8) (1:100 dilu- tion) were obtained from Biolegend. Antibody against CD44 (#553134; IM7) (1:150 dilution) was obtained from BD Biosciences. Antibody against GATA6 (#26452; Depletion of gut commensal bacteria. Gut commensal bacteria were depleted D61E4) (1:50 dilution) was obtained from Cell Signaling Technology. SYTOX using a method modified a protocol as previously described . Mice were provided Green (S7020) was obtained from Thermo Fisher Scientific. For intravital imaging, with ampicillin (1 g/L), vancomycin (0.5 g/L), neomycin (1 g/L), metronidazole we used a very low concentration at 2 μg per mouse of each antibody. Each anti- (1 g/L), and ciprofloxacin (0.2 g/L) in drinking water. All antibiotics were obtained body was diluted in PBS as appropriate. from Sigma-Aldrich. Antibiotics-mixed water was started from E14.5 and con- tinued until the experiments. In vivo treatment. Peritoneal macrophages were depleted by intraperitoneal admin- istration of 100 μL/mouse (0.69 mol/L) clodronate liposome (#CP-020-020; clo- Sterile inflammation induced by focal necrotic injury. Mice were anesthetized dronateliposomes.org, Vrije Universiteit, Netherlands) 4 days prior to the experiment. with isoflurane and a small midline laparotomy was made to exteriorize the colon. For The same dose of PBS-liposome (#CP-020-020) was used for the control experiment. imaging, a single focal injury was induced on the serosal surface of the colon to a depth For adhesion molecule blocking experiments, mice were administered either 50 μganti- of 80 μm using the tip of a heated 30-gauge needle mounted on an electrocautery CD44 (#CL8944AP; KM81, Cedarlane) monoclonal antibody or isotype control anti- device. Injury from the mucosal side was performed by puncturing the needle from the body 1 h prior to the injury. Apyrase treatment was performed by intraperitoneal opposite side, and the puncture hole was sutured closed. After the induction of thermal administration of 25 U apyrase (#A2230; Sigma-Aldrich). For blocking ATP receptor, injury, the abdominal incision was sutured closed and mice were allowed to recover for mice received 10 μg P2X7 antagonist (#A438079; R&D) intraperitoneally 1 h prior to imaging of later time points (2, 6, 24, 48 h). For sham experiments, mice underwent the theinjury. Hyaluronidasetreatment wasperformed by intraperitoneal administration of same surgical procedure, but no injury was induced. 100 U hyaluronidase (#H3506; Sigma-Aldrich) just after the intestinal injury. Each reagent was diluted in PBS as appropriate. Immunocytochemistry. Peritoneal macrophages were obtained from peritoneal exudates of mice, and erythrolysation was performed. Peritoneal macrophages were Preparation of the mouse intestine for intravital imaging. Mice were anesthe- cultured in low-glucose DMEM supplemented with 2% FBS and 1% penicillin/strep- tized by s.c. injection of 200 mg/kg ketamine (Bayer Animal Health) and 10 mg/kg tomycin. Cytospin smears were prepared by placing 1 mL (1.0 × 10 cells/ml) cultured xylazine (Bimeda-MTC). After anesthesia, the right jugular vein was cannulated to fluid in the cytospin funnel with filter paper being placed between the funnel and the administer fluorescent dyes and additional anesthetic. For intestinal imaging, a slide, followed by centrifugation at 800 rpm for 5 min resulting in the formation of a midline incision followed by a left lateral incision along the costal margin to the monolayered sheet of cells within a small circumference. After drying the sections, midaxillary line was performed to expose the intestine. To image the intestinal wall sheeted peritoneal macrophages were fixed with 1% paraformaldehyde for 5 min. from the serosal site, the oral side of the intestine was ligated with 5-0 silk string. Subsequently, the sections were incubated with blocking solution (1.0% bovine serum PBS (200 μL/20 mm of the colon) was introduced into the intestinal lumen using a albumin in TBS) for 20 min at room temperature (RT), then reacted with primary syringe with a 30-gauge needle, and the anal side of the intestine was ligated with antibodies. The sections probed with primary antibodies were incubated with secondary 5-0 silk string. Mice was placed in a left lateral position and the intestine was placed antibodies. Antibodies used in the immunofluorescence (IF) method are listed in on a glass coverslip and imaged if blood flow was normal. We defined “muscularis” Supplementary Table 2. IF sections were mounted using a mounting medium con- as 10–30 μm of depth, “submucosa” as 30–50 μm of depth, and “lamina propria” as taining 4′,6-diamidine-2′-phenylindole dihydrochloride (DAPI) (#SCR-38448; Dianova 50–80 μm of depth by imaging from serosa. Mice were placed on a heating pad to GmbH, Hamburg, Germany). NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 hi + Fig. 5 Peritoneal cavity macrophages facilitate intestinal repair. a Time-lapse images showing F4/80 large macrophage pulling off SYTOX green particles within the intestinal injury site. Elapsed time is shown. Scale bar, 20 μm. b Representative images of SYTOX green cells within intestinal injury +/+ RFP/RFP + site at 48 h post-injury in PBS-L- or CLL-treated CCR2 and CCR2 mice. Scale bars, 100 μm. c Quantification of SYTOX green area within injury in each group. n = 6/group. d Representative images of intestinal injury site 48 h after injury in PBS-L- or CLL-treated mice. Mice were administered an anti- CD31 (red) antibody intravenously to visualize vasculature. The white dashed line highlights the original injury border. Scale bars, 100 μm. e Quantification of revascularization (CD31 area within injury). n = 6/group. f Representative images of collagen (purple) within intestinal injury site 72 h post-injury in PBS-L- or CLL-treated mice. Scale bars, 50 μm. g Quantification of collagen deposition. n = 4/group. Data represent mean ± SEM. *p < 0.05, **p < 0.01, ****p < 0.0001. P values were calculated with two-tailed unpaired Student t-test (e, g) and one-way ANOVA followed by Tukey’s post hoc test (c). Source data are provided as a Source Data file. 10 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE Immunohistochemistry. Specimens were fixed at RT for from 24 to 48 h in 10% for antigen retrieval. After cooling down to RT, the sections were treated with 0.3% formalin (#134-10047; FUJIFILM Wako Pure Chemical, Corp., Osaka, Japan) and H O in methanol (30 min at RT) and were subsequently incubated with the 2 2 embedded in paraffin (#7810; Sakura Finetek Japan Co., Ltd., Tokyo, Japan). To blocking solution for 20 min at RT, following which the sections were reacted with prepare the paraffin-embedded cell block specimens, obtained peritoneal macro- primary antibodies. The sections were then incubated with horseradish peroxidase phages were fixed in 10% neutral buffered formalin. Then cells were suspended in (HRP)-labeled secondary antibodies. Antibodies used in the immunoenzyme 1% sodium alginate and solidified by the addition of 1 M calcium chloride. Finally, method are listed in Supplementary Table 2. Immunoreactions were visualized gelatinous specimens containing macrophages were embedded in paraffinin a using the diaminobenzidine (DAB) substrate kit (#425011, Nichirei Biosciences). routine manner. Specimens were then sectioned at 3 µm. Sections were pretreated For double-immunostaining, sections visualized with DAB in the first NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 11 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 Fig. 6 Peritoneal macrophages accumulate to the colon in DSS-induced colitis. a Representative stitch images of the colon in LysM-eGFP mice 5 days after the start of 4% DSS-containing water. Higher magnification of the indicated area was shown in right. Scale bars, 50 μm. b Quantification of the + hi number of large LysM F4/80 cells per FOV in LysM-eGFP mice that were treated with normal water or DSS-containing water. n = 4/group. c Schematic protocol for peritoneal cell transfer experiments (from LysM-eGFP mice to C57BL/6 mice). d Representative images of the colon in C57BL/6 mice that was administered peritoneal cells obtained from LysM-eGFP mice intraperitoneally 24 h before the imaging. Images were taken 5 days after the start + hi of 4% DSS-containing water. Anti-F4/80 antibody (red) was applied topically to the serosa. Arrows indicate LysM F4/80 cells. Scale bars, 50 μm. + hi e Quantification of the number of large LysM F4/80 cells per FOV in WT mice that were treated with normal water or DSS-containing water. Mice were pretreated by intraperitoneal administration of peritoneal cells from LysM-eGFP mice 24 h before the imaging. n = 4/group. f Representative images of the GFP/+ RFP/+ colon in Cx3cr1 Ccr2 mice 5 days after the start of 4% DSS-containing water. Anti-F4/80 antibody (purple) was applied topically. Scale bars, 50 μm. Data were representative of three independent experiments. g Immunofluorescence staining of the colon harvested from control or 5 days after the start of 4% DSS-containing water with hyaluronic acid binding protein (HABP). Scale bars, 500 μm. h Quantification of the HABP-positive area (%). n = 5 (normal water) and 10 (DSS water). i Representative images of the colonic muscularis in control and Abx mice 5 days after the start of 4% DSS-containing hi water. Anti-F4/80 antibody (red) was applied topically to the serosa. Scale bars, 50 μm. j Quantification of the number of large F4/80 cells per FOV. n = 5/group. Data represent mean ± SEM. *p < 0.05, ****p < 0.0001. P values were calculated with a two-tailed unpaired Student t-test (b, e, h, j). Source data are provided as a Source Data file. immunostaining were then used to perform the second immunostaining, with the run using a flow cytometer (FACSCanto; BD Biosciences) and analyzed using reaction visualized using HistoGreen (#E109; LINARIS Biologische Produkte FlowJo software (Tree Star). GmbH, Dossenheim, Germany). Sections were mounted using malinol (Muto Pure Chemicals Co., Ltd, Tokyo, Japan). Measurement of colon tissue chemokine concentration. Quantification of concentrations of chemokines was performed using the validated Luminex bead- Image processing, cell counting, and area measurement for immunostained based assay from R&D Systems (Minneapolis, MN) according to the manu- facturer’s instructions. Briefly, a 20-mm-long colon tissue sample was placed into sections. Immunostained sections were photographed using a microscope (BX51, 1 mL PBS, and they were mechanically disrupted. The mixture was then cen- Olympus Corporation, Tokyo, Japan). To quantify immunostaining of Gr-1 by cell trifuged at 20,124 × g for 5 min three times and the supernatant was transferred to counting, two pathologists (D.Y. and Y.K.), blinded to mouse information eval- uated sections. To quantify immunostaining by calculating areas of immunostained a 0.22 μm PVDF DuraPore centrifugal filter (EMD Millipore, Billerica, MA) to remove any particles from the solution. Filtered samples were then incubated with sections, we used BZ-X800 (Keyence corp., Osaka, Japan). Immunostained sections were photographed and then entire fields were reconstructed into one picture. a capture bead cocktail on a 96-well plate in the dark for 2 h at room temperature. After incubation, wells were washed with wash buffer, incubated with a biotin IBA1 and CD3 area proportions of the colonic mucosa and muscularis for each antibody cocktail for 1 h next before another round washing, and incubated for entire field were measured, respectively. HABP area proportion for the entire field another 30 min with Streptavidin-PE. Following washes, the plate was read using a was also measured. Luminex 200 apparatus (Applied Cytometry Systems, UK) and analyzed with StarStation V.2.3 (Applied Cytometry Systems, UK). Total protein concentration of Peritoneal cell transfers. Peritoneal lavage was performed as described before filtered samples was measured using Bio-Rad Protein Assay (Bio-Rad Labora- using sterile PBS . Cells were washed with cold PBS twice and resuspended in tories). Results were normalized against the amount of total protein extracted from 100 μL PBS. 5 × 10 cells isolated from LysM-eGFP mice were directly transferred the colon tissues or the weight of the colon tissue. into naïve mice. Both intraperitoneal and intravenous transfers were performed 1 h prior to injury induction. In the 4% DSS-induced colitis model, peritoneal cells Bacterial DNA extraction and amplification. DNA was extracted from fecal pellet isolated from LysM-eGFP mice were transferred 4 days after the start of DSS- using a QIAamp Fast DNA Stool Mini Kit (QIAGEN) according to manufacturer’s containing water and the colon was imaged 24 h later. instructions and 16 S rRNA genes were amplified and sequenced using an Illumina MiSeq (Illumina, San Diego, CA, U.S.A.) as described previously . The 16 S rRNA SYTOX green staining of intestinal content. Bacterial load of intestinal contents operational taxonomic units (OTUs) were selected from the combined reads using was measured as previously described . Briefly, a fecal pellet was collected from a de novo OTU picking protocol clustered at 97% identity using the Quantitative control or an Abx-treated mouse and each sample was added 500 μL 4% paraf- Insights Into Microbial Ecology (QIIME) pipeline software. To investigate the ormaldehyde and mixed. After the incubation for 30 min at room temperature, bacterial diversity of each sample, the number of OTUs, Chao 1, and Shannon were 100 μLof fixed content was resuspended in 800 μL sterile PBS and 1 μL of SYTOX calculated, and a rarefaction curve was generated using QIIME. Differences of green (100 μg/mL) was added. After the incubation for 30 min in the dark, samples microbial communities were evaluated using phylogeny-based unweighted or were centrifuged and 50 μL of supernatant was spread on a slide. The image was weighted UniFrac distance matrices. Principal coordinate analysis (PCoA) graph recorded by microscopy in a green fluorescence channel using a 10× objective. and Unweighted Pair Group Method with Arithmetic mean (UPGMA) tree graph were depicted using QIIME. Detection of intestinal collagen by multi-photon microscopy. Intestinal fibrillar collagen was imaged using second-harmonic generation (SHG). In brief, focal DSS colitis model. Mice were given 4.0% DSS (molecular mass, 36,000–50,000, injured colons were removed, maintained in cold PBS, and imaged using a BX61WI MP Biomedicals, LLC) in the drinking water for continuous 5 days. Body weight upright microscope and FV1000MPE (Olympus, Tokyo, Japan) laser-scanning and disease activity score was observed daily in PBS-L treated control and CLL- microscope system equipped with a MaiTai HP Deep See femtosecond-pulsed laser treated mice. Disease activity index was evaluated according to the previous (Spectra-Physics, Santa Clara, CA, USA). Colonic tissue was visualized using a report and was described as follows; (a) general appearance: normal = 0; combination of multi-photon fluorescence and SHG using 890 nm excitation . piloerection = 1; lethargy and piloerection = 2; motionless, sickly = 4; (b) weight The signal was detected by the external non-descanned detectors (505 nm mirror loss: no change = 0; < 5% = 1; 6–10% = 2; 11–20% = 3; > 20% = 4; (c) feces and band-pass emission filters at 465–485 nm for SHG). Z-stacks were recorded at consistency: normal = 0; pasty, semi-formed = 2; liquid, sticky, or unable to 0.5 μm intervals with an Olympus XLPLN 25X WMP (water immersion; numerical defecate after 5 min = 4; and (d) rectal bleeding: no blood = 0; visible blood in aperture, 1.05; working distance 2.0 mm) objective lens. rectum = 2; visible blood on fur = 4. At 5 days after the start of DSS-containing water, colon length was measured, and the pathological findings were evaluated in a blinded fashion using a previously reported scoring system: (a) inflammatory Flow cytometry. After mice were anesthetized, colon or peritoneal lavage were cell infiltrate (severity, extent): mild, mucosa = 1; moderate, mucosa and sub- collected and placed in PBS on ice. Colonic cells were obtained from intestinal mucosa = 2; marked, transmural = 3; (b) intestinal architecture (epithelial changes, biopsies of the uninjured or injured area. After the homogenization, single-cell mucosal architecture): focal erosions = 1; erosions ± focal ulcerations = 2; extended suspensions were generated by mechanical disruption through a 40-µm nylon ulcerations ± granulation tissue ± pseudopolyps = 3. mesh (BD Bioscience). Residual red blood cells were lysed using ACK lysing buffer (Invitrogen). The cells were blocked using anti-CD16/32 antibody (#BE0307; 2.4G2 clone; Bio X Cell) (1:100 dilution) for 30 min. Then, cells were stained for 30 min Statistical analysis. Data were expressed as mean ± SEM. Unpaired Student t-test or with antibodies for specified markers. Nonviable cells were identified using viability Mann–Whitney U-test was used to compare between two groups as appropriate. One- TM dye efluor 780 (#65-0865-14; eBioscience) (1:1000 dilution) or Ghost Dye Red way ANOVA was used to compare more than two groups, followed by Tukey’spost 710 (#13-0871-T100; TONBO biosciences) (1:1000 dilution). A Foxp3 staining hoc test. Percent body weight change or disease activity index in PBS-L- or CLL-treated buffer set (eBioscience) was used for intracellular GATA6 staining. Samples were mice with 4% DSS colitis were compared by two-way repeated-measures ANOVA. 12 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE Fig. 7 Peritoneal macrophages exert a tissue repair function in DSS-induced colitis. a Body weight change of PBS-L treated control and CLL-treated mice in 4% DSS-induced colitis. n = 5 (normal water group treated by PBS-L or CLL) and 20 (DSS water group treated by PBS-L or CLL). Data were pooled from three independent experiments. b Disease activity index of PBS-L treated control and CLL-treated mice in 4% DSS-induced colitis. n = 5 (normal water group treated by PBS-L or CLL) and 10 (DSS water group treated by PBS-L or CLL). c Macroscopic findings of the colon 5 days after the start of 4% DSS- containing water (left) and quantification of colon length (right). n = 5 (normal water group treated by PBS-L or CLL) and 10 (DSS water group treated by PBS-L or CLL). d Representative H&E staining of the colon 5 days after the start of 4% DSS-containing water and e quantification of histomorphological scores. n = 5 (normal water group treated by PBS-L or CLL) and 10 (DSS water group treated by PBS-L or CLL). Scale bars, 100 μm. f Schematic protocol for peritoneal cell (PC) transfer experiments in DSS-induced colitis. g Body weight change and h disease activity index of PBS-treated control and PC-treated mice in 4% DSS-induced colitis. n = 5/group. i Macroscopic findings of the colon 5 days after the start of 4% DSS-containing water (left) and quantification of colon length (right). n = 5/group. j Representative H&E staining of the colon 5 days after the start of 4% DSS-containing water and k quantification of histomorphological scores. n = 5/group. Scale bars, 100 μm. Data represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. P values were calculated with two-tailed unpaired Student t-test (c, i), Mann–Whitney U-test (e, k), and two-way repeated-measures ANOVA (a, b, g, h). Source data are provided as a Source Data file. NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 13 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 Fig. 8 GATA6 peritoneal cavity macrophages infiltrate an injured intestine via a direct peritoneal route and contribute tissue repair. A scheme + + + showing the mechanism of GATA6 peritoneal macrophage recruitment and bloodstream-derived CCR2 monocyte to CX3CR1 monocyte/macrophage conversion in intestinal injury/inflammation. Experimental findings were reproduced at least twice to ensure consistency. A p value 9. Okabe, Y. & Medzhitov, R. Tissue-specific signals control reversible program <0.05 was considered statistically significant. All tests were two-tailed. All statistical of localization and functional polarization of macrophages. Cell 157, 832–844 analyses were performed using GraphPad Prism v8.0 software (GraphPad Software Inc., (2014). La Jolla, CA). 10. Rosas, M. et al. The transcription factor Gata6 links tissue macrophage phenotype and proliferative renewal. Science 344, 645–648 (2014). 11. Huber, S., Hoffmann, R., Muskens, F. & Voehringer, D. Alternatively activated Reporting summary. Further information on research design is available in the Nature macrophages inhibit T-cell proliferation by Stat6-dependent expression of Research Reporting Summary linked to this article. PD-L2. Blood 116, 3311–3320 (2010). 12. Martinez, F. O. & Gordon, S. The M1 and M2 paradigm of macrophage Data availability activation: time for reassessment. F1000Prime Rep. 6, 13 (2014). The sequence data generated in this study have been deposited in the NCBI repository 13. Hooper, L. V. & Macpherson, A. J. Immune adaptations that maintain GenBank Nucleotide under accession numbers OK665944-OK666378. The first is homeostasis with the intestinal microbiota. Nat. Rev. Immunol. 10, 159–169 accessible and the others can be accessed by editing the accession number in the (2010). hyperlink. Further data that support the findings of this study are available from the 14. Brown, E. M., Sadarangani, M. & Finlay, B. B. The role of the immune system corresponding author upon reasonable request. Source data are provided with this paper. in governing host-microbe interactions in the intestine. Nat. Immunol. 14, 660–667 (2013). 15. Niess, J. H. & Adler, G. Enteric flora expands gut lamina propria CX3CR1+ Received: 28 November 2020; Accepted: 29 November 2021; dendritic cells supporting inflammatory immune responses under normal and inflammatory conditions. J. Immunol. 184, 2026–2037 (2010). 16. Bain, C. C. & Jenkins, S. J. The biology of serous cavity macrophages. Cell Immunol. 330, 126–135 (2018). 17. Buechler, M. B. et al. A stromal niche defined by expression of the References transcription factor WT1 mediates programming and homeostasis of cavity- 1. Mowat, A. M., Scott, C. L. & Bain, C. C. Barrier-tissue macrophages: resident macrophages. Immunity 51, 119–130 e115 (2019). functional adaptation to environmental challenges. Nat. Med. 23, 1258–1270 18. Ghosn, E. E. et al. Two physically, functionally, and developmentally distinct (2017). peritoneal macrophage subsets. Proc. Natl Acad. Sci. USA 107, 2568–2573 (2010). 2. Cerovic, V., Bain, C. C., Mowat, A. M. & Milling, S. W. Intestinal macrophages 19. Gautier, E. L. et al. Gene-expression profiles and transcriptional regulatory and dendritic cells: what’s the difference? Trends Immunol. 35, 270–277 pathways that underlie the identity and diversity of mouse tissue (2014). macrophages. Nat. Immunol. 13, 1118–1128 (2012). 3. Zigmond, E. & Jung, S. Intestinal macrophages: well educated exceptions from 20. Louis, C. et al. Specific contributions of CSF-1 and GM-CSF to the dynamics the rule. Trends Immunol. 34, 162–168 (2013). of the mononuclear phagocyte system. J. Immunol. 195, 134–144 (2015). 4. Bain, C. C. et al. Constant replenishment from circulating monocytes 21. Cassado Ados, A., D’Imperio Lima, M. R. & Bortoluci, K. R. Revisiting mouse maintains the macrophage pool in the intestine of adult mice. Nat. Immunol. peritoneal macrophages: heterogeneity, development, and function. Front. 15, 929–937 (2014). Immunol. 6, 225 (2015). 5. Bain, C. C. et al. Resident and pro-inflammatory macrophages in the colon 22. Kim, K. W. et al. MHC II+ resident peritoneal and pleural macrophages rely represent alternative context-dependent fates of the same Ly6Chi monocyte on IRF4 for development from circulating monocytes. J. Exp. Med. 213, precursors. Mucosal Immunol. 6, 498–510 (2013). 1951–1959 (2016). 6. Honda, M. et al. Perivascular localization of macrophages in the intestinal 23. Barth, M. W., Hendrzak, J. A., Melnicoff, M. J. & Morahan, P. S. Review of the mucosa is regulated by Nr4a1 and the microbiome. Nat. Commun. 11, 1329 macrophage disappearance reaction. J. Leukoc. Biol. 57, 361–367 (1995). (2020). 24. Stengel, S. et al. Peritoneal level of CD206 associates with mortality and an 7. Wang, J. & Kubes, P. A reservoir of mature cavity macrophages that can inflammatory macrophage phenotype in patients with decompensated rapidly invade visceral organs to affect tissue repair. Cell 165, 668–678 (2016). cirrhosis and spontaneous bacterial peritonitis. Gastroenterology 158, 8. Gautier, E. L. et al. Gata6 regulates aspartoacylase expression in resident 1745–1761 (2020). peritoneal macrophages and controls their survival. J. Exp. Med. 211, 25. Zhang, N. et al. Expression of factor V by resident macrophages boosts host 1525–1531 (2014). defense in the peritoneal cavity. J. Exp. Med. 216, 1291–1300 (2019). 14 NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27614-9 ARTICLE 26. Deniset, J. F. et al. Gata6(+) pericardial cavity macrophages relocate to the licensed under a Creative Commons Attribution 3.0 Unported License (https:// injured heart and prevent cardiac fibrosis. Immunity 51, 131–140 e135 (2019). creativecommons.org/licenses/by/3.0/). The source of the mouse icons used in Figs. 2h, 27. Gautier, E. L., Ivanov, S., Lesnik, P. & Randolph, G. J. Local apoptosis 4a, 6c, 7f, and Supplementary Figs. 6a, 6d, 10a are Free vector graphics on Pixabay. mediates clearance of macrophages from resolving inflammation in mice. Content on Pixabay is licensed under Pixabay License (https://pixabay.com/service/ Blood 122, 2714–2722 (2013). terms/#license). We also thank Jeremy Allen, Ph.D., from Edanz Group (https://en- 28. Atarashi, K. et al. Treg induction by a rationally selected mixture of Clostridia author-services.edanzgroup.com/ac) for editing a draft of this manuscript. M.H. is strains from the human microbiota. Nature 500, 232–236 (2013). supported by the Research Fellowship of the Uehara Memorial Foundation and grants 29. Atarashi, K. et al. Induction of colonic regulatory T cells by indigenous from the Ministry of Education, Culture, Sports, Sciences and Technology of Japan Clostridium species. Science 331, 337–341 (2011). (KAKENHI 19H03716). 30. Constantinides, M. G. et al. MAIT cells are imprinted by the microbiota in early life and promote tissue repair. Science 366, eaax6624 (2019). Author contributions 31. Furusawa, Y. et al. Commensal microbe-derived butyrate induces the M.H. designed and performed experiments, analyzed data, and wrote the manuscript. differentiation of colonic regulatory T cells. Nature 504, 446–450 (2013). M.K. performed experiments, analyzed data, and wrote the manuscript. D.Y. performed 32. Ruiz-Alcaraz, A. J. et al. Characterization of human peritoneal monocyte/ experiments and analyzed pathological specimens. Y.K. and T.H. provided material macrophage subsets in homeostasis: phenotype, GATA6, phagocytic/oxidative support and reviewed the manuscript. activities and cytokines expression. Sci. Rep. 8, 12794 (2018). 33. Ng, S. C. et al. Worldwide incidence and prevalence of inflammatory bowel disease in the 21st century: a systematic review of population-based studies. Competing interests Lancet 390, 2769–2778 (2018). The authors declare no competing interests. 34. Honda, M. & Kubes, P. Neutrophils and neutrophil extracellular traps in the liver and gastrointestinal system. Nat. Rev. Gastroenterol. Hepatol. 15, Additional information 206–221 (2018). Supplementary information The online version contains supplementary material 35. Seo, D. H. et al. Interleukin-33 regulates intestinal inflammation by modulating available at https://doi.org/10.1038/s41467-021-27614-9. macrophages in inflammatory bowel disease. Sci. Rep. 7, 851 (2017). 36. Liu, T. et al. Treatment of dextran sodium sulfate-induced experimental colitis Correspondence and requests for materials should be addressed to Masaki Honda. by adoptive transfer of peritoneal cells. Sci. Rep. 5, 16760 (2015). 37. Dal-Secco, D. et al. A dynamic spectrum of monocytes arising from the in situ Peer review information Nature Communications thanks the anonymous reviewer(s) for reprogramming of CCR2+ monocytes at a site of sterile injury. J. Exp. Med. their contribution to the peer review of this work. Peer reviewer reports are available. 212, 447–456 (2015). 38. Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH image to imageJ: 25 Reprints and permission information is available at http://www.nature.com/reprints years of image analysis. Nat. Methods 9, 671–675 (2012). 39. McDonald, B. et al. Intravascular danger signals guide neutrophils to sites of Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in sterile inflammation. Science 330, 362–366 (2010). published maps and institutional affiliations. 40. Macpherson, A. J., Geuking, M. B., Kirundi, J., Collins, S. & McCoy, K. D. Gnotiobiotic and Axenic Animals. In Encyclopedia of Microbiology (ed. Moselio, S.) 237–246 (Elsevier, 2009). Open Access This article is licensed under a Creative Commons 41. Gailhouste, L. et al. Fibrillar collagen scoring by second harmonic microscopy: Attribution 4.0 International License, which permits use, sharing, a new tool in the assessment of liver fibrosis. J. Hepatol. 52, 398–406 (2010). adaptation, distribution and reproduction in any medium or format, as long as you give 42. Omori, M. et al. Fecal microbiome in dogs with inflammatory bowel disease and intestinal lymphoma. J. Vet. Med. Sci. 79, 1840–1847 (2017). appropriate credit to the original author(s) and the source, provide a link to the Creative 43. Taylor, B. C. et al. TSLP regulates intestinal immunity and inflammation in Commons license, and indicate if changes were made. The images or other third party mouse models of helminth infection and colitis. J. Exp. Med. 206, 655–667 material in this article are included in the article’s Creative Commons license, unless (2009). indicated otherwise in a credit line to the material. If material is not included in the 44. Erben, U. et al. A guide to histomorphological evaluation of intestinal article’s Creative Commons license and your intended use is not permitted by statutory inflammation in mouse models. Int J. Clin. Exp. Pathol. 7, 4557–4576 (2014). regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. Acknowledgements We would like to thank Servier for providing Servier Medical Art, which was used for the © The Author(s) 2021 creation of figures (Fig. 8 and Supplementary Fig. 7a). Servier Medical Art by Servier is NATURE COMMUNICATIONS | (2021) 12:7294 | https://doi.org/10.1038/s41467-021-27614-9 | www.nature.com/naturecommunications 15

Journal

Nature CommunicationsSpringer Journals

Published: Dec 15, 2021

There are no references for this article.