Infiltrating CCR2+ monocytes and their progenies, fibrocytes, contribute to colon fibrosis by inhibiting collagen degradation through the production of TIMP-1

Intestinal fibrosis is a serious complication in inflammatory bowel disease (IBD). Despite the remarkable success of recent anti-inflammatory therapies for IBD, incidence of intestinal fibrosis and need for bowel resection have not significantly changed. To clarify the contribution of haematopoietic-derived cells in intestinal fibrosis, we prepared bone marrow (BM) chimeric mice (chimeras), which were reconstituted with BM cells derived from enhanced green fluorescent protein (EGFP)-transgenic mice or CC chemokine receptor 2 (CCR2)-deficient mice. After 2 months of transplantation, BM chimeras were treated with azoxymethane/dextran sodium sulphate. During chronic inflammation, CCR2+ BM-derived monocyte and fibrocyte infiltration into the colon and CC chemokine ligand 2 production increased, leading to colon fibrosis in EGFP BM chimeras. In CCR2-deficient BM chimeras, monocyte and fibrocyte numbers in the colonic lamina propria significantly decreased, and colon fibrosis was attenuated. In colon tissue, mRNA expression of tissue inhibitor of metalloproteinase (TIMP)-1 but not of collagen I, transforming growth factor-β1 or matrix metalloproteinases was significantly different between the two chimeras. CCR2+ monocytes and fibrocytes showed high Timp1 mRNA expression. Our results suggest that infiltrating CCR2+ monocytes and their progenies, fibrocytes, promote colon fibrosis by inhibiting collagen degradation through TIMP-1 production.


Repeated administration of DSS leads to chronic inflammation and fibrosis in the colon.
To facilitate the observation of BM-derived cells infiltrating the injured colon, we generated EGFP BM chimeras. At 2 months after BM transplantation, the BM chimeras were treated with AOM/DSS. Chronic colitis was induced by repeated administration of 1% DSS. Disease activity index (DAI) increased after the administration of 1% DSS, and the score was higher in the third cycle than in the first cycle (Fig. 1A). Treated BM chimeras were sacrificed on day 17 (recovery 1) from the start of DSS administration and on day 59 (recovery 3). The colon length was significantly shorter, and the colon weight to length ratio was significantly higher in mice after the third cycle of DSS compared with the untreated control BM chimeras (Fig. 1B,C).
Histological analysis of colon sections with haematoxylin and eosin (HE) stains revealed areas with severe mucosal and submucosal inflammatory cell infiltration and mucosal thickening but not dysplasia or cancer on day 59 (recovery 3) (Fig. 1D). Colon sections were stained with Sirius red and Masson's trichrome to investigate the colon fibrosis in the chronic DSS-treated mice. Collagen deposition, which was shown as a percentage of Sirius red-positive red areas and Masson's trichrome-positive blue areas, significantly increased in the colon tissue of mice after the third cycle of DSS compared with the untreated control mice (Fig. 1D-F). The distribution of type I and type III collagen was determined in Sirius red-stained colon sections using a polarised light microscope, in which type I and type III collagen were observed as red and green colours, respectively (Fig. 1G). Although the percentage of type I collagen significantly increased in the colon tissue of mice after the third cycle of DSS compared with the untreated control BM chimeras, that of type III collagen did not change during the development of colitis. (A) Enhanced green fluorescent protein bone marrow (EGFP BM) chimeras were treated with a single intraperitoneal injection of AOM, followed by three cycles of 1% DSS for seven days in drinking water. Control mice received regular drinking water. The disease activity index (DAI) scores in DSS-treated (n = 11) and control mice (n = 3) were monitored three times per week. (B,C) The colon length and weight to length ratio were measured on days 17 (recovery 1) and 59 (recovery 3) from the start of DSS. The sample size for each group was n = 6. (D) Paraffin-embedded colon sections obtained from untreated (control) mice (n = 3) and chronically DSS-treated (recovery 3) mice (n = 3) were stained with haematoxylin and eosin (HE; upper), Sirius red (middle) and Masson's trichrome (lower). Original magnification, ×100. (E,F) Quantification of Sirius red-and Masson's trichrome-positive areas. (G) Sections stained with Sirius red were analysed using polarised microscopy to quantify type I and type III collagen. Quantification of type I and type III collagen-positive areas. The histograms show the mean percentage of the staining area per total colon area. The experiments were performed at least twice, yielding similar results. Data are expressed as the mean ± SD. * p < 0.05, ** p < 0.01 and *** p < 0.001 by Kruskal-Wallis test followed by Dunn's multiple comparison test.
immunohistochemical staining of frozen colon sections from EGFP BM chimeras was performed to investigate the infiltration of BM-derived cells into the colon. EGFP + cells were mainly detected in the mucosa of the colon in untreated control mice and the mucosa and submucosa of the colon in mice after the third cycle of DSS treatment. The number of EGFP + cells and EGFP + CD45 + haematopoietic cells increased two-fold after the third cycle of DSS treatment compared with the untreated control mice ( Fig. 2A-C). EGFP + CD45 + procollagen I + , EGFP + CD45 − procollagen I + and EGFP + α-SMA + cells are considered BM-derived fibrocytes, BM-derived fibroblast-like cells including myofibroblasts and BM-derived myofibroblasts, respectively. Although numbers of these three fractions in the untreated control mice were low, EGFP + CD45 + procollagen I + fibrocytes and EGFP + α-SMA + cells significantly increased after the third cycle of DSS ( Fig. 2D-I). However, a vast majority of α-SMA + myofibroblasts in the colonic lamina propria (LP) were resident cells, and there was no change in the total number (BM-derived EGFP + α-SMA + and resident EGFP − α-SMA + cells) between the untreated control mice and mice after the third cycle of DSS treatment (Fig. 2I).
BM-derived monocytes and fibrocytes significantly increased in the inflamed colon. Next, we examined which fraction of the BM-derived cells infiltrated the colonic LP after chronic DSS treatment using flow cytometry ( Fig. 3A-F). After the third DSS treatment, CD45 + F4/80 − Ly6G − CD11c − CD11b + Ly6C + monocytes and CD45 + CD11b + Col I + fibrocytes increased 16.6-fold and 3.8-fold, respectively, but other fractions increased by less than 2.5-fold in the colonic LP of EGFP BM chimeras (Fig. 4A). There was no difference in the number of CD45 − Col I + cells, which include fibroblasts and myofibroblasts, between the untreated control mice and mice after the third cycle of DSS treatment (Fig. 4B).
Most CD45 − Col I + cells were negative for EGFP, suggesting the origin from resident cells (Fig. 3F). On the other hand, fibrocytes detected in the colonic LP were mostly derived from BM haematopoietic cells because the frequency of EGFP + cells was 95%-97% in the untreated control and chronically treated mice (data not shown). The number of EGFP + CD45 + CD11b + Col I + fibrocytes in the colonic LP of chronically treated EGFP BM chimeras significantly increased in association with the development of colon inflammation and were markedly reduced 2 months after cessation of DSS treatment (chronic phase). The number of EGFP + CD45 − Col I + cells was much lower than the number of EGFP + CD45 + CD11b + Col I + fibrocytes and did not change during the development of colitis (Fig. 4C).
The chemokine receptor expression of monocytes and macrophages in the colonic LP and fibrocytes in the colonic LP as well as PB and chemokine production in the colon were analysed to investigate which chemokine influences the infiltration of BM-derived monocyte-lineage cells into the colonic LP. Of note, monocyte-lineage cells express chemokine receptors, such as CCR2, CX3CR1 and CXCR4 [49][50][51] . Therefore, the expression of CCR2 and CX3CR1 was examined using CCR2 RFP/+ CX3CR1 GFP/+ hybrid mice, in which CCR2 + cells express red fluorescent protein and CX3CR1 + cells express green fluorescent protein 52 . Additionally, CXCR4 expression was assessed using an antibody against CXCR4 during the first cycle of DSS treatment.
First, the number of each cell type was counted on days 0, 3, 9 and 16 after DSS treatment initiation. The total monocytes and fibrocytes in the colonic LP increased in a time-dependent manner (days 9-16) and peaked at day 16. Macrophages peaked on day 9 and subsequently declined to approximately half of their peak value on day 16. Fibrocytes in PB increased in a time-dependent manner (days 3-9) and were markedly reduced from day 9 to day 16 ( Fig. 5A-D).
In monocytes, CXCR4, CCR2 and CX3CR1 were highly expressed before DSS treatment. Although the expression level of both CXCR4 and CCR2 did not change, CX3CR1 low/− cells increased with the development of colitis and approximately half of the monocytes were negative for CX3CR1 on day 16 (Fig. 5A). In comparison with monocytes, macrophages expressed a higher level of CX3CR1 and a similar level of CXCR4 before DSS treatment. CCR2 was bimodally expressed. Although the expression level of CXCR4 did not change, almost all macrophages became highly positive for CCR2 with the development of colitis. The percentage of CX3CR1 − cells in macrophages transiently increased on day 9 after DSS treatment initiation but decreased from days 9 to 16 (Fig. 5B). The expression pattern of CXCR4 and CX3CR1 on fibrocytes in the colonic LP was similar to that on monocytes and macrophages before DSS treatment. The expression level of CCR2 on fibrocytes was lower than that on monocytes and macrophages. The CXCR4 expression did not change, however, CX3CR1 − fibrocytes gradually increased with the development of colitis. The CCR2 expression became bimodal (CCR2 − and CCR2 + ) after DSS treatment (Fig. 5C). Fibrocytes in the PB only expressed CXCR4, but the expression level was lower than that on fibrocytes in the colonic LP. The expression of both CCR2 and CX3CR1 remained almost negative even after DSS treatment (Fig. 5D). These results indicate the possibility that fibrocytes in the colonic LP are composed of two types of cells: CCR2 + cells, which are phenotypically similar to monocytes and macrophages in the colonic LP, and CCR2 − cells, which are probably fibrocytes in the PB.
Chemokine production in colon tissue before and after DSS treatment was analysed by quantitative real-time reverse transcription polymerase chain reaction (RT-PCR). The mRNA expression of Ccl2 significantly increased in a time-dependent manner after chronic injury (Fig. 5E). Although the expression of Cxcl12 significantly increased after the first cycle of DSS treatment, it subsequently declined during further treatment. The expression of Cx3cl1 was not altered during the experiment.
Previous reports have demonstrated that uptake of secreted Col I by haematopoietic cells accounted for the fibrocyte population, and the presence of intracellular Col I does not necessarily indicate Col I production by fibrocytes [53][54][55] . Therefore, we sorted colonic LP cells, which were isolated from pooled colons of seven non-transplanted C57BL/6J-Ly5.1 mice following three cycles of treatment with 1% DSS, into four populations of CD45 + CD11b + cells based on Col I and CCR2 staining and examined Col1a1 expression (Fig. 6F). Col1a1 was detected in (a) Col I + CCR2 + cells and (b) Col I + CCR2 − cells as well as (c) Col I − CCR2 + cells and the expression level was higher in Col I + CCR2 + and Col I + CCR2 − cells than in Col I − CCR2 + cells.
Col I − CCR2 + cells were mainly composed of monocytes and macrophages. These results suggest that some CCR2 + monocytes and macrophages in the colonic LP can produce Col I and are precursors of CCR2 + Col I + fibrocytes. Therefore, fibrocytes in the colonic LP are thought to consist of two subsets: CCR2 + infiltrating monocyte-derived fibrocytes and CCR2 − circulating fibrocytes.
CCR2 deficiency attenuates the development of colon fibrosis. Based on the above observations, it was considered that the CCL2/CCR2 axis is essential to recruit Ly6C + CCR2 + monocytes to the injured colon and induce colon fibrosis via the accumulation of CCR2 + fibrocytes. Therefore, we prepared WT BM and CCR2 RFP/ RFP BM chimeras and treated them with three cycles of DSS treatment. Although a significant increase in the DAI score was observed in WT BM and CCR2 RFP/RFP BM chimeras, no difference was found between both mice (Fig. 7A). The colon length decreased in both BM chimeras after DSS treatment. Shortening of the colon length was dampened in the CCR2 RFP/RFP BM chimeras compared with that in the WT BM chimeras; however, the colon length after the third cycle of DSS treatment was shorter than that before DSS treatment in the CCR2 RFP/RFP BM www.nature.com/scientificreports www.nature.com/scientificreports/ chimeras (Fig. 7B). The histological inflammation score increased after chronic DSS treatment in the WT BM and CCR2 RFP/RFP BM chimeras and was not significantly different between both BM chimeras (Fig. 7C,D).
Colon fibrosis was assessed by Sirius red staining. Chronic DSS treatment led to a mild colonic inflammation and submucosal oedema with extensive deposition of collagen in WT BM chimeras (Fig. 7C). In contrast, CCR2 RFP/RFP BM chimeras showed a significant reduction of colon fibrosis (Fig. 7C,E). The percentage of type I collagen but not that of type III collagen significantly decreased in the CCR2 RFP/RFP BM chimeras compared with the WT BM chimeras after the third DSS treatment (Fig. 7F).
After chronic DSS administration, the number of monocytes in the colonic LP markedly increased in WT BM chimeras compared with untreated control conditions. However, there were almost no monocytes in both untreated and chronically treated CCR2 RFP/RFP BM chimeras (Fig. 7G). Although a marked increase of fibrocytes in the colonic LP was seen in chronically treated WT BM chimeras compared with corresponding control mice, no difference was observed in the number of fibrocytes between untreated and chronically treated CCR2 RFP/ RFP BM chimeras (Fig. 7H). The number of monocytes and fibrocytes in the colonic LP of chronically treated CCR2 RFP/RFP BM chimeras were extremely lower than those in the colonic LP of WT BM chimeras. However, there was no difference in the number of fibroblasts in the colonic LP before and after chronic DSS treatment in www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/ both BM chimeras (Fig. 7I). These results indicate that targeted deletion of CCR2 in BM-derived cells attenuated colon fibrosis by inhibiting the accumulation of monocytes and fibrocytes in the inflamed colon.

BM-derived CCR2 + monocytes and fibrocytes promote colon fibrosis through the production of TIMP-1. Tissue fibrosis occurs when collagen synthesis surpasses collagen degradation, which is regulated
by MMPs and TIMPs 7,34,35,40 . Our present results demonstrate that BM-derived CCR2 + monocytes and fibrocytes play an essential role in the development of colon fibrosis. To clarify how these cells contribute to colon fibrosis, we analysed the expression of several genes involved in collagen synthesis and collagen degradation within DSS-treated colon tissues of both EGFP BM and CCR2 RFP/RFP BM chimeras.
First, we analysed the expression of Col1a1, Tgfb1, Timp1, Mmp1a, Mmp8 and Mmp13 in the rectal tissues. Although the expressions of Col1a1 and Tgfb1 were upregulated after chronic DSS treatment in EGFP BM and CCR2 RFP/RFP BM chimeras, there were no significant differences between both chimeras (Fig. 8A). A significant reduction of colon fibrosis, which was observed in CCR2 RFP/RFP BM chimeras, was not caused by the inhibition of TGF-β1-induced Col I synthesis in the inflamed colon.
The expression of Timp1 was upregulated after DSS treatment in both EGFP BM and CCR2 RFP/RFP BM chimeras, but the expression level in EGFP BM chimeras was significantly higher than in CCR2 RFP/RFP BM chimeras (Fig. 8A). Moreover, the number of TIMP-1 + cells in the colon significantly increased after chronic DSS treatment in EGFP BM chimeras (Fig. 8C,D). However, the amount in CCR2 RFP/RFP BM chimeras was similar to the untreated control EGFP BM chimeras (Fig. 8D). Although the number of BM-derived EGFP + TIMP-1 + cells significantly increased after chronic DSS treatment, resident EGFP − TIMP-1 + cells did not change (Fig. 8C,E).
Next, we isolated CD45 + CD11b + Col I + fibrocytes, CD11b + CCR2 + Col I − cells, CD11b + CCR2 − Col I − cells and CD45 − Col I + fibroblasts (including myofibroblasts) from the colonic LP cells of the EGFP BM chimeras after three cycles of DSS treatment to identify the cell populations producing TIMP-1 in the colon. The expression of Timp1 was detected in fibrocytes, CD11b + CCR2 + Col I − cells and fibroblasts (Fig. 8F). These results demonstrate that TIMP-1 produced by CCR2 + cells, such as fibrocytes, monocytes and macrophages but not by resident fibroblasts within the colonic LP, may be involved in the development of colon fibrosis.

Discussion
Recent murine and human studies have demonstrated that fibrocytes accumulate in the colonic submucosa and directly augment fibrosis by producing Col I 26,27 . However, the detailed functional analysis of fibrocytes has not been entirely performed. In the current study, we used BM chimeras, which were transplanted with BM-TNCs obtained from EGFP or CCR2 RFP/RFP mice, to investigate which cells are precursors of fibrocytes in the colon, how they migrate into the inflamed colon and how they contribute to colon fibrosis.
We identified two types of fibrocytes, CCR2 + and CCR2 − fibrocytes, in the colonic LP of mice chronically treated with DSS. Although the former cells were Ly6C − F4/80 + and this phenotype was similar to that of macrophages, the latter cells were Ly6C high F4/80 − and similar to monocytes. These two types of fibrocytes are thought to have a different origin. CCR2 − fibrocytes are likely derived from circulating fibrocytes. It is assumed that CCR2 + fibrocytes are differentiated from infiltrating CCR2 + Ly6C high monocytes since: (1) CCL2 was the major chemokine produced in the injured colon in both acute and chronic injury; (2) CD45 + CD11b + CCR2 + Col I − monocytes/macrophages, as well as fibrocytes in the colonic LP of injured mice, were positive for Col1a1; (3) adoptively transferred CD45.2 + CCR2 + Ly6C + Col I − monocytes differentiated into CD45.2 + CD11b + Col I + fibrocytes in the inflamed colon; and (4) the total monocytes and fibrocytes in the colonic LP after chronic injury significantly decreased, and colon fibrosis was attenuated in CCR2 RFP/RFP BM chimeras compared with WT BM chimeras. Shi et al. demonstrated that CCR2 + CD11b + Ly6G int monocytic myeloid-derived suppressor cells, which were positive for Kruppel-like factor 4, differentiated into fibrocytes 57 . Wang et al. reported that intramuscular CCR2 + Ly6C low/− F4/80 + fibrocytes originated from infiltrating monocytes/macrophages, differentiated within injured muscles and played a pathological role in maintaining chronic inflammation and driving progressive fibrosis in a mouse model for Duchenne muscular dystrophy 58 . Their findings are very similar to ours. Although the two types of fibrocytes in our study may play different roles during wound repair and fibrosis, CCR2 − fibrocytes might not be associated with the development of colon fibrosis as a substantial number of them remained in CCR2 RFP/RFP BM chimeras in which colon fibrosis was significantly reduced.
Recently, BM-derived fibrocytes have been reported to differentiate into fibroblasts and myofibroblasts and participate in tissue wound repair and fibrosis 16,[20][21][22]59 . Suga et al. reported that BM-derived fibrocytes could contribute to the myofibroblast population at the wound site in the acute but not in the chronic phase after injury 20 . Uehara et al. also described that α-SMA + myofibroblasts appeared during days 7-14 in a murine model of DSS-induced colitis 26 . However, the origin of α-SMA + myofibroblasts were not clarified; the authors suggested that BM transplantation studies using the Y chromosome or green fluorescent protein as a marker of donor cells would be necessary. In our study, a considerable number of EGFP + fibrocytes existed in the injured colon, but there were only a few EGFP + fibroblasts and myofibroblasts during the fibrosis process. Some researchers have demonstrated that the contribution of fibrocytes to myofibroblasts is small in the lung, kidney and muscle fibrosis murine models 10,58,60,61 .
Some authors reported that the number of fibroblasts increased in both the mucosa and submucosa of the colon of mice chronically treated with DSS compared with the untreated control mice 62,63 . Fibrocytes in the (2019) 9:8568 | https://doi.org/10.1038/s41598-019-45012-6 www.nature.com/scientificreports www.nature.com/scientificreports/ colonic LP significantly increased during chronic DSS treatment. However, in our study, there was no change in the number of fibroblasts, most of which were resident cells, in EGFP BM chimeras. Although CCR2 deficiency did not affect the number of fibroblasts in the colonic LP of the chronically treated mice, it reduced the number The gating strategy was as follows: fibrocytes; CD45 + CD11b + Col I + cells, CD11b + CCR2 + cells; CD45 + CD11b + Col I − CCR2 + cells, CD11b + CCR2 − cells; CD45 + CD11b + Col I − CCR2cells, fibroblasts; CD45 − Col I + cells. The sample size for each group was n = 6. The experiments were performed at least twice, yielding similar results. Data are expressed as the mean ± SD. *** p < 0.001 by Mann-Whitney U test and Kruskal-Wallis test followed by Dunn's multiple comparison test.
www.nature.com/scientificreports www.nature.com/scientificreports/ of fibrocytes and the development of colon fibrosis. Tsukui et al. also reported that the number of fibroblasts in the lungs remained constant throughout bleomycin-induced lung fibrosis, and excessive deposition of ECM during lung fibrosis cannot be accounted for by an increase in fibroblast numbers 61 . Furthermore, they supposed that fibrocytes were the subsets most likely to contribute to fibrotic responses in the injured lung because CCR2-deficient mice were protected from fibrosis. Their results and ours suggest that the development of tissue fibrosis is significantly associated with the accumulation of CCR2 + cells, especially Ly6C high monocytes and their progenies, fibrocytes.
Although there was no significant difference in the expression level of Col1a1 and Tgfb1 in the colon between EGFP BM and CCR2 RFP/RFP BM chimeras after three cycles of DSS treatment, the fibrosis area assessed by Sirius red staining was significantly reduced in the CCR2 RFP/RFP BM chimeras. These results evoke the possibility that the reduction of Col I deposition may be caused by the stimulation of Col I degradation rather than the inhibition of Col I synthesis.
When we compared the expression level of Mmp1a, Mmp8, Mmp13 and Timp1 within the colon tissues after three cycles of DSS treatment between EGFP BM and CCR2 RFP/RFP BM chimeras, only Timp1 expression was significantly reduced in CCR2 RFP/RFP BM chimeras. An increased level of TIMP-1 protein has been reported in inflammatory and fibrotic lesions in Crohn's disease and a murine model of chronic inflammation-induced intestinal fibrosis 43,44 . Breynaert et al. reported that in the chronic colitis model, reduced colonic inflammation, lower tissue collagen levels and less fibrosis were observed in TIMP-1 knockout mice compared with WT mice 64 . Therefore, TIMP-1 may be a promising therapeutic target to prevent the progression of colon fibrosis. Macrophages, neutrophils, eosinophils, fibroblasts and myofibroblasts reportedly produce TIMPs within the colonic LP 34,44 . We found that fibrocytes, CD11b + CCR2 + cells and fibroblasts, but not CD11b + CCR2 − cells expressed Timp1, and TIMP-1 + cells significantly decreased in CCR2 RFP/RFP BM chimeras compared with EGFP BM chimeras after chronic DSS treatment. These results show that TIMP-1, which is synthesised by CCR2 + cells, such as monocytes, macrophages and fibrocytes, plays a crucial role in the development of colon fibrosis.
Recently, Friedman et al. reported that cenicriviroc, a dual antagonist of CCR2 and CCR5, showed a significant benefit in patients with non-alcoholic steatohepatitis with liver fibrosis 65 . Furthermore, angiotensin II type 1 receptor blockers, which are globally used for the treatment of hypertension, were also reported to act as CCR2 antagonists 66 . In the future, these drugs might be useful for the treatment of patients with colon fibrosis.
In conclusion, our results suggest that circulating CCR2 + Ly6C high monocytes migrate into the inflamed colon via the CCL2/CCR2 axis, differentiate into CCR2 + Ly6C − F4/80 + fibrocytes and contribute to the development of colon fibrosis through the production of TIMP-1. Therefore, blocking the migration of CCR2 + Ly6C high monocytes to the inflamed colon may be a potential new therapeutic option to prevent colon fibrosis in patients with IBD. Although colon fibrosis was attenuated with the reduction of CCR2 + monocyte-derived cells, it was not completely inhibited, as seen in Fig. 7B. As multiple intestinal cell types other than CCR2 + monocyte-derived cells may contribute to the development of colon fibrosis, further experiments are necessary to clarify the mechanism underlying colon fibrosis and develop more effective and targeted therapeutic strategies for colon fibrosis.

Induction of colitis.
We confirmed that colon length did not change between unirradiated and irradiated C57BL/6-Ly5.1 mice at 2 months after BM transplantation (data not shown), in agreement with the report by Followill et al., who reported that single doses of less than 20 Gy produced acute crypt cell depletion followed by successful regeneration, repopulation and restoration of the colonic mucosa 68 . Therefore, AOM/DSS treatment was started at 2 months after BM transplantation. To establish the appropriate concentration of DSS in BM chimeras, different amounts (1% and 2%) of DSS (MW 36-50 kDa, MP Biomedicals, Santa Ana, CA) dissolved in drinking water were administered for 7 days, followed by water alone for 2 weeks. Almost all mice died immediately after 2% DSS administration; thus, colitis was induced using 1% DSS. Mice were injected with 10 mg/kg AOM (Wako Pure Chemical Industries, Osaka, Japan) intraperitoneally 1 week before DSS treatment, and DSS treatment was repeated for 3 cycles. Healthy control mice received only water. AOM/DSS-induced colitis was scored as DAI, which was the combined score of weight loss, stool consistency and bleeding 69 . DAI score was monitored three times per week. Non-transplanted CCR2 RFP/+ CX3CR1 GFP/+ mice were administered 2% DSS because their DAI scores were significantly lower than those of BM chimeras when both groups of mice were treated with 1% DSS.
www.nature.com/scientificreports www.nature.com/scientificreports/ Tissue preparation. Mice were euthanised by cervical dislocation after anaesthesia with isoflurane, and their colons were resected and fixed in 4% phosphate-buffered paraformaldehyde for 1 h at room temperature. Some tissue blocks were embedded in paraffin after dehydration in a graded alcohol series. Other tissue blocks were fixed in 4% phosphate-buffered paraformaldehyde for another 5 h at room temperature, embedded in Tissue-Tek OCT medium (Sakura Finetek USA, Torrance, CA), rapidly frozen by plunging into liquid nitrogen and stored at −80 °C. Tissue blocks were cut to 5 μm sections using a microtome or a cryostat.
Histological analysis. Serial sections were stained with HE, and the histological inflammation score was determined as previously described 69 . Each sample was randomly selected from three perspectives, and average scores were calculated. The sections were stained with Sirius red and Masson's trichrome to evaluate the presence of fibrosis. Three images per section were captured at 100× magnification using an Olympus BX41 microscope (Olympus, Tokyo, Japan) equipped with a 10×/0.40 numerical aperture objective lens and an Olympus Camedia C-5060 camera. Image J software (NIH, Bethesda, MD) was used for evaluating the percentage of red staining for Sirius red and that of blue staining for Masson's trichrome in the whole area of each image, with the dyed area indicating the presence of collagen fibres in the tissue. Sections stained with Sirius red were analysed using an Olympus BX50 microscope fitted with polarising filters to quantify type I and type III collagen. Photomicrographs were obtained in 100× magnification in three representative fields for each sample. The counting of collagen bundles was performed using Image J software and the percentages of red staining for type I collagen and of green staining for type III collagen in the whole area of each image were calculated. (2019) 9:8568 | https://doi.org/10.1038/s41598-019-45012-6 www.nature.com/scientificreports www.nature.com/scientificreports/ RNA isolation. After cutting the colon in longitudinally, 0.5 cm of the rectum was collected from the anal side, placed in RNAlater (Invitrogen) and stored at 4 °C. RNA was extracted using the RNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacturer's protocol.
RNA was isolated from colonic LP cells following intracellular antibody staining and fluorescence-activated cell sorting according to a previously reported method 71 . After sorting, cells were pelleted by centrifugation at 2000 × g for 15 min at 4 °C. Total RNA was isolated from the pellet using the RecoverAll Total Nucleic Acid Isolation Kit (Ambion, Austin, TX), beginning with the protease digestion step of the manufacturer's recommended protocol. The isolation procedure was modified as follows: cells were incubated in digestion buffer at 50 °C for 3 h, cell lysates were frozen at −80 °C overnight and RNA was isolated according to the manufacturer's instructions.
Quantitative RT-PCR. Complementary DNA was generated from total RNA using the SuperScript III Statistics. Data are expressed as the mean and standard deviation. Two experimental groups were compared