Research ArticleTISSUE REPAIR

Emergence of immunoregulatory Ym1+Ly6Chi monocytes during recovery phase of tissue injury

See allHide authors and affiliations

Science Immunology  05 Oct 2018:
Vol. 3, Issue 28, eaat0207
DOI: 10.1126/sciimmunol.aat0207

Monocytes mediate repair

Different subsets of macrophages contribute to inflammation during tissue injury or resolution of inflammation and tissue repair, but the distinct subsets of macrophages that support tissue repair are not well defined. Ikeda et al. now define a role for immunoregulatory Ym1+Ly6Chi monocytes during the recovery phase of tissue injury. Bone marrow–resident precursors of Ly6Chi monocytes express Ym1 during the final stage of differentiation, and these Ym1+Ly6Chi monocytes expand in the bone marrow during late phases of inflammation. These monocytes have an immunoregulatory phenotype and are required for recovery from tissue injury in a mouse model of colitis. These findings highlight a role for bone marrow–originated Ym1+Ly6Chi monocytes in tissue repair.

Abstract

Ly6Chi monocytes migrate to injured sites and induce inflammation in the acute phase of tissue injury. However, once the causes of tissue injury are eliminated, monocyte-derived macrophages contribute to the resolution of inflammation and tissue repair. It remains unclear whether the emergence of these immunoregulatory macrophages is attributed to the phenotypic conversion of inflammatory monocytes in situ or to the recruitment of bone marrow–derived regulatory cells de novo. Here, we identified a subpopulation of Ly6Chi monocytes that contribute to the resolution of inflammation and tissue repair. Ym1+Ly6Chi monocytes greatly expanded in bone marrow during the recovery phase of systemic inflammation or tissue injury. Ym1+Ly6Chi monocytes infiltrating into an injured site exhibited immunoregulatory and tissue-reparative phenotypes. Deletion of Ym1+Ly6Chi monocytes resulted in delayed recovery from colitis. These results demonstrate that a distinct monocyte subpopulation destined to act in immunoregulation is generated in bone marrow and participates in resolution of inflammation and tissue repair.

INTRODUCTION

When tissue is injured by microbial invasion or sterile injury, several kinds of immune cells infiltrate damaged tissue and induce acute inflammation in a coordinated manner. The inflammatory response is critical for the elimination of microorganisms. However, once the causes of tissue injury are eliminated, the acute inflammatory response should shift to tissue repair and regeneration processes. Macrophages play critical roles in both the propagation and resolution of inflammation (1). In the acute inflammatory phase, tissue-resident macrophages and infiltrating monocytes coordinately induce inflammation by secreting inflammatory cytokines and chemokines. In contrast, these cells also contribute to tissue repair and regeneration during the recovery phase of inflammation (2). For example, in a mouse model of liver injury, the depletion of CD11b+ macrophages late in the course of inflammation impairs tissue regeneration (3). Newly infiltrated monocyte-derived macrophages exacerbate ischemia-induced kidney injury, whereas blockade of the in situ proliferation of these macrophages inhibits tubular repair (4). These findings suggest the emergence of phenotypically distinct macrophages that contribute to tissue repair and regeneration after tissue injury. However, the origin of these macrophages with distinct phenotypes is not fully understood. In particular, it remains unclear whether these tissue repair–driving macrophages are derived from circulating precursors or result from the phenotypic conversion of proinflammatory macrophages in situ (5).

It is generally considered that the monocyte-macrophage lineage has high plasticity and can readily change phenotypes in response to environmental cues (6, 7). This concept is supported in part by observations that bone marrow–derived macrophages exhibit immunologically polarized states in vitro (5, 8, 9), that is, two distinct macrophage subsets are differentiated from monocyte-macrophage precursor cells in bone marrow by different sets of cytokines. Classically activated macrophages, also known as M1 macrophages, are induced by lipopolysaccharide (LPS; with or without interferon-γ) and produce proinflammatory cytokines and reactive oxygen intermediates, and these macrophages exhibit microbicidal and tumoricidal activities. Alternatively activated macrophages, also known as M2 macrophages, differentiate in response to interleukin 4 (IL-4; with or without IL-10) stimulation. M2 macrophages show anti-inflammatory activity and enhanced expression of genes related to tissue repair and remodeling. On the basis of this classification of macrophage subsets, much effort has been made to understand the phenotypes and functions of tissue-resident or infiltrating macrophages in a variety of pathological conditions. Although tissue macrophages exhibit properties of M1 or M2 under certain conditions in vivo (2, 9, 10), most macrophages exhibit mixed or atypical phenotypes that do not strictly fit M1 or M2 phenotypes, suggesting that a greater complexity of macrophage differentiation or activation occurs in vivo (1113). Unlike M2 macrophages that are generated in vitro, details remain unknown about the origin and function of immunoregulatory or tissue repair–promoting macrophages in vivo.

Here, we show that the resolution of inflammation prompts precursors of Ly6Chi monocytes in bone marrow to differentiate into Ym1-expressing monocytes endowed with the ability to accelerate tissue regeneration. These findings indicate that the emergence of immunoregulatory macrophages in the recovery phase of tissue injury is attributed to not only the phenotypic conversion of macrophages in situ but also the migration of Ym1+ monocytes produced de novo in bone marrow.

RESULTS

Identification of Ym1-expressing cells in peripheral blood and other tissues

Ym1 is known to be expressed in myeloid cell lineages and generally used as a marker for alternatively activated M2 macrophages. However, it remains obscure what kinds of cell populations and under which conditions those cells actually express Ym1 in vivo. To visualize Ym1-expressing cells in mice, we generated mice that harbor the Venus gene in the Chil3 (Ym1-coding gene) loci (Ym1-Venus mice; fig. S1, A and B). We first sought to determine what kinds of tissue-resident macrophages or dendritic cells expressed Ym1-Venus in various tissues of these mice. As shown in Fig. 1A, tissue-resident macrophages and dendritic cells were Ym1-Venus–negative in lymph nodes, spleen, colon, liver, and peritoneal cavity. The only exception was resident macrophages in the lung. Lung-resident macrophages consisted of CD11c+CD64+ alveolar macrophages and CD11b+MHCII+ interstitial macrophages, and only the former expressed Ym1-Venus. We next sought to identify Ym1-Venus–positive cells in the peripheral blood of these mice. It is reported that neutrophils expressed Ym1 (14, 15), and consistent with these reports, most of Ly6G+ neutrophils were Ym1-Venus–positive (Fig. 1, B and C). We also found that neutrophils expressed Ym1-Venus in the spleen (Fig. 1D). In bone marrow, the expression levels of Ly6G are gradually up-regulated, whereas those of c-Kit are down-regulated in the course of neutrophil differentiation (16). As shown in fig. S2A, not only Ly6Ghi mature neutrophils but also c-Kit dull Ly6G-negative to -low immature neutrophils expressed Ym1-Venus. Immature neutrophils expressed higher levels of Ym1 mRNA than mature neutrophils, indicating that Ym1 is expressed from the early stages of the neutrophil differentiation (fig. S2B). Expression of Ym1-Venus was not detected in eosinophils, B cells, T cells, natural killer (NK) cells, or NKT cells in peripheral blood or the spleen (Fig. 1, B to D). In addition to neutrophils, we noticed that a small population (5%) of CD115+Ly6Chi peripheral blood and splenic monocytes expressed Ym1-Venus, whereas CD115+Ly6Clo monocytes did not express Ym1-Venus (Fig. 1, B to D).

Fig. 1 Identification of Ym1-expressing cells in peripheral blood and other tissues.

Flow cytometry analysis of various tissues (A), peripheral blood (B and C), and spleen (D) from WT (shaded area in blue) or Ym1-Venus (shaded area in red) mice. Left panels of each tissue indicate the gating strategy. Experiments were independently repeated at least two times. Mφ, macrophage; Mo, monocyte; DC, dendritic cell; LPM, large peritoneal macrophage; SPM, small peritoneal macrophage; Neu, neutrophil; Eos, eosinophil; MHCII, major histocompatibility complex II; SSC, side scatter. Average values are shown with SD in (C).

The above results suggested that the Ym1-Venus–expressing CD115+Ly6Chi monocytes may be a unique subpopulation. To evaluate the phenotype and function of this potential monocyte subpopulation, we examined Ym1-Venus expression during monocyte differentiation. Similar to observations in peripheral blood cells, 4% of CD115+Ly6Chi monocytes expressed Ym1-Venus in bone marrow, whereas no Venus-positive cells in CD115+Ly6Clo monocytes were detected (Fig. 2A). It has been reported that CD115+Ly6Chi monocytes can be delineated into three different subpopulations based on expression levels of Flt3 and CD11c (17). Ym1-Venus+ monocytes exclusively belonged to the CD115+Ly6ChiFlt3CD11c subpopulation (Fig. 2B).

Fig. 2 Flow cytometry analysis of bone marrow cells in Ym1-Venus mice.

Flow cytometry analysis of bone marrow cells from WT (blue-shaded area) or Ym1-Venus (red-shaded area) mice. Ym1-Venus expression in leukocytes (A), three subpopulations of Ly6Chi monocytes (B), myeloid progenitors (C), and monocyte progenitors (D) are shown. Representative fluorescence-activated cell sorting (FACS) plots from two independent experiments are shown. Lin, B220, CD19, CD3, Ly6G, MHCII, NK1.1, and Ter119 (B); B220, CD11b, CD3, Gr-1, NK1.1, and Ter119 (C); B220, CD3, Ly6G, NK1.1, and Ter119 (D).

We further evaluated the Ym1-Venus expression in various hematopoietic progenitors in Ym1-Venus mice and found that common myeloid progenitor (CMP), granulocyte-macrophage progenitor (GMP), and megakaryocyte-erythroid progenitor (MEP) did not express Ym1-Venus (Fig. 2C). GMP gives rise to macrophage-dendritic cell progenitor (MDP), and monocytes are differentiated from MDP through the common monocyte progenitor (cMoP) (18). We found that neither MDP nor cMoP expressed Ym1-Venus, and Ym1-Venus expression was detected in monocytes only after these cells lost c-Kit expression (R4, Fig. 2D). These results indicate that monocytes express Ym1 at the final stage of their differentiation in bone marrow under steady-state conditions.

Expansion of Ym1+Ly6Chi monocyte subpopulation in the late phase of inflammation

Ly6Chi monocytes infiltrate into an inflammatory site in response to microbial invasion or tissue injury and regulate inflammation. Thus, we sought to determine the dynamics of Ym1-Venus–positive cells in CD115+Ly6Chi monocytes during inflammation. Serum concentrations of inflammatory cytokines, including tumor necrosis factor–α (TNFα) and IL-6, increased 2 hours after LPS injection and rapidly returned to baseline levels by 24 hours (fig. S3). Both the number and proportion of blood neutrophils increased 24 hours after LPS injection and returned to normal levels after 72 hours (Fig. 3A). In contrast, the proportion of Ly6Chi monocytes remained unchanged 24 hours after LPS injection but increased and occupied about 6% of blood leukocytes after 48 hours (Fig. 3A). The Ym1-Venus+Ly6Chi monocyte subpopulation expanded markedly during the late phase of inflammation (Fig. 3, A and B), such that the proportion of the Ym1-Venus–positive subpopulation in Ly6Chi monocytes increased to 10% 24 hours after LPS injection and reached about 50% after 48 hours. The Ym1-Venus–positive subpopulation decreased to 24% in Ly6Chi monocytes, and 5% of Ly6Clo monocytes became Ym1-Venus–positive 72 hours after LPS injection, suggesting that Ym1-Venus+Ly6Chi monocytes may become Ly6Clo monocytes. We also confirmed that 54% of the Ly6Chi monocytes expressed Ym1-Venus in the spleen 48 hours after LPS injection (Fig. 3C).

Fig. 3 Expansion of Ym1+Ly6Chi monocyte subpopulation in the late phase of inflammation.

Ym1-Venus mice were injected intravenously with LPS (100 μg per mouse). (A and B) Ym1-Venus expression in peripheral blood myeloid cells. Blood leukocytes from Ym1-Venus mice were prepared at indicated time points after LPS challenge. (A) Representative FACS plots of three mice are shown. (B) Kinetics of Ym1-Venus–positive cells in Ly6Chi blood monocytes at indicated time points. Average values are shown with SD. Frequencies of Venus(+) monocytes at indicated time points were compared with baseline level by one-way ANOVA, n = 3 mice. (C to E) Ym1-Venus expression in leukocytes from spleen (C) and bone marrow (D and E). Leukocytes were prepared 48 hours (C and E) or 24 and 48 hours (D) after systemic LPS injection. (E) Ym1-Venus expression in three subpopulations of Ly6Chi monocytes. Representative FACS plots from two independent experiments are shown (C to E). (F) Chil3 (Ym1-coding gene) mRNA level in monocytes. CD115+Ly6Chi monocytes were prepared from either naïve or LPS-injected (48 hours) Ym1-Venus mouse bone marrow. Monocytes from LPS-injected mice were further sorted into Ym1-Venus–positive or Ym1-Venus–negative fraction. Chil3 mRNA level in each fraction was determined by qRT-PCR and is shown as fold change relative to naïve Ly6Chi monocytes. Average values are shown with SD. One-way ANOVA, n = 3. *P < 0.05, **P < 0.01, ****P < 0.0001. n.s., not significant. Blue-shaded area indicates WT control.

Consistent with these findings in peripheral blood and spleen, the proportion of Ym1-Venus–positive cells in the Ly6Chi monocytes started to increase in bone marrow 24 hours after LPS injection and reached about 40% after 48 hours (Fig. 3D). These Ym1-Venus+Ly6Chi monocytes belonged to the Flt3CD11c monocyte subpopulation, the same as that observed in the bone marrow of naïve mice (Fig. 3E). Quantitative polymerase chain reaction (PCR) analysis revealed that these Ym1-Venus+Ly6Chi monocytes actually exhibited higher Ym1 mRNA expression compared with Ym1-Venus–negative monocytes in bone marrow (Fig. 3F). Together, we concluded that the Ym1-expressing Ly6Chi monocyte subpopulation expands markedly in the bone marrow during the late phase of inflammation.

Ym1+Ly6Chi monocytes exhibit immunoregulatory phenotype

We next sought to characterize the Ym1-Venus+Ly6Chi monocyte subpopulation that expanded in the late phase of systemic inflammation. Both Ym1-Venus+Ly6Chi monocytes and Ym1-VenusLy6Chi monocytes showed a large monolobular nucleus, which is a typical morphological feature of monocytes (Fig. 4A). Flow cytometry analysis revealed that both subpopulations shared several Ly6Chi monocytes surface markers, including CD62L, CCR2, and CD204 (Fig. 4B). As a portion of Ly6Chi monocytes differentiate into Ly6Clo monocytes in peripheral blood (1822), we next examined whether transferred Ym1-Venus+Ly6Chi monocytes or Ym1-VenusLy6Chi monocytes became Ly6Clo monocytes. When separately transferred into naïve wild-type (WT) mice, both Ym1-Venus+Ly6Chi monocytes and Ym1-VenusLy6Chi monocytes decreased their surface expression levels of Ly6C after 48 hours, indicating that both monocyte subpopulations have the intrinsic capability to become Ly6Clo monocytes (Fig. 4C). From these results, we concluded that Ym1-Venus+Ly6Chi monocytes and Ym1-VenusLy6Chi monocytes share several common characteristics of Ly6Chi monocytes. To further characterize Ym1-Venus+Ly6Chi monocytes, we compared global mRNA expression profiles between Ym1-Venus+Ly6Chi monocytes and Ym1-VenusLy6Chi monocytes from LPS-treated mice (Fig. 4, D and E). In agreement with the shared characteristics described above, there were small differences between these subsets in terms of the expression levels of inflammatory genes, transcription factors (fig. S4, A and B), or monocyte/macrophage signatures (Fig. 4E). One of the few macrophage-associated genes highly expressed in Ym1-Venus–positive monocytes was Marco, a scavenger receptor required for combating pneumococcal infection (Fig. 4, E and F) (23). Ym1 is considered a marker for alternatively activated (M2) macrophages. Ym1-Venus–positive monocytes expressed higher levels of Ym1 mRNA than Ym1-Venus–negative monocytes (Fig. 4, E and F). However, none of other M1- or M2-related gene expression was down- or up-regulated in Ym1-Venus–positive monocytes when compared with Ym1-Venus–negative ones (Fig. 4E). Principal components analysis (PCA) showed much greater distance between the Ym1+ or Ym1 monocytes and neutrophils than between the naïve monocytes (Fig. 4D). To highlight the difference between Ym1+ and Ym1 monocytes, we further performed PCA between three monocyte subsets (fig. S4C). This analysis extracted several neutrophil-related genes as loading factors that discriminate Ym1+ and Ym1 monocytes. The expression of several neutrophil-related genes including Ltf (lactoferrin-coding gene) and Mmp9 was enhanced in Ym1-Venus–positive monocytes (Fig. 4, E and F). The expression levels of other neutrophil-related genes, such as Ly6g5b and Retnlg (resistin-like gamma-coding gene), were similar between two subpopulation (Fig. 4, E and F), thereby excluding the possibility of neutrophil contamination in sorted Ym1-Venus–positive monocytes. These results suggest that Ym1-Venus–positive monocytes are indeed a subpopulation of monocytes but have some characteristics of neutrophils.

Fig. 4 Characterization of Ym1+Ly6Chi monocytes in mice injected with LPS.

Ym1-Venus mice were injected intravenously with LPS (100 μg per mouse). Ym1-Venus+Ly6Chi monocytes and Ym1-VenusLy6Chi monocytes were analyzed in bone marrow (A and C to F) or peripheral blood (B) of LPS-injected (48 hours) Ym1-Venus mice. (A) Morphological features of Ym1-Venus–positive and Ym1-Venus–negative monocytes. Monocytes were sorted from bone marrow of LPS-injected (48 hours) Ym1-Venus mice and stained with Giemsa. Scale bars, 10 μm. (B) Flow cytometry analysis of an array of surface markers on monocytes. Black lines indicate isotype control. (C) Ym1-Venus+Ly6Chi monocytes or Ym1-VenusLy6Chi monocytes were prepared from bone marrow of LPS-injected (48 hours) Ym1-Venus mice. Purified monocytes (4 × 105 cells each) were adoptively transferred into WT congenic mice (CD45.1+) injected with LPS (10 μg per mouse). After 48 hours, peripheral blood leukocytes from the recipient mice were analyzed by flow cytometry. Note that both Ym1-Venus–positive monocytes (red) and Ym1-Venus–negative monocytes (blue) down-regulated the expression level of Ly6C 48 hours after transfer. Representative FACS plots of two independent experiments are shown. (D and E) Gene expression profiles of Ym1-Venus+Ly6Chi monocytes or Ym1-VenusLy6Chi monocytes were globally compared by RNA-seq analysis. PCA (D) and heatmap of the indicated genes (E). (F) mRNA expression levels of indicated genes were quantitated by qRT-PCR and are shown as fold change relative to Ym1-VenusLy6Chi monocytes. Average values are shown with SD. One-way ANOVA, n = 3. Ltf, Mmp9, and Retnlg mRNA expression levels are compared between three monocytes. *P < 0.05, **P < 0.01, ****P < 0.0001.

To further reveal the functional differences between Ym1-Venus–positive and Ym1-Venus–negative monocytes, we purified each monocyte subpopulation from LPS-treated mice and restimulated these cells with LPS ex vivo. As shown in Fig. 5A, Ym1-Venus–positive monocytes produced lower levels of proinflammatory cytokines and a higher level of anti-inflammatory cytokine IL-10 than Ym1-Venus–negative monocytes. The monocytes did not produce detectable levels of cytokines unless restimulated ex vivo. As shown in Fig. 5B, the transfer of Ym1-VenusLy6Chi monocytes increased serum IL-6 concentrations in LPS-injected recipient mice. On the other hand, the transfer of Ym1-Venus+Ly6Chi monocytes did not have any impact on the serum IL-6 concentration in LPS-injected–recipient mice compared with the mice without transferred cells. These results indicate the reduced proinflammatory cytokine production by Ym1-Venus+Ly6Chi monocytes in vivo.

Fig. 5 Ym1-Venus+Ly6Chi monocytes exhibit immunoregulatory phenotype.

For monocyte isolation, Ym1-Venus mice were injected intravenously with LPS (100 μg per mouse). Forty-eight hours later, Ym1-Venus+Ly6Chi monocytes and Ym1-VenusLy6Chi monocytes were purified from the spleen (A) or bone marrow (B and C). (A) Cytokine production level in monocytes. Purified monocytes were stimulated with LPS (1 μg/ml) ex vivo. Concentrations of indicated cytokines in culture supernatant were measured by ELISA. Assays were performed in triplicate, and mean values are shown with SD. Student’s t test, n = 3 to 5. (B) Serum cytokine levels in mice transferred with or without monocytes. Ym1-Venus–positive or Ym1-Venus–negative monocytes (4 × 105 cells) from LPS-injected Ym1-Venus mice were adoptively transferred into WT mice that were injected with LPS (10 μg per mouse). After 4 hours, serum cytokine levels were measured by ELISA. Average values are shown with SD. Student’s t test, n = 4 to 5. (C) Flow cytometry analysis of cultured splenocytes. Splenocytes from WT mice were stimulated ex vivo with anti-CD3 and anti-CD28 in the presence or absence of monocyte culture medium. After 72 hours, cells were stained for CD4 and Foxp3. Numbers indicate percentage of Foxp3+ cells in CD4+ fraction. Average values are shown with SD. One-way ANOVA, n = 4 to 6. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Each symbol represents an individual sample. Iso, isotype control; IgG1, immunoglobulin G1.

It is reported that monocyte-derived alternatively activated macrophages promote the differentiation of CD4+Foxp3+ regulatory T cells (24). Thus, we treated splenic T cells with culture media of Ym1-Venus+Ly6Chi monocytes or Ym1-VenusLy6Chi monocytes prepared from LPS-treated mice and stimulated these cells with anti-CD3 antibody (Ab) and anti-CD28 Ab. As shown in Fig. 5C, compared with Ym1-VenusLy6Chi monocyte culture media, Ym1-Venus+Ly6Chi monocyte culture media promoted the differentiation of activated T cells toward Foxp3+ regulatory T cells. Together, we concluded that Ym1-Venus+Ly6Chi monocytes exhibit the immunoregulatory phenotype.

Ym1+Ly6Chi monocytes are essential for recovery from tissue injury

Because the proportion of Ym1-Venus–positive monocytes increases in the late phase of systemic inflammation, we next examined the dynamics of Ym1-Venus–positive monocytes during local inflammation. The administration of dextran sodium sulfate (DSS) in drinking water induces epithelial injury that manifests as a clinical symptom of acute colitis in mice. This model is considered dependent on innate immunity because colitis develops in lymphocyte-deficient mice. In this experimental model, discontinuation of DSS administration terminates inflammation and initiates the regeneration of the epithelial barrier. Thus, we can analyze both acute and recovery phases of inflammation using this model. We administered DSS in drinking water to Ym1-Venus mice for 5 days, followed by discontinuation of DSS administration, and examined blood cell expression of Ym1-Venus during the course of inflammation. During the early inflammatory phase, neutrophils constantly expressed Ym1-Venus, whereas the proportion of Ym1-Venus+Ly6Chi monocytes remained low (Fig. 6, A and B). However, 3 days after the discontinuation of DSS administration (day 8), a substantial number of Ym1-Venus–positive monocytes appeared in the Ly6Chi population, accounting for ~30% of this monocyte subpopulation during the recovery phase of colitis. Consistent with these findings, a substantial number of Ym1-Venus–positive monocytes also appeared in the Ly6Chi population during the recovery phase of inflammation in spleen (fig. S5) and bone marrow (Fig. 6C). We examined whether these monocytes can infiltrate into sites of inflammation. Gut macrophages and monocytes can be classified on the basis of the differential expression of Ly6C and CD64, such that Ly6ChiCD64lo monocytes decrease their Ly6C expression while progressively increasing CD64 expression as they differentiate into resident macrophages (25). We found that the frequency of Ym1-Venus–positive cells among Ly6ChiCD64lo monocytes in the inflamed colon expanded not on day 5, but on day 9, whereas only a small proportion of Ly6CintCD64int monocytes and none of Ly6CloCD64hi macrophages were Ym1-Venus–positive (Fig. 6D). Given that Ly6ChiCD64lo monocytes are reported to become Ly6CloCD64hi macrophages after they infiltrate into the colon, Ym1 expression may gradually decrease in Ym1-Venus–positive monocytes during developmental maturation in the colon.

Fig. 6 Expansion of Ym1+Ly6Chi monocyte subpopulation during recovery phase of colitis.

Ym1-Venus and WT mice were orally administered 2% DSS in drinking water for 5 days. (A and B) Ym1-Venus expression in blood myeloid cells. Peripheral blood leukocytes were prepared at indicated time points. (A) Representative FACS plots of four mice are shown. (B) Average frequencies of Ym1-Venus–positive cells among Ly6Chi blood monocytes (top) and their absolute numbers (bottom) are plotted and compared with baseline levels. Average values are shown with SD. One-way ANOVA, n = 4. The representative results of two independent experiments are shown. Ym1-Venus expression in bone marrow (C) and colon (D). Myeloid cells in indicated tissues of Ym1-Venus colitis mice were prepared at indicated time points. Frequencies of Ym1-Venus–positive cells among Ly6Chi monocytes (R1) are shown (C and D, right). Average values are shown with SD. One-way ANOVA, n = 3. Representative FACS plots of three mice are shown. (E) Inflammation-related gene mRNA expression levels in colon lamina propria Ym1-Venus+Ly6Chi monocytes or Ym1-VenusLy6Chi monocytes. mRNA levels of indicated genes were determined by qRT-PCR. Average values are shown with SD. Student’s t test, n = 5 to 7. Each symbol represents an individual animal. (F to H) Ym1-DTR mice or WT mice were orally administered 2% DSS in drinking water for 5 days. (F) Depletion of Ym1-Venus–positive cells caused delayed recovery from DSS-induced colitis. DT was injected intraperitoneally on days 8 and 10. Average body weight with SEM at indicated time points is shown. n = 7 WT mice or n = 8 Ym1-DTR mice. Two-way ANOVA with multiple comparisons. (G) Microscopic observation of colons on day 12. Hematoxylin and eosin staining of paraffin sections from WT (left) or Ym1-DTR (right) mice. Scale bar, 100 μm. Original magnification, ×10 (top). Representative images of six mice are shown. Boxed area is observed at a higher magnification (×20, bottom). (H) Il6 and Il1b mRNA expression levels in inflamed colon (DSS, day 10) were determined by qRT-PCR. DT was injected intraperitoneally on days 8 and 9. Average values are shown with SD. Student’s t test, n = 3 mice per genotype. Blue-shaded area indicates WT control (C and D). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Ly6Chi monocytes exhibit inflammatory signature only after infiltrating into inflamed colons of colitic mice (26). Then, we purified Ym1-Venus+Ly6Chi and Ym1-VenusLy6Chi monocytes from the colon in DSS-induced colitis mice (day 9) and compared gene expression between the two populations. We found that Ym1-Venus–positive monocytes in inflamed colon expressed low levels of Il-6 and high levels of Il-10 compared with Ym1-Venus–negative monocytes, indicating the immunoregulatory phenotype of Ym1-Venus–positive monocytes (Fig. 6E). However, Ym1-Venus–positive monocytes showed higher expression of Nos2, a gene signature associated with a proinflammatory phenotype or classically activated (M1) macrophages as well. Although nitric oxide generates reactive oxidants capable of causing collateral tissue injury, nitric oxide also is reported to play a role in tissue regeneration (27, 28). Therefore, Nos2 expression in Ym1-Venus–positive monocytes can be associated with accelerated recovery from DSS-induced colitis in mice. Ym1-Venus–positive monocytes showed higher expression of inflammation-related genes, such as Tlr2 and Nod2, suggesting that these cells exhibit mixed gene expression profile (Fig. 6E). In addition, we found that Ym1-Venus–positive monocytes expressed higher levels of Slpi, which is essential for resolution of inflammation and tissue repair (Fig. 6E) (29).

The expansion timing and gene expression profile of Ym1-Venus–positive monocytes prompted us to consider the possibility that these cells are involved in the resolution of inflammation and tissue regeneration in the inflamed colon. To examine this concept, we generated mice harboring the diphtheria toxin receptor (DTR) gene in the Ym1 loci (Ym1-DTR mice, fig. S6, A and B). Diphtheria toxin (DT) injection deleted only neutrophils and not monocytes in naïve mice (fig. S6, C and D). However, when these mice were injected with DT 48 hours after LPS injection, the proportion of monocytes was decreased from 4 to 2% (fig. S7, A and B), indicating that Ym1+Ly6Chi monocytes were deleted by DT. Selective depletion of Ym1(DTR)-expressing cells was validated by flow cytometry (fig. S7, C and D). Notably, the deletion of Ym1-positive cells in the recovery phase of colitis delayed body weight recovery (Fig. 6F). Histological examination revealed prolonged inflammation and delayed tissue repair in the absence of Ym1-positive cells (Fig. 6G). Tissue cytokine levels were also increased by the deletion of Ym1-positive cells (Fig. 6H). In addition, the number of IL-10–producing T cells in Ym1-DTR mice was lower than that in WT mice (fig. S7E). Because neutrophils are known to perform the critical tasks of tissue repair in a sterile liver injury model (30), we examined the contribution of Ym1+ neutrophils to the recovery from DSS-induced colitis by selectively deleting neutrophils by anti-Ly6G Ab injection in the recovery phase of colitis. Deletion of only neutrophils had no effect on recovery from colitis (fig. S8), indicating that the delayed recovery observed in Ym1-DTR mice was caused not by Ym1+ neutrophil deletion but by Ym1+ monocyte deletion. However, this finding does not exclude the possibility that Ym1+ monocytes and neutrophils coordinately participate in the resolution of the intestinal inflammation. Collectively, these findings demonstrate that Ym1+Ly6Chi monocytes contribute to the resolution of inflammation and regeneration of injured tissue in a mouse model of intestinal inflammation.

Ym1+Ly6Chi monocytes is originated from monocyte precursors

We next sought to explore the kind of precursor cells the Ym1+Ly6Chi monocytes originated from. We found that a small population of cMoP expressed Ym1-Venus in the bone marrow of LPS-injected mice (about 6% of R2 fraction; Fig. 7A), indicating the cMoP origin of Ym1+ cells. This finding prompted us to perform transfer experiments with granulocyte progenitor (GP) or cMoP (18, 31). We first confirmed that GP or cMoP from a CD45.2+ donor became Ly6G+CD115 neutrophils or Ly6GCD115+ monocytes, respectively, in a CD45.1+ recipient (Fig. 7B). Next, we transferred GP or cMoP from CD45.2+ naïve Ym1-Venus mouse bone marrow into LPS-injected CD45.1+ congenic mice and analyzed blood cells in the recipient mice 48 hours later. As shown in Fig. 7C, the transfer of cMoP, but not GP, gave rise to Ym1-Venus+Ly6Chi monocytes. It has been reported recently that monocytes can be differentiated not only from MDP but also directly from GMP (32). Adoptive transfer of GMP or MDP prepared from the bone marrow (fig. S9) of Ym1-Venus mice into WT mice revealed that GMP, but not MDP, has the ability to generate Ym1+ monocytes (Fig. 7D). We examined the possibility that Ym1-Venus–negative monocytes could convert directly into Ym1-Venus–positive monocytes. Ym1-Venus–negative monocytes were isolated from the bone marrow of naïve Ym1-Venus mice (CD45.2+) and transferred into CD45.1+ congenic mice that were preinjected with LPS. CD45.2+ monocytes were detected in the Ly6Chi monocyte population but did not become Ym1-Venus–positive (Fig. 7E), suggesting that Ym1-Venus–positive monocytes are not derived from Ym1-Venus–negative counterpart but may be from monocyte precursor cells probably programed to become Ym1-Venus–positive monocytes.

Fig. 7 Ym1-Venus+Ly6Chi monocytes are derived from monocyte precursors.

(A) Ym1-Venus expression in monocyte progenitors from LPS-injected (48 hours) Ym1-Venus mouse bone marrow. (B) GP (Lin, c-Kit+, CD16/32+, Ly6C+, and CD115) and cMoP (Lin, c-Kit+, CD16/32+, Ly6C+, and CD115+) were purified from Ym1-Venus mice (CD45.2+) bone marrow and transferred into a naïve CD45.1+ WT mouse. Percentages of CD45.2+Ly6G+ neutrophils and CD45.2+CD115+ monocytes were examined by flow cytometry 48 hours after transfer. GP cells (2 × 104) or cMoP cells (1 × 105) were transferred. Representative FACS plots of two independent experiments are shown. (C) Ym1-positive monocytes differentiated from monocyte progenitor in vivo. GP cells (4 × 104) or cMoP cells (1.7 × 105) were adoptively transferred into LPS-injected (50 μg per mouse) CD45.1+ WT congenic mice. Ym1-Venus expression in CD45.2+Ly6G+ neutrophils and CD45.2+CD115+Ly6Chi or CD45.2+CD115+Ly6Clo monocytes was examined by flow cytometry 48 hours after transfer. Representative FACS plots of two independent experiments are shown. (D) Ym1-positive monocytes differentiated from GMP in vivo. GMP cells (3.1 × 104) or MDP cells (2.7 × 104) were adoptively transferred into LPS-injected (50 μg per mouse) CD45.1+ WT congenic mice. Ym1-Venus expression in CD45.2+CD115+Ly6Chi or CD45.2+CD115+Ly6Clo monocytes was examined by flow cytometry 72 hours after transfer. Representative FACS plots of two independent experiments are shown. (E) Ym1-negative monocytes were not converted into Ym1-positive monocytes in vivo. CD45.2+Ly6Chi monocytes from naïve Ym1-Venus mouse bone marrow were transferred into LPS-injected (50 μg per mouse) CD45.1+ WT mice. Ym1-Venus expression in CD45.2+CD115+Ly6Chi monocytes was examined by flow cytometry 14 to 48 hours after transfer. Five hundred thousand monocytes were transferred. Representative FACS plots of two independent experiments are shown. (F and G) GM-CSF and/or IL-3 axis drives Ym1 expression in cMoP. (F) cMoP from naïve Ym1-Venus mouse bone marrow was purified, and 3000 cells were cultured in the presence of indicated cytokines. After 72 hours, the cultured cells were stained with anti-Ly6C and anti-F4/80 Abs. Representative FACS plots of two independent experiments are shown. (G) Average number of total cultured cells (top) or Ym1-Venus–positive cells (bottom) is shown with SD. Representative FACS plots of two independent experiments are shown. Blue-shaded area indicates WT control. n = 3 to 4. iv, intravenous injection.

To further clarify the mechanisms by which cMoP differentiates into Ym1-Venus+Ly6Chi monocytes, we stimulated cMoP from Ym1-Venus mice with various cytokines ex vivo and examined Ym1-Venus expression in these cells. In the presence of macrophage colony-stimulating factor (M-CSF), cMoP proliferated and mainly differentiated into F4/80+ cells that did not express Ym1-Venus (Fig. 7F). In contrast, granulocyte-macrophage CSF (GM-CSF) exposure of cMoP gave rise to F4/80+ cells and Ly6ChiF4/80 cells, which included Ym1-Venus–positive cells (Fig. 7, F and G). M-CSF enhanced the GM-CSF–induced cMoP differentiation into Ly6ChiF4/80 cells, which included Ym1-Venus–positive cells. We also found that IL-3 with M-CSF gave rise to the same differentiation pattern as observed with GM-CSF + M-CSF. Ym1 expression was not induced by IL-4 or IL-10. These results suggest that the GM-CSF and/or IL-3 axis is involved in the differentiation of cMoP into Ym1-positive immunoregulatory monocytes.

DISCUSSION

It is generally considered that monocytes recruited to an injured site exhibit the proinflammatory phenotype and contribute to the propagation of inflammation. In the case of colitis, blockade of CCR2 inhibits monocyte migration into inflamed colon, and consequently ameliorates DSS-induced colitis, demonstrating the proinflammatory roles of migrating monocytes in the early phase of inflammation (26). CCR2 deficiency also results in the amelioration of fibrosis associated with epithelial injury in the lung (33). On the other hand, the roles of monocytes in the resolution of inflammation and tissue regeneration have been highlighted recently. Lucas et al. (34) reported that the deletion of macrophages and monocytes in different phases of a skin injury model yields distinct outcomes. In this model, the deletion of macrophages and recruited monocytes in the tissue formation phase causes severe hemorrhage in the wounded tissue and stagnation of regeneration. Grainger et al. (35) demonstrated that Ly6Chi monocytes in inflamed colon of a Toxoplasma gondii infection model adopted a mixed phenotype, acquiring both inflammatory and regulatory features, and inhibited neutrophil activation by producing prostaglandin E2. In this infection model, CCR2 deficiency caused severe mucosal inflammation by activated neutrophils, demonstrating the regulatory functions of monocytes at the inflamed site. Very recently, atypical monocytes that are involved in detrimental fibrosis of the lung were identified (36). This monocyte population expanded not in the acute inflammatory phase but in the late fibrotic phase of bleomycin-induced lung injury. These reports demonstrated that monocytes play opposite roles in a context- or time-dependent manner. In this study, we identified a subpopulation of Ly6Chi monocytes whose phenotype is biased toward immune regulation. This Ym1+Ly6Chi monocyte subpopulation, expanding during the recovery phase of tissue injury or inflammation, contributes to the resolution of inflammation in a mouse model of colitis. These results demonstrate that monocytes/macrophages responsible for the resolution of inflammation and tissue repair are, at least in part, derived from distinct circulating monocyte subsets generated in bone marrow. We showed that Ym1+Ly6Chi monocytes in the inflamed colon expressed higher levels of IL-10 than Ym1Ly6Chi monocytes, suggesting their intrinsic ability to directly regulate inflammation. However, we cannot exclude the possibility that Ym1+Ly6Chi monocytes have some additional ability to help the resident macrophages heal injured tissue. Very recently, Yáñez et al. (32) reported that monocytes with different characters were derived from bone marrow progenitors via two independent pathways. The relationship between these monocytes and Ym1+ monocytes in this manuscript needs to be clarified.

Under inflammatory conditions, signals from an inflamed site stimulate myeloid cells and their progenitors in bone marrow and alter myelopoiesis both in quantity and in quality to meet the demands in the periphery (3739). For example, granulocyte CSF (G-CSF) is locally produced at an inflamed site in lung infection and acts to enhance the release of neutrophils from bone marrow to periphery (40). During fungal infection, granulopoiesis in bone marrow takes place in a CCAAT/enhancer binding protein (C/EBP)β–dependent manner, whereas C/EBPα is involved in granulopoiesis under steady-state conditions, suggesting that a certain signal produced at the infected sites makes an influence on hematopoiesis in bone marrow (41). In Listeria monocytogenes infection, monopoiesis in bone marrow is enhanced presumably by bacterial components in a Toll-like receptor (TLR)–dependent manner (42). All these findings demonstrate the roles of locally produced cytokines or bacterial components at the inflamed site on emergence of hematopoiesis. In this study, we found that the Ym1+Ly6Chi monocyte subpopulation expanded only in the recovery phase of inflammation or tissue injury. This suggests the involvement of environmental signals from the inflamed site in expanding of this monocyte subpopulation. The in vitro differentiation assay in this study suggests the involvement of GM-CSF/IL-3 signaling in Ym1+Ly6Chi monocyte differentiation in bone marrow (Fig. 7, F and G). Further analysis is needed to identify the endogenous stimulus responsible for the expansion of Ym1+Ly6Chi monocyte subpopulation in vivo under inflammatory conditions.

Here, we analyzed the gene expression profiles of Ym1+Ly6Chi monocytes and found that these cells exhibited overt immunoregulatory and/or tissue reparative phenotype only after infiltrating injured tissue. These results indicate that the phenotype of Ym1+Ly6Chi monocytes is regulated not only by a mechanism that controls the differentiation of myeloid cells in bone marrow but also by environmental signals at injured sites. Macrophages are characterized by considerable diversity and plasticity and have the ability to change their phenotype in response to local signals (4345). Under inflammatory conditions, Ly6Chi monocytes acquire the proinflammatory phenotype through TLR or nucleotide-binding oligomerization domain (NOD) after infiltrating into the inflamed colon (26). Seo et al. (46) also reported that specific kinds of microbiota activate NLRP3 inflammasome in recruited monocytes and induce IL-1β release in infiltrating monocytes in the inflamed colon. All these reports demonstrate that monocytes and macrophages have the ability to alter their phenotype flexibly in response to environmental cues derived from microbiota or damaged cells. In this regard, it is most likely that injured tissues or dying cells produce or release a certain molecule to dictate the phenotype of Ym1+Ly6Chi monocytes during the recovery phase of tissue injury. It is reported that the phagocytosis of cell debris stimulates macrophages to produce Wnt3a and Wnt7a, which are involved in hepatic regeneration by regulating hepatic progenitor cell fate (47). This finding suggests that dead cell clearance may act as a trigger to dictate the phenotype of Ym1+Ly6Chi monocytes. In any case, the molecular mechanisms by which macrophages adopt a specific phenotype during tissue recovery should be revealed to understand the roles of immunoregulatory monocytes in injured tissue. In this study, we could not identify an environmental factor that drives the expansion of this monocyte in vivo. It also remains to be determined whether bone marrow–derived Ym1+ monocytes differentiate into tissue-resident macrophages in inflamed organs.

In conclusion, we identified emerging monocytes that can act in immunoregulation and tissue repair. Elucidating the mechanism by which Ym1+Ly6Chi monocytes control inflammation and tissue repair will enhance our understanding of the overall process of tissue injury and subsequent regeneration.

MATERIALS AND METHODS

Study design

The objective of this study was to identify a monocyte subset with a capacity to promote tissue regeneration in vivo. To achieve this goal, we designed and performed experiments in cellular immunology and animal study. The sample size (n = 3 to 8) for the in vivo experiments was determined to be the optimal size for statistical analysis while using an appropriate number of laboratory mice and allowing for independent repeats.

Mice

C57BL/6J (7 to 12 weeks old) mice were obtained from CLEA Japan Inc. All mice were housed under specific pathogen-free conditions in the Tokyo University of Pharmacy and Life Sciences animal facility. All experiments using the mice described herein were approved by the Tokyo University of Pharmacy and Life Sciences Animal Use Committee (L15-03, L16-14, and L17-24) and performed in accordance with applicable guidelines and regulations.

Reagents

LPS (Escherichia coli, O111:B4) and phorbol 12-myristate 13-acetate (PMA) were purchased from Sigma-Aldrich. Ionomycin was purchased from Calbiochem. 7-Amino-actinomycin D (7-AAD), PerCP/Cy5.5 Streptavidin, Brilliant Violet 421 Streptavidin, fixation buffer, Intracellular Staining Perm Wash Buffer, brefeldin A, and the FOXP3 Fix/Perm Buffer Set were purchased from BioLegend.

4′,6-Diamidino-2-phenylindole (DAPI) was purchased from Dojindo. The Diff-Quik staining set was purchased from Sysmex. Ethidium monoazide bromide (EMA) was purchased from Thermo Fisher Scientific. Abs are summarized in table S1.

Cell isolation using magnetic sorting and cell sorter

For monocyte isolation, bone marrow cells from Ym1-Venus mice were incubated with anti-CD16/32 and then with a cocktail of biotinylated anti-Lin (CD4, CD8, NK1.1, B220, and Ter119) Abs, followed by incubation with anti-biotin microbeads (Miltenyi, Germany). Lin+ cells were depleted by magnetic sorting (AutoMACS Pro, Miltenyi). Lin cells were stained with anti-CD115 and anti-Ly6C Abs and then fractionated by a cell sorter (SH800; Sony, Japan).

GP, cMoP, GMP, and MDP isolation was performed according to a method established previously (18, 31, 32) with slight modification. Bone marrow cells from Ym1-Venus mice were incubated with a cocktail of biotinylated anti-Lin (CD4, CD8, NK1.1, B220, Ter119, Sca-1, CD11b, and Gr-1) Abs, followed by incubation with anti-biotin microbeads (Miltenyi, Germany). Lin+ cells were depleted by magnetic sorting (AutoMACS Pro, Miltenyi). Lin cells were stained with anti–c-Kit, anti-CD16/32, anti-Ly6C, anti-Flt3, and anti-CD115 Abs and Brilliant Violet 421 streptavidin and then fractionated by a cell sorter (SH800; Sony, Japan). GP (Lin, c-Kit+, CD16/32+, Ly6C+, and CD115), cMoP (Lin, c-Kit+, CD16/32+, Ly6C+, and CD115+). Gating strategy of GMP and MDP is shown in fig. S9.

Total RNA extraction for quantitative reverse transcription PCR and RNA sequencing

Total RNA from sorted cells was extracted with the RNeasy Mini or Micro Kit (Qiagen, Netherlands), the FavorPrep Total RNA Extraction Column (Favorgen, Taiwan), or TRIzol LS (Thermo Fisher Scientific, MA) according to the manufacturer’s protocol. For quantitative reverse transcription PCR (qRT-PCR), complementary DNAs (cDNAs) were synthesized using ReverTra Ace (Toyobo, Japan). qRT-PCR was performed on cDNA using the THUNDERBIRD SYBR qPCR Mix (Toyobo, Japan). Expression levels were normalized to 18S ribosomal RNA. Primer sequences are summarized in table S2.

RNA sequencing

Sorted cells were lysed, and their total RNA was extracted with TRIzol LS reagent. Forty nanograms of total RNA was subjected to DNA library preparation for RNA sequencing (RNA-seq) analysis using the SureSelect Strand-Specific RNA Library Prep Kit for Illumina Multiplexed Sequencing (Agilent Technologies, CA). Sequencing was performed on a HiSeq 1500 sequencer (Illumina) in the 50–base pair single-end read mode. Data with the fragments per kilobase of exon per million reads were used for further analysis after mapping of the sequence reads.

In vitro monocyte stimulation

The sorted monocytes were stimulated for 24 hours with LPS (E. coli, O111:B4, 1 μg/ml; Sigma, MO). Concentrations of IL-6, IL-12, TNFα, IL-10, and Ym1 in the culture medium were quantitated with the enzyme-linked immunosorbent assay (ELISA) MAX Standard Kit (IL-6, IL-12, TNFα, and IL-10; BioLegend, CA) or the ELISA system established in our laboratory (Ym1).

DSS-induced colitis

Mice were orally administered 2.0% DSS (36,000 to 50,000 Da; MP Biomedicals, CA) in drinking water for 5 days and in regular drinking water thereafter. Body weight was monitored daily or every other day for 12 days from the administration of DSS. In some experiments, 25 ng/g body weight of DT was intraperitoneally injected on indicated days after the administration of DSS. For the depletion of neutrophils, 50 μg of anti-Ly6G Ab (clone 1A8; Bio X Cell, NH) was intraperitoneally injected on days 8 and 10.

Assay for cMoP differentiation

Sorted cMoP (3000 cells per well in 96-well plates) was cultured with various cytokine combinations: 10% M-CSF (CMG14-12 culture medium), G-CSF (50 ng/ml; Chugai, Japan), recombinant mouse GM-CSF (20 ng/ml; BioLegend), IL-3 (10 ng/ml; Peprotech, NJ), recombinant mouse IL-4 (20 ng/ml; PeproTech, NJ), or recombinant mouse IL-10 (100 ng/ml; Miltenyi, Germany). The cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 1% penicillin-streptomycin (Wako), and 10 mM Hepes. After a 72-hour culture, the cells were analyzed by flow cytometry.

Intracellular flow cytometry staining

For intracellular Foxp3 staining, CD4 T cells were enriched from the spleen by magnetic sorting. Those T cells were stimulated with immobilized anti-CD3e (10 μg/ml) and anti-CD28 (5 μg/ml) in the presence or absence of monocyte culture media for 72 hours in a 96-well plate. After stimulation, T cells were stained with anti-CD4 and EMA and then fixed and permeabilized in Fix/Perm Buffer (BioLegend). For intracellular IL-10 staining, CD4 T cells were enriched from the colon of DSS-administered mice by magnetic sorting. Those T cells were stimulated with PMA (100 ng/ml) and ionomycin (1.5 μg/ml) for 3 hours in the presence of brefeldin A. After stimulation, those cells were fixed in fixation buffer and permeabilized in Intracellular Staining Perm Wash Buffer (BioLegend). Fixed/permeabilized cells were stained with anti-Foxp3 or anti–IL-10 Ab, or isotype-matched control Ab, and analyzed by FACSVerse.

Statistical analysis

Data were analyzed either by analysis of variance (ANOVA), followed by multiple comparison or by the paired or unpaired t test with GraphPad Prism (GraphPad Software, CA). P < 0.05 was considered significant.

SUPPLEMENTARY MATERIALS

immunology.sciencemag.org/cgi/content/full/3/28/eaat0207/DC1

Materials and Methods

Fig. S1. Generation of Ym1-Venus mice.

Fig. S2. Ym1 expression at different stages of neutrophil maturation.

Fig. S3. Inflammatory cytokine levels in LPS-treated mice.

Fig. S4. Hierarchical cluster analysis and principal component analysis of Ym1-Venus+Ly6Chi monocytes.

Fig. S5. Expansion of Ym1+Ly6Chi monocyte subpopulation in the spleen during recovery phase of colitis.

Fig. S6. Generation of Ym1-DTR mice.

Fig. S7. DT injection depletes Ly6Chi monocytes and neutrophils in Ym1-DTR mice under inflammatory conditions.

Fig. S8. Depletion of neutrophils does not affect recovery from colitis.

Fig. S9. Gating strategy of GMP and MDP.

Table S1. Antibodies.

Table S2. Primers for qRT-PCR.

Table S3. Primers for genotyping PCR.

Table S4. Raw data (Excel file).

References (4853)

REFERENCES AND NOTES

Acknowledgments: We thank S. Nagata for pEF-BOS-EX plasmid, A. Kudo for CMG14-12 cells (M-CSF–expressing cell line), B. Diamond for NSObcl2 cells, A. Miyawaki for Venus cDNA, and K. Kohno for Human heparin-binding EGF-like growth factor cDNA. We thank T. Suito and M. Kawana for secretarial assistance. Funding: This work was supported in part by a Grants-in-Aid for Scientific Research (B) (26293089, 17H04068, and 17H04052) from Japan Society for the Promotion of Science, a Grant-in-Aid for Scientific Research on Innovative Areas (Homeostatic regulation by various types of cell death) (26110006) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) in Japan, MEXT-Supported Program for the Strategic Research Foundation at Private Universities (2014–2019) in Japan, the Uehara Memorial Foundation, the Takeda Science Foundation, and the Naito Foundation. Author contributions: N.I. performed most of experiments with help from K.A., K.K., H.I., R.T., T.S., and S.Y. Y.U., M.O., K.A., and G.N. generated gene-targeting mice. C.M.-O., H.F., and T.W. performed RNA-seq and analyzed the data. N.I., K.K., and K.A. performed statistical analysis. N.I., K.A., G.N., and M.T. contributed to study design and figure preparation and wrote the manuscript. Competing interests: The authors declare that they have no competing financial interests. Data and materials availability: The RNA-seq data for this study have been deposited in the Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo/). The accession code for RNA-seq data is GSE118032.
View Abstract

Stay Connected to Science Immunology

Navigate This Article