Differential expression of CD11c defines two types of tissue-resident macrophages with different origins in steady-state salivary glands

Gland macrophages are primed for gland development and functions through interactions within their niche. However, the phenotype, ontogeny, and function of steady-state salivary gland (SG) macrophages remain unclear. We herein identified CD11c+ and CD11c− subsets among CD64+ macrophages in steady-state murine SGs. CD11c− macrophages were predominant in the SGs of embryonic and newborn mice and decreased with advancing age. CD11c+ macrophages were rarely detected in the embryonic period, but rapidly expanded after birth. CD11c+, but not CD11c−, macrophage numbers decreased in mice treated with a CCR2 antagonist, suggesting that CD11c+ macrophages accumulate from bone marrow-derived progenitors in a CCR2-dependent manner, whereas CD11c− macrophages were derived from embryonic progenitors in SGs. CD11c+ and CD11c− macrophages strongly expressed colony-stimulating factor (CSF)-1 receptor, the injection of an anti-CSF-1 receptor blocking antibody markedly reduced both subsets, and SGs strongly expressed CSF-1, indicating the dependency of SG resident macrophage development on CSF-1. The phagocytic activity of SG macrophages was extremely weak; however, the gene expression profile of SG macrophages indicated that SG macrophages regulate gland development and functions in SGs. These results suggest that SG CD11c+ and CD11c− macrophages are developed and instructed to perform SG-specific functions in steady-state SGs.


Results
Phenotype and localization of macrophage subsets in steady-state SGs. Cells of the submandibular glands (SMGs) from wild-type mice were isolated by enzymatic digestion using collagenase and analyzed by flow cytometry. By combining the expression of CD64 and CD11c, CD45 + MHCII + APCs in SMGs were separated into CD64 + macrophages and CD64 − CD11c + cDCs (Fig. 1a). CD64 + macrophages were further classified into the CD11c + and CD11c − subsets. Approximately 90% of APCs in SMGs were macrophages, and CD11c + macrophages were predominant in the SMGs of adult mice. Since the composition of cDCs and CD11c + and CD11c − macrophages in the sublingual and parotid glands was similar to that in SMGs (Fig. 1a,b), we mainly examined SMGs as representative SGs in subsequent experiments.
CD11c + and CD11c − macrophages equally expressed MHCII, CD11b, the macrophage marker F4/80, and CD14 (Fig. 1c). CD11c − macrophages also more strongly expressed MerTK and the mannose receptor CD206 than CD11c + macrophages, while CX3CR1 was strongly expressed in CD11c + macrophages. These results indicated that macrophages exist as predominant APCs in the steady-state SGs of adult mice and that SG macrophages comprise CD11c + and CD11c − subsets, which may have different ontogenies and functions.
To clarify the localization of cDCs and macrophages in SGs, we examined their frequencies in whole SMG, enriched ductal, and remaining interstitial fractions. Most macrophages existed in the interstitial region, while cDCs were enriched by approximately sevenfold in the ductal region (Fig. 2a,b). CD11c + macrophages were predominant in the ductal and interstitial regions; however, the frequency of CD11c − macrophages slightly increased in the ductal region (Fig. 2a,c), indicating that most CD11c + and CD11c − macrophages reside in the interstitial and ductal regions and also that cDCs and some CD11c − macrophages are potently associated with the ductal region. In support of these results, F4/80 + CD11c − and F4/80 + CD11c + macrophages as well as F4/80 − CD11c + cDCs were detected in the ductal and stromal regions by immunohistochemistry (Fig. 2d). Immunohistochemistry also revealed no significant differences in the morphologies of CD11c + and CD11c − macrophages.
Age-dependent transition of macrophage subsets in steady-state SMGs. To examine the ontogeny of SG macrophage subsets, we analyzed macrophages in steady-state SMGs from mice at E13.5, E17.5, and E19 (embryo) and at 3 days (newborn) and 1,2,3,4,8,12, and 24 weeks after birth by flow cytometry. The results obtained showed that CD11c − macrophages were a major population in the SMGs of the embryonic period and newborn mice, and then decreased with advancing age after birth (Fig. 3a). Inversely, CD11c + macrophages were rarely detected in the SMGs of the embryonic period and newborn mice, and then rapidly expanded with advancing age. The proportion of CD11c + macrophages in young (4 weeks old) mice was similar to that in adult (8-24 weeks old) mice (Fig. 3b). The cell numbers of CD11c + and CD11c − macrophages both rapidly expanded concomitantly with the growth of SMGs at 4 weeks (Fig. 3c); however, the number of CD11c − macrophages per tissue weight slightly decreased after birth (Fig. 3d). In addition, the expression of MHCII in CD11c + and CD11c − macrophages increased 2 and 4 weeks after birth, respectively (Fig. 3e). The number of CD11c + cDCs also slightly increased at 2-4 weeks ( Supplementary Fig. S1). Therefore, marked changes in the composition of macrophage subsets and their maturation status occurred with SG development.
HSC dependence of SMG CD11c + macrophages. The result showing that CD11c + macrophages expanded in SMGs after birth suggested that CD11c + macrophages differentiated from BM-derived monocytes. Since the emigration of monocytes from BM requires signals mediated by the chemokine receptor CCR2 22 , we examined the effects of a CCR2 antagonist on the accumulation of macrophages in steady-state SMGs. The treatment of mice with a CCR2 antagonist reduced the percentage of CD11c + macrophages (Fig. 4a,b). The number of CD11c + macrophages was also significantly reduced by the treatment, whereas that of CD11c − macrophages was not (Fig. 4c). Furthermore, 21 days after CD45.1 + BM cell transplantation into sublethal irradiated CD45.2 + mice, differentiated donor-derived macrophages were mainly the CD11c + phenotype (Fig. 4d,e); however, a small number of CD11c − macrophages was detected 7 and 21 days after the BM transplantation. These results suggest that CD11c + macrophages mainly accumulated from HSC-derived progenitors in a CCR2-dependent manner, while CD11c − macrophages were mostly derived from embryonic progenitors and self-renewed in SGs.
Colony-stimulating factor (CSF)-1-dependent accumulation of SMG macrophages. CSF www.nature.com/scientificreports/ in the spleen and capable of maintaining splenic macrophages locally 23 , whereas the production of CSF-2 in the lungs is required for the development of alveolar macrophages 24,25 . To examine whether CSF-1 or CSF-2 influences the development of SG macrophages, SMG CD11c + and CD11c − macrophages and splenic and alveolar macrophages were sorted ( Supplementary Fig. S1), and the expression of Csf1r and Csf2rb mRNA was examined to assess CSF-1 receptor (CSF-1R) and CSF-2R expression in these macrophages. Splenic and alveolar macrophages were used as positive controls for the expression of Csf1r and Csf2rb mRNA, respectively. The results obtained showed that SMG CD11c + and CD11c − macrophages strongly expressed Csf1r, but not Csf2rb (Fig. 5a). We also measured CSF-1 and CSF-2 expression levels in SMGs. Homogenates of the spleen and lungs were used as positive controls for the expression of Csf1 and Csf2 mRNA, respectively. The expression of Csf1 was significantly stronger in SMGs than in the spleen, whereas that of Csf2 was significantly weaker in SMGs than in the lungs (Fig. 5b), suggesting the dependency of SG CD11c + and CD11c − macrophage development on CSF-1.
To confirm this, an anti-CSF-1R blocking antibody was intraperitoneally injected. The results obtained showed that the percentages and total cell numbers of SMG CD11c + and CD11c − macrophages markedly decreased (Fig. 5c,d), and the injection of anti-CSF-1R did not influence the number of SMG cDCs (Fig. 5d) or body and SMG weights (Fig. 5e,f). Collectively, these results clearly indicate that the production of CSF-1 in SGs plays a critical role in the maintenance and survival of CSF-1R-expressing SG resident macrophages.        Figure 4. HSC dependency of macrophage subsets in steady-state SMGs. (a-c) Macrophage subsets in steady-state SMGs from control or CCR2 antagonist-treated mice. Newborn mice were administered the CCR2 antagonist from postnatal day 3 for 4 weeks, and SMG tissues were then harvested. Representative FACS plots (a). Data are representative of three independent experiments. The proportion (b) and absolute numbers (c) of CD11c + and CD11c − macrophages in steady-state SMGs from control or CCR2 antagonist-treated mice. (d,e) Sublethally irradiated female CD45.2 + mice (7 weeks old) mice were adoptively transferred with CD45.1 + BM cells. SMG tissues were harvested at the indicated time points and the percentage of donor-derived (CD45.1 + ) cDCs and macrophage subsets were analyzed by flow cytometry. Representative FACS plots (d) and the percentages of the CD11c + and CD11c − subsets in donor macrophages (e) are shown. Bars represent the mean ± SD (n = 3). *P < 0.05 by the Student's t test with Welch's correction. Results are representative of three independent experiments. Mϕ macrophage, NS not significant.  26 , and the number of CD11c + macrophages increased in mice aged 2-4 weeks (Fig. 3a,b) when they began to be weaned and eat the standard diet, we investigated whether mastication affected the development of CD11c + macrophages. One-week-old mice and their mothers were fed standard pellets or a powdered diet. After 8 weeks, the proportion of CD11c + and CD11c − macrophages in SMGs was analyzed. No significant differences were observed in the percentage or number of CD11c + macrophages between the pellet and powdered diet groups (Fig. 6a-c). Body and SMG weights were also unchanged between the groups (Fig. 6d,e). These results indicated that mastication did not affect the development of SG resident macrophages.
Functional perspectives of SG macrophages. The gene expression profiles of SMG CD11c + macrophages, alveolar macrophages, and splenic macrophages were examined and compared. SMG CD11c − macrophages were not assessed because their number was too small for analysis. Hierarchical clustering demonstrated that SMG CD11c + macrophages did not cluster with splenic and alveolar macrophages (Fig. 7a), suggesting a weak relationship between these macrophages. We then compared the expression of 39 core macrophage signature genes 15 . Normalized relative gene expression demonstrated that SMG CD11c + macrophages expressed www.nature.com/scientificreports/ lower levels of Arsg encoding arylsulfatase G, Mertk encoding MerTK, and Pld3 encoding phospholipase D3 (Fig. 7b). MerTK is a receptor for apoptotic cell uptake 27 , and arylsulfatase G and phospholipase D3 are lysosomal enzymes 28,29 , suggesting that SG macrophages exhibit weak phagocytic activity. An in vitro analysis showed that although SMG CD11c − macrophages phagocytosed BioParticle and apoptotic cells more efficiently than SMG CD11c + macrophages, the phagocytic activities of SMG CD11c + and CD11c − macrophages were markedly weaker than that of peritoneal macrophages (Fig. 7c). Phagocytosis by macrophages was abrogated by the addition of cytochalasin D, the F-actin disrupter 30 , to the culture. A gene ontology (GO) pathway analysis identified 2988 genes that were up-regulated in SMG CD11c + macrophages (Fig. 7d), which displayed enrichment for pathways involved in developmental processes, such as anatomical structure morphogenesis, anatomical structure development, developmental process, tissue development, system development, and single-organism developmental process, and migration processes including locomotion and the regulation of cellular components, cell migration, and cell motility (Fig. 7e). In contrast, 1401 genes were down-regulated in SMG CD11c + macrophages (Fig. 7f), which displayed enrichment for pathways involved in immune system processes, single organism processes, the positive regulation of biological processes, and the positive regulation of cellular processes (Fig. 7g). In comparisons with alveolar and splenic macrophages, 59 and 20 transcription factors were up-and down-regulated by more than fivefold, respectively, in SMG CD11c + macrophages (Fig. 7h,i). Therefore, the transcription profile is unique to SG macrophages, indicating that they are specifically primed to function in SGs.
We also examined the mRNA expression levels of macrophage signature genes in SMG CD11c + and CD11c − macrophages. In addition to splenic and alveolar macrophages, macrophages from the exocrine pancreas (pancreatic stroma) were used as an exocrine control tissue. Consistent with the gene expression profiles shown in Fig. 7b, the mRNA expression levels of Arsg, Mertk, and Pld3 were markedly lower in SMG CD11c + and CD11c − macrophages than in splenic and pancreatic macrophages (Fig. 8). We then assessed canonical pro-(Il1b) and anti-inflammatory gene expression (Arg1, Fizz1, Cd163, Ym1, and Il10) 31 . SMG CD11c − macrophages expressed Arg1, Fizz1, Cd163, Ym1, and Il10, and the expression of Il1b was similar between SMG CD11c + and CD11c − macrophages. Exocrine pancreatic macrophages highly expressed Fizz1 and Cd163, while alveolar macrophages highly expressed Ym1. These results suggest that CD11c − macrophages exhibit immunoregulatory functions in SGs.

Discussion
The present study showed that CD64 + macrophages among CD45 + MHCII + APCs in murine SGs were phenotypically divided into CD11c + and CD11c − subsets. CD11c − macrophages dominantly existed in the SGs of the embryonic period and neonatal mice and decreased after birth. In contrast, CD11c + macrophages appeared after birth in SGs and rapidly expanded with advancing age. Since CD11c + alveolar macrophages develop from CD11c − fetal monocytes in the lungs 19 , SG CD11c − macrophages may give rise to CD11c + macrophages after birth. However, the present study showed that CD11c + macrophages expanded in a CCR2-dependent manner and differentiated from BM-derived progenitors, while CD11c − macrophages did not change in number following a treatment with a CCR2 antagonist and only weakly differentiated from transferred BM cells. A mass cytometric analysis suggested that fetal-derived tissue-resident macrophages in the mammary gland are CD206 Hi cells 32 , and SG CD11c − macrophages preferentially expressed CD206 (Fig. 1c). Therefore, it is conceivable that SG CD11c − macrophages are mainly derived from fetal progenitors and self-renewed in SGs, while CD11c + macrophages differentiated from BM-derived progenitors after birth in SGs.
The present results demonstrated that macrophages in embryo SMGs were CD11c − MHCII − , which is consistent with previous findings showing that embryo-derived cardiac macrophages present at birth were CX3CR1 + MHCII −33 . This study also revealed that the expression of MHCII in SG macrophages started to increase after birth. The expression of MHCII is fine-tuned to function in various parameters, including developmental stages, the activation status, and exposure to extracellular stimuli, and MHCII loads antigenic peptides and guides the development and activation of antigen-specific CD4 + T cells 34,35 . Therefore, it is conceivable that SG macrophages acquire the ability of specific antigen presentation after birth.
The present study showed that most CD11c + and CD11c − macrophages resided in interstitial and ductal regions and that cDCs and some CD11c − macrophages were strongly associated with the ductal region. The result may not be consistent with recent findings showing that CD11c + macrophages were associated with acini and ducts in SGs using the CD11c-yellow fluorescent protein (YFP) reporter strain 36 . We attributed this discrepancy to differences in the gating strategy. We initially separated CD45 + MHCII + APCs in SGs into CD64 + macrophages and CD64 − CD11c + cDCs while they gated CD11c-YFP, which excluded CD11c − macrophages, and then analyzed the expression of CD64.
The present study showed that the development of SG CD11c + and CD11c − macrophages depended on CSF-1R signaling, which supports CSF-1 being released from the serous acinar cells of SGs to function in the survival, recruitment, and proliferation of SG macrophages 4,5,37,38 . The results obtained also showed that SG CD11c + macrophages expanded rapidly in mice aged 2-4 weeks when they began to be weaned and eat the standard diet. Since mechanical stress has been suggested to explain the replacement of embryo-derived macrophages by circulating monocyte-derived macrophages 12,33 , we examined whether mastication influenced the development of SG CD11c + macrophages; however, the powdered diet did not change the proportion of SG resident macrophages. Previous studies also demonstrated that intestinal commensal microbiota controlled macrophage turnover 11,12 , therefore, oral commensal microbiota may promote SG CD11c + macrophage development. In the intestines, TGF-β was identified as indispensable for the differentiation of monocytes into macrophages 39 . Therefore, further studies are needed to identify the SG microenvironmental factors that control macrophage development in order to maintain SG homeostasis.
In mammary glands, which are exocrine glands similar to SGs, macrophages distribute during distinct phases of development and remodeling 40 , and ductal macrophages survey the mammary epithelium and facilitate tissue remodeling 41 , indicating that mammary gland macrophages play an essential role in gland development. However, there is still no clear evidence to show that macrophages contribute to SG development. In mouse SMG morphogenesis, the gland is highly branched by E14, and functional differentiation begins at E15 and continues to birth 42 . The present study showed that CD11c − macrophages were already detected at E13.5 (Fig. 3) and potently associated with the ductal region (Fig. 2). Therefore, CD11c − macrophages may play an essential role in gland morphogenesis in association with the ductal region in fetal SGs. The treatment of SMGs isolated from E12.5 mouse embryos with anti-CSF-1 or anti-CSF-1R reduced the numbers of F4/80 + macrophages and suppressed the development of SMGs; however, the latter effect may be explained by macrophage-independent pathways 43 . The specific depletion of macrophages by novel approaches, such as the use of CD64-diphtheria toxin receptor mice, may contribute to a more detailed understanding of the exact role of macrophages in SG morphogenesis.
In conclusion, the present results indicate that fetal macrophages are present in SGs as early as E13.5 for the development of SGs, and that SG macrophages continues to function after birth in steady-state SGs by inducing proliferation of fetal-derived macrophages and receiving supplies from BM-derived progenitors. CD11c is a key marker that distinguishes between BM-and embryonic-derived SG macrophages. The present study showed that the phagocytic activity of SG macrophages was extremely weak; however, the gene expression profile of SMG macrophages suggested that SG macrophages regulate gland development and functions in SGs. This view is partly supported by recent findings indicating that murine SG macrophages allow the patrolling of tissue by tissue-resident memory CD8 + T cells for homeostatic organ surveillance 36 , and attenuate radiotherapy-induced dry mouth through the activation of the Hedgehog pathway 44 . The methods employed in the present study do not provide the same level of evidence as other experiments, such as lineage tracing protocols or parabiosis. Therefore, the further characterization of SG macrophages is important for evaluating the contribution of SG macrophage subpopulations to SG homeostasis and to the pathology of SG dysfunctions and their therapeutic purposes.

Materials and methods
Mice. CD45.2 + C57BL/6N mice were purchased from CLEA Japan (Tokyo, Japan). Male and female mice were placed into the same cage at 8 p.m. Mice were separated the next morning and regarded as embryonic development day 0.5 (E0.5). Congenic CD45.1 + C57BL/6 mice were provided by RIKEN BRC (Tsukuba, Japan). This study was carried out in compliance with the ARRIVE Essential 10 Guidelines.

Scientific Reports
Preparation of single-cell suspensions. Mice were euthanized with isoflurane and then perfused with 20 ml of phosphate-buffered saline (PBS) through the left ventricle to remove circulating blood. SG cells were purified following previously described protocols 17 , and subjected to a flow cytometric analysis. To enrich ducts, SMGs were digested with collagenase, hyaluronidase, and DNase for 1 h and pipetted with a 25-ml pipette. Ducts were manually selected using a dissection microscope and processed to single cells as described above.
The spleen and lungs were cut into small pieces. Spleen pieces were incubated with 1 mg/ml of collagenase D (Roche, Basel, Switzerland) and 0.1 mg/ml of DNase I in RPMI 1640 medium containing 10% fetal bovine serum (FBS) at 37 °C for 30 min with shaking. Spleen cells were depleted of red blood cells by hypotonic lysis. Lung pieces were incubated with 0.4 mg/ml of collagenase type IV (Sigma-Aldrich, St Louis, MO, USA) in RPMI 1640 medium containing 10% FBS at 37 °C for 45 min with shaking. Resident peritoneal macrophages were isolated by an injection of 5 ml of ice-cold PBS into the peritoneal cavity. Pancreatic cells were isolated according to previously described protocols 45 . Briefly, the retrograde infusion of 1 ml of 1 mg/ml collagenase type IV in Hanks' balanced salt solution (HBSS) into the pancreatic duct was performed. The inflated pancreas was then removed and digested in HBSS at 37 °C for 8.5 min with gentle shaking. After washing with HBSS, the cell suspension was filtered using a 70-μm cell strainer to remove the islets.
Flow cytometry. Flow cytometry was performed as previously described 17 . Cells were treated with an anti-CD16/32 antibody (2.4G2; produced in-house) to block Fc receptors and then stained with antibodies. The antibodies used were listed in Supplementary Table S1. Dead cells were excluded using DAPI (Dojindo, Kumamoto, Japan). Data were acquired on an LSRFortessa cell analyzer (BD Biosciences, San Jose, CA, USA) and analyzed using FlowJo software (Tree Star, Ashland, OR, USA). Absolute cell numbers were measured by CountBright absolute counting beads (Life Technologies, Carlsbad, CA, USA). Cell sorting was performed using a FACSAria II cell sorter (BD Biosciences). Splenic macrophages were sorted as live gated CD45 + B220 − MHCII + F4/80 + CD64 + CD11c int cells. Alveolar macrophages were sorted as live gated CD45 + MHCII low Siglec-F + CD64 + F4/80 + CD11c + CD11b − cells. Macrophages from the exocrine pancreas (pancreatic stroma) were sorted as live gated CD45 + B220 -MHCII + CD64 + F4/80 + cells ( Supplementary Fig. S3).   CCR2 antagonist treatment in vivo. Newborn mice were administered 50 mg/kg of the CCR2 antagonist propagermanium (Sigma-Aldrich) via daily oral gavage and 0.1 mg/ml propagermanium in drinking water from postnatal day 3 for 4 weeks. The vehicle control group was given an equal volume of tap water. Body and SMG weights were measured and SMG cells were subjected to flow cytometry.

Anti-CSF-1R treatment in vivo.
Anti-CSF-1R mAb was purified from the ascites of nude mice injected with AFS98 hybridoma cells (Riken BRC, Japan). C57BL/6 female mice were injected with 2, 0.5, and 0.5 mg of anti-CSF-1R mAb in 300 μl sterile PBS by an intraperitoneal injection on days 0, 1, and 2, respectively, and then analyzed on day 5. To examine the effects of the early-in-life depletion of SG macrophages, one-week-old mice were intraperitoneally injected with 0.1, 0.025, 0.025, 0.025, and 0.025 mg/g body weight of anti-CSF-1R mAb in 10 μl/g body weight sterile PBS on days 0, 1, 2, 14, and 28, respectively, and then analyzed on day 35. , and Il1b were analyzed by quantitative RT-PCR. Expression levels were normalized to β-actin levels and expressed as relative expression based on the value of splenic macrophages. Bars represent the mean ± SD (n = 2 or 3). If two or more sets of triplicate samples were lower than detection limit, results were expressed as "not detected (ND)". S splenic macrophages, A alveolar macrophages, P pancreatic macrophages, 11c + SMG CD11c + macrophages, 11c − SMG CD11c − macrophages. **P < 0.01, *P < 0.05 by a one-way ANOVA using Tukey's multiple comparisons tests. www.nature.com/scientificreports/ cocultured at 37 °C for 1 h. To inhibit phagocytosis, 5 μg/ml cytochalasin D (Wako Pure Chemical, Osaka, Japan) was added to culture media for 45 min as a negative control. Thymocytes were stained with the CellTrace Far Red Cell Proliferation kit (Thermo Fisher Scientific) and then incubated with 1 μM dexamethasone at 37 °C for 6 h to induce apoptosis. Apoptotic cells and SMG CD45 + cells (1 × 10 6 ) were co-cultured at a 1:1 ratio at 37 °C for 3 h. Cells were further analyzed by flow cytometry after the incubation.

Microarray analysis.
Macrophage subsets were sorted from the SMG, lung, and splenic single-cell suspensions of C57BL/6N mice using the FACS Aria II cell sorter (BD Biosciences). Gating strategies are shown in Supporting Information Fig. S1. The total RNA of macrophages was then extracted using a RNeasy Mini Kit (Qiagen). Cyanine-3-labeled cRNA was amplified from 50 ng of total RNA using a Low Input Quick Amp Labeling Kit (Agilent Technologies, Santa Clara, CA, USA). Gene expression profiles were analyzed using Sure-Print G3 Mouse GE 8x60K ver.2.0 (Agilent Technologies). Microarray slides were scanned using a Microarray Scanner (Agilent Technologies), and raw fluorescence intensities were quantified and normalized using Feature Extraction software 12.0.3.1 (Agilent Technologies). Raw signal data were processed using Feature Extraction software 12.0.3.1 (Agilent Technologies). Processed data were analyzed using GeneSpring ver. 13.1.1 (Agilent Technologies) or R (R Foundation for Statistical Computing, Vienna, Austria). The microarray data analysis was supported by Hokkaido System Science (Sapporo, Japan). SMG macrophages from eighteen mice or the remaining macrophage subsets from six mice were pooled as one experimental sample, and three independent samples were performed in the microarray analysis.