Report

DNGR-1 in dendritic cells limits tissue damage by dampening neutrophil recruitment

See allHide authors and affiliations

Science  19 Oct 2018:
Vol. 362, Issue 6412, pp. 351-356
DOI: 10.1126/science.aan8423

The absence of DNGR-1 is dangerous

Conventional type 1 dendritic cells (cDC1s) can sense tissue damage via DNGR-1, which binds F-actin exposed by necrotic cells. DNGR-1 activation favors cross-presentation, the process by which extracellular antigens are processed and presented to CD8+ T cells via major histocompatibility complex class I molecules. Del Fresno et al. studied mice lacking DNGR-1 and found that DNGR-1 also has anti-inflammatory effects (see the Perspective by Salazar and Brown). It inhibits the secretion of the chemokine CXCL2 by cDC1s, which, in turn, limits neutrophil recruitment. Thus, DNGR-1 connects cell-death sensing with a mechanism of damage control.

Science, this issue p. 351; see also p. 292

Abstract

Host injury triggers feedback mechanisms that limit tissue damage. Conventional type 1 dendritic cells (cDC1s) express dendritic cell natural killer lectin group receptor-1 (DNGR-1), encoded by the gene Clec9a, which senses tissue damage and favors cross-presentation of dead-cell material to CD8+ T cells. Here we find that DNGR-1 additionally reduces host-damaging inflammatory responses induced by sterile and infectious tissue injury in mice. DNGR-1 deficiency leads to exacerbated caerulein-induced necrotizing pancreatitis and increased pathology during systemic Candida albicans infection without affecting fungal burden. This effect is B and T cell–independent and attributable to increased neutrophilia in DNGR-1–deficient settings. Mechanistically, DNGR-1 engagement activates SHP-1 and inhibits MIP-2 (encoded by Cxcl2) production by cDC1s during Candida infection. This consequently restrains neutrophil recruitment and promotes disease tolerance. Thus, DNGR-1–mediated sensing of injury by cDC1s serves as a rheostat for the control of tissue damage, innate immunity, and immunopathology.

After sterile or infectious insults, injured tissues expose or release alarm signals that are detected by specific innate immune receptors on myeloid cells (1). This triggers an inflammatory response, which promotes the recruitment of myeloid cells into the damaged organ. This innate immune response must be tightly regulated to avoid additional tissue damage (2).

Among myeloid cell sensors of tissue damage, dendritic cell natural killer lectin group receptor-1(DNGR-1; Clec9a gene) is a C-type lectin receptor (CLR) that detects F-actin exposed by damaged cells (3, 4). DNGR-1 is mainly expressed by mouse and human conventional type 1 dendritic cells (cDC1s), including CD103+CD11b DCs in peripheral tissues (5, 6). DNGR-1 favors the cross-presentation of dead cell–associated antigens to CD8+ T cells (79). However, whether DNGR-1 plays any role in innate immunity is unknown. To address this issue, we used a mouse model of caerulein-induced acute necrotizing pancreatitis (Fig. 1A), which results in massive acinar cell death, leading to the infiltration of myeloid cells. This, in turn, triggers further pathology and edema (10). Upon caerulein treatment, there was increased pancreatic infiltration by neutrophils, but not monocytes, in DNGR-1–deficient (Clec9agfp/gfp) mice compared with that observed in wild-type (WT) mice (Fig. 1B). Neutrophil numbers in the bone marrow (BM) and blood were similar in both genotypes (fig. S1), ruling out an effect of DNGR-1 deficiency on neutrophil ontogeny and suggesting a local recruitment effect.

Fig. 1 DNGR-1 regulates neutrophil infiltration and tissue damage during acute pancreatitis.

(A) Acute pancreatitis was induced by intraperitoneal injection of caerulein hourly for 6 hours in WT and DNGR-1–deficient (Clec9agfp/gfp) mice. Phosphate-buffered saline (PBS) injection was used as a control. Animals were analyzed 12 hours after the last injection. (B) Infiltrating neutrophils (left) and monocytes (right) in pancreas quantified by flow cytometry. (C and D) Anti–DNGR-1 (α-DNGR-1) or isotype control antibodies were intraperitoneally injected into WT (C) or WT and Batf3−/− (D) mice on the day before (day −1) and the day of (day 0) caerulein injection. Infiltrating neutrophils in pancreas were quantified by flow cytometry. (E) Concentrations of lipase were detected in serum from peripheral blood. U/L, units per liter. (F) Hematoxylin and eosin (H&E) staining in pancreatic sections (left). Representative images of n = 5 pancreata per experimental condition. Percentages of edematous area (right). (G and H) Rag1−/− and Rag1−/−Clec9agfp/gfp mice were subjected to pancreatitis as indicated in (A). Infiltrating neutrophils in pancreas quantified by flow cytometry (G) and serum lipase concentrations (H) are shown. In (B) to (H), each dot represents a single mouse. Data are means ± SEM of a representative experiment (N ≥ 2 independent experiments). Significance was assessed by unpaired Student’s t test between genotypes (B), (E), (F), (G), and (H) or treatments (C) and (D); *P < 0.05, and **P < 0.01.

As a nongenetic approach, we used a DNGR-1–blocking antibody (7, 11). Receptor blockade phenocopied the exacerbated pancreatic infiltration of neutrophils (Fig. 1C) but not monocytes (fig. S2A). Enhanced neutrophilia upon DNGR-1 blockade was lost in Batf3−/− mice (Fig. 1D and fig. S2B), which lack functional cDC1s (12), indicating that cDC1s are the key mediators. Pancreatic CXCR2-mediated neutrophil infiltration is pathological in acute pancreatitis (13). Consistently, caerulein-treated Clec9agfp/gfp mice displayed exacerbated pancreatitis with increased serum lipase concentrations (Fig. 1E) and extended pancreatic edema (Fig. 1F).

The rapid kinetics of neutrophil infiltration suggested the involvement of an innate immune response. To test this, Rag1−/− (lacking B and T cells) and Rag1−/−Clec9agfp/gfp mice were subjected to caerulein-induced acute pancreatitis. Notably, the absence of DNGR-1 resulted in enhanced neutrophil infiltration (Fig. 1G and fig. S2C) and increased circulating lipase concentrations (Fig. 1H) in B and T cell–deficient mice. Thus, after tissue damage, DNGR-1 expressed on cDC1s regulates the recruitment of neutrophils without the involvement of B and T cells.

A reduction of neutrophil-mediated immunopathology is associated with disease tolerance upon infection, which limits the impact of damage-generating infectious challenges on host fitness without affecting pathogen burden (14, 15). To test whether DNGR-1 affects disease tolerance, we used systemic Candida albicans infection, which generates extensive renal tissue necrosis (16). DNGR-1–deficient mice showed increased morbidity and mortality upon systemic candidiasis (Fig. 2, A and B), despite having a similar fungal burden (Fig. 2C). Extended pathology in the absence of DNGR-1 correlated with increased neutrophil infiltration in the kidney (Fig. 2, D and E). Neutrophil numbers in BM or blood of WT and Clec9agfp/gfp mice were similar (fig. S3). DNGR-1 blockade in infected mice phenocopied increased neutrophilia (Fig. 2F), which was prevented in BATF3-deficient mice (Fig. 2G), indicating that cDC1s mediate the effect. Of note, Rag1−/−Clec9agfp/gfp mice also showed increased renal neutrophil numbers (Fig. 2H) and reduced survival after infection (Fig. 2I). Monocyte recruitment into Candida-infected kidneys was not significantly increased in any of the DNGR-1–deficient conditions (right panel of Fig. 2D and fig. S4). Thus, DNGR-1 dampens the recruitment of neutrophils to damaged tissues in both sterile and infectious settings in a B and T cell–independent manner.

Fig. 2 DNGR-1 controls neutrophil recruitment and pathology associated with systemic candidiasis.

C. albicans was intravenously injected into WT and Clec9agfp/gfp mice. (A and B) Weight loss (A) and survival rate (B) were recorded. The dotted line in (A) shows the no-weight-loss reference. (C to E) After 6 days postinfection, renal fungal burden was detected (C), renal infiltrated neutrophils (left) and monocytes (right) were quantified by flow cytometry (D), and H&E staining was performed (E) (arrows indicate neutrophil accumulation). d.l., detection limit; CFU, colony forming unit. (F and G) Anti–DNGR-1 or isotype control antibodies were intraperitoneally injected in WT (F) or WT and Batf3−/− (G) mice on day −1 and daily after infection; infiltrating renal neutrophils were analyzed by flow cytometry. (H and I) Rag1−/− and Rag1−/−Clec9agfp/gfp mice were infected as indicated. Renal neutrophils (day 6 postinfection) were quantified by flow cytometry (H), and survival rate was determined (I). (J) 4′,6-diamidino-2-phenylindole (DAPI) and TUNEL staining in renal sections (left) and the percentage of TUNEL-positive cells (right). (K and L) Serum creatinine concentrations (K) and KIM-1 relative expression in total kidney (L) in WT and Clec9agfp/gfp Candida-infected mice. β-actin expression was used for normalization. (M and N) Number of renal neutrophils (M) and KIM-1 expression in total kidney (N) at the indicated times postinfection. In (A), (C), (D), (F), (G), (H), and (M), data are means ± SEM of a representative experiment (N ≥ 2 independent experiments), including at least five mice per condition. (B) and (I) show representative experiments (N ≥ 2 independent experiments) with n ≥ 9 mice per genotype. In (C), (D), (F), (G), (H), (K), and (M), each dot represents a single mouse. (E) and (J) show representative images of n ≥ 5 kidneys per condition. In (K), (L), and (N), data are means ± SEM of ≥2 pooled experiments (n ≥ 5 mice per experimental condition in each independent experiment). Significance was assessed by two-way analysis of variance (ANOVA) with Bonferroni post hoc test (A), log-rank (B) and (I), or unpaired Student’s t test between genotypes (C), (D), (H), (J), (K), (L), (M), and (N) or treatments (F) and (G); *P < 0.05, **P < 0.01, and ***P < 0.001.

Neutrophil-mediated renal immunopathology causes acute kidney failure and mortality during systemic candidiasis (17, 18). Consistently, C. albicans–infected Clec9agfp/gfp mice showed exacerbated kidney damage, with increased terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL)–positive cells (Fig. 2J), increased concentrations of serum creatinine (Fig. 2K), and enhanced expression of kidney injury molecule-1 (KIM-1) (Fig. 2L). Kidney neutrophilia was increased in Clec9agfp/gfp mice three days after infection (Fig. 2M), along with enhanced KIM-1 expression (Fig. 2N). Thus, exacerbated renal damage caused by neutrophils could underlie increased pathology in Candida-infected Clec9agfp/gfp mice.

We tested whether DNGR-1–regulated neutrophilia drives tissue damage in sterile pancreatitis (Fig. 3A). Partial depletion of neutrophils with an antibody against Ly6G (anti-Ly6G, or 1A8 antibody) (fig. S5) reverted the enhanced edematous lesions found in isotype-treated Clec9agfp/gfp mice upon caerulein treatment (Fig. 3B). Assessing the impact of neutrophils in C. albicans infection is more complex, because the depletion of neutrophils is lethal (19). To circumvent this, we first used fungizone to eliminate the fungus starting at day 3 postinfection (Fig. 3C and fig. S6A), after initial tissue damage by the infection (Fig. 2N). Removal of C. albicans did not affect the exacerbated neutrophil infiltration (Fig. 3D) or renal damage (Fig. 3E) observed in Clec9agfp/gfp mice. Thus, after the initial damage, the presence of fungus was not essential for the DNGR-1–dependent effect. Neutrophil depletion with 1A8 antibody in the presence of fungizone (Fig. 3, C and F, and fig. S6B) prevented the enhanced renal damage found in isotype-treated Clec9agfp/gfp mice (Fig. 3, G and H), even though fungal burden was equivalent between genotypes (Fig. 3I). Thus, neutrophil influx is the cellular mechanism driving the pathology in Candida-infected Clec9agfp/gfp mice.

Fig. 3 DNGR-1 restrains tissue damage by dampening neutrophil-mediated immunopathology.

(A) Representative scheme of pancreatitis induction. 1A8 neutrophil-depleting antibody or isotype (iso) control were administered intraperitoneally, as indicated. (B) H&E staining in pancreatic sections (left) and percentages of edematous area (right). Representative images of n ≥ 6 pancreata per experimental condition. (C) Representative scheme of C. albicans infection; fungizone or PBS together or alone with 1A8 antibody or isotype control were intraperitoneally administered as indicated. (D to I) C. albicans was intravenously injected in WT and Clec9agfp/gfp mice. After 6 days postinfection, renal-infiltrating neutrophils were quantified by flow cytometry (D) and (F), and KIM-1 expression in total kidney was measured (E) and (G). (H and I) Serum creatinine concentrations (H) and renal fungal burden (I). In (B) and (D) to (I), data are means ± SEM of a representative experiment (N ≥ 2 independent experiments), including at least five mice per condition. Each dot represents a single mouse. Significance was assessed by unpaired Student’s t test between genotypes; *P < 0.05, **P < 0.01, and ***P < 0.001.

To decipher the mechanisms underlying the regulatory role of DNGR-1 on cDC1s in neutrophil infiltration, we used F-actin–myosin II complexes as DNGR-1 ligand (DNGR-1L) to robustly trigger the receptor (20). Plated DNGR-1L triggered signaling through the DNGR-1–SYK axis in B3Z-NFAT reporter cells (7) in a dose- (fig. S7A) and DNGR-1–dependent manner (fig. S7B). Then, we used a cDC1 cell line (MutuDC) (21) that expresses DNGR-1 as well as Dectin-1 (fig. S8A), a CLR critically involved in C. albicans recognition (22). Stimulation of MutuDCs with the Dectin-1 agonists whole β-glucan particles (WGP) or heat-killed C. albicans (23) induced the expression of proinflammatory factors such as Tnf, Cxcl2, and Egr2. This was reduced by concomitant exposure to DNGR-1L (Fig. 4A and fig. S8, B and C). Consistently, DNGR-1 triggering attenuated phospholipase Cγ2 (PLCγ2) phosphorylation and IκB degradation in response to WGP (Fig. 4B). DNGR-1 triggering had no impact on the response to toll-like receptor 9 (TLR9) ligand CpG (Fig. 4C and fig. S8, D and E), indicating specificity in the pathways modulated. Using a blocking antibody (fig. S8F), we confirmed that the effect elicited by DNGR-1L was DNGR-1 dependent (fig. S8G).

Fig. 4 DNGR-1 activates SHP-1 and controls neutrophil infiltration by dampening MIP-2 expression in cDC1s.

(A to E) MutuDCs untreated or exposed to DNGR-1 ligand (DNGR-1L), were stimulated with WGP or CpG where indicated. Tnf, Cxcl2, and Egr2 expression was measured by quantitative polymerase chain reaction after 4 hours of stimulation (A); fold induction versus nonstimulated cells is shown. Immunoblot analysis was performed with the indicated antibodies (B). P-PLCγ2, phosphorylated PLCγ2. In (C), Tnf and Cxcl2 expression were measured as in (A). Egr2 expression was not induced in response to CpG. MutuDCs were exposed to DNGR-1L and analyzed by immunoblot (D). P-SHP-1, phosphorylated SHP-1. In (E), Tnf, Cxcl2, and Egr2 expression was measured in cells either preincubated with the SHP-inhibitor NSC (+) or not (−) and stimulated as in (A). (F to J) Mice were intravenously infected with C. albicans. Renal infiltrating neutrophils were quantified after 6 days in CD11cΔSHP-1 and WT littermates (F) and pepducin or control peptide-treated WT and Clec9agfp/gfp mice (G). In (H), relative Cxcl2 expression by immune cells in the kidney was measured 60 hours postinfection. In (I) and (J), renal infiltrating neutrophils were quantified at day 6 post infection. In (I), lethally irradiated B6/SJL CD45.1 recipient mice were reconstituted with a mixture of 50% BATF3-deficient BM cells (CD45.2, producing MIP-2 but lacking cDC1s) and 50% MIP-2–deficient BM cells (CD45.2, Cxcl2−/− mice). Thus, Batf3−/−:Cxcl2−/− chimeric mice are defective for MIP-2 production only in cDC1s compared with control BM chimeras (Batf3−/−:WT). Anti–DNGR-1 or isotype control antibodies were intraperitoneally injected on day −1 and daily after infection. In (J), neutrophil numbers are shown for the following mixed BM chimeric mice, generated as in (I): (i) Batf3−/−:WT control chimeras; (ii) Batf3−/−:Clec9agfp/gfp, which generate cDC1s lacking DNGR-1; (iii) Batf3−/−:Cxcl2−/−, which produce MIP-2–deficient cDC1s; and (iv) Batf3−/−:Clec9agfp/gfp Cxcl2−/−, which generate cDC1s lacking both DNGR-1 and MIP-2. For (I) and (J), all cDC1s in the kidney were of donor origin (fig. S12, A and C), and the number of reconstituted cDC1s were equal in the different chimeric mice (fig. S12, B and D). In (A), (C), (E), and (H), data are means + SEM of pooled experiments, including N ≥ 3 individual cultures or ≥4 mice per condition (N ≥ 4 independent experiments). For (B) and (D), N ≥ 2 representative immunoblots. In (F), (G), (I), and (J), data are means ± SEM of two pooled experiments. Each dot represents a single mouse. Significance was assessed by paired Student’s t test between DNGR-1L–treated or untreated (A), (C), and (E) or between genotypes (H) or unpaired between genotypes (F), (G), (I), and (J); *P < 0.05, **P < 0.01, and ***P < 0.001.

Regulatory phosphatases can couple to some immunoreceptor tyrosine-based activation motif (ITAM)–containing receptors (24, 25). As DNGR-1 bears a hemi-ITAM (hemITAM) motif (7), we tested phosphatase activation upon DNGR-1L sensing. DNGR-1L induced SHP-1 phosphorylation (Fig. 4D) without affecting other CLR-related regulatory mechanisms (26, 27) (fig. S8H). Treatment with the SHP inhibitor NSC-87877 (NSC) abolished the regulatory effect of DNGR-1L on responses elicited by WGP (Fig. 4E). Moreover, mice with SHP-1 depletion in the CD11c+ compartment (CD11cΔSHP-1) (28), including cDC1s, phenocopied the exacerbated neutrophil infiltration observed in DNGR-1–deficient mice (Fig. 4F and fig. S9). These observations are consistent with an involvement of SHP-1 in the molecular mechanism that adjusts inflammatory responses in cDC1s after DNGR-1 engagement.

MIP-2 (encoded by Cxcl2) is a CXCR2 ligand fundamental for neutrophil mobilization from the BM (29) and local recruitment to C. albicans–infected tissues (30). We hypothesized that the MIP-2–CXCR2 axis could be mediating the boosted neutrophilia in the absence of DNGR-1. Administration of pepducin, a peptide that inhibits CXCR2 signaling (13), reverted the enhanced renal neutrophil recruitment observed in Clec9agfp/gfp upon Candida infection (Fig. 4G). To dissect the contribution of DNGR-1 to the MIP-2–mediated process in vivo, we infected mice with C. albicans. After 60 hours, we measured Cxcl2 expression in the renal immune infiltrate (fig. S10). Of all 10 immune populations tested, only neutrophil frequencies were increased in Clec9agfp/gfp mice (fig. S11). Cxcl2 was expressed by neutrophils, macrophages, cDC2s, and cDC1s, but expression was enhanced only in cDC1s in Clec9agfp/gfp mice (Fig. 4H). This suggests that DNGR-1 limits Cxcl2 expression in cDC1s during C. albicans infection.

To investigate the relevance of this increased MIP-2 production by cDC1s on neutrophil recruitment under DNGR-1–deficient conditions, we generated mixed BM chimeric mice with specific MIP-2 deficiency in cDC1s (Batf3−/−:Cxcl2−/−) (see methods and fig. S12). After infection with C. albicans, DNGR-1 blockade generated an exacerbated renal neutrophil recruitment in Batf3−/−:WT control chimeras (Fig. 4I). This boosted neutrophilia was lost in Batf3−/−:Cxcl2−/− chimeras (Fig. 4I), thus relying on MIP-2 produced by cDC1s.

Furthermore, we crossed Clec9agfp/gfp and Cxcl2−/− mice to further generate chimeric mice with cDC1s lacking both DNGR-1 and MIP-2 (Batf3−/−:Clec9agfp/gfp Cxcl2−/−). Upon systemic candidiasis, Batf3−/−:Clec9agfp/gfp chimeras showed an exacerbated neutrophil infiltration into the kidney compared with Batf3−/−:WT control chimeras (Fig. 4J). Notably, this boosted neutrophilia was lost in Batf3−/−:Clec9agfp/gfp Cxcl2−/− mice (Fig. 4J). Thus, in the absence of DNGR-1, MIP-2 produced by BATF3-dependent cDC1s is a key mediator for the enhanced neutrophil recruitment.

Infiltration of immune cells within injured tissues must balance pathogen control with increased damage caused by the inflammatory response. In particular, early infiltration by neutrophils to damaged tissues must be carefully regulated, because these cells can cause further tissue destruction (13, 17, 18). Disease tolerance to infections comprises mechanisms involved in the control of tissue damage. This concept of “tissue damage control” is not restricted to infections and can also be applied to the regulation of damage from sterile inflammation (14, 31). Notably, mediators involved in tissue damage control under both sterile and infectious conditions can be shared (31). Our data suggest that DNGR-1 acts as a shared checkpoint for sterile and infectious tissue damage control. Detection of tissue damage by cDC1s through DNGR-1 would act as a checkpoint for neutrophil infiltration and further immunopathology. Deficient sensing of tissue damage in the absence of DNGR-1 leads to higher production of MIP-2 by cDC1s. This increased MIP-2 production can ignite neutrophil infiltration that drives immunopathology within the damaged organ (fig. S13). Thus, DNGR-1 acts as a necrosis-sensing receptor that, depending on the inflammatory context, may promote a regulatory tissue damage–control mechanism by cDC1s or may contribute to cross-priming during adaptive immunity–related responses (79). This capacity to develop two different host protective functions and the regulation and implications of this dual role remain to be investigated.

Supplementary Materials

www.sciencemag.org/content/362/6412/351/suppl/DC1

Materials and Methods

Figs. S1 to S13

References

References and Notes

Acknowledgments: We are grateful to C. Reis e Sousa and C. A. Lowell for sharing essential reagents. We thank C. Reis e Sousa and members of the D.S. laboratory for discussions and critical reading of the manuscript. We appreciate R. Mota for his advice on the pancreatitis model. We thank the staff at the CNIC facilities for technical support. Funding: C.d.F. is supported by the AECC Foundation as the recipient of an “Ayuda Fundación Científica AECC a personal investigador en cancer.” P.S.-L. is funded by grant BES-2015-072699 from the Spanish Ministerio de Ciencia, Innovación y Universidades (MCIU). M.E. is the recipient of a CNIC International Ph.D. Programme fellowship “la Caixa”-Severo Ochoa OSLC-CNIC-2013-04. S.K.W. is supported by a European Molecular Biology Organization (EMBO) Long-Term Fellowship (grant ALTF 438-2016) and a CNIC-International Postdoctoral Programme Fellowship (grant 17230-2016). Work in the D.S. laboratory is funded by the CNIC and grant SAF2016-79040-R from MCIU, Agencia Estatal de Investigación, and Fondos Europeos de Desarrollo Regional (FEDER); B2017/BMD-3733 Immunothercan-CM from Comunidad de Madrid; RD16/0015/0018-REEM from FIS-Instituto de Salud Carlos III, MICINN, and FEDER; Acteria Foundation; Constantes y Vitales prize (Atresmedia); La Marató de TV3 Foundation (201723); the European Commission (635122-PROCROP H2020); and the European Research Council (ERC-2016-Consolidator Grant 725091). The CNIC is supported by the MCIU and the Pro-CNIC Foundation and is a Severo Ochoa Center of Excellence (SEV-2015-0505). Author contributions: Conceptualization: C.d.F., P.S.-L., and D.S.; methodology, investigation, analysis, and validation: C.d.F., P.S.-L., M.E., S.K.W., S.M.-C., N.B.-M., O.S., M.G., and F.M.-M.; resources: O.S., E.C., and A.P.; writing of original draft: C.d.F., P.S.-L., and D.S.; editing of draft: all authors; supervision, project administration, and funding acquisition: D.S. Competing interests: The authors declare no competing interests. Data and materials availability: All data needed to understand and assess the conclusions of this research are available in the main text and supplementary materials. Batf3−/− mice were obtained from K. M. Murphy under a material transfer agreement with Washington University and the Howard Hugues Medical Institute. The MuTu1940 DC cell line was obtained from H. Acha-Orbea under a material transfer agreement with the University of Laussanne.
View Abstract

Navigate This Article