Antigen Transmission by Replicating Antigen-Bearing Dendritic Cells Jun Diao, Erin Winter, Wenhao Chen, Feng Xu and Mark S. Cattral This information is current as of October 7, 2021. J Immunol 2007; 179:2713-2721; ; doi: 10.4049/jimmunol.179.5.2713 http://www.jimmunol.org/content/179/5/2713 Downloaded from Supplementary http://www.jimmunol.org/content/suppl/2008/03/12/179.5.2713.DC1 Material References This article cites 43 articles, 24 of which you can access for free at: http://www.jimmunol.org/content/179/5/2713.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on October 7, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Antigen Transmission by Replicating Antigen-Bearing Dendritic Cells1

Jun Diao,*† Erin Winter,*† Wenhao Chen,* Feng Xu,* and Mark S. Cattral2*†

During steady-state conditions, conventional spleen dendritic cells (DC) turn over every 2–3 days. Recent evidence indicates that in situ proliferation of DC arising from immediate conventional DC precursors is an important contributor to their homeostasis. In this study, we report that replication-competent conventional DC precursors and DC can internalize and transfer model particulate and soluble Ags directly to their DC progeny during cell division. Real-time confocal microscopy and flow cytometry indicated that Ag transmission to progeny was symmetrical, and suggested that other mechanisms of inter-DC Ag transfer were not involved. Soluble protein Ags inherited by DC progeny were presented effectively to Ag-specific T lymphocytes. Furthermore, we show that the number of DC, and the proportion that are actively proliferating, expands several-fold during an immune response against a viral infection. Our results point to an unanticipated mechanism in which DC are continuously replaced from Ag-bearing replication-competent precursor cells that pass Ag molecules onto their progeny through successive cell divisions. Our Downloaded from findings help explain how Ag may persist in a population of DC despite the brief lifespan of individual mature DC. The Journal of Immunology, 2007, 179: 2713–2721.

endritic cells (DC)3 are specialized APCs that have a key incomplete. In this study, we investigate the hypothesis that DC role in initiating and regulating adaptive immune re- replication provides a novel mechanism for Ag transmission to D sponses and maintaining self-tolerance (1). The esti- their DC progeny, and in this way, sustains a population of DC http://www.jimmunol.org/ mated lifespan of conventional DC in spleen during steady-state capable of presenting Ag to circulating lymphocytes. Our findings conditions is Ͻ2–3 days (2–4). The traditional view of the DC reveal that replication-competent Ag-bearing cDCp and DC pass lifecycle holds that circulating precursors migrate to lymphoid and Ag molecules onto their progeny directly through successive cell parenchymal tissues and differentiate into nonreplicating immature divisions. Furthermore, we show that this mechanism of Ag trans- DC (1, 5, 6). Recent evidence indicates, however, that local pro- mission is operational during an inflammatory response induced by liferation of spleen DC contributes substantially to their homeosta- adenoviral infection, contributing to increased numbers of DC. sis (7). We and others have shown that the majority of these di- viding spleen DC arise from a distinct population of immediate Materials and Methods conventional DC precursors (cDCp) (8, 9). cDCp are CD11cϩ, Mice by guest on October 7, 2021 MHC class IIϪ, lineage (CD4, CD8, B220, DX5, F4/80, CD19, Ϫ ϩ ϩ Male C57BL/6 mice were purchased from Charles River Laboratories. CD3) , and generate exclusively CD11c MHC class II DC that C57BL/6.SJL congenic, C57BL/6-transgenic (ACTbEGFP)OPsb/J, OT-II can continue to replicate for several generations. and OT-1 trangenic mice were purchased originally from Taconic Farms or Classic models of DC biology are based on the premise that DC The Jackson Laboratory, and bred in our animal facility. Mice were main- tained in pathogen-free conditions in accordance with institutional guide- have little or no capacity for cell division (1, 10). Accordingly, the lines, and used at 6–8 wk of age. entire process of Ag internalization, processing, and presentation is considered a function of single DC. The potency of DC in stimu- Antibodies lating immune responses has been attributed, at least partly, to Anti-CD11c (clone HL3), I-Ab (KH74, 25-9-17), CD3 (17A2), CD4 (CT- their inability to degrade internalized Ags rapidly, which increases CD4), CD8␣ (53-6.7), CD19 (1D3), pan-NK (DX5), GR-1 (RB6-8C5), the duration of Ag presentation (11, 12). In light of their apparent CD11b (M1/70), B220 (RA3 6B2), CD45 (30-F11), CD45.2 (104), CD45.1 (A20), CD40 (3/23), CD80 (16-10A1), CD86 (GL1), CD44 (IM7), and brief lifespan, however, this explanation for Ag persistence seems CD16/32 (2.4G2) were purchased from BD Pharmingen. Anti-F4/80 (A3-1) was purchased from Serotec. These Abs were either unlabeled or conjugated to FITC, PE, allophycocyanin, or biotin, as indicated. Unla- *Toronto General Hospital Research Institute, University Health Network, and †De- beled Abs were revealed with PE-conjugated goat anti-mouse Ig and bio- partment of Surgery, University of Toronto, Ontario, Canada tinylated Abs with allophycocyanin, PC5, or PC7. Received for publication November 30, 2006. Accepted for publication June 12, 2007. Cell isolation The costs of publication of this article were defrayed in part by the payment of page DC and cDCp were isolated from bone marrow and spleen, as described charges. This article must therefore be hereby marked advertisement in accordance (8). Briefly, spleens of 10–15 mice were minced, digested with collagenase with 18 U.S.C. Section 1734 solely to indicate this fact. and DNase for 0.5 h at 37°C, and incubated with EDTA. Mononuclear cells 1 This work was supported by the Canadian Institutes for Health Research and As- were isolated from spleen and bone marrow by Lympholyte-M (Cedarlane tellas (to M.S.C.). Presented in part at Immunology 2007, 94th Annual Meeting of the Laboratories) density-gradient centrifugation (13), and enriched for American Association of Immunologists, Miami Beach, FL, May 18, 2007. CD11cϩ cells by positive selection using MACS (Miltenyi Biotec) and ϩ 2 Address correspondence and reprint requests to Dr. Mark S. Cattral, Toronto Gen- CD11c -immunomagnetic beads. Cells retained in the column were eluted eral Hospital, Robert McEwen Building, 11c-1247, 585 University Avenue, Toronto, and labeled with anti-I-Ab FITC, anti-CD11c PE, and anti-lineage markers Ontario, Canada M5G 2N2. E-mail address: [email protected] (anti-CD3, anti-CD4, anti-CD8␣, anti-CD19, anti-B220, anti-F4/80 and anti- ϩ 3 ␤ pan NK/biotin-allophycocyanin or PC7) mAbs. Lineage-negative CD11c Abbreviations used in this paper: DC, dendritic cell; Adv- gal, adenovirus encoding Ϫ ϩ ϩ bacterial ␤-galactosidase; cDCp, conventional DC precursors. MHC II cells (i.e., cDCp) and lineage-negative CD11c MHC II DC fractions were sorted on a MoFlo High-Speed Cell Sorter using Summit Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 acquisition and analysis software (DakoCytomation). The purity of the cell www.jimmunol.org 2714 Ag TRANSMISSION BY REPLICATING DC populations used was routinely Ն99% based on reanalyzed samples. OT-1 and OT-II T cells were isolated from spleen and lymph nodes by negative selection (Miltenyi Biotec). BrdU labeling of phagocytic DC CD45.2 mice were injected with 50 ␮l of 2.7% YG-latex beads in 0.3 ml of saline (0.5 ␮m; Polysciences). Eight hours later, BrdU (1 mg; Sigma- Aldrich) was administered i.p. every 6 h until cell recovery to ensure that it was continuously available to dividing cells (14). Purified CD45.1 DC (1 ϫ 106 per mouse) were injected i.v. into these mice at the start of BrdU treatment. Bone marrow and spleen were recovered for analysis of bead uptake and BrdU incorporation by flow cytometry, as described (8). The proportion of beads captured during the 8 h before BrdU injection was calculated by subtracting the percentage of the adoptively transferred CD45.1 DC that contained beads from the total percentage of spleen DC that contained beads at 32 h. In some experiments, beadϩ DC and beadϪ DC were sorted for analysis of BrdU by immunohistochemistry (BD Bio- sciences) and confocal microscopy (Olympus FluoView 1000 LSCM). Stromal coculture Sorted cDCp (5 ϫ 104) were pulsed with OVA (200–500 ␮g/ml; Sigma- Aldrich) or dextran conjugated with Alexa 647 or Texas Red (1 mg/ml; m.w., 10,000; Molecular Probes) for 1–2 h, labeled with CFSE (Molecular Downloaded from Probes), and cultured on a confluent monolayer of irradiated (25 Gy) stro- mal cells derived from S17 cells (a gift from K. Dorshkind, University of California, Los Angeles, CA (15)). The cells were cultured in RPMI 1640 supplemented with 10% FBS, 50 ␮M 2-ME, 1 mM sodium pyruvate, 10 mM nonessential amino acids, 50 U/ml penicillin, and 50 ␮g/ml strepto- mycin (complete medium) in the presence of GM-CSF (1000 U/ml; BD

Pharmingen), as described (8). At 3 days, the monolayer was disrupted http://www.jimmunol.org/ with 0.25% trypsin/1 mM EDTA and repeated pipetting. The recovered cells were washed and stained with mAbs for flow cytometry or cell sorting. Real-time confocal microscopy cDCp from GFP-transgenic mice were pulsed with dextran conjugated with Texas Red and cocultured on S17 with complete medium in LabTek FIGURE 1. Ag-bearing DC replicate in vivo. a, Mice were immunized chamber slides. Live cell microscopy was performed using a multichannel with OVA or PBS (control) and treated with BrdU for 48 h before cell confocal laser-scanning microscope (Olympus FluoView 1000 LSCM; ϩ Ϫ ϫ recovery. CD11c lineage spleen DC were isolated by FACS and ana- PlanApo 60 oil-immersion objective; NA 1.4) and accompanying imag- by guest on October 7, 2021 ing analysis software (FV10-ASW 1.5; Olympus). The microscope is lyzed by flow cytometry. Dot plots indicate percentage of cells expressing BrdU and MHC class II. b, Mice were injected with OVA conjugated with housed in chamber maintained at 37°C in an atmosphere of 5% CO2,al- ϩ Ϫ lowing cells to remain viable for several days. Images were acquired every FITC or PBS (control). Two hours later, OVA-FITC and OVA-FITC ϩ Ϫ 10 s over several hours. The peak excitation/emission wavelengths for GFP spleen CD11c lineage cells were sorted using the indicated gates and and Texas Red were 488/526 and 516/615 nm, respectively. stained with propidium iodide; the percentage of cells in the S/G2/M phases of the cell cycle was determined by flow cytometry and is indicated in the Cell cycle analysis histograms. c, Sorted OVA-FITCϩ spleen CD11cϩ lineageϪ cells from FACS-purified OVAϪ DC and OVAϩ DC pooled from three mice were CD45.1 mice were transferred into CD45.2 mice that had received either fixed in 75% ethanol at 4°C for 16 h, and stained with propidium iodide (50 CFSE-labeled OT-I or OT-II T cells 24 h earlier. Recipient mice were ␮g/ml) in PBS containing 0.1% Triton X-100 and 0.2 mg/ml RNase for 30 treated with BrdU. Dot plots indicate percentage of CD45.1 DC that in- min at room temperature. DNA content was determined by flow cytometry corporated BrdU (left) at 24 h, and CFSE dilution of OT-1 and OT-II at 3 using the doublet discrimination unit and analyzed by ModFit LT software days. Results are representative of at least three independent experiments. (Verity Software). Ag presentation assays DC arising from OVA-pulsed cDCp were purified by FACS and irradiated. Adenovirus infection Graded numbers of irradiated cells were incubated with CD8ϩ OT-1 T Adenovirus encoding bacterial ␤-galactosidase (Adv-␤gal) was injected cells (purity Ն95%, 5 ϫ 104/well) for 3 days in a humidified atmosphere i.v. into B6 mice (1 ϫ 1010 virus particles per mouse). Preparation, puri- of 5% CO in air at 37°C. Proliferation was assessed by [3H]thymidine 2 fication, and titration of Adv-␤gal have been described previously (17). incorporation, as described (16). Latex beads were injected at the same time or on the first, third, or sixth day Adoptive transfer studies of infection. BrdU labeling was initiated 8 h after bead injection and con- tinued over 12 h. Spleen mononuclear cells were isolated, and the total OVA-pulsed and CFSE-labeled cDCp (0.5–1 ϫ 106) were injected i.v. into number of DC and the proportion that were beadϩ and BrdUϩ were de- mice that had received 1–2 ϫ 106 CFSE-labeled OT-1 CD8ϩ T cells 24 h termined by flow cytometry. earlier. CFSE-labeled syngenic B cells were cotransferred with cDCp in some experiments, which served as a control for nondividing cells. Spleens Statistics and lymph nodes were collected 3 days later, and DC and T cell prolifer- Continuous variables are expressed as mean Ϯ SD and analyzed by two- ation in vivo was assessed by CFSE dilution, as described (8). tail Student’s t test. A p value Ͻ0.05 is considered statistically significant. To assess the stimulatory capacity of the in vivo DC progeny derived from OVA-pulsed cDCp, 2 ϫ 106 cDCp were injected directly into the spleen, and 3 days later, undivided CD45.2 DC, divided CD45.2 DC, and Results endogenous CD45.1 DC were sorted by FACS. The recovered cells were Ag-bearing DC replicate in vivo irradiated, and 2.5 ϫ 103 were incubated with 1 ϫ 105 OT-1 CD8ϩ T cells in triplicate for 3 days. T cell proliferation was evaluated by [3H]thymidine We first injected i.v. the commonly studied model Ag OVA or incorporation. PBS control into B6 mice and treated them with BrdU for 48 h. The Journal of Immunology 2715 Downloaded from http://www.jimmunol.org/ FIGURE 2. Labeling of replicating DC with latex bead in vivo. a, Experimental protocol. CD45.2 congenic mice were injected with YG- latex beads i.v., and 8 h later were treated with BrdU for 24 h. CD45.1 congenic DC were injected i.v. into each recipient at the start of BrdU treatment. b, Detection of YG-latex beads in spleen cDCp and CD11cϩ FIGURE 3. Progeny of Ag-bearing cDCp can stimulate T cell prolifer- MHC class IIϩ DC 2 h after injection of YG-labeled beads. Numbers in ation. OVA-pulsed and CFSE-labeled bone marrow cDCp from CD45.2 histograms indicate percentage of cells. c, Analysis of bead uptake in mice were injected into CD45.1 mice that had received naive CFSE-labeled adoptively transferred CD45.1 DC. d, Analysis of BrdU incorporation OT-1 T cells 24 h earlier. a, Flow cytometric analysis of CFSE staining and ϩ ϩ Ϫ

and MHC class II expression by bead CD11c lineage spleen cells. MHC class II expression in CD45.2 DC at 3 days. b, CFSE dilution and by guest on October 7, 2021 Data from control mice that received unlabeled beads and no BrdU BrdU incorporation of cotransferred OVA-pulsed cDCp (left) and synge- treatment are shown in the left dot plots. e, Proportion of beadϩ and neic B cells (right). c, CFSE dilution of OT-1 T cells at 3 days in mice beadϪ CD11cϩ lineageϪ cells in spleen and that incorporated BrdU. injected with OVA-pulsed (left) or BSA-pulsed cDCp (right) at 3 days. d, Error bars are mean Ϯ SD of four independent experiments with three The undivided and divided DC progeny of OVA-pulsed and CFSE-labeled mice per experiment. f, Immunofluorescent microscopy of FACS-puri- CD45.2 cDCp and CD45.1 endogenous DC were FACS purified using fied spleen DC showing beads inside DC. The nucleus of BrdUϩ DC is the gates indicated 3 days after injection, irradiated, and mixed with stained brown. Original magnification, ϫ600. g, Confocal microscopy naive OT-1 T cells. Proliferation was measured by thymidine uptake. of a beadϩ/BrdUϩ DC. Results are representative of three independent experiments. Error bars are mean Ϯ SD.

The frequency of spleen DC that incorporated BrdU was 46 Ϯ 6% and 31 Ϯ 4% ( p Ͻ 0.05) in OVA-treated and control mice, re- spectively (Fig. 1a). This finding supported the hypothesis that 15% of the transferred CD45.1 DC/cDCp incorporated BrdU at Ag-bearing DC or their immediate precursors could be dividing in 24 h (Fig. 1c). At 3 days, most of the transferred OT-1 and OT-II ϩ vivo. Because BrdU uptake alone cannot establish the identity of T cells had divided, confirming that OVA-FITC DC/cDCp a dividing cell population, we analyzed the cell cycle status of carried Ag. OVA-bearing CD11cϩ lineageϪ spleen cells, which are comprised of DC and cDCp, 2 h after i.v. injection of OVA-conjugated with Labeling of replicating DC with latex beads in vivo FITC or PBS control. OVA-FITC was internalized by 30% of To further confirm that Ag-bearing DC are dividing in vivo, spleen DC, consistent with previous studies (2), and by similar mice were injected i.v. with YG-labeled latex beads, which percentage of MHC class IIϪ cDCp (Fig. 1b). OVA-FITCϩ and served as a model particulate Ag, and treated with BrdU (Fig. OVA-FITCϪ spleen DC and cDCp were sorted by FACS, fixed, 2a). The proportion of spleen DC and cDCp that contained and stained with propidium iodide. We found that 3.7 Ϯ 1.5% (n ϭ beads 2 h after injection was 26–32% (Fig. 2b). cDCp ac- 4) of OVA-FITCϩ DC/cDCp and 4.0 Ϯ 0.7% of OVA-FITCϪ counted for 18% of the total number of bead-positive CD11cϩ Ϫ DC/cDCp were in the S/G2/M phases of the cell cycle (Fig. 1b). lineage cells in spleen. BrdU was initiated 8 h after bead in- These rates were similar to those in control mice injected with jection to allow for clearance of beads from the circulation (18), PBS. To verify that these OVA-FITCϩ cells had the capacity to and continued for 24 h before analysis. We adoptively trans- replicate, we injected sorted OVA-FITCϩ DC and cDCp from ferred DC from congenic CD45.1 mice into each CD45.2 re- CD45.1 mice into CD45.2 congenic mice and treated them with cipient at the start of BrdU treatment, which provided a surro- BrdU. Naive OT-1 T cells or OT-II T cells were injected i.v. into gate measure of bead reuptake during the 24-h BrdU-labeling the recipient mice 24 h before DC/cDCp injection. We found that period. Previous studies established that adoptively transferred 2716 Ag TRANSMISSION BY REPLICATING DC

FIGURE 4. Internalized Ag is transmitted directly to the progeny of replicating DC. a, Experimental proto- col. CD45.2 cDCp were pulsed with dextran conjugated with a fluorophore or OVA, labeled with CFSE, mixed with control CD45.1 cDCp, and cocultured on S17 for 3 days. b, Flow cytometric analysis of dextran-Alexa 647 expression by DC progeny of CD45.2 cDCp (dark line) in each division cycle as defined by CFSE staining in- tensity; the amount of intracellular dextran was quanti- fied by flow cytometry according to mean fluorescence intensity, as indicated in the histograms. The mean flu- orescence intensity of CD45.1 cells (light line) in each Downloaded from histogram was Յ0.8. c, Confocal microscopy of FACS- purified undivided and divided DC from CD45.2 cDCp pulsed with dextran-Texas Red and control CD45.1 DC. Original magnification, ϫ600. d, Divided and undivided progeny of OVA-pulsed CD45.2 cDCp and CD45.1 DC were purified by FACS using the indicated gates, irra- http://www.jimmunol.org/ diated (25 Gy), and mixed with naive OT-1 T cells for 3 days. DC pulsed with BSA served as a negative con- trol. Cell proliferation was measured by thymidine up- take. Results are representative of four independent ex- periments. Error bars are mean Ϯ SD. by guest on October 7, 2021

DC have the same capacity to divide and phagocytose beads and MHC class II, and 50–60% had divided, as determined by CFSE dextran as compared with endogenous DC (data not shown). dilution and BrdU uptake (Fig. 3a). Similar results were obtained Analysis of the adoptively transferred CD45.1 DC revealed that with sorted spleen cDCp. Cotransferred CFSE-labeled B cells a small percentage (3–4%) were beadϩ with most containing a provided a negative control for cell division. As anticipated, single bead; this finding confirmed that the vast majority of beads transferred mature DC were not detectable in spleen beyond were rapidly cleared from circulation (Fig. 2c). The proportion of 48 h (data not shown). OT-1 T cells had completed several endogenous spleen DC that contained beads 32 h after bead injec- rounds of division at 3 days, indicating that the DC arising from tion was 28.1 Ϯ 3.7% (n ϭ 4; Fig. 2d). BrdU was incorporated by cDCp could present OVA (Fig. 3b). Cross-presentation of OVA 15 Ϯ 4% and 21 Ϯ 3% of beadϩ and beadϪ spleen DC, respec- by endogenous DC was not involved because OT-1 T cell pro- tively (Fig. 2e). Fluorescent and confocal microscopy of sorted DC liferation was not stimulated by OVA-pulsed cDCp from Bm1 verified that the beads were inside BrdUϩ DC and not simply mice (data not shown). attached to their cell surface (Fig. 2, f and g). Collectively, these findings indicate that internalized beads captured by replicating DC progeny of Ag-bearing cDCp stimulate T cell proliferation DC can be transmitted to their daughters. Although our findings indicate that Ag-bearing cDCp generate dividing DC, it remained unclear whether their progeny could Replicating Ag-bearing DC arise from cDCp stimulate T cells. We therefore sorted divided and undivided Our finding that cDCp could internalize i.v. injected OVA and CD45.2 DC based on CFSE expression 3 days after adoptive latex beads suggested that they were the source of dividing Ag- transfer. The cells were irradiated, and 2.5 ϫ 103 were cultured bearing DC in vivo. To test this possibility, sorted bone marrow or with 1 ϫ 105 OT-1 CD8ϩ T cells. We found that both popu- spleen cDCp and DC from CD45.2 mice were labeled with CFSE, lations were capable of cross-presenting OVA peptide (Fig. 3c). pulsed with OVA, and injected i.v. into nonirradiated CD45.1 mice As expected, DC recovered from mice that received BSA- that had received CFSE-labeled OT-I T cells 24 h earlier. At 3 pulsed CFSE-labeled CD45.2 cDCp and endogenous CD45.1 days, all of the cells arising from bone marrow cDCp expressed DC did not stimulate OT.1 CD8ϩ T cells. The Journal of Immunology 2717

FIGURE 5. cDCp create interact- ing DC networks. a, Live confocal microscopy of DC division. cDCp from GFP-transgenic mice were pulsed with dextran conjugated with Texas Red and cultured on S17. On the third day of culture, the cells were imaged by real-time confocal micros- copy. Time course of DC division: upper panel, GFP; middle panel, dex- tran-Texas Red; lower panel, differ- ential interference contrast. Origi- nal magnification, ϫ1000. Also see Downloaded from Movie S1.4 b, DC colony that origi- nated from a single dextran-pulsed cDCp. Left upper quandrant, merged; right upper quandrant, GFP; left lower quadrant, dextran-Texas Red; right lower quandrant, differential in- http://www.jimmunol.org/ terference contrast. Original magnifi- cation, ϫ600. Also see Movie S2.4 by guest on October 7, 2021

Replicating DC transmit internalized Ag directly to their contained dextran nor acquired the CFSE stain. These results progeny indicate that dextran internalized by proliferating DC is trans- DC acquire Ags primarily by endocytosis and phagocytosis (19, mitted efficiently and directly to their progeny. 20). Ags can also be transferred between DC through various We next sorted undivided and divided progeny of OVA-pulsed CD45.2 cDCp and control CD45.1cDCp after they were cocul- mechanisms, including uptake of vesicular exosomes and apopto- ϩ tic bodies, exchange of cell membrane, and transfer of organelles tured together for 3 days, and incubated them with naive CD8 and peptides via tunneling nanotubules and gap junctions (21–25). T cells from OT-I transgenic mice. Gates were adjusted on the To further define how daughter DC acquire Ag, we used an in vitro FACS to ensure that the overlap between undivided and divided Ͻ coculture system developed in our laboratory that promotes sur- cells was 2%, and was confirmed by postsort flow cytometric vival and proliferation of DC from cDCp (8). cDCp from CD45.2 analysis. The progeny of CD45.2 cDCp stimulated OT-1 T cell mice were pulsed with dextran conjugated with a fluorophore proliferation, although the magnitude of the response was (Alexa 647 or Texas Red), labeled with CFSE to monitor cell slightly less than with equivalent numbers of undivided CD45.2 division, and cultured on S17 stromal cells in the presence of DC (Fig. 4d). By contrast, control CD45.1 DC were poor stim- GM-CSF for 3 days (Fig. 4a). Dextran is a model polysaccha- ulators of OT-1 T cells. This finding suggests that little or no ride Ag that is endocytosed efficiently by cDCp and DC (26). To free OVA is released into the medium and that acquisition of assess dextran acquisition from the culture medium and from OVA from apoptotic CD45.2 DC or other mechanisms of in- other inter-DC transfer mechanisms, we mixed cDCp from tercellular Ag and peptide transfer are not significant factors CD45.1 mice with the dextran-pulsed CD45.2 cDCp at a 1:1 (22, 23, 25, 27, 28). Thus, dividing DC transmit internalized Ag ratio before placing them on the S17 monolayer. Confocal mi- directly to their progeny, which in turn can present inherited Ag croscopy confirmed that CD45.1 DC were contiguous with to Ag-specific T cells. CFSE-stained CD45.2 DC on the monolayer (data not shown). Flow cytometry and fluorescent microscopy of recovered Real-time confocal microscopy of DC division CD45.2 DC showed that internalized dextran was passed on to To further investigate how internalized Ags are transferred to DC progeny. The amount of intracellular dextran decreased by progeny, we used time-lapse confocal microscopy to visualize di- about half with each division cycle, as determined by mean vision of DC in real time. For these studies, cDCp were isolated fluorescence intensity (Fig. 4, b and c). By contrast, the vast from GFP trangenic mice and pulsed with dextran conjugated with majority of control CD45.1 cDCp and their progeny neither Texas Red. DC division proceeded through several distinct stages 2718 Ag TRANSMISSION BY REPLICATING DC Downloaded from http://www.jimmunol.org/ by guest on October 7, 2021

FIGURE 6. DC replication increases during adenovirus infection. a, Experimental protocol. Mice were infected with 3 ϫ 1010 virus particles of Adv-␤gal. On the day indicated, mice were injected i.v. with YG-latex beads, and 8 h later were treated with BrdU for 12 h before cell recovery. b, Total number of conventional spleen DC on days 0, 1, 3, and 6 of infection. c, Percentage of beadϩ and beadϪ spleen DC that incorporated BrdU over 12 h on days 0, 1, 3, and 6 after infection. d, Percentage of sorted spleen DC in the S/G2/M phase of the cell cycle. e, Frequency of DC that phagocytose beads on days 0, 3, and 6 of infection. f, Flow cytometric analysis of MHC class II and CD86 expression among gated CD11cϩ lineageϪ DC on day 0. g, Fold increase of mean fluorescence intensity (MFI) of MHC class II expression by spleen DC as compared with control. h, Fold increase of MFI of CD86 expression by spleen DC as compared with control. Results are representative of three independent experiments. Error bars are mean Ϯ SD.

(Fig. 5a and Movie S1).4 First, migration on the monolayer symmetrically, producing two daughters of equal size. Endo- stopped. Second, the cells became spherical with loss of dendrites somes containing dextran were also distributed evenly to each and veils; this shape change was the most reliable indicator for daughter, consistent with our flow cytometry data in Fig. 4b. identifying DC that were about to divide. Third, the cells divided Fourth, the daughters remained in close contact for ϳ30–40 min. Finally, the daughters became increasingly motile, separated, 4 The online version of this article contains supplemental material. and acquired characteristic DC morphology and migratory activity. The Journal of Immunology 2719

DC replication resulted in the generation of network of interacting motile DC, each containing a similar amount of Ag (Fig. 5b and Movie S2).4

Spleen DC replication rate increases during adenovirus infection It has been reported that DC turnover in lymphoid tissues in- creases during immune responses to antigenic and inflammatory stimuli (3, 29, 30). To determine whether the mechanism of Ag transfer by replicating DC is operational during an inflamma- tory process, mice were infected with Adv-␤gal, which is highly immunogenic and rapidly induces an adaptive immune response (31). On day 0, 1, 3, or 6, these mice were injected FIGURE 7. Model of Ag transmission by replication-competent cDCp with YG-latex beads and 8 h later were treated with BrdU for and DC. 12 h to identify dividing phagocytic DC (Fig. 6a). The rate of Supplemental Movie Legends bead reuptake during the 12-h BrdU-labeling period was Ͻ2%, Movie S1. DC Replication. Live imaging shows active cell division of as determined by adoptively transferred DC. The number of a DC with transfer of intracellular dextran to progeny. Images were ac- spleen DC decreased by 50% on day 0, and was accompanied quired every 10 s. Left upper quandrant (merged), right upper quandrant by a similar decrease in the proportion that incorporated BrdU (GFP), left lower quandrant (dextran-Texas Red), right lower quandrant ϫ Downloaded from (Fig. 6, b and c). Spleen DC numbers returned to baseline on (DIC). Original magnification, 1200. Movie S2. Replication-competent DC form network of motile Ag-bear- day 1, and subsequently increased 3-fold on days 3 and 6 of ing DC. cDCp from GFP-transgenic mice were pulsed with dextran con- infection. The rate of BrdU incorporation also rebounded, in- jugated with Texas Red and cultured on S17 for 3 days. Live imaging was creasing 2-fold on day 1, 4-fold on day 3, and 5-fold on day 6, performed on 3rd day of culture. Left upper quandrant (merged), right up- and was associated with an increased proportion of spleen DC per quandrant (GFP), left lower quadrant (dextran-Texas Red), right lower in cell cycle (Fig. 6d). Bead phagocytosis and surface expres- quandrant (DIC). Original magnification, ϫ600. sion of MHC class II and CD86 by DC were enhanced on the http://www.jimmunol.org/ first day of infection, and returned to control values by day 6 (Fig. 6, e–h). Remarkably, the proliferation rate of beadϩ DC Ϫ ␤ consistently paralleled that of bead DC. -galactosidase was cDCp was at 5 days, with some DC persisting for 8–10 days. detected in spleen cells at all time points (data not shown), These values are in agreement with those reported by Liu et al. which implied that viral-encoded Ags would be available to (32) in parabionts. Thus, Ags internalized by cDCp have the proliferating DC. We confirmed that an adaptive T cell immune potential to persist via their DC progeny for 10 days, which is response to adenovirus had developed by day 7 (data not significantly longer than the 2- to 3-day lifespan of DC pre- shown). These results highlight that DC proliferation in spleen dicted by previous BrdU incorporation studies. by guest on October 7, 2021 is dynamic, and that phagocytic DC continue to replicate and The relative importance of the DC progeny of replicating transmit internalized beads to progeny during an inflammatory Ag-bearing DC in activating T cells in vivo is unclear. The process. kinetics of T cell proliferation during immune responses would Discussion suggest that they have a minor role in initial T cell activation as compared with mature nondividing DC. Ag transmission by di- Our study documents several previously unknown attributes of viding DC, however, may enable Ag to persist in a population cDCp and DC. First, replication-competent cDCp can capture of DC despite rapid turnover of individual DC, and extend the soluble Ag and particulates; second, cDCp generate proliferat- duration of cognate interactions with T cells. It is interesting to ing DC that transmit internalized Ags to successive generations note that DC expressing low density of MHC-peptide com- of their progeny during cell division; and third, Ag inherited by DC progeny is presented effectively to Ag-specific T cells (Fig. plexes on the surface seem to be important in the generation of 7). Our novel findings have been corroborated by Liu et al. (32), memory T cell responses (34, 35). who reported that MHC-peptide complexes can be detected on We speculate that internalized Ags contained in endosomes the surface of DC that have incorporated BrdU. Furthermore, and lysosomes are distributed randomly to each daughter during we report for the first time that DC replication is augmented cell division, although the study by Liu et al. (32) suggests that significantly during adenovirus infection, suggesting that this MHC/peptide complexes located in the cell membrane might mechanism of Ag transfer may be a natural component of the also be passed on to daughter DC. The inability of DC lyso- response to infection and inflammation. somes to rapidly degrade internalized Ag most likely facilitates cDCp are a distinct population of immediate DC precursors this process (11, 12). In the absence of a continuous supply of that exist in bone marrow and all lymphoid tissues (8). Evi- new Ag, two concurrent processes are envisaged to affect the dence indicates that they are the chief source of conventional amount of Ag in DC progeny: 1) the number of cell divisions DC in spleen, at least during steady-state conditions (8, 9, 33). (dilution); and 2) the rate of Ag degradation, which is influ- It has been suggested that cDCp might arise from a local pro- enced by lysosomal proteolytic activity and the susceptibility of genitor population in spleen, but parabiotic studies by Liu et al. the Ag to proteolysis (12). Previous studies of bone marrow- (32) indicate that they originate in blood, presumably migrating derived DC have shown that immature DC degrade internalized from bone marrow, where they outnumber mature DC by a 3:1 protein Ags slowly (36, 37). Because replicating DC are im- ratio (8). Adoptive transfer studies in our laboratory indicate mature, based on their surface expression levels of MHC class that DC progeny of bone marrow and spleen cDCp are present II and costimulatory molecules and stimulatory capacity of al- for at least 7 days in spleen. Similarly, Naik et al. (9) found that logeneic lymphocytes (16), the number of cell divisions may be the peak production of DC from adoptively transferred spleen the dominant factor that determines their Ag content. 2720 Ag TRANSMISSION BY REPLICATING DC

Our finding that the in vitro and in vivo DC progeny of rep- 3. Kamath, A. T., S. Henri, F. Battye, D. F. Tough, and K. Shortman. 2002. De- licating OVA-pulsed cDCp could stimulate OT-1 T cells indi- velopmental kinetics and lifespan of dendritic cells in mouse lymphoid organs. Blood 100: 1734–1741. cates that this mechanism of Ag transmission is functionally 4. Ruedl, C., P. Koebel, M. Bachmann, M. Hess, and K. Karjalainen. 2000. Ana- relevant. To ensure that there was no overlap between divided tomical origin of dendritic cells determines their life span in peripheral lymph and undivided DC, we waited 3 days until several generations nodes. J. Immunol. 165: 4910–4916. 5. Austyn, J. M., J. W. Kupiec-Weglinski, D. F. Hankins, and P. J. Morris. 1988. of divided DC were detectable, and then discarded the first gen- Migration patterns of dendritic cells in the mouse: homing to T cell-dependent eration of divided DC by FACS. Thus, the divided DC used in areas of spleen, and binding within marginal zone. J. Exp. Med. 167: the T cell stimulation assays are comprised predominately of 646–651. 6. Austyn, J. M. 1998. Dendritic cells. Curr. Opin. Hematol. 5: 3–15. second and third generation progeny. The T cell proliferative 7. Kabashima, K., T. A. Banks, K. M. Ansel, T. T. Lu, C. F. Ware, and J. G. Cyster. response induced by equivalent numbers of divided DC was 2005. Intrinsic lymphotoxin-␤ receptor requirement for homeostasis of lymphoid slightly lower than that by undivided DC. This difference might tissue dendritic cells. Immunity 22: 439–450. 8. Diao, J., E. Winter, C. Cantin, W. Chen, L. Xu, D. Kelvin, J. Phillips, and result from the lower levels of MHC class II expressed by di- M. S. Cattral. 2006. In situ replication of immediate dendritic cell (DC) precur- vided DC as compared with undivided DC (data not shown). sors contributes to conventional DC homeostasis in lymphoid tissue. J. Immunol. Alternatively, it might reflect differences in their Ag content. 176: 7196–7206. 9. Naik, S. H., D. Metcalf, A. van Nieuwenhuijze, I. Wicks, L. Wu, M. O’Keeffe, cDCp cultured on a supportive stromal monolayer create net- and K. Shortman. 2006. Intrasplenic steady-state dendritic cell precursors that are works of Ag-bearing DC with vigorous probing activity that distinct from monocytes. Nat. Immunol. 7: 663–671. resemble those visualized in lymph nodes by two-photon mi- 10. Banchereau, J., F. Briere, C. Caux, J. Davoust, S. Lebecque, Y. J. Liu, B. Pulendran, and K. Palucka. 2000. Immunobiology of dendritic cells. Annu. croscopy (38). It is hypothesized that such networks promote Rev. Immunol. 18: 767–811. inter-DC Ag transfer and increase the available surface area of 11. Trombetta, E. S., M. Ebersold, W. Garrett, M. Pypaert, and I. Mellman. 2003. MHC-peptide complexes needed to stimulate T cells effec- Activation of lysosomal function during dendritic cell maturation. Science 299: Downloaded from 1400–1403. tively. We found that Ag distribution to DC progeny was sym- 12. Delamarre, L., M. Pack, H. Chang, I. Mellman, and E. S. Trombetta. 2005. metrical, suggesting that each daughter of a divided DC should Differential lysosomal proteolysis in antigen-presenting cells determines antigen be equally effective in Ag presentation. Our coculture experi- fate. Science 307: 1630–1634. 13. Vremec, D., and K. Shortman. 1997. Dendritic cell subtypes in mouse lym- ments with dextran- and OVA-pulsed cDCp also indicated that phoid organs: cross-correlation of surface markers, changes with incubation, DC progeny acquired little or no Ag through inter-DC Ag trans- and differences among thymus, spleen, and lymph nodes. J. Immunol. 159: fer even though frequent interactions were clearly evident by 565–573. http://www.jimmunol.org/ 14. Passegue, E., A. J. Wagers, S. Giuriato, W. C. Anderson, and I. L. Weissman. live confocal microscopy (23–25). Whether the low level of DC 2005. Global analysis of proliferation and cell cycle gene expression in the reg- apoptosis (data not shown) or the density of DC on the S17 ulation of hematopoietic stem and progenitor cell fates. J. Exp. Med. 202: monolayer accounts for the apparent absence of inter-DC Ag 1599–1611. 15. Collins, L. S., and K. Dorshkind. 1987. A stromal cell line from myeloid long- transfer is unclear. term bone marrow cultures can support myelopoiesis and B lymphopoiesis. J. Im- Expansion of the number of DC in spleen has been observed munol. 138: 1082–1087. in a variety of microbial infections, and has been attributed 16. Diao, J., E. Winter, W. Chen, C. Cantin, and M. S. Cattral. 2004. Characterization of distinct conventional and plasmacytoid dendritic cell-committed precursors in mostly to influx of new DC and increased DC survival (39–42). murine bone marrow. J. Immunol. 173: 1826–1833. Our findings with adenovirus infection indicate that increased 17. Diao, J., J. A. Smythe, C. Smyth, P. B. Rowe, and I. E. Alexander. 1999. Human by guest on October 7, 2021 DC proliferation is also an important factor. Increased replica- PBMC-derived dendritic cells transduced with an adenovirus vector induce cy- totoxic T-lymphocyte responses against a vector-encoded antigen in vitro. Gene tion of spleen DC could reflect influx of new replication-com- Ther. 6: 845–853. petent cDCp from the circulation, alteration of the cytokine 18. Tacke, F., F. Ginhoux, C. Jakubzick, N. van Rooijen, M. Merad, and milieu in spleen, or both. Whether the increased DC replication G. J. Randolph. 2006. Immature monocytes acquire antigens from other cells in the bone marrow and present them to T cells after maturing in the periphery. rate during infection involves those that have captured patho- J. Exp. Med. 203: 583–597. gen-derived Ags is unclear. In this regard, a previous report 19. Mellman, I., and R. M. Steinman. 2001. Dendritic cells: specialized and regulated indicates that the nature of the phagocytosed cargo can influ- antigen processing machines. Cell 106: 255–258. 20. Trombetta, E. S., and I. Mellman. 2005. Cell biology of antigen processing in ence DC function (43). It will also be important to investigate vitro and in vivo. Annu. Rev. Immunol. 23: 975–1028. how cognate interactions with T cells influence DC prolifera- 21. Morelli, A. E., A. T. Larregina, W. J. Shufesky, M. L. Sullivan, D. B. Stolz, tion. Finally, although we speculate that Ag transmitted through G. D. Papworth, A. F. Zahorchak, A. J. Logar, Z. Wang, S. C. Watkins, et al. 2004. Endocytosis, intracellular sorting, and processing of exosomes by dendritic DC replication could enhance host defenses, it is noteworthy cells. Blood 104: 3257–3266. that this mechanism might also contribute to the replication and 22. Harshyne, L. A., S. C. Watkins, A. Gambotto, and S. M. Barratt-Boyes. 2001. propagation of retroviruses in the immune system. Dendritic cells acquire antigens from live cells for cross-presentation to CTL. J. Immunol. 166: 3717–3723. 23. Rustom, A., R. Saffrich, I. Markovic, P. Walther, and H. H. Gerdes. 2004. Nano- tubular highways for intercellular organelle transport. Science 303: 1007–1010. Acknowledgments 24. Watkins, S. C., and R. D. Salter. 2005. Functional connectivity between immune We thank N. Iscove, M. Julius, R. Gorczynski, D. Grant, A. Kapus, and cells mediated by tunneling nanotubules. Immunity 23: 309–318. members of the Multiorgan Transplantation Program for helpful advice and 25. Neijssen, J., C. Herberts, J. W. Drijfhout, E. Reits, L. Janssen, and J. Neefjes. 2005. Cross-presentation by intercellular peptide transfer through gap junctions. critical reading of the manuscript; C. Cantin for FACS; and M. Shi for Nature 434: 83–88. immunofluorescent microscopy. 26. Lutz, M. B., P. Rovere, M. J. Kleijmeer, M. Rescigno, C. U. Assmann, V. M. Oorschot, H. J. Geuze, J. Trucy, D. Demandolx, J. Davoust, and P. Ricciardi-Castagnoli. 1997. Intracellular routes and selective retention of an- Disclosures tigens in mildly acidic cathepsin D/lysosome-associated membrane protein-1/ MHC class II-positive vesicles in immature dendritic cells. J. Immunol. 159: The authors have no financial conflict of interest. 3707–3716. 27. Wolfers, J., A. Lozier, G. Raposo, A. Regnault, C. Thery, C. Masurier, C. Flament, S. Pouzieux, F. Faure, T. Tursz, et al. 2001. Tumor-derived exosomes References are a source of shared tumor rejection antigens for CTL cross-priming. Nat. Med. 1. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control of 7: 297–303. immunity. Nature 392: 245–252. 28. Albert, M. L., B. Sauter, and N. Bhardwaj. 1998. Dendritic cells acquire antigen 2. Kamath, A. T., J. Pooley, M. A. O’Keeffe, D. Vremec, Y. Zhan, A. M. Lew, A. from apoptotic cells and induce class I-restricted CTLs. Nature 392: 86–89. D’Amico, L. Wu, D. F. Tough, and K. Shortman. 2000. The development, mat- 29. De Smedt, T., B. Pajak, E. Muraille, L. Lespagnard, E. Heinen, P. De Baetselier, uration, and turnover rate of mouse spleen dendritic cell populations. J. Immunol. J. Urbain, O. Leo, and M. Moser. 1996. Regulation of dendritic cell numbers and 165: 6762–6770. maturation by lipopolysaccharide in vivo. J. Exp. Med. 184: 1413–1424. The Journal of Immunology 2721

30. Ingulli, E., A. Mondino, A. Khoruts, and M. K. Jenkins. 1997. In vivo detection mal compartments of dendritic cells is regulated by inflammatory stimuli. J. Exp. of dendritic cell antigen presentation to CD4ϩ T cells. J. Exp. Med. 185: Med. 191: 927–936. 2133–2141. 37. Turley, S. J., K. Inaba, W. S. Garrett, M. Ebersold, J. Unternaehrer, 31. Roberts, D. M., A. Nanda, M. J. Havenga, P. Abbink, D. M. Lynch, B. A. Ewald, R. M. Steinman, and I. Mellman. 2000. Transport of peptide-MHC class II com- J. Liu, A. R. Thorner, P. E. Swanson, D. A. Gorgone, et al. 2006. Hexon-chi- plexes in developing dendritic cells. Science 288: 522–527. maeric adenovirus serotype 5 vectors circumvent pre-existing anti-vector immu- 38. Lindquist, R. L., G. Shakhar, D. Dudziak, H. Wardemann, T. Eisenreich, nity. Nature 441: 239–243. M. L. Dustin, and M. C. Nussenzweig. 2004. Visualizing dendritic cell networks 32. Liu, K., C. Waskow, X. Liu, K. Yao, J. Hoh, and M. Nussenzweig. 2007. Origin in vivo. Nat. Immunol. 5: 1243–1250. of dendritic cells in peripheral lymphoid organs of mice. Nat. Immunol. 8: 39. Sponaas, A. M., E. T. Cadman, C. Voisine, V. Harrison, A. Boonstra, A. O’Garra, 578–583. and J. Langhorne. 2006. Malaria infection changes the ability of splenic dendritic cell populations to stimulate antigen-specific T cells. J. Exp. Med. 203: 33. Shortman, K., and S. H. Naik. 2007. Steady-state and inflammatory dendritic-cell 1427–1433. development. Nat. Rev. Immunol. 7: 19–30. 40. Serbina, N. V., and E. G. Pamer. 2006. Monocyte emigration from bone marrow 34. Catron, D. M., L. K. Rusch, J. Hataye, A. A. Itano, and M. K. Jenkins. 2006. ϩ during bacterial infection requires signals mediated by chemokine receptor CD4 T cells that enter the draining lymph nodes after antigen injection partic- CCR2. Nat. Immunol. 7: 311–317. ipate in the primary response and become central-memory cells. J. Exp. Med. 41. Hou, W. S., and L. Van Parijs. 2004. A Bcl-2-dependent molecular timer regu- 203: 1045–1054. lates the lifespan and immunogenicity of dendritic cells. Nat. Immunol. 5: 35. Bullock, T. N., D. W. Mullins, and V. H. Engelhard. 2003. Antigen density 583–589. presented by dendritic cells in vivo differentially affects the number and avidity 42. Chen, M., Y. H. Wang, Y. Wang, L. Huang, H. Sandoval, Y. J. Liu, and J. Wang. ϩ of primary, memory, and recall CD8 T cells. J. Immunol. 170: 1822–1829. 2006. Dendritic cell apoptosis in the maintenance of immune tolerance. Science 36. Inaba, K., S. Turley, T. Iyoda, F. Yamaide, S. Shimoyama, C. Reis e Sousa, 311: 1160–1164. R. N. Germain, I. Mellman, and R. M. Steinman. 2000. The formation of im- 43. Blander, J. M., and R. Medzhitov. 2006. Toll-dependent selection of microbial munogenic major histocompatibility complex class II-peptide ligands in lysoso- antigens for presentation by dendritic cells. Nature 440: 808–812. Downloaded from http://www.jimmunol.org/ by guest on October 7, 2021