of November 29, 2015.This information is current as

Cell-Based VaccinesMonkeys: Implications for DendriticMonocyte-Derived Dendritic Cells in Maturation and Trafficking of

Albert D. DonnenbergCapuano III, Michael Murphey-Corb, Louis D. Falo, Jr. andHarshyne, E. Michael Meyer, Simon C. Watkins, Saverio Simon M. Barratt-Boyes, Michael I. Zimmer, Larry A.

http://www.jimmunol.org/content/164/5/2487doi: 10.4049/jimmunol.164.5.2487

2000; 164:2487-2495; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/164/5/2487.full#ref-list-1

, 25 of which you can access for free at: cites 36 articlesThis article

Subscriptionshttp://jimmunol.org/subscriptions

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/ji/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/cgi/alerts/etocReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2000 by The American Association of9650 Rockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on Novem

ber 29, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on N

ovember 29, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

Maturation and Trafficking of Monocyte-Derived DendriticCells in Monkeys: Implications for Dendritic Cell-BasedVaccines1

Simon M. Barratt-Boyes,2* Michael I. Zimmer,* Larry A. Harshyne,* E. Michael Meyer, †

Simon C. Watkins,‡ Saverio Capuano III,§ Michael Murphey-Corb, § Louis D. Falo, Jr.,¶ andAlbert D. Donnenberg*†

Human dendritic cells (DC) have polarized responses to chemokines as a function of maturation state, but the effect of maturationon DC trafficking in vivo is not known. We have addressed this question in a highly relevant rhesus macaque model. We dem-onstrate that immature and CD40 ligand-matured monocyte-derived DC have characteristic phenotypic and functional differencesin vitro. In particular, immature DC express CC chemokine receptor 5 (CCR5) and migrate in response to macrophage inflam-matory protein-1a (MIP-1a), whereas mature DC switch expression to CCR7 and respond exclusively to MIP-3b and 6Ckine.Mature DC transduced to express a marker gene localized to lymph nodes after intradermal injection, constituting 1.5% of lymphnode DC. In contrast, cutaneous DC transfected in situ via gene gun were detected in the draining lymph node at a 20-fold lowerfrequency. Unexpectedly, the state of maturation at the time of injection had no influence on the proportion of DC that localizedto draining lymph nodes, as labeled immature and mature DC were detected in equal numbers. Immature DC that trafficked tolymph nodes underwent a significant up-regulation of CD86 expression indicative of spontaneous maturation. Moreover, imma-ture DC exited completely from the dermis within 36 h of injection, whereas mature DC persisted in large numbers associated witha marked inflammatory infiltrate. We conclude that in vitro maturation is not a requirement for effective migration of DC in vivoand suggest that administration of Ag-loaded immature DC that undergo natural maturation following injection may be preferredfor DC-based immunotherapy. The Journal of Immunology,2000, 164: 2487–2495.

M igration is an essential part of dendritic cell (DC)3

function, as DC acquire Ag in the periphery and thenmigrate to lymph nodes, where they present processed

Ag to T cells (1). It is well established that migration of DC in vitrois tightly regulated as a function of maturation (2–4). However, theinfluence of DC maturation on trafficking in vivo remains ill de-fined, particularly as it relates to in vitro-derived DC currently usedin immunotherapy protocols.

Immature DC express CC chemokine receptor 1 (CCR1),CCR2, and CCR5 that bind inflammatory chemokines, which isbelieved to promote the accumulation of Ag-acquiring immatureDC at sites of tissue injury. In contrast, mature DC switch receptorexpression to CCR7 and migrate in response to constitutive che-

mokines that are expressed in lymphatics and lymphoid tissues.This interaction induces migration of Ag-stimulated and maturingDC to lymph nodes (2–7). Based on this pattern of receptor ex-pression it is thought that DC-based vaccination protocols shouldfocus on using ex vivo-matured DC, as without maturation DCmay not traffic to lymph nodes and, hence, may not contribute toinduction of T cell responses. However, in vivo studies by us andothers in mice and chimpanzees suggest that the state of DC mat-uration before administration does not accurately predict the ca-pacity of DC to migrate to lymph nodes (8–12). One explanationfor this discrepancy is that immature DC undergo maturation uponinjection in skin, resulting in responsiveness to constitutive che-mokines that induce homing to lymph nodes.

To test this hypothesis in a preclinical setting, we have studiedthe role of maturation on DC trafficking in the rhesus macaquemodel. We first characterized the phenotype and function of im-mature and CD40L matured monocyte-derived DC and correlatedthis with chemokine receptor expression and chemotactic re-sponses in vitro. We then used rare event analysis by flow cytom-etry to quantify DC trafficking to lymph nodes following intrader-mal (i.d.) injection, using both membrane dye and genetransduction to mark cells. To provide relevance to other methodsof DC-based vaccination, we compared the migration of gene-transduced DC with that of cutaneous DC transfected in situ withgene gun.

Materials and MethodsAnimals

Normal adult rhesus macaques (Macaca mulatta) were used in this study.Animals were housed and all in vivo experiments were performed at the

*Department of Infectious Diseases and Microbiology, Graduate School of PublicHealth, and Departments of†Medicine, ‡Cell Biology and Physiology,§MolecularGenetics and Biochemistry, and¶Dermatology, School of Medicine, University ofPittsburgh, Pittsburgh, PA 15261

Received for publication October 14, 1999. Accepted for publication December21, 1999.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby markedadvertisementin accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by National Institutes of Health Grants RR00119 (toS.M.B.-B.), AI41408 (to A.D.D.), and AI43664 (to S.M.B.-B., S.C.W., M.M.-C.,L.D.F., and A.D.D.) and by the University of Pittsburgh Medical Center CompetitiveMedical Research Fund (to S.M.B.-B.).2 Address correspondence and reprint requests to Dr. Simon M. Barratt-Boyes, De-partment of Infectious Diseases and Microbiology, Graduate School of Public Health,University of Pittsburgh, Pittsburgh, PA 15261. E-mail address: smbb1@pitt.edu3 Abbreviations used in this paper: DC, dendritic cells; CCR, CC chemokine receptor;CD40L, CD40 ligand; MIP, macrophage inflammatory protein; i.d., intradermal; DiD,1,19-dioctadecyl-3,3,39,39-tetramethylindodi-carbocyanine perchlorate; EGFP, en-hanced green fluorescence protein.

Copyright © 2000 by The American Association of Immunologists 0022-1767/00/$02.00

by guest on Novem

ber 29, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

Infectious Disease Primate Research Facility, University of Pittsburgh(Pittsburgh, PA), using protocols approved by the institutional animal careand use committee.

Media and reagents

The medium used to culture DC was RPMI 1640 medium supplementedwith 10% FCS (BioWhittaker, Walkersville, MD),L-glutamine, sodiumpyruvate, nonessential amino acids, penicillin-streptomycin, 10 mMHEPES buffer (Life Technologies, Grand Island, NY), and 2-ME (Sigma,St. Louis, MO). Recombinant human GM-CSF and IL-4 were provided bySchering-Plough (Kenilworth, NJ), and trimeric human CD40L was pro-vided by Immunex (Seattle, WA). Recombinant human MIP-1a, 6Ckine(BioSource, Camarillo, CA), and MIP-3b (R&D Systems, Minneapolis,MN) were used in chemotaxis assays. DQ Green BSA was a gift fromMingjie Zhou (Molecular Probes, Eugene, OR). The lipophilic carbocya-nine dye 1,19-dioctadecyl-3,3,39,39-tetramethylindodi-carbocyanine per-chlorate (DiD; excitation/emission spectra5 644 nm/663 nm) used in la-beling studies was purchased from Molecular Probes.

DC culture

PBMC were isolated from heparinized blood by centrifugation throughsodium diatrizoate and Ficoll (Sigma), and CD141 cells were positivelyselected using Ab-coated microbeads and magnetic separation (MiltenyiBiotec, Auburn, CA). The purity of CD141 cells using this technique wasconsistently.90%. Cells were cultured at a density of 33 105 cells/ml insix-well plates. GM-CSF (1000 U/ml) and IL-4 (1000 U/ml) were added atthe initiation of culture, and media and cytokines were replenished at reg-ular intervals for 7 days, depending on the experiments. CD40L (3mg/ml)was added on day 5 of culture for 24–48 h to mature cells, and cells wereharvested following incubation with 10 mM EDTA.

Flow cytometric analysis

DC were stained for flow cytometric analysis as previously described (13).Cross-reactive mAb specific for human HLA-A, -B, and -C (cloneB9.12.1); CD83 (HB15A; both from Coulter-Immunotech, Miami, FL);HLA-DR (L243), CD14 (mfP9), CD80 (L307.4; all from Becton Dickin-son, San Jose, CA); CD86 (FUN-1; PharMingen, San Diego, CA); CD40(EA-5; Ancell, Bayport, MN); and CCR5 (FAB182; gift from Frank Mor-tari, R&D Systems) were used as conjugates with FITC or PE. In someexperiments biotinylated CD86 was used with streptavidin-Cy5 (Amer-sham, Arlington Heights, IL). The CCR7 mAb was a gift from Lijun Wu(LeukoSite, Cambridge, MA) and was used in association with a goat anti-mouse secondary Ab (BioSource). Analyses related to the characterizationof immature and mature DC were performed on a FACSCalibur flow cy-tometer (Becton Dickinson) using CellQuest software. For analyses of DCand lymph node suspensions in trafficking studies, 1.53 106 events wereacquired in triplicate on a FACS Vantage flow cytometer (Becton Dick-inson), using a second argon laser emitting at 647 nm to excite DiD flu-orescence. Data were acquired and analyzed using LYSIS II and WinListsoftware, respectively. Fluorescence and side scatter measurements for invivo data are shown in four-decade log scales.

Assessment of Ag uptake

For analysis of macropinocytosis by DC we used DQ Green BSA, a self-quenched dye conjugate of BSA. Degradation of protein in this reagentresults in dequenching and elaboration of green fluorescence (14). DC wereincubated with 10mg/ml DQ Green BSA at 4 or 37°C for 30 min with andwithout prior incubation with 10mg/ml cytochalasin D (Sigma) followedby labeling with mAb to CD86. Fluorescence in this assay is indicative ofuptake and proteolytic cleavage of BSA.

Analysis of cytokine production

Supernatants were collected from cultures on days 4 and 7, with and with-out addition of CD40L (3mg/ml) on day 5, and stored at270°C untilassayed. IL-12 (p40 and 70 forms) and TNF-a were detected using ELISAkits specific for rhesus monkey (BioSource) according to the manufactur-er’s instructions.

Chemotaxis assays

Cell migration was measured using a 96-well chemotaxis chamber with a5-mm pore polycarbonate membrane (Neuro Probe, Gaithersburg, MD).Thirty microliters of chemokine in RPMI, 0.5% BSA, and 10 mM HEPESmedium was added to appropriate wells, which were then covered by themembrane. Twenty-five thousand cells at a density of 13 106 cells/ml inthe same medium were placed on the membrane. Plates were incubated for

90 min at 37°C in a humidified incubator, and cells remaining on thetopside of the membrane were then removed by gentle washing with PBSand wiping. Migrated cells that had attached to the underside of the mem-brane were fixed in 2.5% glutaraldehyde and stained with 0.1% toluidineblue (Fisher Scientific, Pittsburgh, PA). To quantify migration, cells in fournonoverlapping areas were counted under3400 magnification andsummed. Results are expressed as the mean6 SEM for duplicate wells.

Transduction of monkey DC using recombinant adenovirusencoding a marker gene

DC were transduced with a replication-defective recombinant adenoviralvector expressing enhanced green fluorescence protein (EGFP) under thecontrol of the CMV immediate-early promoter/enhancer (15), which wasprovided by Andrea Gambotto (University of Pittsburgh). Briefly, viruswas added at a multiplicity of infection of 100 directly to wells containingday 4 cells and incubated for 1 h at room temperature. Medium was thenreplenished, and cells were matured with addition of CD40L for 24 h.

DC trafficking in vivo

Immature day 4 or mature CD40L-treated day 7 DC were labeled with DiDas previously described (12) or were transduced with recombinant adeno-virus and resuspended in 400ml of PBS for injection. Between 2.73 106

and 5.23 106 DC were injected i.d. into anesthetized animals from whichcells were derived in a region lateral to the proximal inguinal lymph nodechain. In some experiments EGFP gene was administered via gene gun toskin before injection of labeled DC. Animals were anesthetized 36 h later,and inguinal and axillary lymph nodes and skin were biopsied. Lymphnode tissues were disrupted using digestion with collagenase D as previ-ously described (16). Unseparated lymph node cell suspensions were la-beled with mAb to CD83 and CD86 before flow cytometric analysis. Skinbiopsies were treated with 4% paraformaldehyde and 30% sucrose infusionbefore freezing in isopentane. Sections were cut and stained with hema-toxylin and eosin or were examined directly for fluorescent cells as pre-viously described (12).

DNA-based immunization

Genetic immunizations were performed by biolistic delivery as previouslydescribed (17) using the pEGFP-C2 plasmid (Clontech, Palo Alto, CA),which contains the EGFP gene under the control of the CMV promoter.Animals received a total of four shots given to two overlapping regions ofshaved skin in the inguinal region. In all cases the gene gun was appliedbefore and at least 0.5 in. away from injection of cells. In experiments inwhich the gene gun was applied, injected DC were labeled with DiD so asto distinguish in situ transfected cells (green) from injected DC (red).

ResultsMonkey DC undergo phenotypic and functional maturation upontreatment with CD40L.

To establish the monkey DC system we performed extensive phe-notypic and functional studies in vitro. DC cultured from CD141

peripheral blood monocytes of rhesus macaques had characteristicdendritic morphology (data not shown), similar to that of humanCD14-derived DC (18) and DC generated by us and others fromPBMC of the chimpanzee and macaque (13, 19). DC cultured for4 days in GM-CSF and IL-4 retained expression of CD14, but werenegative for the mature DC marker CD83 (Fig. 1). The day 4 DCexpressed moderate levels of costimulatory molecules CD80,CD86, and CD40 and high levels of MHC class I and class II (Fig.1). In contrast, monkey DC treated by ligation of CD40, which isknown to induce rapid maturation of DC in the human system (20,21), uniformly expressed CD83 and high levels of costimulatorymolecules CD80, CD86, and CD40 (Fig. 1). Expression of MHCclass I and class II was unaltered, whereas expression of CD14 wasdown-regulated. Cells cultured for 7 days without CD40 ligationhad an intermediate phenotype (Fig. 1).

To examine functional changes associated with CD40L treat-ment we tested the capacity of DC to exhibit macropinocytosis andsecrete proinflammatory cytokines, properties of immature andmature DC, respectively (21, 22). DC cultured for 4 days rapidlyinternalized and cleaved DQ Green BSA when incubated at 37°C,

2488 MATURATION AND TRAFFICKING OF DENDRITIC CELLS IN MONKEYS

by guest on Novem

ber 29, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

but not at 4°C (Fig. 2). This required actin polymerization and,hence, was an active process, as treatment with cytochalasin Dsubstantially inhibited uptake (22). In contrast, DQ Green BSAfluorescence was almost completely absent at 37°C when CD40L-matured day 7 cells were used (Fig. 2). Cells cultured for 7 days inthe absence of CD40L had an intermediate function. The loss ofability to internalize and cleave protein upon maturation was cor-related with increased expression of CD86, as can be seen by thesmall proportion of day 4 DC that did not process DQ Green BSAand expressed high levels of CD86 (Fig. 2). Hence, in this systemCD86 up-regulation is a reliable indicator of functional maturation.In cytokine release assays, day 4 and day 7 DC secreted negligiblelevels of IL-12 and TNF-a, whereas CD40L-treated day 7 DCmarkedly up-regulated secretion of these cytokines (Table I).

Based on these results, in subsequent chemotaxis and traffickingexperiments we classified cells harvested on day 4 of culture asimmature DC and cells harvested on day 7 after CD40L treatmentas mature DC.

Immature and mature monkey DC have distinct patterns ofchemokine receptor expression and chemotactic responses

We next determined the chemokine receptor expression and che-motactic responsiveness of monkey DC at different stages of mat-uration. Chemokine receptor expression of human DC is highlypolarized, with immature DC expressing receptors for inflamma-tory chemokines and switching expression to receptors for consti-tutive chemokines upon maturation (2–5, 23). Consistent with the

FIGURE 1. CD40L induces maturation of monkey monocyte-derivedDC. DC grown for 4 or 7 days in GM-CSF and IL-4, with and withouttreatment with CD40L on day 5, were stained with cross-reactive mAb andanalyzed by flow cytometry. CD40L induces expression of CD83, loss ofCD14, and increased expression of CD80, CD86, and CD40. Expression ofMHC classes I and II remained high and unchanged by CD40L treatment.

FIGURE 2. Day 7 monkey DC treated withCD40L lose the capacity to internalize and processsoluble protein, correlated with an increase in ex-pression of CD86.a, Green fluorescence of DCfollowing incubation with 10mg/ml DQ GreenBSA for 30 min. Fluorescence results from pro-teolytic cleavage of the protein. Incubations wereperformed at 4 or 37°C with and without pretreat-ment with cytochalasin D (CCD).b, Expression ofCD86 vs green fluorescence by DC incubated with10 mg/ml DQ Green BSA for 30 min at 37°C.

Table I. CD40L-treated monkey DC secrete high levels of IL-12 andTNF-a

Cell Type

Cytokine (pg/ml)a

IL-12b TNF-a

Day 4 DC 59.2 ,2.0Day 7 DC 18.6 ,2.0Day 7 DC/CD40L 33,142.7 455.0

a Means of duplicate samples. Standard errors were always,10% of the means.b p70 and free p40 subunits. Results are representative of three experiments.

2489The Journal of Immunology

by guest on Novem

ber 29, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

human DC literature, immature monkey DC expressed low levelsof CCR5, but not CCR7. Conversely, CD40L-matured DC werenegative for expression of CCR5, but expressed CCR7 at highlevels (Fig. 3). When monkey DC were tested in chemotaxis as-

says, day 4 immature DC migrated in response to CCR5 ligandMIP-1a, but not at all to the CCR7 ligands 6Ckine and MIP-3b.Upon maturation, chemotactic responses were completely switched toMIP-3b and 6Ckine (Fig. 3). The weaker response of immature DCto chemokine correlated with the low level expression of CCR5 onthese cells. These data indicate that in vitro chemotactic responses ofmonkey monocyte-derived DC are tightly regulated as a function ofmaturation.

FIGURE 4. Mature (day 6/CD40L) DC expressing EGFP via adenovi-rus transduction localize in lymph nodes following i.d. injection.Left,EGFP expression of DC immediately before injection and EGFP expres-sion in draining lymph node cells 36 h after injection of 5.23 106 DC,showing the region (R1) used to identify injected DC in lymph nodes.Right, CD83 vs CD86 profile of draining lymph node cells and injectedEGFP1 DC that localized to lymph nodes. The area delineated by perpen-dicular lines identifies CD831CD861 lymph node DC. Numbers representthe mean 6 SEM of triplicate samples, describing injectedEGFP1CD831CD861 DC in draining lymph nodes as a percentage of totallymph node DC. The proportion of injected cells recovered from the drain-ing lymph node was 0.096 0.01%. SSC, side scatter.

FIGURE 5. Biolistic delivery of EGFP gene-coated particles leads to acute epidermal inflammation and transfection of cells. Skin was biopsied 36 hafter gene gun application.a, Hematoxylin and eosin staining, showing acute inflammation of the epidermis. Original magnification,3100.b, Fluorescenceand differential interference contrast image of skin. An EGFP-transfected cell with intracellular bead (arrow) is present in the superficial dermis. Originalmagnification,3200.

FIGURE 3. Monkey DC undergo a switch in chemokine receptor ex-pression and chemotactic responsiveness following maturation withCD40L. a, DC were stained with mAb to CCR5 or CCR7 or with anisotype control Ab and analyzed by flow cytometry.b, Cell migration mea-sured using a 96-well chemotaxis chamber. Tweny-five thousand DC wereplaced on top of a membrane containing 5-mm pores, separated from wellscontaining chemokine. Plates were incubated for 90 min at 37°C. Cellsremaining on top of the membrane were then removed, and cells that hadmigrated to the underside were fixed, stained, and enumerated in four highpower fields (hpf). Results are expressed as the mean6 SEM for duplicatewells.

2490 MATURATION AND TRAFFICKING OF DENDRITIC CELLS IN MONKEYS

by guest on Novem

ber 29, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

Mature monkey DC transduced with EGFP ex vivo traffic tolymph nodes in greater numbers than cutaneous DC transfectedin situ with EGFP

To test the capacity of in vitro-derived DC to traffic to lymphnodes, we first transduced DC ex vivo with a gene expressing thecytoplasmic Ag EGFP using recombinant adenovirus and then ma-tured the cells with CD40L. Transduced cells were injected i.d.into the inguinal region, and the draining lymph node was excised36 h later. Approximately 99% of DC expressed EGFP beforeinjection, and a proportion of these cells could subsequently beidentified in the draining lymph node (Fig. 4). No EGFP fluores-cence was detected in control axillary lymph nodes (data notshown). When lymph node cells were stained for expression ofCD83 and CD86, a discrete population of double-labeled cells con-stituting;1% of the total lymph node was identified, representinginterdigitating DC (24, 25). The EGFP1 DC that localized in thelymph node were also CD831CD861, as expected from our invitro analyses of mature DC, and represented 1.5% of total lymph

node DC (Fig. 4). A proportion of CD832 cells was also identifiedwithin the region describing EGFP1 cells in the lymph node. Thismay reflect endogenous cells that have a level of autofluorescenceoverlapping that of EGFP1 cells. To place this result in the contextof other DC-based vaccine protocols, in additional experiments weadministered EGFP gene via gene gun to the inguinal skin of mon-keys and excised lymph nodes 36 h later. Skin DC transfected insitu with EGFP gene were not detected in the draining lymphnodes of two animals using flow cytometric analysis. This is notsurprising, as in previous murine studies fluorescence microscopyhas been required to identify the small number of transfected cellsin lymph nodes (17, 26). When draining lymph nodes from fourother animals were sectioned in their entirety and examined, oc-casional EGFP1 cells were identified (data not shown). Whereasfew EGFP1 cells could be detected in draining lymph nodes, cellscontaining gold particles and expressing EGFP could be identifiedin the superficial dermis (Fig. 5). An acute inflammatory responsewas present in the epidermis as a result of gene gun bombardment(Fig. 5).

Migration of monkey DC to lymph nodes is not dependent onprior in vitro maturation

The above experiments indicated that mature Ag-stimulated mon-key DC could efficiently traffic to lymph nodes after i.d. injection,as was predicted from our in vitro chemotaxis assays. We nextwanted to directly compare the migration of immature and matureDC in vivo. We elected to switch labeling techniques from theadenovirus system to eliminate any potential for inadvertent mat-uration of immature DC before injection. We therefore used aninert lipophilic membrane dye to label cells as in our previousstudies (12). In separate experiments, DiD-labeled immature andmature DC were injected i.d. into the inguinal region of donormonkeys, and the draining inguinal lymph nodes were excised 36 hlater and analyzed by flow cytometry. Labeled DC were stained forexpression of CD83 and CD86 immediately before injection toconfirm the maturation state of cells. Immature cultured DC wereCD832CD861, as expected by our previous analyses (Fig. 6). Incontrast, mature cultured DC were CD831 CD86bright, expressing

FIGURE 6. Immature and mature monkey DC traffic to lymph nodeswith similar efficiency. Immature (day 4) or mature (day7/CD40L) DCwere labeled with DiD and injected i.d. in the inguinal region of donormonkeys. Immature (2.73 106) and mature (3.73 106) DiD labeled DCwere administered to different animals. Draining inguinal lymph nodeswere biopsied 36 h later, and cell suspensions were stained with mAb toCD83 and CD86. One and a half million lymph node cells were acquiredin triplicate by flow cytometry.Left, Representative staining of culturedDC with DiD. Right, CD83 vs CD86 expression of labeled DC immedi-ately before injection (top), total draining lymph node cells (middle), andDiD1 cells in draining lymph nodes (bottom). Perpendicular lines delineateCD831CD861 lymph node DC. Numbers represent the mean6 SEM oftriplicate samples, describing injected DC in draining lymph nodes as apercentage of total lymph node DC. The proportions of injected immatureand mature DC recovered from the draining lymph node were 0.126 0.01and 0.116 0.01%, respectively. Results are representative of two exper-iments for each cell type. SSC, side scatter.

FIGURE 7. CD86 expression is up-regulated on immature DC anddown-regulated on mature DC following trafficking to lymph nodes. Themean fluorescence intensity (MFI) of CD86 expression by cultured DC andDC that localized to lymph nodes following i.d. injection is shown. Sym-bols represent the mean6 SEM of CD86 expression from two experimentsfor immature DC and three experiments for mature DC. Cells were labeledwith DiD (immature and mature DC) or were transduced with Ad-EGFP(mature DC only) to track cells. Error bars are obscured by symbols forimmature DC.p, Significantly different in Student’st test,p 5 0.02.

2491The Journal of Immunology

by guest on Novem

ber 29, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

significantly higher levels of CD86 than immature DC (Fig. 6).Following injection, DiD1 cells were identified in the draininglymph node of animals receiving both immature and mature DC(Fig. 6). No DiD1 cells could be identified in control axillarylymph nodes (data not shown). The DiD1 cells constituted;1%of total lymph node DC regardless of the in vitro maturation state(Fig. 6). The proportion of injected cells that were identified in thelymph node at 36 h ranged from 0.07 to 0.12% over fiveexperiments.

Evidence for regulation of DC maturation in vivo

From our in vitro data we know that immature DC do not migratein response to constitutive lymphoid chemokines MIP3b and6Ckine, yet the in vivo trafficking results clearly indicate migrationof these cells to lymph nodes in significant numbers. One expla-nation for this difference is that upon injection the immature DCundergo spontaneous maturation. To evaluate this possibility wecompared expression of CD86 on immature DC before injectionwith these cells following trafficking to lymph nodes. CD86 up-regulation is a valid marker of maturation, as we have shown inexperiments in vitro (Fig. 2). The mean fluorescence intensity ofCD86 on immature DC that localized to lymph nodes was 3.6-foldhigher than that of immature DC before injection, suggestive of invivo maturation (Fig. 7). Surprisingly, CD86 expression on matureDC that had localized in lymph nodes was, on the average, 3.9times lower than that of the same cells before injection. As a result,expression of CD86 on DC that localized to lymph nodes was

essentially the same regardless of the maturation state at the timeof injection (Fig. 7).

Complete migration of immature, but not mature, DC from skin

To determine the relative efficiency of migration of immature andmature DC from the dermal injection site, we examined skin bi-opsied 36 h after injection of cells. Injection of immature DC re-sulted in a minor localized acute inflammatory response. No fluo-rescently labeled cells could be identified at this time, suggestingcomplete migration of immature DC from the injection site (Fig.8). In marked contrast, a severe acute inflammatory infiltrate waspresent at the site of injection in two of three animals that receivedinjections of mature DC. A large number of fluorescently labeledmature DC were detected in the dermis at 36 h in these animals(Fig. 8). The accumulation of DC was not dependent on labelingwith DiD, as mature DC that expressed EGFP by adenoviral trans-duction were also identified in large numbers in a separate exper-iment (data not shown).

DiscussionThe purpose of this study was to determine in a primate model theextent to which in vitro maturation of DC influences the capacityof DC to traffic in vivo. We were able to reproducibly generateimmature and mature DC from purified CD141 monocytes frommonkey blood using the cytokines GM-CSF and IL-4 with CD40Las a maturation factor (20, 21). Immature DC were characterized asCD832CD861 cells that could rapidly pinocytose and process soluble

FIGURE 8. Mature, but not im-mature, DC are retained in the dermisfollowing i.d. injection into monkeys,associated with an acute inflamma-tory response. Skin sections were ex-amined 36 h after DC injection. He-matoxylin and eosin staining of skininjected with immature (day 4) DClabeled with DiD (a) and red fluores-cence of a serial section of the sameskin (c) are shown. Hematoxylin andeosin staining of skin injected withmature (day7/CD40L) DC labeledwith DiD (b) and red fluorescence ofa serial section of the same skin (d)are shown. Small pockets of inflam-mation are noted in the dermis fol-lowing injection of immature DC, butno fluorescent cells are detected. Incontrast, a prominent infiltration ofneutrophils is noted in the entire der-mis following injection of matureDC, and large numbers of fluorescentcells are present. Original magnifica-tion, 3100 (insets,3200).

2492 MATURATION AND TRAFFICKING OF DENDRITIC CELLS IN MONKEYS

by guest on Novem

ber 29, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

protein but produce negligible quantities of proinflammatory cyto-kines. Mature DC were characterized as CD831CD86bright cells thatwere ineffective at processing soluble Ag but produced large quan-tities of IL-12 and TNF-a. These are recognized properties of im-mature and mature DC as defined in the human system (1). Mon-key DC had the expected chemokine receptor profiles, switchingfrom expression of CCR5 on immature cells to CCR7 followingmaturation with CD40L (2–4, 23). Accordingly, immature DC mi-grated exclusively in response to the inflammatory chemokineMIP-1a, whereas mature DC migrated to constitutive chemokines6Ckine and MIP-3b that are important in attracting DC to T cellareas of lymph nodes (5–7). The clear in vitro similarities betweenmonocyte-derived DC in the human and monkey suggest that con-clusions drawn from in vivo experiments using this model cansafely be extrapolated to the human system.

For the in vivo trafficking experiments, we focused on flow cy-tometric methods to accurately determine the number and pheno-type of DC that migrated to lymph nodes following labeling andinjection. Unexpectedly, we found that immature and CD40L ma-tured DC trafficked to lymph nodes with similar efficiency despitethe polarized chemokine receptor expression and in vitro chemo-tactic responses of these cells. In several experiments using twodifferent labeling techniques, the proportion of lymph node DCthat were derived from the injected population was;1% at 36 hpostinjection regardless of the state of maturation at the time ofinjection. Given this relatively large number, it is likely that Ag-loaded DC will readily contact and stimulate Ag-specific T cellsthat migrate through the paracortex, indicating that DC-based vac-cination should be effective at stimulating immune responses inlarge animals and humans. Using histological techniques in a sim-ilar chimpanzee model, we have previously demonstrated that in-jected DC localize to the lymph node paracortex, with a peak mi-gration occurring from 24–48 h post injection (12). Injected DCintimately associate with T cells in this region and maintain highlevels of HLA-DR, CD40, and CD86 expression similar to inter-digitating DC (12). In preliminary studies in monkeys we notedthat immature and mature DC both localized to T cell-rich areas oflymph nodes when administered together in equal numbers, usingdifferent fluorochromes to track cells (data not shown). From theseanalyses we can conclude that labeled DC detected in tissues byflow cytometry have actively trafficked to lymph nodes and takenup residence as interdigitating DC.

Our findings are consistent with data from the mouse using pu-rified splenic DC of the myeloid lineage (CD8a2), which are anal-ogous to human monocyte-derived DC. These cells traffic to lymphnodes following s.c. injection in the absence of any in vitro mat-uration stimulus, representing about 1% of total lymph node DC(27). A recent report in the murine system indicates that immaturebone marrow-derived DC migrate poorly to lymph nodes follow-ing s.c. injection, relative to migration of GM-CSF transfectedimmature DC (28). Interestingly, GM-CSF transfection did notinduce maturation of DC in this system, based on phenotypic andfunctional analysis (28). When the migration of immature andCD40L matured DC was compared in the same system, injected DCthat trafficked to lymph nodes constituted from 1–2% of lymph nodeDC 48 h after s.c. injection regardless of maturation state (11).

Two potential explanations exist for the trafficking of immatureDC to lymph nodes. The first is that immature DC respond to as yetunknown chemokines that do not use the CCR7 receptor. This isunlikely, as 6Ckine, which binds CCR7, appears to be required forDC homing to lymph nodes. Mice lacking expression of 6Ckinehave a paucity of DC in lymph nodes (7), and administration of Abto 6Ckine blocks the migration of mature DC to lymph nodes wheninjected s.c. into mice (5). A more likely explanation is that im-

mature DC undergo spontaneous maturation upon injection intoskin. Consistent with this hypothesis was the finding that labeledDC that trafficked to lymph nodes had a statistically significantincrease in CD86 expression compared with cultured immatureDC, which is suggestive of functional DC maturation on the basisof our in vitro experiments and other reports in mice (29). A mildinflammatory response was present in the skin at the site of injec-tion at 36 h, and hence injected DC would be exposed to proin-flammatory cytokines that induce maturation (30, 31). Interest-ingly, while CD86 expression was up-regulated on immature DCfollowing localization to lymph nodes, a decrease in expression ofthis costimulatory molecule was noted on mature DC that migratedto lymph nodes, although this change was not statistically signif-icant. Moreover, the relative changes in CD86 expression of in-jected immature and mature DC resulted in an almost identicalexpression of CD86 by these cells in lymph nodes even though thedata were collected from different animals in different experiments.Removal of DC from in vitro exposure to GM-CSF and IL-4 is notlikely to be responsible for the decrease in CD86 expression bymature DC, given the up-regulation of CD86 by immature cellssubjected to the same conditions. However, it is possible that with-drawing cells from high dose exogenous CD40L resulted in mod-ulation of CD86 expression. The data raise the possibility thathomeostatic mechanisms exist in vivo that regulate the degree ofactivation of exogenously supplied DC upon localization to thelymph node.

A second unexpected finding from the in vivo studies was thatimmature DC migrated completely from skin following i.d. injec-tion in monkeys, whereas mature DC tended to remain in the der-mis. Our previous studies in the chimpanzee also demonstratedthat in vitro-derived DC, labeled with DiD but administered with-out Ag or prior maturation, migrated away from an s.c. injectionsite within 48 h of injection (12). Fossum (32) reported that puri-fied lymph DC injected into footpads of mice mostly were retainedat the site of injection. Lymph DC represent a maturing DC pop-ulation (33), and hence these findings may be consistent with ourresults. More recently, Ag-stimulated blood-derived human DCused therapeutically in human cancer patients were only partiallycleared from an i.d. injection site and apparently not at all follow-ing s.c. injection (34). Labeur et al. (11) also found that the ma-jority of mouse DC remained at the site of s.c. injection regardlessof the in vitro maturation state before administration. The differ-ences between our report and others may be due in part to the typesof labels used and the different species studied. However, our find-ings raise the question of whether there is a fundamental differencebetween the capacity of immature and mature DC to be clearedfrom a skin injection site. Mature DC produce large quantities ofproinflammatory cytokines and probably induced the severe in-flammatory response observed in the dermis. Therefore, while amild inflammatory response may induce maturation of immatureDC and promote migration, it is possible that a more severe in-flammatory response impairs migration of mature DC out of skin.

Despite the finding that 1% of lymph node DC were derivedfrom the injected population, migration was relatively inefficient,as only ;0.1% of injected cells could be accounted for in thedraining lymph node at 36 h postinjection. Similar quantitativestudies in the murine system also indicate that only 0.3% of in-jected DC can subsequently be detected in the draining lymphnode (27). The reasons for this apparent inefficiency are notknown. A large number of mature DC in our study could be ac-counted for in the dermis; however, the fate of immature DC thatmigrated completely from the skin but were not detected in thedraining lymph node is uncertain. We are able to rule out thepossibility that DC localized in substantial numbers to other nodes

2493The Journal of Immunology

by guest on Novem

ber 29, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

in the inguinal chain, as secondary nodes were collected and foundto contain few, if any, labeled cells (data not shown). It is possiblethat injected cells trafficked via blood to distant lymphoid sites.Labeled cells were not detected in venous blood samples at 30 minand 36 h postinjection in several experiments (data not shown),although massive dilution by circulating leukocytes would makesuch cells very difficult to detect.

Our studies with membrane-labeled DC are limited by the factthat the dye does not represent a model Ag, and hence the traf-ficking results do not relate directly to DC-based immunotherapy.However, we showed, using a recombinant adenovirus, that DCtransduced in vitro with EGFP traffic to lymph nodes relativelyefficiently and retain high level expression for at least 36 h. Thesedata clearly support the application of adenoviral vectors as ameans of loading DC with Ag for vaccination protocols (35–37).By comparison, DC transfected in situ with EGFP using a genegun were not detectable in draining lymph nodes by flow cytom-etry, although single EGFP-transfected cells could be detected inskin and lymph node by fluorescence microscopy. This would sug-gest a frequency of EGFP1 DC that migrated to lymph nodesfollowing gene gun administration to be,0.05% of total lymphnode DC, at least 20-fold lower than that of ex vivo transducedDC. Although the number of directly transfected DC that localizein lymph nodes is relatively small, as has been shown by others inthe murine system (26), we and others have shown that this methodclearly represents a potent means of inducing Ag-specific immuneresponses (17, 26). Our data indicate that it is possible to generatesubstantially larger numbers of gene-transfected DC in draininglymph nodes using ex vivo manipulation compared with gene gunadministration. Whether this difference translates into a strongerimmune response is yet to be determined.

Our results suggest that DC-based vaccine protocols can be sub-stantially more flexible with respect to in vitro maturation than waspreviously thought. It may be preferable to administer immatureDC immediately following Ag pulsing, especially when Ag is sup-plied as exogenous peptide that may have a limited half-life. Fac-tors in the dermis will induce maturation of injected cells, leadingto responsiveness to constitutive chemokines and migration intolymphatics and lymph nodes. In contrast, our finding that matureDC remain at the site of injection in such large numbers suggeststhat a majority of in vitro matured DC may not contribute effec-tively to the immune response. It will be important to determine inthis model the relative immunogenicities of vaccines using imma-ture and mature DC as well as cutaneous immunization via genegun to establish the optimal DC-based vaccine system forhuman use.

AcknowledgmentsWe thank Anita Trichel, Dawn McClemens, Barbara Murray, and CieloCastillo for technical assistance; Adriana Larregina for evaluation of skinsections; and Ciprian Calmonte for computer assistance. We are grateful toLijun Wu for the gift of the CCR7 mAb, to Andrea Gambotto for the giftof the recombinant adenoviral vector, and to Joe Ahearn for provision ofthe FACSCalibur flow cytometer.

References1. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control of

immunity. Nature 392:245.2. Lin, C. L., R. M. Suri, R. A. Rahdon, J. M. Austyn, and J. A. Roake. 1998.

Dendritic cell chemotaxis and transendothelial migration are induced by distinctchemokines and are regulated on maturation.Eur. J. Immunol. 28:4114.

3. Sozzani, S., P. Allavena, G. D’Amico, W. Luini, G. Bianchi, M. Kataura, T. Imai,O. Yoshie, R. Bonecchi, and A. Mantovani. 1998. Differential regulation of che-mokine receptors during dendritic cell maturation: a model for their traffickingproperties.J. Immunol. 161:1083.

4. Sallusto, F., P. Schaerli, P. Loetscher, C. Schaniel, D. Lenig, C. R. Mackay,S. Qin, and A. Lanzavecchia. 1998. Rapid and coordinated switch in chemokinereceptor expression during dendritic cell maturation.Eur. J. Immunol. 28:2760.

5. Saeki, H., A. M. Moore, M. J. Brown, and S. T. Hwang. 1999. Cutting edge:secondary lymphoid-tissue chemokine (SLC) and CC chemokine receptor 7(CCR7) participate in the emigration pathway of mature dendritic cells from theskin to regional lymph nodes.J. Immunol. 162:2472.

6. Ngo, V. N., H. L. Tang, and J. G. Cyster. 1998. Epstein-Barr virus-inducedmolecule 1 ligand chemokine is expressed by dendritic cells in lymphoid tissuesand strongly attracts naive T cells and activated B cells.J. Exp. Med. 188:181.

7. Gunn, M. D., S. Kyuwa, C. Tam, T. Kakiuchi, A. Matsuzawa, L. T. Williams, andH. Nakano. 1999. Mice lacking expression of secondary lymphoid organ che-mokine have defects in lymphocyte homing and dendritic cell localization.J. Exp. Med. 189:451.

8. Morikawa, Y., M. Furotani, N. Matsuura, and K. Kakudo. 1993. The role ofantigen-presenting cells in the regulation of delayed-type hypersensitivity. II.Epidermal Langerhans’ cells and peritoneal exudate macrophages.Cell. Immu-nol. 152:200.

9. Kupiec-Weglinski, J. W., J. M. Austyn, and P. J. Morris. 1988. Migration patternsof dendritic cells in the mouse. Traffic from the blood, and T cell-dependent and-independent entry to lymphoid tissues.J. Exp. Med. 167:632.

10. Ingulli, E., A. Mondino, A. Khoruts, and M. K. Jenkins. 1997. In vivo detectionof dendritic cell antigen presentation to CD41 T cells.J. Exp. Med. 185:2133.

11. Labeur, M. S., B. Roters, B. Pers, A. Mehling, T. A. Luger, T. Schwarz, andS. Grabbe. 1999. Generation of tumor immunity by bone marrow-derived den-dritic cells correlates with dendritic cell maturation stage.J. Immunol. 162:168.

12. Barratt-Boyes, S. M., S. C. Watkins, and O. J. Finn. 1997. In vivo migration ofdendritic cells differentiated in vitro: a chimpanzee model.J. Immunol. 158:4543.

13. Barratt-Boyes, S. M., R. A. Henderson, and O. J. Finn. 1996. Chimpanzee den-dritic cells with potent immunostimulatory function can be propagated from pe-ripheral blood.Immunology 87:528.

14. Jones, L. J., R. H. Upson, R. P. Haugland, N. Panchuk-Voloshina, and M. Zhou.1997. Quenched BODIPY dye-labeled casein substrates for the assay of proteaseactivity by direct fluorescence measurement.Anal. Biochem. 251:144.

15. Ranieri, E., W. Herr, A. Gambotto, W. Olson, D. Rowe, P. D. Robbins,L. S. Kierstead, S. C. Watkins, L. Gesualdo, and W. J. Storkus. 1999. Dendriticcells transduced with an adenovirus vector encoding Epstein-Barr virus latentmembrane protein 2B: a new modality for vaccination.J. Virol. 73:10416.

16. Vremec, D., and K. Shortman. 1997. Dendritic cell subtypes in mouse lymphoidorgans: cross-correlation of surface markers, changes with incubation, and dif-ferences among thymus, spleen, and lymph nodes.J. Immunol. 159:565.

17. Condon, C., S. C. Watkins, C. M. Celluzzi, K. Thompson, and L. D. Falo, Jr.1996. DNA-based immunization by in vivo transfection of dendritic cells.Nat.Med. 2:1122.

18. Zhou, L. J., and T. F. Tedder. 1996. CD141 blood monocytes can differentiateinto functionally mature CD831 dendritic cells.Proc. Natl. Acad. Sci. USA 93:2588.

19. O’Doherty, U., R. Ignatius, N. Bhardwaj, and M. Pope. 1997. Generation ofmonocyte-derived dendritic cells from precursors in rhesus macaque blood.J. Im-munol. Methods 207:185.

20. Caux, C., C. Massacrier, B. Vanbervliet, B. Dubois, C. Van Kooten, I. Durand,and J. Banchereau. 1994. Activation of human dendritic cells through CD40cross-linking.J. Exp. Med. 180:1263.

21. Cella, M., D. Scheidegger, K. Palmer-Lehmann, P. Lane, A. Lanzavecchia, andG. Alber. 1996. Ligation of CD40 on dendritic cells triggers production of highlevels of interleukin-12 and enhances T cell stimulatory capacity: T-T help viaAPC activation.J. Exp. Med. 184:747.

22. Sallusto, F., M. Cella, C. Danieli, and A. Lanzavecchia. 1995. Dendritic cells usemacropinocytosis and the mannose receptor to concentrate macromolecules in themajor histocompatibility complex class II compartment: downregulation by cy-tokines and bacterial products.J. Exp. Med. 182:389.

23. Dieu, M. C., B. Vanbervliet, A. Vicari, J. M. Bridon, E. Oldham, S. Ait-Yahia,F. Briere, A. Zlotnik, S. Lebecque, and C. Caux. 1998. Selective recruitment ofimmature and mature dendritic cells by distinct chemokines expressed in differentanatomic sites.J. Exp. Med. 188:373.

24. Clarke, G. J., and D. N. J. Hart. 1999. Phenotypic characterization of dendriticcells. In Dendritic Cells: Biology and Clinical Applications. M. T. Lotze, andA. W. Thomson, eds. Academic Press, London, p. 555.

25. Zhou, L. J., R. Schwarting, H. M. Smith, and T. F. Tedder. 1992. A novel cell-surface molecule expressed by human interdigitating reticulum cells, Langerhanscells, and activated lymphocytes is a new member of the Ig superfamily.J. Im-munol. 149:735.

26. Porgador, A., K. R. Irvine, A. Iwasaki, B. H. Barber, N. P. Restifo, andR. N. Germain. 1998. Predominant role for directly transfected dendritic cells inantigen presentation to CD81 T cells after gene gun immunization.J. Exp. Med.188:1075.

27. Smith, A. L., and B. Fazekas de St. Groth. 1999. Antigen-pulsed CD8a1 den-dritic cells generate an immune response after subcutaneous injection withouthoming to the draining lymph node.J. Exp. Med. 189:593.

2494 MATURATION AND TRAFFICKING OF DENDRITIC CELLS IN MONKEYS

by guest on Novem

ber 29, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

28. Curiel-Lewandrowski, C., K. Mahnke, M. Labeur, B. Roters, W. Schmidt,R. D. Granstein, T. A. Luger, T. Schwarz, and S. Grabbe. 1999. Transfection ofimmature murine bone marrow-derived dendritic cells with the granulocyte-mac-rophage colony-stimulating factor gene potently enhances their in vivo antigen-presenting capacity.J. Immunol. 163:174.

29. Inaba, K., M. Witmer-Pack, M. Inaba, K. S. Hathcock, H. Sakuta, M. Azuma,H. Yagita, K. Okumura, P. S. Linsley, S. Ikehara, et al. 1994. The tissue distri-bution of the B7-2 costimulator in mice: abundant expression on dendritic cellsin situ and during maturation in vitro.J. Exp. Med. 180:1849.

30. Roake, J. A., A. S. Rao, P. J. Morris, C. P. Larsen, D. F. Hankins, andJ. M. Austyn. 1995. Dendritic cell loss from nonlymphoid tissues after systemicadministration of lipopolysaccharide, tumor necrosis factor, and interleukin 1.J. Exp. Med. 181:2237.

31. Cumberbatch, M., and I. Kimber. 1992. Dermal tumour necrosis factor-a inducesdendritic cell migration to draining lymph nodes, and possibly provides one stim-ulus for Langerhans’ cell migration.Immunology 75:257.

32. Fossum, S. 1988. Lymph-borne dendritic leucocytes do not recirculate, but enter

the lymph node paracortex to become interdigitating cells.Scand. J. Immunol.27:97.

33. Bujdoso, R., J. Hopkins, B. M. Dutia, P. Young, and I. McConnell. 1989. Char-acterization of sheep afferent lymph dendritic cells and their role in antigen car-riage.J. Exp.Med. 170:1285.

34. Morse, M. A., R. E. Coleman, G. Akabani, N. Niehaus, D. Coleman, andH. K. Lyerly. 1999. Migration of human dendritic cells after injection in patientswith metastatic malignancies.Cancer Res. 59:56.

35. Song, W., H. L. Kong, H. Carpenter, H. Torii, R. Granstein, S. Rafii,M. A. Moore, and R. G. Crystal. 1997. Dendritic cells genetically modified withan adenovirus vector encoding the cDNA for a model antigen induce protectiveand therapeutic antitumor immunity.J. Exp. Med. 186:1247.

36. Dietz, A. B., and S. Vuk-Pavlovic. 1998. High efficiency adenovirus-mediatedgene transfer to human dendritic cells.Blood 91:392.

37. Zhong, L., A. Granelli-Piperno, Y. Choi, and R. M. Steinman. 1999. Recombi-nant adenovirus is an efficient and non-perturbing genetic vector for human den-dritic cells.Eur. J. Immunol. 29:964.

2495The Journal of Immunology

by guest on Novem

ber 29, 2015http://w

ww

.jimm

unol.org/D

ownloaded from