Induction of Targeted Cell Migration by CutaneousAdministration of a DNA Vector Encoding aBiologically Active Chemokine CCL21Ahmad Jalili1, Mikhail Pashenkov1,3, Ernst Kriehuber1, Christine Wagner1, Hideki Nakano2, Georg Stingl1

and Stephan N. Wagner1

Skin inflammation can induce local expression of CCL21, which is subsequently drained to lymph nodes (LNs)influencing their cellular composition. To determine whether the same can be achieved by dermaladministration of a plasmid DNA (pDNA) encoding CCL21, we generated a pDNA-based gene constructallowing high-level expression of CCL21. Expression and secretion of biologically active CCL21 were confirmedin vitro by immunohistochemistry, western blot analysis, ELISA, and transwell chemotactic assays. In vivoexperiments showed cellular expression of transgenic CCL21 after particle-mediated gene gun delivery ofpDNA into skin. CCL21 was expressed in the epidermis, consequently secreted into the upper dermis, andtransported into the draining LNs, which resulted in increased CCL21 concentration, total cell number, andfrequencies of CD11cþ DCs and CD4þ /CD62Lþ naıve, CD4þ /CD62L�, and CD8þ /CD62L� effector memory T-cells (expressing CCL21 receptors CCR7 or CXCR3), as well as retention of adoptively transferred T-lymphocytes,in the draining LNs of plt/plt mice (lacking endogenous expression of CCL21). Our studies show thatbiologically active CCL21 can be overexpressed by genetic means in vitro and in vivo. This strategy allowsreconstitution of a genetic defect and colocalization of different cell types in the secondary lymphoid organs,an important prerequisite for targeted cell migration.

Journal of Investigative Dermatology (2010) 130, 1611–1623; doi:10.1038/jid.2010.31; published online 25 February 2010

INTRODUCTIONThe skin and its draining peripheral lymph nodes (PLNs)comprise a distinct immunological compartment, continu-ously exposed to a potentially pathogenic exogenousenvironment. Within this immunological compartment,skin-residing dendritic cells (DCs) function as sentinels thatacquire antigens and initiate antigen-specific adaptive

immune responses as diverse as peripheral toleranceand immunity in the draining PLNs where they meet naıveT-lymphocytes (Steinman and Nussenzweig, 2002; Bancher-eau et al., 2003; Steinman, 2003).

Chemokines are a group of small (B8–14 kDa) andstructurally related molecules that have originally beendescribed to regulate cell trafficking through interactionswith a subset of seven transmembrane, G protein-coupledreceptors (Rossi and Zlotnik, 2000; Colditz et al., 2007).These chemokine–chemokine receptor interactions have afundamental role in the development, homeostasis, andinduction of innate and adaptive immune responses (Zlotnikand Yoshie, 2000). Chemokines have been broadly classifiedinto ‘‘inflammatory’’ and ‘‘homeostatic’’ chemokines on thebasis of the site of production and induction of stimuli(Zlotnik and Yoshie, 2000; Laing and Secombes, 2004).Whereas inflammatory chemokines share several receptorsand are induced to high-level expression by inflammatorystimuli, homeostatic chemokines are usually receptor-restricted and are primarily involved in the homeostasis ofleukocyte traffic and lymphoid cell localization.

Secondary lymphoid-organ chemokine (CCL21) has longbeen regarded as a homeostatic chemokine as it isconstitutively expressed in mice by (i) stromal cells in theT-cell-rich zones of PLNs (Gunn, 2003); (ii) afferentlymphatic endothelial cells; and (iii) high endothelial venules

& 2010 The Society for Investigative Dermatology www.jidonline.org 1611

ORIGINAL ARTICLE

Received 13 June 2009; revised 17 October 2009; accepted 20 December2009; published online 25 February 2010

1Division of Immunology, Allergy and Infectious Diseases, Department ofDermatology, Medical University of Vienna, Allgemeines Krankenhaus,Vienna, Austria and 2Laboratory of Respiratory Biology, National Institute ofEnvironmental Health Sciences, National Institutes of Health, ResearchTriangle Park, North Carolina, USA

Correspondence: Ahmad Jalili or Stephan N. Wagner, Department ofDermatology, Division of Immunology, Allergy and Infectious Diseases(DIAID), Medical University of Vienna, Allgemeines Krankenhaus, WahringerGurtel 18-20, Vienna A-1090, Austria. E-mail: ahmad.jalili@meduniwien.ac.at or stephan.wagner@meduniwien.ac.at

3Current address: Laboratory of Clinical Immunology, Institute ofImmunology, Moscow, Russia.

Abbreviations: Ab, antibody; ANOVA, analysis of variance; BMDC, bonemarrow-derived dendritic cell; CFSE, 5,6-carboxyfluorescein diacetatesuccinimidyl ester; DC, dendritic cell; HEV, high endothelial venule; LPS,lipopolysaccharide; mCCL21, murine CCL21; pDNA, plasmid DNA; PLN,peripheral lymph node; plt, paucity of lymph node T-cells; rmCCL21,recombinant murine CCL21 chemokine; RT, room temperature; SLO,secondary lymphoid organ

(HEVs). In humans, CCL21 is expressed by stromal cells in theT-cell-rich zones of PLNs and subsequently exposed on HEVs(Carlsen et al., 2005). The CCL21 receptor, CCR7, isexpressed on mature DCs, naıve T-cells, and a subset ofmemory T-cells (Ebert et al., 2005; Yoneyama et al., 2005). Inthe steady state, CCL21 is essential for coordination ofCCR7þ cell migration, that is, directional migration andpositioning of lymphocytes and DCs within the T-cell zone ofsecondary lymphoid organs (SLOs). The importance of thischemokine–chemokine receptor interaction under homeo-static conditions is best illustrated in studies of mice lackingeither this chemokine (Gunn et al., 1999) or its receptor(Forster et al., 1999). plt/plt mice harbor a spontaneousmutation that abolishes CCL21’s expression in lymphatictissues (Gunn et al., 1999; Luther et al., 2000). As a result,antigen-presenting DCs and naıve T-cells fail to enter theT-cell zones of SLOs (Gunn et al., 1999). A similarphenomenon can be observed in CCR7-deficient mice(Forster et al., 1999). In these mice, the overall micro-architecture of SLOs is altered with prominent B-cell follicleswith considerably enlarged germinal centers and significantreduction or even absence of T-cell areas (Forster et al., 1999).These alterations are thought to be caused by the impairedentry and retention of CCR7-expressing, antigen-presentingDCs and CD4þ /CD62Lþ naıve/central memory T-lympho-cytes (Forster et al., 1999). Consistently, CCR7 has beenimplicated in the migration of chronic lymphocyticleukemia cells into the LNs and the development of disease-associated peripheral lymphadenopathy in humans (Till et al.,2002).

Colocalization of DCs and naıve/central memory T-cells inthe PLNs is a pre-requisite for the induction of immuneresponses (Banchereau et al., 2003). It has been previouslyshown that when recombinant CCL21 was injected intracu-taneously into plt/plt mice, the chemokine entered theafferent lymph vessels and accumulated in the draining PLNsleading to luminal presentation of CCL21 in a subset of HEVs(Stein et al., 2000).

In this study we show that gene gun-mediated dermaladministration of a plasmid DNA (pDNA) encoding abiologically active CCL21 chemokine leads to the expressionof the chemokine in the epidermis and its subsequentdrainage into PLNs, leading to accumulation of CCR7-expressing DCs and naıve/central memory T-cells, as wellas CXCR3-expressing CD4þ /CD62L� and CD8þ /CD62L�

effector T-cells, in these organs. Additionally, transgenicCCL21 shows biological activity on DCs, including inductionof antigen uptake and upregulation of CD40 surfaceexpression (previously described to be responsible forsurvival of migrating DCs).

RESULTSTransfection with murine CCL21 pDNA induces the expressionand secretion of CCL21

After translation in the endoplasmic reticulum, chemokinesare transported to the Golgi complex and subsequently tosecretory granules (Nickel, 2003) where they have beenshown to be stored. Upon stimulation with inflammatory

cytokines such as tumor necrosis factor, signaling pathwayssuch as mitogen-activated protein kinase-signaling pathwaysextracellular signal-regulated kinase and p38 are activatedthat induce rapid transport of stored chemokines to the cellsurface and secretion into the pericellular space (Chensue,2001; Wong et al., 2006). To determine whether transfectionwith pVR1012-mCCL21-IRES-EGFP (Figure 1a) andpVR1012-mCCL21/EGFP (Figure 1b) pDNA induces anintracellular expression and trafficking pattern of CCL21comparable to physiological expression levels, transfectedCOS-7 cells were monitored for expression, intracellulardistribution, and secretion of transgenic murine CCL21(mCCL21).

In western blot analysis of protein lysates derived fromCOS-7 cells transfected with pVR1012-mCCL21-IRES-EGFP,the expected 15-kDa band of mCCL21 (Christopherson et al.,2001) could be detected (Figure 1c). Transfection withpVR1012-mCCL21/EGFP pDNA encoding a chimericmCCL21/EGFP resulted in the expression of a protein witha higher molecular weight of around 75 kDa, correspondingto the 15-kDa molecular weight of mCCL21 plus the 60-kDamolecular weight of enhanced green fluorescent protein(EGFP) (Shigaki et al., 2005; Figure 1c).

Immunocytochemistry of COS-7 cells transfected withpVR1012-mCCL21-IRES-EGFP uncovered a cytoplasmic ex-pression pattern of mCCL21 (Figure 1d). As expected,confocal laser-scanning microscopy detected a similarexpression pattern of the separately translated EGFP (Figure1e). To study the subcellular distribution of transgenicmCCL21 in more detail, we transfected COS-7 cells withthe pVR1012-mCCL21/EGFP pDNA encoding a chimericmCCL21/EGFP protein. Confocal microscopy showedmCCL21/EGFP to be particularly enriched in two ways:juxtanuclearly, probably in the Golgi complex, and juxta-membraneously in the secretory granules (Figure 1f). Thesedata are in accordance with the intracellular expression andsecretion pathway described for chemokines (Nickel, 2003).

mCCL21 could also be detected in the supernatants ofCOS-7 cells transfected with pVR1012-mCCL21-IRES-EGFPpDNA. As measured by ELISA 48 hours after transfection,0.5�106 cells transfected with 5 mg of pVR1012-mCCL21-IRES-EGFP were able to secrete around 60 ng ml�1 ofmCCL21, whereas supernatants from cells transfected withcontrol pVR1012 pDNA did not contain mCCL21 atdetectable amounts (o10 pg ml�1) (data not shown).

pDNA-encoded murine CCL21 is functionally active

First, we determined whether transgenic CCL21 binds to cellsexpressing CCR7. Bone marrow-derived dendritic cells(BMDCs) were generated as described earlier (Jalili et al.,2004). Around 55% of these cells express CCR7 asdetermined by flow-cytometric analysis (Figure 2a3). COS-7cells were transfected with pVR1012-mCCL21/EGFP, encod-ing chimeric mCCL21/EGFP, and supernatants were collected48 hours later. As indicated by flow cytometry, approximately46% of BMDCs (incubated with above mentioned super-natants) bound to chimeric mCCL21/EGFP (Figure 2b2).Binding could be completely blocked by pre-incubation of

1612 Journal of Investigative Dermatology (2010), Volume 130

A Jalili et al.Chemokine-Induced Targeted Cell Migration

supernatants with anti-mCCL21 neutralizing antibody (Ab)(Figure 2b3).

We next determined whether plasmid-encoded murineCCL21 also mediates the chemotaxis of CCR7-expressingmurine cell populations in standard transwell migrationassays. In these assays, supernatants of COS-7 cells werecollected 48 hours after transfection with pVR1012-mCCL21-IRES-EGFP and compared with different concentrations ofrecombinant mCCL21 (rmCCL21) protein for their inductionof chemotaxis of splenocytes (containing both DCs andnaıve/central memory T-cells) isolated from wild-type BALB/cmice. As indicated by flow-cytometric analysis, approxi-mately 50% of BALB/c splenocytes express CCR7 and

approximately 25% express CD4/CD62L representing naıve/central memory T-cells (data not shown). rmCCL21 proteininduced concentration-dependent chemotaxis of splenocytes(Figure 2c) as did supernatants of COS-7 cells transfectedwith pVR1012-mCCL21-IRES-EGFP (supernatants of COS-7cells transfected with pVR1012-mCCL21-IRES-EGFP contain-ing B60 ng ml�1 of CCL21 were as effective as 1 mg ml�1 ofrmCCL21 presumably because of post-translation modifica-tion of the transgenic CCL21, making it biologically moreactive, a phenomenon absent in the case of rmCCL21;Figure 2c, Po0.001, analysis of variance (ANOVA) test withTukey post-test). In contrast, chemotaxis could not beobserved with supernatants derived from control pVR1012-

Stop condon

pVR1012-mCCL21-IRES-EGFP

Stop condon

mCCL21 IRES EGFP mCCL21 EGFP

MCS MCS

pCMV BGH pA pCMV BGH pA

Kanamycinresistance

pUC18 Kanamycinresistance

pUC18

pVR1012-mock

pVR1012-mCCL21-IRES-EGFP pVR1012-mCCL21/GFP

pVR

1012

-mC

CL2

1/E

GF

P

pVR

1012

-mC

CL2

1-IR

ES

-EG

FP

pVR

1012

-moc

k

CO

S-7

cel

ls u

ntra

nsfe

cted

~ 75 kDa

~ 15 kDa

Figure 1. mCCL21 is efficiently expressed and secreted by genetic means in vitro. (a) pVR1012-mCCL21-IRES-EGFP, a bicistronic pDNA vector expressing

mCCL21 and EGFP under the pCMV promoter, with two transgenes separated by an IRES sequence allowing efficient translation of the second transgene

upon transfection into cells. MCS indicates the multiple cloning site of the pDNA. (b) pVR1012-mCCL21/EGFP expresses mCCL21, which is tagged at its

N-terminus to EGFP. (c) Western blot analysis of protein lysates derived from COS-7 cells transfected with pVR1012-mCCL21/EGFP showing expression of

mCCL21 with a molecular weight of about 60 kDa higher than the non-chimeric mCCL21 expressed by pVR1012-mCCL21-IRES-EGFP, confirming successful

tagging of mCCL21 with EGFP. (d, e) Transfection of COS-7 cells shows efficient expression of both mCCL21 and EGFP as shown by mCCL21

immunocytochemistry and fluorescence microscopy for EGFP, respectively. (f) Immunofluorescence microscopy of COS-7 cells transfected with pVR1012-

mCCL21/EGFP shows subcellular localization of mCCL21/EGFP chimera juxtanuclearly, probably in the Golgi complex, and juxtamembraneously in secretory

granules. Bar¼ panel d: 100 mm; panels e and f: 10mm.

www.jidonline.org 1613

A Jalili et al.Chemokine-Induced Targeted Cell Migration

IRES-EGFP-transfected COS-7 cells (Figure 2c). Chemotaxis ofsplenocytes toward supernatants derived from COS-7 cellstransfected with pVR1012-mCCL21-IRES-EGFP could beblocked by anti-mCCL21 neutralizing Ab (Figure 2c). Recom-binant mCCL21 protein also induced a dose-dependentenrichment for CCR7-expressing cells in the transmigratedsplenocyte population, as indicated by flow-cytometric analy-sis for CD4þCD62Lþ naıve/central memory T-cells (40% ascompared with 20% of input splenocytes). The same phenom-enon could be observed by supernatants of COS-7 cellstransfected with pVR1012-mCCL21-IRES-EGFP (Figure 2d).

CCL21 induces FITC-dextran uptake by BMDCs and upregulatesCD40 expression on BMDCs without evidence of maturation

To study whether CCL21 is able to induce the maturation ofDCs, BMDCs were treated for 24 hours with rmCCL21 (doserange 0.1–5 mg ml�1) as well as with supernatants of COS-7cells collected 48 hours after transfection with pVR1012-mCCL21-IRES-EGFP (containing B60 ng ml�1 mCCL21 as

measured by ELISA). Lipopolysaccharide (LPS)-treatedBMDCs served as positive control (Hemmi and Akira,2005). While LPS upregulated the surface expression ofCD80, CD86, and CCR7; induced morphological changesassociated with maturation such as increase in the cell-surface area and decrease in the granularity; and enhancedthe allostimulatory capacity of BMDCs in mixed-leukocytereaction assays (data not shown) (Granelli-Piperno et al.,2004), rmCCL21 or mCCL21 derived from the supernatants ofCOS-7 cells did not show any of these effects (data notshown). In contrast, we observed significant upregulation(Po0.001, ANOVA test with Tukey post-test) of the surfaceexpression of CD40 (Figure 3b) upon treatment withrmCCL21 or mCCL21 derived from the supernatants ofCOS-7 cells, a phenomenon also observed by treatment ofthese cells with LPS (Figure 3c). These in vitro data weresupported in vivo by using the pVR1012-mCCL21 pDNAvector expressing mCCL21 administered to the skin of BALB/c mice using the helium-based gene gun technique. There,

FL3

-hei

ght

FL3

-hei

ght

FL3

-hei

ght

FL3

-hei

ght

FL3

-hei

ght

FL3

-hei

ght

CD11C CCR7

FL3-height EGFPEGFP

Splenocytes

0 10,000 20,000 30,000 40,000 50,000 0 10 20 30 40

Total number of cells migrated % of CD4+ /CD62L+ cells migrated

pVR1012-IRES-EFP control

pVR1012-mCCL21-IRES-EGFP

rmCCL21 0.1 µg ml–1

rmCCL21 1 µg ml–1

pVR1012-mCCL21-IRES-EGFP &5 µg/ml anti-mCCL21

104

103

102

101

100

104103102101100

104

103

102

101

100

104103102101100

104

103

102

101

100

104103102101100

104

103

102

101

100

104103102101100

104

103

102

101

100

104103102101100

104

103

102

101

100

104103102101100

Figure 2. pDNA-encoded murine CCL21 is functionally active. (a) Cultured BMDCs from wild-type BALB/c mice express CD11c and CCR7 as measured by

flow cytometry (2 and 3, respectively). The isotype-matched Ab shows no unspecific binding (1). (b) These BMDCs bind to chimeric mCCL21/EGFP (2) when

incubated with the supernatants of COS-7 cells transfected with pVR1012-mCCL21/EGFP, as shown by flow cytometric analysis for EGFP fluorescence. No

fluorescence signal could be observed when BMDCs were incubated with supernatants derived from COS-7 cells transfected with the control pVR1012-IRES-

EGFP (1). Pre-incubation of supernatants of COS-7 cells transfected with pVR1012-mCCL21/EGFP with neutralizing anti-mouse CCL21 Ab leads to complete

inhibition of binding (3). One representative of two independent experiments is shown. (c) In transwell migration assays, freshly isolated splenocytes from

wild-type BALB/c mice migrated toward the supernatants of COS-7 cells transfected with pVR1012-mCCL21-IRES-EGFP as well as rmCCL21. Supernatants

collected from pVR1012-mCCL21-IRES-EGFP-transfected COS-7 cells for 48 hours were as effective as 1 mg ml�1 rmCCL21. The neutralizing anti-mCCL21 Ab

completely blocked migration. (d) Note the preferential enrichment of the transmigrated cell population for CD4þ /CD62Lþ T-cells, known to express CCR7,

by transgenic mCCL21 at levels comparable to recombinant mCCL21. Data collected from three experiments; the error bars indicate the SD.

1614 Journal of Investigative Dermatology (2010), Volume 130

A Jalili et al.Chemokine-Induced Targeted Cell Migration

we observed significant upregulation of surface expression ofCD40 on CD11cþ DCs derived from pVR1012-mCCL21-treated mice (Figure 3g) as compared with mock pDNA(pVR1012)-treated or untreated mice (Figure 3e and f,respectively, Po0.001, ANOVA test with Tukey post-test).rmCCL21, as well as mCCL21 derived from supernatants ofCOS-7 cells, significantly induced the uptake of FITC-dextranby BMDCs (Figure 3d, Po0.001, ANOVA test with Tukeypost-test).

Gene gun-mediated administration of mCCL21 pDNA to theskin leads to the expression of CCL21 in the epidermis, withsubsequent secretion into the dermis and accumulation in thedraining PLNs

Next, we investigated whether transgenic CCL21 expressioncan be achieved in vivo upon gene gun-mediated adminis-tration of a CCL21-encoding plasmid. To identify cells thatare targeted by this approach, we administered the pVR1012-mCCL21-IRES-EGFP construct into mice and performed anexpression analysis of transgenic EGFP, which, after transfec-tion is known to remain intracellularly and not to be secretedinto the surroundings of a transfected cell (Ogura et al.,2004). Laser-scanning microscopy performed 24 hours aftercutaneous administration of pVR1012-mCCL21-IRES-EGFP

showed EGFP expression in the epidermis (Figure 4a),including areas surrounding hair follicles, but not in celltypes residing in the dermis. In time-course experiments,expression of transgenic EGFP was highest at 24 hours,gradually decreased on days 3 and 7 after gene gun-mediatedadministration, and was completely lost on day 14 (Figure 4aand data not shown, respectively). At neither time point wasEGFP expression detected in the dermis. These data argue foran in vivo expression pattern of pVR1012-mCCL21-IRES-EGFP, with restriction to keratinocytes of the epidermis, andagainst a significant transfection of other cell types residing inthe dermis.

We then determined whether administration of mCCL21pDNA leads to expression of the encoded transgene in theskin. Using the helium-based gene gun technique theabdominal area skin of plt/plt BALB/c mice received threenon-overlapping shots of 1 mg pVR1012-mCCL21-IRES-EGFPor control pVR1012-mock and was analyzed for theexpression of mCCL21 24 hours later. As expected, pro-nounced intracellular immunoreactivity could be observed inthe keratinocytes of the epidermis after administration ofpVR1012-mCCL21-IRES-EGFP but not control pVR1012-mock (Figure 4b). Interestingly, diffuse immunostaining wasalso present throughout the dermis after administration of

Untreated

d

Untreated

Untreated

Fold increase inFITC-dextran

uptake at37/4 ˚C

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

Cou

nts

PE CD40 PE CD40 PE CD40

0

10

20

30

40

50

0

10

20

30

40

50

0

10

20

30

40

50

PE CD40 PE CD40 PE CD40

6

4

2

0

20

15

10

5

0

20

15

10

5

0

20

15

10

5

0

100 101 102 103 104

100 101 102 103 104 100 101 102 103 104100 101 102 103 104

100 101 102 103 104 100 101 102 103 104

CCL21

CCL21

pVR1012-mCCL21pVR1012 control

52%26%19%

****

LPS

LPS

42% 65% 89%

M1 M1

M1 M1M1

M1

Figure 3. CCL21 induces the expression of CD40 on DCs. (a–c) CD40 expression by BMDCs generated in the presence of GM-CSF and IL-4. (b, c)

rmCCL21 (1 mg ml�1) induces the expression of CD40 by BMDCs, as does LPS. (d) rmCCL21 (1 mg ml�1) increased the uptake of FITC-dextran by BMDCs in

contrast to treatment with LPS (1 mg ml�1). Data collected from three experiments; the error bars indicate the SD; **Po0.001, ANOVA test with Tukey post-test.

For each graphical plot, identical results were obtained using mCCL21 derived from supernatants of COS-7 cells transfected with pVR1012-mCCL21-IRES-EGFP

(data not shown). (e–g) CD40 expression by skin CD11cþ DCs. Gene gun-mediated administration of pVR1012-mCCL21 to the skin led to significant

upregulation of CD40 surface expression by skin CD11cþ DCs (g) in contrast to administration of pVR1012-mock (n¼ 3). (e) Untreated skin.

www.jidonline.org 1615

A Jalili et al.Chemokine-Induced Targeted Cell Migration

pVR1012-mCCL21-IRES-EGFP (Figure 4b), secretion of thetransgenic CCL21 from the transfected keratinocytes. Again,this phenomenon could not be observed after administrationof control pVR1012-mock (Figure 4b), thereby excluding thepossibility that endogenous mCCL21 may have been inducedby the gene gun technique itself.

We further analyzed whether in vivo transfected kerati-nocytes are able to secrete transgenic CCL21. Single-cellsuspensions of total skin were prepared 24 hours after genegun-mediated administration of pVR1012-mCCL21-IRES-EGFP or control pVR1012-IRES-EGFP and cultured for72 hours. In contrast to supernatants of keratinocytes derivedfrom control pDNA-transfected skin areas, those frompVR1012-mCCL21-IRES-EGFP-transfected sites containedmCCL21 (Figure 4c, Po0.001, ANOVA test with Tukeypost-test).

To investigate whether CCL21 transgenically expressed inthe skin is transported to the draining LNs, we administeredpVR1012-mCCL21-IRES-EGFP or pVR1012-IRES-EGFP plasmidsinto plt/plt mice, collected the PLNs draining the sites of plasmid

administration, and measured CCL21 in PLN homogenates byELISA. The PLNs draining the skin transfected with pVR1012-mCCL21-IRES-EGFP contained dramatically higher amount ofmCCL21 as compared with those draining the pVR1012-IRES-EGFP-transfected or untreated skin (Figure 4d, Po0.001,ANOVA test with Tukey post-test).

Taken together, these data argue for expression of genegun-administered CCL21 pDNA in epidermal keratinocytes,subsequent secretion into the dermis and drainage intoloco-regional LNs.

Gene gun-mediated administration of mCCL21 pDNA to theskin induces the in vivo migration of DCs, naıve/centralmemory, and effector T-cells to the draining LNsKeratinocyte-derived cytokines are secreted into the dermis,where they are collected by the lymphatics and subsequentlytransported to the draining PLNs (Palframan et al., 2001). Wetherefore hypothesized that upon pVR1012-mCCL21-IRES-EGFP administration to the skin, (i) transgenic CCL21 shouldnot only be expressed in keratinocytes and secreted into the

pVR1012-mCCL21-IRES-EGFP

pVR1012-mCCL21-IRES-EGFP

pVR1012-mCCL21-IRES-EGFP LNs

mCCL21 ELISA

**

****

CC

L21/

LN(n

g)

7

6

5

4

3

2

1

0

**

Epidermis

Epidermis

Dermis

Dermis

mCCL21 ELISA

Untreated

Untreated LNs

pVR1012-IRES-EGFP

pVR1012-IRES-EGFP LNs

Pg

ml–

1

30

20

10

0

pVR1012-mock

Figure 4. Epidermal expression of transgenic CCL21 in the skin upon gene gun-mediated administration of pVR1012-mCCL21-IRES-EGFP in vivo and its

subsequent drainage into PLNs. (a) Expression of EGFP is restricted to the epidermis, and no fluorescence is detected after application of the pVR1012-mock

vector. Skin sections were obtained 24 hours after pDNA administration. (b) Day-1 cryosections were immunostained using a polyclonal goat anti-mouse CCL21

Ab. APAAP method: hematoxylin counterstaining. Note that CCL21 immunoreactivity in the epidermis is similar to the EGFP expression pattern. In addition,

CCL21 immunoreactivity can be observed in the dermis, most probably representing CCL21 secreted from the epidermis into the dermis. No immunoreactivity

after application of pVR1012-mock vector. Bar¼ panels a and b: 100 mm. One representative of three independent experiments is shown. (c) In vitro cultured

whole single-cell suspension of skin 3 days after pVR1012-mCCL21-IRES-EGFP application. Secretion of mCCL21 as measured by mCCL21 ELISA in

supernatants. The data are from three experiments (n¼ 3 per group); the error bars indicate the SD; **Po0.001, ANOVA test with Tukey post-test. (d) Inguinal

LNs were isolated 3 days after pVR1012-mCCL21-IRES-EGFP or control pVR1012-IRES-EGFP application into the abdominal skin of plt/plt mice. LNs were

homogenized and the concentration of mCCL21 was measured. The total concentration of CCL21 per LN is expressed in nanograms, n¼ 3; the error bars

indicate the SD; **Po0.001, ANOVA test with Tukey post-test.

1616 Journal of Investigative Dermatology (2010), Volume 130

A Jalili et al.Chemokine-Induced Targeted Cell Migration

dermis, but also be drained to the loco-regional LNs and, as aresult, (ii) skin-derived transgenic CCL21 should promote thein vivo migration of CCR7þ mature CD11cþ DCs andCD4þ /CD62Lþ naıve/central memory T-lymphocytes intothe PLNs. To test whether skin-derived CCL21 can influencecell migration in the PLNs, pVR1012-mCCL21-IRES-EGFPand control pVR1012-IRES-EGFP were administered to theabdominal skin of plt/plt mice and draining, that is, inguinalPLNs were collected at different time points thereafter. At day1 after administration, PLNs draining the pVR1012-mCCL21-IRES-EGFP administration sites contained significantly highertotal cell numbers than those draining the control pVR1012-IRES-EGFP administration sites or untreated skin (P-value¼0.067 and 0.0033, respectively; Figure 5a). Thiseffect was lost within 3–7 days after pDNA administration(Figure 5a). As an additional measure of the biologicalactivity of transgenic mCCL21, we analyzed PLNs fordifferences in the cellular composition with respect toCCR7-expressing cell types, namely CD11cþ DCs andCD4þ /CD62Lþ naıve/central memory T-cells. PLNs drainingthe pVR1012-mCCL21-IRES-EGFP administration sites con-tained a significantly higher percentage of CD11cþ DCs thanthose draining the control pVR1012-IRES-EGFP administra-tion sites or untreated skin (P-value¼ 0.041 and 0.019,respectively; Figure 5b). The same phenomenon couldalso be observed for CD4þ /CD62Lþ naıve/central memoryT-cells (P-value o0.001; Figure 5c). The highest increase inthe percentage of CD11cþ DCs and CD4þ /CD62Lþ naıve/central memory T-cells was observed on day 3 afteradministration of pVR1012-mCCL21-IRES-EGFP (Figure 5b

and c, respectively). This effect was lost within 7–14 daysafter pDNA administration (data not shown). We observedalso a concomitant increase in the number of CD4þ /CD62L�

and CD8þ /CD62L� cells (Figure 6a and b).As an additional independent measure of induction of

targeted cell migration to LNs by cutaneous CCL21 pDNAadministration, we performed adoptive cell transfer experi-ments using plt/plt mice. In wild-type BALB/c mice, theCD62Lþ subpopulation of splenocytes is known to migrateinto the PLNs after adoptive transfer into the tail vein (Steinet al., 2000), a phenomenon not observed in plt/plt mice(Stein et al., 2000). We therefore administered pVR1012-mCCL21-IRES-EGFP and control pVR1012-IRES-EGFP to theabdominal skin of plt/plt mice and after 24 hours—the timepoint of peak-level expression of the transgene in the skin—the injected syngeneic wild-type BALB/c-derived 5,6-carbox-yfluorescein diacetate succinimidyl ester (CFSE)-labeledsplenocytes into the tail vein of these animals. Three hourslater, the PLNs and spleen were analyzed for the presence ofCFSE-labeled splenocytes. The PLNs draining the pVR1012-mCCL21-IRES-EGFP administration site contained a signifi-cantly higher number of CSFE-labeled cells than thosedraining the control pVR1012-IRES-EGFP administration siteor untreated skin (P-value o0.05; Figure 7a). In non-drainingSLOs such as the spleen, induction of cell migration asindicated by increased CFSE-labeled cell numbers could notbe observed (P-value 40.05; Figure 7b).

We also analyzed whether DCs are capable of migratingout of skin samples expressing transgenic CCL21. Skinsamples were isolated 1 day after in vivo gene-gun-mediated

x 10

4 cel

ls/L

N

45

30

15

0Untreated

Untreated

pVR1012-IRES-EGFP

pVR1012-IRES-EGFP

0.710.55

% o

f tot

al c

ells

% o

f tot

al c

ells

2

1

0

NS

1.35

1.81

CD11c+ DC CD4+/CD62L+ naive T cells

6

4

2

0

1.96

3.60* *

*

*

pVR1012-mCCL21-IRES-EGFP

pVR1012-mCCL21-IRES-EGFP

Untreated pVR1012-IRES-EGFP

pVR1012-mCCL21-IRES-EGFP

Day 1 Day 3 Day 7

Figure 5. Impact of skin-expressed CCL21 on the cellular composition of draining PLNs. (a) PLNs (inguinal LNs) of plt/plt mice (n¼ 4 per group) draining

the application sites of pVR1012-mCCL21-IRES-EGFP, control pVR1012-IRES-EGFP, and those from untreated mice were isolated at days 1, 3, and 7 after

pDNA administration, and the total number of cells was counted. The PLNs draining skin treated with pVR1012-mCCL21-IRES-EGFP contained significantly

higher total cell numbers as compared with those draining untreated or control-vector treated skin; *Po0.05, ANOVA test with Tukey post-test. The data

are from three experiments; the error bars indicate the SD. (b, c) Flow cytometric analysis shows preferential recruitment of CD11cþ DCs and CD4þ /CD62Lþ

naıve/central memory T-lymphocytes into PLNs draining the administration site of pVR1012-mCCL21-IRES-EGFP. The data are from three experiments

(n¼4 per group); the error bars indicate the SD; *Po0.05, Kruskall–Wallis ANOVA with Dunns’s post-test.

www.jidonline.org 1617

A Jalili et al.Chemokine-Induced Targeted Cell Migration

administration of pVR1012-mCCL21-IRES-EGFP or controlpVR1012-IRES-EGFP and cultured for 72 hours in vitro. Incontrast to supernatants of skin samples derived from controlpDNA-transfected skin areas, those from pVR1012-mCCL21-IRES-EGFP-transfected sites contained significantly highernumber of CD11cþ DCs (see Supplementary Figure S1 online).

DISCUSSIONGene gun application of pDNA expressing CCL21 is an easy,practical, and cost-effective approach for dermal administra-tion of CCL21 with a good bioavailability. Our data show thatskin-derived CCL21 (by gene gun application of pDNAexpressing CCL21) crucially influences the cellular composi-tion of the draining PLNs. This strategy leads to preferentialaccumulation of mCCL21 in—and recruitment of CCR7-expressing, CD11cþ DCs and CD4þ /CD62Lþ naıve/centralmemory T-cells, as well as CXCR3-expressing, CD4þ /CD62L� and CD8þ /CD62L� effector memory T-cells,into—the PLNs. We postulate that this is a direct functionof the expression kinetics of CCL21 after application ofCCL21–pDNA into the skin. Increased migration of DCs withupregulated surface expression of CD40 and naıve/centralmemory T-cells (principal players executing the adaptiveimmune responses) to the PLNs would lead to the delivery ofhigher amounts of antigens. This would be followed by

stronger and longer-lasting (prolonged DC survival) cellularand humoral (through B-cell activation and Ig class switch-ing) immune responses. The expression of CD62L by naıve/central memory T-cells argues for the migration of these cellsfrom the blood through HEVs rather than enhanced recruit-ment through afferent lymphatics from the skin. In accor-dance with this, we observed enrichment of CCR7-expressingsplenocytes in the PLNs draining the CCL21-expressing skinafter intravenous adoptive transfer, but no recruitment ofsplenocytes to the site of CCL21 expression in the skin (datanot shown). Consistently, CCL21 has been described to bepresented on the luminal surface of a subset of PLN HEVsafter intracutaneous application, where it promotes lympho-cyte function-associated protein-1-mediated T-cell adhesion(Stein et al., 2000). We did not observe accumulation of DCsat the site of CCL21 expression in the skin, which ispresumably because of continuous influx of DC precursorsinto the skin and emigration into draining the LNs, the twoprocesses balancing each other (data not shown).

Interestingly, we observed that the frequencies ofCD4þ /CD62L� and CD8þ /CD62L� effector memoryT-lymphocytes were also significantly increased in the PLNsdraining the VR1012-mCCL21-IRES-EGFP administration site.These cell populations may represent effector memory T-cellsthat are known to downregulate CD62L expression after

Untreated

1.2

10

8

5

3

0

1.1

6.1**

****

**

CD4+/CD62L–effector T cells CD8+/CD62L–effector T cells

1.3

0.20.3

3

2

1

0

% o

f tot

al c

ells

% o

f tot

al c

ells

pVR1012-IRES-EGFP

pVR1012-mCCL21-IRES-EGFP

Untreated pVR1012-IRES-EGFP

pVR1012-mCCL21-IRES-EGFP

a b

Figure 6. Increased frequency of CD4þ /CD62L� and CD8þ /CD62L� lymphocytes in the draining LNs of pDNA-CCL21-administered skin sites.

(a, b) Flow cytometric analysis shows increased frequency of CD4þ /CD62L� and CD8þ /CD62L� effector T-lymphocytes in the PLNs draining the

administration site of pVR1012-mCCL21-IRES-EGFP. The data are from three experiments (n¼4 per group) as in Figure 5; the error bars indicate the SD;

**Po0.001, Kruskall–Wallis ANOVA with Dunns’s post-test.

Average number of CFSE+

cells in inguinal LN

Num

ber

of C

FS

E+

cel

ls/

1-m

in a

cqui

sitio

n

Num

ber

of C

FS

E+

cel

ls/

1-m

in a

cqui

sitio

n

* *

Untreated pVR 1012-IRES-EGFPcontrol

pVR 1012-mCCL21-IRES-EGFP

Untreated pVR 1012-IRES-EGFPcontrol

pVR 1012-mCCL21-IRES-EGFP

140

105

70

35

0

Average number of CFSE+

cells in spleens3,000

2,000

1,000

0

Figure 7. Preferential recruitment of adoptively transferred splenocytes into SLOs draining pVR1012-mCCL21-IRES-EGFP-transfected skin. Splenocytes

isolated from wild-type BALB/c mice were stained with 0.2 mM CFSE. (a) plt/plt mice (n¼ 3 per group) were treated with pVR1012-mCCL21-IRES-EGFP or control

pVR1012-IRES-EGFP vectors, and 1 day thereafter CFSE-labeled splenocytes were adoptively transferred into the tail vain. Note the significantly increased

recruitment of adoptively transferred splenocytes to the SLOs draining the pVR1012-mCCL21-IRES-EGFP-administered skin sites; (b) in contrast, no

preferential recruitment of CFSEþ cells into non-draining SLOs, that is, spleen, could be seen. Every sample was acquired for 1 minute using FACScan

with a steady-state flow rate, and the number of CFSEþ cells was counted. The data are presented as average number per 1-minute acquisition. The data are

from two experiments; the error bars indicate the SD; *Po0.05, ANOVA test with Tukey post-test.

1618 Journal of Investigative Dermatology (2010), Volume 130

A Jalili et al.Chemokine-Induced Targeted Cell Migration

priming in SLOs (Roberts et al., 2005). These cells expressalso CXCR3, recently identified to induce chemotaxis towardCCL21 in murine and human microglial cells (Soto et al.,1998; Rappert et al., 2002; Dijkstra et al., 2004). It has beendocumented that these cells express CXCR3 (Mohan et al.,2005). Rappert et al. (2002) has recently shown that CCL21has the capability to bind to CXCR3 and induce chemotaxisin microglia. CD62L (L-slectin), is an adhesion moleculefound on most leukocytes, and has a primary role inmediating initial leukocyte interaction with the activatedvascular endothelium of HEVs and subsequent extravasationof these cells (Springer, 1995). Lack of CD62L expression byCD4þ /CD62L� and CD8þ /CD62L� effector memory cellspossibly excludes their LN entry through HEVs. We speculatethat these cells have presumably entered the PLNs throughafferent lymphatics under a CCL21 gradient. Migration ofeffector T-cells through this path has been documentedpreviously (Bromley et al., 2005).

These data show that, within this distinct immunologicalcompartment, the skin directly influences the cellularcomposition of the draining PLNs with regard toCD11cþ DCs and T-cells through expression and secretionof CCL21. As CCL21’s expression is dramatically inducedduring skin inflammation and as the numbers of CCR7þ cellsin the PLNs seem to be a direct function of the amountof skin-expressed CCL21, CCL21 may constitute a direct linkof skin inflammation and induction of antigen-specificimmunity.

Our data show that CCL21 induces CD40 expression butnot maturation of DCs as defined by morphology (granularityand size), phenotype (CD80/86-, CCR7- expression), andfunction (mixed-leukocyte reaction). To our knowledge this ispreviously unreported. CD40 ligation has originally beenreported to induce the expression of CD80 and CD86 co-stimulatory molecules on DCs (Caux et al., 1994; Cella et al.,1996; Koch et al., 1996; Flores-Romo et al., 1997), toenhance DC survival (Quezada et al., 2004), migration(Moodycliffe et al., 2000), and to avoid suppression of theimmune responses (Grohmann et al., 2001; Serra et al.,2003). In this regard, the expression and function of CD40seem to accompany features associated with maturation ofDCs. However, a distinctive role for CD40 is indicatedin vitro, where BMDCs show better induction of T-cellresponses by CD40 ligation than by other maturation stimuli(Labeur et al., 1999; Kelleher and Beverley, 2001). This issupported by the in vivo a-GalCer model for DC maturation,where expression of the major histocompatibility complex–-peptide complex and CD86 does not lead to the induction ofCD4þ T-helper cell-I- and CD8þ cytolytic T-lymphocyte-mediated immunity in the absence of the CD40–CD40 ligandinteraction (Fujii et al., 2004). Consistently, increasedexpression of CD80 and CD86 on DCs alone did not leadto immunity in the absence of CD40 ligation in tumornecrosis factor-a-induced skin inflammation (O’Sullivan andThomas, 2003; Fujii et al., 2004). Most recently, CD40 hasbeen described to be essential for CD4þ memory T-cellexpansion and effector cell differentiation, but dispensablefor CD4þ memory T-cell survival and proliferation (MacLeod

et al., 2006). These results may help to explain why plt/pltmice can mount CD4þ T-cell responses that may be ofslower kinetics but higher magnitude than those in wild-typemice (Mori et al., 2001; Junt et al., 2002). Besides the uniquecapability of DCs to activate naıve T-cells, these highlyversatile antigen-presenting cells have been shown to directlyinduce proliferation and IgM secretion, as well as isotypeswitch to g and a, in CD40-activated naıve B-cells (Duboiset al., 1997; Johansson et al., 2000).

Understanding the mechanisms that control the functionand migration of DCs is crucial for the development of newtreatment approaches for immunological skin diseases asdiverse as autoimmune diseases, skin tumors, skin infections,and allergic diseases. Maturation stimuli such as microbialproducts and agonistic/antagonistic Abs are prime candidatesin the development of novel adjuvants for induction ormodification of immune responses. Induction of antigenuptake and CD40 expression on DCs and spatial andfunctional segregation of these cells and naıve/centralmemory T-cells in a defined immunological compartment(such as PLNs) is a pre-requisite for induction of properimmune responses. Expression of CCL21 through administra-tion of pDNA could have important implications on thedesign of future immunotherapeutic strategies, includingthose using DC-based cellular cargo delivery.

MATERIALS AND METHODSAnimals and cell lines

plt/plt mice (Gunn et al., 1999) on BALB/c background were bred

and maintained under specific pathogen-free conditions in the

animal facility of the Medical University of Vienna. Wild-type BALB/c

and C57BL/6 mice were purchased from Harlan Winkelmann GmbH

(Borchen, Germany). Mice aged 6–10 weeks were used in all the

experiments. All animal experiments were approved by the local

ethics committee on animal care and use as well as the Austrian

Ministry of Science and Transportation.

COS-7 cells (African green monkey kidney cells) were obtained

from the American Type Culture Collection (ATCC, Rockville, MD)

and grown in DMEM medium (Invitrogen, San Diego, CA)

supplemented with 10% fetal calf serum (Invitrogen) at 371C in a

humidified atmosphere containing 5% CO2.

Construction of plasmid vectors encoding mCCL21Plasmid pVR1012-mCCL21-IRES-EGFP, allowing separate

translation of mCCL21 and EGFP from a single bicistronic mRNA

(Figure 1a), was constructed as follows: mCCL21 cDNA

(GenBank accession number: AF001980) was amplified by reverse

transcription–PCR with primers containing BglII restriction sites

(VBC-Genomics, Vienna, Austria; 50 primer: GAAGATCTATGGCT

CAGATGATGACTCT and 30 primer: GAAGATCTCTATCCTCTTG

AGGGCTGT). After restriction digestion with BglII (Roche Diag-

nostics GmbH, Mannheim, Germany), the purified PCR product was

inserted into the BglII-cut plasmid pVR1012 (provided by B Zaugg,

Vical, San Diego, CA) (Hartikka et al., 1996) using the Rapid DNA

Ligation kit (Roche Diagnostics) thereby generating plasmid

pVR1012-mCCL21. The IRES-EGFP cDNA was amplified by PCR

with primers containing XbaI restriction sites (VBC-Genomics; 50

primer: GCTCTAGAGCCACCAAAATCAACGGGA and 30 primer:

www.jidonline.org 1619

A Jalili et al.Chemokine-Induced Targeted Cell Migration

GCTCTAGAGCTCAGGTTCAGGGGGAG) using pIRES2-EGFP (In-

vitrogen) as template. After restriction digestion with XbaI (Roche

Diagnostics), the purified PCR product was inserted into the XbaI-cut

plasmid pVR1012-mCCL21 using the Rapid DNA Ligation kit (Roche

Diagnostics), downstream from the mCCL21 transgene, thereby

generating plasmid pVR1012-mCCL21-IRES-EGFP (Figure 1a).

Correct cDNA sequences were confirmed by DNA sequencing

(VBC-Genomics).

Plasmid pVR1012-mCCL21/EGFP, allowing expression of chi-

meric mCCL21/EGFP (Figure 1b), was constructed as follows:

mCCL21 cDNA (see above) was amplified by reverse transcrip-

tion–PCR with primers containing BglII restriction sites. The 30 primer

contained a mismatch base to replace the stop codon TAG by AAG

(VBC-Genomics; 50 primer: GAAGATCTATGGCTCAGATGAT

GACTCT and 30 primer: GAAGATCTCTTTCCTCTTGAGGGCTGT).

After restriction digestion with BglII (Roche Diagnostics), the purified

PCR product was inserted into the BglII-cut plasmid pVR1012 (Vical)

using the Rapid DNA Ligation kit (Roche Diagnostics). The EGFP

cDNA was amplified by PCR using primers containing XbaI

restriction sites (VBC-Genomics; 50 primer: GCTCTAGAGCC-

CATGGTGAGCAAGGG and 30 primer: GCTCTAGAGCTCAGGTT

CAGGGGGAG) using the plasmid pIRES2-EGFP (Invitrogen) as

template. After restriction digestion with XbaI (Roche Diagnostics),

the purified PCR product was inserted into the XbaI-cut plasmid

pVR1012-mCCL21 (with modified stop codon of mCCL21) directly

downstream from the mCCL21 transgene thereby generating plasmid

pVR1012-mCCL21/EGFP (Figure 1b). Correct cDNA sequences were

confirmed by DNA sequencing (VBC-Genomics).

Plasmids were grown in XL1-Blue competent cells (Stratagene, La

Jolla, CA), prepared, and purified using the Qiagen Endotoxin-Free

Plasmid DNA Isolation kit (Qiagen, Hilden, Germany). The quality

and quantity of purified pDNA was assessed by optical densitometry

at 260 and 280 nm and agarose gel electrophoresis. Expression of

mCCL21, EGFP, and chimeric mCCL21/EGFP was confirmed by

immunocytochemistry, flow cytometry, and immunofluorescence

microscopy after transfection into COS-7 cells (as described under

the following section). Parental plasmids pVR1012 and pVR1012-

IRES-EGFP served as controls.

Transfection, immunocytochemistry, and ELISA

Transient transfection of COS-7 cells with pDNA (pVR1012-

mCCL21-IRES-EGFP, pVR1012-mCCL21/EGFP, pVR1012, and

pVR1012-IRES-EGFP) was performed using the Metafectene trans-

fection reagent (Biontex, Martinsried/Planegg, Germany) according

to the manufacturer’s instructions. A 5-mg weight of plasmid DNA

per 5� 105 cells was used. The cells and supernatants were

harvested 48 hours after transfection.

For immunocytochemistry, 1� 105 cells were transferred onto

glass slides (cytospin preparations), fixed in 4% paraformaldehyde

(Merck Chemicals, Darmstadt, Germany) in phosphate-buffered

saline (PBS) (Invitrogen), washed in PBS/0.5% Triton X-100 (Merck

Chemicals), and unspecific binding sites were blocked with PBS/

10% goat serum (Invitrogen)/0.1% natrium azide (Merck Chemicals,

later referred to as blocking solution) for 1 hour. Subsequently, slides

were incubated with a goat anti-mouse CCL21 polyclonal Ab

(2mg ml�1; R&D Systems, Minneapolis, MN) in blocking solution for

1 hour at room temperature (RT) and alkaline phosphatase-con-

jugated rabbit anti-goat antiserum (Molecular Probes, San Diego,

CA) for 30 minutes at RT. Reactions were developed using Fuchsin

Substrate-Chromogen (Dako, Glostrup, Denmark) according to the

manufacturer’s protocol. Expression of EGFP was analyzed in non-

fixed, non-permeabilized living cells by confocal laser-scanning

microscopy (Axiovert 200M; Carl Zeiss MicroImaging GmbH, Jena,

Germany).

For western blotting, transfected COS-7 cells were washed with

cold PBS (Invitrogen), lysed by treatment with radioimmunopreci-

pitation assay buffer (50 mM Tris base (Merck Chemicals), 150 mM

NaCl (Merck Chemicals), 1% NP-40 (ThermoFisher Scientific,

Waltham, USA), 0.25% sodium deoxycholate (Merck Chemicals),

1 mM EDTA (Sigma-Aldrich, Vienna, Austria)) in the presence of a

protease inhibitor cocktail (Complete, Mini Protease Inhibitor

Cocktail Tablets; Roche Diagnostics). Protein concentration was

measured using a bicinchoninic acid protein assay (Pierce, Rockford,

IL). Equal amounts of proteins were separated on a 12% SDS-

polyacrylamide gel, transferred onto polyvinylidene difluoride

membranes (Sigma-Aldrich), and blocked with PBST (PBS/0.05%

Tween-20; Sigma-Aldrich) containing 5% non-fat dry milk (Bio-Rad,

Hercules, CA; blocking buffer). The membranes were incubated with

0.2 mg ml�1 of goat anti-mouse CCL21 polyclonal Ab (R&D Systems)

in blocking buffer overnight at 41C. After extensive washing with

PBST, the membranes were incubated for 45 minutes at RT with a

horseradish peroxidase-conjugated rabbit anti-goat Ab (Bio-Rad)

diluted 1:40,000 in blocking buffer. The membranes were subse-

quently washed in PBST and proteins were visualized using ECL

Western Blotting Detection Reagents (GE Healthcare Europe,

Munich, Germany).

Murine CCL21 concentrations in the cell culture supernatants of

COS-7 cells transfected with pDNA (pVR1012-mCCL21-IRES-EGFP

and pVR1012-IRES-EGFP) were analyzed by mouse CCL21/6Ckine

DuoSet ELISA (R&D Systems) according to the manufacturer’s

instructions.

Gene gun-mediated administration and in vivo expression ofpDNA

Gene gun bullets carrying pDNA were prepared as described by

Tuting et al. (1999). Briefly, pDNA was precipitated onto 1-mm gold

particles (Bio-Rad) at a density of 2 mg of DNA per microgram of

particles. The gold particles and DNA were resuspended in 100ml of

0.05 M spermidine (Sigma-Aldrich) and DNA was precipitated by

addition of 100ml of 1 M CaCl2 (Merck Chemicals). The particles

were washed with ethanol (VWR International, Vienna,

Austria) to remove H2O, resuspended in ethanol (VWR International)

containing 0.1 mg ml�1 polyvinylpyrrolidone (Sigma-Aldrich), and

coated onto the inner surface of Tefzel tubes using a tube loader

(Bio-Rad).

Six- to eight-week-old plt/plt mice were anesthetized by

intraperitoneal injections of Ketalar/Rompun (Hollabrunn mixture;

Bayer AG, Leverkusen, Germany) and three non-overlapping shots of

1mg pVR1012-mCCL21-IRES-EGFP, and control plasmid, respec-

tively, were applied to the shaved skin of the abdomen using an

Accell helium pulse gun (Bio-Rad) at a pressure of 300 pounds per

square inch (psi). After 1, 3, 7, and 14 days of plasmid administration

mice were killed and the abdominal area skin was taken and snap-

frozen in liquid nitrogen. Cryosections measuring 10mm were

prepared and subjected to immunohistochemical staining or direct

immunofluorescence microscopic analysis.

1620 Journal of Investigative Dermatology (2010), Volume 130

A Jalili et al.Chemokine-Induced Targeted Cell Migration

For immunohistochemistry, the cryosections were transferred

onto glass slides, fixed in 4% paraformaldehyde in PBS

(Invitrogen), washed in PBS/0.5% Triton X-100 (Merck Chemicals),

and unspecific binding sites were blocked using a blocking

solution for 1 hour at RT. Subsequently, the slides were incubated

with a goat anti-mouse CCL21 polyclonal Ab (2mg ml�1; R&D

Systems) in blocking solution overnight at 41C and alkaline

phosphatase-conjugated rabbit anti-goat antiserum (Molecular

Probes) for 30 minutes at RT. Reactions were developed using

Fuchsin Substrate-Chromogen (Dako) according to the

manufacturer’s protocol and the sections were counterstained with

hematoxylin. Endogenous alkaline phosphatase was blocked by

incubation with 0.24 mg ml�1 of levamisole (Sigma-Aldrich). In

negative controls, the primary Ab was substituted by isotype-

matched IgG.

For immunofluorescence microscopic analysis of EGFP expres-

sion, the cryosections were directly mounted onto glass slides in

Prolong Gold anti-fade reagent (Molecular Probes) and subsequently

evaluated by confocal laser-scanning microscopy.

In vivo expression of transgenic mCCL21 was also confirmed by

ELISA in single-cell suspensions derived from pDNA-treated skin.

The abdominal area skin was collected 24 hours after administration

of pVR1012-mCCL21-IRES-EGFP and pVR1012-IRES-EGFP, minced

in PBS without Caþ þ /Mgþ þ (Invitrogen), digested with 2mg ml�1 of

collagenase/dispase (Roche Diagnostics) in PBS without Caþ þ /

Mgþ þ (Invitrogen) for 60 minutes at 371C and with light agitation,

passed through a 70-mm nylon mesh, and centrifuged at 1,500 r.p.m.

for 10 minutes. The cells were resuspended in a culture medium and

cultured for 72 hours in 12-well plates (eBioscience, San Diego,

USA). Thereafter, supernatants were collected and concentrations of

mCCL21 were analyzed by ELISA (R&D Systems) according to the

manufacturer’s instructions.

CCL21’s concentration in the draining inguinal LNs of plt/plt

mice after pVR1012-mCCL21-IRES-EGFP or control pVR1012-IRES-

EGFP application into the abdominal skin was measured using a

mouse CCL21/6Ckine DuoSet ELISA (R&D Systems) according to the

manufacturer’s instructions.

For flow cytometric analysis of CD40 expression by skin

CD11cþ DCs, the abdominal skin was collected 24 hours after

administration of pVR1012-mCCL21 and pVR1012, and a single-cell

suspension was prepared as described earlier. The cells were

resuspended in 2% PBS/BSA supplemented with 10% mouse serum

and 5% goat serum (both from Invitrogen) and incubated for

15 minutes at 41C to block FcgR. Surface expression of CD40

by DCs was determined by flow cytometric analysis using the

following mAbs: FITC-CD11c (HL3) and PE-CD40 (3/23) (both

Ab and the corresponding isotype controls (Armenian hamster

FITC-IgG1 and PE-rat IgG2a) were from BD Biosciences, Vienna,

Austria). Autofluorescent cells were excluded using the FL3

channel.

Cellular composition analysis of PLNs draining pDNAadministration sites

After 3, 7, and 14 days of cutaneous gene gun-mediated adminis-

tration of pVR1012-mCCL21-IRES-EGFP draining, that is, inguinal

PLNs were collected. The tissues were minced, digested with

2mg ml�1 of collagenase/dispase (Roche Diagnostics) in PBS without

Caþ þ /Mgþ þ (Invitrogen) for 60 minutes at 371C and under light

agitation, passed through a 70-mm nylon mesh, and centrifuged at

1,500 r.p.m. for 10 minutes. The cells were resuspended in 2% PBS/

BSA supplemented with 10% mouse serum and 5% goat serum (both

from Invitrogen), and incubated for 15 minutes at 41C to block FcgR.

A predetermined number of calibration beads (CaliBRITE Beads; BD

Biosciences) were then added to each sample to allow normalization

of cell counts in different samples and cells were counted using a

FACScan (BD Biosciences) as described by Jahrsdorfer et al. (2005).

The cellular composition was determined by flow cytometric

analysis using the following mAbs: PE-Cy5-CD4 (L3T4), PE-Cy5-

CD8 (53-6.7), FITC-CD11c (HL3), PE-CD62L (MEL-14) (all Abs and

the corresponding isotype controls (PE-rat IgG2a and Armenian

hamster FITC-IgG1) were from BD Biosciences). Autofluorescent

cells were excluded using the FL3 channel.

Adoptive transfer studies

Red blood cell-depleted splenocytes derived from wild-type BALB/c

mice were labeled with CFSE (at a final concentration of 0.2 mM;

Molecular Probes), washed extensively in PBS/5% fetal calf serum,

and resuspended in RPMI/25 mM HEPES (Invitrogen). In general,

labeling efficacy was more than 99%; significant toxicity could be

excluded by staining with propidium iodide (Sigma-Aldrich) and

FACScan analysis (BD Biosciences).

pVR1012-mCCL21-IRES-EGFP and its control vector pVR1012-

IRES-EGFP were administered as described above and 24 hours later

15� 106 CFSE-labeled splenocytes were injected into the tail vein of

syngeneic plt/plt mice. The spleens and PLNs draining the plasmid

administration site were obtained 3 hours later. Single-cell suspen-

sions were prepared as described above and analyzed for the

presence of CFSEþ cells by FACScan analysis (BD Biosciences).

Total cells were acquired for 1 minute with steady-state flow rate

followed by counting of CFSEþ cells within the acquired cell

populations. Dead cells were excluded by staining with propidium

iodide (Sigma-Aldrich) and autofluorescent cells by using the FL3

channel.

Statistical analysis

Statistical analysis was performed by unpaired t-test, one-way

ANOVA tests with Tukey post-tests, and Kruskall–Wallis ANOVA

with Dunns’s post-test using the GraphPad-Instat software

(www.graphpad.com).

CONFLICT OF INTERESTThe authors state no conflict of interest.

ACKNOWLEDGMENTSWe are grateful to Robert Zaugg (Vical Incorp., San Diego, CA) for providingthe VR1012 pDNA vector and Dr Kirsten Merz for critically reading the paper.This work was supported by a grant (20040) from the Center for MolecularMedicine (CeMM) of the Austrian Academy of Sciences (Vienna, Austria).

SUPPLEMENTARY MATERIAL

Supplementary material is linked to the online version of the paper at http://www.nature.com/jid

REFERENCES

Banchereau J, Fay J, Pascual V et al. (2003) Dendritic cells: controllers of theimmune system and a new promise for immunotherapy. Novartis FoundSymp 252:226–35; discussion 235–8, 257–67

www.jidonline.org 1621

A Jalili et al.Chemokine-Induced Targeted Cell Migration

Bromley SK, Thomas SY, Luster AD (2005) Chemokine receptor CCR7 guidesT cell exit from peripheral tissues and entry into afferent lymphatics. NatImmunol 6:895–901

Carlsen HS, Haraldsen G, Brandtzaeg P et al. (2005) Disparate lymphoidchemokine expression in mice and men: no evidence of CCL21 synthesisby human high endothelial venules. Blood 106:444–6

Caux C, Massacrier C, Vanbervliet B et al. (1994) Activation ofhuman dendritic cells through CD40 cross-linking. J Exp Med 180:1263–72

Cella M, Scheidegger D, Palmer-Lehmann K et al. (1996) Ligation of CD40 ondendritic cells triggers production of high levels of interleukin-12 andenhances T cell stimulatory capacity: T–T help via APC activation. J ExpMed 184:747–52

Chensue SW (2001) Molecular machinations: chemokine signals inhost–pathogen interactions. Clin Microbiol Rev 14:821–35

Christopherson KW II, Campbell JJ, Hromas RA (2001) Transgenic over-expression of the CC chemokine CCL21 disrupts T-cell migration. Blood98:3562–8

Colditz IG, Schneider MA, Pruenster M et al. (2007) Chemokines at large:in-vivo mechanisms of their transport, presentation and clearance.Thromb Haemost 97:688–93

Dijkstra IM, Hulshof S, van der Valk P et al. (2004) Cutting edge: activity ofhuman adult microglia in response to CC chemokine ligand 21.J Immunol 172:2744–7

Dubois B, Vanbervliet B, Fayette J et al. (1997) Dendritic cells enhancegrowth and differentiation of CD40-activated B lymphocytes. J Exp Med185:941–51

Ebert LM, Schaerli P, Moser B (2005) Chemokine-mediated control of T celltraffic in lymphoid and peripheral tissues. Mol Immunol 42:799–809

Flores-Romo L, Bjorck P, Duvert V et al. (1997) CD40 ligation on humancord blood CD34+ hematopoietic progenitors induces their prolife-ration and differentiation into functional dendritic cells. J Exp Med 185:341–9

Forster R, Schubel A, Breitfeld D et al. (1999) CCR7 coordinates the primaryimmune response by establishing functional microenvironments insecondary lymphoid organs. Cell 99:23–33

Fujii S, Liu K, Smith C et al. (2004) The linkage of innate to adaptive immunityvia maturing dendritic cells in vivo requires CD40 ligation in addition toantigen presentation and CD80/86 costimulation. J Exp Med199:1607–18

Granelli-Piperno A, Golebiowska A, Trumpfheller C et al. (2004) HIV-1-infected monocyte-derived dendritic cells do not undergo maturation butcan elicit IL-10 production and T cell regulation. Proc Natl Acad Sci USA101:7669–74

Grohmann U, Fallarino F, Silla S et al. (2001) CD40 ligation ablates thetolerogenic potential of lymphoid dendritic cells. J Immunol 166:277–83

Gunn MD (2003) Chemokine mediated control of dendritic cell migration andfunction. Semin Immunol 15:271–6

Gunn MD, Kyuwa S, Tam C et al. (1999) Mice lacking expression ofsecondary lymphoid organ chemokine have defects in lymphocytehoming and dendritic cell localization. J Exp Med 189:451–60

Hartikka J, Sawdey M, Cornefert-Jensen F et al. (1996) An improved plasmidDNA expression vector for direct injection into skeletal muscle. HumGene Ther 7:1205–17

Hemmi H, Akira S (2005) TLR signalling and the function of dendritic cells.Chem Immunol Allergy 86:120–35

Jahrsdorfer B, Wooldridge JE, Blackwell SE et al. (2005) Immunostimulatoryoligodeoxynucleotides induce apoptosis of B cell chronic lymphocyticleukemia cells. J Leukoc Biol 77:378–87

Jalili A, Makowski M, Switaj T et al. (2004) Effective photoimmunotherapy ofmurine colon carcinoma induced by the combination of photodynamictherapy and dendritic cells. Clin Cancer Res 10:4498–508

Johansson B, Ingvarsson S, Bjorck P et al. (2000) Human interdigita-ting dendritic cells induce isotype switching and IL-13-dependentIgM production in CD40-activated naive B cells. J Immunol 164:1847–54

Junt T, Nakano H, Dumrese T et al. (2002) Antiviral immune responses in theabsence of organized lymphoid T cell zones in plt/plt mice. J Immunol168:6032–40

Kelleher M, Beverley PC (2001) Lipopolysaccharide modulation of dendriticcells is insufficient to mature dendritic cells to generate CTLs from naivepolyclonal CD8+ T cells in vitro, whereas CD40 ligation is essential.J Immunol 167:6247–55

Koch F, Stanzl U, Jennewein P et al. (1996) High level IL-12 production bymurine dendritic cells: upregulation via MHC class II and CD40molecules and downregulation by IL-4 and IL-10. J Exp Med 184:741–6

Labeur MS, Roters B, Pers B et al. (1999) Generation of tumor immunity bybone marrow-derived dendritic cells correlates with dendritic cellmaturation stage. J Immunol 162:168–75

Laing KJ, Secombes CJ (2004) Chemokines. Dev Comp Immunol 28:443–60

Luther SA, Tang HL, Hyman PL et al. (2000) Coexpression of the chemokinesELC and SLC by T zone stromal cells and deletion of the ELC gene in theplt/plt mouse. Proc Natl Acad Sci USA 97:12694–9

MacLeod M, Kwakkenbos MJ, Crawford A et al. (2006) CD4 memory T cellssurvive and proliferate but fail to differentiate in the absence of CD40.J Exp Med 203:897–906

Mohan K, Cordeiro E, Vaci M et al. (2005) CXCR3 is required for migration todermal inflammation by normal and in vivo activated T cells: differentialrequirements by CD4 and CD8 memory subsets. Eur J Immunol35:1702–11

Moodycliffe AM, Shreedhar V, Ullrich SE et al. (2000) CD40–CD40ligand interactions in vivo regulate migration of antigen-bearingdendritic cells from the skin to draining lymph nodes. J Exp Med 191:2011–20

Mori S, Nakano H, Aritomi K et al. (2001) Mice lacking expression of thechemokines CCL21-ser and CCL19 (plt mice) demonstrate delayed butenhanced T cell immune responses. J Exp Med 193:207–18

Nickel W (2003) The mystery of nonclassical protein secretion. A currentview on cargo proteins and potential export routes. Eur J Biochem270:2109–19

O’Sullivan B, Thomas R (2003) Recent advances on the role ofCD40 and dendritic cells in immunity and tolerance. Curr Opin Hematol10:272–8

Ogura M, Sato S, Nakanishi K et al. (2004) In vivo targeted gene transferin skin by the use of laser-induced stress waves. Lasers Surg Med 34:242–8

Palframan RT, Jung S, Cheng G et al. (2001) Inflammatory chemokinetransport and presentation in HEV: a remote control mechanism formonocyte recruitment to lymph nodes in inflamed tissues. J Exp Med194:1361–73

Quezada SA, Jarvinen LZ, Lind EF et al. (2004) CD40/CD154 interactions atthe interface of tolerance and immunity. Annu Rev Immunol 22:307–28

Rappert A, Biber K, Nolte C et al. (2002) Secondary lymphoid tissuechemokine (CCL21) activates CXCR3 to trigger a Cl� current andchemotaxis in murine microglia. J Immunol 168:3221–6

Roberts AD, Ely KH, Woodland DL (2005) Differential contributions of centraland effector memory T cells to recall responses. J Exp Med 202:123–33

Rossi D, Zlotnik A (2000) The biology of chemokines and their receptors.Annu Rev Immunol 18:217–42

Serra P, Amrani A, Yamanouchi J et al. (2003) CD40 ligation releasesimmature dendritic cells from the control of regulatory CD4+CD25+ Tcells. Immunity 19:877–89

Shigaki T, Vyzasatya RR, Sivitz AB et al. (2005) The Cre-loxP recombination-based reporter system for plant transcriptional expression studies. PlantMol Biol 58:65–73

Soto H, Wang W, Strieter RM et al. (1998) The CC chemokine 6Ckine binds theCXC chemokine receptor CXCR3. Proc Natl Acad Sci USA 95:8205–10

Springer TA (1995) Traffic signals on endothelium for lymphocyte recircula-tion and leukocyte emigration. Annu Rev Physiol 57:827–72

Stein JV, Rot A, Luo Y et al. (2000) The CC chemokine thymus-derivedchemotactic agent 4 (TCA-4, secondary lymphoid tissue chemokine,

1622 Journal of Investigative Dermatology (2010), Volume 130

A Jalili et al.Chemokine-Induced Targeted Cell Migration

6Ckine, exodus-2) triggers lymphocyte function-associated antigen 1-mediated arrest of rolling T lymphocytes in peripheral lymph node highendothelial venules. J Exp Med 191:61–76

Steinman RM (2003) Some interfaces of dendritic cell biology. Apmis111:675–97

Steinman RM, Nussenzweig MC (2002) Avoiding horror autotoxicus: theimportance of dendritic cells in peripheral T cell tolerance. Proc NatlAcad Sci USA 99:351–8

Till KJ, Lin K, Zuzel M et al. (2002) The chemokine receptor CCR7 and alpha4integrin are important for migration of chronic lymphocytic leukemiacells into lymph nodes. Blood 99:2977–84

Tuting T, Gambotto A, DeLeo A et al. (1999) Induction of tumor antigen-specific immunity using plasmid DNA immunization in mice. CancerGene Ther 6:73–80

Wong CK, Tsang CM, Ip WK et al. (2006) Molecular mechanisms for therelease of chemokines from human leukemic mast cell line (HMC)-1cells activated by SCF and TNF-alpha: roles of ERK, p38 MAPK, andNF-kappaB. Allergy 61:289–97

Yoneyama H, Matsuno K, Matsushimaa K (2005) Migration of dendritic cells.Int J Hematol 81:204–7

Zlotnik A, Yoshie O (2000) Chemokines: a new classification system and theirrole in immunity. Immunity 12:121–7

www.jidonline.org 1623

A Jalili et al.Chemokine-Induced Targeted Cell Migration