Chemokine programming dendritic cell antigen response: part I – select chemokine programming of antigen uptake even after maturation

Authors


Correspondence: James D. Bryers, Department of Bioengineering, University of Washington, Seattle, WA 98195, USA

Email: jbryers@u.washington.edu

Senior author: James D. Bryers

Summary

Here, we report on the successful programming of dendritic cells (DCs) using selectively applied mixtures of chemokines as a novel protocol for engineering vaccine efficiency. Antigen internalization by DCs is a pivotal step in antigen uptake/presentation for bridging innate and adaptive immunity and in exogenous gene delivery used in vaccine strategies. Contrary to most approaches to improve vaccine efficiency, active enhancement of antigen internalization by DCs as a vaccine strategy has been less studied because DCs naturally down-regulate antigen internalization upon maturation. Whereas chemokines are mainly known as signal proteins that induce leucocyte chemotaxis, very little research has been carried out to identify any additional effects of chemokines on DCs following maturation. Here, immature DCs are pre-treated with select chemokines before intentional maturation using lipopolysaccharide (LPS). When pre-treated with a mixture of CCL3 and CCL19 in a 7 : 3 ratio, then matured with LPS, chemokine pre-treated DCs exhibited 36% higher antigen uptake capacity than immature DCs and 27% higher antigen-processing capacity than immature DCs treated only with LPS. Further, CCL3 : CCL19 (7 : 3) pre-treatment of DCs modulated MHC molecule expression and secretion of various cytokines of DCs. Collectively, DC programming was feasible using a specific chemokine combination and these results provide a novel strategy for enhancing DC-based vaccine efficiency. In Part II, we report on the phenotype changes and antigen presentation capacity of chemokine pre-treated murine bone marrow-derived DCs examined in long-term co-culture with antigen-specific CD4+ T cells.

Introduction

Dendritic cells (DCs) bridge innate and adaptive immunity in the host immune response. As professional antigen-presenting cells (APCs), immature DCs (iDCs) undergo maturation upon encountering pathogens or endogenous stimuli.[1] Mature DCs (mDCs) then migrate via the afferent lymphatics to draining lymph nodes to present the previously internalized and processed antigens, in the context of MHC Class molecules, to T and B cells that are subsequently activated in adaptive immunity.[2] Due to these potent features, DCs have recently been employed in emerging immunotherapy vaccines.[3, 4] For instance, combined with appropriate adjuvants that induce DC maturation, specific antigens derived from certain cancer tumors or infected cells can be loaded ex vivo into DCs, then these mDCs can be returned to hosts to stimulate T cells in vivo, thereby inducing adaptive immunity through T-cell activation.[5-7] There are intense research efforts into delivering genes (mRNA or DNA) into DCs that encode for specific antigens.[8-10] Unfortunately, enhancement of the intrinsic endocytic (antigen internalization) process by DCs has not received as much attention as these other strategies. One reason for investigating enhanced endocytosis by DCs is that endocytosis is the critical step in the delivery of a myriad of emerging therapeutic agents (antigens or genes) delivered by in vitro, ex vivo or in vivo methods.[11-14] For example, polymer scaffolds that continuously stimulated DCs by releasing both granulocyte–macrophage colony-stimulating factor (GM-CSF; known to enhance phagocytosis in macrophages and DCs) and cationic polymer condensed DNA led to a 20-fold increase in gene expression, and high levels of expression persisted for a period of 10 days, in vitro.[15]

As defined by Mukherjee et al.,[16] the term endocytosis in this study includes phagocytosis, pinocytosis and receptor-mediated endocytosis. Platt et al.[17] recently reported that mDCs still use endocytic receptors to capture and present antigens while they down-regulate pinocytosis. In this way, iDCs actively internalize antigens, then once matured, DCs naturally down-regulate their overall endocytosis, which encompasses phagocytosis, pinocytosis and receptor-mediated endocytosis. Then they migrate to lymph nodes, where the mDCs effectively process and present antigens to lymphocytes. Various efforts have been made to induce effective antigen loading or gene delivery to DCs; such as: by mannose-decorated pDNA polyplexes[18]; direct antigen fusion with single chain Fv antibody against DC phagocytic receptor, DEC-205[19]; and DEC-205 monoclonal antibody targeted nanoparticles.[20] Most efforts to date are limited by the natural DC maturation process, which down-regulates subsequent internalization of antigens to a certain level,[17, 21] thus significantly reducing levels of further uptake and processing of antigens. Most vaccines are less than ideal because accompanying adjuvants can actually activate iDCs before antigen uptake; thus reducing overall antigen uptake and vaccine efficacy.[12] Very few, if any, studies have been carried out that attempt to manipulate the natural process by which mDCs internalize antigens.

‘Chemokines’ are low-molecular-weight cytokines and their primary biological activity is to promote chemotaxis of leukocytes.[22] Among the many chemokines identified and elucidated for their biological functions, C-C motif ligand 3 (CCL3) and CCL19 are generally considered the most important in DC trafficking because of their selective regulation of iDCs and mDCs, respectively.[23] Immature DCs in the peripheral tissue express C-C chemokine receptor 1 (CCR1) and CCR5 that recognize the ligand, CCL3. When the host response is activated by injury or pathogens, CCL3 is secreted from inflammatory cells, so inducing chemotaxis of iDCs. Once iDCs internalize antigens and mature, they down-regulate both CCR1/CCR5 receptor expression and antigen uptake while up-regulating CCR7 receptor expression. CCR7 receptor recognizes the chemokine CCL19, which promotes DC trafficking from the peripheral tissue to secondary lymph organs.[24]

Most studies of chemokine influence on the host immune response have focused on DC and/or T-cell migration to a specific site and the subsequent T-cell activation and proliferation.[25-27] Other than migration and chemotactic effects, it is increasingly clear that chemokines are also involved in angiogenesis,[28] haematopoiesis,[29] or regulation of DC maturation and T-cell activation.[30] Marsland et al.,[31] reported that DCs pre-treated with the chemokine CCL19 induced T helper type 1 (Th1) rather than Th2 polarization. Further, they found CCL19 and CCL21 to act as potent natural adjuvants for terminal activation of DCs, which suggests that chemokines not only orchestrate DC migration but also regulate their immunogenic potential for the induction of T-cell responses.

Whereas the process of naturally programmed DCs directing the polarization of T cells has been extensively investigated,[32, 33] the effects of artificially programmed DCs have received only minimal attention. However, a few studies have reported that artificially programmed DCs exhibited remarkable changes in phenotype. Immature DCs pre-treated with dexamethasone and subsequently stimulated with tumor necrosis factor-α (TNF-α) exhibited an endocytic capacity four times higher (at maximum dexamethasone concentration) than iDCs treated with only TNF-α.[34] Clingan et al.,[35] reported that pre-treatment of iDCs with either interleukin-4 (IL-4) or interferon-γ (IFN-γ) inhibited the migration of iDCs in response to CCL3. Coincidentally, they observed that when IL-4 or IFN-γ pre-treated DCs were incubated with FITC-dextran in the presence of CCL3 for 2 min, dextran uptake capacity of the DCs was significantly enhanced by approximately fourfold (IFN-γ) or fivefold (IL-4) versus without CCL3. Yanagawa and Onoe,[36] found that CCL3 and CCL19 rapidly (in less than an hour) and selectively enhanced the internalization ability of iDCs and mDCs, respectively, when dextran and chemokines were added simultaneously to the cell culture. They also noted that CCL19 induced an actin-reorganization related to the endocytic behaviour of mDCs.[37] Moreover, the synergistic effects of combinations of cytokines have been shown on the expansion of blood progenitors,[38] on the endocytic pathway in insulin-producing cells,[39] and on the migration and development of other phenotypes in endothelial cells.[40]

Hence it may be possible, using selected chemokines or their combinations, to artificially program iDCs, thereby controlling their phenotypes and maturation status in order to enhance antigen uptake and presentation. We report here the first study to engineer DC phenotypes with select chemokine application to enhance antigen uptake and processing capacity of DCs, which can directly affect antigen presentation and DC-based vaccine efficiency in future. Dendritic cells were pre-treated with the individual chemokines CCL3, CCL19, or their combination in various ratios. Then, 24 hr later, DCs were exposed to lipopolysaccharide (LPS), [a Toll-like receptor 4 (TLR4) ligand], to induce maturation. We demonstrate that when DCs are pre-treated with a chemokine combination of CCL3 : CCL19 in a specific ratio then subsequently stimulated with LPS, certain phenotypic changes arise that are significantly different from those of iDCs or DCs stimulated only with LPS. Dendritic cells programmed with a specific chemokine combination (CCL3 : CCL19 = 7 : 3) retained antigen uptake capacity and exhibited antigen-processing capacity, even after subsequent LPS maturation stimulus, at levels higher than iDCs (36%), and iDCs treated only with LPS (27%), respectively. Along with antigen uptake, this chemokine programming of DCs also modulated expression of MHC molecules and various cytokine responses of DCs even after maturation of DCs. Results here suggest chemokine programming may be a new tool for enhancing ex vivo and in vivo immunotherapy vaccine strategies.

Materials and methods

Dendritic cell culture

The JAWSII DC line was purchased from the American Type Culture Collection (ATCC, Manassas, VA) and cultured in α minimum essential medium (α-MEM) (Invitrogen, Carlsbad, CA) supplemented with sodium bicarbonate (1·5 g/l) (Sigma, St Louis, MO), L-glutamine (4 mm) (Invitrogen), penicillin/streptomycin (1%, volume/volume) (Invitrogen), GM-CSF (5 ng/ml) (Peprotech, Rocky Hill, NJ) and 20% heat-inactivated fetal bovine serum (FBS) (Invitrogen) at 37° in a 5% CO2 atmosphere. Then sequential treatments of these prepared JAWS II iDCs and examination of them were performed as described in the Results section.

Antigen uptake assay

The effects on DCs of chemokine pre-treatment followed by LPS stimulus (to initiate maturation) were assessed by measuring levels of endocytic ability. To quantify endocytic ability, DCs collected on Day 1 (24 hr after no treatment or the described chemokine treatment) and on Day 2 (24 hr after subsequent LPS treatment) were resuspended in medium (without phenol red) at 1 × 106 cells/ml. Then, each sample received 3·33 μg/ml fluorescent Alexa Fluor 488-ovalbumin (OVA) (a model antigen) (Invitrogen) for 30 min at 37°. After incubation, any excess fluorochrome bound to the cell surface was quenched for 3–4 min on ice using a 0·5% Trypan Blue/2% FBS/1× PBS solution. After two repetitive quenching steps, cells were thoroughly washed using ice-cold FACS buffer (2% FBS/1× PBS) and then immediately examined using a FACS Canto (BD Biosciences, San Jose, CA). Negative control DCs were separately prepared by incubation of DCs with the model antigen on ice. The mean fluorescence intensity (MFI) of the ice control cells was subtracted from that of cells incubated at 37° with OVA per treatment or control. Data were analysed using the FlowJo Software (Tree Star Inc., Ashland, OR).

Antigen degradation (processing) assay

The model antigen (OVA) degradation (processing) by DCs was also examined using flow cytometry. Here, DCs were treated with BODIPY-conjugated DQ-OVA (Molecular Probes/Invitrogen), a self-quenched conjugate of OVA that exhibits bright green fluorescence only upon proteolytic cleavage releasing the dye molecule from the OVA. To quantify antigen degradation kinetics, this assay was carried out at 30 min, 1 hr and 2 hr after OVA incubation. DQ-OVA was applied at the concentration identical to the OVA of the antigen uptake assay above. Briefly, after DCs were collected on Day 1 and Day 2, DCs from control or sample wells were divided into three groups and resuspended in medium (without phenol red) at 1 × 106 cells/ml per group. Then, each group was incubated with 3·33 μg/ml of DQ-OVA for 30 min, 1 hr, or 2 hr at 37°. After each time-point, cells were extensively washed using PBS and then fixed with 2% paraformaldehyde (diluted from Cytofix; BD Pharmingen, San Jose, CA) for 10 min at room temperature. Fixed cells were washed twice using ice-cold FACS buffer, then examined using a FACS Canto (BD Biosciences). Negative untreated control cells were separately prepared by incubation of DCs with DQ-OVA on ice followed by fixation. The MFI of the ice control cells was subtracted from that of cells incubated at 37° with OVA per treatment or control. Data were analysed using the FlowJo Software (Tree Star).

Visualizing antigen internalization

Endocytic behaviour and morphology of DCs treated with chemokines and/or subsequent LPS were examined by confocal laser scanning microscopy. Briefly, DCs were collected on Day 1 and Day 2 post-treatment and resuspended in medium (without phenol red) at 1 × 106 cells/ml. Then, each sample was incubated with 5·8 μg/ml of fluorescent Alexa Fluor 488-Ovalbumin (OVA) (a model antigen) (Invitrogen) or 0·5 mg/ml Lucifer Yellow (LY) (Invitrogen) for 30 min at 37°. OVA is known to be internalized by DCs by a combination of receptor-mediated endocytosis and fluid-phase macropinocytosis[17] whereas LY is internalized by only fluid-phase macropinocytosis.[34] After incubation, any excess fluorochrome bound to cell surfaces was quenched for 3–4 min on ice using 0·5% Trypan Blue/2% FBS/1× PBS solution. After two sequential quenching steps, cells were washed three times using 1% BSA/PBS solution, resuspended in complete medium (without phenol red) at 1 × 106 cells/ml, then the cell suspension was used to submerge a glass cover slip and allowed to incubate for 4 hr at 37° to induce cell attachment to the cover glass. After incubation and another washing, cells were fixed with 2% paraformaldehyde for 10 min at room temperature, and permeabilized with 0·05% Triton-X 100 (Sigma) for 15 min at room temperature. Then, cells were washed three times, and incubated with Texas red-X phalloidin (Invitrogen) at 0·165 μm in 1% BSA/PBS solution for 20 min at room temperature. Cells were then washed and permanently mounted using Fluoromount G (SouthernBiotech, Birmingham, AL). Microscopic images were acquired with a Zeiss 510 META confocal laser scanning microscope (Zeiss, Thornwood, NY) using 100× /1·4 NA oil objective. For this analysis, at least seven cells were examined per treatment condition. Each cell was ‘optically sectioned’ by collecting x–y plane images or slices at 12–14 different z-direction altitudes through the cell (x-y slices were collected every Δ= 507 nm). A single x-y slice was selected from the middle of the z-stack of images (middle of the cell) for reporting here.

Surface marker expression

To measure expression levels of DC surface markers, cells were resuspended in FACS buffer, blocked with anti-mouse Fcγ III/II receptor monoclonal antibody (clone 2.4G2; IgG2bκ) (BD Pharmingen), and stained with saturating concentrations of fluorescently conjugated rat or mouse anti-mouse monoclonal antibodies against CD86 (clone GL1; IgG2aκ), MHC Class I (H-2Kb) (clone AF6-88.5; IgG2aκ) and MHC Class II (I-2Ab) (clone AF6-120.1; IgG2aκ) (all from BD Pharmingen) for 30 min at 4° in the dark. After staining, cells were extensively washed three times using ice-cold FACS buffer and then, analysed immediately with 10 000 events per sample using FACS Canto (BD Biosciences). An isotype per antibody against each surface marker was used as the negative control per group of DC treatment or control, and the MFI of the isotype control cells was subtracted from that of cells stained with the respective antibody per surface marker. Data were analysed using the FlowJo Software (Tree Star).

Cytokine release

Cell culture supernatant was saved after DC treatment with chemokines (Day 1) and subsequent LPS (Day 2). Culture supernatant was analysed for TNF-α, IL-1β, IL-4, IL-10, IL-12p70 (all from Peprotech) and IL-23 (R&D systems, Minneapolis, MN) using standard ELISA kits. All ELISAs followed the manufacturer's protocol, with small modifications; for colour development following a detection antibody incubation, the original combination of avidin–horseradish peroxidase and 2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) liquid substrate was replaced with a combination of streptavidin-horseradish peroxidase and tetramethylbenzidine substrate. At the time of supernatant collection, cell numbers were quantified using the CyQuant NF Cell Proliferation Assay Kit (Invitrogen) as per the manufacturer's protocol, using a TECAN Safire™ fluorimeter (MTX Lab Systems, Vienna, VA). ELISA data in pg/ml were normalized to a total cell number per unit sample volume.

Statistical analysis

Statistical analysis of all data was performed by comparison of each cell treatment with control iDCs or mDCs (LPS only) per experiment. A one-tail paired t-test was used when data were normalized to iDCs (= 1) (all data except the cytokine release results), whereas Mann–Whitney U-test was used when data were not normalized to the control (the cytokine release results). For both statistical methods, the GraphPad Prism (Version 5·04, La Jolla, CA) was used. If not indicated, P value ≤ 0·05 was considered to be significant.

Results

Treatment of DCs

The sequential treatment of iDCs with chemokines then LPS was carried out over a total of 4 days with cells and their surrounding medium analysed on the last 2 days. To clarify, one series of cells and their supernatant were analysed 24 hr after chemokine treatment (Day 1) and a second series of cells and their medium were analysed 24 hr after subsequent LPS treatment (Day 2) (Fig. 1). Briefly, cells were plated at 5 × 105 cells/ml (2 ml/well) in 12-well plates (Corning, NY) and then, after 24 hr, spent medium was replaced with fresh medium. After addition of the new medium, one set of cells was left untreated (iDC); one set of cells was treated with murine CCL3 (at 30, 50 or 70 ng/ml); one set of cells was treated with murine CCL19 (by 30, 50 or 70 ng/ml); and finally one set of cells received either a combination of CCL3 (50 ng/ml) and CCL19 (50 ng/ml) (ratio of 5 : 5), a combination of CCL3 (30 ng/ml) and CCL19 (70 ng/ml) (ratio of 3 : 7), or a combination of CCL3 (70 ng/ml) and CCL19 (30 ng/ml) (ratio of 7 : 3) (Peprotech). After another 24 hr (Day 1 after treatment), the contents of each well were divided equally; one portion was used for cell assays or secreted cytokine determination, whereas the other half of the cells was transferred to a new 24-well plate (1 ml/well at 5 × 105 cells/ml) and LPS (100 ng/ml) (Escherichia coli 0111:B4, Sigma) was added. For the original untreated negative control iDCs, after cell transfer to a new 24-well plate, one well still remained untreated whereas the other well was treated with LPS. After another 24 hr (Day 2), contents of each well were collected for either cell or cytokine assays. After DC treatment with chemokines (Day 1) and subsequent LPS stimulus (Day 2), cell viability was also determined using Trypan blue exclusion. All treatments and controls exhibited at least 90% viable cells (data not shown).

Figure 1.

Schematic representation of overall procedure of the JAWS II DC line treatment with chemokines and subsequent lipopolysaccharide (LPS). On Day 0, dendritic cell (DC) wells for a single chemokine treatment received 30, 50, or 70 ng/ml of CCL3 or CCL19 whereas DC wells for chemokine combinations received 50 ng/ml (CCL3) and 50 ng/ml (CCL19) for 5 : 5 ratio, 30 ng/ml (CCL3) and 70 ng/ml (CCL19) for 3 : 7 ratio, or 70 ng/ml (CCL3) and 30 ng/ml (CCL19) for 7 : 3 ratio. The endocytic capacity and phenotype changes were measured after cell and supernatant collection on Day 1 and Day 2.

Chemokine combination of CCL3 and CCL19 (7 : 3) induces endocytic capacity retained at level higher than iDCs, even after maturation by LPS

Hereafter, any combination of CCL3 and CCL19 at a specific ratio will adhere to the nomenclature: CCL3 + 19 (ratio).

To measure the endocytic capacity of DCs upon chemokine or subsequent LPS treatment, DCs were incubated with fluorescently labelled OVA 24 hr after chemokine treatment (Day 1) or 24 hr after subsequent LPS treatment (Day 2) and the amount of OVA internalized by DCs was determined using flow cytometry. Immature DCs treated with individual chemokines or chemokine combinations exhibited endocytic capacity comparable to untreated iDCs (Fig. 2a). As expected, upon subsequent LPS maturation, iDCs treated only with LPS reduced their endocytic capacity significantly compared with untreated iDCs. However, iDCs pre-treated for 24 hr (Day 1) with individual chemokines or an equal combination of CCL3 + 19 (5 : 5), then subsequently treated with LPS exhibited an endocytic capacity similar to untreated iDCs. Surprisingly, even after subsequent LPS treatment, iDCs pre-treated with CCL3 + 19 (7 : 3) showed an endocytic capacity 36% higher than untreated iDCs, whereas iDCs pre-treated with CCL3 + 19 (3 : 7) exhibited a 30% lower endocytic capacity than untreated iDCs (Fig. 2a,b). When endocytic capacity (MFIs by flow cytometry) was recalculated, now normalized to the value of endocytic capacity for untreated iDCs on Day 1, iDCs pre-treated with CCL3 + 19 (7 : 3) retained 57% of their endocytic capacity, even after subsequent LPS treatment. Conversely, the normalized endocytic capacity of untreated iDCs or iDCs treated with only LPS was reduced to 44% or 15%, respectively (Fig. 2c). Even though there is no direct evidence explaining why the endocytic capacity of untreated iDCs decreased over time, this natural decrease was presumably attributed to effects of the GM-CSF in the culture media[41-43] and of the cell transfer (Fig. 1)[41] on minimal maturation of these DCs during the 3-day culture in this study.

Figure 2.

Endocytic uptake of control and treated dendritic cells (DCs). Mean fluorescence intensity (MFI) of flow cytometry analysis of endocytic capacity [ovalbumin (OVA) uptake]: (a) upon DC treatment with chemokine and subsequent lipopolysaccharide (LPS); (b) representative histograms of OVA uptake on day 2 (after subsequent LPS treatment) shown with ice control, untreated immature DCs (iDCs), iDCs treated only with LPS, and iDCs pre-treated with CCL3 + 19 (7 : 3); and (c) retained OVA uptake capacity over subsequent LPS treatment (all these capacities on both Day 1 and Day 2 were normalized to untreated iDC on Day 1 to compare capacity changes from Day 1 through Day 2). Ratios to the untreated iDCs are shown with mean ± SD, n = 4 (four independent trials). * 0·05, significantly different as compared with untreated iDCs. A single chemokine was added only as 50 ng/ml whereas chemokine combinations varied with 30, 50, or 70 ng/ml per chemokine.

Cell morphology and endocytic behaviour are modulated by pre-treatment with chemokine combinations

To further quantify endocytic behaviour and the morphology of DCs upon pre-treatment with chemokines and subsequent LPS treatment microscopically, DCs were pulsed with either OVA (internalized by a combination of receptor-mediated endocytosis and fluid-phase macropinocytosis)[17] or LY (internalized by only fluid-phase macropinocytosis of DCs),[34] followed by staining with a phalloidin-specific antibody to visualize the actin cytoskeleton of cells.[37] As shown in Figs 3 and 4, upon iDC treatment with chemokine combinations of CCL3 + 19 (3 : 7) or (7 : 3), iDCs exhibited extensively ruffled membranes (Figs 3b,c and 4b,c) whereas untreated iDCs did not (Figs. 3a,d and 4a,d). Subsequent LPS treatment induced large extended veils[44] in addition to ruffled morphologies (Figs 3e–g and 4e–g).

Figure 3.

Confocal laser scanning microscopy images of ovalbumin (OVA) internalized by immature dendritic cells (iDCs) treated with chemokines and/or subsequent lipopolysaccharide (LPS). (a) OVA (green), Phalloidin (red), and merged images per sample of untreated iDCs; (b) iDCs treated with CCL3 + 19 (3 : 7); or (c) CCL3 + 19 (7 : 3) for Day 1 (before LPS treatment) are shown. For Day 2 (24 hr after subsequent LPS treatment), images in the same format per sample of (d) untreated iDCs, (e) iDCs treated only with LPS, (f) iDCs pre-treated with CCL3 + 19 (3 : 7), or (g) CCL3 + 19 (7 : 3) are shown. White scale bars indicate 10 μm.

Figure 4.

Confocal laser scanning microscope images of Lucifer Yellow (LY) internalized by immature dendritic cells (iDCs) treated with chemokines and subsequent lipopolysaccharide (LPS). (a) LY (green), Phalloidin (red), and merge images per sample of untreated iDCs, (b) iDCs treated with CCL3 + 19 (3 : 7), or (c) CCL3 + 19 (7 : 3) for Day 1 (before LPS treatment) are shown. For Day 2 (24 hr after subsequent LPS treatment), images in the same format per sample of (d) untreated iDCs, (e) iDCs treated only with LPS, (f) iDCs pre-treated with CCL3 + 19 (3 : 7), or (g) CCL3 + 19 (7 : 3) are shown. White scale bars indicate 10 μm.

Before LPS treatment, untreated iDCs or iDCs treated with both chemokine combinations exhibited spots or speckles of fluorescent OVA[45, 46] or LY[47] dispersed in large areas in the cell (Figs. 3a–c and 4a–c). However, after subsequent treatment with LPS, iDCs pre-treated with CCL3 + 19 (3 : 7) exhibited reduced areas of OVA or LY fluorescence, similar to iDCs treated with only LPS (Figs 3e,f and 4e,f). Remarkably, after subsequent LPS treatment, iDCs pre-treated with CCL3 + 19 (7 : 3) still exhibited OVA or LY spots or speckles showing much brighter accumulations in addition to faint green, indicating more internalized OVA or LY,[48] compared with all other DCs treated with LPS (Figs 3e–g and 4e–g). The morphologies and the endocytic behaviours of iDCs pre-treated with individual chemokines or CCL3 + 19 (5 : 5) were also examined but they did not exhibit morphologies different from iDCs pre-treated with CCL3 + 19 shown in Figs 3 and 4 or endocytic behaviours different from untreated iDCs or iDCs treated only with LPS (data not shown).

Surface marker expression for DC maturation is governed by subsequent LPS treatment, whereas MHC Class molecule expression is not

Co-stimulatory molecule (CD86), MHC Class I and MHC Class II expression on DCs 24 hr after chemokine treatment (Day 1) or 24 hr after subsequent LPS treatment (Day 2) were measured by flow cytometry to assess the DC phenotypic changes.

We originally tried to quantify the immunofluorescence results of surface marker (CD86 and MHC Class I and II) expressions on DCs upon programming and/or subsequent LPS treatment. However, as a result of unexpected variations of minimal response of the negative control (untreated iDCs) between independent trials (data not shown), results observed in this study needed to be normalized to untreated iDCs per trial for further discussion of statistical significance. Also, MFI normalization can represent normalization of the positive cell quantification based on a 5% preset background of each isotype in flow cytometry histograms (data not shown) for each surface molecule examination. For these reasons, we present data in percentage or ratio changes relative to the negative control of untreated iDCs, as ultimately the statistical significance of resultant DC behaviours is investigated, independently from the varying minimal response of immature DCs, upon DC programming by our new protocol.

Interestingly, iDCs treated with CCL3 + 19 (3 : 7) or (7 : 3) exhibited CD86 expression levels slightly lower than untreated iDCs before LPS treatment (Fig. 5a). However, after subsequent treatment with LPS, iDCs pre-treated with individual chemokines and their combinations showed CD86 expression levels that were not significantly different from iDCs treated only with LPS. As shown in Fig. 5(b), MHC Class I molecule expression for all treatments and controls was not significantly different from that of untreated iDCs before LPS treatment. After subsequent LPS treatment, none of the treatments and controls induced MHC Class I molecule expression levels that were significantly different from those of iDCs treated only with LPS. However, MHC Class II molecule expression was significantly affected by chemokine pre-treatment (Fig. 5c). Before LPS treatment, iDCs treated with CCL3, CCL19 or CCL3 + 19 (5 : 5) had significantly reduced expression levels (~30%) of MHC II, compared with untreated iDCs. After subsequent LPS treatment, both untreated iDCs and iDCs treated with CCL3 + 19 (7 : 3) exhibited levels of MHC Class II that were significantly lower (≥ 30%) than those of iDCs treated only with LPS.

Figure 5.

Cell surface markers in response to treatments. Mean fluorescence intensity (MFI) from flow cytometry analysis on surface marker expression of (a) CD86, (b) MHC Class I, and (c) MHC Class II upon dendritic cell (DC) treatment with chemokine and subsequent lipopolysaccharide (LPS). Ratios to untreated immature DCs (iDCs) are shown with mean ± SD, n = 3 (MHC molecules) or n = 4 (CD86) (three or four independent trials). Before LPS treatment (day 1), * 0·05, significantly different compared with untreated iDCs. After LPS treatment (Day 2), * 0·05, significantly different compared with iDCs treated only with LPS. A single chemokine was added only with 50 ng/ml whereas chemokine combinations varied with 30, 50, or 70 ng/ml per chemokine.

Individual chemokine versus chemokine combination pre-treatment

Since the specific combination of chemokines (CCL3 + 19 at 7 : 3) induced DC antigen uptake capacity at levels higher than untreated iDCs even after LPS treatment, we repeated the assays to assess whether individual chemokines at the same concentrations would induce similar responses. For this, a single chemokine of CCL3 or CCL19, at concentrations of 30, 50 or 70 ng/ml, was added into iDCs then LPS was added, as before. As seen in Fig. 6, 24 hr after subsequent LPS treatment (Day 2), individual CCL3 or CCL19 treatments at any concentration did not induce the DC antigen uptake enhancement induced by the chemokine combination of CCL3 + 19 (7 : 3), although they all induced DC antigen uptake capacities that were still significantly higher than iDCs treated only with LPS. In addition, CD86 and MHC Class II expression by iDCs pre-treated with all individual chemokines was not significantly different relative to untreated iDCs before LPS treatment, whereas CD86 and MHC Class II expression levels on the same DCs significantly increased at levels comparable to iDCs treated only with LPS after subsequent LPS treatment (Fig. 6b,d). After subsequent LPS treatment, only iDCs pre-treated with CCL19 at 70 ng/ml reduced MHC Class I molecule expression to levels significantly less than iDCs treated only with LPS (Fig. 6c).

Figure 6.

Dendritic cells (DCs) were treated with only a single chemokine (CCL3 or 19) by 30, 50, or 70 ng/ml and subsequent lipopolysaccharide (LPS). Mean fluorescence intensity (MFI) of flow cytometry analysis on (a) endocytic capacity [ovalbumin (OVA) uptake], (b) surface marker expression of CD86, (c) MHC Class I, and (d) MHC Class II upon DC treatment with chemokine and subsequent LPS. Ratios to untreated immature DCs (iDCs) are shown with mean ± SD, n = 3 (three independent trials). Before LPS treatment (Day 1), * 0·05, significantly different compared with untreated iDCs. After LPS treatment (Day 2), * 0·05, significantly different compared with iDCs treated only with LPS.

Proteolytic degradation of a model antigen by DCs was enhanced by pre-treatment with the chemokine combinations before subsequent LPS treatment

To examine the intracellular degradation (processing) of antigens by DCs upon treatment with chemokines and subsequent LPS, DQ-OVA was incubated with DCs and for various time periods (30 min, 1 hr, 2 hr). The intracellular degradation signal for all DCs was measured by flow cytometry; all data were normalized to the proteolytic degradation level of untreated iDCs seen after a 30-minute incubation with DQ-OVA (Fig. 7). Twenty-four hours after all chemokine pre-treatments, DCs exhibited essentially no statistical difference versus untreated iDCs in OVA degradation for the three time-points. As expected, once treated with LPS, mDCs exhibited enhanced antigen degradations compared with untreated iDCs. Interestingly, after subsequent LPS treatment, only the chemokine combination of CCL3 + 19 (7 : 3) induced antigen degradation after a 2-hr incubation at a statistically significant level, 27% higher than only LPS treatment. However, pre-treatment with individual chemokines at 50 ng/ml or combinations of CCL3 + 19 (5 : 5) or (3 : 7) did not induce antigen degradation levels that were statistically different from those seen after only LPS treatment.

Figure 7.

Mean fluorescence intensity (MFI) of flow cytometry analysis on model antigen [DQ-ovalbumin (OVA)] proteolytic degradation in a time course of 30 min, 1 hr and 2 hr after DQ-OVA was added to dendritic cells (DCs) previously treated with (a) chemokines and (b) subsequently lipopolysaccharide (LPS). Ratios to untreated immature DCs (iDCs) after a 30-minute incubation with DQ-OVA are shown with mean ± SD, n = 3 (three independent trials). * 0·05, significantly different compared with iDC treated only with LPS. A single chemokine was added only with 50 ng/ml whereas chemokine combinations varied with 30, 50, or 70 ng/ml per chemokine.

Cytokine secretion in response to DC pre-treatment with chemokines then LPS

Upon pre-treatment with chemokines or subsequent treatment with LPS, profiles of cytokines (IL-1β, TNF-α, IL-12p70, IL-23, IL-10 and IL-4) released into the supernatants of DCs were measured by ELISA. After subsequent LPS treatment, iDCs pre-treated with individual chemokines or chemokine combinations secreted IL-1β (Fig. 8a) and TNF-α (Fig. 8c) at levels that were statistically no different from iDCs treated only with LPS. Only the combination of CCL3 + 19 (7 : 3) induced IL-1β secretion at a level higher (50%) than untreated iDCs before LPS treatment, whereas TNF-α was below detectable limits for all DCs before LPS treatment.

Figure 8.

Cytokine expression in response to treatment. Expression of (a) interleukin-1β (IL-1β), (b) IL-10, (c) tumor necrosis factor-α (TNF-α), (d) IL-12p70, (e) IL-23 releases into supernatant upon dendritic cell (DC) treatment with chemokine and subsequent lipopolysaccharide (LPS). TNF-α, IL-12p70, or IL-23 release before LPS treatment (Day 1) was not detectable (below the detection limit) for all controls and treatments and results are shown only with data obtained after LPS treatment (Day 2). Cytokine concentrations (pg/ml) normalized to the total cell number per well (control or treatment) are shown with mean ± SD, n = 3 (three independent trials). Before LPS treatment (Day 1), * 0·05, significantly different compared with untreated immature DCs (iDCs). After LPS treatment (Day 2), * 0·05, significantly different compared with iDCs treated only with LPS. A single chemokine was added only with 50 ng/ml whereas chemokine combinations varied with 30, 50, or 70 ng/ml per chemokine.

Secretion levels of both IL-12p70 and IL-23 were below detectable limits for all DCs after just chemokine treatment (Fig. 8d,e). However, after subsequent LPS treatment, individual CCL3 or CCL19 or a combination of CCL3 + 19 (5 : 5) induced IL-12p70 secretion at levels lower than iDCs treated only with LPS, whereas only the combination of CCL3 + 19 (7 : 3) induced IL-23 secretion at a level higher than iDCs treated only with LPS.

While combinations of CCL3 + 19 (3 : 7) or (7 : 3) induced IL-10 secretion at a level higher than untreated iDCs before LPS treatment, all the treatments of iDCs exhibited IL-10 secretion levels similar to iDCs treated only with LPS after subsequent LPS treatment (Fig. 8b).

In addition to these cytokines, IL-4 secretion was also measured but IL-4 secretion levels of all DCs for both cases before and after LPS treatment were not detectable (data not shown).

Discussion

Results here indicate that chemokine pre-treatment can program DCs to internalize and process antigen, even after DC maturation by LPS. The pre-treatment of DCs with CCL3 + 19 (7 : 3) for 24 hr followed by subsequent LPS treatment for another 24 hr induced the endocytic capacity of DCs at levels 96% higher than iDCs that were only exposed to LPS. Our finding differs from that reported for the simultaneous application of antigen or dextran and chemokines, which enhanced DC endocytic capacity but only for less than an hour after treatment.[36, 49] Our results indicate that prolonged presence of chemokines in the cell culture well can modulate DC phenotypes against subsequent TLR stimulation.

Chemokines are known for their role in chemotaxis; inducing DC migration to the secondary lymphoid organs to present antigens to T cells, thereby initiating the adaptive immune response. Besides this conventional role, chemokines have recently been recognized for their additional influence on DC function; such as cell marker expression, cytokine production, dendrite extension, apoptosis, or phagocytosis.[30] So, quantifying chemokine impact on DC phenotype could provide grounds for new immunotherapeutic strategies.

Podosomes are generally described as dynamic assemblies of actin molecules,[50] and iDCs readily form actin-rich podosomes that play a role in extracellular matrix degradation and migration of DCs through tissues.[51, 52] A disassembly of DC podosomes coincides with increases in DC endocytosis while fully matured DCs do not form podosomes.[53] Chemokine (CCL3) induces chemotaxis of iDCs in association with complete remodelling of the actin cytoskeleton, which leads to dissolution of podosomes and to a change of DC morphology.[54] Actin cytoskeleton remodelling depending on chemokines also suggests that the disappearance of podosomes and the acquisition of migratory ability by DCs are linked.[54] Moreover, CCL3 enhances endocytic behaviour of iDCs rapidly within a few minutes, although the exact mechanism still remains unclear.[35, 36]

Cell division control protein 42 (Cdc42) is a small GTPase (an enzyme that hydrolyses guanosine triphosphate) that controls actin cytoskeleton remodelling[55] and regulates endocytosis of DCs; whereas blockage of Cdc42 reduces endocytosis in iDCs. Transfection of this molecule in mDC enhanced their endocytic capacity.[56] In addition, disassembly of podosomes is independent of Cdc42 activation status,[53] and when mDCs are exposed to CCL19, the Cdc42 activation and the endocytic capacity of mDCs increases rapidly within a few minutes.[36] Yanagawa and Onoe[57] also found that CCL19 induces the extension of dendrites in mDCs. From these observations, we can postulate that DC treatment with select chemokines may activate Cdc42 in iDCs or mDCs, which affects actin cytoskeleton reorganization and endocytic behaviour of DCs.

Ovalbumin is internalized by iDCs through a combination of mannose receptor-mediated endocytosis and fluid-phase macropinocytosis, and when the mannose receptor is blocked, OVA internalization of iDCs is reduced by ~20%.[17] These findings suggest that macropinocytosis contributes to OVA internalization by iDCs more than mannose receptor-mediated endocytosis. Upon maturation of DCs, expression of mannose receptors on the cell surface is down-regulated[58] and DCs cease macropinocytosis.[47] Yanagawa and Onoe[36] reported that when CCL19 is added to mDCs, CCL19 does not increase macropinocytosis in mDCs. Here, CCL3 or CCL19 or their combinations were added to iDCs for 24 hr, and then DCs were intentionally matured with LPS for another 24 hr in the presence of chemokines. Hence, it is conceivable that low levels of CCL19 (30 ng/ml) in the chemokine cocktail, induced more OVA internalization (Figs 2 and 6a) mainly by inducing DC macropinocytosis at high levels, even after LPS treatment.

This increased macropinocytosis was attributed to active membrane ruffling driven by actin cytoskeleton reorganization of DCs,[47] which is supported by the cell morphology changes observed by confocal laser scanning microscopy (Figs 3 and 4). First of all, iDCs pre-treated with the chemokine combinations of CCL3 + 19 (3 : 7) or (7 : 3) (before LPS treatment) exhibited active membrane ruffling associated with actin cytoskeleton reorganization. Once subsequently treated with LPS, iDCs pre-treated with chemokines exhibited extended veils, still retaining the previously-formed membrane ruffling. Following DC endocytosis, whereas peptides derived from antigen proteins are transported to the DC surface by MHC Class II molecules, impermeable compounds such as LY accumulate in the cell.[47] In line with this, iDCs pre-treated with CCL3 + 19 (7 : 3) then treated with LPS exhibited dispersed OVA and accumulated LY in green brighter than other DCs (Figs 3g and 4g). This indicates that higher amounts of OVA or LY were internalized by iDCs pre-treated with CCL3 + 19 (7 : 3), then subsequently treated with LPS compared with other DCs. Qualitative evidence therefore suggests that pre-treatment of iDCs with CCL3 + 19 (7 : 3) induces DC endocytic (including macropinocytosis) capacity at a higher level even after subsequent LPS treatment.

Whereas CCL3 does not induce DC maturation,[54] CCL19 is known as a potent natural adjuvant inducing full maturation of DCs.[31] Upon maturation by TLR agonist such as LPS, DCs express cell surface markers of MHC Class II and CD86 and secrete cytokines of TNF-α, IL-6, IL-1β, IL-12 and IL-10 at high levels.[31, 47, 59, 60] However, DCs cannot be fully matured (so called semi-maturation of DCs) when DCs are exposed to specific stimulants or conditions.[59, 60] Interestingly, semi-matured DCs are non-responsive to subsequent TLR stimulation[61] or resist LPS-induced maturation.[62] In this study, iDCs pre-treated with CCL3 + 19 (7 : 3) secreted IL-1β and IL-10 at levels higher than iDCs before LPS treatment (Fig. 8a,b) but they expressed CD86 or MHC Class II molecules at levels lower or similar to iDCs, before LPS treatment (Fig. 5a,c). Moreover, even after subsequent LPS treatment, DCs pre-treated with CCL3 + 19 (7 : 3) still expressed MHC Class II molecules at levels significantly lower than iDCs treated only with LPS, thus appearing not to respond to LPS treatment. Hence, this chemokine combination (more CCL3 and less CCL19) seemingly induces DCs into a condition very similar to semi-maturation before LPS treatment, and then presumably suppresses or delays MHC Class II expression on DCs after exposure to LPS.

Results shown in Fig. 7 imply that both antigen uptake and processing by DCs after maturation can be enhanced at the same time through DC programming by CCL3 + 19 (7 : 3). Moreover, CD86 expression up-regulated following subsequent LPS treatment additionally supports the theory that chemokine programming may prime DCs for processing intracellular peptides derived from antigens and co-stimulatory molecules to stimulate T cells.

However, MHC Class II molecule expression of iDCs pre-treated with CCL3 + 19 (7 : 3) was not enhanced before (compared with untreated iDCs) and after LPS treatment (compared with iDCs treated only with LPS) (Fig. 5c). This observation indicates that even though the programmed DCs continue to internalize and process antigens, chemokine pre-treatment may delay up-regulating peptide–MHC II complexes on the cell surface, thereby failing to effectively present antigens to T cells. Hence, in Part II of this study, we are quantifying the antigen presentation capacity of these programmed DCs and the subsequent T-cell response.

In addition to higher levels of IL-1β and IL-10 secretions from iDCs programmed by CCL3 + 19 (7 : 3) versus untreated iDCs before subsequent LPS treatment, programmed DCs secreted IL-23, after subsequent LPS treatment, at higher levels (44%) than iDCs treated with only LPS. These differential outcomes of various cytokines secreted from DCs also suggest that chemokine programming has a multifunctional impact on modulating the adaptive immunity by signals other than antigens or co-stimulatory molecules. For example, IL-1β and IL-23 secreted from the programmed DCs can accumulate until after subsequent TLR stimulation, and then induce Th17 polarization,[63] which plays a critical role in autoimmune diseases or anti-microbial immunity. Hence, hypothetically chemokine programming of DCs could provide immunomodulating strategies for both innate and adaptive immunity against various pathologies.

As the chemokine combination of CCL3 + 19 (7 : 3) induced DC endocytic capacity retained at high levels even after subsequent LPS treatment, we have examined how the chemokine receptor expressions on the DC surface are modulated upon treatment of DCs with chemokines and subsequent LPS. In this examination, DCs were pre-treated with single CCL3 (70 ng/ml), CCL19 (30 ng/ml), or their combination (7 : 3), and then chemokine receptor expressions on the DC surface were measured using flow cytometry and fluorescently labelled antibodies against mouse CCR5 or CCR7 on Day 1 and Day 2 schedules, as shown in Fig. 1.

Unexpectedly, it was not possible to observe any statistically meaningful data of CCR expressions between DC treatments. Also, CCR5 expressions on JAWSII DC line surface were at very low levels (data not shown). Possibly because of the DC line's unknown immunobiological functions, which are not exactly the same as the primary DCs,[64] we could not determine how CCR5 or CCR7 expressions are modulated upon pre-treatments of this DC line with individual chemokines or their combination. However, we found that CCR5 expressions on untreated iDCs decreased or CCR7 expressions on untreated iDCs increased upon DC maturation (data not shown). Therefore, we can conclude, at least, that even though this JAWSII DC line up-regulates CCR5 or CCR7 at low levels, this cell line still expresses these two chemokine receptors that respond to DC maturation in the same way as other DCs in the literature. Further study using other measurements (e.g. intracellular CCR protein expressions) should be performed to better understand the CCR expressions depending on the chemokine pre-treatment and subsequent LPS treatment.

Numerous DC-based vaccine strategies have emerged as new immunotherapeutics[3, 4, 65]: nanoparticles delivering specific antigen in vivo to DCs[66]; DCs programmed in vivo by cytokines released from an implant biomaterial scaffold[14]; or by in vivo pre-injection of cytokines.[67] Interestingly, when DCs are pre-treated with glucocorticoids (dexamethasone) in vitro, the endocytic capacity and the expression levels of receptors for endocytosis after DC maturation by TNF-α, remained higher than control DCs (no dexamethasone), but CD86 expression was suppressed before and after TNF-α stimulation.[34]

Certainly, chemokine programming of DCs appears a feasible way to directly or indirectly control adaptive immunity. To further confirm the multifunctional impacts of chemokine programming, we are currently quantifying the interaction of the programmed primary bone marrow-derived DCs and T cells.

We demonstrate here that two different chemokines, each of which is selectively recognized by iDCs or mDCs, have a synergistic impact on programming DCs to retain their endocytic capacity, even after DC maturation. Further, we show that this programming induces multifunctional effects on the DC phenotype. These results suggest that DC-based vaccine strategies could be modified by overcoming the natural limit (significant reduction of antigen uptake and processing upon DC maturation) of the host immune response. For instance, ex vivo transfection of DCs can be enhanced by chemokine containing medium, whereas in vivo programming of DCs could be possible using implanted biomaterials releasing chemokines and antigen sequentially or chemokine/antigen targeting iDCs residing in lymphoid organs.[68] In this way, even though iDCs may be accidently pre-matured by an adjuvant before internalizing antigens, they would still retain their endocytic capacity at a certain level, which would increase the overall vaccine efficiency.

Acknowledgements

This work was generously supported by the National Institutes of Health: NIAID R01AI074661 and NIDCR R01DE018701.

Disclosure

The authors declare no competing interests.

Ancillary