Chemokine Up-Regulation and Activated T Cell Attraction by Maturing Dendritic Cells

See allHide authors and affiliations

Science  30 Apr 1999:
Vol. 284, Issue 5415, pp. 819-822
DOI: 10.1126/science.284.5415.819


Langerhans' cells migrating from contact-sensitized skin were found to up-regulate expression of macrophage-derived chemokine (MDC) during maturation into lymph node dendritic cells (DCs). Naı̈ve T cells did not migrate toward MDC, but antigen-specific T cells rapidly acquired MDC responsiveness in vivo after a subcutaneous injection of antigen. In chemotaxis assays, maturing DCs attracted activated T cells more strongly than naı̈ve T cells. These studies identified chemokine up-regulation as part of the Langerhans' cell maturation program to immunogenic T cell–zone DC. Preferential recruitment of activated T cells may be a mechanism used by maturing DCs to promote encounters with antigen-specific T cells.

Skin DCs, or Langerhans' cells (LCs), form an extensive network in the epidermis, and upon exposure to noxious stimuli such as lipopolysaccharide (LPS) or to contact sensitizers such as fluorescein isothiocyanate (FITC), they enter lymphatics and migrate into T cell zones of lymph nodes to become interdigitating DCs (1). For mature DCs to function in an immunogenic manner, it is important that they rapidly interact with antigen-specific T cells. One mechanism that may enhance encounters between DCs and T cells is for DCs to produce T cell–attracting chemokines. Although DCs express several chemokines (2–5), it has not been established whether immature DCs such as LCs up-regulate chemokine expression during in vivo maturation to immunogenic DCs.

Macrophage-derived chemokine (MDC) is highly expressed in lymphoid tissues (5–8). To characterize its expression in more detail, we isolated the mouse MDC homolog and used it to probe lymphoid tissues by in situ hybridization (9). Multiple strongly hybridizing cells were identified in T cell zones of lymph nodes (Fig. 1B), whereas only occasional cells hybridized in the spleen (Fig. 1A). Frequently, the MDC-expressing cells were concentrated in regions of the T cell zone proximal to follicles (Fig. 1B), which are areas rich in DCs (1), and under higher magnification, they demonstrated a dendritic morphology (Fig. 1C). Consistent with the in situ data and with findings by others (5–8), Northern (RNA) blot analysis showed that MDC was highly expressed in lymph node DCs (Fig. 1D) and was undetectable in DC-depleted lymph nodes or in peritoneal macrophages (Fig. 1D). To test for the production of MDC protein, we incubated purified lymph node DCs in vitro to allow chemokine secretion, and we tested the culture supernatant for the attraction of mouse CC-chemokine receptor 4 (CCR4)–transfected cells (10–13). A chemotactic response was observed with transfected but not with control cells (Fig. 1E). In addition to making MDC, we found, in preliminary studies (14), that lymph node DCs express thymus- and activation-regulated chemokine (TARC), which is a second CCR4 ligand (15), and it is possible that TARC contributes to the CCR4 cell attractant activity made by DCs.

Figure 1

MDC expression by DCs in lymph nodes. (A through C) Bright-field micrographs showing hybridization of the mouse spleen (A) and lymph node (B and C) with an antisense RNA MDC probe. Objective magnifications are ×5 (A and B) and ×40 (C). (D) Northern blot analysis of MDC mRNA in total lymph node (LN), lymph node cell suspension depleted of CD11c+ DCs (LN – DC), metrizamide-enriched lymph node DCs (DC), purified B and T cells, and day 7 thioglycollate-elicited peritoneal macrophages (Mac.). Elongation factor–1 (EF-1) hybridization indicates relative amounts of RNA in each lane. (E) Chemotactic response of CCR4-transfected cells but not vector control cells to MDC or lymph node DC culture supernatant (DC s/n) that was prepared as described (21).

After the exposure of skin to contact sensitizers, LCs migrate to lymph nodes but do not enter circulation or migrate to the spleen (1,16, 17), raising the possibility that the MDC-expressing DCs in lymph nodes might be LC derived. We therefore tested whether exposure to the contact sensitizer FITC resulted in changes in lymph node MDC expression. One day after skin was painted with FITC (18), MDC expression in draining lymph nodes was increased threefold (Fig. 2A), and it remained elevated over the next 2 days (Fig. 2A). In contrast, no increase in MDC occurred in the spleen after an intravenous injection of LPS (Fig. 2A), a treatment that promotes the migration of DCs from the splenic marginal zone to the T cell zone (19). In agreement with previous studies (20), FITC+ major histocompatability complex (MHC) class II+ DCs could be detected among cells from lymph nodes draining a site of skin painted with FITC, and these DCs were absent from nondraining nodes (Fig. 2B). A microscopic analysis of isolated DCs showed that the FITC was concentrated within intracellular compartments (Fig. 2C). When day one draining lymph nodes were sectioned and stained to identify B and T cell zones, small numbers of brightly fluorescent FITC+cells could be identified within the T cell zone but not in follicles (Fig. 2D). Combining this technique of tracking migrating DCs with in situ hybridization for MDC (9), we found that it was possible to establish that many of the FITC-bearing cells in the T cell zone were strongly positive for MDC expression (Fig. 2E). Reverse transcriptase–polymerase chain reaction (RT-PCR) analysis of sorted FITC+ lymph node DCs (21) confirmed that MDC was expressed in these cells (Fig. 3A). In contrast, EBI-1 ligand chemokine (ELC), a chemokine expressed constitutively by a subset of T cell zone DCs (4), was not detectable among FITC+ DCs (Fig. 3A). Thus, many LCs express MDC but not ELC as they migrate into lymph node T cell zones to become mature DCs.

Figure 2

Up-regulation of MDC by DCs migrating from FITC-sensitized skin to the T cell zone of lymph nodes. (A) Northern blot analysis of MDC in draining lymph nodes at days 0, 1, 2, and 3 after FITC skin painting (18) and in the spleen 1 day after intravenous injection of phosphate-buffered saline (–) or 35 μg of LPS (+). Top panels show Northern blot analysis probed with MDC and EF-1, and the bottom panel displays the relative change in MDC expression in comparison to untreated mice. (B) Flow cytometric analysis of draining lymph node cells 1 day after FITC skin painting (right) and from the contralateral node (left) showing FITC fluorescence and MHC class II expression. Profiles shown are gated for B220 cells. Numbers indicate the percentage of total cells within the overlying gate. (C) Dual-color fluorescence microscopy of isolated FITC+ lymph node DCs stained in red with anti-MHC class II biotin and streptavidin-Cy3. (D) Fluorescence microscopy of a day 1 draining lymph node section showing B220-stained B cell areas (red) and FITC fluorescence by DCs within the T cell zone. Objective magnification is ×10. (E) Combined bright-field and fluorescence microscopy of a day 1 draining lymph node section after MDC in situ hybridization showing MDC expression (black) and FITC fluorescence. Objective magnification is ×10 in the main panel and ×40 in the inset.

Figure 3

Presence of MDC but not ELC expression by FITC+ lymph node DCs and cultured LCs. (A) RT-PCR analysis for MDC and ELC in total lymph node cells (LN), sorted MHC class II+ LCs that had been cultured for 2 days (2d LC), and sorted FITC+ MHC class II+B220 DCs from draining lymph nodes 1 day after skin painting (FITC+ DC). Actin RT-PCR is shown as a loading control. (B) Northern blot analysis of MDC and ELC expression in RNA from freshly prepared epidermal cells (0 hours) and from epidermal cells that were cultured for 2, 4, 6, 24, and 48 hours. RNA from the lymph node is included for comparison.

To determine whether MDC was already expressed by LCs in the skin or whether expression was up-regulated upon maturation, we prepared cell suspensions from epidermal sheets and cultured them to allow LC maturation (21, 22). About 3% of epidermal cells are MHC class II+ DCs (22), a frequency similar to the DC frequency in lymph nodes (1). Northern blot analysis detected an MDC signal in lymph node RNA, but it only detected a weak signal in RNA from freshly isolated epidermal cells (Fig. 3B), establishing that LCs do not constitutively express substantial amounts of MDC. However, after 4 hours of in vitro maturation, MDC expression began to increase (Fig. 3B), and by 2 days, a 150-fold up-regulation had occurred (Fig. 3B). High MDC expression was also detected by RT-PCR in LCs sorted from day 2 epidermal cell cultures (Fig. 3A). In contrast, ELC was not up-regulated in either the epidermal cell suspension or the sorted LCs (Fig. 3, A and B). The kinetics of MDC up-regulation were similar to the time required in vivo for maturing LCs to reach the T cell zone of draining lymph nodes (20), supporting the notion that MDC contributes to the function of LCs that have matured to T cell–zone DCs.

The expression of MDC by DCs draining into lymph nodes suggested a role in attracting T cells into contact with antigen-bearing DCs. Chemotaxis studies have demonstrated that MDC is not chemotactic for naı̈ve or memory T cells but that MDC can attract T cells that have been activated in vitro and allowed to rest for several days, with repeated rounds of activation and rest improving the response (6–8,23). To better define if T cells may be able to respond to MDC during the early phase of an immune response, we tested MDC responsiveness of in vivo activated ovalbumin (OVA)–specific T cells (24, 25). A striking up-regulation in chemotaxis among OVA-specific T cells in draining lymph nodes was observed 2 days after immunization (Fig. 4A), which is a time when many of the cells have been activated (25), whereas OVA-specific T cells in nondraining nodes did not respond (Fig. 4A). Although the activated T cells remained responsive for 5 days after immunization (Fig. 4B), the response was not increased in comparison to that at day 2, indicating that the acquisition of MDC responsiveness is an early event in in vivo T cell activation.

Figure 4

(A and B) Rapid acquisition of MDC responsiveness in T cells after in vivo activation by antigen. OVA-specific TCR transgenic cells were transferred, and mice were immunized as described (24). Response at day 2 of draining and nondraining (ctrl) lymph node (LN) OVA-specific cells to increasing concentrations of MDC (A). Response of OVA-specific draining lymph node cells to no chemokine (–) or to MDC (10 μg/ml) (+) at days 1 (d1), 2 (d2), 3 (d3), and 5 (d5) after immunization (B). (C) Chemotaxis of activated T cells to factors secreted by FITC+ DC. To the lower transwell chambers, we added supernatant from cultures of no cells (–), sorted FITC+DCs (F) (21), or lymph node stromal cells (S) (21). Responder cells that were added to the upper well in the absence (–) or presence (M) of MDC (1 μg/ml) were day 2 draining lymph node cells that had been cultured with IL-2 (24). The percentage of input OVA-specific CD4 T cells that migrated from the upper to lower chamber was measured. (D) Chemotaxis of naı̈ve T cells to factors secreted by lymph node stromal cells. Culture supernatants added to lower wells were as in (B). Responder cells were unstimulated lymph node CD4 T cells. Results in (C) and (D) are representative of five experiments.

To determine whether MDC was representative of the overall chemoattractant profile of LC-derived DCs, we tested the ability of these cells to attract naı̈ve and activated T cells. FITC+ DCs from lymph nodes draining FITC-painted skin were cultured in vitro for 1 day to allow chemokine secretion (21), and the culture supernatant was used in chemotaxis assays. A chemotactic response was observed among activated T cells (Fig. 4C), whereas naı̈ve T cells responded weakly or not at all (Fig. 4D). The addition of MDC to the activated T cells to desensitize CCR4 caused a partial reduction in their chemotactic response, indicating that MDC (and possibly TARC) makes a contribution to the chemokine activity but also that the DCs make other chemokines that attract activated T cells (Fig. 4C). To establish whether our assay system was able to detect the production of naı̈ve cell attractants, we tested supernatants from cultures of lymph node stromal cell preparations (21) that contained cell types constitutively expressing known attractants of naı̈ve cells, including secondary lymphoid-tissue chemokine (SLC) and ELC (26). These supernatants were active in attracting naı̈ve CD4 T cells (Fig. 4D), whereas they had little activity for activated cells (Fig. 4C).

Our findings add up-regulation of T cell–attracting chemokine expression to the program of events that accompany LC maturation to immunogenic T cell–zone DCs. It is likely that other types of immature DCs up-regulate MDC during maturation as MDC is detected within mesenteric lymph nodes (14) that do not receive LCs from skin but that do contain DCs from intestinal epithelium (27). We propose that there are two phases of chemokine involvement in the early part of the T cell response in lymphoid tissues. In the first phase, maturing DCs and naı̈ve T cells are attracted into the T cell zone by constitutively expressed chemokines such as SLC and ELC (26). Brought together in this way, antigen-loaded DC can prime rare antigen-specific T cells. In the second phase, activated and dividing T cells migrate more efficiently than the surrounding naı̈ve T cells toward other newly arriving antigen-bearing DCs that produce chemokines such as MDC. This second phase should enhance the chance of encounter between antigen-specific T cells and antigen-bearing DCs. A similar mechanism of activated T cell attraction is likely to be used by antigen-bearing B cells (8). Expression of chemokines that differentially attract subsets of activated CD4 T cells (3, 7,23) may also help different DC types to amplify polarized T cell responses.

  • * To whom correspondence should be addressed. E-mail: cyster{at}


View Abstract

Navigate This Article