β-Defensins: Linking Innate and Adaptive Immunity Through Dendritic and T Cell CCR6

See allHide authors and affiliations

Science  15 Oct 1999:
Vol. 286, Issue 5439, pp. 525-528
DOI: 10.1126/science.286.5439.525


Defensins contribute to host defense by disrupting the cytoplasmic membrane of microorganisms. This report shows that human β-defensins are also chemotactic for immature dendritic cells and memory T cells. Human β-defensin was selectively chemotactic for cells stably transfected to express human CCR6, a chemokine receptor preferentially expressed by immature dendritic cells and memory T cells. The β-defensin–induced chemotaxis was sensitive to pertussis toxin and inhibited by antibodies to CCR6. The binding of iodinated LARC, the chemokine ligand for CCR6, to CCR6-transfected cells was competitively displaced by β-defensin. Thus, β-defensins may promote adaptive immune responses by recruiting dendritic and T cells to the site of microbial invasion through interaction with CCR6.

Defensins, a family of small (3.5 to 4.5 kD) cationic antimicrobial peptides with three to four intramolecular cysteine disulfide bonds, are found in mammals, insects, and plants (1–4). On the basis of the position and bonding of six conserved cysteine residues, defensins in vertebrates are divided into two categories, designated as α- and β-defensins (1, 2). Unlike α-defensins that are produced by neutrophils and intestinal Paneth cells, β-defensins are primarily expressed by epithelial cells of the skin, kidneys, and trachea-bronchial lining of nearly all vertebrates, where they can be released upon microbial invasion or up-regulated by stimulation with lipopolysaccharide and tumor necrosis factor–α (TNF-α) (2,5, 6). Two types of human β-defensins (HBDs), β-defensins 1 and 2 (HBD1 and HBD2, respectively), have been identified (6, 7). Inactivation of the antimicrobial activity of HBDs is reported to contribute to the recurrent airway infections in patients with cystic fibrosis (8, 9).

The α-defensins may also have roles in protecting the host, based on their capacity to chemoattract T cells (10), to promote host immunity (11), and to activate the classical complement pathway (12). Because β-defensins are released upon microbial invasion and are located at the host-environment interface, such as mucosal surfaces and skin (2, 5, 6), they may also function to alert the adaptive immune system of vertebrates. We therefore investigated whether HBDs could chemotactically mobilize human dendritic cells (DCs), monocytes, and T cells (13,14). Psoriatic skin-derived pure HBD2 (skin HBD2), synthetic HBD2 (sHBD2), and recombinant HBD2 (rHBD2), all induced substantial in vitro migration of CD34+ progenitor–derived DCs in a dose-dependent manner, with optimal concentrations of HBD2 usually at 1000 ng/ml (Fig. 1A). Liver and activation-regulated chemokine (LARC, also known as MIP-3α), which selectively acts on immature DCs (15), was also chemotactic (Fig. 1A). Fluorescence-activated cell sorting (FACS) analyses (16) revealed that these DCs expressed very little CD83, low amounts of CD1a and CD86, and a moderate amount of human leukocyte antigen–DR (HLA-DR) (Table 1), which are phenotypic characteristics of immature DCs (17). Mature DCs (14) demonstrated high expression of CD1a, CD83, CD86, and HLA-DR (Table 1) and did not migrate in response to HBD2 (18). Examination of other mononuclear cell types revealed that sHBD2 chemoattracted only the memory subset of human peripheral blood T (CD4+/CD45RO+) cells (Fig. 1B); naı̈ve T (CD4+/CD45RA+) cells or monocytes were not chemoattracted (18). Moreover, migration of DCs was inhibitable by treatment with pertussis toxin (PTX) (18), suggesting that HBDs might use one or more seven–transmembrane domain receptor or receptors coupled to Giα protein. Thus, HBDs might use one or more of the chemokine receptors (19).

Figure 1

Chemotaxis by HBDs of human immature DCs and memory T cells. (A) Migration of CD34+progenitor–derived DCs in response to skin HBD2 (▪), rHBD2 (⧫), sHBD2 (▴), and LARC (•). (B) Chemoattraction of human peripheral blood memory T cells by LARC (dotted bars), sHBD2 (hatched bars), and SDF-1α (black bar). The migration of immature DCs and memory (CD4+/CD45RO+) T cells (14) in response to HBDs and chemokines was examined by chemotaxis assay (13). Filter membranes of 5-μm pore size, uncoated or coated with fibronectin, were used to assay the migration of immature DCs and memory T cells, respectively. A representative of two to three experiments is shown. The results are presented as the number of cells per high power field (HPF). Error bars, SD of triplicated wells.

Table 1

Surface marker expression of DCs. Immature and mature DCs were generated from CD34+ progenitor cells as described (14) and analyzed by FACScan (15). At least 2500 events were collected for each sample. The mean ± SD of two separate experiments is given. MFI, median fluorescence intensity.

View this table:

Only four of the chemokine receptors (including CXCR4, CCR1, CCR5, and CCR6) are expressed on CD34+ progenitor–derived immature DCs (15, 20–22). We therefore tested whether HBDs could attract human embryonic kidney 293 (HEK293) cell lines stably transfected to express these receptors (23–26). Skin HBD2 induced the migration of CCR6-expressing HEK293 (CCR6/293) cells in a dose-dependent fashion but did not induce directional migration of CXCR4/293, CCR1/293, CCR5/293, or parental HEK293 cells (Fig. 2A). Skin HBD2, rHBD2, sHBD2, sHBD1, and LARC were chemotactic for CCR6/293 cells in a typical bimodal manner (Fig. 2B). LARC was more potent than HBDs in chemoattracting immature DCs, memory T cells (Fig. 1), and CCR6/293 cells (Fig. 2B), although HBDs and LARC showed similar efficacy. This may be attributable to the smaller size of HBD peptides. Because sHBD2 is indistinguishable from natural HBD2 with respect to its antimicrobial activity and structure (27) and because it showed chemotactic potency and efficacy similar to that shown by other HBDs (Figs. 1 and 2B), we used sHBD2 for most of our subsequent experiments. As expected, CCR6/293 cell migration induced by sHBD2 and LARC was blocked by 30 min before treatment with PTX (100 ng/ml) at 37°C (Fig. 2C).

Figure 2

Selective induction of CCR6/293 cell migration by HBDs. (A) Migration of HEK293 (▾), CXCR4/293 (▪), CCR1/293 (▴), CCR5/293 (⧫), or CCR6/293 (•) cells to skin HBD2. (B) Migration of CCR6/HEK293 cells in response to LARC (⧫), skin HBD2 (▴), rHBD2 (•), sHBD2 (▪), and sHBD1 (▾). Cell migration was investigated by chemotaxis assay (13), with a 10-μm filter membrane precoated with rat-tail collagen (Sigma). Spontaneous cell migration varied, but it was typically in the range of 30 to 50 cells per HPF. Error bars are omitted for clarity. (C) Sensitivity of sHBD2-induced CCR6/293 cell migration to PTX. The cells were incubated at 37°C for 30 min without (white bar) or with (black bar) PTX (100 ng/ml) (Sigma) before chemotaxis assays. Similar results were obtained from at least two separate experiments. Error bars, SD of triplicated wells.

To test for chemokinesis, we performed a simplified checkerboard analysis by comparing sHBD2 and LARC, the sole chemokine ligand identified so far for CCR6 (15, 18, 19, 22). Synthetic HBD2 and LARC were chemotactic for CCR6/293 cells because equal concentrations of these agents in the upper and lower wells of the chemotaxis chamber did not increase cell migration over medium control (Fig. 3A, compare bars 3 and 5 with bar 1). The presence of LARC in the upper wells largely attenuated sHBD2-induced CCR6/293 migration and vice versa. Consequently, sHBD2 and LARC appeared to compete with each other for use of CCR6. This possibility was further tested by examining whether sHBD2 could competitively inhibit the binding of LARC to CCR6/293 cells. Although less potent than unlabeled LARC, sHBD2 displaced the binding of iodinated LARC (125I-LARC) to CCR6/293 cells in a dose-dependent manner with a median effective concentration of ∼700 ng/ml (Fig. 3B). This difference in affinity correlates with the data showing that HBDs were less potent than LARC in attracting immature DCs, memory T cells, and CCR6/293 cells (Figs. 1 and 2B).

Figure 3

Mediation by CCR6 of cell migration induced by HBDs. (A) Cross-desensitization of CCR6/293 cell chemotaxis by sHBD2 and LARC. LARC (100 ng/ml) and sHBD2 (1000 ng/ml) were added into the lower wells, and CCR6/293 cells in CM or in CM containing LARC (100 ng/ml) or sHBD2 (1000 ng/ml) were added into the upper wells of a chamber. The cell migration was evaluated as described (13). Error bars, SD of triplicated wells. (B) Binding inhibition. Competitive binding was performed in triplicate by adding a constant amount of 125I-LARC and increasing amounts of sHBD2 (▪) or LARC (•) to individual 1.5-ml microfuge tubes, each containing 2 × 106 CCR6/293 cells suspended in CM. After incubation at 24°C with constant mixing for 1 hour, the mixture was centrifuged through a 10% sucrose/PBS cushion, and the cell-associated radioactivity was measured with a 1227 Wallac gamma counter. CPM, counts per minute; error bars, SD of triplicated tubes. (C) Dependence on CCR6 of sHBD2-induced immature DC migration. Chemotaxis of DCs in response to sHBD2 (100 ng/ml) (dotted bars) or LARC (10 ng/ml) (hatched bars) in the absence (–) or presence (+) of antibodies (50 μg/ml) as specified was examined by chemotaxis assay (13). Antibodies were added at identical concentrations into the upper and lower wells of a chamber. DCs were pretreated with antibodies at 24°C for 30 min before they were added into the upper wells of a chamber. Error bars, SD of triplicated wells.

Next, the role of CCR6 in DC migration induced by HBDs (Fig. 1A) was investigated. About 50% of immature DCs generated (14) were shown by FACS analysis to express CCR6 on their surface (28). Moreover, DC migration induced by a suboptimal dose (100 ng/ml) of sHBD2 was inhibited by 50 μg/ml of antibodies to CCR6 [immunoglobulin G2b (IgG2b) (R&D Systems, Minneapolis, MN)], but it was not inhibited by 50 μg/ml of isotype-matched antibodies to CCR5 (IgG2b) (Fig. 3C, dotted bars). This inhibition was antibody dose-dependent (28). Additionally, antibody to CCR6 alone had no influence on background DC migration, but it did inhibit DC migration induced by suboptimal dose (10 ng/ml) of LARC (Fig. 3C, hatched bars). Therefore, CCR6 expressed by DCs was involved in the migration of DCs in response to HBDs, but the possibility that β-defensins also use other receptors cannot be ruled out.

Collectively, the results indicate that β-defensins use CCR6, at least, as a receptor. Although β-defensins show no sequence homology with LARC, they presumably have similar tertiary CCR6 binding sites and act as “microchemokines.” This is analogous to data showing that a number of viral genome-encoded proteins, some of which, although apparently sharing no sequence homologies, can nevertheless interact with chemokine receptors (29). The affinity of HBDs for CCR6, which is lower than that of LARC, may be compensated for by the availability of HBDs in higher concentrations at the sites of microbial invasion. Because DCs and T cells are important in adaptive immune responses (17), HBDs in vivo may, through their interaction with CCR6, function to recruit immature DCs and memory T cells to cutaneous or mucosal sites of microbial invasion. Therefore, we propose that β-defensins have developed the capacity to play important roles in both innate and adaptive immune responses against microbial invasion.

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


View Abstract

Navigate This Article