The skin is constantly exposed to commensal microflora and pathogenic microbes. The stratum corneum of the outermost skin layer employs distinct tools such as harsh growth conditions and numerous antimicrobial peptides (AMPs) to discriminate between beneficial cutaneous microflora and harmful bacteria. How the skin deals with microbes that have gained access to the live part of the skin as a result of microinjuries is ill defined. In this study, we report that the chemokine CXCL14 is a broad-spectrum AMP with killing activity for cutaneous Gram-positive bacteria and Candida albicans as well as the Gram-negative enterobacterium Escherichia coli. Based on two separate bacteria-killing assays, CXCL14 compares favorably with other tested AMPs, including human β-defensin and the chemokine CCL20. Increased salt concentrations and skin-typical pH conditions did not abrogate its AMP function. This novel AMP is highly abundant in the epidermis and dermis of healthy human skin but is down-modulated under conditions of inflammation and disease. We propose that CXCL14 fights bacteria at the earliest stage of infection, well before the establishment of inflammation, and thus fulfills a unique role in antimicrobial immunity.

The skin is in constant contact with the environment and is permanently challenged by a range of threats including potentially pathogenic microorganisms. Therefore, the existence of an elaborate local immune surveillance system is critical for immune defense and skin integrity. The outermost layer of healthy skin consists of a stratified epidermis that is sealed on the outside by the largely impermeable stratum corneum. In addition to this physical barrier, it has been long recognized that the skin, i.e., the epidermis and the underlying dermal tissue, harbors a variety of highly specialized immune cells, including dendritic cells (DCs),4 macrophages, mast cells, and lymphocytes, each involved in the maintenance of tissue integrity and cutaneous immunity (1).

Our recent studies of the immune surveillance system in human skin are based on the paradigm that links the control of cell migration with immune cell function (2). Tissue-specific traffic of immune surveillance cells is dependent on the constitutive expression of homeostatic chemokines (2, 3). Initial investigations led us to the chemokine receptor CCR8, which we found preferentially expressed on the majority of lymphocytes in healthy human skin (4, 5). Its selective ligand CCL1 is constitutively expressed in the dermal microvasculature and scattered cells within the epidermis. This chemokine system likely represents a mechanism controlling the traffic and/or positioning of peripheral immune surveillance T cells within healthy human skin.

By contrast to CCL1, a second chemokine, CXCL14, is much more highly expressed in healthy human skin, notably in keratinocytes and dermal fibroblasts, but is also present in other epithelial tissues such as gut and kidney (6, 7, 8). Chemoattractant activity for peripheral blood monocytes led us to propose that CXCL14 contributes to the homeostasis of cutaneous DCs, including dermal DCs and Langerhans cells, by attracting DC precursors to distinct niches within the skin where their final differentiation into bona fide APCs occurs (9). Indeed, in a tissue model with human epidermal equivalents, CXCL14-responsive monocytes were found to develop into APCs with Langerhans cell characteristics (9). This chemokine was also shown to act on immature monocyte-derived DCs, indicating an additional role in DC localization within CXCL14-producing tissues (10, 11). Mice deficient in CXCL14 do not show any gross abnormalities in the tissue distribution and function of DCs, suggesting that murine CXCL14 is not involved in the maintenance of the peripheral DC compartments, that CXCL14 is not the only chemokine involved in this process (functional redundancy), or that this defect was neutralized through compensatory mechanism(s) (12). A detailed analysis is hampered by the striking breeding defect of homozygous CXCL14-null mice, pointing to a role for CXCL14 in reproduction and/or embryogenesis. It will be important to identify the CXCL14 receptor, because this information will lead us to the physiological target cells and, thus, to the function of this ill-defined chemokine in immune processes.

The abundant expression of CXCL14 in normal human skin, notably in the epidermis, is unmatched by any other chemokine and suggested to us a chemokine-atypical function in local immune defense, which is the subject that we have investigated in the present study. In addition to the physical barrier and the repertoire of immune surveillance cells, the skin also harbors a variety of antimicrobial peptides (AMPs) that constitute the third type of weapon in the cutaneous arsenal against the constant onslaught of microbes (13). AMPs are a diverse family of small, mostly cationic polypeptides that have killing activity against a wide spectrum of bacteria, yeast, and some enveloped viruses (14, 15, 16). The major AMPs expressed in human skin are defensins, cathelicidin LL-37, and psoriasin (13, 17, 18), which are produced and released under either steady-state or inflammatory conditions by keratinocytes. Of interest, and in agreement with their role in acute infections, many AMPs also induce migration and effector responses in leukocytes (19, 20). Also, vice versa, a few chemokines were shown to have AMP activity (21, 22, 23, 24, 25, 26). However, our understanding of their involvement in antimicrobial defense is rudimentary as compared with their primary role in leukocyte traffic (3) and with the vast literature describing nonchemokine AMPs (15, 27, 28).

In this study, we describe CXCL14 as a highly active AMP with antimicrobial activity against Gram-positive and Gram-negative bacteria and also Candida albicans. Constitutive as opposed to inflammation-dependent expression supports the view that CXCL14 fulfills a critical function in the maintenance of skin integrity by keeping potentially infectious microbes at bay. This chemokine may be less relevant in fighting microbes during established infections, where a range of inducible AMPs and other innate mechanisms are mobilized.

Chemokines were chemically synthesized as described (29). Human β-defensin (HBD) 2 was from Bachem, anti-human CXCL14 and CCL20 from R&D Systems, isotype controls from BD Biosciences and Sigma-Aldrich, and anti-HBD2 was provided by T. Ganz (University of California, Los Angeles, CA). Clinical isolates of oxacillin-susceptible and -resistant strains of coagulase-negative Staphylococcus spp. (S. coag.neg. spp.) and Staphylococcus aureus, Escherichia coli, Propionibacterium spp., and C. albicans were used. Bacteria and C. albicans were grown on agar plates at 37°C in ambient air with the exception of Propionibacterium spp., which were cultured under anaerobic conditions.

Tu-138 is a human squamous cell carcinoma line (30), PamLy an immortalized BALB/c keratinocyte cell line, and SCC7 a spontaneous squamous cell carcinoma line derived from C3H mice. DNA sequences corresponding to the complete coding regions (including signal peptides) from human or murine CXCL14 were generated by RT-PCR. Human CXCL14 was subcloned into the pDNR1 vector (Clontech) containing Lox P sites and transferred to the pLP-IRESneo expression vector using Cre recombinase. Murine CXCL14 was cloned directly into the expression vector pIRESneo 3 (Clontech). Restriction site-modified primers used for PCR amplification were 5′-AGGATCCCCTCCCCATGTCCCTGC-3′ (human sense); 5′-GAAAGCTTCTATTCTTCGTAGACCCTGCG-3′ (human antisense); 5′-GTACCGGTCTCCTTGCCTCCCTGCTC-3′ (mouse sense); and 5′-CCGAATTCATCGTCCACCCTATTCTTCGTA-3′ (mouse antisense). Tu-138 cells were stably transfected with human CXCL14, whereas PamLy and SCC7 were stably transfected with murine CXCL14 using Lipofectamine 2000 (Invitrogen). Clones were selected with 400 μg/ml G418 antibiotic. CXCL14 in culture supernatants was quantified using a commercially available ELISA kit (R&D Systems).

Normal human skin samples from abdominal or mammary reduction surgery were digested with 1.25 U/ml Dispase II for 15–30 min at 37°C (Roche Diagnostics), before collecting the epidermis. The epidermis was stimulated with 500 ng/ml LPS and 10 ng/ml TNF-α for 9h in RPMI 1640 (Invitrogen). RNA was extracted using the RNAzol B method (AMS Biotechnology), purified using an RNeasy mini kit (Qiagen), and treated with DNase (Qiagen). RNA integrity was checked with a Bioanalyzer 2001 (Agilent Technologies). cRNA preparation and microarray hybridization were conducted according to the supplier using GeneChip HG-U133A (Affymetrix). Hybridized GeneChips were scanned with an Affymetrix microarray scanner 3000 and fluorescence intensity raw data were collected using GeneChip MAS 5.0. Data analysis was performed with GeneSpring 7.2 software (Silicon Genetics). Statistical comparison between stimulated and normal skin was performed using the Welch t test with log transformed data. Genes that passed this test with p < 0.05 and that showed changes in expression >2-fold were assumed to be regulated.

Skin biopsy procedure was approved by the local ethics committee and study participants gave written informed consent before obtaining skin specimens. Skin samples were frozen in Tissue-Tek OCT compound (Sakura Finetek) for the staining of CXCL14 or were formaldehyde fixed and embedded in paraffin for CCL20 and HBD2. Immunohistochemistry was performed as described (12) with the exception of a 10-min acetone fixation for CXCL14 and epitope retrieval for 40 min in boiling 1 mM EDTA (pH 8) for CCL20.

Antimicrobial activities were evaluated by the radial diffusion assay (21) or the colony-forming assay. Radial diffusion units (RDUs) were defined as follows: [diameter of clear zone in millimeters − diameter of the well] × 10. This assay was also used to determine the antimicrobial activity of cell culture supernatants from CXCL14-transfected cells cultured in antibiotic-free medium for 24 h. The relationship between the net zone diameters and the log10 of the concentration of peptide that was introduced into the well was calculated by the method of least mean squares. The minimal effective concentration (MEC) was defined as the x-intercept specified by this function (21). The colony-forming assay was performed with exponentially growing bacteria (OD620 = 0.55) in Luria broth medium. Bacteria were washed twice with 10 mM potassium phosphate buffer (pH 7.4) supplemented with 1% (v/v) Luria broth medium and then diluted to a final concentration of 1 × 105 CFU/ml in the same buffer. One hundred microliters of this bacterial suspension was incubated with peptide or culture supernatant for 3 h at 37°C under rotation (250 rpm). Growth inhibition was analyzed after plating serial dilutions of the bacterial suspensions on Luria broth agar plates and culturing for 16–18 h as described (31) and is expressed as the ratio of colonies counted to the number of colonies on a control plate [1 − (CFU after peptide incubation)/(CFU after control incubation)] × 100.

Constitutive expression in healthy human skin has been taken to suggest a role for CXCL14 in local immune homeostasis, notably in the steady-state turnover of epidermal Langerhans cells (9). However, production under homeostatic conditions does not a priori exclude a contribution of this ill-defined chemokine in inflammatory processes that may include acute infections and chronic skin diseases. To approach this question, we performed immunohistochemical analysis with healthy and diseased human skin (Fig. 1,A). As expected, CXCL14 protein was highly present in healthy human skin, notably in the epidermis and scattered cells of the dermis, but was only produced in isolated patches in skin affected by either psoriasis or atopic dermatitis (large CXCL14-negative areas are not shown in Fig. 1,A). CXCL14 expression is regulated differently than two other skin AMPs, the chemokine CCL20 and HBD2, which showed just marginal staining in healthy skin but were significantly up-regulated in psoriatic and atopic dermatitis lesions (32, 33). HBD2 expression in atopic dermatitis, however, is less prominent, and insufficient HBD2 production is believed to be the reason why these patients are more prone to infections (34). Next, we conducted a global gene expression analysis with RNA from whole skin (not shown) and isolated epidermis tissue from healthy individuals as well as TNF-α- and LPS-treated epidermis tissue to mimic acute infection (Fig. 1,B). Of note, the TNF-α/LPS treatment reduced the expression of CXCL14 by 31-fold. This is in sharp contrast to other chemokines or AMPs, which were either markedly up-regulated like CCL20 (356-fold), CXCL8 (180-fold), CXCL2 (8-fold), S100A7/psoriasin (15-fold), and HBD2 (3-fold), or remained unchanged like CXCL9, CXCL12, and HBD1 (Fig. 1 B). LL-37, another well-characterized human AMP in skin, was absent and not affected by the inflammatory stimuli, consistent with a previous report (35). Together, these results strongly argue for a function of CXCL14 in healthy human skin and underscore the distinct expression profile that separates CXCL14 from other skin chemokines and AMPs.

FIGURE 1.

CXCL14 expression is down-regulated in inflamed human skin. A, Immunohistochemical analysis of tissue sections from normal skin and skin affected by psoriasis and atopic dermatitis demonstrates differential expression of CXCL14, CCL20, and HBD2. Insets, isotype control; scale bar, 50 μm. B, Unlike many chemokines and AMPs, CXCL14 mRNA was significantly down-regulated in inflamed skin as measured by Affymetrix GeneChip analysis. mRNA levels were compared between epidermis of healthy (untreated) human skin and epidermis stimulated in vitro with TNF-α and LPS for 9 h. Expression changes of >2-fold with p < 0.05 (dotted lines) were taken as significant.

FIGURE 1.

CXCL14 expression is down-regulated in inflamed human skin. A, Immunohistochemical analysis of tissue sections from normal skin and skin affected by psoriasis and atopic dermatitis demonstrates differential expression of CXCL14, CCL20, and HBD2. Insets, isotype control; scale bar, 50 μm. B, Unlike many chemokines and AMPs, CXCL14 mRNA was significantly down-regulated in inflamed skin as measured by Affymetrix GeneChip analysis. mRNA levels were compared between epidermis of healthy (untreated) human skin and epidermis stimulated in vitro with TNF-α and LPS for 9 h. Expression changes of >2-fold with p < 0.05 (dotted lines) were taken as significant.

Close modal

CXCL14 is not only highly expressed in skin but also in other epithelial tissues with barrier functions such as gut or kidney (7). This substantial and epithelial tissue-selective expression raised the question whether CXCL14 may display antimicrobial activity and thereby contribute to the prevention of local bacterial infections. The antimicrobial effect of CXCL14 was investigated with Gram-negative and Gram-positive bacterial species including typical commensals of the skin and pathogens colonizing the skin and the mucosa as well as the fungus C. albicans by two distinct methods, the radial diffusion assay measuring the distance of the clearing zone from the antibiotic deposition to the edge of bacterial growth (Fig. 2,A) and the colony-forming assay measuring growth inhibition in liquid culture as a function of antibiotic concentration (Fig. 2,B). The selected microbes are generally benign but become pathogenic as a result of epithelial injury. CXCL14 exhibited strong antimicrobial activity toward Gram-negative E. coli, Gram-positive species S. coag.neg. spp., Gram-positive S. aureus, and Gram-positive Propionibacterium spp., as well as the fungus C. albicans. We quantified and compared the potency of CXCL14 by determining MECs that were calculated by least mean squares regression at the x-intercept of the relationship between zone diameters vs log10 peptide concentration (21) (Fig. 2,A and Table I). The antimicrobial activity of CXCL14 showed a certain degree of species variation, being most effective against E. coli, C. albicans, and S. coag.neg. spp. (MECs of 0.32, 0.56, and 1.02 μM, respectively). We next included inducible skin peptides such as HBD2, CCL20, and CCL27 and the acute phase chemokine CXCL8 in the radial diffusion assay to gauge the level of antimicrobial activity obtained with CXCL14. CCL20 had similar antimicrobial properties as CXCL14 whereas CCL27 and CXCL8 were considerably less potent (against E. coli,) or even inactive (against Gram-positive bacteria and C. albicans). HBD2 only inhibited growth of E. coli and S. coag.neg. spp. The colony-forming assay confirmed these findings by showing strong growth inhibition of CXCL14 and CCL20 against E. coli and S. coag.neg. spp. (Fig. 2,B). HBD2 was bacteriostatic against E. coli, but not against S. coag.neg. spp. (Fig. 2 B; data not shown).

FIGURE 2.

CXCL14 is a strong and broad-spectrum antimicrobial peptide against Gram-positive and Gram-negative bacteria. A, Antimicrobial activity of CXCL14 was compared with selected chemokines and HBD2 against Gram-negative E. coli, Gram-positive S. coag.neg. spp., Gram-positive S. aureus, Gram-positive Propionibacterium spp., and the yeast C. albicans. Data from two to four independent experiments (±SD) represent inhibition of microbial growth. The MEC corresponds to the x-intercept of a linear least mean square regression of the relationship between zone diameters vs log10 peptide concentration. Peptide refers to the type of AMP tested (see legend to right of E. coli graph). B, Bacteriostatic capacity of CXCL14, compared with CCL20 and HBD2, was analyzed with the colony-forming assay against E. coli and S. coag.neg. spp. The percentage of growth inhibition is expressed as the ratio of colonies counted in the presence of a peptide to the number of colonies on a control plate as described in Materials and Methods (±SD of triplicate experiments).

FIGURE 2.

CXCL14 is a strong and broad-spectrum antimicrobial peptide against Gram-positive and Gram-negative bacteria. A, Antimicrobial activity of CXCL14 was compared with selected chemokines and HBD2 against Gram-negative E. coli, Gram-positive S. coag.neg. spp., Gram-positive S. aureus, Gram-positive Propionibacterium spp., and the yeast C. albicans. Data from two to four independent experiments (±SD) represent inhibition of microbial growth. The MEC corresponds to the x-intercept of a linear least mean square regression of the relationship between zone diameters vs log10 peptide concentration. Peptide refers to the type of AMP tested (see legend to right of E. coli graph). B, Bacteriostatic capacity of CXCL14, compared with CCL20 and HBD2, was analyzed with the colony-forming assay against E. coli and S. coag.neg. spp. The percentage of growth inhibition is expressed as the ratio of colonies counted in the presence of a peptide to the number of colonies on a control plate as described in Materials and Methods (±SD of triplicate experiments).

Close modal
Table I.

CXCL14 is a potent AMP

MECsa (μM)
E. coliS. coag.neg. sppPropionibacterium spp.S. aureusC. albicans
CXCL14 0.32 1.02 1.16 5.17 0.56 
CCL20 0.46 1.58 5.23 5.54 0.09 
CCL27 0.67 >80 >80 >80 >80 
CXCL8 9.9 >150 >150 >150 >150 
HBD2 2.54 5.83    
MECsa (μM)
E. coliS. coag.neg. sppPropionibacterium spp.S. aureusC. albicans
CXCL14 0.32 1.02 1.16 5.17 0.56 
CCL20 0.46 1.58 5.23 5.54 0.09 
CCL27 0.67 >80 >80 >80 >80 
CXCL8 9.9 >150 >150 >150 >150 
HBD2 2.54 5.83    
a

MEC was calculated by least mean squares regression at the x-intercept of the relationship between zone diameters vs log10 peptide concentration.

Next, we examined the effect of CXCL14 on the growth of primary isolates of both oxacillin-resistant and oxacillin-sensitive strains of S. coag.neg. spp. and S. aureus. Radial diffusion assays using two different chemokine concentrations (15 and 30 μM) demonstrated that growth inhibition by CXCL14 was largely independent of the state of antibiotic resistance (Fig. 3). We noted a slightly lower (although statistically insignificant) sensitivity in oxacillin-resistant isolates. Collectively, our findings identify CXCL14 as a highly potent AMP with broad-spectrum activity that includes clinical isolates of Gram-negative and Gram-positive bacteria.

FIGURE 3.

Antimicrobial activity of CXCL14 against oxacillin-susceptible and oxacillin-resistant S. coag.neg. spp. and S. aureus. The antimicrobial activity of 30 μM (filled column) and 15 μM (open column) CXCL14 against oxacillin-sensitive (s1–s5) and oxacillin-resistant (r1–r5) strains of S. coag.neg. spp. (left panel) and S. aureus isolates (right panel) was tested by the RDU assay. Representative results for each bacterial strain of two experiments are shown.

FIGURE 3.

Antimicrobial activity of CXCL14 against oxacillin-susceptible and oxacillin-resistant S. coag.neg. spp. and S. aureus. The antimicrobial activity of 30 μM (filled column) and 15 μM (open column) CXCL14 against oxacillin-sensitive (s1–s5) and oxacillin-resistant (r1–r5) strains of S. coag.neg. spp. (left panel) and S. aureus isolates (right panel) was tested by the RDU assay. Representative results for each bacterial strain of two experiments are shown.

Close modal

Human skin harbors a slightly acidic pH (<7.0) and a variable salt milieu (36). The antimicrobial activity of CCL20 (23, 25) and HBD2 (37) is greatly reduced or absent at physiological salt concentrations (150 mM NaCl). We then compared the salt sensitivity of CXCL14 with CCL20 and HBD2 at increased NaCl concentrations but at levels where antimicrobial activity was not completely abrogated. Using the CFU assay as described for Fig. 2,B, our results show that CXCL14 inhibited the growth of E. coli by 65% at 100 mM NaCl (Fig. 4). Growth of S. coag.neg. spp. was less inhibited, but still 70 and 45% inhibition were seen at 50 and 100 mM NaCl, respectively. These experiments show that CXCL14 is less salt sensitive than CCL20 and HBD2. We also tested the inhibition of E. coli growth by CXCL14, CCL20, and HBD2 at pH 5.5, but no significant effect was observed (data not shown).

FIGURE 4.

Robust antimicrobial activity of CXCL14 in increased sodium chloride concentrations. Bacteriostatic activities of 0.53 μM CXCL14, 0.62 μM CCL20, and 2.33 μM HBD2 (corresponding to 5 μg/ml CXCL14 and CCL20 or 10 μg/ml HBD2) in the presence of 10, 50, and 100 mM sodium chloride were determined with the colony-forming assay as described in Materials and Methods. ∗, p < 0.05 by Student’s t test.

FIGURE 4.

Robust antimicrobial activity of CXCL14 in increased sodium chloride concentrations. Bacteriostatic activities of 0.53 μM CXCL14, 0.62 μM CCL20, and 2.33 μM HBD2 (corresponding to 5 μg/ml CXCL14 and CCL20 or 10 μg/ml HBD2) in the presence of 10, 50, and 100 mM sodium chloride were determined with the colony-forming assay as described in Materials and Methods. ∗, p < 0.05 by Student’s t test.

Close modal

Thus far, all antimicrobial experiments were conducted with synthetic CXCL14. To clarify whether natural, cell-derived CXCL14 also had AMP activity, we investigated the culture supernatants of squamous cell lines that were transfected with CXCL14-encoding plasmids. Although keratinocytes in human skin are a rich source of CXCL14 (Fig. 1,A), production of this chemokine in primary cultures of keratinocytes (or related cell lines) is moderate to nondetectable (6, 7, 8, 9). A CXCL14-transfected human squamous carcinoma-derived cell line (Tu-138) secreted 4–5 ng/ml CXCL14 into the culture medium, whereas this chemokine was below detection in supernatants of parental cells. Of note, culture supernatants of CXCL14-transfected Tu-138 cells displayed strong antimicrobial activity against E. coli and S. coag.neg. spp., whereas supernatants from control cultures (parental cells or vector-only transfected cells) were considerably less active (Fig. 5,A). The residual activity was probably due to alternative AMPs. To assign the observed inhibition of bacterial growth to CXCL14, we included CXCL14 neutralizing and control Abs in our assays. In control experiments with synthetic CXCL14, addition of anti-CXCL14 Abs reduced the antimicrobial activity by 88%, whereas the control Ab had no effect (Fig. 5,B). Importantly, the same anti-CXCL14 treatment reduced RDU values in culture supernatants of CXCL14-transfected Tu-138 cells to the level of the background activity seen with supernatants of control cells (Fig. 5,C). Due to considerable variation in the RDU values obtained with cell culture supernatants from different experiments, the antimicrobial activity was expressed as the percentage of reduction in the RDU values in the presence of anti-CXCL14 or control Abs as compared with untreated cell culture supernatants (100% antimicrobial activity). Similar results were obtained with CXCL14 DNA-transfected murine squamous carcinoma (SCC7) or immortalized BALB/c keratinocytes (PamLy) cell lines (Fig. 6). This is not surprising, because human and murine CXCL14 display very close structural homology (12). Collectively, our data demonstrate that native CXCL14 works as well as synthetic CXCL14 in growth inhibition of bacteria. Resistance to neutralization by several factors, including cell-associated and serum proteases, chemokine-binding proteins, and exposure to prolonged tissue culture suggest that natural CXCL14 in healthy skin may exist long enough to provide protection against infection by local microbes under steady-state conditions.

FIGURE 5.

CXCL14 in culture supernatants of epidermal cell lines exhibits strong antimicrobial activity. A, Culture supernatants of CXCL14-transfected Tu138 cells reveal strong antimicrobial activity against E. coli (filled columns) and S. coag.neg. spp. (open columns) as determined by RDU assay. Data ± SEM of triplicate experiments; ∗∗∗, p < 0.001). B, Antimicrobial activity of synthetic CXCL14 (5 μM) for E. coli was neutralized by anti-CXCL14 Ab (Mab866) but not by IgG isotype control Ab. C, The antimicrobial activity of diluted culture supernatants of CXCL14-transfected Tu138 cells was mainly due to CXCL14 as assessed for E. coli (filled columns) and S. coag.neg. spp. (open columns) with either Mab866 or control Ab. Respective antimicrobial activity of synthetic CXCL14 (B) and CXCL14-Tu138 culture supernatants (C) was taken as 100%, and inhibition by anti-CXCL14 or IgG isotype Abs was calculated using the log-linear relationship between zone diameter and peptide concentration (B) or the difference between zone diameters (C) (mean percentage ± SEM of triplicate experiments). ∗, p < 0.05; ∗∗∗, p < 0.001; revealed by Student’s t test.

FIGURE 5.

CXCL14 in culture supernatants of epidermal cell lines exhibits strong antimicrobial activity. A, Culture supernatants of CXCL14-transfected Tu138 cells reveal strong antimicrobial activity against E. coli (filled columns) and S. coag.neg. spp. (open columns) as determined by RDU assay. Data ± SEM of triplicate experiments; ∗∗∗, p < 0.001). B, Antimicrobial activity of synthetic CXCL14 (5 μM) for E. coli was neutralized by anti-CXCL14 Ab (Mab866) but not by IgG isotype control Ab. C, The antimicrobial activity of diluted culture supernatants of CXCL14-transfected Tu138 cells was mainly due to CXCL14 as assessed for E. coli (filled columns) and S. coag.neg. spp. (open columns) with either Mab866 or control Ab. Respective antimicrobial activity of synthetic CXCL14 (B) and CXCL14-Tu138 culture supernatants (C) was taken as 100%, and inhibition by anti-CXCL14 or IgG isotype Abs was calculated using the log-linear relationship between zone diameter and peptide concentration (B) or the difference between zone diameters (C) (mean percentage ± SEM of triplicate experiments). ∗, p < 0.05; ∗∗∗, p < 0.001; revealed by Student’s t test.

Close modal
FIGURE 6.

Potent antimicrobial activity of CXCL14 produced by murine epidermis-derived cell lines. A, The antimicrobial activity of culture supernatants from the murine CXCL14-transfected or the vector-transfected parental BALB/c keratinocyte cell line PamLy were tested in the RDU assay (left panel). Data are representative of three experiments. The contribution of secreted CXCL14 to the antimicrobial activity was determined by the addition of either anti-CXCL14 Ab (Mab866) or isotype control Ab (right panel); antimicrobial activity of diluted CXCL14-transfected PamLy culture supernatants was assumed as 100% and inhibition by Ab or IgG isotype control calculated as described in Fig. 5 C. A representative experiment is shown. ***, p < 0.001. B, Exactly as in A except that culture supernatants of the murine CXCL14-transfected squamous carcinoma cell line SCC7 were examined. *, p < 0.05.

FIGURE 6.

Potent antimicrobial activity of CXCL14 produced by murine epidermis-derived cell lines. A, The antimicrobial activity of culture supernatants from the murine CXCL14-transfected or the vector-transfected parental BALB/c keratinocyte cell line PamLy were tested in the RDU assay (left panel). Data are representative of three experiments. The contribution of secreted CXCL14 to the antimicrobial activity was determined by the addition of either anti-CXCL14 Ab (Mab866) or isotype control Ab (right panel); antimicrobial activity of diluted CXCL14-transfected PamLy culture supernatants was assumed as 100% and inhibition by Ab or IgG isotype control calculated as described in Fig. 5 C. A representative experiment is shown. ***, p < 0.001. B, Exactly as in A except that culture supernatants of the murine CXCL14-transfected squamous carcinoma cell line SCC7 were examined. *, p < 0.05.

Close modal

Our study provides compelling evidence of a role for CXCL14 in antimicrobial immunity. This conclusion is drawn from data showing that synthetic and cell culture supernatant-derived CXCL14 is highly potent in inhibiting the growth of a variety of microbes that include species typically associated with the human skin microflora such as Staphylococci species, Propionibacteria, and the fungus C. albicans as well as the prototype enterobacterium E. coli. Antimicrobial activity of CXCL14 has been reported in a previous publication (23) where numerous chemokines including CXCL14 were screened for antimicrobial activity against E. coli. The reported activity was moderate and was not further examined in that study. In the present study we report that the antimicrobial potency of CXCL14 compares favorably to that of known AMPs such as HBD2 or the chemokine CCL20. We further show that CXCL14 is a constitutive chemokine with exceptionally high-level expression in healthy human skin, an organ that is constantly exposed to environmental microbes (18). In fact, this abundance allowed us to isolate the natural form of CXCL14 and to confirm that its structure and function are identical with those of the synthetic form we routinely use in our experiments (9). We also find that the expression of CXCL14 mRNA and protein within human skin is substantially inhibited by inflammatory stimuli. This finding is remarkable because inflamed skin (like other diseased tissues) is a rich source of both chemokines that participate in controlling the recruitment and localization of immune cells (3) and AMPs that support the innate immune system in its fight against infectious microbes. Inhibition of chemokine expression by inflammatory stimuli has not been reported for any other chemokine and points to a role for CXCL14 in homeostatic as opposed to inflammation-driven immune processes and may include bacterial killing at early stages of infection (before the establishment of inflammatory conditions; see below).

Previously characterized chemokines with bactericidal activity include CCL20, which is abundantly expressed in inflamed human epidermis and has broad-spectrum antimicrobial activity for both Gram-positive and Gram-negative bacteria (23, 25). Interestingly, CCL20 shares with β-defensins its selectivity for CCR6 (20). However, it is not clear how the targeting of CCR6+ T cells and CCR6+ DCs is linked with the AMP activity of the CCR6 ligand, especially so because a preference for CCR6+ immune cells in response to infections has not been demonstrated. Many AMP-type chemokines require extremely high concentrations to kill bacteria in in vitro assays. These concentrations greatly exceed their optimal concentrations for the induction of leukocyte recruitment and may not be generated in peripheral tissues, thus arguing against a major role in bacterial killing. Collectively, currently known chemokines with AMP activity are formidable chemoattractants for target cells irrespective of their relevance to infections and are produced under inflammatory settings, i.e., under conditions where the sites of infections are dominated by an inflammatory infiltrate.

Defensins are prominent AMPs in human skin and include HBD1, which is constitutively expressed, and HBD2 and HBD3, which are absent in healthy skin but highly up-regulated under inflammatory conditions (18). Chemokines display structural similarities with defensins, including abundance of cationic residues, intramolecular disulfide bonds, and tertiary structure. A prominent feature of chemokines and defensins is the formation of large, positively charged patches on the surface of the molecule, and it is suggested that the positive charges interact with negatively charged bacterial membrane components such as lipopolysaccharide or teichoic acid, leading to permeabilization and subsequent death of the bacteria (23, 28). CXCL14 is a highly cationic protein with an estimated isoelectric point (pI) of 9.9 and a net charge of +13 at pH 7 and displays an amphipathic character, thereby revealing all of the important physicochemical properties of an AMP. However, the most abundant AMP in human skin is psoriasin (38), also known as Ca2+-binding protein S100A7. Keratinocytes in the outermost layer of human skin, in proximity to the stratum corneum and apical parts of hair follicles, are responsible for the strikingly focal expression of this AMP, a characteristic fitting with a role in skin homeostasis as opposed to infection (see below). The killing activity of psoriasin is attributed in part to its ability to sequester Zn2+, one of several cofactors that bacteria require for combating oxidative stress. Two other abundant AMPs in the stratum corneum, which are proposed to control the local microflora under homeostatic conditions, are RNase 7 and lysozyme (18). Our data support a model that features CXCL14 as a novel AMP with a distinct function at an early stage of infection (Fig. 7). The microflora in the outermost layer of human skin is controlled by several factors, including a physical barrier made up by the corneocytes and extracellular matrix of the stratum corneum, harsh physiochemical conditions (acidic pH and salt and temperature fluctuations), and water deprivation as well as antimicrobial proteins (18, 39). The combination of these factors ensures the maintenance of a healthy microflora and, at the same time, provides a defense against skin-extraneous microbes. For instance, psoriasin does not harm the numerous Gram-positive bacterial species present in healthy human skin but instead is highly efficient in killing the enterobacterium E. coli (38). In this regard, the antimicrobial “climate” in healthy human skin supports the local microflora by favoring skin-typical bacteria and yeast over potentially pathogenic species. Similar to the commensals in the gut microflora, the host may benefit from the cutaneous microflora due to its secretion of antibiotics and factors supporting tissue regeneration and control of inflammation (40). CXCL14 is abundantly expressed in the epidermis and dermis of healthy human skin (9); however, this chemokine is not described in skin wash fluids and extracts from stratum corneum tissue samples (38) and, thus, is unlikely to play a major role in the homeostasis of the skin microflora. Instead, we propose that CXCL14 fulfills its antimicrobial function in the very first moment of skin injury when microbes pass the stratum corneum and reach the live part of the epidermis (or the upper dermis) (Fig. 7). Such microinjuries may result from frequent skin abrasions and superficial cuts that generally remain unnoticed. Due to the location in the skin and the broad-range AMP activity, CXCL14 is an ideal candidate for immediate involvement in antimicrobial defense against cutaneous and foreign microbes. In this role, CXCL14 may be supported by HBD1, which is also produced in the skin under steady-state conditions (15). HBD1 is also active against E. coli but, in contrast to CXCL14, appears to be less effective against members of the cutaneous microflora that are also known to cause disease upon tissue penetration (40). We propose that CXCL14 kills bacteria before they manage to establish macroscopic foci of infection and inflammation. In fact, the expression of CXCL14 is drastically reduced under inflammatory conditions, ruling out a role for this AMP in late-stage infections characterized by the participation of innate immune cells, including neutrophils and monocytes (Fig. 5). This stage features the production of a new wave of inducible AMPs, including several defensins, cathelicidins, and chemokines that synergize in the killing of local microbes and in the recruitment of antimicrobial immune cells. In our model, CXCL14 plays a unique part in the antimicrobial defense by acting on microbes at the site of entry well before the mobilization of inflammatory cells and mediators. Finally, our model may not be restricted to the skin, because CXCL14 is also present in human intestinal and colonic tissue (7, 8). It will be essential to investigate the mechanism(s) by which CXCL14 is inhibiting bacterial growth. Such studies may inform about the feasibility of transforming this information into much needed alternative antibiotics.

FIGURE 7.

Model of CXCL14 participation in acute skin infection. CXCL14 is constitutively expressed by keratinocytes (K) in the epidermis and fibroblasts (F) in the dermis. CXCL14 is not found in the healthy stratum corneum (SC or S. corneum) and, therefore, is not involved in the homeostasis of the commensal skin microflora. Instead, this model predicts that CXCL14 kills bacteria that have penetrated into the live part of the skin, for instance as a result of microinjuries. At late stages of infections, after an inflammatory milieu has been established, antimicrobial defense is taken over by inducible AMPs, chemokines, and recruited effector cells. Mφ, Macrophage; T, T cell; N, neutrophil.

FIGURE 7.

Model of CXCL14 participation in acute skin infection. CXCL14 is constitutively expressed by keratinocytes (K) in the epidermis and fibroblasts (F) in the dermis. CXCL14 is not found in the healthy stratum corneum (SC or S. corneum) and, therefore, is not involved in the homeostasis of the commensal skin microflora. Instead, this model predicts that CXCL14 kills bacteria that have penetrated into the live part of the skin, for instance as a result of microinjuries. At late stages of infections, after an inflammatory milieu has been established, antimicrobial defense is taken over by inducible AMPs, chemokines, and recruited effector cells. Mφ, Macrophage; T, T cell; N, neutrophil.

Close modal

We thank Ursula Ackermann for helping with the microbe cultures and Tomas Ganz for providing anti-HBD2 Abs.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by Staatssekretariat für Bildung und Forschung Grant 03.04412 and European Framework Programme No. 6 Grant 518167.

4

Abbreviations used in this paper: DC, dendritic cell; AMP, antimicrobial peptide; HBD, human β-defensin; MEC, minimal effective concentration; RDU, radial diffusion unit; S. coag.neg. spp., coagulase-negative Staphylococcus spp.

1
Bos, J. D., M. L. Kapsenberg.
1993
. The skin immune system: progress in cutaneous biology.
Immunol. Today
14
:
75
-78.
2
Schaerli, P., B. Moser.
2005
. Chemokines: control of primary and memory T-cell traffic.
Immunol. Res.
31
:
57
-74.
3
Moser, B., M. Wolf, A. Walz, P. Loetscher.
2004
. Chemokines: multiple levels of leukocyte migration control.
Trends Immunol.
25
:
75
-84.
4
Schaerli, P., L. Ebert, K. Willimann, A. Blaser, R. S. Roos, P. Loetscher, B. Moser.
2004
. A skin-selective homing mechanism for human immune surveillance T cells.
J. Exp. Med.
199
:
1265
-1275.
5
Ebert, L. M., S. Meuter, B. Moser.
2006
. Homing and function of human skin γδ T cells and NK cells: relevance for tumor surveillance.
J. Immunol.
176
:
4331
-4336.
6
Sleeman, M. A., J. K. Fraser, J. G. Murison, S. L. Kelly, R. L. Prestidge, D. J. Palmer, J. D. Watson, K. D. Kumble.
2000
. B cell- and monocyte-activating chemokine (BMAC), a novel non-ELR α-chemokine.
Int. Immunol.
12
:
677
-689.
7
Kurth, I., K. Willimann, P. Schaerli, T. Hunziker, I. Clark-Lewis, B. Moser.
2001
. Monocyte selectivity and tissue localization suggests a role for breast and kidney-expressed chemokine (BRAK) in macrophage development.
J. Exp. Med.
194
:
855
-861.
8
Frederick, M. J., Y. Henderson, X. Xu, M. T. Deavers, A. A. Sahin, H. Wu, D. E. Lewis, A. K. El-Naggar, G. L. Clayman.
2000
. In vivo expression of the novel CXC chemokine BRAK in normal and cancerous human tissue.
Am. J. Pathol.
156
:
1937
-1950.
9
Schaerli, P., K. Willimann, L. M. Ebert, A. Walz, B. Moser.
2005
. Cutaneous CXCL14 targets blood precursors to epidermal niches for Langerhans cell differentiation.
Immunity
23
:
331
-342.
10
Shellenberger, T. D., M. Wang, M. Gujrati, A. Jayakumar, R. M. Strieter, M. D. Burdick, C. G. Ioannides, C. L. Efferson, A. K. El-Naggar, D. Roberts, et al
2004
. BRAK/CXCL14 is a potent inhibitor of angiogenesis and a chemotactic factor for immature dendritic cells.
Cancer Res.
64
:
8262
-8270.
11
Shurin, G. V., R. L. Ferris, I. L. Tourkova, L. Perez, A. Lokshin, L. Balkir, B. Collins, G. S. Chatta, M. R. Shurin.
2005
. Loss of new chemokine CXCL14 in tumor tissue is associated with low infiltration by dendritic cells (DC), while restoration of human CXCL14 expression in tumor cells causes attraction of DC both in vitro and in vivo.
J. Immunol.
174
:
5490
-5498.
12
Meuter, S., P. Schaerli, R. S. Roos, O. Brandau, M. R. Bosl, U. H. von Andrian, B. Moser.
2007
. Murine CXCL14 is dispensable for dendritic cell function and localization within peripheral tissues.
Mol. Cell. Biol.
27
:
983
-992.
13
Braff, M. H., A. Bardan, V. Nizet, R. L. Gallo.
2005
. Cutaneous defense mechanisms by antimicrobial peptides.
J. Invest. Dermatol.
125
:
9
-13.
14
Brown, K. L., R. E. Hancock.
2006
. Cationic host defense (antimicrobial) peptides.
Curr. Opin. Immunol.
18
:
24
-30.
15
Ganz, T..
2003
. Defensins: antimicrobial peptides of innate immunity.
Nat. Rev. Immunol.
3
:
710
-720.
16
Izadpanah, A., R. L. Gallo.
2005
. Antimicrobial peptides.
J. Am. Acad. Dermatol.
52
:
381
-390.
17
Niyonsaba, F., H. Ogawa.
2005
. Protective roles of the skin against infection: implication of naturally occurring human antimicrobial agents β-defensins, cathelicidin LL-37 and lysozyme.
J. Dermatol. Sci.
40
:
157
-168.
18
Schroder, J. M., J. Harder.
2006
. Antimicrobial skin peptides and proteins.
Cell Mol. Life Sci.
63
:
469
-486.
19
Yang, D., A. Biragyn, L. W. Kwak, J. J. Oppenheim.
2002
. Mammalian defensins in immunity: more than just microbicidal.
Trends Immunol.
23
:
291
-296.
20
Hoover, D. M., C. Boulegue, D. Yang, J. J. Oppenheim, K. Tucker, W. Lu, J. Lubkowski.
2002
. The structure of human macrophage inflammatory protein-3α /CCL20. Linking antimicrobial and CC chemokine receptor-6-binding activities with human β-defensins.
J. Biol. Chem.
277
:
37647
-37654.
21
Cole, A. M., T. Ganz, A. M. Liese, M. D. Burdick, L. Liu, R. M. Strieter.
2001
. Cutting edge: IFN-inducible ELR- CXC chemokines display defensin-like antimicrobial activity.
J. Immunol.
167
:
623
-627.
22
Krijgsveld, J., S. A. Zaat, J. Meeldijk, P. A. van Veelen, G. Fang, B. Poolman, E. Brandt, J. E. Ehlert, A. J. Kuijpers, G. H. Engbers, et al
2000
. Thrombocidins, microbicidal proteins from human blood platelets, are C-terminal deletion products of CXC chemokines.
J. Biol. Chem.
275
:
20374
-20381.
23
Yang, D., Q. Chen, D. M. Hoover, P. Staley, K. D. Tucker, J. Lubkowski, J. J. Oppenheim.
2003
. Many chemokines including CCL20/MIP-3α display antimicrobial activity.
J. Leukocyte Biol.
74
:
448
-455.
24
Hieshima, K., H. Ohtani, M. Shibano, D. Izawa, T. Nakayama, Y. Kawasaki, F. Shiba, M. Shiota, F. Katou, T. Saito, O. Yoshie.
2003
. CCL28 has dual roles in mucosal immunity as a chemokine with broad-spectrum antimicrobial activity.
J. Immunol.
170
:
1452
-1461.
25
Starner, T. D., C. K. Barker, H. P. Jia, Y. Kang, P. B. McCray, Jr.
2003
. CCL20 is an inducible product of human airway epithelia with innate immune properties.
Am. J. Respir. Cell Mol. Biol.
29
:
627
-633.
26
Linge, H. M., M. Collin, P. Nordenfelt, M. Morgelin, M. Malmsten, A. Egesten.
2008
. The human CXC-chemokine granulocyte chemotactic protein 2 (GCP-2)/CXCL6 possesses membrane disrupting properties and is antibacterial.
Antimicrob. Agents Chemother.
52
:
2599
-2607.
27
Gallo, R. L., M. Murakami, T. Ohtake, M. Zaiou.
2002
. Biology and clinical relevance of naturally occurring antimicrobial peptides.
J. Allergy Clin. Immunol.
110
:
823
-831.
28
Zasloff, M..
2002
. Antimicrobial peptides of multicellular organisms.
Nature
415
:
389
-395.
29
Clark-Lewis, I., L. Vo, P. Owen, J. Anderson.
1997
. Chemical synthesis, purification, and folding of C-X-C and C-C chemokines.
Methods Enzymol.
287
:
233
-250.
30
Liu, T. J., W. W. Zhang, D. L. Taylor, J. A. Roth, H. Goepfert, G. L. Clayman.
1994
. Growth suppression of human head and neck cancer cells by the introduction of a wild-type p53 gene via a recombinant adenovirus.
Cancer Res.
54
:
3662
-3667.
31
Wu, Z., D. M. Hoover, D. Yang, C. Boulegue, F. Santamaria, J. J. Oppenheim, J. Lubkowski, W. Lu.
2003
. Engineering disulfide bridges to dissect antimicrobial and chemotactic activities of human β-defensin 3.
Proc. Natl. Acad. Sci. USA
100
:
8880
-8885.
32
Ali, R. S., A. Falconer, M. Ikram, C. E. Bissett, R. Cerio, A. G. Quinn.
2001
. Expression of the peptide antibiotics human β defensin-1 and human β defensin-2 in normal human skin.
J. Invest. Dermatol.
117
:
106
-111.
33
Nakayama, T., R. Fujisawa, H. Yamada, T. Horikawa, H. Kawasaki, K. Hieshima, D. Izawa, S. Fujiie, T. Tezuka, O. Yoshie.
2001
. Inducible expression of a CC chemokine liver- and activation-regulated chemokine (LARC)/macrophage inflammatory protein (MIP)-3 α/CCL20 by epidermal keratinocytes and its role in atopic dermatitis.
Int. Immunol.
13
:
95
-103.
34
Ong, P. Y., T. Ohtake, C. Brandt, I. Strickland, M. Boguniewicz, T. Ganz, R. L. Gallo, D. Y. Leung.
2002
. Endogenous antimicrobial peptides and skin infections in atopic dermatitis.
N. Engl. J. Med.
347
:
1151
-1160.
35
Schauber, J., R. A. Dorschner, K. Yamasaki, B. Brouha, R. L. Gallo.
2006
. Control of the innate epithelial antimicrobial response is cell-type specific and dependent on relevant microenvironmental stimuli.
Immunology
118
:
509
-519.
36
Schittek, B., R. Hipfel, B. Sauer, J. Bauer, H. Kalbacher, S. Stevanovic, M. Schirle, K. Schroeder, N. Blin, F. Meier, et al
2001
. Dermcidin: a novel human antibiotic peptide secreted by sweat glands.
Nat. Immunol.
2
:
1133
-1137.
37
Singh, P. K., H. P. Jia, K. Wiles, J. Hesselberth, L. Liu, B. A. Conway, E. P. Greenberg, E. V. Valore, M. J. Welsh, T. Ganz, et al
1998
. Production of β-defensins by human airway epithelia.
Proc. Natl. Acad. Sci. USA
95
:
14961
-14966.
38
Glaser, R., J. Harder, H. Lange, J. Bartels, E. Christophers, J. M. Schroder.
2005
. Antimicrobial psoriasin (S100A7) protects human skin from Escherichia coli infection.
Nat. Immunol.
6
:
57
-64.
39
Elias, P. M..
2007
. The skin barrier as an innate immune element.
Semin. Immunopathol.
29
:
3
-14.
40
Cogen, A. L., V. Nizet, R. L. Gallo.
2008
. Skin microbiota: a source of disease or defense?.
Br. J. Dermatol.
158
:
442
-455.