Original Article

Introduction

Keratinocytes, the main constituent of the epidermis, not only form a passive barrier between external environment and internal organs, but upon external stimuli, such as trauma, bacterial and viral infections, chemical substances, or UV irradiation, they produce various cytokines and chemokines (Uchi et al., 2000; Gröne, 2002). Therefore, keratinocytes may represent the first line of defense against pathogens in the skin.

The innate immune system has evolved to recognize a broad spectrum of pathogens and is often extremely successful as a first line of defense. Innate immune cells also provide specific immune cells with information required for a most effective second line of immunity, which is indispensable when innate immunity fails (Medzhitov, 2001). Recognition of pathogens by innate immune cells is mediated by pattern-recognition receptors that recognize conserved pathogen-associated molecular patterns (PAMPs). One major group of pattern-recognition receptors is formed by the Toll-like receptors (TLRs) (Akira et al., 2001; Medzhitov, 2001), which transduce signals leading to the activation of NF-B (Muzio et al., 1998) that subsequently drive the transcriptional induction of several cytokine, chemokine, and adhesion molecule genes (Siebenlist et al., 1994). TLR2 either forms a heterodimer with TLR1 or TLR6 resulting in the distinction between bacterial- and mycoplasma-derived lypoproteins (Brightbill et al., 1999; Takeuchi et al., 2000). TLR3 recognizes viral double-stranded RNA (Alexopoulou et al., 2001), whereas single-stranded RNA is recognized by TLR7 and TLR8 (Diebold et al., 2004; Heil et al., 2004). The receptor for lipopolysaccharide (LPS) is TLR4 (Tapping et al., 2000) and bacterial flagellin is recognized by TLR5 (Hayashi et al., 2001). TLR9 mediate signals from unmethylated CpG motifs present in bacterial DNA (Hemmi et al., 2000). The ligand that is recognized by TLR10 is still unknown (Chuang and Ulevitch, 2001) and TLR11 recognizes uropathogenic bacteria (Zhang et al., 2004). This wide variety of pattern-recognition receptors suggest that the discrimination of the pathogen by TLRs, and the subsequent production of a specific subset of cytokines and chemokines, may be the first point at which the immune system tailors its response to different classes of pathogens.

Not only the secretion of cytokines but also chemokines are of importance for the adequate immune response to different pathogens. Numerous in vitro studies have documented the increased expression of mRNA and/or protein for several chemokines by human keratinocytes (Uchi et al., 2000; Gröne, 2002). In particular, keratinocytes produce in response to proinflammatory cytokines, such as IFN- and tumor necrosis factor (TNF)-, the CXC chemokines IFN-inducible protein-10 (CXCL10) (Boorsma et al., 1999), monokine-induced by IFN- (CXCL9) (Albanesi et al., 2000), IFN-inducible T-cell -chemoattractant (CXCL11) (Tensen et al., 1999), IL-8 (CXCL8) (Li et al., 1996), the CC chemokines monocyte chemoattractant protein-1 (CCL2) (Barker et al., 1990; Nakamura et al., 1995), macrophage inflammatory protein-3 (CCL20) (Charbonnier et al., 1999), and cutaneous T-cell-attracting chemokine (CCL27) (Morales et al., 1999). CXCL9/monokine-induced by IFN-, CXCL10/IFN-inducible protein-10, and CXCL11/IFN-inducible T-cell -chemoattractant specifically attract activated CXCR3⁺ T-helper type 1 (Th1) cells (Bonecchi et al., 1998) and CXCL8/IL-8 selectively attracts CXCR1⁺2⁺ neutrophils (Baggiolini et al., 1995). CCL2/monocyte chemoattractant protein-1 attracts monocytes (Matsushima et al., 1989), dendritic cells (DCs) and Langerhans cells (Nakamura et al., 1995), memory T cells (Carr et al., 1994), and natural killer cells (Allavena et al., 1994) via CCR2 (Boring et al., 1996). CCL20/macrophage inflammatory protein-3 attracts both memory T cells (Liao et al., 1999) and Langerhans cells into the skin (Charbonnier et al., 1999) via CCR6 receptor. CCL27/cutaneous T-cell-attracting chemokine is exclusively and constitutively expressed by human keratinocytes and binds to the receptor CCR10. This chemokine preferentially attracts skin-homing cutaneous lymphocyte-associated antigen (CLA⁺CCR10⁺) memory T cells (Morales et al., 1999; Homey et al., 2000).

Although it has long been recognized that keratinocytes are the main producers of cytokines and chemokines in the skin (Uchi et al., 2000; Gröne, 2002), the knowledge of the mechanisms of this secretion in response to microbial stimuli remained fragmented. Therefore, we examined the expression of TLR1–10 by normal human keratinocytes and questioned whether these pattern-recognition receptors played a role in cytokine and chemokine production in response to different PAMPs. We demonstrate that human keratinocytes constitutively express mRNA for TLR1, 2, 3, 4, 5, 6, 9, and 10, but not for TLR7 or 8, confirming previously published studies on the expression of TLR1, 2, 3, 4, 5, and 9 in keratinocytes (Kawai et al., 2002; Song et al., 2002; Baker et al., 2003; Mempel et al., 2003; Kariko et al., 2004). The functionality of TLR2 and 4 has been addressed previously (Kawai et al., 2002; Song et al., 2002; Mempel et al., 2003; Baker et al., 2003), but little information is available of the functional expression of TLR3, 5, and 9 (Kollisch et al., 2005). Therefore, we focused on the response of keratinocytes to these respective ligands using the response to the TLR4 ligand LPS, as a positive control. We show that human keratinocytes express functional TLR3, 5, and 9. The fact that keratinocytes respond to TLR3, 4, 5, and 9 ligands indicates that keratinocytes may play a critical role in alerting the immune system in the presence of pathogens so that an immediate response can be mounted to contain the infection. This suggests that keratinocytes function as a link between the innate and specific immunity.

Results

Human keratinocytes express TLR1, 2, 3, 4, 5, 6, 9, and 10, but not TLR7 and 8

TLRs play a central role in innate immunity by mediating recognition of PAMPs (Akira et al., 2001). To gain insight into keratinocyte expression of TLRs, we examined the mRNA expression of all known TLRs (TLR1–10) by reverse transcription (RT)-PCR. The purity of the keratinocytes was verified by the expression of the epithelial marker, cytokeratin, and the fibroblast-specific marker (ASO2) (data not shown), and was 99% (data not shown). As depicted on Figure 1a, keratinocytes constitutively express mRNA for TLR1, 2, 3, 4, 5, 6, 9, and 10, but not for TLR7 and 8. As positive control, TLR7 and 8 expressions by plasmacytoid DCs and monocyte-derived DCs, respectively, is shown (Figure 1b).

Figure 1.

Detection of TLR mRNA by RT-PCR in cultured human keratinocytes. Human keratinocytes express constitutively TLR1, 2, 3, 4, 5, 6, 9, and 10, but not TLR7 and 8. (a) For RT-PCR, cDNAs were amplified for 45 cycles and were separated on a 1% agarose gel containing ethidium bromide. Plasmacytoid dendritic cells (pDC) express TLR7 and monocyte-derived DC (moDC) express TLR8. (b) For iCycler RT-PCR (TLR7 and 8), cDNAs were amplified for 40 cycles and were separated on a 1% agarose gel containing ethidium bromide. The data shown are representative of four independent experiments.

Full figure and legend (39K)

Chemokine/cytokine release by TLR-triggered keratinocytes

As the defense against different pathogens requires different types of immune responses, we questioned whether PAMPs, that trigger different TLRs, induce different patterns of expression of cytokines and chemokines that are associated with appropriate immune responses. We first examine whether TLR expression triggering resulted in chemokine and cytokine production by human keratinocytes. To this aim, we stimulated human keratinocytes with four different TLR ligands, for example, TLR3 ligand, double-stranded RNA (poly-I:C), the TLR5 ligand, bacterial flagellin, CpG-oligodeoxynucleotide (ODN), which ligate TLR9, and as a positive control, the TLR4 ligand LPS. Subsequently, we analyzed the expression of CXCL8, TNF-, and type I IFNs. Figure 2a shows the kinetics of CXCL8 and TNF- production by keratinocytes in response to different concentrations of poly-I:C, LPS, flagellin, or CpG (2006 or 2216). Poly-I:C-stimulated keratinocytes induced CXCL8 and TNF- in a time- and dose-dependent manner with the optimal concentration of 10 g/ml. However, keratinocytes secreted higher levels of these factors 48 hours after stimulation (P<0.01). After 72 hours, the release of these factors decreased. LPS also induced the release of CXCL8 and TNF- by keratinocytes in a time- and dose-dependent manner. In this case, also higher amounts of CXCL8 and TNF- were reached with the higher concentration of this TLR4 ligand (100 g/ml) 48 hours after stimulation (P<0.01). The TLR5 ligand, flagellin, showed the same pattern as the TLR3 and TLR4 ligands, although CXCL8 production by keratinocytes reached higher levels 72 hours after stimulation (P<0.01).

Figure 2.

Kinetics of chemokine and cytokine production by keratinocytes stimulated with different concentrations of TLR3, 4, 5, and 9 ligands. (a) Keratinocytes were cultured with poly-I:C (1, 10, and 100 g/ml), LPS (1, 10, and 100 g/ml), flagellin (1, 10, and 100 pg/ml), or CpG-ODN 2006 or 2216 (1 and 10 M) for 4, 8, 24, 48, and 72 hours, and the concentrations of CXCL8 and TNF- in the cell-free supernatants were measured with specific ELISA. Results, expressed as meanSD of triplicate cultures are from one experiment representative of five with different donors. (b) Keratinocytes were cultured with non-CpG controls, CpG 2216c and 2006c (both 10 M), or poly-I:C (10 g/ml) for 48 hours, and the concentrations of CXCL8 and TNF- in the cell-free supernatants were measured with specific ELISA. Data were analyzed for statistical significance using analysis of variance followed by Dunnett's multiple comparisons test, ^*P<0.05, ^**P<0.01. (c) TNF-, IFN-, and IFN- expression by human keratinocytes. Keratinocytes were cultured for 4 hours with poly-I:C (10 g/ml), or CpG-ODN 2006 (10 M) or 2216 (10 M), and the cells lyzed for RT-PCR analysis. The data shown are representative of four experiments.

Full figure and legend (159K)

Recent reports indicate that in humans TLR9 is primarily expressed in plasmacytoid DCs and B cells (Krieg et al., 1995; Kadowaki et al., 2001). Here we show that human keratinocytes express functional TLR9 without the addition of TGF- (Miller et al., 2005). Stimulation of keratinocytes with TLR9 ligands, CpG-ODN 2006 or 2216, resulted in the production of CXCL8 and TNF- in time- and dose-dependent manner. CXCL8 production upon stimulation of keratinocytes with 1 M CpG 2006 reached significance levels after 8 hours (P<0.01), 24 hours (P<0.01), 48 hours (P<0.05), and 72 hours (P<0.05). The same pattern was seen when the keratinocytes were stimulated with 10 M CpG 2006 (8 hours, P<0.01; 24 hours, P<0.05; 48 hours, P<0.01; 72 hours, P<0.05). CpG 2216 induced significant levels of CXCL8 production in a different manner than CpG 2006. Stimulation of keratinocytes with the lowest concentration of CpG 2216 (1 M) led to significant levels of CXCL8 production only after 8 hours (P<0.05) and 72 hours (P<0.05), whereas the stimulation with the highest concentration (10 M) led to significant levels of CXCL8 production after 8 hours (P<0.05), 24 hours (P<0.05), and 72 hours (P<0.05). Stimulation of keratinocytes with 1 M CpG 2006 led to significant levels of TNF- production after 24 hours (P<0.01), 48 hours (P<0.01), and 72 hours (P<0.01), whereas 10 M CpG 2006 led to significant levels of TNF- production after 8 hours (P<0.01), 24 hours (P<0.01), and 48 hours (P<0.01). Only the highest concentration of CpG 2216 (10 M) led to significant levels of TNF- production (after 8 hours, P<0.01; 24 hours, P<0.01; 48 hours, P<0.01; 72 hours, P<0.01). As expected, control non-CpG-containing oligonucleotides (Figure 2b) did not lead to the production of CXCL8, whereas CpG-containing oligonucleotides significantly induced the production of CXCL8. Altogether, these data indicate that keratinocytes are able to respond to CpG motifs, consistent with the presence of TLR9.

Since CpG-ODN 2216 is known to induce the expression of type I IFNs, IFN-, and IFN- in plasmacytoid DCs (Krug et al., 2001), we investigated whether these cytokines are also expressed by keratinocytes upon activation with CpG-ODNs. Type I IFNs were undetectable by standard ELISA in all groups (data not shown). Therefore, we investigated whether CpG-ODNs were able to induce mRNA for IFN- and IFN-, by RT-PCR, using TNF- expression as control cytokine and the TLR3 ligand poly-I:C as positive control. As expected, both TLR9 ligands CpG-ODN 2006 and 2216 induced the expression of TNF- mRNA in 4-hour-stimulated keratinocytes (Figure 2c). However, RT-PCR analysis readily revealed mRNA expression for IFN- (and not IFN-), in both CpG-ODN 2006- and 2216-activated keratinocytes, whereas poly-I:C also induced expression of mRNA for TNF-, IFN-, and IFN- (Figure 2c).

In a next set of experiments, we analyzed the chemokine expression by keratinocytes upon differential TLR activation. As shown in Figure 3, all the TLR ligands induced significant amounts of CCL2 and CCL20 production, whereas CCL27 production was exclusively induced by poly-I:C and flagellin. It is of particular interest that each TLR ligand induced the release of CCL2, CCL20, or CCL27 with a different kinetics and/or pattern. In this respect, both poly-I:C and LPS induced the release of higher levels of CCL2 by keratinocytes 48 hours after stimulation (for both poly-I:C and LPS, P<0.01). After 72 hours, the release of this factor decreased. Flagellin-stimulated keratinocytes released CCL2 with the same pattern as observed for poly-I:C and LPS only at the concentration of 10 ng/ml (P<0.01). Stimulation with 100 ng/ml of flagellin induced higher amounts of CCL2 after 72 hours (P<0.01). Higher concentrations of CCL2 were observed after 48 hours stimulation with CpG-ODN (2006 or 2216), irrespective of the concentration used (P<0.01).

Figure 3.

Kinetics of CCL2, CCL20, and CCL27 production by keratinocytes stimulated with different concentrations of TLR3, 4, 5, and 9 ligands. Keratinocytes were cultured with poly-I:C, LPS, flagellin, or CpG-ODN 2006 or 2216, as stated in Materials and Methods, and the concentrations of chemokines in the cell-free supernatants were measured with specific ELISA. Results, expressed as meanSD of triplicate cultures, are from one experiment representative of five.

Full figure and legend (76K)

The highest production of CCL20 upon poly-I:C (optimal concentration of 10 g/ml) or LPS (optimal concentration of 100 g/ml) stimulation was observed after 48 hours (both P<0.01), whereas upon flagellin (optimal concentration of 100 ng/ml) or CpG-ODN (both at the optimal concentration of 10 M) stimulation, the highest concentration of this chemokine was observed after 72 hours (both P<0.01).

CCL27 release was irrespective of the concentration of poly-I:C or flagellin used to stimulate the keratinocytes. These two TLR ligands only differ in that CCL27 release after 72 hours was dramatically decreased after stimulation of poly-I:C, whereas after flagellin stimulation, the release of CCL27 did not change significantly.

Altogether, these data demonstrate that human keratinocytes respond to different TLR ligands/PAMPs by showing different patterns of chemokines that are associated with the attraction of different immune cells.

Release of the Th1-attracting chemokine CXCL9 is exclusively induced by the TLR3 ligand poly-I:C, whereas CXCL10 is also induced by CpG-ODNs but not LPS or flagellin

The selective recruitment of Th cell subsets to an inflammatory site contributes strongly to the class of the specific immune responses. Therefore, we questioned whether the activation of human keratinocytes by the different TLR ligands would result in differential patterns of Th1- (CXCL9, CXCL10, and CXCL11) or Th2- (CCL22 and CCL17)-associated chemokine production (Bonecchi et al., 1998). We could not detect CCL22 and CCL17 mRNA expression by human keratinocytes, in response to the TLR ligands tested (data not shown). In addition, the levels of CCL22 and CCL17 were below the detection limit of the ELISAs (data not shown), as reported previously (Albanesi et al., 2001). Interestingly, only the TLR3 ligand poly-I:C induced production (Figure 4) of the Th1-associated chemokine CXCL9 in a time- and dose-dependent manner (optimal concentration of 10 g/ml, higher release 48 hours after stimulation, P<0.01). Whereas CXCL10 release by keratinocytes after poly-I:C stimulation required minimal amounts of this TLR ligand (1 g/ml), stimulation with CPG-ODNs only led to CXCL10 release with the highest concentration (10 M).

Figure 4.

Kinetics of the Th1-attracting chemokines CXCL9 and CXCL10 production by keratinocytes stimulated with different concentrations of TLR3, 4, 5, and 9 ligands. Keratinocytes were cultured with poly-I:C, LPS, flagellin, or CpG-ODN 2006 or 2216, as stated above, and the concentrations of chemokines in the cell-free supernatants were measured with specific ELISA. Results, expressed as meanSD of triplicate cultures, are from one experiment representative of five.

Full figure and legend (49K)

TLR ligands poly-I:C, LPS, and flagellin induce different patterns of surface molecule expression by keratinocytes

To study the effects of different TLR ligands on the expression of surface molecules by keratinocytes, keratinocytes were cultured for 48 hours in the absence or in the presence of the optimal concentration of TLR ligands: poly-I:C (10 g/ml), LPS (100 g/ml), flagellin (100 ng/ml), and CpG-ODN 2006 or CpG-ODN 2216 (both 10 M). As shown in Figure 5, poly-I:C and flagellin were the only ligands that induced the expression of ICAM-1 and major histocompatibility complex class I molecules (HLA-ABC). In contrast, major histocompatibility complex class II (HLA-DR) expression was induced by poly-I:C and LPS. None of the TLR ligands induced significant changes in FasR and CD40 expression, except for poly-I:C that moderately, but not significantly, upregulated the expression of these two surface molecules. These results demonstrate that human keratinocytes are able to respond to selected TLR ligands by enhancing the expression of certain surface molecules.

Figure 5.

Keratinocytes modulate the expression of cell surface molecules in response to different TLR ligands. Keratinocytes were stained with antibodies to ICAM-1, HLA-DR, HLA-ABC, FasR, or CD40 48 hours after exposure to the indicated TLR ligands (poly-I:C, 10 g/ml; LPS, 100 g/ml; flagellin, 100 ng/ml; CpG-ODN 2006, 10 M; CpG-ODN 2216, 10 M) or to medium alone, and analyzed by FACS. Results are expressed as mean of duplicate cultures. Mean fluorescence intensity represents the difference between the various stainings and the isotype control. Data are representative of four experiments with different donors with similar results.

Full figure and legend (27K)

TLR3, 4, 5, and 9 triggering results in phosphorylation of IB

TLR ligation activates NF-B and results in the transcription of NF-B-dependent genes (Muzio et al., 1998). Therefore, we tested whether TLR3, 4, 5, and 9 triggering can give rise to a typical TLR-related signaling response, for example, phosphorylation of IB. Human keratinocytes were stimulated with two different concentrations of poly-I:C, LPS, flagellin, or CpG-ODNs (2216 or 2006) as indicated in Figure 6, and whole-cell lysates were prepared and analyzed by Western blot. Phosphorylation of IB was determined using a specific antibody. As shown in Figure 6, no phosphorylated-IB was observed in unstimulated keratinocytes. On the contrary, poly-I:C at a concentration of 10 g/ml already induces the expression of phosphorylated-IB, as indicated by a clear band on the blot in lane 2. This expression of phospho-IB was even more pronounced using 100 g/ml of poly-I:C. LPS at a concentration of 10 g/ml only show just a visible band of phospho-IB, which is potently enhanced using 100 g/ml. The same pattern of expression of phospho-IB is observed in flagellin-stimulated keratinocytes. The CpG-ODN 2216 induces some expression of phospho-IB, whereas CpG 2006 at 1 M clearly induces phosphorylation of IB, which is not further enhanced by a higher concentration of CpG 2006. Thus, TLR3, 4, 5, or 9 ligand engagement leads to the activation of NF-B, indicating the presence of intact TLR3, 4, 5, and 9 signaling in human keratinocytes.

Figure 6.

Phosphorylation of IB by TLR3-, 4-, 5-, or 9-triggered human keratinocytes. Keratinocytes were stimulated with different concentrations of poly-I:C (1 and 10 g/ml), LPS (10 and 100 g/ml), CpG-ODN 2006, or CpG-ODN 2216 (1 and 10 M) for 4 hours. Cell lysates were fractionated by SDS-PAGE and then analyzed by Western blotting using antibodies against phospho-IB. A single representative experiment is shown from three different experiments.

Full figure and legend (30K)

Activation of TLR3, 4, 5, and 9 lead to nuclear translocation of the NF-B subunit p65

To further substantiate the activation of keratinocytes via TLR3, 4, 5, or 9 ligands, we tested whether triggering with these PAMPs results in the translocation to the nucleus of the NF-Bp65. To this aim, human keratinocytes were cultured in slide chambers in the presence or in the absence of different concentrations of TLR3, 4, 5, or 9 ligands and stained for the NF-Bp65 with a 4,6-diamidino-2-phenylindole counterstain to indicate the nucleus as described in Materials and Methods. The slides were then analyzed by confocal laser scanning microscopy (Figure 7a) or by fluorescence microscopy (Figure 7b–q). Figure 7a shows the separate stainings of the nucleus (Figure 7a, upper panel, 4,6-diamidino-2-phenylindole), NF-Bp65 staining, (Figure 7a, middle panel, Alexa-488), as well as the overlay (Figure 7a, lower panel). In unstimulated keratinocytes, the stainings for the NF-Bp65 are mainly visible in the cytoplasm. In contrast, when keratinocytes are stimulated with poly-I:C, LPS, flagellin, or CpG-ODNs (2006), a clear nuclear translocation of the p65 subunit is observed, indicated by the light green in single color pictures and light blue staining in the overlay pictures (Figure 7a, arrows). Since in our experimental conditions, lower concentrations of LPS did not lead to the release of cytokines/chemokines by keratinocytes (data not shown), we asked whether this fact was due to no translocation to the nucleus of the NF-Bp65. Indeed, only a few cells stimulated with 100 ng/ml LPS show nuclear translocation of the p65 subunit (Figure 7e, arrows), whereas at lower concentrations (Figure 7c and d), no nuclear translocation is observed. As expected, keratinocytes cultured with non-CpG controls, CpG 2006c (Figure 7m) and 2216c (Figure 7p), did not result in nuclear translocation of the p65 subunit. Altogether, these data indicate that NF-B participates in TLR3, 4, 5, and 9 signaling in response to the respective PAMPs.

Figure 7.

Detection of nucleus translocation of the NF-B subunit p65 expression in TLR3, 4, 5, and 9 ligand-stimulated keratinocytes. Detection by (a) confocal laser scanning microscopy (original magnification, 64, bar=20 m) or (b--q) fluorescence microscopy (original magnification, 40, bar=20 m). (a) The upper panel shows the 4,6-diamidino-2-phenylindole staining (blue) indicating the nucleus; the middle panel shows the NF-B staining (Alexa-488, green); and the lower panel the overlay. NF-Bp65 present in the nucleus is indicated by arrows. The data shown are representative of four experiments. (b–q) As in (a) NF-Bp65 present in the nucleus is indicated by arrows. (b) Unstimulated; (c) LPS (1 ng/ml); (d) LPS (10 ng/ml); (e) LPS (100 ng/ml); (f) LPS (1 g/ml); (g) LPS (10 g/ml); (h) LPS (100 g/ml); (i) poly-I:C (10 g/ml); (j) flagellin (100 ng/ml); (k) CpG 2006 (1 M); (l) CpG 2006 (10 M); (m) control-CpG 2006 (10 M); (n) CpG 2216 (1 M); (o) CpG 2216 (10 M); (p) control-CpG 2216 (10 M); and (q) control Ig.

Full figure and legend (330K)

Discussion

In the current study, we described that human keratinocytes constitutively express mRNA for TLR1, 2, 3, 4, 5, 6, 9, and 10, but not for TLR7 and 8. Activation of keratinocytes with different PAMPs resulted in differential expression patterns of chemokines and cytokines, and cell surface molecules. These data indicate the importance of keratinocytes in the defense against invading pathogens of the skin as they express functional TLRs. Moreover, ligation of different TLRs resulted in the release of different patterns of cytokines and chemokines that contribute to the initiation of appropriate immune responses.

A major challenge to innate immune cells is the discrimination of foreign pathogens. Innate immune cells possess germline-encoded pattern-recognition receptors that recognize and are triggered by evolutionary conserved molecules essential for pathogen function but absent in the host (Akira et al., 2001; Medzhitov, 2001). The activation of TLRs by innate immune cells has been extensively studied in monocytes, in macrophages, and in different DC subsets (Akira et al., 2001; Kadowaki et al., 2001). Recently, it was reported that epithelial cells also express TLRs (Cario and Podolsky, 2000; Cario et al., 2000; Wolfs et al., 2002). Thus far, in keratinocytes the expression of TLR1, 2, 3, 4, 5, and 9 has been documented (Kawai et al., 2002; Song et al., 2002; Baker et al., 2003; Mempel et al., 2003; Kariko et al., 2004; Kollisch et al., 2005). In this study, we show that human keratinocytes functionally express TLR3, 4, 5, and 9.

TLR activation results in the activation of NF-B (Muzio et al., 1998) that induces a genetic program that is essential for host defense, including the induction of a subset of surface molecules expression, cytokines and chemokines (Siebenlist et al., 1994). NF-B is present in the cytoplasm as an inactive complex, which is rapidly translocated into the nucleus upon stimulation (Siebenlist et al., 1994). To examine the molecular mechanisms of cytokine production in TLR3-, 4-, 5-, or 9 ligand-activated human keratinocytes, we investigate whether their ligation results in the phosphorylation of I-B and NF-B translocations. Indeed, we demonstrate the phosphorylation of IB and the translocation to the nucleus of the NF-B subunit p65 upon keratinocyte activation with TLR3-, 4-, 5-, or 9 ligands, indicating that signaling via these receptors led to NF-B activation and subsequent gene transcription of NF-B-sensitive genes. These findings support the notion that NF-B is one of the most important cellular factors involved in the regulation of the host innate anti-microbial response (Siebenlist et al., 1994).

Keratinocytes secrete various soluble mediators that orchestrate the immune response (Uchi et al., 2000; Gröne, 2002). Thus, chemokines released from keratinocytes may determine whether and which types of immune cells are attracted to the skin epithelium. We found that all the TLR ligands used induced statistically significant levels of the neutrophil chemoattractant CXCL8, as well as CCL2, which may promote the immigration of monocytes (Matsushima et al., 1989), DCs and Langerhans cells (Nakamura et al., 1995), memory T cells (Carr et al., 1994), and natural killer cells (Allavena et al., 1994) into the skin in case of most pathogen infections. The recruitment of immature DCs into the epidermis is a key step in the development of specific immunity. It was recently reported that keratinocyte-derived CCL20, induced by proinflammatory cytokines TNF- and IL-1 (Dieu-Nosjean et al., 2000), is important not only in recruiting precursors of Langerhans cells to the epidermis (Charbonnier et al., 1999) but also memory T cells (Liao et al., 1999). Here we demonstrate that CCL20 can also be released by keratinocytes in response to the TLR ligands poly-I:C and flagellin. In addition, our data confirm the specificity of flagellin as one of the major bacterial PAMPs that induces the production of CCL20 by epithelial cells, in contrast to LPS (Sierro et al., 2001). CCL27 is a skin-restricted chemokine that has an important role in both homeostasis and initiation of inflammation (Morales et al., 1999; Homey et al., 2002). We show that CCL27 production by keratinocytes was selectively induced by the TLR3 ligand poly-I:C and the TLR5 ligand flagellin. CCR10⁺CLA⁺ T cells that may be attracted to the skin by keratinocyte-derived CCL27 may encounter their specific antigen and release effector mediators, which induce more and different chemokines to sustain a state of inflammation that finally leads to efficient clearance of the virus or bacteria, respectively.

The selective recruitment of Th cell subsets to an inflammatory site is of particular importance in determining the character of the immune responses. CXCL9, CXCL10, and CXCL11 are CXCR3 ligands that attract mainly activated Th1 cells. Recently, we have reported that poly-I:C-activated keratinocytes promote Th1 responses by inducing a type-1-promoting phenotype in DCs (Lebre et al., 2003). Here we show that poly-I:C further promotes local Th1 responses by inducing the expression of Th1-attracting chemokines in keratinocytes. The experiments with poly-I:C, as model for virus infection, thus suggest that keratinocytes may be important players in cutaneous anti-viral responses by participating in both initiation phase of the immune response, by inducing the Th1-promoting phenotype, and, in the effector phase, by attracting polarized Th1 cells to the site of virus entry. Moreover, we show for the first time that, in addition to plasmacytoid DCs (Kadowaki et al., 2001) and B cells (Hornung et al., 2002), CpG-ODNs can also induce the release of CXCL10 by human keratinocytes.

The immune system uses TLR9 to detect the presence of unmethylated CpG motifs (Hemmi et al., 2000). In humans, the expression of TLR9 is thought to be restricted to B cells and plasmacytoid DCs (Krieg et al., 1995; Kadowaki et al., 2001). Although a recent study by Kollisch et al. (2005) failed to detect TLR9 expression by human keratinocytes and only nonsignificant responses to CpG-ODN, we here clearly demonstrate the expression of TLR9-specific mRNA in keratinocytes, as well as a CpG-ODN response as determined by the induction of cytokines and chemokines, the phosphorylation of IB and the nuclear translocation of NF-B RelA. These findings are in line with earlier published data showing that activation of keratinocytes with CpG-ODN resulted in the upregulation of various acute-phase response genes (Mirmohammadsadegh et al., 2002) and that TLR9-specific mRNA is detectable in lysates of human keratinocytes (Mempel et al., 2003).

Remarkably, the TLR9 ligands CpG-ODN 2006 or 2216 did not induce detectable protein levels of IFN- and IFN-. However, mRNA analysis revealed that, in addition to poly-I:C, CpG-ODN 2006- or 2216-activated keratinocytes show upregulated mRNA expression for both TNF- and IFN-. Compared to TNF- (Uchi et al., 2000; Gröne, 2002), type I IFNs are rapidly expressed by keratinocytes (Fujisawa et al., 1997) and participates in the innate immune response both as a signal for the presence of bacterial and viral infections and as an effector molecule inhibiting the spread of infection (Bogdan, 2000).

The early expression of certain cell surface molecules, chemokines, and cytokines is essential in shaping the innate and specific immune responses. This study strongly suggests that keratinocytes are not merely a barrier but actively contribute to the induction of the immune response through the differential activation of TLRs, resulting in differential patterns of expression of genes involved in the inflammatory responses.

Materials and Methods

Keratinocyte cultures

Primary keratinocyte cultures were prepared from plastic surgery skin obtained from healthy subjects. Donation of skin biopsies by healthy controls followed approval by the ethical committee of the Academic Medical Center (Amsterdam, The Netherlands) and this study was conducted according to the Declaration of Helsinki Principles. All normal subjects gave written informed consent. Briefly, the skin was incubated with dispase (0.3%; Boehringer, Mannheim, Germany) for 16 hours at 4°C. Epidermal sheets were removed from the dermis and single-cell suspensions were obtained by placing epidermal sheets in trypsin (0.025%; Life Technologies, Paisley, UK) for 5 minutes at 37°C. After neutralizing with equal volumes of fetal calf serum (HyClone, Logan, UT), stratum corneum debris was removed and then sieved through sterile nylon gauze to obtain a single-cell suspension. Isolated epidermal cells were seeded at the density of 8–10 10⁴ cells/cm². At 70–80% confluence, keratinocytes were detached with 0.025% trypsin, 2 mM EDTA for 5 minutes at 37°C, and subcultured or frozen. Keratinocyte cultures were maintained in keratinocyte serum-free medium (Gibco, Paisley, UK).

RNA isolation and cDNA synthesis

Keratinocytes were plated at the concentration of 1.5 10⁵ cells/3 ml (six-well plates, Costar, Cambridge, MA), and cultured at least for 48 hours before the cells were lyzed. Keratinocyte total RNA was purified by using the NucleoSpin^® RNA II kit (Macherey-Nagel, Düren, Germany) according to the manufacturer's instructions. cDNA was generated using the first-strand cDNA synthesis kit for RT-PCR (MBI Fermentas, St Leon-Rot, Germany). To anneal the primer to the RNA, 9 l of total RNA, 1 l oligo(dT)₁₈, and 1 l D(N)₆ were added. This mix was then heated for 5 minutes at 94°C.

RT-PCR and iCycler PCR analysis

The primers sequences used are shown in Table 1. cDNAs were amplified for 45 cycles (RT-PCR) or 40 cycles (iCycler) and were analyzed on a 1% agarose gel containing ethidium bromide. A 100 bp DNA ladder standard (MBI Fermentas) was used as a size marker. iCycler PCR for TLR7 and 8 was performed with SYBR Green (AB gene, Hamburg, Germany). For each iCycler PCR reaction, the same amount of reverse-transcribed RNA was used and input cDNA was normalized according to glyceraldehydes-3-phosphate dehydrogenase as internal control gene.

Table 1 - Primer sequences.

Full table

Keratinocyte stimulation for cytokine/chemokine measurements

Keratinocytes were plated at the concentration of 0.1 10⁴ cells/200 l (96-well plates, Costar), and cultured at least for 48 hours before conducting experiments. To induce cytokine and chemokine production, keratinocytes were cultured for 4, 8, 24, 48, or 72 hours in the presence of different concentrations of poly-I:C (1, 10, 100 ng/ml or 1, 10, 100 g/ml; Sigma-Aldrich, St Louis, MO), or HPLC-purified LPS (1, 10, 100 g/ml; from E. coli 0111:134, Sigma-Aldrich), or flagellin (1, 10, 100 ng/ml; a kind gift from Dr A.T. Gewirtz), or the CpG-ODN 2006 (5'-GGGGGGACGATCGTC GGGGGG-3') (Krieg et al., 1995) or CpG-ODN 2216 (5'-TCGTCG TTTTGTCGTTTTGTCGT-3') (Krug et al., 2001) (1, 10 M; Biosource International, Nivelles, Belgium). The following methylated CpG-ODNs were used as controls: ODN 2117 (5'-TQGTQGTTTTGTQ GTTTTGTQGTT-3'; to 2006) and ODN 2243 (5'-ggGGGAGCAT GCTCgggggG-3'; control to 2216). The concentration of the specific TLR ligands was determined in dose-finding experiments (data not shown).

Determination of cytokine and chemokine production by ELISA

Measurements of CXCL8 and TNF- were performed by ELISA using pairs of specific mAbs and recombinant standard obtained from BioSource International (Camarillo, CA). CCL2 was determined using the Ab pair, rat polyclonal 20521D for coating and rabbit polyclonal 20532D for detection and recombinant CCL2 19781T (BD Pharmingen, San Diego, CA). CCL20 was determined using a DuoSet Elisa purchased by R&D Systems (Abingdon, Oxon, UK). CXCL10, CXCL9, and CCL27 were determined using pairs of specific Abs and recombinant standard obtained from R&D Systems. The limits of detection of these ELISA are as follows: CXCL8, 30 pg/ml; TNF-, 20 pg/ml; CCL2, 40 pg/ml; CCL20, 60 pg/ml; CXCL10, 80 pg/ml; CXCL9, 40 pg/ml, and CCL27, 80 pg/ml.

Analysis of cell surface molecule expression

Keratinocytes were plated at the concentration of 1.5 10⁵ cells/3 ml (six-well plates, Costar), and cultured at least for 48 hours. Keratinocytes were cultured in the presence of optimal concentrations of synthetic double-stranded RNA poly-I:C (50 g/ml), or LPS (100 g/ml), or flagellin (100 ng/ml), or the CpG-ODN 2006 (Krieg et al., 1995) or CpG-ODN 2216 (5 M). After 48 hours, the cells were detached from the wells by incubating them with 1 ml of the cell dissociation solution (Sigma-Aldrich) for 15 minutes at 37°C. Trypsin neutralization was performed by adding equal volumes of fetal calf serum. After spinning for 10 minutes at 800 r.p.m., the single-cell suspension was stained with mouse anti-human mAbs against the following surface molecules: ICAM-1 (15.2, IgG1; Central Laboratory of the Netherlands Red Cross Blood Transfusion Service, CLB, Amsterdam, The Netherlands), HLA-DR (L234, IgG2a; Becton Dickinson, San Jose, CA), HLA-ABC (IgG2a; Halan Sera-Lab Ltd, Belton, UK), FasR/CD95 (ZB4, IgG1; Immunotech, Marseille, France), and CD40 (5D12, IgG1; ATCC, Manassas, VA). FITC-coupled goat F(ab')₂ anti-mouse IgG and IgM (Jackson ImmunoResearch Laboratories, West Grove, PA) was used as a secondary reagent. Samples were analyzed on a FACScan (Becton Dickinson).

Western blotting

Keratinocytes were plated at the concentration of 1.5 10⁵ cells/ml (24-well plates, Costar), and cultured at least for 48 hours before the cells were stimulated. Four hours after stimulation with different concentration of poly-I:C, LPS, flagellin, and CpG 2006 or CpG 2116, keratinocytes were washed twice with ice-cold phosphate-buffered saline to remove all serum proteins and then lyzed in 1 SDS-PAGE sample buffer. The proteins were separated by SDS-PAGE on a 12% gel, using Rainbow-colored protein molecular weight markers (Amersham, Little Chalfont, UK) as a reference, and transferred onto a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA). The membrane was blocked in Tris-buffered saline containing 2% non-fat dry milk (Bio-Rad) and 0.05% Tween 20 (Merck-Schuchardt, Hohenbrunn, Germany) during 1 hour. Detection of the phosphorylated protein was performed by incubating (overnight at 4°C) the membranes with a mouse against human phospho-IB (Ser32/36) primary antibody (Cell Signaling Technology, Beverly, MA), and visualized the next day by horseradish peroxidase-conjugated goat anti-mouse IgG (H+L) antibody (Bio-Rad) (diluted in Tris-buffered saline containing 2% non-fat dry milk and 0.05% Tween 20), enhanced chemiluminescence system (LumiGLO, Cell Signaling Technology) and exposure to a Fuji Medical X-ray film (Fuji, Tokio, Japan).

Detection of the NF-Bp65 expression by confocal laser scanning microscopy and fluorescence microscopy

Keratinocytes were plated at the concentration of 0.1 10⁴ cells/200 l (Lab-Tek^® chamber slides; Life Technologies), and cultured at least for 48 hours before conducting experiments. To induce nuclear translocation of p65, keratinocytes were cultured for 4 hours in the presence of different concentrations of poly-I:C (10 and 100 g/ml), or LPS (10 and 100 g/ml), or flagellin (10 and 100 ng/ml), or CpG-ODN 2006 (1 and 10 M), or CpG-ODN-2216 (1 and 10 M). After removal of the culture medium, the cells were fixed with acetone (-20°C) for 10 minutes After washing in phosphate-buffered saline, the slides were incubated (30 minutes at room temperature) with a rabbit anti-human NF-Bp65 antibody (2 g/ml, sc-109, Santa Cruz Biotechnology, Heidelberg, Germany) diluted in phosphate-buffered saline/1% fetal calf serum/1% human serum. The cells were then washed with phosphate-buffered saline and incubated (30 minutes at room temperature) with an Alexa 488-conjugated goat anti-rabbit (Molecular Probes Europe BV, Invitrogen, Karlsruhe, Germany). After washing the cells with phosphate-buffered saline, the slides were mounted in Vectashield Hard Set with 4,6-diamidino-2-phenylindole in order to visualize nuclear staining (Vector Laboratories, Burlingame, CA). The slides were analyzed by confocal laser scanning microscopy using a Leica SP2 confocal microscope (Leica, Wetzlar, Germany) with a 40 Plan Apochromat objective and edited with Simulator Leica confocal software (version 2.5 build 1347, Leica). When stated, the slides were also analyzed using a fluorescence microscope (Leica DMRA, Wetzlar, Germany) coupled to a CCD camera and Image-Pro Plus software (Media Cybernetics, Dutch Vision Components, Breda, The Netherlands).

Statistical analysis

Data are expressed as meanSD. Data were analyzed for statistical significance with the GraphPad InStat^® software (version 3.00; GraphPad InStat Inc., San Diego, CA) using analysis of variance followed by Dunnett's multiple comparisons test. A P-value <0.05 was considered as the level of significance (^*P<0.05, ^**P<0.01, ^***P<0.001) compared to unstimulated cultures.

Conflict of Interest

The authors state no conflict of interest.

References

1. Akira S, Takeda K, Kaisho T (2001) Toll-like receptors: critical proteins linking innate and acquired immunity. Nat Immunol 2:675–680 | Article | PubMed | ISI | ChemPort |
2. Albanesi C, Scarponi C, Sebastiani S, Cavani A, Federici M, Girolomoni G et al. (2000) IL-4 enhances keratinocytes expression of CXCR3 agonistic chemokines. J Immunol 165:1395–1402 | PubMed | ISI | ChemPort |
3. Albanesi C, Scarponi C, Sebastiani S, Cavani A, Federici M, Sozzani S et al. (2001) A cytokine-to-chemokine axis between T lymphocytes and keratinocytes can favor Th1 cell accumulation in chronic inflammatory skin diseases. J Leukocyte Biol 70:617–623 | PubMed | ISI | ChemPort |
4. Alexopoulou L, Holt AC, Medzhitov R, Flavell RA (2001) Recognition of double-stranded RNA and activation of NF-kappaB by Toll-like receptor 3. Nature 413:732–738 | Article | PubMed | ISI | ChemPort |
5. Allavena P, Bianchi G, Zhou D, Van Damme J, Jilek P, Sozzani S et al. (1994) Induction of natural killer cell migration by monocyte chemotactic protein-1, -2 and -3. Eur J Immunol 24:3233–3236 | PubMed | ISI | ChemPort |
6. Baggiolini M, Loetscher P, Moser B (1995) Interleukin-8 and the chemokine family. Int J Immunopharmacol 17:103–108 | Article | PubMed | ISI | ChemPort |
7. Baker BS, Ovigne JM, Powles AV, Corcoran S, Fry L (2003) Normal keratinocytes express Toll-like receptors (TLRs) 1, 2 and 5: modulation of TLR expression in chronic plaque psoriasis. Br J Dermatol 148:670–679 | Article | PubMed | ISI | ChemPort |
8. Barker JNWN, Sarma V, Mitra RS, Dixit VM, Nickoloff BJ (1990) Marked synergism between tumor necrosis factor- and interferon- in regulation of keratinocyte-derived adhesion molecules and chemotactic factors. J Clin Invest 85:605–608 | PubMed | ISI | ChemPort |
9. Bogdan C (2000) The function of type I interferons in antimicrobial immunity. Curr Opin Immunol 12:419–424 | Article | PubMed | ISI | ChemPort |
10. Bonecchi R, Bianchi G, Bordignon PP, D'Ambrosio D, Lang R, Borsatti A et al. (1998) Differential expression of chemokine receptors and chemotactic responsiveness of type 1 T helper cells (Th1s) and Th2s. J Exp Med 187:129–134 | Article | PubMed | ISI | ChemPort |
11. Boorsma DM, Flier J, van den Brink EN, Sampat S, Walg HL, Willemze R et al. (1999) IP-10 mRNA expression in cultured keratinocytes is suppressed by inhibition of protein kinase-C and tyrosine kinase and elevation of cAMP. Cytokine 11:469–475 | Article | PubMed | ChemPort |
12. Boring L, Gosling J, Monteclaro FS, Lusis AJ, Tsou CL, Charo IF (1996) Molecular cloning and functional expression of murine JE (monocyte chemoattractant protein 1) and murine macrophage inflammatory protein 1alpha receptors: evidence for two closely linked C-C chemokine receptors on chromosome 9. J Biol Chem 271:7551–7558 | Article | PubMed | ISI | ChemPort |
13. Brightbill HD, Libraty DH, Krutzik SR, Yang RB, Belisle JT, Bleharski JR et al. (1999) Host defense mechanisms triggered by microbial lipoproteins through toll-like receptors. Science 285:732–736 | Article | PubMed | ISI | ChemPort |
14. Cario E, Podolsky DK (2000) Differential alteration in intestinal epithelial cell expression of toll-like receptor 3 (TLR3) and TLR4 in inflammatory Bowel disease. Infect Immun 68:7010–7017 | Article | PubMed | ISI | ChemPort |
15. Cario E, Rosenberg IM, Brandwein SL, Beck PL, Reinecker H-C, Podolsky DK (2000) Lipopolysaccharide activates distinct signaling pathways in intestinal epithelial cell lines expressing toll-like receptors. J Immunol 164:966–972 | PubMed | ISI | ChemPort |
16. Carr MW, Roth SJ, Luther E, Rose SS, Springer TA (1994) Monocyte chemoattractant protein 1 acts as a T-lymphocyte chemoattractant. Proc Natl Acad Sci USA 91:3652–3656 | Article | PubMed | ChemPort |
17. Charbonnier A-S, Kohrgruber N, Kriehuber E, Stingl G, Rot A, Maurer D (1999) Macrophage inflammatory protein 3 is involved in the constitutive trafficking of epidermal Langerhans cells. J Exp Med 190:1755–1767 | Article | PubMed | ISI | ChemPort |
18. Chuang T, Ulevitch RJ (2001) Identification of hTLR10: a novel human Toll-like receptor preferentially expressed in immune cells. Biochim Biophys Acta 1518:157–161 | Article | PubMed | ISI | ChemPort |
19. Diebold SS, Kaisho T, Hemmi H, Akira S, Reis E Sousa (2004) Innate antiviral responses by means of TLR7-mediated recognition of single-stranded RNA. Science 303:1529–1531 | Article | PubMed | ISI | ChemPort |
20. Dieu-Nosjean M-C, Massacrier C, Homey B, Vanbervliet B, Pin J-J, Vicari A et al. (2000) Macrophage inflammatory protein 3 is expressed at inflamed epithelial surfaces and it is the most potent chemokine known in attracting Langerhans cell precursors. J Exp Med 192:705–717 | Article | PubMed | ISI | ChemPort |
21. Fujisawa H, Kondo S, Wang B, Shivji GM, Sauder DN (1997) The expression and modulation of IFN-alpha and IFN-beta in human keratinocytes. J Interferon Cytokine Res 17:721–725 | PubMed | ISI | ChemPort |
22. Gröne A (2002) Keratinocytes and cytokines. Vet Immunol Immunopathol 88:1–12 | PubMed | ISI | ChemPort |
23. Hayashi F, Smith KD, Ozinsky A, Hawn TR, Yi EC, Goodlett DR et al. (2001) The innate immune response to bacterial flagellin is mediated by Toll-like receptor 5. Nature 410:1099–1103 | Article | PubMed | ISI | ChemPort |
24. Heil F, Hemmi H, Hochrein H, Ampenberger F, Kirschning C, Akira S et al. (2004) Species-specific recognition of single-stranded RNA via toll-like receptor 7 and 8. Science 303:1526–1529 | Article | PubMed | ISI | ChemPort |
25. Hemmi H, Takeuchi O, Kawai T, Kaisho T, Sato S, Sanjo H et al. (2000) A Toll-like receptor recognizes bacterial DNA. Nature 408:740–745 | Article | PubMed | ISI | ChemPort |
26. Homey B, Alenius H, Muller A, Soto H, Bowman EP, Yuan W et al. (2002) CCL27–CCR10 interactions regulate T cell-mediated skin inflammation. Nat Med 8:157–165 | Article | PubMed | ISI | ChemPort |
27. Homey B, Wang W, Soto H, Buchanan ME, Wiesenborn A, Catron D et al. (2000) The orphan chemokine receptor G protein-coupled receptor-2 (GPR-2, CCR10) binds the skin-associated chemokine CCL27 (CTACK/ALP/ILC). J Immunol 164:3465–3470 | PubMed | ISI | ChemPort |
28. Hornung V, Rothenfusser S, Britsch S, Krug A, Jahrsdorfer B, Giese T et al. (2002) Quantitative expression of toll-like receptor 1–10 mRNA in cellular subsets of human peripheral blood mononuclear cells and sensitivity to CpG oligodeoxynucleotides. J Immunol 168:4531–4537 | PubMed | ISI | ChemPort |
29. Kadowaki N, Ho S, Antonenko S, Malefyt RW, Kastelein RA, Bazan F et al. (2001) Subsets of human dendritic cell precursors express different toll-like receptors and respond to different microbial antigens. J Exp Med 194:863–869 | Article | PubMed | ISI | ChemPort |
30. Kariko K, Bhuyan P, Capodici J, Weissman D (2004) Small interfering RNAs mediate sequence-independent gene suppression and induce immune activation by signaling through toll-like receptor 3. J Immunol 172:6545–6549 | PubMed | ISI | ChemPort |
31. Kawai K, Shimura H, Minagawa M, Ito A, Tomiyama K, Ito M (2002) Expression of functional Toll-like receptor 2 on human epidermal keratinocytes. J Dermatol Sci 30:185–194 | Article | PubMed | ISI | ChemPort |
32. Kollisch G, Kalali BN, Voelcker V, Wallich R, Behrendt H, Ring J et al. (2005) Various members of the Toll-like receptor family contribute to the innate immune response of human epidermal keratinocytes. Immunol 114:531–541 | Article |
33. Krieg AM, Yi AK, Matson S, Waldschmidt TJ, Bishop GA, Teasdale R et al. (1995) CpG motifs in bacterial DNA trigger direct B-cell activation. Nature 374:546–549 | Article | PubMed | ISI | ChemPort |
34. Krug A, Rothenfusser S, Hornung V, Jahrsdorfer B, Blackwell S, Ballas ZK et al. (2001) Identification of CpG oligonucleotide sequences with high induction of IFN-alpha/beta in plasmacytoid dendritic cells. Eur J Immunol 31:2154–2163 | Article | PubMed | ISI | ChemPort |
35. Lebre MC, Antons JC, Kalinski P, Schuitemaker JH, van Capel TM, Kapsenberg ML et al. (2003) Double-stranded RNA-exposed human keratinocytes promote Th1 responses by inducing a Type-1 polarized phenotype in dendritic cells: role of keratinocyte-derived tumor necrosis factor alpha, type I interferons, and interleukin-18. J Invest Dermatol 120:990–997 | Article |