|Home | About | Journals | Submit | Contact Us | Français|
Staphylococcus aureus is a significant human pathogen which can colonize the skin. Neutrophils are well known to be involved in clearance of the bacterium. This study focused on exploring the role human keratinocytes play as first responders to bacterial challenges. IL-1α and IL-1β increased mRNA production and protein secretion of the neutrophil chemotactic CXCL1, CXCL2 and IL-8 in keratinocytes. S. aureus and the bacterial cell wall components lipoteichoic acid (LTA) and peptidoglycan (PGN) induced similar expression profiles in a Toll-like receptor (TLR)-2 dependent manner. Interestingly, the S. aureus induced mRNA levels peaked at later time-points than those induced by IL-1. The S. aureus activated chemokine production was preceded by significant IL-1α and IL-1β secretion. Expression of IL-1α was significantly higher than that of IL-1β. Inhibition of IL-1RI using neutralizing antibodies revealed that S. aureus derived LTA and PGN induced chemokine expression requires IL-1RI engagement. Surprisingly, we further found that chemokine secretion is dependent upon endocrine IL-1α, but not IL-1β, signaling. Our data demonstrate that the innate immune response of keratinocytes is regulated differently than those of other cell types. This may represent a fail-safe system, which protects the host against genetic variation and immune evasion mechanisms developed by pathogens.
The Gram-positive Staphylococcus aureus bacterium causes a wide spectrum of human inflammatory diseases ranging from benign abscesses, through necrotizing fasciitis to potentially fatal pneumonia and sepsis. When challenged by a microorganism the host must be able to immediately initiate elimination processes. These are in part activated by the innate immune system and the TLRs. TLRs distinguish self from non-self by recognizing pathogen-associated molecular patterns (PAMPs) present on, or in, the microorganism but not the host (reviewed in (Ishii et al., 2008)). Such motifs include flagellin from the flagellum of motile bacteria (recognized by TLR5), single and double stranded RNA generated during viral life-cycles (sensed by TLR7/8 and TLR3, respectively), and lipopolysaccharide and lipoteichoic acid (LTA) from bacteria (bound by TLR4 and TLR2, respectively). TLRs are expressed on innate immune cells such as macrophages, neutrophils and dendritic cells but also on the adaptive B and T cells. Furthermore, the receptors can be found on fibroblasts and several types of epithelial cells (Akira et al., 2006). Engagement of the TLRs activates complex intracellular signaling pathways which through increased production of pro-inflammatory cytokines and type I IFNs regulate down-stream innate and adaptive immune responses (Akira et al., 2006).
The S. aureus cell wall contains two potential TLR activators: LTA and peptidoglycan (PGN). LTA and PGN from many types of Gram-positive bacteria are typically recognized by TLR2 (Akira et al., 2006) and several studies have independently implicated TLR2 in the immune response against S. aureus. TLR2 deficient mice have a significantly higher mortality rate than wild-type mice following i.v. administration of S. aureus (Takeuchi et al., 2000) and production of TNF-α and IL-6 is lower in TLR2−/− than TLR2+/+ macrophages treated with heat-killed S. aureus (HKSA) (Takeuchi et al., 2000). TLR2 deficient mice are more readily colonized in the nasal cavity than wild-type mice (Gonzalez-Zorn et al., 2005) and establishment of subcutaneous infections requires 10-fold more CFU in TLR2+/+ than in TLR2−/− mice (Kristian et al., 2003). Curiously, neutrophil recruitment and clearance of subcutaneous bacteria appear to be TLR2 independent in mice (Miller et al., 2006). These phenotypes are instead dependent upon IL-1β and IL-1R type I (IL-1RI) signaling (Miller et al., 2006; Miller et al., 2007).
IL-1 is a pluri-potent pro-inflammatory cytokine (reviewed in (Dinarello, 2009; O'Neill, 2008)). There are two classical isoforms of IL-1, IL-1α and IL-1β. While both proteins initially are expressed as intracellular proteins they are released from cells in distinct manners. The IL-1α precursor is active and is released from dying cells (Dinarello, 2009). In contrast the IL-1β precursor requires processing by the inflammasome and caspase-1 before it is secreted by cells in a purine receptor P2X7 dependent manner (Franchi et al., 2009). IL-1α and IL-1β active expression of many immunologically important genes through the transmembrane receptor complex, comprising IL-1RI and IL-1R accessory protein (IL-1RAcP), which is present on most, if not all, cell types (O'Neill, 2008). The intracellular domain of IL-1RI is homologous to the intracellular domains of the TLRs (O'Neill, 2008). Due to this similarity these distinct receptor classes activate largely the same intracellular signaling pathways and transcription factors, including MAP kinases, AP-1 and NF-κB (O'Neill, 2008). Consequently, such diverse phenotypes as fever, vasodilation, proliferation, differentiation, adhesion and migration (Akira et al., 2006; Dinarello, 2009) are synchronized to ensure an appropriate immune response.
It has been established that keratinocytes, the main constituents of the epidermal layer of the skin, express TLRs and produce cytokines when the cells are exposed to TLR ligands ((Lebre et al., 2006) and refs. therein). However, the specific involvement of keratinocytes in initiating and orchestrating immune responses to skin infections, including S. aureus colonization, is largely unknown. We here report how S. aureus stimulates production of neutrophil chemoattractive cytokines in human keratinocytes.
Neutrophils play an essential role in eliminating S. aureus from the host (DeLeo et al., 2009; Mölne et al., 2000; Verdrengh and Tarkowski, 1997). It was recently demonstrated that IL-1β, but not IL-1α, is required for recruitment of neutrophils to a subcutaneous S. aureus infection (Miller et al., 2007). Due to the experimental model this study did not address the role of the epidermis and specifically the keratinocytes. Given the capability of S. aureus to colonize the skin and many infections start as abrasions of the skin we were interested in evaluating the role of human keratinocytes in recruiting neutrophils. Using microarray analyses of chemokine mRNA expression in response to IL-1 we have previously observed that mRNAs encoding the neutrophil targeting chemokines CXCL1 and CXCL2 were up-regulated by IL-1 (Sanmiguel et al., 2009). However, these observations were not further examined.
To validate the CXCL1 and CXCL2 specific array data primary human neonatal keratinocytes were treated with increasing concentrations of IL-1β. After 1.5, 3, 6 and 24 hours production of mRNA and protein was determined using real-time RT-PCR and ELISA, respectively. Expression levels of CXCL1 and CXCL2 mRNA and protein were induced in time- and IL-1β concentration-dependent manners (Fig. 1a and b). Interestingly, after only 1.5 hours rapid bursts in CXCL1 and CXCL2 mRNA levels were observed in IL-1β treated keratinocytes (Fig. 1a). Subsequently levels of the CXCL1 and CXCL2 mRNAs gradually declined. At the 1.5-hour time-point the CXCL1 and CXCL2 mRNA levels were approximately 30- and 15-fold higher (respectively, p < 0.01) in 5 ng/ml IL-1β treated keratinocytes than in medium only treated cells (Fig. 1a).
Significantly increased secretion of CXCL1 (p < 0.05) and CXCL2 (p < 0.01) could be detected already at 1.5 hours post-IL-1β treatment (Fig. 1b). Despite declining mRNA levels after 1.5 hour, CXCL1 and CXCL2 protein production continued to increase throughout the duration of the experiment (Fig. 1b). The differences in mRNA and protein expression profiles likely reflect differential stabilities of these molecules. At the 24-hour time-point levels of CXCL1 and CXCL2 in medium from cells treated with 5 ng/ml IL-1β were approximately 25- and 5-fold, respectively, higher (p < 0.01) than in medium from cells not receiving IL-1β.
It has previously been shown that keratinocyte production of IL-8 (also known as CXCL8) is induced by S. aureus (Mempel et al., 2003; Sasaki et al., 2003). Due to the proximity of the microarray targets for mRNAs encoding IL-8 and the highly expressed Toll-interacting protein, TOLLIP, our previous expression analyses could not evaluate the expression of IL-8 (Sanmiguel et al., 2009). Expression of IL-8 mRNA and protein was examined in this current study as described above. In contrast to the expression of the CXCL1 and CXCL2 mRNAs (Fig. 1a), IL-8 mRNA levels peaked 3–6 hours post-stimulation. Approximately 25-fold increases (p < 0.01) in IL-8 mRNA expression were observed in 5 ng/ml IL-1β stimulated keratinocytes compared to untreated cells after 3 and 6 hours (Fig. 1c). Levels of IL-8 protein in medium from cells treated with IL-1β increased in a time- and concentration dependent manner throughout the experiment (Fig. 1c). At the 24-hour time-point approximately 12-fold (p < 0.05) higher levels of IL-8 were observed in medium from cells treated with 5 ng/ml IL-1β than cells treated with medium only (Fig. 1c).
We have previously demonstrated that IL-1α and IL-1β activate gene expression in the same manner in keratinocytes (Sanmiguel et al., 2009). In this study IL-1α and IL-1β also induced similar CXCL1, CXCL2 and IL-8 mRNA and protein expression profiles (data not shown). IL-1 induced expression of CXCL1, CXCL2 and IL-8 was confirmed using several independent batches of neonatal and adult primary keratinocytes and the stable cell lines HEK001 and KERTr (data not shown).
Bacteria are recognized by the host TLRs (Ishii et al., 2008) which activate an intracellular signaling cascade shared by IL-1RI (O'Neill, 2008; Akira et al., 2006; Boraschi and Tagliabue, 2006). We therefore wondered if S. aureus would activate expression of the same CXC chemokines as IL-1 (Fig. 1). It is known that keratinocytes express antimicrobial peptides, including the anti-staphylococcal β-defensin 3, which is up-regulated by S. aureus (Menzies and Kenoyer, 2006). To ensure a consistent degree of keratinocyte activation throughout our experiments we chose to use heat-killed bacteria. Primary keratinocytes were treated with increasing amounts of HKSA and expression of CXCL1, CXCL2 and IL-8 examined as described above after 1, 3, 6 and 24 hours (Fig. 2). Time- and concentration-dependent increases in all three CXC chemokine mRNA levels were observed following HKSA stimulation (Fig. 2a). Curiously, HKSA induced mRNA levels peaked later (Fig. 2a) than in response to IL-1 (Fig. 1a and c). The CXCL1 mRNA levels reached maximum induction (2- to 5-fold, p < 0.05) after 3 hours with all 3 amounts of HKSA (Fig. 2a). Levels of the CXCL2 and IL-8 mRNAs peaked (9- and 32-fold induction, respectively, p < 0.01) at 6 hours with the highest concentration of HKSA (Fig. 2a).
In line with the observed shift in maximum CXC chemokine mRNA induction by HKSA compared to IL-1, significantly increased protein expression was also slightly delayed (Fig. 2b compared to Fig. 1b and c). The highest levels of CXCL1, CXCL2 and IL-8 induced by HKSA after 24 hours (Fig. 2b) were comparable to those produced in response to IL-1 (Fig. 1).
LTA and PGN from S. aureus are known to activate gene expression in macrophages (Takeuchi et al., 1999; Hashimoto et al., 2006). To evaluate if the LTA and PGN, derived from S. aureus, activate CXC chemokine expression in keratinocytes in a manner similar to that induced by the bacterium, cells were in addition to HKSA (Fig. 2) treated with LTA and PGN in increasing amounts (Fig. S1). RNA and protein expression was evaluated after 1, 3, 6 and 24 hours. Both LTA and PGN led to increased CXCL1, CXCL2 and IL-8 (Fig. S1 a–c) mRNA expression. The mRNA levels peaked at 3–6 hours (Fig. S1 a–c) in agreement with that observed for HKSA (Fig. 2a). Secretion of CXCL1, CXCL2 and IL-8 (Fig. S1 d–f) protein was also induced by both LTA and PGN. While LTA activated gene expression at lower concentrations than PGN, PGN gave rise to a higher degree of induction than LTA (Fig. S1). Similar observations were made using several independent batches of neonatal and adult primary keratinocytes, and the stable cell lines HEK001 and KERTr (data not shown). These observations demonstrate that S. aureus can activate neutrophil targeting CXC chemokine production in keratinocytes possibly through the bacterial cell wall components LTA and PGN.
It has previously been demonstrated that S. aureus activates expression of antimicrobial peptides (Menzies and Kenoyer, 2006) and IL-8 (Mempel et al., 2003; Sasaki et al., 2003) in human keratinocytes through TLR2 (Mempel et al., 2003; Menzies and Kenoyer, 2006). Curiously, it has also been reported that neutrophil recruitment to subcutaneous S. aureus infection sites in mice was independent of TLR2 (Miller et al., 2006).
To test if TLR2 is involved in driving CXC chemokine expression in keratinocytes exposed to S. aureus and its subcomponents, KERTr cells were incubated with a monoclonal (MAb) human TLR2 specific antibody (Fig. 3a–c). Control cells received isotype matched mouse IgG2a. After two hours medium, HKSA, LTA or PGN was added to the medium and cells incubated for an additional 20 hours after which CXC chemokine secretion into the medium was examined. A borderline significant (p = 0.07) 32% decrease in HKSA induced CXCL1 expression (Fig. 3a) was observed in samples from MAb hTLR2 incubated cells compared to control cells. CXCL1 production in response to LTA and PGN was significantly (p < 0.01 and p < 0.05, respectively) reduced by approximately 50% in the presence of MAb hTLR2 compared to control IgG2a (Fig. 3a). HKSA, LTA and PGN induced production of CXCL2 (Fig. 3a) and IL-8 (Fig. 3a) by cells incubated with MAb hTLR2 was all also significantly (p < 0.01 or p < 0.05) reduced by 30–50% compared to cells incubated with mouse IgG2a.
To verify the involvement of TLR2 the above experiments were repeated using a polyclonal (PAb) antibody against hTLR2 (PAb, Fig. 3d–f). While increased expression (p < 0.01) of CXCL1 (Fig. 3d), CXCL2 (Fig. 3e) and IL-8 (Fig. 3f) was observed in response to HKSA, LTA and PGN in the presence of normal rat IgG there was no induction of these chemokines in the presence of the PAb hTLR2 (Fig. 3d–f).
Lower concentrations of the antibodies (MAb and PAb and appropriate control IgGs) led to less dramatic, but significant, inhibition of CXC chemokine expression (data not shown). Similar data was obtained using primary keratinocytes and HEK001 cells (data not shown). These observations demonstrate that TLR2 regulates CXC chemokine expression in keratinocytes exposed to S. aureus and cellular components thereof.
Our previous microarray analyses of inflammatory gene expression in keratinocytes identified IL1A and IL1B as putative target genes regulated by IL-1 (Sanmiguel et al., 2009). In analogy with the similar induction of CXC chemokines by IL-1 (Fig. 1) and S. aureus (Fig. 2 and Fig. S1) we speculated that IL-1α and IL-1β production could be regulated in response to HKSA, LTA and/or PGN. To test this hypothesis expression of IL-1α and IL-1β mRNA and protein was examined (Fig. 4 and Fig. S2) as described above. Only a modest 1.5-fold increase in IL-1α (p < 0.05, Fig. 4a) and no change in IL-1β (Fig. 4b) mRNA levels were observed in cells treated with the highest amount of HKSA after 1 hour. Surprisingly, dramatic increases in IL-1 secretion were observed after only 1 hour (Fig. 4c and d). An approximately 6-fold increase in IL-1α (p < 0.01, Fig. 4c) and a 15-fold increase in IL-1β (p < 0.01, Fig. 4d) protein levels were observed at the 1-hour time-point in medium from cells treated with HKSA. Similar effects were observed with LTA and PGN (Fig. 4 and Fig. S2). Given the lack of a comparable mRNA induction at the same early time-point these observations strongly suggest that pre-formed and intracellular stored pools of IL-1α and IL-1β protein are released from keratinocytes immediately following exposure to S. aureus.
Significantly increased (1.5- to 3.5-fold, p < 0.05) IL-1α and IL-1β mRNA levels were observed at the 3, 6 and/or 24 hour time-points following HKSA, LTA and PGN treatment (Fig. 4 and Fig. S2) compared to medium only treated cells. These increases in mRNA levels correlated with 1.5- to 3.5-fold elevated protein levels (p < 0.05, Fig. 4 and Fig. S2), although a transient decrease (compared to the 1 hour time-point) in protein levels was observed at the 3-hour time-point (Fig. 4c and d).
Within individual samples approximately 10-fold more IL-1α than IL-1β was detected (Fig. 4 c and d). Based on the RT-PCR CT values IL-1α mRNA levels were approximately 8-times higher than IL-1β mRNA levels in untreated cells (data not shown).
In summary IL-1α is expressed at higher levels than IL-1β at both the mRNA and protein level in keratinocytes. Following exposure to S. aureus, or its subcomponents, keratinocytes rapidly, within 1 hour, release IL-1α and IL-1β into the culture medium. Subsequently IL-1α and IL-1β mRNA levels are increased and de novo synthesized protein is secreted.
The observed CXC chemokine mRNA expression profiles induced by IL-1 (Fig. 1) and S. aureus (Fig. 2 and Fig. 3) differed with respect to the time at which the mRNA levels peaked (before 3 hours for IL-1 and between 3 and 6 hours for HKSA, LTA and PGN). Furthermore, increased IL-1 secretion was present within an hour of HKSA, LTA or PGN stimulation (Fig. 4 and S2). These observations led us to hypothesize that IL-1 secretion and subsequent activation of IL-1RI on the keratinocytes contribute to the S. aureus induced chemokine expression.
The involvement of IL-1RI was examined by incubating primary adult keratinocytes with polyclonal antibodies specific for human IL-1RI (PAb hIL-1RI) prior to treatment with medium only, LTA or PGN (Fig. 5a–c). Control cells were incubated with normal goat IgG. Efficacy of the antibody was verified by evaluating inhibition of IL-1α and IL-1β induced production of CXCL1, CXCL2 and IL-8 (Fig. 5d).
Pre-incubation of cells with PAb hIL-1RI led to an approximately 50% reduction (p < 0.01) in constitutive expression of CXCL1 (Fig. 5a), CXCL2 (Fig. 5b) and IL-8 (Fig. 5c) compared to cells receiving the control IgG. Cells treated with PAb hIL-1RI completely failed to respond to PGN with increased production of CXCL1 and CXCL2 (Fig. 5a and b). PGN induced production of IL-8 in these cells amounted to only approximately 80 pg/ml above levels generated by PAb hIL-1RI and medium only treated cells. In comparison normal goat IgG and PGN treated cells produced approximately 250 pg/ml more IL-8 than normal goat IgG and medium only treated cells (Fig. 5c). While PAb hIL-1RI treated cells did respond to LTA the levels of increased production were significantly lower than those induced in appropriately matched control cells (Fig. 5a–c). For example normal goat IgG and LTA treated cells secreted approximately 5 ng/ml more CXCL1 than cells treated with normal goat IgG and medium only (Fig. 5a). In comparison only approximately 0.9 ng/ml more CXCL1 was released from PAb hIL-1RI and LTA treated cells than the PAb hIL-1RI and medium only reference cells (Fig. 5a). Similar observations were made using HEK001 and KERTr cells (data not shown). The above data demonstrate that LTA and PGN mediated CXC chemokine expression requires IL-1RI engagement.
To investigate further the mechanism whereby IL-1RI mediates S. aureus induced CXC chemokine production we explored whether the IL-1 secreted by the keratinocytes contributed to the gene activation. Since we observed that inhibition of IL-1RI signaling reduced constitutive production of the CXC chemokines (Fig. 5a–c) we initially examined the effect of IL-1α and IL-1β on constitutive CXCL1, CXCL2 and IL-8 expression. Cells were treated with polyclonal antibodies specific for human IL-1α (PAb hIL-1α) or a monoclonal antibody recognizing human IL-1β (MAb hIL-1β). Control cells were incubated with appropriate species and isotype matched normal IgG. While neutralization of IL-1α reduced chemokine expression by approximately 50% (p < 0.01) the IL-1β specific antibody did not affect chemokine expression (Fig. 6a).
To confirm the activities and specificities of the antibodies, cells were pre-treated with PAb hIL-1α, MAb hIL-1β, or appropriate control IgGs. After 2 hours cells were further treated with medium only, IL-1α or IL-1β and chemokine expression determined after an additional 20 hour incubation. As expected PAb hIL-1α blocked IL-1α induced chemokine expression (p < 0.01) but had no effect upon IL-1β signaling (Fig. 6b). The MAb hIL-1β was confirmed to efficiently (p < 0.01) neutralize 10 ng/ml IL-1β (Fig. 6b), a concentration significantly exceeding that produced by the keratinocytes (Fig. 4d).
One may speculate that the apparent effect of the PAb hIL-1α could be due to a non-specific recognition of an unrelated protein. To confirm the role of IL-1α in regulating constitutive production of CXC chemokines cells were treated with a monoclonal mouse anti-human IL-1α antibody (MAb hIL-1α) or normal mouse IgG2a before analyses of CXCL1, CXCL2 and IL-8 expression (Fig. 6c). In agreement with the observations made with the polyclonal antibody chemokine expression by MAb hIL-1α treated cells were approximately 30% (p < 0.01) of that observed with mouse IgG2A treated cells (Fig. 6c). Similar data were obtained using the stable cell lines HEK001 and KERTr (data not shown). The above observations establish that IL-1α, but not IL-1β, regulate constitutive CXC chemokine secretion by keratinocytes.
Variation in constitutive production of the CXC chemokines was observed between different batches of both neonatal (Fig. 1–Fig. 2) and adult primary keratinocytes (Fig. 6–Fig. 7) and the stable cell lines KERTr (Fig. 3) and HEK001 (data not shown). These lot-to-lot disparities likely reflect gene polymorphisms in CXCL1, CXCL2 or IL8 and/or genes regulating expression of these, e.g. IL1A.
Studies of subcutaneous S. aureus infections in mice have indicated that IL-1RI and IL-1β, but not IL-1α, are required for neutrophil recruitment (Miller et al., 2006; Miller et al., 2007). Since keratinocytes secrete elevated levels of IL-1 in response to HKSA, LTA and PGN (Fig. 4 and Fig. S2) we wondered if this could be an autocrine mechanism regulating increased production of the neutrophil targeting chemokines CXCL1, CXCL2 and IL-8. We initially performed neutralization experiments as described above testing the involvement of IL-1β. Surprisingly, we found that the functionally active (Fig. 6b) MAb hIL-1β did not affect chemokine induction (Fig. 7a). In contrast, both antibodies directed against IL-1α significantly (p < 0.05) inhibited HKSA, LTA and PGN activated CXCL1, CXCL2 and IL-8 expression (Fig. 7b and c). The levels of increased production of the three chemokines in the presence of the IL-1α neutralizing antibodies were decreased to 25–50% (p < 0.05) of that observed with control cells treated with normal IgG (Fig. 7b and c). Similar observations were made using HEK001 cells (data not shown). Our data demonstrate that both constitutive and S. aureus induced neutrophil targeting CXC chemokine expression in keratinocytes is regulated by an endocrine signaling pathway involving IL-1α, but not IL-1β.
It is well established that neutrophils play an important role in eliminating S. aureus from the host in general ((DeLeo et al., 2009; Verdrengh and Tarkowski, 1997) and refs. therein) and from the skin specifically (Mölne et al., 2000). S. aureus infections may initially involve colonization of the skin and/or the nasal cavity and compromised epithelial barrier function (van Belkum, 2006). When the skin is injured the keratinocytes are likely to be the first to encounter an invading potential pathogen. To ensure that the challenge is cleared before the microorganism can multiply and spread the keratinocytes must send danger signals into the surrounding tissue to summon for example neutrophils. The data reported here demonstrate that S. aureus exposed keratinocytes produce chemokines (CXCL1, CXCL2 and IL-8, Fig. 2 and and3)3) which are chemotactic to neutrophils. Interestingly, the TLR2 dependent (Fig. 3) chemokine induction requires an endocrine IL-1α, but not IL-1β, signaling loop (Fig. 6–7). Initially it may be puzzling that although the keratinocytes produce both IL-1α and IL-1β only IL-1α has a functional effect. However, it should be noticed that the keratinocytes secrete significantly (p < 0.01) more IL-1α (Fig. 4c) than IL-1β (Fig. 4d) and the levels of IL-1β synthesis may be insufficient to significantly affect gene expression.
It has previously been reported that IL-1 signaling, via IL-1RI, is required for neutrophil recruitment to a subcutaneous S. aureus challenge in mice (Miller et al., 2006). It was furthermore demonstrated that IL-1β, but not IL-1α, was the functional isoform (Miller et al., 2007). We chose to examine specifically the role keratinocytes play as first responders to microbial challenges. Our data demonstrate that chemokine expression in keratinocytes is regulated differently than in the (un-indentified) cell types involved in the in vivo model, i.e. it is TLR2 and IL-1α, but not IL-1β, dependent. The different outcomes (IL-1α versus IL-1β dependence) of using various models may reflect repetition, yet variation, in the immune system which is required to ensure proper protection against immune evasion mechanisms developed by pathogenic microorganisms and genetic variation in the host. S. aureus has acquired several virulence factors aimed at inhibiting neutrophil chemotaxis (reviewed in (Kraus and Peschel, 2008; Rooijakkers et al., 2005)). A multi-layered IL-1 system involving keratinocytes and IL-1α (as reported here) as initial responders in the skin and IL-1β as the activator utilized by other cell types in the epidermis and underlying tissues (Miller et al., 2007) may provide a fail-safe scheme where one mechanism can substitute for the other should this fall short of its task. It should be noted that we do observe expression of IL-1β in keratinocytes following activation by S. aureus (Fig. 4). The levels of IL-1β expressed by the keratinocytes appear to be too low to have a functional effect in our model (Fig. 6 and and7).7). However, there may be situations where IL-1α is inactive, e.g. mutations in the host, or becomes depleted. In these situations IL-1β may substitute for IL-1α in the epidermis.
The review literature on TLR and IL-1RI signaling often mention that these receptors activate the same (Akira et al., 2006; Boraschi and Tagliabue, 2006) or similar signaling pathways (O'Neill, 2008). Not surprisingly, we observed that the TLR2 ligands (HKSA, LTA and PGN, Fig. 2–Fig. 3) increased expression of the same CXC chemokine genes activated by IL-1 (Fig. 1). Given that the used TLR2 ligands also led to secretion of IL-1 (Fig. 4) we anticipated that IL-1 would be partially responsible for the TLR2 ligand induced CXC chemokine expression. Surprisingly, we found that TLR2 ligand induced CXC chemokine expression could be completely blocked by antibodies directed against IL-1RI (Fig. 5). This implies that TLR2 signaling does not directly regulate CXC chemokine expression, but requires IL-1 secretion and signaling as an intermediate step (TLR2 → IL-1 → IL-1RI → CXC chemokine). This may further indicate that the IL-1RI and TLRs do not strictly activate the same signaling pathways. The intracellular regions of IL-1RI and the TLRs contain a conserved protein-protein interaction domain involved in binding the adapter protein MyD88 (Akira et al., 2006; O'Neill, 2008; Dinarello, 2009). One known difference between IL-1RI and TLR2 signaling is the specific requirement for the adapter protein toll-interleukin 1 receptor (TIR) domain containing adaptor protein (TIRAP) in TLR2, but not IL-1RI, signaling. TIRAP facilitates the recruitment of MyD88 to the TLR2 receptor complex (Akira et al., 2006; O'Neill, 2008). It has been demonstrated that knockout of TIRAP impairs TLR2 (and TLR4) mediated cytokine expression in macrophages and dendritic cells. Induction of the same cytokines by other TLR ligands or IL-1 was not affected (Horng et al., 2002; Yamamoto et al., 2002). It is possible that TIRAP, through an unknown mechanism could facilitate a differential gene expression profile compared to those induced by MyD88 restricted pathways. An alternative, or perhaps complementary, hypothesis is that scaffolds play an essential role in regulating which genes are activated in response to certain stimuli. Such scaffolds could involve the Pellino proteins as previously described (Jensen and Whitehead, 2003a, b), or some of the many proteins involved in regulating MAP kinase signaling (reviewed in (Morrison and Davis, 2003)).
TLR2 requires a co-receptor, TLR1 or TLR6, for ligand binding; however, the specific functional involvement of these co-factors remains controversial (Fournier and Philpott, 2005). We observed that hTLR2 MAb had only a borderline significant effect upon CXCL1 expression (Fig. 3a) whereas the hTLR2 PAb abolished CXCL1 expression (Fig. 3d). It is likely that the used mono- and polyclonal antibodies have differential effects upon heterodimer formation between TLR2 and the co-factors TLR1 and TLR6. Furthermore, the co-factors may have diverging roles in down-stream signaling. Further studies are required to determine if TLR1 and TLR6 have specific functions in terms of the range of chemokines expressed in response to distinct S. aureus derived ligands.
Interestingly, we observed variations in the constitutive levels of the three neutrophil chemotactic chemokines CXCL1, CXCL2 and IL-8 (compare Fig. 1–Fig. 2 and Fig. 6–Fig. 7, medium only treated samples) between lot numbers of primary cells. This could be due to genetic variations in the CXC chemokine genes or genes regulating their expression. Differences in constitutive expression of chemokines, and other innate immune factors, may determine whether an individual becomes an intermittent or chronic carrier of S. aureus. Furthermore, as the starting degree of expression may influence the levels during inflammation, polymorphisms in the CXC chemokine genes, or their regulators, may also affect the outcome of an infection. Further, comprehensive studies involving genome wide analyses of gene expression and polymorphisms may reveal patterns associated with S. aureus carrier status and infections. Analyses of such profiles in the context of further characterization of the multifaceted immune system activated in response to S. aureus may reveal mechanisms whereby effective prophylactic and/or treatment strategies can be achieved.
Neonatal foreskin (pooled) and adult torso human primary keratinocytes were obtained from Invitrogen (Carlsbad, CA). The human keratinocyte cell lines HEK001 (CRL-2404) and CCD 1106 KERTr (KERTr, CRL-2309) were obtained from American Type Culture Collection (Manassas, VA). Cells were maintained and treated in Defined Keratinocyte Serum Free Medium supplemented with 50 µg/ml gentamicin (Invitrogen).
Cell cultures were grown to approximately 80–90% confluence and treated with medium only, IL-1α (National Cancer Institute, Frederick, MD), IL-1β (PeproTech, Rocky Hill, NJ), heat-killed Staphylococcus aureus (HKSA, Invivogen, San Diego, CA), purified lipoteichoic acid (LTA, Invivogen) and peptidoglycan (PGN, Invivogen) as indicated in figures. The HKSA, LTA and PGN were all derived from S. aureus strain ATCC 6538. Bacteria were purchased heat-killed from Invivogen. Briefly, bacteria were killed by one or more 15 min cycles of autoclaving at 121°C. Cells were pelleted by centrifugation and resuspended in water. Lack of viability was confirmed by solid and liquid state cultures for 5 days.
Mouse monoclonal IgG2a (MAb hTLR2) and polyclonal rat (PAb hTLR2) antibodies directed against human TLR2 were obtained from Invivogen. Polyclonal rabbit antibody against human IL-1α (500-P21A) was from PeproTech. Monoclonal mouse antibodies specific for human IL-1α (IgG2a, MAB200) and IL-1β (IgG1, MAB601) and polyclonal goat antibodies against human IL-1RI (AF269) were obtained from R&D Systems (Minneapolis, MN). Purified normal goat IgG, mouse IgG1, mouse IgG2A, rabbit IgG and rat IgG were obtained from R&D Systems. For inhibition of TLR2 signaling, cells were pre-incubated with 50 µg/ml MAb hTLR2, 100 µg/ml PAb hTLR2, normal mouse IgG2a or normal rat IgG for 2 hours. In IL-1 signaling neutralization experiments cells were pre-incubated with 10 µg/ml antigen specific antibodies or species and isotype matched control antibodies for 2 hours. After pre-incubation medium only, HKSA (109 cells/ml final conc.), LTA (10 µg/ml final conc.) or PGN (250 µg/ml final conc.) was added to the antibody containing medium and incubation continued for an additional 20 hours before chemokine expression was determined by ELISA. All experiments were performed at least three times with similar outcomes.
Total RNA was extracted using the RNeasy purification system according to the manufacturer’s instructions (Qiagen, Valencia, CA). Reverse transcription and real-time PCR was performed as described elsewhere (Sanmiguel et al., 2009). In brief, 1 µg total RNA was reverse transcribed using AMV reverse transcriptase (Promega, Madison, WI) and oligo(dN)6 primer (GE Healthcare). Real-time RT-PCR was performed using RT2 Real-Time SYBR Green PCR Master Mix (SABiosciences Corp.) on an Opticon2 instrument (Bio-Rad, Hercules, CA). Primer pairs (Table S1) specific for individual mRNA/cDNAs were designed such that PCR products (80–100 bp) span exon-exon junctions thereby preventing amplification of genomic DNA. Assays were validated using serial dilutions and confirmation of equal amplification efficiencies of the cDNA of interest and the GAPDH cDNA. Fold differences in expression were calculated using the Comparative CT method by standardizing against GAPDH expression and comparing expression in cytokine or TLR ligand treated cells to expression in cells treated with medium only.
CXCL1 and IL-1β concentrations in culture medium from stimulated cells were determined using the appropriate DuoSet ELISA Development Kits (R&D Systems). CXCL2, IL-1α and IL-8 levels were determined using ELISA Development assays from PeproTech.
Data are shown as mean values and standard deviations from one representative experiment of at least three independent experiments. Data were analyzed using the Student’s t test when appropriate.
This work was supported by National Institutes of Health Grant AR053672, National Scientist Development grant 0535212N from the American Heart Association and the Transdisciplinary Program in Translational Medicine and Therapeutics at the University of Pennsylvania.
CONFLICT OF INTEREST
The authors state no conflict of interest.