|Home | About | Journals | Submit | Contact Us | Français|
Yersinia enterocolitica is a food-borne pathogen that preferentially infects the Peyer's patches and mesenteric lymph nodes, causing an acute inflammatory reaction. Even though Y. enterocolitica induces a robust inflammatory response during infection, the bacterium has evolved a number of virulence factors to limit the extent of this response. We previously demonstrated that interleukin-1α (IL-1α) was critical for the induction of gut inflammation characteristic of Y. enterocolitica infection. More recently, the known actions of IL-1α are becoming more complex because IL-1α can function both as a proinflammatory cytokine and as a nuclear factor. In this study, we tested the ability of Y. enterocolitica to modulate intracellular IL-1α-dependent IL-8 production in epithelial cells. Nuclear translocation of pre-IL-1α protein and IL-1α-dependent secretion of IL-8 into the culture supernatant were increased during infection with a strain lacking the 70-kDa virulence plasmid compared to the case during infection with the wild type, suggesting that Yersinia outer proteins (Yops) might be involved in modulating intracellular IL-1α signaling. Infection of HeLa cells with a strain lacking the yopP gene resulted in increased nuclear translocation of pre-IL-1α and IL-1α-dependent secretion of IL-8 similar to what is observed with bacteria lacking the virulence plasmid. YopP is a protein acetylase that inhibits mitogen-activated protein kinase (MAP kinase)- and NF-κB-dependent signal transduction pathways. Nuclear translocation of pre-IL-1α and IL-1α-dependent secretion of IL-8 in response to Yersinia enterocolitica infection were dependent on extracellular signal-regulated kinase (ERK) and p38 MAP kinase signaling but independent of NF-κB. These data suggest that Y. enterocolitica inhibits intracellular pre-IL-1α signaling and subsequent proinflammatory responses through inhibition of MAP kinase pathways.
There are three species of Yersinia pathogenic for humans, including the two enteric pathogens Y. enterocolitica and Y. pseudotuberculosis as well as Y. pestis, the causative agent of plague (20). Y. enterocolitica and Y. pseudotuberculosis are both food-borne pathogens that infect the Peyer's patches and mesenteric lymph nodes, causing a self-limiting infection (11, 12, 20). Initially, the bacteria attach to and invade M cells, which make up a specialized intestinal epithelium that overlays the Peyer's patches (27, 31). In rare cases, often in the context of immune compromise, systemic infections involving most body systems can occur (12). Yersinia infection is characterized by an acute inflammatory response that is initiated by proinflammatory cytokines, leading to the recruitment and activation of neutrophils and macrophages (14–16, 21–23). Ultimately, a CD4+ T-helper type 1 response clears the infection (1–3).
Using animal models and cell culture, we and others demonstrated that interleukin-1 (IL-1) plays a critical role in initiating the inflammatory response to Y. enterocolitica infection (5, 6, 23). The IL-1 family consists of proinflammatory cytokines and includes a number of molecules important for the host response to Y. enterocolitica infection, such as IL-1α, IL-1β, and IL-18 (5–8, 19, 23, 41). These cytokines are produced as preproteins that require proteolytic cleavage to remove the propiece prior to secretion. IL-1 family members are differentially processed, with IL-1β and IL-18 being substrates of caspase-1 and the inflammasome and IL-1α being cleaved by calpain (19, 41). Mature IL-1 family members are secreted from cells, and they subsequently act to initiate inflammatory signaling on a variety of cell types. Unlike IL-1β and IL-18, pre- and pro-IL-1α are biologically active, utilizing a nuclear localization sequence (NLS) at amino acids 79 to 86 to translocate from the cytoplasm to the nucleus, where IL-1α enhances the transcription of other proinflammatory cytokines, such as IL-8 (17, 38). Nuclear pre-IL-1α is known to interact with proteins associated with the transcriptional machinery, including necdin, GAL4, and histone acetyltransferase (13, 26, 37). It is now hypothesized that the predominant role of IL-1α is as an intracellular signaling molecule. In addition to IL-1α being a nuclear factor, translocation of IL-1α to the nucleus may serve as a means of limiting inflammation during necrosis, when pro-IL-1α can function as a danger-associated molecular pattern (DAMP) molecule.
Even though Yersinia infection leads to acute inflammation as part of the host response, Y. enterocolitica has evolved numerous mechanisms to temper the host's inflammatory response (20). Immune evasion molecules utilized by Yersinia are encoded on both the chromosome and the 70-kDa virulence plasmid (pYv). Certain strains of Y. enterocolitica encode three distinct type three secretion systems (TTSS), including chromosomal and flagellar TTSS, but the best-studied immune modulating mechanisms are associated with the pYv-encoded TTSS and associated effector proteins (18, 24, 40). TTSS allow Yersinia to directly secrete effector proteins from the bacteria directly into the cytoplasm of host cells. The TTSS effector proteins known as Yops are enzymes that mimic host proteins such as phosphatases, kinases, GTPase-activating proteins (GAPs), acetylases, and proteases that impact host cell physiology by disrupting signal transduction pathways and the cytoskeleton (18). YopP (YopJ in Y. pestis and Y. pseudotuberculosis) is a protein acetylase that ultimately inhibits NF-κB, extracellular signal-regulated kinase (ERK), p38, and Jun N-terminal protein kinase (JNK) signal transduction pathways by acetylating activating serine and threonine residues on the activating kinases in these pathways (9, 33, 34). The action of YopP has a variety of consequences depending on the cell type being infected, but YopP can lead to the inhibition of proinflammatory cytokine production (tumor necrosis factor alpha [TNF-α]) and in macrophages promotes apoptosis (4, 9, 32, 35). In mouse models of Y. enterocolitica infection, deletion of YopP has a measurable impact on virulence, but it is not an essential virulence factor in the highly mouse virulent serogroup 0::8 strains (36).
Infection of human epithelial cells with Y. enterocolitica leads to the secretion of IL-8, and following infection in the mouse model, there is a rapid increase in the mouse IL-8 homologues KC and Mip-1α (28). More recently, it was shown that during a Chlamydia trachomatis infection, IL-8 production was dependent on pre-IL-1α intracellular signaling (17). Based on these observations, we investigated the hypothesis that pre-IL-1α was responsible or partially responsible for the IL-8 observed after Y. enterocolitica infection of epithelial cells and that pre-IL-1α intracellular signaling might be a target for Yersinia-mediated immune modulation.
Yersinia enterocolitica strains used in this study are derivatives of the serogroup 0::8 strain 8081 and have been described previously (21, 22). JB-580v is the virulent, wild-type (WT) strain, and JB-580c is an isogenic avirulent derivative of JB-580v lacking the pYv virulence plasmid. The yopP mutant has an in-frame deletion of yopP on the JB-580v background. Bacteria were grown in LB broth containing 20 μg/ml nalidixic acid at 28°C overnight, diluted into fresh media, and cultured at 37°C for 2 h prior to use.
HeLa cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS) at 37°C in a 5% CO2 atmosphere. Cells were cultured in 24-well culture dishes on glass coverslips prior to analysis. Cells were infected with bacteria at a multiplicity of infection (MOI) of approximately 5 for 30 min. Cells were then washed to remove nonadherent bacteria, fresh media containing 100 μg/ml gentamicin were added for 30 min to kill extracellular bacteria, and then cells were maintained in media with 8 μg/ml gentamicin.
We have described the pre-IL-1α–red fluorescent protein (RFP) construct previously (17). ERK 1/2, p38, and IL-1α short hairpin RNAs (shRNAs) and control shRNAs were purchased from OriGene. For transfection, HeLa cells were transfected using Lipofectamine 2000 with 500 ng of pre-IL-1α–RFP and/or 500 ng of shRNA plasmids 24 or 48 h before use. Efficiency of the RNA interference (RNAi) was determined by reverse transcription-PCR (RT-PCR) (23). HeLa cells were infected with adenovirus at an MOI of 6 for 48 h prior to use. We have described the IκB superrepressor-expressing virus and green fluorescent protein (GFP)-expressing viruses and their use previously (10).
Signal transduction inhibitors that selectively block p38 mitogen-activated protein kinase (MAPK) (SB 202190), MEK (U0126), c-RAF-1, and IκB kinase (IKK)–NF-κB (IKK-2 inhibitor) were purchased from Calbiochem and dissolved in dimethyl sulfoxide (DMSO). DMSO was used as a solvent control in all inhibitor studies. Cells were treated with inhibitor SB-202190 (20 μM), U0126 (10 μM), IKK-2 (100 μΜ), c-RAF-1 (9 nM), or DMSO for 2 h prior to infection and maintained in the media for the course of the experiment. Cells were monitored for cytotoxicity by lactate dehydrogenase (LDH) release assay, and the concentrations of signal transduction inhibitor used in these studies did not induce any detectable signs of cytotoxicity.
HeLa cells were grown and treated as described above before being scrapped from the dish. Cells were washed once with ice-cold phosphate-buffered saline (PBS) and collected at 2,600 rpm for 6 min. Cells were resuspended in 450 μl buffer A (100 mM HEPES, pH 7.9, 100 mM KCl, 0.5 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol [DTT], 1 mM phenylmethylsulfonyl fluoride [PMSF]) and vortexed four times for 30 s. Extracts were collected by centrifugation 8,000 rpm for 8 min. The supernatant was the cytoplasmic fraction. The pellet was then resuspended in 100 μl buffer C (200 mM HEPES, pH 7.9, 500 mM KCl, 5.0 mM EDTA, 1.0 mM EGTA, 1 mM DTT, 1 mM PMSF) by vortexing 3 times for 30 s. This solution was collected by centrifugation 10,000 × g for 12 min at 4°C. The supernatant was the nuclear fraction. Protein concentrations in the fractions were estimated by Bradford assay, and the purity of the fractions was tested by Western blotting for tubulin and lamin to define the cytoplasmic and nuclear fractions, respectively.
Endogenous pre-IL-1α was detected by indirect immunofluorescence. Briefly, HeLa cells were plated on coverslips and treated/infected as indicated. At the indicated time points, the cells were fixed in 2% paraformaldehyde for 20 min prior to being placed in blocking/staining buffer (PBS, pH 7.4, containing 0.025% saponin, 1% bovine serum albumin [BSA], and 0.1% nonfat dry milk). IL-1α was detected using the mouse monoclonal anti-human IL-1α antibody clone MAB4145 from R&D Systems at 1:1,000 and a goat anti-mouse Alexa Fluor 568 secondary antibody at 1:2,000. The anti-IL-1α antibody binds an epitope in proprotein corresponding to the region that is processed to release the mature/secreted protein. Coverslips were counterstained with Hoechst dye to visualize the nucleus. Staining was detected with a Zeiss Axioscope-2 microscope equipped with a digital camera and analyzed with the Axiovision suite of software. To quantify IL-1α nuclear localization, 4 random fields (magnification, 40×)/coverslip were imaged in the red and blue channels. The number of cells in the field with nuclear localization of IL-1α was determined (red nucleus). Then, the total numbers of cells in the field (blue nuclei) were determined. The ratio of cytoplasmic to nuclear IL-1α was obtained and averaged for each field. Each experiment was done in duplicate with at least three independent repeats. Data are presented as the average ± the standard deviation.
An enzyme-linked immunosorbent assay (ELISA) using a DuoSet ELISA kit from R&D Systems determined the levels of IL-8 secreted into the culture supernatant as we have described previously (21). Extracellular IL-1α was assayed for using a DuoSet ELISA kit from R&D Systems. Cellular concentrations of IL-1α were determined using ELISA and expressed as pg cytokine/ml/μg total cellular protein as determined by Bradford assay. Each experiment was done in triplicate with at least two to four independent repeats. Data are presented as averages ± standard deviations. In each figure, the IL-8 ELISA is from the same experiment as the corresponding IL-1α nuclear translocation data.
Data are presented as averages ± standard deviations, and statistical significance was determined by two-tailed analysis of variance (ANOVA) with P values of ≤0.05 considered significant.
IL-1α is made as a proprotein containing a nuclear localization signal (NLS) located from amino acids 79 to 86 in the propiece (Fig. 1A) (38). Infection of HeLa cells with Y. enterocolitica lacking the pYv virulence plasmid (JB-580c) or treatment with lipopolysaccharide (LPS) led to a significant (P ≤ 0.05) increase in nuclear IL-1α relative to that in the nontreated cells or cells infected with wild-type Y. enterocolitica (JB-580v; Fig. 1B and C). At the time points tested, using ELISA, no extracellular IL-1α was detected in the cell culture supernatants in any of the experiments (not shown).
HeLa cells infected with JB-580v were most similar to the untreated controls, suggesting that genes carried on the pYv virulence plasmid might be involved in modulating IL-1α nuclear translocation (Fig. 1B and C). To evaluate the functional significance of nuclear IL-1α, we measured the secretion of IL-8 into the culture supernatant following infection with Y. enterocolitica (Fig. 1D). Levels of IL-8 secretion following Y. enterocolitica infection strongly correlated with nuclear IL-1α; significantly more IL-8 was secreted from cells infected with JB-580c or treated with LPS than from nontreated cells or cells infected with JB-580v (Fig. 1D). Nontreated cells did not secrete detectable levels of IL-8, and throughout the study, IL-1α was not detected in the cell supernatant in any experimental condition tested (data not shown). Altogether, these data suggest that genes carried on the pYv virulence plasmid significantly decrease the amount of IL-1α translocating to the nucleus and the amount of IL-8 secreted into culture supernatant.
The data presented in Fig. 1 strongly suggest that genes carried on pYv modulate intracellular IL-1α signaling. pYv contains a number of genes that encode Yops, secreted virulence factors which modulate the host immune response during infection (18). YopP is well known to inhibit proinflammatory cytokine production by inhibiting MAP kinase- and NF-κB-dependent pathways (34). To determine if the total IL-1α protein concentration changed after infection with Y. enterocolitica, ELISA was performed on HeLa cell extracts infected with various Y. enterocolitica strains. As shown in Fig. 2A, cells infected with a strain lacking pYv or yopP contained more IL-1α/μg cellular protein than cells infected with the WT strain. HeLa cells infected with a Y. enterocolitica strain having an in-frame deletion of the yopP gene had significantly more IL-1α (P ≤ 0.05) in the nucleus than the cells infected with the isogenic wild-type strain JB-580v (Fig. 2B). Interestingly, the levels of nuclear IL-1α corresponding to the yopP mutant and the strain lacking the pYv virulence plasmid (JB-580c) were similar (Fig. 2B), suggesting that YopP modulates nuclear IL-1α translocation. HeLa cells infected with JB-580c or the yopP mutant secreted significantly more IL-8 into the culture supernatant than cells infected with JB-580v (Fig. 2C). To test the compartmentalization of IL-1α after infection by another means, we performed cellular fractionation and measured IL-1α concentrations in the cytoplasmic and nuclear fractions. As shown in Fig. 3A, relatively pure fractions were used in the analysis as determined by α-tubulin for the cytoplasmic fraction and lamin A for the nuclear fraction. Consistent with our immunofluorescence-based assay, cellular fractionation revealed significantly more (P < 0.0005) nuclear IL-1α when cells were treated with LPS or infected with JB-580C or the yopP mutant (Fig. 3B). Likewise, the WT strain, JB-580v, had significantly more (P values of 0.01 to <0.005) cytoplasmic IL-1α than cells infected with JB-580c or the yopP mutant or those treated with LPS (Fig. 3B). Altogether, these data strongly suggest that YopP modulates the nuclear translocation of IL-1α and the subsequent IL-1α-dependent production of IL-8.
Data presented in Fig. 2 and and33 could also be interpreted to suggest that there is no correlation between IL-1α and IL-8 and that bacteria lacking Yops or YopP in particular just make more IL-1α and IL-8. Therefore, to test if nuclear localization of IL-1α following Y. enterocolitica infection was correlated to IL-8 secretion, we employed RNAi technology. Plasmids encoding short hairpin RNAs (shRNAs) directed against the IL-1α transcript were purchased from OriGene and screened by RT-PCR. As shown in Fig. 4A, HeLa cells transfected with an IL-1α overexpression plasmid and shRNA for 24 h showed varying levels of IL-1α-specific RNA inhibition relative to the control shRNA. Specific shRNA 2 (Fig. 4A, lane 4) showed the best inhibition and was used in all subsequent experiments. Treatment of HeLa cells with IL-1α-specific shRNA for 24 h prior to infection significantly (P ≤ 0.05) reduced the levels of IL-8 secreted into the culture supernatant relative to cells treated with control shRNA (Fig. 4B). Longer treatment with shRNA (for 48 h) resulted in a slightly improved inhibition with significantly (P < 0.0005 for all groups) reduced levels of IL-1α and IL-8 compared to those in cells treated with control shRNA (Fig. 5A and B). Using transient transfection of IL-1α-specific shRNA, we were never able to completely remove IL-1α protein from the HeLa cells, but we could consistently reduce it by two-thirds, which corresponded to a two-thirds decrease in IL-8. Altogether, these data suggest that IL-8 production following Y. enterocolitica infection is partly dependent on intracellular IL-1α.
Many proinflammatory cytokines are dependent on transcriptional activation by NF-κB. Indeed, the expression of IL-1 is partially controlled by NF-κB, and NF-κB is often a target of pathogen-mediated immune modulation; for example, Y. enterocolitica uses YopP to inhibit NF-κB through the acetylation of IκB (33, 34). To investigate the role of NF-κB in pre-IL-1α intracellular signaling, we first evaluated cellular responses to infection when cells were pretreated with an IKK-2 inhibitor. Blocking NF-κB activation via chemical inhibitor had no impact on the levels of nuclear IL-1α. As shown in Fig. 6A and B, cells treated with IKK-2 inhibitor and those treated with solvent had nearly identical percentages of nuclear IL-1α and concentrations of IL-8 secreted into the culture supernatant at 6 h postinfection. Further, inhibitor treatment or solvent treatment had no impact on the patterns of pro-IL-1α nuclear localization or IL-8 secretion. For example, irrespective of treatment with inhibitor, cells infected with JB-580c or the yopP mutant or treated with LPS still had more nuclear IL-1α and secreted more IL-8 into the supernatant than cells infected with JB-580v or untreated cells. To determine if the IKK-2 inhibitor was functioning as expected, we treated differentiated U937 macrophages with IKK-2 inhibitor or solvent and then stimulated them with LPS. As expected, macrophages treated with IKK-2 inhibitor produced significantly less TNF-α, an NF-κB-dependent cytokine, than the macrophages treated with solvent, indicating that inhibitor was working as expected (Fig. 6C).
This result was unexpected; therefore, to test the role of NF-κB in pro-IL-1α signaling in another way, we utilized adenoviruses expressing the IκB superrepressor or GFP as previously described (10). Cells were infected for 48 h with nonreplicating adenoviruses expressing the IκB superrepressor or, as a control, GFP prior to being infected with Y. enterocolitica. Consistent with our results from the IKK-2 inhibitor experiments, expression of the IκB superrepressor or GFP had no impact on pre-IL-1α-dependent nuclear localization or IL-8 secretion into the culture supernatant (Fig. 7A and B). To determine if the adenoviruses were functioning as expected, we infected differentiated U937 macrophages with adenoviruses expressing the IκB superrepressor or GFP and then stimulated them with LPS. As expected, macrophages infected with adenoviruses expressing the IκB superrepressor produced significantly less TNF-α than the macrophages infected with a virus expressing GFP, indicating that the viruses were working as expected (Fig. 7C). Altogether, these data strongly suggest that NF-κB plays a minimal role in pre-IL-1α-dependent intracrine signaling events.
As shown in Fig. 2 and and3,3, modulation of nuclear localization and the associated secretion of IL-8 are modulated by genes carried on pYv and more specifically YopP. However, as shown in Fig. 6 and and7,7, pro-IL-1α signaling events are independent of NF-κB in HeLa cells. In addition to NF-κB, YopP also inhibits signaling through MAP kinase pathways. Previous work suggests that YopP/J works on the ERK and p38 pathways to block expression of TNF-α (9, 35). Upstream of p38 and ERK is the c-Raf kinase. To test if c-Raf activation leads to increased nuclear IL-1α and IL-8 secretion, HeLa cells were treated with a c-Raf inhibitor, and then pre-IL-1α nuclear localization and IL-8 production were evaluated. Inhibition of c-Raf led to a 50% decrease in both nuclear pre-IL-1α and IL-8 secretion into the culture supernatant (data not shown). These data suggest involvement of MAP kinase pathways in IL-1α intracrine signaling.
c-Raf is not a target of YopP, and the effects of the c-Raf inhibitor are likely indirect effects on the YopP targets ERK 1/2 and p38. To test the role of kinases downstream of c-Raf, we investigated the ERK 1/2 and p38 MAPK pathways. Involvement of p38 MAP kinase in pre-IL-1α signaling was tested through the use of the selective p38 MAPK inhibitor SB-202190 or DMSO as a solvent control followed by infection with Y. enterocolitica. p38 MAPK inhibition potently and significantly (P, ≤0.05 to ≤0.005) inhibited pro-IL-1α nuclear translocation, reducing levels of nuclear IL-1α to baseline levels, whereas DMSO had no effect (Fig. 8A). Treatment of cells with SB-202190 also significantly (P ≤ 0.005) reduced IL-8 production by approximately 50% compared to the DMSO controls (Fig. 8B). These data were in good agreement with studies using shRNA directed against p38 MAPK, demonstrating significant reductions (P ≤ 0.005 [control versus specific shRNA]) in both nuclear IL-1α and IL-8 (Fig. 8C and D). The reduction in IL-8 following p38 MAPK shRNA or SB-202190 treatment was similar in magnitude to what was observed in cells treated with IL-1α shRNA, suggesting that an IL-1α–p38 MAP kinase-independent pathway contributes to the residual IL-8 observed during infection (Fig. 5 and and8).8). Altogether, these data suggest that p38 MAP kinase is an important signaling molecule leading to pre-IL-1α nuclear translocation.
The involvement of the downstream p38 MAPK in pre-IL-1α nuclear translocation raised the possibility that the MEK 1/2 kinases might also be involved in the activation of pre-IL-1α signaling via ERK 1/2. To test the roles of MEK 1/2 in pre-IL-1α signaling, we treated HeLa cells with the MEK 1/2 inhibitor U0126 and measured IL-1α nuclear translocation and IL-8 production. Similar to what was observed in the p38 MAPK inhibitor studies, treatment of cells with U0126 led to significant (P, ≤0.05 to ≤0.005) decreases in the percentage of nuclear IL-1α and secretion of IL-8 following infection of these cells (Fig. 9A and B). Consistent with the p38 MAPK shRNA analysis, treating cells with shRNA directed at ERK 1/2 reduced the levels of nuclear IL-1α and IL-8 (Fig. 9C and D). Although MEK 1/2 inhibition did not reduce IL-1α nuclear localization to the levels observed with p38 inhibition, these data suggest that both the ERK and p38 pathways contribute to IL-1α intracrine signaling.
Treatment of cells with both p38 MAPK and MEK 1/2 inhibitors or p38 and ERK 1/2 shRNA resulted in significant (P < 0.005) decreases in nuclear IL-1α and IL-8 production following infection with Y. enterocolitica (Fig. 10A to D).
Our data suggest that YopP inhibits nuclear translocation of IL-1α through modification of p38 MAPK and MEK 1/2. These data would also predict that cells pretreated with p38 or MEK 1/2 inhibitors would have more IL-1α in the cytoplasm and less in the nucleus than cells treated with solvent controls when infected with the yopP mutant. To test this directly, HeLa cells were pretreated with p38 MAPK inhibitor (SB-202190), MEK 1/2 inhibitor (U106), p38 plus MEK inhibitor, or DMSO as a solvent control for 1 h. Cells were then treated with LPS or infected with JB-580v or the yopP mutant for 6 h before being fractionated into cytoplasmic and nuclear fractions. As shown in Fig. 11A, there was significantly (P ≤ 0.005) less IL-1α in both fractions when cells were treated with inhibitors, as inhibition of these signaling pathways does decrease the overall levels of IL-1α. More importantly, there was an increase in the nuclear-to-cytoplasmic IL-1α ratio when cells were treated with p38 or MEK 1/2 inhibitors and then infected with the yopP mutant, suggesting an accumulation of cytoplasmic IL-1α.
Altogether, these data suggest that MAPK signaling cascades are involved in IL-1α intracrine signaling, leading to subsequent IL-8 secretion, and that YopP action impacts MAPK signaling to partially modulate these intracrine signaling events (Fig. 8, ,9,9, ,10,10, and and1111).
Yersinia enterocolitica infection causes self-limiting disease characterized by a robust inflammatory response in the Peyer's patches and mesenteric lymph nodes (14). Interestingly, the extensive number of virulence factors that Y. enterocolitica uses to modulate the host's immune response likely mutes the observed inflammatory response. Early during infection, members of the IL-1 family, including IL-1α, IL-1β, and IL-18, are important mediators of the host response (5, 6, 8, 23). Previously, we showed that following Y. enterocolitica infection of the mouse, IL-1α is critical for the generation of gut inflammation (23). Moreover, a number of other studies using the mouse model have highlighted the role of IL-1 family members in the control of Y. enterocolitica infection (5, 6, 8). However, these studies predominantly examined the secreted forms of these proteins, and while IL-1β and IL-18 function as secreted cytokines, IL-1α has a variety of other roles during the host response.
IL-1α has a complex biology, and currently it is accepted that the major role of this protein is as an intracellular signaling molecule and as a DAMP. As diagramed in Fig. 1A, the propiece of IL-1α contains a nuclear localization signal, and Werman et al. demonstrated that this piece of the protein was sufficient to translocate heterologous proteins from the cytoplasm to the nucleus (37). Within the nucleus, IL-1α interacts with the transcriptional machinery (necdin and histone acetyltransferase) (13, 26). Nuclear IL-1α can facilitate the expression of PAI-1 in endothelial cells (29, 30), IL-8 and IL-6 in epithelial cells (17), and presumably other proinflammatory cytokines. The full repertoire of IL-1α-elicited intracrine responses is currently unknown.
IL-1α intracrine signaling is an important part of the host response to infection, leading to the production of the potent neutrophil chemoattractant IL-8. Cheng and colleagues previously demonstrated the ability of IL-1α intracrine responses to lead to increased expression of IL-6 and IL-8 during infection of HeLa cells with the intracellular pathogen Chlamydia trachomatis (17). Indeed, the work presented in this study is complementary to what was observed with Chlamydia and illustrates how Yersinia modulates this aspect of the immune response. Interestingly, following the initial invasion of epithelial cells, Y. enterocolitica is predominantly an extracellular pathogen, making neutrophils prominent cells in the host response to infection. In this report, we demonstrate that YopP tempers intracrine IL-1α signaling and IL-8 production during a wild-type Y. enterocolitica infection. We previously reported that infection of mice with a Y. enterocolitica rovA mutant, which does not induce IL-1α expression, also leads to a dramatic reduction in the levels of neutrophil-mediated inflammation that may also be linked to reductions in nuclear IL-1α and the expression of the IL-1α-dependent cytokines KC, Mip-1, and IL-6 (23, 25).
The first host cells encountered by Y. enterocolitica are intestinal epithelial cells, and presumably epithelial cells are the initial sites of Yop-mediated immune evasion. The binding to and invasion of epithelial cells by Y. enterocolitica can trigger IL-8 production in an invasin-dependent manner (28). Although we did not identify the pathogen-associated trigger of IL-1α nuclear localization and IL-8 production, based on previous work (27), it would be reasonable to speculate that invasin-mediated integrin signaling is involved.
Epithelial cells are an important part of the innate immune system, and as such, they are capable of responding rapidly to microbial insult. In fact, epithelial cells play a central role in the pathogenesis of many gut pathogens, such as Shigella, Salmonella, and Yersinia species. The attachment to and invasion of epithelial cells by Y. enterocolitica likely represent the first opportunity for the host to detect and respond to infection, and epithelial cells also represent an initial site of pathogen-mediated innate immune modulation. Several studies have shown that the TTSS effector protein YopP/J can negatively regulate proinflammatory cytokine expression and impact cell survival in a variety of cell types (4, 35). YopP/J impacts many of the signaling pathways leading to proinflammatory cytokine production, including NF-κB, ERK 1/2, p38, and JNK by acetylating the activation serine or threonine on the activating or terminal kinases in these pathways (34). This conservation of biochemical action allows the virulence factor to inactivate a variety of signaling pathways at once or to block signaling in situations in which different cell types use pathways preferentially. Interestingly, in epithelial cells, signaling through p38 and ERK 1/2 seems to be responsible for pro-IL-1α-dependent IL-8 production, whereas NF-κB does not appear to play a role in this process. The fact that inhibition of both ERK 1/2 and p38 MAPK led to partial reductions in IL-8 and nuclear IL-1α suggests that both pathways, along with yet to be identified pathways suggested by the residual IL-1α and IL-8 present following MAPK inhibition, may be involved in this response. YopP-mediated inhibition of p38 might function partially through posttranslational mechanisms, since p38 can be important for cytokine mRNA stabilization (39), but this remains to be tested during Y. enterocolitica infection.
The mechanisms underlying the ability of YopP to interfere with IL-1α nuclear translocation are unknown. However, it has been reported that IL-1α interacts with histone acetyltransferases in the nucleus (26), and YopP is an acetylase as well, raising the possibility of a direct interaction. Another possibility that is more likely is that MAPKs phosphorylate a protein or proteins that promote IL-1α nuclear translocation. Further investigation will illuminate the mechanisms underlying control of intracrine signaling following Yersinia infection.
Altogether, these data expand our understanding of the signaling pathways and effector molecules impacted by YopP. Moreover, these data give insight into possible mechanisms of IL-1α intracrine signaling during the response to infectious insult. Further investigations will determine how IL-1α intracrine signaling is triggered in response to Y. enterocolitica infection and how YopP modulates these responses to maintain pre-IL-1α in the cytoplasm.
The National Institutes of Health, through grants AI067716 and AI060789 awarded to P.H.D. and AI083387 awarded to S.B. supported this work.
Published ahead of print 14 November 2011