Transforming Growth Factor- (cid:1) -Smad Signaling Pathway Negatively Regulates Nontypeable Haemophilus influenzae -induced MUC5AC Mucin Transcription via Mitogen-activated Protein Kinase (MAPK) Phosphatase-1-dependent Inhibition of p38 MAPK*

In contrast to the extensive studies on the role of transforming growth factor- (cid:1) (TGF- (cid:1) ) in regulating cell proliferation, differentiation, and apoptosis over the past decade, relatively little is known about the exact role of TGF- (cid:1) signaling in regulating host response in infectious diseases. Most of the recent studies have suggested that TGF- (cid:1) inhibits macrophage activation dur-ing infections with pathogens such as Trypanosoma cruzi and Leishmania , thereby favoring virulence. In certain situations, however, there is also evidence that TGF- (cid:1) has been correlated with enhanced resistance to microbes such as Candida albicans , thus benefiting the host. Despite these distinct observations that mainly focused on macrophages, little is known about how TGF- (cid:1) regulates host primary innate TGF- signaling pathway by NTHi-derived TGF- like via a independent of TGF- 1, 2, and 3 autocrine negative regulation of NTHi-induced transcription. our data do rule out of latent TGF- (cid:1) s in extracellular activated NTHi T RII. family the molecular identity of NTHi-derived NTHi the latent TGF- extracellular matrix in of TGF- (cid:1) signaling. In how TGF- up-regulates MKP-1 expression explored using biochemical and genetic approaches. the role of MKP-1 talk between TGF- (cid:1) -Smad pathway and the p38 pathway in confirmed by using MKP-1 knock-out mice. deepen our of the role of TGF- signaling in regulating innate

The TGF-␤ 1 pathway represents a key signaling pathway participating in regulation of diverse biological processes (1)(2)(3)(4)(5)(6)(7). In contrast to the extensive studies on the role of TGF-␤ in regulating cell proliferation, differentiation, and apoptosis, relatively little is known about the exact role of TGF-␤ in regulating host response in infectious diseases. In review of the role of TGF-␤ signaling in infections, most studies have focused on pathogens that infect host macrophages such as Trypanosoma cruzi and a variety of Leishmania species (8). These studies have demonstrated that excessively produced TGF-␤ upon infection inhibits macrophage activation, thereby favoring virulence (9 -11). In certain situations, however, there is also evidence that TGF-␤ has been correlated with enhanced resistance to microbes such as Candida albicans, thus benefiting the host (12). Despite these distinct observations that mainly focused on macrophages, little is known about how TGF-␤ regulates host innate defensive responses, such as upregulation of mucin, in the mucosal epithelial cells of airways (13)(14)(15)(16)(17).
Mucins, the major component of mucous secretions, are high molecular weight and heavily glycosylated proteins synthesized by the mucosal epithelial cells lining the middle ear, trachea, and digestive and reproductive tracts (12)(13)(14)(15)(16)(17). They protect the epithelial surface by binding and trapping inhaled infectious particles, including bacteria and viruses, for mucociliary clearance, at least in part because of the extraordinary diversity of their carbohydrate side chains (13,14). Therefore, up-regulation of mucin in infectious diseases represents an important host innate defensive response to microbes (13). However, in patients with otitis media with effusion and chronic obstructive pulmonary diseases whose mucociliary clearance mechanisms have become defective, excessive production of mucin will lead to airway obstruction in chronic obstructive pulmonary disease and conductive hearing loss in otitis media with effusion (18 -24). To date, 18 mucin genes have been identified (14 -17). Among these, at least MUC2, MUC5AC, and MUC5B have been shown to play an important role in the pathogenesis of respiratory infectious diseases (14,15,(17)(18)(19)(20)(21). We have demonstrated recently that TGF-␤-Smad * This work was supported by National Institutes of Health Grants DC004562, DC005843, and HL070293 (to J. D. L.), CA24321 (to Y. S. K. and J. G.), and GM63773 (to X.-H. F.) and a grant from the Department of Veterans Affairs Medical Research Service (to J. G. and Y. S. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
** To whom correspondence should be addressed: House Ear Inst., 2100 West Third St., Los Angeles, CA 90057. E-mail: jdli@hei.org. signaling pathway cooperates with NF-B to mediate nontypeable Haemophilus influenzae-induced MUC2 mucin transcription (24). Still unknown is whether or not the TGF-␤-Smad signaling pathway regulates MUC5AC, another key member of the mucin superfamily, in a similar manner. Understanding how TGF-␤ signaling mediates up-regulation of MUC5AC mucin may not only bring new insights into the novel role of TGF-␤ signaling in attenuating host primary innate defensive responses and but may also open up novel therapeutic targets for these diseases.
In addition to the TGF-␤-Smad signaling pathway, p38 MAPK, consisting of four isoforms, ␣, ␤, ␥ and ␦, together with its upstream MAPK kinases (MKK3/6), comprises another important signaling pathway mediating diverse cellular responses (25). Recent studies have shown that p38 MAPK is not only regulated by its immediate upstream kinases such as MKK3/6 but is also regulated by other signaling pathways such as TGF-␤ signaling pathway (26,27). In contrast to the relatively extensive studies on the positive regulation of p38 by TGF-␤ signaling, relatively little is known about the negative regulation of p38 MAPK by TGF-␤ signaling, especially in the pathogenesis of bacterial infectious diseases.
Because of the important role of TGF-␤ signaling in mediating diverse cellular responses and our recent observations showing the activation of TGF-␤ and p38 signaling by nontypeable H. influenzae (NTHi) as well as the reported interaction between TGF-␤ and p38 pathways, we hypothesized that the TGF-␤-Smad signaling pathway interacts with p38 to mediate up-regulation of MUC5AC mucin in response to NTHi in human epithelial cells. Here, we showed that NTHi, a major human bacterial pathogen of otitis media and chronic obstructive pulmonary disease, strongly induces up-regulation of MUC5AC mucin (28 -33), a primary innate defensive response for mammalian airways, via activation of multiple signaling pathways. The activation of the TLR2-MyD88-dependent p38 MAPK pathway is required for NTHi-induced MUC5AC transcription, whereas activation of TGF-␤ receptor-mediated signaling leads to down-regulation of p38 MAPK by inducing MAPK phophatase-1 (MKP-1) expression, thereby acting as a negative regulator for MUC5AC induction. These studies bring new insights into the novel role of TGF-␤ signaling in attenuating host primary innate defensive responses and enhance our understanding of the negative cross-talks between the TGF-␤-Smad and p38 MAPK signaling pathways.
Bacterial Strains and Culture Conditions-NTHi strain 12, a clinical isolate, was used in this study (21,24,33). Bacteria were grown on chocolate agar at 37°C in an atmosphere of 5% CO 2 . For making NTHi crude extract, NTHi were harvested from a plate of chocolate agar after overnight incubation and incubated in 30 ml of brain heart infusion broth supplemented with NAD (3.5 g/ml). After overnight incubation, NTHi were centrifuged at 10, 000 ϫ g for 10 min, and the supernatant was discarded. The resulting pellet of NTHi was suspended in 10 ml of phosphate-buffered saline and sonicated. Subsequently, the lysate was collected and stored at Ϫ70°C. NTHi lysates (5 g/ml) were used in all the experiments. We chose to use NTHi lysates because of the following reasons. First, NTHi has been shown to be highly fragile and has the tendency to autolyze. Its autolysis can be triggered in vivo under various conditions including antibiotic treatment (21,24,33). Therefore, using lysates of NTHi represents a common clinical condition in vivo, especially after antibiotic treatment.
Cell Culture-Human colon epithelial cell line HM3 was maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Invitrogen) (21,24). Human cervix epithelial cell line HeLa and human middle ear epithelial cell line HMEEC-1 were maintained as described (17,28). Primary human airway epithelial (NHBE) cells were purchased from Clonetics (San Diego, CA). NHBE cells were maintained in Clonetics' recommended bronchial epithelial growth medium, which includes supplements of bovine pituitary extract, hydrocortisone, human recombinant epidermal growth factor, epinephrine, transferrin, insulin, retinoic acid, triiodothyronine, gentamicin, and amphotericin B (Clonetics), to a confluence of 60 -80% (37°C, 5% CO 2 ). The media were replaced every other day. Only cells of passage 4 were used for experiments. Wild type mink Mv1Lu cells and mutant cell lines, DR26 that are derived from Mv1Lu and lack functional TGF-␤ receptor type II, were kindly provided by Dr. Joan Massague (Memorial Sloan-Kettering Cancer Center, New York), and were maintained as described previously (24,34). All media received additions of 100 units/ml penicillin and 0.1 mg/ml streptomycin.
Reverse Transcription-PCR Analysis of MKP-1-Total RNA was isolated from human epithelial cells using a Qiagen kit (Valencia, CA) following the manufacturer's instruction. For the reverse transcription reaction, the Moloney murine leukemia virus preamplification system (Invitrogen) was used. PCR amplification was performed with Taq gold polymerase (PerkinElmer Life Sciences). The oligonucleotide primers were: MKP-1, 5Ј-GCT GTG CAG CAA ACA GTC GA-3Ј and 5Ј-CGA TTA GTC CTC ATA AGG TA-3Ј, and cyclophilin, 5Ј-CCG TGT TCT TCG ACA TTG CC-3Ј and 5Ј-ACA CCA CAT GCT TGC CAT CC-3Ј.
Real Time Quantitative Reverse Transcription-PCR Analysis-Total RNA was isolated from human epithelial cells using TRIzol® Reagent (Invitrogen) following the manufacturer's instruction. For the real time quantitative reverse transcription-PCR, Pre-Developed TaqMan Assay Reagents (Applied Biosystems) were used. Synthesis of cDNA from total RNA samples was performed with MultiScribe TM Reverse Transcriptase. To normalize MUC5AC expression relative to cDNA, we used primers and a TaqMan probe corresponding to cyclophilin. Expression of MUC5AC was measured relative to cyclophilin. Primers and the TaqMan probe for MUC5AC were designed by using Primer Express software (Applied Biosystems) and synthesized by Applied Biosystem Customer Oligo Synthesis Service (Applied Biosystems). TaqMan probes were labeled with 6-carboxyfluorescein on the 5Ј end and 6-carboxytetramethylrhodamine on the 3Ј end. The primers and probes for MUC5AC were: forward primer, 5Ј-GTTCTATGAGGGCTGCGTCTTT-3Ј; reverse primer, 5Ј-GGCTGGAGCACACCACATC-3Ј; and TaqMan probe, 5Ј-FAM-ACCGGTGCCACATGACGGACCT-TAMARA-3Ј. The reactions were amplified and quantified using an ABI 7700 sequence detector and the manufacturer's software (Applied Biosystems). The relative quantity of MUC5AC mRNA was obtained using the comparative CT method (for details, see user Bulletin 2 for the ABI PRISM 7700 Sequence Detection System under www.appliedbiosystems.com/ support/tutorials) and was normalized using Pre-Developed TaqMan Assay Reagent Human Cyclophilin as an endogenous control (Applied Biosystems).
Plasmids, Transfections, and Luciferase Assays-The expression plasmids fp38␣(AF), fp38␤(AF), MyD88 DN, hTLR2 DN and wild type, ⌻␤RII DN and wild type, Smad3 DN and wild type, Smad4 DN and wild type were previously described (24,33,35). The expression plasmids of the wild type and antisense MKP-1 were kindly provided by Dr. N. Tonks (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY) and Dr. C. Desbois-Mouthon (INSERM U-402, Faculté de Médecine Saint-Antoine, Paris, France), respectively (36). The reporter construct, human MUC5AC regulatory region, was subcloned into upstream of a TK-32 promoter luciferase vector. 2 SBE-luc was previously described (24,37). All of the transient transfections were carried out in triplicate using TransIT-LT1 reagent (Mirus, Madison, WI) following the manufacturer's instructions, unless otherwise indicated. In all co-transfections, an empty vector was used as a control. The transfected cells were pretreated with or without chemical inhibitors including SB203580, cycloheximide, and Ro-31-8220 for 1 h. NTHi or recombinant TGF-␤1 was then added to the transfected cells 42 h after transfection. After 5 h, the cells were harvested for luciferase assay. In experiments using neutralization TGF-␤ antibody, NTHi lysates were pretreated with either TGF-␤ neutralization antibody or control antibody for 1 h before being added to the transfected cells for 5 h.
Western Blot Analysis-In all transfections with either a wild type or a dominant-negative mutant of signaling molecules, an empty vector was used as a control. Transfected cells were treated with or without NTHi 42 h after transfection. Antibodies against phospho-p38 (Thr-180/ 182) and p38 were purchased from Cell Signaling (Beverly, MA). Antibodies against phospho-T␤RII (Tyr-336) and T␤RII were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Phosphorylation of p38 and T␤RII were detected as described and following the manufacturer's instructions (24,33).
Immunofluorescent Staining-The cells were cultured on four-chamber microscope slides. After NTHi or TGF-␤1 treatment, the cells were fixed in paraformaldehyde solution (4%) and incubated with mouse anti-Smad4 monoclonal antibody for 1 h (Santa Cruz Biotechnology, Inc.). Primary antibody was detected with fluorescein isothiocyanateconjugated goat anti-mouse IgG (Santa Cruz Biotechnology, Inc.). The samples were viewed and photographed using a Zeiss Axiophot microscope.  (21). The signaling mechanisms underlying the NTHi-mediated activation of p38 MAPK that leads to up-regulation of MUC5AC, however, have yet to be defined. To investigate how p38 MAPK is regulated by NTHi, we first confirmed the effect of NTHi on MUC5AC transcription and the requirement for p38 MAPK in MUC5AC induction. As we showed previously, NTHi up-regulated MUC5AC expression at mRNA level in human epithelial HeLa and HM3 cells ( Fig 1A) as assessed by performing real time quantitative PCR analysis. In addition, NTHi-induced MUC5AC expression was also observed in human middle ear epithelial HMEEC cells as well as in primary human bronchial epithelial NHBE cells. Moreover, as shown in Fig. 1B, the pyridinyl imidazole SB203580, a highly specific inhibitor for p38 MAPK, greatly inhibited NTHiinduced MUC5AC expression. Consistent with this finding, SB203580 and overexpression of a dominant-negative mutant of p38␣ or p38␤ also inhibited NTHi-induced MUC5AC transcription as assessed using a human MUC5AC promoter-luciferase reporter construct (Fig. 1C), thus confirming the requirement for p38 MAPK in MUC5AC induction.

TLR2-
We next sought to determine the receptor-mediated signaling pathway upstream of p38 that mediates NTHi-induced MUC5AC transcription. Based on our recent finding that NTHi induces p38-dependent activation of NF-B via TLR2 (24, 33), we therefore determined the involvement of TLR2 signaling in MUC5AC induction. Interestingly, co-transfection of HM3 cells with a dominant-negative mutant of TLR2 inhibited NTHiinduced MUC5AC transcription, whereas overexpression of a wild type TLR2 enhanced MUC5AC induction ( Fig. 2A). Similarly, overexpressing a dominant-negative mutant MyD88, a key adapter protein downstream of TLR2 (38), also inhibited MUC5AC induction. Concomitantly, overexpression of dominant-negative mutants of TLR2 and MyD88 also abrogated MUC5AC induction at the mRNA level as assessed by performing real time quantitative PCR analysis (Fig. 2B). To address whether the TLR2-MyD88 signaling pathway also mediates NTHi-induced MUC5AC in primary human bronchial epithelial cells, we next studied the effects of overexpressing dominant-negative mutants of TLR2 and MyD88 on MUC5AC induction in NHBE cells. As shown in Fig. 2C, NTHi-induced MUC5AC expression at mRNA level was also abolished by these treatments in primary NHBE cells. These data indicate that the TLR2-MyD88 signaling pathway is required for MUC5AC induction.
To further determine whether TLR2-MyD88 acts upstream of p38 MAPK, we assessed the effect of overexpressing the same dominant-negative mutants of TLR2 and MyD88 on NTHi-induced p38 phosphorylation by using anti-phosphorylated p38 MAPK antibody. Indeed, co-transfecting the HM3 cells with both dominant-negative mutants inhibited NTHi-induced p38 phosphorylation (Fig. 2D). Thus, it is evident that the TLR2-MyD88 signaling pathway mediates NTHi-induced up-regulation of MUC5AC via activation of p38 MAPK. induced transcription of MUC2 (8,24), another key member of mucin superfamily, we were interested in determining whether TGF-␤-receptor signaling also mediates NTHi-induced up-regulation of MUC5AC mucin, a primary innate defensive response for host respiratory mucosa (13). We first examined the effects of overexpressing the dominant-negative mutant and wild type of T␤RII on NTHi-induced MUC5AC expression (40,41). Surprisingly, co-transfecting the HM3 cells with a dominant-negative mutant of T␤RII greatly enhanced NTHi-induced MUC5AC expression at the mRNA level, whereas overexpressing the wild type T␤RII attenuated MUC5AC induction (Fig. 3A). We next confirmed the negative involvement of T␤RII signaling in MUC5AC induction in primary human bronchial epithelial NHBE cells. As shown in Fig. 3B, co-transfecting NHBE cells with a dominant-negative T␤RII also enhanced MUC5AC induction at the endogenous mRNA level. Thus, these data indicate that T␤RII signaling negatively regulates NTHi-induced MUC5AC expression in human epithelial cells.
To further confirm whether T␤RII signaling indeed acts as a negative regulator for MUC5AC induction, we took advantage of the available DR26 cell, a lung epithelial cell line that is derived from the wild type Mv1Lu cells and that lacks functional T␤RII (24,34). We first assessed the effect of TGF-␤1 on SBE-dependent promoter activity in Mv1Lu and DR26 cells, respectively. As expected, no TGF-␤1-induced promoter activity was observed in mutant DR26 cells, whereas the wild type Mv1Lu cells showed potent induction of SBE-dependent promoter activity by TGF-␤1 (Fig. 3C, left panel). We next examined the effects of NTHi on MUC5AC promoter activity in the same cells, respectively. As shown in Fig. 3C (right panel), almost no MUC5AC induction by NTHi was observed in wild type Mv1Lu cells, whereas NTHi induced MUC5AC promoter activity in mutant DR26 cells by ϳ2.5-fold. In accordance with these results, co-transfecting the wild type Mv1Lu cells with a dominant-negative mutant T␤RII abolished the TGF-␤1-induced SBE response, whereas co-transfecting the mutant DR26 cells with a wild type T␤RII rescued the SBE response to TGF-␤1 (Fig. 3D, upper panels). In contrast to the SBE response to TGF-␤1, overexpressing a dominant-negative mutant T␤RII in wild type Mv1Lu cells rendered the cells MUC5ACresponsive to NTHi, whereas co-transfecting the mutant DR26 cells with a wild type T␤RII abolished the MUC5AC response to NTHi (Fig. 3D, lower panels). Thus, these data suggest that T␤R signaling is a negative regulator for NTHi-induced MUC5AC mucin transcription in human epithelial cells.
Because the negative involvement of T␤R signaling in MUC5AC induction was determined mainly by using overexpression of T␤R expression plasmids and the T␤R mutant cell lines, we next sought to confirm whether NTHi indeed activates TGF-␤ signaling like TGF-␤ does. We first evaluated the effect of NTHi on phosphorylation of T␤RII by using an antibody against phosphorylated T␤RII. As shown in Fig. 3E (left  panel), NTHi, like TGF-␤1, induced phosphorylation of T␤RII. We next investigated whether NTHi activates TGF-␤-Smad signaling by evaluating NTHi-induced nuclear translocation of Smad4, a key step for the Smad3/4 complex to exert its transcriptional activity (2). Fig. 3E (right panel) shows that NTHi potently induces nuclear translocation of Smad4. As expected, Smad4 translocation was also induced by TGF-␤1. To further confirm whether NTHi activates TGF-␤-Smad-dependent transcriptional activity, we assessed the effect of NTHi on SBE-dependent promoter activity by using SBE luciferase reporter in HM3 cells (45). When we exposed the transfected cells to NTHi or TGF-␤1, SBE-driven luciferase activity increased in cells treated with NTHi or TGF-␤1 (Fig. 3E, right panel). Taken together, these results confirm that NTHi, like TGF-␤1, indeed activates TGF-␤-Smad signaling pathway, which in turn leads to the inhibition of NTHi-induced MUC5AC transcription.
Although we have demonstrated that TGF-␤ signaling acts as a negative regulator for NTHi-induced MUC5AC transcription, it is still unclear how TGF-␤ receptor-mediated signaling is activated by NTHi in the regulation of MUC5AC transcription. Based on our recent findings that NTHi did not induce any detectable increase in three major TGF-␤ family members, TGF-␤1, 2, and 3 in the conditioned medium of HM3 cells and that the early NTHi-induced T␤RII phosphorylation was observed at 5 min upon treatment (24), it is likely that NTHi may activate TGF-␤-Smad signaling via a TGF-␤ autocrine-independent mechanism. To determine whether NTHi-derived TGF-␤-like factor is responsible for the negative regulation of NTHi-induced MUC5AC transcription mediated by TGF-␤ signaling, we assessed the effect of NTHi lysates pretreated with TGF-␤ neutralization antibody or control antibody on the transcriptional activity of MUC5AC promoter. As shown in Fig. 3F, NTHi-induced MUC5AC promoter activity was enhanced by TGF-␤ neutralization antibody treatment, whereas NTHi-induced MUC5AC transcription remained unchanged upon treatment with control antibody. Collectively, these data suggest that TGF-␤ receptor-mediated signaling is likely activated by a NTHi-derived TGF-␤-like factor via a mechanism independent of TGF-␤1, 2, and 3 autocrine signaling in the negative regulation of NTHi-induced MUC5AC transcription. It should be noted that our data do not preclude the involvement of the latent TGF-␤s stored in the extracellular matrix that might be activated by NTHi and then cross-talk with T␤R. In addition, it is still unclear whether other TGF-␤ family members may be involved in mediating the negative regulation of NTHi-induced MUC5AC transcription in an autocrine-dependent manner.
Because of the importance of Smads in transducing TGF-␤ receptor-mediated signals into the nucleus (5, 37, 39), we next sought to determine the involvement of Smad3, one of the key receptor-activated Smads, and Smad4, the Co-Smad (common partner Smad). As shown in Fig. 4A, overexpression of a dominant-negative mutant of either Smad3 or Smad4 enhanced NTHi-induced MUC5AC expression, whereas co-transfection with the wild type form of either Smad3 or Smad4 inhibited MUC5AC induction at the mRNA level as assessed by performing real time quantitative PCR analysis in HM3 cells. In addition, activation of TGF-␤-Smad signaling by co-expression of wild type Smad3 and Smad4 markedly abrogated MUC5AC induction at the mRNA level (Fig. 4B). Similar to their inhibitory effect on MUC5AC induction at the mRNA level, coexpression of wild type Smad3 and Smad4 also abolished NTHi-induced MUC5AC at the transcriptional level as evaluated by co-transfecting the HM3 cells with a MUC5AC promoter luciferase reporter construct (Fig. 4C). Collectively, these results demonstrated that TGF-␤ type II receptor-Smad3/4 signaling is negatively involved in NTHi-induced MUC5AC transcription, thereby inhibiting the primary innate defensive response in host airway mucosa.
T␤RII-Smad3/4 Signaling Pathway Negatively Mediates NTHi-induced MUC5AC via a Negative Cross-talk with p38 MAPK-Having identified TLR2-MyD88-p38 MAPK as a positive pathway and T␤R-Smad3/4 as a negative pathway involved in NTHi-induced MUC5AC transcription, it is still unknown whether or not there is a negative cross-talk between these two signaling pathways. To test the hypothesis that TGF-␤ signaling negatively regulates NTHi-induced MUC5AC transcription via down-regulating the p38 MAPK activity, we first investigated the effect of overexpressing a dominant-negative mutant of T␤RII on NTHi-induced p38 phosphorylation. As shown in Fig. 5A, NTHi-induced p38 phosphorylation was greatly enhanced in HM3 cells transfected with a dominantnegative mutant of T␤RII. The enhanced p38 phosphorylation was observed in HM3 cells treated with NTHi for various times. The highest level of the enhanced p38 phosphorylation was observed at 60 min. In addition, the enhancement of NTHiinduced p38 phosphorylation by inhibition of TGF-␤ signaling were also observed in primary bronchial epithelial NHBE cells (Fig. 5B). Similarly, co-transfecting the HM3 cells with a dominant-negative mutant of either a Smad3 or Smad4 also markedly enhanced the NTHi-induced p38 phosphorylation (Fig.  5C). In agreement with these results, addition of exogenous TGF-␤1 in parallel with NTHi attenuated NTHi-induced p38 phosphorylation in both HM3 and primary NHBE cells (Fig. 5,  D and E). Moreover, exogenous TGF-␤1 also inhibited NTHiinduced MUC5AC transcription in a dose-dependent manner (Fig. 5F). These results thus confirmed our hypothesis that T␤R-Smad3/4 signaling indeed acts as a negative regulator for NTHi-induced MUC5AC transcription via inhibiting p38 activation.
TGF-␤-Smad Signaling Negatively Regulates NTHi-induced MUC5AC Induction via a MAPK Phosphatase-1-dependent Inhibition of p38 MAPK-One key issue that has yet to be addressed is how TGF-␤-Smad signaling down-regulates NTHiinduced p38 activation. We first determined the possible involvement of de novo protein synthesis in NTHi-induced MUC5AC expression at the mRNA level by using the protein synthesis inhibitor cycloheximide. As shown in Fig. 6A, NTHiinduced up-regulation of MUC5AC at the mRNA level was further enhanced in HM3 cells pretreated with cycloheximide, indicating that de novo protein synthesis is negatively involved NTHi-induced MUC5AC expression. Based on this finding and the evidence that T␤R signaling down-regulates NTHi-induced p38 phosphorylation, it is logical that T␤R-Smad3/4 signaling may be involved in up-regulation of an inhibitor for p38 MAPK. In review of the known inhibitors for p38, MKP-1, a member of a novel class of dual specificity phosphatases collectively termed MAPK phosphatases, represents a key protein phosphatase that dephosphorylates and inactivates p38 MAPK (36,42). To determine whether T␤R-Smad3/4 signaling is involved in the NTHi-induced MKP-1 expression, we next evaluated the effect of overexpression of a dominant-negative mutant of T␤RII on NTHi-induced MKP-1 expression at the mRNA level. Interestingly, NTHi greatly induced MKP-1 expression at the mRNA level in a time-dependent manner, and the MKP-1 induction was greatly inhibited by overexpressing a dominantnegative mutant of T␤RII (Fig. 6B). A similar result was also observed in primary bronchial epithelial NHBE cells (data not shown). Therefore, T␤R signaling appears to be involved in NTHi-induced MKP-1 expression, which in turn leads to downregulation of p38-dependent MUC5AC transcription.
To confirm the negative involvement of MKP-1 in NTHiinduced MUC5AC up-regulation, we then assessed the effect of Ro-31-8220, a chemical inhibitor for MKP-1 expression (36), on MUC5AC induction by NTHi in HM3 cells. As expected, Ro-31-8220 indeed enhanced NTHi-induced up-regulation of MUC5AC in a dose-dependent manner (Fig. 5C). To further confirm whether MKP-1 is indeed negatively involved in MUC5AC induction, we next investigated the effects of overexpressing an antisense and a wild type full-length MKP-1 construct on NTHi-induced MUC5AC expression (36,42,43). As shown in Fig. 6D, overexpression of the antisense MKP-1 construct enhanced MUC5AC induction, whereas co-expression of a wild type MKP-1 attenuated MUC5AC induction. Taken together, our data demonstrated that TGF-␤-Smad signaling is negatively involved in NTHi-induced MUC5AC induction via MKP-1-dependent inhibition of p38 MAPK.

DISCUSSION
In conclusion, our studies have demonstrated that NTHi, a major human bacterial pathogen of otitis media and chronic obstructive pulmonary disease (28 -33), strongly induces upregulation of MUC5AC mucin, a primary innate defensive response for mammalian airway (13), via activation of multiple signaling pathways (Fig. 7). The activation of TLR2-MyD88-dependent p38 MAPK pathway is required for NTHi-induced MUC5AC transcription, whereas activation of TGF-␤ receptor-Smad3/4 signaling, however, leads to down-regulation of p38 MAPK by inducing MKP-1 expression, thereby acting as a negative regulator for MUC5AC induction.
A major finding in this study is the experimental evidence for the negative regulation of host mucosal defensive response to bacterial pathogen by TGF-␤ signaling. In review of the role of TGF-␤ signaling in infectious diseases (8), most studies have focused on pathogens that infect host macrophages such as T. cruzi and a variety of Leishmania species. These studies have demonstrated that excessively produced TGF-␤ upon infection inhibits macrophage activation, thereby favoring virulence (9 -11). In certain situations, however, there is also evidence that TGF-␤ has been correlated with enhanced resistance to microbes such as C. albicans (12), thus benefiting the host. Despite these distinct observations that mainly focused on macrophages, little is known about how TGF-␤ regulates host innate defensive responses, such as up-regulation of mucin (13), in the mucosal epithelial cells of airway. Therefore, our study may bring new insights into the novel role of TGF-␤ signaling in attenuating host primary innate defensive responses to respiratory bacterial pathogens and may open up novel therapeutic targets for treatment of airway infectious diseases.
Another interesting finding in this study is the negative cross-talk between the TGF-␤-Smad3/4 signaling pathway and the p38 MAPK pathway. Experimental evidence over the past few years has suggested that TGF-␤-Smad pathway may signal through interactions with other signaling pathways such as p38 MAPK. Although most of these studies have demonstrated an important role of TGF-␤ signaling in activation of p38 MAPK (26,27), the negative regulation of p38 MAPK by TGF-␤-Smad signaling still remains largely unknown. In the present study, we provided first hand evidence that activation of TGF-␤-Smad3/4 signaling by bacterium NTHi leads to attenuation of p38 MAPK via up-regulation of MKP-1 in human airway epithelial cells. These observations should further enhance our understanding of the signaling mechanisms underlying the functional cross-talk between the TGF-␤-Smad signaling pathway and the p38 MAPK pathway.
Finally, interesting evidence was also provided for the possible involvement of NTHi-derived TGF-␤-like factor in activation of the TGF-␤-Smad signaling pathway, which in turn leads to the negative regulation of NTHi-induced MUC5AC transcription. Several lines of evidence support this notion. First, the NTHi-induced T␤RII phosphorylation was observed at as early as 5 min (24). Given such an early phosphorylation of T␤RII, it is likely that the early phosphorylation of T␤RII may occur as a result of direct activation of T␤R signaling by NTHi rather than NTHi-induced TGF-␤ autocrine signaling. Although T␤RII is generally known as a serine/threonine kinase, there is also evidence that T␤RII can function as a dual specificity kinase, and tyrosine phosphorylation may have an important role in T␤R signaling (44). Thus, NTHi-induced tyrosine phosphorylation of T␤RII at 5 min may be at least interpreted as a T␤R-mediated early response to NTHi factors. Second, pretreatment of NTHi lysates with the TGF-␤ neutral- ization antibody enhanced its ability to induce the transcriptional activity of MUC5AC promoter as compared with the NTHi lysates treated with control antibody. Finally, NTHi did not induce any detectable increase in three major TGF-␤ family members, TGF-␤1, 2, and 3, in the conditioned medium of HM3 cells. Taken together, these data suggest that TGF-␤-Smad signaling pathway is likely activated by NTHi-derived TGF-␤like factor via a mechanism independent of TGF-␤1, 2, and 3 autocrine signaling in the negative regulation of NTHi-induced MUC5AC transcription. However, our data do not rule out the possible involvement of the latent TGF-␤s stored in the extracellular matrix that might be activated by NTHi and then cross-talk with T␤RII. In addition, it is still unclear whether other TGF-␤ family members are involved in mediating the negative regulation of NTHi-induced MUC5AC transcription in an autocrine-dependent manner. Future studies will focus on determining the molecular identity of NTHi-derived TGF-␤like factors and whether NTHi also activates the latent TGF-␤ stored in the extracellular matrix that in turn leads to the activation of TGF-␤ signaling. In addition, how TGF-␤ signaling up-regulates MKP-1 expression via a Smad3/4-dependent mechanism will be further explored using biochemical and genetic approaches. Finally, the role of MKP-1 in mediating the negative-cross talk between TGF-␤-Smad pathway and the p38 pathway in vivo will also be confirmed by using MKP-1 knockout mice. These studies should deepen our understanding of the role of TGF-␤ signaling in regulating host innate defensive response to respiratory bacterial pathogens.