Intracellular Bacterial Infection-Induced IFN-γ Is Critically but Not Solely Dependent on Toll-Like Receptor 4-Myeloid Differentiation Factor 88-IFN-αβ-STAT1 Signaling1

Infection of murine bone marrow-derived macrophages (BMMφ) with Chlamydia pneumoniae induces IFN-αβ-dependent IFN-γ secretion that leads to control of the intracellular bacterial growth. Enhanced growth of C. pneumoniae in Toll-like receptor (TLR) 4−/− and myeloid differentiation factor (MyD) 88−/− (but not TLR2−/−, TLR6−/−, or TLR9−/−) BMMφ is shown in this study. Reduced accumulation of IFN-α and IFN-γ mRNA was also observed in TLR4−/−- and MyD88−/−-infected cells. IL-1R and IL-18R signaling did not account for differences between MyD88−/− and wild-type BMMφ. Surprisingly, infection-induced NF-κB activation as well as TNF-α, IL-1, or IL-6 mRNA expression were all normal in TLR4−/− and MyD88−/− cells. Phosphorylation of the transcription factor STAT1 during bacterial infection is IFN-αβ dependent, and necessary for increased IFN-γ mRNA accumulation and chlamydial growth control. Signaling through common cytokine receptor γ-chain and RNA-dependent protein kinase both mediated IFN-αβ-dependent enhancement of IFN-γ mRNA levels. Accumulation of IFN-γ mRNA and control of C. pneumoniae growth required NF-κB activation. Such NF-κB activation was independent of IFN-αβ, STAT1, and RNA-dependent protein kinase. In summary, C. pneumoniae-induced IFN-γ expression in BMMφ is controlled by a TLR4-MyD88-IFN-αβ-STAT1-dependent pathway, as well as by a TLR4-independent pathway leading to NF-κB activation.

M ammalian Toll-like receptors (TLRs) 3 constitute a family of closely related transmembrane, primary signal-transducing proteins that respond to an array of microbial products (1). They recognize different pathogen-associated molecular patterns such as LPS, flagellin, unmethylated CpG motifs in DNA, dsRNA, mycobacterial lipoarabinomannan, yeast zymosan, and bacterial lipoproteins (1). Upon pathogen-associated molecular pattern recognition, the intracellular domains of all known TLRs interact with the adaptor molecule myeloid differentiation factor (MyD) 88 and initiate a common signaling cascade that leads to nuclear translocation of the transcription factors NF-B and AP-1. This signaling cascade can also be activated upon IL-1 or IL-18R engagement (2). TLR can induce IFN-␣␤ in the presence of MyD88 (3)(4)(5) or in its absence (6,7) by using the adaptor named Toll-IL-1R domain-containing adaptor molecule/ TIR domain-containing adaptor inducing IFN-␤ (8 -10).
IFN-␣␤ are key immunoregulatory cytokines produced directly after cell exposure to many pathogens. IFN-␣␤ interfere with virus replication through induction of, e.g., double-stranded RNA-dependent protein kinase (PKR) and 2Ј-5Ј-oligoadenylate synthetase (2Ј5ЈOAS). Besides the antiviral effect exerted by inhibition of eukaryotic protein synthesis, PKR plays a catalytic or structural role to activate I-B kinase or directly phosphorylate I-B. Stressactivated protein kinases p38 and c-Jun kinase are also regulated by PKR in a pathway that leads to production of proinflammatory cytokines (11).
Binding of IFN-␣␤ to their cellular receptor results in phosphorylation of STATs by receptor-associated Janus kinases, resulting in formation of homodimeric (STAT1 ⅐ STAT1), heterodimeric (STAT1 ⅐ STAT2), or heterotrimeric (STAT1 ⅐ STAT2 ⅐ IFN regulatory factor-9) protein complexes. These multimeric complexes translocate to the nucleus, where they bind to distinct DNA elements, leading to expression of IFN-inducible genes (12). The key macrophage-activating cytokine, IFN-␥, mediates resistance against intracellular bacteria and protozoa via activation of antimicrobial effector molecules (13). Although involved in different responses, both IFN-␣␤R and IFN-␥R signaling share in common STAT1. In turn, both IFNs have partially overlapping biological effects.
The obligate intracellular Gram-negative bacterium Chlamydia pneumoniae is a common cause of acute respiratory disease (14) and has been associated with development of atherosclerosis (15). Chlamydia are internalized by macrophages, but by avoiding phagolysosomal fusion are able to replicate intracellularly. IFN-␥ is central in resistance to this pathogen both in vivo and in vitro (reviewed in Ref. 16). Several studies show that, besides NK and T cells, myeloid cells such as macrophages, dendritic cells (DCs), and neutrophils can also express IFN-␥ (17). In accordance, we have shown that mouse bone marrow-derived macrophages (BMM) infected in vitro with C. pneumoniae express IFN-␥ that in turn protects these cells against chlamydial growth. This IFN-␥ production was IL-12 independent, but required IFN-␣␤ (18).
DCs, smooth muscle cells, macrophages, endothelial cells, and PBMCs are all activated by chlamydial infection or acellular chlamydial components in a TLR2-or TLR4-dependent way (19 -22). However, further details of such activation, in particular with regards to IFN-␥ ⑀ expression, are unknown.
IFN-␣␤-dependent induction of IFN-␥ in splenocytes and T cells has been suggested to be STAT4 dependent (23,24). However, IFN-␣␤ might also activate IFN-␥ in an IL-15-mediated way: Expression of IL-15 depends on presence of a functional IFN-␣␤R (25,26), and IFN-␥ secretion by splenic DCs and macrophages was markedly increased after treatment with IL-15 (27). In this study, we have explored how C. pneumoniae recognition and the ensuing signaling pathways lead to protective IFN-␥ expression in BMM.

Generation of mouse BMM
Mouse BMM were obtained from 6-to 10-wk-old mice, as described (18). Mice were euthanized, and the femur and tibia of the hind legs were dissected. Bone marrow cavities were flushed with 5 ml of cold, sterile PBS. The bone marrow cells were washed and resuspended in DMEM (Sigma-Aldrich, St. Louis, MO) containing glucose and supplemented with 2 mM L-glutamine, 10% FCS, 10 mM HEPES, 100 g/ml streptomycin, 100 U/ml penicillin (all from Sigma-Aldrich), and 20 -30% L929 cellconditioned medium (as a source of M-CSF). Bone marrow cells were passed through a 100-m cell strainer, plated in six-well plates (1.2 ϫ 10 7 cells/well, 2 ϫ 10 6 cells/ml), and incubated for 7 days at 37°C, 5% CO 2 . Before use, BMM cultures were washed vigorously to remove nonadherent cells. Cells from several wells were also harvested and counted by trypan blue exclusion. Typically, bone marrow cells from one mouse yielded 2-3 ϫ 10 6 BMM after 7 days in culture. We have previously shown by immunofluorescence staining that these BMM are F4/80 ϩ , CD14 ϩ , and Mac-3 ϩ (18).

Generation of fibroblasts from mouse lung
Primary fibroblast cultures were generated, as described (40). The pulmonary vasculature was perfused, and lungs were aseptically removed, excised into small pieces, and subjected to collagenase digestion at 37°C for 15 min under agitation. The resulting cell suspension was collected. Enzymatic digestion and collection of cell suspensions were repeated twice. Cell suspensions were then pooled, passed through a 100-m cell strainer, washed, and plated in tissue culture flasks with IMDM supplemented with 2 mM L-glutamine, 5% FCS, 10 mM HEPES, 100 g/ml streptomycin, and 100 U/ml penicillin at 37°C, 5% CO 2 . Fresh medium was added every 7 days until fibroblast cultures attained confluence, ϳ7-14 days later. Lung primary fibroblasts were then washed vigorously, trypsinated, and replated onto six-well plates (2.5 ϫ 10 5 cells/well, 0.42 ϫ 10 5 cells/ml).

Infection and infectivity assay
Mycoplasma-free C. pneumoniae isolate Kajaani 6 (41) was propagated in HEp-2 cells. Bacteria were stored in small aliquots in sucrose-phosphateglutamate (SPG) solution at Ϫ70°C until further use. The infectivity as measured by inclusion-forming units (IFU) of bacterial preparation was determined in HEp-2 cells, as described below.
Cultures of BMM were infected with C. pneumoniae by centrifugation for 1 h, 500 ϫ g at 35°C. A multiplicity of infection of 3 was used for both  GAC TCA TCT GCT GCT TGG AAT GCA ACC CTC  GAC TCA CTC CTT CTC CTC ACT CAG TCT TGC C  IFN-␥  TCG ATC TTG GCT TTG CAG CTC TTC CTC ATG GC  TGC ACC TGT GGG TTG TTG ACC TCA AAC TTG GC  IL-1␣  ATG GCC AAA GTT CCT GAC TTG TTT  CCT TCA GCA ACA CGG GCT GGT C  IL-6  ATG AAG TTC CTC TCT GCA AGA GAC T  CAC TAG GTT TGC CGA GTA GAT CTC  IL-15 CAT

Competitive RT-PCR assay
Cultures of C. pneumoniae-infected BMM were disrupted in RNAzol B (Nordic Biosite, Täby, Sweden), and total RNA was isolated according to the instructions of the manufacturer and reversed transcribed into cDNA, as described (42). Specific primer pairs for IFN-␣, IFN-␥, inducible NO synthase (iNOS), PKR, 2Ј5ЈOAS, IL-1␣, IL-6, IL-15, TNF-␣, and ␤-actin were used to amplify cDNA. Amplified cDNAs were visualized in ethidium bromide-stained 2% agarose gels or quantified by competitive PCR assays. Briefly, competitor fragments with a different length, but using the same primers as the target cDNA, were constructed using composite primers and an exogenous DNA fragment (43). Three-to 4-fold serial dilutions of the competitor were amplified in the presence of a constant amount of cDNA. Reactions were conducted for 23-41 cycles in a thermal cycler (PerkinElmer, Cetus, CT) using an annealing step of 65°C for 2Ј5ЈOAS; 62°C for IL-15, 60°C for IFN-␣, IFN-␥, iNOS, IL-1␣, IL-6, and ␤-actin; and 54°C for PKR and TNF-␣. All primer sequences are given in Table I.

STAT1 is necessary for enhanced IFN-␥ mRNA levels and controls growth of C. pneumoniae in BMM
Biological effects of IFN-␣␤ and IFN-␥ are in part mediated via STAT1 (30). Phosphorylation of STAT1 is noted in C. pneumoniae-infected WT BMM, but not in uninfected controls (Fig.   4). The level of phosphorylated STAT1 was relatively diminished in infected IFN-␥R Ϫ/Ϫ and undetectable in IFN-␣␤R Ϫ/Ϫ BMM (Fig. 4). Growth of C. pneumoniae was higher (Fig. 5A) and IFN-␥ mRNA accumulation lower (Fig. 5C) in STAT1 Ϫ/Ϫ BMM compared with WT controls. PKR, 2Ј5ЈOAS, and iNOS mRNA levels as well as NO in culture supernatants were all reduced in STAT1 Ϫ/Ϫ BMM, while mRNA levels of IL-1␣, IL-6, and TNF-␣ were not or only slightly affected as compared with WT BMM (Fig. 5D, and data not shown). Primary cultures of lung fibroblasts generated from STAT1 Ϫ/Ϫ mice also displayed enhanced C. pneumoniae growth compared with WT controls (Fig.  5B).

Discussion
We show in this study that macrophage infection with C. pneumoniae induces IFN-␥ mRNA accumulation and bacterial growth control in a TLR4-MyD88-IFN-␣␤-STAT1-dependent manner (Fig. 10). PKR and ␥ c R signaling participate downstream of IFN-␣␤R in C. pneumoniae infection-induced IFN-␥ expression and bacterial growth control. However, accumulation of IFN-␥ mRNA does not solely depend on signals via TLR4-IFN-␣␤ pathway. NF-B activation was observed in C. pneumoniae-infected macrophages and DCs (19,21). We now show that NF-〉 activation after C. pneumoniae infection is required for IFN-␥ expression and chlamydial growth control, but can take place in a TLR4-and IFN-␣␤R-independent way. C. pneumoniae-induced proinflammatory cytokine transcripts are neither reduced in MyD88 Ϫ/Ϫ nor TLR4 Ϫ/Ϫ BMM (Fig. 10).
In MyD88 Ϫ/Ϫ macrophages, nuclear translocation of NF-B and phosphorylation of the mitogen-activated protein kinases (MAPK) in response to LPS are somewhat delayed (32,49). In contrast, TLR4 Ϫ/Ϫ macrophages show no NF-〉 and MAPK activation after exposure to LPS (34). Likewise, no activation of NF-〉 or c-Jun kinase was observed in MyD88 Ϫ/Ϫ macrophages in response to other TLR ligands such as peptidoglycan, lipoprotein, CpG DNA, or the imidazoquinolines (reviewed in Ref. 1).
Activation of NF-〉 can, however, occur in a TLR4-and MyD88-independent manner in C. pneumoniae-infected BMM. Thus, a TLR-independent signaling pathway probably also participates in bacterial induced NF-〉 activation. It is also unlikely that other TLR mediate MyD88-independent proinflammatory cytokine expression (e.g., via Toll-IL-1R domain-containing adaptor molecule/TIR domain-containing adaptor inducing IFN-␤) in our experimental model, because expression of proinflammatory cytokines in response to different specific TLR ligands is all reduced in MyD88 Ϫ/Ϫ cells (32,49). However, TLR2 has been shown to be involved in TNF-␣ and IL-6 secretion in Chlamydia trachomatisinfected macrophages (50) and C. pneumoniae-infected DCs (19).
Activation of NF-〉 is needed for IFN-␥ expression, and thereby for growth control of C. pneumoniae. In line with this, NF-〉-binding elements have been identified in the IFN-␥ promoter important for enhancement of gene transcription (51,52). However, an IL-18-independent role for NF-B in IFN-␥ gene expression has not been previously reported.
Together, our results thus suggest that Chlamydia can induce both TLR-dependent and -independent pathways that cooperate for IFN-␥ expression. TLR-analogous detection systems for microorganisms inside cells have been described (53). Nucleotide-binding oligomerization domain proteins in the cytoplasm are candidate among the molecules for such detection systems inside cells and recognize products from both Gram-positive and -negative bacteria (53)(54)(55). We plan to study the role for nucleotide-binding oligomerization domain proteins in IFN-␥ secretion and chlamydial growth control.
In other system, listerial infection and LPS stimulation were shown to activate IFN regulatory factor 3 inducing the synthesis of IFN-␤ and IFN-␣4 (7,56). IFN regulatory factor 3 and IFN-␣␤ can be induced after activation of certain TLR in both MyD88dependent and -independent ways (3, 4, 6, 10). However, MyD88independent signaling was not sufficient for enhanced IFN-␣ mRNA accumulation in C. pneumoniae-infected BMM.
Consistent with the importance of the Janus kinase-STAT pathway in mediating the actions of IFN-␣␤, mice lacking either STAT1 (30,31) or STAT2 (57) have impaired IFN-␣␤-regulated responses and are highly sensitive to viral infection. However, IFN-␣␤ can also control cellular functions in a STAT1-independent fashion (26, 58 -61): STAT1 Ϫ/Ϫ mice are thus more resistant to virus infection than mice lacking expression of both IFN-␣␤R and IFN-␥R (26, 58 -61). We found that STAT1 was critical for the control of C. pneumoniae by BMM. IFN-␣␤R signaling fully  accounted for C. pneumoniae-induced STAT1 phosphorylation. Moreover, IFN-␣␤-dependent STAT1 signaling was necessary for IFN-␥ secretion. To our knowledge, our report is the first indicating the latter. Paradoxically, STAT4 activation was shown to be the critical intermediary in the induction of IFN-␥ in IFN-␣␤stimulated T cells (24), while STAT1 was found to be a negative regulator of IFN-␥ in the same system (62).
PKR, responsible for IFN-␣␤-dependent antiviral effects, also functions as a signal transducer in the proinflammatory response to many agents (11). Our results suggest that PKR is activated in C. pneumoniae-infected BMM in an IFN-␣␤-dependent manner and that it plays a role in IFN-␥ expression and control of infection. However, such a protective role(s) of PKR (and IFN-␣␤) does not depend on the NF-B-activating properties of the enzyme. Whether PKR mediate control of chlamydial infection via MAPK activation remains to be explored.
In summary, our results indicate that TLR4-MyD88-dependent and TLR4-MyD88-independent signaling are both critical and complementary for IFN-␥ expression in macrophages after intracellular bacterial infection. Our results indicate the simultaneous presence of high diversity and nonredundancy in the signals required for this process. A novel pathway of IFN-␥ induction mediated by STAT1 activation is also demonstrated by our data.