|Home | About | Journals | Submit | Contact Us | Français|
To investigate the respective contributions of TLR versus IL-1R mediated signals in MyD88 dependent control of Mycobacterium tuberculosis, we compared the outcome of M. tuberculosis infection in MyD88, TRIF/MyD88, IL-1R1, and IL-1β-deficient mice. All four strains displayed acute mortality with highly increased pulmonary bacterial burden suggesting a major role for IL-1β signaling in determining the MyD88 dependent phenotype. Unexpectedly, the infected MyD88 and TRIF/MyD88-deficient mice, rather than being defective in IL-1β expression, displayed increased cytokine levels relative to wild-type animals. Similarly, infected mice deficient in caspase-1 and ASC, which have critical functions in inflammasome-mediated IL-1β maturation, showed unimpaired IL-1β production and importantly, were considerably less susceptible to infection than IL-1β deficient mice. Together our findings reveal a major role for IL-1β in host resistance to M. tuberculosis and indicate that during this infection the cytokine can be generated by a mechanism that does not require TLR signaling or caspase-1. The Journal of Immunology, 2010, 184: 3326–3330.
Innate resistance to Mycobacterium tuberculosis is a key factor in disease outcome as only a small percentage of individuals exposed to this pathogen become actively infected (1). Mice deficient in the TLR/IL-1R family receptor adaptor molecule MyD88 have been shown to be highly susceptible to infection with M. tuberculosis, supporting a major role for this pathway in innate defense against the bacterium (2, 3). Nevertheless, there has been considerable controversy as to whether the requirement for MyD88 reflects the involvement of a single or multiple TLRs or instead cytokine signaling via IL-1R or IL-18R (4-7). Despite convincing evidence for the importance of TLR/TLR-ligand interactions in the response to M. tuberculosis in vitro, a number of studies have indicated that IL-1R1 signaling may play a more dominant role in MyD88-dependent resistance in vivo (8-10).
The two known ligands for IL-1R1 are IL-1α and IL-1β. These cytokines, although both induced by NF-κb dependent pathways, have different posttranslational processing requirements. Expression of the highly proinflammatory cytokine IL-1β is tightly controlled through the combination of two distinct triggers: signal 1 that induces the expression of pro-IL-1β and signal 2 that results in IL-1β protein maturation via caspase-1 (11). Signal 1 typically involves triggering by a TLR ligand and IL-1 itself can serve as positive feedback regulator of this pathway. Signal 2, in contrast, involves a number of poorly defined recognition events that lead to the assembly of a protein complex known as the inflammasome that activates caspase-1, thereby cleaving the proform of IL-1β into mature IL-1β (12). The apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC) is a critical component of most inflammasomes.
In this paper, in addition to confirming the major role of IL-1R1 signaling in MyD88-dependent host resistance to M. tuberculosis, we have now identified IL-1β as a critical ligand in this pathway. Unexpectedly, our experiments reveal that although IL-1β induction by M. tuberculosis in vitro depends on TLR triggering and the inflammasome for signals 1 and 2 respectively, both triggers are dispensable for IL-1β production in vivo in mice infected with the pathogen.
C57BL/6 mice were purchased from Taconic Farms (Germantown, NY). All knockout strains were on a full C57BL6 background or were backcrossed for a minimum of six generations. MyD88−/− mice and IL-18−/− were originally obtained from S. Akira (Osaka University, Osaka, Japan) and IL-1β−/− mice from Y.Iwakura (University of Tokyo,Tokyo, Japan).N6 backcrossed IL-1R1−/− mice were first purchased from The Jackson Laboratory (Bar Harbor, ME). Results from these mice were later confirmed using fully backcrossed IL-1R1−/− mice generously provided by Drs. Dirk Schlueter and Horst Bluethmann (Pharmaceutical Research Gene Technology, F. Hoffmann-La Roche, Basel, Switzerland). TLR2/9−/− and TRIF/MyD88−/− were generously supplied by D. Golenbock and E. Lien (University of Massachusetts, Worcester, MA). The TLR2/9−/− mice were fully backcrossed. ASC−/− mice were originally derived at Millennium Pharmaceuticals, Cambridge, MA (13). Casp1−/− mice were generated as described and subsequently backcrossed to C57BL6 (N6) (14). All animals (8–15 wk old of both sexes) were bred in an AALAC-accredited animal facility at the National Institute of Allergy and Infectious Diseases, National Institutes of Health (Bethesda, MD).
Low-dose (50–100 CFU/mouse) aerosol infection with the H37Rv strain of M. tuberculosis and determination of bacterial growth in tissues was performed as previously described (3, 4). Bone marrow-derived macrophages (BMDMs), were generated as described (3, 4) and exposed to H37Rv at a multiplicity of infection of one (MOI:1) in complete media for 48 h after which the supernatants were harvested for ELISA.
Bronchoalveolar lavage fluid (BALF) was obtained by washing the lung airways with 1 ml sterile PBS/mouse. Aliquots were taken for bacterial load determination and after centrifugation the cell-free supernatants assayed by ELISA. Lungs were then perfused with 10 ml sterile PBS and mechanically disrupted by passing through a 100-μm cell strainer in a total volume of 4 ml PBS 1% FCS. An aliquot was removed for determining bacterial load. After centrifugation, the cell-free lung suspensions were analyzed by ELISA. IL-1β was measured using ELISA kits from eBioscience (San Diego, CA) and R&D Systems (Minneapolis, MN) with similar results. IL-1α, IFNγ, TNFα, and IL-1R antagonist were determined with R&D DuoSet kits. High-mobility group box 1 (HMGB1) was detected using the HMGB1 ELISA Kit II from IBL Hamburg (Hamburg, Germany). Total nitrite was measured using the nitrate reduction assay from R&D Systems.
Lungs were fixed by inflating the tissues with 4% formaldehyde, sectioned and stained with H&E or by the Ziehl-Neelsen method to visualize acid-fast mycobacteria.
Total RNA was isolated from lungs, and real-time RT-PCR was performed as previously described (3, 4). The following primer pairs were used: for hprt: 5′-GTTGGTTACAGGCCAGACTTTGTTG-3′ (forward) and 5′-GAGGGTAGGCTGGCCTATAGGCT-3′ (reverse); il-1r1: 5′-TGTAATGAAAGACGGCACACC (forward) and 5′-TCTTCTTTGGGTATTGCTTGG-3′ (reverse)
Cell-free lung suspensions from 4-wk infected wild-type (WT), IL-1β−/−, and casp1−/− mice were loaded in equal volumes and separated on a 12% polyacrylamide NuPAGE gel (Invitrogen, Carlsbad, CA), and the proteins transferred to a Hybond-P PVDF membrane (GE Healthcare, Piscataway, NJ). A goat anti-mouse polyclonal IgG Ab (R&D Systems) was used to detect IL-1β.
To help clarify the respective roles of upstream signals determining MyD88-dependent host resistance to M. tuberculosis, we performed an extensive series of experiments directly comparing the susceptibility of seven different mouse strains deficient in specific TLR/IL-1R family receptors, adaptor molecules or cytokines. Confirming published data (3, 8), MyD88−/− and IL-1R1−/− mice succumbed rapidly to low−dose aerosol infection, displaying 2–3 log increases in bacterial loads in lung and BALF (Fig.1A, 1B) and extensive pulmonary necrosis (Fig.1C, 1D and data not shown) 4 wk postinfection (p.i). No further increase in susceptibility was evident in mice doubly deficient in both MyD88 and the alternative TLR adaptor TRIF (Fig.1A, 1B). Also as previously reported, IL-18−/− and TLR2/9−/− mice succumbed to infection earlier than WT animals (Fig.1A) (4, 15). However, their increase in susceptibility was minor relative to the acute mortality seen in the MyD88/IL-1R1 deficient strains. Importantly, mice deficient in IL-1β alone recapitulated the phenotype observed in the MyD88−/− mice in terms of survival, bacterial load and pulmonary pathology (Fig.1).
The induction of reactive nitrogen species mediated by IFNγ and TNFγ signaling has been demonstrated to be critical for the control of M. tuberculosis infection in mice (1). Despite their increased susceptibility, infected IL-1β−/− mice expressed even higher levels of these mediators in lungs than WT animals (Fig. 2A). In contrast, IL-12/23 p40 expression was reduced in IL-1β−/− lungs (Fig. 2B), indicating that the increase in IFNγ is not due to elevated p40 expression and suggesting that defective IL-1β/IL-1R1 signaling may contribute to the previously reported defect in p40 production in MyD88−/− mice (2-4). Importantly, IL-1α levels were markedly elevated in the infected IL-1β−/− animals, arguing that in the absence of IL-1β, IL-1α is not sufficient to mediate IL-1R1 dependent host resistance (Fig. 2C). Moreover, HMGB1, which, in common with IL-1α, is released by dyingnecrotic cells (16-18), was significantly increased in the lungs of infected IL-1β−/− mice (Fig. 2C). Expression of the endogenous IL-1R antagonist was also highly upregulated in the lungs of both WT and IL-1β−/− mice (Fig. 2D), which could explain why in the latter animals IL-1α is not protective.
The above findings support the hypothesis that IL-1R1 rather than TLR or IL-18 signaling underlies the critical role of MyD88 in early host resistance to M. tuberculosis and for the first time assign to IL-1β a major nonredundant function in acute control of this pathogen. Based on our data, the loss in acute bacterial control observed in IL-1β−/− mice is not due to a downstream defect in the induction of the host-protective mediators IFNγ, TNFα, and NO. In addition, infected IL-1 β−/− mice showed no impairment in M. tuberculosis-specific CD4 T cell priming measured by MHC class II tetramer staining and IFNγ cytokine production after cognate peptide stimulation (K. Mayer-Barber, D. Barber, and A. Sher, manuscript in preparation). Further information concerning the cellular sources and targets of IL-1β in the lungs of M. tuberculosis infected mice and the downstream effects of IL-1β signaling will be important in identifying the specific host-protective pathway involved.
MyD88 could regulate IL-1β/IL-1R1 dependent resistance to M. tuberculosis through its downstream role in IL-1R1 signal transduction and/or its function in upstream TLR/NFκB dependent il1b gene induction as signal 1 in the expression of the cytokine. To address this point, we compared IL-1β levels 4 wk p.i. in lungs and BALF of the MyD88 and IL-1β/IL-1R1 deficient mouse strains analyzed above. Importantly, we found that IL-1β expression was significantly induced in MyD88−/−, TRIF/MyD88−/−, and IL-1R1−/− mice at levels that actually exceeded those seen in WT animals (Fig. 3A). Consistent with the latter findings, significant induction of IL-1β mRNA was also observed in the lungs of IL-1R1−/− and MyD88−/− mice (Fig. 3B). In direct contrast to these in vivo observations documenting MyD88 independent IL-1β synthesis, and consistent with previous work (19), a major defect in IL-1β production measured by the same ELISA assay, was seen when BMDMs from MyD88−/− or TLR2/9−/− mice were stimulated in vitro with M. tuberculosis (Fig. 3C, data not shown).
Together, these observations suggest that MyD88 regulates host resistance to M. tuberculosis at the level of IL-1β/IL-1R1 signaling rather than TLR-mediated triggering of cytokine production and unexpectedly show that IL-1β can be induced in vivo by an unidentified pathway that requires neither TLR/TRIF/MyD88 signaling nor IL-1R1/MyD88 as a positive feedback loop.
Nevertheless, our findings do not rule out a nonessential role of TLR signaling for IL-1β induction in WT mice, particularly at earlier stages of infection. One scenario is that in the knockout mice proinflammatory/necrotic mediators override the usual role of TLR stimulation in IL-1β expression. One major candidate is TNFα, which we found to be highly expressed in lungs of infected TRIF/MyD88−/−, MyD88−/−, and IL-1R1−/− mice (data not shown) and which has recently been demonstrated to induce IL-1β in vitro in the absence of TLR signaling (12, 17).
To examine the requirement for caspase-1 and inflammasome-mediated IL-1β expression (i.e., signal 2) in IL-1β mediated host resistance, we assessed the response of mice deficient in ASC and caspase-1 (casp1−/−) to M. tuberculosis infection. BMDMs from both ASC−/− and casp1−/− mice showed major defects in IL-1β production in response to M. tuberculosis stimulation in vitro (Fig. 4A). In striking contrast, IL-1β expression was unimpaired in M. tuberculosis infected ASC−/− and casp1−/− mice (Fig. 4B) in both cell-free lung suspensions and BALF. Consistent with their unexpectedly normal levels of IL-1β, the majority of both ASC−/− and casp1−/− animals survived the acute phase of infection displaying approximate mean survival times of 6 and 3 mo, respectively (Fig. 4C). When compared with WT animals, pulmonary bacterial burdens were unchanged in ASC−/− mice and moderately but significantly elevated in casp1−/− animals (Fig. 4D). Western blotting of cell-free lung suspension, 4 wk p.i. confirmed that IL-1β is cleaved into its 17 kD mature form in both WT, as well as casp1−/− mice (Fig. 4E).
In contrast to findings in other infectious disease models (20-22), our data argue that the production of mature host-protective IL-1β during M. tuberculosis infection can occur independently of caspase-1 activation and ASC-containing inflammasomes. This process likely involves either a caspase-1 independent pathway for IL-1β processing in macrophages and dendritic cells, and/or a distinct cellular source, such as neutrophils, which can process IL-1β without a requirement for caspase-1 (12, 23, 24). It should be noted that a partial loss in host resistance was observed in both the casp1−/− and ASC−/− mice. This could reflect a contribution of the inflammasome in the processing of IL-1β and/or IL-18, and/or functions of caspase-1 and ASC distinct from cytokine cleavage acting late in infection (15, 25, 26).
Thus, although our findings establish IL-1β as a critical component in innate resistance to M. tuberculosis, they also indicate that the pathways for expression and regulation of the cytokine in M. tuberculosis infection are complex and may involve mechanisms that do not fit the classical paradigms of TLR recognition and inflammasome-mediated caspase-1 processing seen in other infections or for that matter in the response to M. tuberculosis in in vitro systems.
We are grateful to Drs. Y. Iwakura and S. Akira for generously sharing access to genetically deficient mouse strains, originally derived in their laboratories. We also thank Dr. C.G. Feng for scientific discussion and review of the manuscript and the staff of the National Institute of Allergy and Infectious Diseases animal BSL3 facility for technical assistance.
This work was supported by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health. R.A.F. is an Investigator of the Howard Hughes Medical Institute.
The authors have no financial conflicts of interest.