|Home | About | Journals | Submit | Contact Us | Français|
Primary Mycobacterium tuberculosis infection results in granuloma formation in lung tissue. A granuloma encapsulates mycobacterium-containing cells, thereby preventing dissemination and further infection. Tumor necrosis factor alpha (TNF-α) is a host-protective cytokine during M. tuberculosis infection due to its role in promoting and sustaining granuloma formation. TNF activity is regulated through the production of soluble TNF receptors (sTNFRI and sTNFRII). Therefore, we examined the potential production of endogenous sTNFRs during M. tuberculosis infection. Using the murine model of aerosol M. tuberculosis infection, we determined that levels of sTNFR production were elevated in bronchoalveolar lavage fluid 1 month following infection. An investigation of M. tuberculosis cell wall components identified that the known virulence factor mannose-capped lipoarabinomannan (ManLAM) was sufficient to induce sTNFR production, with sTNFRII being produced preferentially compared with sTNFRI. ManLAM stimulated the release of sTNFRs without TNF production, which corresponded to an increase in TNF-α-converting enzyme (TACE) activity. To determine the relevance of these findings, serum samples from M. tuberculosis-infected patients were tested and found to have an increase in the sTNFRII/sTNFRI ratio. These data identify a mechanism by which M. tuberculosis infection can promote the neutralization of TNF and furthermore suggest the potential use of the sTNFRII/sTNFRI ratio as an indicator of tuberculosis disease.
Mycobacterium tuberculosis infects millions of people worldwide, and the World Health Organization (WHO) estimates that nearly one-third of the world's population are currently infected (56). Upon entering the airways, M. tuberculosis is taken up via complement and mannose receptors on alveolar macrophages. The bacteria inhibit the acidification and maturation of the endosome, thus preventing catalysis and allowing bacterial survival within the macrophage. Infected host alveolar macrophages aggregate to form giant multinucleated cells and can differentiate into epithelioid macrophages. M. tuberculosis-infected macrophages are often characterized by the presence of lipid droplets containing both host and bacterial products in their cytoplasm. The infected macrophages secrete cytokines and chemokines that attract lymphocytes and produce fibrous, extracellular matrix proteins that form the granuloma (21, 37). It is thought that a granuloma functions to encapsulate mycobacterium-containing cells, thereby reducing dissemination and further infection (48).
Tumor necrosis factor alpha (TNF-α) is necessary for the formation and maintenance of a granuloma by the host as well as for inducing antimicrobial responses by macrophages and supporting T cell effector function (3). The dissemination of M. tuberculosis occurs when the integrity of the granuloma is lost or with fluctuations in immune infiltrates in the lung, which are typically due to a reduction in the level of bioactive TNF (46, 48). TNF knockout mice have a reduced capacity to contain infection (6), and patients who receive anti-TNF therapies for autoimmune and inflammatory diseases have been shown to be susceptible to a reactivation and/or dissemination of M. tuberculosis infection (33). TNF is also beneficial to the host through its ability to induce apoptosis in infected macrophages, thereby facilitating the cell-mediated destruction of the bacterium (5).
For most inflammatory conditions, TNF functions primarily to augment cellular activation, cytokine production, and tissue destruction (2). Some of this effect is related to the ability of TNF and TNF receptor I (TNFRI) to provide costimulatory signals during T cell activation and differentiation (20, 57). The regulation of TNF activity is normally accomplished through the production of soluble TNF receptors (sTNFRs). There are two primary soluble TNF receptors, sTNFRI (p55) and sTNFRII (p75). There is conflicting evidence regarding the ability of sTNFRI and sTNFRII to block versus augment TNF-induced signaling through a dose-dependent stabilization of TNF and surface-bound TNF receptor interactions (1, 22). Nevertheless, it seems that soluble TNFRs function primarily as anti-inflammatory components through the neutralization of TNF bioactivity (36). Soluble TNFRII is the biologically active component of etanercept (Enbrel), a therapeutic TNF inhibitor used for rheumatoid arthritis and Crohn's disease, which, among other TNF inhibitors, has also been associated with reactivations of latent M. tuberculosis infection (33, 55).
Both TNF and its receptors TNFRI and TNFRII are membrane-bound proteins that are released from the plasma membrane following cleavage by the TNF-α-converting enzyme (TACE) (ADAM17) (9). Upon activation, TACE cleaves TNF family members between Ala and Val residues at the juxtamembrane domain (8). TACE activity is regulated by the tissue inhibitor of matrix metalloproteinases 3 (TIMP3) (4). TIMP3 production is induced following phosphatidylinositol 3-kinase (PI3K)/AKT pathway activation (45), and following secretion, it binds to the extracellular domain of TACE to inhibit enzymatic activity (4). A recent study demonstrated that mitogen-activated protein kinase (MAPK) pathway activation, specifically p38 and extracellular signal-regulated kinases 1 and 2 (ERK1/2), induces TACE activation by promoting the monomerization of the enzyme and the dissociation of TIMP3 from TACE (58).
The ratio of TNF levels to soluble TNF receptor levels is critical for the regulation of TNF activity at sites of inflammation. Consistent with the requirement of bioactive TNF, Tsao et al. demonstrated previously that imbalances in TNF/sTNFR ratios can occur in the pleural cavity of M. tuberculosis-infected patients (53). In order to investigate this further, we evaluated the production of TNF and soluble TNFRI and TNFRII in the lungs in a mouse model of M. tuberculosis infection. Previous studies have shown that pathogenic M. tuberculosis blocks the apoptosis of host macrophages via the release of sTNFRII, thereby resulting in the inactivation of TNF (5). We confirmed that M. tuberculosis infection induced elevated levels of soluble TNFRs, with a predominance of sTNFRII, in mouse lung lavage fluid at 1 month postinfection.
To address a potential mechanism for the induction of sTNFRs, we tested the abilities of different M. tuberculosis antigens to promote sTNFR production. One virulence factor identified for M. tuberculosis, mannose lipoarabinomannan (ManLAM) (51), was shown previously to modulate cytokine production by lymphocytic and monocytic cells via signaling pathway alterations (38, 39, 50). Based on this evidence, and evidence that ManLAM is an accessible antigen during live M. tuberculosis infection (47), we hypothesized that ManLAM could induce sTNFR production. Here we show that ManLAM alone was sufficient to induce sTNFR production by both mouse and human leukocytes and in an in vivo mouse model. We indentify that the mechanism for this effect is through a reduction in TIMP3 levels with a concomitant hyperactivation of TACE, which we determined is mediated by ManLAM-induced ERK1/2 activation. This process represents a potential mechanism used by M. tuberculosis to limit the effects of TNF. Furthermore, these findings indicate that the ratios of TNF family members in serum might be used as indicators of M. tuberculosis infection.
M. tuberculosis antigens were obtained through NIH Tuberculosis Vaccine Testing and Research Materials Agreement contract no. HHSN266200400091C. All in vitro testing with antigens was performed with 100 ng/ml unless a different dose is noted.
Eight- to twelve-week-old female C57BL/6J mice (Jackson Laboratory stock number 000664) were housed in the animal medicine facility at the University of Massachusetts Medical School. Experimental protocols were approved by the Institutional Animal Care and Use Committee. Mice were infected with 100 CFU M. tuberculosis Erdman (Trudeau Institute Mycobacterial Culture Collection) per mouse by using a Glas-Col inhalation exposure system. Two mice were culled at 24 h postinfection to confirm the delivered dose in every experiment. At 1 month postinfection, bronchoalveolar lavage fluid (BALF) was collected and stored at −20°C.
Eight- to twelve-week-old female C57BL/6J mice (Jax 000664) were housed at the Laboratory Animal Science Center at the Boston University School of Medicine. Experimental protocols were approved by the Institutional Animal Care and Use Committee. Mice were intratracheally instilled with the antigens in a volume of 100 μl. A dose of 25 μg was used, unless otherwise indicated. Twenty-four hours following instillation, blood and BALF were collected and stored at −20°C.
Bone marrow-derived macrophages were generated by culturing bone marrow samples isolated from C57BL/6J mice in complete RPMI containing 20% L929 culture supernatant for 1 week. A total of 4 × 105 cells were plated into 24-well plates and were infected with M. tuberculosis Erdman or Mycobacterium smegmatis at a multiplicity of infection (MOI) of 5. Cell culture supernatants and cell lysates were collected at 6, 24, and 48 h postinfection and were sterile filtered and frozen prior to enzyme-linked immunosorbent assay (ELISA) or enzymatic activity assay analysis.
Healthy donors consented to donate blood in accordance with guidelines set forth by the Boston University Institutional Review Board (IRB) and the National Institutes of Health. The human whole-blood model was performed as previously described (17, 18, 26). Briefly, blood was collected into heparinized syringes, aliquoted into Eppendorf tubes, and stimulated at 37°C on a rotor for the times indicated. Plasma was collected following centrifugation, and red blood cell lysis was performed on pellets to obtain white blood cells. All samples were stored at −80°C prior to analysis.
Serum from a cohort of M. tuberculosis-infected patients in South Korea was obtained according to the Declaration of Helsinki and was banked at −80°C. Samples had been classified based on the disease state but were deidentified and thus were IRB exempted. Filter-sterilized aliquots were shipped to the University of Massachusetts Medical School and were stored at −80°C prior to analysis.
TNF, sTNFRI, sTNFRII, TIMP3, and TACE ELISAs (R&D Systems) were performed according to the manufacturer's protocol and as previously described (43).
An in-house multiplex ELISA was developed by using a Piezoarray noncontact spotter and commercially available antibody pairs (R&D Systems). Monoclonal antibodies were spotted in quadruplicate onto the bottom of a 96-well plate in an array format. Plates were blocked with Odyssey LiCor buffer, and a standard cocktail of recombinant proteins and samples was incubated for 2 h at ambient temperature. Biotinylated secondary antibody cocktails were incubated for 2 h, and streptavidin-conjugated infrared dye (IR 800; LiCor) was incubated in the dark for 30 min. Plates were read on an Odyssey LiCor instrument, and data were analyzed with SearchLight software, version 2.1. Previously reported cross-reactivity and sensitivity studies with this system demonstrated that both the human and mouse multiplex panels exhibited no cross-reactivity and that the multiplex format had a larger range of detection than conventional singleplex colorimetric ELISAs (35).
Human TACE activity was determined by using an Innozyme TACE activity kit from Calbiochem according to the manufacturer's instructions. For mouse TACE activity assays, opaque 96-well plates were coated with anti-TACE antibody (R&D Systems), washed, and blocked with 1% bovine serum albumin (BSA) prior to the addition of samples, which had been diluted in 1% BSA. Samples were incubated at room temperature for 1 h, the plates were then washed, and fluorogenic peptide (R&D Systems) was added in assay buffer (Calbiochem) and incubated at 37°C with shaking for 4 h prior to the reading of the plate at an excitation wavelength of 320 nm (ex320)/emission wavelength of 405 nm (em405). All TACE activity is reported as raw fluorescence units (RFU)/mg protein.
Cells were Fc blocked with anti-CD16/32 antibodies (eBioscience) for 10 min prior to staining with biotinylated TNFRI (clone MABTNFR1-B1; BD Pharmingen) and streptavidin-phycoerythrin (PE) (BD Biosciences), or PE-conjugated TNFRII (clone hTNFR-M1; BD Pharmingen) antibodies and appropriate isotype controls for 30 min on ice. Samples were washed and run on a Becton Dickinson LSR II instrument. Data were collected with FACSDiva software, version 6.0. All analyses were performed with FlowJo software v7.2.2 (Treestar).
Quantitative real-time PCR (qRT-PCR) was performed by using an iScript One-Step PCR kit with Sybr green (Bio-Rad). Primers for 18S rRNA, TNF, TNFRI, TNFRII, and TACE were purchased from SA Biosciences. Real-time PCRs were performed on a Bio-Rad iQ5 LightCycler. Delta threshold cycle (CT) values were calculated and are expressed as fold values compared to values for the housekeeping gene (18S rRNA).
All data are presented as means ± standard errors of the means (SEM). Student's paired t test and analysis of variance (ANOVA) with Tukey's posttests (95% confidence intervals [CI]) were performed with GraphPad Prism software, version 5.0. Statistical significance was determined to be a P value of <0.05.
To investigate the potential effect of M. tuberculosis infection on TNFR expression, we measured cytokine and soluble cytokine receptor levels, using an in-house multiplex ELISA system, in serum and bronchoalveolar lavage fluid (BALF) samples from mice that had been infected for 1 month with M. tuberculosis Erdman. Assessments of the BALF indicated that TNF, sTNFRI, and sTNFRII levels were all elevated in M. tuberculosis-infected mice compared to uninfected controls (P < 0.05 and P < 0.01) (Fig. 1a to toc).c). Baseline levels of sTNFRII were 10-fold higher than those of sTNFRI and were increased 10-fold with infection, while infection induced only a 5-fold increase in sTNFRI levels (Fig. 1). Serum levels of TNF were significantly elevated in M. tuberculosis-infected mice; however, the levels of neither sTNFRI nor sTNFRII was significantly different from those in uninfected controls (P < 0.05) (Fig. 1d to tof).f). However, there was a trend toward less sTNFRI with more sTNFRII in serum following infection.
Based on our finding that M. tuberculosis infection induced elevated levels of sTNFRs in BALF, we next examined if there was a specific M. tuberculosis component responsible for promoting sTNFR production. Mannose lipoarabinomannan (ManLAM) is a known M. tuberculosis virulence factor that has been shown to have immunomodulatory effects (13, 14, 28, 38, 41). Therefore, we hypothesized that this component of the M. tuberculosis cell wall was in part responsible for the observed effects. To investigate this, mice were intratracheally administered 25 μg ManLAM or an equal volume of phosphate-buffered saline (PBS) vehicle control. At 24 h postinstillation, mice were sacrificed, and blood and BALF were collected. A dose of 25 μg of ManLAM was chosen to correspond to doses used in mouse antigen challenge models and to approximately correspond to levels of ManLAM detected in sputum samples from M. tuberculosis-infected patients (29, 42, 44). As shown in Fig. 2a to toc,c, only the level of sTNFRII was found to be significantly higher in mouse BALF (P < 0.05) following treatment.
To identify the cellular source for the production of soluble TNF receptors in the mouse lung, we tested mouse primary leukocytes and mouse alveolar cell lines. Immune cells, specifically monocytes/macrophages, were shown previously to be a significant source of sTNFRs (27, 30, 36). We found that pooled mouse lymph node and spleen cell preparations stimulated with 100 ng/ml ManLAM were able to produce both sTNFRs, with little to no production of TNF, in response to ManLAM treatment (P < 0.01) (Fig. 2d to tof).f). Mouse alveolar cell lines did not produce any significant TNF or sTNFRs (data not shown). This corresponds with data from previous studies which demonstrated that sTNFR production is derived primarily from alveolar macrophages following intratracheal challenge (16).
To investigate the specificity of the ManLAM effect, other components that are similar to ManLAM were also tested. Arabinose-capped lipoarabinomannan (AraLAM) is another M. tuberculosis cell wall component that lacks the mannose cap sugars characteristic of ManLAM (14, 34). Cells stimulated with a comparable dose of AraLAM did not demonstrate an induction of any TNF family member proteins (data not shown). The stimulation of splenocytes using another M. tuberculosis cell wall component, hexamannosylated phosphatidyl inositol (PIM6), which contains acyl tail structures and 4 to 6 mannose sugar residues but lacks arabinose motifs, induced the production of sTNFRII but only with concomitant TNF production (data not shown). Irradiated M. tuberculosis H37Rv (virulent strain) was used as a control to test the effects of the intact (but replication-incompetent) bacterium on the production of TNF family members. Irradiated H37Rv cells were able to induce TNF production but not sTNFRs (data not shown). This observation is consistent with data from previous studies which demonstrated that intact, killed bacteria can lead to the production of TNF by the host (40). Taken together, these data indicate that different cell wall components of M. tuberculosis have differential effects on the production of TNF and TNFR.
We next investigated whether the ManLAM induction of sTNFRs in murine cells was predictive of the effects of ManLAM on human cells. Human white blood cells (approximately 5 × 106 cells) were stimulated with 100 ng/ml ManLAM ex vivo for 24 h. We found that, as with mouse cells, ManLAM induced a significant increase in the level of production of sTNFRII with lower levels of detectable sTNFRI and little to no induction of TNF (P < 0.05) (Fig. 3a to toc),c), which is consistent with data from a previous report (15). AraLAM and PIM6 were unable to induce the production of sTNFRs by human white blood cells (Fig. 3).
We also used the same approach to test a panel of M. tuberculosis components for TNF and sTNFR production. From the components tested, we found that ManLAM was the only lipoglycan capable of inducing sTNFRII without the production of TNF. None of the components tested significantly induced sTNFRI over levels in controls. We also tested a human alveolar cell line and found that, as with mouse cells, none of the M. tuberculosis antigens were able to induce sTNFR production (data not shown).
The increase in levels of sTNFRs could correspond to the loss of surface receptors induced by ManLAM-mediated cell activation. To determine this, stimulated cells were analyzed by flow cytometry following ManLAM treatment. The production of sTNFRI and sTNFRII did coincide with a loss of the receptors from the cell surface, indicating the loss of membrane (Fig. 3d and andee).
Since the production of soluble TNF receptors usually correlates with elevated TNF levels as a mechanism for regulating inflammation (23, 54), we wanted to investigate the selective effects of ManLAM on the induction of sTNFRs without inducing the production of TNF. We began by examining mRNA levels of TNF, TNFRI, TNFRII, and TACE. We found that there were no significant differences in transcript levels for any of these products in human cells treated with ManLAM over a 48-h time course (Fig. 4).
As there was no difference in mRNA levels, we next examined whether ManLAM stimulation resulted in the induction of TACE activity. We determined that there was elevated TACE activity between 2 and 6 h after stimulation with ManLAM (Fig. 5a). TIMP3 is the natural inhibitor of TACE; therefore, lower TIMP3 levels would increase the enzymatic activity of TACE, thereby leading to the increased cleavage of TNF family members from the surface of cells. As shown in Fig. 5b, we found that there was a significant decrease in TIMP3 levels at 4 h in plasma from ManLAM-treated blood (P < 0.05) (Fig. 5a), which recovered by 24 h poststimulation. The TACE protein levels from cell pellets remained consistent between conditions at all time points (data not shown).
We wanted to determine how ManLAM was altering TIMP3 levels and TACE activity. It was demonstrated previously that ManLAM stimulation can alter PI3K/AKT pathway activation (32); therefore, we hypothesized that ManLAM may induce a signaling event that affects TACE activity. A recent report demonstrated that TACE activity is induced following MAPK pathway activation (58). Western blotting of primary human white blood cells demonstrated that ManLAM can induce ERK1/2 phosphorylation (Fig. 5c).
We wanted to determine the relevance of our findings about ManLAM-mediated changes in TACE activity in a live infection. Bone marrow-derived macrophages were infected at an MOI of 5 with M. tuberculosis Erdman or M. smegmatis, an avirulent mycobacterial strain that does not express ManLAM. We found that M. smegmatis, but not M. tuberculosis, was able to induce detectable levels of TNF production (Fig. 6a). Interestingly, there was a decrease in the amount of sTNFRI produced following infection with either mycobacterial species compared with medium controls (P = 0.0001) (Fig. 6b), with a concomitant increase in the level of sTNFRII production compared to the vehicle control (P < 0.0001) (Fig. 6c).
An assessment of changes in TIMP3 levels demonstrated that both species could induce decreases in levels of TIMP3 at 6 h postinfection, although M. tuberculosis infection resulted in significantly less TIMP3 than did M. smegmatis infection (Fig. 6e). TIMP3 levels were undetectable at 24 and 48 h postinfection, possibly due to binding to TACE or some other mechanism of clearance from the system. Interestingly, we found that TACE activity levels were increased following infection with either species, but only M. tuberculosis infection resulted in sustained TACE activity at a low level above the baseline up to 48 h postinfection (P = 0.004 by an unpaired t test) (Fig. 6f).
It was reported previously that TACE can be activated through an interleukin-10 (IL-10)-dependent, TIMP3-independent mechanism (11); therefore, we also measured levels of IL-10 production following infection. We found that M. smegmatis was able to induce IL-10 production (P < 0.0001) (Fig. 6d).
The in vivo and in vitro studies indicated that M. tuberculosis infection resulted in an increased protein ratio of sTNFRII to sTNFRI without much concomitant TNF production. We next investigated whether this observation could be identified in human tuberculosis disease. To determine whether altered sTNFRII/sTNFRI ratios could be detected in patients, we assessed TNF and sTNFRs in patient serum from a small cohort (Table 1). While there was no association between clinical parameters (as detailed in Table 1) and absolute levels of TNF family members, the ratios of sTNFRII/TNF approached statistical significance, and alterations in sTNFRII/sTNFRI ratios were statistically significant compared to ratios in sera from either patients with COPD or healthy controls (P < 0.05) (Fig. 6). These data suggest a mechanism for M. tuberculosis-induced sTNFRs similar to that seen in the mouse model and support the concept of using this ratio as a biomarker for disease.
Despite the fact that approximately one-third of the world's population is infected, the mechanisms involved in M. tuberculosis infection and pathogenesis are not fully understood. A number of viruses and bacteria have developed mechanisms to evade immunosurveillance and promote a proinfection environment. One possible mechanism utilized by M. tuberculosis is through the modulation of TNF, which can function in the maintenance of lung granulomas associated with infection. In this study, we demonstrate that M. tuberculosis infection in mice results in the production of sTNFRs in the lung, as determined by measuring levels in BALF (Fig. 1). Based on its reported immunomodulatory effects, we hypothesized that mannose-capped lipoarabinomannan (ManLAM), a major component of the M. tuberculosis cell wall, may induce the production of sTNFR as a mechanism to reduce circulating TNF levels (13, 28, 38, 50). We found that the intratracheal instillation of the M. tuberculosis virulence factor ManLAM alone resulted in the production of sTNFRII in the BALF of naïve mice without any induction of TNF (Fig. 2). This indicates the possibility that ManLAM is capable of altering the balance between TNF and sTNFRs whereby less unbound TNF is available for granuloma maintenance.
In the live-infection model, levels of TNF as well as both sTNFRs were elevated in BALF. Relative production, however, greatly favors sTNFRII, as infection induced a 10-fold increase in the sTNFRII level with only a 5-fold increase in the sTNFRI level. These findings are similar to those presented previously by Dai et al., who showed that the level of TNF produced by alveolar macrophages in response to lipopolysaccharide (LPS) stimulation in patients with sarcoidosis and extrinsic allergic alveolitis was closely related to sTNFRII levels (16). In our study, TNF levels were elevated from the threshold of detection to approximately 200 pg/ml with infection; however, this concentration of TNF was significantly lower than the 4,000-pg/ml concentrations of sTNFRII detected.
Having established the effects of ManLAM stimulation using murine cells, we tested human cells to determine whether the effects were similar. We found that ManLAM treatment did induce sTNFRII production without concomitant TNF production (Fig. 3). To further investigate the mechanism, we used this human whole-blood model to examine the effects of ManLAM on RNA and protein levels as well as the enzymatic activity of TACE (Fig. 4 and and5).5). We found that ManLAM treatment resulted in the hyperenzymatic activation of TACE with a maximal effect at 4 h, which corresponds to a decrease in TIMP3 levels (Fig. 5). This effect is likely due to ERK1/2 activation induced by ManLAM (Fig. 5). Despite being a cell wall component of M. tuberculosis, during active infection, ManLAM is secreted from infected host macrophages and dendritic cells in the form of lipid bodies (7) and can subsequently be incorporated into lipid rafts on peripheral blood mononuclear cells (28). It is the incorporation of ManLAM into cell membranes that is believed to cause alterations in cell signaling (38, 50), although the exact mechanism has not been clearly elucidated.
We examined changes in TIMP3 levels following infection with M. smegmatis (Fig. 6) and found that this bacterium induces the shedding of TNF with the sTNFRs; therefore, it was likely that this occurred through a different mechanism. It was reported previously that TACE can be activated through an IL-10-dependent, TIMP3-independent mechanism (11); therefore, we also measured IL-10 production following infection. We found that M. smegmatis, but not M. tuberculosis, was able to induce large amounts of IL-10 (P < 0.0001) (Fig. 6d). It is likely that antigens unique to virulent M. tuberculosis are responsible for the selective release of sTNFRII and the mechanism of induction of TACE activation which we observed. It is possible that a combination of virulence factors is contributing to the enhanced sTNFRII release during M. tuberculosis infection, however, so future studies could examine other virulence factors that could contribute to this phenotype by performing infections with genetically modified strains and monitoring sTNFR, TNF, and TIMP3 levels and bacterial burden or disease outcome. This might provide some insight into how ManLAM is capable of preferentially inducing the release of sTNFRII but not TNF. It would also be interesting to determine if other virulent pathogens alter sTNFR production through a similar mechanism.
We demonstrate that following ManLAM treatment, TACE preferentially cleaves TNFRII and, to a lesser extent, TNFRI from the cell surface. While limited compared with the effects on sTNFRII, the effects of ManLAM on the production of sTNFRI could also represent a mechanism to modulate the immune response. It was reported previously that membrane-bound TNF on the surface of T cells is sufficient to promote granuloma formation and control acute M. tuberculosis infection regardless of soluble bioactive TNF (49). However, another study demonstrated that membrane-bound TNFRI can provide a costimulatory signal for T cells. Therefore, the cleavage of TNFRI from the surface of T cells could attenuate T cell proliferation as well as granuloma development and maintenance (20), even in the presence of membrane-bound TNF. The elevated levels of both sTNFRs found in our infected mice may facilitate, through two different mechanisms, modulating effects on the immune response.
It was shown previously that there are imbalances in TNF and TNFRs in the BALF of TB patients (53). Based on our mouse model of M. tuberculosis infection, we wanted to test whether or not there were altered ratios of TNF and sTNFRs in the sera of M. tuberculosis-infected patients. We tested serum samples from a small cohort of patients and found that there was a statistically significant higher ratio of sTNFRII to sTNFRI than that found for sera of either COPD patients or healthy controls (Fig. 7). A previous report by Juffermans et al. showed that levels of sTNFRI and sTNFRII were elevated in M. tuberculosis patient serum samples (31). While we did not find statistically significant increases in absolute values of sTNFRI and sTNFRII, our results suggest that the use of a ratio of sTNFRs might prove to be a useful biomarker of M. tuberculosis infection. Larger cohorts and longitudinal studies with M. tuberculosis patient serum samples are required to establish the clinical relevance of these findings and would provide a basis for an examination of TNF/sTNFR ratios as potential biomarkers of M. tuberculosis infection and disease severity, reactivation, or dissemination (31). Furthermore, TACE can cleave other molecules off the cell's surface, including ICAM-1 and Notch (12, 52). Therefore, future studies of M. tuberculosis biomarkers could examine other targets of TACE to enhance the predictive power of sTNFRs.
ManLAM is secreted from M. tuberculosis-infected macrophages (7), and ManLAM levels in serum, urine, and sputum samples increase over time in patients infected with M. tuberculosis (10, 31, 44). Diagnostic tests to detect the presence of ManLAM in human sputum and urine samples for the rapid diagnosis of M. tuberculosis infection are under development (24, 44). The pathological relevance of ManLAM has been demonstrated by in vivo infection studies in mice and in human patients, where endogenous anti-ManLAM antibodies have been identified (25, 44). These antibodies appear to be beneficial, as higher circulating antibody levels correlate with decreased dissemination, lower bacterial loads, prolonged survival, and better disease outcomes (19, 25). Systemic TNF levels were not measured in those studies, and therefore, a direct correlation cannot be made; however, disease outcomes would suggest that a reduction of ManLAM levels is beneficial. In association with most inflammatory conditions, sTNFRs are generated in tandem with TNF as a mechanism for regulating the proinflammatory effects of TNF. Our studies indicate that ManLAM stimulation in vivo or the in vitro stimulation of immune cells results in the selective production of sTNFRs, preferentially sTNFRII, without TNF production.
The overall supposition of these studies is that a reduction of TNF levels creates an environment that would promote bacterial infection and/or dissemination. The increase in levels sTNFRs is one mechanism that would accomplish this. Our data suggest that ManLAM shed or secreted from M. tuberculosis-infected cells can stimulate TACE activity, resulting in the generation of sTNFRs, which would function to reduce bioactive TNF levels, thereby reducing the ability of the host to contain the infection. While we did not definitively conclude from our studies that ManLAM directly alters sTNFR ratios in M. tuberculosis patient samples or mice infected with M. tuberculosis, as this would require the presence of a selective ManLAM inhibitor, the beneficial effects reported in anti-ManLAM antibody studies (19, 25) support the overall concept. Our data strongly suggest that ManLAM is involved in the process and identify a potential mechanism for this effect. Additional studies are required to correlate levels of ManLAM and/or anti-ManLAM antibodies with TNF/sTNFR ratios to more fully understand the immune response and disease state in M. tuberculosis infections.
This work was supported by NIH grants RO1 HL081149 and HL064884 (to H.K.), NIH grant RO1 ES013528 (to D.G.R.), and NIH grant RO1 CA1227373 (to W.W.C.). J.M.R. was supported by a fellowship from the National Science Foundation (grant 0538608). All TB materials were obtained through a collaboration with Colorado State University under NIH NIAID contract no. HHSN266200400091C, entitled Tuberculosis Vaccine Testing and Research Materials.
We thank Yun Seong Kim of Pusan National University School of Medicine, Yangsan, South Korea, for providing M. tuberculosis patient samples. We also thank Michael Ieong for assistance with TACE activity assays, Zachary Hunter for assistance with qRT-PCR data analysis, Matthew Blahna and Daniel Ferrari for assistance with alveolar cell line culture, and John Bernardo for thoughtful comments on interpreting patient sample data. All flow cytometric data were acquired using equipment maintained by the Boston University Medical Campus Flow Cytometry Core Facilities.
We declare no competing financial interests.
Published ahead of print 27 August 2012