Although much is understood regarding the role of B7/CD28 family of costimulatory molecules in regulating host resistance in the context of several pathogens, analogous information with Mycobacterium tuberculosis is lacking. To address the requirements of B7-mediated costimulation in host resistance against tuberculosis, mice deficient in both B7.1 and B7.2 (B7DKO) were aerosol infected with M. tuberculosis Erdman and disease progression was monitored. We report herein that B7DKO mice are initially able to contain the bacterial load in the lung, but exhibit enhanced susceptibility during chronic infection. Despite the early control of bacterial replication, B7DKO mice essentially start off with compromised Th1 immunity and slower granulomatous response in the lung, characterized by markedly reduced lymphocytic infiltration. As the infection progresses from acute phase to the chronic phase, the nascent granulomas in the B7DKO lungs never fully achieve the architecture of granulomas developing in wild-type mice. Instead, lesions spread progressively to involve much of the lung in the B7DKO mice, ultimately leading to necrosis. Thus, early control of M. tuberculosis growth in the lung can occur in the absence of B7 costimulation and is less dependent on Th1 immunity and formation of a granulomatous structure. However, B7 costimulation is critical for long-term containment of infection within lung granulomas. These findings suggest that the use of costimulation-based immunomodulators may have significant repercussions on the induction of host protective immunity against tuberculosis.
Aproductive interaction between APCs and naive T cells is dependent on the successful transmission of a second signal, which is provided by costimulatory molecules present on the APCs via ligation of cognate receptors on T cells (1, 2). The fact that enhanced expression of costimulatory molecules is a hallmark feature of a mature APC suggests the potential role these molecules may play in the generation and regulation of Mycobacterium tuberculosis-specific effector T cell responses and the subsequent development of protective immunity in the lung.
The B7 family of ligands and receptors is a growing family of immune molecules critical for T cell homeostasis, activation, and tolerance (for reviews see Ref. 3, 4, 5, 6). Interaction of B7.1 (CD80) and B7.2 (CD86) with CD28 was the first of these receptor ligand pairs to be discovered, followed by studies showing that CTLA4 (CD152) could also engage B7.1 and B7.2, but to initiate negative, rather than positive, signaling in activated T cells (reviewed in Ref. 7 , 8). Other receptor:ligand pairs that were subsequently identified include ICOS:ICOSL and PD-1:PD-L1/PD-L2, with the latter, like CTLA-4, inducing inhibitory signaling in target cells (reviewed in Ref. 9). ICOS, although initially identified as a positive regulator of T cell responses (10), can also down-regulate T cell functions dependent on the activation status of the T cell (11, 12). Recently, B7.1 was shown to interact with PD-L1, and like CTLA4, the interaction initiates inhibitory signaling (13, 14). Thus, the outcome of T cell activation is the integration of positive and negative costimulation from several ligand:receptor pairs and a balance between these opposing signals is necessary for generating a successful immune response against pathogens and at the same time maintaining self tolerance in the host.
B7/CD28 pathway is probably the most critical in initiating and regulating T cell responses and has been extensively studied in animal models of autoimmune and infectious diseases using approaches that have included blocking B7 or CD28 via Abs or by using animals deficient in these molecules (reviewed in Ref. 7 , 15 , 16). B7.1 and B7.2 have dichotomous functions in Th subset development, but distinct roles in regulating the development of a specific subset is context dependent. For instance, blocking B7.1 activity in SJL mice inhibited the induction of pathogenic Th1 response and abrogated susceptibility to experimental autoimmune encephalomyelitis, while blocking B7.2 functions inhibited Th2 responses and led to disease exacerbation (17, 18). In other models of murine autoimmune disease, such as diabetes in NOD mice (19) and Sjogren’s syndrome in NFS/sld mice (20), blocking B7.2, in contrast, prevented disease development. In the murine model of mercury-induced autoimmunity where both Th1 and Th2 responses are involved, B7.1 and B7.2 were found to regulate different manifestations of the autoimmune syndrome, and absence of either one did not completely protect from disease (21). However, B7.1 and B7.2 can also have overlapping functions in Th subset development. For instance, C57BL/6 mice lacking either B7.1 or B7.2 produce autoreactive Th1 cells and are equally susceptible to experimental autoimmune encephalomyelitis (22).
The influence of B7.1 and B7.2 to the development of Th1 and Th2 subsets in response to infectious challenge has also been extensively probed (23). For example, granuloma formation and Th2 response in mice infected with the nematode parasite Schistosoma mansoni require B7.2 (24), while Th2 induction by the gastrointestinal nematode, Nematanthus brasiliensis is B7-independent (25). Using CTLA4Ig to block B7/CD28 interaction during Leishmania major infection in BALB/c mice, Corry and colleagues (26) showed that priming of Th2 cells was more dependent on B7/CD28 pathway than the priming of Th1 cells. Subsequently, it was shown that B7.2 is important for Th2 induction in the susceptible BALB/c mice, and absence of B7.2 alone or the combined absence of B7.1 and B7.2 induced resistance to leishmaniasis in these mice (27, 28). However, BALB/c mice deficient in B7.1 remain as susceptible as wild-type (WT)4 BALB/c mice (27). Interestingly though, absence of B7.2 had no influence on disease phenotype in the resistant 129 mouse strain (27).
Despite our extensive understanding of B7 costimulation to Th development in many disease conditions, it is surprising that little is known regarding the contribution of this pathway to M. tuberculosis-specific Th1 cell induction, differentiation, and expansion, except that B7 molecules are up-regulated on dendritic cells in response to M. tuberculosis infection (29, 30). Given the fact that modulation of B7/CD28 costimulation is being pursued as a therapeutic strategy to regulate autoimmune diseases and chronic inflammation in humans (31, 32, 33, 34, 35), it becomes essential to determine whether this pathway is necessary for host resistance against primary and reactivation tuberculosis. In this study, we therefore examined the role of B7 costimulation in the induction of Th1 immunity and host resistance to tuberculosis following aerosol infection with the virulent M. tuberculosis Erdman.
Materials and Methods
WT C57BL/6 mice were purchased from the National Cancer Institute. Breeding pairs of double knockout (B7DKO) mice were obtained from Dr. Arlene Sharpe (Brigham Women’s Hospital, Harvard Medical School, Boston, MA). The KO mice were bred and maintained under pathogen-free conditions at the University of Medicine and Dentistry of New Jersey animal facility. M. tuberculosis-infected mice were housed in BSL3 facility. Animal protocols used in this study were approved by the university Institutional Animal Care and use Committee.
Animal infection, colony forming unit (CFU) determination, and isolation of lymph node and lung cells
Aerosol infection with M. tuberculosis Erdman (Trudeau Institute, Saranac, NY) of age-matched WT and B7DKO female mice (6–8 wk old) was conducted in a closed-air aerosolization system (In-Tox Products). Mice were exposed for 20 min to nebulized M. tuberculosis at a density that was optimized to deliver a low dose of ∼20–50 CFU or high dose of ∼200–500 CFU to the lungs. At each time interval studied, infected animals were sacrificed by cervical dislocation and the left lobe of the lung was homogenized in PBS containing 0.05% Tween 80. Serial dilutions of the homogenates were plated onto 7H11 agar. The plates were incubated at 37°C and colonies counted after 21 days. The right lobes were reserved for tissue gene expression and histological studies. For obtaining single cell suspensions, the lungs were perfused with 5 ml sterile PBS and harvested. Mediastinal lymph nodes were also harvested and both tissues were processed separately to obtain single cell suspensions. Lungs were cut into small pieces and incubated with 1 mg/ml collagenase D (Roche) for 30 min. The digestion was stopped by adding 5 mM EDTA. The digested tissue was transferred to a 40-μm nylon cell strainer and disrupted using a syringe plunger to obtain single cell suspensions. Lymph node tissue was processed similarly, but without collagenase digestion. RBCs were lysed with ACK lysing buffer and viable cell number was determined by trypan blue dye exclusion.
The following mAbs were used for the study: RM4–5 (anti-CD4), Ly-2 (anti-CD8), RB6–8CF (anti-GR-1), CL:A3-1 (F4/80), 2C11 (anti-CD3), 37.51 (anti-CD28), UC10-4F10-11 (anti-CTLA-4), H1.2F3 (anti-CD69), IM7 (anti-CD44), and XMG1.2 (anti-IFN-γ). Ab to F4/80 was purchased from AbD Serotec and the rest of the Abs were purchased from BD Biosciences. All Abs were directly conjugated to flurochromes and isotype controls were included for each Ab type. Surface immunofluorescence staining was conducted using standard procedures. For intracytoplasmic detection of IFN-γ, 1 × 106 lymph node and lung cells were stimulated with plate-bound anti-CD3 and anti-CD28 Abs for 5 h in the presence of monensin (3 μM). The cells were harvested from the plates, pelleted, washed in FACS buffer, and resuspended in buffer containing appropriate concentrations of fluorochrome-conjugated anti-CD4 and CD8 Abs. After a 30 min incubation at 4°C, the cells were washed again in FACS buffer, fixed with 4% paraformaldehyde for 30 min and then permeabilized with 0.2% saponin and 2% FCS in PBS. After additional washes in FACS buffer, the cells were reacted with 1 μg/ml fluorescently conjugated anti-IFN-γ Ab for 30 min, washed with FACS buffer and subjected to a three-color flow cytometric analysis. The cells were acquired on a FACSCalibur instrument and data were analyzed using FlowJo (Tree Star).
BrdU and annexin staining
In situ proliferation and apoptosis of T cells in lungs of infected mice was measured as described previously (36). At several time intervals during the course of infection, mice were given 1 mg BrdU (Sigma-Aldrich) i.p. Sixteen hours later the mice were sacrificed and lungs were harvested and processed as described above to obtain single cell suspensions. For measuring T cell proliferation, 1 × 106 lung cells were reacted with flurochrome-conjugated anti-CD4 and CD8 Abs, fixed, permeabilized, and reacted with fluorescently conjugated anti-BrdU Ab (BD Biosciences). For quantitation of T cell apoptosis in the lung, another aliquot of 1 × 106 cells were reacted with flurochrome-conjugated anti-CD4 and CD8 Abs, stained with Annexin V and 7-aminoactinomycin D (7-AAD) (BD Biosciences), and then fixed. Cells were acquired on FACSCalibur and data were analyzed using FlowJo.
ELISPOT assay to detect IFN-γ-producing effector cells from infected lymph nodes and lungs was conducted as described previously (36). Ninety-six-well Millipore filter plates (MultiScreen HTS) were coated with 6 μg/ml anti-IFN-γ Ab (clone R4-6A2, BD Biosciences). The plates were washed and seeded with 25,000 and 50,000 each of lymph node cells or lung cells obtained from infected animals at each time interval studied. For ex vivo restimulation of the lymph node and lung cells, bone marrow-derived dendritic cells pulsed overnight with M. tuberculosis (multiplicity of infection of three) served as APCs. The ratio of T cell to dendritic cell was maintained at 2:1 and the cells were cocultured for 40 h at 37°C. The plates were subsequently washed and treated sequentially with biotinylated secondary Ab (Clone XMG1.2, BD Biosciences) and streptavidin HRP. The spots were developed by treating the plates with 3-amino-9-ethyl-carbazole and spot forming units were counted using an ELISPOT plate reader (Cellular Technology).
RNA isolation and RT PCR
Lungs and lymph nodes were homogenized in 2 ml of TRIzol reagent (Invitrogen) and immediately stored at −80°C. Total RNA was then extracted and purified using RNAeasy column (Qiagen). Total RNA obtained was reverse transcribed using Superscript II enzyme (Invitrogen), as directed by the manufacturer. Real-time PCR was performed using the Mx3000P system (Stratagene). The primer sequences used were as follows: IFN-γ (forward, 5′-TGC ATC TTG GCT TTG CAG CTC-3′, reverse, 5′-CTT GCT GTT GCT GAA GAA GG-3′), NOS2 (forward 5′-TGC CCC TTC AAT GGT TGG TA-3′, reverse 5′-ATT TGG CTG GTC CCT CCA GT-3′), TNF-α (forward, 5′-GACGTGGAACTGGCAGAAGA-3′, reverse, 5′-CTCATTCCTGCTTGTGGCAG-3′), and β-actin (forward, 5′-CCG TGA AAA GAT GAC CCA GAT C-3′, reverse 5′-ACG TAC CCA TCC AGG CTG TG-3′). The reactions were performed to generate cycles (Ct). Relative gene expression was calculated as 2−ΔΔCt, where ΔCt = Ct (gene of interest) – Ct (normalizer = β-actin) and the ΔΔCt = ΔCt (sample) – ΔCt (calibrator). Calibrator was total RNA from uninfected lung or lymph node. Data are expressed as mean ± SD.
Histopathological analysis was performed on lungs that were fixed in 4% paraformaldehyde in PBS for 1 wk and subsequently paraffin embedded. Five to seven micrometer sections were cut and stained using a standard H&E protocol. The slides were evaluated by the pathologist at the University of Medicine and Dentistry of New Jersey Histology Core Facility.
Statistical significance was determined by 2-way ANOVA for repeated measure comparisons and Student’s t test for pair wise comparisons using the GraphPad Prism 4 software (Jandel Scientific Software).
B7DKO mice exhibit enhanced susceptibility to aerosol infection with M. tuberculosis
To evaluate the role of B7 costimulation in host resistance against M. tuberculosis infection, WT and B7DKO mice were aerosol-infected with M. tuberculosis and disease progression was monitored over time. We first examined disease progression in the B7DKO mice following a low dose infection of ∼20 CFU of M. tuberculosis. Comparing the bacterial burden in the B7DKO mice with WT indicated that M. tuberculosis burden in the lungs did not differ between the two groups during the acute phase of infection (up to 4 wk). However, as infection progressed there was a steady and significant increase in bacterial burden in the B7DKO lungs, while the WT mice, as expected, were able to contain infection and maintain a steady bacterial load (Fig. 1⇓A). By week 15, there was a two log increase in bacterial burden in B7DKO mice compared with WT (Fig. 1⇓A), and during this period the B7DKO mice also began to exhibit enhanced morbidity.
Earlier studies had suggested that the magnitude of the immune response and the ability to control M. tuberculosis infection is dependent on the initial inoculum. For example, mice lacking CD40 were more susceptible to a lower inoculum, while they were able to bypass the need for CD40 costimulation with higher infectious inoculum (37). We argued that an increase in Ag concentration could perhaps enhance host immunity and compensate for the lack of B7.1 and B7.2. To investigate this possibility, we conducted a high dose infection using 500 CFU of M. tuberculosis. Similar to what was observed with low dose infection, B7DKO were not able to contain bacterial burden in the lung during the chronic phase following a high dose of infection (Fig. 1⇑B). The only difference between the two doses of infection was that B7DKO mice receiving a higher inoculum of M. tuberculosis were able to contain infection for a longer period (up to 7 wk), after which they lost the ability to hold a steady M. tuberculosis number in the lung.
Lack of B7 signaling leads to altered granuloma architecture
We next extensively characterized the development of lung pathology in H&E stained sections of infected B7DKO and WT. Below, we describe the histological evaluation of lung sections prepared from mice receiving a low dose infection. At 4 wk postinfection, a substantial granulomatous inflammation was present in the WT lungs (Fig. 2⇓A), while in the B7DKO lungs, the foci of granulomatous inflammation was smaller (Fig. 2⇓C). A magnified view of Fig. 2⇓A, showed that the granulomatous inflammation in the WT lung consisted generally of macrophages and lymphocytes and occasional karyorrhectic nuclei of indeterminate origin scattered through the lesions (Fig. 2⇓B). In contrast, a magnified view of Fig. 2⇓C showed that the granulomatous inflammation in B7DKO mice presented with fewer lymphocytes but more frequent karyorrhectic nuclei (Fig. 2⇓D). At 7 wk, prominent clumps of cells appeared within the granulomatous structure in the WT lungs (Fig. 2⇓E). The clumps consisted mainly of lymphocytes with occasional foamy macrophages when examined under high magnification (Fig. 2⇓F). In contrast, in the B7DKO lungs, there was an absence of cell clumps and, in fact, the granulomatous inflammation occupied a larger area of lung tissue (Fig. 2⇓G). Visualization at higher magnification showed that fewer lymphocytes were present, but the number of foamy macrophages was increased in the granulomatous inflammation (Fig. 2⇓H). Overall, the B7DKO mice exhibited minimal early inflammatory response to M. tuberculosis and presented fewer T cells in the lung as compared with WT mice. Between 12 and 16 wk, the granuloma architecture in the WT lungs, while stable in size, contained more lymphocytes (3, A and E). In agreement with previous studies (38, 39), many of the granulomas presented as a central region of densely packed lymphocytes surrounded by foamy macrophages (Fig. 3⇓, B and F). In contrast, the B7DKO lungs became consolidated with infiltrates that progressively spread to involve much of the lung (Fig. 3⇓, C and G). The infiltrates consisted almost exclusively of foamy macrophages with numerous pockets of karyorrhectic nuclei and associated cellular debris (Fig. 3⇓, D and H). Similar changes in lung histopathology were seen in B7DKO mice infected with 500 CFU of M. tuberculosis (data not shown). Overall, during the chronic phase, B7DKO mice never developed the characteristic granulomatous structure exhibited by WT mice.
Characterization of the cellular infiltrates of WT and B7DKO lungs
We next performed flow cytometric analyses of single cell suspensions from lung digests to determine the kinetics of CD4 and CD8 T cell influx into the lungs of WT and B7DKO mice. At each time interval studied, there was significantly decreased accumulation of both CD4 (Fig. 4⇓A) and CD8 T cells (Fig. 4⇓B) in the lungs of infected B7DKO mice in comparison to WT. Despite differences in CD4 and CD8 T cell accumulation, the total number of cells recovered from WT and B7DKO lungs, was similar (Fig. 4⇓C). Further characterization of immune cell types contributing to the lung infiltrates during infection revealed that Gr1high F4/80− cells (neutrophils) were significantly increased in B7DKO lungs compared with WT at 4 and 12 wk postinfection, while at 7 wk the neutrophil numbers were similar in the two groups (Fig. 5⇓). F4/80+ cells (macrophages) were present in equivalent numbers in the two groups at 4 wk, but increased significantly at 7 and 12 wk postinfection in the B7DKO (Fig. 5⇓). In contrast, there was a significant decrease in CD19+ B cells at 7 and 12 wk postinfection in the B7DKO lungs.
We further investigated whether the reduced number of CD4 and CD8 T cells observed in B7DKO lungs could have resulted from reduced proliferation of these cell types in the lungs, or alternatively due to increased apoptosis in the absence of B7 costimulation. To obtain a measurement of proliferating cells in the lungs, BrdU was injected 16 h before harvesting the lungs from the infected mice. At indicated time intervals, lungs were harvested and an aliquot of cells from the same lungs were reacted with anti-BrdU, anti-CD4, and anti-CD8 Abs and the percentage of BrdU+ T cells present within the CD4 and CD8 population was calculated. For apoptosis determination, another aliquot of cells from the same lungs was reacted with annexin V, 7-AAD, anti-CD4, and anti-CD8 Abs and the percentage of annexin and 7-AAD double positive cells was determined within the CD4 and CD8 gates. Fig. 6⇓, A, B, E, and F are examples of FACS plots for each staining condition. We found that the percentage of BrdU+CD4+ (Fig. 6⇓C) and BrdU+CD8+ (Fig. 6⇓D) T cells was equivalent in the lungs of WT and B7DKO mice at all the three time points tested, indicating that T cell turnover in the lungs of the two groups of mice is similar. At weeks 4 and 7, the percentage of T cells positive for Annexin and 7-AAD was not different between the two groups (Fig. 6⇓, G and H), but at week 12 postinfection, a significant increase in apoptosis of CD4+ and CD8+ T cells was seen in both WT and B7DKO mice (Fig. 6⇓, G and H). However, the percentage of apoptotic CD4+ and CD8+ T cells numbers was significantly more in the B7DKO lung at week 12 compared with WT, possibly as a consequence of the enhanced bacterial burden in the lung at this time.
Absence of B7 affects Th1 cell priming in the draining lymph nodes and effector Th1 cell recruitment to the lungs
We next examined whether reduced T cell numbers in the lungs of B7DKO resulted from deficient priming and differentiation of naive T cells to IFN-γ-secreting Th1 subset. To test this possibility, equivalent numbers of lymph node and lung cells obtained from infected lungs of WT and B7DKO were stimulated ex vivo with M. tuberculosis-pulsed dendritic cells. Forty hours later, the frequency of M. tuberculosis-specific IFN-γ secreting cells was measured by ELISPOT. A significantly reduced frequency of IFN-γ-secreting cells was present throughout infection in the lymph nodes (Fig. 7⇓A) and lungs (Fig. 7⇓B) of B7DKO mice as compared with WT. These data indicate that there was suboptimal Th1 priming in the lymph nodes of M. tuberculosis-infected B7DKO mice resulting in less recruitment of IFN-γ-secreting cells to the lung. To determine whether IFN-γ secretion from both CD4 and CD8 T cells was equally affected by the absence of B7 costimulation, lung cells obtained from infected animals were stimulated ex vivo with anti-CD3 and anti-CD28 Abs and intracytoplasmic expression of IFN-γ was detected in CD4 and CD8 T cells by flow cytometry. At each time interval studied, we observed reduced number of IFN-γ-expressing CD4 (Fig. 7⇓C) and CD8 (Fig. 7⇓D) T cells in lungs of B7DKO mice in comparison to WT. These data further confirm that B7DKO lungs have significantly reduced IFN-γ-secreting effector T cells.
The reduced T cell priming was reflected in fewer activated T cells in the lymph nodes of B7DKO mice. As shown in Fig. 7⇑E, the percentage of lymph node CD4+ T cells expressing the activation markers, CD69, CD44, and CTLA-4 is significantly reduced in B7DKO compared with WT.
Curiously, IFN-γ gene expression in the lungs of WT and B7DKO mice at 4 and 7 wk postinfection was similar, and at 15 wk was actually significantly higher in the B7DKO mice (Fig. 8⇓A). TNF-α expression was also significantly higher in B7DKO at 4 and 15 wk while expression was barely detectable in WT mice (Fig. 8⇓B). In contrast, NOS2 expression was detected only at 7 and 15 wk and was equivalent between the two groups (Fig. 8⇓C).
In the present study, we identified B7 as an essential costimulatory molecule in the generation of M. tuberculosis-specific Th1 response and formation of the tubercle granuloma. Interestingly, although the absence of B7 resulted in early defects in Th1 priming, the consequences of the defect to control of bacterial replication and maintenance of host resistance was manifested only later in infection. B7DKO mice exhibited slower granulomatous response in the lung, characterized by markedly reduced lymphocytic infiltration. As infection progressed from acute to chronic phase, the granulomas in the B7DKO lungs never fully achieved the architecture of granulomas developing in the WT mice. Instead, lesions spread steadily to involve much of the lung in the B7DKO mice, ultimately leading to necrosis.
Recruitment of other immune cell types may have contributed to the ability of B7DKO mice to resist acute infection, despite lowered T cell numbers in the lung. Neutrophils are present in high numbers during the early phase of M. tuberculosis infection (38), and have been shown to participate in controlling bacillary growth in a phagocytosis-independent mechanism by either enhancing IFN-γ production (40) or via transfer of granules containing antimicrobial peptides to macrophages (41). Neutrophils have also been shown to regulate early granuloma formation in the lung of M. tuberculosis-infected mice via CXCR3-signaling chemokines (42). Consistent with the role of neutrophils to provide protection, we observe increased neutrophil recruitment in B7DKO lungs, particularly during early infection. In contrast, macrophage numbers were similar in the two groups during acute infection, but increased significantly in B7DKO in chronic infection, suggesting that this could be the consequence of increased bacterial load present in the lung at this time.
We were surprised to find, notwithstanding the reduced T cell priming, that B7DKO mice expressed mRNA levels of IFN-γ and TNF-α that were equivalent to WT mice at earlier time points. Furthermore, during chronic infection, B7DKO mice continued to express significantly elevated mRNA levels of IFN-γ and TNF-α while gene expression of these cytokines had waned in WT mice. B7DKO mice have reduced T regulatory cells (43). Therefore, a plausible explanation for the similar mRNA levels for the cytokines in the two groups of infected mice could be that, despite reduced T cell priming, each T cell in the B7DKO mice produces more message and protein in the absence of basal B7. Indeed, there is evidence that if OVA-transgenic T cells are primed with OVA-pulsed WT dendritic cells in B7DKO mice, the primed T cells secrete higher levels of cytokine than if priming was conducted in WT mice, because of reduced numbers of T regulatory cells in B7DKO mice (43). The increased expression of IFN-γ and TNF-α mRNA at week 15 in the B7DKO could be a manifestation of the enhanced bacterial load in the lungs of these mice at that time. However, the exact mechanism behind increased cytokine mRNA expression in the B7DKO, despite suboptimal T cell priming, needs further investigation.
The ability of B7DKO mice to resist acute infection suggests that the early granulomatous response in the lung is not necessary to control acute disease, but critical for maintaining a stable chronic state in the host and preventing reactivation of chronic disease. Consistent with this, mice lacking ICAM-1 where migration of effector cells is disrupted (44) or mice lacking CD4 T cells (45), control M. tuberculosis growth in the absence of any appreciable granuloma formation during early infection, but succumb later with highly elevated bacterial load in the lungs.
An important question that still remains unanswered is whether the initial defect in Th1 priming is directly responsible for the lack of granuloma maturation in B7DKO mice or alternatively does B7 costimulation have additional roles in modulating the granulomatous response. Our finding that recruitment of B cells, a cell type that has recently been shown to contribute significantly to bacterial containment in the lung (46), was reduced in the B7DKO lungs, specifically during chronic infection, suggests that the reduced T cell numbers in the B7DKO lungs may alter the chemokine microenvironment of the lung and down-modulate recruitment of specific cell subsets to the developing granuloma. Alternatively, B7 costimulation may also be critical for maintaining T cell viability in the granuloma, because earlier studies have reported that Th1 cells undergo apoptosis if stimulated in the absence of costimulation (47, 48). It is therefore tempting to speculate that B7 costimulation has multiple roles in the induction of host immunity against tuberculosis. Besides being necessary for optimal Th1 priming in the lymph nodes, B7 costimulation may also be necessary for up-regulation of chemokine expression and enhanced cellular recruitment into the granuloma, and for sustaining T cell viability within the granuloma. Further analysis is required to evaluate the contribution of B7 in regulating the chemokine network and maintaining T cell viability in a developing tubercle granuloma.
The two possible outcomes in individuals exposed to M. tuberculosis are rapid progression to active tuberculosis or entry into latent infection with the risk of endogenous reactivation tuberculosis later in life. For the most part the immunological forces that maintain a latent infection remain ill defined, except for the knowledge that TNF-α is a major player (49). Absence of TNF-α leads to reactivation disease as has been shown in murine models (50), and importantly in humans receiving anti-TNF-α therapy for chronic inflammatory diseases (51). The present findings that B7 is important for granuloma integrity would predict that inhibiting B7 function during chronic infection would lead to disease reactivation, akin to that seen with anti-TNF-α therapy. Interestingly, blocking B7, via CTLA4Ig, did not result in reactivation disease in mice chronically infected with M. tuberculosis (52). With the advent of sophisticated tools for mycobacterial genotyping there is now accumulated evidence that exogenous reinfections occur in individuals with previously ascertained latent infection and in individuals cured from an episode of tuberculosis (53, 54, 55). Importantly, this suggests that CTLA4Ig and similar costimulation modulators, despite having minimal effect on reactivation disease, can still inhibit B7/CD28 signaling pathway to attenuate Th1 cell induction and mycobactericidal lung immunity during an initial response to M. tuberculosis. Further studies are clearly warranted to establish whether such modulators will increase progression to active disease following a primary infection or exogenous reinfection with M. tuberculosis.
We thank Arlene Sharpe for providing the B7DKO mice and David Lagunoff for help with the histopathological evaluations.
The authors have no financial conflict of interest.
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.
↵1 This work was supported by National Institutes of Health Grant AI49778 and The Potts Memorial Foundation.
↵2 Current address: Beth Israel Deaconess Medical Center, Boston, Massachusetts 02215.
↵3 Address correspondence and reprint requests to Dr. Padmini Salgame, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, 185 South Orange Avenue, MSB Room A-902, Newark, NJ 07101. E-mail address:
↵4 Abbreviations used in this paper: WT, wild type; CFU, colony forming unit; Ct, cycle threshold; 7-AAD, 7-aminoactinomycin D.
- Received September 9, 2008.
- Accepted January 16, 2009.
- Copyright © 2009 by The American Association of Immunologists, Inc.