Published online before print June 30, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1,
,
,2
* Department of Microbiology and Divisions of
Infectious Diseases and
Internal Medicine and
Center for Microbial Interface Biology, The Ohio State University, Columbus, Ohio, USA
2 Correspondence at current address: Chemical and Biological Defense Directorate, Defense Threat Reduction Agency, 8725 John J. Kingman Rd., Fort Belvoir, VA, 22060, USA. E-mail: paula.bryant{at}dtra.mil
ABSTRACT
Antigen presentation by class II MHC molecules in the uninfected host is a multi-step process involving key functions provided by specific cathepsins (Cat) and the peptide editor DM. Herein, we examined the requirement for each of these components in mice to control a low-dose aerosol infection with Mycobacterium tuberculosis (MTB). Mice lacking Cat B, -L, or -S were similar to wild-type in their ability to control the growth and dissemination of MTB. In contrast, DM–/– mice failed to limit MTB growth and showed 
100-fold higher bacterial burden in the lung and spleen (5–6 weeks postinfection) as compared with wild-type and Cat-deficient mice. Histopathology revealed impaired cellular recruitment and altered granuloma formation in the lungs of MTB-infected DM–/– mice. Moreover, despite impaired thymic selection in Cat L–/– and DM–/– mice, MTB-specific CD4+ T cells were elicited only in the former. The lower numbers of MTB-specific CD4+ T cells available in Cat L–/– mice as compared with wild-type animals were sufficient to control MTB growth and dissemination. In addition, DM–/– macrophages infected with MTB in vitro were unable to stimulate pathogen-specific T cells. The data indicate that the majority of antigens derived from MTB are loaded onto nascent class II MHC molecules via the classical DM-dependent pathway.
Key Words: class II MHC • antigen presentation • CD4+ T cell
INTRODUCTION
IFN-
-producing CD4+ T cells are critical for control of Mycobacterium tuberculosis (MTB) infection [1
]. MTB-specific CD4+ T cells are elicited upon recognition of class II MHC molecules presenting mycobacterial-derived antigens on the surface of infected cells. Accordingly, mice lacking class II MHC molecules or CD4+ T cells are unable to control a primary aerosol infection with MTB and succumb to disease [2
3
4
]. The pathway leading to the generation of nascent peptide-loaded class II MHC molecules in "uninfected" APCs is a multi-step process involving key functions provided by the chaperone invariant chain (Ii), specific cathepsins (Cat), and the peptide-editor DM. However, the individual contribution of these accessory molecules to the immune response to MTB in vivo remains undefined.
MTB invades and replicates in phagocytic cells {i.e., macrophages, dendritic cells (DCs) [5
]} of the host. The pathogen and/or its antigenic products are digested into class II-presentable peptides by proteases (i.e., Cat), while traversing the endocytic route. Class II MHC molecules intersect these peptides via association in the endoplasmic reticulum with Ii, which delivers them directly to the same endocytic vesicles [6
]. A segment of the lumenal region of Ii, designated class II-associated Ii-derived peptides (CLIP), fills the peptide-binding groove of class II. Aspartyl and cysteine proteases progressively degrade Ii via a series of defined intermediates (i.e., 
β-Iip22, β-Iip10) until CLIP is all that remains in the peptide-binding groove [7
]. The rate-limiting step of class II peptide-loading, cleavage of β-Iip10 into β-CLIP, is performed by Cat S in bone marrow (BM)-derived (professional) APCs [8
9
10
]. Accordingly, Ii breakdown and subsequent peptide loading are delayed significantly in the periphery of Cat S-deficient mice [8
9
10
11
]. As Cat S is not expressed in cortical thymic epithelial cells (cTECs) that mediate positive selection, Cat S–/– mice exhibit normal numbers of CD4+ T cells. Instead, Cat L is required to cleave Iip10 to CLIP in cTECs. Thus, mice lacking Cat L contain 50–75% fewer CD4+ T cells as compared with wild-type mice [12
].
The exchange of CLIP for antigenic peptides is catalyzed by DM. Moreover, DM functions as a peptide editor, selecting for peptides, which when bound in the cleft of class II molecules, create kinetically stable ligands for CD4+ T cells [13
, 14
]. In DM-deficient mice, 99% of class II molecules (I-Ab) are predominantly loaded with CLIP instead of the usual diverse array of self-peptides [15
]. As a result, CD4+ T cell selection is impaired, resulting in 50–75% less CD4+ T cells in the periphery as compared with wild-type mice, similar to Cat L-deficient animals. In contrast to Cat L–/– mice, the CD4+ T cells in DM–/– mice are selected in the thymus by a single class II-peptide combination (i.e., β-CLIP), yielding a CD4+ T cell repertoire that lacks some specificities typical of wild-type mice [15
16
17
].
MTB attempts to avoid hydrolysis and the "classical" class II antigen-presentation pathway outlined above by preventing the fusion of its phagosome with late-endocytic vesicles (reviewed in ref. [18
]), an evasion tactic that is overcome upon IFN- activation of macrophages [19
]. In turn, the pathogen impairs a subset of IFN- signaling pathways in the macrophage, including IFN--induced class II MHC transcription. Interference with IFN- signaling was shown to be the result of chronic (>24 h) TLR2 stimulation by mycobacterial lipoproteins [20
21
22
23
]. Nonetheless, a class II-restricted, MTB-specific CD4+ T cell response capable of controlling MTB replication and dissemination is elicited in most individuals infected with MTB, albeit it is not eradicating. An alternative Cat S- and DM-independent pathway of class II MHC peptide loading has been described in early [24
] and recycling [25
26
27
28
29
] endosomes, which could account for the effective CD4+ T cell response elicited against MTB. Thus, in the midst of MTB interference, we asked whether class II antigen presentation in the MTB-infected host relies on individual components of the classical class II peptide-loading pathway described above. We show that Cat B, Cat L, and Cat S are not required to mount a successful CD4+ T cell-mediated immune response in vivo against MTB. In contrast, DM is required to control a primary aerosol infection with the pathogen. The data indicate that the majority of antigens derived from MTB are processed and loaded onto newly synthesized class II molecules via the classical DM-dependent pathway. Furthermore, the ability of Cat L–/– animals to control MTB growth and dissemination suggests that low numbers of MTB-specific CD4+ T cells are sufficient to control pathogen replication, at least during the early stages following infection.
MATERIALS AND METHODS
Mice
C57BL/6, (129xC57BL/6)F1, and (129xC57BL/6)F2 mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). Cat B–/– [30
], Cat L–/– [31
, 32
], Cat S–/– [9
], and DM–/– (backcrossed to C57BL/6 for more than or equal to nine generations) [33
] mice have been described. All mice were bred and maintained in conformity with the institutional guidelines for animal research at the Ohio State University (OSU; Columbus, OH, USA). All experiments involving mice were performed under protocols approved by the Institutional Animal Care and Use Committee at OSU in a BSL3 facility.
MTB infection of mice
MTB Erdman (ATCC #35801) was obtained from American Type Culture Collection (Manassas, VA, USA) and grown in Proskauer-Beck liquid medium containing 0.05% Tween 80 to mid-log phase and frozen in aliquots at –80°C. Mice were infected aerogenically with a low dose of MTB Erdman using the Glas-Col (Terre Haute, IN, USA) inhalation exposure system. Briefly, the nebulizer compartment was filled with a suspension of bacteria calculated to deliver between 50 and 100 viable bacteria into the lung per mouse during a 40-min exposure. The numbers of viable bacteria in the lungs and spleens were determined at various time-points postinfection by plating serial dilutions of organ homogenates onto Middlebrook 7H11 agar (Becton Dickinson, San Diego, CA, USA), supplemented with oleic acid-albumin-dextrose-catalase (Becton Dickinson), and enumerating colonies after 21 days of incubation at 37°C, 5% CO2. For each experiment, the initial inoculum delivered to the mice was determined by homogenizing and plating the lungs of four to five individual mice 1 day after the aerosol exposure (data not shown). Data are expressed as the mean log10 numbers of CFUs recovered per organ (n=4–5 mice). Statistical significance was determined using the Prism 4 software (GraphPad Software, San Diego, CA, USA). A one-way ANOVA with Tukey’s post-test was used to compare results obtained from an individual group (i.e., knockout mouse strain) with wild-type. Differences were considered significant where P < 0.05.
Generation of single-cell suspensions from infected lungs
Mice were killed by CO2 asphyxiation, and their thoracic cavities were opened aseptically. The lungs were cleared of blood by injecting 5–10 ml saline containing 50 U/ml heparin (Sigma-Aldrich, St. Louis, MO, USA) into the right ventricles. The lobes were harvested in complete RPMI [i.e., RPMI 1650, 10% heat-inactivated FCS, 100 U/ml penicillin-streptomycin, and 2 mM L-glutamine (all Gibco, Grand Island, NY, USA)]. The lung lobes were sliced into small pieces (1 mm3) and incubated in RPMI containing collagenase XI (0.7 mg/ml, Sigma-Aldrich) and type IV bovine pancreatic DNase (30 µg/ml, Sigma-Aldrich) for 30 min at 37°C. The digested lungs were then gently dispersed through a nylon screen. The RBCs were lysed with ACK lysis buffer (0.15 M NH4Cl, 10 mM KHCO3, 0.1 mM Na2EDTA, pH 7.2), and the remaining lung cells were resuspended in complete RPMI and counted.
Isolation and purification of CD4+ T cells
Spleens were harvested from MTB-infected mice and dispersed through a nylon screen. Homogenates from four to five spleens from the same strain of mice (i.e., wild-type, Cat L–/–, or DM–/–) were pooled, the RBCs were lysed in ACK lysis buffer, and the remaining spleen cells were resuspended in complete RPMI and counted. The T cells were enriched over a nylon wool column and then incubated with anti-CD8 (clone TIB 210) followed by rabbit complement to deplete CD8+ T cells. The purified CD4+ T cells were run over a lympholyte M gradient to eliminate nonviable cells, washed twice in complete RPMI, counted, and diluted to the required concentration. The number of CD8+ cells in the final preparation was less than 1% as determined by flow cytometry (data not shown).
Preparation of BM-derived macrophages (BM-macs)
BM was harvested from the femurs of 8- to 12-week-old uninfected mice and differentiated into macrophages by culturing the BM precursors in bacterial-grade dishes for 6 days in RPMI 1640 (10% FCS+10% L-glutamine+10 penicillin/streptomycin) in the presence of 10 ng/ml recombinant murine (rm)GM-CSF (PeproTech, Rocky Hill, NJ, USA). On Days 3 and 5 of the culturing period, 70% of the culture supernatant containing nonadherent cells was removed and replaced with fresh media containing 10 ng/ml rGM-CSF. On Day 6, the loosely adherent and nonadherent cells (representing granulocytes and DCs [34
]) were removed by vigorous washing. The remaining adherent macrophage population was harvested by incubating the cells for 5 min in PBS + 10 mM EDTA followed by gentle scraping with a rubber policemen (Starstedt, Newton, NC, USA). The harvested BM-macs were resuspended in complete RPMI 1640 lacking rGM-CSF and plated in tissue-cultured, treated 96-well, flat-bottom plates (1x105/well) in the presence of 100 U/ml rmIFN- (PeproTech) for 48 h prior to use in the antigen-presentation assays described below.
Antigen presentation
Frozen aliquots of MTB Erdman were thawed, pelleted, washed with infection medium (antibiotic-free RPMI supplemented with 10% heat-treated FCS), and subjected to two, 20-s rounds of sonication to remove any large clumps. A single-cell bacterial suspension was confirmed by microscopy. The IFN--activated BM-macs seeded onto 96-well plates as described above were mock-infected or infected with varying concentrations of the viable, clump-free MTB Erdman for 6 h, followed by three washes to remove nonphagocytosed bacilli. Acid-fast staining revealed that a multiplicity of infection (MOI) of 20 bacilli to one BM-mac resulted in the infection of >85% of the total BM-macs plated with two to three bacilli phagocytosed per cell. The viability of the infected cells was confirmed by trypan blue exclusion. Whole lung cells, lymph node cells, or purified, splenic CD4+ T cells (isolated from uninfected and infected animals) were added (2x105 cells per well) to the uninfected and infected BM-macs and incubated for 24 h (to measure IL-2) or 72 h (to measure IFN-). Following each incubation period, the plates were centrifuged briefly and then frozen at –80°C until ELISAs were performed. Data points are means of triplicate wells ± SD. A one-way ANOVA with Tukey’s post-test was used to compare results obtained from an individual group (i.e., knockout mouse strain) with wild-type. Differences were considered significant where P < 0.05.
Cytokine ELISAs
Ninety-six-well plates (Immunoblot) were coated with anti-mouse IL-2 or anti-mouse IFN- overnight at 4°C. The wells were blocked with 1% BSA (Fisher, Fairlawn, NJ, USA) for 2 h at room temperature. The culture supernatants (50 µl) from the antigen-presentation assays described above or 50 µl media containing standard concentrations of rIL-2 or IFN- were added to the wells and incubated overnight at 4°C. Biotinylated anti-IL-2 or anti-IFN- detection antibody was added to each well for 60 min followed by the addition of streptavidin-alkaline phosphatase for 30 min. The wells were washed extensively between each step. Finally, substrate (p-nitrophenyl phosphate) was added to each well, and the color-change intensity was measured at 405 nm. Standard curves were prepared, and the amount of IL-2 and IFN- was measured for each sample. Unless otherwise indicated, all reagents were purchased from BD PharMingen (San Diego, CA, USA).
Histology
The right caudal lung lobe was isolated from each individual mouse and inflated with and stored in 10% formal-buffered saline. The lung tissue was processed, sectioned, and stained with H&E for light microscopy with lobe orientation designed to allow for maximum surface area of each lobe to be seen. Serial sections were stained with Ziehl Neelsen acid-fast stain to visualize bacilli.
RESULTS
DM-deficient mice fail to control MTB growth and dissemination
To determine the requirements for individual components of the class II MHC antigen-presentation pathway in the immune response against MTB, we examined the abilities of Cat B-, Cat S-, Cat L-, and DM-deficient mice to control a low-dose (100 CFUs) aerosol infection with MTB as compared with wild-type animals. Although an essential role for Cat B in Ii breakdown and class II peptide loading has not been defined, this enzyme may participate in the liberation of key T cell epitopes from mycobacterial-derived antigens, and thus, its absence could influence the effectiveness of the immune response against the pathogen. All mice showed equivalent bacterial growth (i.e., no statistical differences) in the lungs and spleens 21 days postinfection (data not shown). By 6 weeks postinfection, bacterial growth in the lungs and spleens was under control in wild-type, Cat B–/–, Cat S–/–, and Cat L–/– mice, indicating that the functions of these cysteine proteases, at least when considered independently, are not required for a successful immune response against MTB (Fig. 1A
). In contrast to wild-type and Cat-deficient mice, MTB replication continued in the lungs and spleens of DM–/– mice, resulting in significantly higher bacterial burdens at 6 weeks postinfection (Fig. 1A)
. At this time-point, DM–/– lungs contained 107 CFUs as compared with the 106 CFUs in wild-type lungs. Moreover, the bacterial load in DM–/– spleens approached 107 CFUs, almost 2 log higher than that in wild-type spleens. Similar results were observed 5 weeks following a low-dose aerosol infection in a separate experiment (Fig. 1B)
. Thus, the role of DM in class II-restricted antigen presentation appears indispensable in the immune response against MTB.
 View larger version (34K): [in a new window] |
Figure 1. DM–/– mice are unable to control MTB growth in vivo. Eight- to 10-week-old wild-type, Cat- and DM-deficient mice were infected via aerosol to achieve 100 CFUs of MTB Erdman per lung. Lungs and spleens were harvested 3 weeks (data not shown) and 6 weeks postinfection (p.i.) in the first experiment (A) and 5 weeks postinfection in a repeat experiment (B), and homogenates were plated to determine CFUs. The data are expressed as the log10 value of the mean number of CFUs recovered from four to five individual animals. A one-way ANOVA with Tukey’s post-test was used to determine statistical significance of bacterial count between an individual group (i.e., knockout mouse strain) with wild-type mice. Error bars represent means of four to five mice ± SD. (A) The bacterial count in DM–/– lungs and spleens at 6 weeks postinfection was significantly (***, P<0.001) higher than in wild-type and Cat-deficient organs. (B) The bacterial count in DM–/– lungs (**, P<0.01) and spleens (***, P<0.001) at 5 weeks postinfection was significantly higher than wild-type and Cat-deficient organs.
|
Granuloma formation is altered in the lungs of MTB-infected DM–/– mice
The formation and maintenance of lymphocyte-rich granulomas are essential for the control of mycobacterial infections [35
]. As DM–/– mice were unable to control MTB replication, tissue sections from the lungs of wild-type and DM-deficient mice 6 weeks postinfection were examined for histopathology. H&E staining showed that the lungs of wild-type mice contained multifocal granulomatous lesions normally observed within the lungs of MTB-infected, wild-type mice [36
] (Fig. 2 A-C
). Lymphoid aggregates and perivascular and peribronchial cuffing were observed, indicative of chronic inflammation. Acid-fast bacilli were observed in wild-type lungs, which appeared in clusters that were well-contained within granulomatous lesions (Fig. 2G)
. In marked contrast, the lung tissue from DM–/– mice exhibited minimal lung involvement or granuloma formation. The majority of lung tissue appeared relatively healthy, with negligible perivascular and peribronchiolar cuffing (Fig. 2D)
. Interestingly, where cellular infiltrates were observed, the granulomas were dominated by macrophages with few lymphocytes present (Fig. 2 E and F)
. Consistent with the high CFUs observed, acid-fast staining showed that there were more clumps of bacilli per unit area in DM–/– lungs than in wild-type lungs (Fig. 2H)
. More significantly, numerous macrophages containing acid-fast bacilli could be observed within the alveoli (Fig. 2I)
. Thus, although the lung appeared relatively healthy with minimal signs of inflammation, acid-fast bacilli were observed in macrophages within the alveoli, without any overt recruitment of inflammatory cells.
 View larger version (147K): [in a new window] |
Figure 2. Granuloma formation is disrupted in MTB-infected DM–/– mice. Lungs (right caudal lobes) were isolated from wild-type and DM–/– mice 6 weeks postinfection, buffered in 10% formalin, sectioned, subjected to H&E staining (A–F) and Ziehl Neelsen staining (G–I), and examined by microscopy. (A–C) Wild-type lung showed granuloma formation with perivascular and peribronchial cuffing. Acid-fast bacilli (arrows) were located in clusters within distinct granuloma (G). In contrast, (D–F) DM–/– lung exhibited minimal granuloma formation, few lymphocytes, and minimal perivascular cuffing. (H) Acid-fast bacilli were located in small clusters throughout the small granuloma. (I) Macrophages infected with acid-fast bacilli were also seen in DM–/– alveoli without any recruitment of inflammatory cells. (A and D) Original magnification, 2x; (B and E) 10x original magnification; (C and F–H) 40x original magnification; (I) 100x original magnification.
|
 View larger version (22K): [in a new window] |
Figure 3. MTB-specific T cells are present in the lungs of infected Cat L–/– but not DM–/– mice. IFN--stimulated BM-macs were prepared from naïve B6 mice and infected in vitro with varying MOIs of MTB for 6 h and washed. The macrophages were then incubated for 24 h with whole lung cells prepared from B6, L–/–, and DM–/– mice 4 weeks post-MTB infection. The supernatants were collected and analyzed for IL-2 by ELISA. The results shown for each strain are representative of at least four individual mice per experiment and two separate experiments. One-way ANOVA with Tukey’s post-test was used to compare the results obtained from an individual group (i.e., knockout mouse strain) with wild-type. Error bars represent mean ± SD. All groups (i.e, wild-type lung response, L–/– lung response, and DM–/– lung response) were significantly different (P<0.001) from each other. The response observed for DM–/– lung cells when incubated with MTB-infected, wild-type macrophages was not significant (ns) when compared with that seen with no APCs, or unprimed B6, or uninfected APCs.
|
by mycobacterium-specific CD4+ T cells is required for the activation of macrophages and granuloma formation in vivo [35
, 37
], we examined the ability of CD4+ T cells purified from the infected mutant mice to produce IFN- upon incubation with MTB-infected macrophages in vitro. The use of equal numbers of CD4+ T cells "rescued" the deficiency exhibited by Cat L–/– cells considerably, although the response was still significantly lower than wild-type (Fig. 4
). Consistent with the results thus far, equivalent numbers of purified CD4+ T cells isolated from infected DM–/– mice did not respond significantly to MTB-infected macrophages (Fig. 4)
. Thus, these results suggest that DM–/– mice are unable to elicit MTB-specific CD4+ T cells following an aerosol infection.
 View larger version (17K): [in a new window] |
Figure 4. IFN--producing, MTB-specific CD4+ T cells are present in the spleens of infected Cat L–/– but not DM–/– mice. IFN--stimulated BM-macs prepared from naïve B6 mice were infected in vitro with varying MOIs of MTB for 6 h and washed. The macrophages were then incubated for 24 h with purified CD4+ splenic T cells isolated from B6, Cat L–/–, and DM–/– mice 4 weeks post-MTB infection. The supernatants were collected and analyzed for IFN- by ELISA. The results shown for each strain are representative of at least three individual mice per experiment and three separate experiments. One-way ANOVA with Tukey’s post-test was used to compare results obtained from an individual group (i.e., knockout mouse strain) with wild-type. All groups (i.e., B6 splenic CD4+ T cell response, L–/– splenic CD4+ T cell response, and DM–/– splenic CD4+ T cell response) were significantly different (P<0.001) from each other. The response observed for DM–/– CD4+ T cells when incubated with MTB-infected, wild-type macrophages was not significant when compared with that seen with no APCs or uninfected APCs.
|
-stimulated BM-macs generated from wild-type, Cat L–/–, and DM–/– mice were infected in vitro with viable MTB for 6 h and then fixed. Whole lung cells isolated from MTB-infected, wild-type mice were incubated with the infected, wild-type and mutant macrophages, and their production of IL-2 and IFN- was examined. B6 and Cat L–/– macrophages were indistinguishable in their ability to process MTB bacilli and present MTB antigens to the pathogen-specific T cells, as evidenced by similar levels of IL-2 (Fig. 5A
) as well as IFN- production (Fig. 5B)
. In marked contrast, the MTB-infected DM–/– macrophages did not stimulate pathogen-specific lung cells (Fig. 5 A and B)
. Similar results were obtained when mediastinal lymph node T cells and splenic T cells isolated from MTB-infected, wild-type mice were used as responders (data not shown). These data demonstrate a critical role for DM within the macrophage and suggest that the majority of MTB-derived antigens are processed and presented by newly synthesized class II molecules via the classical DM-dependent pathway, as opposed to the recycling pathway.
 View larger version (30K): [in a new window] |
Figure 5. DM is required for macrophages to present MTB antigens. IFN--stimulated BM-macs prepared from naïve B6, Cat L–/–, and DM–/– mice were infected in vitro with varying MOIs of MTB for 6 h, washed, and fixed. The BM-macs were then incubated for 24 h with whole lung cells isolated from wild-type mice 4 weeks postinfection. The supernatants were analyzed for IL-2 (A) and IFN- (B) by ELISA. The results are representative of three separate experiments. One-way ANOVA with Tukey’s post-test was used to compare results obtained from an individual group (i.e., knockout macrophage) with wild-type (***, P<0.001). DM–/– macrophages were unable to significantly stimulate pathogen-specific lung cells to produce IL-2 or IFN-.
|
The success of the immune response against MTB requires that infected APCs process MTB-derived antigens for presentation by class II MHC molecules to CD4+ T cells. In vivo, class II antigen presentation is required for the activation and differentiation of naive T cells in the draining mediastinal lymph nodes and for stimulating effector cells arriving in the lungs. The generation of peptide-loaded class II molecules relies heavily on functions provided by cysteine proteases and the peptide editor DM. Previous studies have examined the course of MTB infection in mice completely lacking class II or CD4 T cells [2 3 4 ]. However, no study to date has examined MTB infection in mice in which the class II genes and their transcription/translation are intact, but the peptide-loading machinery is defective. In this new study, we show that whereas Cat B, Cat S, and Cat L are dispensable, DM is required to control a primary aerosol infection with MTB.
Cat B, -S, and -L represent the predominant cysteine proteases expressed in professional APCs. The role of Cat S and -L in Ii breakdown was discussed above. In addition, Cat B, -S, and -L have been shown to participate in the processing of internalized antigens into class II-presentable T cell epitopes [7 , 38 39 40 ]. Our data show that despite the role of Cat S for efficient Ii breakdown in professional APCs, its absence did not interfere with the ability of mice to gain control of MTB replication by 6 weeks postinfection. In addition, the data indicate that none of the enzymes examined is absolutely required to liberate T cell epitopes from MTB antigens during the acute phase of infection. Moreover, the T cell repertoires selected in these mutant mice appear to possess the necessary specificities to mount an effective antigen-specific response against MTB (although further studies are warranted to determine the role of these enzymes in generating an effective memory T cell response).
Most strikingly, despite reduced numbers of CD4+ T cells available in the periphery of Cat L–/– mice (50–75% fewer than wild-type [12 ]), these animals were able to control a primary aerosol infection with MTB similar to wild-type mice. Our data showed that MTB-specific CD4+ T cells were elicited in Cat L–/– animals, although the MTB-specific cellular response in Cat L–/– mice was significantly lower than that of wild-type mice. In addition, the data demonstrated that Cat L was dispensable for the generation of MTB-derived CD4+ T cell epitopes, as Cat L–/– macrophages were as efficient as wild-type macrophages at processing MTB bacilli for presentation of MTB antigens via class II MHC molecules to pathogen-specific T cells. These data suggest that it is not the quantity of pathogen-specific CD4+ T cells elicited that dictates the course of MTB infection.
Unlike Cat-deficient mice, DM–/– animals were unable to control MTB replication and dissemination following a low-dose aerosol infection. A striking feature of MTB-infected DM–/– mice was the scarcity of granulomas and the lack of any obvious inflammatory response in the lung, despite the presence of macrophages harboring readily observable bacilli. The MTB-infected macrophages present in DM–/– lungs were not surrounded by inflammatory cells or lymphocytes typically seen in a successful granulomatous response. Those granulomas that were found in DM–/– lungs consisted predominantly of macrophages. Few lymphocytes were observed. These findings are similar to Saunders et al. [37 ], who described mild macrophage accumulations, with few lymphocytes, in the lungs of CD4–/– mice (40 days postinfection). As infection progressed, CD4–/– mice developed pyogranulomatous lesions, again, with relatively few lymphocytes. Our studies did not extend beyond 6 weeks of infection, and longer time-points would be required to determine whether DM–/– mice eventually succumb to infection or simply maintain the high bacterial burden. However, given the similarities between DM–/– mice and CD4–/– mice, we would anticipate that DM–/– animals would exhibit reduced survival relative to the wild-type control mice.
Critical to the formation of an effective granulomatous response are the elicitation and recruitment of pathogen-specific CD4+ T cells to the lung. Similar to Cat L–/– mice, DM–/– mice contain 50–75% fewer CD4+ T cells than wild-type [15 ]. The reduced numbers of peripheral CD4+ T cells available in DM–/– mice unlikely contributed to their susceptibility to MTB infection, as similar numbers were sufficient in Cat L–/– mice. Instead, our data showed that T cells isolated from MTB-infected DM–/– mice failed to respond ex vivo when incubated with wild-type macrophages infected with MTB (in vitro), suggesting MTB-specific T cells were not elicited in infected DM–/– mice. In addition to impaired CD4+ T cell activation, a failure in leukocyte recruitment to the lung may contribute to the susceptibility of DM–/– mice, a possibility that warrants further investigation.
Consistent with the lack of pathogen-specific CD4+ T cells and altered granuloma formation in infected DM–/– animals, DM–/– macrophages infected with MTB in vitro were unable to stimulate primed CD4+ T cells isolated from infected B6 mice. As the activation of CD4+ T cells is essential for the formation of granulomas during MTB infections [35 , 37 ], these data indicate that the failure to elicit MTB-specific CD4+ T cells in DM–/– mice and form "effective" granulomas in the lung was likely a result of the requirement for DM in class II peptide loading in peripheral APCs. Moreover, the data suggest that most, if not all MTB antigens, are loaded onto class II molecules in a DM-dependent manner. Thus, the pathway used by infected macrophages to process and present MTB-derived antigens appears to be that involving newly synthesized class II (DM-dependent) as opposed to recycled class II. The latter involves internalization of class II molecules expressed on the cell surface and exchange of peptide in the recycling endosome, a process shown not to require DM [24 25 26 27 28 29 ].
The DM–/– mice used in this study were H-2b haplotype. I-Ab forms unusually stable complexes with CLIP and thus, relies heavily on DM-mediated CLIP-peptide exchange—at least in late endocytic compartments [15 , 24 ]. Other class II alleles, with weaker affinities for CLIP, vary to the degree at which they require DM to mediate CLIP release and exchange for antigenic peptides [41 , 42 ]. Nonetheless, these alleles are still dependent on the function of DM as a peptide editor [13 , 14 ]. When DM-deficient mice with less-stringent dependency on DM for peptide loading were infected with Leishmania major, the CD4+ T cell response was skewed away from the immunodominant epitope typically seen in L. major-infected, DM-sufficient mice [43 ]. Thus, even if a class II allele does not need DM to release CLIP from the peptide-binding groove, the "editing" function of DM that selects the immunodominant epitope(s) required to elicit an effective CD4+ T cell response may be essential. Further studies using different haplotype mice deficient for DM are needed to assess the role DM plays in generating an effective CD4+ T cell response against potential immunodominant, MTB-derived antigens.
Although a "wild-type" immune response to MTB is effective at controlling pathogen replication and disease progression, this response is not eradicating. Also unanswered is why it takes 21 days for significant numbers of MTB-specific, effector CD4+ T cells, capable of controlling MTB replication, to arrive in the lungs following infection [44
]. This "delay" may be a result of defects in the activation of naïve T cells by DCs in the draining lymph nodes or in the recruitment and/or trafficking of the resulting effector T cells to the lung. Numerous studies have characterized mechanisms by which MTB attempts to interfere with the antigen-presentation functions of macrophages, including avoidance of phagosome-lysosome fusion, alteration of intravesicular pH, and impairment of IFN--induced signaling pathways that drive class II transcription, each of which may contribute to a failure to eradicate the pathogen completely. More recent studies have shown that MTB also interferes with the proteolytic environment of macrophages. Previously, we showed that when murine macrophages were infected with MTB or Mycobacterium avium in vitro, maturation and activity of Cat L were inhibited, and Cat S activity remained largely unaffected [45
]. Another study showed that infection of human macrophages with Mycobacterium bovis bacillus Calmette-Guérin inhibited Cat S activity via the induction of IL-10, resulting in the reduced expression of class II MHC peptide complexes at the cell surface [46
]. Whether Cat S activity would also be impaired in DCs participating in the immune response against MTB or whether this inhibition occurs in vivo is not known. If MTB does indeed inhibit Cat S activity in vivo, the fact that Cat S–/– and wild-type mice were indistinguishable in our experiments in their ability to control a low-dose aerosol infection with MTB may not be surprising.
In summary, our new data show that the immune response to MTB in vivo and antigen presentation by macrophages infected with MTB in vitro are dependent on DM function, implying that peptide-loading of nascent class II molecules follows the classical pathway (i.e., nonrecycling/early endosome) in infected APCs. The data obtained comparing Cat L–/– with DM–/– mice, both containing similar reduced numbers of CD4+ T cells, indicate that the quantity of CD4+ T cells available in the periphery does not determine the ability to mount an effective immune response against MTB. Instead, the integrity of the class II pathway in peripheral APCs appears indispensable.
ACKNOWLEDGEMENTS
This work was supported in part by the OSU Seed Grant. R. M. N. was sponsored by a Fulbright Scholarship.
FOOTNOTES
1 Current address: Department of Immunology, University of Toronto, Toronto, Canada. 
Received December 23, 2007; revised April 24, 2008; accepted June 5, 2008.
REFERENCES
, yet succumb to tuberculosis J. Immunol. 162,5407-5416 interferon-regulated HLA-DR and FcgR1 on human macrophages through Toll-like receptor 2 Infect. Immun. 71,4487-4497(–/–) mice show the importance of major histocompatibility complex-bound peptide in cardiac allograft rejection J. Exp. Med. 192,31-40 receptors Eur. J. Immunol. 31,1592-1601[CrossRef][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
