Dr. Joanne Turner, 1010 BRT, 460 West 12th Avenue, Columbus, OH 43210, USAFax: +1-614-292-9616 e-mail: firstname.lastname@example.org
In CBA/J mice, susceptibility to Mycobacterium tuberculosis (M.tb) is associated with low interferon-gamma (IFN-γ) responses to antigens (Antigen 85 (Ag85) and early secreted antigenic target-6 (ESAT-6)) that have been defined as immunodominant. Here, we asked whether the failure of CBA/J mice to recognize Ag85 is a consequence of M.tb infection or whether CBA/J mice have a general defect in generating specific T-cell responses to this protein antigen. We compared CBA/J mice during primary M.tb infection, Ag85 vaccination followed by M.tb challenge, or M.tb memory immune mice for their capacity to generate Ag85-specific IFN-γ responses and to control M.tb infection. CBA/J mice did not respond efficiently to Ag85 in the context of natural infection or re-infection. In contrast, CBA/J mice could generate Ag85-specific IFN-γ responses and protective immunity when this antigen was delivered as a soluble protein. Our data indicate that although M.tb infection of CBA/J mice does not drive an Ag85 response, these mice can fully and protectively respond to Ag85 if it is delivered as a vaccine. The data from this experimental model suggest that the Ag85-containing vaccines in clinical trials should protect M.tb susceptible humans.
Up to one-third of the world's population may be infected with the bacterium Mycobacterium tuberculosis (M.tb), of which a small fraction develop pulmonary tuberculosis (TB) disease []. The underlying mechanisms of susceptibility to M.tb and TB disease progression in most adult human patients are poorly understood []. It is possible that subtle differences in the responses to immunodominant antigens may predispose the host to M.tb susceptibility and TB disease progression. Here, we focus on immune responses to Antigen 85 (Ag85) using a mouse model of M.tb aerosol infection.
Ag85 is detectable in TB patients [] and is an immuno-dominant antigen recognized by TB patients [[3, 4]]. Contacts of TB patients also respond variably to Ag85 [] but the cause(s) or significance of the variation are unknown. There is evidence that TB patients [] do not respond as strongly to Ag85 as do individuals who control M.tb infection. Together, these data show variability in human immune responses to Ag85 and associate weak Ag85 immune responses with M.tb susceptibility and TB disease. It is not known whether poor Ag85-specific interferon-gamma (IFN-γ) responses lead to TB disease in humans, or whether poor Ag85-specific responses are a marker of increased risk for TB disease. Studies in mice suggest the latter [[6, 7]].
Responses to Ag85 are important to understand because it is a vaccine component in human clinical trials []. The efficacy results from human Ag85-containing vaccine clinical trials are not available, but whether M.tb susceptible humans respond efficiently to Ag85 vaccination and M.tb challenge is unknown. We address these questions experimentally to determine whether the context of Ag85 exposure (natural infection/re-infection versus soluble protein) affects Ag85-specific IFN-γ responses and protective immunity.
We use a low-dose aerosol infection of M.tb susceptible CBA/J mice [[9, 10]] to model M.tb susceptible humans. We have shown that CBA/J mice have low IFN-γ responses to other M.tb antigens as compared with C57BL/6 mice [] and to Ag85 during early M.tb infection []. Additionally, M.tb susceptible C3H substrains ([], our unpublished observations) and DBA/2 mice (our unpublished observations) also respond poorly to Ag85 as determined by IFN-γ production during primary M.tb infection. Here, we extend our previous findings to show that CBA/J mice have low Ag85-specific IFN-γ responses throughout early and chronic M.tb infection and to M.tb re-infection. Given the very low IFN-γ responses to Ag85 during infection, we questioned whether CBA/J mice were capable of recognizing and responding to Ag85. Delivery of Ag85 as a soluble antigen (vaccine plus adjuvant (Adj) format) successfully generated Ag85-specific IFN-γ responses and protected CBA/J mice against M.tb to the same degree as C57BL/6 mice. These results indicate that although CBA/J mice can generate IFN-γ secreting Ag85-specific responses if initial encounter is in soluble, but not native, form.
Overall, our results suggest that CBA/J mice may model M.tb susceptible humans who do not respond strongly to Ag85 []. A low Ag85-specific response during natural M.tb infection, however, does not correspond to a failure to respond to Ag85 vaccination during challenge infection. Therefore, we predict that the Ag85-containing vaccines in clinical trials will afford some protection for M.tb susceptible humans.
Ag85-specific IFN-γ production from lung cells and blood of M.tb-infected mice
To determine whether CBA/J mice generated Ag85-specific IFN-γ responses during primary M.tb infection, we cultured lung and blood cells with M.tb Ag85, concanavalin-A (Con-A) or ovalb-umin (data not shown), and compared the responses to C57BL/6 mice, which are resistant to M.tb [[9, 10]] and respond vigorously to Ag85 [[6, 7]]. Similar to M.tb culture filtrate protein and early secreted antigenic target-6 (ESAT-6) [], lung cells from CBA/J mice produced significantly less Ag85-specific IFN-γ than did lung cells from C57BL/6 mice (Fig. 1A). In both mouse strains, peak Ag85-specific IFN-γ production occurred at day 21 of M.tb infection (Fig. 1A), which corresponds to Ag85 gene expression [], peak Ag85-specific T-cell numbers [], and stabilization of M.tb growth [[11, 12]]. CBA/J mice harbor elevated M.tb burdens [] and their lungs contain more Ag85 protein (Supporting Information Fig. 1) than C57BL/6 mice. It seems unlikely, therefore, that the low Ag85-specific IFN-γ responses in CBA/J mice result from limited antigen availability due to inadequate amounts of Ag85 in vivo. Peripheral blood mononuclear cells (PBMCs) from CBA/J mice also produced less Ag85-specific IFN-γ than did C57BL/6 PBMCs, but there was no peak of Ag85-specific IFN-γ from CBA/J PBMCs (Fig. 1B). Our positive control, the potent T-cell activator Con-A, showed that cells from CBA/J mice were not inherently defective in IFN-γ production (Fig. 1C and 1D).
Overall, CBA/J mice had poor Ag85-specific IFN-γ responses throughout early and chronic asymptomatic primary M.tb infection. The same is true for other M.tb antigens due to low frequencies of responding T cells []. We therefore investigated the numbers and potency of Ag85-specific IFN-γ-producing cells from CBA/J mice.
Numbers and potency of Ag85-specific IFN-γ-producing cells from M.tb-infected mice
ELISPOTs were used to determine whether low Ag85-specific IFN-γ production by CBA/J mice reflected a reduced frequency of responding cells or reduced secretion by individual cells. The lungs and blood of CBA/J mice contained significantly lower frequencies (Fig. 1E and G) and absolute numbers (data not shown) of Ag85-specific IFN-γ-secreting cells than did M.tb resistant C57BL/6 mice. The individual Ag85-specific IFN-γ-secreting cells from CBA/J mice also produced significantly less than the cells from C57BL/6 mice (Fig. 1F and H). Together, these results identified two reasons for poor Ag85-specific IFN-γ responsiveness in CBA/J mice: a 50- to 100-fold reduction in the frequency of cells and a twofold reduction in Ag85-specific IFN-γ output at the single-cell level. The mechanisms underlying poor Ag85-specific IFN-γ responses in CBA/J mice are unknown, but likely involve H-2k haplotype-dependent and -independent effects [].
We next asked whether CBA/J mice could respond to Ag85 by delivering exogenous Ag85 as a preexposure vaccine and whether CBA/J mice could generate Ag85-specific IFN-γ memory recall responses as a result of prior exposure to M.tb.
Exogenous delivery of Ag85-induced IFN-γ responses in CBA/J mice
We delivered Ag85 as a soluble protein with an Adj to determine whether Ag85 could be recognized by M.tb-challenged CBA/J mice in vivo. We hypothesized that the poor capacity of CBA/J mice to respond to Ag85 during primary infection would impair their ability to respond to Ag85 as a soluble protein when challenged with M.tb.
CBA/J mice were vaccinated with phosphate-buffered saline (PBS), Adj, Ag85 plus Adj, or Mycobacterium bovis bacillus Calmette–Guérin (BCG) and challenged with M.tb by aerosol at least 10 weeks after the first injection. Mice were euthanized 28 days later. Contrary to our expectations, but similar to M.tb resistant mice [[13, 14]], vaccination increased Ag85-specific IFN-γ responses in M.tb challenged CBA/J mice. In the lungs, the frequency of Ag85-specific IFN-γ-secreting cells in Ag85-vaccinated CBA/J mice was significantly higher than all other groups (Fig. 2A). Ag85 vaccination did not affect the frequency of circulating Ag85-specific IFN-γ-producing cells (Fig. 2C), nor did Ag85 vaccination affect IFN-γ output by individual lung or blood cells (Fig. 2B and D). M. bovis BCG vaccination of CBA/J mice increased IFN-γ output by individual Ag85-specific IFN-γ-producing cells (Fig. 2B).
These results show that CBA/J mice can successfully prime Ag85-specific cells when Ag85 is administered exogenously. Together with the previous results, this suggests that CBA/J mice can process and present Ag85 protein, but they do not recognize Ag85 efficiently in the context of an in vivo M.tb infection. Interestingly, M. bovis BCG increased the potency of responding Ag85-specific IFN-γ cells, suggesting that CBA/J mice recognize and respond to Ag85 from M. bovis BCG better than Ag85 from M.tb.
Ag85 vaccination protected CBA/J mice against pulmonary M.tb challenge
To determine whether exogenous Ag85 plus Adj provided protection, CBA/J mice were vaccinated, challenged, and M.tb burden quantified in the lungs and mediastinal (lung-draining) lymph nodes 28 days later. Ag85 plus Adj vaccination significantly reduced M.tb lung burden as compared with the controls (PBS and Adj) (Fig. 3A), but protection was not as good as M. bovis BCG vaccination (Fig. 3C). In the mediastinal lymph nodes, Ag85 plus Adj vaccination had little effect on M.tb burden, while M. bovis BCG vaccination significantly reduced M.tb load (Fig. 3B and D).
These results show that CBA/J mice can successfully develop lung-specific protection against M.tb due to exogenous Ag85 plus Adj, possibly mediated by increased numbers of Ag85-specific IFN-γ-producing cells. M. bovis BCG vaccination also protected CBA/J mice, as has been shown in other M.tb susceptible mice [].
Immune cells and granulomas in the lungs of vaccinated CBA/J mice
Vaccination reduces the size of M.tb granulomas [], but few Ag85-containing vaccine and challenge studies in mice have quantified immune cell subpopulations or granulomas. This was an important question to address because in addition to reducing M.tb burden an optimal vaccine should minimize nonspecific inflammation.
We enumerated and phenotyped the lung immune cells in vaccinated and challenged CBA/J mice. The flow cytometry gating strategy is shown in Supporting Information Fig. 2. Adj and Ag85 plus Adj groups had fewer total cells (Fig. 4A) and fewer immune cells (Fig. 4B–H) than did PBS and M. bovis BCG groups. Vaccination with M. bovis BCG resulted in similar cellularity and T-cell activation profiles as did PBS (Fig. 4A–G). Major histo-compatibility complex (MHC)II expression was enhanced on CD11c+ antigen-presenting cells (APCs) in mice that received BCG (Fig. 4H).
We next evaluated the lungs of vaccinated and challenged CBA/J mice to determine whether the differences in immune cell numbers were associated with differences in the spatial distribution of cells to granulomas or to lymphocytic cuffs. This was important to address because distribution of cells could impact pulmonary function or control of M.tb. Immune cells were dist-ributed differently in the lungs of each group (Fig. 5A–D) despite similarities in absolute numbers (Fig. 4). The two control groups, PBS and Adj, were similar in granuloma size and number but showed moderate differences in lymphocytic cuffs, suggesting that the Adj affects lymphocyte trafficking to or within the lungs. Comparison of Adj to Ag85 plus Adj showed that Ag85 vaccination increased the size and number of granulomas and increased the number of lymphocytic cuffs, although none of the differences were statistically significant. M. bovis BCG-vaccinated mice formed numerous, small granulomas. Representative lung granulomas and lymphocytic cuffs are shown (Fig. 5E–H).
In all groups at high magnification (40×), granulomas were composed of macrophages, lymphocytes, and neutrophils (data not shown); and there were small foci of necrotic/apoptotic cells (data not shown). As expected, intracytoplasmic acid-fast M.tb bacilli were present within macrophages in granulomas (data not shown).
Overall, these results show that lung immune cells were dist-ributed differently during primary and vaccine-challenge M.tb infections. There were also differences between Ag85 plus Adj and M. bovis BCG groups, showing that the type of vaccination affected the spatial organization of immune cells. Delivery of Ag85 as a soluble protein increased three of four parameters measured by morphometry: granuloma size, lymphocytic cuff size, and number of lymphocytic cuffs. These data suggest that the total cellularity should have been increased in Ag85 plus Adj groups. However, by flow cytometry, the cell numbers were reduced (Fig. 4). The cause of this discordance is not known. One possibility is that the large granulomas and/or lymphocytic cuffs in Ag85 plus Adj-vaccinated mice contained cells that were lost during processing due to increased fragility or adherence to plastic.
M.tb memory responses to Ag85 and protection in CBA/J and C57BL/6 mice
Our previous results show that CBA/J mice responded poorly to Ag85 during primary M.tb infection, but were fully capable of responding protectively with IFN-γ production when Ag85 was delivered exogenously. We next asked whether Ag85-specific resp-onses generated during primary M.tb infection of CBA/J mice (Fig. 1) were enhanced in a re-infection model. C57BL/6 mice were included in these experiments as controls for robust Ag85-specific IFN-γ responses (Fig. 1 and []). Mice were left untreated (primary infection), vaccinated with Ag85 plus Adj, or exposed to M.tb by aerosol (memory) as shown in the timeline (Fig. 6E). After 30 days of M.tb infection in the memory mice, all groups were treated with rifampicin (RIF) and isonicotinic acid hydrazine (INH) which reduced M.tb burdens in memory mice to undetectable levels (Fig. 6A). Upon completion of the drug treatment, all groups were exposed to M.tb by aerosol and euthanized 28 days later. Ag85-specific IFN-γ responses and protection in the lungs were determined.
Compared with primary infection, Ag85 plus Adj significantly increased Ag85-specific IFN-γ production from lung cells of both C57BL/6 and CBA/J mice (Fig. 6B). Lung cell cultures from primary or Ag85-vaccinated CBA/J mice produced less IFN-γ than did cultures from identically treated C57BL/6 mice (Fig. 6B). We anticipated that M.tb memory-immune mice would have enhanced Ag85-specific IFN-γ responses due to their prior exposure to M.tb. However, very little Ag85-specific IFN-γ was produced in memory mice of both mouse strains, possibly due to accelerated kinetics of memory antigen-specific IFN-γ responses. Alternatively, the low number of Ag85-specific cells generated from the primary infection may not expand or function protectively during recall responses.
To determine whether Ag85 vaccination and M.tb memory immunity afforded protection, lungs were harvested from separate groups of identically treated CBA/J and C57BL/6 mice, and the M.tb burdens were calculated at day 28 of infection. In both mouse strains, Ag85 vaccination significantly decreased the M.tb burden as compared with unvaccinated mice (Fig. 6C), and afforded about 1 Log10 of protection (Fig. 6D). M.tb memory-induced responses were superior in reducing the M.tb load (Fig. 6C), affording about 1.5 Log10 of protection in both mouse strains (Fig. 6D).
Together, these results associate increasing Ag85-specific IFN-γ responses with control of M.tb growth in Ag85 plus Adj-vaccinated CBA/J and C57BL/6 mice. Importantly, protection was equivalent in both mouse strains. The results from these experiments therefore demonstrate that a relative increase, rather than the absolute levels of Ag85-specific IFN-γ responses, may enhance M.tb control.
M.tb Ag85 includes three related proteins (A, B, C) that share structure and function []. Ag85 components are immunogenic in humans [] and in M.tb-resistant mice [[9, 10]] including C57BL/6 [[14, 19-23]], C57BL/10 [], C57BL/6 × BALB/c F1 hybrids [[13, 24]], and BALB/c mice [[19-21, 25, 26]]. Ag85 has been intensely investigated in animal models as a vaccine against M.tb in many formulations, including peptides or proteins [], fusion proteins [[13, 23, 24, 27]], recombinant M. bovis BCG [], recombinant vaccinia virus [], antigen-pulsed cell-based vaccines [[26, 30]], and DNA vaccines [[14, 20, 31, 32]]. Ag85 is also a component of the vaccines in human clinical trials against M.tb [[8, 29]]. Efficacy results are not yet available, and it is not known whether these Ag85-containing vaccines will protect humans. It is promising that in many [[13, 20, 23-26, 32]] but not all [[30, 31, 33]] experimental animal models, Ag85 vaccination enhances IFN-γ responses and protects against M.tb.
We are interested in understanding antigen-specific responses of M.tb susceptible CBA/J mice [] to model M.tb susceptible immunocompetent humans. M.tb susceptible C3H mice also have low responses to Ag85 during early primary M.tb infection [[6, 7]] due to haplotype-dependent T-cell responses []. Little is known, however, about the ability of any M.tb susceptible strains to respond to Ag85 throughout chronic asymptomatic M.tb infection or in the context of M.tb re-infection or vaccination followed by M.tb challenge. Here, we showed that CBA/J mice had low IFN-γ resp-onses to Ag85 throughout chronic primary infection (Fig. 1), yet this did not prevent CBA/J mice from recognizing and responding to Ag85 when it was delivered exogenously. In fact, delivery of Ag85 in an adjuvanted, preexposure vaccine reduced the M.tb burden by two- to tenfold in CBA/J lungs (Fig. 3). Similar to other mouse strains [[14, 25]], Ag85-induced protection in CBA/J mice was associated with more Ag85-specific IFN-γ-producing cells in the lungs (Fig. 2 and data not shown). As in other mouse strains [[23, 24]], M. bovis BCG vaccination of CBA/J mice provided superior protection, perhaps due to antigenic persistence or diversity. Thus, despite relatively low responses to Ag85 during primary infection or re-infection, M.tb susceptible CBA/J mice were fully capable of generating Ag85-specific protective immunity when the antigen was delivered in an appropriate context, that is as an exogenous soluble protein. Additional data from our laboratory also show that the contradictory responses of CBA/J mice to Ag85 in vivo during M.tb infection (low) versus responses induced by exogenous exposure and challenge (high and protective) are not limited to Ag85. Indeed, when CBA/J mice were exposed to ESAT-6 plus Adj and challenged with M.tb, they developed strong ESAT-6-specific IFN-γ responses and protective immunity in the lungs that was equivalent to protection in C57BL/6 mice (Supporting Information Fig. 3).
Our results demonstrate that low Ag85-specific IFN-γ responses from CBA/J mice resulted from two mechanisms: a low frequency of Ag85-specific IFN-γ-producing cells, and low Ag85-specific IFN-γ output by individual responding cells. CBA/J mice do not have general deficiencies in IFN-γ production (Fig. 1). The relative importance of frequency versus output by individual Ag85-responsive cells has not been determined, but we speculate that the frequency of responding cells would have a larger effect on M.tb control. Additional studies are needed to determine why CBA/J mice generate low IFN-γ responses to Ag85 in the context of M.tb infection. The mechanisms are likely complex, involving antiinflammatory cytokines [], presentation by the H-2k haplotype [[6, 7]], or diminished T-cell co-stimulation in vivo, as IFN-γ output by single cells is robust when triggered with anti-CD3/anti-CD28 []. Specific studies are needed to identify the Ag85 peptides that fit H-2k homo- and heter-dimeric MHC molecules, how those peptides influence the generation and function of Ag85-specific T cells, and whether the Ag85 peptides present during M.tb infection (primary or re-infection) are different than the Ag85 peptides generated by Ag85 vaccination.
Ag85-containing vaccines are currently in human clinical trials [[8, 29]]. Our experimental results in CBA/J mice may predict that Ag85-containing vaccines will protect M.tb susceptible humans. Important limitations to be addressed in future mouse experiments include determining preexposure responses to exogenous Ag85, and long-term experiments to assess whether Ag85 vaccination improves TB disease outcome in CBA/J mice.
Overall, our results demonstrate four main findings. First, CBA/J mice did not respond efficiently to Ag85 in the context of M.tb primary infection or re-infection, but could produce abundant IFN-γ in response to positive controls. This suggests that CBA/J mice may recognize uncharacterized M.tb antigens, or that immunogenic epitopes of Ag85 are not available during in vivo infection. Further studies may identify other immunodominant M.tb antigens in CBA/J mice and possibly develop those antigens as vaccine candidates for M.tb susceptible humans. Second, CBA/J mice could generate robust protective immune responses to exogenous Ag85. This indicates that vaccination can be successful despite inefficient Ag85-specific IFN-γ responses during natural (primary or re-infection) M.tb infection. Third, in our memory model, M.tb-induced protection was not associated with enhanced Ag85 responses in either M.tb-resistant mice (C57BL/6) or M.tb susceptible mice (CBA/J). This surprising finding may be consistent with memory immune responses that precede M.tb control. Alternatively, Ag85 exposure during natural M.tb infection may not generate robust long-term protective Ag85-specific memory immune cells. Finally, Ag85 vaccination induced immune cell infiltrates that were larger than cell infiltrates in M. bovis BCG-vaccinated CBA/J mice. Whether this has a negative impact on pulmonary function or disease progression is unknown and should be addressed in future studies.
Materials and methods
Specific pathogen free 8-week-old female C57BL/6 and CBA/J mice (Charles River Laboratories, Wilmington, MA, USA) were maintained in ventilated cages within Biosafety Level-3 facilities and provided with sterile food and water ad libitum. All protocols were approved by The Ohio State University's Institutional Laboratory Animal Care and Use Committee.
M.tb Erdman (ATCC no. 35801) and M. bovis BCG (ATCC no. 35734) were obtained from American Type Culture Collection (Manassas, VA, USA). Stocks were grown in Proskauer–Beck liquid medium containing 0.05% Tween 80 to mid-log phase and frozen in 1 mL aliquots at −80°C. One aliquot was used for each experimental M.tb infection.
Mice received 50–100 CFU of M.tb Erdman using an Inhalation Exposure System (Glas-col, Terre Haute, IN, USA). M.tb dose and lung burden were determined as previously described []. Mice in long-term experiments (Fig. 1) were monitored, euthanized, and excluded from experiments if TB-associated morbidity developed [].
Vaccination protocols were modified from Deitrich et al. [] and Andersen []. A total of 250 μg/mL of synthetic monophosphoryl lipid A (MPL) (Avanti Polar Lipids, Inc., Alabaster, AL, USA) in 0.2% triethylamine in PBS was heated at 70°C and then sonicated for 30 s. A total of 2.5 mg/mL of dimethyldioctadecyl ammonium bromide (DDA) (Sigma-Aldrich, Co., St. Louis, MO, USA) in PBS was heated at 80°C for 10 min and then cooled. M.tb antigens at 100 μg/mL (Ag85) or equivalent volumes of PBS were added to the cooled DDA solution. Finally, equivalent volumes of the MPL solution and DDA solutions (with or without antigen) were mixed. Vaccines were agitated immediately prior to injection. Ag85 vaccinated mice were injected subcutaneously interscapularly three times at 2-week intervals with 200 μL (containing 25 μg MPL, 250 μg DDA, and 10 μg of antigen). Control groups received 200 μL of PBS or 200 μL of Adj (MPL + DDA). In some experiments, mice received one injection of 2.5 × 105M. bovis BCG bacilli in 200 μL of PBS. Mice were rested and challenged with M.tb Erdman by aerosol at least 10 weeks after the first injection.
Mice received 250 mg/L of isoniazid (isonicotinic acid hydrazine; INH, Sigma) and 100 mg/L of rifampin (rifampicin; RIF, Sigma) in the drinking water for 12 weeks. The protocol was modified [[36-39]] and based on consumption of 6 mL of water per mouse per day []. The targeted doses were 25–50 mg/kg/day for INH and 10–20 mg/kg/day for RIF, which are below the carcinogenic [] or toxic [] levels. INH was dissolved in a small volume of water at 50 mg/mL. RIF was dissolved in a minimal volume of DMSO (Sigma) at 100 mg/mL. MQ water was added to achieve the final concentrations and the pH was adjusted to neutral.
Lung cell isolation
Cells were isolated from the lungs []. Briefly, lungs were perfused with heparinized PBS, diced into 1 mm2 pieces, digested with collagenase/DNase, and single-cell suspensions created by pressing the pieces through cell strainers. Red blood cells were lysed. The leukocytes were counted and re-suspended in media, or fixed with 0.1% sodium azide.
Heparinized blood was collected and diluted in media as previously described [].
Bone marrow-derived APCs
Bone marrow cells from the tibiae and femora of age- and sex-matched noninfected C57BL/6 and CBA/J mice were differentiated into APCs [] using GM-CSF-containing media derived from GM-EL4 cells []. The only modification was that 5 μg/mL of M.tb Ag85 (Colorado State University's NIH NIAID Contract No. HHSN266200400091C) was added prior to the culture with lymphocytes.
Lung and blood antigen-specific responses by ELISA
One hundred microliters of lung cells containing 2 × 105 cells or 100 μL of diluted blood from each individual mouse were cultured in duplicate or triplicate as previously described []. The only modification was that M.tb Ag85 (5 μg/mL) was added to the cultures.
Lung and PBMC antigen-specific responses by ELISPOT
Purified CD4+ T cells from the lungs of mice or PBMCs were cultured with antigen-pulsed bone marrow-derived APCs as previously described []. The only modification was that APCs were pulsed with 5 μg/mL of Ag85.
Conjugated and unconjugated antibodies and isotype controls were purchased from BD Biosciences, including Fc Block™ (clone 2.462), PerCP-Cy5.5 anti-CD3ε (145-2C11), allophycocyanin-Cy7 anti-CD4 (GK1.5), PE-Cy7 anti-CD8 (53–6.7), allophycocyanin anti-TCR-β chain (H57597), PerCP anti-CD8 (53–6.7), FITC anti-CD44 (IM7), FITC rat IgG2bκ, FITC anti-I-A/I-E (2G9), and FITC rat IgG2aκ. The staining protocol and results analyses were performed as previously described [].
Histology and histomorphometry
Five-micrometer sections from formalin-fixed, paraffin-embedded sections were stained with hematoxylin and eosin or for acid-fast bacilli. Blind evaluation was performed by a board-certified veterinary pathologist (PS). Histomorphometry was performed using Aperio ImageScope v8.2.5 by a second board-certified veterinary anatomic pathologist (GB) [].
Statistical analyses were performed using Prism 4 software (GraphPad Software, San Diego, CA, USA). Multigroup comparisons were analyzed with one-way ANOVA using Dunnett's posttest or Tukey's posttest. Pairwise comparisons used the Student's t-test. Statistical significance was defined as *p < 0.05, **p < 0.01, ***p < 0.001.
Support was provided by NIH R01 (AI064522; J.T.), NIH T32 (RR07073; G.L.B.) and NIH K08 (AI071111; G.L.B.). M.tb Ag85 and anti-Ag85 antibodies were provided by Colorado State University's TBVTRM Contract HH5N266200400091c NIH N01AI40091. Histology services were provided by The Ohio State University's Department of Veterinary Biosciences.
Conflict of interest: The authors declare no commercial or financial conflicts of interest.