INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

In developing countries, tuberculosis has become the most common opportunistic infection associated with human immunodeficiency virus infection. With the subsequent emergence of drug-resistant strains of Mycobacterium tuberculosis, there has been a strong impetus to develop an improved vaccine against this disease, considered the leading cause of mortality from a single infectious agent in adults (24). Recently, interleukin-12 (IL-12) has been implemented as a coadjuvant with experimental vaccines to accelerate the immune response against a variety of viruses and intracellular pathogens, including M. tuberculosis (1, 16, 19). Understanding how IL-12 influences the immune response to mycobacterial infection should facilitate the search for more-effective vaccines.

The development of a T-helper cell type 1 (Th1) response to M. tuberculosis is augmented by IL-12, one of the main cytokines produced by macrophages infected with M. tuberculosis (8, 20). IL-12 is known to induce gamma interferon (IFN-), which serves as the major activator of macrophages and as a promoter of Th1 cell development, thus contributing to the host's control and containment of the mycobacteria. The importance of IFN- for host resistance to mycobacteria is supported by studies of IFN--deficient mice, produced by targeted gene disruption. Mice lacking IFN- develop disseminated infection when inoculated with a high titer of M. tuberculosis (6, 11). Recently, IL-12 production was shown to be required for an effective IFN- immune response to M. tuberculosis with IL-12-knockout mice (7). Splenocytes from IL-12-knockout mice were incapable of producing IFN- in response to M. tuberculosis antigen challenge in vitro. These mice demonstrated unrestricted growth of bacteria in all organs tested, including lungs, liver, and spleen, thereby reinforcing the importance of IL-12 in controlling mycobacterial growth.

IL-12 has also been implicated in the response to the vaccine strain of Mycobacterium bovis, bacillus Calmette-Guérin (BCG). IL-12 gene expression is upregulated in human monocyte-derived macrophages when stimulated with BCG in vitro (21). Production of IL-12 in vitro by murine bone marrow-derived macrophages infected with BCG is highly dependent on other cytokines such as IFN- and tumor necrosis factor alpha (TNF-) (10). However, the importance of IL-12 in the in vivo response to BCG has not yet been investigated.

The macrophage's primary response to intracellular pathogens includes the production of cytokines-chemokines that influence the microenvironment of the lymph nodes which drain the infected tissues. Antigen-specific T cells migrating to the lymph nodes are influenced by the mediators present. In this way, the macrophage impacts on the development of the adaptive T-helper cell response to a given pathogen and serves as a bridge between the innate and adaptive immune systems. The innate response to several mycobacterial species including M. bovis BCG, Mycobacterium lepraemurium, and Mycobacterium intracellulare (12, 14, 25) in mice has been shown to be controlled by a single dominant gene on chromosome 1 which is present in two allelic forms in inbred strains of mice, Bcg^r (resistant) and Bcg^s (susceptible) (14). The gene, now renamed Nramp1, encodes an integral phagosomal membrane protein expressed in macrophages (13, 32). An increase in mycobacterial susceptibility is correlated with a single nucleotide substitution in the Nramp1 gene in the susceptible strains of mice (31). Using gene targeting to create Nramp1 null mice, Vidal et al. (30) have demonstrated that the deletion of this gene made the mice susceptible to M. bovis BCG infection. Furthermore, Nramp1 null mice are also susceptible to Leishmania donovani and Salmonella typhimurium infection during the early phases of the immune response. Thus, the Nramp1 gene, expressed in macrophages, was formally proven to be located at the Bcg/Lsh/Ity locus.

Since IL-12 is one of the first cytokines produced by macrophages during infection (21), and macrophages are the key mediators of innate resistance, it is of interest to further delineate the role of IL-12 in the innate phase of immunity to mycobacteria. Although the study of the IL-12-knockout mice by Cooper et al. (7) clearly demonstrates the pivotal role of IL-12 in helping the host mount an effective T-cell response to mycobacteria, it has not been documented to what extent IL-12 contributes to the innate resistance to mycobacterial infection observed in strains of mice which carry the resistant (Bcg^r) or susceptible (Bcg^s) allele of the Nramp1 gene. One limitation of using knockout mice is that cytokine ablation may affect the ontogeny of the immune system. Furthermore, cytokine gene-disrupted mice frequently have a heterogeneic genetic background as a result of the use of 129/Sv (Bcg^r) embryonic stem cells and C57BL/6 (Bcg^s) blastocysts during knockout construction, resulting in the mixed genotypes of the knockout mice (including the Bcg locus). We have chosen to use blocking antibodies to deplete cytokine activity in mice with a clearly defined genetic background at the Bcg locus to avoid several of the problems inherent in the knockout experiments. The aim of this study is to investigate the role of IL-12 in animals differing in their innate immunity to infection with the Montreal strain of M. bovis BCG. Using two congenic strains of mice differing at the Bcg locus for susceptibility to BCG infection, B10.A (Bcg^r) (resistant) and B10.A (Bcg^s) (susceptible), we demonstrate that the response of the innately resistant Bcg^r animals is not affected by IL-12 ablation. However, the susceptible Bcg^s animals respond to blockage of IL-12 with a significant increase in recoverable bacteria after 3 weeks of infection. These data reinforce the concept that adaptive immunity but not innate immunity to BCG is a T-cell-dependent phenomenon as shown by Gros et al. with athymic nude mice (15).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals. Female and male B10.A (Bcg^s) mice were purchased from the National Cancer Institute (Frederick, Md.). The B10.A (Bcg^r) mice were bred in the Montreal General Hospital Research Institute animal facility under specific-pathogen-free conditions. Mice were maintained in isolators and provided sterile water and food ad libitum. All mice were 6 to 15 weeks of age when the experiments were initiated. All experiments with mice were carried out by protocols reviewed and approved by the McGill University Animal Care Committee.

Bacterial cultures and infections. M. bovis BCG (Montreal strain) was cultured in Dubos' albumin liquid medium (Difco, Detroit, Mich.) for 14 days with one passage at day 7. Growing cultures were filtered at the end of 14 days through a 5-µm-pore-size filter and stored at 4°C for up to 4 days. Mice were infected with 0.5 × 10⁵ to 1 × 10⁵ bacteria via lateral tail vein injection. At various time points, spleens were removed and homogenized in phosphate-buffered saline with a mortar and pestle. The number of viable bacteria per organ was assessed by plating serial 10-fold dilutions of homogenate in Dubos agar culture supplemented with 10% OADC enrichment (Difco). Colonies were counted after 21 days of culture at 37°C. The results are presented as means of three animals ± standard errors. Statistical analysis was performed by Student's t test. Results were considered significantly different at P < 0.05.

Antibodies. The purified rat anti-mouse IL-12 p40 monoclonal antibody 10F6 was generously provided by David Presky of Hoffmann-La Roche (Nutley, N.J.). The rat antibody to murine IFN-, XMG1.2, was prepared from ascites fluid and purified by ammonium sulfate precipitation (4). The purified isotype control rat monoclonal antibody (rat immunoglobulin G, product no. I-4131) was purchased from Sigma (St. Louis, Mo.). Sterile solutions of antibodies (0.5 mg per mouse) were injected intraperitoneally 1 day prior to infection and twice weekly for the duration of the experiment. We have consistently seen no significant difference in bacterial growth between animals infected with BCG and treated with the isotype control antibody and infected animals receiving no antibody treatment. To test the efficacy of the 10F6 anti-IL-12 antibody, we treated animals with an equivalent amount of antibody in a lipopolysaccharide (LPS)-BCG challenge experiment as described by others (33). Briefly, BALB/c mice were injected with BCG and challenged intravenously (i.v.) with 1 mg of LPS on day 14 to induce IFN- production. The 10F6 anti-IL-12 antibody was administered intraperitoneally 12 and 1 h prior to LPS challenge. Animals were sacrificed 5 h after LPS injection, and serum was collected. The efficacy of anti-IL-12 treatment was confirmed by the finding that IFN- production in the serum decreased almost 10-fold from 208 ± 102 to 29 ± 38 U/ml in animals treated with the 10F6 antibody (P < 0.05).

Isolation, purification, and analysis of total RNA. Spleens and lungs were removed from animals under aseptic conditions and rapidly frozen with liquid nitrogen. RNA was isolated by homogenization in guanidinium isothiocyanate and by cesium chloride centrifugation (5). RNA pellets were suspended in RNase-free water, and aliquots were stored at 80°C until use. The RNase protection assay was performed according to the manufacturer's instructions with the RiboQuant kit (Pharmingen, San Diego, Calif.). Results are expressed as relative absorbance obtained by scanning autoradiograms on a U.S. Biochemical SciScan 5000 scanner with internal standard absorbance of glyceraldehyde-3-phosphate dehydrogenase for each RNA sample.

Histological analysis. Lungs were inflated with 10% buffered formalin, embedded in paraffin, and cut as 5-mm sections. A standard hematoxylin-and-eosin staining was used to identify granulomas. Acid-fast mycobacteria were identified by Ziehl-Neelsen stain.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Antibodies to IL-12 increase bacterial growth in B10.A (Bcg^s) mice but not B10.A (Bcg^r) mice. Since IL-12 induces a Th1-type T-cell response to intracellular pathogens (27), we examined whether abrogation of IL-12 activity by intraperitoneal administration of purified anti-IL-12 during infection with BCG could alter bacterial growth in congenic B10.A mice differing at the Bcg locus. Twenty-four hours prior to inoculation with 10⁵ viable bacteria of the Montreal strain of BCG, Bcg^r and Bcg^s mice were treated with 0.5 mg of blocking antibodies to IL-12 per mouse. Antibody was administered twice weekly until the termination of the experiment. At weekly intervals, the growth of bacteria in the spleens of infected Bcg^r and Bcg^s mice was assessed by agar assay. As shown in Fig. 1, the growth of bacilli in Bcg^s mice treated with antibodies to IL-12 was not significantly different from that in the Bcg^s mice treated with isotype control antibody at day 7 and day 14. However, by day 21 the isotype control-treated Bcg^s mice were in the process of resolving the infection while treatment with anti-IL-12 abrogated this response, resulting in a two- to threefold (P = 0.039) increase in bacterial load. As expected, the bacterial counts in the resistant (Bcg^r) mice, remained below or near the inoculum titer regardless of treatment with anti-IL-12 throughout the time course of the experiment.

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Growth of BCG in the spleens of Bcgr and Bcgs mice inoculated i.v. with 105 bacteria. Mice were treated with antibodies to IL-12 (stippled bars) or an isotype control (striped bars) prior to and during the course of infection. The Bcgs mice showed a significant increase in bacterial load following treatment with anti-IL-12 at the 21-day time point. These results are the means and standard deviations for three mice per time point and are representative of two independent experiments. The double asterisks indicate a significant difference between the anti-IL-12- and control-treated mice (P < 0.05).

Effects of long-term treatment with anti-IL-12. It is known that susceptible mice can eventually resolve a BCG infection after 6 weeks (14). To determine if long-term depletion of IL-12 alters the response of susceptible mice to BCG, we carried the experiment out to 12 weeks after infection with BCG. Depletion of IL-12 in Bcg^s mice did not have a significant effect on recoverable bacteria at week 1 and week 2 compared to mice treated with control antibody (Fig. 2). However, at week 3 of infection Bcg^s mice showed a threefold increase in viable bacteria compared to isotype control-treated mice (P < 0.05). This was followed by a drop in bacterial growth in both groups in subsequent weeks. The anti-IL-12-treated group cleared infection more slowly, with four-, five-, and eightfold increases in viable bacteria compared to controls at weeks 4, 6, and 7, respectively (Fig. 2).

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Effect of anti-IL-12 treatment for 12 weeks. Bcgs mice were inoculated with 105 BCG bacteria and treated with either anti-IL-12 (open squares) or an isotype control (closed squares) for as long as 12 weeks. Data represent the means and standard deviations of BCG colonies per spleen from three mice per time point. These results are representative of three independent experiments up to week 6 and one experiment to week 12. Asterisks represent a significant difference with a P value of <0.05 between mice treated with isotype control antibody and mice treated with anti-IL-12 at all time points after 3 weeks.

View larger version (174K): [in this window] [in a new window]   FIG. 3.   BCG-infected Bcgs mice have defective lung granuloma formation following treatment with anti-IL-12. Lung tissues from BCG-infected mice treated with an isotype control antibody (a to c) or anti-IL-12 (d to f) were fixed in formalin and stained with Ziehl-Neelsen stain to detect bacteria. Shown is one granuloma from each treatment group (isotype control [a] and anti-IL-12 [d], hematoxylin-eosin) with higher magnifications of the same granuloma shown in panels b and e (hematoxylin-eosin) and panels c and f (Ziehl-Neelsen stain). The arrows in panel f point to acid-stained bacteria. The regions of cell infiltration shown were typical of three different mice receiving the same treatment 6 weeks following BCG infection. Bars, 100 µm (d), 50 µm (e), and 10 µm (f).

Blocking IFN- leads to a significant increase in bacterial replication in both Bcg^r and Bcg^s strains of mice. The lack of effect of anti-IL-12 treatment on Bcg^r mice indicates that the Bcg^r mice do not require IL-12-dependent IFN- to resist BCG infection during the early response to the mycobacterium. Although data from IFN--knockout mice support the hypothesis that macrophages cannot be mobilized against M. tuberculosis infection in the absence of IFN- (6, 11), we wanted to determine if IFN- was needed by resistant Bcg^r mice for an optimal immune response to BCG. Our results, shown in Fig. 4, demonstrate that treatment with anti-IFN- results in a marked increase in bacterial growth in both animal strains, with a three- to fourfold increase in bacterial growth in the first 2 weeks of infection and a six- to sevenfold increase in anti-IFN--treated mice after 3 weeks of infection. Even when IFN- was blocked, the Bcg^r mice controlled the infection better than did the Bcg^s mice, further demonstrating the importance of the Nramp1 gene in controlling infection even when IFN- is compromised. Nevertheless, these data underline the importance of IFN- in both the innate response of Bcg^r mice to BCG and the adaptive immunity of the Bcg^s mice.

View larger version (25K): [in this window] [in a new window]   FIG. 4.   IFN- is critical for control of bacterial growth in both Bcgs and Bcgr mice. Mice were inoculated with 105 BCG bacteria i.v. and injected intraperitoneally with anti-IFN- twice weekly. The mean numbers of bacteria recovered per spleen from three mice per group are shown with the standard deviations. Both groups of mice demonstrated a significant difference (P < 0.05) when anti-IFN- was administered.

Induction of cytokine gene expression in resistant and susceptible mice during BCG infection. A number of cytokines are associated with inflammation and the immune response to infection. IFN- is associated with the T-helper cell response referred to as the Th1 response, whereas IL-4 and IL-10 correlate with and promote the Th2 immune response (22). IL-6 is associated with inflammation (26), and IL-15 induces the proliferation of a class of T cells associated with primary responses to infection (3). To examine whether the neutralization of IL-12 could influence the induction of any of these cytokines in the Bcg^r or Bcg^s mice, we analyzed cytokine expression at the RNA level by RNase protection analysis. We isolated total RNA from spleens taken directly from mice at the indicated time points in an effort to obtain a nonbiased picture of gene expression actively occurring in vivo. We chose to use this RNase protection method since it provides a highly sensitive, specific, and quantitative way of detecting multiple cytokine mRNAs. As can be seen in Fig. 5, there was a profound difference in the expression of relevant cytokine mRNAs between the two strains of mice, with the most significant differences occurring during the early and late stages of infection with BCG. Interestingly, we have found IFN-, IL-10, IL-6, and IL-15 mRNA levels to be significantly higher in the spleens of Bcg^r mice at day 0 and day 1 of infection. However, during the second phase of the infection mRNA levels for these cytokines drop in these mice. IL-4 mRNA levels remained below our detection limits under these experimental conditions. The early induction of IFN- in the resistant strain following infection is consistent with our previous reports utilizing semiquantitative PCR techniques (18). At day 1 of infection in the Bcg^r mice, a clear decrease in IFN- mRNA could be observed in animals treated with anti-IL-12. No significant difference in IFN- mRNA levels could be seen at later time points. In contrast, the IFN- mRNA expression in the Bcg^s mice was low early in the infection but was elevated during the second phase of infection, as bacterial load increased at 3 weeks. No significant difference could be seen on day 21 of infection between IFN- mRNA levels in Bcg^s mice treated with antibody to IL-12 and levels in mice treated with an isotype control (Fig. 5A), although mice treated with anti-IL-12 did show significantly higher bacterial growth after 3 weeks (21 days) of infection, as illustrated in Fig. 1. We attempted to measure IFN- in the serum of these animals by enzyme-linked immunosorbent assay. However, due to the low dose of BCG used for infection we were unable to detect significant amounts of IFN- (data not shown). Using semiquantitative PCR, we have been unable to detect any significant change in IL-12 (p40) mRNA production in either mouse strain either with or without anti-IL-12 treatment (data not shown). Only the Bcg^s mice showed any significant increase in IL-12 mRNA after 21 days of BCG infection, as would be predicted for the peak of infection (data not shown). Overall, the treatment with anti-IL-12 did not change the differential pattern of most cytokine expression observed in Bcg^r and Bcg^s mice infected with M. bovis BCG.

View larger version (28K): [in this window] [in a new window]   FIG. 5.   There is a profound difference in cytokine gene expression between Bcgr and Bcgs mice. RNA was extracted from frozen spleens taken from mice at the time points indicated. RNase protection assay analysis for IFN- (A), IL-10 (B), IL-6 (C), and IL-15 (D) was performed with a RiboQuant kit. Bcgr mice treated with isotype control antibody () or anti-IL-12 antibody () were compared with Bcgs mice treated with isotype control antibody () or anti-IL-12 antibody (). The data points represent the averages of duplicate mice on days 0, 1, and 3 and single mice on days 7, 14, and 21. Results are expressed as relative absorbance obtained by scanning autoradiograms with internal standard absorbance of glyceraldehyde-3-phosphate dehydrogenase for each individual RNA sample. KIOD, optical density/1,000.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Macrophages, primary effectors in the host's innate immune system, control the development of the T-cell immune response through cytokine production as well as through antigen presentation to T cells. Thus, the macrophage exists as a common link between the innate and adaptive immune responses to pathogens. We speculate that strain differences observed in the host's innate immune response to various pathogens may in part reflect a variance in the efficacy of macrophage stimulation (i.e., cytokine production) and interaction with T cells (i.e., antigen presentation). IL-12 is one of the major cytokines produced by macrophages that can direct the development of a Th1 T-cell response against intracellular pathogens (29). Factors which may affect macrophages' innate response to a given pathogen may modify the amount or kinetics of IL-12 production, ultimately impacting on the T-cell response to a pathogen. For example, complete sterilization immunity to Listeria monocytogenes requires a combination of both an effective innate response and optimal IL-12 production by macrophages leading to a pathogen-specific Th1 response (17).

Our results indicate that innate resistance to BCG infection exhibited by Bcg^r mice is not dependent on optimal IL-12 activity. The macrophages in mice homozygous for the Bcg^r allele can control infection with BCG when IL-12 is neutralized. This is consistent with the observations for the Nramp1-deficient mice, in which early successful antimycobacterial response to BCG is mostly dependent on Nramp1 gene function and less dependent on T-cell-mediated factors (30). Nevertheless, mice with the Bcg^s allele displayed a dependence on optimal IL-12 during the late immune phase of BCG infection. IL-12 seems to be important to control bacterial growth in the spleens between the second and third week postinfection at the time when the T-cell response to the pathogen is starting. We can conclude this for only the latter time points since we were unable to accurately determine the number of bacteria during the first week of infection. The anti-IL-12 treatment seems to slow down the clearance of the bacteria in the Bcg^s mice. Although the bacterial load in Bcg^s mice decreases with time even in the presence of blocking antibodies to IL-12, at every time point tested there were still four- to eightfold more bacteria in anti-IL-12-treated mice than in control-treated mice. Therefore, IL-12 seems to be most important during the peak of the T-cell immune response in the mice with the allele for susceptibility to BCG. Since macrophages in Bcg^s strains of mice do not express mature Nramp1 protein (32), one could speculate that without Nramp1 these mice are totally dependent on cytokines such as IL-12 and IFN- to develop a strong Th1 response required to effectively overcome BCG infection. Expression of the Nramp1 gene does not seem to be necessary for the development of this Th1 response as long as optimal levels of IL-12 are present.

The data of Cooper et al. (7) clearly suggest that IL-12 is required for resistance to M. tuberculosis. Indeed, IL-12 is known to enhance the production of some cytokines such as IFN- and together with IFN- to induce a Th1 phenotype in some experimental models (28). The relative level of IFN- mRNA in the Bcg^r mice is already high prior to and early during BCG infection, indicating an environment conducive for Th1 development in these BCG-resistant mice. The early induction of IFN- in the Bcg^r mice is not affected by anti-IL-12 treatment. It is possible that other factors such as IFN- inducing factor may be able to induce IFN- production in these animals independent of IL-12 (23). During BCG infection, optimal IFN- production in the Bcg^r mice seems to be important to control bacterial growth since we demonstrated that in vivo treatment with antibodies to IFN- effectively increased BCG bacterial load in the Bcg^r mice. It seems that the immediate induction of IFN- may be critical for innate resistance but clearly does not depend only on optimal IL-12 levels to be effective. Indeed, the data of Flesch et al. (10), with a model of in vivo sensitization followed by in vitro stimulation with BCG, suggest that IL-12 production is dependent on IFN- and TNF- and that the initial production of IFN- may be independent of IL-12. Our data indicate that the innate resistance of the Bcg^r mice does not require an enduring IL-12-dependent Th1 response for effective clearance of bacteria even during the later stages of the infection, although IFN- is needed for optimal innate immunity. Together, these two studies suggest that IFN- and TNF- may be more important for successful clearance of BCG than is IL-12. Our findings are in contrast to the innate resistance of the severe combined immunodeficient (SCID) mouse to Lyme disease, which is dependent on IL-12 even in the absence of a T-cell response in these mice (2). The mechanism by which the innate response of the Bcg^r animal can handle the mycobacterial infection without IL-12 and the nature of the cells (either CD8⁺ or NK cells) that are producing IFN- in these mice during early infection and their relationship to the expression of the resistant allele of the Nramp1 gene have not been elucidated yet.

Other genes that may affect the immune response to mycobacteria could code for factors or cytokines that might replace or affect the activity of the two key cytokines studied here, IFN- and IL-12. These factors and others that affect the host's T-cell response may heavily influence whether the host develops an appropriate primary and effective secondary response to mycobacteria. A delicate balance may exist between the response of the macrophage during its primary encounter with the mycobacterium and its ability to promote an effective T-cell response during virulent infection. An animal that can mount an effective innate response to a pathogen may be unable to eradicate a virulent form of the pathogen if unable to activate the T-cell adaptive immune system through appropriate cytokine expression and Th1 induction. IL-12 may be effective in boosting a Th1-cell response to a nonvirulent vaccine form of a pathogen such as M. bovis BCG. In an individual with an effective innate macrophage response to the BCG vaccine, IL-12 may be needed to boost the T-cell arm of the immune response and effectively acquire memory cells.

A more complete understanding of the differences between the activation and response of macrophages with Bcg^r or Bcg^s allele expression may guide the design of effective vaccines against mycobacteria with the goal of developing appropriate T-helper cell populations and stable T-cell memory responses to mycobacteria. IL-12 is being considered for use as a coadjuvant for vaccines (1, 16, 19). Deciphering how IL-12 modulates a host's initial response to vaccination and/or infection may be useful in developing new vaccine strategies.