Role of Nitric Oxide in Defense of the Central Nervous System against Mycobacterium tuberculosis

Abstract

Murine models of tuberculous meningitis (TBM) have not reflected the severity of disease in humans. Based on reports that activated murine microglial cells, but not human microglial cells, express inducible nitric oxide synthase (iNOS), the objective of this study was to determine whether iNOS-knockout (iNOS^−/−) mice would provide such a model. iNOS^−/− mice infected with M. tuberculosis developed serious clinical manifestations and granulomatous lesions containing tubercle bacilli throughout the meninges, all of which were absent in wild-type mice. This study underscores the importance of nitric oxide in defense against TBM and suggests that iNOS^−/− mice are an appropriate model for human TBM.

Central nervous system (CNS) disease is an ominous complication of Mycobacterium tuberculosis infection, which is almost always fatal when not treated [1]. Although the neuropathogenesis of M. tuberculosis is incompletely understood, autopsy findings and results from animal models carried out in the early 20th century demonstrated that, unlike other forms of bacterial meningitis, tuberculous meningitis (TBM) begins in the brain parenchyma, with the rupture of a tuberculoma into the subarachnoid space [2]. Recent studies of human microglia—the resident macrophages of the brain parenchyma—and astrocytes in our laboratory suggested that microglia are the main cell type infected by tubercle bacilli, and that infection of these brain macrophages elicits the production of inflammatory cytokines and chemokines [3].

The role of microglia in defense of and injury to the brain following CNS infections has been a subject of increased research interest [4]. Although the importance of cytokine-mediated up-regulation of macrophage expression of inducible nitric oxide synthase (iNOS) in the killing of intracellular pathogens, such as M. tuberculosis, has been clearly demonstrated in rodent models, the contribution of this antimicrobial system in human macrophages has been controversial [5, 6]. Lee et al. [7] were the first to report that in contrast to astrocytes, iNOS expression was undetectable in human microglia following in vitro stimulation by interleukin (IL)–1 or interferon (IFN)–γ. This finding was corroborated in studies in our laboratory that showed robust production of the reactive nitrogen intermediate (RNI) nitric oxide (NO) by activated murine microglia but not by human microglia [8]. Subsequently, we demonstrated that activation of human microglia with IFN-γ resulted in up-regulation of many genes involved in the inflammatory response, but iNOS mRNA was not among them [9].

In recent years, the development of a rabbit model that mirrors human CNS tuberculosis has proved valuable for studying the neuropathogenesis and treatment of M. tuberculosis [10]. However, the absence of a murine model that parallels the serious nature of the disease in humans has hampered progress in this field. In 2002, Mazzolla et al. [11] reported that intracerebral inoculation of BALB/c and DBA mice with M. bovis BCG Montreal resulted in a mononuclear cell infiltration of the brain, microglial cell activation, and an increase in the number of bacteria in the CNS through day 21 after infection. Similar histopathological findings were reported in 2007 by van Well et al. [12] after intracerebral inoculation of C57Bl/6 mice with M. tuberculosis H37Rv. In addition to recovering viable bacilli from brain homogenates, these investigators detected an up-regulation of IL-1 β, tumor necrosis factor (TNF)–α, IL-6, and IFN-γ in the CNS. However, evidence of clinical signs of disease and mortality rates were not reported for either of these murine models. This surprising absence of clinical disease was also observed in unpublished studies (M.R.O.) performed in our laboratory after intracerebral inoculation of FVB mice (n = 8) with M. tuberculosis H37Rv, which were then monitored for 60 days after infection. Although an inflammatory infiltrate was identified histopathologically [1], none of the mice showed clinical signs of disease or succumbed to infection.

Based on knowledge of the pivotal role that macrophages play in tuberculosis and of the reports regarding the species difference with respect to iNOS expression in microglia, the present study was designed to test the hypothesis that mice lacking the iNOS gene would develop CNS disease that would more closely resemble TBM in humans.

MATERIALS AND METHODS.

Mice.

C57BL/6N iNOS^−/− mice and wild-type C57BL/6N mice (Taconic) were housed in individual high-efficiency particulate air–filtered isolator units. Animal care and procedures were undertaken in accordance with the University of Minnesota Institutional Animal Care and Use Committee.

Bacterial challenge.

M. tuberculosis H37Rv (ATCC) was cultured for 2 weeks in standard Dubos broth medium supplemented with 10% oleic acid–albumin-dextrose complex (Becton Dickinson) [3]. Cultures were then filtered and resuspended in culture medium (Dulbecco’s modified Eagle medium, 10% fetal bovine serum, and 0.025% Tween 80), and the bacterial concentration was determined before the cultures were frozen at −80°C. Bacilli (1 × 10⁶ cfu suspended in 3 μL saline) were inoculated intracerebrally in anesthetized iNOS^−/− mice and wild-type mice 1 mm lateral and posterior to the bregma, at a depth of 2 mm with a 27-gauge disposable needle. iNOS^−/− control mice received 3 μL saline delivered in the same manner. The intracerebral location of this route of administration was verified by using Evan’s blue, followed by gross and histological examination of the CNS.

Clinical signs.

Wild-type mice and iNOS^−/− mice were monitored daily for neurological signs (limb weakness, listlessness, nucchal rigidity, and/or uncoordinated movement around the cage) and for survival for 60 days after infection.

Pathology studies.

Mice were deeply anesthetized with ketamine and xylazine and euthanized by cervical dislocation. A standard necropsy was performed. The brains were removed from the animals and placed immediately in formalin for a minimum of 72 h for fixation. Fixed brains were analyzed for gross pathology. Tissue, including the brain and parenchyma, were paraffin-embedded, sectioned, and stained with hematoxylin and eosin. For immunocytochemical analyses, macrophage and lymphocyte infiltration of brain tissue from infected and control mice was evaluated after staining with anti–MAC-2 (activated-macrophage specific) (Cederlane) and anti-CD3 (lymphocyte specific) (Serotec) antibodies. In brief, paraffin-embedded tissues were deparaffinized and steamed in Tris-EDTA for antigen retrieval. Tissues were subsequently incubated in 3% H₂O₂ to block endogenous peroxidase, protein blocked with undiluted rodent block (Biocare), and stained with either anti-CD3 (1: 15,000 dilution) or anti–MAC-2 (1:10,000 dilution) antibodies. For the detection of primary antibodies, tissues were incubated with undiluted rat probe and polymer and then chromogen substrate was used for detection (Dako). All tissues were counterstained with Mayer’s Hematoxylin (Dako).

Data analysis.

Survival after intracerebral inoculation with M.tuberculosis H37Rv or saline was analyzed by the log rank test, and P < .05 was considered significant.

RESULTS

Clinical manifestations.

Wild-type mice (n = 8) and iNOS^−/− mice(n = 8) were monitored daily for 60 days after infection for clinical signs indicative of CNS tuberculosis. Five (63%) of the 8 iNOS^−/− mice demonstrated signs of nucchal rigidity, listlessness, or uncoordinated movement around the cage by 35–58 days after infection. All 5 of these mice died or were sacrificed in a moribund state (figure 1). None of the wild-type mice or control iNOS^−/− mice that received saline (n = 8) mice displayed any evidence of clinical disease during the 60-day observation period.

Figure 1.

Survival rate of mice after challenge with Mycobacterium tuberculosis H37Rv. Wild-type and inducible nitric oxide synthase knockout (iNOS^−/−) mice were inoculated intracerebrally with 10⁶ cfu M. tuberculosis H37Rv; iNOS^−/− control mice received saline intracerebrally (there were 8 mice in each group). **P < .01, for comparison with other groups.

Pathological findings.

The brains of iNOS^−/− mice and wild-type mice inoculated with M. tuberculosis were examined for gross and histopathological lesions. In 3 iNOS^−/− mice, a reddish opacity of the meninges was present over the major brain depressions, which was especially evident over the transversal encephalic fissure, sulci, and fissures of the cerebellum, brain pedunculus, and pons. Histologically, multifocal coalescing epithelioid macrophage aggregations surrounded by more sparse lymphocyte infiltration (noncaseating granulomas) and a few neutrophils were found disseminated throughout the brain and meninges in all of the iNOS^−/− mice challenged with M. tuberculosis (figure 2A and 2B). The brain lesions were characterized by an extensive granulomatous chronic meningitis, ventriculitis, and plexus choroiditis, commonly associated with small vessels. Focal granulomas were also found in the thalamus and mesencephalon. Caseating tuberclomas were uncommon (they occurred in only 1 of the mice examined), and Langhan’s multinucleated giant cells were absent in all animals. Acid-fast bacilli were readily apparent in macrophages in affected areas of the brain (figure 2C).

Figure 2.

Histopathological and immunological findings in the brains of inducible nitric oxide synthase knockout mice inoculated intracerebrally with Mycobacterium tuberculosis H37Rv. A, Brain section at (–) 3.80 mm caudal to the bregma showing meninges expanded by a severe granulomatous inflammation. B, Magnification of noncaseating granulomatous meningitis (from section shown in rectangle in A) showing aggregation of epithelioid macrophages surrounded by a peripheral infiltration of lymphocytes. C, Acid-fast stained section demonstrating numerous tubercle bacilli within the granuloma. D, Immunostaining with MAC-2 (section sequence of A) demonstrating a predominant histiocytic infiltration expanding the meninges. E, Immunostaining with anti–MAC-2 showing activated macrophages within a granuloma. F, Immunostaining with anti-CD3 (section sequence of E) showing lymphocytes at the periphery of a granuloma.

In addition to these profound histopathological findings, the iNOS^−/− mice had evidence of macrophage and T lymphocyte infiltration in the brain. Immunocytochemical analysis for macrophages (anti–MAC-2) (figure 2D and 2E) and lymphocytes (anti-CD3) (figure 2F) revealed activated macrophages and lymphocyte infiltration throughout the meninges and in multiple localized areas in the brain parenchyma. In addition to these neuropathologic findings, examination of the peripheral organs of the iNOS^−/− mice revealed granulomas in the liver and spleen that contained bacilli.

In sharp contrast to the marked gross, histopathological, and immunocytochemical abnormalities in the brains of the iNOS^−/− mice and the lesions in their peripheral organs, no pathological abnormalities were detected in the wild-type mice or in the brains of the iNOS^−/− control animals that received saline intracerebrally, all of which were sacrificed at day 60 after infection.

DISCUSSION

CNS tuberculosis is a life-threatening manifestation of M. tuberculosis infection [1]. The neuropathogenesis of CNS tuberculosis remains incompletely understood, in part due to the absence of a murine model that displays the characteristics of the human disease. While a rabbit model that mirrors human TBM has proved useful for studying the neuropathogenesis and treatment of CNS tuberculosis [10], the lack of immunological reagents and genetically modified animals limits the ability to define the neuroimmunological characteristics of TBM in rabbits. While the complementary immunological tools and genetically modified animals are present for murine models, previous murine models suggest that mice are resistant to the development of the spectrum of clinical and pathological manifestations of TB meningitis [11, 12; and M.R.O., unpublished findings].

The present study tested the hypothesis that iNOS^−/− mice that were infected intracerebrally with the virulent strain M. tuberculosis H37Rv would develop clinical manifestations that resembled TBM in humans. In sharp contrast to wild-type mice, iNOS^−/− mice demonstrated clinical manifestations, a mortality rate, and histopathological and immunopathologic abnormalities similar to those reported in humans with TBM [1, 2]. Although the clinical and pathological findings in this study are consistent with CNS tuberculosis in humans, the results of this study—in which bacilli were inoculated directly into the brain parenchyma—do not reflect the course of natural infection in humans and thus must be interpreted with caution.

In addition to serving as a murine model that mirrors the clinical and neuropathologic findings of human TBM, we believe the findings in this study of iNOS^−/− mice may help explain the conundrum arising from previous studies in which virulent tubercle bacilli were delivered intracerebrally, yet the mice appeared resistant to clinically manifest CNS infection, as was the case with the wild-type mice in the present study. Within the brain parenchyma, microglia are the first responders to invading pathogens [4], and in the case of murine but not human microglia, these brain macrophages up-regulate iNOS [9] with production of RNIs which have potent microbicidal activity. We postulate that up-regulation of iNOS in murine but not in human microglia underlies, at least in part, the relative resistance of mice and the susceptibility of humans to TBM. The role of iNOS in defense of the mouse brain has been shown with respect to other intracellular pathogens [13, 14], and it would be interesting to extend studies with iNOS^−/− mice to these and other neurotropic pathogens. In addition, on the basis of recent studies demonstrating that the antimicrobial peptide cathelicidin is responsible for the killing of M. tuberculosis in activated human macrophages [15], it would be of interest to determine whether this antimicrobial system is operative in activated microglia from iNOS^−/− mice.