INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Mycobacterium avium is the most common cause of disseminated bacterial infection among human immunodeficiency virus (HIV)-infected individuals in the United States and Europe (2, 8, 13). Infection typically occurs late in the course of the AIDS disease, when CD4⁺-T-cell counts are below 100/mm³ (7, 25). Successful treatment of M. avium infection is difficult, not only because of the suppressed state of the host's immune system but also because of a lack of effective antibiotics to treat the infection and the concurrent development of resistance to currently available drugs (17, 18, 22).

Infection of HIV-positive individuals with M. avium is thought to occur through the respiratory or gastrointestinal tract, with systemic spread being a common feature of this disease (7, 19). In the HIV-negative population M. avium generally presents as a pulmonary disease, and both systemic and pulmonary infection cause disabling disease in infected individuals (9, 15). M. avium infection typically leads to the generation of large, diffuse granulomatous lesions within infected tissue, such as the lung and lymph nodes. Characteristically, these lesions are filled with foamy macrophages, neutrophils, and a smaller percentage of lymphocytes. Tissue necrosis and fibrosis are common features of pulmonary M. avium infections (11, 20).

Recent studies in mice have suggested that the inflammatory response generated after infection with several intracellular pathogens, including Mycobacterium tuberculosis, is controlled in part by T cells (10, 27, 29). While they constitute only a minor T-cell population (1 to 5%) within lymphoid organs, T cells are a predominant T-cell population within epithelial tissues, including the skin, gut, and airways (16). They accumulate in infected tissue, including lesions from leprosy patients, and in vitro they produce an extensive array of cytokines and chemokines after antigenic stimuli (3, 6). T cells recognize a range of mycobacterial proteins in vitro, including low-molecular-weight proteins and nonprotein ligands, often without the need for previous antigen processing and presentation (4, 6, 30). In addition, T cells taken from patients with AIDS that are infected with mycobacteria recognize mycobacterial antigens in vitro (28). A recent epidemiological study found that T-cell numbers were increased in HIV-M. avium- coinfected individuals but not in patients with HIV-M. tuberculosis coinfections, thus suggesting a role for T cells in response to M. avium infection in patients with AIDS (24).

To investigate the hypothesis that T cells control the influx of inflammatory cells into the lung after M. avium infection, gene-disrupted (-K/O) mice were infected with M. avium, and the disease progression was monitored. In these experiments the mice were infected with either M. avium 724 or a smooth transparent isolate of M. avium 2-151 (2-151 SmT). Strain 724 is a highly virulent M. avium isolate that grows progressively within mice (12). Strain 2-151 SmT is also a virulent isolate, which induces strong protective immunity resulting in a chronic infection (12). Mice were infected by aerosol exposure to a low bacterial dose in order to mimic one of the natural routes of infection in humans. We report here that the contribution of T cells to immunity to M. avium varies depending upon the infective strain used. Infection of -K/O mice with M. avium, while having only a transient influence upon the growth of bacteria, significantly affected the inflammatory response generated within the lungs. -K/O mice developed lesions containing a higher proportion of lymphocytes and, in the case of infection with M. avium 724, they did not show the extensive neutrophil influx nor the development of caseated lesions seen in wild-type-infected mice.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Mice. Breeding pairs of T-cell receptor C gene mutant mice (C57BL/6J-Tcrd^tm/mom and JR2120) (23) were obtained from the Jackson Laboratories (Bar Harbor, Maine) and bred in the Laboratory Animal Resources Center at Colorado State University. Wild-type controls (C57BL/6J) were obtained from the Jackson Laboratories as required. Age- and sex-matched mice were kept under barrier conditions in the ABL-3 biohazard facility throughout the experiment. The specific-pathogen-free nature of each colony was shown by testing sentinel animals; these were determined to be negative for 12 known mouse pathogens.

Bacteria and infection. M. avium 724 and 2-151 SmT were grown from laboratory stocks in Proskauer-Beck liquid medium to mid-log phase, aliquoted, and then frozen at 70°C. Mice were infected with approximately 500 bacteria by using a Middlebrook Airborne Infection apparatus (Middlebrook, Terre Haute, Ind.) as previously described (26). The numbers of viable bacteria in the lung, spleen, and liver were determined at various time points by plating serial dilutions of organ homogenates on nutrient Middlebrook 7H11 agar and counting the bacterial colonies after 14 days of incubation at 37°C. The data are expressed as the log₁₀ value of the mean number of bacteria recovered per organ (n = four animals).

Histology. Tissues from four mice per experimental group were infused with fresh 10% formaldehyde in phosphate-buffered saline. Sections made from paraffin blocks were stained with hematoxylin and eosin. Consecutive sections were stained for acid-fast bacilli by the Kinyoun staining procedure. Sections were examined by a veterinary pathologist without prior knowledge of the experimental groups.

Statistical analysis. Differences between the mean of experimental groups were analyzed by using the Student t test. Differences were considered significant when P was <0.05.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Variation in the growth of two strains of M. avium in -K/O mice. Strains 724 and 2-151 SmT are virulent serotype-2 strains of M. avium that have been extensively studied in this laboratory. Both generate protective T-cell immunity early during the course of the infection, and as a result the growth of 2-151 SmT is restrained, giving rise to a chronic disease state. In contrast, however, mice infected with 724 gradually lose their expression of acquired immunity (9a), for reasons as yet completely unknown, allowing the infection to grow progressively.

View larger version (24K): [in this window] [in a new window]   FIG. 1.   Bacterial growth in -K/O mice infected with M. avium 724 or 2-151 SmT. -K/O mice and their wild-type littermate controls were infected with approximately 500 bacilli via aerosol. Bacterial load was assessed up to 120 days postinfection. Data are expressed as the mean ± the standard deviation (n = 4). *, P < 0.05.

Differences in the inflammatory response generated in -K/O mice. The lungs of wild-type and -K/O mice infected with M. avium had significant differences in lesion type, with the severity differing between the two bacterial strains (Table 1). Wild-type mice infected with M. avium 724, while showing progressive bacterial growth, did not mount a strong inflammatory response to the invading bacteria during the first 60 days of infection. There was some thickening of the alveolar septae by lymphocytes and macrophages along with scattered small foci of lymphocytes, but no large rafts of macrophages or lymphocytes were present (Fig. 2A). By day 90, the lungs of the wild-type mice had extensive granulomas with severe caseation surrounded by a thick wall of degenerative neutrophils that were in turn surrounded by a rim of epithelioid macrophages (Fig. 2C, enlarged in Fig. 3A to C). Acid-fast staining of consecutive lung sections revealed extensive numbers of bacilli within the epithelioid macrophage layer surrounding the wall of degenerative neutrophils (Fig. 3D).

View this table: [in this window] [in a new window]   TABLE 1.   Lung histology in wild-type and -K/O mice infected aerogenically with M. avium 724 or 2-151 SmT

View larger version (165K): [in this window] [in a new window]   FIG. 2.   Representative low-magnification photomicrographs of lung tissue from wild-type and -K/O mice infected with M. avium 724. (A) Wild-type mice at 60 days postinfection. Note the limited cellular infiltrates composed of scattered lymphocyte foci, including apparent bronchial alveolar lymphatic tissue expansion. The lungs of -K/O mice were similar in appearance at this time point. (B) -K/O mice at 90 days postinfection. There are multiple small granulomas composed predominantly of macrophages and lymphocytes with few neutrophils. (C) Wild-type mice at 90 days postinfection. A large focus of necrosis with central caseation can be seen, with lesions surrounded by neutrophils and then cuffed with macrophages (enlarged in Fig. 3). (D) -K/O mice at 120 days postinfection. Large granulomatous lesions composed primarily of macrophages with a few foci of neutrophils are evident. Lymphocyte numbers appear to have waned since day 90. Hematoxylin and eosin staining was used. Bar, 100 µm.

View larger version (161K): [in this window] [in a new window]   FIG. 3.   Representative high-magnification photomicrograph of lung tissue from mice infected with M. avium 724. (A) Section of a caseated lesion in a wild-type mouse at 90 days postinfection. Note the concentrated laminated rings of degenerative neutrophils (middle) and epithelioid macrophages (top) surrounding the central caseation (bottom). The thick-bar area is magnified in panel B; the thin-bar area is magnified in panel C. (B) Higher magnification of the neutrophilic lamina (bottom) with the epithelioid macrophages (top). (C) The center of the caseated area contains amorphous debris. (D) Acid-fast bacilli in macrophages. This is a magnification of the area of panel A just above the thick bar. (E) Section of a noncaseated lesion in a -K/O mouse at 90 days postinfection. Note that the lesion is composed primarily of macrophages and lymphocytes. (F) Acid-fast bacilli in macrophages from a -K/O mouse at 90 days postinfection. The panel is a magnification from an area similar to that shown in panel E. Panels A, B, C, and E were stained with hematoxylin and eosin; panels D and F were stained with Kinyoun's stain. Bar, 10 µm.

View larger version (93K): [in this window] [in a new window]   FIG. 4.   Representative moderate-magnification photomicrograph of liver tissue from M. avium 724-infected mice. (A) Wild-type mice at 120 days postinfection. Cellular infiltrates resulting in the formation of multifocal granulomatous hepatitis are shown. (B) -K/O mice at 120 days postinfection. This image is similar to panel A, with cellular infiltrates resulting in the formation of multifocal granulomatous hepatitis. Panels were stained with hematoxylin and eosin. Bar, 100 µm.

View larger version (168K): [in this window] [in a new window]   FIG. 5.   Representative low-magnification photomicrographs of lung tissue from M. avium 2-151 SmT-infected mice. (A) Wild-type mice at 60 days postinfection. Cellular infiltrates resulting in the formation of primary granulomas composed predominantly of macrophages with few lymphocytes are shown. This response was prominent earlier than the one shown in Fig. 2. (B) -K/O mice at 60 days postinfection. The lesions are similar to those in Fig. 4A, but the proportion of lymphocytes is higher than in the wild type. (C) Wild-type mice at 90 days postinfection. The lymphocytic granulomas are larger than at 60 days postinfection (Fig. 4A). (D) -K/O mice at 90 days postinfection. The K/O mice also showed an increase in granuloma formation. These are composed predominantly of macrophages and lymphocytes, with the proportion of lymphocytes being higher than in the wild type (Fig. 4C). Panels were stained with hematoxylin and eosin. Bar, 100 µm.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This study shows that, when bacterial growth remains unchecked, as in the M. avium 724 infection, the presence of T cells in the host can be detrimental and appears to accelerate the destructive pathogenic response. If, however, the bacterial infection remains chronic, as in the 2-151 SmT infection, then the T cells appear to play no significant role in disease progression.

The key observation reported here concerns the necrotic nature of the lesions in strain 724-infected mice. In M. avium-infected individuals a high bacterial burden within the lungs is usually accompanied by extensive tissue necrosis and fibrosis (2, 11, 13, 15, 20). Like the lesions seen in humans, wild-type mice infected with M. avium 724 developed lesions composed predominantly of neutrophils and macrophages. Perhaps as a consequence of the high bacterial numbers in this animal model, extensive tissue necrosis was also seen, with some necrotic lesions progressing to a caseated state.

An indication of the mechanism involved in the generation of this pathogenic response is provided by the observations reported here. Specifically, while the bacterial loads in the wild-type and -K/O mice were equivalent, the lesion development was significantly different. In contrast to the extensive degenerating lesions seen in the wild-type mice, the lung lesions of the -K/O mice infected with strain 724 consisted of small granulomas containing mixtures of macrophages and lymphocytes. By comparing histological samples from different time points it was evident that progression of the pathogenic response was slower in the -K/O mice. By day 120 of the experiment, lung lesions in the K/O mice had increased to a size similar to those of the wild-type mice 1 month earlier. However, the lesions were composed of vast fields of epithelioid macrophages, with a few scattered aggregates of lymphocytes and with no overt necrosis evident.

While the absence of T cells diminished the necrosis and tissue damage seen within the strain 724-infected mice, this was not the case in M. avium 2-151 SmT-infected animals. Both wild-type and -K/O mice infected with 2-151 SmT developed similar lymphocytic lesions, which were composed predominantly of macrophages and lymphocytes, except that lesions in the -K/O mice appeared to be more lymphocytic in nature.

One possible explanation for the differences in lesion formation induced by the two strains of M. avium may be in the immune response each one induces. Protective immunity to M. avium requires an inflammatory cell influx to surround and contain infected macrophages and a protective T-cell response to activate macrophages to kill infecting bacilli. M. avium 2-151 SmT induces both the influx of inflammatory cells to surround infected macrophages and, as previously shown, a strong acquired immune response (14) that controls bacterial growth such that a chronic disease state then ensues. Mice aerogenically infected with strain 724 show no ability to contain bacterial growth. Early in the infection only a mild interstitial pneumonia was seen within the lungs of infected mice, and cells were not recruited to surround infected macrophages. It appears, therefore, that a protective acquired immune response was not generated. Indeed M. avium 724 grows identically in both CD4 K/O and wild-type mice (28a). Studies in this laboratory (9a) found that intravenous infection with 724 does appear to generate acquired immunity during the early course of the infection but, for reasons that are currently unclear, this specific resistance is then gradually lost, thus allowing the infection to grow progressively and eventually kill the animal.

We postulate that in mice aerogenically infected with strain 724, once the bacterial load reaches a critical threshold, T cells, driven by recognition of mycobacterial antigens (4, 5, 31) and/or recognition of damaged or infected self (21), produce a cytokine-chemokine response that stimulates the influx of inflammatory cells, particularly macrophages into infected tissue. These macrophages are detrimental to the mouse since they serve as host cells for the pathogen and, because of the absence of protective T cells, they remain inactivated and gradually degenerate, causing local tissue damage and an influx of neutrophils. In the absence of T cells, this inflammatory cell influx is diminished and less tissue damage is seen.

In M. avium 2-151 SmT-infected mice, where a protective gamma interferon (IFN-)-producing T-cell response is generated (1), macrophages entering infected tissues become activated and are able to control bacterial growth. In the absence of T cells, more lymphocytes were seen in lesions in the lung, suggesting that the presence of T cells dampens the recruitment of lymphocytes to the infected tissue. In 2-151 SmT-infected mice, because of the presence of a protective acquired immune response, the macrophages recruited by T cells become activated and are able to control bacterial growth. As the bacterial growth is controlled the amount of tissue damage and mycobacterial debris is also reduced. In this situation therefore, unlike in the strain 724-infected tissue, there was a much lower stimulus for the recruitment of neutrophils and the development of necrosis and caseation.

In support of this hypothesis, studies investigating the role of T cells in other infections have shown repeatedly that T cells, while often not directly affecting the growth of the infective pathogen, still control inflammatory processes in response to infection (10, 27, 29). In the absence of T cells, infected mice generally developed larger more-diffuse lesions, with an increase in tissue necrosis and abscess formation being common.

Indeed, in M. tuberculosis infection the absence of T cells led to the development of a pyogranulomatous inflammatory response with lesions containing increased numbers of neutrophils and large foamy macrophages, distinct from the lymphocytic granulomas formed in the wild-type mice (10). While this data appears to contradict our data, we would contend that it is the recruitment of macrophages by T cells into a nonimmune site that leads to the potential for larger more necrotic lesions. Thus, in strain 724-infected mice, T cells stimulate a macrophage influx, but no protective T-cell response is present to activate these macrophages, and as a result bacterial growth continues and the macrophages degenerate, stimulating a neutrophil influx which increases the tissue damage seen. In contrast, in M. tuberculosis-infected mice, T cells still stimulate a macrophage influx; however, these cells are activated by the protective T-cell response, and bacterial growth is controlled. In the absence of T cells, macrophage recruitment is diminished and so we see the rapid growth of M. tuberculosis in a low macrophage environment which results in rapid tissue damage and the subsequent recruitment of neutrophils.

Finally, this animal model, if extrapolated, may have implications for HIV-positive patients infected with M. avium. The implication here is that a failure of acquired cellular immunity expressed by CD4 T cells allows an opportunistic M. avium infection to disseminate and grow. Meanwhile, the T-cell response inadvertently promotes this process by amplifying the recruitment of macrophages into lesions, but the absence of IFN--secreting CD4 T cells results in astronomic bacterial loads and pathogenic lesions. This model is therefore similar to the disease progression in HIV-positive patients with disseminated M. avium infections and may be useful in identifying mechanisms of pathogenesis.