Analysis of Cellular Phenotypes That Mediate Genetic Resistance to Tuberculosis Using a Radiation Bone Marrow Chimera Approach

Numerous association and linkage studies with humans (1-4, 10, 11, 21, 28), as well as segregation and genetic mapping studies with mice (13-15, 18, 19, 24), have demonstrated that susceptibility to and severity of tuberculosis (TB) infection are under the control of several interacting genetic loci. Although the nature of epistatic interactions within these complex genetic networks is poorly understood, there is little doubt that it is the host genetics that determine the particular place which any given individual or inbred mouse strain occupies on the wide scale of TB manifestations, from asymptomatic control of mycobacteria to rapidly progressive fatal pulmonary disease. In the experimental system which we have studied for many years, polar manifestations of TB severity are represented by the two mouse strains I/St (H-2^j) and A/Sn (H-2^a). Following intravenous or intratracheal TB challenge, I/St mice rapidly lose body weight, develop severe lung pathology, and die, whereas in A/Sn mice, all these complex quantitative disease phenotypes (QDP) develop significantly later (6, 22). Whether TB progression is rapid or slow is determined by at least four interacting quantitative trait loci, mapped within the mouse genome by means of a genome-wide microsatellite scan (14, 24); particular genes involved in disease control have not been identified yet.

At the cellular level, analysis of more-subtle phenotypes that underlie the QDP demonstrated that, compared to A/Sn, infected I/St mice are characterized by (i) a significantly higher T-cell proliferation but lower level of gamma interferon (IFN-γ) production in the lungs (6, 16), (ii) a decreased capacity for interstitial lung macrophages to inhibit mycobacterial growth (17), and (iii) a prolonged peak of neutrophil accumulation in the lung tissue, accompanied by a greatly increased capacity for neutrophils to ingest (but not to kill) mycobacteria (5). The last phenotype might be important in creating a reservoir of mycobacteria hidden from the macrophage bacteriostatic response, given that the I/St neutrophils appeared to be unusually resistant to apoptotic death, which, in turn, explains their accumulation in the lungs (5). It remains unclear, however, whether these numerous differences between TB-susceptible and -resistant mice in response to infection totally depend upon the cells of immune system per se or, alternatively, whether some of the phenotypes result from the interaction of bone marrow-derived cells with stromal elements of the lung and lymphoid organs. Since susceptibility to and severity of TB infection in the I/St-A/Sn system are inherited as a recessive QDP, with (I/St × A/Sn)F₁ hybrids being resistant (6, 14), we applied a radiation bone marrow chimera approach to directly analyze in vivo the TB resistance phenotypes that are adoptively transferred to irradiated I/St recipients with bone marrow cells from F₁ mice.

To this end, female I/St recipients were irradiated at 9.5 Gy from a ⁶⁰Co source and, within 6 h, restored with bone marrow cells freshly isolated from femurs of F₁ (experiment) or I/St (control) female donors (one donor-to-one recipient transfer, 9 × 10⁶ to 12 × 10⁶ cells/mouse). Five control mice which did not receive protective cell transfer died on days 9 through 11 following irradiation with signs of acute bone marrow radiation disease. Recipients were rested for 3 months to allow the replacement of the majority (∼90%) of radiation-resistant, long-living tissue macrophages with cells of the donor origin (as determined by complement-dependent cytotoxicity with A/Sn anti-I/St and I/St anti-A/Sn reciprocal hyperimmune antisera, using purified interstitial lung macrophages [see reference 17] as target cells). After the resting period, mice were infected either intravenously (i.v.) (10⁵ CFU/mouse) or intratracheally (i.t.) (10² CFU/mouse) with Mycobacterium tuberculosis H37Rv substrain Pasteur as described previously (6, 16). The challenge doses for both types of infection were diminished fivefold compared to those used for infection of nonirradiated mice (6, 16), since preliminary experiments demonstrated that even for mice restored with a high dose (15 × 10⁶/mouse) of syngeneic bone marrow cells, postinfection time to death is significantly shorter than that for nonirradiated controls. Separate groups of mice protected by syngeneic (control) or semiallogeneic (experiment) adoptive transfers served as a source of interstitial lung macrophages for evaluation of their effector function in vitro at week 4 post-i.t. challenge. The following parameters of the antimycobacterial response were assessed: (i) time to death and CFU counts in lungs and spleens (serial whole-organ 10-fold dilutions were plated onto Dubos agar and incubated for 18 to 20 days at 37°C for CFU counting); (ii) fluorescence-activated cell sorter-based assessment of accumulation of CD4⁺ (monoclonal antibody [mAb] clone CT-CD4) and CD8⁺ (mAb clone CT-CD8a) T cells, Mac-3⁺ (mAb clone CI:A3-1) macrophages, and Ly-6G⁺ (mAb clone RB6-8C5) neutrophils in the lung tissue (all antibodies were from Caltag, Burlingame, CA); (iii) production of proinflammatory, apparently protective (reviewed in references 8 and 23), cytokines (interleukin 12 [IL-12], IL-6, tumor necrosis factor alpha, and IFN-γ) by lung cells (enzyme-linked immunosorbent assay kits purchased from BD-PharMingen, San Diego, CA); (iv) capacity for interstitial lung macrophages to restrict mycobacterial growth following infection in vitro, as measured by [³H]uracil uptake by mycobacteria. All experimental procedures were previously described in detail (5, 6, 16, 17). Two experiments with i.v. challenge and one experiment with i.t. challenge were performed with consistent results.

As shown in Fig. 1, adoptive transfer of bone marrow cells from TB-resistant F₁ mice into TB-susceptible I/St recipients resulted in significant prolongation of postinfection survival time, compared to recipients that received syngeneic I/St bone marrow cells, in both i.v. (P < 0.02, log rank test) and i.t. (P < 0.001) challenge models. The differences in lung and spleen CFU counts between F₁→I/St chimeric mice and I/St→I/St mice reached a significant level (P < 0.05, Mann-Whitney U test) only late in the infection course (week 5 post-i.v. challenge [Fig. 1]; the corresponding time point has not been assessed in the i.t. challenge model due to the death of control mice). Thus, the expression of the resistance phenotype in F₁→I/St chimeras was much more obvious in survival than in bacterial loads in the tissues. The most likely interpretation of this difference is that lung pathology, the hallmark of TB infection and the main cause of mortality (8, 23), may develop after a relatively weak triggering event (i.e., relatively low number of mycobacteria) if the host is genetically predisposed to respond detrimentally to this trigger.

The degree of lung pathology can be evaluated indirectly by enumerating lymphoid cells that infiltrate lung tissue in the course of the disease. The results of corresponding experiments obtained using the i.t. challenge model are displayed in Fig. 2. In agreement with previously reported characteristics of inflammatory responses in the lungs of I/St, A/Sn, and F₁ mice (6), the total cellularity and the numbers of CD4⁺ and CD8⁺ T cells and Ly-6G⁺ neutrophils increased more prominently in the susceptible I/St→I/St mice during the fourth week of infection. The only cells whose accumulation did not differ between I/St→I/St and F₁→I/St mice were Mac-3-positive lung macrophages. In the i.v. challenge model, similar results regarding T-cell accumulation and the total increase in lung cellularity were obtained, although no differences in the numbers of phagocytes were found between the two groups of mice (data not shown).

Another characteristic feature that distinguishes the immune response to mycobacteria in I/St and A/Sn mice is a lower level of IFN-γ production in lungs of mice of the susceptible I/St strain irrespective of the challenge route (6, 16). As shown in Fig. 3, significantly (P < 0.05) more IFN-γ was produced in the antigen-specific manner by lung cells extracted at week 3 postinfection from F₁→I/St chimeras than by lung cells from I/St→I/St mice, both after i.v. (Fig. 3A) and i.t. (Fig. 3B) challenges. In addition, tumor necrosis factor alpha production in response to mycobacterial sonicate was also elevated in F₁→I/St mice following i.v. infection. On the other hand, the production of regulatory cytokines IL-6 and IL-12 did not differ between the two groups, and their levels did not respond to stimulation with mycobacterial antigens in vitro, which is in good agreement with the results obtained previously (6).

Another important difference between TB-susceptible I/St and TB-resistant A/Sn mice is a more effective inhibition of mycobacterial growth by lung macrophages from the latter mice (17). To test this phenotype in radiation chimeras, we used a surrogate [³H]uracil assay which provides an accurate evaluation of mycobacterial growth in phagocytic cells (17). Infection of lung macrophages with mycobacteria in vitro at various multiplicities of infection (MOI) showed that F₁→I/St cells display a significantly (P, <0.01 to 0.001 for different MOI) greater bacteriostatic capacity than I/St→I/St cells (Fig. 4). Importantly, when exogenous IFN-γ was added to macrophage-mycobacterium cocultures, the capacity for inhibition of mycobacterial growth increased profoundly in both groups and became identical. Thus, a deficiency in IFN-γ production by lung T cells which gradually develops in I/St mice following TB infection (6) may explain their increased susceptibility in terms of lung CFU counts at the advanced stage but not at the early stage of the disease (Fig. 1) (see also references 6 and 17).

Taken together, these results indicate that in our I/St-A/Sn-F₁ experimental system replacement of bone marrow-derived cells in TB-susceptible recipients with their counterparts from resistant animals is sufficient for the expression of all cellular phenotypes underlying the major resistant QDP in the lung and for the significant prolongation of survival time. We made an attempt to evaluate possible effect of stroma on the expression of these phenotypes by incorporating F₁→F₁ control mice in our experiments. Such mice were prepared and challenged, and they showed a somewhat more resistant phenotype than F₁→I/St chimeras (not shown). However, this fact could be explained by at least four different factors: (i) the influence of stromal elements; (ii) a radioresistance of F₁ that is intrinsically higher than that of I/St recipients (irradiation has a general deleterious effect on TB resistance); (iii) partial rejection of semisyngeneic F₁ cells by the I/St recipients before their T-cell function was completely abrogated by irradiation, i.e., within the first 2 to 3 days postgrafting; (iv) a wider T-cell repertoire in F₁→F₁ animals, as selection of T cells in their thymi allows survival of both H-2^a- and H-2^j-restricted T cells, contrary to the case for F₁→I/St animals. To address these complicated issues, a more sophisticated immunological experimental design is required. Nevertheless, we feel that a direct in vivo evidence of the role which T cells, macrophages, and neutrophils play in pulmonary TB susceptibility and pathology in the mouse model demonstrated here is important for further studying corresponding molecular mechanisms. Microarray gene expression profiling provides a powerful insight into the genetic regulation of host-parasite interactions (7, 9, 12, 25-27). However, RNA extracted from whole lung tissue does not provide a straightforward link between gene expression profiles and biochemical pathways, given a complex cellular composition and polyfunctional physiology of the lung. Simultaneous differential expression of hundreds of genes significantly complicates straightforward interpretations. Thus, simplified in vitro systems may be extremely useful for microarray analysis (25, 26) if they adequately reflect the situation in vivo. On the other hand, the use of in vitro systems (e.g., infected macrophage cell lines) without confirmation that the chosen cells behave similarly to ex vivo tissue macrophages may lead to artifacts. In this regard, it should be emphasized that in our system, the purified interstitial lung macrophages strictly follow the genetic pattern of TB susceptibility (17) and this is readily confirmed by the radiation chimera approach (Fig. 4). This strengthens our hopes that the recently accomplished gene expression profiling in naïve and mycobacterium-infected I/St and A/Sn lung macrophages (Orlova et al., unpublished data) will provide really important new information about mycobacterium-host interactions.

• Open in new tab
• Download powerpoint

FIG. 1.

Survival curves and mycobacterial multiplication in organs in semiallogeneic F₁→I/St chimera and syngeneic I/St→I/St control mice following i.v. (A) and i.t. (B) TB challenge. Mice were infected with either 10⁵ (i.v.) or 10² (i.t.) M. tuberculosis H37Rv CFU. Mortality was monitored daily, and CFU counts were estimated at indicated time points postinfection by plotting serial dilutions of the whole-left-lung homogenate onto Dubos agar. CFU counts in individual mice obtained in one of two similar experiments are present on a log scale, with indications of the means ± standard deviations (SD). *, significant differences in CFU counts (P < 0.05, Mann-Whitney U test).

• Open in new tab
• Download powerpoint

FIG. 2.

Infiltration of lung tissue with CD4⁺ and CD8⁺ T lymphocytes and Ly-6G⁺ neutrophils is less severe in F₁→I/St chimera than in I/St→I/St control mice at week 4 postinfection. At indicated time points following i.t. challenge, right lungs from three to four mice per group were enzymatically disrupted, single-cell suspensions stained with fluorescein isothiocyanate- or phycoerythrin-labeled antibodies against surface markers and analyzed individually by fluorescence-activated cell sorter exactly as described previously (6, 16). Results are present as means ± SD. *, significant differences in cell numbers per organ (P, <0.05 to 0.01, Mann-Whitney U test).

• Open in new tab
• Download powerpoint

FIG. 3.

Cytokine production by lung cells from F₁→I/St chimera and I/St→I/St control mice at week 3 postinfection via i.v. (A) or i.t. (B) route. Production of indicated cytokines in the presence or absence of 10 μg/ml mycobacterial sonicate was measured in supernatants of 48-h cultures of lung cells (2 × 10⁶/ml) individually for three to four mice per group in enzyme-linked immunosorbent assay format exactly as described previously (6, 16). Data are expressed as means ± SD. *, significant differences between experimental and control groups (P < 0.05, Mann-Whitney U test); TNF-α, tumor necrosis factor alpha.

• Open in new tab
• Download powerpoint

FIG. 4.

Inhibition of mycobacterial growth by interstitial lung macrophages extracted from F₁→I/St chimera and I/St→I/St control mice. The rate of mycobacterial growth was measured by [³H]uracil uptake at the indicated MOI at day 3 after establishing cocultures. One microcurie/well [³H]uracil was added for the last 18 h of incubation. The wells containing mycobacteria alone at numbers corresponding to each MOI served as controls. One hundred units/ml of IFN-γ was added to some cocultures. Results are expressed as mean counts per minute ± SD for triplicate cultures; statistically significant (P < 0.01, Mann-Whitney U test) differences between macrophages from F₁→I/St and I/St→I/St mice are marked with asterisks.

• Copyright © 2005 American Society for Microbiology