Latency-Associated Peptide of Transforming Growth Factor β Enhances Mycobacteriocidal Immunity in the Lung during Mycobacterium bovis BCG Infection in C57BL/6 Mice

Abstract

Latency-associated peptide of transforming growth factor β (TGF-β) (LAP) was used to determine whether in vivo modulation of TGF-β bioactivity enhanced pulmonary immunity to Mycobacterium bovis BCG infection in C57BL/6 mice. LAP decreased BCG growth in the lung and enhanced antigen-specific T-cell proliferation and gamma interferon mRNA expression. Thus, susceptibility of the lung to primary BCG infection may be partially mediated by the immunosuppressive effects of TGF-β.

Tuberculosis is a leading infectious disease worldwide and responsible for significant morbidity and mortality (4, 15). Although innate and acquired T-cell responses are necessary for containment of mycobacterial growth during Mycobacterium tuberculosis infection, host responses are not sufficiently mycobacteriocidal and bacilli survive sequestered within granulomas in the lung (3, 8, 19, 22). Thus, host immune responses may be permissive for growth and survival of mycobacteria in the lung, which is particularly susceptible to aerosol infection with M. tuberculosis (18).

Transforming growth factor β (TGF-β) is a product of mycobacterial antigen-activated macrophages (12, 14), lung epithelial cells (21), and other inflammatory cells and is secreted as a homodimer noncovalently bound to its latency-associated peptide (LAP) (7). Extracellular dissociation of TGF-β from LAP releases biologically active TGF-β. TGF-β deactivates macrophages (24), suppresses T-cell functions (17), and has been detected in granulomas during active tuberculosis in humans (23) and in murine lungs after intratracheal M. tuberculosis H37Rv infection (11). In addition, TGF-β renders T cells hyporesponsive to antigen stimulation and impairs mycobacteriocidal activity of M. tuberculosis-infected monocytes (14, 24). In vitro modulation of TGF-β bioavailability with neutralizing antibody or LAP enhances mycobacteriocidal functions of monocytes and improves the hyporesponsiveness of T cells isolated from patients with tuberculosis (12, 14). In vivo, LAP enhances hepatocyte regeneration and reduces fibrosis in TGF-β-transgenic mice (1). Since TGF-β exerts negative regulatory effects on macrophage and T-cell functions, we hypothesized that TGF-β expression contributes to the growth and survival advantage of Mycobacterium bovis BCG in the lung.

It has been shown previously that intratracheal BCG infection in C57BL/6 mice is characterized by maximal bacterial growth and T-cell recruitment and activation in the lung (10) after 28 days of infection. Although expression of gamma interferon (IFN-γ) mRNA and protein closely parallels growth and clearance of BCG in the lung, a low-level steady-state bacterial burden persists 10 to 12 weeks after infection. Thus, this model mimics primary infection in humans who do not develop progressive disease and is useful for studying mechanisms of protective immunity expressed in the lung. In the current study, LAP was used to modulate TGF-β bioactivity in vivo and to determine if neutralization of TGF-β expression enhances mycobacteriocidal host immune responses during primary pulmonary BCG infection.

The experimental design was based on previous studies demonstrating systemic delivery of LAP and inhibition of TGF-β bioactivity (1). First, 10- to-12-week-old pathogen-free female C57BL/6 mice (Charles River Laboratories, Wilmington, Mass.) were anesthetized intraperitoneally (i.p.) with tribromoethanol (180 mg/kg of body weight) and infected either intratracheally or by aerosol with 0.5 × 10⁵ to 1.0 × 10⁵ or 100 to 500 CFU of BCG as described previously (10, 20). Osmotic minipumps (Alzet 1002; Alza Corporation, Palo Alto, Calif.) filled with 0.0125 mg of recombinant human LAP (R & D Systems, Minneapolis, Minn.) or phosphate-buffered saline (PBS) were implanted i.p. at the time of infection and replaced after 14 days of treatment. No surgical-wound infection or mortality was observed. After 14 or 28 days of infection, mice were euthanized and tissue samples were processed for numbers of BCG CFU, cytokine expression, and T-cell proliferation, as previously described (10). In each experiment, 3 to 5 mice per group were used. Mice were housed in microisolator cages and were fed a standard rodent diet and water ad libitum.

To assess the effects of LAP treatment on pulmonary immune defenses, BCG growth, cytokine expression, and T-cell proliferation were examined in different lung compartments: bronchoalveolar spaces, lung parenchyma, and mediastinal lymph node. We hypothesized that LAP treatment would enhance mycobacteriocidal immune responses in the lung. Figure 1 shows that LAP treatment significantly reduced BCG growth in the lung and lymph node by 40 and 60%, respectively, after 28 days of an aerosol infection, compared to that in controls. However, after 14 days of infection, numbers of BCG CFU in the lung were similar in control (424.2 ± 126.9) and LAP-treated (437.1 ± 91.3) mice, suggesting that LAP treatment does not appear to affect early growth of BCG in lung parenchyma. Since bronchoalveolar lavage (BAL) and lymph node CFU were at or below the level of detection on day 14, we were unable to determine if growth in these compartments was altered by LAP treatment.

FIG. 1.

LAP treatment decreases BCG growth in bronchoalveolar spaces (BAL), lung parenchyma (LUNG), and mediastinal lymph node (NODE). First, i.p. pumps containing PBS or LAP (0.0125 mg) were aseptically implanted in mice that had been aerosol infected with 500 to 1,000 CFU of BCG. After 2 weeks, osmotic pumps were replaced, and LAP (and PBS) treatment continued an additional 2 weeks. BAL cells and lung and lymph node cell homogenates were prepared from individual mice (n = 5) and CFU were counted. The results show mean (±SEM) CFU and are representative of three independent experiments. The Wilcoxon rank sum test was used to compare numbers of CFU for control and treated mice ∗, P ≤ 0.05.

A similar decrease in BCG growth on day 28 (43.4% ± 8.9% reduction compared to controls [mean ± standard error of the mean {SEM}, n = 9]) also was observed in BAL. However, the result was not statistically significant, owing to higher variability of CFU measurements on BAL samples. This downward trend in BCG CFU in BAL was paralleled by a 30.6% (±10.9% [mean ± SEM, n = 9]) reduction in BAL cell number compared to that of controls. Examination of BAL cell cytospins (25,000 cells) stained with Diff-Quik (Fisher, Pittsburgh, Pa.) did not reveal changes in the percent distribution of lymphocytes, neutrophils, or macrophages (data not shown). Thus, decreased BCG CFU in bronchoalveolar spaces during LAP treatment might be due to enhanced innate defenses of alveolar macrophages. This hypothesis is consistent with other studies demonstrating enhanced killing of intracellular mycobacteria by human mononuclear phagocytes treated with LAP (12). However, additional studies are needed to determine if alveolar macrophage function and TGF-β activity can be modulated by LAP in vivo.

In contrast to infection on day 14, reduced BCG growth was measured after 28 days of infection. This result suggested that the effect exerted by LAP may have depended upon mononuclear cells which migrate to the lung after 14 days of infection (10). To determine whether increased T-cell recruitment coincided with reduced growth of BCG in the lung, parenchymal T cells were quantified by flow cytometery. As described previously, 5 × 10⁵ cells were stained with phycoerythrin-conjugated anti-CD3 and either fluorescein isothiocyanate-conjugated anti-CD4 or anti-CD8 monoclonal antibodies (#01085A, #01064A, and #01044A; Pharmingen, San Diego, Calif.) and compared to lung cells stained with isotype control antibodies (10). In four experiments, a reduced number of BCG CFU in the lungs of LAP-treated mice was not associated with changes in either CD4⁺ or CD8⁺ T-cell count, compared to the counts for PBS-treated control mice (data not shown). In addition, histologic sections of the left lung and right upper lung lobe were stained with hematoxylin and eosin and examined microscopically. After 28 days of infection, we observed typical peribronchial and perivascular lymphocytic infiltrates and an alveolitis composed primarily of activated epithelioid macrophages, as previously described for BCG and M. tuberculosis (10, 22). Although significant accumulation of inflammatory cells has been described in TGF-β gene-disrupted mice (2, 6), LAP treatment did not detectably affect cellular composition or distribution in infected lungs (data not shown). Thus, LAP treatment appeared to enhance effector cell functions rather than cell recruitment or granuloma formation.

Next, we determined if control of BCG growth in the BAL and lung correlated with specific cytokine expression. Using the OptEIA assay kit (Pharmingen), we measured IFN-γ protein in BAL fluids. In BAL a 31.5% (± 8.5% [mean ± SEM, n = 9]) reduction in IFN-γ coincided with reduced BCG CFU, but the result was not statistically significant (P > 0.05). However, we do not know whether LAP treatment resulted in enhanced IFN-γ expression which preceded day 28 and was not detectable when CFU had decreased. Thus, it is not known if BAL IFN-γ expression is a correlate of protective immunity in the bronchoalveolar space. We also examined both naturally processed (endogenous, bioactive) and acid-activated (total) TGF-β in BAL fluids obtained after 14 and 28 days of infection to determine if LAP affected BAL TGF-β. Levels of naturally processed TGF-β in BAL were similar in LAP- and PBS-treated mice (data not shown). However, BAL proteins may have nonspecifically interfered with acid activation and precluded measurement of total TGF-β (data not shown). Thus, we developed a direct enzyme-linked immunosorbent assay to measure endogenous LAP as a surrogate marker of activated TGF-β. Briefly, Immulon 4 plates (Dynex Technologies, Chantilly, Va.) were coated with BAL fluid samples or recombinant human LAP as a standard (R & D Systems). Plate-bound LAP was detected using a biotinylated goat anti-human LAP polyclonal antibody (R & D Systems) and streptavidin alkaline phosphatase (DAKO, Carpinteria, Calif.). The sensitivity of the assay was less then 1 ng/ml, but we were unable to adequately discriminate between LAP-treated and control mice. In addition, preliminary studies have suggested that BAL nitrite concentrations (range, 876.0 to 1,389.2 pM) were similar in both PBS- and LAP-treated mice after 28 days. Thus, while additional studies are needed to understand protective innate and T-cell-mediated immunity in bronchoalveolar spaces, subtle changes in TBG-β bioavailability may have been sufficient to enhance immunity and decrease BCG growth in bronchoalveolar spaces.

We also analyzed cytokine expression in lung and lymph node cells using the RNase protection assay (25). In brief, cells were solubilized in Tri-Reagent (Molecular Research Center, Cincinnati, Ohio) and total RNA was isolated according to the manufacturer's instructions. By use of the mCK-3b Multi-Probe Template Set (Pharmingen), 2.5 μg of RNA was hybridized to a cocktail of [³²P]UTP (DuPont)-labeled RNA probes specific for TNF-α, interleukin-6, IFN-γ, IFN-β, lymphotoxin-β, and TGF-β1, -β2, and -β3, as well as glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and ribosomal structural protein L32 as housekeeping genes. Protected RNA hybrids were electrophoresed through a 5% denaturing polyacrylamide gel and were identified by using their migration position relative to that of unhybridized cytokine probes. Bands were quantitated using a phosphorimager (Bio-Rad, Hercules, Calif.) and normalized to GAPDH. As shown in Fig. 2, LAP treatment resulted in a twofold increase in IFN-γ mRNA expression in the lung on day 28 of infection compared to that in control mice (P < 0.05, Wilcoxon rank sum analysis). In contrast, expression of IFN-γ mRNA at day 14 was not significantly altered (Fig. 2) or associated with reduced BCG growth. Expression of TNF-α, TGF-β (isoforms 1 to 3), interleukin-6, and lymphotoxin-β mRNA were unchanged on days 14 and 28 (data not shown). Enhanced expression of IFN-γ correlated with reduced BCG growth in the lungs of LAP-treated mice. These results suggest that LAP probably down-modulated TGF-β bioactivity and enhanced T-cell immunity in lung parenchyma, although the role of additional cytokines has not been excluded.

FIG. 2.

Increased IFN-γ mRNA expression is associated with a reduction in BCG infection in lung parenchyma. Mice were infected intratracheally with 0.5 × 10⁴ CFU of BCG and treated i.p. with PBS or LAP (0.0125 mg). After 2 and 4 weeks, lung homogenates from three mice were prepared, total RNA was extracted, and IFN-γ mRNA was measured using RNase protection. Radioactivity was quantitated using a phosphorimager, and IFN-γ-specific mRNA was quantitated relative to GAPDH. The data show the ratio (mean ± SEM) of IFN-γ to GAPDH (n = 3 mice) of a representative experiment. Similar results were observed in mice infected by aerosol.

Since TGF-β has been shown to impair T-cell proliferation (12, 13), we also examined spontaneous and antigen-driven T-cell proliferation of lymph node and lung cells isolated from LAP-treated BCG-infected mice. Lymph node cells were cultured at 2 × 10⁵/well, in the presence or absence of BCG (10⁵/ml), for 72 h in complete RPMI 1640 medium (10). Cells were labeled with 1 μCi of [³H]thymidine (ICN Chemicals, Costa Mesa, Calif.)/well and harvested onto glass microfibers after 6 h. Incorporated radioactivity was measured using liquid scintillation (Packard Instruments, Meriden, Conn.). As shown in Fig. 3, proliferation of both spontaneous and BCG-stimulated lymph node T-cells was significantly higher (P < 0.05) in LAP-treated mice, suggesting that in vivo modulation of TGF-β bioactivity enhanced T-cell responsiveness to antigen stimulation. Enhanced proliferation was observed after both 14 and 28 days. However, increased IFN-γ mRNA expression was associated with reduced BCG growth only after 28 days, when maximal T-cell numbers were measured in the lung. To estimate responsiveness of CD4⁺ T cells, which respond primarily to soluble antigens, 2 × 10⁵ lung cells were cultured with and without mycobacterial purified protein derivative (20 μg/ml). As with BCG-stimulated lymph node cells, we observed a significant increase (P = 0.02) in proliferation of lung cells from LAP-treated mice compared to those from controls (2,541.0 ± 414.0 versus 1,637.0 ± 186 cpm, n = 5 mice/group). Thus, a LAP-mediated decrease in TGF-β bioactivity appeared to enhance T-cell responsiveness and mycobacteriocidal immunity. These results suggest that during primary pulmonary BCG infection, induction and activation of TGF-β expression may impair immune cell activation and clearance of bacilli. In complementary studies using a guinea pig model, Dai and McMurray have shown that recombinant TGF-β inhibits T-cell activation and enhances growth of virulent M. tuberculosis (5).

FIG. 3.

LAP enhances spontaneous and antigen-specific T-cell proliferation. Mice were infected intratracheally with 500 to 1,000 CFU of BCG and treated with either LAP or PBS as described. After 14 and 28 days, mediastinal lymph node cells were prepared from individual mice (n = 5) and triplicate cultures were incubated for 72 h in the presence (antigen-specific proliferation) or absence (spontaneous proliferation) of 10⁵ BCG/ml. The data shown (mean counts per minute ± SEM) are representative of three independent experiments. The Wilcoxon rank sum test was used to compare LAP and PBS treatments. ∗, P ≤ 0.05 for spontaneous responses; †, P ≤ 0.05 for BCG-stimulated responses.

Innate immunity and acquired T-cell responses are necessary for control of mycobacterial infection in the lung (3, 9, 16, 19). However, mycobacteriocidal immunity is incomplete, and organisms persist within macrophages and granulomas (22). We hypothesized that the susceptibility of the lung to BCG infection was partly due to immunosuppressive effects of TGF-β. We utilized LAP to modulate the bioavailability of TGF-β and test this hypothesis. Although the evidence for blocking TGF-β in vivo by LAP was indirect, we observed a significant reduction in BCG growth in the lung that coincided with increased IFN-γ mRNA expression and enhanced T-cell responsiveness. A downward trend in BCG growth, cellular recruitment, and IFN-γ was observed in BAL. In other experimental animal models, administration of TGF-β increases mycobacterial growth and decreases proliferative responses of T cells. Thus, local TGF-β expression in the lung may impair innate and T-cell-mediated immunity, resulting in a lung microenvironment permissive for BCG growth and persistence.