Immune biology of macaque lymphocyte populations during mycobacterial infection

Abstract

Immune responses of lymphocyte populations during early phases of mycobacterial infection and reinfection have not been well characterized in humans. A non-human primate model of Mycobacterium bovis bacille Calmette–Guerin (BCG) infection was employed to characterize optimally the immune responses of mycobacteria-specific T cells. Primary BCG infection induced biphasic immune responses, characterized by initial lymphocytopenia and subsequent expansion of CD4⁺, CD8⁺ and γδ T cell populations in the blood, lymph nodes and the pulmonary compartment. The potency of detectable T cell immune responses appears to be influenced by the timing and route of infection as well as challenge doses of BCG organisms. Systemic BCG infection introduced by intravenous challenge induced a dose-dependent expansion of circulating CD4⁺, CD8⁺ and γδ T cells whereas, in the pulmonary compartment, the systemic infection resulted in a predominant increase in numbers of γδ T cells. In contrast, pulmonary exposure to BCG through the bronchial route induced detectable expansions of CD4⁺, CD8⁺ and γδ T cell populations in only the lung but not in the blood. A rapid recall expansion of these T cell populations was seen in the macaques reinfected intravenously and bronchially with BCG. The expanded αβ and γδ T cell populations exhibited their antigen specificity for mycobacterial peptides and non-peptide phospholigands, respectively. Finally, the major expansion of T cells was associated with a resolution of active BCG infection and reinfection. The patterns and kinetics of CD4⁺, CD8⁺ and γδ T cell immune responses during BCG infection might contribute to characterizing immune protection against tuberculosis and testing new tuberculosis vaccines in primates.

INTRODUCTION

It is estimated that one-third of the world population is latently infected with Mycobacterium tuberculosis[1]. While the majority of infected people develop acquired immunity to tuberculosis, 5–10% of infected individuals go on to develop active tuberculosis. Identification of immune responses during the early phase of M. tuberculosis infection may facilitate exploring the immune control of tuberculosis and, possibly, the rational design of an effective tuberculosis vaccine. Studies in mice have demonstrated that CD4⁺ T cells play an important role in host defence against tuberculosis [2–7]. Recent studies in HIV-infected humans and SIV-infected macaques also suggest that CD4⁺ T cells are of central importance for immunity to tuberculosis and AIDS-related mycobacterial diseases [8–12]. Moreover, CD8⁺ T cells may also contribute to antimycobacterial immune responses [3,13–15]. CD8⁺ T cells with effector function of cytokine production and cytotoxic activity have been identified in mice infected with M. tuberculosis[14,16,17]. CD8⁺ T cells specific for peptides and lipid antigens of M. tuberculosis have also been detected in the blood of healthy purified protein derivative-positive (PPD⁺) people and tuberculous patients [18–25]. Finally, human γδ T cells may contribute to antituberculosis immune responses [26–28]. It has been recently shown that Vγ2Vδ2⁺ T cells, a unique γδ T cell subset that exists only in primates, can mount adaptive immune responses and contribute to immunity to mycobacterial infections including fatal tuberculosis in monkeys [26]. While all T cell populations are likely to contribute to immune responses to M. tuberculosis infection in humans, how these cell populations act in concert to develop protective immunity has not been well characterized. Although immune responses have been described in humans primed with bacille Calmette–Guerin (BCG) [29–34], the kinetics and magnitudes of mycobacterium-driven immune responses of T cell populations have not been examined formally during the early phases of primary infection and reinfection with mycobacteria including M. tuberculosis.

Non-human primates may provide an important model system for characterizing protective immunity against tuberculosis. Rhesus and cynomolgus monkeys have been used as models for evaluating this disease and vaccine protection against tuberculosis [26,35–40]. The pathology and disease process in M. tuberculosis-infected macaques resemble human tuberculosis [38,39,41]. BCG vaccination has been shown to confer upon rhesus and cynomolgus macaques some protection against tuberculosis, although the degree of BCG-mediated protection in rhesus macaques is controversial [38,39]. The degree of BCG protection against tuberculosis in macaques appears to be consistent with the BCG vaccine efficacy against a fatal form of tuberculosis in humans [26,35,37,40,42,43]. However, immune correlates in the context of BCG protection against M. tuberculosis infection remain poorly characterized both in macaques and humans [39,44]. It is therefore important to identify and compare the immune responses in the settings of BCG infection/vaccination and virulent M. tuberculosis infection in primates. Such comparative immunology between the avirulent and virulent mycobacterial infections might enhance our understanding of fundamental mechanisms underlying the pathogenesis of tuberculosis. Furthermore, with new vaccine candidates emerging for testing a better-than-BCG tuberculosis vaccine, BCG-induced immune responses should provide important standards for comparing and evaluating the vaccine efficacy and surrogate markers for immune protection against tuberculosis. As an important step for characterizing the comparative immunology of BCG and M. tuberculosis infections, we have initiated a series of studies to identify immune responses during BCG infection in macaques. In the present studies, we have assessed the occurrence and magnitude of immune responses of lymphocyte populations as well as host factors influencing those responses during BCG infection and reinfection in different anatomical compartments of macaques.

MATERIALS AND METHODS

Animals

Twenty rhesus (Macaca mulatta) and two pigtailed (Macaca nemestrina) macaques, 3–8 years of age, were used in these studies. These animals were maintained in accordance with the guidelines of the Committee on Animals for Harvard Medical School and the Guide for the care and use of laboratory animals (National Academy Press, 1996).

Mycobacterium bovis BCG infection and reinfection

Mycobacterium bovis BCG (Pasteur strain) was stored in liquid nitrogen and thawed immediately before inoculation. Three different sizes of BCG inocula, 10⁸, 10⁶ and 10³ colony-forming units (CFU), were used for the intravenous and bronchial inoculations. Pulmonary BCG infection was established by the bronchoscope-guided spread of 3 ml phosphate buffered saline (PBS) containing BCG into the caudal or cranial lobe of the right lung. Three to six months after the first BCG inoculation, the macaques received a second inoculation of 10⁸ CFU BCG for reinfection through the same route as inoculation for the primary infection. Following the first or second BCG inoculation, the macaques were followed clinically for any signs of diseases.

Isolation of lymphocytes from blood, lymph nodes and pulmonary alveoli

Peripheral blood lymphocytes (PBL) were isolated from EDTA-anticoagulated blood of the monkeys using Ficoll/diatrizoate gradient centrifugation. Peripheral lymph nodes (some of them were inguinal lymph nodes, whereas others were axillary ones) were obtained by standard biopsy procedures before and after BCG inoculation, and were teased carefully to generate single-cell suspensions. Lymphocytes in pulmonary alveoli were obtained from bronchial alveoli lavage (BAL) fluid. BAL was performed using a paediatric bronchoscope. A total of 60 ml of sterile saline solution was instilled and recovered through the biopsy channel of the scope. The recovery rate of lavaging saline solution was up to 75–80%. The lymphocytes in BAL fluid were collected by Ficoll gradient centrifugation. Total numbers of cells in BAL fluid were determined by microscope using trypan blue and adjusted in 60 ml of BAL fluid (2–30 × 10⁶ cells).

Fractionation of lymphocyte populations from blood, lymph nodes and BAL fluid

To characterize a single T cell population for BCG specific responses, CD4⁺ and CD8⁺ T cell were selected negatively by immunomagnetic beads for ELISPOT assays [45]. CD4⁺ T cells were enriched using anti-CD8 antibody-conjugated Dynabeads (Dynal, Inc., Great Neck, NY, USA) and CD8⁺ PBL by means of anti-CD4 antibody-beads. Peripheral blood mononuclear cells (PBMC) or lymph node cells were incubated with these immunomagnetic beads for 30 min at room temperature, and then selected in two cycles with a magnetic particle concentrator. These procedures were repeated for an additional time to increase the purity of CD4 or CD8 enriched PBL. CD4⁺ or CD8⁺ T cells enriched by these methods contained less than 5% CD8⁺ or CD4⁺ T lymphocytes.

Monoclonal antibodies (MoAbs) and flow cytometric analysis

The MoAbs used were as follows: pan anti-TCR Cδ (anti-TCRδ1) [46], antihuman Vγ2 (7A5), antihuman Vδ1Jδ1/Jδ2 (Ts8), antihuman Vδ2 (15D) (all from Pierce, Rockford, IL, USA), fluoroisothyocyanate (FITC)-conjugated antihuman CD3 (Pharmigen, CA, USA), phycoerythrin (PE)-conjugated antirhesus monkey CD3 (FN18, Biosource, Camarillo, CA, USA), PE-conjugated antihuman CD4 (Ortho Diagnostic Systems, Raritan, NJ, USA), Cy5-conjugated antihuman CD8 (Dako Corporation, Carpinteria, CA, USA) and FITC-conjugated antihuman CD20 (all from Becton Dickinson, CA, USA). Staining panels were as follows: CD3/CD4/CD8; Cδ/CD3/CD8; CD3/CD20; Vδ1/CD3; Vγ2/Vδ2. Two- or three-colour flow cytometric analyses were performed on an XL flow cytometer (Coulter, Hialeah, FL, USA). Lymphocytes were gated by means of forward and side scatter, and up to 20 000 gated cells were analysed. Absolute numbers of T cell populations in the blood were calculated based on the flow data and complete blood count (CBC) analyses. CBCs were performed on a Coulter T 540 haematological analyser (Coulter, Hialeah, FL, USA).

Proliferation assay

A conventional proliferation assay was carried out as described previously [10]. Briefly, unfractionated macaque PBL (1 × 10⁵ cells per well) enriched CD4⁺ or CD8⁺ T cells were cultured in triplicate in 96-well plates in the presence of BCG PPD (1, 5 or 25 µg/ml), Con A (5 µg/ml), bovine serum albumin (BSA 3 µg/ml) or medium alone. Five days later, cells were pulsed with [³H]-thymidine at 1·0 µCi per well and uptake was measured 8 h later using a 1450 Microcbeta scintillation counter (Wallac, Gaithersburg, MD, USA). Stimulation index was defined as the ratio of the mean CPM of PPD- or concanavalin-A (Con-A)-stimulated wells relative to the mean CPM of control wells (medium alone).

ELISPOT assays

To estimate the numbers and IFN-γ production capacity of mycobacteria-specific T cells, PBL were assessed for their specific recognition of PPD antigens using ELISPOT assays. In these experiments, 96-well nitrocellulose-backed plates were coated with 1 µg antihuman IFN-γ (B27) by overnight incubation at room temperature; 2 × 10⁵ cells were plated in duplicate in each well with or without those BCG antigens. Eighteen h after culture, the plates were washed and incubated for 2 h at room temperature with 500 ng of the biotinylated anti-IFN-γ. Spots were developed by adding HRP-streptavidin and freshly prepared substrate. The numbers of ELISPOT-forming cells in response to the BCG antigens were calculated and verified by PBL counts. To estimate the relative number of IFN-γ-producing CD4⁺ T cells in response to PPD stimulation, PBL depleted of CD8⁺ cells were used for ELISPOT assays. Similarly, PPD-specific, IFN-γ-producing CD8⁺ T cells were analysed by ELISPOT using PBL depleted of CD4⁺ T cells.

In vitro stimulation and expansion of γδ T cells with non-peptide phosphoantigens from M. fortuitum and BCG

This was performed as described previously [26]. PBL obtained from macaques were stimulated in culture medium with or without partially purified phosphoantigens (prenyl pyrophosphoate) (proposed structure 2-methyl-4-hydroxybutanal-pyrophosphate; also termed 3-formyl-1-butyl-pyrophosphate) from M. fortuitum or purified BCG phosphoantigen. On day 3 IL-2 was added and on days 10–12 the cells were counted and analysed by flow cytometry using γδ TCR-specific MoAbs as described above. BCG phosphoantigens were prepared from 1 g of BCG using CHl₃/MeOH-based extraction followed by chromatographic isolation, as described previously [26]. A 1 : 1000 dilution of M. fortuitum supernatant and 5 functional units of BCG phosphoantigen/ml medium were used for in vitro stimulation (one functional unit is the dose that results in detectable expansion of Vγ2Vδ2⁺ T cells in 1 ml culture containing 10⁶ PBL). In control experiments, purified BCG phosphoantigen stimulate a specific expansion of Vγ2Vδ2⁺ T cells but not αβ T cells in PBL from BCG-infected macaques.

BCG colony counts for the quantification of mycobacteria

The viable BCG counts in the blood and BAL were calculated by the quantification of BCG CFUs in cell lysates from the blood and BAL cells, as described previously [11]. The red blood cell-depleted cell pellets from 0·5 ml of blood were lysed with 10% saponin to release the intracellular BCG. Fivefold dilutions of the lysate were plated in duplicate on Middlebrook 7H10 agar plates (Difco, Detroit, MI, USA). The CFUs were counted after a 3-week incubation at 37°C.

Statistical analysis

Student's t-test and non-parametric tests, as described previously [11], were employed to examine whether any differences in numbers of T cell populations, or BCG loads identified after BCG infection and reinfection were statistically significant. In addition, a correlation coefficient was calculated with Prism [11] to determine the correlation between changes in BCG loads and numbers or function of T cell populations.

RESULTS

Primary BCG infection induced initial lymphocytopenia and subsequent expansions of T cell populations in the blood and lymph nodes of macaques

To evaluate global immune responses during mycobacterial infection, macaques were infected intravenously and bronchially with different sizes of BCG inocula. Primary BCG infection induced biphasic responses of lymphocyte populations in the blood and lymph nodes of the macaques. The initial response was characterized by profound lymphocytopenia in the BCG-infected macaques. A major decline of CD4⁺ T cells was identified in the macaques inoculated with different sizes of BCG inocula, although the down-regulation also involved CD8⁺ T cells, γδ, B and CD8⁺ non-T lymphocyte populations in the blood. A one- to two-fold decline of CD3⁺, CD4⁺ and CD8⁺ T cells was noted in the blood. The down-regulation of those lymphocyte populations lasted for 2–3 weeks after BCG inoculation (Fig. 1a). Following the initial down-regulation, these lymphocyte populations recovered gradually to baseline levels or expanded one- to two-fold in response to the primary BCG infection (Fig. 1a). While no consistent increase in numbers of B and non-T lymphocytes was seen, T lymphocyte populations exhibited a dose-dependent expansion in the blood after BCG inoculation (Fig. 1a). The systemic BCG infection introduced by intravenous inoculation with 10⁸ BCG organisms resulted in about a two-fold increase in CD4⁺ and CD8⁺ T cells (means ± s.e.m. of CD4⁺ from 1486 ± 75 to 2980 ± 880; CD8⁺ from 923·5 ± 116·6 to 1525 ± 163·2) at 5 weeks and an up to five-fold expansion of γδ T cells (from 70 ± 21 to 423 ± 92) in the blood (Fig. 1a,Fig. 1b). In contrast, a systemic BCG infection generated by 10³ BCG organisms or a pulmonary infection by bronchoscope-guided inoculation induced only subtle changes in numbers of circulating T cells (Fig. 1a). The expanded T cell populations returned to baseline levels approximately 6–8 weeks after the BCG inoculation (Fig. 1a). Similar to what was seen in the circulating lymphocytes, the systemic BCG infection induced noticeable changes in the lymphocyte populations in the lymph nodes. Expansion of CD4⁺ and CD8⁺ T cells but not γδ T cells was noted in the lymph nodes of the macaques infected with BCG through the intravenous but not pulmonary route (Fig. 1a). The biphasic alteration in the numbers of lymphocyte populations was induced by the active BCG infection, because the control infection with non-pathogenic M. fortuitum did not result in significant changes in numbers of lymphocyte populations (Fig. 1a). These results therefore demonstrated that primary BCG infection induced initial lymphocytopenia and subsequent expansion of T cell populations in macaques.

Fig. 1.

Primary BCG infection induced dose-dependent immune responses, characterized by initial down-regulation of lymphocyte populations and subsequent expansion of T cells in the blood (PBL) and lymph nodes (LN) of the macaques. (a) Absolute numbers of lymphocytes in the blood and lymph nodes after intravenous or bronchial BCG inoculation. The sizes of BCG inocula are indicated at the top panel of the figure. Four different groups, four macaques in each group, received the BCG challenge with different sizes of BCG inocula. The control group (two monkeys) was inoculated with 10⁶ CFU of M. fortuitum.

Fig. 1.

Continued (b) Changes in the percentage of CD4⁺, CD8⁺ and γδ T cells in CD3⁺ T cell population in the blood and LN of BCG-infected monkeys. Note that intravenous BCG inoculation favoured the accumulation of γδ T cells; iv.: for intravenous BCG inoculation; bi: for bronchoscope-guided inoculation. Shown are mean values with the error bars of s.e.m. P < 0·05: CD3, CD4 and CD8 T cells in week 5 of the group inoculated with 10⁸ BCG; CD3, CD4 and CD8 LN cell-expansion in week 4; P < 0·01: γδ T cell expansion in weeks 2, 3 and 4 after inoculation with 10⁶ or 10⁸ BCG.

Pulmonary immune responses of T cell populations were induced by both bronchial and systemic challenges with BCG

We then sought to examine lymphocyte population responses in the lung in the settings of systemic and pulmonary BCG infection. The pulmonary BCG infection following BCG inoculation through the bronchial route resulted in increases in numbers of CD4⁺, CD8⁺ and γδ T cells in the BAL fluid (Fig. 2a), although this pulmonary challenge did not induce an apparent expansion of lymphocyte populations in the blood (Fig. 1a). An expansion of these T cells populations was seen 4–5 weeks after pulmonary BCG infection introduced by bronchoscope-guided inoculation (Figs 2a,b). Interestingly, systemic BCG infection generated by intravenous BCG inoculation induced increases in numbers of T cell populations in the pulmonary compartment. The expansion of pulmonary T cell populations was detected 2–5 weeks after intravenous BCG inoculation. The expansion of γδ T cell population was most evident in the BAL fluid collected from the macaques inoculated intravenously and bronchially with BCG. Up to 10-fold expansions of γδ T cell population was noted in the BAL fluid (65750 ± 17220 versus 788800 ± 199800), whereas a less than twofold increase in numbers of CD4⁺ (115800 ± 301600 versus 1589000 ± 425500) T cells was seen (Fig. 2). No major expansion of B and CD8⁺ non-T cells was seen the BAL fluid from the BCG-infected macaques (Fig. 2). These results therefore demonstrate that the pulmonary compartment participates actively in immune responses of T cell populations in the settings of both systemic and local infections with mycobacteria.

Fig. 2.

Primary BCG infection introduced through both intravenous and bronchial routes induced T cell expansion in the pulmonary compartment of the infected macaques. (a) Absolute numbers of lymphocytes in BAL fluid following intravenous or bronchoscope-guided inoculation with BCG. The numbers of animals in each group were the same as described in the legend of Fig. 1a. (b) Changes in the percentage of CD4⁺, CD8⁺ and γδ T cells in the CD3⁺ T cell population in BAL fluid after intravenous and bronchoscope-guided inoculation of BCG (10⁶ CFU). Note that intravenous BCG inoculation appeared to favour the accumulation of γδ T cells in BAL fluid. P < 0·05: γδ T cells in weeks 3 and 4 after intravenous BCG inoculation; all T cell populations in week 5 after bronchoscope-guided inoculation.

BCG reinfection induced potent recall responses of T lymphocyte populations in macaques

Because mycobacterial diseases can occur as a result of reinfection, it was important to determine the recall immune responses of T cell populations during mycobacterial reinfection. To this end, we made use of BCG reinfection as a model to examine the recall or memory immune responses of T cells in the macaques reinfected with BCG. BCG reinfection through the intravenous route induced marked recall responses of T cell populations but not B or non-T cells in the blood and lymph nodes. The recall expansion of T cells was evident 1 week after the second BCG inoculation; the magnitude and duration of the expansion were greater than the primary responses seen after the first BCG inoculation (Figs 1a and 3). As much as a three-fold increase in CD4⁺ or CD8⁺ T cells was seen in the blood after the second intravenous inoculation of BCG (Fig. 3). Major expansions of total γδ T cells were identified in the blood and BAL fluid following the BCG reinfection by the intravenous route. Similarly, pulmonary BCG reinfection introduced by bronchoscope-guided inoculation induced potent recall immune responses of T cell populations in the lung, although there was only a subtle expansion of circulating T cell populations after the pulmonary reinfection. More than a 10-fold expansion of CD4⁺ and CD8⁺ and γδ T cells was noted for the peak recall responses in the BAL fluid collected from the BCG-inoculated lung (Fig. 3). The marked recall responses of these T cells persisted for more than 4 weeks in the BAL fluid from the lung reinfected with BCG (Fig. 3). These results demonstrated the recall or memory immune responses of T cell populations during the systemic and pulmonary reinfection of macaques with a mycobacterium.

Fig. 3.

BCG reinfection induced a marked recall expansion of T cell populations in the blood and pulmonary compartment. Shown are absolute numbers of T cells in the blood and BAL fluid. Two groups, four monkeys for each, were inoculated again either intravenously or intrabronchially with 10⁸ BCG organisms, 3–5 months after the first BCG inoculation. P < 0·01 for all T cell populations in blood and BAL fluid 1–8 weeks after intravenous or intrabronchial inoculation with BCG.

Antigen-specific T cells were identified at the time when the in vivo expansion of T cell populations was seen in the BCG-infected macaques

We then sought to determine whether the T cell populations that expanded in vivo were able to recognize mycobacterial antigens in vitro. Because human T cells from PPD⁺ individuals have been shown to recognize mycobacterial peptides and phosphoantigens [24,27,47,48], these antigens were used for the in vitro stimulation of macaque T cell populations. T cell recognition and proliferation in response to BCG PPD stimulation were detected readily in PBL 3–4 weeks after the intravenous and pulmonary BCG inoculations (Fig. 4a). In the pulmonary compartment, PPD-specific T cell responses were also detected, but delayed for 1–2 weeks in terms of the magnitude of proliferation (Fig. 4a). Consistently, a large number of PPD-specific, IFN-γ-producing T cells was detected in PBL of the monkeys during the primary infection and reinfection with BCG (Fig. 4b). Finally, the antigen specificity of γδ T cells could be identified at the time in vivo expansion of those cells was seen during BCG infection. The BCG phosphoantigens stimulated a remarkable expansion of total γδ T cell population (Fig. 4c). These functional experiments suggest therefore that the in vivo expansion of T cell populations was indeed driven by the BCG antigen-specific stimulation.

Fig. 4.

BCG antigen-specific T cells were detected during the in vivo expansion of T cell populations in the BCG-infected macaques. The first and second BCG inoculations were indicated. The doses for the first and second inoculations were 10⁶ and 10⁸ CFU, respectively. The legends for inoculation routes were same as those described in those of Fig. 1a. (a) Proliferative responses of PPD (5 µg/ml)-specific T cells in the blood and BAL fluid after the first and second BCG inoculation. (b) PPD-specific, IFN-γ-producing T cells in unfractionated (left) and fractionated (right) PBMC using ELISPOT assays. The ELISPOT-forming cells shown in each time point are the numbers derived from subtracting the values seen in negative controls by the original ones. The data shown in (a,b) were collected from four animals in each group and shown as mean values with the error bars of s.e.m. The second BCG inoculation was performed 5 months after the first BCG infection. (c) BCG phosphoantigen (5 functional units/ml) stimulated in vitro expansion of γδ T cells. Data were generated using PBL obtained from four monkeys 4 weeks after BCG-inoculation. P≤ 0·01: all data at time points after week 4 in the first BCG infection; all data at time-points after week 2 in the second BCG infection.

Global expansion of T cell populations coincided with a decrease in the numbers of detectable BCG organisms

Finally, we sought to examine if the expanded T cell populations contributed to the antimycobacterial immunity. To test this presumption, changes in BCG CFU counts during the primary infection and reinfection were measured and assessed for their temporal correlation in time course with the expansion of T cell populations. The decline of detectable BCG CFU counts in the lymph node cells coincided with the initial expansion of CD3⁺ T cells comprised of CD4⁺, CD8⁺ and γδ T cells in the blood of the macaques inoculated intravenously with BCG (Fig. 5). Similarly, the decline of detectable BCG in the lung coincided with the occurrence of pulmonary T cell expansion and PPD-specific proliferation (Fig. 5). The decline of detectable BCG CFU counts appeared to be associated more closely with the recall expansion of T cell populations following the BCG reinfection (Fig. 5). The association of emerging T cell responses with the decreasing trend of active BCG infection implied that the expanded T cell populations contributed to the containment of active BCG infection.

Fig. 5.

The T cell expansion was associated with the decreasing trend of active BCG infection. Shown are the kinetics of BCG CFUs in LN cells or BAL cells and CD3⁺ T cell expansion in the blood (left) and BAL fluid (right) following the first and second BCG inoculations. Data shown are the means of values with the error bars of s.e.m. from four infected animals of each group.

DISCUSSION

In the present studies, we have found profound down-regulation of all macaque lymphocyte populations prior to their expansions during the early phase of BCG infection. The down-regulation of lymphocytes can last for as long as 2–3 weeks after the primary systemic or pulmonary BCG infection. Transient down-regulation of lymphocytes is also seen after the BCG reinfection. The CD4⁺ lymphocytopenia is most striking among the all cell populations, as CD4⁺ T cells constitute the major lymphocytes in the blood and lymph nodes. In fact, clinical case studies have reported that the patients with active M. tuberculosis infection can exhibit CD4⁺ lymphocytopenia [49]. The CD4⁺ lymphocytopenia in the tuberculous patients can be reversed after antituberculosis treatment [50]. Our studies have demonstrated that down-regulated responses in acute mycobacterial infection involves not only CD4⁺ T cells but also other lymphocyte populations in BCG-infected macaques. Moreover, the down-regulation of lymphocytes is associated with the active stage of the BCG infection and reinfection as well as transient undetectable proliferative responses of T cells after the second BCG inoculation. It is likely that high levels of inflammatory cytokines produced during active BCG or M. tuberculosis infection contribute to the down-regulation of lymphocyte populations.

Our data provide in vivo evidence that CD8⁺ T cells are able not only to mount dose-dependent immune responses during primary BCG infection, but also contribute greatly to recall or memory immune responses in the reinfection. The longitudinal studies of CD8⁺ T cell immune responses in acute M. tuberculosis or BCG infection have not been described well in humans. Using the macaque animal model, we have identified adaptive immune responses of CD8⁺ T cells during mycobacterial infection. While a dose-dependent increase in numbers of CD8⁺ T cells is seen during primary BCG infection, a recall expansion of CD8⁺ T cell population is striking in the macaques reinfected with BCG. M. tuberculosis-specific CD8⁺ T cells have been characterized in vitro in the blood of PPD⁺ people and TB patients [18–25]. The demonstration of CD8⁺ T cell expansion during mycobacterial infection in monkeys suggest that characterizing the fine specificity and immune function of these responding T cells would be useful.

Interestingly, an expansion of the γδ T cell population appears to occur earlier in time and to a greater magnitude than that of CD4⁺ and CD8⁺ T cells. The γδ T cell expansion is detectable 2–3 weeks after the first intravenous BCG inoculation, the time when the numbers of CD4⁺ and CD8⁺ T cells usually remain low or rebound to baseline levels. The magnitude of γδ T cell expansion is 2–5 times greater than that of CD4⁺ or CD8⁺ T cell population during the primary infection and reinfection with BCG, although the pure increase in total cell numbers of CD4⁺ and CD8⁺ T cell populations is more striking. The expanded γδ T cells are comprised predominantly of a single cell subpopulation that expresses a Vγ2Vδ2 TCR heterodimer [11]. It is likely that the Vγ2Vδ2⁺ T cell subpopulation is stimulated by a single BCG phosphoantigen. Such a magnitude of γδ T cell expansion is remarkable when compared with a single peptide-specific CD4⁺ or CD8⁺ T cell subpopulation. The frequencies of PPD-specific IFN-γ-producing αβ T cells were less than 0·5% of PBL in BCG-infected monkeys (Fig. 4). The fact that γδ T cells expand faster and to a greater extent than αβ TCR lymphocytes may reflect the unique features of these cells in immune response to mycobacterial antigens. In vitro studies have demonstrated that the recognition of non-peptide phosphoantigens by Vγ2Vδ2 T cells does not require the intracellular processing or restriction by MHC prior to antigen presentation [27]. Bypassing a requirement for antigen processing and MHC restriction may help to explain why the in vivo expansion of the γδ T cells occurs so rapidly after infection. We certainly cannot exclude the possibility that the cytokines produced in the initial infection favour the proliferation and clonal expansion of γδ T cells when compared with the αβ T cells.

In addition to the timing and challenge doses, the route of BCG infection impacts on the global and local immune responses. The pulmonary BCG infection introduced through the bronchial route does not induce significant expansion of T cell populations in the blood, although the peripheral T cells, after the pulmonary BCG challenge, clearly acquire the ability to proliferate and expand in vitro in response to PPD or phosphoantigen stimulation. The major BCG infection is localized in the lung after pulmonary challenge with BCG, as no or few viable BCG organisms can be detected in the blood and lymph nodes after bronchial BCG inoculation (data not shown). Interestingly, while the intravenous BCG inoculation induces subtle changes in the numbers of CD4⁺ and CD8⁺ T cells in the lung, a marked expansion of γδ T cells can be identified in the pulmonary compartment during systemic infection and reinfection with BCG. The explanation and significance for the selected expansion of pulmonary γδ but not CD4⁺ or CD8⁺ T cells during systemic BCG infection are currently not apparent. It is likely that chemokine receptor-mediated trafficking in response to chemokines is responsible for transendothelial migration of immune cells to the lung from the circulation and/or lymphoid tissues. In fact, it has recently been shown that human Vδ2⁺, but not Vδ1⁺ or αβ TCR⁺, T cells express high levels of CCR5 and CXCR3 as well as related C-C chemokines [51,52]. Furthermore MIP-1α, MIP-1β and RANTES, the ligands for CCR5, have been shown to facilitate γδ T cell migration in an in vitro migration system [53]. These chemokines may have roles in transendothelial chemotaxis of Vγ2Vδ2⁺ T cells into the lungs or other organs [54].

Current studies provide in vivo data suggesting that CD4⁺, CD8⁺ and γδ T cells all contribute to adaptive immunity against mycobacterial infection. The major expansion of those T cells is associated with the challenge doses of BCG organisms for active BCG infection and reinfection in the macaques. Such correlations support the possibility that the expanded T cells are involved in the protective immunity against BCG infection, although the results do not provide direct cause–effect evidence. Importantly, our studies demonstrate that pulmonary T cells exhibit characteristics of adaptive immune responses. Although pulmonary T cell populations increase only at moderate levels during primary pulmonary BCG infection, BCG reinfection in the lung drives a remarkable recall increase in numbers of CD4⁺, CD8⁺ and γδ T cells in the pulmonary compartment. The results from pulmonary BCG infection imply that a recall expansion of pulmonary T cell populations can be localized predominantly in the pulmonary compartment after mycobacterial reinfection through the same bronchial route. Such faster and greater recall responses of T cell populations may provide an advantage in mounting an adaptive immune response during mycobacterial reinfection.

Thus, our studies imply that the ability of CD4⁺, CD8⁺ and γδ T cells to mount faster and stronger recall immune responses may be important mechanisms underlying the immune protection against mycobacterial infection. These results also suggest that vaccine efforts might be focused on maximizing the adaptive immune responses of T cells. The patterns and kinetics of CD4⁺, CD8⁺ and γδ T cell immune responses during BCG infection may facilitate exploring immune correlates of protection against virulent M. tuberculosis infection and evaluating new tuberculosis vaccines in primates.