Abstract
The possibility that CD4+ T cells can act as “innate-like” cells to contain very early Mycobacterium tuberculosis dissemination and function as master helpers to sustain multiple effector functions of CD8+ T cells and CD3− lymphocytes during development of adaptive immunity against primary tuberculosis (TB) has not been demonstrated. We showed that pulmonary M. tuberculosis infection of CD4-depleted macaques surprisingly led to very early extrapulmonary M. tuberculosis dissemination, whereas CD4 deficiency clearly resulted in rapid TB progression. CD4 depletion during M. tuberculosis infection revealed the ability of CD8+ T cells to compensate and rapidly differentiate to Th17-like/Th1-like and cytotoxic-like effectors, but these effector functions were subsequently unsustainable due to CD4 deficiency. Whereas CD3− non–T lymphocytes in the presence of CD4+ T cells developed predominant Th22-like and NK-like (perforin production) responses to M. tuberculosis infection, CD4 depletion abrogated these Th22-/NK-like effector functions and favored IL-17 production by CD3− lymphocytes. CD4-depleted macaques exhibited no or few pulmonary T effector cells constitutively producing IFN-γ, TNF-α, IL-17, IL-22, and perforin at the endpoint of more severe TB, but they presented pulmonary IL-4+ T effectors. TB granulomas in CD4-depleted macaques contained fewer IL-22+ and perforin+ cells despite the presence of IL-17+ and IL-4+ cells. These results implicate a previously unknown innate-like ability of CD4+ T cells to contain extrapulmonary M. tuberculosis dissemination at very early stage. Data also suggest that CD4+ T cells are required to sustain multiple effector functions of CD8+ T cells and CD3− lymphocytes and to prevent rapid TB progression during M. tuberculosis infection of nonhuman primates.
Introduction
Tuberculosis (TB) remains one of the leading causes of morbidity and mortality among infectious diseases worldwide, and it has become increasingly prevalent and deadly as a result of the HIV/AIDS pandemic and the emergence of multidrug resistant/extensively multidrug resistant strains of Mycobacterium tuberculosis (1, 2). There is a pressing need to elucidate essential components of anti-TB immunity and develop better TB vaccines and immunotherapeutics for global TB control because bacillus Calmette–Guérin, the current TB vaccine, inconsistently protects against TB in adults (3). Although most people can develop immunity against active TB after exposure to M. tuberculosis (4), little is known about how immune cells control primary M. tuberculosis infection in humans. Although HIV infection increases susceptibility to TB (5–8), in-depth studies are needed to determine whether the increased susceptibility is attributed to HIV-induced immune suppression and/or CD4+ T cell decline. Essentially, clinical studies have not shown sound correlation between human Th1 cells/cytokines and anti-TB immunity (9, 10), despite that murine IFN-γ and TNF-α play a role in protection against M. tuberculosis infection (11–13). Mechanistic studies in animal TB models may help to elucidate precise elements of anti-TB immunity and immune mechanisms for protection in humans.
Studies in mice indicate that CD4+ T cells are important for immunity against M. tuberculosis infection (14–22). The most established role for CD4+ T cells in immunity against TB is to evolve into Th1 effector cells and produce IFN-γ/TNF-α to directly activate macrophage for controlling infection or killing M. tuberculosis–infected macrophages (23). Whereas multifunctional CD4+ T cells simultaneously producing IFN-γ/TNF-α/IL-2 are detectable in M. tuberculosis infection (24, 25), CD4+ T cells can also function as cytotoxic cells and produce anti–M. tuberculosis perforin and granulysin in vitro. Murine CD4+ T cells can directly kill M. tuberculosis–infected cells in vivo, although the mechanisms remain unclear (26, 27). The role of IL-17– and IL-22–producing CD4+ T cells has been investigated in animal models and whether these cells are protective or detrimental needs to be further investigated (28–30). Whereas murine IL-17 produced by CD4+ T cells helps to recruit IFN-γ–producing cells and to control M. tuberculosis infection (29), Th17 cells could correlate with either anti–M. tuberculosis response or TB severity (31–33). IL-22–producing cells can limit intracellular M. tuberculosis growth in vitro (30), but their role in M. tuberculosis–infected individuals needs to be further investigated. The role of IL-4–producing Th2 cells is elusive in TB, although an increase in IL-4–producing T cells is correlated with active TB patients (34). Alternatively, CD4+ T cells can indirectly contribute to immune control of M. tuberculosis infection by helping other immune components. It has been reported that CD4+ T cells are required for cytotoxic CD8+ T cell development in vivo in mice (25).
Despite direct and indirect roles of CD4+ T cells, however, precise mechanisms by which CD4+ T cells mediate immunity to primary M. tuberculosis infection remain incompletely understood. Mechanistic studies in CD4-deficient mice were mainly focused on changes in production of Th1 cytokines, which were measured after in vitro potent panstimulation of all T cells by CD3/CD28 Abs (14, 21, 25). It was found that IFN-γ production in M. tuberculosis infection was comparable between wild-type and CD4-deficient mice (14, 21, 25). Although CTL precursors in CD4 knockout mice were evaluated after M. tuberculosis restimulation in culture or triggered by CD3/CD28 Ab (25), how CD4+ T cells impacted initial and subsequent effector functions for producing perforin/IFN-γ/IL-17/IL-22 by CD8+ T cells during M. tuberculosis infection was not directly evaluated in vivo. Moreover, most studies of CD4+ T cell functions in mice and HIV-infected humans were focused on the endpoint pathology or advanced stage of M. tuberculosis infection (14, 21, 35), and little is known about whether early activation of CD4+ T cells after pulmonary M. tuberculosis exposure can help contain M. tuberculosis replication or dissemination. Thus, further studies are needed to address how early CD4+ T cell immunity against M. tuberculosis occurs and how CD4+ T cells induce subsequent adaptive immunity against primary M. tuberculosis infection directly and indirectly.
Given the possibility that CD4+ T cells have broad immune functions, including hypothetical “innate-like” defense in M. tuberculosis infection, we hypothesize that CD4+ T cells after M. tuberculosis exposure can rapidly act as innate-like cells to contain early pulmonary M. tuberculosis replication/infection, whereas subsequent development of an adaptive CD4+ T cell response is of central importance for mediating immunity against TB. We also hypothesize that CD4+ T cells may serve as master helper cells to promote and sustain systemic and pulmonary anti-TB responses of CD8+ T cells and non–T lymphocytes in direct and indirect fashions. To test these hypotheses, we examined whether in vivo depletion of CD4+ T cells in macaques compromised early innate-like defense and subsequent adaptive immunity against TB, and whether CD4 depletion impacted development and pulmonary recruitment of CD8+ CTLs, NK-like, and other Th-like effector cells. We demonstrated that depletion of CD4+ T cells led to very early extrapulmonary M. tuberculosis dissemination and subsequently rapid TB progression after pulmonary M. tuberculosis infection of macaques. We also found that CD4+ T cell deficiency abrogated the ability of the immune system to sustain systemic and pulmonary immune responses of CD8+ T effector and non–T (innate) effector populations constitutively producing Th1, Th17, Th22, and CTL cytokines during M. tuberculosis infection. The findings suggest that CD4+ T cells possess previously unappreciated an early innate-like defense for containing M. tuberculosis replication, suggesting that CD4+ T cells are needed to sustain multieffector cellular responses and to mount adaptive immunity against TB. Data also implicate potential mechanisms relevant to CD4+ T cell–mediated protection against primary M. tuberculosis infection.
Materials and Methods
Animals
Twelve adult cynomolgus macaques, 3–9 y old, were used in CD4+ T cell depletion during M. tuberculosis infection. Another nine cynomolgus macaques with similar ages served as controls receiving saline treatment without CD4+ T cell depletion during M. tuberculosis infection. All macaques were naive prior to M. tuberculosis infection, with negative results for tuberculin skin tests, IFN-γ ELISOPOT assays, and thoracic radiographs. M. tuberculosis–infected macaques were housed at the Biologic Research Resources Annex BSL3 nonhuman primate facilities at the University of Illinois at Chicago. Daily or weekly clinical follow-up were taken to ensure that animals were not suffering from severe coughing, respiratory distress, depression, refusing to take food, body weight loss, or other potential life-threatening signs. Humane euthanization procedures were immediately taken when those signs occurred progressively. M. tuberculosis–infected macaques were sacrificed at 9 wk postinfection as we previously did (36, 37). Macaques treated with depleting CD4 Ab exhibited progressive clinical deteriorations and became moribund at ∼7 wk after M. tuberculosis infection, requiring humane euthanization. Bursting bacillus burdens and more severe TB lesions seen at necropsy in these CD4-depleted macaques suggested a rapid progression of TB after M. tuberculosis infection. All animal experimental procedures and protocols were approved by the University of Illinois Chicago Animal Care Committee.
M. tuberculosis bacilli and pulmonary M. tuberculosis infection
M. tuberculosis strain Erdman stocks (3.5 ± 1.5 × 108 CFU/ml) were provided by Dr. Bill Jacobs at the Albert Einstein College of Medicine and Dr. Mike Brennan at the U.S. Food and Drug Administration. An aliquot was diluted to the desired concentration for infection in sterile saline. Each monkey was infected with 500 CFU M. tuberculosis Erdman by the bronchoscope-guided spread of the inoculum into the right caudal lobe as previously described (36, 38, 39). The inoculum used for infection was diluted and plated on 7H11 agar plates (BD Biosciences) to further confirm the bacterial CFU dose for inoculation.
Bronchoalveolar lavage
This was done essentially the same as previously described (36–39). Briefly, after overnight fasting, animals were tranquilized i.m. with 1-2 mg/kg xylazine (Ben Venue Laboratories, Bedford, OH) and 10 mg/kg ketamine HCl, and then given 0.05 mg/kg atropine (Phoenix Scientific, St. Joseph, MO) i.m. as an anticholinergic for bronchoalveolar lavage (BAL) while being restrained in an upright position. A pediatric feeding tube was inserted down into the trachea through direct visualization with a laryngoscope and further into the right or left bronchus at the level of the carina. Saline (10 ml) was instilled into the bronchus and immediately withdrawn and repeated a maximum of three times until a total of 12–15 ml BAL fluid (BALF) was retrieved. This procedure generally gave rise to fluid compositions and volumes that were comparable to the bronchoscope-guided BAL (data not shown).
Administration of depleting anti-CD4 Ab for CD4+ T cell depletion
Depleting anti-CD4+ mAb (CDR-OKT4A-human IgG1, provided by Dr. Keith Reimann, Biovest International) was supplied as sterile solution in borosilicate glass bottles at a concentration of 10.5 mg/ml with 1000 mg and 200 mg in two different package sizes in PBS with 0.02% polysorbate 80 final buffer (pH 6.5). The CD4-depleting Ab is CDR-OKT4A-human IgG1. It is an Ab that combines the CDR of murine OKT4A (not OKT4) with the V framework and the constant regions of human κ L chain and IgG1 H chain (40). OKT4A is different from OKT4, as OKT4A binds to an epitope on the CD4 molecule distinct from that recognized by OKT4, and these Abs do not cross-block each other (40, 41). Moreover, OKT4A is IgG1 and OKT4 is IgG2. To determine CD4 depletion efficacy, we used the clone L200 Ab for surface CD4 staining before and after treatment with OKT4A CD4-depleting Ab. Depleting anti-CD4+ Ab was administered i.v. at 50 mg/kg (5 ml/kg) without further dilution at days −5, 9, and 24 after M. tuberculosis infection, respectively. The doses were calculated on macaque body weight every time, held in sterile plastic syringes, and injected slowly over a period of at least 10 min. The administration of depleting Ab led to almost complete depletion of CD4+ T cells in vivo during M. tuberculosis infection (36, 40, 42, 43).
Isolation of lymphocytes from blood, BALF, and lung tissues
This procedure was done exactly the same as described before (38). Briefly, PBMCs were isolated from freshly collected EDTA blood by Ficoll-PaquePlus (Amersham Biosciences, Piscataway, NJ) density gradient centrifugation. Fresh BALF was filtered through 40-μm cell strainers (BD Biosciences) followed by 5 min at 1500 rpm centrifugation. Cell pellets were treated with 5 ml RBC blood lysis buffer (Sigma-Aldrich) for 10 min or waiting until the suspension became clear and then washed once with 5% FBS-PBS. Lung tissues were minced with sharp scissors and squeezed with sterile copper mesh in Petri dish, and suspensions were filtered through 40-μm cell strainers and further purified by Ficoll-PaquePlus density gradient centrifugation
Antibodies
Direct intracellular cytokine staining assay to measure T effector cells and CD3− cells capable of constitutively producing various cytokines without in vitro Ag stimulation
This procedure was done exactly the same as we previously described (36, 38, 44). Briefly, 106 PBMCs or 106 or 3–5 ×105 lymphocytes (depending on availability) from BALF or lung tissues were used in each reaction (round-bottom 96-well plate) to measure T cells and CD3− lymphocytes that could constitutively produce IFN-γ, TNF-α, IL-17, IL-22, IL-4, and perforin without Ag stimulation in vitro. Lymphocytes were incubated for 1 h with medium in the presence of CD28 (1 μg/ml) and CD49d (1 μg/ml) mAbs in a 200 μl final volume in round-bottom 96-well plates at 37°C, 5% CO2, followed by a 5-h incubation in the presence of brefeldin A (GolgiPlug; BD Biosciences). After a total of 6 h of incubation, cells in 96-well plates were transferred into 5 ml polystyrene round-bottom tubes (BD Biosciences) for surface and intracellular staining. Cells were washed once with 2% FBS-PBS and stained at room temperature for at least 15–30 min with surface marker Abs (CD3, CD4, and CD8) and washed twice with 2% FBS-PBS. Cells were permeabilized for 45 min (Cytofix/Cytoperm; BD Biosciences) and washed twice by Perm buffer (BD Biosciences) and then stained another 45 min for IFN-γ Ab, TNF-α Ab, IL-17 Ab, IL-4 Ab, biotinylated IL-22, or biotinylated perforin and repeated with Perm wash twice. Cells stained with biotinylated IL-22 or biotinylated perforin were further stained for 45 min with streptavidin–Pacific Blue conjugate followed by final twice Perm buffer wash. Finally, cells were resuspended in 2% formaldehyde-PBS (Protocol Formalin, Kalamazoo, MI) and subjected to flow cytometry analysis. To ensure the specific immune staining in direct or indirect intracellular cytokine staining (ICS), matched normal serum or isotype IgG served as negative controls for staining cytokines or surface markers. As another control setting, PBLs were collected weekly or biweekly for up to 3 mo from two uninfected healthy macaques and four simian HIV–infected macaques, and we assessed de novo production of IFN-γ, IL-17, and IL-22 over time using the direct ICS approach without Ag stimulation in vitro. T cells from these controls gave rise to no or very low levels of cytokines stained by this direct ICS method (36, 38, 44).
Flow cytometry analysis
Fixed lymphocytes were run on a CyAn ADP flow cytometer (DakoCytomation, Carpinteria, CA) for analysis of PBMCs or cells from tissues, as we previously described (36, 38, 44). Briefly, lymphocytes were gated based on their forward scatter, side scatter, and pulse-width characteristics. We adjusted the voltage to optimally display lymphocytes according to the forward and side scatters. We also gated out cell aggregates according to the pulse width. At least 40,000 gated events were analyzed using Summit data acquisition and analysis software created by DakoCytomation. Cells stained with different color–conjugated Abs alone were used as controls and to estimate the amount of compensation needed for the different color combinations. Further gating and determination of quadrant position for analysis were based on specific Ab staining (positives) together with negative background determined by either unstained cells or isotype control Ab–stained cells.
To clearly demonstrate CD4 depletion impacts on each of effector populations, we employed CD3 positive-focused or CD3 negative-focused displays of cytokine differences in flow histograms between CD4-depleted and nondepleted groups (representative flow histograms in Figs. 2–6), as we are dealing with multiple cytokines in each of four different sections (blood CD3+ T, blood CD3− non–T, lung CD3+ T, and lung CD3− non–T subpopulations). This method of displays yields similar frequencies or numbers compared with those without focusing.
Analyses of various T effector cells and CD3− non–T effector lymphocytes
In this study, CD3+ lymphocytes that are capable of producing IFN-γ, TNF-α, IL-17, IL-22, IL-4, and perforin are defined as T effector cells, CD3− lymphocytes that were capable of producing IFN-γ, TNF-α, IL-17, IL-22, and IL-4, and perforin are defined as CD3− effector lymphocytes. Because we gated out large size and granularity of cells, most of CD3− lymphocytes were mainly composed of NK cells. All of these effector cells were measured by direct ICS without in vitro Ag stimulation. Percentages of CD3+ and CD3− effector cells in saline group at week 1 are percentages in lymphocytes. At 2.5 and 5.5 wk postinfection, we collected the blood from the CD4-depleted group of macaques ∼2–3 d earlier than that from the control group. Significantly large numbers of macaques were employed in each group, and therefore sampling dates at 2.5 or 5.5 wk could not be precisely matched between the groups due to manpower and Biologic Research Resources Annex BSL3 handling limits. For data expression, results were compared between the groups at 2.5 and 5.5 wk, respectively. Our justification for comparing data between the groups at 2.5 or 5.5 wk despite a difference in 2–3 d was that M. tuberculosis infection was slow without rapid changes within 2–3 d, and data trends could be judged by referring to results at 1 and 4 wk, respectively.
Determination of bacterial CFU counts from blood, BALF, and lung tissues
Whole blood cells (500 μl) or 1 ml 4% NaOH-decontaminated, filtered BALF of each M. tuberculosis–infected animal was first treated with 5 ml RBC blood lysis buffer (Sigma-Aldrich) for 10 min or waiting until the suspension became clear, washing once with 5% FBS-PBS, and then directly plating on 7H11 agar plates (BD Biosciences) as we previously described (36, 38). To measure bacilli counts in lung tissues, a half of the cut sections of the right caudal, right middle, or the left caudal lobes from each animal was taken for CFU determination after the extensive gross pathologic evaluation was accomplished. When there were TB lesions in the respective lobe, a half of the lung tissue containing ∼50% lesions was taken. When no visible lesions were seen in the respective lobe, a random half of tissue was taken for evaluation. Tissue homogenates were made, 5-fold serially diluted (5, 25, and 125×), and plated as described previously. The plates were incubated in a 37°C incubator for 3 wk, and CFU were counted (36, 39).
Immunohistochemistry analysis of IL-17+, IL-22+, IL-4+, perforin+ lymphocytes in lung tissue sections containing TB granulomas
This procedure was done as we previously described (36–38). Briefly, lung tissue sections in Trilogy solution were heated for epitope retrieval at 121°C for 15 min and incubated at room temperature for 1 h with rabbit anti-human IL-17 (sc-7927, Santa Cruz Biotechnology), anti-human IL-22 (CI0144, Capralogics), anti-human IL-4 (RPA05112, Reprokine), or mouse anti-human perforin mAb (5B10, Abcam) at 4 μg/ml concentration each. Sections were washed and then incubated for 30 min with peroxidase-labeled polymer-conjugated goat anti-rabbit or anti-mouse IgG. Isotype rabbit serum IgG or mouse IgG1 was used as a control to match Abs against IL-17, IL-22, IL-4, or perforin. As the “normal uninfected control,” lung tissue sections from normal uninfected macaques were also stained with Abs against above cytokines. No positive stainings were seen for isotype control or normal uninfected control in immunohistochemistry analyses.
Gross pathologic analyses of TB lesions and scoring systems
This procedure was done essentially the same as previously described (36, 37, 39, 44, 46). Briefly, animals were euthanized by i.v. barbiturate overdose and immediately necropsied in a biological safety cabinet. Standard gross pathologic evaluation procedures were followed by the blinded pathologist and associates, with each step recorded and photographed. Lung lobes, bronchial, mesenteric, axillary, and inguinal lymph nodes, tonsils, and other major organs were collected and labeled. Multiple specimens from all tissues with gross lesions and remaining major organs were harvested. Gross observations including but not limited to the presence, location, size, number, and distribution of lesions were recorded. The scoring system was excised to calculate gross pathology scores for TB lesions as previously described (36, 37, 39).
Statistical analysis
Statistical analysis was done by using GraphPad Prism software (GraphPad Software, La Jolla, CA). A normality test was first performed to decide whether the data were normally distributed. Data that passed the normality test were analyzed by Student t test (parametric method) in this study. Data that did not pass the normality test were analyzed by a Mann–Whitney U test (nonparametric method). A p value <0.05 was considered statistically significant. Only p values <0.05 were shown in the text.
Results
M. tuberculosis infection of CD4-depleted macaques induced very early extrapulmonary M. tuberculosis dissemination and, subsequently, rapid progression to more severe TB
The possibility that CD4+ T cells might act as innate-like cells to play a role in very early phase of M. tuberculosis infection has not been evaluated, because most studies in mice and in HIV-infected or uninfected humans are focused on pathologic consequences at the endpoint or advanced stage of M. tuberculosis infection (14, 21, 35). To address this possibility, macaques were depleted of CD4+ T cells using the humanized mAb CDR-OKT4A-human IgG1 (40, 42, 43) administered i.v. three times at an interval of every 2 wk at a dose of 50 mg/kg during the study and evaluated for changes in early infection and progression status. The CD4 depletion regimen was considerably efficient, as almost complete depletion of blood CD4+ T cells and marked lymphocytopenia occurred after the treatment and lasted for >5 wk (Fig. 1A). A steady, significant decline of CD4+ T cells in BAL cells was also observed despite the fact that the progressive M. tuberculosis infection in lungs could potentially increase pulmonary recruitment of CD4+ cells (Fig. 1A).
M. tuberculosis infection of CD4-depleted macaques induced very early extrapulmonary M. tuberculosis dissemination and, subsequently, rapid progression to more severe TB. Macaques were treated with either depleting anti-CD4 mAb (36) or saline (control) at days −5, 9, and 24 after M. tuberculosis infection. Saline served as control in this study because earlier studies showed that treatment of macaques with IgG isotype control Ab did not significantly cause changes in numbers of CD4+ or CD8+ T cells (39). (A) Flow cytometry data (CD3 gated) show percentages (left) and absolute numbers (middle) of CD4+ T cells in the blood and percentages of CD4+ T cells in BAL cells (right) over time after pulmonary M. tuberculosis infection. Data are mean numbers with error bars of SEM derived from 12 (CD4-depleted group) and 9 (control group) cynomolgus macaques. *,#p < 0.05, **,##p < 0.01, ***p < 0.001. (B) Graphic data of absolute numbers of blood lymphocytes indicate severe lymphocytopenia in CD4-depleted macaques compared with control (saline-treated) animals. Data were derived from CBC differentials of 12 (CD4 T cell–depleted group) and 9 (control group) cynomolgus macaques, respectively. (C) CD4-depleted macaques exhibited M. tuberculosis bacteremia (left) and increases in M. tuberculosis CFU counts in BALF (middle) and in lung homogenates derived from right caudal lung lobe (infection site) when compared with control macaques. Data are means ± SEM of CFU counts/10 ml blood, BALF, and right caudal lung tissue homogenates derived from individual macaques for each group. *p < 0.05. (D) Mean TB pathology score in CD4-depleted macaques was significantly worse than that in control macaques. Please see pathology scoring system for evaluation of TB lesions (36, 39). *p < 0.05.
Surprisingly, profound CD4 depletion resulted in an early extrapulmonary M. tuberculosis dissemination after pulmonary M. tuberculosis infection, as bacilli bacteremia was clearly detected at week 1 and still evident at week 3 (Fig. 1B). In fact, one of the CD4-depleted macaques exhibited an extremely high level of bacteremia, with blood M. tuberculosis CFU counts being too many to count in culture plates. In contrast, control macaques without CD4 depletion did not exhibit detectable bacteremia within 5 wk after pulmonary M. tuberculosis infection (Fig. 1C). The occurrence of early mycobacteremia in macaques depleted of CD4+ T cells suggests that CD4+ T cells are needed to limit M. tuberculosis replication and extrapulmonary dissemination at the very early stage of pulmonary M. tuberculosis infection.
Consistently, CD4-depleted macaques exhibited much higher levels of M. tuberculosis CFU counts in BAL fluid at weeks 5 and 6 than did the control macaques (Fig. 1C). The CD4-depleted macaques rapidly progressed to moribund TB conditions such as 20–25% weight loses, reduced alertness, and marked lymphopenia (Supplemental Fig. 2) that required euthanasia at ∼7 wk after M. tuberculosis infection. At the necropsy, CD4-depleted macaques had ∼2 log higher mean M. tuberculosis CFU counts in lung tissue homogenates than did the control macaques (Fig. 1B), with M. tuberculosis spreading from infection site to other lung lobes (Supplemental Fig. 1). Furthermore, when pathology scores for TB lesions were calculated individually using the scoring system as we and others previously described (36, 39, 45) and compared between the groups, the CD4-depleted group exhibited significantly worse TB lesions than did the control group (Fig. 1D). All CD4-depleted macaques displayed TB lesions in extrapulmonary organs whereas only ∼15% of controls had the gross TB lesions in kidneys, livers, spleens, or other organs.
Thus, M. tuberculosis infection of CD4-depleted macaques led to very early extrapulmonary M. tuberculosis dissemination and rapid progression to moribund condition, with much worse TB lesions and higher levels of M. tuberculosis infection than in immune-competent individuals.
CD4 depletion during M. tuberculosis infection diminished the ability of CD8+ T cells to sustain Th17-like/Th1-like and cytotoxic-like effector responses
We then sought to address the question as to whether a loss of CD4+ T cell–mediated immunity coincided with any changes in Th1, Th2, Th17, Th22, and cytotoxic effector functions in the adaptive T cell pool during M. tuberculosis infection. This question has been incompletely addressed, because CD4 depletion studies in mice appeared to be focused mostly on changes in Th1 cytokines measured by ICS after vigorous in vitro CD3/CD28 Ab stimulation of TCR expressed by all T cells (14, 21, 25), and because CTL precursors in CD4 knockout mice were evaluated in vitro either by coculture with M. tuberculosis–infected dendritic cells or stimulated vigorously by CD3/CD28 Ab (25). To examine in vivo activation, development and differentiation details of CD8+ T effectors and Th-like effectors after CD4 depletion, we employed direct ICS methods to directly detect T effector cells constitutively producing cytokines without in vitro Ag stimulation or TCR ligation as described by us and others (38, 47, 48). Because almost complete depletion of circulating CD4+ T cells was seen after CD4 depletion, our studies of blood T cells were focused on two major aspects: 1) we examined whether CD8+ T cells could compensate for the CD4+ T cell loss and mount Th1/Th2/Th17/Th22-like effector functions; and 2) we measured perforin production directly as a surrogate marker for CTL effector function, because CD8+ CTL killing activity against M. tuberculosis–infected cells is mainly mediated by perforin (49, 50).
To our surprise, CD4-depleted macaques exhibited earlier and remarkable increases in T effector cells constitutively producing cytotoxic molecule perforin in the T cell pool at weeks 1–3 after M. tuberculosis infection when compared with controls (Fig. 2A, 2B). Perforin-producing T cells increased up to 18% at weeks 1–3 from baseline <3% when the data were expressed by percentages of CD3+ T cells (Fig. 2A, 2B). When total numbers of perforin-producing CD8+ T cells were calculated in the blood, we found that CD4 depletion resulted in ∼200-fold increases in absolute numbers of CD8+ T effector cells constitutively producing perforin (Fig. 2C). This was considered significant because such early major increases in CD8+ T effectors occurred at the times of lymphocytopenia (Figs. 1B, 2C) after CD4 depletion, and because the increases in absolute numbers were much more dramatic than those of saline-treated controls without CD4 depletion or lymphocytopenia (Fig. 2C). However, without CD4+ T cells, perforin-producing CD8+ T effector cells were subsequently unsustainable, as they apparently declined at weeks 4 and ∼6 after M. tuberculosis infection (Fig. 2B, 2C).
CD4 depletion during M. tuberculosis infection diminished the ability of CD8+ T cells to sustain Th17-like/Th1-like and cytotoxic-like effector responses. Note that scales in y-axes for perforin+ cells in subfigures are much higher than those for other cytokine subfigures. (A) Representative flow cytometry histograms show percentage numbers (upper right quadrants) of T effector cells constitutively producing IL-17, IL-22, and perforin in CD4-depleted and control macaques. Macaque IDs are indicated in bottom left quadrants. Data are gated on CD3 and derived from PBLs collected at 3 wk from CD4-depleted macaques and at 4 wk from controls after M. tuberculosis infection. (B) Percentages of different T effector cells among CD3+ T cells in the blood during M. tuberculosis infection of control and CD4-depleted (CD4 mAb-treated) macaques. Data are mean numbers with error bars of SEM from up to 12 CD4-depleted and 9 control macaques. Because CD4+ T cells were almost completely depleted in the blood of CD4 Ab-treated macaques (Fig. 1A), various CD3+ T effector cells were temporarily interpreted as CD3+CD8+ T effector cells. *p < 0.05, **p < 0.01, ***p < 0.001. (C) Graph data show mean absolute numbers of CD8+ T effector cells producing IFN-γ, perforin, IL-17, and IL-22 in PBLs from up to 9 control and 12 CD4-depleted macaques during M. tuberculosis infection. Note the initial increases in CD8+ T cells producing perforin and IL-17 and their subsequent contractions at week 4 or 6 after M. tuberculosis infection of CD4-depleted monkeys. *p < 0.05, **p < 0.01.
Interestingly, we also found detectable increases in circulating T effector cells constitutively producing IL-17, IFN-γ, and IL-22 without a need for in vitro Ag stimulation (Fig. 2A, 2B). In fact, we detected 7- and 3-fold increases in absolute numbers of IL-17–producing and IFN-γ–producing CD8+ T effector cells, respectively, at early weeks after M. tuberculosis infection of CD4-depleted macaques when compared with controls (Fig. 2C). Similarly, the numbers of Th1/Th17-like CD8+ T effector cells were only transiently increased at some early weeks, but not sustained in later time points, presumably due to the absence of CD4+ T cells (Figs. 1B, 2C). In contrast, control macaques without CD4 depletion exhibited no or subtle increases in CD8+ T effector cells constitutively producing perforin, IFN-γ, or IL-17 at early weeks after M. tuberculosis infection (Fig. 2C) (38, 39).
Thus, these results demonstrated that CD4 depletion during M. tuberculosis infection of macaques uncovered the ability of CD8+ T cells to compensate and to rapidly differentiate to cytotoxic-like effectors and Th17-like/Th1-like effectors, but, without CD4+ T cells, such compensating effector functions were subsequently unsustainable in the blood circulation.
CD4 depletion during M. tuberculosis infection abrogated the ability of CD3− lymphocytes to mount CD3−IL-22+ and CD3−perforin+ effector responses, but left CD3−IL-17+ effectors unaffected
We then sought to examine whether CD4 depletion might impact effector functions of CD3− innate lymphocytes during M. tuberculosis infection. Using an approach for initial identification of broad ranges of innate or non–T immune cells, we evaluated CD3− lymphocytes that could constitutively produce IL-17, IL-22, IL-4, TNF-α, IFN-γ, and perforin. Interestingly, CD3− non–T lymphocytes in control macaques without CD4 depletion were able to mount predominant IL-22+ and NK-like (perforin) responses in M. tuberculosis infection because percentage and absolute numbers for CD3− IL-22+ or NK-like effector cells were much higher than other innate-like effector cells producing IL-17 IL-4, TNF-α, and IFN-γ (Fig. 3). However, CD4 depletion during M. tuberculosis infection apparently reduced the ability of CD3− non–T lymphocytes to produce perforin and IL-22, as both percentage and absolute numbers of perforin- and IL-22–producing CD3− lymphocytes were not increased anymore at 4–6 wk after M. tuberculosis infection (Fig. 3). Surprisingly, CD4 depletion during M. tuberculosis infection increased the ability of CD3− non–T lymphocytes to produce IL-17 (Fig. 3). The IL-17–producing CD3− lymphocyte subpopulation increased from ∼0.2% at baseline to 5% at week 1, peaked at ∼8% at week 3, and remained at 2–3% at weeks 4–6 after M. tuberculosis infection of CD4-depleted macaques (Fig. 3A, 3B). Consistently, absolute numbers for the IL-17–producing CD3− lymphocyte subpopulation increased ∼100-fold at week 3 and ∼30-fold at week 5, respectively, in CD4-depleted macaques after M. tuberculosis infection (Fig. 3C). The CD3− non–T lymphocyte subset producing IL-17 after CD4 depletion increased much more dramatically than did those CD3+ T effector cells producing IL-17 (Figs. 2C, 3C). In contrast, numbers of CD3− lymphocytes producing IL-17 in control macaques remained undetectable or very low during the entire course of M. tuberculosis infection (Fig. 3). Of note, there were increases in percentage and absolute numbers of IL-4–producing CD3− lymphocytes after M. tuberculosis infection during CD4 depletion (Fig. 3B, 3C). CD3− lymphocytes producing IFN-γ increased but remained a minor population compared with IL-17+ and IL-4+ cells after CD4 depletion (Fig. 3B, 3C).
CD4 depletion during M. tuberculosis infection depressed the ability of CD3− lymphocytes to mount CD3−IL-22+ and CD3−perforin+ effector responses, but left CD3−IL-17+ effectors unaffected. Note that y-axis scales for perforin and IL-17 are much higher than other cytokines in subfigures, whereas y-axis scale for IFN-γ is much lower. (A) Representative flow cytometry histograms show percentages (upper left quadrants) of CD3− lymphocytes constitutively producing IL-17, IL-22, and perforin in CD4-depleted and control macaques. Data are gated on CD3− lymphocyte populations in PBLs collected at 3 wk from CD4-depleted macaques and at 4 wk from controls after M. tuberculosis infection. (B) Graphs show percentage numbers of different CD3− effector populations among blood CD3− lymphocytes during M. tuberculosis infection of CD4-depleted and control macaques. Data are mean percentage numbers in CD3− lymphocytes with error bars of SEM derived from up to 12 CD4-depleted and 9 control macaques. *p < 0.05, **p < 0.01, ***p < 0.001. (C) Graphs show mean absolute numbers of IL-17–, IL-22–, and IL-4–producing CD3− lymphocytes in blood of CD4-depleted and control groups during M. tuberculosis infection. Of note, CD4 depletion drove increases in absolute numbers of IL-17–producing CD3− lymphocytes during M. tuberculosis infection. Data are derived from up to 12 CD4-depleted and 9 control macaques. *p < 0.05, **p < 0.01, ***p < 0.001.
Thus, whereas CD3− lymphocytes in the presence of CD4+ T cells mounted predominant Th22-like and NK-like effector responses to M. tuberculosis infection, CD4 depletion depressed these effector functions and increased the ability of CD3− lymphocytes to produce IL-17.
CD4-depleted macaques exhibited no or few pulmonary T effector cells capable of constitutively producing IFN-γ, TNF-α, IL-17, IL-22, or perforin at the endpoint
Given that M. tuberculosis infection in CD4-depleted macaques rapidly progressed to life-threatening TB and a loss of compensative T effector cells, we sought to examine whether CD4 deficiency in more severe TB led to no or little recruitment of T effector cells in the pulmonary compartment. Toward this aim, we similarly measured T effector cells in BALF cells or lung cells using direct ICS without in vitro Ag stimulation at the time when CD4-depleted macaques developed clinical moribund TB that required euthanasia at ∼7 wk after M. tuberculosis infection. CD4-depleted macaques exhibited a loss of the ability to recruit pulmonary T effector subpopulations constitutively producing IFN-γ, TNF-α, IL-17, IL-22, or perforin, as these effector subpopulations were almost undetectable, with mean frequencies being <0.3%, in BAL cells at the time when more severe TB occurred (Fig. 4A, 4B). In contrast, up to 15% of IL-4–producing T effector cells were detected in BAL cells. Consistently, lymphocytes isolated from lung tissues of CD4-depleted macaques did not accommodate any detectable T effector cells producing Th1, Th17, or Th22 cytokines in (Fig. 4C) when these animals developed more severe TB lesions coincident with marked body weight loss, deteriorated general conditions, and lymphocytopenia. These findings appeared to be in contrast with those seen in control macaques that began to exhibit body weight loss and severe TB lesions at week 9 after M. tuberculosis infection. In fact, the control macaques displayed great numbers of T effector cells producing IFN-γ, TNF-α, IL-17, IL-22, or perforin in BAL cells (Fig. 4A, 4B). Large numbers of Th1/Th17/Th22-like T effector cells were also detected at the severe TB endpoint in lung tissues from the control macaques without CD4 depletion (Fig. 4C, 4D). These results suggested that CD4 depletion decreased pulmonary recruitment of T effector cells producing IFN-γ, TNF-α, IL-17, IL-22, and perforin at the endpoint of M. tuberculosis infection, despite pulmonary presence of IL-4+ T effector cells.
CD4-depleted macaques exhibited no or few pulmonary T effector cells capable of constitutively produced IFN-γ, TNF-α, IL-17, IL-22, and perforin at the endpoint. (A) Representative histograms show different T effector cell populations from control and CD4-depleted macaques at the endpoint. Percentages and macaque IDs are indicated similarly as in Fig. 2A. (B) Bar graphs show mean numbers of different T effector cells in controls and CD4-depleted macaques. Data were derived from up to 12 CD4-depleted and 9 control macaques. Data were gated on CD3+ T lymphocytes. *p < 0.05, **p < 0.01, ***p < 0.001 when comparing numbers detected in TB endpoints between controls and CD4-depleted macaques. (C) Representative histograms show IFN-γ–, IL-17–, and IL-22–producing T effector cells in lymphocytes isolated from lungs of control and CD4-depleted macaques at the endpoint. The percentages and macaque IDs are indicated similarly as above. Lung lymphocytes were derived from right middle/caudal lung lobes collected at necropsy. Owing to limited numbers of lymphocytes recovered from lung lobes of CD4-depleted macaques, other T effector cells in lung tissues could not be evaluated. (D) Bar graphs show mean percentage numbers of different T effector cell populations among CD3+ T cells in lungs from controls and CD4-depleted macaques. Data were derived from four CD4-depleted and up to nine control macaques. Data were gated on CD3+ T lymphocytes. *p < 0.05, **p < 0.01, ***p < 0.001.
CD4-depleted macaques showed no or few CD3−perforin+ and CD3−IL-22+ cells in the pulmonary compartment at the endpoint, despite the presence of CD3−IL-17+ and CD3−IL-4+ cells
Because CD4 depletion abrogated IL-22+ and NK-like effector functions and favored IL-17 production by CD3− lymphocytes (Fig. 3), we examined whether CD4 depletion impacted the ability of CD3− effector lymphocyte subpopulations to accumulate in the pulmonary compartment at the endpoint of more severe TB. CD4 depletion did not appear to reduce pulmonary presence of CD3− lymphocytes producing TNF-α, IFN-γ, IL-17, or IL-4 (Fig. 5). In contrast, CD3− lymphocytes constitutively producing perforin (NK-like cells) were almost undetectable in the BAL cells of CD4-depleted macaques. Similarly, <1% of CD3− lymphocytes producing IL-22 were detected in BAL cells from CD4-depleted macaques (Fig. 5). Interestingly, control macaques without CD4 depletion exhibited large numbers of NK-like effector cells accounting for ∼30% of total CD3− lymphocytes in BAL cells at the endpoint of severe TB (Fig. 5B). These control macaques also showed >5% IL-22+ effector cells detected in CD3− lymphocytes in BALF cells (Fig. 5B). Thus, these results demonstrated that CD4-depleted macaques exhibited no or few CD3− effector cells producing perforin and IL-22 in BAL cells at the endpoint of more severe TB, despite the presence of CD3− lymphocytes producing IL-17 and IL-4. The occurrence of pulmonary IL-17+/IL-4+ non–T effector cells consisted of IL-17 and IL-22 responses of CD3− lymphocytes in the blood of CD4-depleted macaques (Fig. 3).
CD4-depleted macaques showed no or few CD3−perforin+ and CD3−IL-22+ cells in the pulmonary compartment at the endpoint, despite presence of CD3−IL-17+ and CD3−IL-4+ cells. Note that the y-axis scale for perforin is much higher than other cytokines in subfigures, whereas the y-axis scale for IL-17 is much lower. (A) Representative histograms show IL-17– and IL-22–producing CD3− lymphocytes in the BALF cells from control and CD4-depleted macaques at the endpoint. Percentages and macaque IDs are indicated similarly as in Fig. 3A. (B) Bar graphs show mean percentages of different CD3− lymphocyte effectors among CD3− lymphocytes in the BALF cells at the endpoint from control and CD4-depleted macaques. Data were gated on CD3− lymphocytes. Data are mean percentage numbers in CD3− lymphocytes with error bars of SEM derived from up to 12 CD4-depleted and 9 control macaques *p < 0.05, **p < 0.01, ***p < 0.001.
TB granulomas in CD4-depleted macaques appeared to contain fewer IL-22+ and perforin+ cells despite the presence of IL-17+ and IL-4+ cells
Our direct ICS-based approaches allowed us to show that CD4 depletion led to remarkable changes in IL-17+, IL-22+, and perforin+ cells in systemic and pulmonary compartments during M. tuberculosis infection. We therefore employed immunohistochemistry analysis to examine whether CD4 depletion could impact distribution of effector lymphocytes producing IL-17, IL-22, and perforin in TB granuloma at the endpoint of more severe TB. IL-17+ lymphocytes were detectable in TB granulomas in CD4-depleted macaques, but they appeared rare in those from control animals without CD4 depletion (Fig. 6). The distribution of these IL-17+ lymphocytes in TB granulomas in CD4-depleted macaques consisted of increases in CD3−IL-17+ lymphocytes (Figs. 3, 5), although T effector cells producing IL-17 were decreased (Fig. 4). Alternatively, CD4-depleted macaques appeared to exhibit few IL-22+ and perforin+ lymphocytes in TB granulomas, whereas IL-22+ or perforin+ effector cells were readily detected in TB granulomas of control macaques at their endpoint of severe TB (Fig. 6). IL-4+ effector cells were readily detected in lung TB granulomas from both CD4-depleted and control macaques (Supplemental Fig. 3) perhaps due to the fact that both groups had similar levels IL-4+ cells in BALF cells (Fig. 5). Overall, the in situ staining results were similar to those changes in IL-17+, IL-22+, and perforin+ T cells or non–T cells as seen by direct ICS-based flow cytometry analysis (Figs. 2–5), suggesting that TB granulomas in CD4-depleted macaques contained few IL-22+ and perforin+ cells but displayed detectable IL-17+ and IL-4+ cells.
TB granulomas in CD4-depleted macaques appeared to contain fewer IL-22+ and perforin+ cells despite the presence of IL-17+ and IL-4+ cells. Shown are immunohistochemistry staining images in lung sections from representative CD4-depleted and control macaques, with animal ID and original imaging magnification scales indicated. Cytokines-producing cells were stained brown. Lung tissue sections were prepared from the right caudal lung lobes (infection site) collected at the endpoint. Similar data were seen from other macaques in each group. CN, Cynomolgus macaque.
Discussion
The present study allowed us to make, to our knowledge, a previously undiscovered observation suggesting an innate-like role of CD4+ T cells in containing acute M. tuberculosis replication and extrapulmonary dissemination at the very early stage of pulmonary M. tuberculosis infection. This new role of CD4+ T cells in primary M. tuberculosis infection has not been demonstrated in humans or mice. Murine TB models do not appear to have a power to reveal this new role of CD4+ T cells owing to the fact that pulmonary M. tuberculosis infection even in immune-competent mice usually leads to systemic M. tuberculosis infection (51). Likewise, in vivo studies in HIV− persons have not been undertaken to address this role of CD4+ T cells in containing very early extrapulmonary M. tuberculosis dissemination in primary M. tuberculosis infection. Although cross-sectional clinical studies reported mycobacteremia or extrapulmonary TB in some of HIV-infected patients (52, 53), it remains to be determined whether M. tuberculosis dissemination is attributed to HIV-induced immune suppression and/or depletion of CD4+ T cells. It is noteworthy that pulmonary infection of macaques with medium- to low-dose (500 CFU) M. tuberculosis reproducibly induces tuberculosis in nonvaccinated macaques (54), and that CD4 depletion starting from initial M. tuberculosis infection can lead to greater increases in M. tuberculosis burdens in the very early stage in which innate immune response takes place. Our finding suggests that CD4+ T cells can act as innate-like cells contributing to early innate-like resistance to M. tuberculosis infection, whereas CD4+ T cells subsequently function as major adaptive immune cells. Thus, the results from the present study support the view that CD4+ T cells may be needed for early prevention from extrapulmonary M. tuberculosis dissemination. From this standpoint, our observation revealing this role for CD4+ T cells appears to be novel, and therefore it is relevant to one of the possible mechanisms underlying CD4+ T cell–induced resistance to TB.
The mechanistic experiments using direct ICS without in vitro TCR stimulation yield interesting in vivo data, suggesting that primate CD4+ T cells may play a critical role in adaptive immunity against TB during primary pulmonary M. tuberculosis infection of HIV/SIV− macaques. A combined loss of both “CD4-driven early resistance to M. tuberculosis” and CD4+ T cell adaptive immune responses certainly can lead to very early and continuous increases in M. tuberculosis burden, and subsequently worsen TB disease and pathology. The rapid progression to more severe TB during M. tuberculosis infection of CD4-depleted macaques indeed confirms those previous reports for CD4+ T cell–mediated adaptive immunity against TB (14–22).
The present study demonstrates that CD4+ T cells play an important role in sustaining multiple CD8+ T cell effector functions in primary M. tuberculosis infection. After M. tuberculosis infection of CD4-depleted macaques, CD8+ T cells tend to temporarily compensate for CD4+ T cell deficiency and rapidly differentiate to high levels of Th1-like, Th2-like, Th17-like, Th22-like, and CTL-like effector cells for a short time. However, without CD4+ T cells, CD8+ T cells cannot sustain CTL effector function and Th1/Th17/Th22-like responses after they compensate CD4+ T cell loss for 2–3 wk. Because CD4+ T cells were depleted before M. tuberculosis infection, our data were consistent with previous reports of viral infections, suggesting that CD4+ T cells are not necessary for the priming and differentiation of effector CD8+ T cells for the short term during acute infections but are required to sustain effector CD8+ T cells in the long term during chronic infections (55). Earlier studies in mice reported that the ability of CD8+ T cells to produce IFN-γ in M. tuberculosis infection was comparable between wild-type and CD4-depleted mice and that CTL precursors were impaired in CD4 knockout mice (25). Because the murine studies employed approaches to stimulate and examine the entire T cell population using ICS after potent in vitro TCR stimulation of all T cells by CD3/CD28 Ab (14, 21, 25), our results appear to complement the mouse study, as we used direct ICS methods to detect effector cells without in vitro Ag stimulation or TCR ligation. Our approaches may be close to measuring in vivo M. tuberculosis–driven T effector cells. Note that at some time points, PBMCs were also stimulated with purified protein derivative for 6 h prior to the ICS/surface staining (37). After 6 h of purified protein derivative stimulation, frequencies of CD3+IL-17+ cells, but not CD3+IL-22+ cells, were further increased when compared with direct ICS staining, but CD3+IL-22+ cells remained the same (data not shown).
Furthermore, our study provides new information regarding detailed changes in Th1, Th2, Th17, Th22, and perforin responses over time after CD4 depletion, which were not addressed in mouse studies (14, 15, 20, 21, 25, 48, 56). Our results therefore represent new findings and suggest that CD8+ T cells not only could remarkably potentiate short-term CTL effector function in an absence of CD4+ T cells during M. tuberculosis infection, but also could rapidly mount Th17-like, Th22-like, Th1-like, and even Th2-like responses. Thus, although CD4 depletion during M. tuberculosis infection eliminates CD4-dependent innate-like early defense and subsequent adaptive anti-TB immunity, deficiency of CD4+ T cells also indirectly leads to a loss of protective CD8+ T cell responses (39). In fact, a critical role of CD8+ T cells in protection against TB is implicated in macaque TB models (39).
The precise mechanism whereby CD4+ T cells help sustain CD8+ T effector function responses of CD8+ T cells and CD3– lymphocytes in M. tuberculosis–infected nonhuman primates is currently not known. It is noteworthy that CD4+ T cell help is essential to generate long-lived antiviral CD8+ T cell memory (57). Progressive loss of CD4+ T cells in HIV, hepatitis C virus, and hepatitis B virus infections has been associated with dysfunction of virus-specific CD8+ T cells and ineffective containment of these chronic viral infections, although the nature of CD4 help for controlling chronic human infections remains unclear. In mice, recent studies indicate that CD4+ T cell production of IL-21 is required for maintenance of CD8+ T cell function in persistent but not resolving viral infections, including lymphocytic choriomeningitis virus infection (58–60). The role of IL-21 in helping virus-specific CD8+ T cells is also implicated in chronically HIV-1–infected humans and SIV-infected macaques (61, 62).
It remains to be determined whether subsequent downregulation of CD8+ T effector cells in CD4-depleted macaques is relevant to cell exhaustion as a result of Tim-3 or PD-1 expression. Of note, our earlier studies demonstrated that these “exhaustion” molecules are highly expressed during early mycobacterial infection of macaques (unpublished data and Refs. 63, 64), despite that T cell responses are detected. New studies from us and others show that Tim-3+ T cells exhibit effector memory phenotypes and stronger anti-TB effector functions in M. tuberculosis–infected humans (65, 66), suggesting that Tim-3 expression and signaling may act differently between M. tuberculosis and HIV-1 infections.
Another interesting observation is that CD4 depletion during M. tuberculosis infection leads to no or very few T effector cells constitutively producing IFN-γ, TNF-α, IL-17, IL-22, and perforin in the pulmonary compartment at the endpoint of more severe TB. In contrast, control macaques without CD4 depletion exhibit appreciable numbers of these pulmonary T effector cells at the endpoint of severe TB. Undetectable occurrence of pulmonary T effector cells in CD4-depleted macaques appears to be disconnected with the high-level M. tuberculosis infection and more severe TB in lungs, as lung TB might help to drive marked influx of inflammatory cells, including T effector cells to the pulmonary compartment. The changes in the cytokine environment after CD4 depletion may impact pulmonary trafficking of these T effector cells. Consistently, CD4+ T cell deficiency caused a shortage in total cell recovery from lungs in a murine model of M. tuberculosis infection (56). Therefore, CD4+ T cells are needed to sustain the local pulmonary immune response. Alternatively, CD4+ T helper cells and/or their cytokines in the pulmonary compartment may also be important for maintaining the capabilities of T effector cells to produce IFN-γ, TNF-α, IL-17, IL-22, and perforin in lungs or granulomas once they are recruited to these tissues. IFN-γ produced by CD4+ T cells but not by other cells has been shown to be required to enhance the effector function of CD8+ T cells in a murine model of M. tuberculosis infection (56). It is likely that a loss of IFN-γ production by CD4+ T cells in our models directly affects the maintenance of optimal capabilities of T effector cells in the lung. Recent studies have suggested that T effector cells in the infection site need a second activation step to maintain their fully competent function, proliferation, and survival (67). Continuous in vivo activation of CD4+ T cells has been shown to be required for the IFN-γ production by CD8+ T cells in the lungs in a murine model of M. tuberculosis infection (48). Our data thus provide supporting evidence that CD4+ T cells are needed for the presence of multifunctional effector T cells in the local infection site. Essentially, the marked decreases in pulmonary recruitment and accumulation of anti-TB T effector cells after CD4 depletion and M. tuberculosis infection of macaques therefore suggest a new role of CD4+ T cells and a potential mechanism for CD4+ T cell–induced immunity against TB. Notably, we have recently demonstrated that rapid clonal expansion and pulmonary trafficking of IFN-γ–producing CD4+ and CD8+ T effector cells is critical for vaccine-induced protection against M. tuberculosis infection in macaques (24). It is noteworthy that anti-TB cytokines include IFN-γ (68), TNF-α (69), and perforin (70), and that primate IL-22–producing CD4+ T cells can also mount anti-Mycobacterium effector function (30). Thus, marked decreases in pulmonary trafficking and accumulation of T effector cells constitutively producing IFN-γ, TNF-α, IL-22, and perforin after CD4 depletion would help to explain a loss of CD4-mediated anti-TB immunity in macaques.
The present study also found a surprising discovery by conceiving effector functional responses of CD3− lymphocytes during M. tuberculosis infection. Although useful direct ICS methods reveal that CD3− lymphocytes from control macaques clearly have the ability to mount Th1-like, Th2-like, Th17-like, Th22-like, and NK-like (perforin) effector functions in M. tuberculosis infection, the selected NK-like and Th22-like effector functions appear to be impressive and quite dominant among all detectable effector cells. Notably, CD4 depletion during M. tuberculosis infection leads to a dramatic decrease or loss of these CD3− NK-like or Th22-like effector functions, particularly in the pulmonary compartment at the endpoint. These results implicate that CD3− lymphocytes capable of producing perforin and IL-22 are predominantly involved in non–T immune responses to M. tuberculosis infection, and that these CD3− NK-like and Th22-like effector functions depend on CD4+ T helper cells. This finding is considered significant because little is known about in vivo responses of CD3− NK-like and Th22-like effector cells and their relationship with CD4+ T helper cells during primary M. tuberculosis infection of humans. Given the possibility that these CD3− effector lymphocytes contribute to protection against TB, a loss of these CD3− effector functions after CD4 depletion might be somehow relevant to abrogation of CD4+ T cell–induced anti-TB immunity.
Surprisingly, during CD4 depletion and M. tuberculosis infection, CD3− lymphocytes could mount overreacting IL-17 response at least in early time points, and CD8+ T cells have short-term appreciable effector function of IL-17 production. Essentially, absolute numbers of CD3− lymphocytes producing inflammatory IL-17 in CD4-depleted macaques are indeed ∼100-fold greater than the combined total of IL-17+ T cells and IL-17+CD3− lymphocytes in control macaques (Fig. 3). Such changes consist of more IL-17+ cells and fewer IL-22+ and perforin+ lymphocytes in TB granulomas. Overproduction of IL-17 after CD4 depletion may result from depletion of CD4+ regulatory T cells, as regulatory T cells and Th17 are mutually competitive against each other (71). Studies have been done to evaluate whether IL-17 and IL-22 or cells producing them play a protective and/or detrimental role in TB. Whereas IL-17 produced by CD4+ T cells in vaccinated mice helps recruit Th1 cells and control M. tuberculosis infection after challenge (29), Th17 cells or Th22 cells may correlate with either an anti–M. tuberculosis response or TB severity (28, 30–33). Notably, an enhanced IL-17 response is associated with higher M. tuberculosis burden in multidrug-resistant TB (72). The enhanced IL-17 response similar to what we found in the present study may not be protective against TB (Figs. 3, 6) (28, 73) and instead may negatively influence protective Th1 responses (75). Additionally, overreacting IL-17+ cells would favor TB inflammation and immunopathology, as Th17 cells play a role in TB inflammation and autoimmune diseases (28, 75). These inflammatory effects of overproducing IL-17 may help to exacerbate M. tuberculosis infection and TB lesions, and therefore contribute to the loss of CD4+ T cell–mediated immunity against TB after CD4 depletion.
Thus, CD4+ T cell depletion during M. tuberculosis infection leads to early extrapulmonary M. tuberculosis dissemination, a subsequent loss of systemic/pulmonary responses of multiple effector cells, and rapid progression to more severe TB in macaques. These results implicate previously unknown innate-like ability of CD4+ T cells to contain M. tuberculosis dissemination at very early stage. Data also suggest that CD4+ T cells are needed to prevent rapid TB progression and to sustain multiple effector functions of CD8+ T and NK-like cells during M. tuberculosis infection of nonhuman primates.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank Dr. Keith Reimann for providing the depleting anti-CD4 mAb and other Chen Laboratory staff for technical assistance.
Footnotes
This work was supported by National Institutes of Health Grants R01 OD015092 (RR13601), R01 HL64560, and RO1 AI106590 (all to Z.W.C.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
- BAL
- bronchoalveolar lavage
- BALF
- bronchoalveolar lavage fluid
- ICS
- intracellular cytokine staining
- TB
- tuberculosis.
- Received May 23, 2013.
- Accepted December 23, 2013.
- Copyright © 2014 by The American Association of Immunologists, Inc.
References
In this issue
