Mycobacterium tuberculosis antigen-specific T-cell responses in smear-negative pulmonary tuberculosis patients

Abstract

Despite recent improvements in microbial detection, smear-negative TB remains a diagnostic challenge. In this study, we investigated the potential discriminatory role of polychromatic flow cytometry of M. tuberculosis antigen-specific T cells to discriminate smear-negative TB from health controls with or without latent TB infection, and non-TB respiratory illnesses in an endemic setting. A cross-sectional study was conducted on HIV negative, newly diagnosed smear-positive PTB (n = 34), smear-negative/GeneXpert negative PTB (n = 29) patients, non-TB patients with respiratory illness (n = 33) and apparently healthy latent TB infected (n = 30) or non-infected (n = 23) individuals. The expression of activation (HLA-DR, CD-38), proliferation (Ki-67), and functional (IFN-γ, TNF-α) T-cell markers using polychromatic flow cytometry was defined after stimulation with PPD antigens. Sputum samples were collected and processed from all patients for Mtb detection using a concentrated microscopy, LJ/MGIT culture, and RD9 typing by PCR. Our study showed CD4 T cells specific for PPD co-expressed activation/proliferation markers together with induced cytokines IFN-γ or TNF-α were present at substantially higher levels among patients with smear-positive and smear-negative pulmonary TB than among healthy controls and to a lesser extent among patients with non-TB illness. Our study conclude that smear-negative TB can be distinguished from non-TB respiratory illness and healthy controls with a flow cytometric assay for PPD-specific T cells co-expressing activation/proliferation markers and cytokines.

Substantially high level of activation markers was observed among smear positive pulmonary TB patients. Flow cytometry could be a potential tool to distinguish smear negative pulmonary TB from healthy control and non TB respiratory illness.

Graphical Abstract

Graphical Abstract.

Introduction

Despite global efforts to control tuberculosis (TB), the disease continues to be a diagnostic and treatment challenge [1]. TB diagnosis has been difficult either due to poor sensitivity of tools to diagnose Mycobacterium tuberculosis (Mtb) [1–5], cost or feasibility [6–8]. In Ethiopia, TB diagnostics is mainly done using direct sputum Ziehl–Neelsen acid-fast smear microscopy or GeneXpert, with a suboptimal yield to detect cases and many are left mis- or undiagnosed [1, 8–11]. Several reports have shown that about a third of notified pulmonary TB (PTB) cases are diagnosed based on clinical and radiological grounds without bacteriological confirmation [1, 10, 12].

Smear-negative TB is a major public health problem in high TB and HIV settings [13, 14], and underdiagnosis can contribute to transmission, even in higher income countries. Indeed Tostmann et al. demonstrated 13% of TB transmission in the Netherlands was due to smear-negative TB patients [15]. Similar studies in Ethiopia showed a high prevalence of undiagnosed SNTB cases among the prison communities, and the majority (93%) of the acid-fast bacilli (AFB) smear-negative microscopy cases in the country were not etiologically explained during routine diagnostic services [10, 16]. The GeneXpert assay has improved diagnosis, but cannot detect all cases of culture proven TB, and smear negative, culture-negative PTB diagnosed on clinical grounds remains problematic.

Recent studies have shown the diagnostic potential of T-cell activation molecules such as HLA-DR, and CD38 and the intracellular proliferation marker (KI-67) as good markers integrated with TB-specific intracytoplasmic cytokine flow cytometry [17]. The role of these biomarkers in the diagnosis of smear-negative PTB patients has not been explored in-depth in endemic countries. In this study, we evaluated in the Ethiopian setting the diagnostic potential of this approach for smear negative TB in comparison with smear positive and latent TB infection (LTBI), and TB suspects found ultimately to have other non-TB respiratory illness using this flow cytometry approach.

Methodology

Study setting: In this cross-sectional study, a total of 149 adult and HIV negative individuals were enrolled, including 29 AFB smear-negative and 34 AFB smear-positive pulmonary TB patients, recruited from health centers in Addis Ababa (Girar Health Center, Alem Bank Health Center, Koteba Health Center, Lomi-meda Health Center, Akaki Health Center, Kaliti Health Center, and Bole Health Center), in a high TB burden setting [18]. We adopted the Ethiopian TB screening and diagnostic algorithm [19], and operationally defined smear-positive PTB cases as those who were newly identified TB patients diagnosed using acid-fast bacilli (AFB) smear microscopy or LJ/Löwenstein–Jensen or MGIT/Löwenstein–Jensen/Mycobacterium growth indicator tube culture positive or using Mtb (RD9) typing by PCR. Patients who were clinically diagnosed with TB but negative by AFB smear microscopy and gene expert and positive by LJ or MGIT culture positive or using Mtb (RD9) typing by PCR were defined as smear-negative PTB patients. Additionally, we recruited a few comparator groups: 33 patients with non-TB respiratory illness (defined as those who were confirmed negative either by AFB smear microscopy, GeneXpert, or culture, but responded to first line broad-spectrum antibiotics treatment) and 52 apparently healthy controls, who were further categorized as latently infected (QFT positive, n = 30) and uninfected (QFT negative, n = 23). The latter groups had no clinical sign or symptoms for TB.

Laboratory procedure

Sampling: All participants were clinically screened for signs, symptoms and chest X-ray suggestive of TB and TB diagnoses. Some were bacterialogically confirmed following tests with AFB smear microscopy, and GeneXpert MTB/RIF assay using spot-morning-spot sputum samples. Further laboratory confirmations were done at the Armauer Hansen Research Institute (AHRI) laboratory using Löwenstein–Jensen (LJ)/Mycobacterium growth indicator tube (MGIT) culture and RD9 typing using PCR. Approximately 20 ml of venous blood was collected into heparinized blood collection tubes from all participants and transported to AHRI laboratory at ambient temperature for flow cytometry analyses. Aliquots of blood were also collected into Quantiferon blood collection tubes from the apparently healthy groups for the IFN-γ-release assay (IGRA).

Cell isolation and flow cytometry: Peripheral blood mononuclear cells (PBMC) were separated from heparinized whole blood by density centrifugation technique over Ficoll–Hypaque as previous described [20]. Briefly, diluted whole blood was layered on Ficoll–Hypaque density gradient medium and centrifuged at 1500 rpm for 15 min at room temperature. Cells from the interface were separated, washed in RPMI-1640 media, supplemented with 10% fetal calf serum, 1% l-glutamine (L-G) and 1% penicillin/streptomycin (P/S) solution and counted. Cell concentrations were adjusted to 1–2 million cells per well and fresh PBMCs were stimulated overnight (for 18 h) in 96-well microtiter plates with PHA (5 μg/ml) or PPD (Statens Serum Institute, Denmark, 10 μg/ml). Protein secretion inhibitor Brefeldin A (BD, USA) was added at 30 min for PHA and 2 h after M. tuberculosis stimulations for PPD. Cells were then harvested from replicate 96-well plate stained with cocktail of surface anti-human monoclonal antibodies containing 2.5 μl CD4-BV510 (BD, USA), 2.5 μl CD8-APC-Cy7 (BD, USA), 2.5 μl CD38-BV421 (BD, USA), and 2.5 μl HLA-DR-PE-Cy7 (BD, USA) within FACS polystyrene tubes, incubated for 30 min on ice and washed two times with FACS buffer (PBS with 1 mM EDTA and 1% fetal calf serum). Cells were then fixed and permeabilized with Cytofix/perm (BD, USA) and stained with a cocktail of intracellular anti-human monoclonal antibodies which contained 5 μl Ki-67-PerCP-C^y5.5 (BD, USA), 5 μl IFN-γ-FITC (BD, USA), and 5 μl TNF-α-APC (BD, USA). After 30 min incubation, cells were washed two times with Perm Wash (BD, USA). Finally, cells were fixed with 300 μl of 2% paraformaldehyde (PFA), washed and resuspended with 500 μl of FACS buffer until acquisition with BD FACSCanto II using FACSDIVA software. All flow cytometry data analyses were done using FlowJo analytical software (version 9.9.6).

Quantiferon TB gold plus assay: It was performed using QFT Gold plus (Qiagen, Germany). Briefly, 1 ml of blood was taken in each of the four tubes precoated with TB-antigen (TB1 and TB2), phytohemagglutinin for the positive control or no antigen for the negative control and incubated for 24 h at 37°C. Following centrifugation at 2500 rpm for 10 min, supernatants (plasma) was collected from each tube and assayed for IFN-γ measurement using QFT-plus ELISA (Qiagen, catalog number 622120). Optical density was read with 450 nm filter and 620 nm reference filter. Test interpretations were analyzed using QuantiFERON-TB Gold Analysis Software (softmax^Rpro7.013) according to the manufacturer’s instruction.

Bacteriological assessments: 5–10 ml of morning–morning productive sputum was collected from TB patients and patients with non-TB respiratory disease according to the standard sputum collection and transportation protocol. Sputum samples were digested and processed with the N-acetyl l-cysteine-sodium hydro oxide method (NALC-NaOH) and the sediment inoculated into the LJ/ MGIT 960 media (lot number, 0059457). All cultures were incubated at 35–37 °C until growth was observed or discarded as negative after 8 weeks of follow-up. DNA extraction using a DNA extraction kit (69504 and 69506) was done from heat killed culture isolates or sediment (using the QBD2 grant heat block at 95 °C for 20 min). Finally M. tuberculosis identification was done using RD9 PCR (RD-9 REV-RTPCR—5ʹ-CACTGCGGTCGCCATTG-3,TM-57-60OC, GC: 64.7%, 17 mer, RD9-FW-RTPCR-5-TGCGGGCGGACAACTC-3,TM-56-86OC,GC = 68.75%, 16mer, Eurofins genomics, H8223-24498/11). In these cases, we incorporated positive control (RH37V) and negative control (RNase-free water). All processing and decontamination of sputum samples were done in a SAFE FAST classic level 3 biosafety cabinet. Both known positive and negative controls were used to ensure the quality of individual experiments.

Data analysis

At the AHRI data management center, sociodemographic and clinical data were coded, double-entered, and cleaned before being entered into the database. SPSS version 25 was used to code, input, and analyze all socio-demographic variables. FlowJo 9.9.6 was used to analyze all flow cytometry data. The frequency of PPD-specific cytokine-producing CD4 T cells was calculated by subtracting the frequency of cytokine-producing CD4 cells cultured with medium alone from paired cultures stimulated with PPD. The frequency of PPD-specific HLA-DR+ cytokine-producing CD4 T cells was determined by subtracting the frequency of HLA-DR+ cytokine-producing cells cultured with medium alone from paired cultures stimulated with PPD. The same calculation was done to determine the frequency of PPD-specific CD-38+, and PPD-specific Ki-67+ cytokine-producing CD4+ T cells. To determine the fraction of activated or proliferating cells among PPD reactive T cells, we divided the above frequency of PPD-specific HLA-DR or CD-38 or Ki-67-positive cytokine producing by the aforementioned total PPD-specific cytokine-producing cells. Note that the term fraction here is mathematically equivalent to percentage/100.

Further statistical analysis was performed with GraphPad prism version 6 software. The non-parametric Mann–Whitney two group comparison statistical test was used to analyze the data. The percentage of IFN-γ+ and/or TNF-α+ producing CD4+ T cells was always at least two-fold higher in the PPD stimulated samples than the unstimulated (negative control) sample response. Also, a positive control (PHA) was included for quality control purposes. statistical significance was defined as a P-value of <0.05.

Result

Flow cytometric gating strategy

As shown in Fig. 1, forward and side scatter features of lymphocytes were used as an initial gate, followed by exclusion of doublets defined by light scatter. After gating CD4+ and CD8+ cells, cells were evaluated for the expression of markers of activation (HLA-DR, CD-38), proliferation (KI-67), and function (TNF-α, IFN-γ). Thresholds for cytokine positive cells were defined after comparison of PPD-stimulated with media control-stimulated cells. The expression of activation, proliferation, and functional markers on CD8⁺ T cells was all relatively low in this study, hence we focused on CD4 T cells.

Figure 1:

flow cytometry gating strategy in the TB biomarker study, Addis Ababa, Ethiopia.

Cytokines markers profile on treatment naïve smear positive and negative PTB patients from unstimulated CD4 T cells

First we evaluated the spontaneous cytokine response of PBMC cultured in the absence of PPD antigens. When compared with smear-negative TB patients, non-TB respiratory disease patients, QFT-positive and -negative study participants, the overall magnitude of IFN-γ+ on CD4+ T cells from smear-positive TB patients was significantly lower, whereas the magnitude of TNF-α+ response on CD4+ T cells did not show any significant differences across the study groups. In smear-positive and negative TB patients, the level of double-positive cytokine (TNF-α+ IFN-γ+) producing CD4+ T cells was considerably lower than in QFT-positive study participants. Furthermore, the frequency of double-positive cytokine (TNF-α+ IFN-γ+) producing CD4+ T cells in non-TB respiratory illness cases was significantly higher than in smear-negative TB patients (Fig. 2).

Figure 2:

the overall magnitude of IFN-γ+ and TNF-α+ production in unstimulated PBMC from study cohorts. PBMC were cultured without PPD and intracytoplasmic detection performed as described in Materials and Methods. Data analysis was done using the non-parametric Mann–Whitney test with significant P values indicated. Legend labels “Smear Pos” refer to smear-positive pulmonary TB patients, “Smear Neg” to smear-negative pulmonary TB patients, “NTB” to patients with non-TB respiratory disease, “QFT^+” to quantiferon positive (LTBI⁺), “QFT^−: to quantiferon negative (LTBI⁻).

Cytokine markers profile on treatment naïve smear positive and negative PTB patients from PPD-specific CD4 T cells

Smear-positive PTB patients had a considerably higher frequency of PPD-specific IFN-γ⁺CD4⁺ T, TNF-α⁺CD4⁺T cells, and IFN-γ⁺TNF-α⁺CD4⁺T cells than smear negative PTB patients, but the difference did not reach at significant level (Fig. 3). However, when compared with QFT negative and non-TB respiratory disease study participants, the over all frequency of IFN-γ⁺CD4⁺ T, TNF-α⁺CD4+T cells, and IFN-γ⁺TNF-α⁺CD4⁺ T cells from smear-negative PTB patients was much higher (Fig. 3). Furthermore, the level of expression of IFN-γ⁺CD4⁺ T, TNF-α⁺CD4+T cells, and IFN-γ⁺TNF-α⁺CD4⁺ T cells among smear positive and negative PTB patients were comparable with QFT-positive study participants.

Figure 3:

the overall magnitude of PPD-specific IFN-γ⁺ and TNF-α⁺ production of CD4⁺ T among different cohorts. PBMC were stimulated overnight with PPD and evaluated for cytokine production as described in Materials and Methods. The percentage of PPD-specific cytokine responses from CD4+T cells was calculated by subtracting the frequencies of cytokine-producing CD4+ cells of unstimulated cultures from that of paired cultures of cytokine-producing CD4+ cells in PPD-stimulated cultures. Refer to Fig. 2 for descriptions of statistical analysis and legend labels.

The overall magnitude of IFN-γ⁺, TNF-α⁺, and IFN-γ⁺TNF-α⁺ responses from PPD-specific CD4⁺ T cells was significantly higher among QFT-positive study participants than among QFT-negative study participants (Fig. 3). In addition, when compared with QFT-negative study participants, the expression of IFN-γ⁺CD4⁺T cells and IFN-γ⁺TNF-α⁺ CD4⁺T cells among non-TB respiratory disease was much higher.

Furthermore, the overall magnitude of PPD-specific CD4⁺ T cells that expressed IFN-γ⁺, TNF-α⁺, and IFN-γ⁺TNF-α⁺ was significantly higher among smear-positive PTB patients compared with QFT-negative study subjects and non-TB respiratory disease (Fig. 3).

Activation and proliferation markers profile on CD4 T cells from unstimulated samples

We initially compared the expression of activation antigens without stimulation to determine whether clinical groups differed with respect to presumed in vivo activation. When compared with QFT-positive and -negative study participants, the level of HLA-DR expression on bulk CD4⁺ T lymphocytes among both smear-positive and smear negative TB patients was considerably greater in the unstimulated sample. Similarly, Ki-67 expression on CD4⁺T cells was considerably higher in smear-positive TB patients than in QFT positive and non-TB respiratory disease while smear-negative TB cases did not have Ki-67 expression above that of QFT-positive study participants. Patients with non-TB respiratory disease also showed significantly higher levels of HLA-DR but not CD38 among CD4 T cells compared with LTBI-positive and -negative individuals (Fig. 4).

Figure 4:

the overall magnitude of activation (HLA-DR, CD-38) and proliferation (Ki-67) markers expression on CD4⁺ T cells in study cohorts. Unstimulated cells were stained for expression of markers. Refer to Fig. 2 for descriptions of statistical analysis and legend labels.

Activation and proliferation marker profile among PPD-specific cytokine-producing cells

Activation and proliferation marker expression together with cytokine detection was analyzed with and without PPD-stimulation in vitro. Smear-positive PTB patients had significantly higher frequencies of activation (HLA-DR) markers on PPD-specific IFN-γ⁺ and TNF-α⁺ producing CD4+T cells (Fig. 5) than patients with non-TB respiratory disease, QFT-positive and -negative study participants. However, smear-negative PTB had considerably higher levels of activation markers (CD-38, HLA-DR) and proliferation (Ki-67) markers on PPD-specific IFN-γ⁺CD4⁺T and TNF-α⁺ CD4⁺T cells when compared with QFT-negative study participants.

Figure 5:

activation marker co-expression on PPD-specific IFN-γ⁺CD4⁺T cells and TNF-α⁺ CD4⁺ T cells in multiple cohorts. PBMC from multiple cohorts were cultured overnight with or without PPD antigen, stained and analyzed. The percentage of PPD-specific CD4 T cells expressing activation or proliferation markers was calculated by subtracting the frequencies of CD4⁺T cells co-expressing activation/proliferation and cytokine markers of unstimulated cultures from that of paired PPD stimulated cultures. Refer to Fig. 2 for descriptions of statistical analysis and legend labels.

Furthermore, smear-negative PTB patients had significantly more PPD-specific Ki-67⁺IFN-γ⁺CD4⁺T cells than QFT-positive study participants. In addition, HLA-DR⁺IFN-γ⁺CD4⁺T cells, CD38⁺ IFN-γ⁺CD4⁺T cells, and CD38⁺ TNF-α⁺ CD4⁺T cells were considerably greater in QFT-positive study participants compared with QFT-negative study participants (Fig. 5).

Furthermore, cytokine-positive cells (TNF⁺, IFN-γ⁺) expressing HLA-DR and CD-38 marker were higher in smear-positive and negative PTB patients than in non-TB respiratory disease; however, the Ki-67 marker in combination with TNF⁺ and IFN-γ⁺ did not distinguish TB disease from non-TB disease (Fig. 5).

The overall expression of activation (HLA-DR, CD-38) and proliferation (Ki-67) markers from TNF⁺CD4⁺Tcells and IFN-γ⁺ CD4⁺Tcells except HLA-DR⁺TNF-α⁺ CD4⁺ T cells among non-TB respiratory disease patients was considerably higher than QFT-negative subjects (Fig. 5).

The fraction of activation or proliferation marker expressing cells among PPD-specific cytokine-producing CD4+ T cells

In smear positive PTB patients, the fraction of activation marker (HLA-DR, CD-38) expression among cytokine producing (IFN-γ⁺, TNF-α⁺⁾ PPD-specific CD4+T cells was considerably greater than in QFT-positive and -negative study participants. Also, compared with QFT-positive and -negative study participants, the fraction of proliferation marker positive cells among TNF-α⁺-producing PPD-specific CD4+T cells in smear positive PTB patients was significantly higher. However, the fraction of Ki-67+ cells among PPD-specific IFN-γ⁺CD4+T cells was higher than QFT-positive study participants only.

In smear positive PTB patients, the fraction of activation or proliferation marker positive cells among PPD-specific cytokine-producing CD4+T cells was higher than in non-TB respiratory disease patients. However, this only reached statistical significance for the fraction of HLA-DR+ cells among PPD-specific IFN-γ⁺ CD4+T cells.

Furthermore, compared with QFT-positive and -negative study participants, the fraction of HLA-DR-positive cells among PPD-specific IFN-γ-producing CD4+T cells and the fraction of Ki-67+ cells among TNF-α +CD4+T cells in smear negative PTB patients was significantly higher. However, smear negative PTB patients exhibited a significantly higher fraction of activated or proliferating cells among PPD-specific cytokine-producing cells compared with QFT-positive study participants. Only the fraction of Ki-67+ of PPD-specific Ki-67+ IFN-γ⁺CD4+T cells differ substantially between smear-positive and -negative PTB in this study.

Moreover, higher fraction of activation or proliferation markers were observed from PPD-specific cytokine-producing (IFN-γ⁺, TNF-α⁺) CD4+T cells in non-TB respiratory illness patients compared with QFT positive and negative, However, only the fraction of Ki-67+ IFN-γ⁺CD4+T cells and HLA-DR+ TNF-α⁺CD4+T cells was significantly higher in QFT-positive study participants.

Comparable fraction of activation and proliferation from cytokine-producing PPD-specific CD4+T cells were observed between QFT-positive and -negative study participants (Fig. 6).

Figure 6:

the fraction of activation marker expressing cells among PPD-specific IFN-γ⁺CD4⁺T cells and TNF-α⁺ CD4⁺ T cells in multiple cohorts. PBMC from multiple cohorts were cultured overnight with or without PPD antigen, stained and analyzed. These fractions were calculated as described in Materials and Methods. Refer to Fig. 2 for descriptions of statistical analysis and legend labels.

Discussion

The ability of M.tb-specific cytokine (in particular IFN-γ) responses from T cells to define different stages of TB has long been recognized and has contributed to the replacement of the tuberculin skin test with Interferon-Gamma Release Assays, which have better sensitivity and specificity but fail to distinguish latent TB infection from active TB disease due to comparable levels of antigen-specific cytokine-producing cells among apparently healthy individuals [21, 22]. Adekambi et al. have found a high frequencies of antigen-specific T cells which displayed markers suggestive of recent in vivo activation and/or proliferation in smear positive pulmonary TB patients, and those markers were able to discriminate active from latent TB with a high level of sensitivity and specificity [17]. Many recent studies on active pulmonary TB patients who were smear positive have supported this finding [23–26]. However, data on paucibacillary TB, in particular smear-negative PTB cases is scarce, and this is particularly relevant due to the difficulties in diagnosing paucibacillary disease, challenges. We therefore tested the concept of presumed recent in vivo activation in smear-negative TB by evaluating activation (HLA-DR, CD-38) and proliferation (Ki-67) markers on antigen-specific T cells. Smear-negative pulmonary tuberculosis, which is a major public health problem in high TB and HIV endemic settings such as Ethiopia [10] [16] [27, 28]. In this research context, we define smear-negative PTB is defined as patients who were smear microscopy negative, GeneXpert negative, but bacteriologically confirmed by growth on LJ or MGIT or RD9 PCR. In our study, 12 of the 29 smear and GeneXpert negative PTB patients became LJ or MGIT culture media positive, while the remaining 17 were confirmed by RD 9 PCR. As comparator groups we included patients with non-TB respiratory illness (observed to be smear negative, GeneXpert negative, LJ or MGIT culture-negative, and clinically resolved with broad-spectrum antibiotics), because we reasoned that pre-activated T cells were potentially present in the latter group, especially given their presentation with an acute infectious disease. In addition, we also included comparator cohorts of healthy controls, with or without LTBI.

In our study, we observed higher level of activation markers expressed from M.tb-specific TNF-α⁺CD4⁺T cells and IFN-γ⁺CD4+T cells in smear negative PTB patients compared with patients with non-TB respiratory illness, but this did not reach at significance level. The detectable levels of activation markers and cytokines among M.tb-specific T cells in non-TB respiratory illness could be due to bystander T-cell activation. This possibility was supported by our findings of enhanced expression of activation and proliferation markers among unstimulated cells in patients with non-TB compared with healthy controls with or without latent TB. The fact that the percentage of activated cells was in the 5–10% range both for non-TB as well as smear-negative and -positive TB, while frequencies of antigen-specific CD4 T cells measured in TB and other infectious diseases is far lower suggests that the majority of such in vivo activated cells were not specifically activated to the etiological pathogen but by other mechanisms, including a bystander effect, possibly by cytokines generated during the immune response. In a clinical setting, therefore, the expression of activation antigens on PPD reactive cells among patients with non-TB respiratory infectious, though lower than that of smear negative TB patients, could complicate the interpretation of the future diagnostic potential of this approach.

Smear negative PTB Patients also showed higher fraction of activation (HLA-DR, CD-38) and proliferation (Ki-67) markers expressed from M.tb-specific TNF-α+CD4+T cells and IFN-γ+CD4+T cells in comparison with QFT-positive and -negative healthy controls. Smear-negative PTB and latent TB, on the other hand, displayed comparable frequency of cytokine response (IFN-γ, TNF-α, IFN-γ+TNF-α+) from M.tb-specific CD4+T cells. These findings highlight the importance of using an antigen-specific cytokine production assay in combination with markers reflecting presumed in vivo activation. The high level of activated and proliferated cells from antigen-specific cytokine-producing CD4+T cells in smear-negative PTB could therefore have diagnostic potential for smear-negative PTB as other studies have shown for other TB states [17, 23–25, 29].

Smear positive PTB Patients had much higher levels of cytokine (IFN-γ+ TNF-α, IFN-γ+TNF-α+), activation (HLA-DR, CD-38), and proliferation (Ki-67) markers from Mtb-specific TNF-α+CD4+T cells and IFN-γ+CD4+T cells than smear negative as well as patients non-TB respiratory illness. The higher load of bacilli in smear positive compared with smear-negative TB could explain this finding. Consistent results were observed by Adekambi et al., who found a decrease in activation markers on TB-specific cells after therapy of smear-positive TB patients [17].

Furthermore, when compared with non-TB respiratory illness and QFT-negative study participants, pulmonary TB patients (smear positive and negative) had a higher level of IFN-γ+TNF-α+ CD4+T cells in the present study. Although the relevance of those cytokines in clearing the bacilli is controversial, a high number of PPD-specific CD4+T cells co-expressed IFN-γ+TNF-α+ could be associated to T-cell response to the M. tuberculosis antigen [30, 31].

The frequency of activated or proliferating cells which co-express cytokines in response to PPD, as described above, is a statistic which nonetheless is dependent on the frequency of reactive cells. Thus, equivalent results would be obtained with one individual who had high percentages of activated cells among few PPD reactive cytokine-producing cells, and another individual who had low percentages of activated cells among a much higher frequency of PPD-specific cells. Given varying rates of LTBI globally, it is likely that different populations would express varying frequencies of PPD reactivity, and this in turn could significantly impact an assay which depends on the frequency of PPD reactivity. This issue was overcome in this study by an additional analysis in which PPD-specific cytokine-producing CD4 T cells were first gated and then analyzed for the fraction of activated or proliferating T cells among such gated cells. Although this approach requires at least some PPD reactive cells to gate, the final statistic is in essence independent of the antigen-specific cytokine-producing frequency. With this approach, in particularly using IFN-γ as the readout cytokine for PPD reactivity and HLA-DR as the activation antigen, we were able to successfully discriminate smear-negative TB from non-TB respiratory infections and LTBI.

In our study we by using PPD as a stimulatory antigen preparation, we may have detected some cross reactive responses with non-tuberculous mycobacterial infections. Our rationale in focusing on PPD-specific responses was that such responses are more robust than those to TB-specific peptides, and as such would serve as a more useful starting point for a resource limited setting. Furthermore, we observed approximately 5 to 10-fold differences in the frequencies of PPD-specific response in LTB positive vs negative subjects, suggesting that the majority of the PPD-specific responses measured in this study were due to truly TB-specific T cells. However, it is clear our findings require confirmation with more highly selective TB peptides such as those from ESAT and CFP10 molecules. Finally, we have not fully explored all the potential combinations of markers to distinguish between smear-negative TB and other clinical cohorts. For example, we observed, in the absence of activation or proliferation markers, the highest discriminatory power using co-expression of IFN-γ and TNF-α as cytokines. Further work will be needed to determine whether additional combinations of cytokines or activation or proliferation markers can improve on the resolution observed here.

In conclusion, we found a higher level of activation markers (HLA-DR, CD-38) expression from M.tb-specific IFN-γ+CD4+T and TNF-α+CD4+T cells in smear-negative PTB than non-active TB disease, suggest that such flow cytometry may have promise as a diagnostic approach.

Glossary

Abbreviations

AFB

acid-fast bacilli

AHRI

Armauer Hansen Research Institute

IFN-γ

interferon gamma

IGRA

IFN-γ-release assay

LJ

Lowenstein–Jensen

LTB

latent TB

LTBI

latent TB infection

MGIT

mycobacterium growth indicator tube

PBMC

peripheral blood mononuclear cells

PTB

pulmonary TB

SNPTB

smear-negative pulmonary tuberculosis

SPPTB

smear-positive pulmonary tuberculosis

TB

tuberculosis

TNF-α

tumor necrosis α

Patient Consent

Before being enrolled in the study, all individuals provided their informed consent. The study participants’ information was kept private, and no personal identifiers were recorded. All data collection sheets and accompanying papers were kept in a secured cabinet and password-protected computers at the AHRI data management center, and only study personnel had access to the information entered on data collection sheets.

Funding

This work was supported by a partial fund received from NIH/Fogarty International Center Global Infectious Diseases (grant no. D43TW009127, A.E.) and a core fund from the Armauer Hansen Research Institute (AHRI), received from SIDA and NORAD. The funders had no direct role in the study design, data collection, or analyses.

Conflict of Interests

We declared that there were no conflicts of interest.

Author Contributions

AE collected data did the laboratory work, analyzed the data, and wrote the first draft of the manuscript; T.A., A.M., J.E., and J.R. analyzed the data, edited the manuscript, and supervised the work; D.M. and S.N. did the lab work and analyzed the data; L.W. and R.H. conceived and designed the study, analyzed the data, supervised the study and edited the manuscript. All authors approved the final version of the manuscript.

Ethical Approval

Before recruiting study participants, the research proposal was examined by the Institution Review Boards (IRB) of AAU-CHS and the AHRI/ALERT ethical review committee, and ethical approval was obtained. Before the study, letters of support from the Addis Ababa Health Bureau and other appropriate institutions were received. During the recruitment process, all safety protocols and proper data handling were followed.

Data Availability

Data can be accessed by contacting the corresponding author and the AHRI data management center.