Neutrophils from tuberculosis patients are polarized toward pro-inflammatory and anti-inflammatory phenotypes according to the disease severity

Abstract

Neutrophils are the first line of defense against pathogens, combating them by using several antimicrobial mechanisms. These cells display a remarkable plasticity that can be molded by the different environments that neutrophils confront to protect the host, therefore presenting diverse phenotypes. Actually, pro- and anti-inflammatory neutrophils populations (N1- and N2-like phenotypes) have been described in cancer and inflammatory disorders. However, the identification of N1/N2 neutrophil subtypes in human intracellular bacterial diseases remains unexplored. Here, we characterized neutrophils from tuberculosis (TB) patients presenting distinct immunological status according to their disease severity. TB patients were classified as high or low responders (HR or LR) in accordance with their immunity against Mycobacterium tuberculosis (Mtb). Interestingly, by analyzing the phenotypic and functional characteristics of neutrophils from the two groups of TB patients we demonstrated that HR patient’s neutrophils display a pro-inflammatory N1-like phenotype, whereas LR patient’s neutrophils show an anti-inflammatory N2-like phenotype. Remarkably, whereas neutrophils from both groups of patients phagocytized MtbH37Rv strain equally, HR TB’s neutrophils displayed a significantly increased ability to kill pathogenic Mtb as compared to neutrophils from LR TB patients that presented a diminished capacity of bacterial elimination. Together, our findings suggest the existence of different subtypes of neutrophils in TB patients according to their immune response to Mtb and disease severity, indicating that neutrophils might be promising targets for TB host-directed therapy.

Introduction

Tuberculosis (TB) is an infectious respiratory disease that caused 1.3 million deaths in 2022, thus remaining as a main health problem worldwide.¹ The successful establishment of Mycobacterium tuberculosis (Mtb) infection depends on the early interaction of the bacteria with the host's innate immune cells.² Macrophages and neutrophils are professional phagocytes that contribute to control TB infection. Neutrophils are the first cells that migrate to the site of infection and predominate in TB patients’ lungs where they phagocytose the pathogen.³ However, the mechanisms of Mtb elimination by neutrophils are still controversial. Current evidence has demonstrated that neutrophils display superior functional diversity than previously appreciated and seem to have a key role during non-infectious diseases as well, mainly in cancer. In fact, the terminology N1 and N2 phenotypes in tumor-associated neutrophils (TAN) refers to the classification of tumor-associated macrophages (TAM), with differential activation/differentiation states.^4,5 Anti-tumorigenic N1 phenotype is distinguished by the expression of increased immune-activating cytokines and chemokines and an enhanced capacity for in vitro tumor cell killing in comparison to the pro-tumorigenic N2 phenotype.⁵ Conversely, N2 neutrophils show augmented expression of pro-tumor factors that aid in promoting tumor growth.^5,6 Interestingly, the use of N1/N2 phenotypes was extended beyond tumor immunity. In fact, in the acute phase of visceral leishmaniasis, neutrophils show an increased expression of the immunosuppressive molecules IL-10 and arginase.⁷ Furthermore, for free-living extracellular protozoa, the tolerant human host response is characterized by a Th17 immunity involving N2 neutrophils and other immune cells.⁸ Besides, during heart or brain infarction, the pro-inflammatory effect of ischemia triggers the polarization of neutrophils towards an N1 profile,^9–11 represented by high levels of inflammatory markers and elevated NETosis activity.^9,10 Additionally, N1 and N2 neutrophil heterogeneity has been described in burn mouse models,^12,13 human periodontitis,^14,15 sepsis,¹⁶ and several parasitic infections,⁸ among other diseases. However, the existence of N1- and N2-like neutrophil subpopulations during intracellular bacterial infections such as human TB has not been elucidated.

Recent studies have demonstrated the existence of neutrophil subpopulations that are differentially abundant during Mtb infection. In particular, neutrophils can be separated into normal (NDNs) and low-density neutrophils (LDNs).¹⁷ LDNs are present in higher numbers during active TB disease as compared to healthy controls.^18,19 Furthermore, Deng et al. observed that TB patients with advanced disease exhibited notably superior levels of LDNs than those with moderate disease.¹⁹ Besides, it was reported that high Reactive Oxygen Species generation, low levels of banded neutrophils, and elevated levels of IL-10-expressing CD16^dimCD62L^lo neutrophils were associated with reduced TB pathology.^20,21 Therefore, neutrophils can be contemplated as potential early indicators of TB severity. Considering the increasing evidence for the existence of neutrophil subpopulations with distinct phenotypic and functional characteristics,¹⁶ and given that the study of neutrophil heterogeneity has never been extended to pro and anti-inflammatory (N1- and N2-like) phenotypes during active TB, we study neutrophils from TB patients presenting distinct immunological status and disease severity.

Materials and methods

Subjects

Patients with TB (n = 74) were diagnosed at the División Tisioneumonología, Hospital F.J. Muñiz (Buenos Aires, Argentina), based on clinical and radiological data, together with the identification of acid-fast bacilli (AFB) in sputum. All participants had received less than 1 wk of anti-TB regular therapy. Peripheral blood from each participant was collected in heparinized tubes after obtaining informed consent. The protocols conducted in this study were approved by the Ethical Committee of F.J. Muñiz Hospital (ethical protocol number 1542/19). All methods were carried out in accordance with relevant guidelines and regulations.

Inclusion/exclusion criteria and classification of patients

All individuals were 18 to 60 yrs old and had no previous history of medical conditions known to compromise the immune system, such as human immunodeficiency virus (HIV) infection, cancer diagnosis, ongoing treatment with immunosuppressive medications, hepatic or renal diseases, pregnancy, and positive serology for various viral infections (i.e. hepatitis A, B, or C) or bacterial infections (i.e. leprosy and syphilis). Individuals with preexisting bleeding disorders or those currently undergoing anticoagulant therapy, which might present an elevated risk of bleeding during the procedure of obtaining the sample, were excluded from the study.

TB patients were classified as high responders (HR TB) or low responders (LR TB), based on their in vitro lymphocyte responses to a whole cell lysate of M. tuberculosis (Mtb-Ag) as previously described.²² Briefly, HR TB patients are individuals displaying significant proliferative responses, IFN-γ production and an increased percentage of SLAMF1⁺ CD3⁺ cells after Mtb-Ag stimulation, whereas LR TB patients exhibit low proliferative responses, IFN-γ release and SLAMF1⁺CD3⁺ cells. LR TB patients had more severe pulmonary disease compared with HR individuals. Cut-off values to differentiate between LR TB and HR TB were established previously by Pasquinelli et al.²². The fulfillment of 2 out of the 3 criteria was enough to assign a patient to the corresponding group.

Proliferation index

The proliferation index was calculated either by determining [³H]TdR incorporation or monitoring carboxyfluorescein diacetate succinimidyl ester (CFDA-SE) staining by flow cytometry on a FACSAria II flow cytometer (BD, San Jose, California, USA). Thereby, fresh peripheral blood mononuclear cells (PBMC) were stimulated with Mtb-Ag for 5 d and afterward, cells were pulsed with [³H]TdR (1 μCi/well) and harvested 16 h later. [³H]TdR incorporation was measured in a liquid scintillation counter as counts per minute (c.p.m). Proliferation index for each individual was calculated as: (c.p.m. after Mtb-Ag stimulation)/(c.p.m. after culturing with medium). On the other hand, fresh PBMC were stained with CFDA-SE (BioLegend, 423801) for 20 min, and then the reaction was stopped by adding fetal bovine serum (FBS). After washing, cells were stimulated with or without Mtb-Ag and after 5 d, cells were fixed and analyzed by flow cytometry. Proliferation index for each individual was calculated as: % proliferation of Mtb-stimulated cells – % proliferation of cells cultured with medium.

SLAMF1 expression analysis

Fresh PBMC were stimulated with Mtb-Ag for 5 d and then cells were collected. Afterward, the expression of SLAMF1⁺ (BioLegend, 306306) on CD3⁺ lymphocytes was determined by flow cytometry.

IFN-γ measurement

Fresh PBMC were stimulated with Mtb-Ag for 48 h and then supernatants were obtained. The levels of IFN-γ were evaluated in supernatants using a commercial ELISA kit (BioLegend, 430101). The index for each individual was calculated as: (pg/ml IFN-γ after Mtb-Ag stimulation)/(pg/ml IFN-γ after culturing with medium).

Antigen

In vitro stimulation of cells was performed with a cell lysate from the pathogenic Mycobacterium tuberculosis strain H37Rv, prepared by mechanic disruption (referred to as Mtb-Ag) and employed at a final concentration of 10 μg/ml (obtained from BEI Resources, NIAID, NIH: Mycobacterium tuberculosis, Strain H37Rv, whole cell lysate, NR-14822).

Cell preparations and culture conditions

PBMC were isolated by centrifugation over Ficoll-Hypaque (GE Healthcare, 17-1440-03). Furthermore, neutrophils were purified from the bottom phase by using dextran sedimentation (Sigma, 31392) and hypotonic lysis.²³ Purity and viability of neutrophils was tested as previously described.²⁴ In fact, Fig. S1A shows the gating strategy for the analysis of flow cytometry experiments. In addition, we analyzed CD14 and CD56 expression to determine monocyte and NK cell contamination (Fig. S1B). Cells were suspended at 2 × 10⁶ /ml in RPMI 1640 (Gibco, 11875085) supplemented with L-glutamine (2 mM; Sigma Aldrich, G6392), penicillin (100 U/ml; Gibco, 15,140,122), streptomycin (100 µg/ml; Gibco, 15,140,122), and 10% fetal bovine serum (FBS; NATOCOR) and used immediately after isolation. For human Neutrophil Extracellular Traps (NETs) detection, RPMI without phenol red was used (Gibco, 11835030).

Flow cytometry

Staining of cell surface receptors

To determine the expression of immune receptors on neutrophils, cells were stained for surface expression of the diverse markers with fluorophore-conjugated antibodies against CD150 (SLAMF1, BioLegend, 306306), CD95 (Fas, BioLegend, 305606), CD54 (ICAM-1, BioLegend, 353112), CD182 (CXCR2, BioLegend, 320706), CD69 (BioLegend, 310906), CD274 (PD-L1, BioLegend, 393609), CD11b (BioLegend, 101208 and 301330), CD14 (Biolegend, 301804); CD56 (BD Pharmingen, 555517).

Intracellular staining of endogenous LC3

Neutrophils were washed with PBS and incubated with paraformaldehyde (PFA) 4% for 30 min at room temperature (RT). Afterward, permeabilization was performed by using PBS containing 0.05% saponin during 1 h. Then, cells were blocked with blocking solution at 4ºC. After 16 h, cells were incubated with rabbit anti-human LC3B antibody (Cell Signaling Technology, 2775S) for 60 min at RT, rinsed with PBS, incubated with anti-rabbit antibody (Alexa Fluor^® 488 Goat Anti-Rabbit IgG H&L; Abcam, 150077) for 45 min at RT and finally washed twice with PBS.

Detection of apoptotic cells by Annexin-V staining

Apoptosis was assessed by phosphatidylserine detection on the cell membrane using FITC-conjugated Annexin-V (BioLegend, 640906) in combination with a Viability Kit, following the manufacturers’ instructions (Zombie NIR™ Fixable Viability Kit, BioLegend, 423105). Samples were then analyzed by flow cytometry.

ELISA

Cell-free culture supernatants from neutrophils stimulated or not with Mtb-Ag were obtained to evaluate cytokine levels by ELISA. Therefore, IL-8 (BD Biosciences, 555244) and CXCL10 (IP-10; BioLegend, 439907) secretion levels were measured by ELISA following the manufacturers’ instructions. Moreover, PBMC stimulated or not with Mtb-Ag were also obtained to determine IFN-γ and TNF (BD OptEIA, 555212) levels by ELISA according to the manufacturers’ instructions.

ROS determination

After 60 min of stimulation with Mtb-Ag, neutrophils were incubated with 2ʹ,7ʹ-dichlorofluorescein diacetate (DCFDA, 10 µM; Invitrogen, D399) for 15 min at 37 °C. Finally, DCFDA fluorescence was evaluated by flow cytometry to determine the production of ROS.

Confocal microscopy

Immunofluorescence was performed as previously described.²⁴ Briefly, after indicated treatment, neutrophils were fixed and permeabilized with cold methanol, then blocked and LC3 primary antibody was added (Cell Signaling Technology, 2775S) and incubated for 16 h at 4 °C. Then, cells were washed with blocking buffer and incubated with the secondary antibody (Alexa Fluor^® 488 Goat Anti-Rabbit IgG H&L; Abcam, 150077) for 2 h at room temperature.

Human NETs detection

Determination of NETs was carried out as described previously.²⁵ Briefly, after 4 h of Mtb-Ag stimulation, a rapid treatment with DNase I was performed, and the supernatants were analyzed for DNA release using SYTOX green (Invitrogen, S34860) and myeloperoxidase (MPO) activity using 3,3′,5,50-tetramethylbenzidine (TMB) reagent. Images of NETs (DAPI-DNA staining) were acquired by confocal microscopy (60X oil objective).

Bacterial growth conditions

Mycobacterium tuberculosis H37Rv (Mtb H37Rv) was grown in Middlebrook 7H9 broth (BD, 271310) or on 7H11 agar (BD, 283810) with 0.05% Tween 80 (Biopack, 2003.08), 0.2% glycerol, and 10% albumin-dextrose-catalase supplement (ADC; 50 g/l BSA, 8.5 g/l NaCl, 20 g/l dextrose, 0.03 g/l catalase). Cultures were harvested at an exponential growth phase at 37 °C. Then culture was washed with PBS, centrifuged for 10 min at 2000 rpm, and resuspended in PBS. Bacterial growth in all experiment involving the pathogenic strain Mtb H37Rv was performed in BSL3 security cabinets at the Instituto de Agrobiotecnología y Biología Molecular, INTA-CONICET, Buenos Aires, Argentina.

In vitro infection of neutrophils with Mtb H37Rv

Neutrophils (2.5 × 10⁶ cells/mL) were infected with Mtb H37Rv (MOI = 5) at 37 °C with 5% CO₂ for 2 h. Afterward, cells were washed with PBS to remove extracellular bacteria. The cells were then incubated at 37 °C with 5% CO₂ for 24 h in a 24-well plate. Infected neutrophils were washed with PBS and then were lysed by addition of 0.05% Triton-X in PBS. Lysates were 10-fold serially diluted in PBS and plated (in triplicate) on Middlebrook 7H11 agar plates (BD Biosciences, 283810). After 3 to 4 wks at 37 °C, colonies were counted from dilutions yielding 10 to 100 visible colonies.

Statistical analysis

Statistical analysis was performed with the GraphPad Prism software 6. Mann-Whitney U test and Wilcoxon rank sum test were used for the analysis of unpaired and paired samples respectively. Fisher exact test was used to analyze the association between 2 qualitative variables. P values of < 0.05 were considered statistically significant.

Results

Demographic and clinical variables of patients with active TB included in the present study are listed in Table 1. Out of the 74 patients with active pulmonary tuberculosis that were included in this study; 50 were classified as LR patients, while 24 were identified as HR patients. No differences in terms of age, sex, ethnicity, leukocytes, monocytes and neutrophils counts were detected between LR and HR patients. However, HR patients had significantly higher percentages of total lymphocytes as compared to LR patients. On the other side, LR individuals had been ill longer (months of disease evolution since symptoms began until patient hospital admission) than HR individuals.²⁶ In addition, increased numbers of bacilli in sputum (AFB) and more severe pulmonary lesions were observed in LR patients as compared to HR patients. Besides, we detected an increase in Neutrophil/Lymphocyte Ratio (NLR) in individuals with more severe TB, as compared to subjects with strong immunity against Mtb (Table 1), in line with previous reports.²⁷ Although both groups of TB patients presented hypoalbuminemia (albumin ≤ 3.5 g/dL), the lowest albumin levels were found in LR patients’ blood (Table 1). Accordingly, lower plasma protein levels have been shown to correlate with worse clinical outcomes in critically ill patients.^28,29 Furthermore, Mtb-Ag stimulated PBMC from HR TB patients secreted significantly higher levels of TNF as compared to LR TB patients (Fig. S2A). All these findings are consistent with greater severity in LR TB patients.

Fridlender and collaborators adopted the N1/N2 terminology for the first time by describing that Tumor-associated neutrophils (“TANs”), like macrophages, may acquire an N1 anti-tumorigenic/pro-inflammatory or N2 pro-tumorigenic/anti-inflammatory phenotypes and might be classified according to their activation state, cytokine production and effects on tumor growth.^5,30 Initially, we analyzed the expression of several typical N1/N2 surface markers.³¹ Therefore, we studied the basal levels of the receptors ICAM -1 and Fas (as typical N1 markers) and CXCR2 (as typical N2 marker), on recently purified neutrophils from TB patients. As shown in Fig. 1, HR TB patient’s neutrophils displayed significantly higher levels of ICAM-1 and Fas as compared to those from LR TB patients (Fig. 1A, B). In addition, neutrophils from HR patients expressed significantly reduced CXCR2 levels as compared to neutrophils from LR (Fig. 1C).

We next assessed whether the expression of these markers could be modulated by Mtb-Ag stimulation. We observed that ICAM-1 and Fas expression were not modified by Ag-stimulation on LR TB patients’ neutrophils. In contrast, Mtb-Ag stimulation markedly upregulated the expression of both ICAM-1 and Fas on neutrophils from HR TB patients (Fig. 1D, E). Interestingly, the levels of ICAM-1 and Fas on Mtb-Ag-stimulated neutrophils from HR TB patients were significantly greater than the levels of those receptors on LR’s neutrophils (Fig. 1D, E). Besides, Mtb-Ag stimulation significantly decreased CXCR2 expression on neutrophils from both groups of TB patients, although CXCR2 levels remained higher in neutrophils from LR TB patients as compared to HR TB patients (Fig. 1F).

Considering the described activated state of N1-polarized neutrophils as compared to N2-like neutrophils,³¹ we next investigated additional markers implicated in cell activation. Thus, we first analyzed the levels of the signaling lymphocytic activation molecule family member 1 (SLAMF1), a type I glycoprotein belonging to the SLAM family of immune receptors that participates in neutrophil autophagy during active TB²⁴. As shown in Fig. 2, we observed a differential expression of SLAMF1 on neutrophils from HR and LR TB patients. Although both groups of patients increased the percentage of SLAMF1⁺ cells in response to Mtb-Ag, we detected significantly more SLAMF1⁺ neutrophils in HR TB patients than in LR individuals (Fig. 2A). Similar results were obtained by measuring other activation markers like the CD69 glycoprotein and the early activation receptor CD11b (Fig. 2B, C). Additionally, programmed death ligand 1 (PD-L1), was shown to be upregulated upon activation in bone marrow cells, lymphocytes, normal epithelial cells, cancer cells and neutrophils.^23,32 It was demonstrated that PD-L1 expression might be induced by low doses of IFN-γ in monocytes/macrophages.³³ Moreover, it has been reported that IFN-γ causes PD-L1 upregulation in ovarian cancer cells, which is responsible for disease progression.³⁴ Therefore, we next investigated the expression of PD-L1 on neutrophils from HR and LR TB patients. In contrast to our findings on the expression of SLAMF1, CD69 and CD11b, Mtb-Ag stimulation upregulated PD-L1 levels only on neutrophils from HR TB patients (Fig. 2D).

Ohms et al. reported that N1 and N2 populations of neutrophils were primarily defined by their functional phenotype since no specific cell surface markers had been identified.³¹ Therefore, despite the differences we detected in the expression of several receptors in neutrophils from the two groups of TB patients, we then analyzed the secretion of cytokines and chemokines associated with N1 and N2 phenotypes. To do so, we obtained cell-free culture supernatants from in vitro Mtb-Ag stimulated neutrophils. As shown in Fig. 3, we observed that neutrophils from both groups of TB patients secreted IL-8 and CXCL10 in response to Mtb-Ag. However, Ag-stimulated neutrophils from LR TB patients produced significantly greater levels of IL-8 as compared to HR TB patients (Fig. 3A). In contrast, Ag-stimulated neutrophils from HR TB patients secreted significantly higher amounts of the pro-inflammatory chemokine CXCL10 as compared to LR TB patients (Fig. 3B).

At the site of infection, neutrophils eliminate invading pathogens using a combination of ROS, cytotoxic granule components, and NETs.³⁵ Therefore, we also investigated the effect of Mtb-Ag on NETs formation and ROS production in both groups of TB patients. Initially, we analyzed free DNA on supernatants of purified Mtb-Ag stimulated neutrophils. Figure 4A showed that after 4 h of Mtb-Ag stimulation, neutrophils from HR patients were able to release significantly more DNA as compared to neutrophils from LR patients. This increase in DNA expulsion was accompanied by a higher secretion of the MPO enzyme (as a parameter of azurophilic granules release) by neutrophils from HR TB patients (Fig. 4B). Together, our findings would be in line with reports indicating that NET components, including neutrophil elastase, myeloperoxidase (MPO), and cell-free DNA, cause a pro-inflammatory response.¹⁴

To explore other microbicidal mechanisms of neutrophils, we analyzed the generation of ROS by neutrophils from HR and LR TB patients. Previously, we had reported that Mtb-Ag triggered ROS production in healthy donors' neutrophils.²⁴ Here, we demonstrated a significant augment in ROS secretion in Mtb-Ag stimulated neutrophils from HR TB patients as compared to LR TB patients (Fig. 4D). Our data are in agreement with previous reports showing that neutrophil phenotypes and their capacity to generate ROS are associated with disease severity in patients with active TB.²⁰

Autophagy participates in several effector functions of neutrophils, such as NETs release, granule formation, cytokine and chemokine production, degradation, bactericidal activity and inflammation control. Aiming to evaluate whether autophagy could be differentially modulated by Mtb-Ag in neutrophils from both groups of TB patients, we analyzed the levels of membrane-anchored LC3, a widely used marker of autophagy. We observed that stimulation with Mtb-Ag significantly augmented the percentage of LC3B⁺ neutrophils in both groups of TB patients as measured by flow cytometry (Fig. 5A). Nevertheless, HR patients displayed markedly higher autophagy levels as compared to LR TB patients (Fig. 5A, C). Furthermore, we were able to confirm by confocal microscopy an accumulation of localized LC3B foci in Mtb-Ag-treated neutrophils from HR and LR patients with active TB (Fig. 5B). Taken together, our findings suggest that neutrophils from HR patients could be displaying an N1-like pro-inflammatory phenotype, whereas neutrophils from LR patients would be presenting an N2-like anti-inflammatory phenotype.

Finally, to investigate the killing capacity of neutrophils from TB patients against Mtb, we infected neutrophils from HR and LR TB patients with the pathogenic H37Rv Mtb strain for 24 h. Afterward, we determined the bacterial load by counting colony-forming unit (CFU) in the infected cells. After uptake of bacteria, we observed that neutrophils from both groups of TB patients were able to phagocytize Mtb H37Rv strain to the same extent (Fig. 6A). Nevertheless, neutrophils from HR’s patients showed significantly lower CFUs as compared to LRs’ neutrophils. These data indicate a decreased Mtb-killing activity by neutrophils from TB patients displaying the weakest immunity (LR TB), cells that exhibit an anti-inflammatory N2-like phenotype according to our present findings (Figs. 1–5). In contrast, neutrophils from TB patients with strong immunity (HR TB) and showing a pro-inflammatory N1-like phenotype (Figs. 1–5) had a greater microbicidal capacity (Fig. 6A).

It has been demonstrated that Mtb infection reduces the viability of neutrophils.³⁶ Furthermore, the clearance of dead cells becomes crucial for resolving immunopathology during active TB.³⁷ Accordingly, both humans and murine models connect neutrophil necrosis with pathogen virulence, suggesting that this pattern of cell death would be associated with disease severity.³⁸ In contrast, the elimination of cells by apoptosis during active TB is considered advantageous for the host's defense.³⁹ Therefore, we investigated the viability of Mtb-infected neutrophils in the 2 population of TB patients. We observed that the proportion of live cells (Annexin-V⁻/viability stain⁻) was equal in both groups of TB patients (Fig. 6B). However, our findings showed a markedly higher percentage of necrotic neutrophils in LR patients (Annexin-V⁺/viability stain⁺) as compared to HR (Fig. 6B). Interestingly, we also found that HR patients displayed an evident greater percentage of early apoptotic cells (Annexin-V⁺/dead stain⁻) as compared to LR (Fig. 6B). Thus, considering that LR TB patients present more severe disease,²⁶ N2 neutrophils in these patients might reflect an advanced senescent state that could explain why these cells are less responsive and less effective in controlling Mtb growth (Fig. 6A). On the other side, HR TB patient’s N1 neutrophils containing Mtb would be in an apoptotic state, and therefore they might be efferocytosed by macrophages benefiting the host.

Discussion

We previously reported that common clinical parameters analyzed in patients with active TB in Argentina paralleled immunological parameters studied in those individuals, allowing the classification of patients in HR and LR individuals.²² Later on, we extended those findings, demonstrating that the proportion of CD4⁺IFN-γ⁺IL-17⁺ lymphocytes expanded in response to Mtb-Ag might be associated with the severity of the disease.²⁶ More recently, we reported that IL-17A augments autophagy of Mtb-infected monocytes only in HR TB patients.⁴⁰ Moreover, here we observed that stimulation of PBMC with sonicated Mtb H37Rv-Ag significantly increased not only IFN-γ but also TNF production from HR tuberculosis patients. However, Ag stimulation only slightly increased the levels of these cytokines in LR tuberculosis patients (Fig. S2). TNF plays a major role in the initial and long-term control of tuberculosis, in part by augmenting T cell responses through promoting macrophage phagolysosomal fusion and apoptosis.⁴¹ We previously also showed that both HR and LR patients produce similar IL-10 levels against Mtb-Ag.⁴² Then, HR patients that secreted IL-10 but produced high levels of IFN-γ and TNF against Mtb-Ag, might be creating a Th1-like pro-inflammatory microenvironment, whereas in LR patients, Mtb-Ag induced IL-10 production but very low levels of IFN-γ and TNF, generating a predominantly Th2-like anti-inflammatory environment. In the present work, we expanded our previous studies by characterizing neutrophils from HR and LR TB patients. Interestingly, our findings indicate that neutrophils from HR patients seem to display a pro-inflammatory N1-like phenotype, whereas neutrophils from LR patients appear to show an anti-inflammatory N2-like phenotype.

Studies of N1/N2 neutrophil subsets in cancer and other diseases such as myocardial infarction,⁹ periodontitis,¹⁴ and thermal injury,¹² have allowed the differentiation of two cell subpopulations displaying distinctive characteristics. Briefly, in the pro-inflammatory N1 phenotype, neutrophils are highly activated and show high expression of ICAM-1 and Fas on their surface.³¹ Moreover, N1 neutrophils exhibit enhanced cytotoxicity against tumor cells and possess an immune-activating potential that boosts the production of high amounts of TNF and CXCL10.^14,43 Importantly, N1 neutrophils induce NETosis, high levels of ROS production and reduce arginase expression.^5,44 In contrast, anti-inflammatory N2 neutrophils are less activated, express high levels of CXCR2 and secrete high amounts of IL-8.^5,6,31

Remarkably, we detected significantly lower basal levels of CXCR2 in neutrophils from patients with robust immunity against Mtb (HR individuals). Furthermore, upon Mtb-Ag stimulation, HR patients showed significantly less CXCR2 expression as compared to LR patients (Fig. 1). CXCR2 is a key stimulant of immune cell migration and recruitment, especially of neutrophils. Reducing excessive neutrophil accumulation and infiltration could prevent prolonged tissue damage in inflammatory disorders.⁴⁵ Accordingly, patients with more severe TB displayed higher numbers of neutrophils in their lungs as compared to patients with mild/moderate disease.¹⁷ Therefore, the high levels of CXCR2 detected in fresh neutrophils from LR TB patients could be stimulating the migration of these cells to their lungs and causing the characteristic damage observed in these patients. Besides, CXCR2 is rapidly internalized after stimulation,^46,47 which could explain at least in part, the decrease in the levels of CXCR2 in Mtb-stimulated neutrophils from TB patients (Fig. 1).

We also observed that TB HR patients presented significantly higher levels of Fas and ICAM-1 as compared to LR (Fig. 1). The adhesion-glycoprotein ICAM-1 has several implications on neutrophils’ functions. For example, Vignarajah et al. found that ICAM-1^high neutrophils display elevated Staphylococcus aureus phagocytic capacity.⁴⁸ In addition, Nourshargh et al. demonstrated that ICAM-1 expressing neutrophils generate higher amounts of ROS.⁴⁹ In our study, HR TB patient’ neutrophils display significantly elevated ICAM-1 levels, a higher microbicidal capacity –without changes in the phagocytic ability as compared to LR patients and higher ROS generation after Mtb-Ag stimulation. Also, ICAM-1 interacts with CD11b/CD18 triggering the adhesion between stimulated neutrophils and the endothelial cells.⁵⁰ Here we observed that in addition to the higher expression of ICAM-1, neutrophils from HR TB patients also display greater levels of CD11b compared to LR patients (Fig. 2C), which could have implications on neutrophil adhesion and extravasation.

Fas (CD95/Apo-1) -initially described as a death receptor- is involved in several nonapoptotic cell responses that promote inflammation and carcinogenesis.⁵¹ Martin-Villalba et al. demonstrated that Fas is required for myeloid cell recruitment in in vivo animal models of inflammation.⁵² Moreover, CD95-deficient myeloid cells exhibit impaired bacterial clearance in an animal model of sepsis.⁵² In the same line, Xuetao Cao et al. reported that Fas ligation on dendritic cells promotes chemokine production, endocytosis by neutrophils and T cell activation.⁵³ Here, we found that HR TB patient’s neutrophils present higher levels of Fas as compared to LR TB patient’s neutrophils, which support the N1 pro-inflammatory profile characterization.

SLAMF1 is a costimulatory molecule involved in host immunity regulation of innate and adaptive responses. We have demonstrated that SLAMF1 contributes to Th1 cytokine responses in human TB.²² However, SLAMF1 is also a bacterial sensor that helps to remove Gram-negative bacteria by macrophages.⁵⁴ Furthermore, we demonstrated that stimulation of SLAMF1 promotes neutrophil autophagy induced by Mtb, suggesting that it might participate in the anti-mycobacterial human response.²⁴ Therefore, SLAMF1 emerges as an attractive target for host-directed therapies.²⁴ In the present work, we further characterized neutrophil phenotypes in HR and LR TB patients by analyzing the expression of SLAMF1. Interestingly, we observed significantly higher SLAMF1 levels on neutrophils from HR individuals as compared to those from patients with more severe TB (Fig. 2). So, we hypothesized that stimulation of SLAMF1 might augment the functions of neutrophils from LR TB patients, increasing the ability of these patients to eliminate mycobacteria. Other activation markers like CD69 and CD11b were also upregulated in Mtb-Ag stimulated neutrophils from HR TB patients, clearly indicating an increased activation state of these cells as compared to neutrophils from LR patients. Similar to our data, Pokkali et al. previously demonstrated that neutrophils from TB patients displayed elevated levels of CD69 after 3 h of Mtb infection.⁵⁵ The present findings are in line with studies in other pathologies that reported an activation state of pro-inflammatory N1 neutrophils.¹⁴

PD-L1 is not only a central receptor in mediating tumor immune escape, but it has been also demonstrated to regulate the development of inflammation.³² We previously demonstrated that Mtb-Ag stimulation significantly augmented the levels of PD-L1 on T cells from HR and LR TB patients, though significantly higher levels of the receptor were detected in HR patients compared with LR individuals. Moreover, blockage of PD-Ls: PD-L1, PD-L2 enhanced the specific degranulation of CD8⁺ T cells and the percentage of specific IFN-γ-producing lymphocytes against the pathogen, demonstrating that the PD-1: PD-Ls pathway inhibits T cell effector functions during active Mtb infection.⁵⁶ Furthermore, we and others have previously shown that PD-L1 is upregulated in human neutrophils.^23,32 Moreover, Zhu et al. demonstrated that neutrophil specific knockout or blockage of PD-L1 reduced lung injury in acute respiratory distress syndrome (ARDS), suggesting that anti-PD-L1 antibody administration may be a promising therapeutic strategy.³² Our present findings demonstrate that Mtb stimulation upregulated PD-L1 levels only on neutrophils from HR TB patients (Fig. 2D). Therefore, we postulate that blocking PD-L1 on neutrophils from TB patients might increase these cell’s effector functions such as in T cells from TB patients⁵⁶ and in patients with ARDS.³²

By analyzing the secretion of cytokines and chemokines related to N1/N2 phenotypes of neutrophils we observed that Ag-stimulated neutrophils from patients with robust immunity against Mtb (HR individuals) secreted higher concentration of CXCL10 as compared to LR patients (Fig. 3). CXCL10 is involved in trafficking of Th1 lymphocytes to areas of inflammation where it binds to CXCR3. CXCL10 differs from other chemokines as it targets lymphocytes and monocytes specifically and has no activity on neutrophils.⁵⁷ High levels of CXCL10 have been found in vivo in lymph nodes and lung tuberculous granulomas⁵⁸ and in pleural effusions and plasma of TB infected patients.^59,60 Thus, our findings could be indicating that the levels of CXCL10 secreted by N1-like neutrophils from HR patients, might allow the recruitment of Th1 cells to the site of infection, contributing to granuloma formation and disease containment.^61–63 These functions could be impaired in LR patients, whose N2-like neutrophils secrete low amounts of CXCL10, experimenting difficulties to contain the pathogen and suffering a more severe disease.

On the other hand, neutrophils from patients with weak immunity against Mtb (LR individuals) produce elevated amounts of IL-8 as compared to HR patients (Fig. 3). IL-8 is a chemokine that mainly induces neutrophil migration but also plays a central role in the recruitment of other leukocytes to the areas of granuloma formation. Therefore, the high levels of IL-8 produced by LR TB patients’ neutrophils might be associated with the extensive lung damage observed in these individuals. Accordingly, higher IL-8 plasma levels were detected in patients who died from TB as compared to survivors.⁶⁴

Compelling evidence has demonstrated that neutrophils play a crucial role in the immune response and that NETs can capture and kill pathogens, mainly big microorganisms such as Candida albicans and M. bovis aggregates, which result difficult to swallow.^65,66 Besides, it has been reported that although Mtb triggers NET formation that trapped mycobacteria, they are unable to kill the pathogen.^67,68 Nevertheless, apoptotic neutrophils can impart macrophages with molecules with antimicrobial capacity,⁶⁹ helping to kill intracellular pathogens such as Mtb.⁷⁰ In the present work we demonstrated that neutrophils from HR TB patients produced significantly higher NETosis as compared to LR patients (Fig. 4). Since HR TB’s neutrophils display a pro-inflammatory N1-like phenotype, the augmented NETosis observed in patients with strong immunity to Mtb could ultimately contribute to the control of the bacteria.

In addition, the production of ROS during TB plays a fundamental role, as demonstrated in patients with chronic granulomatous disease who present mutations that alter NADPH oxidase and lead to recurrent or disseminated TB.^71,72 In this work, we observed that HR TB patients presented significantly higher production of ROS as compared to LR patients (Fig. 4C). Among other reasons, the superior production of ROS by these patients’ neutrophils could be related to the greater expression of SLAMF1 (Fig. 2), as we previously demonstrated.²⁴ In fact, SLAMF1 is able to recruit the class III phosphatidylinositol kinase Vps34, which through the phosphorylation of the lipid phosphatidyl-3'-phosphate (PI3P), allows the assembly of the classical phagocytic NADPH oxidase (Nox2).^73–75 The increase of ROS either by signaling through SLAMF1 or other mechanisms is of huge importance considering that phagocytes play a key role in bacterial elimination. Furthermore, ROS signaling impacts multiple processes, including migration, adhesion, and chemotaxis.⁷⁶

Additionally, we previously demonstrated that SLAMF1 activation in Mtb-Ag-stimulated neutrophils augments endogenous LC3B aggregation in a ROS-dependent manner.²⁴ Here, we showed that neutrophils from less severe TB patients (HR) display higher SLAMF1 and LC3B levels as compared to LR patients (Figs. 2 and 5). Considering that autophagy protects against excessive inflammation during Mtb infection,⁷⁷ the diminished levels of autophagy detected in neutrophils from patients with weak immunity against Mtb might be related to the typical detrimental inflammatory reactions that occur in these patients that suffer a more severe TB.

Neutrophils from HR patients (N1-like phenotype) were able to eliminate pathogenic Mtb H37Rv more effectively than those from LR patients (N2-like phenotype). In line with our findings, Ohms and collaborators demonstrated that N2-polarized neutrophils showed a significantly reduced ability to eliminate Leishmania donovani, allowing the parasite to persist in the phagocyte.³¹ In addition, Tsuda et al. polarized murine neutrophils employing a systemic inflammatory response syndrome (SIRS) model. The authors showed that mice adoptively transferred with PMN-I (N1-like) cells resisted the infection with methicillin-resistant Staphylococcus aureus (MRSA), in sharp contrast with PMN-II (N2-like)-transferred animals that did not survive.¹²

High human neutrophil plasticity has been demonstrated since these cells were shown to acquire an anti-tumorigenic N1 or a pro-tumorigenic N2 phenotype according to culture conditions with polarizing cocktails.³¹ In the present work, we demonstrated phenotypical and functional differences between Mtb-stimulated neutrophils obtained from HR and LR TB patients without using polarizing cocktails. Still, considering that even fully differentiated mature neutrophils maintain their plasticity and can acquire typical phenotypic characteristics, anti-inflammatory N2-like neutrophils from LR TB patients might be polarized to a beneficial pro-inflammatory N1-like phenotype that exerts an augmented capacity to kill intracellular Mtb, ultimately contributing to improve the outcome of the disease.

During the innate immune response, neutrophils operate as part of the first responders against Mtb. Increased circulating IL-1 secreted by macrophages during Mtb infection stimulate neutrophil production via the IL-17-G-CSF axis.⁷⁸ Thus, neutrophils promote efficient Mtb clearance through phagocytosis and the secretion of destructive granules, contributing to innate resistance.⁷⁸ Then, phagocytosis of apoptotic neutrophils by macrophages during active tuberculosis led to destruction of intracellular bacteria. This finding represents a bridge of defense between neutrophils and macrophages in response to Mtb intracellular infection.⁷⁰ Moreover, during the adaptative immune response, neutrophils modulate communication with lymphocytes to promote a strong pro-inflammatory response and to mediate the containment of Mtb through the production of granulomas. Th1 cells secrete IFN-γ which upregulate macrophage response to control Mtb infection, whereas Th17 cells induce neutrophilic inflammation and tissue damage, serving as a mediator of Mtb pathology.⁷⁹ Neutrophils might play a protective role in early granulomas by functioning to eliminate Mtb, but a pathological role in late granulomas by damaging adjacent tissue. Generally, neutrophils play a role in the progression of TB infection, with prognosis dependent on neutrophil viability, the severity of the infection and location of infection (pulmonary or systemic). More severe infections occur especially when neutrophils have not been properly cleared by macrophages.⁸⁰ In line with the above reports, our present findings showed that LR TB patients, individuals with more severe disease,²⁶ display anti-inflammatory N2 like neutrophils that might contribute to TB progression. Interestingly, we observed that LR TB patients display an enhanced percentage of necrotic neutrophils (Fig. 6B), what could difficult the proper clearance by macrophages and contribute to tissue damage. On the other hand, HR TB patients, individuals with strong immunity against Mtb, show pro-inflammatory N1 like neutrophils that could promote efficient Mtb clearance (Fig. 6A), contributing to innate resistance. Furthermore, HR TB patients display an augmented proportion of early apoptotic neutrophils (Fig. 6B), what might contribute to phagocytosis by macrophages leading to destruction of Mtb.

In summary, in the present work, we describe for the first time to our knowledge, the existence of N1/N2 neutrophil subpopulations in TB patients with different immune response to Mtb and disease severity. Further studies on the diversity of neutrophil subpopulations during Mtb infection is crucial to identify new targets to be used in host-directed therapies.

Acknowledgments

We thank Eva Danti for providing technical support. We also thank Guillermo Piazza for his constant support and technical assistance. We also acknowledge Nancy Tateosian and Nicolás Amiano for valuable comments and suggestions.

Abbreviations

AFB: acid-fast bacilli

CFDA-SE: Carboxyfluorescein Diacetate Succinimidyl Ester

DCFDA: 2ʹ,7ʹ-dichlorofluorescein diacetate

FBS: fetal bovine serum

HR: High responder

IFN-γ: interferon gamma

IL-8: Interleukin 8

LR: Low responder

MPO: myeloperoxidase

Mtb-Ag: Mycobacterium tuberculosis, Strain H37Rv, whole cell lysate

NETs: neutrophils extracellular traps

Number: N

PBMC: Peripheral Blood Mononuclear Cells

ROS: reactive oxygen species

SLAMF1: signaling lymphocytic activation molecule family member 1

TB: Tuberculosis

Author contributions

M.P.M. and V.E.G. designed the study. M.P.M. and C.M. performed the experiments. M.P.M. and C.M. were responsible for processing samples and performing data management and analysis. F. Blanco and F. Bigi provided methodological support for some experiments. L.C., R.M., A.R.M., G.C. and D.J.P. were in charge of patients’ diagnosis, obtained the blood samples and contributed with the analysis of the clinical data. M.P.M. and V.E.G. wrote the manuscript. M.P.M., C.M., J.M.P., and V.E.G. have contributed to data interpretation and revision of the report. All authors have read and approved to the published version of the manuscript.

Supplementary material

Supplementary material is available at The Journal of Immunology online.

Funding

This work was supported by the Agencia Nacional de Promoción Científica y Tecnológica (PICT-2019-01617 to V.E.G; PICT-2020-03211 to M.P. M.; PICT-2021-367 to V.E.G.) Secretaría de Ciencia y Técnica, Universidad de Buenos Aires (20020220300133BA to V.E.G.).

Conflicts of interest

The authors reported no potential conflict of interest.

Data availability

The data underlying this article are available in the article and in its online supplementary material.

References

Author notes