Elevated Natural Killer Cell-Mediated Cytotoxicity Is Associated with Cavity Formation in Pulmonary Tuberculosis Patients

Cavitation is a major pathological feature of pulmonary tuberculosis (TB). The study is aimed at investigating the mechanism of natural killer (NK) cells participating the cavity formation during Mycobacterium tuberculosis (MTB) infection. Human peripheral blood samples were donated by pulmonary TB patients with cavity or not. Real-time quantitative PCR and enzyme-linked immunosorbent assay were performed to analyze the expression of cytokines secreted by NK cells. And the cytotoxicity of NK cells was compared between two groups. Our data showed that NK cells were more abundant in cohorts of cavity. Increased abundance of granzyme A and granzyme B was observed in culture supernatants of NK cells isolated from cavitary TB patients, which also resulted in a higher level of nonviable MTB-infected monocytes. Our data firstly demonstrates that NK cells participate in cavity formation in pulmonary TB patients. The elevated level and increased cytotoxicity of NK cells accelerate the cavitary formulation.


Introduction
Tuberculosis (TB), caused by Mycobacterium tuberculosis (MTB) complex, is currently the leading cause of death from a single infectious pathogen [1]. In 2018, this disease was responsible for an estimated 10.0 million incident cases and 1.4 million deaths worldwide [1]. Cavitation is a major pathological feature of human pulmonary TB and is the site of very large bacillary burden [2,3]. The presence of pulmonary cavitation is associated with poor treatment outcomes, such as prolonged sputum culture conversion, relapse after treatment, and acquisition of drug resistance [4,5]. In addition, the patients with pulmonary cavity are more infectious than those with noncavitary disease, recognized as a principle source for TB transmission in the community [2,6]. The contribution of cavitation to the increased risk of person-toperson TB transmission makes it an important target for therapeutic invention [7].
Evidence from clinical cohorts demonstrated that the rates of cavitary tuberculosis at the time of diagnosis range from 29% to 87% [8][9][10]. A better understanding of the pathogenesis of cavity formation is of great significance in reduction of adverse outcome for pulmonary TB patients and prevention of TB transmission. There is a widely held view that cavitation is originated from granulomas coalesce and break down via liquefactive necrosis [11,12]. Thus, the complex interaction between tubercle bacilli and host immunity constitutes the underlying mechanisms for cavity formation. In this aspect, the uncontrolled immune response against MTB leads to clearance of MTB-infected cells, accompanied by accelerated tissue and organ damage. Previous studies demonstrated that the T-helper-2 cytokine adaptive immune response predominated in cavitary TB patients [13,14], indicating that the attenuation of the hostprotective regulatory T cell response reduces their ability to control multiplication of MTB and leads to cavity formation. However, investigators still debate the directionality of this association, given that pulmonary TB patients with cavitary disease had higher amounts of the proinflammatory cytokines in bronchoalveolar lavage, such as tumor necrosis factor-(TNF-) α, interleukin-(IL-) 6, and IL-1β [15,16]. Therefore, an alternative hypothesis proposes that the cavitation is the result of a progressive immune response to MTB despite little is known.
Natural killer (NK) cells are cytotoxic innate lymphocytes that produce high levels of proinflammatory cytokines in response to pathogens [17]. These cells have the capability to recognize and induce apoptosis in target cells via the release of perforin and granzymes [18]. Based on the existing data, both circulating and tissue-resident NK cells contribute to host protection, as well as possible adverse effects in the tissue. In view of the remarkable cytotoxic activity of NK cells, the question is raised as to whether their biased inflammatory response correlates with cavity diseases. The study is aimed at investigating the mechanism of NK cells participating the cavity formation during MTB infection.

Patients and Ethics Statement.
Human peripheral blood samples were donated by pulmonary TB patients with cavity or not and healthy uninfected donors. All pulmonary TB patients accepted primary treatment. The subjects with other acute or chronic diseases, including diabetes, hepatitis, tumor, and immune dysfunction, were excluded for analysis. Pulmonary TB patients were defined with positive sputum smears, cultures, and/or GeneXpert MTB/RIF (Cepheid, USA) for Mycobacterium tuberculosis with pulmonary imaging features. Fourteen pulmonary TB patients with cavity and eleven pulmonary TB patients with noncavity were recruited to analyze the cytokines secreted by NK cells after cultivation with MTB H37Rv protein lysates. In addition, nine pulmonary TB patients with cavity and eight pulmonary TB patients with noncavity were recruited for NK cell cytotoxicity assay.
The information of all individuals involved in the study was anonymized. Written informed consent was obtained from all individuals. All procedures performed on human samples were carried out according to the tenets of the World Medical Association's Declaration of Helsinki and approved by the Ethical Committee of Beijing Tuberculosis and Thoracic Tumor Institute.

Isolation and
Culture of NK Cells. Peripheral blood mononuclear cells (PBMCs) were isolated from peripheral blood samples of pulmonary TB patients with cavity or not by density gradient centrifugation using lymphocyte separation medium (TBD, Tianjin, China). NK cells were isolated with human NK cell isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's instructions, and negatively selected cells were >95% CD16+/CD56+ cells as measured by flow cytometer. Isolated NK cells were plated in 24-well plates (Corning, USA) at 2 × 10 5 /well in 500 μl of RPMI 1640 medium (HyClone, Waltham, USA) containing 10% fetal bovine serum (Gibco, USA) and incubated with 50 μg/ml MTB H37Rv (ATCC 27294) protein lysates for 24 h at 37°C in a humidified 5% CO 2 atmosphere.

Cytotoxicity of NK Cells against MTB H37Rv-Infected
Monocytes. As NK cells can release cytotoxic granule contents such as granzyme A, granzyme B, and perforin to trigger programmed cell death and kill intracellular pathogens [18], NK cells play an important role in TB progression. Flow cytometry analysis (Figure 3) showed that increased nonviable MTB-infected CD14+ monocytes were observed after coculture with MTB protein lysates treated NK cells which were isolated from pulmonary TB patients with cavity (8:99 ± 5:82) when compared with those from patients with-out cavity (4:43 ± 2:15, P = 0:01). The results indicate that NK cells from TB patients with pulmonary cavity may have more potent cytotoxic ability to trigger MTB-infected monocyte cell death.

Discussion
Despite innate immunity plays a major role in the host's response to MTB infection, more attention has been paid to macrophages that are recognized as the first line of defense to control multiplication of MTB [20]. In this study, our data firstly demonstrated that NK cells participate in cavity formation in pulmonary TB patients (Figure 4). On the one hand, pulmonary TB patients with cavitation had elevated NK cell levels compared with those without cavitation. Early studies suggested that NK cells rapidly expand as a consequence of viral infection. Similar results were observed animal model in response to MTB infection, suggesting an important role   Journal of Immunology Research for NK cells in controlling intracellular multiplication of tubercle bacilli. Current evidence demonstrates that the vast majority of NK cells in the lung tissue are circulating between the organ and peripheral blood [21]. Thus, the increased peripheral NK cell levels in cavitary patients could likely contribute not only to host protection but also to adverse effects in host tissue. This increase may reflect a higher output of new NK cells from the bone marrow and/or a high proportion of    Figure 3: Cytotoxicity of NK cells against MTB H37Rv-infected monocytes. Freshly isolated NK cells from nine pulmonary TB patients with cavity and eight pulmonary TB patients with noncavity were cocultured with MTB H37Rv-infected monocytes to analyze the ability of cytotoxicity. Representative dot plots show the proportion of nonviable CD14+ monocytes infected with MTB H37Rv after coculture with NK cells isolated from pulmonary TB patients with (a) cavity or (b) not. (c) Comparison of the proportion of nonviable MTB H37Rv-infected CD14+ monocytes after coculture with NK cells from pulmonary TB patients with cavity or not. P value was derived using a Mann-Whitney U test. 5 Journal of Immunology Research long-lived NK cells. There is strong evidence that interleukin-15 (IL-15) acts as an important cytokine not only to trigger NK cell proliferation but also to promote the survival of proliferating NK cells [22]. Thus, further studies are warranted to elucidate the potential effect of IL-15 on the NK cell homeostasis in pulmonary TB patients with cavity.
On the other hand, the increased cytotoxicity was noted in peripheral NK cells from cavitary patients. Despite tradi-tionally NK cells being considered as nonspecific components of innate immunity, this suggests an MTB-specific NK cell response, which may accelerate the immune activation against tubercle bacilli. This hypothesis is supported experimentally in the animal models that robust antigenspecific NK cell can be identified after both infection and vaccination [23,24]. Of note, the NK cells from these cavitary donors could induce higher expression of granzymes A and B rather than perforin upon stimulation with MTB antigens, which in turn promotes apoptotic cell death. It has long been recognized that perforin is essential for enabling granzyme-induced apoptosis. We speculate that the high transcriptional level of perforin at baseline warrants entrance of granzymes into targeted cells.
Another interesting finding of our study was that the CD8+ T cell level of cavitary patients was significantly lower than that of noncavitary patients. The low level of CD8+ T cells by flow cytometry suggested one possible interpretation that these peripheral immune cells may be recruited to the tissues of higher antigen concentrations. Early studies have described an important role for the CD4+ T cell in the pathogenesis of cavitation [13,14]. Decreased CD4+ lymphocytes were reported in patients with cavitary TB compared with noncavitary TB [13]. However, this dogma has been challenged by findings in patients coinfected with HIV that the deficiency of CD4+ T cells is accompanied by a low frequency of cavitary disease [25]. In addition, the comparative studies in animal models confirmed that the mice with little measurable hypersensitivity response do not form true caseous necrosis and cavitation in response to MTB infection. Taken together, the data from HIV patients and animal models suggests that the potent host immune response is essential for formulation of cavitary diseases, although the exact role of adaptive immunologic triggers and cascades are yet to be fully examined.
Our findings have important hints for clinical management of pulmonary TB patients to reduce the occurrence of cavitation. The disease severity of pulmonary TB is not only caused by increasing bacterial load, but also to hypersensitivity to MTB antigens. The excessive activation of NK cells increases the risk of cavitary TB diseases. The multiplication of peripheral NK cells could be used as a surrogate maker for predicting this stage of disease. In addition, given the two-edged sword role of NK cells in TB pathogenesis, the delicate regulation of their function which is aimed at limiting the adverse overactivation of the immune system is necessary after elimination of intracellular tubercle bacilli. Monitoring NK cell dynamics will provide additional benefit in subsets of pulmonary TB patients with early-stage highrisk cavitation.
Our study suffered from several obvious limitations. First, despite enrollment of a large number of patients for determining immune cell composition, the cross-section design hampered the detailed analysis of the effect of disease duration on cavity formation, which majorly caused the lack of correlation between cavitation size and NK cell level. Second, only peripheral blood rather than bronchial lavage fluid was explored. Our findings thus require further validation by analysis of lymphoid compartments of bronchial lavage  Journal of Immunology Research fluid. Third, despite exploring the transcriptional levels of the major activation receptors for NK cells, no significant difference was noted between TB patients with cavity or noncavity. Further experiments are urgently required to elucidate the molecular mechanisms that trigger human NK cell-mediated cytotoxicity. Fourth, previous experimental data demonstrated that the virulence of MTB could influence the expansion of the NK cell population, which was not taken into consideration in our analysis. Finally, Th1/Th2 homeostasis is an important immune factor to protect host from MTB infection. Thus, it is worth immunophenotyping T helper cell subsets to investigate whether the percentage of Th1 or Th2 cells changed in the TB patients with cavity although the overall percentage helper T cells remained unchanged.
In conclusion, our data firstly demonstrates that NK cells participate in cavity formation in pulmonary TB patients. The elevated NK cell level and increased cytotoxicity of NK cells accelerate the cavitary formulation. These results underscore the need for immunotherapy interventions to limit the adverse overactivation of NK cells. Monitoring NK cell dynamics will provide additional benefit in subsets of pulmonary TB patients with early-stage highrisk cavitation.

Data Availability
The data used to support the findings of this study are included within the article.

Ethical Approval
This study was approved by the Ethics Committee of Beijing Chest Hospital, Capital Medical University, and the informed consent of all participating subjects was obtained.

Disclosure
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of manuscript.