OUP user menu

Differential upregulation of chemokine receptors on CD56+ NK cells and their transmigration to the site of infection in tuberculous pleurisy

Supriya Pokkali, Sulochana D. Das, Anbalagan Selvaraj
DOI: http://dx.doi.org/10.1111/j.1574-695X.2008.00520.x 352-360 First published online: 1 April 2009

Abstract

Chemokines and their receptors orchestrate leukocyte recruitment and confer immunity during Mycobacterium tuberculosis infection. The immunoregulatory and cytotoxic activities of natural killer (NK) cells are essential at the site of infection during tuberculous pleurisy. The frequency, subtypes, and expression of phenotype markers and chemokine receptors on NK cells were assessed by flow cytometry in tuberculous (TB) and nontuberculous (NTB) pleural fluid (PF). Chemotaxis was also shown in response to chemokines. A significant decrease in CD56dim with no change in CD56bright NK cells was observed, while a significant increase in activation markers and Toll-like receptors (TLRs) was observed on TB-PF CD56bright NK cells. Significantly increased expression of chemokine receptors CCR1, CCR2 and CCR7 on CD56bright and CCR5 on CD56dim NK cells was observed in the TB group. Transmigration of TB-PF NK cells was significantly high in response to IL-8, IP-10, MCP-1 and SLC. Transmigrated TB-NK cells showed a significant increase in CXCR2, CCR2 and CCR7 expression. The study suggests that CD56bright NK cells may recognize M. tuberculosis directly using TLRs, HLA-DR and express CD69 as an early activation marker. In addition, CC chemokines induce activation signals in chemokine receptors mediating differential NK cell migration to the site. Thus, NK cells act as first direct sensors and effectors in mycobacterial infection.

Keywords
  • tuberculous pleurisy
  • NK cells
  • chemokine
  • chemokine receptor

Introduction

A combination of innate and adaptive immune responses to Mycobacterium tuberculosis or its antigens at the pleural space represents an intense host response in clearing the pathogen. Tuberculous pleuritis serves as a good model to study the individual components of a protective host immune response at the site (Shimokata et al., 1991; Schierloh et al., 2005). Tuberculous effusions are self-limiting and known to resolve without therapy. However, the factors underlying the self-resolving nature of this disease are still unknown.

Although T cells are predominant in tuberculous pleural effusions, the contribution of other recruited immune cells cannot be ruled out. It is well documented that natural killer (NK) cells do accumulate at inflammatory sites and exert their effector functions (Okubo et al., 1986,; 1987,; 1988). Many studies have highlighted the role of NK cells in cytotoxicity and direct antigen recognition during pulmonary tuberculosis (Schierloh et al., 2005,; 2007). Earlier, a low percentage of NK cells in both tuberculous (TB) and nontuberculous (NTB) pleural effusions was reported (Jalapathy et al., 2004).

NK cells are distinct populations of CD16+CD56+ lymphocytes further divided into CD56dim and CD56bright subsets. These subsets represent different stages of NK cell maturation. The CD56dim NK cells are cytotoxic in nature and mediate the target cell lysis through the release of perforin and granzyme, whereas CD56bright NK cells are immature cells as they lack perforin and granzyme. These cells are recruited as an early host defense. They identify the downregulation or absence of the host major histocompatibility complex (MHC) class-1 molecule and kill infected cells specifically, thereby sparing normal cells (Berahovich et al., 2006). The ability of NK cells to interact directly with various microorganisms has also been reported by others, but the receptor engaged is still not known (Nylen et al., 2003; Chalifour et al., 2004). Although T cells dominate the defense, initially, NK cells of the innate immune arm activate and prime antigen-presenting cells (APC), thus shifting adaptive immunity towards T-helper type 1 (Th1) response. NK cells are unique in bridging the innate and adaptive immunity during M. tuberculosis infection, but their role at the site of infection is still not completely understood.

The orderly recruitment of immune cells to inflammatory sites during tuberculosis is purely by chemokines and their receptors. A known function of chemokines is to activate immune cells (Loetscher et al., 1996). Chemokines are multi functional and divided into inflammatory, constitutive, dual-function chemokines and many more (Roda et al., 2006). It is well known that human NK cells are capable of synthesizing and secreting several chemokines that recruit T cells, B cells, neutrophils and other activated NK cells (Somersalo et al., 1994; Hedrick et al., 1997; Fehniger et al., 1999). The molecular basis of NK cell chemokine responsiveness and its subsequent recruitment into tissues is still being elucidated.

Earlier, we demonstrated significantly high levels of Th1 cytokines, chemokines and a protective T-cell response compartmentalized at the site of infection during tuberculous pleurisy (Jalapathy et al., 2004; Prabha et al., 2007; Pokkali et al., 2008). It is now evident that elevated chemokine levels and their respective receptor expression drive the immune cells to the site of infection; therefore, in the present study, an attempt was made to assess the percentage of NK cell subsets at the site, their differential expression of activation markers (CD69 and HLA-DR), antigen recognition receptors (Toll-like receptors [TLR]) and chemokine receptors (CXCR2, CCR1, CCR2, CCR5 and CCR7) in tuberculous pleural effusion in comparison with nontuberculous effusion. Further, to identify the chemokines involved in NK cell compartmentalization and the receptor used for the same, in vitro migration of NK cells in response to various recombinant human chemokines was analyzed and the respective chemokine receptor expressions on migrated NK cells were enumerated.

Materials and methods

Selection of patients

A total of 62 patients were recruited before the start of the treatment from Government General Hospital (GGH), Chennai. The tuberculous pleuritis group (TB) comprised of 38 patients with exudative pleural effusions with lymphocytic predominance. The remaining 24 patients had a nontuberculous etiology [malignant (n=9), liver failure (n=6), parapneumonic (n=5), cardiac heart failure (CHF) (n=3) and renal failure (n=1)] and were hence grouped as nontuberculous control (NTB). The mean age of the study patients was 34 years (range 18–60 years). Written and informed consent was obtained from each patient. All the patients were sero-negative for HIV and the study followed the ethical guidelines of GGH, Chennai.

The diagnosis of tuberculosis was based on the smear, culture and PCR positivity (IS6110 specific) of the sputum or the pleural fluid (PF), together with the clinical picture of the chest X-ray. These patients showed positivity in at least any two of the above criteria and hence were categorized as the TB group. All these patients were first time diagnosed as having TB and were not relapsed cases. They responded well to antituberculous treatment and were followed for the first 3 months. In the NTB group, patients with clinical evidence of heart failure, liver cirrhosis and renal failure had transudative effusions whereas malignant patients had exudative effusions.

Specimen collection and processing

Diagnostic thoracentesis was performed to collect the PF and processed immediately to separate the cell-free PF, which was subsequently stored at −70 °C for future assays. The pelleted PF cells were washed and layered on a Ficoll-Hypaque (Amersham Biosciences) density gradient centrifugation to obtain pleural fluid mononuclear cells (PFMC). The cells were washed twice in 1 × HBSS (Sigma Chemicals, St. Louis). A final suspension of 1 × 106 cells mL−1 was made in RPMI-1640 (Sigma Chemicals) supplemented with 10% fetal calf serum (Nissinen et al., 2004).

Immune cell profile, percentage of NK cells and their subset determination

Various immune cell percentages in TB and NTB-PFMC were studied by flow cytometric analysis. One hundred microliters of the cell suspension containing 105 cells was washed with phosphate-buffered saline (PBS). A panel consisting of unstained cells, isotype control [mouse immunoglobulin G conjugated to fluorescein thiocyanate (FITC) and phycoerythrin (PE)], T cell/B cell marker (CD3 FITC/CD19 PE), and helper T cell/cytotoxic T cell marker (CD4 FITC/CD8 PE), respectively, was chosen. CD56+ NK cell subset populations were determined by flow cytometry. Finally, to identify the NK cell population in the lymphocyte gate, cells were stained with CD3 FITC/CD16+56 PE (T cells/NK cells). To differentiate CD56bright and CD56dim NK cell subsets further, the expression of CD56 based on fluorescent intensities on CD3 gated lymphocytes was analyzed. Coexpression of CD3+ CD56+ cells depicts the presence of an NKT cell population. However, the NKT cell population was not included for analysis in the present study. Fluorescence compensation on the flow cytometry was adjusted using unstained and isotype control to minimize the overlap of the FITC and PE signals. Briefly, 1 × 105 cells per tube were stained with combinations of antibodies for 20 min at 4 °C. Cells were then washed twice with PBS and fixed with 4% paraformaldehyde in PBS and analyzed on a FACS Calibur flow cytometer (Becton Dickinson). For each sample, lymphocytes were gated based on forward and side scatter parameters and a total of 10 000 gated events were collected for each sample. Data were analyzed using cellquest pro software (Becton Dickinson). The data were expressed as percentage of total lymphocytes. Enumeration of NK cell subsets was performed using antihuman PE-Cy5 conjugated CD56 (BD Pharmingen, CA) as described elsewhere (Cooper et al., 2001b).

Phenotyping of NK cell activation marker, TLRs and chemokine receptor

The NK cell subsets were further characterized for the surface expression of CD69, HLA-DR, TLR (TLR-2, TLR-4 and TLR-9) α- and β-chemokine receptors (CXCR2, CXCR3, CCR1, CCR2, CCR5 and CCR7) using FITC-labeled mouse antihuman CCR5 and TLR9, PE-labeled mouse antihuman CCR1, CCR2, CCR7, TLR-2 and TLR4 and Allo-PhycoCyanin (APC)-labeled mouse antihuman CXCR2 (R&D Systems, Minneapolis, MN) antibodies, and processed as above.

Transmigration assay of NK cells

The transmigration assay was the modified protocol as described elsewhere (Borchers et al., 2002). Purified NK cells were obtained by negative selection using magnetic sorting (Schierloh et al., 2005). Briefly, 20 × 106 PFMCs were treated with magnetically labeled anti-CD3, anti-CD14, anti-CD19 and anti-CD16 for 30 min at 4 °C and purified using MACS mini columns (Milteny Biotech, Germany). The purified PFMC NK cells of TB and NTB patients were further used for the in vitro migration assay in response to selected chemokines and investigated using a 3 µm polycarbonate membrane ThinCert™ (Greiner Bio-one, Germany) in 24-well tissue culture plates (Costar, NY). The inserts were preincubated with media (RPMI 1640 containing 1% bovine serum albumin) for 1 h and then placed into the individual wells containing 500 µL media alone or media containing chemokines (20 ng mL−1 of each IL-8, IP-10, MCP-1 and SLC). Culture conditions such as the concentrations of chemokines loaded, cell number and incubation time were optimized earlier. A single-cell suspension of 1 × 106 cells in a total volume of 200 µL media was loaded onto the inserts. The plates were incubated at 37 °C in 5% CO2 for 3 h. At termination, the nonmigrated cells present in the suspension of the upper chamber were removed. The number of cells that had migrated was determined as the sum of the cells in the lower chamber plus the cells that had migrated, but remained attached to the bottom of the insert. The cells that had migrated but remained attached to the insert were then recovered by placing the bottom of the inserts in PBS containing 0.02% EDTA and tapping gently. This cell suspension was added to the corresponding lower chamber migrated cells into a microcentrifuge tube. The absence of adherent cells to the undersurface of the insert was ascertained by light microscopy. Quantification of input and output cells was performed using FACS analysis in a time-triggered acquisition setting. The cells were then stained as mentioned above for the cell phenotype marker along with the cognate chemokine receptor and acquired on a flow cytometer for 30 s and analyzed (Becton Dickinson). Cell migration is expressed as a percentage migration, which is the ratio of the number of cells migrating in response to chemokines relative to the number of cells migrating in response to media alone × 100. The assay was always performed in triplicate.

Statistical analysis

The data were subjected to statistical analysis using the graphpad prism software (version 4.0), and a two-tailed Student's independent-sample t-test was performed. Comparisons between the two groups were performed using a Mann–Whitney U-test. The data are expressed as mean ± standard error of the mean (SEM). A P-value <0.05 was considered statistically significant when compared with the respective dim or bright population of NK cells.

Results

Immune cell profile during pleurisy

The in vivo correlates of protective immunity were earlier demonstrated by studying the immune cell profile using flow cytometry (Jalapathy et al., 2004). The immunological cell architecture of TB and NTB-PF in the study patients is listed in Table 1. A significantly high T-cell percentage (both CD3 and CD4) was observed in TB-PF than in NTB-PF, supporting the earlier observation.

View this table:
Table 1

Immune cell profile of the site of infection during pleurisy

TB-PFMCNTB-PFMC
CD3+73.20 ± 6.0665.31 ± 5.05
CD4+48.42 ± 5.6034.89 ± 4.51
CD8+24.92 ± 4.1626.27 ± 3.39
CD19+4.69 ± 1.333.94 ± 1.28
CD16/56+10.62 ± 1.5216.27 ± 2.91
CD3+ CD16/56+5.73 ± 0.894.0 ± 0.41
  • Mononuclear cells from the tuberculous and nontuberculous pleural fluid were stained with respective fluorescent labeled antibodies to enumerate immune cell architecture at the site of infection. Results are expressed as mean values ± SEM.

  • P<0.05 was considered to be statistically significant.

Profile of NK cells at the site of infection

Initially, the percentage of total CD3CD16/56+ NK cells present in the lymphocyte gate was enumerated by dual immuno-staining using flow cytometry (Fig. 1). The total NK cells were high in NTB compared with TB, but the difference was not statistically significant. NK cells exhibited a difference in the surface expression of CD56 and divided into CD56dim and CD56bright subsets (Cooper et al., 2001b). The CD56bright and CD56dim NK cell subsets, based on CD56 fluorescent intensities on CD3-gated lymphocytes, were analyzed as shown in Fig. 1b. The NK cell subset analysis revealed a significant decrease (*P<0.05) in the CD56dim NK cell population in TB-PF when compared with NTB, but there was no change in the CD56bright NK cells in both the groups. Also, the percentage of NK cell subsets in blood and PF from TB and NTB patients was evaluated and is shown in Table 2. However, a significant increase in CD56bright NK cells was observed in TB and NTB-PFMC in comparison with their respective PBMC values. This highlights the fact that immunoregulation for amplifying the inflammatory response at the site is specifically mediated by bright NK cells, which further supports our earlier observation of an increased cytokine response at the site during pleurisy (Jalapathy et al., 2004; Pokkali et al., 2008).

Figure 1

Profile of NK cell subsets in pleural fluid (PF) of TB (N=38) and NTB (N=24) patients and flow cytometric analysis of dot plot (b). Initially, the PF cells were double stained using CD3-FITC and CD16/56-PE for enumerating the total CD3 CD16/56+ NK cells in the lymphocyte gate and acquired on flow cytometry. To distinguish the subsets of dim and bright NK cells, the cells were stained with CD56 PE and CD3-FTIC. The CD56+ NK cells were further divided into CD56bright and CD56dim based on CD56 fluorescence intensity and by region analysis as shown in the dot plot. The region gate depicts the pure NK cells (1) and its subsets (2) in PF from TB patients. Coexpression of CD3+ CD56+ cells depicts the presence of the NKT cell population. However, in the present study, the focus was on CD56dim and CD56bright cells and NKT cells were not used for analysis. Results are expressed as mean values ± SEM. *P<0.05 was considered to be significant using an independent Student's t-test. An isotype control was used to set the fluorescent compensation and to minimize the overlap of fluorochrome signals or background staining.

View this table:
Table 2

NK cell subset in blood and pleural fluid during pleurisy

TBNTB
PBMCPFMCPBMCPFMC
CD56dim5.59 ± 1.936.5 ± 1.0111.49 ± 2.4813.77 ± 1.65
CD56bright0.17 ± 0.123.46 ± 0.460.34 ± 0.322.82 ± 1.17
  • CD56+ NK cell subset population were determined by two color flow cytometry. Briefly, mononuclear cells were isolated and stained with anti-CD3-FITC and anti-CD56-PE Abs to differentiate NK cell subsets. Results are expressed as mean values ± SEM.

  • P<0.05 was considered to be statistically significant when comparison made between TB- and NTB PFMC and

  • Represents the comparison between PBMC and PFMC within TB and NTB group.

Analysis of activation markers on PF NK cell subsets

We further characterized the frequency of activation marker and phenotype expression on NK cell subsets. The expressions of HLA-DR and CD69 on these subsets in TB and NTB are depicted in Fig. 2. There was a significant increase (@P<0.05) in the expression of these markers on both the subsets of NK cells in TB compared with NTB patients. Within TB, the CD56bright NK cells showed a significant increase (*P<0.05) in the expression of HLA-DR and CD69 compared with the CD56dim NK cells.

Figure 2

Phenotype characterization of pleural fluid (PF) NK cell subsets in TB (N=38) and NTB (N=24) patients. Cells were double stained and analyzed by flow cytometry for coexpression of HLA-DR and CD69 on the subsets of NK cells in both the study groups. Results are expressed as mean values ± SEM. *@P<0.05 was considered to be significant using an independent Student's t-test and the Mann–Whitney U-test for comparison between groups. An isotype control was used to set the fluorescent compensation and to minimize the overlap of fluorochrome signals or background staining.

Expression profile of CC-chemokine receptors on PF NK cell subsets

Normally, PF is devoid of cells, but in TB pleuritis, as a result of a DTH response to mycobacterial antigens, there is an accumulation of immune cells in response to chemokines. Hence, we profiled chemokine receptor expression on NK cell subsets and the results are shown in Fig. 3. We observed a significant increase (*P<0.05) in the expression of all the CC chemokine receptors (CCR1, CCR2 and CCR7), except for CCR5 on the CD56dim NK cell subset in the TB-PF group (Fig. 3a). In NTB-PF, only CCR7 was found to be significantly high (*P<0.05) on the CD56dim NK cell subset, suggesting its role in host immune surveillance as shown in Fig. 3b.

Figure 3

Differential expression of chemokine receptors on the subsets of NK cells in TB (a) and NTB (b) PF. Coexpression analysis for chemokine receptors (CXCR2, CCR1, CCR2, CCR5 and CCR7) was performed on TB and NTB-PF NK cells. Results are expressed as mean values ± SEM. *P<0.05 was considered to be significant using an independent Student's t-test. An isotype control was used to set the fluorescent compensation and to minimize the overlap of fluorochrome signals or background staining.

Expression profile of TLR on TB-PF NK cell subsets

The cytotoxicity of NK cells is highly specific to lyse the identified altered self-cell and infected cells at the site, which is principally through TLRs. We analyzed the profile of TLR expression on the subsets of NK cells in TB patients, and the results are shown in Fig. 4. A significant increase (*P<0.05) in TLR-2, TLR-4 and TLR-9 expression was observed on CD56bright compared with CD56dim NK cells of TB-PF.

Figure 4

Expression profile of TLRs on the subsets of TB-PF NK cells. Results are expressed as mean values ± SEM. *P<0.05 was considered to be significant using an independent Student's t-test. An isotype control was used to set the fluorescent compensation and to minimize the overlap of fluorochrome signals or background staining.

In vitro chemotaxis of NK cells

To demonstrate the migration of total NK cells, chemotaxis was performed and the results are shown in Fig. 5. Recombinant human chemokines, such as IL-8, IP-10 and MCP-1, which were also high in TB-PF, were selected along with SLC to assess the migration of NK cells. We observed a significant increase (*P<0.05) in the migration of total NK cells of TB compared with NTB-PF in response to all the chemokines tested. However, this in vitro migratory capacity was not reflected in the ex vivo percentage of total NK cells in TB-PF as shown in Fig. 1.

Figure 5

Demonstrates the chemotaxis of NK cells from TB and NTB pleuritis patients, in response to recombinant human chemokines. The percentage of transmigrated cells was analyzed in each study group. *P<0.05 was considered to be significant using an independent Student's t-test. The cells that migrated in response to media or HBSS alone were considered to migrate passively and were reduced from chemokine-induced migration.

Expression of chemokine receptors on the trans-migrated NK cells

NK cells responded to the recombinant human chemokines, and so we wanted to analyze the receptors it uses to facilitate the trans-migration. Hence, the migrated NK cells were stained for their cognate chemokine receptor expression in both the study groups. All chemokine receptors (CXCR2, CCR2 and CCR7), except CXCR3, were found to be significantly increased (*P<0.05) in TB-PF NK cells as shown in Fig. 6. This indicates that IP-10 was not potent in stimulating the NK cell migration through CXCR3 to become compartmentalized at the inflammatory site. Other chemokines such as IL-8, MCP-1 and SLC are effective in stimulating NK cells with increased expression of CXCR2, CCR2 and CCR7.

Figure 6

Chemokine receptor expression pattern on transmigrated NK cells. A time-dependent acquisition was performed on the double-stained transmigrated NK cells and subsequently analyzed for chemokine receptor expression. *P<0.05 was considered to be significant using an independent Student's t-test. An isotype control was used to set the fluorescent compensation and to minimize the overlap of fluorochrome signals or background staining.

Comparison of ex vivo and in vitro chemokine receptor expression on TB-PF NK cells

To ensure the stimulating effect of these chemokines, we compared the ex vivo and in vitro chemokine receptors on TB-PF NK cells, and the results are shown in Fig. 7. There was a significant increase (*P<0.05) in CXCR2, CCR2 and CCR7 chemokine receptor expressions on in vitro migrated TB-PF NK cells compared with their in vivo levels.

Figure 7

Comparison of ex vivo and in vitro chemokine receptor expressions on TB NK cells. *P<0.05 was considered to be significant using an independent Student's t-test.

Discussion

Although tuberculous pleural effusions are rich in protective effector T-cells, an array of other immune cells also become localized at the foci. Many studies have highlighted the role of NK cells in cytotoxicity and direct antigen recognition during pulmonary tuberculosis (Nirmala et al., 2001; Gerosa et al., 2002; Wang et al., 2004). In this study, we focused on NK cells at the site of infection to study their functional role and mechanism of action against M. tuberculosis.

NK cell cytotoxicity is exhibited by the CD56dim population through the action of perforin and granzyme. The significantly low levels of CD56dim NK cells observed in TB compared with NTB pleural effusion may be indicative of the limited cytotoxicity by these cells at the infection foci. Earlier, the low levels of CD56dim NK cells in tuberculous pleural effusion were attributed to apoptosis of these cells by the inflammatory mediators present in the milieu (Schierloh et al., 2005). This cytotoxicity check at the site may be important for the host to sustain the protective immune response. Otherwise, the outflow of cytotoxic granules would result in host tissue damage at the site, which would be more deleterious than the M. tuberculosis infection.

An important attribute of CD56bright NK cell in immunoregulation is mediated through abundant production of type 1 and type 2 cytokines (Cooper et al., 2001b). Elevated levels of CD56bright NK cells in TB pleural effusions compared with blood were reported earlier, highlighting the selective recruitment of these cells (Ota et al., 1990; Lande et al., 2003). The comparison between TB and NTB pleural effusions in the present study shows c. 3–4% of CD56bright NK cells in both the groups. However, CD56bright NK cells may contribute toward immunoregulation only in TB effusions, as a protective T-helper response predominates during the pathology (Jalapathy et al., 2004). Moreover, CD56bright NK cells secrete IL-21, which is known to act like IL-2 and induce T-cell proliferation, thus, contributing towards proliferation and enrichment of antigen-specific T-cells in TB effusion and emphasizing the indirect role of NK cells in host protection.

Sensing of the antigen/mitogen stimuli directly at the site is aided by HLA-DR expression on CD56bright NK cells with simultaneous upregulation of CD69, an early activation marker (Craston et al., 1997; Gerosa et al., 2002; Junqueira-Kipnis et al., 2003; Cao & He, 2005; Schierloh et al., 2005). The increased CD69 and HLA-DR expressions on CD56bright NK cells in TB-PF may probably be in response to the recognition of mycobacterial antigens. These activated CD56bright NK cells become compartmentalized selectively at the pleural space and differentiate either to effector CD56dim or may aid in producing a protective response through immunoregulation.

Earlier studies highlight that NK cell subsets express chemokine receptors varying from 1% (e.g. CCR9 and CXCR5 on CD56dim NK cells) to a vast majority (e.g. CCR7 on CD56bright NK cells). Certain chemokine receptors were exclusively present on very minor subsets of the CD56dim/CD56bright NK cell population (Campbell et al., 2001; Cooper et al., 2001a; Inngjerdingen et al., 2001; Berahovich et al., 2006). In our earlier study, we have reported significantly high levels of CXC (IL-8, MIG and IP-10) and CC (MIP-1α and MCP-1, but not regulated on activation normal T cell expressed and secreted (RANTES) chemokines in tuberculous pleural effusion (Pokkali et al., 2008). These pools of chemokines may aid homing of NK cells to the infected site via increased CC receptors (except CCR5) on CD56dim cells in tuberculous effusions (Campbell et al., 2001; Berahovich et al., 2006). This upregulation of receptors may be perhaps due to the differentiation of already recruited immature CD56bright to mature CD56dim NK cells through CC chemokine activation. The other possibility is that mature CD56dim NK cells are directly recruited to the site to help other cells in reducing the antigen burden. NK cells are known to respond well to MIP-1α and RANTES through CCR1 and CCR5. The differential expression of CCR1 and CCR5 on CD56dim NK cells may be the result of low RANTES in tuberculous PF.

All immune cells express CCR7 for their continuous recirculation to recognize antigens or the altered self-cells and their elimination (Kim et al., 1999). An interesting observation of increased CCR7 in CD56dim NK cells of both the groups indicated their primary role in immunological surveillance. As CCR7+ CD56dim NK cells are produced against the primary pathophysiology, their increase in TB effusions may be antigen specific.

NK cells play a vital role in sensing the antigens via TLR expression and help in the elimination of infected cells during tuberculous pleurisy. In our earlier study, we observed increased expressions of TLR-2 and TLR-4 on CD16/56+ NK cells from TB PF (Prabha et al., 2008). Similar increased expressions of TLR-2, TLR-4 and TLR-9 on immature CD56bright NK cells was observed in this study. This may suggest that NK cells elicit a protective Th1 immune response by activating interferon-γ production at the site and indirectly help macrophages to eliminate M. tuberculosis during tuberculosis (Batoni et al., 2000,; 2005; Esin et al., 2004). Their further interactions with the antigen may cause these immature cells to differentiate to mature cells.

The chemokines IL-8, IP-10, MCP-1 and SLC predominantly attract neutrophils, T-cells, monocytes and other immune cells by engaging CXCR2, CXCR3, CCR2 and CCR7 to aid their transmigration. All these chemokines have been studied extensively in M. tuberculosis infection (Flynn & Chan, 2001; Algood et al., 2003; Peters & Ernst, 2003). The increased migration of purified TB-PF NK cells to these recombinant human chemokines demonstrated their spontaneous responsiveness because of their prior in vivo activation. This increased in vitro migration also underlines the effectiveness of the individual chemokine in attracting the antigen-specific activated NK cells to the site. However, in vitro migratory capacity was not reflected in the ex vivo percentage of the total NK cells observed in TB-PF. It can be assumed that the total NK cell number observed in TB-PF is due to the chemotactic stimuli provided by the chemokine pool present at the site and not by individual chemokines as observed in in vitro experiments.

NK cell migration is directly proportional to the expression of chemokine receptors. Our results highlighted increased transmigration of NK cells from TB-PF via CXCR2, CCR2 and CCR7 and not via CXCR3. This suggests that the significantly high levels of IP-10 in TB-PF did not recruit NK cells, but specifically induced T-cell migration. It is well known that IL-8, being proinflammatory, amplifies the immune response by recruiting NK cells via CXCR2. Secondary to monocyte recruitment, MCP-1 also induces NK cell recruitment through CCR2 to release the monocytes-activating cytokines, reflecting the initial immune responses specific to M. tuberculosis. Immune response regulation during M. tuberculosis infection at the site is a consequence of various factors primarily elicited by SLC and CCR7 together (Kim et al., 1999; Batoni et al., 2000,; 2005). This allows the continuous recirculation of antigen-specific NK cells between the infected site and general circulation.

Earlier studies have demonstrated chemokine receptor expression either on different NK cell subset populations (such as CD16+, CD16, NKT cells) enriched with human peripheral blood or using NK cell lines (using NKL, KHYG-1 and NK-92) (Campbell et al., 2001; Inngjerdingen et al., 2001; Berahovich et al., 2006). All these studies demonstrated that NK cells have the capacity to migrate to inflammatory sites with varying expression levels of CCR1, CCR2, CCR5, CXCR2 and CXCR3. Interestingly, all the studies highlighted the role of CCR7 in immunosurveillance.

Overall, it can be stated that NK cell recruitment to the pleural space is essential for an early protective host immune response. Apart from inducing transmigration, chemokines and their receptors are probably involved in differentiation of subtypes at the site. As first direct sensors and effectors in mycobacterial infections, NK cells may represent of an underlying factor for the self resolving nature as an effect of an innate immune arm.

Acknowledgements

Ms Supriya Pokkali thanks the Indian Council of Medical Research (ICMR) for providing a Junior Research fellowship.

Footnotes

  • Editor: Nicholas Carbonetti

References

View Abstract