Chemokine receptor expression and modulation by Mycobacterium tuberculosis antigens on mononuclear cells from human lymphoid tissues

Abstract

Chemokine receptor switching on lymphoid cells is an important factor regulating migration and homing, but little is known about the expression of such molecules during Mycobacterium tuberculosis infection in humans. We describe CCR2, CCR5 and CCR7 expression on human cells from blood, spleen and pulmonary hilar lymph nodes (PHLN) stimulated by M. tuberculosis antigens. CCR2 was not expressed by CD3⁺ cells regardless of the presence of antigen, but was highly expressed on CD14⁺ CD63⁺ monocytes/macrophages. CCR2 decreased on splenic monocytes/macrophages by nearly 50% in culture, independent of antigen, but remained high in blood and PHLN. CCR5 was low in CD3⁺ cells and was down-regulated by M. tuberculosis antigens on blood and splenic cells but not in PHLN. CCR5 was highly expressed on monocytes/macrophages and was down-regulated by M. tuberculosis antigens at 48 hr only in blood. Less than 15% of CD3⁺ cells from spleen and PHLN were CCR7⁺, whereas nearly 40% from blood expressed this receptor on primary isolation. However, CCR7 in PHLN increased in culture, independent of antigen. Monocytes/macrophages did not express CCR7. Thus, we characterize, for the first time, chemokine receptor expression and differential modulation by M. tuberculosis antigens on human mononuclear cells from spleen, blood and PHLN. Knowledge of chemokine receptor switching in human lymphoid tissue provides novel insight into mechanisms of the immune response to M. tuberculosis with potential effects on directing cell trafficking.

Introduction

Containment of Mycobacterium tuberculosis infection is associated with cell-mediated immune response. This response involves T cells and starts with the capturing of mycobacterial antigens by both macrophages and immature dendritic cells (DC) in peripheral tissues. The function of macrophages, in contrast to DC which migrate to T-cell areas within secondary lymphoid organs after antigen capture, has been associated with local inflammation and effector antimicrobial responses.^1–4

Phagocytosis of M. tuberculosis by macrophages rapidly induces cellular activation and production of microbicidal factors that limit the growth of intracellular mycobacteria.⁵ Such activation also results in chemokine production with subsequent recruitment of additional leucocytes, such as effector memory T cells, from the peripheral circulation.^6,7 This cell recruitment is likely to be orchestrated by interaction of these chemokines with their receptors.⁸ Chemokine receptors (CCR) are G-protein-coupled receptors expressed on lymphoid cells such as T cells and macrophages, according to the state of cell activation.⁹ Directed cell migration to the inflammation site is induced by inflammatory chemokines such as IL-8, monocyte chemoattractant protein-1 (MCP-1), macrophage inflammatory protein-1α (MIP-1α), MIP-1β, and regulated on activation, normal T-cell expressed and secreted (RANTES), produced by local epithelial and immune cells.^7,10–14 Cells from the periphery respond to gradients of these chemokines through expression of chemokine receptors such as CCR2, the receptor for MCP-1, and CCR5 the receptor for MIP-1α, MIP-1β, and RANTES.^15,16 Interaction of pathogens with migrating cells induces switching of chemokine receptor expression which permits either homing of cells at the inflammatory foci to facilitate granuloma formation^2,4 or travelling to regional lymph nodes (LN) where antigen presentation and development of adaptive immune responses take place.^17,18 The orchestration of such events is clearly crucial in the host response to the mycobacterial infection. Permanent residency of macrophages at an inflammatory focus is probably the result of both marked down-regulation of CCR7, or rather an absence of expression on the cell surface^2–4 and, at the same time, persistence of expression of receptors for inflammatory chemokines such as CCR2 and CCR5.^19,20

Regulation of chemokine receptor expression and chemokine production by lymphoid cells in response to M. tuberculosis has been described.^21–26 For example, M. tuberculosis cell wall components have been shown to induce the production of MCP-1 and MIP-1α by highly purified human blood monocytes, and to down-regulate expression of CCR2.²¹ CCR5 was reported to be also down-regulated in whole blood monocytes²⁷ and monocyte-derived macrophages²⁸ from humans after stimulation with M. tuberculosis cell wall products. However, Fraziano et al. showed that CCR5 is up-regulated on human monocyte-derived macrophages and alveolar macrophages after in vitro infection with M. tuberculosis H37Rv.²⁹ In contrast to CCR2 and CCR5, CCR7 was neither expressed by murine macrophages nor induced by M. tuberculosis.²

Effective control of M. tuberculosis infection mainly depends on successful interaction between infected macrophages and T lymphocytes. The protective effect of CD4⁺ T cells is principally derived from the production of cytokines, such as interferon-γ, after stimulation with M. tuberculosis antigens.⁵ T cells are recruited to the site of infection through expression and modulation of chemokine receptors such as CCR2 and CCR5.^8,19 In this regard, CCR5 expression was reported to be up-regulated on CD4⁺ T lymphocytes after both lipopolysaccharide and lipoarabinomannan stimulation.²⁴ However, another study reported that circulating peripheral blood γδ T cells had a higher expression than αβ T cells of CC- and CXC-chemokine receptors including CCR2, CCR5, and CXCR1 but not that of CCR7. Such expression was differentially affected by non-virulent M. tuberculosis, and that of CCR5 was diminished.²³

Thus, previous studies have shown that the expression of chemokine receptors and their regulation by M. tuberculosis may vary according to the nature of the antigen and cell type. However, little is known about the modulation of human chemokine receptor expression by M. tuberculosis antigens on cells from lymphoid tissues, such as spleen and LN, crucial sites of antigen presentation and development of an adaptive immune response.

Therefore, we characterized CCR2, CCR5 and CCR7 expression by mononuclear cells from three human lymphoid compartments, i.e. spleen, pulmonary hilar lymph nodes (PHLN) and blood, and studied their modulation by M. tuberculosis antigens. Our data demonstrate the existence of differential chemokine receptor expression patterns within different lymphoid compartments, and that modulation of such receptors by M. tuberculosis antigens differs according to the receptor and cell type. These data go some way to explaining lymphoid cell trafficking in the human immune response to M. tuberculosis.

Materials and methods

Origin of lymphoid tissues

Spleen and LN from the pulmonary hilar region were obtained from human cadaveric donors in the Transplantation Unit at the University Hospital San Vicente de Paúl, Medellín, Colombia, taken during surgery to obtain organs for transplantation. Surgically removed spleens are sent to our Laboratory for human leucocyte antigen typing and cross-match cytotoxic analysis to screen for potential recipients. Donors are routinely tested for human immunodeficiency virus, hepatitis B and C viruses, cytomegalovirus, Chagas' disease and syphilis, and are not older than 60 years.

Venous blood from healthy donors recruited from Laboratory personnel was obtained to purify mononuclear cells (MNC) with the aim of comparing CCR expression in blood with that in spleen and LN. Tuberculin status was tested in healthy controls by in vitro cell proliferation response to purified protein derivative (PPD). All such individuals provided written consent.

Ethical approval was obtained from the ethics committee of both the University Hospital ‘San Vicente de Paul’, and the Facultad de Medicina of the Universidad de Antioquia.

Tissue processing and blood collection

Following surgical removal, tissues were quickly chilled to 4° in RPMI-1640 (Gibco, Grand Island, NY) supplemented with 10 mm HEPES, 2 mm l-glutamine, 100 IU/ml penicillin, 100 μg/ml streptomycin, and 10% fetal calf serum (HyClone, South Logan, UT), called hereafter complete medium (CM), and processed, under sterile conditions, within 10 hr of removal from the cadaver. Tissues were placed on 100-mm Petri dishes with 40 ml cold CM and were thoroughly minced using forceps and scalpel. The tissue suspension was transferred to a 50-ml tube and placed vertically for 30 seconds to allow debris to sediment. Supernatant was centrifuged at 400 g for 10 min, and cells were harvested and re-suspended in room temperature phosphate-buffered saline (PBS) for MNC separation by gradient centrifugation on lymphocyte separation medium (BioWhittaker, Walkersview, MD). The layer of MNC was recovered, washed twice in PBS, and then passed through a 70-μm cell strainer (Becton-Dickinson, San Diego, CA) to remove remaining debris. After this procedure, cells were re-suspended in CM and kept on ice until phenotyping or freezing. Cell viability after this procedure was consistently >90%, as indicated by trypan blue exclusion.

Sixty millilitres of peripheral blood were obtained by venepuncture from healthy donors, and MNC were obtained by gradient centrifugation. The MNC were used on same day of separation.

Freezing procedure and cell storage

Cells obtained from lymphoid tissues were phenotyped after separation but most of the cells were frozen in liquid nitrogen until further use, maintained in freezing medium (20% dimethyl sulphoxide, 30% fetal calf serum, and 50% CM) with not more than 10⁷ cells per ampoule. Viability of cells after thawing was consistently more than 88% as judged by trypan blue exclusion. Cell samples below this viability were discarded. Although no experiments were performed to define whether the freezing procedure would affect cell surface CCR expression, a study by Campbell et al. showed that such a procedure did not affect the pattern of CCR expression either by cells from blood or bronchoalveolar lavages.³⁰

Cell culture

Either fresh or thawed MNC from lymphoid tissues were cultured in CM, supplemented with 10% AB pooled human serum (PHS, BioWhittaker) in U-bottomed 96-well plates (NUNC, Rochester, NY) at 2 × 10⁵ cells/well in a final volume of 200 μl in a humidified incubator at 37° and 5% CO₂. The cells were stimulated with 10 μg/ml PPD from M. tuberculosis (Statens Serum Institute, Copenhagen, Denmark), or 50 ng/ml phorbol 12-myristate 13-acetate (PMA) plus 1 μm ionomycin both from Sigma (St Louis, MO). Untreated cells were used as negative controls. At different time points, supernatants were collected and stored at −70° for further quantification of chemokines, and the cells were used for CCR immunostaining.

Immunofluorescence and flow cytometry

For phenotyping of the principal MNC populations, fresh MNC from spleen and LN were resuspended in cold binding buffer (PBS, pH 7·2–7·4 supplemented with 2% PHS plus 0·1% sodium azide), and washed twice. Then, cells were resuspended in 100 μl binding buffer, and incubated at 4° for 30 min with specific monoclonal antibodies (mAb) for the main MNC populations (CD3, CD14, CD19, CD56, and CD11C⁺ DR⁺) labelled with either fluorescein isothiocyanate (FITC) or phycoerithrin (PE). Table 1 defines the antibodies used in immunofluorescence and flow cytometry. Then, cells were washed twice with 1 ml cold PBS and fixed with 400 μl 2% paraformaldehyde (Fisher Scientific, Fair Lawn, NJ) in PBS. Staining for CC receptors was performed as follows: either fresh or stimulated MNC from the three tissues studied (blood, spleen and LN) were resuspended in cold binding buffer and washed twice. Subsequently, cells were doubled stained with PE-labelled anti-CCR2, anti-CCR5 and anti-CCR7 mAbs, and FITC-labelled anti-CD63 (blood) and anti-CD14 (spleen and LN), for monocytes/macrophages, and anti-CD3 mAb for T cells. Isotype control antibodies, as described in Table 1, were used to control for non-specific binding. The cells were washed, incubated and fixed as explained above. Ten thousand events were acquired in a Beckman-Coulter Epics XL (Hialeah, FL) or a FACSort Becton-Dickinson (San Jose, CA) cytometer, and the samples were analysed using the CellQuest software version 3·3 (Becton-Dickinson, PaloAlto, CA).

Table 1.

Antibodies

Specificity	Clone	Isotype	Source
CD3-FITC/Cy	UCHT1	IgG1	BDP
CD11C-PE	B-Ly6	IgG1	BDP
CD14-FITC/RDE	My4	IgG2b	BC
CD19-FITC	HIB19	IgG1	BDP
CD56-FITC	B159	IgG1	BDP
CD63-FITC	H5C6	IgG1	BDP
HLA-DR-FITC	G46-6 (L243)	IgG2a	BDP
CCR2-PE	48607·211	IgG2b	R&D
CCR5 (CD195)-PE	3A9	IgG2a	BDP
CCR7-FITC/PE	150503	IgG2a	R&D
Mouse IgG-FITC/PE	MOPC-21	IgG1	BDP
Mouse IgG-FITC/PE	MPC-11	IgG2b	BC
Mouse IgG-PE	20102	IgG2a	R&D

BDP, BD Pharmingen, San Diego, CA; BC, Beckman Coulter, Galway, Ireland; R&D, R&D Systems, Minneapolis, MN.

Cell proliferation assay

Triplicate cultures of MNC from the different lymphoid tissues (1 × 10⁵ cells/200 μl) were stimulated for 5 days in CM (supplemented with 10% PHS) with 10 μg/ml PPD, or with 5 μg/ml phytohaemagglutinin (PHA, ICN Biomedicals, Aurora, OH) in round-bottomed 96-well plates. Cell proliferation was estimated by incorporation into the DNA of [³H]TdR (specific activity 185 GBq/mmol, 5·0 Ci/mmol; Amersham Biosciences, Buckinghamshire, UK). The cells were pulsed with 0·5 μCi [³H]TdR/well 18 hr before harvesting, and counts per min (c.p.m.) were counted in a liquid scintillation β-counter (Beckman LS 6500, Albertville, Minnesota). The proliferation response was calculated as the mean ± SD of the c.p.m. from three replicates.

Enzyme-linked immunosorbent assay (ELISA)

Determination of RANTES, MIP-1α, and MIP-1β CC-chemokines released into cell culture supernatants from stimulated and non-stimulated MNC was performed by ELISA protocols recommended by the manufacturers (R & D Systems, Minneapolis, MN). Coating of ELISA plates (MaxiSorp, Nalge Nunc International, Rochester, NY) with 100 μl of a dilution of capture antibody was carried out overnight at room temperature. Then, plates were washed three times with 300 μl wash buffer (0·05% Tween-20 in PBS, pH 7·4), and dried by tapping on absorbent paper. Blocking was performed by filling the wells with blocking buffer (PBS containing 1% bovine serum albumin, 5% sucrose and 0·05% NaN₃) and incubating for 2 hr. After washing, samples and standards were diluted in diluent buffer [0·1% bovine serum albumin, 0·05% Tween-20 in Tris-buffered saline pH 7·3 (20 mm Trizma base, 150 mm NaCl)] and 100 μl of these dilutions was added to wells, and incubated for 2 hr at room temperature. After washing, 100 μl diluted biotinylated detecting antibody was dispensed into each well, and plates were incubated for 2 hr at room temperature. Plates were then washed and 100 μl of a 1/200 dilution of horseradish peroxidase–streptavidin conjugate (R & D Systems) in PBS, pH 7·2–7·4 plus 1% bovine serum albumin, was added into each well and incubated for 20 min. Plates were washed and 100 μl tetramethylbenzidine (Pierce-Endogen, Woburn, MA) was dispensed into each well and incubated at room temperature for 12 min in darkness. The reaction was stopped using 50 μl/well of stop solution (0·18 m H₂SO₄), and optical densities were read immediately at 450 nm on a plate reader (Bio-Tek Instruments, Kimpton, UK). Results were calculated by interpolating to non-linear regression standard curves. The limit of detection was 7 pg/ml for MIP-1α and MIP-1β, and 4 pg/ml for RANTES.

Statistical analyses

Analyses were performed using GraphPad Prism version 4·00 (GraphPad Software, San Diego, CA). Comparison between treatments and effect of time was performed by two-way analysis of variance (anova) with Bonferroni's post-test to compare treatments against non-stimulated controls. Comparison between two groups was performed using Student's t-test. Differences were considered significant at P-values <0·05.

Results

Demographic and clinical characteristics of cadaveric donors

Lymphoid tissue was surgically harvested from 33 donors and processed as described above. Six donors were females (mean age 29·2 ± 14·8 years, range 6–59 years). Most of donors died of encephalo-cranial trauma and haemorrhagic cerebrovascular disease, and the time between trauma and death was 1·7 ± 3·2 days (range 0·03–18 days). Drugs received by donors during intensive care were predominantly inotropes. Few donors received plasma expanders, and only four donors received blood or plasma transfusions. The characteristics of 11 donors (see below) are described in Table 2.

Table 2.

Demographic and clinical characteristics of cadaveric donors

						PPD status (SI)3
Donor no1	Sex	Age	Time from trauma to death (days)	Cause of death and clinical findings	Drugs received	Spleen	PHLN
3	F	21	0·7	ECT2	Mannitol, dopamine, desmopressin	8·8	3·8
6	M	25	0·2	ECT	Mannitol, dopamine, adrenaline, desmopressin	4·2	1·3
9	M	24	1·1	Multiple trauma	Dopamine, mannitol, dipyrone, ranitidine, tetanol, blood (2 U)	7·7	0·91
12	M	59	0·2	ECT	Ranitidine, metoclopramide, tramadol, mannitol, furosemide, dopamine, adrenaline, potassium chloride	18·0	5·1
13	F	56	0·04	ECT	Dopamine, adrenaline	4·5	4·1
14	M	24	0·4	ECT	Dopamine, adrenaline, desmopressin	17·3	1·3
15	M	24	0·5	ECT	Dipyrone, ranitidine, tramal, metoclopramide, mannitol, dopamine	0·9	nd4
20	M	17	0·2	ECT	Dopamine, desmopressin, mannitol, lidocaine, ranitidine, potassium chloride, blood (1 U)	0·6	1·1
31	M	40	0·2	ECT	Dopamine, vasopressin	1·6	10·8
32	F	50	1·2	ECT	Dopamine, vasopressin	2·8	nd
33	M	22	0·2	Neck trauma	Dopamine, vasopressin, adrenaline, blood (3 U)	6·0	nd

¹The number corresponds to the order in which organs from cadaveric donors arrived to the Laboratory;

²ECT, encephalo-cranial trauma;

³SI, stimulation index = c.p.m. PPD-treated cells/mean c.p.m. unstimulated cells;

⁴nd, not determined.

Tuberculin status of donors was determined by cell proliferation assays using total MNC from both spleen and PHLN. Cells were stimulated with PPD, and PHA was used as a positive control of cell stimulation. The stimulation index (SI) was calculated to determine if a donor was PPD-positive (PPD⁺). This SI was calculated as the ratio between the mean c.p.m. of PPD-treated cells and the mean c.p.m. of unstimulated cells. Donors were considered PPD⁺ when SI ≥ 3 (Table 2). Twenty-three spleens and PHLN from the same number of donors were tested for PPD response, and from these, 11 donors were selected to study PPD-induced modulation of chemokine production and chemokine receptor expression (Table 2). PPD-negative donors were used as controls of PPD response. It is important to note that the PPD response was not always the same in spleen and PHLN cells from the same donor. For some individuals, only cells from spleen proliferated in response to PPD, such as those from donors 6, 9 and 14, whereas for others, such as donor 31, only cells from PHLN responded to PPD (Table 2). Response was considered positive in a donor when cells from either one or more tissues gave an SI ≥ 3. Parameters determined in cells from spleen and PHLN were also measured in MNC from peripheral blood obtained from five healthy PPD⁺ donors.

MNC populations in spleen and lymph nodes

MNC from 33 spleens and PHLN were phenotyped. The composition of the MNC populations differs between these two lymphoid compartments (Table 3). Whereas B cells account for 42·9 ± 10·5% of spleen MNC, these cells reach only 26·9 ± 10·5% in PHLN (P < 0·0001). In contrast, T cells predominate in PHLN (67·1 ± 12·6%) whereas they account for only 31·9 ± 8·9% in spleen (P < 0·0001). CD14⁺ cells compose 1 ± 0·6% of PHLN cells, however, this cell population is larger in spleen (8·2 ± 3·2%, P < 0·0001). The CD56⁺ natural killer cell population was also larger in spleen (9·4 ± 4·7% spleen vs. 2·9 ± 1·3% PHLN, P < 0·0001). When cells were double-stained for CD3 and CD56, a double-positive CD3⁺ CD56⁺ population was found in considerable amounts in spleen (6·3 ± 3·4%), and in far fewer numbers in PHLN (1·3 ± 0·5%, P = 0·003, Table 3). This population may correspond to the natural killer T cells.

Table 3.

Mononuclear cell populations in human spleen and pulmonary hilar lymph nodes

	Spleen		PHLN
Cell type	Mean % ± SD (range)	n	Mean % ± SD (range)	n	P-value1
CD3+	31·9 ± 8·9 (19·3–56·8)	19	67·1 ± 12·6 (38·3–87·0)	23	< 0·0001
CD14+	8·2 ± 3·2 (2·1–14·0)	27	1·0 ± 0·6 (0·1–2·3)	24	< 0·0001
CD19+	42·9 ± 10·5 (19·2–62·3)	27	26·9 ± 10·5 (10·0–43·5)	23	< 0·0001
CD56+	9·4 ± 4·7 (2·8–18·3)	17	2·9 ± 1·3 (1·0–5·6)	22	< 0·0001
CD3+ CD56+	6·3 ± 3·4 (2·7–13·2)	13	1·3 ± 0·5 (0·9–2·3)	6	0·003
CD11c+ DR+	1·0 ± 0·6 (0·3–2·2)	10	0·6 ± 0·2 (0·4–0·9)	5	ns2

¹Student's t-test;

²ns: not significant.

The presence of CD11c⁺ DR⁺ myeloid DC was also tested in a reduced number of donors. This population was determined within total MNC by gating large cells from the forward and side scatter profiles. Then, these large cells were analysed by two-colour staining using anti-CD11c and anti-DR antibodies, and defining the DC population by gating on the CD11c^+high and DR^+high cells, which have been described as the myeloid DC population. Table 3 demonstrates that this population of cells is very small within both lymphoid tissues, accounting for only 1% of MNC.

CCR2 expression by CD3⁺ and CD14⁺/CD63⁺ cells and modulation by mycobacterial antigens

The kinetics of CCR2 expression by MNC from blood, PHLN and spleen was tested in culture after cell stimulation with M. tuberculosis antigens. PMA plus ionomycin was used as a positive control. Flow cytometry analysis was performed to determine CCR2 expression by CD3⁺ T cells, and CD63⁺ (blood) and CD14⁺ (PHLN and spleen) monocytes/macrophages using both gating according to forward and side scatter profiles, and markers for such cell populations.

CD14 is rapidly down-regulated in blood monocytes in culture whereas CD63 expression is stable. However, this latter molecule, although highly expressed on fresh monocytes/macrophages from spleen and PHLN, is rapidly down-regulated in vitro. For this reason, analysis of the monocyte/macrophage population in blood was performed using the CD63 marker, whilst CD14 was used as the marker in spleen and PHLN, both within the monocyte gate.

Fresh CD3⁺ cells from all three lymphoid compartments did not express CCR2, and this molecule was not induced on these cells under any culture conditions (data not shown), consistent with published data.³¹ In contrast, nearly 100% of CD63⁺ cells from blood, and one-third of CD14⁺ cells from spleen and PHLN expressed CCR2 (blood: 95·9 ± 2·3; spleen: 33·1 ± 12·6; PHLN: 28·8 ± 5·8; P < 0·0001). The percentage of unstimulated CD63⁺/CD14⁺ CCR2⁺ cells varied over time in culture according to lymphoid tissue type. Indeed, while the percentage of CD63⁺ CCR2⁺ cells from blood remained constant over time, spleen CD14⁺ CCR2⁺ cells decreased more than 50% after 24 hr and then maintained stable numbers (Fig. 1). In contrast, CD14⁺ CCR2⁺ cells from PHLN increased nearly 100% after 2 hr of culture and retained these levels over time. Contrary to the number of CCR2⁺ cells, the density of CCR2 molecules per cell, denoted by the mean fluorescence intensity (MFI), was significantly affected over time, but only on cells from blood (percentage of CCR2 MFI change with respect to 100% at 0 hr: 2 hr: 177·7 ± 20·9; 8 hr: 91·6 ± 18·6; 24 hr: 79·3 ± 22·4; 48 hr: 211·3 ± 27·9; P < 0·0009). Expression of CCR2 by these cells was decreased by PPD stimulation both at percentage level (Fig. 1; P < 0·0001) and MFI level (2 hr: PPD 138·9 ± 37·1; 8 hr: PPD 54·8 ± 12·6; 24 hr: PPD 38·5 ± 7·5; 48 hr: PPD 63·7 ± 22·3, P < 0·0003). The inhibitory effect of PPD was more evident after 48 hr of stimulation (P < 0·001). No effect of PPD on CCR2 expression by cells from spleen and PHLN was observed, either in cell percentage (Fig. 1) or MFI (data not shown), even though there was clear down-regulation of its expression by PMA/ion (Fig. 1).

Figure 1.

CCR2 expression and its modulation by M. tuberculosis antigens on CD14⁺/CD63⁺ cells from human lymphoid tissues. (a) CCR2 expression was tested on CD14⁺/CD63⁺ cells in freshly isolated MNC obtained from human lymphoid tissues by double-staining with anti-FITC-CD14 or -CD63, and anti-PE-CCR2 labelled mAb. Flow cytometry analysis was performed within the monocyte gate. Insets show staining with the isotype immunoglobulin G2b-PE antibody, which was used as a control of non-specific binding for CCR2. (b) Total mononuclear cells (2 × 10⁵) from three different lymphoid tissues obtained from PPD⁺ donors were cultured in a final volume of 200 μl complete medium in the presence or absence of 10 μg/ml PPD from M. tuberculosis H37Rv. As a positive control of stimulation, 50 ng/ml PMA plus 1 μm ionomycin was used. At different time-points, as indicated, the cells were double-stained with anti-CD14 or anti-CD63-FITC and anti-CCR2-PE mAb and analysed by flow cytometry. CCR2 expression is shown as the mean ± SEM of the percentage increase or decrease, taking expression at 0 hr as 100%. Statistical differences were calculated by two-way anova. Bonferroni's post-test was used to compare PPD against non-stimulated cells, and this difference is shown as: *P < 0·05, **P < 0·001, ***P < 0·0001.

CCR5 expression by CD3⁺ and CD14⁺/CD63⁺ cells and modulation by mycobacterial antigens

CCR5 expression was relatively low in the CD3⁺ lineage (blood: 21·6 ± 9·8%; spleen: 23·1 ± 21·3%; LN: 9·0 ± 5·2%), and there was no difference between the three lymphoid tissues. CCR5 expression had decreased by around 50% after 48 hr of culture. Its expression was strongly down-regulated on CD3⁺ cells from blood (P < 0·0001) in response to PPD, reaching 90% inhibition after 8 hr post-stimulation, and completely abolished at 24 hr (Fig. 2a). A similar but less dramatic effect was also observed in spleen (P < 0·01), because CCR5 was still observed after 48 hr of stimulation. PPD also significantly decreased CCR5 MFI on CD3⁺ cells from blood but not on those from spleen (percentage of CCR5 MFI decrease with respect to 100% at 0 hr: 2 hr: NS 94·4 ± 13·3, PPD 74·7 ± 13·4; 8 hr: NS 71·8 ± 8·6, PPD 38·0 ± 8·2; 24 hr: NS 69·2 ± 9·2, PPD 34·8 ± 8·4; 48 hr: NS 71·5 ± 7·6, PPD 35·6 ± 9·6; P < 0·0001) Interestingly, PPD had no effect on CCR5 expression by CD3⁺ cells from PHLN. Of note, treatment of cells with PMA/ion strongly down-regulated CCR5 on CD3⁺ cells, and was more remarkable in blood where the receptor is undetectable as early as 2 hr post-treatment.

Figure 2.

CCR5 expression and its modulation by M. tuberculosis antigens on CD3⁺ and CD14⁺/CD63⁺ cells from human lymphoid tissues. CCR5 expression was tested on CD3⁺[ a (i)] and CD14⁺/CD63⁺ cells [ b (i)] in freshly isolated MNC obtained from human lymphoid tissues by double-staining with anti-FITC-CD3, -CD14 or -CD63, and anti-PE-CCR5-labelled mAb. Flow cytometry analysis was performed within the lymphocyte gate for CD3⁺ cells, and monocyte gate for CD14⁺/CD63⁺ cells. Insets show staining with the isotype immunoglobulin G2a-PE antibody, which was used as a control of non-specific binding for CCR5. Total mononuclear cells (2 × 10⁵) from three different lymphoid tissues obtained from PPD⁺ donors were cultured and stimulated as described in Fig. 1. At different time-points, the cells were double-stained with [a(ii)] FITC-labelled anti-CD3 or [b(ii)] anti-CD14/anti-CD63 plus PE-labelled anti-CCR5 mAb, and analysed by flow cytometry. CCR5 expression is shown as the mean ± SEM of the percentage increase or decrease, taking expression at 0 hr as 100%. Statistical differences were calculated by two-way anova. Bonferroni's post-test was used to compare PPD against non-stimulated cells, and this difference is shown as: *P < 0·05, **P < 0·001, ***P < 0·0001.

In contrast to CD3⁺ cells, CCR5 expression by cultured monocytes/macrophages from the three lymphoid compartments exhibited a different pattern. Cells from blood had higher CCR5 before culture than those from spleen and PHLN (blood: 70·4 ± 21·0; spleen: 27·6 ± 17·9; PHLN: 26·0 ± 14·5; P < 0·009). Whereas CCR5 expression decreased over time on CD3⁺ cells, that of CD63⁺/CD14⁺ cells was high from time zero and remained relatively constant or even increased over time. PPD did not affect CCR5 on CD14⁺/CD63⁺ cells from any of the three lymphoid tissues studied (Fig. 2b).

CCR7 expression by CD3⁺ and CD14⁺/CD63⁺ cells and modulation by mycobacterial antigens

The number of fresh CD3⁺ CCR7⁺ cells was consistently low in spleen and PHLN when compared with blood (spleen: 8·6 ± 7·1; PHLN: 12·8 ± 15·9; blood: 38·7 ± 16·0; P < 0·009). CCR7 expression on CD3⁺ cells from PHLN but not from spleen or blood increased spontaneously in culture, both as a percentage (Fig. 3; P < 0·002) and as MFI level (MFI ± SEM 0 hr: 100%; 2 hr: 197·5 ± 37·0%, 8 hr: 272·8 ± 72·2%, 24 hr: 515·8 ± 158·8%, 48 hr: 419·5 ± 91·5%; P < 0·01). PPD stimulation had no effect on CCR7 expression either in cell number (Fig. 3) or MFI (data not shown). Of note, PMA/ion dramatically reduced CCR7 expression on blood CD3⁺ cells but had no effect on CD3⁺ cells from spleen and PHLN. CD63⁺/CD14⁺ monocytes from any tissue did not appear to express significant levels of CCR7.

Figure 3.

CCR7 expression and its modulation by M. tuberculosis antigens on CD3⁺ cells from human lymphoid tissues. (a) CCR7 expression was tested on CD3⁺ cells in freshly isolated MNC obtained from human lymphoid tissues by double-staining with anti-CD3 and anti-CCR7 mAb. Flow cytometry analysis was performed within the lymphocyte gate. Insets show staining with the isotype immunoglobulin G2a-FITC/PE antibody, which was used as a control of non-specific binding for CCR7. (b) Total mononuclear cells (2 × 10⁵) from three different lymphoid tissues obtained from PPD⁺ donors were cultured and stimulated as described in Fig. 1. At different time-points, the cells were double-stained with FITC- or CyChrome-labelled anti-CD3 plus FITC- or PE-labelled anti-CCR7 mAb, and analysed by flow cytometry. CCR7 expression is shown as the mean ± SEM of the percentage increase or decrease, taking expression at 0 hr as 100%. Statistical differences were calculated by two-way anova. Bonferroni's post-test was used to compare PPD against non-stimulated cells, and this difference is shown as: *P < 0·05, **P < 0·001, ***P < 0·0001.

Interestingly, a specific population of CD3^– lymphocytes from all tissues examined that did not express CCR7 at time zero, dramatically up-regulated expression over time up to more than 100-fold. This up-regulation was independent of the presence of M. tuberculosis antigens. It is expected that this population represents B cells, as it is known that B cells strongly express CCR7 levels (data not shown).

Table 4 summarizes the expression of chemokine receptors over time on MNC from the three lymphoid tissues studied, and their modulation by M. tuberculosis antigens.

Table 4.

Chemokine receptor expression and their modulation by Mtb antigens on CD3⁺ and CD63/CD14⁺ cells from human lymphoid tissues

		Blood			Spleen			PHLN
Phenotype	Time (hr)	NS	PPD2	PMA/ion3	NS	PPD	PMA/ion	NS	PPD	PMA/ion
CD63+/ CD14+CCR2+4	2	100·41 (1·0)	95·3 (1·9)	63·4 (4·4)	118·8 (42·8)	120·8 (34·6)	47·0	184·5 (29·6)	159·3 (31·7)	180·0
	8	98·1 (0·7)	92·2 (2·3)	47·4 (7·5)	98·5 (25·3)	63·7 (24·4)	24·0	153·0 (38·2)	148·8 (41·5)	74·0
	24	89·1 (2·4)	71·3 (7·7)	19·4 (10·9)	34·7 (7·5)	32·3 (22·3)	0·00	145·5 (57·0)	143·0 (42·9)	28·0
	48	95·4 (2·6)	66·0 (10·2)	4·1 (4·1)	54·5 (21·7)	45·2 (19·0)	0·00	146·0 (78·2)	138·5 (59·1)	25·0
CD63+/ CD14+CCR5+	2	143·9 (26·4)	121·1 (23·3)	64·2 (26·4)	113·3 (18·9)	117·3 (18·1)	109·0 (0·0)	149·0 (15·6)	173·0 (34·2)	55·0
	8	141·0 (24·1)	111·9 (22·4)	44·6 (27·4)	85·0 (16·3)	81·7 (16·8)	130·0 (57·0)	187·3 (14·4)	197·3 (21·4)	48·0
	24	141·2 (24·9)	106·9 (23·4)	16·8 (9·2)	150·0 (25·2)	103·0 (23·3)	96·0 (0·0)	283·3 (42·4)	281·3 (12·2)	31·0
	48	140·9 (26·8)	120·8 (18·1)	7·6 (7·6)	177·7 (55·8)	153·0 (28·7)	109·0 (0·0)	303·3 (40·4)	247·3 (67·4)	56·0
CD3+CCR5+	2	97·0 (4·4)	83·8 (9·0)	5·5 (2·3)	197·8 (76·3)	124·2 (40·9)	74·8 (31·9)	104·5 (7·6)	109·3 (12·4)	66·0 (24·0)
	8	77·6 (9·5)	26·6 (9·3)	6·7 (3·9)	97·1 (19·7)	60·4 (23·0)	45·4 (13·9)	95·5 (20·2)	85·0 (9·8)	48·5 (18·5)
	24	55·2 (4·8)	7·1 (3·8)	0·0 (0·0)	86·2 (30·8)	19·2 (7·8)	27·0 (9·8)	65·0 (4·9)	56·7 (4·0)	27·0 (0·0)
	48	58·9 (6·4)	8·2 (2·7)	0·0 (0·0)	56·1 (16·8)	20·3 (8·3)	24·0 (11·3)	45·5 (8·6)	37·7 (14·2)	24·0 (0·0)
CD3+CCR7+	2	95·3 (4·2)	103·0	65·9	120·4 (15·9)	116·2	120·0	192·0	188·7	328·0
			8·5	9·7		12·6	8·0	33·3	24·4
	8	116·1	124·4	5·2	145·0	154·6	159·5	134·7	200·7	161·0
		17·0	17·8	3·6	17·8	19·1	7·5	38·2	62·4
	24	122·0	141·0	17·8	186·0	194·8	159·5	352·0	424·3	279·0
		14·8	20·5	12·4	37·6	52·4	30·5	152·3	153·0
	48	125·7	150·6	22·1	178·8	189·8	188·0	218·3	206·3	296·0
		23·6	24·2	22·1	16·0	18·3	0·0	54·0	58·9

¹Values are the mean ± (SEM), and are calculated as the percentage of increase or decrease of CCR expression at 0 hr taking such expression as 100%;

²PPD was used at 10 μg/ml;

³PMA and ionomycin were used at 50 ng/ml and 1 μm, respectively;

⁴CCR2 and

⁵CCR5 expression was tested on blood CD63⁺ cells, and on spleen and PHLN CD14⁺ cells.

CC-chemokine production by MNC from lymphoid tissues after stimulation with mycobacterial antigens

It has been shown previously that CC-chemokines binding CCR5, such as MIP-1α, MIP-1β and RANTES, have the capacity to induce ligand-induced receptor internalization.³² Since CCR5 down-regulation on CD3⁺ cells induced by PPD was observed in cells from blood and spleen, the question arose as to whether CC-chemokines, ligands of CCR5, were produced by MNC from the three lymphoid tissues in the presence of PPD. To answer this question, three CC-chemokines, CCL3/MIP-1α, CCL4/MIP-1β, and CCL5/RANTES, were tested by ELISA in cell culture supernatants from the same CCR kinetics experiments, and also in cell cultures from PPD-negative donors. Figure 4 shows that MNC from the three lymphoid tissues spontaneously produced these three chemokines. However, PPD stimulation induced higher levels of MIP-1α and MIP-1β in blood and spleen but not in PHLN, whereas PPD induced RANTES in all tissues (Fig. 4). It is noteworthy that in blood PPD was as potent a stimulus as PMA/ion. PPD-induced chemokine production by cells from PPD-negative donors was not significantly different to that of unstimulated cells (data not shown).

Figure 4.

Effect of Mtb antigens on the kinetics of CC chemokine production by mononuclear cells from human lymphoid tissues. Total mononuclear cells (2 × 10⁵) from three different lymphoid tissues were obtained from PPD⁺ donors and cultured as described in Fig. 1. At different time-points, the cell culture supernatants were collected and the CC chemokines (a) MIP-1α, (b) MIP-1β and (c) RANTES were tested by ELISA. Chemokine production is expressed as the mean of pg/ml ± SEM on a log scale. Statistical differences were calculated by two-way anova. Bonferroni's post-test was used to compare PPD against non-stimulated cells, and this difference is shown as: *P < 0·05, **P < 0·001, ***P < 0·0001.

Discussion

We have defined chemokine receptor expression and its in vitro modulation by M. tuberculosis antigens on MNC isolated from three human lymphoid tissues: blood, spleen and PHLN.

There were significant differences, before culture, in CCR2, CCR5 and CCR7 expression both at cell lineage and type of lymphoid tissue. CD3⁺ T cells from any tissue showed no CCR2 expression, whereas this receptor was highly expressed on the monocyte/macrophage lineage, and those from blood expressed threefold more CCR2 than cells from spleen and PHLN. In contrast, CCR5 was expressed by both T cells and monocytes/macrophages. However, whereas there was no difference in CCR5 expression on T cells between tissues, its expression on monocytes/macrophages from blood was around twice of that in spleen and PHLN. CCR7 was expressed by CD3⁺ cells from all tissues but was hardly detectable on monocytes/macrophages. Expression by CD3⁺ cells was threefold higher in blood than in spleen and up to fivefold higher than in PHLN. Such differential expression of chemokine receptors may have functional significance in terms of positioning lymphocytes and monocytes/macrophages within lymphoid compartments. The higher expression of these receptors in cells from blood may render cells more responsive to chemokine gradients induced either in inflammatory sites, or in secondary lymphoid tissues. Interestingly, CD3⁺ cells from PHLN but not from blood and spleen dramatically enhanced CCR7 expression in culture. It is possible that EBI1 ligand chemokine, one of the CCR7 ligands, is produced in high amounts in PHLN in vivo, maintaining low levels of CCR7 expression, and that these levels are highly induced once the cells are removed from their microenvironment.³³

Antigens for M. tuberculosis modulated expression of chemokine receptors in vitro. Such modulation was specific for receptor type, cell lineage and type of lymphoid tissue. They down-regulated expression of CCR5 and CCR2 whereas expression of CCR7 was unaffected. CCR5 inhibition by M. tuberculosis antigens was specific of CD3⁺ cells, was observed only in cells from blood and spleen but not PHLN, and was more profound and rapid in cells from blood than from spleen. CCR2 was only marginally down-regulated on monocytes/macrophages, and such an effect was only observed in cells from blood. It is noteworthy that PMA/ion strongly inhibited expression of the receptors regardless of cell lineage and lymphoid tissue. Nevertheless, expression of CCR5 on CD14⁺ cells from spleen, and that of CCR7 on CD3⁺ cells from spleen and PHLN was unaffected by PMA/ion.

Expression of chemokine receptors, such as CCR2 and CCR5, has been implicated in the recruitment of phagocytic cells to the site of M. tuberculosis invasion.^15,29,34 Such receptors are expressed on monocytes/macrophages and DC, as well as activated T lymphocytes, and induce their migration and localization at the site of inflammation. The ligands of these receptors are chemokines such as MCP-1 for CCR2, and MIP-1α, MIP-1β and RANTES for CCR5. Production of these chemokines is induced by M. tuberculosis infection on local epithelial,¹¹ mesothelial³⁵ and inflammatory cells.^25,26,36 Migration of monocytes/macrophages to the site of inflammation is important for initiation of the immune response against M. tuberculosis. These cells phagocytose M. tuberculosis and provide an important cellular niche during infection, and a source of antigenic stimulation. Furthermore, accumulation of monocytes/macrophages at the site of infection, and their subsequent activation enhance production of inflammatory chemokines that recruit additional cells facilitating granuloma formation. We showed here that CCR5 is refractory to down-regulation on the monocytes/macrophages after M. tuberculosis antigen stimulation, and that CCR2 is only marginally modulated (Figs 1 and 2). Such persistence of CCR5 and CCR2 expression was more remarkable given the levels of MCP-1 (data not shown), and MIP-1α, MIP-1β and RANTES (Fig. 4) produced in response to PPD. These results may be relevant because it has been established that CC-chemokines can inhibit expression of receptors such as CCR5 through ligand–receptor interaction.^32,37,38 Although we do not know the mechanisms that explain this resistance to receptor down-regulation, persistence of CCR2 and CCR5 on monocytes/macrophages may contribute to homing and retention of these cells at the inflammatory site.^15,19 After homing, these cells would actively participate in controlling M. tuberculosis facilitating granuloma formation, and inducing recruitment of additional inflammatory and effector cells through production of chemokines.¹⁵

The observation that monocytes/macrophages expressed little if any CCR7, involved in migration of cells out of inflammation sites towards secondary lymphoid tissues³⁹ reinforces the concept that monocytes/macrophages are programmed to remain in target tissues. These data are particularly relevant because the monocytes/macrophages that we studied were isolated from three different human lymphoid organs, including regional PHLN where actual M. tuberculosis antigen presentation takes place.

We show here that CCR5 expression in fresh CD3⁺ cells from blood, spleen and PHLN was relatively low, 9·0–23·1%. In addition, it has been described that CCR2 and CCR5 expression is restricted to a specific population of CD4⁺ memory T cells.³¹ Although we did not study CD3⁺ subsets, the possibility that the cells that down-regulated expression of CCR5 after PPD stimulation correspond to memory T cells is very high, because those cells obtained from PPD-negative donors did not down-regulate CCR5 (data not shown). In this regard, increased expression of CCR5 by blood CD4⁺ cells in patients with tuberculosis has been described before.⁴⁰ More recently, a higher number of CD4⁺ CCR5⁺ memory T cells was reported in bronchoalveolar lavages and peripheral blood from patients with tuberculosis, when compared to healthy controls.⁴¹ Nonetheless, and in agreement with the results shown here, although the proportion of CD4⁺ CCR5⁺ memory T cells was higher in the lungs of tuberculosis patients, the density of CCR5 molecules per cell, as indicated by mean fluorescence intensity, was down-regulated in comparison to that of healthy controls,⁴¹ a phenomenon which may be explained by local production of CCR5 ligands. Since it has been reported that polarized T helper type 1 (Th1) cells preferentially express CCR5,^42,43 such chemokine receptor expression would facilitate efflux of effector cells from regional LN and subsequent positioning at the site of inflammation,^40,41 where high gradients of CCR5 ligands are induced by infection. Once such cells have been recruited to these sites, Th1 cells produce interferon-γ and consequently activate local macrophages optimizing mycobacterial control.

Mechanisms that control down-regulation of CCR5 induced by PPD are unknown, however, such inhibition may be the result of ligand-induced internalization³² induced by MIP-1α, MIP-1β and RANTES produced at the site of infection. In this context, the difference in CCR5 down-regulation on T cells from the three different lymphoid tissues could be related to differential CC-chemokine production induced by PPD. Figure 4 shows that PPD induced higher levels of CC-chemokines by cells from blood than spleen and PHLN. There is a maximum of MIP-1α and MIP-1β production close to log 5, which is reached by the non-specific treatment PMA/ion. PPD only achieves this level of stimulation in blood MNC, with an intermediary level in spleen, and a lower level in PHLN. Interestingly, this difference in chemokine production by these lymphoid tissues closely parallels the differential effect of PPD on CCR5. Also, these data demonstrate that cells from blood are, in this respect, more susceptible to the effect of PPD.

Other studies have also demonstrated CCR5 modulation on T cells by M. tuberculosis antigens.^23,24 Glatzel et al. showed that CCR5 is highly expressed by γδ T cells, but not by αβ T cells, and that such expression was down-regulated by heat-killed M. tuberculosis H37Ra. Inhibition of CCR5 was associated with chemokine production because antibodies against RANTES, MIP-1α and MIP-1β restored its expression.²³ On the other hand, Juffermans et al. reported that lipoarabinomannan, a mycobacterial cell wall component, induced CCR5 up-regulation by CD4⁺ T cells.²⁴ Discrepancy between these reports may be the result of either a different source of antigens (non-virulent versus virulent M. tuberculosis), or the type of antigens used. Mechanisms of cell activation by total M. tuberculosis extract are clearly different to those by lipoarabinomannan. We have used PPD from virulent M. tuberculosis, which is a source of antigens known to induce Th1 responses.

It is worth emphasizing that in this study, blood cells were obtained from healthy individuals, whereas PHLN and spleen cells were obtained from cadaveric donors. Therefore, differences in CCR expression and chemokine production observed between blood cells and the other two lymphoid tissues may have been influenced by a number of additional factors. Cadaveric donors are in general individuals who have suffered major trauma and received potent drugs, which may themselves affect the immune response. Also, the death process may itself induce profound immune modulation.

In conclusion, this work describes, for the first time, in vitro regulation of chemokine receptors and chemokine production induced by M. tuberculosis antigens on MNC from three human lymphoid organs. The differential expression of chemokines and chemokine receptors on such lymphoid cells goes someway to explaining programmed cell migration in human M. tuberculosis infection, responsible for containment of the microorganism, and development of the host immune response. Better understanding of the intricate network of cytokines and chemokines involved in activation and migration of cells within lymphoid tissues will help to develop newer and better methods of controlling tuberculosis through both new drugs and vaccines.

Abbreviations

CCR

CC chemokine receptor

CM

complete medium

DC

dendritic cells

FITC

fluorescein isothiocyanate

LN

lymph node

MCP-1

monocyte chemoattractant protein-1

MFI

mean fluorescence intensity

MIP-1

macrophage inflammatory protein-1

MNC

mononuclear cells

PE

phycoerythrin

PMA

phorbol 12-myristate 13-acetate

PBS

phosphate-buffered saline

PHA

phytohaemagglutinin

PHS

pooled human serum

PHLN

pulmonary hilar lymph nodes

PPD

purified protein derivative

RANTES

regulated on activation, normal T cell expressed and secreted

Th1

T helper type 1