Monocyte chemoattractant protein-1 (MCP-1) inhibits the intestinal-like differentiation of monocytes

Abstract

Monocytes (MO) migrating into normal, non-inflamed intestinal mucosa undergo a specific differentiation resulting in a non-reactive, tolerogenic intestinal macrophage (IMAC). Recently we demonstrated the differentiation of MO into an intestinal-like macrophage (MAC) phenotype in vitro in a three-dimensional cell culture model (multi-cellular spheroid or MCS model). In the mucosa of patients with inflammatory bowel disease (IBD) in addition to normal IMAC, a reactive MAC population as well as increased levels of monocyte chemoattractant protein 1 (MCP-1) is found. The aim of this study was to investigate the influence of MCP-1 on the differentiation of MO into IMAC. MCS were generated from adenovirally transfected HT-29 cells overexpressing MCP-1, macrophage inflammatory protein 3 alpha (MIP-3α) or non-transfected controls and co-cultured with freshly elutriated blood MO. After 7 days of co-culture MCS were harvested, and expression of the surface antigens CD33 and CD14 as well as the intracellular MAC marker CD68 was determined by flow-cytometry or immunohistochemistry. MCP-1 and MIP-3α expression by HT-29 cells in the MCS was increased by transfection at the time of MCS formation. In contrast to MIP-3α, MCP-1 overexpression induced a massive migration of MO into the three-dimensional aggregates. Differentiation of IMAC was disturbed in MCP-1-transfected MCS compared to experiments with non-transfected control aggregates, or the MIP-3α-transfected MCS, as indicated by high CD14 expression of MO/IMAC cultured inside the MCP-1-transfected MCS, as shown by immunohistochemistry and FACS analysis. Neutralization of MCP-1 was followed by an almost complete absence of monocyte migration into the MCS. MCP-1 induced migration of MO into three-dimensional spheroids generated from HT-29 cells and inhibited intestinal-like differentiation of blood MO into IMAC. It may be speculated that MCP-1 could play a role in the disturbed IMAC differentiation in IBD mucosa.

Introduction

Intestinal macrophages (IMAC) in normal non-inflamed mucosa represent a specific non-reactive tolerogenic macrophage (MAC) population. Several activation-associated monocyte (MO)/MAC-specific surface antigens such as CD14, CD16, CD11b or CD11c are down-regulated in these cells [1–7]. The expression of T cell co-stimulatory molecules CD80 and CD86 is reduced [8], resulting in an inability to induce antigen-specific T cell expansion.

In addition, we have shown recently that the expression of the receptors for bacterial wall products such as lipoteichoic acid (LTA) or lipopolysaccharide (LPS) and the Toll-like receptors (TLRs) 2 and 4 is also down-regulated in IMAC on transcriptional and translational levels [9]. Typical MO/MAC functions such as the generation of superoxide radicals (oxidative burst reaction) are down-regulated or absent in normal mucosal IMAC, due to a lack of nicotinamide adenine dinucleotide phosphate (NADPH)-oxidase subunit expression [10]. Taken together, these data indicate that normal IMAC are involved in the induction of tolerance, rather than in the induction of immune responses.

When the differentiation of MO into IMAC is impaired, a reactive MAC population may persist followed by permanent inadequate immune responses leading to chronic inflammation. In the healthy gut an inadequate immune response needs to be avoided and tolerance against antigens and commensal bacteria needs to be achieved. In inflammatory bowel disease (IBD) this situation is disturbed and strong immune responses against autoantigens, commensal bacteria or food components are found. IMAC populations with high expression of activation-associated antigens (CD14, CD16, CD80 and CD86) can be detected in the mucosa of IBD patients [3,8,11]. Recently we demonstrated that the lysosomal aspartate protease cathepsin D is selectively up-regulated in IBD IMAC [12]. The presence of this reactive IMAC population is considered to be one reason for the chronification of the inflammatory process.

A number of proinflammatory cytokines and chemokines are found to be increased in the mucosa of IBD patients, such as interleukin (IL)-8, IL-1, tumour necrosis factor (TNF) as well as monocyte chemoattractant protein (MCP)-1, 2 and 3 and macrophage inflammatory protein 1 alpha (MIP-1α), 1β and 3α, among others [13]. They are highly expressed by different cell populations [13]. The presence of MCP-1 is considered to be responsible for the massive invasion of blood MO and granulocytes to the inflamed tissue [14,15].

As tissue-specific MAC differentiation depends on specific local conditions such as cytokine milieu [16,17] or the composition of extracellular matrix [18–22], we examined the influence of the chemokines MCP-1 and MIP-3α on the intestinal-like differentiation of blood MO in the three-dimensional multi-cellular spheroid (MCS) model.

Materials and methods

MO isolation

Primary blood MO were obtained by leukapheresis of healthy donors, followed by density gradient centrifugation over Ficoll/Hypaque as described previously [21]. The study was approved by the ethics committee of the University of Regensburg. Informed consent was obtained from all MO donors.

Cell culture

The intestinal epithelial cell line HT-29 was cultured in Dulbecco’s modified essential medium (DMEM) supplemented with 10% fetal calf serum (FCS), 1% sodium-pyruvate, 1% non-essential amino acids and 1% penicillin/streptomycin under standard tissue culture conditions. The cell line QBI-293 A (Qbiogene, Heidelberg, Germany) for transfection of the AdEasy™ plasmid was cultured in DMEM high glucose (4·5 g/l) supplemented with 10% FCS, 1% sodium pyruvate and 1% penicillin/streptomycin.

Generation of MCS

MCS were generated according to the liquid overlay culture technique [21]; 4 × 10³ cells (either transformed with the MCP-1, MIP-3α virus, the empty AdEasy construct or non-transformed control cells) suspended in 0·2 ml medium/well were seeded in agarose-coated wells of 96-well plates and cultured under static conditions. The MCP-1 adenovirus construct was provided by Meenhard Herlyn (Wistar Institute, Philadelphia, PA, USA) [23]. The MIP-3α adenovirus construct was generated as described below. Seeded cells formed small aggregates within 3 days, which enlarged during further culture. After 3 days half the medium was replaced by fresh medium. After 3 more days of culture, MCS had formed and were used for experiments.

MCS supernatants (0·1 ml) were replaced by freshly isolated MO in 0·1 ml medium supplemented with 2% of human antibody serum. Co-cultures of MCS with MO were harvested after 24 h, 3 days and 7 days for immunohistochemical and flow cytometric analysis. Basal expression of CD14 by MO was monitored in each experiment by fluorescence activated cell sorter (FACS). No CD14 negative monocytes were detectable.

Generation of the MIP-3α adenovirus construct

The MIP-3α-overexpression adenovirus was constructed with the AdEasy™ system [24]. In brief, the MIP-3α cDNA was cloned into a transfer vector. The resulting plasmid was linearized with PmeI and co-transformed into Escherichia coli strain BJ5183 together with the viral DNA plasmid pAdEasy-1 by electroporation. The recombinant AdEasy™ plasmid was amplified in DH5α competent cells by electroporation, selected with kanamycin and screened by restriction enzyme analysis. The recombinant adenoviral construct was cleaved with PacI to expose its inverted terminal repeats and transfected into QBI-293 A cells to produce viral particles.

Transfection of HT-29 cells

HT-29 cells (1 × 10⁶) were grown in six-well culture plates in normal tissue culture medium until they showed about 80% confluence. First, the optimum multiplicity of infection (moi) for HT-29 cells with the Ad5 viral particles was determined. Cells were incubated with different numbers of viral particles for 48 h. Incubation of the cells with moi = 10 or 5 resulted in cell lysis after 48 h; with moi = 1 the cells were viable after 48 h and were used for further experiments.

When cells showed about 80% confluency the Ad5_MCP-1 or Ad5_MIP-3α viral constructs, the empty control virus (moi = 1) or combinations of two of these constructs (moi = 2) were added in 0·5–1·0 ml cell culture medium containing 5% FCS. After 1·5 h of incubation 2 ml of normal tissue culture medium were added. Cells were harvested and used for generation of MCS after 48 h. Cell supernatants were collected and used for enzyme-linked immunosorbent assay (ELISA) to confirm transfection efficiency.

To block the MCP-1 effect in the MCS model a neutralizing antibody (R&D systems, Wiesbaden, Germany; cat. number AB-279-NA) was added to the cultures at days 1, 3 and 5 of the culture period.

ELISA

Successful transfection with the MCP-1 and MIP-3α viral particles and consecutive protein expression was confirmed by ELISA. Cell culture supernatants from transfected HT-29 cells were collected 48 h after virus addition. Supernatants were centrifuged and used for determination of MCP-1 and MIP-3α by commercially available ELISA kits (R&D Systems).

Immunohistochemistry

Immunohistochemical staining was carried out according the standard alkaline phosphatase anti-alkaline phosphatase (APAAP)-technique [25]. The following monoclonal antibodies against MO/MAC-antigens were used: anti-CD68 (clone: KP1, Dako, Hamburg, Germany), anti-CD11b (clone: BEAR1, Immunotech, Hamburg, Germany), anti-CD11c (clone: BU15, Immunotech), anti-CD14 (clone: RMO52, Immunotech) and anti-CD16 (clone: 3G8, Immunotech).

Flow cytometry

Flow cytometry was performed using a Coulter EPICS® XL-MCL (Coulter, Krefeld, Germany). Cells were double-stained with a fluorescein isothiocyanate (FITC)-conjugated anti-CD14 antibody (clone Tük4; Coulter) and a phycoerythrin (PE)-conjugated anti-CD33 antibody (clone MY9; Coulter) as described previously.

Data acquisition and analysis were performed using win-mdi software (see: http://facs.scripps.edu/help/html/).

Statistical analysis was performed using SigmaStat software. Appropriate tests were used as indicated.

Results

Transfection of HT-29 cells

HT-29 cells were transfected with different numbers of adenoviral particles to find the optimal moi for this cell line. The optimal moi for transfection of HT-29 cells with Ad5_MCP-1 and Ad5_MIP-3α viral particles was 1; transfection with a higher moi [5,10] resulted in cell lysis. After 48 h of incubation of HT-29 cells with viral particles, an increased production of the chemokine of interest into the culture supernatants was detected. Analysis of the cell culture supernatants by ELISA showed an approximately 50-fold increase of MCP-1 levels in monolayer cultures 48 h after transfection with the Ad5_MCP-1 viral construct (1932 ± 1631 pg/ml) compared to non-transfected control cultures (40 ± 34 pg/ml). After generation of spheroids and co-culture of the spheroids with MO for 7 days, MCP-1 levels were increased spontaneously to 1016 ± 970 pg/ml. Supernatants of MCS containing Ad5_MCP-1-transfected HT-29 cells still had significantly higher MCP-1 concentrations (4645 ± 100 pg/ml, Fig. 1a). MIP-3α levels in culture supernatants 48 h after transfection with Ad5_MIP-3α reached from 27 to 85 pg/ml compared to 0–18 pg/ml in non-transfected controls (58·1 ± 23·9 versus 8·7 ± 7·9 pg/ml). After co-culture of the spheroids with MO for 7 days MIP-3α levels in the supernatants were about twice as high in MCS, consisting of transfected cells as in MCS from non-transfected cells (448 ± 353 pg/ml compared to 243 ± 214 pg/ml, Fig. 1b).

Fig. 1.

Chemokine levels in supernatants of transfected cells. (a) Monocyte chemoattractant protein (MCP)-1 levels in the supernatants of non-transfected controls after 48 h of culture were 40 ± 34 pg/ml compared to 1932 ± 1631 pg/ml in HT-29 cells 48 h after Ad5_MCP-1 transfection. In supernatants of spheroids cocultured with monocyte/macrophage (MO/MAC) for 7 days the mean value of MCP-1 in the supernatant was 1016 ± 970 pg/ml in non-transfected controls compared to 4645 ± 100 pg/ml in experiments with Ad5_MCP-1-transfected cells. ▪, Control; □, Ad5_MCP-1 transfected. (b) The mean value of macrophage inflammatory protein 3 alpha (MIP-3α) in the supernatants of HT-29 cells before generation of spheroids was 8·7 ± 7·9 pg/ml in non-transfected control cells compared to 58·1 ± 23·9 pg/ml in MIP-3α-transfected cells. MIP-3α values in supernatants of multi-cellular spheroid (MCS) after 7 days of co-culture with MO/MAC were 243 ± 214 pg/ml in aggregates consisting of non-transfected HT-29 cells and 448 ± 353 pg/ml in aggregates consisting of Ad5_MIP-3α-transfected cells. ▪, Control; □, Ad5_MCP-3α transfected.

MCP-1 but not MIP3α induced migration of MO into MCS

After 7 days of incubation the three-dimensional aggregates were analysed by flow cytometry. MO/MAC inside the different MCS were identified by the MO/MAC specific surface antigen CD33. Figure 2 shows the FACS dot-plots of representative experiments. In aggregates of non-transfected control cells, Ad5_Null- and MIP-3α-transfected HT-29 cells only a small percentage of all cells could be identified as MO/MAC (Fig. 2a), whereas transfection of HT-29 cells with Ad5_MCP-1 or combinations of Ad5_Null and Ad5_MIP-3α with Ad5_MCP-1 resulted in a massive migration of MO/MAC into the three-dimensional aggregates (Fig. 2b).

Fig. 2.

Macrophages (MAC) inside the aggregates were identified by positive staining for the monocyte (MO)/MAC specific surface marker CD33. (a) CD33⁺ MO/MAC cultured in non-transfected, Ad5_Null- and macrophage inflammatory protein 3 alpha (MIP-3α)-transfected HT-29 multi-cellular spheroid (MCS) for 7 days. Only a low number of CD33⁺ cells can be detected. (b) CD33⁺ MO/MAC in Ad5_MCP-1-, Ad5_Null/Ad5_MCP-1- and Ad5_MIP-3α/Ad5_MCP-1-transfected HT-29 MCS after 7 days of culture. Clearly, many more CD33⁺ cells can be isolated from the MCS if monocyte chemoattractant protein (MCP)-1 was overexpressed.

Evaluation of all experiments showed that 4·5 ± 3·3% CD33⁺ MO/MAC migrated into MCS from non-transfected HT-29 cells (n = 11) compared to 30·9 ± 24·7% CD33⁺ MO/MAC inside Ad5_MCP-1-transfected aggregates (n = 10). The number of invading cells was statistically different (P = 0·022, Mann–Whitney rank sum test). In HT-29 aggregates transfected with the empty Ad5 (Ad5_Null) 5·1 ± 3·3% of all cells expressed CD33 (n = 4) compared to 37·9 ± 8·0% CD33⁺ cells in Ad5_Null/Ad5_MCP-1 co-transfected MCS (n = 4). Statistical analysis (Mann–Whitney rank sum test) indicated that these values were significantly different (P = 0·026). Transfection of HT-29 cells with Ad5_MIP-3α resulted in 4·9 ± 6·1% MO/MAC inside the MCS (n = 7) (not statistically different to non-transfected MCS), whereas transfection with Ad5_MIP-3α/Ad5_MCP-1 resulted in a significantly higher number of MO/MAC inside the aggregates (36·1 ± 19·2% CD33⁺ cells, n = 4, P = 0·005, t-test) (Fig. 3).

Fig. 3.

Percentage of CD33⁺ monocyte/macrophage (MO/MAC) in multi-cellular spheroid (MCS) after 7 days of co-culture. In control experiments with spheroids from non-transfected HT-29 cells 4·5% of total cells could be identified as MO/MAC (CD33⁺ cells) after 7 days (n = 11) compared to 30·9% MO/MAC of total cells in aggregates from Ad5_MCP-1-transfected cells (n = 10). In spheroids generated from HT-29 cells transfected with the empty control virus 5·1% of total cells were MO/MAC (n = 4), co-transfection of HT-29 cells with Ad5_Null and Ad5_MCP-1 resulted in 37·9% of MO/MAC inside the aggregates (n = 4). Experiments with MCS from HT-29 cells transfected with Ad5_MIP-3α revealed 4·9% CD33⁺ cells inside the aggregates after 7 days of co-culture (n = 7), co-transfection with Ad5_macrophage inflammatory protein 3 alpha (MIP-3α) and Ad5_MCP-1 resulted in 36·1% of CD33⁺ cells inside the aggregates. □, Alone; ▪, with monocyte chemoattractant protein (MCP)-1.

MCP-1 inhibits differentiation of IMAC

Differentiation of MO into intestinal-like MAC was determined by the expression of CD14, which is present on blood MO but not detectable on IMAC from normal, non-inflamed mucosa and down-regulated in control HT-29 MCS, as established recently [21]. Immunohistochemical analysis of HT-29-MCS co-cultured with MO/MAC for 7 days showed CD68 positive MAC inside non-transfected control spheroids (Fig. 4a), with no expression of the activation-associated MO/MAC-specific antigen CD14 (Fig. 4b). The vast majority of the cells in the non-transfected MCS were epithelial cells, as demonstrated with the epithelial cell marker EP4 (Fig. 4c). Isotype controls supported the specificity of these results (Fig. 4d). In MCP-1-transfected spheroids many more CD68 positive cells could be detected than in the control aggregates (Fig. 4e), most of them expressing CD14 (Fig. 4f). Again, HT-29 cells were stained with EP4 (Fig. 4g). Specificity of the staining was controlled with isotype antibody incubation (Fig. 4h).

Fig. 4.

Immunohistochemical staining of multi-cellular spheroid (MCS) co-cultured with monocytes (MO) for 7 days. Aggregates were stained for the intracellular MO/macrophage (MAC) specific marker CD68 and the differentiation and activation associated surface antigen CD14. (a–d) Non-transfected control MCS; (e–h) Ad-MCP-1-transfected MCS. (a) In non-transfected control MCS CD68⁺ MAC could be identified inside the aggregates after 7 days of co-culture; (b) no expression of CD14 could be detected in theses cells; (c) staining of EP4 revealed the epithelial character of the vast majority of cells; (d) isotype control staining; (e) transfection of HT-29 cells with Ad_MCP-1 resulted in a higher rate of monocyte migration into the aggregates; (f) a much higher number compared to non-transfected MCS expressed CD14; (g) EP4 staining; (h) isotype control (original magnification ×400).

Analysis of MCS by flow cytometry confirmed the results obtained by immunohistochemistry. Figure 5 shows dot-plots of CD33 (PE, x-axis) versus CD14 (FITC, y-axis) fluorescence. Clearly, there is a low number of cells in the upper right quadrant in Fig. 5a (non-transfected, Ad_Null and Ad_Mip3α-transfected MCS), indicating a low number of CD33⁺/CD14⁺ MACs. In contrast, in Fig. 5b (always transfection with Ad_MCP-1) almost half the cells found in the MCS were CD33⁺/CD14⁺. Whereas in Fig. 5a in the lower right quadrant the intestinal-like MAC population can be found (CD33⁺/CD14^–), this population is virtually absent in the dot-plots of Fig. 5b.

Fig. 5.

Dot-plots of CD33/CD14 double-positive cells. (a) In non-transfected control spheroids 1·3% of total cells showed expression of CD33 and CD14. Transfection of HT-29 cells with Ad5_Null also resulted in 1·3% CD33⁺/CD14⁺) cells. In spheroids consisting of HT-29 cells transfected with Ad5_macrophage inflammatory protein 3 alpha (MIP-3α) 1·2% of total cells were positive for CD33 and CD14. A small population of CD33⁺/CD14^– intestinal-like macrophages was present in all analyses (lower right quadrant, arrows). (b) In spheroids from Ad5_MCP-1-transfected cells 26·2% of total cells showed expression of CD33 and CD14 after 7 days. Co-transfection with Ad5_Null and Ad5_MCP-1 resulted in similar effects with 30·8% of total cells co-expressing CD33 and CD14. Co-transfection with Ad5_MIP-3α and Ad5_MCP-1 resulted in 34·6% of total cells expressing CD33 and CD14. No CD33⁺/CD14^– population of cells can be detected.

Quantification revealed 1·3 ± 0·9% CD14⁺ cells in non-transfected control aggregates (n = 11). The same result was obtained with MCS, consisting of HT-29 cells transfected with the empty Ad5 virus with 1·3 ± 0·7% CD14⁺ cells in these experiments (n = 4). In aggregates from Ad5_MIP-3α-transfected HT-29 cells 1·2 ± 1·1% CD14⁺ cells could be detected after 7 days of culture (n = 7). In all these control experiments with no Ad5_MCP-1 transfection invading MO/MAC showed a mainly non-reactive intestinal-like phenotype after a 7-day incubation period.

Transfection of HT-29 cells with the Ad5_MCP-1 virus resulted in a reduced differentiation of MO/MAC into the intestinal-like phenotype inside these aggregates. In HT-29, MCS transfected with the Ad5_MCP-1 virus 26·2 ± 23·3% of all cells showed expression of CD14 after an incubation period of 7 days (n = 10). In MCS co-transfected with the Ad5_Null and Ad5_MCP-1, 34·6 ± 8·8% of total cells were positive for CD14 (n = 4). Co-transfection of HT-29 cells with Ad5_MCP-1 and Ad5_MIP-3α resulted in 30·8 ± 17·3% of CD14⁺ cells inside the three-dimensional aggregates after 7 days (n = 7).

The percentage of CD14⁺ cells with regard to CD33⁺ cells was significantly decreased in non-transfected control spheroids, with 38·3 ± 21·3% compared to 74·3 ± 23·8% in Ad5_MCP-1-transfected aggregates (P = 0·002, t-test) (Fig. 6). Transfection of HT-29 cells with Ad5_Null resulted in 40·0 ± 18·1% of CD14⁺ cells with respect to CD33⁺ cells compared to 85·5 ± 9·9% CD14⁺ inside aggregates transfected with Ad5_Null/Ad5_MCP-1 (P = 0·001, t-test). Similar results were obtained in experiments with Ad5_MIP-3α- and Ad5_MCP-1/Ad5_MIP-3α-transfected HT-29 cells. Aggregates consisting of Ad5_MIP-3α-transfected HT-29 cells included 43·4 ± 21·6% of CD14⁺ MAC, whereas co-transfection with Ad5_MIP-3α/Ad5_MCP-1 resulted with 82·6 ± 8·4% in a significantly higher rate of CD14⁺ MAC inside the aggregates after 7 days of culture (P = 0·013, t-test) (Fig. 6).

Fig. 6.

Percentage of CD14⁺ cells with respect to CD33⁺ cells after 7 days of co-culture. The amount of CD14⁺ cells among CD33⁺ cells was significantly higher in Ad5_MCP-1-transfected spheroids compared to aggregates from non-transfected control cells (P = 0·002, t-test). The same result was obtained with monocytes (MO) cultured in Ad5_Null (40·0%) and AD 5_Null/Ad5_MCP-1 (85·5%) transfected aggregates (P = 0·001, t-test). Also comparison between MO cultured inside Ad5_macrophage inflammatory protein 3 alpha (MIP-3α) and Ad5_MIP-3α/MCP-1 aggregates revealed a significantly higher percentage of CD14⁺ MO/macrophage (MAC) among CD33⁺ MO/MAC when HT-29 cells were transfected with Ad5_MCP1 (43·4%versus 82·6%, P = 0·013, t-test). □, Alone; ▪, with monocyte chemoattractant protein (MCP)-1.

Blockade of the of MCP-1 effect by a neutralizing antibody (R&D, AB-279-NA) almost completely abrogated the recruitment of monocytes into the MCS (Fig. 7). Whereas there was a high number of CD14⁺ cells after transfection of HT-29 cells with Ad5_MCP-1 (Fig. 7a), addition of a neutralizing antibody almost completely abrogated the presence of CD14⁺ cells (Fig. 7b). Immunohistochemical analysis proved the presence of CD68/CD14⁺ cells after Ad5_MCP-1 transfection (Fig. 7c), whereas no CD68/CD14 double-positive but also almost no CD68 or CD14 single-positive cells could be observed after neutralization of MCP-1 (Fig. 7d).

Fig. 7.

Neutralization of monocyte chemoattractant protein (MCP)-1 by the addition of anti-MCP-1 antibodies to the multi-cellular spheroid (MCS) model. A MCP-1 neutralizing antibody (R&D systems, cat. number AB-279-NA) was added to the cultures at days 1, 3 and 5 of the culture period. MCS were analysed by FACS analysis (a, b) and by immunohistochemistry (c, d). Whereas there was a high number of CD14 positive cells after transfection of HT-29 cells with Ad5_MCP-1 (a) neutralization of MCP-1 almost completely abrogated the presence of CD14⁺ cells in the MCS (b). For immunohistochemistry CD14 (Pharmingen, Heidelberg, Germany) was stained with APAAP (red) and CD68 (Dako; clone: KP1) was stained with BDHC (blue). The presence of CD68/CD14⁺ cells after Ad5_MCP-1 transfection was obvious (c), whereas no CD68/CD14 double-positive and almost no CD68 or CD14 single-positive cells could be observed after neutralization of MCP-1 (d). The experiment shown was conducted in triplicate and is representative of two further experiments.

Discussion

In this study we determined the influence of the chemokines MCP-1 and MIP-3α on the intestinal-like differentiation of blood MO in a three-dimensional cell culture model. Freshly elutriated blood MO co-cultured with three-dimensional spheroids consisting of IEC differentiate into an intestinal-like phenotype during a 7-day incubation period. Our results indicate that transfection of HT-29 cells with an Ad5_MCP-1 adenovirus dramatically increases the number of MO invading the MCS and disturbs the differentiation of MO cultured inside the spheroids, whereas transfection with an Ad5_MIP-3α adenovirus has no influence on the intestinal-like differentiation of MO.

Flow cytometrical and immunohistochemical analysis showed a strong expression of the MO specific activation-associated surface antigen CD14, which is down-regulated during intestinal-like differentiation, in MO cultured in Ad5_MCP-1-transfected spheroids. MO cultured inside spheroids generated from Ad5_MIP-3α-transfected HT-29 cells showed the same differentiation pattern as MO cultured in non-transfected control spheroids. The migration of MO into the three-dimensional aggregates was also increased by Ad5_MCP-1 but not by Ad5_MIP-3α. Neutralization of MCP-1 was followed by an almost complete abrogation of monocyte migration into the MCS model.

Elevated levels of MCP-1 have been found in atopic reactions such as allergic rhinitis or allergic asthma and in atherosclerotic plaques, where MAC and lymphocytes are the main inflammatory cells [13]. During IBD the proinflammatory chemokine MCP-1 is up-regulated on RNA [26,27] and protein level [28] in intestinal mucosa; however, it is not always accompanied by increases in chemokine serum levels [29]. MCP-1 induces strong migration of MO into the inflamed tissue and is one of the factors responsible for the maintenance of the inflammation [30].

An increase in MIP-3α protein production was observed in HT-29 and Caco-2 cells after stimulation with TNF and IL-1β. MIP-3α protein levels were also elevated in primary intestinal epithelial cells from patients with IBD. The increased production of MIP-3α in epithelial cells may also play an important role in lymphocyte activation and recruitment to the colonic epithelium in IBD [31,32]. In a recent study we have shown a differentiation-dependent induction of Mip-3α expression in intestinal macrophages and a co-localization of Mip-3α/CD68⁺ (macrophages) and CD45R0⁺ cells (memory T cells) in the lamina propria [32]. Mip-3α overexpression was associated with an increased recruitment of memory T cells in the spheroid model, whereas neutralization of Mip-3α abolished this effect [32]. These data indicate clearly that under identical conditions MIP-3α is chemotactic for T cells but not for monocytes/macrophages.

Our results indicate that MCP-1 might play a very important role for the recruitment of MO into the intestinal mucosa, as overexpression was followed by a very dramatic increase in the recruitment of MO into the MCS and neutralization of MCP-1 almost completely prevented invasion of MCS by MO. In addition, the differentiation of MO into anergic IMAC taking place in the healthy gut is disturbed by the presence of the proinflammatory chemokine MCP-1 but not by MIP-3α.

The presence of MCP-1 in the inflamed tissue could directly inhibit the intestinal-like differentiation of MO. Tabata et al. showed that MCP-1 is involved in the inflammatory process of atherosclerosis through the induction of scavenger receptor expression via the ERK pathway and differentiation of MO into foam MAC [33].

On the other hand, the presence of increased numbers of MO inside the MCP-1-transfected aggregates could influence the differentiation process, hallmarking an indirect MCP-1 effect. Inhibition of MO/MAC differentiation is not due to transfection of HT-29 cells with adenoviral constructs, as transfection with an empty control virus or the Ad5_MIP-3α virus did not disturb the differentiation process.

One explanation of the observed effects could be that in the healthy gut only a subpopulation of blood MO is able to differentiate into IMAC and that only this subpopulation of cells migrates into the intestinal mucosa, where the differentiation process takes place. In an inflamed environment with strong expression of proinflammatory and chemotactic mediators such as MCP-1 a large number of unspecific blood MO is attracted to the inflamed tissue, and the subpopulation of MO able to differentiate into intestinal MAC is too small to be detected.

There was a significant production of MCP-1 in non-transfected spheroids after 7 days that did not seem to influence IMAC differentiation. A dose effect might be one explanation for this observation. There is a low physiological production of MCP-1 in the intestinal epithelium that may be necessary for the permanent recruitment of monocytes to the mucosa. It might be speculated that this concentration is just sufficient to recruit a primed subpopulation of monocytes, which is resembled in the spheroid setting. On the other hand, it should be kept in mind that the MCP-1 production in non-transfected spheroids normally is not observed at days 1 and 3, when the monocytes already have entered the spheroids. Therefore, during the early phase of IMAC differentiation MCP-1 is not present if no transfection has been performed.

Our results indicate that the differentiation of MO into anergic IMAC taking place in the healthy gut is disturbed by the presence of the proinflammatory chemokine MCP-1 but not by MIP-3α, both elevated in the inflamed tissue from patients with IBD.