• Dendritic cells;
  • Epithelial cells;
  • Intestinal immunity;
  • Tolerance


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods

A network of dendritic cells (DC) can be detected in close proximity to the epithelial cells overlying Peyer's patches in the gut. Intestinal DC show distinct phenotypes as compared to DC from the systemic lymph nodes (relatively low MHC and costimulatory molecules and high IL-10 and TGFβ) and may play a role in maintaining tolerance to enteric antigens. We show that a similar phenotype is induced in the presence of a polarised epithelial cell monolayer in vitro. Monocyte-derived DC were co-cultured with Caco-2 intestinal epithelial monolayers for 24 h. Co-culture resulted in DC with reduced expression of MHC class II, CD86, and CD80, and poor T cell stimulatory capacity. Cytokine profiles showed reduced levels of inflammatory cytokine production, and co-cultured DC were less sensitive to stimulation via Toll-like receptors (TLR2, 4, and 6) as a result of increased levels of autocrine TGFβ production. However, phenotypic changes in co-cultured DC could not be blocked by removal of apoptotic cells or addition of anti-TGFβ antibodies, suggesting that other soluble factors are involved in DC modulation. Thus, polarised epithelial cell monolayers create a ‘tolerogenic’ environment which modulates the activity of DC. These results highlight the regulatory importance of the epithelial microenvironment at mucosal surfaces.


7-aminoactinomycin D






macrophage-activating lipopeptide-2




mean fluorescence intensity


trans-epithelial resistance


thymic stromal lymphopoietin


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods

Constant exposure to antigens derived from food and resident microorganisms has meant that the gastrointestinal (GI) immune system has had to develop a complicated set of mechanisms to ensure that any immune responses are tightly controlled. Increasingly, evidence is pointing towards dendritic cells (DC) as key to this regulation. DC are generally found at potential sites of pathogen entry and continually sample the local microenvironment for antigens 1. Upon encountering invasive microorganisms or inflammatory cytokines, these cells undergo ‘maturation’ and migrate to the local draining lymph nodes to efficiently present captured antigens to naive T lymphocytes 2. DC can also traffic to the lymph nodes under non-inflammatory conditions, and presentation of antigens in this context normally results in the induction of tolerance and the generation of regulatory T cells 35. Indeed, enhancement of oral tolerance following expansion of DC subsets, in vivo, suggests that DC are important mediators of homeostatic immune responses and tolerance induction 6, 7. The crucial factor governing whether they do this or mount an inflammatory immune response could well be the status of the tissue microenvironment. In support of this, GI tract DC appear to have unique properties distinguishing them from DC in other tissues, and an emerging consensus suggests that the tissue microenvironment plays an important role in shaping DC function 8, 9.

At least four different subsets of DC (based on CD11b, CD11c and CD8α expression) have been identified in the GI tract of mice, and distinct DC subsets are also present in humans 1014. They are often found in close association with epithelial cells and tend to be concentrated at sites of antigen entry. Large numbers of myeloid, MHC class II+ cells can be detected in the subepithelial dome of Peyer's patches, immediately adjacent to the follicle-associated epithelium 11. Their location is thought to allow efficient sampling of incoming antigen as it passes across the epithelium via M cells 9, 15. In addition, numerous DC populate the lamina propria and recent studies suggest that these cells gain access to the intestinal lumen by extruding dendritic processes between epithelial cells 14, 16, 17. Further evidence for the close interactions between DC and epithelial cells comes from observations made in rat models in which DC leaving the mucosa (via the lymphatics) under steady-state conditions were found to contain apoptotic fragments derived from apoptotic epithelial cells 18.

Importantly, unlike DC from other tissues, murine DC from Peyer's patches and the lamina propria express elevated levels of the immunomodulatory cytokines IL-10 and TGFβ, and skew T lymphocytes towards a T helper type 2 response (Th2), which is vital for the generation of IgA 10, 12, 19. In addition, GI DC have the ability to selectively switch on the expression of ‘gut-homing’ chemokine receptors (CCR9) and adhesion molecules (α4β7integrin) in the naive T lymphocytes they stimulate – possibly via a mechanism involving retinoic acid 2023. These adaptations to the gut environment thus enable DC to maintain tissue homeostasis without generating a damaging immune response to lumenal antigens (commensals and food) under steady-state conditions.

In view of these studies, the maturational state of the resident DC population within the GI tract is currently regarded as a crucial factor governing the type of immunological response to antigen. We aimed to investigate which factors might govern the conditioning of GI tract DC. We hypothesised that the key factor controlling the maturational state of the GI tract DC population might be the local cytokine microenvironment generated by the epithelial cells themselves. For this reason, we have attempted to replicate the in vivo dynamics of DC-epithelial cell interactions in an in vitro model. Since myeloid DC are continually recruited to mucosal surfaces, we have investigated the effects of co-culturing monocyte-derived, immature DC in the presence of an established epithelial cell monolayer.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods

Generating stable Caco-2 epithelial cell monolayers

We aimed to develop stable intestinal epithelial cell monolayers which would best reflect those found in a healthy individual. To this end, we used the well-characterised Caco-2 colonic carcinoma cell line, cultured on the surface of 3-µM pore Costar transwell inserts for 14 days and monitored every other day for trans-epithelial resistance (TER) using a Millicell-ERS voltmeter (Fig. 1a). Monolayer integrity reached a peak (TER ∼1500 Ohms/cm2) at day 11, and Caco-2 cells at this stage were well differentiated with microvilli on their apical surfaces (not shown). Monolayers were thus used for co-culture experiments between day 10 and day 12, while those failing to reach a TER >1000 Ohms/cm2 by this time were excluded from further experiments. By modifying a previously described model 24, 25, we investigated differences mediated by close cellular interactions (cell contact) or by soluble factors (cell-separated) (Fig. 1b).

thumbnail image

Figure 1. Intestinal epithelial monolayer development. (A) Caco-2 intestinal epithelial cells were grown as monolayers on the surface of 3-µM pore Costar transwell inserts for 14 days to form stable, polarised monolayers. Monolayer integrity was measured by TER (Ohms/cm2). Monolayer integrity reached a peak between days 10 and 13. Mean of 12 wells ± SD. (B) Diagram of co-culture system setup. Monolayers were either cultured in the upper chamber or on the underside of the transwell insert, and DC were added to the upper chamber.

Download figure to PowerPoint

Co-culturing intestinal epithelial cells and DC does not adversely affect either population

Adding large numbers of DC or co-culturing for longer than 24 h resulted in nutrient depletion and subsequent disruption of the epithelial cell monolayer. Time course and titration experiments indicated that 2.5 × 105 DC for up to 24 h was the optimal number of cells that could be co-cultured without subsequent monolayer disruption (optimisation not shown). After co-culture, monolayers were analysed for changes in TER (Fig. 2a) and DC were analysed for viability by annexin V and 7-aminoactinomycin D (7-AAD) staining (Fig. 2b). Epithelial monolayers showed no changes in TER following co-culture, while DC were routinely found to be >90% viable after co-culture.

thumbnail image

Figure 2. Effect of co-culture monolayer integrity and DC viability. (A) Mean TER of monolayers following DC co-culture ± SD (n = 10). The presence of 2.5 × 105 immature DC for 24 h did not disrupt epithelial integrity. (B) Annexin V-APC and 7-AAD staining of co-cultured DC. Co-culture conditions did not adversely affect DC viability. Data is representative of five experiments.

Download figure to PowerPoint

Co-cultured DC down-regulate expression of MHC class II and costimulatory molecules

Day 5 monocyte-derived, immature DC were co-cultured with stable Caco-2 monolayers or supernatants for 24 h. Immature DC cultured for 24 h in the absence of epithelial cells were CD14neg, MHC class IIhigh, CD80low, CD86low, CD40high and CD11chigh (Fig. 3a). DC incubated with epithelial cells expressed reduced levels of MHC class II and CD86. The effect was most potent when DC were co-cultured with the epithelial cell monolayers in the cell contact system. Expression of CD40 and CD11c also appeared to be slightly reduced in some experiments. Because we found considerable heterogeneity in the phenotypes of monocyte-derived DC from different individuals, we performed the same experiment using DC derived from the PBMC of ten healthy volunteers (Fig. 3b). Analysis of this data showed that DC in the contact co-culture system exhibited the most profound alterations in phenotype, with MHC class II being significantly reduced by 60.9±23.1% and CD86 by 30.9±8.2%. Minor changes were observed in levels of CD40 and CD11c, but these were not statistically significant. At no point did we observe up-regulation of MHC class II or costimulatory molecules indicative of DC maturation. Additional experiments using unpolarised Caco-2 cells and supernatants failed to induce significant amounts of MHC class II down-regulation (data not shown), suggesting that polarised epithelial cell monolayers delivered the most potent modulatory signal.

thumbnail image

Figure 3. Effect of intestinal epithelial cell co-culture on surface phenotype of co-cultured DC. (A) Flow cytometric analysis of DC cultured in the presence or absence of intestinal epithelial cell monolayers for 24 h showing phentoypic alterations upon co-culture. Filled histograms are isotype controls, unfilled histograms show staining for the named antigens. Numbers indicate mean fluorescence intensities (MFI) of a representative experiment (representative of ten experiments). (B) MFI for DC derived from ten healthy volunteers expressed as percentage of the MFI of DC cultured alone ± SD. Accumulated results indicate that co-cultured DC express reduced levels of MHC class II (p <0.005) and the costimulatory molecules CD80 and CD86 (p <0.05 in both systems). CD40 and CD11c expression was relatively unaffected. ** p <0.005, * p <0.05 by Wilcoxon matched pairs signed rank test.

Download figure to PowerPoint

Co-cultured DC modulate their cytokine profile

We next investigated the production of two inflammatory and two anti-inflammatory cytokines from co-cultured DC (Fig. 4a). IL-8, IL-10, and TGFβ proteins were found to be secreted by both the epithelial cells (very low) and DC (∼10-fold higher levels). However, cytokine production by epithelial monolayers remained unchanged following co-culture, indicating DC as the main source of IL-8, IL-10, and TGFβ. TNF-α protein could not be detected in any of the supernatants, although we could detect mRNA for this cytokine in all samples. IL-8 and IL-10 levels were reduced following DC co-culture (cell contact system), but were unaffected by culture with epithelial supernatants (cell-separated system). By contrast, elevated TGFβ protein was found in supernatants from co-cultured DC (cell contact system). Thus, co-cultured DC exhibited a modulated phenotype (reduced MHC and costimulatory molecule expression) and possessed a cytokine profile dominated by TGFβ.

thumbnail image

Figure 4. Effect of intestinal epithelial cell co-culture on DC cytokine expression. Detection of cytokines in supernatants of co-cultured DC by capture ELISA. (A) IL-8 (p <0.05) and IL-10 were reduced in co-cultured DC, while TGFβ was up-regulated (p <0.05). The most pronounced effects were induced in the cell contact system. Samples were analysed in triplicate and data are means ± SD (n = 3). * p <0.05 (Wilcoxon). (B) Detection of IL-12p70 and IL-10 in co-cultured DC stimulated for 24 h in the presence of CD40L-expressing mouse L cells. Co-cultured DC (cell contact system) show a trend towards reduced IL-12p70 and increased IL-10 production. Unfilled bars represent responses to wild-type mouse L cells not expressing CD40L. Samples were analysed in triplicate and data are means ± SD. Representative of three experiments.

Download figure to PowerPoint

Co-cultured DC release elevated IL-10 and reduced IL-12p70 upon CD40L stimulation

Co-cultured DC were next analysed for their ability to release cytokines in response to CD40L stimulation (Fig. 4b). IL-12 is released by activated DC during stimulation of naive T cells in the draining lymph nodes and generally promotes a Th1 cytokine profile. Conversely, IL-10 induces a Th2 cytokine profile and inhibits IL-12 production. A tightly controlled balance between these two cytokines thus influences the type of T cell response stimulated by DC leaving the tissues. All unstimulated DC released barely detectable amounts of IL-12 and IL-10. However, upon CD40L stimulation, co-cultured DC (cell contact system) produced reduced levels of IL-12p70 but elevated levels of IL-10, as compared to DC cultured alone or with epithelial monolayer supernatants only (cell-separated system).

Co-cultured DC show specific impairment in their Toll-like receptor responses

Co-cultured DC were also analysed for their ability to respond to Toll-like receptor (TLR) agonists [Pam3CSK4 – TLR2; LPS – TLR4; macrophage-activating lipopeptide-2 (MALP-2) – TLR2/6], the NOD2 ligand muramyl-dipeptide (MDP-Nopia) or CD40L stimulation. Stimulation for 24 h with all of the above induced readily detectable TNF-α responses from DC cultured alone, with maximal response seen with CD40L stimulation (Fig. 5a). Similar responses to CD40L and MDP stimulation were elicited from co-cultured DC (cell contact system), but we observed a profound impairment in the response to all TLR agonists (also seen with IL-8; data not shown). These results suggested that the epithelial monolayer microenvironment might modulate the sensitivity of DC to TLR agonists such as LPS without affecting other stimulatory signalling pathways such as those utilised by CD40L and MDP. Since co-cultured DC produced elevated levels of TGFβ, we investigated the effects of this cytokine on DC responses to LPS (TLR4) by co-culturing DC (cell contact system) in the presence of anti-TGFβ blocking antibodies (Fig. 5b). As before, co-cultured DC showed reduced cytokine responses to LPS. However, those DC co-cultured in the presence of anti-TGFβ antibodies displayed minimal impairment in their cytokine responses to LPS (as compared to DC co-cultured in the presence of an isotype-matched control antibody). These results suggest that autocrine production of TGFβ by co-cultured DC is important in the regulation of TLR-mediated pathogen responses.

thumbnail image

Figure 5. Effect of intestinal epithelial cell co-culture on DC responses to stimulation. (A) Co-cultured DC (cell contact system) and DC cultured alone were incubated for a further 24 h in the presence of 10 ng/mL ultrapure LPS, 10 ng/mL Pam3CYSK4, 10 ng/mL MALP-2, 100 ng/mL MDP-Nopia, or 2 × 105 CD40L-expressing mouse L cells. Supernatants were analysed for TNF-α by capture ELISA. Co-cultured DC showed impaired responses to the TLR agonists, but not to MDP or CD40L stimulation. Data are representative of three experiments. (B) DC were cultured alone or co-cultured (cell contact system) in the presence of either 10 μg/mL IgG1 control antibody (MOPC21) or 10 μg/mL anti-TGFβ antibody. Cells were harvested and incubated for a further 24 h in the presence of 10 ng/mL ultrapure LPS. Supernatants were analysed for TNF-α by capture ELISA. DC co-cultured in the presence of anti-TGFβ antibody recovered their TLR sensitivity. Data is representative of three experiments.

Download figure to PowerPoint

Co-cultured DC show reduced stimulatory capacity in mixed lymphocyte reactions

The functional relevance of the observed modulatory effects resulting from epithelial cell co-culture was investigated using an allogeneic mixed lymphocyte reaction (MLR). DC cultured alone or in co-culture were recovered and used to stimulate a population of allogeneic CD4+ T cells for 5 days. Proliferation was then measured by [3H]thymidine incorporation. Co-cultured DC were as good as, if not better than, DC cultured alone (Fig. 6a). We hypothesised that this was because co-cultured DC could recover their expression of MHC, costimulatory molecules, and cytokines, once removed from the epithelial cell microenvironment – perhaps upon encountering T cells. To test this, DC cultured alone or in co-culture were metabolically fixed before being used in the allogeneic MLR. When treated in this way, co-cultured DC were found to be profoundly impaired in their ability to stimulate the proliferation of CD4+ T cells, as compared to DC cultured alone (Fig. 6b). This data indicated that the modulatory effect of intestinal epithelial cells on DC was likely to be reversible.

thumbnail image

Figure 6. Effect of intestinal epithelial cell co-culture on T cell stimulatory capacity of DC. (A) Co-cultured DC were used as stimulators in a 5-day MLR using allogeneic CD4+ T cells as responders. Co-cultured DC were slightly more stimulatory than DC cultured alone, irrespective of the co-culture system. (B) Effect of metabolic fixation (to prevent phenotypic changes) on stimulatory capacity. Co-cultured DC were severely impaired in their ability to stimulate CD4+ T cells as compared to DC cultured alone. Black bars, cells cultured alone; grey bars, co-cultured cells (contact); unfilled bars, co-cultured cells (separated). All samples were analysed in triplicate. Data are means ± SD. Results are representative of three experiments.

Download figure to PowerPoint

Mechanisms behind epithelial monolayer-mediated modulation of DC activity

We next sought to identify the mechanism by which intestinal epithelial cells mediate their modulatory function. A number of groups have shown that the phagocytic uptake of apoptotic cells and fragments by myeloid cells results in their differentiation towards TGFβ- and IL-10-secreting, ‘regulatory’ antigen-presenting cells 2630. We hypothesised that capture of apoptotic Caco-2 cells in the co-culture systems might have the same effects on our DC. To test this, we grew Caco-2 cells in a transwell system on inserts with a 0.4-µM pore filter (to block transfer of large apoptotic bodies) and analysed the phenotype of co-cultured DC (cell contact system only) compared to DC cultured on a 3-µM pore filter or DC cultured alone. MHC class II and CD86 were down-regulated by approximately 60% and 35%, respectively, when using both the 3-µM pore and 0.4-µM pore filters, indicating that transfer of large particles was not essential for epithelial cell modulation of DC (data not shown).

We next investigated the possibility that epithelium-derived TGFβ might be the modulatory signal, since TGFβ is known to have wide-ranging immunomodulatory properties and is an important factor in the generation of ‘tolerising’ DC and regulatory T cells 3133. Co-culture experiments were thus carried out in the presence of anti-TGFβ blocking antibodies and an isotype-matched IgG1 control. In all cases, the presence of these factors was not enough to prevent down-regulation of MHC class II or CD86 (not shown). To confirm these results, DC from all test conditions were fixed and used as stimulators in an allogeneic MLR (CD4+ T cells as responders). As before, DC co-cultured with epithelial cells on a 3-µM pore filter were impaired in their ability to stimulate allogeneic T cells (Fig. 7). This was also true for DC co-cultured with epithelial cells on a 0.4-μM pore filter as well as for co-cultures performed in the presence of anti-TGFβ blocking antibodies. Taken together, this data indicates that an intact intestinal epithelial cell monolayer can modulate the phenotype and function of immature DC by a mechanism that does not involve uptake of apoptotic cells or phenotypic modulation by TGFβ. However, co-culture induces the production of TGFβ by DC and this appears to act in an autocrine manner to regulate TLR responses.

thumbnail image

Figure 7. Effect of blocking apoptotic cell transfer or TGFβ on capacity of co-cultured DC to stimulate allogeneic T cells. Stimlatory capacity of metabolically fixed, co-cultured DC (cell contact system), using 3-μM pore or 0.4-μM pore filter inserts, or in the presence of anti-TGFβ antibodies. Allogeneic CD4+ T cells were used as responders in a 5-day MLR. Black bars represent results for cells cultured alone, unfilled bars are for co-cultured cells (cell contact system). All samples were analysed in triplicate. Data are means ± SD. Results are representative of three experiments.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods

Many investigators have noted the unique properties of GI tract DC and have speculated upon their development and function 34. Here, we present data suggesting that the tissue microenvironment, provided by a healthy polarised epithelium, has important immunomodulatory effects on immature myeloid DC. This appears to be the case even when commensal bacteria – which are also thought to deliver modulatory signals to GI lymphoid tissue – are absent 3538.

Upon co-culture with intact epithelial cell monolayers, DC down-regulated MHC class II and costimulatory molecules (CD86 and CD80). The changes tended to be more profound in the cell contact system but were also present in the supernatant-only system, suggesting the presence of a soluble factor. It is important to note that DC cultured in the upper chamber were in a total volume of 200 µL while those in the lower chamber were in 1 mL. This difference results in a fivefold dilution factor and may account for some of the phenotypic differences between the two systems. In both cases, the mediating factor(s) are likely to be soluble molecules released by epithelial cells. Co-culture for 24 h did not induce any maturational changes or inflammatory cytokine production in any of our DC samples, but did result in decreased IL-8 and IL-10 production and increased TGFβ levels in the supernatants. The observed phenotype of co-cultured DC correlates well with GI tract DC phenotypes observed in situ. DC expressing low levels of surface MHC class II and costimulatory molecules can readily be detected in the lamina propria and the subepithelial dome of Peyer's patches 14, 39, 40. Also, the production of increased amounts of TGFβ is significant, since this factor is thought to be a crucial regulatory cytokine in the GI tract and TGFβ knockout mice develop intestinal inflammation 41. TGFβ also inhibits the proliferation and development of effector function in naive T cells and may promote the development of regulatory T cells 4244. Thus, production by homeostatic GI tract DC could potentially contribute to the maintenance of the anti-inflammatory microenvironment.

Co-cultured DC also expressed very little IL-12p70 or IL-10, a phenotype matching that of unstimulated DC extracted from healthy intestinal tissues 39, 45. However, when the co-cultured DC were stimulated for a further 24 h in the presence of CD40L-expressing fibroblasts, we observed a clear increase in IL-10 production, which was mirrored by a decrease in IL-12p70 production as compared to DC cultured alone. This also corresponds with observations made ex vivo, in which intestinal DC stimulated with commensal bacteria up-regulated IL-10 but not IL-12 39. Myeloid DC from murine Peyer's patches have also been shown to express higher levels of IL-10 in response to stimulation, as compared to DC from the spleen 10, 12, 39. These DC also stimulated naive T cells to produce high levels of the Th2 cytokines IL-4 and IL-10 (as opposed to high IFN-γ for Th1), suggesting that DC derived from the gut microenvironment have different properties to those found elsewhere (i.e. the spleen).

In addition to phenotypic alterations, co-cultured DC also showed changes in their ability to respond to bacterial ligands via TLR. Co-cultured DC showed impaired sensitivity to TLR2, 4, and 6 agonists, but preserved sensitivity to MDP (via NOD2) and CD40L stimulation. We were able to show that TLR sensitivity in co-cultured DC is partly regulated by TGFβ. This data adds to observations made in TGFβ-deficient mice in which uncontrolled inflammation is associated with increased levels of mRNA for TLR4 and endotoxin hypersensitivity 46. While TLR2 and TLR4 have been shown to be significantly enhanced in those with Crohn's disease, expression by intestinal DC is low in healthy individuals 47. It is tempting to speculate that, in the absence of inflammation, soluble factors derived from a healthy intestinal epithelium promote TGFβ production by co-cultured DC, which then acts in an autocrine manner to down-regulate TLR expression or signalling and DC sensitivity to bacterial ligands. Modulation of DC responses to bacterial ligands is emerging as an important factor in inducing tolerance and homeostasis in the epithelial microenvironment. Indeed, mouse models indicate that functional TLR pathways are actually essential for intestinal epithelium integrity 48.

The functional outcome of intestinal epithelial co-culture was also confirmed in our system when DC were fixed (preventing further phenotypic changes) immediately after co-culture and used as stimulators in a 5-day MLR. If the DC were not fixed prior to the MLR, they were found to be as stimulatory, if not more so, than DC cultured alone. This result indicates that co-cultured immature DC can recover their stimulatory capacity once they leave the epithelial cell microenvironment – perhaps upon encountering maturational signals from T cells. This data corresponds with observations made in mice in which CD11c+ cells isolated from murine Peyer's patches were found to be poor stimulators in a primary allogeneic MLR, but recovered their stimulatory capacity following short-term culture 39, 40. This response also makes sense in evolutionary terms since, in addition to their homeostatic role, DC need to be able to respond promptly to truly pathogenic microorganisms when the need arises. Clearly it will be important to classify the type (Th1/Th2) of T cell response induced by our co-cultured DC.

Finally, we investigated the possible mechanism by which the epithelium might alter DC phenotype and function. A number of studies have shown that mucosal DC can acquire apoptotic epithelial cells and that this can modulate cell cytokine production, producing IL-10 and TGFβ rather than IL-12 2730. We hypothesised that this might be occurring in our co-culture systems and used monolayers cultured on 0.4-µM pore filters to block possible uptake of large apoptotic bodies by the DC. This measure failed to prevent the phenotypic and functional changes observed after co-culture. Many of the phenotype changes we observed can also be induced by TGFβ and this cytokine is released in small amounts by Caco-2 epithelial cells 33. However, blocking TGFβ was not sufficient to prevent the observed phenotypic changes and only served to restore TLR responses of co-cultured DC. Further work is required to understand the nature of the modulatory signals produced by a healthy epithelial monolayer; however, recent reports indicate thymic stromal lymphopoietin (TSLP) could be a likely candidate – although there are likely to be others 49. Significantly, TSLP is produced by intestinal epithelial cells in situ and is down-regulated in patients with Crohn's disease 49. TSLP is also produced by keratinocytes 50 and airway epithelium 51 and appears to contribute to the development of Th2 immune responses by modulating the phenotype of DC. In this regard, it will be interesting to investigate the immunomodulatory properties of epithelial cells from other mucosal surfaces, since this type of modulation may result from factors released by all healthy epithelial cells. Indeed, preliminary experiments in our lab indicate that other colonic epithelial cell lines (such as HT29) possess DC immunomodulatory effects, particularly when grown as stable monolayers (M. Butler, unpublished observations).

In conclusion, we have described an in vitro model of epithelial-dendritic cell interactions at the mucosal surface. Our data suggests that polarised intestinal epithelial monolayers might contribute to intestinal homeostasis by providing an anti-inflammatory microenvironment for the development and maintenance of myeloid DC. Part of this conditioning involves TGFβ-mediated modulation of TLR responses to bacterial ligands. Commensal bacteria can also mediate a tolerising effect 49, 52, and it will be important to understand how the presence of pathogenic or ‘friendly’ bacteria in our model might affect this process. We believe that the identification of epithelial cell-derived tolerising factors will greatly extend our insight into immune regulation at mucosal surfaces.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods

Caco-2 monolayers

Caco-2 human colonic carcinoma cells were maintained in DMEM (Gibco Life Technologies, UK) supplemented with 10% foetal calf serum (FCS; Sera Laboratory International Ltd., UK) and incubated at 37°C in a humidified atmosphere of 5% CO2. Monolayers were grown in 24-well Corning Costar Transwell plates (Corning Inc., USA) with either 3-µM pore or 0.4-µM pore size filter inserts. In the cell-separated system (Fig. 1b), 100 µL of cells (3–4 × 105 cells) was added to the upper chamber and the lower chamber was filled with 1 mL medium. In the cell contact system, inserts were removed and inverted in tissue culture dishes. Cells (100 µL) were then added to the exposed filter membrane and incubated overnight. Inserts were returned to their original 24-well plates, containing 1 mL medium in the lower chamber, the following day. Monolayers were maintained for 10–12 days (changing medium every other day) or until they consistently gave TER readings of >1000 Ohms/cm2 (measured using a Millicell-ERS voltmeter; Millipore Corp., USA) before being used in co-culture experiments. Monolayers failing to reach a TER of >1000 Ohms/cm2 by this time were excluded from further experiments.

Monocyte purification and DC differentiation

Fresh peripheral blood (from healthy volunteers) was collected in 25-mL aliquots and then made up to 50 mL with sterile PBS containing 10 U heparin. PBMC were separated by centrifugation on a Ficoll-Hypaque (Nycomed, UK) density gradient (according to the manufacturer's instructions). The upper layer was collected and then washed with cold sterile PBS. CD14+ monocytes were isolated by positive selection using MACS beads (Miltenyi Biotech, Germany) as per the manufacturer's instructions. Cells were transferred to 6-well tissue culture plates at a concentration of 3–5 × 105 cells/mL in 3 mL. To induce differentiation, cells were given 25 ng/mL IL-4 and 50 ng/mL GM-CSF (First Link Ltd., UK) on day 0 and day 4. Cells were used in co-culture experiments between days 6 and 9.

Co-culture experiments

At 24 h before the start of the experiment, the medium on the Caco-2 monolayers was replaced with X-VIVO 15 serum-free medium (Biowhittaker, USA). On the day of the experiment, DC were harvested, washed and resuspended at 2.5 × 106 cells/mL of X-VIVO 15. Of cells, 100 µL (2.5 × 105 cells) was added either to the lower chamber (cell-separated system) or to the upper chamber (cell contact system) and the volume was made up to 200 μL (top chamber) or 1 mL (bottom chamber) with fresh medium. As a control, DC were also plated out at the same concentration and volume in tissue culture plates. For cytokine blocking experiments, 10 μg/mL IgG1 control (MOPC21) or anti-TGFβ antibodies (R&D Systems, USA) were introduced at the same time as DC. All cultures were incubated for a maximum of 24 h.

Cell viability and flow cytometry

DC viabilty was monitored by staining with annexin V-allophycocyanin (APC) (Becton Dickinson, USA) and 7-AAD (Sigma) as per the manufacturer's instructions immediately following co-culture. For flow cytometry, DC were harvested and washed three times in cold FACS buffer (PBS, 3% FCS, 0.02% NaN3) and then stained with the appropriate antibodies – anti-HLA-DR, -DP, -DQ (DakoCytomation, UK), anti-CD80 (Pharmingen, Becton Dickinson), anti-CD86 (Becton Dickinson), anti-CD40 (Diaclone, UK), anti-CD11c (DakoCytomation) – at a final concentration of 10 µg/mL for 30 min on ice and in the dark. Isotype control antibodies were mouse IgG1 (MOPC21; Sigma-Aldrich) and mouse IgG2a (Sigma-Aldrich). Cells were analysed immediately by flow cytometry in a FACSCalibur (Becton Dickinson).

Cytokine ELISA

Supernatants from the basolateral chambers (bottom chamber in the cell-separated system and upper chamber in the cell contact system), or DC cultured alone, were collected after co-culture and were spun to remove cell debris. For direct comparison, supernatants from the upper chamber were made up to 1 mL with fresh medium. Thus, in all cases results are for 2.5 × 105 cells/mL. For IL-8 and TNF-α ELISA, maxisorp 96-well flat-bottom plates (Nunc, UK) were coated with 1 µg/mL anti-IL-8 or anti-TNF-α capture antibodies (BD Pharmingen) overnight at 4°C. Plates were washed with PBS/0.1% Tween-20 and blocked for 2 h with 100 µL PBS/1% BSA. Samples were diluted in X-VIVO 15 and added to the plate at 100 µL per well. A standard curve was generated using serial dilutions of recombinant IL-8 or TNF-α (Peprotech, UK) at a top dilution of 4 ng/mL. Captured cytokine was detected using 0.5 µg/mL of the appropriate biotinylated detection antibodies for IL-8 and TNF-α (BD Pharmingen). Finally, avidin-horseradish peroxidase (Biosource Int., USA) was added at 0.1 μg/mL for 30 min before TMB (Zymed, USA) was used as a substrate. IL-10, IL-12p70, and TGFβ1 were detected using duoset kits (R&D Systems, USA) as per the manufacturer's instructions. Colour development was stopped by the addition of 0.5 M H2SO4 and plates were read for optical density at 450 nm (ELISA plate reader; Titertek Multiskan). Samples were analysed in triplicate and expressed as mean concentrations ± standard deviation (SD). Data were analysed for statistical significance using the Wilcoxon matched pairs signed ranks test.

Stimulation of DC with TLR agonists, MDP, and CD40L

After epithelial cell co-culture, DC were plated at 105/well of a 96-well culture plate and stimulated with either 10 ng/mL ultrapure LPS (gel-purified LPS from Escherichia coli 055:B5, previously shown to lack NOD2 agonist activity53), 10 ng/mL Pam3CSK4 (Invivogen, USA), 10 ng/mL MALP-2 (Invivogen), 100 ng/mL MDP (synthetic pharmaceutical grade MDP-Lys18/romurtide; kind gift from Daiichi Pharmaceutical Company to D.v.H.), or 2 × 105 CD40L-expressing fibroblasts in a total volume of 200 µL X-VIVO 15 serum-free medium. After 24 h, supernatants were harvested and analysed for cytokine production.

Mixed lymphocyte reaction

DC were harvested and either fixed (2 h in 0.1% NaN3/PBS at 37°C) or left unfixed. They were then plated out in 96-well round-bottom plates at 10 000, 5000, or 1000 cells per well in 100 µL X-VIVO 15 serum-free medium. Allogeneic CD4+ T cells (isolated from PBMC of healthy volunteers by negative selection) were used as responders and added at 1 × 105 cells/well in 100 µL medium. Plates were incubated for 5 days. At 18 h prior to harvesting, 1 µCi of tritiated methyl thymidine (Amersham Int., UK) was added to each well. Cells were harvested onto glass fibre filter mats (Wallac, UK) and [3H]thymidine incorporation was measured by liquid scintillation spectroscopy in a beta counter (Wallac).

  • 1


  • 2


  • 3


  • 4


  • 5


  • 6


  • 7


  • 1
    Sallusto, F., Cella, M., Danieli, C. and Lanzavecchia, A., Dendritic cells use macropinocytosis and the mannose receptor to concentrate macromolecules in the major histocompatibility complex class II compartment: Downregulation by cytokines and bacterial products. J. Exp. Med. 1995. 182: 389400.
  • 2
    Guermonprez, P., Valladeau, J., Zitvogel, L., Thery, C. and Amigorena, S., Antigen presentation and T cell stimulation by dendritic cells. Annu. Rev. Immunol. 2002. 20: 621667.
  • 3
    Steinman, R. M., Hawiger, D., Liu, K., Bonifaz, L., Bonnyay, D., Mahnke, K., Iyoda, T. et al., Dendritic cell function in vivo during the steady state: A role in peripheral tolerance. Ann. N. Y. Acad. Sci. 2003. 987: 1525.
  • 4
    Hawiger, D., Inaba, K., Dorsett, Y., Guo, M., Mahnke, K., Rivera, M., Ravetch, J. V., Steinman, R. M. and Nussenzweig, M. C., Dendritic cells induce peripheral T cell unresponsiveness under steady state conditions in vivo. J. Exp. Med. 2001. 194: 769779.
  • 5
    Steinman, R. M., Hawiger, D. and Nussenzweig, M. C., Tolerogenic dendritic cells. Annu. Rev. Immunol. 2003. 21: 685711.
  • 6
    Viney, J. L., Mowat, A. M., O'Malley, J. M., Williamson, E. and Fanger, N. A., Expanding dendritic cells in vivo enhances the induction of oral tolerance. J. Immunol. 1998. 160: 58155825.
  • 7
    Williamson, E., Westrich, G. M. and Viney, J. L., Modulating dendritic cells to optimize mucosal immunization protocols. J. Immunol. 1999. 163: 36683675.
  • 8
    Kelsall, B. L. and Rescigno, M., Mucosal dendritic cells in immunity and inflammation. Nat. Immunol. 2004. 5: 10911095.
  • 9
    Kapsenberg, M. L., Dendritic-cell control of pathogen-driven T-cell polarization. Nat. Rev. Immunol. 2003. 3: 984993.
  • 10
    Iwasaki, A. and Kelsall, B. L., Freshly isolated Peyer's patch, but not spleen, dendritic cells produce interleukin 10 and induce the differentiation of T helper type 2 cells. J. Exp. Med. 1999. 190: 229239.
  • 11
    Iwasaki, A. and Kelsall, B. L., Localization of distinct Peyer's patch dendritic cell subsets and their recruitment by chemokines macrophage inflammatory protein (MIP)-3alpha, MIP-3beta, and secondary lymphoid organ chemokine. J. Exp. Med. 2000. 191: 13811394.
  • 12
    Iwasaki, A. and Kelsall, B. L., Unique functions of CD11b+, CD8 alpha+, and double-negative Peyer's patch dendritic cells. J. Immunol. 2001. 166: 48844890.
  • 13
    MacDonald, T. T., The mucosal immune system. Parasite Immunol. 2003. 25: 235246.
  • 14
    Bell, S. J., Rigby, R., English, N., Mann, S. D., Knight, S. C., Kamm, M. A. and Stagg, A. J., Migration and maturation of human colonic dendritic cells. J. Immunol. 2001. 166: 49584967.
  • 15
    Kraehenbuhl, J. P. and Neutra, M. R., Epithelial M cells: Differentiation and function. Annu. Rev. Cell Dev. Biol. 2000. 16: 301332.
  • 16
    Rescigno, M., Identification of a new mechanism for bacterial uptake at mucosal surfaces, which is mediated by dendritic cells. Pathol. Biol. (Paris) 2003. 51: 6970.
  • 17
    Niess, J. H., Brand, S., Gu, X., Landsman, L., Jung, S., McCormick, B. A., Vyas, J. M. et al., CX3CR1-mediated dendritic cell access to the intestinal lumen and bacterial clearance. Science 2005. 307: 254258.
  • 18
    Huang, F. P., Platt, N., Wykes, M., Major, J. R., Powell, T. J., Jenkins, C. D. and MacPherson, G. G., A discrete subpopulation of dendritic cells transports apoptotic intestinal epithelial cells to T cell areas of mesenteric lymph nodes. J. Exp. Med. 2000. 191: 435444.
  • 19
    Sato, A., Hashiguchi, M., Toda, E., Iwasaki, A., Hachimura, S. and Kaminogawa, S., CD11b+ Peyer's patch dendritic cells secrete IL-6 and induce IgA secretion from naive B cells. J. Immunol. 2003. 171: 36843690.
  • 20
    Mora, J. R., Bono, M. R., Manjunath, N., Weninger, W., Cavanagh, L. L., Rosemblatt, M. and Von Andrian, U. H., Selective imprinting of gut-homing T cells by Peyer's patch dendritic cells. Nature 2003. 424: 8893.
  • 21
    Mora, J. R. and von Andrian, U. H., Retinoic acid: An educational “vitamin elixir” for gut-seeking T cells. Immunity 2004. 21: 458460.
  • 22
    Johansson-Lindbom, B. and Agace, W. W., Vitamin A helps gut T cells find their way in the dark. Nat. Med. 2004. 10: 13001301.
  • 23
    Iwata, M., Hirakiyama, A., Eshima, Y., Kagechika, H., Kato, C. and Song, S. Y., Retinoic acid imprints gut-homing specificity on T cells. Immunity 2004. 21: 527538.
  • 24
    Gullberg, E., Leonard, M., Karlsson, J., Hopkins, A. M., Brayden, D., Baird, A. W. and Artursson, P., Expression of specific markers and particle transport in a new human intestinal M-cell model. Biochem. Biophys. Res. Commun. 2000. 279: 808813.
  • 25
    Kerneis, S., Bogdanova, A., Kraehenbuhl, J. P. and Pringault, E., Conversion by Peyer's patch lymphocytes of human enterocytes into M cells that transport bacteria. Science 1997. 277: 949952.
  • 26
    Sauter, B., Albert, M. L., Francisco, L., Larsson, M., Somersan, S. and Bhardwaj, N., Consequences of cell death: Exposure to necrotic tumor cells, but not primary tissue cells or apoptotic cells, induces the maturation of immunostimulatory dendritic cells. J. Exp. Med. 2000. 191: 423434.
  • 27
    Fadok, V. A., Bratton, D. L. and Henson, P. M., Phagocyte receptors for apoptotic cells: Recognition, uptake, and consequences. J. Clin. Invest. 2001. 108: 957962.
  • 28
    Fadok, V. A., McDonald, P. P., Bratton, D. L. and Henson, P. M., Regulation of macrophage cytokine production by phagocytosis of apoptotic and post-apoptotic cells. Biochem. Soc. Trans. 1998. 26: 653656.
  • 29
    Fadok, V. A., Bratton, D. L., Konowal, A., Freed, P. W., Westcott, J. Y. and Henson, P. M., Macrophages that have ingested apoptotic cells in vitro inhibit proinflammatory cytokine production through autocrine/paracrine mechanisms involving TGF-beta, PGE2, and PAF. J. Clin. Invest. 1998. 101: 890898.
  • 30
    Albert, M. L., Death-defying immunity: Do apoptotic cells influence antigen processing and presentation? Nat. Rev. Immunol. 2004. 4: 223231.
  • 31
    Weiner, H. L., Induction and mechanism of action of transforming growth factor-beta-secreting Th3 regulatory cells. Immunol. Rev. 2001. 182: 207214.
  • 32
    Weiner, H. L., Oral tolerance: Immune mechanisms and the generation of Th3-type TGF-beta-secreting regulatory cells. Microbes Infect. 2001. 3: 947954.
  • 33
    Strobl, H. and Knapp, W., TGF-beta1 regulation of dendritic cells. Microbes Infect. 1999. 1: 12831290.
  • 34
    Bilsborough, J. and Viney, J. L., Gastrointestinal dendritic cells play a role in immunity, tolerance, and disease. Gastroenterology 2004. 127: 300309.
  • 35
    Jiang, H. Q., Thurnheer, M. C., Zuercher, A. W., Boiko, N. V., Bos, N. A. and Cebra, J. J., Interactions of commensal gut microbes with subsets of B- and T-cells in the murine host. Vaccine 2004. 22: 805811.
  • 36
    Cebra, J. J., Influences of microbiota on intestinal immune system development. Am. J. Clin. Nutr. 1999. 69: 1046S–1051S.
  • 37
    Talham, G. L., Jiang, H. Q., Bos, N. A. and Cebra, J. J., Segmented filamentous bacteria are potent stimuli of a physiologically normal state of the murine gut mucosal immune system. Infect. Immun. 1999. 67: 19922000.
  • 38
    Cebra, J. J., Periwal, S. B., Lee, G., Lee, F. and Shroff, K. E., Development and maintenance of the gut-associated lymphoid tissue (GALT): The roles of enteric bacteria and viruses. Dev. Immunol. 1998. 6: 1318.
  • 39
    Rigby, R. J., Knight, S. C., Kamm, M. A. and Stagg, A. J., Production of interleukin (IL)-10 and IL-12 by murine colonic dendritic cells in response to microbial stimuli. Clin. Exp. Immunol. 2005. 139: 245256.
  • 40
    Ruedl, C., Rieser, C., Bock, G., Wick, G. and Wolf, H., Phenotypic and functional characterization of CD11c+ dendritic cell population in mouse Peyer's patches. Eur. J. Immunol. 1996. 26: 18011806.
  • 41
    Shull, M. M., Ormsby, I., Kier, A. B., Pawlowski, S., Diebold, R. J., Yin, M., Allen, R. et al., Targeted disruption of the mouse transforming growth factor-beta 1 gene results in multifocal inflammatory disease. Nature 1992. 359: 693699.
  • 42
    Fu, S., Zhang, N., Yopp, A. C., Chen, D., Mao, M., Zhang, H., Ding, Y. and Bromberg, J. S., TGF-beta induces Foxp3+ T-regulatory cells from CD4+ CD25 precursors. Am. J. Transplant. 2004. 4: 16141627.
  • 43
    Chen, W., Jin, W., Hardegen, N., Lei, K. J., Li, L., Marinos, N., McGrady, G. and Wahl, S. M., Conversion of peripheral CD4+CD25 naive T cells to CD4+CD25+ regulatory T cells by TGF-beta induction of transcription factor Foxp3. J. Exp. Med. 2003. 198: 18751886.
  • 44
    Gorelik, L. and Flavell, R. A., Transforming growth factor-beta in T-cell biology. Nat. Rev. Immunol. 2002. 2: 4653.
  • 45
    Hart, A. L., Lammers, K., Brigidi, P., Vitali, B., Rizzello, F., Gionchetti, P., Campieri, M. et al., Modulation of human dendritic cell phenotype and function by probiotic bacteria. Gut 2004. 53: 16021609.
  • 46
    McCartney-Francis, N., Jin, W. and Wahl, S. M., Aberrant Toll receptor expression and endotoxin hypersensitivity in mice lacking a functional TGF-beta 1 signaling pathway. J. Immunol. 2004. 172: 38143821.
  • 47
    Hart, A. L., Al-Hassi, H. O., Rigby, R. J., Bell, S. J., Emmanuel, A. V., Knight, S. C., Kamm, M. A. and Stagg, A. J., Characteristics of intestinal dendritic cells in inflammatory bowel diseases. Gastroenterology 2005. 129: 5065.
  • 48
    Rakoff-Nahoum, S., Paglino, J., Eslami-Varzaneh, F., Edberg, S. and Medzhitov, R., Recognition of commensal microflora by toll-like receptors is required for intestinal homeostasis. Cell 2004. 118: 229241.
  • 49
    Rimoldi, M., Chieppa, M., Salucci, V., Avogadri, F., Sonzogni, A., Sampietro, G. M., Nespoli, A. et al., Intestinal immune homeostasis is regulated by the crosstalk between epithelial cells and dendritic cells. Nat. Immunol. 2005. 6: 507514.
  • 50
    Watanabe, N., Hanabuchi, S., Marloie-Provost, M. A., Antonenko, S., Liu, Y. J. and Soumelis, V., Human TSLP promotes CD40 ligand-induced IL-12 production by myeloid dendritic cells but maintains their Th2 priming potential. Blood 2005. 105: 47494751.
  • 51
    Ying, S., O'Connor, B., Ratoff, J., Meng, Q., Mallett, K., Cousins, D., Robinson, D. et al., Thymic stromal lymphopoietin expression is increased in asthmatic airways and correlates with expression of Th2-attracting chemokines and disease severity. J. Immunol. 2005. 174: 81838190.
  • 52
    Rimoldi, M., Chieppa, M., Larghi, P., Vulcano, M., Allavena, P. and Rescigno, M., Monocyte-derived dendritic cells activated by bacteria or by bacteria-stimulated epithelial cells are functionally different. Blood 2005. 106: 28182826.
  • 53
    Inohara, N., Ogura, Y., Fontalba, A., Gutierrez, O., Pons, F., Crespo, J., Fukase, K. et al., Host recognition of bacterial muramyl dipeptide mediated through NOD2. Implications for Crohn's disease. J. Biol. Chem. 2003. 278: 55095512.