A distinct subset of intestinal dendritic cells responds selectively to oral TLR7/8 stimulation

Authors


Abstract

The intestinal innate immune system continually interacts with commensal bacteria, thus oral vaccines should induce extra/alternative activation of DC, potentially through TLR. To examine this we collected intestinal lymph DC (iL-DC) under steady-state conditions and after feeding resiquimod (R-848), a synthetic TLR7/8 ligand, which we showed induces complete emptying of gut DC into lymph. iL-DC are heterogeneous with subset-specific functions. In this study we determined the kinetics of iL-DC subset release, activation and cytokine secretion induced by R-848. We show that L-DC comprise three distinct subsets (CD172ahigh, CD172aint and CD172alow) present with similar frequencies in intestinal but not hepatic lymph. No iL-DC express TLR7 mRNA, and only CD172a+ iL-DC express TLR8. However, after oral R-848 administration, output of all three subsets increases dramatically. CD172ahigh DC release precedes that of CD172alow DC, and the increased frequency of CD25high iL-DC is restricted to the two CD172a+ subsets. After feeding R-848 only CD172ahigh iL-DC secrete IL-6 and IL-12p40. However, CD172aint and CD172ahigh DC secrete similar but markedly lower amounts when stimulated in vitro. These results highlight the importance of in vivo approaches to assess adjuvant effects on DC and give novel insights into the subset-specific effects of an oral TLR ligand on intestinal DC.

Abbreviations:
CoeLNX:

coeliac lymphadenectomised

hL-DC:

hepatic lymph DC

iL-DC:

intestinal lymph DC

LP:

lamina propria

MLN:

mesenteric LN

MLNX:

mesenteric lymphadenectomised

pDC:

plasmacytoid DC

PP:

Peyer's patch

R-848:

resiquimod

SS:

steady state

TDL:

thoracic duct leukocyte

Introduction

DC are heterogeneous, and multiple subsets exist in rodent lymphoid tissues 1. Importantly, this phenotypic heterogeneity correlates directly with functional specialization, e.g. uptake of dying cells, secretion of cytokines and skewing of T helper cell differentiation 25. DC migrating via afferent lymph from peripheral tissues to draining LN both activate naive CD4+ T cells and transduce information that determines the differentiation of the activated cells. Migrating DC are thought to be conditioned by microenvironmental cues in peripheral tissues. Thus to understand how different peripheral conditions regulate T cell differentiation, it is crucial to characterize DC that have arrived in LN via lymph. DC in LN have not, however, all reached the node from afferent lymph. For example, in addition to plasmacytoid DC (pDC), some conventional DC enter directly from blood (reviewed in 1). Thus, it is difficult to define the specialized functions of DC that have arrived in LN via lymph, as there is no definitive way to identify lymph-derived DC amongst DC extracted from LN. In addition, DC extracted from LN will have arrived in the nodes over a relatively long time period, and recent studies have shown that extracting DC from lymphoid tissues leads to functional maturation 6. This makes it difficult to identify and characterize those DC that have responded to an added peripheral stimulus. Therefore, in many instances DC thought to have arrived via lymph to LN in mice have been described as a homogeneous population of interstitial DC 7, 8.

To selectively study DC that are actively migrating from the gut or the liver, we collect these cells from thoracic ducts of rats that have previously been mesenteric or coeliac lymphadenectomised (MLNX and CoeLNX, respectively) 9. This enables us to study, with minimal in vitro manipulation, DC that have very recently left the intestine or the liver. Using this system we have shown that intestinal lymph DC (iL-DC) collected directly from rat lymph in steady state (SS) are not a homogeneous population of cells but represent at least two phenotypically distinct subsets 10. This is also true for afferent lymph DC in other species 1113. Importantly, we have also shown that this phenotypic heterogeneity correlates with different functions in vivo, as iL-DC expressing low levels of CD172a selectively carry apoptotic material derived from intestinal epithelial cells to the T cell areas of the mesenteric lymph nodes (MLN) 2.

It is currently thought that the induction of active immune responses depends on activation of the innate immune system and that recognition of pathogen-associated molecules, particularly via TLR, is crucial to this activation. Even in SS conditions, however, the intestinal innate immune system is interacting with commensals, and this interaction is essential for intestinal homeostasis 14. How pathogens evoke different responses is at present unclear. An effective oral vaccine will thus need to stimulate the innate immune system differently from commensal organisms and change the properties of migrating DC. The key changes in DC properties induced following TLR-dependent activation of the innate immune system are up-regulation of costimulatory molecule expression and cytokine secretion. The expression of TLR by DC is not, however, homogeneous. For example, murine DC subsets differentially express TLR7, and these expression patterns may result in subset-dependent responses after i.v. administration of TLR7/8 ligands such as the synthetic adjuvant resiquimod (R-848) 15, 16. With respect to migrating DC, especially their subsets, very little is known about expression of TLR and their response to specific TLR ligands in vivo.

In this study, we have characterized DC subsets migrating from the intestine in terms of the expression of possible ligands, including TLR, to find potential targets for oral adjuvants. In a recent study, we demonstrated that feeding the adjuvant R-848 to rats and mice results in complete emptying of DC from the lamina propria (LP) of the small intestine into afferent lymph 17. We now examined the expression of TLR7/8 by iL-DC subsets and their in vivo response to oral R-848. We show that iL-DC represent a more complex population than previously appreciated, comprising three subsets (CD172ahigh, CD172aint and CD172alow) present with similar frequencies. In marked contrast, we show that the CD172ahigh subset is virtually absent from hepatic lymph. No iL-DC express TLR7, but subsets do express TLR8 at different levels. After feeding R-848 the subsets enter lymph and express activation markers with different kinetics. Finally, we show that only the CD172ahigh subset of iL-DC secretes pro-inflammatory cytokines in response to oral R-848 administration. These results extend our knowledge of migrating intestinal DC subsets and provide information that will be crucial in understanding the outcome of DC subset interactions with naive T cells and thus in improving vaccine design.

Results

Migrating DC comprise three subsets with different frequencies in intestinal and hepatic lymph

DC migrating from the gut and liver in rats comprise two phenotypically distinct subsets differing in expression of CD172a and CD4 9, 10. Here we have further analyzed these subsets following their collection from the thoracic ducts of MLNX and CoeLNX rats. CD103+ thoracic duct leukocyte (TDL) were enriched using OX62 MACS beads. All large MHC-IIhigh TDL are also CD103+ (Fig. 1A and 9). The CD103 staining was therefore omitted, and large MHC-IIhigh cells (L-DC) were further examined by FACS for expression of CD172a and CD11b/c. This labelling defined three distinct subsets of intestinal lymph DC present with similar frequencies (Fig. 1B). Further analysis of the three subsets showed that they also expressed different levels of CD32, identifying CD172ahighCD11b/clowCD32low (CD172ahigh), CD172aintCD11b/chighCD32high (CD172aint) and CD172alowCD11b/cintCD32int (CD172alow) subsets (Fig. 1B). When the CD32 staining was exchanged for CD103, all subsets expressed similar high levels of this marker (Fig. 1C). The same surface markers also identified three subsets of hepatic lymph DC (hL-DC) (Fig. 1D). However, in marked contrast to intestinal lymph DC, the frequency of the CD172ahigh subset was very low (1–3% of hL-DC), the frequency of CD172alow DC was only 15–25% and the CD172aint subset represented 65–75% of hL-DC. These experiments show that DC migrating under SS conditions in both hepatic and intestinal lymph can be separated into three subsets by expression of CD11b/c, CD32 and CD172a but that the frequency of these subsets differs dramatically between intestine and liver.

Figure 1.

Surface phenotype of intestinal and hepatic L-DC. (A) CD103 and MHC-II expression by large TDL collected from MLNX rats. CD11b/c, CD172a and CD32 expression on L-DC (gated on large MHC-IIhigh CD103-enriched TDL) from cannulated MLNX (B, C) or CoeLNX (D) rats. Numbers in density plots and histograms represent percentages (of CD172ahigh, CD172aint and CD172alow L-DC) and the mean fluorescent index (MFI), respectively.

Oral R-848 stimulates differential release of intestinal DC

We have shown that intravenous administration of the TLR4 ligand LPS results in increased numbers of DC exiting both the intestine and liver 9, 18. For the intestine this is due to complete emptying of DC from the LP 9. We have recently observed that feeding rats with R-848, a TLR7/8 ligand, also results in a dramatic but transient increase in the numbers of DC in intestinal and liver lymph (17 and unpublished observations). To investigate the effects of oral R-848 on the kinetics of DC subset release, cannulated lymphadenectomised rats were fed R-848 and lymph collected over 2-h intervals. This kinetic study showed that early after feeding R-848 (2–6 h), CD172ahigh iL-DC constituted 40–45% of total iL-DC (Fig. 2A, B). However, by 6–8 h this subset was outnumbered, as total numbers of DC of the other two subsets of iL-DC had increased dramatically (Fig. 2C). In particular, CD172alow iL-DC constituted roughly half of all iL-DC recovered at the peak of output (8–10 h) (Fig. 2C). The total output of DC then gradually waned (Fig. 2C) and returned to normal levels by 24–36 h (data not shown).

Figure 2.

Effects of giving oral R-848 on the output of L-DC subsets from MLNX and CoelNX rats. CD11b/c and CD172a expression by iL-DC (A–C) or hL-DC (D–F) collected over 2-h periods after feeding R-848 (50 μg/rat). Numbers above dot plots represent the collection period in hours, and the numbers in the dot plots represent the frequency of each subset. Graphs show the mean frequency (B, E) or the mean total number (C, F) of each subsets of L-DC over the collection periods. The total number of L-DC was determined using the gates shown (A, D), the total number of TDL collected at each collection period is presented as mean ± SEM. These results were obtained from three independent experiments with a total of six MLNX rats and two independent experiments with a total of three CoeLNX rats.

In hepatic lymph the output of CD172aint DC (the most abundant) increased after feeding R-848 both as a proportion and in total numbers (Fig. 2D–F). The proportion of CD172alow hL-DC decreased after feeding (Fig. 2D, E), but this was primarily due to a large increase in the numbers of CD172aint DC, and by 8–12 h the output of CD172alow DC had also increased (Fig. 2F). The output of CD172ahighCD11b/clow DC, present at very low frequency in SS hepatic lymph, increased only marginally in frequency and numbers after feeding R-848 (Fig. 2D–F).

These experiments show that after feeding R-848, iL-DC expressing CD172a, in particular the CD172ahigh subset, are initially released in higher numbers into intestinal lymph but then become rapidly outnumbered by the CD172alow subset. In addition they show that in hepatic lymph, even after feeding R-848, the CD172ahigh subset constitutes only a very minor proportion of total hL-DC.

Oral R-848 induces differential up-regulation of activation markers on iL-DC subsets

To determine if the differential release of DC subsets is accompanied by differences in activation, DC collected over different intervals were stained for activation markers. Initially, total iL-DC from pooled samples at the peak of output (6–12 h post-feeding) were compared to iL-DC collected from control animals (Fig. 3A, B). No increase in either CD80 or CD86 could be detected, but CD25 expression was up-regulated on iL-DC. When CD172a was included in the staining, a selective increase in the proportion of CD25high cells was observed among CD172a+ iL-DC (Fig. 3B). Increased expression of CD25 has been associated with DC activation in both rats and mice 1921. To determine the activation of the three subsets of iL-DC described in Fig. 1A, four-colour FACS analysis of iL-DC subsets was performed (Fig. 3C). This analysis showed that a large proportion of CD172ahigh iL-DC rapidly increased CD25 expression after feeding of R-848, with a peak at 4–8 h post-feeding (Fig. 3C, D). An increase in the frequency of CD25highCD172aint DC was also observed, but this was not as prominent and was delayed compared to CD172ahigh iL-DC, with a peak at 8–10 h post-feeding (Fig. 3C, D). A minor shift in CD25 expression could be observed in some experiments, but no changes in the frequency of CD25high cells was seen in CD172alow iL-DC at any time after feeding R-848.

Figure 3.

Subset-specific activation of iL-DC after oral R-848 administration. (A) CD25, CD80 and CD86 expression by total iL-DC (gated on large MHC-II+ TDL) collected 6–12 h after feeding R-848 (50 μg/rat; filled histogram) or PBS (thick line). Dotted lines show staining with the isotype control. (B) CD25 and CD172a expression by iL-DC (gated on OX62-enriched large MHC-II+ TDL) collected 6–12 h after feeding PBS or R-848. CD25 expression by (C) iL-DC subsets at 0–2 and 6–8 h or (E) hL-DC subsets at 0–2 and 8–10 h after feeding R-848 (subsets gated as indicated by the regions and arrows). The numbers in the histograms represent the frequency CD25-expressing cells in each subset. Graphs show the mean frequency of each iL-DC (D) or hL-DC (F) subset that expresses CD25 over each collection period (gated as in C and E). The total numbers of L-DC were determined using the gates shown in (C) and (E), the total numbers of TDL obtained during each collection period are presented as mean ± SEM (N.D. = not determined). These results were obtained from three independent experiments with a total of six MLNX rats and two independent experiments with a total of three CoeLNX rats.

Analysis of DC exiting the liver after oral R-848 administration also showed that the frequency of CD25+CD172aint hL-DC increased gradually to reach a peak at 8-12 h after feeding (Fig. 3E, F). As observed for CD172alow iL-DC, expression of CD25 on CD172alow hL-DC did not increase significantly at any time after feeding R-848 (Fig. 3E, F). As the frequency of CD172ahigh hL-DC was very low in control rats and did not increase significantly after feeding R-848 (Fig. 2D–F), the total numbers of CD172ahigh hL-DC were too small to permit a proper kinetic analysis of CD25 expression. At the peak of output these cells did, however, express increased levels of CD25 (data not shown).

Thus, following feeding of R-848, the frequency of CD25high DC is not increased among CD172alow iL-DC or hL-DC, whereas it is significantly increased in the CD172a-expressing subsets. For iL-DC, the increase in frequency of CD25high DC is most rapid and prominent in CD172ahigh iL-DC.

Orally administered R-848 induces IL-6 and IL-12p40 secretion exclusively by CD172ahigh iL-DC

To examine cytokine secretion by iL-DC, the three subsets were sorted by MoFlo from normal rats or rats fed R-848 18 h previously. The sorted DC from both sets of rats were then cultured alone or in the presence of R-848 and/or CD40L (Fig. 4A), and the supernatants assayed for IL-12p40 and IL-6 after 24 h. No DC from untreated rats secreted significant amounts of IL-12 p40 or IL-6 following culture in media alone or with CD40L. Following in vitro stimulation of DC from control rats with R-848, both populations of CD172a+ iL-DC secreted moderate amounts of both cytokines, whereas CD172alow DC secreted very low amounts of the cytokines (Fig. 4A). CD172ahigh DC consistently secreted more cytokines than CD172aint DC, though this difference was diminished in some experiments if CD40L was included in the cultures. However, when iL-DC subsets were isolated from rats that had been fed R-848 18 h previously, CD172ahigh DC secreted large amounts of both cytokines in the absence of any further stimulation. In striking contrast, the other DC subsets isolated from R-848-fed rats did not secrete significant amounts of these cytokines. The amounts of IL-12p40 and IL-6 secreted by CD172ahigh DC from R-848-fed animals were 15-fold and 4-fold higher, respectively, than from DC isolated from normal rats that had been stimulated in vitro with R-848. None of the iL-DC subsets, however, secreted detectable amounts of IL-12p70, and in preliminary experiments the levels of secreted IL-10 did not differ significantly between the subsets (data not shown).

Figure 4.

Subset-specific cytokine production by L-DC subsets after R-848 stimulation in vitro or in vivo. L-DC from MLNX (A, B) or CoeLNX (B) rats were fed either PBS or R-848 (50 μg/rat) and lymph collected for 18 h. DC were sorted into three subsets by FACS as indicated by the gates in Fig. 3. Sorted DC were then cultured for 24 h without further stimulation or with human recombinant CD40L trimer (5 μg/mL) or R-848 (1 μg/mL) or a mixture of both. Following culture of (A) 7.5 × 104 or (B) 9 × 104 sorted L-DC for 24 h as indicated, the supernatants were assayed for cytokines by ELISA (N.D. = not determined). The results are representative of three independent experiments in (A) and two in (B).

CD172aint hL-DC from R-848-fed rats secreted IL-12p40 and IL-6 in amounts similar to CD172aint iL-DC (Fig. 4B). We could not obtain enough cells from the CD172ahigh hL-DC to assay for cytokines in these experiments. After stimulation in vitro with R-848, CD172alow hL-DC, similar to the same subset of iL-DC, did not secrete detectable amounts of IL-12p40 (data not shown).

These experiments show that CD172aint and CD172ahigh DC but not CD172alow DC secrete IL-6 and IL-12p40 after stimulation in vitro with R-848. Moreover, they show that the CD172ahighCD11b/clow subset and only this subset secretes very high levels of these cytokines after feeding of R-848, much higher levels than detected following stimulation in vitro.

Differential expression of TLR7 and TLR8 by iL-DC subsets

To determine if the subset-specific secretion of IL-12p40 and IL-6 by iL-DC following feeding of R-848 or stimulation with R-848 in vitro relates to differential expression of TLR7 and 8, the three subsets were isolated and mRNA expression examined by PCR. None of the iL-DC subsets expressed detectable levels of TLR7 (Fig. 5A). Both CD172a+ iL-DC populations did, however, express TLR8, while no transcripts could be detected in the CD172alow subset (Fig. 5A). In preliminary quantitative PCR experiments, the level of TLR8 mRNA in these two subsets was not significantly different (two-tailed paired t-test, p=0.355).

Figure 5.

Expression of TLR7 and TLR8 by iL-DC subsets. (A) iL-DC subsets were sorted as in Fig. 3, and the mRNA expression of TLR7 and TLR8 by each subset was detected by reverse transcription and PCR. Amplified products were visualized on an agarose gel with the housekeeping gene Cyclophilin B. Data are representative of four independent experiments. (B) Sorted CD172ahigh cells (4 × 104) were cultured for 24 h with R-848 (1 μg/mL) or Loxoribine (0.5 mM) as indicated, and supernatants were screened for IL-12p40 by ELISA. The results are representative of two independent experiments.

To determine if the lack of detectable expression of TLR7 mRNA resulted in unresponsiveness to TLR7 ligands, CD172ahigh iL-DC were stimulated in vitro with Loxoribine, a TLR7-specific ligand 22, and secreted IL-12p40 was measured. Very little if any IL-12p40 was secreted after stimulation with Loxoribine, while co-incubation with R-848 resulted in significant secretion of the cytokine (Fig. 5B). Stimulation of the other subsets of iL-DC by Loxoribine in vitro did not induce any secretion of IL-12p40 but did result in high levels of cytokine secretion by splenic pDC (data not shown).

These experiments strongly suggest that iL-DC do not express functional TLR7. Hence, the lack of cytokine secretion by CD172alow iL-DC could be explained by the lack of TLR7 and 8. However, the subset-specific cytokine secretion by the CD172ahigh cells following feeding of R-848 could not be completely accounted for by differences in expression of TLR7 or TLR8.

Discussion

An effective oral vaccine will need to stimulate the innate intestinal immune system in a manner different from stimulation received via commensal bacteria. How this stimulation is to be achieved most effectively by vaccines is unclear. What is generally accepted, however, is 1) that information from the innate immune system needs to be transduced from the periphery to the sites of activation of naive T cells, in this case Peyer's patches (PP) or MLN; 2) that the DC that are actively involved in antigen and information transfer are those migrating in afferent lymph; 3) that TLR are important receptors for innate immune cell activation, and therefore their ligands are potential strong adjuvants. However, little is known about TLR expression on and responsiveness of migrating DC (and potential subsets thereof), as they cannot be studied easily in mice or humans. In this study we have examined these migrating intestinal DC directly in lymph using our unique model 23. We extend our previous studies and show that DC migrating from the small intestine comprise three distinct subsets. We also show that the migration of distinct DC subsets in lymph is not confined to the intestine but is also true for the liver. There is, however, subset tissue-specificity, as only two of these subsets can be identified in hepatic lymph. Whether this relates to the distinct nature of immune responses induced via the liver remains to be determined. All iL-DC are negative for TLR7 expression, whereas two of the subsets express TLR8. All three subsets are released, but with different kinetics, from the gut following feeding the TLR7/8 ligand R-848. Importantly, we show that secretion of IL-6 and IL-12p40 is restricted to the subset of iL-DC that is absent from hepatic lymph.

DC migrating from peripheral tissues are heterogeneous, but the functional significance of this heterogeneity is uncertain. We have previously shown that iL-DC comprise two subsets differing in expression of CD172a (SIRPα) and CD4. We now show that CD172a+ DC can be further divided by expression of CD11b/c and the level of CD172a expression. In intestinal lymph these three subsets, CD172ahigh, CD172a int and CD172a low, are present in roughly equal numbers, whereas hepatic lymph contains very few CD172ahigh(CD11b/clow) DC. The presence of a CD172ahigh DC population in intestinal lymph is intriguing, and given its functional properties, it is important to know its anatomical origin. It is unlikely that these DC are migrating from PP. We have shown that feeding R-848 leads to a marked increase in output of all three DC subsets into lymph, accompanied by complete emptying of DC from the LP. However, by immuno-histochemistry and FACS analysis, there is no decrease in PP DC numbers (17 and unpublished observation). That DC do not migrate from the PP to the MLN under SS conditions has recently been suggested by studies in minipigs 11.

Oral administration of R-848 dramatically and rapidly stimulates the entry of L-DC into intestinal 17 and hepatic lymph. Output of CD172ahigh DC increases before that of the other subsets, but the release of CD172alow DC in both intestinal and hepatic lymph also increases after feeding of R-848; indeed, the CD172alow subset constitutes the majority of iL-DC at the peak of increased output (8–10 h post-feeding). To determine if these differences reflect direct or indirect stimulation of DC, we examined TLR7 and TLR8 expression by migrating DC. No iL-DC express TLR7, and CD172alow iL-DC do not express TLR8. These data show that even though the CD172alow subset cannot respond directly to the TLR7/8 ligand, it is stimulated to migrate by cytokines released by other TLR7/8-expressing cells. It is possible that the rapid increase in output of CD172a+ DC represents direct stimulation by TLR8 on these cells, whereas the delayed release of CD172alow DC that do not express TLR8 represents an indirect effect. The indirect effect could be mediated by TNFα, as it is in mice where the bulk of this cytokine is secreted by pDC 17.

Activation/maturation of DC is a complex and controversial area. DC migrating in lymph under SS conditions express very high levels of surface MHC class II but low levels of CD25, CD80 and CD86. We have recently shown that the majority of iL-DC released from the LP following oral R-848 administration have up-regulated CD25 but not CD86, whereas MLN DC express increased levels of CD25 and CD86 17. CD25 expression can be increased on rat and murine DC during in vitro culture 21, 24, and the receptor on rat lymph DC binds IL-2 21 and may have a functional role 19. Here we show that after feeding rats R-848, there is an increased frequency of CD25high DC among CD172a+ iL-DC, particularly in the CD172ahigh subset. In marked contrast to both CD172a+ subsets, the frequency of CD25highCD172alow DC in liver or intestinal lymph did not change after feeding R-848. The up-regulation of costimulatory molecules on DC in vivo appears to be similar on splenic and LN DC purified from TLR+ or TLR mice following TLR ligand administration 25, 26. Our study suggests, however, that CD25 expression on lymph DC may distinguish those DC that have been stimulated directly via TLR from those released via indirect stimulation, as only the TLR8+CD172aint and CD172ahigh subsets express CD25. When DC reach the draining lymph node, however, they will be exposed to locally secreted cytokines and will up-regulate CD25 and costimulatory molecules such as CD86.

In contrast to the up-regulation of costimulatory molecules by DC after administration of TLR ligands, secretion of cytokines depends on expression of the appropriate TLR by the DC 26. Importantly, TLR-dependent cytokine secretion correlates directly with the induction of cytokine-secreting effector T cells 25, 26. When sorted subsets of iL-DC were cultured with R-848 in vitro, the two CD172a+ TLR8-expressing subsets secreted similar amounts of IL-12p40 and IL-6, whereas TLR8CD172alow iL-DC did not secrete either cytokine. However, when iL-DC were purified from rats that had been fed R-848, the CD172ahigh subset, and only this subset, produced large amounts of IL-12p40 and IL-6 without further in vitro stimulation. The difference in cytokine secretion between CD172ahigh and CD172aint DC is probably not due to differential expression of TLR8, as neither the level of expression by the DC nor the secretory response in vitro differed significantly. In addition, we have not been able to detect any modulation in expression of CD11b/c or CD172a in vitro by sorted iL-DC after various stimuli, including R-848 (data not shown). It is possible that CD172ahigh and CD172aint populations differ in their anatomical distribution in the gut, with CD172ahigh iL-DC being more accessible to R-848. Indeed, it has been shown in histochemical analysis of the GALT of minipigs that CD11b and CD172a can be used to distinguish DC from different areas of the gut 11. In rats, CD172ahigh DC may be located in the proximal part of the ileum, a region where murine DC can extend processes into the lumen of the gut 27. DC present in this part of the intestine have been shown to actively transcribe IL-12p40 and IL-12p19 28. In preliminary experiments, however, we have not detected an increase in IL-12p19 gene expression or any significant increase in IL-12p70 secretion by CD172ahigh iL-DC after feeding R-848.

We have previously shown that CD172alow iL-DC selectively transport apoptotic intestinal epithelial cells to T cell areas of MLN 2. ssRNA, both viral and self (29, 30 and C. Reis e Sousa, personal communication), is the natural ligand for TLR7 or 8, depending on the species. The subset-specific lack of TLR7/8 expression could be important for tolerance to self antigens, as the CD172alow iL-DC lacking TLR7 and 8 expression will be less likely to respond to their cargo of cellular debris, which includes ssRNA. That this could be important for self tolerance is supported by murine studies showing that CD8α+ DC, which share the selective ability to take up dying cells in vivo2, 3, do not express TLR7, whereas CD8α DC do 16.

In this study we have shown that DC migrating from the intestine comprise a heterogeneous population of cells that can be divided into three subsets, CD172a high, CD172aint and CD172alow, and that the frequencies of these subsets differ in DC originating from the liver. Moreover, we show that the CD172ahigh subset, which is virtually absent in liver lymph, selectively secretes large amounts of immunomodulatory cytokines after oral administration of a TLR7/8 ligand. This subset-specific response cannot simply be explained by different levels of TLR7 and 8, as all iL-DC are TLR7, and while the CD172alow subset is TLR8, the two CD172a+ subsets express comparable levels of TLR8 mRNA. These experiments provide novel information regarding the regulation of migration, activation and cytokine secretion by intestinal DC subsets in response to TLR stimulation and highlight the importance of complementing analysis of TLR expression with in vivo functional studies when assessing potential oral adjuvants.

Materials and methods

Animals and surgical procedures

PVG (RT1c) rats were bred and maintained under specific pathogen-free conditions at Sir William Dunn School of Pathology (Oxford, UK). Rats used were males 12–24 wks of age. Mesenteric and coeliac lymphadenectomy (MLNX and CoeLNX, respectively) and thoracic duct cannulation were performed as described previously 23, 31. All procedures were carried out in accordance with Home Office guidelines.

Reagents

Cells were cultured in IMDM with 5% FCS, 50 μM β-mercaptoethanol, 100 μg/mL penicillin, 50 U/mL streptomycin (IMDM-5%; all Invitrogen Life Technologies). R-848 and Loxoribine (InvivoGen, San Diego, CA) were dissolved in water. Human recombinant CD40L trimer was provided by Amgen.

mAb

mAb to the rat antigens Igκ (OX12), CD6 (OX52), CD8α (OX8), CD11b/c (OX42), CD103 (OX62) and CD172a (OX41) were purified from cell culture supernatants and used for depletions or conjugated to biotin, FITC or Cy5. Purified anti-CD32 (D34–485) (BD Pharmingen, Oxford, UK) was biotinylated in house, and PE-labelled anti-CD25 (OX39), PerCP-labelled anti-MHC-II (OX6) and Streptavidin-PE and -allophycocyanin (APC) were purchased from BD Pharmingen. PE-labelled anti-CD80 (3H5) and -CD86 (24F) were purchased from Serotec (Oxford, UK).

Isolation of cells

Thoracic duct leukocytes (TDL) were collected on ice in PBS with 10 mM EDTA and 20 U/mL Heparin. The TDL were then passed through a 70-μM cell strainer (BD Biosciences), and RBC were lysed with ACK lysis buffer and used for FACS analysis. In some experiments the L-DC were enriched using OX62-MACS beads and AutoMACS (Miltenyi Biotech, Germany). In functional studies of L-DC subsets, the MACS-enriched CD103+ cells were purified further by cell sorting using a MoFlo (Cytomation, Fort Collins, CO). Purity of the cell populations was greater than 95%.

FACS

Surface staining for FACS was performed in PBS with 2% FCS and 10 mM EDTA for 15 min on ice after blocking in 10% rat serum. Cells were then fixed in 2% paraformaldehyde and analysed using a FACScalibur (BD Biosciences).

PCR

Cells were suspended and the RNA extracted in Trizol (Invitrogen). RNA was treated with the DNAfree kit (Ambion, Austin, TX) and RNasin (Promega, Madison, WI) and reverse transcribed using the Reverse-IT RT system (ABgene, Epsom, UK). The amount of cDNA was normalised to Cyclophilin B by non-saturating PCR. Sequences of the primers were: TLR7, forward CTGTGTGGTTTGTCTGGTGG, reverse CACTTTGACCTTTGTGTGCG; TLR8, forward GGCATTTACACGCTCACAGA, reverse CCATCATTTGCATTCCACAG; cyclophilin B, forward CAAGACCTCCTGGCTAGACG, reverse GCTGTCCGTCTTGGTGTTCT. PCR reactions were run with the following parameters: denaturation at 95°C for 30 s, annealing at 59°C for 30 s and extension at 72°C for 20 s repeated for 35 cycles to ensure maximum sensitivity (detection limit ⩽200 copies). Reactions were analysed by gel electrophoresis.

Quantitation of cytokine production

ELISA for rat IL-6 (BD Pharmingen) and IL-12p40 and p70 (Biosource International, Nivelles, Belgium) were performed according to the manufacturers’ protocols.

Acknowledgements

U. Y. is a Wellcome Trust International fellow. This work was financed by the BBSRC and the E. P. Abrahams Trust. We thank Dr. R. Josien for his critical review of the manuscript, Dr. F.-X. Hubert for measurements of IL-12p70 and Nigel Rust for excellent assistance with cell sorting.

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

Ancillary