Immune development in jejunal mucosa after colonization with selected commensal gut bacteria: A study in germ-free pigs

Veterinary Immunology and Immunopathology 119 (2007) 243–253 www.elsevier.com/locate/vetimm

Immune development in jejunal mucosa after colonization with selected commensal gut bacteria: A study in germ-free pigs Karin Haverson a,*, Zuzana Rehakova b,1, Jiri Sinkora b,2, Lidija Sver c, Michael Bailey a a

Department of Clinical Veterinary Science, University of Bristol, Langford BS40 5DU, UK b Academy of Science of the Czech Republic, Novy Hradek, Czech Republic c University of Zagreb, Croatia

Received 26 January 2007; received in revised form 17 March 2007; accepted 31 May 2007

Abstract The immunological structure of the porcine jejunal lamina propria in germ-free piglets was compared with that of their counterparts associated with two strains of commensal Escherichia coli, A0 34/86 serotype O83:K24:H31 and the O86 E. coli strain, up to 20 days post-colonization. In the antigen-presenting compartment, both dendritic cells (DC) and cells expressing CD163, probably macrophages were investigated. In addition we also assessed the number of CD2+/CD3+ (T) cells. In contrast to some previous reports, we show a total lack of both DC and T cells for germ-free animals in the diffuse lymphoid tissue of villi and crypts of the jejunum. Association with either strain of commensal E. coli had a profound effect on the immune structure and resulted in extensive recruitment of DC to the lamina propria and of T cells to epithelium and lamina propria. The data suggest that the earliest immigrant cells were monocytes, which soon acquired the phenotype of mucosal DC. T cells migrated in at a slightly slower rate. Nevertheless, the response could be extremely rapid: within 3 days of colonization with O83, the magnitude of this response was comparable to that observed 20 days post-colonization. # 2007 Elsevier B.V. All rights reserved. Keywords: Germ-free; Gnotobiotic; Jejunum; Dendritic cells; T cells; Macrophages

1. Introduction There is epidemiological and experimental evidence that colonization with commensal bacteria is essential for many functional processes in the gut including the development of a balanced and regulated immune system (Hooper and Gordon, 2001; Hooper et al., 2001). The ‘hygiene hypothesis’ suggests that this regulation may be disturbed in the Western world (Wold, 1998; Liu and Murphy, 2003), accounting for the recent epidemic * Corresponding author. Tel.: +44 117 9289289; fax: +44 117 9289505. E-mail address: [email protected] (K. Haverson). 1 Address: University of Defense, Hradec Kralove, Czech Republic. 2 Address: DakoCytomation AG, Brno, Czech Republic. 0165-2427/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.vetimm.2007.05.022

in atopic diseases in the Western world. However, it has proved extremely difficult to identify the detailed mechanisms involved. Microbial colonization of the gut involves a huge number of complex and interactive populations of microorganisms, most of which are still largely unknown. Microbial composition is highly variable, both with time as well as between individuals, and even varies between litter mates reared under identical environmental conditions. Therefore, the role of individual components in such a complex system is difficult to elucidate. Nevertheless, there is considerable evidence that certain subpopulations of intestinal microorganisms have desirable effects on gut health and the immune system, as so-called probiotics. Lactobacilli and bifidobacteria, certain strains of commensal Escherichia coli and other microorganisms

244

K. Haverson et al. / Veterinary Immunology and Immunopathology 119 (2007) 243–253

have been used as probiotics in humans. In rodents, segmented filamentous bacteria have been shown to affect the mucosal immune system in various ways: Peyer’s patch (PP) germinal centre formation is affected and mucosal epithelial cells as well as intra-epithelial lymphocytes are stimulated and expanded (Okada et al., 1994; Umesaki et al., 1995; Jiang et al., 1998; Talham et al., 1999). The pig provides an exceptional model to study the effect of environmental factors on the development of the immune system, as piglets are born immunologically naı¨ve, without maternal antibody or other maternally transmitted macromolecules, due to their epitheliochorial placenta. Thus, pigs have been used extensively for studies of the effect of gut microflora on the immune system, both in conventional animals as well as in animals raised under germ-free conditions. Colonization with the probiotic bacterium Enterococcus faecium SF68 resulted in decreased colonization by certain pathogenic serovars of E. coli (Scharek et al., 2005). Pre-colonization with an avirulent Salmonella strain protected gnotobiotic piglets against subsequent colonization with the pathogenic strain Salmonella enterica serovar Infantis (Foster et al., 2003). Multiple effects on the immune system have also been reported: Pigs raised under germ-free conditions cannot make serum antibodies to T-dependent and type 2 T-independent antigens (Butler et al., 2002). Colonization diversifies the pre-immune repertoire in mucosal lymphoid tissues (Butler et al., 2000) and expands the T cell receptor (TCR) repertoire in tonsils (Wilson et al., 2005). An increase of blood monocyte oxidative burst activity (Rehakova et al., 1998b) and phenotypic changes in peripheral B cells could also be shown after colonization of germ-free piglets (Sinkora et al., 1998). The two E. coli strains O86 and O83 used to colonize the germ-free (GF) piglets in this experiment are classified as commensals in conventional pigs, some effects on health status and/or the immune system are already known. O83 has been used successfully as an immune-modulator and ‘probiotic’ in human infants (Lodinova-Zadnikova et al., 1991; Dlabac et al., 1995; Lodinova-Zadnikova and Sonnenborn, 1997; Cukrowska et al., 2002; Vancikova et al., 2003). O86 affected the composition of peripheral T cell subsets (Rehakova et al., 1998a) in pigs and specifically and polyclonally stimulated serum and mucosal antibody (Cukrowska et al., 2001). Peyer’s patches are involved in the inductive phase of the mucosal immune response, whereas the so-called diffuse tissue is traditionally thought to be an effector site. However, there is increasing evidence that in

healthy animals, the role of the diffuse lymphoid tissue in the gut is also that of immune regulation (Bailey and Haverson, 2006) and that early environmental events can affect its long-term function. We and others have presented evidence of the highly organized and compartmentalized structure of the diffuse immune tissue of the pig jejunum in mature animals, containing large numbers of immature DC (Haverson et al., 2000; Bimczok et al., 2005, 2006; Haverson and Riffault, 2006). Also present are large numbers of T cells (Vega-Lopez et al., 1993; Rothko¨tter et al., 1994) prone to activation-induced apoptosis (Bailey et al., 1998; Bailey and Haverson, 2006). Immature DC are potentially involved in immune regulation as well as defence and are thought to play a major role in the decision making process between active defence and tolerance. This structure is almost totally lacking at birth and develops gradually over a period of several weeks (Rothko¨tter et al., 1994; VegaLopez et al., 1995). Previous studies with GF piglets have shown reduced T and B cell numbers (Rothko¨tter and Pabst, 1989; Rothko¨tter et al., 1991, 1999) in PP and gut lamina propria (LP). However, no studies have been conducted during the early phase post-colonization with single microbial strains. Therefore, we have investigated the effect of monoassociation of germ-free piglets with these two commensal strains of E. coli on the immune development of the diffuse lymphoid tissue, with particularly emphasis on the antigen-presenting cells. This was done by multicolour immunohistochemistry and pixel-based image analysis, labelling molecules characteristic for DC, monocytes/macrophages and T cells. 2. Materials and methods 2.1. Animals and bacterial association Germ-free Minnesota minipigs reared in Novy Hradek, Czech Republic, were delivered by hysterectomy and raised colostrum-free under germ-free conditions on autoclaved bovine milk as previously described (Mandel and Travnicek, 1987). Bacterial mono-association was done when piglets were between 1 and 4 days of age. GF piglets were mono-associated orally with 107 CFU of the A0 34/86 serotype O83:K24:H31 (O83 hereafter) and the O86 E. coli strain (Sinkora et al., 1998; Cukrowska et al., 1998, 2001), respectively. As controls, one GF piglet was sacrificed aged 5 days and three aged 24 days. The animals were ex-sanguinated under anesthesia aged 24 days.

K. Haverson et al. / Veterinary Immunology and Immunopathology 119 (2007) 243–253

245

Table 1 Monoclonal antibodies and fluorochrome-conjugated secondary reagents Clone MSA3 G7 PPT3 MSA4 MIL17 MIL11 a b

Specificity MHC II DR CD16 CD3 CD2 CD4 Capillary endothelium

Isotype IgG2a IgG1 IgG1 IgG2a IgG2b IgE

Secondary conjugate

Source a

Goat anti-mouse IgG2a–Texas Red Goat anti-mouse IgG1–FITCa Goat anti-mouse IgG1–FITCa Goat anti-mouse IgG2a–Texas Reda Goat anti-mouse IgG2b–FITCa Biotinylated rat anti-mouse IgEa + SA-AMCA b

Dr. J. Lunney Dr. B.Y. Kim Dr. H. Yang Dr. J. Lunney Authors Authors

Southern Biotechnology. Vector.

For practical reasons (i.e. litter size), a balanced experimental design was not always possible: the data shown are built up from several litters, over a period of time and were grouped based on the duration and strain of association. The number of animals per group are indicated on the appropriate figure.

(Yang et al., 1996) and low expression of CD2 occurs on a B cell subset (Sinkora et al., 1998) and on NK cells in pigs. Nevertheless, CD2 and CD3 co-expression is typical for all peripheral ab T lymphocytes and a large subset of porcine gd T cells (Yang et al., 1996). 2.3. Image and statistical analysis

2.2. Immunohistochemistry Three-colour fluorescent immuno-cytochemistry was carried out on cryo-sections, taken either from Peyer’s patches (PP) or from mid-jejunum as described previously (Haverson et al., 2000). Antibodies and secondary reagents used for this study are listed in Table 1. We have shown previously that functional dendritic cells (DC) isolated from the pig jejunum can be characterized by their surface co-expression of CD16 and MHC class II (Haverson et al., 2000), these two molecules were used in a cocktail to characterize DC. CD163, a molecule characteristic for mature monocytes and macrophages and lacking in DC (Sanchez et al., 1999; Buechler et al., 2000), was also used for a more detailed characterization of the APC compartment, in a cocktail with MHC II. In addition, our monoclonal antibody MIL11 has been shown to label endothelium in the pig (Wilson et al., 1996) and is a useful marker for delineating villous structure and to examine MHC II expression on endothelium – unlike in rodents, epithelium is MHC class II-negative in pigs but MHC class II expression on endothelium is found in normal healthy animals. Previous work investigating T cell development in pigs has mostly been carried out using antibodies to CD2 as a T cell marker, as CD3 reagents were not yet available (Vega-Lopez et al., 1993, 1995; Rothko¨tter et al., 1994). As this lack of CD3 reagents has been remedied (Pescovitz et al., 1998) and in order to put this work into context, we used a cocktail of two antibodies to CD2 and CD3 as T cell-associated markers, although it is has recently been proved that CD2 gd T cells exist

Stained slides were examined using a Leica epifluorescence microscope fitted with a combined excitation and emission filter block specific for FITC, Texas Red (TXRD) and 7-amino-4-methylcoumarin-3acetic acid (AMCA). Images were recorded on a digital camera (Photonic Science, Milham Mountfield, UK). Quantitative image analysis was conducted for the diffuse lymphoid areas, including villi and crypts but omitting submucosa. The co-expression of MHCII and other surface molecules was measured by counting pixel areas of single or multiple colour expression as described previously (Rees et al., 2003; Inman et al., 2005). The reference area was the whole villous and crypt area including the epithelium. Please note that the percentages are proportional to area and not to cell numbers. Thus, molecules present in the cytoplasm as well as on the cell surface, such as MHC II, will result in more positive pixels per cell than those present only on the cell surface (such as CD16, CD163, CD2 and CD3). This technique has been validated extensively in our laboratory (Rees et al., 2003; Inman et al., 2005). Between 5 and 10 representative areas for each animal and immunological compartment were analyzed. Means and standard deviations were determined for each group and compared by two-tailed Student’s t-test. 3. Results Fig. 1 illustrates the events immediately post microbial colonization with O86. It is apparent that 1 day post-association (Fig. 1a and d), cells expressing CD163 (arrowhead) and cells expressing MHC II

246

K. Haverson et al. / Veterinary Immunology and Immunopathology 119 (2007) 243–253

Fig. 1. Jejunum of germ-free piglets 1 day (a), 2 days (b) and 20 days (c) post-association with O86. All images show labelling of capillary endothelium with AMCA (blue). In addition, MHC II dR is labelled with Texas Red. The cell surface molecule CD163, characteristic for monocytes and macrophages and absent from DC, is shown by co-labelling with FITC. Images a, b and c have been duplicated as d, e and f, with the red MHC II labelling removed for clarity.

Fig. 2. Jejunum of 24-day-old GF-associated (a–c) and O86-associated (d–f) piglets. All images show labelling of capillary endothelium with AMCA (blue). In addition, MHC II dR is labelled with Texas Red (a, b, d and e). DC are characterized by co-labelling with CD16 (FITC-green; a and d); macrophages by CD163 expression (FITC-green; b and e). T cells are identified by labelling for CD3 with FITC (green) and CD2 with Texas Red (c and f). Images d, e and f have been duplicated as g, h and i, with the red MHC II labelling removed. SM: submucosa.

K. Haverson et al. / Veterinary Immunology and Immunopathology 119 (2007) 243–253

(arrow) are present, but frequently represent different cell populations. One cell co-expressing CD163 and MHC II can also be seen (**). One day later (Fig. 1b and e), many more cells have been recruited, mostly coexpressing both molecules. However, after 20 days (Fig. 1c and f), although large numbers of MHC II+ cells fill the lamina propria, all CD163 expression is lost. A comparison with a corresponding tissue section labelled with CD16 instead of CD163 shows that such MHC II+ areas co-express CD16 and are therefore mucosal DC (Fig. 2d and g). The corresponding photomicrographs for CD2 CD3 for the early time

247

points are not shown, as very few of such cells have been recruited (Fig. 3). Fig. 2 shows representative tissue sections from 24day-old GF- and O86-associated animals. The villi of the diffuse lymphoid tissue of 24-day-old germ-free piglets were profoundly different from those associated with both E. coli strains, with an almost total lack of LP area and a lack of both DC (Fig. 2a), monocytes (Fig. 2b) and T cells (Fig. 2c). The only cells which appeared to be constitutively present in these animals were small numbers of CD16+ MHC II-cells (Fig. 2a) and CD163+ MHC II-cells (Fig. 2b) within the

Fig. 3. Pixel areas, calculated by automated image analysis: CD163-MHC II co-expression (a), CD16-MHC II co-expression (b) and CD2-CD3 coexpression (c). Data are expressed as the proportion of pixels per unit villous and crypt area (means and standard errors). Please note that the percentages refer to the proportion of positive pixels in the whole tissue.

248

K. Haverson et al. / Veterinary Immunology and Immunopathology 119 (2007) 243–253

submucosa and small numbers of CD2 CD3+ gd T cells (Fig. 2c). There was also a small amount of faint MHC II in the LP, but without co-expression of CD16, CD163 or the endothelial marker. This MHC II looked diffuse and ‘spindly’ and did not appear characteristic of a cellular shape. The 5-day-old GF piglet also lacked all immune cells and was indistinguishable from the 24-day-old ones (not shown). In contrast, the diffuse lymphoid jejunal tissue of both groups of mono-associated pigs contained large numbers of cells. The villi were filled with a dense network of DC, characterized by their co-expression of MHC II and CD16. Fig. 2d shows the co-expression of CD16 and MHC II, the CD16+ cellular network in the same tissue section can be seen more clearly in Fig. 2g, where the intense MHC II label has been removed. Fig. 3 shows the results of the quantitative image analysis. A comparison of the 24-day-old animals associated with O83 or O86 showed that both the DC and T cell populations investigated in this study were significantly different from GF animals at the P < 0.001 level. However, there were no significant differences between animals associated for 20 days with the two E. coli strains. Additionally, cell numbers observed 3 days post-association with O83 could be compared to those observed after 20 days: there were no statistically significant differences between these and those 20 days later. However, there was between-pig variability on day 3, two out of three piglets had similar cell numbers to those observed on day 24 and the third pig had lower cell numbers more similar to day 2. The CD163+ MHC IIcells seen in the immediate period following colonization in the crypt and villous areas were absent in all 24-dayold animals. However, some CD163+ MHC II-cells were constitutively present in the submucosa of all germ-free and colonized piglets; bacterial association did not affect their numbers (not shown). Colonization also induced high levels of MHC II on endothelium (Fig. 1d and e). The proportion of endothelium co-expressing MHC II was 2.1  0.9% for germ-free and 72.1  16.9% and 82.7  12.7% for O83- and O86-associated animals, respectively. Again, the difference between GF and associated animals was significant at the P < 0.001 level and not significant between the associated groups. In contrast to the undeveloped status of the diffuse tissue in germ-free pigs, their PP areas showed a distinct immunological structure (not shown). However, we did observe that the villi in the immediate vicinity of PPs contained some leucocytes, which appeared to diminish with increasing distance from the PP itself.

There was a significant difference in morbidity between piglets associated either with E. coli O83 or E. coli O86. GF piglets could be associated with O86 aged 1, 2 or 4 days, without any signs of subsequent morbidity. However, when the O83 strain was used, association at less than 4-day-old led to highly increased mortality/morbidity: two out of two (100%) piglets associated aged 2 days, two out of three (67%) associated aged 3 days and one out of seven (14%) associated aged 4 days had to be sacrificed early. 4. Discussion 4.1. Bacterial colonization Although both O83 and O86 strains of E. coli are classified as ‘commensals’ in normal pigs, we observed significant piglet morbidity and mortality following association of germ-free animals younger than 4 days with the O83 strain, although this strain has been used successfully as a probiotic in infants. Piglets associated with O86 aged 1 or 2 days had 100% survival rates. O83 appears to act as a pathogen in these very young germ-free colostrum-deprived animals, whereas E. coli O86 appears to be truly non-pathogenic in pigs. The difference between the two E. coli strains remains to be investigated further. Germ-free piglets, probably due to their immunologically naive status, may be extremely sensitive indicators of bacterial pathogenicity, more sensitive than human infants who are born with maternally derived antibodies and with small numbers of immune cells in the gut (Howie et al., 1998). Additionally, it is unlikely that human infants have a germ-free status, which seems to be required to turn this commensal into a pathogen. This finding also supports the finding by others that colostrum-deprived germ-free animals respond differently to conventional ones, and suggests that microflora classified as ‘commensal’ in conventional animals may behave as a pathogen in GF ones. 4.2. Dendritic cells High levels of MHC II are a characteristic of DC. We have demonstrated that pig jejunal dendritic cells can be characterized by their high cell-surface co-expression of MHC II and CD16 and other molecules (Haverson et al., 2000) by flow cytometry and functional studies: these cells are potent antigen presenting cells in a mixed lymphocyte reaction. In immunohistochemistry, molecules co-expressed on the cell surface will result in pixel areas carrying both labels. However, we have also

K. Haverson et al. / Veterinary Immunology and Immunopathology 119 (2007) 243–253

shown that these jejunal DC carry high levels of cytoplasmic MHC II, such cytoplasmic molecules will be stained in tissue sections as MHC II+ pixels only. CD16 expression appears restricted to the cell surface, thus, pixel areas positive for CD16 and MHC II as well as areas positive for MHC II alone may reveal DC. Additionally, mature DC may lose cytoplasmic MHC II and surface FcG receptors such as CD16 (Regnault et al., 1999), again resulting in MHC II+ pixel areas. Finally, MHC II may also be expressed on many other activated cells such as B cells, T cells and other leucocytes or stromal cells, albeit at low levels. Thus, MHC II+ pixel areas without either CD16 or endothelium cannot be assigned unambiguously and may represent cytoplasmic MHC II in immature DC, cell surface MHC II of mature DC or other activated cells. Unfortunately, we cannot discriminate clearly between these alternatives in this study. However, it seems reasonable to suggest that much of this high intensity MHC II is associated with resident DC and reflects large amounts of their cytoplasmic MHC II, as the network of CD16+ cells colocalizes closely with the MHC II+ areas. 4.3. T cells and other lymphocytes Most T cells can be characterized by co-expression of CD2 and CD3 in pigs as in other species, reflected in pixel areas positive for both markers. As expected, this expression pattern is found in T cell areas in mesenteric lymph nodes or Peyer’s patches (not shown) and in many cells, predominantly intra-epithelial lymphocytes, in the jejunal tissue (Fig. 2c, f and i). Cells expressing only CD3 and no CD2 are known to be a subset of gd T cells in the pig (Kirkham et al., 1996; Yang and Parkhouse, 1996; Davis et al., 1998, 2001), these cells are well represented in the associated animals (Fig. 3c) and could also be seen in very small numbers in GF animals (Fig. 2c). However, pixel analysis revealed a large amount of CD2 without concomitant CD3, especially in the LP. This lack of CD3 expression with CD2 was surprising and characteristic only for cells in the diffuse gut tissue, it was found rarely in other gut-associated lymphoid tissue such as Peyer’s patches and lymph nodes. This phenomenon may deserve some discussion. Although the TCR-associated molecule CD3 by implication characterizes a cell as a T cell, CD2 expression occurs early during thymic T cell development but is not unique to T cells. CD2 is also found at low intensity on B and NK cells. Therefore, possible identities for these CD2+ CD3 cells we considered were: immature cells undergoing extra-thymic T cell development, NK cells or B cells. Subsequent flow cytometric work on isolated cells confirmed the presence

249

of some CD2+ lymphocytes lacking all other lineagespecific markers including those characteristic for B and NK cells (these cell types are at present under investigation). Thus, we rejected their identity as B or NK cells. However, co-labelling with a cocktail of CD2, CD3 and CD4 antibodies by immunohistochemistry showed that a second larger set of CD2+ CD3-negative cells expressed CD4 (not shown), yet flow cytometric analysis of isolated cells with a similar cocktail had shown that all such jejunal CD4+ cells expressed CD2 and CD3 (Haverson et al., 1999). Additionally, a close inspection of the immunohistochemical images showed that even in cells expressing both CD2 and CD3, the intensity of CD3 staining was very variable and CD3 was frequently located immediately beneath the cell surface CD2, implying CD3 translocation from the cell surface. Such translocation from the cell surface is associated with TCR engagement and immune regulation (Niedergang et al., 1995; Huang et al., 1999; Sanui et al., 2003). Taking together all observations, we considered that we had at least two different CD2+ CD3 cells in this site. The first cell type appeared to be CD2+ but truly CD3-negative cells lacking any other lineage markers, possibly undergoing extra-thymic T cell development. The second cell type appeared to be CD4+ T cells where the epitope recognized by the anti-CD3 antibody was not accessible in the tissue, yet was recognized by the same antibody after cell isolation, possibly because of conformational changes in the epsilon chain in the TCR due to receptor engagement (Gil et al., 2002) or other local causes and reversible upon cell isolation. Unfortunately, the present work did not allow us to discriminate between such lymphocyte subsets and this phenomenon will require further investigation. 4.4. Monocytes/macrophages Although CD163 has been thought to be restricted to monocytes and macrophages in pigs and other species (Sanchez et al., 1999; Chamorro et al., 2000, 2004, 2005), it has recently been shown that CD163 can be expressed by human DC under certain circumstances (Maniecki et al., 2006). The identification of CD163 expression with a macrophage status is therefore not totally unambiguous. Some CD163+ cells were constitutively present in the submucosa of all animals, bacterial association did not affect their numbers. The presence of similar CD16+ cells in the same site and under the same conditions suggested that they were CD16+ CD163+. Interestingly, they lacked any co-expression of MHC class II, which would suggest that at least in this site, they cannot be

250

K. Haverson et al. / Veterinary Immunology and Immunopathology 119 (2007) 243–253

involved in antigen presentation and cannot be DC, their function may be predominantly phagocytic. In addition, we could identify CD163+ (CD16+) MHC II+ cells in the villi of very young piglets, they therefore had a phenotype intermediate between the constitutive CD163+ cells in the submucosa and the DC in the villi of older animals. It has recently been shown that mucosal DC derive from blood monocytes (Varol et al., 2007). Our findings are compatible with this interpretation and show that CD163+ cells, probably blood monocytes, can appear within 1 or 2 days postcolonization, many without co-expression of MHC II. By day 20, they have lost CD163 and express high levels of MHC II, compatible with their transformation into mucosal DC. 4.5. Correlation of bacterial colonization and immune status In some contrast to previous reports, which reported reduced numbers of (CD2+) T cells in Peyer’s patches as well as in diffuse lymphoid tissue of GF pigs, all GF piglets used in this study showed a total lack of DC and T cells in the diffuse tissue of the jejunum at 5 and 24 days of germ-free status. We could confirm this lack of leucocytes in the diffuse tissue by studying six more GF piglets from other experiments conducted in Novy Hradek (not shown). PP tissue was not explored quantitatively in this study, but a clear immunological structure was seen (not shown). An explanation for this discrepancy may be that the GF animals used by other workers were older (49 days) (Rothko¨tter et al., 1991). Strain or other local differences cannot be excluded. As demonstrated in this work, the exact location of the diffuse tissue is also important, villi very close to PP contain DC and T cells. However, these studies demonstrate very clearly that pigs kept 5 or 24 days under germ-free conditions can totally fail to establish an immune system in the diffuse lymphoid tissue of the jejunum. We have previously shown that the number of MHC II+ cells per unit area increase rapidly in newborn conventional piglets and reach levels comparable to adult animals within 4 days (Vega-Lopez et al., 1993). The present study also showed a rapid recruitment of DC, initially as CD163+ MHC II-cells, probably monocytes recruited from blood, as early as 1 day post-colonization with O86, unfortunately, no data are available for this time point for the O83 associated pigs. Some CD163 MHC II+ and CD163+MHC II+ were also seen, probably mature DC and monocytes differentiating into mucosal DC, respectively. One day later, the majority of cells

co-expressed both molecules, i.e. were differentiating into mucosal DC. The speed of this recruitment and differentiation for the O83-associated group was so great that by day 3, cell numbers could be comparable to those observed 20 days later. Unlike the rapid recruitment of APC in conventional animals, we know that T cell numbers increase more slowly and do not reach adult levels until the conventional piglet is approximately 7 weeks old (Vega-Lopez et al., 1995). Surprisingly, the three animals associated with O83 and sacrificed 3 days later already had T cell and other lymphocyte numbers indistinguishable from those found in the 24-day-old associated animals. Thus, mono-association with at least the O83 strain of E. coli led to extremely large and persistent increases in both DC and T cells, reflecting a profound recruiting and activating effect of a single microorganism on the immune system in this tissue. Additionally, a comparison with previous work with farm-reared animals in our laboratory suggested that the immune structure was equivalent to that of much older animals, thus, the rate of DC and especially of T cell recruitment in response to a single E. coli species appeared more rapid than that found in conventional farm piglets exposed to a vastly more complex bacterial flora. Of course, future studies are required for a direct comparison between mono-associated and conventional pigs. The important question as to whether there were different leukocyte recruitment rates between the two different strains remains unanswered, as we lacked enough comparative data for the immediate postassociation period. However, by day 20, no differences were observed. It also remains to be investigated, whether there are more subtle differences in activation status or subset composition. The extent and speed of the cellular recruitment in response to O83 may explain its probiotic effect in humans and suggests mechanisms which may be nonspecific and driven mainly by innate immune mechanisms: it is difficult to conceive that the huge numbers of recruited T cells are specific ones expanded within 3 days in response to a single bacterial species. It appears more likely that epithelial cells, signalling via receptors such as TLRs and NODs, produce chemokines such as IL8 and others (Jung et al., 1995) in response to bacterial products, leading to DC and T cell recruitment. We hypothesize that recruited T cells may well not all be specific for bacterial antigens, but may be activated and expanded non-specifically and characterized by gut homing receptors and memory markers. Complex and transient regulation of homing receptors after recent TCR stimulation have been reported

K. Haverson et al. / Veterinary Immunology and Immunopathology 119 (2007) 243–253

(Sallusto et al., 1999). This interpretation would be compatible with many of the polyclonal effects of mono-association on immune responses observed in the periphery and other tissues. Acknowledgements The support of this work by funding from grant no. 523/03/0186 of the Grant Agency of the Czech Republic (GACR) and also by the European Union (Framework FP 5), as part of a project entitled ‘Defining and validating gut health criteria in young pig, based on digestive physiology, microbiology and mucosal immunology investigations for testing alternative strategies to in-feed antibiotics’, contract number QLK5-LT2000-00522, is gratefully acknowledged, as is the financial support by the British Council of the working visit of Dr. Sver. References Bailey, M., Haverson, K., 2006. The postnatal development of the mucosal immune system and mucosal tolerance in domestic animals. Vet. Res. 37, 443–453. Bailey, M., Plunkett, F., Clarke, A., Sturgess, D., Haverson, K., Stokes, C., 1998. Activation of T cells from the intestinal lamina propria of the pig. Scand. J. Immunol. 48, 177–182. Bimczok, D., Sowa, E.N., Faber-Zuschratter, H., Pabst, R., Rothko¨tter, H.J., 2005. Site-specific expression of CD11b and SIRP alpha (CD172a) on dendritic cells: implications for their migration patterns in the gut immune system. Eur. J. Immunol. 35, 1418– 1427. Bimczok, D., Post, A., Tschernig, T., Rothko¨tter, H.J., 2006. Phenotype and distribution of dendritic cells in the porcine small intestinal and tracheal mucosa and their spatial relationship to epithelial cells. Cell Tissue Res. 325, 461–468. Buechler, C., Ritter, M., Orso, E., Langmann, T., Klucken, J., Schmitz, G., 2000. Regulation of scavenger receptor CD163 expression in human monocytes and macrophages by pro- and anti-inflammatory stimuli. J. Leukoc. Biol. 67, 97–103. Butler, J.E., Sun, J., Weber, P., Navarro, P., Francis, D., 2000. Antibody repertoire development in fetal and newborn piglets. III. Colonization of the gastrointestinal tract selectively diversifies the preimmune repertoire in mucosal lymphoid tissues. Immunology 100, 119–130. Butler, J.E., Weber, P., Sinkora, M., Baker, D., Schoenherr, A., Mayer, B., Francis, D., 2002. Antibody repertoire development in fetal and neonatal piglets. VIII. Colonization is required for newborn piglets to make serum antibodies to T-dependent and type 2 Tindependent antigens. J. Immunol. 169, 6822–6830. Chamorro, S., Revilla, C., Alvarez, B., Lopez-Fuertes, L., Ezquerra, A., Dominguez, J., 2000. Phenotypic characterization of monocyte subpopulations in the pig. Immunobiology 202, 82–93. Chamorro, S., Revilla, C., Gomez, N., Alvarez, B., Alonso, F., Ezquerra, A., Dominguez, J., 2004. In vitro differentiation of porcine blood CD163( ) and CD163(+) monocytes into functional dendritic cells. Immunobiology 209, 57–65. Chamorro, S., Revilla, C., Alvarez, B., Alonso, F., Ezquerra, A., Dominguez, J., 2005. Phenotypic and functional heterogeneity

251

of porcine blood monocytes and its relation with maturation. Immunology 114, 63–71. Cukrowska, B., Trebichavsky, I., Rossmann, P., Rehakova, Z., Sinkora, J., Haverson, K., Lodinova-Zadnikova, R., Tlaskalova-Hogenova, H., 1998. Antigenic stimuli do not influence thymic B lymphocytes: a morphological and functional study in germ-free and conventionally reared piglets. Dev. Immunol. 6, 171–178. Cukrowska, B., Kozakova, H., Rehakova, Z., Sinkora, J., TlaskalovaHogenova, H., 2001. Specific antibody and immunoglobulin responses after intestinal colonization of germ-free piglets with non-pathogenic Escherichia coli O86. Immunobiology 204, 425– 433. Cukrowska, B., Lodinova-Zadnikova, R., Enders, C., Sonnenborn, U., Schulze, J., Tlaskalova-Hogenova, H., 2002. Specific proliferative and antibody responses of premature infants to intestinal colonization with nonpathogenic probiotic E. coli strain Nissle 1917. Scand. J. Immunol. 55, 204–209. Davis, W.C., Zuckermann, F.A., Hamilton, M.J., Barbosa, J.I.R., Saalmuller, A., Binns, R.M., Licence, S.T., 1998. Analysis of monoclonal antibodies that recognize gamma delta T/null cells. Vet. Immunol. Immunopathol. 60, 305–316. Davis, W.C., Haverson, K., Saalmu¨ller, A., Yang, H., Lunney, J.K., Hamilton, M.J., Pescovitz, M.D., 2001. Analysis of monoclonal antibodies reacting with molecules expressed on gd T cells. Vet. Immunol. Immunopathol. 80, 53–62. Dlabac, V., Splichal, I., Rehakova, Z., Trebichavsky, I., Sinkora, J., Kozakova, H., 1995. The influence of colonization of germ-free animals by defined strains E. coli and S. thyphimurium on the development of immune functions and defence mechanisms.In: Abstracts of the 4th International Veterinary Immunology Symposium, Davis, p. 283. Foster, N., Lovell, M.A., Marston, K.L., Hulme, S.D., Frost, A.J., Bland, P., Barrow, P.A., 2003. Rapid protection of gnotobiotic pigs against experimental salmonellosis following induction of polymorphonuclear leukocytes by avirulent Salmonella enterica. Infect. Immun. 71, 2182–2191. Gil, D., Schamel, W.W.A., Montoya, M., Sanchez-Madrid, F., Alarcon, B., 2002. Recruitment of Nck by CD3 epsilon reveals a ligand-induced conformational change essential for T cell receptor signaling and synapse formation. Cell 109, 901–912. Haverson, K., Riffault, S., 2006. Antigen presenting cells in mucosal sites of veterinary species. Vet. Res. 37, 339–358. Haverson, K., Stokes, C.R., Bailey, M., 1999. T cell populations in the pig intestinal lamina propria—memory cells with unusual phenotypic characteristics. Immunology 96, 66–73. Haverson, K., Singha, S., Stokes, C.R., Bailey, M., 2000. Professional and non-professional antigen-presenting cells in the porcine small intestine. Immunology 101, 492–500. Hooper, L.V., Gordon, J.I., 2001. Commensal host–bacterial relationships in the gut. Science 292, 1115–1118. Hooper, L.V., Wong, M.H., Thelin, A., Hansson, L., Falk, P.C., Gordon, J.I., 2001. Molecular analysis of commensal host–microbial relationships in the intestine. Science 291, 881–884. Howie, D., Spencer, J., DeLord, D., Pitzalis, C., Wathen, N.C., Dogan, A., Akbar, A., MacDonald, T.T., 1998. Extrathymic T cell differentiation in the human intestine early in life. J. Immunol. 161, 5862–5872. Huang, J.F., Yang, Y., Sepulveda, H., Shi, W.X., Hwang, I., Peterson, P.A., Jackson, M.R., Sprent, J., Cai, Z.L., 1999. TCR-mediated internalization of peptide–MHC complexes acquired by T cells. Science 286, 952–954.

252

K. Haverson et al. / Veterinary Immunology and Immunopathology 119 (2007) 243–253

Inman, C., Rees, L.E., Barker, E., Haverson, K., Stokes, C.R., Bailey, M., 2005. Validation of computer-assisted, pixel-based analysis of multiple-colour immunofluorescence histology. J. Immunol. Methods 302, 156–167. Jiang, H., Snel, J., Talham, G., Cebra, J., 1998. Segmented filamentous bacteria (SFB) stimulate gut mucosal immune system of formerly germ-free mice. FASEB J. 12, 5280. Jung, H.C., Eckmann, L., Yang, S.K., Panja, A., Fierer, J., Morzyckawroblewska, E., Kagnoff, M.F., 1995. A distinct array of proinflammatory cytokines is expressed in human colon epithelial-cells in response to bacterial invasion. J. Clin. Invest. 95, 55–65. Kirkham, P.A., Takamatsu, H., Yang, H., Parkhouse, R.M.E., 1996. Porcine CD3 epsilon: its characterization, expression and involvement in activation of porcine T lymphocytes. Immunology 87, 616–623. Liu, A.H., Murphy, J.R., 2003. Hygiene hypothesis: fact or fiction? J. Allergy Clin. Immunol. 111, 471–478. Lodinova-Zadnikova, R., Sonnenborn, U., 1997. Effect of preventive administration of a nonpathogenic Escherichia coli strain on the colonization of the intestine with microbial pathogens in newborn infants. Biol. Neonate 71, 224–232. Lodinova-Zadnikova, R., Slavikova, M., TlaskalovaHogenova, H., Adlerberth, I., Hanson, L.A., Wold, A., Carlsson, B., Svanborg, C., Mellander, L., 1991. The antibody-response in breast-fed and non-breast-fed infants after artificial colonization of the intestine with Escherichia coli O83. Pediatr. Res. 29, 396–399. Mandel, L., Travnicek, J., 1987. The minipig as a model in gnotobiology. Nahrung-Food 31, 613–618. Maniecki, M.B., Moller, H.J., Moestrup, S.K., Moller, B.K., 2006. CD163 positive subsets of blood dendritic cells: the scavenging macrophage receptors CD163 and CD91 are coexpressed on human dendritic cells and monocytes. Immunobiology 211, 407–417. Niedergang, F., Hemar, A., Hewitt, C.R.A., Owen, M.J., Dautryvarsat, A., Alcover, A., 1995. The Staphylococcus aureus enterotoxin-B superantigen induces specific T-cell receptor down-regulation by increasing its internalization. J. Biol. Chem. 270, 12839–12845. Okada, Y., Setoyama, H., Matsumoto, S., Imaoka, A., Nanno, M., Kawaguchi, M., Umesaki, Y., 1994. Effects of fecal microorganisms acid their chloroform-resistant variants derived from mice, rats, and humans on immunological and physiological-characteristics of the intestines of ex-germ-free mice. Infect. Immun. 62, 5442–5446. Pescovitz, M.D., Book, B.K., Aasted, B., Dominguez, J., Ezquerra, A., Trebichavsky, I., Novikov, B., Valpotic, I., Sver, L., Nielsen, J., Arn, S., Sachs, D.H., Lunney, J.K., Boyd, P.C., Walker, J., Lee, R., Davis, W., Barbosa, I.R., Zuckermann, F., Saalmuller, A., 1998. Summary of workshop findings for antibodies reacting with porcine T-cells and activation antigens: results from the Second International Swine CD Workshop. Vet. Immunol. Immunopathol. 60, 251–260. Rees, L.E., Ayoub, O., Haverson, K., Birchall, M.A., Bailey, M., 2003. Differential major histocompatibility complex class II locus expression on human laryngeal epithelium. Clin. Exp. Immunol. 134, 497–502. Regnault, A., Lankar, D., Lacabanne, V., Rodriguez, A., Thery, C., Rescigno, M., Saito, T., Verbeek, S., Bonnerot, C., RicciardiCastagnoli, P., 1999. Fc gamma receptor-mediated induction of dendritic cell maturation and major histocompatibility complex class I-restricted antigen presentation after immune complex internalization. J. Exp. Med. 189, 371–380. Rehakova, Z., Sinkora, J., Sinkora, M., Kozakova, H., Splichal, I., Barot-Ciorbaru, R., Trebichavsky, I., 1998a. Effect of controlled

antigenic stimulation on lymphocyte subsets in pigs and pig fetuses. Folia Microbiol. 43, 513–516. Rehakova, Z., Trebichavsky, I., Sinkora, J., Splichal, I., Sinkora, M., 1998b. Early ontogeny of monocytes and macrophages in the pig. Physiol. Res. 47, 357–363. Rothko¨tter, H.J., Ulbrich, H., Pabst, R., 1991. The postnatal-development of gut lamina propria lymphocytes—number, proliferation, and T-cell and B-cell subsets in conventional and germ-free pigs. Pediatr. Res. 29, 237–242. Rothko¨tter, H.J., Kirchhoff, T., Pabst, R., 1994. Lymphoid and nonlymphoid cells in the epithelium and lamina propria of intestinalmucosa of pigs. Gut 35, 1582–1589. Rothko¨tter, H.J., Pabst, R., 1989. Lymphocyte subsets in jejunal and ileal Peyer patches of normal and gnotobiotic minipigs. Immunology 67, 103–108. Rothko¨tter, H.J., Mollhoff, S., Pabst, R., 1999. The influence of age and breeding conditions on the number and proliferation of intraepithelial lymphocytes in pigs. Scand. J. Immunol. 50, 31–38. Sallusto, F., Kremmer, E., Palermo, B., Hoy, A., Ponath, B., Qin, S.X., Forster, R., Lipp, M., Lanzavecchia, A., 1999. Switch in chemokine receptor expression upon TCR stimulation reveals novel homing potential for recently activated T cells. Eur. J. Immunol. 29, 2037–2045. Sanchez, C., Domenech, N., Vazquez, J., Alonso, F., Ezquerra, A., Dominguez, J., 1999. The porcine 2A10 antigen is homologous to human CD163 and related to macrophage differentiation. J. Immunol. 162, 5230–5237. Sanui, T., Inayoshi, A., Noda, M., Iwata, E., Oike, M., Sasazuki, T., Fukui, Y., 2003. DOCK2 is essential for antigen-induced translocation of TCR and lipid rafts, but not PKC-theta and LFA-1, in T cells. Immunity 19, 119–129. Scharek, L., Guth, J., Reiter, K., Weyrauch, K.D., Taras, D., Schwerk, P., Schierack, P., Schmidt, M.F.G., Wider, L.H., Tedin, K., 2005. Influence of a probiotic Enterococcus faecium strain on development of the immune system of sows and piglets. Vet. Immunol. Immunopathol. 105, 151–161. Sinkora, J., Rehakova, Z., Sinkora, M., Cukrowska, B., TlaskalovaHogenova, H., Bianchi, A.T.J., DeGeus, B., 1998. Expression of CD2 on porcine B lymphocytes. Immunology 95 (3), 443–449. Talham, G.L., Jiang, H.Q., Bos, N.A., Cebra, J.J., 1999. Segmented filamentous bacteria are potent stimuli of a physiologically normal state of the murine gut mucosal immune system. Infect. Immun. 67, 1992–2000. Umesaki, Y., Okada, Y., Matsumoto, S., Imaoka, A., Setoyama, H., 1995. Segmented filamentous bacteria are indigenous intestinal bacteria that activate intraepithelial lymphocytes and induce MHC class-II molecules and fucosyl asialo gm1 glycolipids on the small-intestinal epithelial-cells in the ex-germ-free mouse. Microbiol. Immunol. 39, 555–562. Vancikova, Z., Lodinova-Zadnikova, R., Radl, J., Tlaskalova-Hogenova, H., 2003. The early postnatal development of salivary antibody and immunoglobulin response in children orally colonized with a nonpathogenic, probiotic strain of E. coli. Folia Microbiol. 48, 281–287. Varol, C., Landsman, L., Fogg, D.K., Greenshtein, L., Gildor, B., Margalit, R., Kalchenko, V., Geissmann, F., Jung, S., 2007. Monocytes give rise to mucosal, but not splenic, conventional dendritic cells. J. Exp. Med. 204, 171–180. Vega-Lopez, M.A., Telemo, E., Bailey, M., Stevens, K., Stokes, C.R., 1993. Immune cell distribution in the small-intestine of the pig— immunohistological evidence for an organized compartmentalization in the lamina propria. Vet. Immunol. Immunopathol. 37, 49–60.

K. Haverson et al. / Veterinary Immunology and Immunopathology 119 (2007) 243–253 Vega-Lopez, M.A., Bailey, M., Telemo, E., Stokes, C.R., 1995. Effect of early weaning on the development of immune cells in the pig small-intestine. Vet. Immunol. Immunopathol. 44, 319–327. Wilson, A.D., Haverson, K., Southgate, K., Bland, P.W., Stokes, C.R., Bailey, M., 1996. Expression of major histocompatibility complex class-II antigens on normal porcine intestinal endothelium. Immunology 88, 98–103. Wilson, S., Norton, P., Haverson, K., Leigh, J., Bailey, M., 2005. Development of the palatine tonsil in conventional and germ-free piglets. Dev. Comp. Immunol. 29, 977–987.

253

Wold, A.E., 1998. The hygiene hypothesis revised: is the rising frequency of allergy due to changes in the intestinal flora? Allergy 53, 20–25. Yang, H., Parkhouse, R.M.E., 1996. Phenotypic classification of porcine lymphocyte subpopulations in blood and lymphoid tissues. Immunology 89, 76–83. Yang, H., Oura, C.A.L., Kirkham, P.A., Parkhouse, R.M.E., 1996. Preparation of monoclonal antiporcine CD3 antibodies and preliminary characterization of porcine T-lymphocytes. Immunology 88, 577–585.