Despite a growing understanding of the role of cytokines in immunity to the parasitic helminth Trichuris muris, the local effector mechanism culminating in the expulsion of worms from the large intestine is not known. We used flow cytometry and immunohistochemistry to characterize the phenotype of large intestinal intraepithelial lymphocytes (IEL) and lamina propria leukocytes (LPL) from resistant and susceptible strains of mouse infected with T. muris. Leukocytes accumulated in the epithelium and lamina propria after infection, revealing marked differences between the different strains of mouse. In resistant mice, which mount a Th2 response, the number of infiltrating CD4+, CD8+, B220+, and F4/80+ IEL and LPL was generally highest around the time of worm expulsion from the gut, at which point the inflammation was dominated by CD4+ IEL and F4/80+ LPL. In contrast, in susceptible mice, which mount a Th1 response, the number of IEL and LPL increased more gradually and was highest after a chronic infection had developed. At this point, CD8+ IEL and F4/80+ LPL were predominant. Therefore, this study reveals the local immune responses underlying the expulsion of worms or the persistence of a chronic infection in resistant and susceptible strains of mouse, respectively. In addition, for the first time, we illustrate isolated lymphoid follicles in the large intestine, consisting of B cells interspersed with CD4+ T cells and having a central zone of rapidly proliferating cells. Furthermore, we demonstrate the organogenesis of these structures in response to T. muris infection.
Trichuris muris is a natural mouse model of the nematode parasite, Trichuris trichiura, one of the most prevalent human helminth infections worldwide. The range of protective immunity mounted against T. muris in the mouse infection model varies depending upon the background genetics of the inbred strain of mouse (1, 2) and parallels the range of responses observed within an outbred human population exposed to T. trichiura. The majority of mouse strains, such as BALB/c, are resistant to T. muris and quickly expel the parasite, whereas a few strains, such as AKR, are susceptible, allowing the development of fecund adult parasites, culminating in a chronic infection of the cecum and proximal colon.
It is now well established that a Th2-dominated response, characterized by the production of IL-4, IL-5, IL-9, and IL-13, is an absolute requirement for the expulsion of worms by resistant strains of mouse (3, 4, 5, 6). Susceptible strains, rather than failing to respond to T. muris, instead mount an inappropriate Th1 response associated with high levels of IFN-γ and IL-12 (7, 8). Despite this knowledge, the effector mechanism ultimately responsible for the expulsion of T. muris by the host is not understood. Many of the Th2 responses typically associated with helminth infection, such as mastocytosis, eosinophilia, and strong parasite-specific Ab responses, are not essential (9, 10). Resistance can be conferred to immunodeficient SCID mice by the transfer of CD4+ donor cells (9, 11). However, in this model, protective immunity can be abrogated (using a combination of Ab against the gut-homing adhesion molecules mucosal addressin cell adhesion molecule-1, β7 integrin, and CD103 (11)) by blocking T cell migration to the gut (12). This supports the theory that locally acting T cell-dependant effector mechanisms are responsible for the expulsion of T. muris from the large intestine.
Lamina propria leukocytes (LPL)3 and intraepithelial lymphocytes (IEL) are the effector compartments of the gut mucosal immune system (13). By virtue of their anatomical location, IEL have the closest direct contact with foreign Ags derived from the gut lumen and are thought to play a key role in the immune responses to these Ags and in the pathogenesis of a variety of disease states.
Small intestinal IEL have been extensively studied in the mouse. Most are T cells, but compared with peripheral T cells found in secondary lymphoid organs, a high proportion of IEL are CD8+, express TCRγδ, and develop independently of the thymus. Thymus-independent IEL, which can be either TCRαβ+ or TCRγδ+, are relatively abundant and express CD8 in its αα homodimeric form. Contrastingly, the thymus-dependent population expresses TCRαβ and bears either CD4 or CD8 in its more familiar αβ heterodimeric form (14, 15, 16, 17, 18, 19).
However, IEL from the large intestine are seldom studied despite the marked differences in both function and luminal environment between the different regions of the intestine and the development of diseases specific to the large intestine, such as ulcerative colitis and colon cancer. Accordingly, a few studies have shown that IEL of the large intestine have a different phenotype and function than those of the small intestine (20, 21, 22). Within the large intestine, the proportion of CD8+ cells is lower; however, they still constitute a major subset of the T cell pool, with the ratio of CD4+ to CD8+ being approximately equal. Principally, although thymus-independent T cells, characterized by their expression of CD8αα and TCRγδ, predominate in the small intestine, they are much less abundant in the large intestine (20, 21, 22). IEL from the large intestine have much less cytolytic activity in vitro than IEL from the small intestine (20). Furthermore, although similar patterns of IFN-γ production are seen, more of the type 2 cytokines, IL-4 and IL-5, are produced by IEL of the large intestine (21, 23). Therefore, a pronounced regional specialization of epithelial T cells is found in the gut.
Given that T. muris forms syncitial tunnels within the epithelium of the cecum and proximal colon (24), IEL are especially close to the parasites and their Ag. Therefore, IEL may play a major role in the immune response to and, ultimately, the elimination of T. muris. Because Th cells are essential for the expulsion of T. muris, we hypothesize that these cells migrate to the large intestine in resistant strains of mouse in temporal association with the expulsion of worms. In susceptible strains of mouse, the accumulation of inappropriate subsets of leukocytes in the large intestine might underlie their inability to expel the parasite, leading to the development of a chronic infection. This study characterizes large intestinal IEL throughout the infection of resistant and susceptible mice with T. muris.
Materials and Methods
Specific pathogen-free AKR and BALB/c mice were purchased from Harlan U.K. and were maintained in individually ventilated cages. In all experiments, male mice were infected with T. muris when they were 6–8 wk old. SCID mice were used to investigate the influx of macrophages into the large intestine during T. muris infection in the absence of an adaptive immune system and to examine whether isolated lymphoid follicles (ILF)-like structures could develop in the absence of lymphocytes. SCID mice were bred in isolators at the University of Manchester, and male mice used when they were 6–8 wk old. The animal studies were reviewed and approved by the Home Office and were performed under the legal requirements of the Animal (Scientific Procedures) Act (1986).
T. muris was maintained as described previously (24). Mice were infected orally with ∼150 infective eggs. Mice were killed at various time points postinfection (p.i.), and the worm burdens in the large intestine were assessed as described previously (1, 2). T. muris excretory/secretory (E/S) Ag was prepared from adult worms after a 4-h in vitro culture as described previously (24).
Mesenteric lymph nodes (MLN) were removed, and single-cell suspensions were prepared. Total MLN cells were suspended in RPMI 1640 medium supplemented with 5% FCS, 2 mmol/l l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin (all from Invitrogen Life Technologies), and 60 μmol/l monothioglycerol (Sigma-Aldrich). The cells were stimulated with 50 μg/ml T. muris E/S Ag in 48-well plates (5 × 106 cells/well) at 37°C for 24 h. The cell supernatants were harvested and stored at −20°C until they were assayed for cytokines.
Cytokines were analyzed by sandwich ELISA as described previously (25). The following mAb pairs were used: IFN-γ, R4-6A2, and XMG1.2; IL-4, 11B11, and BVD-24G.2; IL-5, TRFK5, and TRFK4 (all from BD Biosciences); IL-9, 249.2 (E. Schmitt, University of Mainz, Mainz, Germany), and DC9302C12 (BD Biosciences); and IL-12 p40, C15.6 (G. Trinchieri, Schering-Plough, Dardilly, France), and C17.8 (BD Biosciences). The detection Ab were biotinylated, and a streptavidin-peroxidase (Roche) system was used in conjunction with the substrate ABTS (Sigma-Aldrich). The samples were quantified using recombinant murine cytokine standards (R&D Systems). The plates were read at 405 nm.
Parasite-specific Ab ELISA
Serum was assayed by capture ELISA for T. muris-specific IgG1 and IgG2a as described previously (26). Briefly, Immulon 96-well plates (Thermo Electron) were coated with 5 μg/ml T. muris E/S Ag and incubated with serum diluted through eight serial 2-fold dilutions from 1/20 to 1/2560. Parasite-specific Ig was detected using either biotinylated anti-murine IgG1 (Serotec) or IgG2a (BD Biosciences).
Isolation of IEL
IEL were isolated by an accepted modification of the method described by Davies and Parrott (27). Briefly, large intestines (cecum and ∼6 cm of proximal colon) were removed, and macroscopically visible lymphoid aggregates on the cecum were cut off and dispensed with. Fat and connective tissues were removed, and the large intestines were opened longitudinally, then washed twice, to remove the feces, in calcium- and magnesium-free HBSS containing 2% FCS (at 4°C). The intestinal tissue from 10 mice was pooled and then cut into 1-cm pieces. This tissue was placed in 50-ml tubes and washed three times in HBSS containing 2% FCS at 4°C. The tissue was transferred to 25-cm3 tissue culture flasks and incubated at 37°C in HBSS containing 10% FCS, 0.2 mmol/l EDTA, 1 mmol/l DTT, 100 U/ml penicillin, and 100 μg/ml streptomycin. After 20 min, the flasks were shaken vigorously for 30 s, and the supernatant containing the IEL was separated from the tissue fragments using a stainless steel sieve. The supernatant was collected and put on ice, the tissue fragments were retuned to the flasks, and the process was repeated. After this process, the tissue pieces were examined microscopically to ensure that the epithelium had been removed and that the characteristic folds and ridges of the lamina propria were still intact. The epithelial cell suspensions from both incubations were pooled, washed, and suspended in RPMI 1640 at 4°C, then passed through nylon wool columns. The cell suspension was collected and suspended in 44% Percoll, which was layered on top of 67.5% Percoll and centrifuged at 600 × g for 20 min at 4°C. The IEL were collected from the interface between the Percoll gradients and prepared for phenotypic analysis by flow cytometry.
IEL were washed in PBS containing Dulbecco’s A and B salts, 0.1% sodium azide, and 2% FCS. Triple staining was performed on samples of 1 × 106 cells using a combination of the following Abs: anti-CD3-PE, biotinylated anti-CD25 used in conjunction with streptavidin-TriColor (BD Biosciences), and one of the following FITC-conjugated Abs: anti-CD4, anti-B220, anti-CD30 (Serotec), anti-CD69, or anti-CD103. Alternatively, triple staining was conducted using anti-CD8α-PE or anti-TCRβ-PE, anti-CD8β-FITC or anti-TCRγδ-FITC, and biotinylated anti-CD25 used in conjunction with streptavidin-TriColor (BD Biosciences). Appropriate isotype controls of irrelevant specificity (rat IgG2a-PE, rat IgG2b-PE, rat IgG2a-FITC, rat IgG2b-FITC, hamster IgG-FITC, and biotinylated rat IgG2a) were included. All Abs were obtained from BD Biosciences unless otherwise stated. All cells were stained for 30 min in the dark on ice and then fixed by the addition of 2% formaldehyde in PBS, 0.1% sodium azide, and 2% FCS. The data were acquired on a FACSCalibur flow cytometer and analyzed using CellQuest Pro software (both from BD Biosciences).
Mice were killed at various time points p.i. with T. muris. Age-matched, uninfected control mice were killed on day 21 p.i. Approximately 6 mm of the proximal colon (juxtaposed to the distal cecum) was removed, trisected, and carefully positioned in OCT embedding medium (R. A. Lamb). The tissue was snap-frozen in liquid nitrogen-chilled isopentane (BDH-Merck), and 6-μm sections were cut using a cryomicrotome. The tissue was air-dried for 1 h to maximize its adhesion to gelatin-coated microscope slides, then fixed using 4% paraformaldehyde (Sigma-Aldrich) in PBS for 10 min at 4°C. Slides were washed in PBS, and endogenous peroxidase activity was quenched using 0.064 mg/ml sodium azide, 1.5 U/ml glucose oxidase, and 1.8 mg/ml d-glucose (Sigma-Aldrich) in PBS for 20 min at 37°C. After another wash in PBS, nonspecific binding sites in the sections were blocked using 10% normal rat serum (Sigma-Aldrich) in PBS for 1 h at room temperature. Endogenous avidin and biotin binding sites were blocked using a commercial kit according to the manufacturer’s instructions (Vector Laboratories). The sections were incubated at room temperature for 1 h with one of the following rat anti-mouse biotinylated mAb: anti-CD4 (5 μg/ml; BD Biosciences), anti-CD8α (10 μg/ml; BD Biosciences), anti-B220 (10 μg/ml; BD Biosciences), anti-CD11b (5 μg/ml; BD Biosciences), anti-F4/80 (5 μg/ml; Caltag Laboratories), or anti-β7 integrin (10 μg/ml; BD Biosciences). Alternatively, a number of sections were incubated in parallel with the appropriate biotinylated isotype control Abs (BD Biosciences). A Vectastain Elite avidin-biotin-peroxidase complex kit, followed by a 3,3′-diaminobenzidine chromagen kit, were then used according to the manufacturer’s instructions (Vector Laboratories). The sections were counterstained in Harris’s hematoxylin solution and mounted in Aquamount aqueous mounting medium (BDH-Merck). The number of positively stained cells per 20 crypt units was assessed in triplicate by light microscopy after randomization and blinding.
In vivo labeling and in situ immunohistochemical visualization of proliferating lymphocytes
Mice were injected i.p. with 10 mg of BrdU, which is taken up by proliferating cells during the S phase of the cell cycle. After 40 min, the mice were killed, and the detection of nuclei that had incorporated BrdU was performed by immunohistochemistry using an anti-BrdU mAb (Mas 250b; Harlan Sera Laboratories) as described previously (28).
Statistical analysis was performed by ANOVA and Tukey’s post-test (using the statistical package GraphPad PRISM 3.0).
BALB/c mice are resistant to T. muris and mount a Th2 response, whereas AKR mice are susceptible to infection and mount a Th1 response
After infection with T. muris, BALB/c mice expelled the majority of the worms from the large intestine before day 21 p.i. and were free of worms by day 35 p.i. In contrast, AKR mice failed to expel the worms and were chronically infected with T. muris (Fig. 1,A). MLN cells from uninfected mice and mice infected with T. muris were stimulated in vitro with T. muris E/S Ag. The cells from infected BALB/c mice produced the Th2 cytokines IL-4, IL-5, and IL-9, whereas AKR mice displayed a Th1-skewed, Ag-specific cytokine response, characterized by higher levels of IFN-γ and IL-12 p40 (Fig. 1,B). Furthermore, the Ag-specific Ab produced by BALB/c mice in response to infection were predominantly IgG1, in contrast to AKR mice, in which high levels of IgG2a were detected (Fig. 1 C). Taken together, these data confirmed that the immune response to T. muris was Th2 and Th1 dominated in resistant BALB/c mice and susceptible AKR mice, respectively.
Reduction in yield of IEL during the infection
Using accepted standard methods, the number of IEL extracted from the large intestine of one uninfected mouse was typically in the range of 1 × 106 to 1.5 × 106 (Fig. 1,D). However, p.i., the yield of IEL decreased (Fig. 1,D) in temporal association with the expulsion of worms from the large intestine (Fig. 1 A). The lowest yield of IEL from BALB/c mice occurred on days 14 and 21 p.i., when the worms were being actively expelled. The IEL yield gradually returned to normal as the mice became free of infection. In AKR mice, a progressive reduction in the yield of IEL was noted as the chronic infection developed. The expulsion of worms or the development of a chronic infection is associated with enteropathy in the large intestine, including crypt hyperplasia, goblet cell hyperplasia, and the hypersecretion of mucus (24). This appears to interfere with our method of IEL extraction from the large intestine, leading to an artificially low yield p.i. The percentage of IEL expressing CD103 (αEβ7 integrin) was ∼85%, and this was unaffected by infection (data not shown).
The number of CD4+ IEL increased p.i.
Using flow cytometry, the percentage of IEL exhibiting a Th cell phenotype (CD3+CD4+) was ∼5% in both BALB/c and AKR mice (Fig. 2,A). However, p.i., dynamic changes in the percentage of CD3+CD4+ IEL occurred that differed between the two strains of mouse. In BALB/c mice, the percentage of CD3+CD4+ IEL increased gradually, and at its peak (21 days p.i.) had risen by >2-fold (Fig. 2,A). There followed a decline in the percentage of CD3+CD4+ IEL in BALB/c mice, approaching preinfection levels by 35 days p.i. (Fig. 2,A). In contrast, the percentage of CD3+CD4+ IEL in AKR mice continued to rise throughout the infection, reaching a 3-fold increase by 35 days p.i. (Fig. 2 A).
CD4 detection by immunohistochemistry allowed the numerical quantification of CD4+ IEL and LPL within the large intestine. Dynamic changes in the number of CD4+ IEL (Fig. 2, B and C) and LPL (Fig. 2, B and D) occurred p.i. with T. muris, revealing differences between the strains of mouse. Initially, in BALB/c mice, the number of CD4+ IEL increased, reaching a peak 21 days p.i., then declined at later time points. In contrast, the number of CD4+ IEL in AKR mice continued to increase throughout the infection (Fig. 2, B and C). These strain-specific patterns of change in CD4+ IEL during infection are essentially similar to those demonstrated by flow cytometry (Fig. 2,A). The number of CD4+ LPL increased in both strains of mouse p.i. (Fig. 2, B and D). There were roughly 15 times more CD4+ cells in the lamina propria than in the epithelium regardless of infection with T. muris (Fig. 2, C and D).
The number of CD8+ IEL increased p.i.
The expression by IEL of both isoforms of CD8 was investigated using flow cytometry. This revealed major differences between the strains of mouse in the relative abundance of both CD8αα+ and CD8αβ+ IEL in the large intestine. Twice the percentage of CD8αα+ IEL were detected in AKR mice compared with BALB/c mice before infection (Fig. 3,A). Conversely, although a significant proportion of IEL expressed CD8αβ in BALB/c mice, the percentage of these cells was negligible in AKR mice (Fig. 3,A). After infection, CD8α+ IEL were more abundant in AKR mice than in BALB/c mice (Fig. 3 B).
Immunohistochemical detection of the CD8 α-chain (which is expressed by all CD8+ cells) enabled all CD8+ IEL and LPL within the large intestine to be quantified. Paradoxically, although CD8+ IEL were found in BALB/c mice by flow cytometry (Fig. 3, A and B), virtually no CD8+ IEL or LPL were detected by immunohistochemistry in uninfected BALB/c mice (Fig. 3,D). After infection, there was a limited influx of CD8+ IEL and LPL into the large intestine in BALB/c mice (Fig. 3, C–E). Compared with BALB/c mice, CD8+ IEL and LPL were relatively abundant in uninfected AKR mice, and the number of these cells was considerably greater p.i. (Fig. 3, C–E). Approximately 20 times more CD8+ leukocytes were found in the lamina propria than in the epithelium in uninfected AKR mice (Fig. 3, D and E).
Analysis of TCR subunits
There were ∼4 times more TCRαβ+ IEL than TCRγδ+ IEL in uninfected mice, as determined by flow cytometry (data not shown). In BALB/c mice, the percentages of TCRαβ+ and TCRγδ+ IEL were 59 and 15%, respectively. No clear pattern of change over time p.i. to TCRαβ+ or TCRγδ+ IEL was evident in either strain of mouse (data not shown).
The number of B220+ IEL increased p.i.
The percentages of T cells (CD3+B220−) and B cells (B220+CD3−) in the IEL compartment were evaluated by flow cytometry. T cells were always more abundant than B cells (data not shown). In BALB/c mice, the percentage of B220+CD3− IEL increased 21 days p.i. (Fig. 4 A). An increase in the percentage of B220+CD3− IEL was associated with a decrease in the percentage of CD3+B220− IEL (data not shown).
B cells in the large intestine were also examined by immunohistochemistry. There were ∼4 times more B220+ IEL than LPL in uninfected mice (Fig. 4, C and D). After infection, there were greater numbers of B220+ IEL and LPL (Fig. 4, B–D). The patterns of this change over time broadly mirrored those found by flow cytometry (Fig. 4,A). That is to say, the number of B cells initially increased p.i. in BALB/c mice and subsequently returned to normal levels, whereas in AKR mice the highest levels of B cells occurred at later time points (Fig. 4, C and D). However, there was also some disparity in the data between the two methods of B cell analysis. Most strikingly, by immunohistochemical analysis, there were significantly more B220+ IEL in AKR mice than in BALB/c mice 21 days p.i. (Fig. 4, B and C), although by flow cytometry the converse was found (Fig. 4 A).
Large influx of macrophages into the large intestine p.i.
Because no macrophage markers are entirely specific, two such markers (F4/80 and CD11b) were used to investigate more clearly the influx of macrophages into the large intestine by immunohistochemistry. In practice, there was little difference between the two methods of analysis. In uninfected mice, F4/80+ and CD11b+ IEL were scarce (Fig. 5, A, B, D, and E), whereas F4/80+ and CD11b+ LPL were relatively abundant (Fig. 5, A, C, D, and F). After infection, there was a significant increase in the number of F4/80+ and CD11b+ IEL (Fig. 5, A, D, B, and E). There was a more striking increase in the number of F4/80+ and CD11b+ cells in the lamina propria p.i., uncovering major differences between the two strains of mouse. In BALB/c mice, the number of F4/80+ and CD11b+ LPL reached a peak 21 days p.i., declining at later time points. In contrast, the numbers of F4/80+ and CD11b+ LPL in AKR mice continued to increase throughout the infection (Fig. 5, C and F). Interestingly, there were twice as many F4/80+ and CD11b+ cells in the lamina propria of BALB/c mice than in AKR mice 21 days p.i. (Fig. 5, A, C, D, and F). In uninfected SCID mice there were 52 ± 11 F4/80+ LPL/20 crypts (data not shown). After infection, the number of F4/80+ LPL continued to increase in SCID mice (139 ± 12 F4/80+ LPL/20 crypts after 21 days and 192 ± 17 F4/80+ LPL/20 crypts after 35 days; data not shown), resembling that found in AKR mice (Fig. 5 C).
Expression of activation markers by IEL does not change p.i.
There were more CD25+ IEL in AKR mice than in BALB/c mice (Fig. 6, A–C). Only a small proportion of CD3+ IEL expressed CD25 in either strain of mouse (Fig. 6,A); accordingly, few CD4+ IEL expressed CD25 (Fig. 6,B). The majority of CD25+ IEL in AKR mice were B cells (Fig. 6,C). In contrast to BALB/c mice, in which few B220+ IEL expressed CD25, in AKR mice most B220+ IEL expressed CD25 (Fig. 6,C). A high proportion of CD3+ IEL expressed the activation marker CD69 (Fig. 6,D). The infection of mice by T. muris caused no discernable difference in the percentage of CD25+ or CD69+ IEL (Fig. 6). Less than 0.5% of the IEL expressed CD30 (an early marker of activation) in uninfected mice or at any time point p.i. (data not shown).
Lymphoid follicles filled mainly with B cells are present in large intestine
Numerous pronounced follicular structures were discovered in the large intestine of both AKR and BALB/c mice. These follicles were comprised primarily of closely packed B cells interspersed by small clusters of CD4+ T cells. CD8+ T cells were much less common, but could occasionally be found at the edge of the follicles. Macrophages were found at the marginal zone of the follicles, and occasionally, individual macrophages could also be detected more centrally. Some cells around the outside of the follicles expressed α4 integrin (Fig. 7,A). BrdU was incorporated by leukocytes in the core of the follicles, suggesting a central zone of proliferating B cells (Fig. 7,B). Infrequently, structures resembling follicles were also found in the large intestine of infected SCID mice, although they consisted of neither B cells nor macrophages (Fig. 7,C). In some mice, multiple follicles were found in the gut sections (Fig. 7,D). In AKR mice, but not in BALB/c mice, there was a significant increase, per mouse in the number of follicles at later time points p.i. (Fig. 7,E). However, the follicles tended to be larger in BALB/c mice than in AKR mice, particularly p.i. (Fig. 7 F).
Typical of previous investigations, only ∼1 × 106 IEL were obtained from the large intestine of an individual uninfected mouse. Consequently, to analyze their phenotype comprehensively, it is common practice to pool IEL from several individuals (20, 21, 22). A considerable percentage of IEL extracted from the large intestine expressed the classical IEL marker CD103 (αEβ7 integrin), confirming the reliability of our preparation technique. Some differences in the phenotype of IEL were apparent between uninfected BALB/c and AKR strains of mouse. CD8+ IEL were more abundant in AKR mice than in BALB/c mice. Furthermore, although CD8αα+ IEL were found in both strains of mouse, CD8αβ+ cells were found only in BALB/c mice. Indeed, there are numerous examples in the literature of phenotypic differences between strains (20, 21, 22).
Regardless of interstrain differences, we were able to compare the present study with previous investigations because BALB/c mice are routinely used. Consistent with previous descriptions of IEL isolated from the large intestine (20, 21), ∼76% were CD3+ T cells, of which 80% were TCRαβ+ and 20% were TCRγδ+. However, a discrepancy with previous reports was evident when the T cell subsets were subjected to a more detailed analysis. In the present study the proportions of CD3+ IEL expressing CD4 and CD8 were 7 and 51%, respectively, and by deduction, the remaining T cells (∼42%) were double negative (CD4−CD8−). Others estimate a higher proportion of CD4+ T cells (between 32 and 72%), with the ratio of CD4+ to CD8+ being approximately equal (20, 21, 22). Indeed, the relative abundance of CD4+ T cells from the large intestine is thought to distinguish them from T cells of the small intestine, where CD8+ cells predominate (20, 21, 22). Furthermore, contrary to previous reports, CD8αα+ cells were found to be more plentiful than CD8αβ+, again resembling the phenotype commonly associated with the small intestine (17, 21, 22). Although double-negative (CD4−CD8−) cells constituted a major T cell subset in the present study, previous reports suggest they are less prevalent (from 1 to 27%) (20, 21, 22). Therefore, in this study we reproducibly define a large intestinal T cell phenotype that contrasts with previous descriptions. It was vital in the present study to use only mice that were free of gastrointestinal infections before infection with the cecum-dwelling nematode T. muris. However, laboratory mice are often chronically infected with the gut-dwelling nematodes Aspiculuris tetraptera and Syphacia obvelata (29), and as we discuss later, infection does alter the balance of different IEL subsets. Thus, our results may differ from those of previous reports in part due to the use of specific pathogen-free laboratory mice housed in individually ventilated cages. Hence, this study challenges previous descriptions of IEL isolated from the large intestine (suggesting that they are phenotypically similar to IEL from the small intestine), and therefore, fundamentally, the potential for CD8-mediated cytotoxicity in the large intestine is greater than described previously.
The number of IEL extracted from the large intestine p.i. appeared to be affected by infection-associated gut enteropathy. It is therefore misleading and does not reflect the actual number of IEL in the large intestine p.i. However, a reliable account of IEL numbers is given by histological examination, demonstrating the marked accumulation of IEL in the large intestine p.i., uncovering differences between the contrasting stains of mouse and reinforcing the value of using both flow cytometry and immunohistochemistry. In BALB/c mice, the number of IEL increased (peaking at ∼21 days p.i.), then reverted toward normal levels, corresponding with the kinetics of worm expulsion. In contrast, the number of IEL in AKR mice increased and remained high as the infection progressed.
Th cells are known to play a pivotal role in the mechanism of T. muris expulsion, because the depletion of CD4+ cells confers a susceptible phenotype to resistant strains of mouse (30). As we confirm, the generation of a Th2 response is essential for the expulsion of worms (3, 4, 5, 6). A locally acting mechanism for Th2 cells has been postulated, yet no previous studies have shown the migration of CD4+ cells into the large intestine. Importantly, the present study demonstrates for the first time that in resistant mice exhibiting a Th2 response, CD4+ Th cells do indeed accumulate in the epithelium of the large intestine around the time of worm expulsion. In susceptible mice that mount a Th1 response, the number of CD4+ Th cells increases more gradually and is greatest during the chronic phase of infection. Recently, several studies have suggested various potential effector mechanisms by which T. muris may be expelled from the gut. These theories include an increased rate of epithelial cell turnover (31) and the release of factors by goblet cells that may impair chemotaxis of the parasite (32). Nevertheless, both these potential mechanisms depend upon the secretion of Th2 cytokines in the large intestine. Because the present study suggests that Th2 cells migrate to the large intestine at the time of worm expulsion, this bridges the gap in our knowledge between the well-characterized afferent immune responses to the putative efferent immune effector mechanisms of worm expulsion.
In contrast to resistant mice, a large population of CD8+ cells infiltrated the mucosa of the large intestine in susceptible mice p.i. However, a recent study in our laboratory shows that the depletion of CD8+ cells in susceptible mice fails to influence the development of a chronic infection (33). Therefore, although they are not essential for the development or maintenance of a chronic infection, the sheer magnitude of CD8+ cell migration into the gut underlines the inability of susceptible mice to mount an appropriate protective immune response to the parasite. A much lower number of CD8+ IEL was found in resistant BALB/c mice p.i. These cells were detected less frequently by immunohistochemistry than by flow cytometric analysis of isolated IEL, perhaps indicating a difference in the sensitivity of the contrasting methods.
The accumulation of B cells in the gut was also noted p.i. B cells were shown by immunohistochemistry to be more numerous in susceptible mice. However, in resistant mice, especially 21 days p.i., the percentage of B cells in the isolated IEL (as shown by flow cytometry) was considerable, somewhat contradicting the immunohistochemical findings. During their extraction from the large intestine, the contamination of IEL by B cells from ILF is inevitable, because these follicular structures are intimately associated with the epithelium. Furthermore, we suggest (with relevance to all previous studies of isolated IEL) that although LPL can be excluded from the preparations, a degree of contamination from ILF is to be expected. Because ILF tended to be larger in BALB/c mice than in AKR mice (particularly p.i.), this is the most likely explanation for the abundance of B cells in IEL preparations from resistant mice. Recently, other authors have considered it inevitable that B cells from germinal centers contaminate IEL suspensions (13). Therefore, the present study highlights the importance of immunohistochemistry as a vital tool to address this problem.
It is difficult to describe the activation state of IEL, because some markers of activation were widely expressed by IEL, and others were expressed by only a small minority of cells; the majority of CD3+ IEL expressed CD69, whereas CD25, in accordance with other studies (20), was expressed by a very small percentage of T cells. Because CD3+ IEL isolated from the large intestine have been shown to express CD25 after TCR stimulation in vitro (20), it is perhaps surprising that there was no change in the frequency of these cells p.i. with T. muris. Interestingly, CD25 was expressed most notably by B cells in AKR mice. A distinct population of CD25+CD4+ cells, namely, regulatory T cells, is thought to play a role in the persistence of infection to the parasite Leishmania major (34). However, a role for regulatory T cells in the immune response to T. muris seems unlikely, because <1% of the IEL were CD25+CD4+.
There was a sizeable influx of leukocytes into the lamina propria of the large intestine p.i. The phenotype of LPL differed markedly from that of IEL. CD4+ cells were up to 10 times more abundant in the lamina propria than in the epithelium. Conversely, CD8+ and B220+ LPL were less numerous than CD8+ and B220+ IEL. Macrophages made up a large fraction of the LPL p.i., whereas they were rarely encountered in the IEL population. This work confirms that IEL and LPL are distinct components of the GALT. Intriguingly, the migration of macrophages into the lamina propria reached a peak around the time of worm expulsion, at which point there were approximately twice as many macrophages in resistant mice than in susceptible mice. Greater numbers of macrophages were observed p.i. in the lamina propria of SCID mice, suggesting that the accumulation of macrophages in the large intestine is in part an innate immune response to the parasite. A recent study in our laboratory shows that mice devoid of the macrophage chemokine CCL2 fail to expel T. muris, and this is associated with fewer macrophages in the lamina propria and an altered Th1/Th2 cytokine balance (35). Taken together, this suggests a potential role for macrophages in the mechanism of worm expulsion from the gut. Future work will investigate the phenotype and role of lamina propria macrophages in this context.
Lymphoid structures known as ILF have been identified in both small and large intestines of humans (36, 37); more recently, they have been discovered and characterized, in some detail, in the small intestine of mice (38, 39). They are composed of a large B cell area, including a germinal center, and like Peyer’s patches, the epithelium overlying the follicles contains M cells, suggesting that they are inductive sites for local IgA responses. Although Hamada et al. (38) and Dohi et al. (40) discovered ∼50 ILF in the large intestine of normal mice in addition to ∼10 colonic patches, they neither described in detail nor showed any photographic examples of these structures (38, 40). We found numerous pronounced follicular structures, invisible from the serosal or mucosal surface, in the large intestine of both AKR and BALB/c mice. Consisting of B cells interspersed with CD4+ T cells and having a central zone of rapidly proliferating cells, these lymphoid aggregations are analogous to ILF. Therefore, the present study extends our knowledge of the GALT, illustrating ILF in the large intestine of mice for the first time. While this manuscript was in revision, it was shown that although colonic patches and ILF of the large intestine both contain M cells, they have a distinct structure and organogenesis (41). Lymphoid aggregations, equivalent in size to ILF, but devoid of B cells, were found in the large intestine of SCID mice. This suggests that the organogenesis of ILF depends on neither B cells nor T cells. Indeed, similar structures have been identified in the small intestine of other immunodeficient mouse models, such as athymic nude mice (nu/nu) and RAG-2 knockout mice (38). It has been shown that ILF are formed in the small intestine in response to normal gut flora in the cecum (39). In the present study there was a significant increase p.i. with T. muris in the number of large intestinal ILF in AKR mice. These findings provide evidence for the organogenesis of ILF in response to luminal stimuli. In our model of T. muris infection, it is intriguing that ILF formation occurs specifically in susceptible AKR mice. It may arise due to the chronic infection of these mice, where the epithelium is exposed for a longer duration to T. muris Ag in the gut. This may represent a diversion of the local immune response away from an effective Th cell-dominated mechanism toward an inappropriate response dominated by B cells. The role of ILF in the pathogenesis of gastrointestinal infections and diseases should be investigated in the future.
In conclusion, in this study we characterize for the first time the accumulation of cells into the epithelium and lamina propria of the large intestine during infection of mice with T. muris. There were marked differences in this respect between resistant and susceptible strains of mouse. In resistant BALB/c mice, the local inflammation was dominated by CD4+ IEL and F4/80+ LPL at the time of worm expulsion, in contrast to susceptible AKR mice, where CD8+ IEL and F4/80+ LPL were predominant during the chronic phase of infection. Therefore, this study reveals the local immune responses underlying the expulsion of worms or the persistence of chronic infection, respectively. Furthermore, we describe and illustrate ILF in the large intestine of mice and demonstrate the organogenesis of these structures in response to T. muris infection.
We are grateful to Neil E. Humphreys for helpful discussions and advice.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Wellcome Trust Grant 044494/Z (to M.C.L., L.V.B., and K.J.E.).
Abbreviations used in this paper: LPL, lamina propria leukocyte; E/S, excretory/secretory; IEL, intraepithelial lymphocyte; ILF, isolated lymphoid follicle; MLN, mesenteric lymph node; p.i., postinfection.