Main

TGF-β is a multifunctional cytokine involved primarily in regulation of cell proliferation and differentiation, stimulation of extracellular matrix synthesis and deposition, and modulation of inflammatory and immune responses (1, 2). In the intestinal mucosa, TGF-β may be involved in the initiation of immune responses by regulation of MHC antigen expression and presentation (3). The major effects of TGF-β on the mucosal immune system seem to be regulation of IgA responses and induction of oral tolerance. For example, TGF-β induces B-cell isotype switching to IgA in murine B cells stimulated with lipopolysaccharide (4). T cells isolated from mesenteric lymph nodes, Peyer's patches, and lamina propria of orally tolerant animals had a 5–10-fold increase in their production of TGF-β together with increased amounts of IL-4 and IL-10 compared with control animals (5).

Several studies have shown that TGF-β is present in milk of various species (6–8). The total concentration of TGF-β in human milk was high in colostrum and relatively low in late lactation (6). ELISA of rat milk found no TGF-β1, whereas TGF-β2 was high in early milk and progressively declined toward weaning (8). TGF-β1, 2, and 3 have been found in normal murine intestinal tissues (9). We showed that TGF-β1 expression in the postnatal rat small intestine was very low in the first week of life and increased toward weaning by d 19 (8), indicating a reciprocal relationship between levels of TGF-β in milk and endogenous production of TGF-β in the neonatal gut. In addition, Letterio et al. (10) showed that labeled TGF-β was distributed intact in various tissues after oral administration to TGF-β1 null mice.

Three distinct TGF-β receptors have been characterized on the cell surface. TβRIII (also known as betaglycan) is a transmembrane proteoglycan containing a short cytoplasmic domain; it may not be directly involved in signaling but binds and presents TGF-β to signaling receptors (11). TβRII and I are transmembrane serine/threonine kinases, which are essential for TGF-β signaling. TβRII determines the ligand specificity, whereas TβRI interacts with TβRII and may not have a ligand-binding specificity by itself (12). TGF-β1 interacts with TβRI and II, forming a high-affinity signaling receptor complex; this interaction does not require betaglycan (13). In contrast, TGF-β2 seems to depend on betaglycan for high-affinity binding to TβRI and II. TβRI and II can distinguish between TGF-β isoforms: they bind TGF-β1 and β 3 far better than they bind TGF-β2 (14, 15). betaglycan, however, binds to TGF-β1 and β2 with equal affinity and presents them to TβRII, subsequently forming the high-affinity signaling receptor complex with TβRI (13). Indeed, by promoting TGF-β2 access to TβRII, betaglycan counteracts the intrinsic low affinity of the signaling receptors toward this isoform, thus rendering cells as sensitive to this isoform as they are to TGF-β1.

TGF-β receptors have been found in a number of intestinal epithelial cell lines as well as in fetal and adult intestinal tissues (16–19). Given the progressive changes in TGF-β levels in milk and endogenous production of TGF-β in the neonatal gut (8), a developmental change in TGF-β receptor expression in the gut may also occur from preweaning to postweaning. However, to date, no studies have assessed TGF-β receptor expression in the postnatal gut and gut-associated lymphoid tissue. The aims of the present study were 2-fold:a) to study the ontogeny of TGF-β receptor expression in the postnatal small intestine;b) to identify which cells express TGF-β receptor(s). Analysis based on these findings should provide additional insight into the role of maternal milk TGF-β in the development of the postnatal intestine and intestinal mucosal immune system.

METHODS

Tissue preparation.

Rat tissue collection was performed with the approval of the Women's and Children's Hospital Animal Ethics Committee as well as the Animal Ethics Committee of the University of Adelaide, Australia. Each litter of Hooded Wistar rat pups was culled to six and weaned at d 21 of age. Jejunum and ileum tissue samples were collected at d 3, 10, 14, 21, and 28 postpartum (four to six pups per age group). Rat pups were killed by decapitation. The intestinal tract was immediately excised and placed onto an ice-cold glass slab. Two 1-cm-long segments were collected from proximal jejunum and ileum, respectively, embedded in optimal cutting temperature (OCT) compound (Miles Inc., Elkhart, IN), frozen in liquid nitrogen, and stored at −70°C until sectioning, 8-μm-thick cryostat sections were cut and mounted onto gelatin-coated slides. As described elsewhere (20, 21), the sections were air dried, fixed in 100% acetone for 10 min at room temperature, then dried and stored at −20°C until use.

Immunofluorescence single labeling.

Before staining, the tissue sections were washed with Tris-buffered saline (TBS), and preincubated with 3% BSA (Sigma Chemical Co., St. Louis, MO) for 30 min to block nonspecific binding of the antibodies. The sections were then incubated with either rabbit anti-human Tβ RI (sc-399, Santa Cruz Biotechnology, Santa Cruz, CA) or Tβ RII (sc-1700, Santa Cruz Biotechnology), or goat anti-human Tβ RIII (AF-242-PB, R&D Systems, Minneapolis, MN) polyclonal antibodies diluted to working concentration (1 μg/mL) at room temperature for 1 h. After three washes, sections were further incubated with indocarbocyanine (Cy3) conjugated F(ab′)2 fragment donkey anti-rabbit IgG (1:800) or donkey anti-goat IgG (1:100) antibodies (Jackson ImmunoResearch, PA) as appropriate for 1 h at room temperature. Sections were then washed three times with TBS and mounted in fluorescence mounting medium (Dako, CA). Control sections were incubated without primary antibodies or IgG from the same species at the same concentration as the primary antibody. The specificity of labeling was evaluated by replacing the primary antibodies above with respective antibody solutions either preincubated with a corresponding blocking peptide (Santa Cruz Biotechnology) or preabsorbed with a recombinant receptor protein (R&D Systems).

Immunofluorescence dual labeling.

Sections were stained with antibodies against Tβ RI, II, or III as above in conjunction with antibodies to rat cell surface determinants (CD) detailed below. The chosen cell markers detect a range of immune cells including T lymphocytes, B lymphocytes, and a number of antigen-presenting cells as well as T-cell activation markers. Both indirect and direct immunofluorescence techniques were used, as some of the antibodies were not available as direct conjugates. During indirect immunofluorescence, the tissue sections were coincubated with polyclonal antibodies against TGF-β receptor molecules and MAb against cell surface markers for 1 h at room temperature. The following mouse anti-rat MAb were purchased from SEROTEC (Serotec, Oxford, England): OX-8, CD8; W3–25, CD4; OX-17, MHC class II; OX-39, IL-2R; ED5, follicular dendritic cell and OX-62, veiled dendritic cell. After initial incubation, the sections were washed thoroughly in TBS/0.05% polyoxyethylenesorbitan monolaurate 20 (Tween 20, BDH Laboratory Supplies, Poole, England) and stained with either Cy3-conjugated F(ab)2 fragment donkey anti-rabbit IgG (Jackson ImmunoResearch) plus FITC-conjugated sheep anti-mouse Ig (DDAF, Silenus, Hawthorn, Australia) or Cy3-conjugated F(ab)2 fragment donkey anti-goat IgG (Jackson ImmunoResearch) plus FITC-conjugated rabbit anti-mouse Ig (Sigma Chemical Co.) preabsorbed with 1% normal rat serum (Sigma Chemical Co.) for another 60 min at room temperature.

During direct immunofluorescence, the sections were first incubated with either rabbit anti-Tβ RI or Tβ RII (Santa Cruz Biotechnology), or goat anti-Tβ RIII (R&D Systems) for 1 h at room temperature. After washing three times in TBS/0.05% Tween 20, a further 60-min incubation with FITC directly conjugated mouse anti-rat ED-1 (macrophage), OX-26 (transferrin receptor), or OX-33 (B cell) antibodies (Serotec) was applied, together with Cy3-conjugated F(ab)2 fragment donkey anti-rabbit or donkey anti-goat IgG antibodies (Jackson ImmunoResearch). The number of double-stained cells per mm2 villus mucosa was counted using a fluorescence microscope (Leica, Wetzlar, Germany) at ×400 magnification. The average cell number was based on the scores of jejunal sections obtained from four to six animals in each group. At least three sections for each piece of tissue, 8–15 different fields of view per section, were counted.

Statistics.

The nonparametric Mann-Whitney test was used for statistical analysis. A 2-sided p< 0.05 was considered significant for differences in dual-labeled cell counts between different age groups.

RESULTS

Ontogeny of TGF-β receptor expression in rat pup jejunum.

Distinctive distribution patterns of TGF-β receptor expression were observed in the postnatal jejunum. Typical immunostaining patterns for Tβ RI, II, and III are represented in Figures 1,2, and 3 and summarized for all animals in each group (n= 6) in Table 1. As seen in Figure 1, both basolateral and apical membranes of the villus and crypt epithelium expressed Tβ RI. Staining on the apical membrane increased with age, whereas, in the basolateral membrane, the staining was relatively constant in all ages assessed. Tβ RII was abundantly expressed in both upper and lower parts of the crypt at all ages (Table 1 and Fig. 2). Staining in the villi appeared after d 10, and the pattern changed with age. At d 10, the apical membrane of the villus epithelium was stained (Fig. 2, A and B); at d 14 and 28 of age, staining was more isolated on individual enterocytes (Fig. 2, C–F). Tβ RIII was broadly expressed on the goblet cells (Fig. 3, A–D), crypt (Fig 3, A and C) and villus (Fig. 3, B and D) epithelial cells, and IEL (Fig. 3, A and C); lamina propria cells (Fig. 3, A, C, and D) and muscularis (Fig. 3C) were also stained. However, Tβ RIII staining in the epithelium and lamina propria was markedly diminished at d 28 (Fig. 3, E and F). In addition, preabsorption of antibodies to respective blocking peptide or recombinant antigen resulted in the abolition of all staining (not shown).

Figure 1
figure 1

Immunofluorescence single labeling of Tβ RI in rat jejunum. Longitudinal sections were stained with rabbit anti-Tβ RI as the primary antibody and Cy3-conjugated donkey anti-rabbit F(ab′)2 antibody as the secondary antibody. Panels A and B show Tβ RI immunostained sections from 10-d-old rat pup (×400);panels C and D, from 14-d-old rat pup (×400);panels E and F, from 28-d-old rat pup (×400). V, villi (with the tip of the villus downward);→, crypts and crypt epithelium;→, staining on apical membrane of villus epithelium;→, staining on basolateral membrane of villus epithelium.

Full size image
Figure 2
figure 2

Immunofluorescence single labeling of Tβ RII in rat jejunum. Longitudinal sections were stained with rabbit anti-Tβ RII as the primary antibody and Cy3-conjugated donkey anti-rabbit F(ab′)2 antibody as the secondary antibody. Panels A and B show Tβ RII immunostained sections from 10-d-old rat pup (×400);panels C and D, from 14-d-old rat pup (×400);panels E and F, from 28-d-old rat pup (×400). V, villi (with the tip of the villus downward or rightward);→, crypts and crypt epithelium;→, staining on apical membrane of villus epithelium;→, Tβ RII-positive enterocytes.

Full size image
Figure 3
figure 3

Immunofluorescence single labeling of Tβ RIII in rat jejunum. Longitudinal sections were stained with goat anti-Tβ RIII as the primary antibody and Cy3-conjugated donkey anti-goat F(ab′)2 antibody as the secondary antibody. Panels A and B show Tβ RIII immunostained sections from 10-d-old rat pup (×400);panels C and D, from 14-d-old rat pup (×400);panels E and F, from 28-d-old rat pup (×400). V, villi (with the tip of the villus downward or rightward);→, crypts and crypt epithelium;M, staining on muscularis mucosa;→, labeled goblet cells;▴, labeled IEL;→, labeled epithelial cells;→, Tβ RIII-positive lamina propria cells.

Full size image
Table 1 Summary of TGF-β receptor I, II, and III expression in the small intestine of rats at different ages Immunofluorescence single labeling was performed as described in “Methods.” The average score (n= 6) is expressed as +++, strong positive staining on most cells; ++, moderate staining on most cells; +, staining on a few cells; −, below detectable level.
Full size table

Expression of Tβ RI, II, and III in rat pup ileum.

Tβ RI, II, and III were also abundantly expressed in the ileum, and overall distribution patterns of immunostaining in the ileum is similar to that observed in the jejunum (Table 1 and Fig. 4). Only minor variations were seen between the jejunum and ileum. For example, Tβ RI staining of muscularis in sections from animals aged d 3 was most intense at the outer muscle layers in ileum, whereas, in jejunum, the inner circular layers were more strongly stained (Fig. 4, A and B).

Figure 4
figure 4

Comparison of immunofluorescence staining of Tβ RI, II, and III in rat jejunum and ileum. Longitudinal sections were stained with either rabbit anti-Tβ RI, Tβ RII, or goat anti-Tβ RIII as the primary antibody and Cy3-conjugated donkey anti-rabbit or donkey anti-goat F(ab′)2 antibody as the secondary antibody. Left panels (A, C, and E) are sections from jejunum (×400), and right panels (B, D, and F) are sections from ileum (×400). A and B are sections from 3-d-old rat pups stained with Tβ RI. C and D are sections from 3-d-old rat pups stained with Tβ RII. E and F are sections from 21-d-old rat pups stained with Tβ RIII. V, villi (with the tip of the villus upward);M, staining on muscularis mucosa;→, labeled crypts and crypt epithelium;→, Tβ RIII-positive lamina propria cells.

Full size image

TGF-β receptor-labeled immune cells in intestinal mucosa.

Double immunofluorescence showed that a number of different types of immune cells in the lamina propria and IEL expressed Tβ RIII between d 10 and 21 of age (Table 2 and Fig. 5). Although Tβ RI and II were expressed by intestinal epithelial cells, no Tβ RI or II dual-labeled immune cells were detected in this study. Controls in which isotypic IgG was used at the same concentration as the primary antibodies were negative compared with specific antibody staining (not shown). Tβ RIII-labeled OX-33 (B cell), W3/25 (CD4 T cell;Fig. 5, A and B), and OX-62 (veiled dendritic cell;Fig. 5, C and D) along with OX-17 (MHC class II) positive cells were detected as early as d 10; dual-labeled OX-8 (CD8) positive cells were detected at d 14 and 21 of age, whereas low levels of follicular dendritic cells (ED5+) expressing Tβ RIII were detected at d 21 (Table 2). Occasionally, IEL were CD8 and Tβ RIII double positive (not shown). Overall, the number of dual-labeled cells in the mucosa increased with age from d 10 to 21; no dual-labeled cells were detected at d 3 and 28 (Table 2). We found that very few immune cells were present in the lamina propria in sections from animals aged d 3. The number of TβRIII-positive B cells, CD4 T cells, veiled dendritic cells, and MHC class II+ cells was relatively low at d 10 of age but significantly increased by d 21, p< 0.05 (Table 2). Loss of double-stained cells in sections from animals aged d 28 reflects the loss of Tβ RIII staining from the epithelium and lamina propria at this age (Fig. 3, E and F).

Table 2 Immune cells dual labeled with TGF-β receptor III in the jejunum of rats at different ages (mean ± SD cells/mm2) The average number of dual-labeled cells was scored as described in “Methods.” Abbreviations: FDC, follicular dendritic cells; VDC, veiled dendritic cells; IL-2R, interleulin-2 receptor; TFR, transferrin receptor; ∼, not detected. Data are expressed as mean ± SD cells/mm2. * Significantly different from the 10-d-old group (p < 0.05).
Full size table
Figure 5
figure 5

Immunofluorescence dual-labeled sections. Sections were first incubated with mouse anti-rat cell surface markers and goat anti-human Tβ RIII antibodies, then FITC-conjugated anti-mouse and Cy3-conjugated anti-goat antibodies were applied as the secondary antibodies. Cell marker and Tβ RIII staining were detected with a 520-nm filter (green) and a 546-nm filter (red), respectively. (A) A section from 14-d-old rat pup labeled with a marker for CD4 T cell (×400). (B) A section identical to (A) labeled with TβRIII (×400). (C) A section from 21-d-old rat pup labeled with a marker for OX-62 veiled dendritic cell (×400). (D) A section identical to (C) labeled with Tβ RIII (×400). →, CD4/Tβ RIII double-staining cells;→, OX-62/TβRIII double-positive cells.

Full size image

DISCUSSION

The present study was carried out to assess the ontogeny of TGF-β receptor expression in the postnatal small intestine. Our results revealed that both Tβ RI and II were expressed in the intestinal epithelium postnatally. Tβ RIII was coexpressed with TβRI and II in the epithelium before weaning but diminished from the epithelium postweaning. The latter is in agreement with the observation that betaglycan (Tβ RIII) is less abundant in adult tissues compared with fetal tissues (22). As an accessory receptor, Tβ RIII may bind and present TGF-β to homo-oligomeric Tβ RII, which is a constitutively active serine/threonine kinase. The TGF-β /Tβ RII complex then recruits TβRI into the complex, after which Tβ RI is phosphorylated by Tβ RII, hence becoming activated, and propagates signals (23, 24). Tβ RIII binds TGF-β1 or TGF-β2 indiscriminately and greatly enhances the otherwise minimal affinity for TGF-β2 binding to the signaling Tβ RII (25). The coexistence of Tβ RI, II, and III in the gut epithelium during the suckling period could allow milk-derived TGF-β, especially TGF-β2, to be functionally active via interaction with the ligand-binding TβRIII and signaling TβRII and I. Expression of Tβ RIII in the suckling rat intestine may reflect a physiologic necessity for milk-derived TGF-β2, the predominant isoform in rat milk (8), to access signaling receptors in the gut. Lack of expression of Tβ RIII postweaning may correlate with the onset of endogenous TGF-β1 production by enterocytes at weaning (8). Norgaard et al. (26) reported that betaglycan mRNA was rapidly down-regulated, whereas Tβ RI and II mRNA were slightly increased after 24 h of treatment with TGF-β1 by small-cell lung cancer cell lines, suggesting that TGF-β1 regulates the expression of its receptors, in particular betaglycan. Unlike TGF-β2, TGF-β1 can directly interact with Tβ RII, which then leads to signal transduction in conjunction with Tβ RI (13). Once endogenous production of TGF-β1 occurs, Tβ RIII may become redundant in terms of ligand binding and presentation.

Distinctive distribution patterns of Tβ RI, II, and III in the gut epithelium were also observed related to the postnatal ages. Tβ RI was expressed on both apical and basolateral membranes of the whole epithelium, and expression on the apical membrane increased with age, whereas Tβ RII expression appeared on the apical membrane of the epithelium at d 10 and became scattered on individual enterocytes along the villi after d 14. Tβ RIII was broadly expressed across the epithelium including goblet cells, epithelial cells, and IEL before d 28. Tβ RI, II, and III coexpressed in the gut epithelium but were not necessarily situated side by side on each single cell. This is consistent with the report that no physical association of Tβ RI with Tβ RIII has been observed (27). Although TβRIII and II can be coimmunoprecipitated when bound to TGF-β (13, 25), double-labeling immunofluorescence experiments of cell surface receptors indicated that Tβ RIII and II are not detectably associated with each other (27). As suggested by others, the hetero-oligomerization between Tβ RIII and II may only be transient and restricted to a small percentage of the receptors (28), and differential regulation of Tβ RII and I expression and their signaling activities may provide the cell with a strategy to modulate TGF-β responsiveness (27).

TGF-β is a potent immunoregulator in the mucosal immune system (3–5). It is expected that TGF-β target cells would express functional TGF-β receptors. We examined various types of immune cells in the gut-associated lymphoid tissue, but none of these cells were dual-labeled with Tβ RI or II, even though Tβ RI- and II-positive cells were present in the lamina propria or gut epithelium. Mussener et al. (29) reported a similar finding in rat synovia in which cells within lymphocytic aggregates such as CD4+ cells lacked expression of Tβ RI and II. In contrast, we found immune cells including CD4+ and CD8+ T cells, veiled dendritic cells, B cells, and a small number of follicular dendritic cells in the lamina propria as well as CD8+ IEL expressing Tβ RIII, and the number of labeled cells increased significantly from d 10 to 21 of age. Tβ RIII can bind TGF-β independently of the other receptors (13, 30). Tβ RIII may function in attracting and concentrating TGF-β on the cell surface for eventual transfer to closely situated signal-transducing receptors (31, 32). It may serve to regulate the binding of TGF-β to its signal-transducing receptors by targeting TGF-β to appropriate locations in the microenvironment of cells.

In summary, there is an ontogenetic development of TGF-β receptor expression in the postnatal rat small intestine. Before weaning, Tβ RIII was coexpressed with Tβ RI and II throughout the epithelium. Over the weaning period, Tβ RI and II expression remained on the epithelium, whereas Tβ RIII gradually disappeared from the epithelium. This unique distribution pattern of TGF-β receptor expression from preweaning to postweaning may correlate with the change in ligand specificity from maternal TGF-β2 to endogenous TGF-β1 in the postnatal intestine.