Maximum Immunobioactivity of Murine Small Intestinal Intraepithelial Lymphocytes Resides in a Subpopulation of CD43⁺ T Cells

Mice

Adult female BALB/c and C57BL/6 mice were purchased from Harlan Sprague Dawley (Houston, TX). IL-10^−/− (B6.129P2-Il10^tm1Cgn/J) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Animals were used in accord with the University of Texas Animal Welfare Guidelines.

Isolation and purification of IELs

Isolation of small intestine IELs was done as previously described (4, 5). Briefly, after tissue digestion with 2 mM DTT and 5 mM EDTA, IELs were separated from epithelial cells by centrifugation over a 40/70% Percoll gradient. For IEL fractionation into S7⁺ and S7⁻ populations, IELs were stained by direct labeling at 4°C with PE-anti-CD43 (S7) and sorted using a Vantage SE Turbo high-speed cell sorter (BD Biosciences, Mountain View, CA). Cell sorting was done by first selecting the IEL population according to parameters of forward- and right-angle scatter, which we have previously determined to include >95% CD3⁺ cells, and second, by gating onto the S7⁺ and S7⁻ subsets.

Abs and flow cytometry

Abs used in this study were as follows: purified NA/LE anti-CD3 (145-2C11); FITC-anti-TCRβ (H57-597); FITC-anti-TCRδ (GL3); CyChrome-anti-CD8α (53-6.7); biotynylated-anti-CD8β (53-5.8); purified, PE-, FITC-, and biotin-anti-CD43 (S7); FITC-CD44 (IM7); purified NA/LE anti-hamster Ab (G94-56); streptavidin-allophycocyanin; strepavidin-CyChrome; streptavidin-PE; FITC-anti-BrdU; GolgiStop (monensin) (BD Pharmingen, San Diego, CA); FITC-anti-IL-4 (BVD6-24G2); FITC-anti-IL-10 (JES5-16E3); PE-anti-IFN-γ (XMG1.2); FITC control (eBRG1); PE control (eBRG1); PE-anti-CD94 (18d3); PE-anti-NK (Ly-49F/I/C/H) (eBioscience, San Diego, CA). Anti-CD43 mAb R2/60 was biotinylated in our laboratory. Alkaline phosphatase anti-rat IgG was purchased from Southern Biotechnology Associates (Birmingham, AL). Flow cytometric analyses were done on a FACSCalibur flow cytometer using CellQuest software (BD Biosciences). Intracellular staining was done using IELs cultured in RPMI 1640 (Sigma-Aldrich, St. Louis, MO) supplemented with 10% FBS (Sigma-Aldrich) for 24 h with or without anti-CD3 stimulation as previously described (4). Monensin (1 μl/ml) was added for the last 5 h of culture. Intracellular staining of IELs from 14-mo-old IL-10^−/− mice was done using cells recovered from the lower one-third of the small intestine, followed by in vitro culture in RPMI 1640 containing 10% FBS and 1 μl/ml monensin for 5 h to allow for intracellular accumulation of cytokines. Cells were collected and stained for CD43 S7, IFN-γ, and IL-4, or IL-10.

Enzyme-linked immunoassays and redirected cell-mediated cytotoxicity assay

Commercial cytokine assay kits for IL-2, IL-4, IL-10, IL-17, and IFN-γ were purchased from eBioscience. S7⁺ and S7⁻ IEL populations were obtained by flow cytometric cells sorting. In vitro anti-CD3 stimulation of IELs was done by coating plates overnight at 4°C with 10 μg/ml anti-hamster Ig in PBS. Wells were washed, and 1 μg/ml anti-CD3 mAb was added in PBS for 2 h at 37°C followed by 20-min blocking with 50% newborn calf serum. Cell-free supernatants were collected after 24 h and assayed for the presence of cytokines. Redirected cell-mediated assays were done as previously reported (4, 5) using S7⁺ and S7⁻ IELs purified by flow cytometric cell sorting. For determining the effects of anti-CD3 stimulation on the cytotoxic activity of S7⁻ and S7⁺ IELs, cell-sorted populations were cultured for 48 h in plates coated with anti-CD3 mAb (4). Viable cells were collected and assayed in the redirected cytotoxicity assay.

Western blot

A total of 10⁷ whole unfractionated IELs, or IELs depleted of S7⁺ cells by MACS (autoMACS; Miltenyi Biotec, Auburn, CA), were solubilized in 50 μl of detergent buffer consisting of 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EGTA, 1 mM NaF, 1 μg/ml aprotinin, 1 μg/ml leupeptin, 1 μg/ml pepstatin, and 1% Nonidet P-40 (Sigma-Aldrich). A volume of 12.5 μl of cell lysates was mixed with 2× sample buffer (Bio-Rad, Hercules, CA) containing 5% 2-ME (Sigma-Aldrich) and boiled for 5 min. A volume of 15 μl of sample was added to precast 7.5% SDS-glycine cells (Bio-Rad). After electrophoresis, proteins were transferred to PVDF membranes (Bio-Rad). Membranes were blocked overnight with 5% nonfat dry milk in PBS, washed, and reacted for 2 h with 1 μg/ml purified S7 mAb, and washed and reacted for 2 h with 1/2000 dilution of alkaline phosphatase anti-rat IgG. Membranes were washed, developed using an ECL protocol (ICN Pharmaceuticals, Costa Mesa, CA), and exposed to Kodak (Rochester, NY) x-ray photographic film.

In vivo BrdU labeling

Mice were injected i.p. twice 4 h apart with 1 mg of BrdU (Sigma-Aldrich) suspended in PBS. Twenty-four hours after the last injection, IELs were recovered and stained for intracellular BrdU uptake using FITC-anti-BrdU Ab with PE-anti-CD43 Ab.

DNA microarrays

IELs, enriched for S7⁺ and S7⁻ cells by high-speed cell sorting, were washed in PBS and lysed with buffer provided in the RNeasy mini-kit 50 (Qiagen, Valencia, CA). Three independent samples were used for S7⁻ gene chip analyses, and two independent samples were used for S7⁺ gene chip analyses. Each sample consisted of cell-sorted IELs from three to four mice. RNAs were extracted and stored at −80°C until a sufficient amount was obtained to yield 8 μg for each sample. cDNAs were prepared from samples. The first-strand cDNA was primed with a T7-(dT)₂₄ primer at 70°C and catalyzed by SuperScript II RT (Invitrogen Life Technologies, Carlsbad, CA) at 42°C. The second-strand cDNA was synthesized by nick translational replacement in the presence of Escherichia coli DNA polymerase, E. coli RNase H, and E. coli DNA ligase (Invitrogen Life Technologies). To synthesize the biotin-labeled cRNA, the dsDNA was used as template for in vitro transcription with Enzo Bioarray High Yield RNA Transcription Labeling kit (Affymetrix, Santa Clara, CA). Fragmentation of the labeled cRNA was done at 95°C before the hybridization mixture was generated; the hybridization mixture was made by adding 10 μg of biotin-labeled cRNA from each sample along with Eukaryotic Hybridization Controls and Control Oligonucleotide B2 (Affymetrix). The mixture was then hybridized to GeneChip Murine Genome U74Av2 (Affymetrix). Chips were washed and stained with streptavidin-PE (Molecular Probes, Eugene, OR). Signals were amplified by hybridizing with biotin-labeled anti-streptavidin-PE (Vector Laboratories, Burlingame, CA) and restained with streptavidin-PE (Molecular Probes). The arrays were scanned with GeneArray Scanner and visualized with Microarray Suite, version 5.0 (Affymetrix). Gene expression levels were averaged for the three S7⁻ samples and for the two S7⁺ samples for determination of fold-change values and p values. Statistical analysis was done using Data Mining Tool, version 3 (Affymetrix); statistical significance was defined as ≥2-fold increase with a p value of ≤0.05, similar to that previously reported (22). Absolute call values for gene expression levels were determined for the five samples using Gene Expression MAS 5.0 software (Affymetrix), using default values provided by the manufacturer.

CD8⁺ IELs express the CD43 core molecule but differentially express a murine CD43 S7 isoform

Studies from our laboratory demonstrated that the CD43 determinant recognized by the S7 mAb is expressed on 85–90% of all CD8⁺ peripheral T cells in mice (16). However, although all IELs expressed CD43 as seen by reactivity with mAb R2/60 (Fig. 1⇓), the S7 isoform was differentially expressed on small intestinal IELs. Thus, the S7 determinant was expressed on 99.2% of the TCRαβ⁺CD8β⁺ IELs (Fig. 2⇓, A and B), and on 80.7% of the TCRαβ⁺CD8α⁺ IELs (A and C). A total of 19.3% of the TCRαβ⁺CD8α⁺ cells were S7⁻ cells (Fig. 2⇓, A and C). In contrast, 42.4% of the TCRγδ⁺CD8a⁺ IELs were S7⁺ cells (Fig. 2⇓, D and F), and 57.6% were S7⁻ cells (D and F). The overall distribution of S7 expression on CD8α and CD8β IELs is shown in Fig. 2⇓, G–I.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Nearly all murine small intestinal IELs express the CD43 core molecule recognized by mAb R2/60. Data are representative of four independent experiments.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

The CD43 S7 Ag is differentially expressed on murine IELs. A–C, The CD43 S7 Ag is expressed on nearly all TCRαβ⁺CD8β⁺ IELs, and most but not all TCRαβ⁺CD8α⁺ IELs. D–F, TCRγδ⁺CD8α⁺ IELs consist of both S7⁺ and S7⁻ subsets. G–I, Three-color staining demonstrating that CD8β⁺ IELs are S7⁺ cells, whereas the S7 determinant is differentially expressed on CD8α⁺ IELs. Data are representative of five independent experiments.

The fact that not all IELs express the S7 Ag by flow cytometry did not formally rule out the possibility that the S7 determinant was expressed on S7⁻ cells, because the lack of staining could be due to a failure of mAb S7 to react with some IELs. To address this, Western blotting was done using the S7 mAb with cell lysates from whole IELs (i.e., cells that included the S7⁺ population), and IELs from which the S7⁺ subset had been removed by negative MACS sorting. As seen in Fig. 3⇓, the S7 mAb was nonreactive by Western blotting for cell lysates from S7-depleted IELs, whereas lysates from unfractionated IELs identified a 114- to 118-kDa component characteristic of that recognized by mAb S7 (21), thereby confirming the lack of expression of the S7 determinant on CD43 S7⁻ IELs.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

CD43 S7⁻ IELs are nonreactive with S7 mAb by Western blotting. Cell lysates from equivalent numbers of whole IELs and IELs depleted of S7⁺ cells by MACS sorting were electrophoresed and blotted with mAb CD43 S7 as described in Materials and Methods. The lack of reactivity of the S7 mAb against S7-depleted cell lysates confirms that the S7 determinant is not expressed on those cells. Data are representative of two independent experiments.

Following anti-CD3 activation, S7⁺ IELs secrete higher levels of cytokines, and both S7⁺ and S7⁻ populations contain Th1, Th2, and Th0 cells

To determine how S7⁺ and S7⁻ IELs differ functionally, purified S7⁺ and S7⁻ IELs were assayed for their ability to secrete five IEL cytokines that are prominently involved in IEL regulation and/or activation (IL-2, IL-4, IL-10, IL-17, and IFN-γ). After isolation and purification by Percoll gradient centrifugation, S7⁺ and S7⁻ IELs were enriched to >98% purity by high-speed flow cytometric cell sorting. Equal numbers of each population were cultured for 24 h in 24-well tissue culture plates coated with anti-hamster Ig plus hamster anti-CD3 mAb or control mAb. Cell-free supernatants were collected and screened for IL-2, IL-4, IL-10, IL-17, and IFN-γ activities by enzyme immunoassay. Findings from these experiments were notable in that S7⁺ cells produced significantly higher levels of all five cytokines compared with S7⁻ IELs (Fig. 4⇓), indicating that there was a major difference in the levels of cytokine production by S7⁺ IELs following anti-CD3 stimulation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

CD43 S7⁺ IELs produce significantly greater levels of IL-2, IL-4, IL-10, IL-17, and IFN-γ than S7⁻ IELs. IELs were enriched into S7⁺ and S7⁻ populations by flow cytometric cells sorting. Purified cells were stimulated with immobilized anti-CD3 mAb for 24 h; cell-free supernatants were collected and assayed for cytokine activity using enzyme-linked immunoassays. Data are mean values of two independent experiments; bars indicate range of values for both experiments.

Because S7⁺ and S7⁻ IELs secreted both Th1 (IL-2 and IFN-γ) and Th2 (IL-4 and IL-10) cytokines, albeit at quantitatively different levels, it was of interest to determine the extent to which activated IELs were Th1 or Th2 cells, or whether they were Th0 cells that produced both Th1 and Th2 cytokines (23, 24). IELs were cultured for 24 h with or without anti-CD3 stimulation. Cells were collected and stained for three-color flow cytometric analysis based on expression of S7, intracellular IFN-γ, and intracellular IL-4 or IL-10. After anti-CD3 stimulation, the majority of cytokine-secreting cells in both the S7⁺ and S7⁻ groups were Th1 cells that produced IFN-γ but not IL-4, although proportionally more Th1 cells were present in the S7⁺ population (Fig. 5⇓A). Additionally, a small proportion of the S7⁺ and S7⁻ cells were Th2 cells that produced IL-4 only following anti-CD3 stimulation (Fig. 5⇓A). Interestingly, some cells in both the S7⁺ and S7⁻ populations were Th0 cells that produced IFN-γ and IL-4; this was particularly true for the S7⁺ cells (Fig. 5⇓A). The pattern for IL-10 staining reflected that observed for IL-4 in that most IL-10-producing S7⁺ cells also produced IFN-γ (Fig. 5⇓B). Cells cultured without anti-CD3 stimulation did not synthesize IFN-γ, IL-4, or IL-10 (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Anti-CD3 stimulation of IELs results in Th1, Th2, and Th0 profiles. A and B, IELs were cultured for 24 h with plate-bound anti-CD3 mAb and stained for three-color flow cytometric analysis for surface CD43 S7, intracellular IFN-γ, and intracellular IL-4 or IL-10. In both the S7⁺ and S7⁻ populations, the majority of IELs were Th1 cells that synthesized IFN-γ. IL-4⁺ and IL-10⁺ cells were present among both the S7⁺ and S7⁻ populations, although proportionally more S7⁺ cells were IL-4⁺ or IL-10⁺ cells. Note the presence of Th0 cells, particularly among the S7⁺ population. Cursor settings are positioned according to staining with FITC-Ig and PE-Ig control Abs. Data are representative of four independent experiments for each group.

CD43 S7⁺ but not S7⁻ IELs in IL-10^−/− mice are the primary source of IFN-γ

Studies involving anti-CD3 stimulation of IELs were relevant for understanding the process of activation as it would occur in response to Ag involving a TCR-mediated signal. It was of interest, however, to also examine the responses of S7⁺ and S7⁻ IELs during pathophysiological conditions of chronic intestinal inflammation. Because most adult IL-10^−/− mice >15 wk of age develop intestinal inflammation in the colon, ileum, and jejunum (25), small intestinal IELs from the ileum of 14-mo-old IL-10^−/− mice with rectal prolapse were isolated and cultured for 5 h without anti-CD3 stimulation in RPMI 1640 supplemented with 10% FBS plus monensin to permit the accumulation of intracellular cytokines generated in situ during the inflammatory process. Cells were harvested and stained for CD43 S7 expression and intracellular IFN-γ and IL-4. As seen in Fig. 6⇓A, IELs from the ileum of normal C57BL/6 mice cultured without anti-CD3 stimulation produced neither IFN-γ nor IL-4 (B). In contrast, a significant proportion of S7⁺ ileal IELs from IL-10^−/− mice without anti-CD3 stimulation spontaneously produced IFN-γ (Fig. 6⇓B). Moreover, IFN-γ-producing cells by S7⁻ IELs was <10% of that of S7⁺ IELs (Fig. 6⇓B). In neither the S7⁺ nor the S7⁻ IEL subset was there a significant number of IL-4-producing cells (Fig. 6⇓B). These findings make an important point in that they indicate that the basic relationship between S7⁺ IELs and activation observed for CD3-stimulated IELs also held true in a model of intestinal inflammation induced by cytokine dysregulation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

IELs from the ileum of IL-10^−/− mice with rectal prolapse spontaneously produce IFN-γ predominantly within the S7⁺ population. A, Ileal IELs from normal C57BL/6 mice cultured for 5 h without anti-CD3 stimulation fail to synthesize IFN-γ or IL-4. B, Ileal IELs from a 14-mo-old IL-10^−/− mice with rectal prolapse cultured for 5 h without anti-CD3 stimulation synthesized IFN-γ in the S7⁺ population but only minimally in the S7⁻ population, indicating that the S7⁺ subset is the likely source of proinflammatory cytokines in IL-10^−/− with active intestinal inflammation. Data are representative of three IL-10^−/− mice.

CD43 S7⁺ cells have higher cytolytic activity and in vivo proliferation than S7⁻ IELs

Small intestinal IELs are known to express cytolytic activity in freshly isolated preparation without overt activation; this can be demonstrated using anti-CD3 or anti-TCR Abs to bridge effector cells to target cells (1, 2, 3, 4, 5). Freshly isolated IELs were enriched by high-speed cell sorting into S7⁺ and S7⁻ populations and tested for lytic activity in the redirected assay. As seen in Fig. 7⇓A, there were ∼4-fold higher levels of cytolytic activity of S7⁺ cells compared with an equivalent number of S7⁻ cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

CD43 S7⁺ IELs are more cytotoxic than S7⁻ IELs. A, IELs from normal mice were purified into S7⁺ and S7⁻ populations by flow cytometric cell sorting, and assayed against P815 target cells in the presence of anti-CD3 mAb. S7⁺ IELs expressed ∼4-fold greater cytotoxic activity than S7⁻ IELs. B, After 48 h of anti-CD3 stimulation, there is ∼2-fold increase in the lytic activity of S7⁻ IELs compared with fresh S7⁻ IELs (A); however, there is ∼16-fold increase in the lytic activity of the S7⁺ cell population, indicating that S7⁺ IELs are more cytotoxic than S7⁻ IELs both before and after CD3-mediated activation. C, A greater proportion of CD43 S7⁺ IELs undergo in vivo than CD43 S7⁻ IELs in normal nonprimed mice. Twenty-four hours after in vivo BrdU injection, IELs were isolated and stained for intracellular BrdU with anti-CD43 S7 staining. Data are representative of two independent experiments.

To determine how CD3-mediated activation might affect the cytotoxic activity of S7⁻ and S7⁺ IELs, S7⁻ and S7⁺ cells were purified by flow cytometric cell sorting and were cultured in plates containing immobilized anti-CD3 mAb. After 48 h, cells were collected and assayed for lytic activity in the redirected cytotoxicity assay. Although the lytic activity of both the S7⁻ and the S7⁺ population increased over that of fresh IELs, there were quantitative differences in the increase in lytic activity of the S7⁻ vs the S7⁺ cells. Whereas the cytolytic activity of S7⁻ IELs increased by ∼2-fold, the activity of the S7⁺ IELs increased by ∼16-fold over that of fresh IELs (Fig. 7⇑, A and B). Additionally, the cytotoxic activity of the S7⁺ IELs was at least 16-fold greater than that of the S7⁻ IELs in CD3-stimulated cultures (Fig. 7⇑B).

Murine IELs have been shown to undergo cell proliferation in vivo as demonstrated by BrdU incorporation (26). Based on the above findings of heightened cytolytic activity and cytokine secretion by S7⁺ IELs, we sought to determine whether S7 expression correlated with the in vivo proliferation capacity of IELs. Normal mice were injected twice with BrdU as described in Materials and Methods. Twenty-four hours after the last injection, animals were sacrificed, IELs were recovered, and cells were stained for intracellular BrdU in conjunction with CD43 S7 expression. As seen in Fig. 7⇑C, significantly more CD43 S7⁺ than S7⁻ IELs incorporated BrdU in vivo, thus indicating that S7⁺ cells are an actively dividing cell population in vivo, either from a homeostatic proliferation process or possibly due to Ag-driven activation in vivo.

Gene array studies point to significant differences in the functional roles of S7⁺ and S7⁻ IELs

To more extensively examine the differences in IELs based on CD43 S7 expression, DNA microarray analyses were done using cell-sorted S7⁺ and S7⁻ IELs. Findings from these studies reinforced the observations that functional differences exist between S7⁺ and S7⁻ IELs. By three criteria (fold increase; value of p < 0.05; and absolute call for gene expression), S7⁺ IELs expressed significantly higher levels of genes for a number of functionally relevant genes. The most notable of these were the CD6, CD44, and CD97 cell surface activation markers; CCR2, CXCR3, and MIP-β gene expression; and the cathepsin L protease gene (Table I⇓). S7⁻ IELs also displayed several prominent features. Perhaps the most significant of these was the strong association between S7⁻ cells and the expression of genes for receptors linked to NK cell activation/inhibition (Ly49E-GE, Ly49G.2, CD94/NKG2, Ly49H, and Ly49C) (Table II⇓). It should be noted that, although the majority of the genes with increased gene expression for the S7⁻ also were expressed by S7⁺ cells (absolute call), gene expression for those nonetheless occurred at significantly higher levels than for S7⁺ cells (fold increase, and value of p < 0.05). Although this likely points to a real difference between those populations at a functional level, it similarly reinforces the tendency of S7⁺ IELs to have more broad-ranging effector activities than S7⁻ cells.

To confirm at the protein level the gene expression patterns observed in Tables I⇑ and II⇑, freshly isolated IELs were stained for expression of CD43 S7, CD44, Ly-49F/I/C/H, and CD94. Consistent with the gene array findings, CD43 S7⁺ cells expressed CD44 at high levels, whereas CD44 was differentially expressed on S7⁻ IELs. Conversely, more than twice as many CD43 S7⁻ IELs expressed the NK markers, Ly-49F/I/C/H and CD94, than S7⁺ IELs (Fig. 8⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

To confirm the validity of gene expression patterns observed in Tables I⇑ and II⇑, freshly isolated IELs were stained according to S7 expression and CD44, Ly-49F/I/C/H, and CD94. Note that nearly all CD43 S7⁺ cells were CD44⁺ cells, whereas CD44 was differentially expressed on S7⁻ IELs. Conversely, a greater proportion of CD43 S7⁻ IELs expressed Ly-49F/I/C/H and CD94 than CD43 S7⁺ IELs, thus confirming the basic pattern observed from the gene array studies. Data are representative of two independent experiments.