Response Directing the Tone of the Intestinal Immune Murine Small ...

Spontaneous and Continuous Cyclooxygenase-2-Dependent Prostaglandin E2 Production by Stromal Cells in the Murine Small Intestine Lamina Propria: Directing the Tone of the Intestinal Immune Response1 Rodney D. Newberry,*† Jacquelyn S. McDonough,‡ William F. Stenson,*† and Robin G. Lorenz2*‡ The mechanisms allowing the gastrointestinal immune system to avoid an inappropriate inflammatory response to nonpathogenic luminal Ags are poorly understood. We have previously described a role for cyclooxygenase (COX)-2-dependent arachidonic acid metabolites produced by the murine small intestine lamina propria in controlling the immune response to a dietary Ag. To better understand the role of COX-2-dependent arachidonic acid metabolites produced by the lamina propria, we examined the pattern of expression and the cellular source of COX-2 and COX-2-dependent PGE2. We now demonstrate that non-bone marrow-derived lamina propria stromal cells have basal COX-2 expression and that COX-2-dependent PGE2 production by these cells is spontaneous and continuous. The other mucosal and nonmucosal lymphoid compartments examined do not share this phenotype. In contrast to the majority of descriptions of COX-2 expression, COX-2 expression by lamina propria stromal cells is not dependent upon exogenous stimuli, including adhesion, LPS signaling via Toll-like receptor 4, or the proinflammatory cytokines TNF-␣, IFN-␥, and IL-1␤. These findings, in conjunction with the known immunomodulatory capacities of PGs, suggest that COX-2 expression by the small intestine lamina propria is a basal state contributing to the hyporesponsiveness of the intestinal immune response. The Journal of Immunology, 2001, 166: 4465– 4472.

T

he alimentary tract represents the physical convergence of bacteria and bacterial products, dietary Ags, and activated cellular components of the immune system (1). The proximity of these elements to each other presents a continual challenge to the immune system because the development of a cellmediated immune response with the production of inflammatory mediators can result in damage to the intestinal mucosa (2– 6). Given this volatile environment, it is likely that compensatory/ regulatory mechanisms exist that limit inflammation in the intestinal mucosa. The majority of recent work has focused on the role of immunomodulatory T cells and the cytokines they produce in controlling this potentially damaging inflammatory response (7–10). However, an additional mechanism for preventing an inflammatory response to dietary Ags has been demonstrated for cyclooxygenase (COX)3-2-dependent arachidonic acid (AA) metabolites produced by the small intestine lamina propria (LP) (11).

*Department of Internal Medicine, †Division of Gastroenterology, and ‡Department of Pathology and Immunology, Washington University School of Medicine, St. Louis, MO 63110 Received for publication July 31, 2000. Accepted for publication January 24, 2001. 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. 1 This work was supported by the Glaxo Wellcome Institute for Digestive Health Basic Research Award (to R.D.N.), National Institutes of Health Grants DK-33165 and DK55753 (to W.F.S.) and DK-02608 (to R.D.N.), and a grant from Washington University School of Medicine Digestive Diseases Research Core Center (P30-DK52574). 2 Address correspondence and reprint requests to Dr. Robin G. Lorenz, Department of Pathology and Immunology, Box 8118, Washington University School of Medicine, 660 South Euclid Avenue, St. Louis, MO 63110. E-mail address: lorenz@pathbox. wustl.edu 3 Abbreviations used in this paper: COX, cyclooxygenase; AA, arachidonic acid; TX, thromboxane; LP, lamina propria; PP, Peyer’s patch; SPL, splenocyte; PC, peritoneal cells; MNC, mononuclear cells.

Copyright © 2001 by The American Association of Immunologists

COXs perform the committed step in the conversion of AA to PGH2, an intermediate metabolite in the pathway to the production of PGs and thromboxanes (TXs) by cell-specific reductases and isomerases (12). Mammalian COXs are known to exist in two isoforms encoded by distinct genes, COX-1, whose expression is constitutive, and COX-2, whose expression is predominantly inducible and transient (13, 14). COX-1 and -2 enzymes perform identical steps in the production of AA metabolites; however, their physiological roles are viewed differently, largely due to differences in the patterns of expression of these isoenzymes. COX-2dependent AA metabolites are rapidly and transiently produced at sites of inflammation (15). This observation, along with the antiinflammatory properties of selective COX-2 inhibitors, led to the concept that COX-2 and COX-2-dependent AA metabolites act predominantly as inflammatory mediators. Conversely, COX-1 and COX-1-dependent AA metabolites are largely viewed as having housekeeping functions based on their constitutive expression and production (16). These concepts are challenged by descriptions of constitutive COX-2 expression in the kidney (17), the brain (18), and in colorectal carcinomas (19, 20) and by descriptions of the immunomodulatory/anti-inflammatory roles of PGs (21, 22). We have previously shown that the small intestine LP produces abundant PGE2 in a COX-2-dependent manner (11). To better understand the role of the COX-2-dependent AA metabolites produced by the small intestine LP, we have identified the source of COX-2-dependent AA metabolites and examined the pattern of COX-2 expression in the small intestine LP. Here we report that non-bone marrow-derived stromal cells in the murine small intestine spontaneously and continuously express COX-2 and abundantly produce a COX-2-dependent AA metabolite, PGE2. Development and maintenance of this phenotype is not dependent on “typical” proinflammatory mediators, including bacterial flora, LPS, TNF-␣, IL-1␤, IFN-␥, or IL-12. We also demonstrate that 0022-1767/01/$02.00

4466

COX-2-DEPENDENT PGE2 PRODUCTION IN THE MOUSE SMALL INTESTINE

that these cells do not produce a soluble mediator with the ability to induce PGE2 production in resident peritoneal macrophages. These observations, combined with previous descriptions of the immunomodulatory role of PGs, suggest that COX-2 expression and COX-2-dependent PGE2 production by the murine small intestine LP is a basal state establishing the tone of the intestinal immune response to luminal Ags.

Materials and Methods Mice All mice used for this study were 6 –12 wk of age and, with the exception of germfree mice, were housed in a specific pathogen-free facility and fed routine chow diet. Germfree mice were housed in a gnotobiotic facility and fed an elemental diet. At the time of sacrifice, cecal cultures from germfree animals were obtained to confirm the absence of bacterial flora. B10.BR/ SgSnJ, C3H/HeJ, C3H/HeSnJ, C57BL/6, TNFR II-deficient, and IL-12 p40-deficient mice on the C57BL/6 background were purchased from The Jackson Laboratory (Bar Harbor, ME). IFN-␥R-deficient mice (23) on the 129/SvImJ background were a gift from R. Schrieber (Washington University School of Medicine, St. Louis, MO). IL-1␤-deficient mice (24) and TNFR I-deficient mice (25) on the C57BL/6 background were a gift from D. Chaplin (Washington University School of Medicine). Germfree and conventionally housed FVB/NJ mice were a gift from J. Gordon (Washington University School of Medicine). COX-2-deficient mice on a mixed C57BL/6 ⫻ 129/SvImJ background were a gift from R. Langenbach (National Institute of Environmental Health Sciences, Research Triangle Park, NC) (15).

Cell isolation and culture LP cells, Peyer’s patch (PP) cells, and splenocytes (SPLs) were isolated as previously described (11). Briefly, single-cell suspensions were prepared from spleens, small intestines, and PPs by dispase and collagenase digestion. Resident peritoneal cells (PC) were harvested by flushing the peritoneal cavity with 15 ml ice-cold PBS. Cells were counted for viability by trypan blue exclusion. Viability was ⬎75% for all cell populations. The typical yield for LP cell isolation was 2 ⫻ 107 viable cells/intestine, and the typical yield for resident PC was 5 ⫻ 106 viable cells/mouse. Isolated cells were cultured in 96-well tissue culture plates at a density of 2.5 ⫻ 106 cells/ml in RPMI 1640 medium (BioWhittaker, Walkersville, MD) containing 2 mM GLUTAMAX I (L-Alanyl-L-Glutamine; Life Technologies, Gaithersburg, MD), 10 mM HEPES, 1 mM sodium pyruvate, 50 U/ml penicillin-50 ␮g/ml streptomycin, 5 ⫻ 10⫺5 M 2-ME, and 10% FCS (HyClone, Logan, UT) at 37°C and 5% CO2. NS-398 (Biomol, Plymouth Meeting, PA) was used to selectively inhibit COX-2, and indomethacin (Sigma, St. Louis, MO) was used as a nonselective COX inhibitor. NS-398 selectively inhibits COX-2 (IC50 ⫽ 1 ␮M) with no inhibition of COX-1 at concentrations up to 100 ␮M (26). LPS from Salmonella typhimurium (Sigma) was used to stimulate resident PC (Table II).

Enrichment of densely adherent cells LP cells were isolated as described above and cultured on polystyrene tissue culture plates at a density of 2 ⫻ 106 cells/ml in RPMI 1640 medium containing 2 mM GLUTAMAX I, 10 mM HEPES, 1 mM sodium pyruvate, 50 U/ml penicillin-50 ␮g/ml streptomycin, 5 ⫻ 10⫺5 M 2-ME, and 0.5% normal mouse serum. After 1 h, nonadherent cells were removed by washing with cold PBS, and the remaining adherent cells were cultured overnight in the medium described above. After overnight culture, the plates were washed again with cold PBS to remove nonadherent cells (adherent 1 h), and the remaining adherent cells were incubated in HBSS (BioWhittaker) containing 2 mM EDTA at 4°C for 15 min to release the remaining adherent cell population (adherent overnight). For experiments examining the profile of PG and TX production (Table I), LP cells were isolated and cultured as described above. After 24 h, nonadherent cells were removed and the remaining adherent cell population was cultured for an additional 24 h, and supernatants were then removed for analysis.

Measurement of PGs and TX For the measurement of PGs and TXB2, cellular populations were isolated and cultured as described above, and supernatants were removed and stored at ⫺80°C until analysis. PGE2, PGF2␣, 6-keto PGF1␣, TXB2, and 11␤PGF2␣ measurements were performed using specific ELISAs (Cayman Chemicals, Ann Arbor, MI). For the measurement of PGD2, samples were treated with methoxime-HCl and analyzed for the presence of PGD2-me-

thoxime using a PGD2-methoxime-specific ELISA (Cayman Chemicals). Each supernatant was measured in duplicate in at least three dilutions using a Bio-Tek microplate reader (Bio-Tek, Winooski, VT).

Western blot assays Western blot analysis for COX-2 production was performed as previously described (15). Briefly, cells were washed twice in PBS, resuspended in a buffer consisting of 100 mM Tris-HCl, 3% SDS, 6% 2-ME, 30% glycerol, and 0.05% bromophenol blue and heated to 90°C for 5 min. Suspensions underwent electrophoresis on a 5–10% gradient Tris-glycine gel (Bio-Rad, Hercules, CA) and were transferred to polyvinylidene difluoride membranes (Tropix, Bedford, MA). Detection of COX-2 protein was performed using Western light detection system (Tropix), rabbit anti-COX-2 Ab (Cayman Chemicals), and goat anti-rabbit IgG alkaline phosphatase-conjugated Ab (Jackson ImmunoResearch, West Grove, PA). Polyvinylidene difluoride membranes were stripped and reprobed with rabbit anti-actin Ab followed by goat anti-rabbit IgG alkaline phosphatase-conjugated Ab to detect actin as a control for protein loading and transfer.

Immunohistochemical analysis Cell culture. Small intestine LP cells were isolated as described above and cultured in chamber slides (Nunc, Naperville, IL) for 72 h until they became densely adherent, and nonadherent cells were removed by washing with PBS. CD45 and CD11b were detected using biotin-conjugated antiCD45 (BD PharMingen, San Diego, CA) and biotin-conjugated antiCD11b (BD PharMingen), respectively, followed by FITC-conjugated streptavidin (BD PharMingen). Cells were then fixed and permeabilized using a Cytoperm/Cytofix kit (BD PharMingen), and COX-2 was detected with goat anti-COX-2 Ab (SC-1747; Santa Cruz Biotechnology, Santa Cruz, CA) followed by Cy3-conjugated donkey anti-goat Ab. Cells were counterstained with Hoescht dye (bis-benzimide, B2883; Sigma) to detect nuclei. To determine the percentage of cells expressing COX-2, CD45, and CD11b, stained cell populations were examined by fluorescent microscopy at ⫻200 magnification. Random fields were selected from each group, and the number of nuclei, COX-2⫹ cells, and double positive (COX-2⫹ and CD45⫹ or CD11b⫹) cells were counted. Random fields were counted until ⬃500 cells had been examined. Percentage of single positive and double positive cells were recorded for each field and used to calculate the mean ⫾ SD of single and double positive cells. Small intestine sections. Two-centimeter sections of small intestines from 6- to 8-wk-old C57BL/6 mice were flushed with ice-cold PBS, embeded in Tissue Tek OCT compound (VWR Scientific Products, West Chester, PA), frozen, and cut into 7-␮m thick acetone-fixed sections. Endogenous peroxidases were inactivated by treating sections with 0.3% hydrogen peroxide followed by treatment with avidin/biotin blocking kit (Vector Laboratories, Burlingame, CA) and blocking of nonspecific staining with PBS containing 1% BSA and 0.2% powdered skim milk. COX-2 protein was identified using goat anti-COX-2 Ab (SC-1747; Santa Cruz Biotechnology) followed by biotin-conjugated donkey anti-goat Ab and streptavidin-conjugated HRP. For detection, we used a Cy3-tyramide signal amplification kit (DuPont/NEN, Boston, MA). Following this signal amplification, peroxidases were again inactivated, and avidin and biotin was blocked as described above. ␣-smooth muscle actin, CD45, and CD11b were detected using FITC-conjugated ␣-smooth muscle actin (clone 1A4; Sigma), FITCconjugated anti-CD45 (BD PharMingen) or biotin-conjugated CD11b (BD PharMingen) followed by FITC-conjugated streptavidin (BD PharMingen). Adjacent sections were stained with appropriate isotype Abs to serve as negative controls. Sections were counterstained with Hoescht dye to detect nuclei. Flow cytometric analysis. Single-cell suspensions of adherent small intestine LP cells were obtained as described above and resuspended in PBS with 1% BSA (Fisher Scientific, Pittsburgh, PA) and 1 mg/ml human IgG (Novartis Pharmaceuticals, East Hanover, NJ) at 2 ⫻ 107 cells/ml. Cells were stained with FITC-conjugated rat anti-mouse CD45 Ab, PE-conjugated rat anti-mouse CD19, biotin-conjugated rat anti-mouse CD11b, biotin-conjugated hamster anti-mouse TCR-␤, (all obtained from BD PharMingen), rat anti-mouse c-kit (Caltag, Burlingame, CA), or appropriate isotype control Abs (BD PharMingen) for 30 min on ice, washed twice, and resuspended in the above-described buffer. To detect CD11b, TCR-␤, and c-kit, cellular suspensions were then stained with streptavidin-PE or FITCconjugated goat anti-rat Ig (both obtained from BD PharMingen) for 30 min on ice. Cellular suspensions were washed twice and fixed with 1% paraformaldehyde in PBS. Flow cytometric analysis was done on a triplelaser flow cytometer (FACScan; Becton Dickinson, Mountain View, CA), and analysis was performed on a Macintosh G3 using the CellQuest program (Becton Dickinson). Dead cells were excluded based on forward and

The Journal of Immunology

4467

Table I. Profile of COX-dependent AA metabolites produced by LP adherent cells

C57BL/6 LP C57BL/6 LP ⫹ C57BL/6 LP ⫹ C57BL/6 LP ⫹ C57BL/6 LP ⫹ COX-2⫺/⫺ LP

PGE2 (ng/ml)

PGF2␣ (ng/ml)

6-keto PGF1␣ (ng/ml)

TXB2 (ng/ml)

PGD2 (ng/ml)

Viable Cells

338.0 ⫾ 60 10.3 ⫾ 1.1 23.2 ⫾ 3.0 9.8 ⫾ 1.0 12.5 ⫾ 1.4 9.7 ⫾ 0.7

7.4 ⫾ 0.7 1.0 ⫾ 0.1 1.6 ⫾ 0.2 0.9 ⫾ 0.09 1.2 ⫾ 0.1 0.6 ⫾ 0.05

14.9 ⫾ 2.7 0.6 ⫾ 0.1 1.5 ⫾ 0.3 0.4 ⫾ 0.08 0.5 ⫾ 0.1 0.2 ⫾ 0.1

11.5 ⫾ 1.4 9.3 ⫾ 1.0 10.7 ⫾ 1.3 9.6 ⫾ 1.1 10.2 ⫾ 1.1 9.1 ⫾ 1.0

3.5 ⫾ 1.0 1.5 ⫾ 0.6 2.1 ⫾ 0.7 1.3 ⫾ 0.6 1.0 ⫾ 0.6 1.4 ⫾ 0.9

86.3 ⫾ 3.9d 83.3 ⫾ 2.3 89.5 ⫾ 3.0 81.8 ⫾ 8.4 87.8 ⫾ 6.2 NDe

a

NS-398 (1 ␮g/ml) NS-398 (0.1 ␮g/ml) indomethacin (2 ␮g/ml) indomethacin (0.2 ␮g/ml)

b

c

Values are weighted mean ⫾ SEM for all metabolites measured. Culture media contained 6.6 ⫾ 0.4 ng/ml TXB2. All supernatants contained ⬍10 pg/ml 11␤-PGF2␣, a breakdown product of PGD2. d Values represent the percentage of viable cells (mean ⫾ SD) as determined by trypan blue exclusion following 24-h culture (n ⫽ 4). e ND, Not done. a b c

side light scatter, and the 10,000 cells from remaining population were analyzed for CD45, CD19, CD11b, TCR-␤, and c-kit expression. Gates for positive staining were defined such that ⬍1% of the analyzed population stained positive with the appropriate isotype control Ab. Stastical analysis. Delta Soft 3 software (BioMetallics, Princeton, NJ) was used to determine the weighted mean ⫾ the SEM of each sample. Weighted mean ⫽ ⌺wi conci/⌺wi; conci is the interpolated mean concentration for dilution i; wi ⫽ 1/(SEMi)2, and SEMi is the SE of mean for conci. Data analysis using an unpaired Student’s t test was performed using GraphPad Prism (GraphPad, San Diego, CA).

Results Murine small intestine LP-adherent cells posses a unique phenotype of COX-2 expression and PGE2 production We have recently described an anti-inflammatory/immunomodulatory role for COX-2-dependent AA metabolites produced by the small intestine LP in the response to a dietary Ag (11). In most systems studied to date, COX-2 expression and the subsequent production of AA metabolites are an inducible and transient event. To better understand the role of COX-2-dependent AA metabolites produced by the small intestine LP, we examined the pattern of COX-2-dependent PGE2 production by small intestine LP cells. Small intestine LP cells display spontaneous and abundant COX2-dependent PGE2 production. This phenotype is stable, persisting for days in culture (Fig. 1).

FIGURE 1. Spontaneous and continuous production of COX-2-dependent AA metabolites by small intestine LP cells. Small intestine LP cells from 6- to 12-wk-old C57BL/6 mice were isolated and cultured at a density of 2.5 ⫻ 106 cells/ml as described in Materials and Methods. Culture supernatants were removed daily, and cells were washed with PBS and cultured with fresh medium. Supernatant PGE2 concentration was measured by ELISA as described in Materials and Methods. 䡺, supernatant PGE2 concentrations from C57BL/6 small intestine LP cells cultured in the absence of a COX inhibitor. Every 3rd day (f), C57BL/6 small intestine LP cells were cultured with the selective COX-2 inhibitor NS-398 at 1 ␮g/ml to document the dependency of PGE2 production on COX-2 activity. Results are displayed as the weighted mean of the concentration of PGE2 ⫾ the SEM in picograms per milliliter. Results are representative of one of two experiments. ⴱ, p ⬍ 0.05 when compared with the PGE2 concentration of the supernatant from the previous day.

The mucosal immune system represents a complex organization of bone marrow-derived cells and supporting stromal cells with distinct phenotypes and localization. It is unknown whether the spontaneous production of COX-2-dependent AA metabolites is a phenotype unique to the mucosal immune system and, in particular, the small intestine LP, or whether this is a more global phenotype shared by mucosal and/or nonmucosal lymphoid organs. To address this question, we isolated cells from the small intestine LP, PP, spleen, and peritoneal cavity and evaluated their spontaneous PGE2 production and COX-2 protein expression. As previously demonstrated in Fig. 1, small intestine LP cells produce abundant PGE2, which can be inhibited by the selective COX-2 inhibitor NS-398. The level of PGE2 production in the presence of NS-398 is equivalent to the level of PGE2 production by small intestine LP cells from COX-2⫺/⫺ mice (compare Fig. 1 and Table II). In contrast, SPLs, resident PC, and PP cells produce little spontaneous PGE2 (Fig. 2A). Evaluation of COX-2 protein expression further demonstrates the differences between these cellular populations (Fig. 2B). Small intestine LP cells, in contrast to the other cell populations, demonstrate spontaneous expression of COX-2 by Western blot analysis. These findings suggest that basal COX-2 expression represents a unique property of the small intestine LP when compared with other mucosal and nonmucosal lymphoid compartments. We have previously demonstrated that the production of COX2-dependent PGE2 by the small intestine LP does not require ␣␤ TCR-bearing T cells (11). Additionally, we have noted that recombination activation gene-deficient mice (lacking T cells and B cells) have small intestine LP PGE2 production that is equal to or greater than wild-type mice (data not shown). To better identify the cellular population responsible for the basal COX-2 expression and PGE2 production, we separated cells based on the property of adherence. As shown in Fig. 3A, strongly adherent populations are significantly enriched for PGE2-producing cells. Others have described the activation of adherent cells in response to physical stimuli, including adherence to plastic or phagocytosis of beads (27). To investigate the possibility that adherence to plastic was inducing COX-2 expression in small intestine LP cells, we cultured small intestine LP cells under conditions that would not allow adherence. As shown in Fig. 3B, adherence was not required for production of COX-2-dependent PGE2, and transferring cells from nonadherent to adherent culture conditions did not augment PGE2 production further. To better characterize the profile of COX-dependent AA metabolites produced by this cellular population, strongly adherent LP mononuclear cells (MNC) were isolated and cultured for 24 h in the presence of varying concentrations of the selective COX-2 inhibitor NS-398 and the nonselective COX inhibitor indomethacin. Supernatants were harvested, and COX-dependent AA metabolites were measured as described in Materials and Methods. As

4468


Table II. Small intestine LP cells do not produce factors that can induce PGE2 production by resident PC Cell Culture Conditions Responding Resident PC

Media alone

Supernatant from C57BL/6 SPLs

Supernatant from C57BL/6 LP

Supernatant from COX-2⫺/⫺ SPLs

Supernatant from COX-2⫺/⫺ LP

Media ⫹ LPS (10 ng/ml)

None C3H/HeJ C3H/HeSnJ

455 ⫾ 95a 394 ⫾ 91

42 ⫾ 77 384 ⫾ 91 436 ⫾ 94

8704 ⫾ 484 10296 ⫾ 2476 8169 ⫾ 489

40 ⫾ 70 374 ⫾ 90 402 ⫾ 92

422 ⫾ 57 596 ⫾ 116 752 ⫾ 149

2197 ⫾ 509 46524 ⫾ 9423

a

Results are displayed as the weighted mean of the concentration of PGE2 ⫾ the SEM in picograms per milliliter in supernatants after 48-h culture.

shown in Table I, the predominant metabolite measured in these supernatants is PGE2, and its production is significantly inhibited by the selective COX-2 inhibitor NS-398 as well as the nonselective COX inhibitor indomethacin, indicating that the vast majority of PGE2 production by these cells is COX-2-dependent. This is consistent with our previous findings demonstrating that LP MNC from COX-1⫺/⫺ mice produce PGE2 levels equivalent to wildtype littermates, whereas LP MNC from COX-2⫺/⫺ mice produce significantly reduced PGE2 levels (11), equivalent to those seen when wild-type LP MNC are treated with NS-398. PGF2␣ and 6-keto PGF1␣, a breakdown product of prostacyclin, were present in moderate amounts in these supernatants, and their production was predominantly COX-2-dependent. TXB2 and PGD2 were produced in predominantly COX-2-dependent manners but were

FIGURE 2. Spontaneous expression of COX-2 and production of PGE2 is a unique property of small intestine LP cells. A, Spontaneous PGE2 production. Small intestine LP cells, PP cells, SPLs, and resident PC from B10.BR/SgSnJ mice were isolated and cultured at a density of 2.5 ⫻ 106 viable cells/ml as described in Materials and Methods. After 48 h, supernatants were removed and the PGE2 concentration was measured by ELISA. Results are displayed as the weighted mean of the concentration of PGE2 ⫾ the SEM in picograms per milliliter. Recovery of viable cells after 48 h of culture, as determined by trypan blue exclusion, was 56.7 ⫾ 1.9%, 89.1 ⫾ 1.7%, 40.9 ⫾ 8.0%, and 58.7 ⫾ 2.3% for SPL, PC, PP, and LP, respectively (mean ⫾ SD; n ⫽ 3). Recovery of 250 ng of exogenously added PGE2 from SPL, PC, PP, and LP cells cultured in the presence of 2 ␮g/ml of indomethacin for 48 h was 97.2, 90.9, 106.4, and 96.2%, respectively, when compared with recovery of PGE2 from spiked medium cultured for 48 h. ⴱ, p ⬍ 0.05 when compared with the PGE2 concentration from LP supernatant. B, COX-2 protein expression. Lysates from 3.5 ⫻ 105 cells from each culture (described above), as well as J774 cells stimulated with LPS (positive control), were assayed for COX-2 and actin protein expression by Western blot analysis as described in Materials and Methods. The closed arrow represents the mobility of a 72-kDa marker, and the open arrow represents the mobility of the 44-kDa marker. Results are representative of one of two experiments.

present in more modest concentrations. Notably, the majority of TXB2 measured in these supernatants was present in medium in the absence of cell culture and was not produced by the adherent LP cell population. Supernatants from all groups contained ⬍10 pg/ml 11␤-PGF2␣, a breakdown product of PGD2, consistent with the finding of low PGD2 production by these cellular populations.

FIGURE 3. Strongly adherent small intestine LP cells spontaneously produce COX-2-dependent PGE2. A, Small intestine LP cells from C57BL/6 mice were isolated and separated into nonadherent, adherent 1 h, and adherent overnight populations as described in Materials and Methods. These populations represented 52.5 ⫾ 11.6%, 31.0 ⫾ 7.8%, and 16.7 ⫾ 4.7%, respectively, of the total viable isolated cell population (n ⫽ 4). Viable cells from each population were cultured at a density of 2.5 ⫻ 106 cells/ml for 48 h after the initial 12-h isolation procedure. Supernatant PGE2 concentration was measured by ELISA as described in Materials and Methods. Results are displayed as the weighted mean of the concentration of PGE2 ⫾ the SEM in picograms per milliliter and are representative of one of two experiments. ⴱ, p ⫽ ⬍0.05 for PGE2 concentration when compared with the supernatant from the adherent overnight population. B, Adherence is not required to induce PGE2 production by small intestine LP cells. Small intestine LP cells from C57BL/6 mice were isolated and cultured in either polystyrene (adherent) or polypropylene (nonadherent) culture tubes. Nonadherence to polypropylene and adherence to polystyrene was confirmed by light microscopy. After 18 h, supernatants were removed and PGE2 concentration was measured by ELISA. Nonadherent cellular populations were recultured in polystyrene culture tubes (nonadherent/adherent) for an additional 18 h, and then supernatants were removed and PGE2 concentration was measured by ELISA. Results are displayed as the weighted mean of the concentration of PGE2 ⫾ the SEM in picograms per milliliter.

The Journal of Immunology

4469

FIGURE 4. Non-bone marrow-derived stromal cells are the predominant COX-2-expressing population in the small intestine LP. Adherent small intestine LP cells (A–C) or frozen sections of small intestine from C57BL/6 mice (D–F) were prepared as described in the Materials and Methods and stained for COX-2 expression (red fluorescence, A, B, D, and E), CD45 expression (green fluorescence, A and D), and CD11b expression (green fluorescence, B and E). Controls for Ab specificity (C and F) were stained with anti-COX-2 Ab preincubated with blocking COX-2 peptide, followed by detection as described in Materials and Methods. All sections were counterstained with Hoescht dye (blue fluorescence) to illustrate nuclei. The absence of colocalization of red and green fluorescence indicates that the majority of COX-2-expressing cells are CD45⫺ and CD11b⫺. Colocalization of COX-2 and CD11b expression (inset panel, B) and CD45 expression (data not shown) could be seen. These cells most closely resemble macrophages (based on morphology, adherence, and cell surface marker expression) and were not the predominant COX-2-expressing cell type. Inset panel A, CD45⫹ COX-2⫺ staining in adherent LP MNC. COX-2⫹ cells, COX-2 and CD45 double positive cells, and COX-2 and CD11b double positive cells represented 51.4 ⫾ 2.8%, 3.7 ⫾ 1.3%, and 4.7 ⫾ 1.9% of the adherent cell population, respectively.

By FACS analysis, this adherent LP cell population was comprised of cells that were 21% CD45⫹, 11% CD11b⫹, 1% TCR-␤⫹, 1% CD19⫹, and ⬍1% c-kit⫹. Non-bone marrow-derived stromal cells are the predominant COX-2-expressing cell types in the murine LP To further characterize the cellular population responsible for the basal expression of COX-2, immunohistochemical analysis of COX-2 expression on LP-adherent cell populations was performed. As shown in Fig. 4, A and B, the predominant population of cells expressing COX-2 were both CD45⫺, a cell surface marker of bone marrow-derived cells, and CD11b⫺, a cell surface marker expressed on macrophages and dendritic cells that could be enriched by adherence. Occasional CD11b⫹ COX-2-expressing cells were seen (inset, Fig. 4B); however, these cells were not the predominant population of COX-2-expressing cells. By immunohistochemical analysis of the adherent population, we found that 51.4 ⫾ 2.8% of nucleated cells expressed COX-2, 3.7 ⫾ 1.3% were both CD45⫹ and COX-2⫹, and 4.7 ⫾ 1.9% were both CD11b⫹ and COX-2⫹. Immunohistochemical analysis of ␣-smooth muscle actin expression was variable in the cultured LP-adherent cell population, with 36.6 ⫾ 23.9% of COX-2⫹ cells expressing ␣-smooth muscle actin. These in vitro findings were further confirmed by immunohistochemical analysis of murine small intestine. As shown in Fig. 4, D and E, COX-2-expressing cells were present in the LP of small intestine villi. These COX-2-expressing cells were preferentially located near the tips of the villi and were predominantly CD45⫺ and CD11b⫺, consistent with the immunohistochemical findings

seen on cultured cells described above. By immunohistochemical analysis, these COX-2-expressing cells did not express ␣-smooth muscle actin. This finding, along with the location of these cells in the villus, suggests that these cells are not subepithelial myofibroblasts (28, 29). These findings demonstrate that the population of COX-2expressing cells in the small intestine LP is a non-bone marrow-derived stromal cell residing adjacent to lymphocytes and monocytes. COX-2 expression and PGE2 production by LP cells does not require exogenous stimuli LP MNC reside in an environment exposing them to multiple inflammatory stimuli, such as LPS and other bacterial components, and mediators of inflammation, such as TNF-␣, IFN-␥, and IL-1␤. These stimuli have been noted to induce the expression of COX-2 in other systems (30 –33). Therefore, we evaluated the requirement for these inflammatory stimuli in the production of PGE2 by small intestine LP cells through the examination of various spontaneous and induced mutant strains of mice, as well as germfree and conventionally housed mice. As shown in Fig. 5A, small intestine LP cells from germfree FVB/NJ mice and conventionally housed FVB/NJ mice produce equivalent levels of PGE2, indicating that the presence of resident bacterial flora is not required to develop or express this phenotype. This finding was expanded and confirmed by examining PGE2 production by small intestine LP cells from C3H/HeJ mice, which contain a spontaneous mutation in the Tolllike receptor 4 gene, resulting in the inability to respond to LPS via the CD14 receptor (34). Small intestine LP cells from C3H/HeJ mice produce levels of PGE2 equivalent to those of wild-type mice (Fig. 5). Notably, PC from C3H/HeJ mice produce ⬍5% of the

4470

COX-2-DEPENDENT PGE2 PRODUCTION IN THE MOUSE SMALL INTESTINE LP cells do not produce a soluble factor with the ability to induce PGE2 production by peritoneal macrophages To address the possibility that a soluble stimulus produced by LP cells induces COX-2 expression and PGE2 production, we evaluated the ability of small intestine LP cell- and SPL-conditioned medium to induce PGE2 production by resident PC, a well studied cell for the induction of COX-2. Conditioned medium was generated using SPLs and small intestine LP cells from wild-type C57BL/6 mice, as well as COX-2⫺/⫺ mice, to reduce background levels of PGE2 in the conditioned medium. To minimize any effect that LPS contaminating the cellular preparations might have on the induction of COX-2 in our responding resident PC, we used resident PC from C3H/HeJ (LPS-nonresponsive) as well as congenic C3H/HeSnJ (LPS-responsive) mice as responding cell populations. Conditioned medium generated from culture of SPLs or LP MNC from C57BL/6 and COX-2⫺/⫺ mice was diluted 1:1 with fresh medium and cultured in the absence of cells (no responder cells) to assess the basal level of PGE2 in the conditioned medium or used to culture resident PC from C3H/HeJ (LPS-nonresponsive) or C3H/HeSnJ (LPS-responsive) mice at a density of 2.5 ⫻ 106/ml for 48 h. After 48 h, supernatants were removed, and the PGE2 concentration was measured by ELISA. As shown in Table II, SPL- and small intestine LP cell-conditioned medium does not contain factors with the ability to induce PGE2 production by resident PC. These findings further suggest that the phenotype of spontaneous COX-2 expression and abundant PGE2 production is not an acute response to a stimulus, but represents a stable phenotype unique to the murine small intestine LP.

Discussion FIGURE 5. Small intestine LP cells do not require bacterial flora, LPS, TNF-␣, IFN-␥, IL-1␤, or IL-12 to produce abundant PGE2. Small intestine LP cells from conventionally housed FVB/NJ mice, germfree housed FVB/NJ mice, conventionally housed C3H/HeJ and C3H/HeSnJ mice (A), and from conventionally housed TNFR I⫺/⫺, TNFR II⫺/⫺, IFN-␥R⫺/⫺, IL-1␤⫺/⫺, IL-12 p40⫺/⫺, and C57BL/6 mice (B) were isolated and cultured as described in Materials and Methods (n ⫽ two mice from each group). Supernatants were removed every 12 h and the PGE2 concentration was measured by ELISA as described in Materials and Methods. Small intestine LP cells from 129/SvImJ mice (controls for IFN-␥R⫺/⫺ mice) produced PGE2 levels equivalent to those from C57BL/6 mice (data not shown for clarity). Results are displayed as the weighted mean of the concentration of PGE2 ⫾ the SEM in picograms per milliliter.

PGE2 when compared with PC from the congenic C3H/HeSnJ (LPS-responsive) mice in response to 10 ng/ml LPS (Table II). This result documents the requirement for LPS signaling through the Toll-like receptor 4 complex for the optimal induction of COX2-dependent PGs in response to LPS in peritoneal macrophages, but not for the continuous production of PGE2 by LP stromal cells. To assess the requirement for inflammatory cytokines in the development and expression of this phenotype, we examined PGE2 production by small intestine LP cells from IL-1␤-deficient (IL-1␤⫺/⫺), IFN-␥R-deficient (IFN-␥R⫺/⫺), TNFR I-deficient (TNFR I⫺/⫺), TNFR II-deficient (TNFR II⫺/⫺), IL-12-deficient (IL-12 p40⫺/⫺), and C57BL/6 mice. As shown is Fig. 5B, all induced mutant strains of mice tested had spontaneous PGE2 production that was equivalent to wild-type mice. These findings convincingly document that bacterial flora, LPS, IL-1␤, IFN-␥, TNF, and IL-12 are not essential for the development or expression of this phenotype.

The gastrointestinal immune system is continuously exposed to luminal Ags in the form of normal microbial flora and nonpathogenic dietary Ags. Crucial to normal gastrointestinal function is the ability of the gastrointestinal immune system to avoid an inflammatory response to nonpathogenic luminal Ags, despite the concurrent exposure to bacterial stimuli known to induce inflammatory responses in more traditional lymphoid organs. Given the context in which the gastrointestinal immune system resides, it is not surprising that the basal tone of the immune response in the intestine favors the production of “anti-inflammatory” cytokines such as IL-10 (35) and TGF-␤ (36), whereas pathogenic states such as inflammatory bowel disease are associated with the production of inflammatory cytokines IL-12, TNF-␣, and IFN-␥ (37– 39). In addition to these observations, we have recently described a role for a COX-2-dependent AA metabolite, PGE2, produced by the small intestine LP in the immune response to a dietary Ag (11). PGE2 is known to have immunomodulatory effects including down-regulation of MHC class II (40) and IL-12R (41) expression, increased production of IL-10 (42), and decreased production of TNF-␣ (43) and IL-12 (44). In this context, PGE2 production by the LP could represent a proximal step in the development of the immunomodulatory tone of the intestine. Here we demonstrate the spontaneous and sustained expression of COX-2 and the resultant production of abundant PGE2 by LP cells in the absence of a known exogenous stimulus. The spontaneous production of COX-2-dependent PGE2 is a unique property of the LP when compared with other lymphoid compartments, including intestinal lymphoid compartments exposed to identical or nearly identical luminal stimuli. We have observed a modest number of COX-2-expressing cells in the subepithelial dome of the PP by immunohistochemistry (our unpublished observations). This observation could explain the low levels of COX-2 expression and PGE2 production seen in PP (Fig. 2). In addition, the pattern of

The Journal of Immunology COX-2 expression in the LP contrasts with the majority of published descriptions of COX-2 expression in other organs, which demonstrate inducible and transient expression in response to an exogenous stimulus. To better understand a potential role for the spontaneous and continuous expression of COX-2 by LP stromal cells, we examined the profile of COX-dependent AA metabolites produced by these cells. Here we demonstrate that the predominant COX-dependent AA metabolite is PGE2, and its production is overwhelmingly COX-2-dependent. Other COX-dependent AA metabolites (PGF2␣, 6-keto PGF1␣, TXB2, and PGD2) were also produced in a COX-2-dependent manner but in more modest amounts. This finding, taken in the context of the previously described immunomodulatory roles of PGE2, suggests that a function of these stromal cells may be to contribute to the immunologic hyporesponsiveness of the intestinal immune response. Given the specialized microenvironment in which LP cells reside, it is intriguing to hypothesize that agents such as LPS or other bacterial products may play a role in the development or expression of this unique phenotype. In this study, we demonstrate that neither bacterial flora nor conventional LPS signaling through the Toll-like receptor 4 are required for the development of small intestine LP cells with this phenotype. In contrast, signaling via Tolllike receptor 4 is required to induce COX-2 expression in peritoneal macrophages in response to LPS (Table II). We also document that other proinflammatory stimuli known to induce COX-2 expression are not essential for the development or expression of this phenotype. It is possible that the lack of individual proinflammatory stimuli in the knockout mice may not lead to loss of PGE2 production due to a redundancy in the inflammatory stimuli inducing COX-2 in this system. However, we further demonstrate that small intestine LP cells do not produce a soluble factor with the ability to induce COX-2-dependent PGE2 production in resident peritoneal macrophages, suggesting that if redundant stimuli are responsible for the COX-2 expression in this system, they are not the typical inflammatory mediators known to induce COX-2 in peritoneal macrophages. Together, these findings suggest that the abundant PGE2 production by these LP cells does not represent an acute and transient response to an inflammatory or secreted stimulus. We have shown that the basal production of COX-2-dependent PGE2 in the small intestine LP is a phenotype of non-bone marrow-derived stromal cells. These COX-2-expressing stromal cells are preferentially located near the tips of the villi and in the proximal small intestine (our unpublished observations). Based on the location of these stromal cells in the villi, their lack of ␣-smooth muscle actin expression by immunohistochemistry of the small intestine, and their spontaneous expression of COX-2, these cells do not resemble previously described intestinal myofibroblasts (28, 29, 45, 46). Multiple studies have documented a tolerant phenotype of LP lymphocytes in response to antigenic stimuli. This phenotype is manifest by decreased production of proinflammatory cytokines, increased production of immunomodulatory cytokines, and predominance of immunoregulatory Th cell subsets. Given previous descriptions of the ability of PGE2 to blunt the production of proinflammatory cytokines, enhance the production of immunomodulatory cytokines, and favor the development of immunomodulatory Th cell subsets, it is intriguing to postulate that a factor in the development of the immunological tolerance in the intestine is the production of PGE2 by resident stromal cells in the LP. In this context, it is logical that the production of PGE2 would be continuous and the source would be a cell that is intrinsic to the intestine and located adjacent to responding lymphocytes. These res-

4471 ident cells, through their production of AA metabolites, could help dictate the tone of the intestinal immune response produced by circulating lymphocytes and monocytes. The physical convergence of bacteria, bacterial products, dietary Ags, and activated components of the cellular immune system in the gastrointestinal tract presents an unusual challenge to the mucosal immune system. The description of the immunomodulatory role of COX-2-dependent AA metabolites in the intestinal immune response adds another agent to the list of mechanisms used by the mucosal immune system. Understanding how these mechanisms allow the mucosal immune system to negotiate this continual challenge will broaden our understanding of basic immunology as well as yield insight into pathophysiologic conditions such as inflammatory bowel disease and celiac disease.

Acknowledgments We thank E. Newberry, S. Amadeus, and the members of the St. Louis Institute of Mucosal Immunology for assistance and advice with manuscript preparation.

References 1. McCracken, V. J., and R. G. Lorenz. 2001. The gastrointestinal ecosystem: a precarious alliance among epithelium, immunity and microbiota. Cell. Microbiol. 3:1. 2. Kontoyiannis, D., M. Pasparakis, T. T. Pizarro, F. Cominelli, and G. Kollias. 1999. Impaired on/off regulation of TNF biosynthesis in mice lacking TNF AUrich elements: implications for joint and gut-associated immunopathologies. Immunity. 10:387. 3. MacDonald, T. T., and J. Spencer. 1988. Evidence that activated mucosal T cells play a role in the pathogenesis of enteropathy in human small intestine. J. Exp. Med. 167:1341. 4. MacDonald, T. T., G. Monteleone, and S. L. Pender. 2000. Recent developments in the immunology of inflammatory bowel disease. Scand. J. Immunol. 51:2. 5. Bhan, A. K., E. Mizoguchi, R. N. Smith, and A. Mizoguchi. 1999. Colitis in transgenic and knockout animals as models of human inflammatory bowel disease. Immunol. Rev. 169:195. 6. Papadakis, K. A., and S. R. Targan. 2000. Role of cytokines in the pathogenesis of inflammatory bowel disease. Annu. Rev. Med. 51:289. 7. Nagler-Anderson, C. 2000. Tolerance and immunity in the intestinal immune system. Crit. Rev. Immunol. 20:103. 8. Asseman, C., S. Mauze, M. W. Leach, R. L. Coffman, and F. Powrie. 1999. An essential role for interleukin 10 in the function of regulatory T cells that inhibit intestinal inflammation. J. Exp. Med. 190:995. 9. Groux, H., and F. Powrie. 1999. Regulatory T cells and inflammatory bowel disease. Immunol. Today 20:442. 10. Groux, H., A. O’Garra, M. Bigler, M. Rouleau, S. Antonenko, J. E. de Vries, and M. G. Roncarolo. 1997. A CD4⫹ T-cell subset inhibits antigen-specific T-cell responses and prevents colitis. Nature 389:737. 11. Newberry, R. D., W. F. Stenson, and R. G. Lorenz. 1999. Cyclooxygenase-2dependent arachidonic acid metabolites are essential modulators of the intestinal immune response to dietary antigen. Nat. Med. 5:900. 12. Smith, W. L., D. L. DeWitt, and R. M. Garavito. 2000. Cyclooxygenases: structural, cellular, and molecular biology. Annu. Rev. Biochem. 69:145. 13. Smith, W. L., R. M. Garavito, and D. L. DeWitt. 1996. Prostaglandin endoperoxide H synthases (cyclooxygenases)-1 and -2. J. Biol. Chem. 271:33157. 14. Smith, W. L., and D. L. Dewitt. 1996. Prostaglandin endoperoxide H synthases-1 and -2. Adv. Immunol. 62:167. 15. Morham, S. G., R. Langenbach, C. D. Loftin, H. F. Tiano, N. Vouloumanos, J. C. Jennette, J. F. Mahler, K. D. Kluckman, A. Ledford, C. A. Lee, et al. 1995. Prostaglandin synthase 2 gene disruption causes severe renal pathology in the mouse. Cell 83:473. 16. Langenbach, R., S. G. Morham, H. F. Tiano, C. D. Loftin, B. I. Ghanayem, P. C. Chulada, J. F. Mahler, C. A. Lee, E. H. Goulding, K. D. Kluckman, et al. 1995. Prostaglandin synthase 1 gene disruption in mice reduces arachidonic acidinduced inflammation and indomethacin-induced gastric ulceration. Cell 83:483. 17. Harris, R. C., J. A. McKanna, Y. Akai, H. R. Jacobson, R. N. Dubois, and M. D. Breyer. 1994. Cyclooxygenase-2 is associated with the macula densa of rat kidney and increases with salt restriction. J. Clin. Invest. 94:2504. 18. Lukiw, W. J., and N. G. Bazan. 1997. Cyclooxygenase 2 RNA message abundance, stability, and hypervariability in sporadic Alzheimer neocortex. J. Neurosci. Res. 50:937. 19. Eberhart, C. E., R. J. Coffey, A. Radhika, F. M. Giardiello, S. Ferrenbach, and R. N. DuBois. 1994. Up-regulation of cyclooxygenase 2 gene expression in human colorectal adenomas and adenocarcinomas. Gastroenterology 107:1183. 20. Sano, H., Y. Kawahito, R. L. Wilder, A. Hashiramoto, S. Mukai, K. Asai, S. Kimura, H. Kato, M. Kondo, and T. Hla. 1995. Expression of cyclooxygenase-1 and -2 in human colorectal cancer. Cancer Res. 55:3785. 21. Allgayer, H., K. Deschryver, and W. F. Stenson. 1989. Treatment with 16,16⬘dimethyl prostaglandin E2 before and after induction of colitis with trinitrobenzenesulfonic acid in rats decreases inflammation. Gastroenterology 96:1290.

4472


22. Gilroy, D. W., P. R. Colville-Nash, D. Willis, J. Chivers, M. J. Paul-Clark, and D. A. Willoughby. 1999. Inducible cyclooxygenase may have anti-inflammatory properties. Nat. Med. 5:698. 23. Huang, S., W. Hendriks, A. Althage, S. Hemmi, H. Bluethmann, R. Kamijo, J. Vilcek, R. M. Zinkernagel, and M. Aguet. 1993. Immune response in mice that lack the interferon-␥ receptor. Science 259:1742. 24. Shornick, L. P., P. De Togni, S. Mariathasan, J. Goellner, J. Strauss-Schoenberger, R. W. Karr, T. A. Ferguson, and D. D. Chaplin. 1996. Mice deficient in IL-1␤ manifest impaired contact hypersensitivity to trinitrochlorobenzone. J. Exp. Med. 183:1427. 25. Matsumoto, M., S. Mariathasan, M. H. Nahm, F. Baranyay, J. J. Peschon, and D. D. Chaplin. 1996. Role of lymphotoxin and the type I TNF receptor in the formation of germinal centers. Science 271:1289. 26. Futaki, N., S. Takahashi, M. Yokoyama, I. Arai, S. Higuchi, and S. Otomo. 1994. NS-398, a new anti-inflammatory agent, selectively inhibits prostaglandin G/H synthase/cyclooxygenase (COX-2) activity in vitro. Prostaglandins 47:55. 27. Shibata, Y. 1995. Prostaglandin E2 release triggered by phagocytosis of latex particles: a distinct association with prostaglandin synthase isozymes in bone marrow macrophages. J. Immunol. 154:2878. 28. Powell, D. W., R. C. Mifflin, J. D. Valentich, S. E. Crowe, J. I. Saada, and A. B. West. 1999. Myofibroblasts. II. Intestinal subepithelial myofibroblasts. Am. J. Physiol. 277:C183. 29. Valentich, J. D., V. Popov, J. I. Saada, and D. W. Powell. 1997. Phenotypic characterization of an intestinal subepithelial myofibroblast cell line. Am. J. Physiol. 272:C1513. 30. Topley, N., M. M. Petersen, R. Mackenzie, A. Neubauer, E. Stylianou, V. Kaever, M. Davies, G. A. Coles, A. Jorres, and J. D. Williams. 1994. Human peritoneal mesothelial cell prostaglandin synthesis: induction of cyclooxygenase mRNA by peritoneal macrophage-derived cytokines. Kidney Int. 46:900. 31. Arias-Negrete, S., K. Keller, and K. Chadee. 1995. Proinflammatory cytokines regulate cyclooxygenase-2 mRNA expression in human macrophages. Biochim. Biophys. Acta 208:582. 32. Berenbaum, F., C. Jacques, G. Thomas, M. T. Corvol, G. Bereziat, and J. Masliah. 1996. Synergistic effect of interleukin-1␤ and tumor necrosis factor ␣ on PGE2 production by articular chondrocytes does not involve PLA2 stimulation. Exp. Cell. Res. 222:379. 33. Pang, L., and A. J. Knox. 1997. Effect of interleukin-1␤, tumour necrosis factor-␣ and interferon-␥ on the induction of cyclo-oxygenase-2 in cultured human airway smooth muscle cells. Br. J. Pharmacol. 121:579. 34. Qureshi, S. T., L. Lariviere, G. Leveque, S. Clermont, K. J. Moore, P. Gros, and D. Malo. 1999. Endotoxin-tolerant mice have mutations in Toll-like receptor 4

35.

36.

37.

38.

39.

40. 41.

42.

43.

44.

45.

46.

(Tlr4). [Published erratum appears in 1999 J. Exp. Med. 189:1518.] J. Exp. Med. 189:615. Braunstein, J., L. Qiao, F. Autschbach, G. Schurmann, and S. Meuer. 1997. T cells of the human intestinal lamina propria are high producers of interleukin-10. Gut 41:215. Gonnella, P. A., Y. Chen, J. Inobe, Y. Komagata, M. Quartulli, and H. L. Weiner. 1998. In situ immune response in gut-associated lymphoid tissue (GALT) following oral antigen in TCR-transgenic mice. J. Immunol. 160:4708. Fuss, I. J., M. Neurath, M. Boirivant, J. S. Klein, C. de la Motte, S. A. Strong, C. Fiocchi, and W. Strober. 1996. Disparate CD4⫹ lamina propria (LP) lymphokine secretion profiles in inflammatory bowel disease: Crohn’s disease LP cells manifest increased secretion of IFN-␥, whereas ulcerative colitis LP cells manifest increased secretion of IL-5. J. Immunol. 157:1261. Reimund, J. M., C. Wittersheim, S. Dumont, C. D. Muller, R. Baumann, P. Poindron, and B. Duclos. 1996. Mucosal inflammatory cytokine production by intestinal biopsies in patients with ulcerative colitis and Crohn’s disease. J. Clin. Immunol. 16:144. Monteleone, G., L. Biancone, R. Marasco, G. Morrone, O. Marasco, F. Luzza, and F. Pallone. 1997. Interleukin 12 is expressed and actively released by Crohn’s disease intestinal lamina propria mononuclear cells. Gastroenterology 112:1169. Snyder, D. S., D. I. Beller, and E. R. Unanue. 1982. Prostaglandins modulate macrophage Ia expression. Nature 299:163. Wu, C. Y., K. Wang, J. F. McDyer, and R. A. Seder. 1998. Prostaglandin E2 and dexamethasone inhibit IL-12 receptor expression and IL-12 responsiveness. J. Immunol. 161:2723. Kalinski, P., C. M. Hilkens, A. Snijders, F. G. Snijdewint, and M. L. Kapsenberg. 1997. IL-12-deficient dendritic cells, generated in the presence of prostaglandin E2, promote type 2 cytokine production in maturing human naive T helper cells. J. Immunol. 159:28. Demeure, C. E., L. P. Yang, C. Desjardins, P. Raynauld, and G. Delespesse. 1997. Prostaglandin E2 primes naive T cells for the production of antiinflammatory cytokines. Eur. J. Immunol. 27:3526. van der Pouw Kraan, T. C., L. C. Boeije, R. J. Smeenk, J. Wijdenes, and L. A. Aarden. 1995. Prostaglandin-E2 is a potent inhibitor of human interleukin 12 production. J. Exp. Med. 181:775. Powell, D. W., R. C. Mifflin, J. D. Valentich, S. E. Crowe, J. I. Saada, and A. B. West. 1999. Myofibroblasts. I. Paracrine cells important in health and disease. Am. J. Physiol. 277:C1. Hinterleitner, T. A., J. I. Saada, H. M. Berschneider, D. W. Powell, and J. D. Valentich. 1996. IL-1 stimulates intestinal myofibroblast COX gene expression and augments activation of Cl-secretion in T84 cells. Am. J. Physiol. 271:C1262.