Intestinal P-glycoprotein exports endocannabinoids to prevent inflammation and maintain homeostasis

Abstract

Neutrophil influx into the intestinal lumen is a critical response to infectious agents, but is also associated with severe intestinal damage observed in idiopathic inflammatory bowel disease. The chemoattractant hepoxilin A3, an eicosanoid secreted from intestinal epithelial cells by the apically restricted efflux pump multidrug resistance protein 2 (MRP2), mediates this neutrophil influx. Information about a possible counterbalance pathway that could signal the lack of or resolution of an apical inflammatory signal, however, has yet to be described. We now report a system with such hallmarks. Specifically, we identify endocannabinoids as the first known endogenous substrates of the apically restricted multidrug resistance transporter P-glycoprotein (P-gp) and reveal a mechanism, which we believe is novel, for endocannabinoid secretion into the intestinal lumen. Knockdown or inhibition of P-gp reduced luminal secretion levels of N-acyl ethanolamine-type endocannabinoids, which correlated with increased neutrophil transmigration in vitro and in vivo. Additionally, loss of CB2, the peripheral cannabinoid receptor, led to increased pathology and neutrophil influx in models of acute intestinal inflammation. These results define a key role for epithelial cells in balancing the constitutive secretion of antiinflammatory lipids with the stimulated secretion of proinflammatory lipids via surface efflux pumps in order to control neutrophil infiltration into the intestinal lumen and maintain homeostasis in the healthy intestine.

Keywords: Gastroenterology; Neutrophils.

Figure 1. HxA3 drives inflammation during DSS colitis.

(A) Mucosal scrapings from C57BL/6 WT mice (n = 10) treated with 3% DSS for 7 days were enriched for lipids, and the amount of HxA3 was quantified by LC-MS/MS. *P = 0.050. (B–I) C57BL/6 mice were treated with 3% DSS for 7 days and sacrificed at day 9. Starting at day 4, daily rectal administration of PBS control (vehicle) or 1 mM probenecid conjugate was performed. **P = 0.001, Mann-Whitney 1-tailed nonparametric U test. All data are shown as mean ± SEM. n = 10 mice per group. (C and D) Paraffin-embedded sections of mid and distal colon were stained for H&E and scored by a trained investigator blinded to sample identity. *P = 0.026. Arrow highlights accumulation of neutrophils in intestinal lumen. Original magnification, ×20. (E and F) MPO activity was measured by ADHP assay over 8 minutes from (E) feces (*P = 0.044) or (F) colonic tissue and slopes calculated by linear regression (P = 0.362). For tissue, slopes were normalized to total protein content. (G–I) Total lamina propria leukocytes were isolated and stained for flow cytometry. Neutrophils were characterized as live/CD45⁺CD11b^hiLy6G⁺. (G) Percentage of neutrophils (NS, P = 0.193), (H) number of neutrophils (NS, P = 0.259), and (I) representative plots of neutrophils in colon tissue.

Figure 2. Epithelial cells secrete P-gp–dependent ECs that inhibit neutrophil migration.

(A) Supernatants from T84 epithelial monolayers were enriched for lipids and tested for the ability to inhibit HxA3-induced migration in a 96-well modified Boyden chamber assay. In order to compare across experiments with different donors, migration values within individual experiments were normalized to enriched HxA3 with vehicle treatment. For A–C, data are shown as mean ± SEM of 3 independent experiments. *P < 0.05: ***P < 0.01, 1-way ANOVA. (B) Performed as in A, with supernatants from cell lines expressing different shRNA constructs to knock down P-gp expression (B4-MDR1 and B5-MDR1). (C) Performed as in A, but prior to use in the migration assays, enriched T84 supernatants were pretreated with FAAH or MAGL at 37°C for 30 minutes. (D) Lipid-extracted T84 supernatants from control or 2 separate MDR1-knockdown cell lines (B4 and B5) were subjected to electrospray ionization MS. AEA-d8 was included in each sample to allow quantitative comparisons between samples and was used to calculate relative units of each EC. Individual ECs were normalized to the relative units in the scrambled control condition.

Figure 3. P-gp–dependent apical secretion of ECs inhibits neutrophil migration.

(A) Commercially available ECs and related compounds were tested in the 96-well migration assay, with AEA, OEA, and α-LEA significantly suppressing neutrophil movement. Compounds were used at the highest concentration at which they were soluble in PBS. Data are shown as mean ± SEM of at least 3 independent experiments. *P < 0.05; **P < 0.01, 1-way ANOVA. (B) Confluent 2D human intestinal organoid cultures were used to examine apical secretion of AEA over a 6-hour time course with and without 50 μM verapamil (apical untreated, apical with verapamil), with levels being reported as average of triplicate measurements (± SD) normalized to time 0 values. Increasing apical compartment levels of AEA over time were suppressed by the presence of 50 μM verapamil. Data are shown as mean ± SD from 3 independent experiments with 12–24 technical replicates in each experiment. *P < 0.01; **P < 0.001, unpaired t test. (C) The levels of AEA present in the basal compartment of confluent 2D human intestinal organoid cultures remained relatively constant over this same 6-hour time course, but were reduced in cultures treated with 50 μM verapamil. Relative changes in the basal compartment AEA levels were measured at 0 and 6 hours, with (white bars) and without (black bars) 50 μM verapamil addition. AEA levels are reported as normalized to time 0 values. Data are shown as mean ± SD from 3 independent experiments with 12–24 technical replicates in each experiment. *P < 0.01, unpaired t test.

Figure 4. AMEND is present in mouse intestine.

(A) Colonic scrapings from 5 WT or Mdr1a^–/– FVB mice were pooled, enriched for lipids, and tested in the 96-well migration assay as in Figure 2. Data are shown as mean ± SEM from 3 independent experiments. **P < 0.01, 1-way ANOVA. (B) Indicated sample was pretreated with FAAH for 30 minutes at 37°C. *P = 0.01. (C) Intestinal mucosal scrapings from WT and Mdr1a^–/– FVB mice were lipid extracted (modified Folch method) and analyzed by MS. Mice were pooled in groups of 3. Data shown are from 3 pooled groups. Semiquantitative analysis was performed as in Figure 2.

Figure 5. Signaling through CB2 contributes to AMEND inhibition of neutrophil migration.

Neutrophils were isolated from the BM of WT or Cnr2^–/– (CB2KO) mice and allowed to migrate across Salmonella-infected epithelial monolayers. SL, Salmonella. (A) Neutrophil migration was tested in the presence of enriched AMEND (AM). *P < 0.05. (B) Neutrophil migration was tested in the presence of purified ECs. *P < 0.05; **P < 0.01. Data are shown as mean ± SD from 3 independent experiments. Comparison of groups was performed by Mann-Whitney nonparametric U test.

Figure 6. CB2-deficient mice are vulnerable to severe intestinal inflammation with increased neutrophil transmigration.

WT or Cnr2^–/– mice were treated with 3% DSS for 7 days and sacrificed at day 9. (A) Weights are shown as percentage of day 0 (D0) weight. n = 25 WT; n = 24 Cnr2^–/– mice. **P < 0.01; ***P < 0.001. (B and C) Histopathology of mid and distal colon as in Figure 1. Arrows highlight accumulation of neutrophils in intestinal lumen. *P = 0.021. Original magnification, ×20. (D) Fecal MPO activity. n = 14 WT; n = 12 Cnr2^–/– mice. **P = 0.008. (E) Tissue MPO activity. n = 13 WT; n = 11 Cnr2^–/– mice. NS, P = 0.471. (F) Number (NS, P = 0.482) and (G) percentage (NS, P = 0.095) of tissue neutrophils by flow cytometry analysis. (H) Representative plots of lamina propria neutrophils. (F–H) n = 13 WT; n = 12 Cnr2^–/– mice. (I–K) WT or Cnr2^–/– mice were lethally irradiated and reconstituted with homologous or heterologous BM (donor→recipient). Following reconstitution for 7 weeks, mice were treated with DSS as above. (I) Weight curve. (J) Colon length. (K) Histopathology score. n = 7 WT→WT; n = 9 KO→WT; n = 9 WT→KO; n = 7 KO→KO mice. For all experiments, data are shown as mean ± SEM. Statistical analysis was performed with Mann-Whitney 1-tailed nonparametric U test. Histopathology scoring, MPO activity measurement, and flow cytometry analyses were performed as in Figure 1.

Figure 7. The antiinflammatory P-gp/EC and proinflammatory MRP2/HxA3 pathways in the intestinal epithelium.

In the homeostatic intestine, P-gp secretes ECs from the epithelial surface. Secreted NAEs act through the CB2 receptor on neutrophils to inhibit migration and maintain an antiinflammatory state. During inflammation, P-gp is downregulated and the MRP2/HxA3 pathway is activated. Phospholipase A2 liberates arachidonic acid from the membrane, and it is converted to HxA3 and secreted into the lumen via surface MRP2. HxA3 forms a concentration gradient that attracts neutrophils across the epithelial layer into the lumen, where they cause inflammatory damage and pathology. Of note, FAAH metabolism of NAEs yields arachidonic acid and may also feed into the proinflammatory MRP2/HxA3 pathway.