Intestinal alkaline phosphatase inhibits the proinflammatory nucleotide uridine diphosphate

Angela K. Moss, Sulaiman R. Hamarneh, Mussa M. Rafat Mohamed, Sundaram Ramasamy, Halim Yammine, Palak Patel, Kanakaraju Kaliannan, Sayeda N. Alam, Nur Muhammad, Omeed Moaven, Abeba Teshager, Nondita S. Malo, Sonoko Narisawa, José Luis Millán, H. Shaw Warren, Elizabeth Hohmann, Madhu S. Malo, Richard A. Hodin


Uridine diphosphate (UDP) is a proinflammatory nucleotide implicated in inflammatory bowel disease. Intestinal alkaline phosphatase (IAP) is a gut mucosal defense factor capable of inhibiting intestinal inflammation. We used the malachite green assay to show that IAP dephosphorylates UDP. To study the anti-inflammatory effect of IAP, UDP or other proinflammatory ligands (LPS, flagellin, Pam3Cys, or TNF-α) in the presence or absence of IAP were applied to cell cultures, and IL-8 was measured. UDP caused dose-dependent increase in IL-8 release by immune cells and two gut epithelial cell lines, and IAP treatment abrogated IL-8 release. Costimulation with UDP and other inflammatory ligands resulted in a synergistic increase in IL-8 release, which was prevented by IAP treatment. In vivo, UDP in the presence or absence of IAP was instilled into a small intestinal loop model in wild-type and IAP-knockout mice. Luminal contents were applied to cell culture, and cytokine levels were measured in culture supernatant and intestinal tissue. UDP-treated luminal contents induced more inflammation on target cells, with a greater inflammatory response to contents from IAP-KO mice treated with UDP than from WT mice. Additionally, UDP treatment increased TNF-α levels in intestinal tissue of IAP-KO mice, and cotreatment with IAP reduced inflammation to control levels. Taken together, these studies show that IAP prevents inflammation caused by UDP alone and in combination with other ligands, and the anti-inflammatory effect of IAP against UDP persists in mouse small intestine. The benefits of IAP in intestinal disease may be partly due to inhibition of the proinflammatory activity of UDP.

  • lipopolysaccharide
  • P2Y6 pyrimidinergic receptor
  • intestinal loop model

the gut mucosal immune system is defined by complex interactions between inflammatory products of commensal and pathogenic luminal bacteria and host factors that serve to promote a defensive inflammatory response and protect the host from excessive inflammation (20, 38). The number of factors involved in this process is vast and relates to a broad range of diseases, including infectious colitis, bacterial peritonitis, and chronic conditions such as Crohn's disease and ulcerative colitis (1, 33). Recently, extracellular nucleotides have been implicated as proinflammatory mediators in intestinal disease (2, 14). Nucleotides play a broad range of roles in biological processes by binding to P2 receptors, of which there are two classes: P2X receptors are ligand-gated channels that respond to ATP, and P2Y receptors are G protein-coupled receptors that differ in their specificity for adenine and uridine nucleotides (44). In inflammatory bowel disease, commensal bacteria are linked to an excessive inflammatory response, resulting in release of large amounts of nucleotides into the extracellular environment (14, 19), which has been shown to aggravate colonic inflammation in rats with chemically induced experimental colitis (2).

The nucleotide UDP is the specific ligand for the P2Y6 pyrimidinergic receptor, which has been directly implicated in immune function and intestinal inflammation (19, 39). P2Y6 receptor mRNA is found widely, including in intestinal epithelium, lymphocytes, and macrophages (12, 22, 47), and is highly expressed in T cells infiltrating active inflammatory bowel disease (IBD) tissue but is absent in T cells of normal bowel, suggesting a role for the P2Y6 receptor in the pathogenesis of IBD-mediated damage (39). In addition, P2Y6 receptor mRNA is greatly increased in colon tissue of dextran sodium sulfate-treated mice, as well as in human biopsies of Crohn's disease and ulcerative colitis tissues (19), and when gut epithelial cells are subjected to inflammatory cytokines, they release UDP into the media and upregulate expression of the P2Y6 receptor (19). Activation of the P2Y6 receptor has been shown to stimulate production of IL-8, a powerful neutrophil and monocyte chemoattractant, in a variety of cell types, including human monocytes, THP-1 monocytic cells, and cultured Caco-2/15 cells (19). The media of UDP-treated monocytes cause chemotaxis of neutrophils (24), which is prevented by treatment with apyrase, a class of nucleotide scavenger that is ubiquitously expressed in eukaryotes and has been found in some prokaryotes (37).

The small intestinal brush border enzyme intestinal alkaline phosphatase (IAP) is known to play a role in the interaction between the host and luminal bacteria (17, 25). IAP is expressed on the small intestinal epithelium and is secreted bidirectionally into the bloodstream and the lumen, maintaining activity throughout the colon and into the stool (4). IAP has been shown to have a broad range of roles (17, 25) involving regulation of lipid absorption (30–32), regulation of bicarbonate secretion and duodenal luminal pH (3, 29), maintenance of normal gut microbiota (28), protection against translocation of luminal bacteria across the gut epithelium (18), and dephosphorylation and subsequent detoxification of bacterial toxins, including LPS (6, 18, 34), flagellin (11), and unmethylated CpG DNA (11). We hypothesized that IAP protects against the proinflammatory nucleotide UDP, possibly explaining the benefit that IAP confers in IBD.


Cell culture and reagents.

Human monocyte THP-1 cells were purchased from American Type Culture Collection (Manassas, VA) and maintained in RPMI 1640 culture medium (Hyclone, Logan, UT) supplemented with 10% heat-inactivated FBS, 2 mM l-glutamine, 1% antibiotic-antimycotic solution (Cellgro, Herndon, VA), and 0.05 mM mercaptoethanol (Sigma, St. Louis, MO). Human colorectal adenocarcinoma HT-29 cells were purchased from American Type Culture Collection and maintained in McCoy's culture medium (Invitrogen, Carlsbad, CA) supplemented with 10% FBS (Lonza, Allendale, PA), 2 mM l-glutamine, and 1% antibiotic-antimycotic solution (Cellgro). Human colorectal adenocarcinoma Caco-2 cells were purchased from American Type Culture Collection and maintained in RPMI 1640 culture medium (Hyclone) supplemented with 20% heat-inactivated FBS and 1% antibiotic-antimycotic solution (Cellgro). Murine macrophage RAW 264.7 cells were purchased from American Type Culture Collection and maintained in DMEM culture medium (Invitrogen, Grand Island, NY) supplemented with 10% heat-inactivated FBS, 2 mM l-glutamine, and 1% antibiotic-antimycotic solution (Cellgro). UDP (uridine 5′-diphosphate sodium salt, >96% purity by HPLC; Sigma), LPS (Escherichia coli serotype 055:B5; Sigma), flagellin (ultrapure flagellin Salmonella typhimurium; Invivogen, San Diego, CA), (S)-(2,3-bis(palmitoyloxy)-(2RS)-propyl)-N-palmitoyl-(R)-Cys-(S)-Ser(S)-Lys4-OH trihydrochloride (Pam3Cys; EMD Biosciences, Gibbstown, NJ), mouse TNF-α (Biolegend, San Diego, CA), potato apyrase (grade III, from Solanum tuberosum; Sigma), polymixin B sulfate (Sigma), Toll-like receptor (TLR) 5 (TLR5) blocking antibody (rat polyclonal; Invivogen), l-phenylalanine (Sigma), and l-homoarginine (Sigma) were reconstituted, divided into aliquots, and stored according to the manufacturers' directions. Bovine IAP (Sigma) was reconstituted in the manufacturer-suggested vehicle (50 mM Tris·HCl, 100 mM NaCl, 10 mM MgCl2, and 1 mM dithiothreitol at pH 7.9) and stored according to the manufacturer's directions.

Determination of free phosphate release by malachite green assay.

The malachite green-based phosphate quantitation assay was modified from Baykov et al. (7) as follows. The working solution [7 μM ammonium molybdate (Sigma), 0.2% Tween 20, and 15 mM malachite green (Sigma)] for the assay was mixed immediately before the experiment. Separately, nucleotides (10–1,000 μM) or LPS (1 mg/ml) was combined with enzymes [1–10 U/ml of IAP, potato apyrase, or Antarctic phosphatase, cloned from psychrophilic Antarctic bacterium strain TAB5 (New England Biolabs, Ipswich, MA), with buffers] and enzyme inhibitors (10 μM phenylalanine or homoarginine) and incubated in 96-well plates (Costar, Corning, NY) at 37°C. At predetermined time points, working solution was added at a ratio of 1:4 malachite green working solution to incubated groups, and optical density at 630 nm was measured after 60 min at room temperature and compared with a phosphate standard curve. IAP dephosphorylation of LPS was used as a positive control, since this reaction has been well established by our laboratory and others (11, 18, 34, 41).

In vitro inflammatory response to UDP and other proinflammatory ligands.

THP-1 cells were plated at a concentration of 1 × 106 cells/ml. HT-29 cells were seeded into 48-well plates and grown until they were 80% confluent. Caco-2 cells were seeded into 48-well plates and grown until 4 days after they reached confluence. UDP (0–1,000 μM final concentration), LPS (100 ng/ml), flagellin (100 ng/ml), TNF-α (100 ng/ml), and Pam3Cys (100 ng/ml) were incubated at 37°C for 1 h with or without IAP (0–400 U/ml) or potato apyrase (1 U/ml) or polymixin B (10 μg/ml). Incubation solutions were then applied directly to cells. In some experiments, a blocking TLR5 antibody (5 μg/ml) was applied to cells 15 min prior to application of inflammatory ligands. After 6 h of incubation, media were collected and measured for IL-8 by ELISA (BD Biosciences, San Diego, CA). UDP dosing was based on previous studies using UDP in cell culture (13, 19, 46), as well as the finding by Grbic et al. (19) that Caco-2 cells exposed to inflammatory-like stress release ∼100 nM UDP.


IAP-knockout (KO; Akp3−/−) mice (Mus musculus C57BL/6) (32) were obtained from the Sanford-Burnham Medical Research Institute and bred at the Massachusetts General Hospital (MGH) animal facility to create homozygous IAP-KO, heterozygous, and wild-type (WT) C57BL/6 littermates. Genotype was confirmed by polymerase chain reaction analysis (32). Animals were maintained in a specific pathogen-free environment at MGH in accordance with the guidelines of the Committee on Animals of Harvard Medical School. All experiments were reviewed and approved by the Institutional Animal Care and Use Committee and carried out according to regulations of the Subcommittee on Research Animal Care of the MGH and the National Institutes of Health (NIH Publication 85-23, 1985).

Intestinal loop model.

We established a novel surgical intestinal loop model to study the impact of IAP on the inflammatory activity of UDP within the small intestine. Mice were anesthetized with 3.0% inhaled isoflurane (Baxter International, Deerfield, IL) and intraperitoneal pentobarbital sodium (40 mg/kg; Akorn, Lake Forest, IL) and kept on a heating pad at 37°C. After midline laparotomy, the proximal jejunum was identified, and a 5-cm loop was isolated by ligation of either end with 5-0 silk suture (Ethicon). A 28-gauge needle was used to instill 100 μl of endotoxin-free water or UDP (1 mM), with or without IAP (1,000 U/ml), into the loop. The incision was closed in two layers using 3-0 silk suture (Ethicon). After 2 h, the incision was reopened and the loop was harvested. The luminal contents were isolated, and the intestinal tissue was snap-frozen in liquid nitrogen and stored at −80°C for later analysis.

In vitro cytokine response to luminal contents and UDP.

To study the impact of endogenous IAP on the inflammatory activity of UDP within the in vivo environment, UDP (1 mM) or endotoxin-free water was instilled into the loop model of IAP-KO mice and WT littermates (5 mice per group) as described above. The luminal contents were collected, proportionally diluted in endotoxin-free water, and centrifuged twice at 13,000 g for 15 min, and the supernatant was collected. RAW 264.7 cells were plated at a concentration of 2 × 106 cells/ml and starved for 8 h in DMEM with 0.5% FBS. Cells were then treated with 100 μl of luminal content supernatant. After overnight incubation, media were collected and assayed for TNF-α concentration by ELISA (eBioscience, San Diego, CA).

In vivo cytokine response to luminal UDP.

To study the action of IAP against UDP-induced inflammation in the small intestine, jejunal loops were created in IAP-KO mice as described above, and UDP (1 mM) or endotoxin-free water was instilled with or without IAP (100 U; 11 mice in each UDP group, 5 mice in each water group). The intestinal loops were harvested after 2 h, homogenized with 10 volumes of ice-cold RIPA buffer (150 mM NaCl, 10 mM Tris·HCl, pH 7.5, 1% sodium deoxycholate, 1% NP-40, 10 mM EDTA, and 0.1% SDS, including protease and phosphatase inhibitor cocktail; Thermo Scientific, Rockford, IL), and then incubated for 30 min on ice. The homogenates were centrifuged twice at 15,000 g for 15 min at 4°C, and the supernatants were collected. Protein concentration was determined using the Coomassie (Bradford) Protein Assay Kit (Thermo Scientific), and TNF-α was quantified by ELISA (eBioscience).

Statistical analyses.

Malachite green assay experiments were carried out in duplicate or triplicate sample wells, expressed as averages with standard error, and repeated three times. Cell culture results were averaged between duplicate or triplicate wells, expressed with standard error, and repeated a minimum of three times. Data are from representative experiments. Animal experiments were performed on a minimum of five mice per group. Values are means ± SE. Statistical significance was determined by unpaired Student's t-test using Excel software (Microsoft Office 2007).


Dephosphorylation of UDP by IAP is efficient, dose-dependent, and specific.

First, we used the malachite green assay for free phosphate measurement to examine the ability of IAP to dephosphorylate UDP. It has previously been shown by other methods that alkaline phosphatases release inorganic phosphate from a variety of organic pyrophosphate substrates (16). Dephosphorylation of UDP and LPS by IAP occurred in a dose-dependent fashion, with a 63-fold increase in free phosphate released from UDP by IAP (10 U/ml) after 60 min (Fig. 1A; P < 0.05) and a 2.9-fold increase from LPS (1 mg/ml) by IAP (10 U/ml, P < 0.05; Fig. 1B). As each mole of UDP contains 2 mol of phosphate, release of >100 μM free phosphate was not a surprise. In addition, treatment with l-phenylalanine (10 mM), a specific inhibitor of IAP (18), prevented IAP (10 U/ml) from dephosphorylating UDP (Fig. 1C; 74% decrease in free phosphate compared with IAP treatment, P < 0.05) and LPS (Fig. 1D; 55% decrease compared with IAP treatment, P < 0.005). Homoarginine, an inhibitor of nonintestinal alkaline phosphatases, did not affect dephosphorylation of UDP (Fig. 1C; P > 0.3) or LPS (Fig. 1D; P > 0.1) by IAP, confirming that the dephosphorylation was specific to IAP, rather than contaminants. Other phosphatases were used as positive controls and effectively, but less efficiently, dephosphorylated UDP. Potato apyrase, which hydrolyzes extracellular nucleoside triphosphates and diphosphates, caused a 20-fold increase in phosphate release from UDP, and Antarctic alkaline phosphatase, a bacteria-derived alkaline phosphatase, caused a 15-fold increase in phosphate release from UDP (P < 0.05 for both). These enzymes were unable to dephosphorylate LPS (P > 0.4 for both), demonstrating a specificity of IAP for LPS.

Fig. 1.

Intestinal alkaline phosphatase (IAP) specifically and efficiently dephosphorylates UDP. A and B: UDP (100 μM) or LPS (1 mg/ml) was combined with increasing doses of IAP or vehicle and incubated for 1 h at 37°C. Malachite green working solution was added at a ratio of 1:4 malachite green working solution to incubation volume, and optical density at 630 nm was read after 60 min. Incubation with IAP resulted in a dose-dependent increase in free phosphate for UDP (63-fold increase from UDP by 10 U/ml IAP) and LPS (2.9-fold increase from LPS by 10 U/ml IAP). *P < 0.05; ***P < 0.001 vs. no IAP. C and D: UDP (100 μM) or LPS (1 mg/ml) was combined with apyrase (5 U/ml), Antarctic phosphatase (5 U/ml), or IAP (5 U/ml) with or without phenylalanine (10 mM) or homoarginine (10 mM). After 1 h of incubation, free phosphate concentration was determined by malachite green reaction. Phenylalanine prevented release of free phosphate by IAP for UDP (∼75% decrease in free phosphate compared with IAP treatment, P < 0.05) and LPS (∼55% decrease compared with IAP treatment, P < 0.005), and neither was significantly affected by homoarginine. Apyrase and Antarctic phosphatase released free phosphate from UDP (20- and 15-fold increase, respectively, P < 0.05 for both) but not significantly from LPS. Results are representative of 3 experiments.

IAP prevents the proinflammatory effect of UDP on immune cells and intestinal cells.

First, we confirmed that UDP causes dose-dependent activation of the inflammatory pathway in undifferentiated THP-1 monocytes, which was shown by Warny et al. (47) but was not reproduced by others (13). Increasing concentrations of UDP resulted in increased release of IL-8, an effect that peaked at 100 μM UDP (Fig. 2A; P < 0.001). At >1,000 μM, we saw significant cell swelling and cell death, indicating toxicity of the nucleotide. Next, on an immune cell line and on two separate colon cancer cell lines, we showed that the proinflammatory effect of UDP was prevented by pretreatment with IAP, again in a dose-dependent fashion. On undifferentiated THP-1 cells (Fig. 2B), treatment with UDP resulted in a 4.7-fold increase in IL-8 release compared with control, and cotreatment with IAP (20 U/ml) resulted in a >50% decrease in IL-8 release (P < 0.001). On HT-29 cells (Fig. 2C), treatment with UDP resulted in a 2.2-fold increase in IL-8 release, and cotreatment with IAP (20 U/ml) decreased IL-8 release by >35% (P < 0.001). On Caco-2 cells (Fig. 2D), UDP treatment caused a 3.4-fold increase in IL-8 release, which was completely inhibited by cotreatment with IAP (20 U/ml; P < 0.001). Potato apyrase, a nucleotide scavenger, was used with each cell line as a positive control and also resulted in statistically significant UDP inhibition (P < 0.01 for all cases).

Fig. 2.

Proinflammatory effects of UDP are prevented by exogenous IAP on 3 separate cell lines. A: increasing doses of UDP were applied to THP-1 monocytic cells for 6 h, and media were assayed for IL-8 by ELISA. UDP increased release of IL-8 in a dose-dependent fashion, which peaked at 100 μM, above which cell toxicity was seen. B–D: treatment with UDP resulted in increased IL-8 release from THP-1 cells (∼5-fold), HT-29 cells (>2-fold), and Caco-2 cells (∼3.5-fold; P < 0.001 for all). Cotreatment with IAP (20 U/ml) decreased IL-8 release by >50% (B), >35% (C), and >70% (D; P < 0.001 for all). Apyrase (1 U/ml) similarly prevented UDP-induced IL-8 release (P < 0.01 for all cell lines). Results are representative of 3 or 4 experiments. ^^^P < 0.001 vs. control. *P < 0.05; ***P < 0.001 vs. UDP stimulation.

IL-8 release due to inflammatory ligands is enhanced by the presence of UDP, and this effect is prevented by IAP.

Others have shown that blocking the P2Y6 receptor decreases IL-8 release induced by LPS (24, 46) and Pam3Cys (8). Here we showed that addition of UDP to THP-1 cell cultures exposed to proinflammatory ligands, including LPS (Fig. 3A; 1.7-fold increase over LPS alone, P < 0.001), flagellin (Fig. 3B; 1.7-fold increase over flagellin alone, P < 0.001), Pam3Cys (Fig. 3C; 1.4-fold increase over Pam3Cys alone, P < 0.001), and TNF-α (Fig. 3D; 3.9-fold increase over TNF-α alone, P < 0.001), resulted in a significant amplification of inflammation as measured by IL-8 release. This enhanced IL-8 release was prevented most notably by IAP (70% decrease for LPS + UDP, 60% decrease for flagellin + UDP, 42% decrease for Pam3Cys + UDP, and 79% decrease for TNF-α + UDP, P < 0.001 for all ligands) but also by known inhibitors of these inflammatory ligands [91% decrease due to polymixin B for LPS + UDP (P < 0.001) and 41% decrease due to TLR5-blocking antibody for flagellin + UDP (P < 0.001)], as well as potato apyrase (72% decrease for LPS + UDP, 22% decrease for flagellin + UDP, 40% decrease for Pam3Cys + UDP, and 66% decrease for TNF-α + UDP, P < 0.001 for all ligands). As we have previously shown (11), IAP was ineffective at blocking the effects of TNF-α or Pam3Cys alone. However, when used in combination with UDP, IAP did exert substantial inhibitory effects.

Fig. 3.

UDP enhances in vitro release of IL-8 due to other proinflammatory ligands, and this effect is prevented by exogenous IAP. UDP (100 μM) or other proinflammatory ligands [100 ng/ml; LPS (A), flagellin (B), Pam3Cys (C), and TNF-α (D)] were preincubated with IAP (200 U/ml) or apyrase (1 U/ml) and then applied directly to THP-1 monocytes. Specific inhibitors of LPS (polymixin, 10 μg/ml) and the flagellin receptor [Toll-like receptor 5 (TLR5) antibody, 5 μg/ml] were also used. After 6 h of incubation, media were assayed for IL-8 by ELISA. Presence of UDP significantly enhanced IL-8 release due to other inflammatory ligands (P < 0.001), most notably LPS and TNF-α. Exogenous IAP and apyrase prevented this increase in IL-8 release (P < 0.001), even when IAP had no effect on IL-8 release caused by the primary inflammatory ligand [Pam3Cys and TNF-α (C and D)]. Results are representative of 4 experiments. ^P < 0.001 vs. UDP alone and LPS, flagellin, Pam3Cys, or TNF-α alone, respectively. *P < 0.001 vs. vehicle.

IAP is highly efficient at preventing inflammation caused by UDP alone, as well as the synergistic effect of LPS and UDP.

In attempts to determine the principal target of IAP in preventing the enhanced inflammation caused by the combination of UDP and other inflammatory ligands, we analyzed the relative effect of increasing doses of IAP on IL-8 release by THP-1 cells treated with LPS and UDP (Fig. 4). We found that, relative to the conditions of cells treated with vehicle only, IAP was the most effective at preventing inflammation caused by UDP alone, with a statistically significant decrease in IL-8 release achieved with 10 U/ml of IAP (P < 0.001) and an ∼70% decrease with 500 U/ml (P < 0.001). The pattern of decreased IL-8 release in the cells treated with LPS + UDP was similar to that in cells treated with UDP alone, with 500 U/ml of IAP resulting in an ∼60% decrease in IL-8 release (P < 0.001), whereas LPS alone was relatively resistant to IAP treatment, with 500 U/ml of IAP causing a 13% decrease in IL-8 release (P < 0.01).

Fig. 4.

Prevention of enhanced IL-8 release due to combination of LPS and IAP is largely due to the effect of exogenous IAP on UDP. UDP (100 μM), LPS (100 ng/ml), or UDP + LPS was incubated with increasing doses of IAP (0–500 U/ml) for 1 h and then applied to THP-1 monocytes. Media were assayed for IL-8 by ELISA after 6 h. Results were normalized to vehicle treatment for each group. While IL-8 release due to LPS was relatively resistant to exogenous IAP, IL-8 release due to UDP and UDP + LPS was efficiently prevented by exogenous IAP treatment, suggesting that the anti-inflammatory action of IAP was weighted toward UDP dephosphorylation. Results are representative of 3 experiments. *P < 0.05; **P < 0.01; ***P < 0.001 vs. vehicle.

Endogenous mouse intestinal IAP prevents UDP-induced inflammation.

To establish that endogenous IAP in the mouse small intestine could inhibit the proinflammatory action of UDP, we incubated UDP in a small intestinal loop model in WT and IAP-KO mice, collected the luminal contents, and applied them to mouse RAW 264.7 macrophage cell culture. We found that luminal contents from WT mice treated with UDP induced a greater than threefold increase in TNF-α compared with luminal contents treated with water alone (P < 0.001; Fig. 5). Interestingly, IAP-KO luminal contents treated with UDP caused 25% more TNF-α release than WT luminal contents treated with UDP (P = 0.01). There was no significant difference in TNF-α release between IAP-KO and WT contents treated with endotoxin-free water (P = 0.19).

Fig. 5.

Luminal UDP is proinflammatory on target cells, with greater response in IAP-KO than WT mouse luminal contents. At 2 h after instillation of 100 μl of UDP (1 mM) or water into a 5-cm jejunal loop in IAP-KO or WT mice (n = 5 mice per group), luminal contents were collected and applied to RAW 264.7 cells. After 16 h, supernatant was assayed for TNF-α by ELISA. UDP-treated luminal contents of WT and IAP-KO mice caused more TNF-α release by target cells than control luminal contents (>3-fold increase for WT mice, P < 0.001). IAP-KO luminal contents treated with UDP caused 25% more TNF-α release than WT luminal contents (P = 0.01). TNF-α release due to IAP-KO and and that due to WT contents treated with endotoxin-free water were similar (P = 0.19). **P < 0.01; ***P < 0.001.

IAP prevents mouse small intestinal inflammation caused by luminal UDP.

To assess the direct proinflammatory properties of UDP on small intestinal tissue, we incubated UDP in the small intestinal loop model using IAP-KO mice, harvested the intestinal loops, washed off the luminal contents, homogenized the tissue, and measured mouse TNF-α by ELISA. We found that treatment with intraluminal UDP caused a greater than twofold increase in intestinal TNF-α (P = 0.02) and that cotreatment with UDP and IAP caused a 37% decrease in TNF-α release compared with UDP alone (P = 0.03), resulting in TNF-α release similar to that achieved with water alone (P = 0.08) and IAP alone (P = 0.46; Fig. 6). Treatment with IAP alone caused no significant change in TNF-α (P = 0.28) compared with treatment with water.

Fig. 6.

Exogenous IAP prevents intestinal inflammation caused by luminal UDP in IAP-KO mice. At 2 h after instillation of 100 μl of UDP (1 mM) or water, with or without IAP (1,000 U/ml), into a 5-cm jejunal loop in IAP-KO mice (n = 11 mice per UDP group, n = 5 mice per water group), intestinal loops were collected and homogenized, and TNF-α content was measured by ELISA. UDP treatment caused a >2-fold increase in intestinal TNF-α release (P = 0.02), and UDP + IAP prevented TNF-α release, resulting in levels similar to those in mice treated with water (P = 0.08) and mice treated with IAP (P = 0.46) alone. TNF-α release in mice treated with IAP alone was similar to that in mice treated with water (P = 0.28). *P < 0.05.


It is well established that IBD in part results from an inappropriate activation of the mucosal immune system driven by the presence of luminal bacteria (20, 33). Several animal models that normally develop spontaneous colitis due to genetic alterations in their immune systems remain colitis-free under germ-free conditions (15, 38). Similarly, in human patients with IBD, clinical improvement is seen with diversion of the fecal stream (21) and broad-spectrum antibiotic treatment (40). The host response to bacterial ligands such as LPS and flagellin is complex and has been characterized in a variety of in vitro and in vivo model systems (36, 43, 45). While it is clear that bacteria are critical to induction of gut inflammation in experimental models of IBD, various bacterial ligands are generally studied in isolation; in reality, however, they interact with each other and with multiple host factors.

We have been interested in exploring the effects of the gut enzyme IAP in the context of intestinal inflammation. Work in our laboratory and previous studies have established a protective role of IAP in animal models of local intestinal inflammation, as well as systemic sepsis secondary to abdominal processes. In acute (9, 42) and chronic (35) rodent models of colitis, exogenous administration of IAP significantly reduced colonic inflammation, and Campbell et al. (10) induced and inhibited endogenous IAP activity in mice with dextran sodium sulfate-induced colitis and showed that disease severity negatively correlated with IAP activity. Additionally, in the first report of IAP for human IBD, Lukas et al. (27) showed that intraduodenal administration of IAP in 21 patients with refractory ulcerative colitis resulted in significant improvement in several objective disease parameters. Importantly, IAP was associated with no significant side effects (27), unlike nearly every other treatment available for IBD (33). Taken together, these studies indicate a protective role for endogenous and exogenous IAP against colitis.

We and others have shown that IAP is able to dephosphorylate and detoxify several bacterial toxins, but its precise mechanism of action in disease models remains poorly understood. The known bacterial ligands that are targets of IAP cause inflammation through binding to specific TLRs, which are located on the cell surface membrane (1). In this study we considered that the process may be more complex than simple dephosphorylation of these toxins. It has been reported that UDP is a known mediator of cytokine release from immune cells and gastrointestinal epithelial cells (19) and also plays a dramatic role in upregulating neutrophil migration in response to bacterial inflammatory ligands (8, 24); therefore, it is conceivable that a cycle exists in which inflammation is initiated by bacterial ligands, leading to release of extracellular nucleotides such as UDP as danger signals and, thus, further amplifying inflammation. Interestingly, Lavoie et al. (26) demonstrated localization of a family of membrane-bound ectonucleotidases, the nucleoside triphosphate diphosphohydrolases (NTPDases), to the epithelium of the foregut, as well as the entire enteric nervous system. Kukulski et al. (23) showed that one member of this family of ectonucleotidases, NTPDase1, controls IL-8 production by human neutrophils through regulation of P2Y receptor activation, demonstrating a clear relationship between dephosphorylation of luminal nucleotides and intestinal inflammation. The NTPDase family of enzymes has a similar function of hydrolysis of nucleotides, as we have demonstrated in the case of IAP, a secreted enzyme that maintains activity throughout the entire small intestine and colon.

Exogenous IAP has been shown to be a potential treatment in human ulcerative colitis (27) and in several animal models of IBD. Although IAP has clear anti-inflammatory properties in a variety of bacteria-induced gastrointestinal diseases, its precise mechanism of action is unclear. Here we show that IAP efficiently dephosphorylates and detoxifies the nucleotide UDP, effectively preventing the primary inflammation caused by UDP on immune and epithelial cells and also preventing the synergistic inflammation that UDP causes in the presence of other inflammatory ligands. This effect persisted in vivo in a jejunal loop model, with more effective detoxification of UDP by WT than IAP-KO luminal contents and with lower levels of TNF-α released by UDP in IAP-KO intestinal tissues than in intestinal tissues in which luminal IAP activity was restored. This demonstrates that the proinflammatory effects of UDP carry over into animal models and that IAP prevents UDP-induced inflammation in vivo, although further studies with inhibitors of IAP, UDP, and the P2Y6 receptor would strengthen this conclusion. While it is well established that IAP detoxifies LPS and other bacterial ligands, high doses of IAP were required to detoxify bacterial ligands. On the contrary, IAP efficiently prevents inflammation due to UDP. This suggests that the anti-inflammatory action of IAP in the complex environment of the intestines could be related to its ability to remove inflammatory signals such as extracellular nucleotides that are derived from the host, rather than from bacteria themselves. Interestingly, IAP prevents the synergistic inflammation that results from the combination of UDP with ligands that are not directly affected by IAP, such as TNF-α and Pam3Cys. It is known that, in response to inflammatory stimuli, host cells release nucleotides, including UDP (19), that cause release of inflammatory cytokines and promote neutrophil migration. Despite the finding that the decreased inflammation provided by IAP against UDP alone is small in absolute terms and that inflammation caused by LPS could only be prevented with high doses of IAP, the anti-inflammatory effect of IAP against LPS + UDP was more profound than the action of IAP against either ligand alone. This supports the conclusions by others that host-derived nucleotides act in synergy with bacterial ligands to regulate inflammation (5, 46) and is the first demonstration of IAP working against a host-derived inflammatory factor.


This study was supported by National Institutes of Health Grants R01 DK-050623 (R. A. Hodin), R01 DK-047186 (R. A. Hodin), T32 CA-009535 (M. M. Bertagnolli), and AI-059010 (H. S. Warren), Shiners Hospital for Crippled Children Grant 8720 (H. S. Warren), and a Grand Challenge Exploration Grant from the Gates Foundation (M. S. Malo).


No conflicts of interest, financial or otherwise, are declared by the authors.


A.K.M., S.R.H., M.M.R.M., M.S.M., and R.A.H. are responsible for conception and design of the research; A.K.M., S.R.H., M.M.R.M., S.R., H.Y., P.P., K.K., S.N.A., N.M., O.M., A.T., and N.S.M. performed the experiments; A.K.M., S.R.H., M.M.R.M., S.N., J.L.M., H.S.W., E.H., M.S.M., and R.A.H. analyzed the data; A.K.M., S.R.H., M.M.R.M., S.N., J.L.M., H.S.W., E.H., M.S.M., and R.A.H. interpreted the results of the experiments; A.K.M., S.R.H., and M.M.R.M. prepared the figures; A.K.M. drafted the manuscript; A.K.M., S.N., J.L.M., H.S.W., E.H., M.S.M., and R.A.H. edited and revised the manuscript; A.K.M., S.R.H., M.M.R.M., S.R., H.Y., P.P., K.K., S.N.A., N.M., O.M., A.T., N.S.M., S.N., J.L.M., H.S.W., E.H., M.S.M., and R.A.H. approved the final version of the manuscript.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
View Abstract