Ectonucleotidases in Inflammation, Immunity, and Cancer

Following the realization that ATP is released together with classical neurotransmitters, including acetylcholine and glutamate, in both the peripheral nervous system and CNS (1–3), ATP was demonstrated to be an extracellular signaling molecule involved in the regulation of a variety of physiological and pathophysiological processes, such as blood flow, cell death, and inflammation (4). Subsequent studies have demonstrated that in addition to ATP, other nucleotides such as ADP, UTP, UDP, and UDP-sugars can also serve as extracellular signaling molecules. Cell stress, cell death, and inflammation are common events that trigger extracellular nucleotide release (5). Nucleoside triphosphate diphosphohydrolases (NTPDases) are a family of enzymesconsisting of eight members, denoted NTPDases 1–8, that hydrolyzes nucleotides (6). Four enzymes, NTPDase1/CD39, NTPDase2/CD39L1, NTPDase3/CD39L3, and NTPDase8, are cell membrane–bound enzymes that face the extracellular matrix and therefore are able to degrade extracellular nucleotides (Fig. 1). Therefore, these enzymes can control the bioavailability of purines and signaling through P2 receptors in the extracellular space (7–16). The final products of NTPDase-mediated nucleotide hydrolysis are nucleoside monophosphates, such as AMP and UMP, which are further metabolized in the extracellular space by ectoenzymes, such as CD73, to adenosine and uridine, respectively (6). Adenosine has tissue-protective, -immunosuppressive, and –anti-inflammatory functions through P1 or adenosine (1–3) receptors signaling (6, 17–25), whereas uridine has no known receptors (26). Thus, membrane surface NTPDases together with CD73 eventually switch the extracellular milieu microenvironment from pro- to anti-inflammatory (27–34). Unlike NTPDases 1 and 3 where ADP is not released from the enzymes during nucleotide catabolism to AMP or UMP, respectively, NTPDase2 and NTPDase8 can release ADP and UDP transiently, which can have at least transient, proinflammatory, and prothrombotic effects through P2 receptors on immune cells and platelets (16, 35). Eventually, this ADP and UDP will be further degraded to AMP and UMP and then adenosine and uridine, respectively, thereby suppressing inflammation.

The first NTPDase identified was CD39, which later was renamed NTPDase1 while other enzymes of the NTPDase family began to be uncovered (7, 8). NTPDases are expressed in all tissues and regulate essential functions such as development, blood flow, and hormone secretion (6). NTPDases have five “apyrase conserved regions” (ACR1 to ACR5) involved in the catalytic cycle (36). NTPDases 1, 2, 3, and 8 have two transmembrane domains, which together with the extracellular domains control substrate specificity during nucleotide binding and hydrolysis (6). Additionally, NTPDases require Ca²⁺ and Mg²⁺ as cofactors and thus alterations of Ca²⁺ and Mg²⁺ concentrations can alter enzyme activity.

NTPDases 4, 5, 6, and 7 are localized in the intracellular space where they are anchored to intracellular organelle membranes or vesicles and have lower affinity to nucleotides when compared with extracellular ones (37) (Fig. 1). NTPDases 5 and 6 are anchored to the organelles by one transmembrane domain, whereas NTPDases 4 and 7 are anchored to the organelles by two transmembrane domains (38, 39).

In this review, we will discuss the role of NTPDases in inflammation, immunity, and cancer. Because the role of NTPDase1 has been already studied and is the subject of many excellent reviews (30, 38, 40–44), our goal is to summarize recent literature on the role of NTPDases 2–8, for which NTPDase1 will serve as a reference point. We first outline which cellular compartments, tissue, and organs express the various extra and intracellular NTPDases. We then discuss the role of NTPDases in regulating infectious and inflammatory diseases and cancer. Finally, we provide a short perspective on the possible therapeutic value of targeting NTPDases for the treatment of infectious and inflammatory diseases and cancer.

Extracellular NTPDases are found in different tissues in which they have distinct functions based on their localization. NTPDase1/CD39 (ATPDase/ectoapyrase) is expressed in human vascular tissue, liver, spleen, peripheral blood leukocytes, and lungs (7, 8). Diverse cell types including vascular endothelial cells, regulatory T cells, macrophages, Langerhans cells, and NK cells express NTPDase1 (45, 46). Several reviews have summarized in detail the role of CD39 in inflammation and immunity (30, 38, 40–43), and in this review, we will primarily focus on the other NTPDases.

NTPDases 1 and 2 are expressed together in blood vessels where NTPDase1 is found primarily on endothelial and smooth muscle cells and NTPDase2 is found on adventitial cells and portal fibroblasts (6). NTPDases 1 and 2 have different enzymatic activities and, therefore, functions. NTPDase1 hydrolyzes ATP directly to AMP, without any significant accumulation of ADP (35), whereas NTPDase2 first produces ADP and then slowly generates AMP (47). In the liver, which has one of the highest ATPase and ADPase activities of all organs, NTPDase1 is expressed on Kupffer and vascular endothelial cells and NTPDase2 is found on portal fibroblasts and activated hepatic stellate cells (16).

Besides the liver, NTPDase2 is highly expressed in the brain, muscles, testis, and at lower levels, in the lungs. In the brain, it was described to be related to neural development and differentiation (48). It is also expressed by type I cells of taste buds, in which it regulates ATP levels and, therefore, gustation (49). Both NTPDase2 and NTPDase3 are expressed on cells of the enteric nervous system (ENS) in both murine and human colon (50). NTPDase3 can be detected in the rat brain on neurons related to feeding and sleep-wake behavior (51). It was also reported that NTPDase3 expression is specific to human pancreatic β-cells (52) and can regulate insulin secretion (53). NTPDase3 together with NTPDase1 is distributed in human airways in which these enzymes have a major role in regulating nucleotide concentrations in the airway surface liquid layer (54, 55). NTPDases 3 and 8 have higher affinity to ATP/UTP over ADP/UDP when compared with NTPDases 1 and 2 (35). NTPDase8 (liver canalicular ecto-ATPase/hATPase) was the last NTPDase characterized (15) and it is expressed mainly in the liver, followed by the intestine and kidney (15, 16). In the liver, NTPDase8 is found in bile canaliculi where it is involved in bile flow regulation and is likely involved in purine reuptake and salvage by hepatocytes (16). Because the liver is a major metabolic and immune organ with important roles in immunological and inflammatory diseases, such as sepsis (56), the function of NTPDade8 in metabolism, immunology, and inflammation will be the subject of future studies. A synthesis of our knowledge of NTPDase cellular localization, tissue expression, and function can be found in Table I.

NTPDases 4, 5, 6, and 7 are localized intracellularly and regulate levels of nucleotides/nucleosides within organelles. NTPDase4 (Golgi UDPase/LALP70) is anchored to the Golgi apparatus, and hence has been named Golgi UDPase (38). A second variant of NTPDase4 was reported to be expressed in lysosomes and autophagic vacuoles (57). NTPDase5 (CD39L4/ER-UDPase) is found in the endoplasmic reticulum (ER) (58). Reglucosylation reactions involved in glycoprotein folding and quality control take place in the ER and UDP is generated in the lumen from the UDP-Glucose that is translocated from the cytosol (59). A soluble form of NTPDase5 called ER-UDPase was described in vitro, and it was found to metabolize UDP, thereby preventing accumulation of inhibitory levels of UDP that could negatively interfere with the reglucosylation process (60). NTPDase5 is also a protooncogene known as PCPH (61). Both NTPDases 5 and 6 are expressed in the rat cochlea and can also be found in the cochlear fluid. NTPDase5 is localized to the spiral ganglion neurons and supporting Deiters cells in the organ of Corti and it appears to have a role in the cochlear response to stress (62). NTPDase6 (CD39L2) that is mostly associated with the Golgi and to a small extent with the cell membrane but can also be found in cell supernatants (63). It is highly expressed intracellularly in the heart and vasculature (57, 64, 65). In the cochlea, NTPDase6 is found in vestibular hair cells and it was proposed to be involved in vestibular sensory transduction (66). NTPDase7 (LALP1) is widely expressed in intracellular vesicles in tissues (38), but its function is poorly understood.

NTPDase1 was first identified in B lymphocytes (45) and later in T lymphocytes (67). Although NTPDase1 function has been extensively studied in immune cells (31, 41, 43, 68, 69), less is known about the role of NTPDases 2–8 in immunity. Similar to NTPDase1 (70), NTPDases 2 and 3 are also expressed in naive peritoneal macrophages (71). NTPDase1 and 3 expression in macrophages varies depending on whether macrophages exhibit a proinflammatory M1 or anti-inflammatory M2 phenotype, whereas no differences in NTPDase2 are noted between these two major macrophage phenotypes (71). In the blood, extracellular NTPDase2 was also found highly expressed in eosinophils (72). An important expression of NTPDase3 is also found in dendritic cells (72). NTPDase5 is found in eosinophils and basophils (72). The expression of NTPDases 4, 6, 7, and 8 has not been reported in immune cells. A synthesis of our knowledge of the NTPDase expression in immune cells is presented in Fig. 2.

Ectonucleotidase activity contributes to the promotion of infection caused by the parasite Leishmania, which can modulate host immune responses to promote disease progression (73). The expression of endothelial cell NTPDases 2 and 3 is increased upon infection with Schistosoma mansoni, which results in increased ATP hydrolysis to ADP. Higher concentrations of ADP increase ICAM-1 expression and mononuclear cell adhesion to the endothelial cells through activating P2Y₁ receptors (74). The authors of this study proposed that increased inflammation secondary to increased monocyte adhesion is likely to contribute to morbidity in Schistosoma (74).

In a model of bacterial infection using primary cultures of human bronchial epithelial cells treated with Pseudomonas aeruginosa LPS, the mRNA level of both NTPDase1 and NTPDase3 was increased 24h after treatment (54). In another study, exposure of airway epithelial cells to supernatant of the mucopurulent material collected from cystic fibrosis patients increased NTPDase3 activity and expression (55).

We demonstrated for the first time that NTPDase1 increased survival in mice suffering from polymicrobial sepsis, which was secondary to decreased inflammation and infection, organ damage and cellular apoptosis (34). The protective role of CD39 in polymicrobial sepsis was confirmed in a subsequent study (75). Another study reported increased ATP but not ADP hydrolysis in neutrophils isolated from rats undergoing polymicrobial sepsis (76). This indicates that the function of distinct NTPDases is differentially regulated in neutrophils during sepsis. The precise roles of NTPDases 2–8 have yet to be determined in bacteremia and sepsis.

NTPDases have emerged as promising targets for the treatment of inflammatory diseases. The role of NTPDases in regulating inflammation in the intestine has received much attention. NTPDases 1, 2, 3, and 8 are all expressed in the gut, where they have all been associated with blood vessels (77). NTPDase1 single-nucleotide polymorphisms tagging low levels of NTPDase1 expression are associated with increased susceptibility to inflammatory bowel disease (IBD) in humans (78). NTPDases 2 and 3 are expressed by both the human and murine ENSs (50). It was demonstrated that the gene expression of NTPDases 1, 3, and 7 was upregulated in human colon areas affected by ulcerative colitis when compared with healthy colon areas (79). Recently, it was demonstrated that primary intestinal epithelial cells cultivated in vitro express NTPDases 3 and 8 (80).

General NTPDase inhibition using the nonselective inhibitor polyoxotungstate-1 augmented intestinal injury in mice treated with dextran sulfate sodium (DSS), indicating a protective role for NTPDases (79). In another murine study, NTPDases 2 and 3 were found predominantly expressed in the ENS (50). Genetic deletion of NTPDase2 but not NTPDase3 increased the severity of DSS-induced colitis, and macrophages in NTPDase2-deficient mice had increased expression of the activation markers CD68 and MHC class II (50). Grubišić and colleagues confirmed the anti-inflammatory role of NTPDase2 in DSS-induced colitis in mice, where NTPDase2 suppressed gut permeability and normalized gut motility (65). It was demonstrated that nonselective NTPDase inhibition exacerbated tissue damage in DSS-induced colitis in mice, which was associated with a lower frequency of IL-22–producing innate lymphoid cell type 3 (ILC3) in the gut. The same study showed that intestinal or splenic ILC3s challenged with extracellular ATP had decreased IL-22 production, whereas apyrase-mediated ATP breakdown increased IL-22 production by splenic ILC3s (79). However, it is unclear which NTPDase was targeted by apyrase in this study.

NTPDase7 deficiency in mice increased levels of ATP in the small intestinal lumen and increased the number of Th17 lymphocytes in the lamina propria (81). This increase in Th17 cells was reversed by the P2 receptor antagonist oxidized ATP as well as antibiotics. The authors speculated that microbially released ATP mediates the increased Th17 cell development in NTPDase7-deficient mice. Consistent with this, NTPDase7-deficient mice were resistant to oral infection with Citrobacter rodentium, a model in which Th17 cells are protective and developed severe experimental allergic encephalomyelitis where Th17 are responsible for pathogenesis (81). Although these are interesting observations, further studies will need to investigate how deficiency of NTPDase7, an intracellular NTPDase, leads to increased ATP levels extracellularly. Recent evidence has demonstrated that NTPDase8 hydrolyzes nucleotides at the apical surface of mouse colonic epithelium and that it is an important player in the protection of the intestine from inflammation (80). NTPDase8-deficient mice presented with a worse outcome in DSS-induced colitis when compared with WT mice. This worse outcome was due to increased P2Y₆ activation causing increased inflammation at the intestinal apical surface during colitis (80).

Although NTPDase3 activity is higher in primary bronchial epithelial cell cultures from patients suffering from chronic lung disease (54), the clinical significance of this finding is unclear at this point. NTPDase2 is expressed on portal fibroblasts in the liver (82). After chronic carbon tetrachloride intoxication, a model of liver fibrosis, NTPDase2-deficient mice exhibit more severe liver fibrosis. In contrast, NTPDase2-deficient mice do not have altered responses to the fibrotic agent 3,5-diethoxycarbonyl-1,4-dihydrocollidine or partial hepatectomy when compared with control, wild-type mice (83).

Multiple sclerosis (MS) is a demyelinating disease, which is caused by an immune response to myelin. Patients with MS have reduced CD39⁺/NTPDase1⁺ T regulatory cell counts in the blood (84). Studies in mice employing the MS model experimental autoimmune encephalomyelitis have indicated that NTPDase1 limits the course of MS (30). A recent study demonstrated that NTPDase2 levels decrease in the spinal cord during experimental autoimmune encephalomyelitis in mice and that this decrease is associated with the severity of disease (85). The precise roles of the various NTPDase isotypes and their potential as targets for the treatment of chronic inflammatory diseases need to be further investigated.

Increasing evidence supports the notion that the presence of a chronic inflammatory environment, characterized by immune cell infiltrates and their secretion of inflammatory mediators is fundamental for tumor progression (86). Purinergic signaling orchestrates immune/inflammatory response and its dysfunction has also been described in a variety of tumors, including breast cancer, melanoma, and glioblastoma (29, 87). Indeed, increased ATP and adenosine concentrations in the tumor microenvironment are emerging as key mediators of cancer-related inflammation, promoting tumor progression as well as immune escape (88). ATP accumulates in the tumor microenvironment as a combination of processes, including release by cell death, as a result of cancer growth/invasion, or following radiochemotherapy treatment (89). These increased levels of ATP can have protumor activities (90).

In glioblastoma, the most common and aggressive type of tumor found in the brain, both NTPDase expression and activity are downregulated when compared with astrocytes, which express abundant levels of NTPDase2 (91). Treatment of glioma cells with apyrase, a nucleotidase isolated from the potato plant that is similar to NTPDases, reduces in vitro glioma proliferation through preventing P2Y₆ and P2X7 stimulation and IL-8 and MCP-1 production (92). Similarly, treatment of glioma cells with apyrase impaired in vivo tumor growth in a preclinical glioblastoma model (93). In contrast, by overexpressing NTPDase2, we demonstrated that ADP derived from NTPDase2 activity stimulates platelet migration to the tumor area and, by regulating angiogenesis and inflammation, promotes glioma progression (94). NTPDase2 overexpression in gliomas also induced peripheral inflammation, as evidenced by higher plasma concentrations of IL-1β/TNF-α/IL-6, increased platelet reactivity, and pathological alterations on lung tissue suggestive of tumor spread (95). Because NTPDase2 produces ADP and apyrase AMP, the differential effects of NTPDase2 and apyrase in glioma are likely related to these different nucleotides.

Another member of the NTPDase family associated with tumor progression is NTPDase5/PCPH. PCPH was originally identified as a protooncogene that is converted to an oncogene by a single base pair deletion, resulting in a mutated protein (PCPH-mt) (61). PCPH is structurally and functionally identical to NTPDase5 (61) and its expression has been reported in a variety of tumors, including breast (96), prostate (97), bladder cancer (98) and glioblastoma (93). Loss of function of ENTPD5 causes higher turnover of hepatocytes and promotes hepatocellular oncogenesis (99). In accordance with its intrinsic ATP diphosphohydrolase activity, PCPH-mt has been shown to deplete intracellular ATP and consequently inhibit c-JUN NH2-terminal kinase-mediated stress signaling pathways. The restoration of ATP levels recovered both stress-response signaling and sensitivity to chemotherapy-induced apoptosis (100). A comparative study of purinergic signaling in the RT4 and T24 cancer cell lines, demonstrated differential NTPDase regulation and nucleotide metabolism (98). In RT4 cells, the hydrolysis of tri- and diphosphate nucleosides was higher than that of monophosphonucleosides. T24 cells, however, presented with low level of hydrolysis of tri- and diphosphate nucleosides and a high level of hydrolysis of monophosphates. This is consistent with observations that T24 cells only expressed the intracellular NTPDase5 mRNA, whereas the RT4 cells expressed both the extracellular NTPDase3 and, to a lesser extent, NTPDase5 mRNA. In a preclinical model of bladder cancer, a progressive decrease of NTPDase3 expression was found (101). Taken together, these data suggest that the impairment of extracellular nucleotide metabolism as a result of NTPDase downregulation and/or the expression of NTPDase5/mt-PCPH can modulate the initiation, establishment, and/or progression of tumors by modulating both cancer cell proliferation and cancer-stroma cross-talk.

Our understanding of the role of NTPDases 2–8 in inflammation, immunity, and cancer has been expanding in recent years. Gene-deficient mouse models are increasingly becoming available for these enzymes, which should facilitate the identification of functions for NTPDases 2–8. In this respect, the role of extracellular nucleotidases 2, 3, and 8 in intestinal chronic inflammatory disease is especially captivating and has attracted much interest (50, 65, 78–80). Because NTPDases 2 (50, 65) and 8 (80) knockout mice presented worse outcomes in experimental models of IBD, enteric administration of these enzymes may present a new therapeutical approach for patients with IBD. Because NTPDase3 activity is higher in bronchial epithelial cells from patients with chronic lung disease (54) and NTPDase2 levels were decreased in the spinal cord of mice with experimental autoimmune encephalomyelitis (85), it will be important to study the role NTPDase3 and NTPDase2 in chronic inflammatory lung disease and MS models, respectively.

Previous data also demonstrated a role of NTPDase2 in glioblastoma progression in vitro and in vivo (92, 94, 95). Mice deficient for NTPDase2 thus will be useful for uncovering more specific cellular mechanisms of NTPDase2 in the context of glioblastoma. These studies could establish NTPDase2 as a target for the treatment of glioblastoma as well as other tumors.

There is also progress in developing specific enzyme inhibitors (102), which will be helpful in characterizing the immunomodulatory functions of the various NTPDases. It will also be important to study in detail which immune cell types express NTPDases 2–8 and how their expression is regulated. Another important gap in the current knowledge is how NTPDases 2–8 function in conjunction with purinergic receptors. Given the enormous promise of targeting NTPDase1 and CD73 in the therapy of immune diseases (103, 104) and inflammation, we anticipate that NTPDases 2–8 will be the focus of future pharmacological approaches to treat inflammatory, immune diseases and cancer.

G.H. owns stock in Purine Pharmaceuticals, Inc. and has a patent to develop NTPDases for the treatment of sepsis. The other authors have no financial conflicts of interest.