Extracellular lysophosphatidate (LPA) signalling is regulated by the balance of LPA formation by autotaxin (ATX) versus LPA degradation by lipid phosphate phosphatases (LPP) and by the relative expressions of six G-protein-coupled LPA receptors. These receptors increase cell proliferation, migration, survival and angiogenesis. Acute inflammation produced by tissue damage stimulates ATX production and LPA signalling as a component of wound healing. If inflammation does not resolve, LPA signalling becomes maladaptive in conditions including arthritis, neurologic pain, obesity and cancers. Furthermore, LPA signalling through LPA1 receptors promotes fibrosis in skin, liver, kidneys and lungs. LPA also promotes the spread of tumours to other organs (metastasis) and the pro-survival properties of LPA explain why LPA counteracts the effects of chemotherapeutic agents and radiotherapy. ATX is secreted in response to radiation-induced DNA damage during cancer treatments and this together with increased LPA1 receptor expression leads to radiation-induced fibrosis. The anti-inflammatory agent, dexamethasone, decreases levels of inflammatory cytokines/chemokines. This is linked to a coordinated decrease in the production of ATX and LPA1/2 receptors and increased LPA degradation through LPP1. These effects explain why dexamethasone attenuates radiation-induced fibrosis. Increased LPA signalling is also associated with cardiovascular disease including atherosclerosis and deranged LPA signalling is associated with pregnancy complications including preeclampsia and intrahepatic cholestasis of pregnancy. LPA contributes to chronic inflammation because it stimulates the secretion of inflammatory cytokines/chemokines, which increase further ATX production and LPA signalling. Attenuating maladaptive LPA signalling provides a novel means of treating inflammatory diseases that underlie so many important medical conditions.
Lysophosphatidate (LPA) was discovered to be released by platelets and in the late 1980s and early 1990s a variety of signalling properties were described . The signalling role of LPA became much more accepted when the first LPA receptor (Edg1, which is now called LPA1) was discovered by the group of Jerold Chun in 1996 . We now know that there are six G-protein-coupled LPA receptors, which belong to two subclasses: three EDG type receptors (LPA1–3) and three receptors (LPA4–6) from the purinergic family of receptors . These receptors are differentially expressed in cells to provide unique responses to signalling by LPA through coupling to heterotrimeric G proteins including Gi, Gs, G12/13 and Gq (Figure 1). Activated LPA receptors are also able to transactivate receptor tyrosine kinases such as the platelet-derived growth factor receptor-β and this involves the activation of phospholipase D2 . Generally, LPA is a survival factor because signalling through these receptors increases cell proliferation, migration, survival and angiogenesis (the development of new blood vessels).
LPA signals through six G-protein-coupled receptors by coupling to the heterotrimeric G proteins, Gi, Gs, G12/13 and Gq as shown to activate multiple signalling cascades
Most extracellular LPA is produced from lysophosphatidylcholine (LPC) by a secreted enzyme called autotaxin (ATX) (Figure 2) . ATX shares ∼45% DNA sequence homology with ENPP1 , which is one of the five members of an ectonucleotide pyrophosphatase/phosphodiesterase family. ATX (ENPP2) was first discovered as a secreted product of melanoma cells that stimulated cell migration [6,7]. The five members of the ENPP family hydrolyse phosphodiester bonds, but it was not until about ten years later that the true substrate of ATX was identified as LPC [8,9] where ATX acts as a lysophospholipase D . There are five splice variants of ATX (ATXα, -β, -γ, -δ and -ε) [11,12], which are all catalytically active. The difference in function of these isoforms is not well understood. One possible role of the different isoforms could be to modify the interactions of ATX with the cell surface through binding to proteoglycans and integrins [13–16]. This binding to the cell surface appears to selectively channel LPA signaling to LPA receptors on cells and this could explain some of the differences in functions of the ATX isoforms. For example, ATXα binds specifically to negatively charged heparin with a high affinity and this increases ATX activity by ∼2-fold . This binding and increased activity is not seen with ATXβ.
Turnover of extracellular LPA
Plasma LPC concentrations of human beings are >200 μM and they are the most prevalent plasma phospholipids. The Km of ATX for LPC is ∼100 μM and the affinity for LPC is ∼10-fold higher than for nucleotide substrates . Essentially, ATX acts as a gatekeeper for extracellular LPA production (Figure 2). Extracellular LPC is produced by high density lipoproteins through the action of lecithin:cholesterol acyltransferase and this LPC becomes associated with plasma albumin . This LPC is mainly saturated since the unsaturated fatty acid of postion-2 in phosphatidylcholine is transferred to cholesterol. Another source of LPC is the liver, and perhaps other organs, which secretes LPC species that are mainly polyunsaturated containing fatty acids such as arachidonate (C20:4) and docosahexaenoate (C22:6) . The secretion of LPC from hepatocytes depends on the presence of albumin in the incubation medium to which it binds rather than being secreted along with phosphatidylcholine and triacylglycerol, which are present mainly in very low density lipoproteins . ATX produces mainly unsaturated LPA species because of its preference for unsaturated LPC. Some saturated extracellular LPA is produced from phosphatidate through the action of secretory phospholipase A2 during inflammation and by Ca2+-independent phospholipase A2 activities .
Extracellular LPA concentrations are normally ∼100 nM, although they can rise to 1 μM or greater in inflammatory conditions including cancers . LPA signalling is terminated by lipid phosphate phosphatases (LPP1-3), which are expressed on the exterior surfaces of cells . This turnover of LPA is rapid with a half-life for plasma LPA of ∼1 min . The balance in activities of ATX versus the LPPs (Figure 2) is a major regulator of LPA signalling, which has a host of important physiological/pathological implications.
LPA signalling plays an important function in adult animals in wound healing [24,25]. ATX secretion is increased by inflammatory cytokines (TNF-α, IL-1β, IL-6 etc) that are produced in injured areas of the body (Figure 3) . These cytokines overcome the normal feedback regulation that LPA exerts on the transcription of ATX. This enables ATX concentrations to rise at the same time as LPA concentrations, thus facilitating increased LPA signalling. LPA increases the migration of fibroblasts, keratinocytes, endothelial cells and smooth muscle cells (SMCs) into the injured area. Proliferation of these cells facilitates wound repair and the angiogenesis necessary to supply blood to the repaired tissue. Activation of fibroblasts by LPA helps to stabilize the affected area through the formation of scar tissue . LPA signalling also increases the exit of leucocytes from the circulation and the conversion of monocytes into macrophages, which is integral to the innate immunity and protection from infection . When the tissue is repaired, inflammation resolves, and secretion of ATX and LPA production fall back to normal levels [27,29].
The autotaxin (ATX)–lysophosphatidate (LPA)–inflammatory cycle
Many common diseases are characterized by chronic inflammation, which leads to sustained ATX production and LPA signalling (Figure 4). These conditions include rheumatoid arthritis, atherosclerosis, aortic valve calcification, obesity (characterized by inflamed adipose tissue), asthma, idiopathic pulmonary fibrosis, scleroderma, hepatitis and inflammatory bowel diseases including Crohn’s disease and ulcerative colitis. It is also significant that hepatitis leads to liver fibrosis (cirrhosis) and hepatocellular cancer and inflammatory bowel diseases often progress to colorectal cancer. Inflammation is a hallmark of cancers, which are often described as “wounds that never heal” [30,31]. ATX production and LPA signalling through LPA1 receptors are responsible for the development of fibrosis in organs such as liver, kidney and lungs [32–34].
Beneficial and detrimental impacts of ATX/LPA signalling
Increased signalling by LPA becomes maladaptive in chronic inflammation . In fact, signalling through LPA promotes a feedforward inflammatory cycle by increasing the production of cyclooxygenase-2 (COX-2), which produces inflammatory eicosanoids (Figure 3) [5,21,27]. In addition, LPA signalling activates the transcription factor, nuclear factor kappa-light-chain-enhancer of activated B cells (NFκB), which increases the synthesis of inflammatory cytokines such as TNF-α, IL-1β and IL-6. These inflammatory mediators in turn increase further secretion of ATX and more LPA signalling.
Clues to the importance of LPA as a signalling molecule came from the study of mice in which ATX was knocked out. These mice die at embryonic day 9.5 after showing extensive defects in blood vessel formation and neural tube development by day 8.5 [36,37]. Consequently, ATX and LPA play an important role in tissue remodelling.
Later work established that the ATX/LPA signalling system is an important driver of cardiovascular diseases including atherosclerosis, thrombosis, hypertension and aortic valve stenosis (abnormal narrowing) [38–41]. A key finding is that the risk of developing coronary artery disease is associated with inherited single-nucleotide polymorphisms in the gene for LPP3 . LPP3 is one of three enzymes that dephosphorylate LPA and its sphingolipid analogue, sphingosine 1-phosphate . A deficiency of LPP3 in the vasculature increases LPA concentrations, which decreases endothelial barrier function and increases adhesion of monocytes to the endothelium, which exacerbates atherosclerosis [38,40–42]. LPA also stimulates division (mitogen) and migration of SMCs [40,41,43,44].
LPA is found in atherosclerotic lesions [40,45] and in mildly oxidized low-density lipoproteins that are atherogenic [46,47]. LPA also initiates thrombus formation by stimulating platelet aggregation [40,46], which can precipitate myocardial infarction and ischemic stroke. Activated platelets release LPA [46,48] and this initiates the coagulation cascade through production of tissue factor by SMCs . Oxidation of low density lipoproteins promotes the production of LPC and conversion to LPA by ATX, which activates the endothelium to release of CXCL1. This induces monocyte adhesion and the progression of atherosclerosis [47,50]. LPA also affects key events in atherosclerotic plaques by promoting conversion of monocytes to macrophages  and stimulating development of foam cells by decreasing expression of scavenger receptor B1, which decreases cholesterol efflux . Unsaturated LPA species cause neointimal formation through the migration and division of vascular SMCs [40,53,54], which is central to the development of atherosclerosis.
Increased vascular tone has implications for blood pressure regulation and can result from increased constriction, endothelial dysfunction leading to impaired relaxation or a combination of these effects. Both LPC and LPA stimulate these responses because LPC is converted to LPA by ATX. LPA induces constriction in the cerebral circulation of unconscious piglets  and increases blood pressure in conscious rats , in part by increasing Ca2+-mobilization in SMCs [41,57,58]. LPC constricts pig coronary arteries and impairs their dilation [59,60]. Inhibition of dilation by LPC was also shown in small resistance-sized mesenteric arteries , which are important for blood pressure regulation . LPA also inhibits endothelial-dependent relaxation in cerebral arteries by decreasing cAMP concentrations . These constriction effects could occur in response to other vasoactive factors produced by LPA signalling, for example, LPA induces production of endothelin-1, a potent vasoconstrictor . In contrast, LPA also acutely produces the potent vasodilator, nitric oxide in endothelial cells , which induces short-acting dilation of isolated pre-constricted mouse aortas through LPA1  and transiently reduces blood pressure in vivo in cats and rabbits . Removal of aortic endothelium leads to constriction that occurs through LPA1-mediated activation of COX-1 in SMCs followed by release and signalling by thromboxane A2, a potent vasoconstrictor .
ATX and LPA are also implicated in development of calcific aortic valve stenosis, a multifactorial disease that primarily afflicts the elderly [38,68]. Mineralization of the valve is promoted by increased activities of phospholipase A2 and ATX that increase levels of LPA and pro-inflammatory conditions. LPA1 receptor activation stimulates an osteogenic response through bone morphogenetic protein 2 , an important mechanism leading to stenosis. A lipoprotein called Lp(a), which carries ATX and has a key role in inflammation, appears to be particularly significant for the presence of aortic valve calcification and stenosis .
LPA also has an adverse metabolic effect by decreasing the production of adiponectin and this effect is associated with insulin resistance, defective glucose homeostasis and obesity . Adiponectin is an anti-inflammatory protein hormone  and this contributes to its effects in counteracting the co-morbidities of the Metabolic Syndrome (obesity, diabetes, dyslipidemias, hypertension etc), which are important components leading to accelerated atherosclerosis . Consequently, ATX/LPA are key mediators of a vicious cycle that leads progressively to the development of atherosclerosis, thrombosis, hypertension and calcific aortic valve stenosis. Circulating levels of ATX and LPA increase also in response to a high fat diet in mice [39,73], which could link lifestyle choices to cardiovascular diseases.
LPA affects both implantation of the placenta and maintenance of pregnancy, in part by promoting prostaglandin synthesis . In rodents, LPA3 receptor signalling is critical for successful implantation of the placenta and spacing of the multiple foetal–placental units by stimulating production of vascularization and factors that prepare the endometrium for pregnancy (decidualization) [75,76]. While studies in rodent models where enzymes and receptors can be knocked out provide a unique opportunity to determine the importance and function with respect to viability and outcomes of pregnancies, direct application to human pregnancies is limited. Studies of human pregnancies are restricted for ethical and other reasons; however, using cultured human endometrial cells it was found that LPA1, ATX and LPP3 are up-regulated during the process of decidualization . The proposed importance of ATX/LPA in human implantation is highlighted by an interesting finding that women with repeated implantation failure after in vitro fertilization have reduced endometrial LPA3 expression . LPA also converts specialized placental cells called trophoblasts to an endovascular phenotype  that infiltrate the uterine spiral arteries, remodelling them to be open conduits for blood supply to the placenta and foetus. Spiral artery remodelling is a critical vascular adaptation that is dysfunctional in pregnancies complicated by preeclampsia (hypertension with other organ system damage) or intrauterine growth restriction of the foetus . Serum ATX levels and LPA receptor expression in the placenta rise throughout normal human pregnancy and peak at term, suggesting an important role in maintaining pregnancy [81,82]. Placental trophoblasts are the likely source of ATX with mRNA levels rising throughout human pregnancy . In preeclampsia, serum ATX levels are higher  and LPA receptors, especially LPA3, are more highly expressed . Preeclampsia is a state of chronic immune activation with elevated inflammatory cytokines  that could be driven by elevated ATX/LPA activity and vice versa. In contrast, Masuda et al. reported serum ATX levels to be decreased in women with pregnancy-induced hypertension , although the sample size was small and the hypertensive women were not categorized as early or late onset. As well, placental ATX mRNA expression is decreased in women with early onset preeclampsia compared with that found in placentas from women with normal pregnancies . These disparate results emphasize the importance of the ATX/LPA signalling axis in normal and abnormal pregnancies, but key questions surrounding mechanisms remain unanswered.
Serum ATX activity rises at the onset of intrahepatic cholestasis (ICP), involving a decrease in bile flow that occurs in the late stages pregnancy. This leads to intense itching without a rash primarily on the hands and feet, but can also affect other parts of the body (pruritus). High ATX levels can be used diagnostically to distinguish ICP from other pregnancy disorders involving the liver including preeclampsia . Spontaneous preterm birth and stillbirths are increased in women with ICP, characterized by high bile salts and transaminases typically at the beginning of the third trimester. Bile acids induce production of inflammatory cytokines through activation of NF-κB and this can trigger the ATX/LPA/inflammatory cycle. This cycle is integral to adipose tissue expansion and metabolic dysfunction [73,88] and may explain the glucose intolerance and dyslipidemia found in patients with ICP . Serum IL-6 levels were also higher in women with ICP compared with controls and elevation of IL-6 and other inflammatory cytokines could drive the observed increase in ATX . However, a crystal structure of ATX shows that bile salts bind allosterically to ATX simultaneously with LPA and decrease ATX activity non-competitively. The authors reconciled these apparently contradictory observations by proposing that bile salts, which can reach >100 μM in cholestasis, bind to ATX, stabilizing and prolonging the existence of the complex in which LPA remains bound to ATX. The authors suggest that this would delay the release of LPA, which could increase the effects of LPA on cells such as sensory neurons that perceive ICP-induced itch .
Neuropathic pain occurs in ∼18% of adults diagnosed with chronic pain and it may occur as a symptom of underlying diseases such as diabetes, cancer, herpes zoster infection, multiple sclerosis, stroke, fibromyalgia and demyelinating diseases [91,92]. Demyelination in the dorsal root spinal nerve and sciatic nerve involves activation of LPA3 receptors. Additional involvement of LPA1 receptors is established since LPA1 receptor deficient mice show no neuropathic pain or demyelination in response to intrathecal injection or nerve injury [91,92]. Significantly, it is LPA species that are high in C20:4 (which are found in brain) and also in C18:1 that show the best correlations with neuropathic pain in human studies . These observations emphasize the role of controlling LPA signalling as a potential treatment for neuropathic pain.
Emerging evidence links LPA signaling to autoimmune diseases such as rheumatoid arthritis . LPA1 receptor activation contributes to the development of arthritis though infiltration of inflammatory macrophages, differentiation of T helper cells to inflammatory Th17 cells, and osteoclast formation . LPA production in synovial fluids plays a critical role in inducing COX-2 expression and production of prostaglandin E2 . Other work showed that inhibition of LPA1 receptor signalling decreases the severity of arthritis by reducing inflammatory mediators and proteins involved in bone remodelling in a mouse model . The severity of rheumatoid arthritis correlates directly with bone erosions that are associated with hyperactivity of osteoclasts. ATX has been identified a novel factor that specifically controls inflammation‐induced bone erosions and systemic bone loss . These combined observations show that blocking the activity of ATX and activation of LPA1 receptors could provide a very promising target for treating rheumatoid arthritis.
Scleroderma and idiopathic pulmonary fibrosis
Scleroderma or systemic sclerosis is a condition in which there is multiorgan fibrosis. Mice with the LPA1 receptor knocked out were relatively resistant to the effects of bleomycin, which was used to induce dermal thickness and increased collagen content. By contrast, mice with a knockout of the LPA2 receptor were just as susceptible as the control mice . In addition, pharmacological inhibition of LPA1 receptors with AM095 or LPA1/3 receptors with Ki16425 also attenuated the bleomycin-induced fibrosis [99,100]. This work was extended using an ATX inhibitor, PAT-048, which attenuated the feedforward amplification loop between the production of IL-6 and ATX (Figure 3) and the consequent increase in LPA signalling . This ATX inhibition decreased the development of dermal fibrosis in systemic sclerosis. These combined results demonstrate that targeting LPA signalling and activation of LPA1 receptors could potentially provide a new therapeutic strategy for the treatment of dermal fibrosis and scleroderma.
Idiopathic pulmonary fibrosis is an inflammatory condition of the lungs where there is gradual destruction of the lung structure and replacement by fibrotic tissue, which decreases the ability of patients to breathe. Bleomycin, has also been used widely as a preclinical model for idiopathic pulmonary fibrosis and these studies demonstrate an important role for ATX and LPA signaling through LPA1 in this condition [102,103]. Bleomycin is also a DNA-damaging radiomimetic agent  and thus it should partially mimic the effects of ionizing radiation and LPA signaling in the lungs, as will be discussed below.
The conclusion that idiopathic pulmonary fibrosis is driven by ATX and LPA signalling in animal studies has been verified in human trials. Significantly, an ATX inhibitor, GLPG1690  and an antagonist of LPA1 receptor signalling, BMS986020  were successful in Phase 2 clinical trials in attenuating the progression of the fibrosis. These trials represent a landmark advance in the introduction of pharmaceutical agents that target LPA signalling into clinical practice.
LPA concentrations are increased in bronchoalveolar lavage fluid from patients with asthmatic lungs, and are strongly induced after airway allergen challenge in subjects with allergic asthma . The polyunsaturated molecular species of LPA (C22:5 and C22:6) are synthesized selectively in airways of asthma subjects following allergen challenge and also in mouse models of allergic airway inflammation. These species of LPA can be used as potential biomarkers for asthma .
LPA enhances the chemotaxis of immune cells including dendritic cells, T cells, eosinophils and monocytes that are involved in the pathogenesis of asthma . In addition, ATX/LPA increases IL-13 expression, mast cell maturation and degranulation, and mitogenesis and contractility of airway SMCs. Overexpression of ATX in transgenic mice aggravates lung inflammation in allergic asthma, whereas this lung inflammation is attenuated in LPA2 knockout and ATX deficient mice or those treated with a selective ATX inhibitor .
Cancers and effectiveness of chemotherapy and radiotherapy
The role of ATX and LPA signalling in wound healing has been hijacked by tumours [5,21]. Tumours produce a chronic state of inflammation, which is maladaptive and contributes to immune evasion wherein cancer cells avoid destruction by the immune system even though tumours contain more macrophages [5,109]. One mechanism for this is provided by the discovery that LPA through LPA5 receptors suppresses the function of CD8-positive cytotoxic T cells by inhibiting intracellular Ca2+-mobilization and ERK activation .
Signalling through LPA receptors also increases the growth and migration of cancer cells leading to increased spread of the tumors to other sites in the body (metastasis). The ATX gene is amongst the 40–50 most overexpressed genes in aggressive metastatic cancers [111–113]. Transgenic mice that overexpress ATX, LPA1, LPA2 or LPA3 receptors develop spontaneous oestrogen receptor-positive breast tumours . High levels of ATX in stromal cells and LPA3 receptors in epithelial cells are associated with aggressiveness of breast cancer in women . In addition to increased ATX levels caused by the inflammatory milieu of the tumour, the ATX gene, ENPP2, is amplified in several cancers .
ATX is secreted by many cancer cells, especially by melanoma, glioblastoma and thyroid cancer cells . This provides an autocrine signalling model in which cancer cells secrete ATX, which provides the LPA to stimulate tumour growth. By contrast, breast cancer cells secrete very little ATX compared with the surrounding breast tissue, which contains a substantial amount of adipose tissue [46,47]. In fact, adipocytes provide ∼40% of the ATX produced by mice and a similar secretion of ATX occurs in human adipose tissue [73,118]. ATX production by adipose tissue could contribute to the association of obesity with ∼30% of breast cancers [119,120].
LPA signalling is also exacerbated by the low activities of LPP1 and LPP3, in breast, lung and ovarian tumours . This not only decreases the ability of cancer cells to dephosphorylate LPA, but LPP1 also attenuates signalling downstream of G-protein-coupled receptors, such as protease activated receptors. This makes the cancer cells hypersensitive to LPA and other agonists that activate G-protein-coupled receptors . The low expression of LPP1 in cancer cells also increases the expression of metalloproteinases and cyclin D1/D3, which are transcribed downstream of the AP1 (c-Fos-c-Jun) complex . Cyclins D1/D3 control cell cycle progression and the increased expression of metalloproteinases is associated with decreased collagen content of the tumours, which could allow cancer cells to leave the tumour, enter the circulation and eventually metastasize to different organs. The resulting increase in metastasis could explain why a low expression of LPP1 in the tumours of breast cancer patients is associated with increased mortality .
Increased ATX/LPA signalling decreases the effectiveness of doxorubicin , tamoxifen  and taxanes , which are commonly used to treat different types of breast cancers. ATX activity also delays carboplatin-induced apoptosis in ovarian cancer cells . In the case of doxorubicin and tamoxifen therapy, these effects have been linked to LPA signalling through LPA1 receptors, which activates phosphatidylinositol 3-kinase and stabilizes Nrf2 [123,124]. This transcription factor activates the antioxidant response element and as a consequence this has wide implications for the treatment for cancers by chemotherapy and radiotherapy. The antioxidant response element is responsible for the transcription of a wide variety of genes that protect cells from oxidative damage caused by xenobiotics, including chemotherapeutic agents . In addition, the antioxidant response element stimulates the synthesis of multidrug-resistant transporters, which expel chemotherapeutic agents and toxic oxidation products from cells . These combined effects contribute to LPA-induced chemoresistance . Thus, inhibiting ATX activity increases the efficacy of doxorubicin and tamoxifen and results in slower tumour growth and less metastasis [123,124]. The other consequence of inhibiting ATX activity in mouse models of breast and thyroid cancers is the anti-inflammatory effect of decreasing the concentrations of >16 inflammatory mediators and the number of CD45+ immune cells in the tumour microenvironment [88,127]. This decrease in inflammation and consequent LPA signalling contributes to the observed decrease in tumour growth and metastasis .
The second mainstay in cancer treatment is the use of RT to eliminate residual cancer cells by causing DNA damage after the surgical removal of tumours. RT results in cell debris and the release of proteins that cause inflammation and enhance the elimination of cells [128–130]. Although inflammation is an integral component of RT, it also activates the wound healing response and increases signalling through ATX/LPA [131,132]. Increased activation of LPA2 receptors provides protection against RT-induced cell death by decreasing the expression of the pro-apoptotic protein, Siva-2 . However, administration of RT to solid tumors, including breast tumours, does not induce apoptosis, but rather causes cancer cells to stop dividing and undergo cytostasis (senescence or polyploid giant-cell formation) .
Approximately 60% of breast cancer patients receive RT that involves 15–16 fractions of ∼2.6 Gy of X-radiation to the postoperative breast after lumpectomy . This means that breast adipocytes are repeatedly damaged by RT. Irradiating human breast adipose tissue at 0.5–5.0 Gy causes DNA damage, activation of NFκB and increased expression of ATX, LPA1/2 receptors and multiple inflammatory cytokines (Figure 3) . When a single fat pad in mice is exposed to one fraction of RT, there is increased ATX in the plasma and in irradiated adipose tissue, whereas increases in inflammatory cytokines are not observed until three fractions of RT are administered . This means that the ATX response appears to be an early event, which is followed by a cytokine surge.
We hypothesized that blocking the ATX response to RT would improve the therapeutic index of the RT by decreasing the division of breast cancer cells and also by diminishing RT-induced fibrosis, which is a common adverse side-effect of RT. In fact, RT is restricted in head and neck, thoracic and pelvic cancers because of the risk of fibrosis. RT-induced fibrosis in breast cancer is less severe, but it occurs in 15–28% of patients [136–138]. Co-administration of the ATX inhibitor, GLPG1690 with five fractions of 7.5 Gy decreased breast cancer cell proliferation in mouse tumours . Previous work showed that the ATX inhibitors, BrP-LPA and PF-8380, increased the sensitivity of heterotopic glioblastomas to RT in mice [140,141]. Unlike GLPG1690, these ATX inhibitors are unlikely to enter clinical practice because of problems of drug bioavailability. However, these combined results indicate the blocking ATX-LPA signaling and the wound healing response increases the efficacy of RT. Part of this effect could be mediated by blocking the LPA-induced stabilization of Nrf2 that protects against oxidative damage and increases DNA repair (Figure 5) .
LPA signalling stabilizes the transcription factor, Nrf2 and activates the anti-oxidant response element (ARE)
Other recent work shows that the glucocorticoid, dexamethasone, produces part of its anti-inflammatory effects through a comprehensive blockade of LPA signalling through decreased ATX secretion in adipose tissue and decreased expression of LPA1 receptors . Dexamethasone also increases the expression of LPP1, which attenuates LPA signalling. Thus, treatment with dexamethasone at the time of administering five fractions of 7.5 Gy to a breast fat pad in mice decreases fibrosis in the breast fat pad and underlying lungs, which occurs after seven weeks . These results indicate that blocking the activation of the ATX-LPA-inflammatory cycle (Figure 3) that is induced by RT  could increase the efficacy of RT in eliminating residual cancer cells and attenuate morbidity from fibrosis. However, glucocorticoids, such as dexamethasone, are not necessarily the best therapeutics to decrease inflammatory effects in the lungs and consequent fibrosis since their actions in this respect do not appear to be long lasting [144–146]. Furthermore, dexamethasone has the potentially adverse action of decreasing immune responses . It will be important to study more direct inhibitors of the ATX-LPA-inflammatory cycle in attenuating RT-Induced fibrosis.
Conclusions and future directions
LPA signalling is controlled by a balance between LPA production by ATX and LPA degradation by LPPs and also by the relative expressions of six LPA receptors on different cells. Our current thinking is that ATX secretion is promoted by inflammation and this establishes a feedforward cycle though LPA that increases further inflammation, which promotes more ATX secretion. This process facilitates wound healing and when this is completed, inflammation normally resolves. When inflammation is not resolved, chronic inflammation becomes maladaptive and sustained LPA signalling contributes to conditions including arthritis, neurologic pain, obesity, cardiovascular diseases, complications of pregnancy, fibrosis and cancers. These associations with disease emphasize the growing importance of ATX and LPA signalling.
This area of research is being actively pursued in academic institutions and the pharmaceutical industry. This work has progressed such that first in class drugs for inhibiting ATX (e.g. GLPG1690) and signalling through LPA1 receptors (e.g. BMS986020) have proved to be effective in Phase 2 clinical trials in attenuating the progression of idiopathic pulmonary fibrosis. This work lays the foundation for more widespread targeting of LPA signalling in the treatment of a wide variety of clinical conditions that are characterized by chronic inflammation as discussed in this review. It will be important to determine how to minimize possible adverse side effects of targeting the ATX-LPA-inflammatory axis and how to maximize the efficacies of different drugs that are used in these treatments for different inflammatory conditions. The treatment of cancers will probably depend on understanding how to control LPA signalling as an adjuvant to improving outcomes from chemotherapy and RT. Further research in how to control the maladaptive ATX-LPA-inflammatory feed forward cycle (Figure 3) at different levels could lead to a better understanding of the optimum therapeutic targets in this cycle that can be used to treat these different inflammatory conditions.
The concentration of extracellular LPA depends mainly on the balance of its production by ATX and the rate of degradation by the LPPs.
LPA signals through six G-protein-coupled receptors.
LPA signalling initiates a feedforward proinflammatory cycle by stimulating the secretion of inflammatory cytokine/chemokines, which increase more ATX secretion and LPA signalling.
This cycle has a beneficial effect in wound healing.
However, the cycle is maladaptive in chronic inflammatory diseases in which the inflammation is not resolved. Consequently, attenuating LPA signalling in conditions such as neuropathic pain, arthritis, obesity, cardiovascular diseases, complications of pregnancy, scleroderma, idiopathic pulmonary fibrosis and cancers could provide novel strategies for the treatments of these conditions.
Future studies should establish that are the best targets in the ATX-LPA-inflammatory cycle to alleviate morbidity and mortality from these conditions.
The authors declare that there are no competing interests associated with the manuscript.
D.N.B. was supported by an Innovative Grants from the Canadian Cancer Society Research Institute (703926)/Canadian Breast Cancer Foundation (300034) and the Women and Children’s Health Research Institute (WCHRI) at the University of Alberta and by a project grant from the Canadian Institutes of Health Research (PJT-169140). D.G.H. was supported by Canadian Institutes of Health Research, Canadian Breast Cancer Foundation (300034), WCHRI and Li Ka Shing Institute of Virology at the University of Alberta.
Both authors contributed to writing and editing the article.
lysophosphatidate at physiological pH but often referred to as lysophosphatidic acid
lipid phosphate phosphatase
nuclear factor kappa-light-chain-enhancer of activated B cells
nuclear erythroid 2-like 2
smooth muscle cell
tumour necrosis factor-α