Autotaxin signaling , purinergic receptors and lung damage

ATX is a secreted enzyme that produces lysophosphatindic acid (LPA) in plasma. For several years investigators have characterized endogenous factors that regulate ATX expression and compartmentalization. However many questions remain unanswered and this article highlights our recent finding that autotaxin (ATX) is readily induced by toxic environmental chemicals. LPA binds G-protein coupled receptors that affect basic cell functions. The interest in the ATX-LPA axis stems from its role in embryogenesis and its association to diseases such as allergic asthma, idiopathic lung fibroses, rheumatoid arthritis, wound healing and several common types of cancer. In our study we used toluene diisocyanate (TDI) and other diisocyanates. Diisocyanates are low-molecular weight industrial chemicals notorious for being respiratory sensitizers and lung toxicants. Mechanisms behind these effects are not sufficiently characterized. We mainly used TDI and found that TDI in the nM range induced a rapid secretion of ATX from respiratory epithelial cells. Two purinergic receptors, P2X4 and P2X7, were implicated in this effect of TDI, suggesting that there is a “P2X-ATX axis” in bronchial epithelium that is sensitive to diisocyanates. We also showed associations between TDI exposures, LPA levels in plasma and symptoms reported by exposed individuals. Thus, our data support a role for the ATX-LPA axis in TDI toxicity. Furthermore, they suggest novel ways to study the regulation of ATX expression. Of particular interest is to understand how ATX expression is affected by purinergic receptors, and to investigate a possible involvement of ATX in asthma induced by diisocyanates and perhaps other low-molecular weight environmental chemicals. Our study also raises questions about current occupational exposure limits for diisocyanates.


Autotaxin (ATX), the ATX-LPA axis, inflammation and xenobiotics
Autotaxin (ATX) is an enzyme, a lysophosphoslipase, which is secreted by many cell types even though most studies on ATX expression have focused inflammatory cells [1]   .ATX has enzyme activity in plasma and converts lysophosphatidylcholine to lysophosphatidic acids (LPA).LPA are agonists for G-protein coupled cell membrane receptors (LPAR1-6), which influence e.g.embryogenesis and basic cell functions including growth, migration, DNA repair etc.Several studies indicate a role for the ATX-LPA axis in inflammatory conditions of the lung such as asthma [2] and idiopathic lung fibrosis [3] , but also in wound healing [4] , rheumatoid arthritis [5,6] and in linking inflammation and carcinogenesis in e.g.mammary cancer [7] .
Lysophosphatidylcholine is found in excess in blood and details about compartmentalization and regulation of ATX activity is thus of interest but insufficiently known.Most

RESEARCH HIGHLIGHT
studies describe endogenous regulatory factors of importance for inflammation such as TNF- and IL-1 [4] .ATX binds to integrins and other structures on the cell surface, which may confer strict localization of LPA delivery to LPARs [8] .
We are interested in the possibility that ATX is a sensitive target for toxic environmental chemicals and we recently reported that low-molecular weight xenbiotics can act as potent inducers of ATX.We show that nano-molar concentrations of toluene diisocyanate (TDI) [9] , a notorious lung toxicant and respiratory sensitizer [10] , induce a sudden release of ATX from lung epithelial cells and we couple ATX to TDI toxicity [9] .

Proof of principal: TDI is a potent activator of a "P2X-ATX" axis in vitro and in humans
Our studies of TDI and other diisocyanates highlight the need for a detailed characterization of the regulation of the ATX-LPA axis.In particular its activation by xenobiotics has been neglected.Diisocyanate exposure is perhaps the most common cause for occupational asthma and the mechanisms for the development of this type of asthma are not adequately known [11] .For example, increased levels of IgE are found in most cases of allergic asthma but relatively rarely in diisocyanate asthma patients, and this lack of IgE antibodies has puzzled researchers for many years [11] .We find that two purinergic receptors, P2X4 and P2X7, are needed for TDI-induced ATX secretion and a subsequent de novo synthesis of ATX in lung epithelial cells [9] .How the initial secretion relates to the subsequent protein synthesis is not known, but our data suggest not yet described control mechanisms for ATX expression; previous work indicate an inhibitory feedback loop between LPA and ATX in e.g.thyroid and melanoma cell lines [4] .Further work showed that P2X4 and P2X7 co-localize in response to TDI, and that ATX is rapidly concentrated in plasma membrane vesicles, presumably as a step preceding the secretion of ATX.Our data also show that the cytokine MCP-1 (also referred to as CCL2) affect the purinergic receptors so that vesicle formation and ATX secretion was facilitated [9] .The low concentrations of TDI needed for ATX activation in our cell models suggested that workers exposed to TDI in their work environment should exhibit signs of increased ATX activity.Although TDI air levels in industry are restricted by very low occupational exposure limits (OELs), it is possible that TDI in blood might rise to nM concentrations [9] .It is also possible that aerosol droplets, which reach epithelial surfaces in the respiratory tree, might contain much higher concentrations of diisocyanates [12] .In any event, in collaboration with Work Environment Medicine in Lund, Sweden, we analyzed LPA levels in workers´ plasma, and found a dose-response relationship between TDI exposure and LPA levels [9] .This dose response corroborates the relevance of our cell data and support the existence of a novel "P2X-ATX axis" in humans.
We could also correlate symptoms among workers to their LPA levels in plasma.This correlation to symptoms indicates a role for purinergic receptors and ATX in diisocyanate toxicity [9] .Thus, the P2X-ATX axis seems to be a very sensitive target in bronchial epithelial cells for inhaled xenobiotics such as TDI.It also stands out as the most sensitive target for diisocyantes so far described in the literature [9] .Importantly, our data indicate a role for the P2X-ATX axis in TDI toxicity at concentrations in inhaled air encountered in today's industrial environments.

Diisocyanates and occupational exposure limits (OELs)
In our study we analyzed plasma samples from individuals exposed to less than 5 ppb TDI (which was the Swedish OEL at the time of sample collection).Although there were no cases of occupational asthma in our study group, LPA levels correlated to other adverse effects.These included sneezing, which is a symptom that can be taken as an early sign of asthma development [9] , so our data indicate that an OEL of 5 ppb over an 8 hour work day is too high.Whether the current Swedish OEL (2 ppb) is protective is not known.However, the ceiling value (with a reference period of 5 min) is set to 5 ppb, and we detected ATX release within shorter time periods than that.So the margin of safety is slim or absent for at least the ceiling value.Of further interest is that it has been claimed that current OELs in USA do not sufficiently protect workers as several cases of diisocyanate asthma were reported in recent years [10] .US OELs are higher than the Swedish OELs [10] so our data support this conclusion.

Future studies of P2X-induced ATX
In future studies we will investigate polymorphisms in genes/proteins implicated in the P2X-ATX axis.The goal is to identify genes that predispose to serious adverse effects such as asthma, so that sensitive individuals can be identified.We will also characterize the P2X-dependent mechanism activating ATX in more detail.We speculated [9] that the efficacy of TDI to engage the two purineric receptors and to trigger ATX secretion depends on the bi-functionality of diisocyanate molecules [13] .In support of this idea we find that low doses of other diisocyanates than TDI activated ATX, whereas the monoisocyanate methyl isocyanate (MIC) did not do so [9] .Our data indicate that TDI rapidly create proximity (< 40 nm) between P2X4 and P2X7 and we assume that TDI in this way create a short-lived (10 min) functional heteromeric receptor.
Several previous studies have indicated functional associations between P2X4 and P2X7.For example, it has been suggested that a P2Xcilia receptor consists of P2X4 and P2X7 subunits [14] .This is of possible importance for the defense against inhaled noxious agents such as TDI as it may contribute to sustained high levels of cilia activity in airway epithelium.The interrelationship between this [14] and our work should thus be investigated.Further, biglycan, an endogenous degradation product from inflamed connective tissue and a ligand for TLR4 and TLR2, activates P2X4 and P2X7 in macrophages and form immune-perceptible heteromeric complexes [15] .Computational studies of low-molecular weight compounds that causes occupational asthma or act as respiratory sensitizers indicate cross-linking capabilities as an important factor [13] .However, cell transfected P2X4 and P2X7 are not cross-linked by disuccinimidyl suberate (DSS) [16] , and this is in line with our data [9] .Alternative models for P2X4-P2X7 interactions have been suggested [17] .
So although it remains to be shown how P2X4 and P2X7 interact in response to TDI, a more general biological role than that previously anticipated for this combination of purinergic receptors [18] is suggested.In further studies these questions will be addressed.The involvement of ATX in wound healing [4] and cell motility [7,19] suggests additional roles for ATX in the bronchial tissue remodeling seen in occupational asthma [20] .So also the remodeling should be studied in this connection.Of impotence for these mechanistic studies is that it can be anticipate that TDI and other diisocyantes will be excellent tools for characterizing P2X4-P2X7 interactions and their role in ATX activation.

Are there other environmental chemicals that activate ATX?
In an earlier study we found that eight chemicals, including TDI, had the capacity to induce ATX.These chemicals were selected because they induced rat pancreatic acinar carcinomas in cancer bioassays performed by the National Toxicology Program (NTP).Based on the observed ATX induction, we suggested that these chemicals facilitated pancreatic carcinogenesis via ATX [21] .Of these eight chemicals, TDI is so far the only one that we have studied in some detail, but the potency of TDI to induce ATX is now a stimulus for further investigations.In particular lung toxicants should be of interest as ATX might be a sensitive target in bronchial cells, and as mentioned above, ATX also has a role in allergic asthma and in idiopathic lung fibrosis [2,3] .In earlier studies ATX has been implicated in lung fibrosis induced by bleomycin, a cancer drug with serious fibrotic side effects in the lung [22] .Considering this background it might be meaningful to test both established and possible or suspected lung toxicants in our environment.In literature surveys bleomycin stands out as the first low-molecular weight xenobiotic shown to induced the ATX-LPA axis [23] .TCDD also belong to this limited group of chemicals and has been shown to induce ATX in the mammary cancer cell line MCF-7.A functional role was indicated by the finding that LPA promoted migration of these cells [19] .

Diisocyanates, purineric receptors, sensitization and innate immunity
Diisocyanates are respiratory sensitizers and contact skin sensitizers, but the mechanisms leading to these effects are not adequately characterized [10,24] .It has been suggested that protein cross-linking might lead to conformational changes in cross-linked proteins and the formation of newly exposed epitopes that provoke immune reactions [13] .Our data so far challenge this notion.They are more in line with indications that diisocyanates activate innate immunity.For example, it has been shown that polymorphism in the CD14 gene, a component in the innate immune system and the first described pattern recognition receptor (PRR), associates with diisocyanate asthma [25] .Furthermore, albumin conjugated to HDI, another diisocyanate, induces chitinase-1 that also participates in PRR activation [26] .ATX [27] and purineric receptors, and in particular P2X4 and P2X7 [28] , have also been implicated in innate immunity.For example and as mentioned above, biglycan has been shown to affect P2X4 and P2X7.This finding connects P2X4 and P2X7 to innate immunity as biglycan is a ligand for Toll-like receptors TLR2 and TLR4 and may act as a "danger signal" (DAMP) in innate immunity [15] .Of further possible interest is that P2X4 is implicated in neurological sensitization seen after cell damage [29] , and P2X7 in the development of skin contact dermatitis and hypersensitivity in mice [30] .In summary, we find that our data have established additional connections between innate immunity and diisocyanate toxicity.
More recent studies on hyperoxic lung damage in mice add further support and also suggest a link to both P2X7 and to ATX.Thus, death resulting from hyperoxic toxicity in mice depends on ATX induction in invariant natural killer cells (iNKT cells), often described as a bridge between innate and adaptive immunity, but also depends on extracellular ATP (that is a ligand for P2X7) [31] .Interestingly, not only hyperoxic injury but also hypoxia induce ATX levels [19] , so cell damage might be the common denominator in these studies [19,31] and, as mentioned above, ATX is implicated in wound healing [4] .

Cell based test models for skin and respiratory sensitizers
The observation that diisocyanates induce skin contact hypersensitivity, and that skin exposure to TDI is employed in experimental models to induce sensitization [32] , is of interest for us.In unpublished work we note that many genes/proteins downstream of ATX signaling are the same as those induced by test substances in cell-based tests for skin sensitizers [33] .We have also seen that other well-known contact sensitizing agents induce ATX in proportion to their sensitizing potency.This raises the question whether ATX is part of the sensitization process for at least some skin contact allergens.Perhaps ATX serves as a DAMP in the toxicity induced by sensitizing agents in dermal cells.This reasoning is in line with the above discussion on innate immunity and with studies on increased ATX levels in skin blisters and in wound healing [34] .A specific role for ATX and LPA in skin affections is further supported by data implicating ATX in pruritus caused by e.g.primary biliary cirrhosis or cholestasis of pregnancy [35] .
There is a strong demand for test models that accurately predict contact sensitizers and respiratory sensitizers [33] .In particular there is a demand for cell based models.Further studies characterizing a possible role of ATX in skin sensitization is thus warranted.They may help clarifying mechanistic aspects of sensitization and may lead to improved test models.A cell based model for testing respiratory sensitizers has recently been suggested [36] .The model was based on a leukemia cell line and TDI was one of the positive controls.However, the concentration used was 40 000 times higher than that eliciting ATX release from respiratory epithelial cells [9] .This difference in concentrations illustrates a perhaps selective sensitivity to diisocyanates in bronchial epithelial cells.It also illustrates the need for mechanistic knowledge when choosing optimal cell types for the design of test models.

Diisocyanates, ATX, cancer and gender
TDI is classified by NTP as "reasonably anticipated to be a human carcinogen" and as "possibly carcinogenic to humans" (2B) by IARC.According to IARC there is "sufficient evidence in experimental animals", but "inadequate evidence in humans".A recent review by Mills and colleagues summarizes evidence indicating a role for increased ATX levels in carcinogenesis.For example, amplifications of the ATX gene, ENPP2, is a common genomic alteration in several human cancer types including mammary and lung adenocarcinoma, and transgenic overexpression of ATX or LPA receptors causes mammary cancer in mice [7] .We would like to highlight gender aspects of ATX expression in connection to its assumed role in human carcinogenesis.Women have higher ATX levels in plasma [37] and this raises the possibility that women are more susceptible to carcinogenic effects of diisocyanates.As discussed in our paper [9] , an increased incidence of lung cancer among women has been observed in diisocyante exposed cohorts, and NTP test data from rat and mouse bioassays indicate more tumors in females than in males (https://ntp.niehs.nih.gov/ntp/roc/content/profiles/toluenediisocyanates.pdf).Of perhaps further interest is that thyroid cancer, which is one of the few cancer types in non-sex organs that is more common in women than in men [38] , has been associated with increased ATX levels [39] .It may thus be suggested that women are focused in future epidemiological studies of carcinogenic effects related to diisocyante exposure.It may also be suggested that ATX or LPA in plasma are used as biomarkers for diisocyanate exposure.

Summary
This research highlight summarizes a recent study suggesting a role for ATX and LPA in lung toxicity induced by inhaled xenobiotics.Our data indicate a novel role for interacting P2X4 and P2X7 purinergic receptors in causing a rapid ATX activation.The regulation of the ATX-LPA axis is insufficiently characterized, and our data suggest new ways to study ATX expression and secretion.Our data also cast new light on mechanisms involved in diisocyanate asthma.Future studies may lead to a better understanding of how this and perhaps other diseases can be avoided.A critical objective is to characterize the toxicity induced by low-molecular weight chemicals and that initiate innate immunity.