ATX-LPA Receptor Axis in Inflammation and Cancer

Abstract

Lysophosphatidic acid (LPA, 1- or 2-acyl-sn-glycerol 3-phosphate) mediates a plethora of physiological and pathological activities via interactions with a series of high affinity G protein-coupled receptors (GPCR). Both LPA receptor family members and autotaxin (ATX/LysoPLD), the primary LPA-producing enzyme, are aberrantly expressed in many human breast cancers and several other cancer lineages. Using transgenic mice expressing either an LPA receptor or ATX, we recently demonstrated that the ATX-LPA receptor axis plays a causal role in breast tumorigenesis and cancer-related inflammation, further validating the ATX-LPA receptor axis as a rich therapeutic target in cancer.

Lysophosphatidic acid – a structurally simple lipid with diverse actions

Lysophosphatidic acid (LPA) is a simple lipid with a single fatty acyl chain, a glycerol backbone and a free phosphate group.¹ Despite the simplicity of its structure, LPA exerts multifaceted bioactivities including stimulation of proliferation, migration, and survival of many cell types. LPA binds to specific cell-surface G protein-coupled receptors (GPCRs). So far, there are six identified as bona fide LPA receptors, LPA1–6,^2–7, and several others which include intracellular receptor PPAR©⁸, putative LPA receptors (GPR87⁹ and p2Y¹⁰). The best-characterized LPA receptors are those of the endothelial differentiation gene (Edg) family including LPA1/Edg2, LPA2/Edg4 and LPA3/Edg7. The other Edg-family proteins, Edg1, 3, 5, 6 and 8, are receptors for a closely-related bioactive lipid, sphingosine 1-phosphate, which by itself, can exert an array of functions both similar and distinct from those of LPA.¹¹

Early evidence suggested that LPA is primarily synthesized in extracellular fluids. We found a marked elevation of lysophosphatidylcholine (LPC) in the plasma of ovarian cancer patients, suggesting that LPA was produced via a lysoPLD activity.¹² Umezu-Goto et al. demonstrated for the first time that the ectoenzyme ATX, possessed lysoPLD activity and is the main source of LPA production, resulting in LPA-mediated growth and motility of cancer cells.¹³ Interestingly, levels of the circulating LPA were decreased ~50% in mice carrying a heterozygous null mutation of Enpp2, the gene encoding ATX.^14,15 It is now established that ATX is a key enzyme that produces LPA in cancer.

LPA receptors are ubiquitously and variably expressed in most cell types. Upon binding to its receptors, LPA can elicit divergent signaling pathways through LPA1–3 via at least three subunits of G proteins, G_q/11, G_i/o and G_12/13.¹⁶ First, LPA induces transient elevation of cytosolic free Ca²⁺ concentration ([Ca²⁺]_i) through a pathway involving G_q/11, phospholipase C and protein kinase C. Second, LPA receptors couple to G_i/o leading to inhibition of adenylate cyclase, activation of the Ras/mitogen-activated protein kinase (MAPK/ERK) pathway and activation of protein kinase B/Akt pathway. Third, LPA stimulates a family of guanine exchange factor RhoA GEFs and RhoA GTPase through G_12/13.¹⁷ Together with the appropriate contexts, these pathways display potential to interact with one another, and in turn, integrate to regulate cell proliferation, survival, differentiation and motility through several immediate, early and late-onset genes.

In order to further and unambiguously identify the specific function of each Edg-family LPA receptor, mice with targeted deletion of the genes encoding LPA1–3 were created. Edg2^−/− mice showed ~50% neonatal lethality, a defect in suckling behavior in neonatal pups and craniofacial dysmorphism.¹⁸ However, Edg2^−/−/Edg4^−/− mice revealed no additional phenotype to those found in Edg2^−/− mice,¹⁹ suggesting that LPA2 may serve similar role to the LPA1 during development. Blastocysts of mice with null mutation of Edg7^−/− developed delayed implantation and improper spacing due to a decrease in prostaglandin-induced uterine contraction.²⁰ Enpp2^−/− mice are embryonic lethal with severe attenuation of blood vessel formation.^14,15 Together, these compelling genetic models point to the critical roles of LPA in fetal development. However, the early embryonic lethality in some of the LPA knockout mice hinders evaluating the significance of LPA signaling in late-onset disease models. Nevertheless, mice lacking LPA2 show a decrease in tumor development induced by colitis.²¹ demonstrating a critical role for LPA receptors and in this case LPA2 for development of adult onset diseases.

ATX-LPA receptor axis in cancer related- inflammation

In breast cancer, epidemiological evidence suggests that inflammation is associated with poor prognosis.²² Release and activation of growth factors and cytokines provide biochemical cues that exert a major influence on tumor cell survival. In addition, inflammatory cells show pro-tumorigenic roles, specifically tumor-associated macrophages and B cells, which are postulated to foster tumor growth and metastasis through release of cytokines and matrix remodeling enzymes.^{23, 24} The role of LPA and LPA receptor function in airway inflammation has been studied extensively revealing that LPA regulates receptor-mediated pro-inflammatory transcriptional factors including NF-κB, AP1, which regulate cytokine and lipid mediator production and secretion.^25–29

Until now, the role of LPA in mammary gland inflammatory diseases had not been defined. We recently demonstrated for the first time that the ATX-LPA receptor axis induces inflammation and tumor formation in the mammary gland by establishing transgenic mouse models expressing human ATX, LPA1, LPA2 or LPA3 under the MMTV-LTR promoter. Individual overexpression of each receptor resulted in a high frequency of chronic mastitis, hyperplasia, mammary intraepithelial neoplasia and invasive and metastatic tumors. Strikingly, inflammation was present in mammary glands whether or not tumors were present in the transgenic mice. Furthermore inflammation was detected earlier than tumorigenesis in the transgenic mice. The high frequency and early onset of chronic mastitis suggests that chronic inflammation could contribute to tumor development in these models.³⁰

We recently demonstrated that over expression of the Edg family of LPA receptors enhances the production of IL-6, IL-8, and VEGF in SKOV-3 and OVCAR-3 cells, while knockdown decreases production.³¹ This establishes a strong correlation between LPA function and production of pro-inflammatory cytokines, at least in ovarian cancer cells. Multiple mechanisms could contribute to LPA stimulating proinflammatory gene expression. NF-κB and ATF-2 both regulation expression of genes contributing to inflammatory reactions.^32–35 LPA enhances NF-κB expression via activation of PKC or AKT pathways and LPA induces ATF-2 production through Rho-CDC42-p38 MAPK pathways^36–38 (Figure 1). LPA stimulation also induces the activation of signal transducer and activator of transcription 3 (STAT3) and STAT5.³⁰ Together, these transcription factors coordinate the production of inflammatory cytokines, chemokines and the production of cyclooxygenase 2 (COX2).^38–41 Subsequently, cytokines bind to their own receptors and activate STAT family members,^{42, 43} further contributing to inflammation and to tumor cell activation^{44, 45} in an autocrine and paracrine feedforward cycle, stabilizing the production of inflammatory factors and the generation of an inflammatory tumor microenvironment. Finally, LPA also transactivates receptor tyrosine kinases (RTKs), such as EGFR,^{26, 46–49} PDGFRβ⁵⁰ and c-Met,^{51, 52} which can contribute to cytokine production.

Figure 1. Synthetic pathways for LPA and major LPA signaling pathways connecting inflammation and cancer.

ATX hydrolyzes lysophospholipids, in particular lysophosphatidylcholine (LPC), to produce bioactive LPA. Newly produced LPA acts on its own GPCRs via at least three distinct classes of heterotrimeric G proteins — Gq, Gi and G12/13 activating multiple downstream pathways and evoke its biological effects, including RAS-ERK pathway through Gi and Gq; PI3K-AKT pathway through Gi; PKC –GSK30β – β-catenin pathway through Gq (and /or Gi); Rho-CDC42 pathway through G12/13; Src-Stat pathway through Gi, which induce expression and activation of multiple transcriptors including STAT3, NF-κB, ATF-2, which induce cell proliferation and production of cytokines. The cytokines bind to their receptors inducing Stat 3 and 5 production and activation.

Understanding the mechanism by which the ATX-LPA receptor axis induces inflammation in the mammary gland could enhance our understanding of breast tumorigenesis, and especially inflammatory breast cancer (IBC). IBC is a rapidly progressing and highly aggressive form of cancer characterized by increased angiogenesis, high expression of VEGF, EGF and their receptors, hyperactivation of NF-κB and overexpression of E-cadherin.⁵³ For those patients who have HER2-positive IBC tumors, trastuzumab has been demonstrated activity but unfortunately this group represents only a small portion of the total number affected. The intracellular signaling signature in mammary tumors of transgenic mice over-expressing either ATX or LPA1, LPA2 or LPA3 receptors in mammary glands is remarkably similar to molecular abnormalities in IBC.³⁰ In the LPAR and ATX transgenic mice, we observed mastitis and mammary tumors³⁰ with increased levels of MIP-2 (the murine homolog of IL-8) and VEGF common observations in IBC. Thus, research is warranted into the potential role of LPA and LPA induced inflammation in IBC.

ATX-LPA receptor axis in breast and ovarian cancer

Although significant progress has been made in improving prognosis for patients with breast and ovarian cancer, unfortunately many patients still succumb to these diseases, which necessitate new molecular therapeutics to further improve outcomes. Enhanced ATX expression,^{54, 55} elevated LPA production,^{56, 57} decreased lipid phosphate phosphatase expression⁵⁸ and increased LPA receptor expression^{59, 60} are found in multiple cancer types, including ovarian and breast cancer. The role of LPA in ovarian cancer has been extensively studied. The most notable finding, which spawned interest in this field, was that LPA is present at high levels in the ascitic fluid of ovarian cancer patients.⁶¹ Interestingly, peritoneal-lining mesothelial cells as well as ovarian cancer cells can also produce LPA,⁵⁶ suggesting that both malignant and benign cells are sources of LPA in the ascites. Consistent with the role of LPA in tumor invasiveness, ovarian cancer cells frequently adhere to the lining cells as well as invade underlying stroma.^{62, 63} Moreover, LPA2 and LPA3 transcripts were increased between 15% and 49% in benign tumors and early and late stage ovarian carcinomas suggesting that both LPA production and action is aberrant in ovarian cancer.⁵⁷ Systemic analyses of the LPA transcriptomes in human epithelial ovarian cancer suggested that LPA regulates several genes associated with cell adhesion and migration, which in turn, correlate with a poor prognosis in patients.⁶⁴

In accord with ovarian cancer, breast cancer cells exhibit aberrant expression of ATX, which promotes tumor aggressiveness.⁶⁵ Expression of ATX is markedly upregulated by the Jun oncoprotein in embryonic chick fibroblasts.⁶⁶ In contrast, transcript levels of ATX were decreased by tumor suppressor CST6 in human breast cancer cells.⁶⁷ A role for LPA receptors has also been suggested in the development and progression of breast cancer.⁶⁸ LPA2 expression was upregulated in patients with invasive mammary ductal carcinoma,⁶⁹ whereas the tumor suppressor Nm23-H1 downregulated LPA1 expression and inhibited metastasis in breast cancer cells.⁷⁰ Inhibition of LPA1 receptor expression suppressed tumor growth and bone metastasis in the xenografts of MDA-BO2 breast cancer cells in mice.⁷¹

Thus, it is clear that the aberrant regulations of ATX-LPA signaling are strongly associated with the pathogenesis of both breast and ovarian cancer. We therefore sought to determine whether ATX-LPA signaling by itself is sufficient to induce tumorigenesis. Strikingly, transgenic mice expressing either ATX or each of the three major Edg family LPA receptors under the MMTV-LTR promoter exhibited chronic mastitis, mammary intraepithelial neoplasia together with invasive and metastatic breast cancer.³⁰ In support of a role for LPA and its receptors in tumor pathophysiology, overexpression of each of the Edg family LPA receptors in SKOV-3 cells that endogenously produce LPA enhanced cell proliferation and invasion in vitro as well as formation of ascites and metastasis in vivo.³¹ Together, our data from animal models of breast and ovarian cancer indicate that LPA production and action and specifically ATX and LPA Edg family receptors are candidates for drug development in breast and ovarian cancer.

Beginning with reverse phase protein arrays (RPPA) functional proteomic analysis on mammary cancers isolated from transgenic mouse models³⁰ and combining the data with published reports, a resultant pathway signature mediated by the ATX-LPA receptor axis is illustrated using ScienceSlides 2008 (Figure 1). ATX-LPA receptor axis in cancer signals are mediated by multiple pathways, including at least three distinct G proteins (G_q/11, G_i/o and G_12/13) coupled to receptors, inflammatory cytokine pathways, transactivation of receptor tyrosine kinase (RTK) signaling, which, in turn, feeds into multiple transcriptional mediators,^72–82 see Figure 1.

Expression of the estrogen receptor (ER) is a key determinant of the pathophysiology of breast cancer.⁸³ ER-positive breast cancers are characterized as luminal subtype associated with a better prognosis and response to hormonal manipulation.^72–82 Primary metastatic sites of the ER-positive cancers are lung and bone, and less frequently liver and brain, whereas ER-negative cancers spread mainly to visceral organs.⁸⁶ The expression of ER was upregulated in a subset of the transgenic mice, providing a novel tumor model for invasive and metastatic ER-positive tumors. Some findings have also suggested that estrogen could regulate the functions of LPA. For example, estrogen has been shown to modulate actions of LPA through LPA1 in C9 hepatic cells.⁸⁷ Moreover, estrogen was found to induce expression of LPA3 receptors in porcine uterine endometrial explants.⁸⁸ However, the of ER in LPA-induced breast cancer progression remains to be elucidated.

Targeting the ATX-LPA receptor Axis in Cancer

The mortality rate for breast cancer in the U.S. has been in decline for over a decade and continues to decrease 1.8% every year.⁸⁹ For patients with this disease, advances in screening, early detection and chemotherapy integrate to contribute to the improved prognosis⁸⁹ Targeted therapeutics provide best hope for improving outcomes for breast cancer.

For patients with HER-2-positive breast cancer, the advent of trastuzumab (Herceptin™) contributed significantly to improved outcomes for breast cancer patients. It has also provided the ideal model for future breast cancer drug development. In certain breast cancers, the amplified and overactive HER-2 receptor at the cell surface continuously signals cellular growth and survival, which ultimately drives disease development and progression. The combination of an approach to identify patients with amplified HER2 (herceptest™) and an effective therapeutic approach trastuzumab, resulted in benefit to patients with amplified HER2.

Potential of LPA pathway-targeted therapeutics for breast cancer

Previous work by our group and others combined with the new transgenic models³⁰ supports a role for LPA, ATX, LPA receptors and lipid phosphate phosphatases in cancer.^{31, 69, 79, 90–94} Thus, there are multiple potential therapeutic targets in this pathway that could function as targets for anti-cancer therapeutics. In fact, the therapeutics under development in both academic and biotech centers encompass several different goals: monoclonal antibodies which bind to and inhibit LPA, stabilized analogues of naturally occurring molecules that inhibit ATX and chemical compounds that function as LPA receptor antagonists.⁹⁵

Applying the current standard of care to breast cancer would necessitate the identification of tumor driven by the LPA pathway, as is already done with hormone receptor and HER2 status at the time of diagnosis. Based on that result, patients with an active LPA pathway would be eligible for therapeutics targeting the LPA production and function.

Targeting LPA receptors

Since most actions mediated by LPA occur through GPCRs, and GPCRs have been fruitful targets, multiple groups have developed inhibitors to LPA receptors.⁹⁵ Each of the Edg family LPA receptors has functional antagonists, although not all are specific to one receptor, many have dual receptor antagonism because of technical obstacles and similarities in receptor binding. For examples, there are number synthetic compounds (e.g. Ki16425,⁹⁶ DGPP 8:0,⁹⁷ PA 8:0,⁹⁷ VPC12249⁹⁸) that are antagonists of both LPA1 and LPA3. Since Edg family LPA receptors retain the capability of inducing similar signaling events and functional outcomes, although each of them preferentially link to particular pathways and functional outcomes, a lack of exquisite receptor selectivity may not be a problem.^{30, 31} Indeed, many current targeting therapeutics are pleiomorphic in activity blocking multiple kinases.

Previous work implicated that LPA1 receptor contributes to metastasis.⁹⁹ Interestingly, siRNA targeting either LPA1 receptors or a pharmacological blockade of LPA1 receptor activity with the LPA1/LPA3 antagonist Ki16425 inhibited metastasis to bone in vivo.⁷¹ In support of this concept, cDNA array analysis demonstrated that the LPA1 receptor plays a major role expression of genes involved in invasion and motility.¹⁰⁰ Furthermore, a comprehensive screen using cDNA overexpression in breast cancer cells expressing inducible HER-2 variants demonstrated that the LPA1 receptor cooperates with HER-2 though a hypothesized mechanism of ligand-independent dimerization of LPA1 with HER-2 facilitating cell proliferation, morphogenesis and migration.⁶⁸ In addition, our studies suggest that the LPA1 receptor may contribute to cancer metastasis.^{30, 31}. Ki16425, DGPP 8:0, PA 8:0, VPC-32183, Darmstoff analogues, tetradecyl-phosphonate, phosphonothioate and fluoromethylene phosphonate ccPA analogues, methylene phosphonate analogues, isoxazole derivative may thus be useful in preventing breast cancer metastasis to bone.

The LPA2 receptor is preferentially and often exclusively expressed in more than half of bowel cancers. In addition, the expression of the LPA2 receptor is correlated with a higher rate of invasion and metastasis⁹³ and is significantly overexpressed in patients with invasive ductal carcinoma, particularly in postmenopausal women.⁶⁹ Multiple cancer cells have been found to expresses abundant LPA2 receptor levels, including colon, ovary, breast, and thyroid. This receptor may regulate LPA-induced stimulation of the chemokine growth regulated oncogene (GROα) which contributes to angiogenesis and tumorigenesis of ovarian cancer cells.¹⁰¹ Expression of the human LPA2 receptor in the breast of multiply pregnant mice was sufficient to result in the development of mammary carcinomas with a high incidence and short tumor-free period compared to other LPA receptors,³⁰ suggesting that the LPA2 receptor may be a particularly attractive target in breast cancer. A number of compounds including Tetradecyl-phosphonate, Darmstoff analogues, FDP, FMP, methylene phosphonate analogues and Compound 35^102–105 have been developed as antagonists of LPA2 receptor.

So far, more antagonists exist for the LPA3 receptor than other LPA receptors because of its unique and advantageous basic residue near the phosphate headgroup (e.g. Ki16425, DGPP 8:0, PA 8:0, VPC-32183, VPC-12249, FAP-12, Darmstoff analogues, tetradecyl- phosphonate, (2R)-TPA 8:0, SDP, FR, FDP, FMP).^102–105 Our demonstration that transgenic expression of the LPA3 receptor is sufficient to result in mammary carcinomas with a high frequency of tumor metastasis, combined with other studies suggests that LPA3 may be a particularly attractive target in breast and ovarian cancer.^30,31

Targeting ATX and LPA

An alternative strategy to LPA receptor antagonism is to inhibit the production of LPA or to immunoneutralize LPA. Inhibiting LPA directly could have all of the effects described above as well as blocking yet unknown or uncharacterized LPA receptors such as LPA4–8 and bypassing potential partial agonist effects against receptor subsets. Thus, directly targeting LPA and/or targeting the enzyme responsible for the bulk of LPA production are logical approaches. Development of monoclonal antibodies targeting LPA has commenced with Lpathomab™⁹⁵ and monocolonal antibodies that bind to S1P have been successfully developed and tested in small animal models, including Asonep (Sphingomab^®)¹⁰⁶. Indeed, Asonep is progressing through phase I clinical trials.

LPA is produced through at least two pathways: 1) in blood, mainly from lysophospholipids by ATX; 2) by inflammatory cells, activated platelets, adipocytes, mesothelial cells and some cancer cells, mainly from phosphatidic acid by phospholipases. ATX plays a critical role in both pathways of LPA production.¹⁰⁷ Cell line and our recent transgenic studies identified potential roles for ATX in cancer progression, tumor cell invasion and metastasis including promotion of tumor angiogenesis¹⁰⁸. LPA and S1P are natural ATX inhibitors providing an approach for the development of ATX inhibitors. These include synthetic analogues of the naturally-occurring ATX-inhibitor cyclic phosphatidic acid (carba analogues of cyclic phosphatidic acid or ccPA),^{109, 110} L-histidine,¹¹¹ VPC8a202,^{112, 113} Darmstoff analogues,¹¹⁴ Thiophosphoric acid O-octadec-9-enyl ester¹¹⁵ and small molecular inhibitors identified using high-throughput screening.¹¹⁶ Additional ATX inhibiting compounds are under development by multiple groups world-wide, attesting to the importance of this enzyme as a therapeutic target in tumorigenesis.

Research in progress and perspective

Aberrant expression of LPA receptors and ATX enhance ER expression in mammary glands and induced ER positive mammary cancer compatible with functional ER signaling.¹ Studies are underway to verify the mechanism by which the ATX-LPA receptors axis enhances ER expression. Both in vitro and in vivo studies suggest that the ATX-LPA receptors axis promote tumor metastasis, but further studies are needed to determine the mechanism. Indeed, the late and variable onset of tumors suggest that other mutational events in the transgenic mice cooperate with LPA signaling to induce tumorigenesis. Alternatively, since LPA is a potent survival factor, LPA signaling may allow cells to survive oncogenic success leading to tumor development. Studies have demonstrated that augmented LPA signaling contributes to cancer related inflammation and breast cancer initiation and progression, indicating that LPA producing enzymes and LPA receptors can be targets for treatment or chemoprevention of inflammation and cancer. Development of compounds targeting LPA production and action has been initiated and hopefully will soon enter clinical trials determining whether the ATX-LPA axis is indeed a useful target in cancer.

Abbreviations

LPA

lysophosphatidic acid

ATX

autotaxin

GPCR

G protein-coupled receptors

Edg

endothelial differentiation gene

LPC

lysophosphatidylcholine

MAPK

mitogen-activated protein kinase

PKB/AKT

protein kinase B

STAT

signal transducer and activator of transcription

ATF-2

activating transcription factor 2

COX2

cyclooxygenase 2

RTK

receptor tyrosine kinase

IBC

inflammatory breast cancer

NF- κB

nuclear factor kappa B

IL

Interleukin

EGFR

epidermal growth factor receptor

PDGF

platelet-derived growth factor

MIP-2

macrophage-inflammatory protein-2

VEGF

vascular endothelial growth factor

RPPA

reverse phase protein arrays

ER

estrogen receptor

GRO

growth regulated oncogene