Phospholipase D and Choline Metabolism

Abstract

Phospholipases D (PLDs) catalyze hydrolysis of the diester bond of phospholipids to generate phosphatidic acid and the free lipid headgroup. In mammals, PLD enzymes comprise the intracellular enzymes PLD1 and PLD2 and possibly the proteins encoded by related genes, as well as a class of cell-surface and secreted enzymes with structural homology to ecto nucleotide phosphatases/phosphodiesterases as typified by autotaxin (ENPP2) that have lysoPLD activities. Genetic and pharmacological loss-of-function approaches implicate these enzymes in intra- and inter- cellular signaling mediated by the lipid products phosphatidic acid, lysophosphatidic acid, and their metabolites, while the possibility that the water-soluble product of their reactions is biologically relevant has received far less attention. PLD1 and PLD2 are highly selective for phosphatidylcholine (PC), whereas autotaxin has broader substrate specificity for lysophospholipids but by virtue of the high abundance of lysphosphatidylcholine (LPC) in extracellular fluids predominantly hydrolyses this substrate. In all cases, the water-soluble product of these PLD activities is choline. Although choline can be formed de novo by methylation of phosphatidylethanolamine, this activity is absent in most tissues, so mammals are effectively auxotrophic for choline. Dietary consumption of choline in both free and esterified forms is substantial. Choline is necessary for synthesis of the neurotransmitter acetylcholine and of the choline-containing phospholipids PC and sphingomyelin (SM) and also plays a recently-appreciated important role as a methyl donor in the pathways of “one-carbon (1C)” metabolism. This review discusses emerging evidence that some of the biological functions of these intra and extracellular PLD enzymes involves generation of choline.

Overview of Choline Metabolism

Figure 1 summarizes the pathways of choline metabolism that are discussed in detail in the following sections. Choline is a ubiquitous metabolite present in tissues, plasma and other biological fluids. Humans can synthesize small amounts of choline by sequential methylation of phosphatidylethanolamine (PE) to phosphatidylcholine (PC) in the liver¹. Hence, choline is considered an essential nutrient that must be obtained from the diet to supplement this smaller endogenously/de novo synthesized pool². Plasma choline is derived from three major sources (1) dietary choline - as free base or as a constituent of phospholipids present within many foods (2) endogenous synthesis – principally in the liver (3) liberation from choline-containing phospholipids, which are major constituents of all cell membranes. The most abundant sources of esterified choline, phosphatidylcholine and sphingomyelin, are essential for both the structural integrity of cellular membranes and for lipid-dependent signaling pathways, both of which are required for cancer cell growth³. Free choline is released from esterified choline lipid sources by hydrolysis catalyzed by phospholipase D enzymes. Dietary choline is absorbed by enterocytes in the lumen of the small intestines mediated by choline transporter proteins³. Free choline enters the portal circulation of the liver, whereas PC enters the circulation via lymph as a phospholipid constituent of triglyceride-rich chylomicrons. Choline can be phosphorylated to phosphocholine or oxidized to form betaine by the gut microbiota, or in some cell types such as hepatocytes, to feed into the choline biosynthetic pathway and one-carbon metabolism⁴. It is also a precursor to the neurotransmitter acetylcholine. Although choline plays essential roles, it is also involved, either directly or indirectly via its metabolites, in chronic non-communicable diseases, cardiovascular diseases, and cancers.

Figure 1. Role of Diet and Phospholipase D enzyme in the complex network of choline pathways and Choline-Mediated 1C metabolism.

Phospholipase D (PLD) catalyzes release of choline from either diet-derived or endogenously synthesized phosphatidylcholine (PC). The free choline product of PLD activity, or obtained from diet, is a substrate for the synthesis of acetylcholine, trimethylamine (TMA), and phosphatidylcholine (PC) via Cytidine diphosphate choline (CDP-choline). Formation of phosphocholine catalyzed by choline kinase α (ChoKα) is the initial step of de novo PC synthesis. Choline can also be oxidized to betaine. The oxidation of choline to betaine links PLD enzyme activity to 1C metabolism through de novo synthesis of methionine from homocysteine. The byproduct of this reaction dimethylglycine (DMG) is also a potential methyl donor. The methionine product is utilized for synthesis of S-adenocylmethionine (SAM) catalyzed by methionine adenosyltransferase 2A (MAT2A). SAM is the universal methyl donor for a wide range of substrates including phosphatidylethanolamine (PE), DNA, RNA and proteins including histones, and generating S-adenosylhomocysteine (SAH). SAH is hydrolyzed to homocysteine which couples 1C metabolism to the transsulfuration pathway involving sequential synthesis of cystathionine and cysteine and ultimately forming glutathione, taurine and hydrogen sulfide (H₂S). SAM can also be decarboxylated to decarboxylated S-adenosylmethionine (dc-SAM) which supports synthesis of the polyamines, spermine and spermidine. The byproduct 50 -methylthioadenosine (MTA) is salvaged back to methionine for SAM generation. The highlighted enzymes PLD, ChoKα and MAT2A are presently attractive therapeutic targets for treatment of cancer.

Dietary sources of choline

Dietary choline can be obtained from a wide variety of foods and supplements. Because de novo synthesis of choline alone is not sufficient to meet human requirements, choline was officially recognized as an essential nutrient by the Institute of Medicine in 1998³. It is available in foods as both water-soluble (free choline, phosphocholine, and glycerophosphocholine) and lipid-soluble forms (phosphatidylcholine and sphingomyelin)⁵. The main dietary sources of choline in the United States are animal products such as meat - especially beef liver, poultry, eggs, fish, and dairy products⁶. Other sources rich in choline include cruciferous vegetables such as cauliflower and broccoli, nuts, seeds, certain beans and whole grains. Additives and supplements containing choline are also readily available. The most common choline-containing food additive, lecithin, is usually derived from sunflower seeds, eggs, or soybeans. Lecithin is used as an emulsifying agent in routinely-consumed processed foods such as salad dressings, gravies and margarine. It can also be purchased as an over-the-counter supplement. Other forms of supplements high in choline include choline chloride, CDP-choline and alpha-GPC. The choline oxidation product, betaine, is also available as a supplement.

The Food and Nutrition Board of the Institute of Medicine recommends daily intake of choline of 425 and 550 mg for adult females and males respectively³. The Tolerable Upper Intake level for adults is estimated at 3.5 g/day. Dietary choline intake is calculated based on estimates of the free choline and phosphatidylcholine content of foods. There are no reliable estimates of the frequency of use or amount of these dietary supplements consumed by individuals in the United States and Canada. However, there are reports suggesting that the average dietary intake of choline in adults could be higher and is underreported⁷. According to the US Dietary Guidelines Advisory Committee (DGAC) 2015–2020 report, eggs provide the most abundant source of choline for individuals whose intake is at or above the recommended levels. Egg yolks provide 680 milligrams of choline per 100 grams².

Choline insufficiency in the diet can lead to liver damage, muscle damage and nonalcoholic fatty liver disease (NAFLD or hepatosteatosis)⁸. Furthermore, choline deficiency is one of the few single-nutrient deficiencies that causes increased spontaneous carcinogenesis⁹. In rats and mice, diets low in choline and other methyl-group donors such as folate, methionine, serine, vitamin B12 and betaine result in spontaneous hepatocarcinomas, or sensitize rodents to hepatic carcinogens such as aflatoxin B126¹⁰. Oral administration of lecithin (containing esterified choline) or free choline for a prolonged period of time has been shown to increase plasma choline levels^11,12. High dietary intake of choline causes adverse effects such as hypotension, fishy body odor, vomiting, excessive sweating and salivation, and liver toxicity^13–16. High choline consumption also increases the production of trimethylamine oxide (TMAO), a choline metabolite that has been linked to a higher risk of cardiovascular disease, in a dose-dependent manner in adults¹⁷. The association between dietary intake of choline and cancers is less clear. Overall, the recognition of dietary association with cancer incidence has only began to be appreciated. For example, the US DGAC concluded in 2010 that there was not sufficient evidence to acknowledge the effects of dietary patterns and cancer risk¹⁸. It was not until very recently in 2015 that DGAC officially recognized diet as a risk factor for colon and postmenopausal breast cancer risk⁶. However, there is no clear consensus on the effects of high choline consumption on cancer.

Choline and phospholipid synthesis

Choline is an important dietary nutrient that supports the synthesis of the most abundant glycerophospholipid PC via the cytidine diphosphate (CDP)-choline pathway. Up to 95% of the total choline pool in most tissues is used in the synthesis of PC^19,20. First described in 1956, Kennedy and Weiss elucidated a pathway for the de novo biosynthesis of PC and PE using rat liver as source of an enzyme activities²¹. Hence, there are two branches of the Kennedy pathway; CDP-choline pathway and the analogous pathway for PE called CDP-ethanolamine pathway²². The CDP-choline pathway consists of three enzymatic steps. Initially, choline kinase catalyzes the ATP-dependent phosphorylation of choline to form phosphocholine. In the second step, CDP-choline is formed from phosphocholine and CTP catalyzed by CTP:phosphocholine cytidylyltransferase. The final step of PC synthesis involves CDP-choline and a lipid anchor such as diacylglycerol (DAG) or alkyl‐acylglycerol (AAG)²². The Kennedy pathway enzymes are found ubiquitously in eukaryotes²³. PC synthesis is required for lipoprotein secretion from the liver²⁴. It is also a major source of the second messengers DAG, phosphatidic acid, lysophosphatidic acid, and arachidonic acid, which can be further metabolized to other signaling molecules²⁴.

Choline and one-carbon metabolism

One-carbon metabolism comprises a complex network of biochemical reactions that facilitates the transfer of 1C moieties in the form of methenyl, formyl, and methyl groups to support the synthesis of molecules that are required for cellular processes^25–27. The carbon unit cycle is essential for multiple physiological processes including cellular biosynthesis (purines and thymidine), amino acid homeostasis (glycine, serine, and methionine), methylation (PC biosynthesis, epigenetic maintenance) and regulation of redox status²⁸. Three pathways are involved; the folate cycle, methionine remethylation, and trans-sulfuration pathways. The most important metabolites through which cells refuel one-carbon metabolism are folic acid, serine, glycine and choline. Essentially, two major one-carbon donors are involved in the biosynthetic reactions: tetrahydrofolate (THF) derived from folate and S-adenosylmethionine (SAM) from methionine²⁹.

In mammalian cells, choline contributes to regeneration of methionine from homocysteine via its oxidation product, betaine, catalyzed by the enzyme betaine-homocysteine methyltransferase (BHMT)³⁰. After donating a methyl group to homocysteine, the byproduct of this reaction, dimethylglycine, can also yield additional 1C units through its utilization in the synthesis of THF³¹. Expression of BHMT is restricted to the liver and kidney. Methionine synthase is expressed globally, thus enabling homocysteine remethylation to occur throughout the body. Methionine is the substrate for S-adenosylmethionine (SAM) synthetase which produces SAM. The reactive methyl carrier, SAM, is the second most common enzymatic cofactor after ATP³². SAM plays a major role in epigenetics; biosynthetic processes including phosphatidylcholine, creatine, and polyamine synthesis; and sulfur metabolism. It is believed that phosphatidylcholine synthesis from phosphoethanolamine using SAM is likely the largest 1C sink in adult mammals³³. During the PC synthesis from PE, SAM donates three methyl groups via phosphatidylethanolamine methyltransferase (PEMT), an enzyme expressed principally in liver. The impact of the intake of the 1C donor choline is less studied and the associations between choline consumption and disease is not clear.

Role of PLD in choline metabolism.

Phospholipase D (PLD) and phospholipase C (PLC) are the major enzymes responsible for the release of choline and phosphocholine, respectively, from the most abundant phospholipid PC. Choline can also be released from the lyso-derivative of PC, lysoPC (LPC), by autotaxin, an enzyme that has lysophospholipase D activity. Although no mammalian PC-specific PLC isoforms have been sequenced or cloned, PLD is well characterized and is highly expressed in breast cancers as well as in colorectal, renal, gastric, ovarian, prostate and non-small cell lung cancers^34,35. Mammalian PLD constitutes an enzyme superfamily with six members, PLD1–6. Each PLD isoform has unique patterns of sub-cellular localization and roles in cellular processes. PLD1 and PLD2 are the most broadly studied PLD isoforms and are expressed in most tissues at varying basal levels. Human PLD1 and PLD2 share 50% amino acid identity and display a similar protein structure³⁶. The PLD1 (120kDa) isoform is mainly found in the inner membranes of mammalian cells including lysosomes, secretory endosomes, the Golgi complex, and endosomes³⁷. It is readily transported to the plasma membrane subsequent to extracellular stimulation³⁸. PLD2 (106 kDa), on the other hand, most commonly localizes to the plasma membrane³⁹. Whereas PLD1 exhibits low intrinsic activity, i.e. does not transduce extracellular signaling in its basal state, PLD2 has been linked to high basal catalytic activity. However, the available evidence shows that PLD2 displays the same enzymatic activity as PLD1, catalyzing the hydrolysis of PC to produce free choline and phosphatidic acid (PA). A single nucleotide polymorphism with increased risk of non-small cell lung cancer has been observed in PLD1, while polymorphisms in PLD2 are associated with the prevalence of colorectal cancer^40,41. Overexpression of PLD1 or PLD2 in murine fibroblasts or in breast cancer xenograft models alters cell growth and leads to primary tumor initiation and metastasis^42,43. The signaling lipid PA has pleiotropic effects⁴⁴. It can facilitate membrane vesicle trafficking, endocytosis/exocytosis and also function as a lipid anchor to recruit PA-binding proteins to localize to sites of signal transduction. PA is also known to activate proteins such as phosphatidylinositol 4-phosphate 5-Kinase (PI4P5K) and mTOR to regulate cellular processes such as cell hypertrophy, survival, and differentiation⁴⁵. Furthermore, PA can be dephosphorylated by Lipin to produce diacylglycerol (DAG) or hydrolyzed by phospholipase A (PLA) to produce LysoPA (LPA)⁴⁶. Because of its significance to disease, PLD has been studied both in vitro using cultured cell models and in animal models. Mice with genetic ablation of PLD1 are viable and fertile with no reported abnormalities⁴⁷. When used in studies on spontaneous intestinal tumorigenesis, PLD1-deficient mice showed significantly reduced intestinal tumorigenesis and increased survival⁴⁸. In contrast, deficiency of the related isoform PLD2 does not significantly reduce tumorigenesis. Similarly, in a syngeneic melanoma cancer model, loss of PLD1 by genetic knock out or use of the PLD1,2 small molecule inhibitor, 5-fluoro-2-indolyl des-chlorohalopemide (FIPI), decreased tumor size and weight, and decreased metastasis⁴⁹. These reports suggest that PLD1 could be a promising target for breast cancer therapy. Furthermore, an interdependency has been reported between the PLD1 isoform and choline kinase α enzymes in terms of protein expression⁵⁰. PLD1 silencing increases CHKA expression and vice versa. Interestingly, novel findings from recent microbiota studies show that several gut microorganisms also express and possibly secrete PLD enzymes that hydrolyze PC as the preferred substrate to produce free choline for downstream metabolism⁵¹. Although Chittim et al.⁵¹ demonstrated in their work that the released choline is utilized in the production of trimethylamine (TMA) by other gut microorganisms, these findings support the idea that gut microbiota PLD-released choline can be a substrate for choline kinase.

Dysregulated choline metabolism in Cancer Cells

Cancer is a dynamic disease with phenotypic and functional heterogeneity arising from genetic changes, environmental differences and reversible changes in cellular properties⁵². Common pathways are rare in cancer. However, dysregulated choline metabolism, a fairly new metabolic hallmark of cancer, is a common feature in nearly all cancers^53,54. Elevated levels of choline, phosphocholine and glycerophosphocholine have been consistently observed in cancer cells and tumor tissues⁵⁵. Comparative studies in cancer cells and rapidly proliferating non-cancerous epithelial cells have identified malignant transformation rather than just cell proliferation as the cause of abnormal choline metabolism in cancer^56,57. Enzymes involved in choline metabolism are proving to be an attractive and effective strategy for cancer treatment. For example, choline kinase α (CHKA), the first enzyme in the CDP-pathway, is expressed in a large diversity of human tumors including breast, lung, colorectal, bladder, prostate, ovary, endometrial, and T-cell lymphoma^58,59. Overexpression of CHKA has oncogenic potential and synergizes with other known oncogenes. As a result, several groups have made attempts to design small molecules as new inhibitors of its enzymatic activity and characterize their activity as anticancer drugs under in vitro and in vivo conditions^60,61. Among these, TCD-717, a second generation of CHKA inhibitors, has completed a first-in-human, Phase I clinical trial (http://clinicaltrials.gov/ct2/show/NCT01215864)⁶². An interdependency has been reported between the PLD1 isoform and choline kinase α enzymes in terms of protein expression that suggests compensatory effects between these two main choline-regulating enzymes⁵⁰. Numerous studies with cells, animal models and humans have established magnetic resonance spectroscopy (MRS) measurements for total choline-containing compounds (tCho) for detection of cancers and determining response to therapy⁵⁷. In addition, dietary manipulations that may impact on choline metabolism, especially methionine restriction, have been shown to attenuate tumor development, suggesting a viable approach to cancer management through nutritional intervention^63–65. Choline metabolites have also been implicated in oncogenesis and tumor progression through roles in phospholipid synthesis and 1C metabolism⁵³.

Pharmacological targeting of PLD and 1C metabolism.

Because of their roles in generating phosphatidic acid (PA), which regulates many cellular processes and is also a precursor to many bioactive lipid signaling molecules, PLD1 and PLD2 have emerged as drug targets for various diseases such as cancer, cardiovascular diseases, infectious diseases and neurodegenerative conditions (Parkinson’s and Alzheimer’s disease). Initially, primary alcohols (for example n-butanol) were the most commonly used molecules to inhibit PLD-catalyzed production of PA by promoting the formation of phosphatidyl alcohols which are presumed to be biologically inactive^66,67. This approach generally results in only incomplete attenuation of PLD dependent PA production. High throughput screening approaches enabled discovery of small molecules that could be used to target the two PLD isoenzymes. The first compounds described were halopemide and its derivatives notably 5-fluoro-2-indolyl des-chlorohalopemide (FIPI)^49,68. Halopemide and its derivatives were initially reported to be PLD2 inhibitors but were later found to also inhibit PLD1 with even greater potency^49,69. Using halopemide as a base, isoenzyme-selective PLD inhibitors with improved ancillary pharmacology and drug metabolism and pharmacokinetics profiles were later developed. A series of highly selective PLD1 inhibitors such as VU0359595 were prepared⁷⁰. VU0359595 is approximately 1,700 times more selective for PLD1 compared to PLD2. On the other hand, the first highly selective PLD2 inhibitor identified by the same group, VU0364739, has 75-fold selectivity for PLD2 versus PLD1⁷¹. In attempts to further develop even more selective PLD2 inhibitors, ML298 and ML395 were synthesized using triazaspirone as the core of the molecules⁷². The halopemide-derived and triazaspirone-based series were potent inhibitors of mammalian PLD enzymes but were inactive against both the bacterial Pseudomonas aeruginosa PLD (PldA) and the endocannabinoid regular NAPE-PLD⁷³. This prompted the use of selective oestrogen receptor modulators (SERMs) which had previously displayed off-target inhibition of PLD enzymes⁷⁴. One of the best characterized SERM, desketoraloxifene, has comparable inhibitory activity against mammalian PLD1 and PLD2, and bacterial PldA⁷³.

As mentioned elsewhere in this review and shown in Fig. 1, the free head group (choline) generated by PLD enzymes can be used for de novo synthesis of phosphatidylcholine or enter 1C metabolism via its oxidative product, betaine. Therefore inhibitors of PC biosynthetic pathway or 1C metabolism might be complementary or synergistic with PLD inhibitors. Two of the enzymes targeted with potential anti-cancer agents are choline kinase α (ChoKα) and methyl adenosyltransferase (MAT2A). ChoKα catalyzes the phosphorylation of choline which is the first step in the choline pathway for synthesis of PC. As with PLD inhibitors, the first inhibitors of ChoKα included alcohols such as propanol and butanol⁷⁵. Other earlier inhibitors of ChoKα to be synthesized were methyl-substituted derivatives of choline. Hemicholinium-3 (HC-3), structural homologs of choline, were later developed and demonstrated potent antitumor effects^57,76. This was followed by a series of bis-quinolinium compounds which are less toxic derivatives of HC-3. The first generation of these compounds was MN58b⁷⁷. With further improvement for tolerability, RSM-932A, a more potent bis-quinolinium with better therapeutic window and improved safety profile was developed⁶². RSM-932A (also called TCD-717) became the first ChoKα inhibitor to be tested in humans when in entered phase I clinical trials⁶². MAT2A catalyzes the biosynthesis of the universal methyl-donor S-adenosylmethionine (SAM) from methionine and ATP. SAM is required for many methyl transfer reactions and polyamine biosynthesis. Inhibitors of MAT2A have been developed that bind to either an allosteric site or to the catalytic subunit of MAT2A. A family of fluorinated N,N-dialkylaminostilbene (FIDAS) agents and their analogs have been developed that directly target the catalytic subunit of MAT2A⁷⁸. On the other hand Pfizer developed a MAT2A inhibitor, PF-9366, that binds and allosterically regulates the activity of MAT2A⁷⁹. The MAT2A inhibitor AG-270 developed by Agios Pharmaceuticals as a first-in-class MAT2A inhibitor that is being tested in patients with advanced solid tumors or lymphoma characterized by the homozygous deletion of the MTAP gene which plays a critical role in methionine salvage and consequently when absent renders cells highly dependent on MT2A dependent methionine synthesis⁸⁰. Collectively, continued development of the classes of inhibitors discussed above is expected to result in compounds with improved specificity, potency and tolerability. These compounds will continue to be valuable tools for analyzing the role of PLD, ChoKα and MAT2A enzymes in normal physiology and disease. These tools can be used to complement genetic knockout approaches in studies targeting distinct PLD isoenzymes and other pathways that utilize products of PLD activity. In addition, these small molecules can be utilized in cases where genetic ablation would be difficult to achieve or in settings where deletion of the aforementioned enzymes would have effects distinct from those of short-term inhibition. Although presently not used in the clinics, it is likely that small molecule inhibitors of PLD isoenzymes will be tested for efficacy in diseases, especially those where currently available therapeutics are limited. PLD inhibitors could be used in combination with novel ChoKα and MAT2A inhibitors, or with standard chemotherapy to target pathways of 1C metabolism that are known to be critical for development and progression of some cancers.

Concluding comments.

Although substantial advances have been made in understanding the molecular mechanisms that drive the aberrant choline metabolism in cancer, the role of diet in this process is unclear. A number of epidemiologic studies examining the relationship between dietary choline intake and cancer have yielded conflicting findings. Therefore, investigating the idea that high consumption of dietary lipids, rich in esterified choline lipids, promotes the abnormal choline metabolism leading to cancer progression and metastasis would be interesting. This can be done using mouse models of cancer and rodent diets formulated with exogenous phosphatidylcholine as the major source of dietary choline. PLD-deficient animal models have been generated and characterized. Development of small molecule inhibitors targeting PLD is on-going. With the interest and rapid increase in gut microbiome research over the last several years, the discovery of widespread PLD activity among gut bacteria provides additional opportunities for therapeutic intervention. Together, these tools present an excellent opportunity to unravel the missing link between PLD and diet in contributing to choline and one-carbon metabolism. It would be possible, for example, to test the hypothesis that PLD expression and higher consumption of dietary PC supports the abnormal choline metabolism in cancer.