Novel Sphingolipid-Based Cancer Therapeutics in the Personalized Medicine Era

Abstract

Sphingolipids are bioactive lipids that participate in a wide variety of biological mechanisms, including cell death and proliferation. The myriad of pro-death and pro-survival cellular pathways involving sphingolipids provide a plethora of opportunities for dysregulation in cancers. In recent years, modulation of these sphingolipid metabolic pathways has been in the forefront of drug discovery for cancer therapeutics. About two decades ago, researchers first showed that standard of care treatments, e.g., chemotherapeutics and radiation, modulate sphingolipid metabolism to increase endogenous ceramides, which kill cancer cells. Strikingly, resistance to these treatments has also been linked to altered sphingolipid metabolism, favoring lipid species that ultimately lead to cell survival. To this end, many inhibitors of sphingolipid metabolism have been developed to further define not only our understanding of these pathways but also to potentially serve as therapeutic interventions. Therefore, understanding how to better use these new drugs that target sphingolipid metabolism, either alone or in combination with current cancer treatments, holds great potential for cancer control. While sphingolipids in cancer have been reviewed previously (Hannun & Obeid, 2018; Lee & Kolesnick, 2017; Morad & Cabot, 2013; Newton, Lima, Maceyka, & Spiegel, 2015; Ogretmen, 2018; Ryland, Fox, Liu, Loughran, & Kester, 2011) in this chapter, we present a comprehensive review on how standard of care therapeutics affects sphingolipid metabolism, the current landscape of sphingolipid inhibitors, and the clinical utility of sphingolipid-based cancer therapeutics.

1. EFFECTS OF CHEMO- AND RADIATION THERAPIESON SPHINGOLIPIDS

1.1. Chemotherapy

Chemotherapeutic or cytotoxic agents are drugs commonly used as standard of care for multiple types of cancers due to their ability to kill or damage highly proliferative cells. These agents are usually divided into classes based on their mechanisms of action, and, strikingly, a wide range of these classes have been described as leading to accumulation of sphingolipids (particularly ceramide) in cells. As shown in Fig. 1, we depict the steady state accumulation of endogenous ceramide via chemo- and radiation therapies as a combination of “faucets” and “drains.” Enzymes underlying sphingomyelin hydrolysis (sphingomyelinased—SMase), de novo sphingolipid synthesis (serine palmitoyltransferased—SPT, dihydroceramide synthased—CerS, dihydroceramide desaturased—DES), and the sphingolipid salvage pathway (glucosylceramidased—GCase, CerS) are the “faucets” that generate ceramide, whereas enzymes that metabolize ceramide (ceramidased—CDase, glucosylceramide synthased—GCS, sphingomyelin synthased—SMS, ceramide kinased—CERK) are the “drains.” In essence, when too much ceramide accumulates and the metaphorical sink overflows the cell dies. Complicating this system, some of these sphingolipids have separate “faucets” or “drains” due to different isoenzymes of sphingolipid enzymes that generate distinct ceramide species with unique biological and biophysical properties. This diversity in ceramide species is due to distinct ceramide synthases with specific fatty acyl-CoA specificities. As a generalization for the major ceramide species observed in cancer, CerS1 preferentially generates C₁₈-ceramide, CerS2 and CerS4 generate long chain C₂₄ species, and CerS5 and CerS6 generate C₁₆ species. The “fluid dynamics” of these enzyme activities determines both the efficacy of, and resistance to, chemo- and radiation therapies. In fact, this more interconnected and compensatory metabolic system is just a new version of the “sphingolipid rheostat,” by which the pro-death actions of ceramide are counterbalanced by the pro-survival functions of sphingosine1-phosphate (S1P) (Cuvillier et al., 1996). A detailed metabolic pathway for sphingolipid regulation in cancer is depicted in Fig. 2 (top panel).

Figure 1.

The accumulation of ceramide can be generated from three main pathways (sphingomyelin hydrolysis, de novo synthesis, or salvage pathway) or exogenously added (in the form of short chain ceramides), whereas the degradation of ceramide is facilitated by four main enzymes (SMS, GCS, CDase, and CerK). At a steady state, the level of ceramide is held stable, as flux through the system remains fairly constant. However, similarly to increasing flow through the faucet or clogging the drain, increasing the generation of ceramide or preventing its degradation can lead to an accumulation of ceramide, which can be selectively toxic to cancer cells.

Figure 2.

This figure depicts a ceramide-centric view of sphingolipid metabolism. The upper panel outlines the major sphingolipid species (blue boxes) and the enzymes that metabolize these lipids (multicolored boxes). Inhibitors and activators of specific enzymes in the pathway are listed in the lower panel using enzyme color-coded “sticky notes.”

Studies performed in yeast showed that under stress conditions, cells responded by generating ceramide, demonstrating that sphingolipid metabolism is an evolutionary response to stress (Cowart & Obeid, 2007). Chemotherapeutics induce stress in cancer cells, and numerous investigations have shown that in response to this stress, ceramide levels are increased, both by sphingomyelin hydrolysis as well as through de novo synthesis of ceramide (Beckham, Cheng, Marrison, Norris, & Liu, 2013). This increase in ceramide levels has been validated in many different classes of chemotherapies including daunorubicin (anthracycline), etoposide (topoisomerase II inhibitor), and gemcitabine (nucleotide analog DNA replication inhibitor). These drugs have all been described as inducers of de novo generation of ceramide, either by activating ceramide synthases or serine palmitoyl transferase (Bose et al., 1995; Chalfant et al., 2002; Jaffrezou et al., 1996; Perry et al., 2000). Inhibition of enzymes important for de novo generation of ceramide decreases the cytotoxicity of these chemotherapeutics, and therefore, decreases their overall efficacy. However, in glioma cells treated with etoposide, the increase in ceramide levels was caused by increased activation of neutral SMase, which is responsible for converting sphingomyelin to ceramide (Sawada et al., 2001). In one clinical trial for head and neck squamous cell carcinoma, administration of a combination of gemcitabine plus doxorubicin increased serum levels of C₁₈-ceramide that correlated with better drug responses (Saddoughi et al., 2011).

Interestingly, altered ceramide levels are not the only biological connection between sphingolipids and chemotherapy; the Cabot laboratory demonstrated that levels of glucosylceramide are increased in breast cancer cell lines (Lavie et al., 1997) and in patients who were resistant to chemotherapy (Lucci et al., 1998). The enzyme that generates glucosylceramide, GCS, is upregulated in several different tumor types such as lung cancer (Zhang et al., 2014), breast cancer (Liu, Patwardhan, Xie, et al., 2011), bladder cancer (Sun et al., 2010), and colorectal cancer (Wang, Liu, Xu, Mu, & Sun, 2014). Given that increased conversion of ceramide to glucosylceramide decreases killing of cancer cells exposed to chemotherapeutic agents and the increased expression/activity of glucosylceramide synthase in several tumor models, focus has shifted to developing inhibitors for this enzyme that could synergize with standard of care drugs.

Cancer cells may also decrease ceramide levels by increasing activity of the CDase enzymes that cleave ceramide to sphingosine, which can then be converted to S1P. Due to this conversion from ceramide, a pro-apoptotic lipid, to S1P, a pro-survival lipid, CDase can modulate cell fate under stress conditions. CDase enzymes are divided in three different types according to their optimum pH: neutral, acid, or alkaline. In cancer research, a vast majority of the studies have been devoted to acid ceramidase (ACDase). This enzyme was reported as overexpressed in a myriad of tumor types including prostate (Saad et al., 2007), head and neck squamous cell (Elojeimy et al., 2007), and acute myeloid leukemia (AML) (Tan, Pearson, Feith, & Loughran, 2017). In vitro and in vivo studies have shown that by overexpressing ACDase, tumors are more aggressive and resistant to chemotherapies (Beckham et al., 2012; Saad et al., 2007). In prostate cancer cells, ACDase overexpression promoted resistance to etoposide, cisplatin (platinum-based agent), and gemcitabine by stopping accumulation of ceramide (Beckham, Cheng, et al., 2013). Interestingly, ACDase has been shown in hepatocellular cells to be upregulated after exposure to daunorubicin, a phenomenon that was not observed in normal hepatocytes, suggesting a specific mechanism for cancer cells to resist cytotoxicity (Morales et al., 2007).

Recently, the use of medical marijuana was legalized to mitigate side effects of chemotherapies in 30 states (http://www.governing.com/gov-data/state-marijuana-laws-map-medical-recreational.html). In 1998, the Sanchez group first reported the relationship between sphingolipids and cannabinoids, the active ingredients of medical marijuana (Sanchez, Galve-Roperh, Canova, Brachet, & Guzman, 1998; Sanchez, Galve-Roperh, Rueda, & Guzman, 1998). They demonstrated that THC (tetrahydrocannabinol) caused an increase in ceramide levels through de novo synthesis, which led to apoptosis in C6 glioma cells. This increase of ceramide levels through THC or other cannabinoids was also shown in prostate and colon cancer cells, as well as mantle cell lymphoma (MCL) and multiple myeloma (Barbado et al., 2017; Cianchi et al., 2008; Gustafsson, Christensson, Sander, & Flygare, 2006; Mimeault, Pommery, Wattez, Bailly, & Henichart, 2003). Further work in MCL showed that cannabinoids induced CerS3 and CerS6 expression. Notably, blocking ceramide metabolism into S1P or glucosylceramide also increased the cytotoxic effect of cannabinoids (Gustafsson, Sander, Bielawski, Hannun, & Flygare, 2009).

Increases in S1P levels at the expense of sphingosine and ceramide in cancer are often manifested by an increase in sphingosine kinase (SphK) expression or activity. The conversion of sphingosine to S1P is executed by two SphK isoenzymes that phosphorylate sphingosine. SphK1 is upregulated in paclitaxel-resistant prostate cancer cell lines (Aoyama et al., 2017) as well as in several tumor xenograft models such as stomach, breast, lung, brain, colon, and kidney (Pyne & Pyne, 2010). Moreover, SphK1 expression was shown to be upregulated in AML patient samples (Dick et al., 2015). Although the specific role of SphK2 in cancer is not as well described, specific SphK2 inhibitors are showing efficacy in the clinic (Orr Gandy & Obeid, 2013). In vitro and in vivo studies in prostate cancer and none small cell lung carcinoma showed that overexpression of SphK1 was responsible for resistance to docetaxel (taxane), camptothecin (topoisomerase inhibitor), and doxorubicin (Akao et al., 2006; Pchejetski et al., 2008; Sauer et al., 2009; Song et al., 2011). However, other studies have also shown that in leukemia cells, treatment with actinomycin D (peptide antibiotic), doxorubicin, or etoposide resulted in reduction of SphK1 due to accumulation of p53 (Taha et al., 2004). SphK1 was also found to be responsible for resistance to cetuximab (a monoclonal antibody chemotherapeutic against Epidermal Growth Factor Receptor (EGFR)) in colorectal cancer both in vitro as well as in patients (Rosa et al., 2013). An interesting mechanism for cell death is to alter the sphingolipid rheostat, not by increasing ceramide levels, but by decreasing S1P levels (French et al., 2003). For example, Oskouian et al. (2006) showed that treating several cancer cell lines with etoposide led to increased levels of S1P lyase and concomitant increases in apoptosis. Taken together, inhibitors of SphK or activators of sphingosine lyase or phosphatases may sensitize patients to conventional therapeutics and prevent drug resistance mechanisms. Although manipulating sphingolipid metabolism as new therapeutic avenue might be enticing, understanding each patient’s sphingolipid and enzymatic profile is vital for interfering with this very intricate, complex, and highly regulated process.

1.2. Radiation

Radiotherapy, as currently used in the clinic, was first developed and tested as a treatment for cancer in 1934 by Henri Coutard. As described by Halperin et al., Coutard developed a fractionized protocol and observed that by administrating lower doses of X-rays over a course of days, he could kill cancer cells while allowing normal cells to recover from the effects of radiation (Halperin, Perez, & Brady, 2008). Contrary to chemotherapeutics, radiotherapy is frequently limited to tumor locations and not systemic. By using high-energy liberation of electrons, it is possible to induce cell death specifically in malignant cells (Aureli et al., 2014). This process occurs by induction of DNA damage, as radiation causes double-strand breaks in DNAs that are lethal to cells (Beckham, Cheng, et al., 2013; Carroll, Donaldson, & Obeid, 2015). Cancer cells can attempt to repair these double-strand breaks, but as repair mechanisms are usually dysregulated in cancer, it leads to errors and subsequent mitosis-associated cell death or p53-induced apoptosis (Aureli et al., 2014; Beckham, Cheng, et al., 2013). Strikingly, the work performed by the Kolesnick lab has shown that even p53-independent cell death mechanisms after radiation exposure are regulated by sphingolipid metabolism (Haimovitz-Friedman et al., 1994; Kolesnick & Fuks, 2003). Among sphingolipid species, ceramide was shown to be the major mediator of cellular stress after radiation exposure (Modrak, Gold, & Goldenberg, 2006).

One of the earliest evidences for the role of ceramide in cell death upon radiation led to the discovery of the rapid hydrolysis of sphingomyelin to ceramide by SMase after exposure to ionizing radiation (Haimovitz-Friedman et al., 1994). The identification of different isoforms of SMase that work under different optimal pH conditions, led to the discovery that sphingomyelin metabolism is triggered by different SMase enzymes in a cell-type specific manner (Beckham, Cheng, et al., 2013; Li & Zhang, 2015). In a very elegant set of experiments, the Kolesnick lab showed that by ablating acid SMase in human type I Niemanne Pick disease lymphoblastoid cells, or in p53 knockout mice, ceramide induced apoptosis after radiation was ablated (Lozano et al., 2001; Santana et al., 1996). Ectopic expression of acid SMase as well as by exogenous treatment with ceramide restored sensitivity to irradiation. These results indicated a critical role of ceramide in cell death induced by radiation. Moreover, SMases can be present extracellularly, and in a small group of cancer patients, fractionated high dose radiation was correlated with an increase of SMase extracellular content and ceramide levels (Oskouian & Saba, 2010).

Sphingomyelin hydrolysis is not the only metabolic pathway to generate ceramide in response to radiation. An in vitro study in radiation-resistant DU145 prostate cancer cells showed that pretreatment with resveratrol (an antioxidant agent) resulted in sensitization to radiation. This shift had occurred by induction of de novo synthesis of ceramide, a finding that was validated when sphingolipid synthesis inhibitors blocked sensitization and reverted DU145 cells to radiation-resistant status (Scarlatti et al., 2007).

Different ceramide chain length species are now known to have different roles in cells. Therefore it is not surprising that different responses to radiation treatments have been correlated with changes in various ceramide species or ceramide synthases. CerS5 and CerS6 are activated following radiation and induce cell death. However, C₂₄ ceramide generated by CerS2, conferred resistance to radiation in the same human cervical cancer (HeLa) cells (Galadari, Rahman, Pallichankandy, & Thayyullathil, 2015). This paradigm of differential activities and cellular roles between C₁₆/C₁₈ versus C₂₄ species is actively being investigated in various other cancer therapeutic models. In addition, the role of saturated versus unsaturated ceramide species of similar chain length is also an area deserving further investigation.

In addition to sphingomyelin hydrolysis and de novo synthesis, ceramide synthase activity has long been known to increase in cells after exposure to radiation (Vit & Rosselli, 2003). The current rationale in the field is that ceramide levels are initially increased by activation of sphingomyelin hydrolysis as described above, and then, between 8 and 24 h after irradiation, the increase in ceramide levels shifts to become driven by increased CerS activity (Carroll et al., 2015). Strikingly, the second wave of ceramide increase seems to be dependent on the first one. In the aforementioned type I Niemanne—Pick disease cells that lack acid SMase, after radiation treatment, there is an increase in CerS activity but not an increase in ceramide levels (Carroll et al., 2015). This dual wave system of cell death was also validated in Jurkat leukemia cells (Ardail et al., 2009).

Another important sphingolipid pathway is the generation of sphingosine by ceramide cleavage via CDases. Hara et al. (2004) reported that in vitro irradiated glioblastoma cells showed increased expression of ACDase. The authors further showed that by inhibiting this enzyme, glioblastoma cells accumulated ceramide on irradiation and increased levels of apoptosis (Hara et al., 2004). A different laboratory working with prostate cancer cells showed that this upregulation of ACDase after radiation exposure was mediated by c-Jun (Cheng et al., 2013). Similarly to glioblastoma cells, reducing ACDase expression in prostate cancer cells resulted in ceramide accumulation and cell death after radiation treatments (Cheng et al., 2013). Considering the evidence above, conversion of ceramide to sphingosine can be a mechanism to counteract the proapoptotic role of ceramide and lead to cell survival and insensitivity to radiation.

2. CURRENT LANDSCAPE OF INHIBITORS OFSPHINGOLIPID METABOLISM

Sphingolipids are vital to normal cell function and thus the dysregulation of the enzymes responsible for their production and maintenance can lead to a number of diseases, including cancer. As such, the ability to regulate the activity of these enzymes using inhibitors or activators is important for developing novel treatments for diseases in which sphingolipid imbalance occurs. Thus, the search for effective inhibitors for the enzymes that decrease ceramide levels are worth exploring. The following section will discuss in vitro and in vivo anti-cancer effects of some of the best-characterized inhibitors and activators of the main sphingolipid enzymes. The lower panel of Fig. 2 provides a comprehensive list of these drugs, their targets, and their impact in sphingolipid metabolism.

2.1. Targeting Sphingomyelin Synthesis

SMSs generate sphingomyelin from two substrates, ceramide and phosphatidylcholine. Tricyclodecan-9-yl-xanthogenate (D609) is a cytotoxic drug that causes selective tumor cell death by inhibiting SMS leading to increased levels of the pro-apoptotic lipid ceramide (Bai et al., 2004; Meng et al., 2004). However, due to the presence of a xanthate group on D609, cells may metabolize the drug and render it useless. Thus, a series of D609 prodrugs have been developed, which are more stable and effective at inhibiting SMS at lower LD₅₀ values (56.6 mM prodrug vs. 117 mM D609) (Bai et al., 2004). Another inhibitor of SMS, MS-209, has been used in numerous studies for its ability to reverse multidrug resistance in tumors (Delgado, Casas, Llebaria, Abad, & Fabrias, 2006; Robert, 2004). Finally, the SMS inhibitor, Dy105, was initially developed as an atherosclerosis treatment but may be useful for future cancer studies (Lou et al., 2014).

2.2. Targeting Ceramide Formation

The first step in the de novo synthesis of ceramide is the condensation of serine and palmitoyl CoA by the enzyme SPT (Hanada, 2003). Myriocin, also known as ISP-1 or thermozymocidin, is the most studied inhibitor of SPT and can inhibit the enzyme at picomolar concentrations (Miyake, Kozutsumi, Nakamura, Fujita, & Kawasaki, 1995; Sauane et al., 2010). More natural products such as sphingofungins, lipoxamycin, and, to a lesser extent, viridiofungins have been shown to inhibit SPT as well (Mandala et al., 1994; Mandala, Thornton, Frommer, Dreikorn, & Kurtz, 1997; Zweerink, Edison, Wells, Pinto, & Lester, 1992). Other previously used less-specific compounds, L-cycloserine and β-chloro-L-alanine, have received some attention as SPT inhibitors but both are known to have off-target effects. Specifically, L-cycloserine also inhibits 3-ketodihydrosphingosine reductase, which is downstream of SPT in the de novo synthesis pathway, whereas β-chloro-L-alanine has off-target effects inhibiting other pyridoxal-5’phosphate—dependent enzymes (Medlock & Merrill, 1988; Sundaram & Lev, 1984).

The CerS family of six isoenzymes is involved in both the de novo synthesis and the recycling of sphingosine into ceramide in the salvage pathway. One of the most well-known general inhibitors of CerS is Fumonisin B1, a mycotoxin produced by Fusarium moniliforme (Bose et al., 1995; Wang, Norred, Bacon, Riley, & Merrill, 1991; Yoo, Norred, Wang, Merrill, & Riley, 1992). A structurally similar toxin, AAL toxin produced by Alternaria alternata, has also been shown to inhibit CerS (Delgado et al., 2006; Winter, Gilchrist, Dickman, & Jones, 1996). Australifungin, a mycotoxin from Sporormiella australis, potently inhibits CerS but has limited use because of its high chemical reactivity (Delgado et al., 2006; Mandala et al., 1995). Interestingly, despite showing efficacy in preclinical development, N-(4-hydroxyphenyl)retinamide (4-HPR or fenretinide) activates CerS but also inhibits DES, the next enzyme in the de novo pathway (Schulz et al., 2006). Fenretinide is now being tested in clinical trials for its potential as a broad anti-tumor inhibitor in several tumor models.

The sphingolipid-based compound, GT11, is a competitive inhibitor of DES that has been modified to generate more effective versions of the compound (Bedia, Triola, Casas, Llebaria, & Fabrias, 2005; Triola, Fabrias, Casas, & Llebaria, 2003; Triola, Fabrias, & Llebaria, 2001). The utility of these compounds that effect multiple sphingolipid metabolites in cancer are undefined.

Desipramine, imipramine, and amitriptyline, which are all tricyclic antidepressants, are also functional aSMase inhibitors that increase aSMase degradation by lysosomal proteases (Albouz et al., 1981; Arimochi & Morita, 2008; Bhabak & Arenz, 2013; Delgado et al., 2006; Huang et al., 2007; Lu et al., 2009). L-alpha-phosphatidyl-D-myo-inositol-3,5-bisphosphate (PtdIns3,5P2) and the more potent phosphatidyl-myo-inositol 3,4,5triphosphate (PtdIns3,4,5P3) are both physiological inhibitors of aSMase that are found in humans (Kolzer et al., 2003; Testai, Landek, Goswami, Ahmed, & Dawson, 2004). A calcium channel blocker that sensitized multidrug resistant cells to doxorubicin, SR33557, also functions to inhibit aSMase (Jaffrezou et al., 1991). NB6, a structurally similar molecule to the tricyclic antidepressants, decreases aSMase activity in a different manner, by inhibiting its transcription (Deigner et al., 2001). α-Mangostin, cowanin, and cowanol, compounds isolated from the bark of Garcinia speciosa, each inhibit aSMase with varying degrees of potency (14.1, 19.2, and 10.9 μM IC₅₀ values, respectively) (Delgado et al., 2006; Okudaira et al., 2000). SMA-7 is another aSMase inhibitor that was shown to inhibit Dextran Sulfate Sodium (DSS)-induced colitis in mice, though other studies have suggested that it may also inhibit nSMase (Canals, Perry, Jenkins, & Hannun, 2011; Sakata et al., 2007; Yokomatsu et al., 2001). A sphingomyelin analogue that was developed to inhibit aSMase, AD2765, increased cellular ceramide levels and cell death in cancer cells in vitro (Darroch et al., 2005). Treatment with fluphenazine, an anti-psychotic drug that induces lysosomal stress and inhibits aSMase activity, led to accumulation of sphingomyelin and subsequent cell death in hypoxic tumors (Klutzny et al., 2017). In contrast to the inhibitors listed above, bismonoacylgycerophopsohate was found to be an activator of aSMase that increased the generation of ceramide (Canals et al., 2011; Linke et al., 2001). Other functional aSMase inhibitors (FIASMAs) that cause aSMase to dissociate from the inner lysosomal membrane and become inactivated include Dextromethorphan, Fluoxetine, Maprotilin, Nortriptyline, Orphenadrine, Sertralin, and Triflupromazine (Kornhuber et al., 2008, 2010).

Numerous studies have explored the effects of inhibition of the SMase family member, neutral SMase (nSMase), utilizing GW4869, a noncompetitive nSMase inhibitor (Luberto et al., 2002; Marchesini, Luberto, & Hannun, 2003; Takahashi, Inanami, Asanuma, & Kuwabara, 2006). The antioxidant glutathione, which is present in mammalian cells at relatively high concentrations (1—20 mM), was reported to almost completely inhibit nSMase at a concentration of 5 mM, suggesting that nSMase is relatively inactive in vivo under normal conditions (Liu & Hannun, 1997). Another nSMase inhibitor, C11AG, which was originally used to inhibit herpes simplex virus-1 replication, was later shown to enhance apoptosis of Jurkat T cell lymphoma cells (Amtmann & Zoller, 2005; Amtmann, Zoller, & Schilling, 2000; Delgado et al., 2006). Macquarimicin A and scyphostatin, two natural products that are primarily nSMase inhibitors, are also weak ACDase inhibitors (Nara, Tanaka, Hosoya, Suzuki-Konagai, & Ogita, 1999; Nara, Tanaka, Masuda-Inoue, et al., 1999; Tanaka et al., 1999). Alutenusin and chlorogentisylquinone are both nSMase inhibitors that were originally purified from culture media of different fungal species (Uchida, Tomoda, Arai, & Omura, 2001; Uchida, Tomoda, Dong, & Omura, 1999). Manumycin A is an antibiotic produced by the bacterium Streptomyces parvulus, which selectively inhibits nSMase, not aSMase, and was also found to be a useful inhibitor of Ras farnesyltransferase (Arenz et al., 2001; Kouchi et al., 1999). Cambinol is another molecule with dual uses, as it was originally discovered as a regulator of SIRT1/2 and was later found to be dramatically more effective for its ability to inhibit nSMase (Figuera-Losada et al., 2015).

2.3. Targeting Glucosylceramide Generation and Degradation

Glucosylceramidase (GCase) hydrolyzes glucosylceramide back to ceramide and glucose. In very early studies, α-gluconolactone was found to inhibit glucosylceramidase, but more potent inhibitors have since been developed (Overkleeft et al., 1998; van Weely, Brandsma, Strijland, Tager, & Aerts, 1993). Deoxynojirimycin is an a-glucosidase inhibitor isolated from mulberry leaves that has been extensively used as a GCase inhibitor, although interestingly, its derivatives, N-nonyl-1-deoxynojirimycin (NNDNJ) and N-Butyldeoxynojirimycin (NBDNJ), have been shown to function as chaperones of GCase that increase its activity (Alfonso et al., 2005; Delgado et al., 2006; Sawkar et al., 2002). Other natural product inhibitors, such as castanospermine, isofagomine, valienamine, and validamine, have also been used to inhibit GCase activity (Delgado et al., 2006; Sasak, Ordovas, Elbein, & Berninger, 1985; Saul, Chambers, Molyneux, & Elbein, 1983).

GCS is the enzyme responsible for adding a glucose molecule to ceramide. A pair of the most widely used GCS inhibitors, D-threo-1-phenyl-2palmitoylamino-3-morpholino-l-propanol (PPMP) and D-threo-1-phenyl2-decanoylamino-3-morpholino-propanol (PDMP), are ceramide analogues (Kovacs, Pinter, & Csaba, 2000; Vunnam & Radin, 1980). PPMP has previously been shown to resensitize multidrug-resistant cancer cells to chemotherapies in vitro and in vivo, whereas PDMP has shown cytotoxicity in AML in vitro (Cakir, Saydam, Sahin, & Baran, 2011; Gouaze et al., 2005). A derivative of PDMP, ethylenedioxy-P4, showed specific inhibition of human GCS but did not inhibit other animal, plant, fungal, or bacterial versions of the enzyme (Hillig, Warnecke, & Heinz, 2005). Miglustat, or NBDNJ, has been shown to reduce the effects of Gaucher disease by inhibiting GCS to prevent glucosylceramide buildup (Cox, 2005; Platt, Neises, Dwek, & Butters, 1994). A structurally similar molecule, N-(n-Butyl)deoxygalactonojirimycin (NBDGJ), also functions to inhibit GCS (Platt, Neises, Karlsson, Dwek, & Butters, 1994).

2.4. Targeting Ceramide-1-Phosphate Formation

CERK, a highly homologous relative of SphK, catalyzes the ATPdependent specific phosphorylation of ceramide to ceramide-1-phosphate. Two main CERK inhibitors have been developed that block its activity: K1 and NVP-231 (Graf et al., 2008; Kim, Inagaki, et al., 2005; Pastukhov et al., 2014). K1 is an isomer of F-12509A, a SphK inhibitor discussed below, that non-competitively inhibits CERK (Kim, Inagaki, et al., 2005). NVP-231 is a competitive CERK inhibitor (Graf et al., 2008) and inhibits breast and lung-cancer cell growth (Pastukhov et al., 2014).

2.5. Targeting Sphingosine Formation

Similar to the three types of SMase enzymes, CDaseare classified according to their optimal pH and thus are named acid ceramidase (ACDase), neutral ceramidase (nCDase), and alkaline ceramidase (alkCDase). B13 (also referred to as D-NMAPPD) is a ceramide analogue that inhibits ACDase and has been used extensively in vitro and in in vivo models of melanoma, pancreatic cancer, and metastatic colon cancer (Raisova et al., 2002; Samsel et al., 2004; Selzner et al., 2001). This has led to the generation of a number of analogs (LCL204, LCL385, LCL521) that have been designed to target ACDase in the lysosomal compartment that B13 struggles to reach (Bai et al., 2017; Holman et al., 2008; Mahdy et al., 2009). LCL 204 has been shown to inhibit in vitro and in vivo models of AML (Tan et al., 2016). N-oleoylethanolamine (NOE) is a synthetic analogue of ceramide that inhibits ACDase that has been used to sensitize cancer cells to chemotherapies and reduce pro-survival signals (Monick et al., 2004; Morales et al., 2007; Strelow et al., 2000; Sugita, Willians, Dulaney, & Moser, 1975). Like NOE, AD2646, SABRAC, and RBMI-12 are other ceramide analogues that were designed to inhibit ACDase (Camacho et al., 2013; Dagan, Wang, Fibach, & Gatt, 2003). In contrast, _D-e-MAPP (or D-MAPP) is a ceramide analogue that inhibits alkCDase (Bielawska et al., 1996). DM102 meanwhile is a specific ACDase inhibitor as it shows no activity toward other CDases and causes cell death in vitro (Gouaze-Andersson et al., 2011). Carmofur, a chemotherapeutic drug used to treat colorectal cancer, was discovered to be an ACDase inhibitor suggesting that its mechanism of action is promoting ceramide buildup in cancer cells (Realini et al., 2013). Cystatin SA, a member of the cystatin family, was identified as one of the first physiological inhibitors of ACDase (Eliyahu et al., 2011). Finally, a very potent ACDase inhibitor, SABRAC (IC₅₀ 52 nM) induces ceramide accumulation and reduces cell growth of PC-3 prostate cancer cells in vitro (Camacho et al., 2013).

C₁₆-urea-ceramide, as well as shorter-chain C₆-urea-ceramide, are potent inhibitors of nCDase (Garcia-Barros et al., 2016; Saied & Arenz, 2014; Usta et al., 2001). C₁₈-Ceramine (ceramine), a synthetic ceramide derivative, is also able to inhibit nCDase (Garcia-Barros et al., 2016; Saied & Arenz, 2014). Two additional inhibitors, Ceranib-1 and Ceranib-2, have been identified that broadly inhibit ACDase, nCDase, and alkCDase (Draper et al., 2011).

2.6. Targeting Sphingosine-1-Phosphate Formation and Degradation

Dimethylsphingosine (DMS), a methylated version of sphingosine, is the first SphK inhibitor that was shown to promote apoptosis in cancer cells (Edsall, Van Brocklyn, Cuvillier, Kleuser, & Spiegel, 1998) and to inhibit growth of xenografted breast cancer tumors in mice (Nava et al., 2000).

Based on the lead compound dihydroxyaurone, which was reported to be a SphK inhibitor that has anti-tumor properties in a mouse model of breast cancer, a group of SphK inhibitors (termed SKI I—V) were developed (French et al., 2003). Of these, SKI-II (also called SKi) has found the most use although it inhibits both SphK1 and SphK2 and possibly other targets, complicating understanding of the mechanism by which it promotes cell death in a variety of cancer types (French et al., 2003; Truman, GarciaBarros, Obeid, & Hannun, 2014). _L-threo-dihydrosphingosine, also known as Safingol, was initially reported to be a protein kinase C inhibitor but was later proven to also inhibit SphKs (Buehrer & Bell, 1992; Tsukamoto et al., 2015). Additionally, a number of SphK inhibitors from various bacterial and fungal sources have been discovered. S-15183a and S-15183b were produced by the fungus Zopfiella inermis, B-5354a, B-5354b, and B-5454c were all derived from a marine bacteria from the genus Ruegeria, and F-12509A was isolated from the fungus Trichopezizella barbata (Kono, Sugiura, & Kohama, 2002; Kono, Tanaka, Mizuno, et al., 2000; Kono, Tanaka, Ogita, Hosoya, & Kohama, 2000; Kono et al., 2001).

SK1-I (BML-258), a sphingosine analogue that is an isozyme-specific SphK1 inhibitor, promoted cell death in a number of in vitro cancer models (Paugh et al., 2008). Two other SphK1 inhibitors, SKI-178 and PF-543, also occupy the substrate-binding site of SphK and thus prevent the phosphorylation of sphingosine (Hengst et al., 2017; Schnute et al., 2012; Wang, Knapp, Pyne, Pyne, & Elkins, 2014).

SG-12 and SG-14 were the first two SphK2 inhibitors discovered, but ABC294640 has been used more frequently. ABC294640 displays activity in vitro and in vivo to prevent cancer growth and is currently in phase II clinical trials (Antoon et al., 2010; Beljanski, Knaak, & Smith, 2010; Kim, Kim, et al., 2005). K145 was recently developed as a more specific and potent SphK2 inhibitor and has shown promise in vivo in murine xenografts for suppressing breast tumor growth (K. Liu et al., 2013).

S1P is not only a bioactive mediator; it is also the ultimate sphingolipid metabolite and the only exit point for metabolism of all sphingolipids. S1P lyase is responsible for the irreversible cleavage of S1P into hexadecenal and ethanolamine phosphate, which are substrates for metabolism in various metabolic pathways. 1-deoxysphinganine-1-phosphate, an S1P analogue, is a naturally occurring phosphorylated sphingoid base that inhibits S1P lyase (Stoffel & Grol, 1974). S1P lyase activity can also be inhibited by 2-vinylsphinganine-1-phosphate (Boumendjel & Miller, 1994). S1P lyase is a pyridoxal phosphate (B6)-dependent enzyme but the often used S1P lyase inhibitor 4-deoxypyridoxine, which interferes with many B6dependent enzymes is generally not used anymore because this lack of specificity results in high levels of toxicity (Serra & Saba, 2010).

In addition, therapeutics can regulate S1P activity by altering S1P receptor expression or activity. As a prime example, FTY720, also known as fingolimod/gilenya, is a myriocin derivative and sphingosine analogue that is phosphorylated in vivo to its active form FTY720-P mainly by SphK2; inhibiting S1P activity through S1P receptor 1, 3, 4, and 5 binding (Brinkmann et al., 2002). In vivo studies have shown that FTY720-P is also able to inhibit S1P lyase activity directly, as S1P lyase mRNA and protein expression were not altered (Bandhuvula, Tam, Oskouian, & Saba, 2005). Thus FTY720, already FDA approved for multiple sclerosis, is now being actively investigated as a cancer therapeutic.

3. DELIVERABLE AND CLINICALLY RELEVANTSPHINGOLIPID-BASED DRUGS FOR CANCER

3.1. Methods for Delivery of Sphingolipids and Mimetics to Cells

Instead of relying on increasing “flow through the faucets or clogging the drains,” we can also simply add water to the sink externally to overflow the system (Fig. 1). Over the past 10 to 15 years, a number of approaches have been developed to enhance delivery of sphingolipids and analogues directly to cells in an effort to alter cell physiology, usually to disrupt the growth of tumor cells. Of these delivered sphingolipids, the most common is ceramide, with numerous chemical modifications and synthetic derivatives created to increase cell permeability, efficacy, and chemotoxicity. One method of increasing specificity of ceramide derivatives for negatively charged cellular compartments (such as the mitochondria) is the introduction of a positive charge on the fatty acid residue by the addition of a pyridine structure. Studies with pyridine-ceramides showed that they localized more readily to the mitochondria and induced the release of cytochrome c, a marker of apoptosis (Novgorodov et al., 2005; Senkal et al., 2006). These positively charged ceramides selectively induced pancreatic cancer cell death (Beckham, Lu, et al., 2013).

Another approach is to shorten the ceramide fatty-acid chain length to C₂, C₆, or C₈ (Liu, Beckman, & Foroozesh, 2013). These synthetic shortchain ceramides have increased cell permeability and have been shown to have potent anti-tumor effects (Obeid, Linardic, Karolak, & Hannun, 1993; Wolff, Dobrowsky, Bielawska, Obeid, & Hannun, 1994). C₆ ceramide has been utilized for the development of a novel nanoliposome formulation termed ceramide nanoliposomes (CNLs). These CNLs can deliver C₆ ceramides effectively as well as synergizing with current chemotherapeuitcs by encapsulating them within the same nanoliposomes (Adiseshaiah et al., 2013; Kester et al., 2015; Ma, Mou, Ding, Zou, & Huang, 2016; Sun, Fox, Adhikary, Kester, & Pearlman, 2008). The efficacy of these CNLs is currently being investigated in a phase 1 clinical trial in patients with solid tumors.

Finally, several sphingosine analogues have been researched as promising sphingolipid drugs, such as safingol (Schwartz et al., 1995, 1997), N,N-DMS (Edsall et al., 1998; Sweeney et al., 1996), fingolimod (Chun & Hartung, 2010; White, Alshaker, Cooper, Winkler, & Pchejetski, 2016), and (R)-FTY720 methyl ether (ROME) (Watson et al., 2013). These sphingosine analogues display anti-tumor effects as inhibitors of SphK enzyme activity and/or S1P receptors signaling.

3.2. Sphingolipid-Based Therapeutics Clinical Pipeline

A search through the clinicaltrials.gov database for various sphingolipid keywords such as S1P, sphingosine, ceramide, and sphingomyelin returned an impressive number of completed, ongoing, or recruiting clinical trials. For the sake of brevity and clarity, this review is focused on the three largest classes of sphingolipid-related drugs currently in clinical trials: synthetic ceramides, S1P antagonists, and SphK inhibitors. While clinical trials for other sphingolipid pathway inhibitors exist (notably amitriptyline, an acid SMase inhibitor), they are not as far along in the clinical pipeline and thus will not be discussed here (ClinicalTrials.gov) (Beckmann, Sharma, Gulbins, Becker, & Edelmann, 2014). Fig. 3 depicts various sphingolipid-based therapeutics that have entered clinical oncology trials.

Figure 3.

An outline of the sphingolipid-based therapeutics currently undergoing clinical trials for cancer treatment. Although there are few sphingolipid therapies currently in the clinic, many other sphingolipid-based therapeutics are showing promise in preclinical models.

Synthetic ceramides are ceramide analogues that have been modified to increase their tumor-specific toxicity, membrane permeability, or efficacy. As mentioned earlier, numerous synthetic ceramides exist, but the most common involve modifications to the hydrocarbon chain length, rendering it shorter. In 2003, topical administration of C₂ and C₆ ceramides were employed in a phase II study against cutaneous breast cancer, with mixed results (Clinical trial #: NCT00008320) (ClinicalTrials.gov). The trial found that while the C₂ and C₆ ceramides had little to no toxicity, patients yielded only a 4% response rate, shelving topical administration of short chain ceramides as a potential chemotherapeutic (Jatoi et al., 2003). However, by encapsulating a C₆ ceramide into stable nontoxic nanoliposomes, researchers found that they could increase chemotoxicity as compared to free ceramide alone in in vivo models (Kester et al., 2015; Sun et al., 2008). These CNLs entered a phase I study beginning in 2017 for patients with advanced solid tumors at the University of Maryland, the University of Virginia, and the Medical University of South Carolina (Clinical trial #: NCT02834611) (ClinicalTrials.gov).

For the past decade, S1P has been the focus of intense study due to its tumorigenic properties. As a result, numerous inhibitors against the enzyme responsible for its generation, SphK, have been advanced as candidates for clinical trials. Of these, the furthest down the clinical pipeline is ABC294640. ABC294640 acts as a specific SphK2 inhibitor, with numerous studies demonstrating its efficacy and utility as a potential chemotherapeutic across a broad range of cancer types (Schrecengost, Keller, Schiewer, Knudsen, & Smith, 2015; Yang et al., 2015). According to clinicaltrials. gov, six separate trials used ABC294640 against a variety of cancers including pancreatic cancer, cholangiocarcinoma, multiple myeloma, and hepatocellular carcinoma. Trademarked as YELIVA, ABC294640 is currently being employed in a single-arm phase IIA clinical study in the treatment of cholangiocarcinoma at the Mayo Clinic and the MD Anderson Cancer Center (Clinical trial #: NCT03377179).

By far the most common sphingosine analogue in clinical trials today, FTY720 (also known as fingolimod and trademarked as gilenya) has been referenced 94 times on clinicaltrials.gov and is an FDA-approved therapy for multiple sclerosis (Ingwersen et al., 2012). Though reports differ, the most universal mechanism for its potential anti-cancer properties is outcompeting sphingosine for the binding pocket of SphK, thus limiting the conversion of sphingosine to S1P (Billich et al., 2003; Lim et al., 2011; Paugh, Payne, Barbour, Milstien, & Spiegel, 2003). The first trial using FTY720 as a potential treatment for cancer began in 2015, when a phase I safety study of FTY720 with radiation and temozolomide against glioblastoma or anaplastic astrocytoma was conducted at Johns Hopkins University (Clinical trial #: NCT02490930) (ClinicalTrials.gov). For more information regarding FTY720’s role in cancer treatment, see the 2016 review by White et al. (2016). FTY720 has shown promise as a cancer therapy partly because of its ability to prevent pro-tumorigenic T cell transport out of the thymus by blocking the S1P receptor (S1PR) on T cells (Matloubian et al., 2004; Segui, Andrieu-Abadie, Jaffrezou, Benoist, & Levade, 2006). In addition, safingol is currently in the clinical setting for a phase 1 clinical trial for metastatic solid tumors in combination with fenretinide. Another clinical trial already completed used safingol combined with cisplatin for metastatic solid tumors. This trial showed that administering safingol with cisplatin was safe for cancer patients and ascertained as a recommended dose for safingol in a future phase 2 trial (Dickson et al., 2011).

Other, glyco- and sphingolipid analogues have also been shown to regulate immune components, which are dysregulated in cancer. α-Galactosylceramide, a sphingolipid produced by marine sponges, activates invariant natural killer T cells, which recruits anti-tumor immune cells (Berkers & Ovaa, 2005) (McEwen-Smith, Salio, & Cerundolo, 2015). A chemical modified α-galactosylceramide, KRN7000, has been used in a phase 1/2 clinical trial for treatment of patients with NCT00698776 myeloma (Berkers & Ovaa, 2005; Ishikawa et al., 2005). Other sphingolipid-targeting drugs are being developed that have immune-regulating properties. The ACDase inhibitor LCL521 decreased viability of the tumor-promoting myeloidderived suppressor cells (MDSCs) in mice (F. Liu et al., 2016; Umansky & Sevko, 2013). In addition, SphK2-dependent dihydrosphingosine-1phosphate also ameliorates MDSCs in mice models of solid tumors (Barth et al., 2013). Finally, GM3 ganglioside, overexpressed in human melanomas, was recently targeted with a preventative vaccine in mice models that resulted in rejection of B16 melanoma tumors in a CD8+ dependent manner (Mazorra, Mesa, Fernandez, & Fernandez, 2008).

4. SPHINGOLIPIDS SYNERGIZE WITH CHEMOTHERAPEUTICS

Recent studies have shown that elevating endogenous levels of ceramide by opening metabolic “faucets” or closing some of the enzymatic “drains” can augment the efficacy of conventional treatments in cancer. In addition, exogenous delivery platforms for sphingolipid mimetics can also elevate endogenous ceramide levels (Fig. 1). This section will review underlying combinatorial/synergistic cytotoxic mechanisms between elevated ceramide levels/decreased S1P levels and chemotherapeutic regulation of signal transduction cascades.

4.1. Synergistic Therapeutic Strategies Against Glucosylceramide Synthase

Perhaps the most prevalent area of research regarding synergy of sphingolipid therapies involves inhibiting GCS to limit conversion of pro-apoptotic ceramide into the less toxic glucosylceramide in cancer cells. Additionally, inhibitors of GCS downregulate the multidrug resistance gene 1 (MDR1) and its protein product P-glycoprotein (P-gp) reducing drug resistance (Gouaze et al., 2005; Morad & Cabot, 2013).

In breast cancer cell lines, an antisense-driven GCS knockdown led to higher ceramide levels and increased caspase-3 activity, which resensitized these cells to doxorubicin (Liu, Han, Giuliano, Hansen, & Cabot, 2000). The same lab later showed that using a pharmacological approach to inhibit GCS could also synergize with doxorubicin (an anthracycline chemotherapeutic). However, the novel, mixed backbone oligonucleotide antisense to GCS (MBO-asGCS), showed greater than 10-fold sensitivity to doxorubicin in doxorubicin-resistant lines of breast and ovarian cancer, through increased drug uptake and C₁₈-ceramide—induced apoptosis (Liu et al., 2004; Patwardhan et al., 2009). The MBO-asGCS was also found to decrease levels of MDR-1, cSrc kinase, nuclear β-catenin, and P-gp that sensitized breast, ovarian, cervical, and colon cancer cells to doxorubicin (Liu et al., 2010). Finally, the sensitization of ovarian cancer cells by PPMP or MBO-asGCS to doxorubicin was shown to only be effective in p53 mutants and could be mitigated by CerS inhibition using FB1 (Liu, Patwardhan, Bhinge, et al., 2011).

Using GCS antisense RNA and paclitaxel together enhanced ceramide formation, but, interestingly, also showed 10-fold higher levels of intracellular drug and 80% less P-gp beginning to suggest the relationship between P-gp and GCS. PPMP also restored sensitivity to vinblastine by enhancing its uptake by threefold, and diminished MDR1 expression by 84% in a vinblastine-resistant cell line (Gouaze et al., 2005). Additional studies using shRNAs against GCS confirmed increased apoptosis following treatment with doxorubicin and vinblastine while also displaying synergy with daunorubicin (Sun et al., 2006) and paclitaxel (Sun et al., 2010).

A study in 2001 that showed the GCS inhibitor DL-threo-phenyl-2hexadecanoylamino-3-pyrrolidino-1-propanol (PPPP) synergizes with vincristine in leukemic cell lines (Olshefski & Ladisch, 2001) paved the way for a more in-depth analysis of 65 human leukemic samples. This study established not only that GCS expression levels correlated with increased multidrug resistance in patients and cell lines but also that PPMP could sensitize leukemic cell lines to doxorubicin (Xie et al., 2008). Counterintuitively, one study in leukemia cell lines showed that by blocking GCS using both PDMP and PPMP, cells were no longer sensitive to daunorubicin due to an anti-apoptotic effect of accumulated galactosylceramides (Grazide et al., 2004).

GCS inhibition using PDMP has also been shown to synergize with new treatment options, including BCL2-inhibitor ABT-737 (Casson et al., 2013) and Imatinib (Baran, Bielawski, Gunduz, & Ogretmen, 2011) in multiple leukemia cell lines, with temozolomide in glioblastoma multiforme (Giussani, Tringali, Riboni, Viani, & Venerando, 2014), with curcumin in melanoma (Yu, Li, Qiu, & Sun, 2012), and with sorafenib in hepatocellular carcinoma (Stefanovic et al., 2016). Efficacy of glucosylceramide synthase inhibition by PPMP and tamoxifen (an anti-estrogen with P-gp antagonistic properties) was also shown to be augmented in combination with C₆ ceramide in neuroepithelioma and ovarian cancer cells, respectively (Chapman, Gouaze-Andersson, Karimi, Messner, & Cabot, 2011; Spinedi, Bartolomeo, & Piacentini, 1998).

4.2. Synergistic Therapeutic Strategies Against Sphingosine Kinase

As mentioned previously, S1P is the most pro-mitogenic sphingolipid. SphK inhibitors decrease S1P levels in cells and increase levels of the more-toxic sphingosine leading frequently to increased ceramides as well.

The first of many synergy studies investigating SphK inhibition in blood cancers was performed in 2006 when a combination of F-12509A, a microbially derived weak inhibitor of SphK and CerK, and doxorubicin or etoposide effectively resensitized AML cell lines, which had developed chemoresistance (Bonhoure et al., 2006). Additionally, decreased SphK1 activity resulting from treatment with etoposide in combination with Perifosine (an Akt inhibitor) was able to increase Fas-receptore—mediated cell death in human T-leukemia—mimicking Jurkat cells (Nyakern, Cappellini, Mantovani, & Martelli, 2006). Another SphK1 inhibitor (Ski), but not a SphK2 inhibitor (ROME), was able to synergistically induce autophagic cell death in T-ALL cells in combination with vincristine (Evangelisti et al., 2014). A different SphK2 inhibitor (Ski-II) was able to synergize with ABT-263 (a BCL2-like protein inhibitor) by causing accumulation of sphingosine and ceramide, leading to in vitro death of multiple human leukemia cell lines (Casson et al., 2013). In myeloma lines, K145 (a specific SphK2 inhibitor) in combination with Bortezomib, a receptor tyrosine kinase (RTK) inhibitor, synergistically induced ER stress and led to apoptosis (Wallington-Beddoe et al., 2017).

This combination of SphK and RTK inhibitors was also successful in solid tumor models. A combination of ABC294640 (a SphK2 inhibitor) and another RTK inhibitor, sorafenib, synergistically induced cytotoxicity in hepatocellular carcinoma (Beljanski, Lewis, & Smith, 2011), pancreatic adenocarcinoma, kidney carcinoma (Beljanski, Knaak, Zhuang, & Smith, 2011), and cholangiocarcinoma (Ding et al., 2016). Finally, there appears to be promise in combining the previously mentioned Ski-II with temozolomide (a monofunctional alkylator) to synergistically increase caspasedependent cell death in glioblastoma multiforme (Noack, Choi, Richter, Kopp-Schneider, & Regnier-Vigouroux, 2014).

4.3. Synergistic Therapeutic Strategies Against Acid Ceramidase

In multiple prostate cancer cell lines, the ACDase inhibitor NOE synergized with the EGFR inhibitor PD135035 to show drastic increases in both the estimated levels of ceramide as well as the percentage of cells undergoing apoptosis (Mimeault, Pommery, & Henichart, 2003). Interestingly, despite the success of NOE in the previous study, another group studying prostate cancer found that NOE did not synergize when combined with another compound, 4-HPR (fenretinide). However, using a different ACDase inhibitor, DM102, with fenretinide synergistically decreased cell viability, enhanced caspase activity, led to a 30-fold increase in ROS levels, and increased dihydroceramide levels. Although these results suggested activation of de novo ceramide synthesis, myriocin was unable to rescue from this cytotoxic effect, whereas surprisingly, antioxidant vitamin E did convey a partial rescue (Gouaze-Andersson et al., 2011).

Further work in head and neck squamous cell carcinoma showed that while overexpression of ACDase promoted resistance to Fas ligand-induced cell death, Fas-induced death was also potentiated by siRNAs against ACDase. Additionally, inhibition of ACDase using LCL 204 in the presence of Fas ligand led to cellular effects including: decreased survivin protein, increased cleaved caspase-8, and increased caspase-3 levels in vitro. The combination of these same two compounds also decreased tumor volume and promoted survival in in vivo xenograft models (Elojeimy et al., 2007). ACDase inhibition additionally showed synergy with daunorubicin in hepatocarcinoma cells. While NOE or siRNAs against ACDase in combination with daunorubicin increased the ceramide:S1P ratio resulting in increased cell death in vitro, the siRNA and daunorubicin combination was also efficacious in decreasing tumor volume in in vivo xenograft models. This synergistic cell death, which did not occur in healthy mouse hepatocytes, was preceded by markers of mitochondrial-induced apoptosis including ultrastructural changes, generation of ROS, release of cytochrome C, and caspase-3 activation (Morales et al., 2007).

4.4. Synergistic Therapeutic Strategies Using Exogenous Sphingolipids

Short-chain ceramides including C₂ and C₆ species have increased solubility and cell permeability compared with longer-chain, more physiological, ceramides. Short-chain ceramides have proven to be synergistic with docetaxel, doxorubicin, methotrexate and pemetrexed (antifolates) in murine melanoma cell lines, and osteosarcoma (Chapman et al., 2011; Feng, Li, Liu, Yang, & Zhang, 2014; Zhai, Sun, Han, Jin, & Zhang, 2015). Combining C₆-ceramide with ACDase inhibitor (DM102), AZD-8055 (mTORC1/2 dual inhibitor), or AT406 (a small molecule IAP antagonist) was synergistically efficacious in breast cancer cell lines, hepatocellular carcinoma lines, or pancreatic cell lines and primary samples, respectively (Flowers et al., 2012; M. Liu, Gu, Guo, & Fan, 2016; Zhao, Sun, Zhang, Zhang, & Zhang, 2016).

Nanotechnology platforms to improve ceramide delivery including CNL have shown synergy with sorafenib, PPMP, FTY720, gemcitabine, vinblastine, and tamoxifen/verapamil/cyclosporine A in in vitro and in vivo models of melanoma, breast, natural killer cell leukemia, pancreatic, hepatocellular carcinoma, and colorectal cancer, respectively (Adiseshaiah et al., 2013; Chapman et al., 2010; Jiang et al., 2011; Tran, Smith, Kester, & Robertson, 2008; Watters et al., 2013). It has been hypothesized that vinblastine and other microtubule destabilizers such as chloroquine and vincristine synergize with C₆ ceramide by reducing autophagosome formation and switching ceramide-induced autophagy into ceramide-induced cell death (Adiseshaiah et al., 2013; Lima, Milstien, & Spiegel, 2017; Young et al., 2016). A recent study showed that the use of CNL combined with adoptive T cell transfer immunotherapy in mice bearing liver tumors was able to reduce the number of tumor-associated macrophages and decreased the ability of these immune cells to suppress the anti-tumor properties (Li et al., 2018).

5. CONCLUSION

In the ever-growing quest to better understand the basis of cancer, numerous tools and collaborative efforts have emerged as powerful initiatives for research and medicine. The Cancer Genome Atlas represents one of the largest such endeavors, chronicling patient samples and organizing them by cancer type, gene expression, RNA profile, and mutation profile, among others. With the increasing sensitivity and selectivity as well as availability of mass spectrometry devices capable of creating full sphingolipid profiles for patients, we would argue that these vital class of lipids should be included when discussing patient diagnosis and treatment plans, as every year new research is published establishing links between sphingolipids and cancer.

As researchers continue to define the essential cellular functions of sphingolipids, their pivotal role in cancer progression is beginning to come into focus. Discoveries that established links between sphingolipid metabolism and carcinogenesis have cemented the need for a more complete understanding of these bioactive lipids. The medical advances in systems biology, big data, genomics, and epigenetics will synergize with advances in lipidomics to further the precision medicine revolution. All of these tools will further advance the utility of sphingolipid-based chemotherapeutics in oncology. Sphingolipid metabolism inhibitors, sphingolipidmodulated immunotherapies, cannabinoids, and sphingolipid analogues hold the promise for entirely new approach to cancer treatments.