Ca2+ as a therapeutic target in cancer

Abstract

Ca²⁺ is a ubiquitous and dynamic second messenger molecule that is induced by many factors including receptor activation, environmental factors, and voltage, leading to pleiotropic effects on cell function including changes in migration, metabolism and transcription. As such, it is not surprising that aberrant regulation of Ca²⁺ signals can lead to pathological phenotypes, including cancer progression. However, given the highly context-specific nature of Ca²⁺-dependent changes in cell function, delineation of its role in cancer has been a challenge. Herein, we discuss the distinct roles of Ca²⁺ signaling within and between each type of cancer, including consideration of the potential of therapeutic strategies targeting these signaling pathways.

1. Introduction

Ca²⁺ is a ubiquitous and dynamic second messenger molecule that regulates a variety of cellular functions including proliferation, cell cycle control, focal adhesion and others (Bong & Monteith, 2018). As such, it is perhaps not surprising that the contribution of Ca²⁺ signals to cancer progression has long been under investigation for many years. Hence, dysregulation of Ca²⁺ signaling is now well established to modulate oncogenic signaling pathways and promote invasive behavior (Bong & Monteith, 2018). The oldest record of the use of Ca²⁺ in cancer treatment is from 1896, when Calcium Carbide (CaC₂) was used as a clinical palliative treatment for uterine cancer (Das, 1896), while the role of Calcium Chloride (CaCl₂) in cancer growth was proposed in 1918 (Cramer, 1918). As of today, there are over 42,000 publications featuring both Ca²⁺ and cancer listed in PubMed, a number that has increased fourfold over the last 20 years.

In considering the role of Ca²⁺ in cancer, it is important to recognize that the expression and function of Ca²⁺ permeable channels vary considerably between cell types, driving considerable heterogeneity across different cancer types. This results in what seems to be mutually exclusive roles in cancer biology; Ca²⁺ can drive proliferation, migration and invasion, but can also induce apoptosis. Overall, this fuels the perception that there is no relationship at all. However, the reality is that Ca²⁺ signals serve distinct roles in different contexts, a point that we hope to clarify over the course of this review.

2. Ca2+ signal generation

Ca²⁺ signals are generated in response to a wide variety of autocrine, paracrine, hormonal, neurocrine and environmental factors. Here, we will briefly discuss some of the major underlying approaches that cells use to generate Ca²⁺ responses.

2.1. Receptor-mediated control of Ca2+ signals

G-Protein Coupled Receptor (GPCR) and Receptor Tyrosine Kinase (RTK) families are widely expressed and facilitate Ca²⁺ responses in virtually all cell types through phospholipase C (PLC). Briefly, as shown in Fig. 1, when Gαq-coupled GPCRs (driving PLCβ activation) or RTKs with PLCγ docking sites are activated, PLC degrades phosphatidylinositol 4, 5-bisphosphate (PIP₂) into Inositol Triphosphate (InsP₃) and diacylglycerol (DAG). InsP₃ diffuses within the cell toward the ER where it encounters InsP₃ Receptors (InsP₃Rs), ER-localized Ca²⁺ channels that mediate ER Ca²⁺ release. The resultant ER Ca²⁺ depletion leads to a highly conserved process known as store-operated Ca²⁺ entry (SOCE) (Parekh, 2008; Parekh & Putney, 2005; Putney, 1986; Takemura & Putney, 1989). SOCE is initiated by Stromal Interaction Molecule (STIM), a single pass membrane protein that acts as an Endoplasmic Reticulum (ER) Ca²⁺ sensor, responding to ER Ca²⁺ depletion by translocating within the ER toward the plasma membrane (PM) where it binds to and activates members of the Orai family of Ca²⁺-selective channels (Putney, 1986; Soboloff, Rothberg, Madesh, & Gill, 2012). STIM1 and Orai1 serve as the predominant mediators of SOCE in most cell types (Soboloff et al., 2012), although the other members of the STIM and Orai families can function analogously, although with distinct characteristics. For example, a decrease in the Ca²⁺ affinity of STIM2’s EF hand causes its activation at near-resting ER Ca²⁺ levels (Bird et al., 2009; Brandman, Liou, Park, & Meyer, 2007; Stathopulos & Ikura, 2013). Given that, it is perhaps not surprising that STIM2 also exhibits a decreased capacity to activate Orai1 (Zhou et al., 2009) which has been attributed to the presence of a leucine in place of a key phenylalanine within the Orai-binding region of STIM-Orai Activating Region (SOAR) (Wang et al., 2014). Orai2 and Orai3 can each be activated by STIM proteins expressed heterologously (DeHaven, Smyth, Boyles, & Putney Jr., 2007; Lis et al., 2007), although the roles of the endogenous proteins are less clear. Transformed cells can exhibit fundamental changes in the expression and activation of all components of these pathways (GPCRs, RTKs, STIM, and Orai) ( Jardin & Rosado, 2016; Nieto Gutierrez & McDonald, 2018; Regad, 2015), making this an important current area in the investigation of Ca²⁺ signaling in cancer cells. As such, this will be discussed considerably throughout this review.

Fig. 1.

Model of Ca²⁺ Homeostasis. At rest Ca²⁺ levels are higher in ECM and in the ER (darker blue), which functions as an intracellular Ca²⁺ store. Upon ligand binding to receptors, including GPCR and RTK, downstream effector PLC is activated which activates TRPC channels for Ca²⁺ entry and facilitates generation of InsP₃. InsP₃ binds to InsP₃R which leads to Ca²⁺ exit from ER, thus depleting the intracellular store. Following Ca²⁺ release from InsP₃R channels, Ca²⁺ either leaves the cytosol through PMCA pumps or is transferred to mitochondria, which orient themselves close to ER in response to elevated intracellular Ca²⁺ levels, through mitochondria IMM channel, MCU. Lower ER Ca²⁺ levels (lighter blue) lead to activation of ER Ca²⁺ sensors, STIM proteins. STIM then binds to PM Orai Ca²⁺ channels leading to SOCE. SOCE leads to increases in cytosolic Ca²⁺, which are then pumped into the ER by SERCA pumps to replete the store. Each step of regulating Ca²⁺ homeostasis plays a role in cancer progression. Cancers affected by each step labeled respectively.

Although SOCE is a major source of Ca²⁺ entry, members of other classes of channels also mediate receptor-dependent Ca²⁺ signals. The Transient Receptor Potential (TRP) channel superfamily of non-selective cationic channels includes 28 different members organized into 6 families expressed in mammals: TRPA, TRPC, TRPM, TRPML, TRPP, and TRPV (Zheng, 2013). The canonical TRP (TRPC) family is particularly relevant in this context since, similar to SOCE, they are activated downstream of PLC-coupled receptors (Vazquez, Wedel, Aziz, Trebak, & Putney, 2004), facilitating Ca²⁺ entry through SOCE-like mechanisms (Cheng, Liu, Ong, Swaim, & Ambudkar, 2011). Hence, DAG directly activates TRPC3/6/7 (Dietrich, Kalwa, Rost, & Gudermann, 2005; Venkatachalam, Zheng, & Gill, 2003) while TRPC1/4/5 activation is thought to be mediated by PIP₂ depletion (Myeong et al., 2018; Trebak et al., 2009). TRPC2 is not expressed in higher mammals (Vazquez et al., 2004) in rodents, it is found primarily within the vomeronasal organ (Mast, Brann, & Fadool, 2010). As discussed below, many TRPC family members exhibit dysregulated expression in cancer, making this area of considerable potential interest.

2.2. Environmental triggers

Several members of the TRPV, TRPM, and TRPA family are widely expressed, facilitate Ca²⁺ entry in a variety of cell types, and respond to a variety of environmental factors (see Table 1). For example, TRPV1–4, TRPM8, and TRPA1 are thermosensitive channels activated by temperature changes, each within a specific range (McKemy, 2005; Wetsel, 2011). In addition, there are several chemical activators of TRPs, which ultimately accounts for the sensory confusion between temperature changes and different sensations.

Table 1.

Environmental triggers of TRP channel activation.

TRP channel	Activators	Cancers affected	Mechanism of activation
TRPV1	≥43°C1	Prostate (S)	1. Enthalpy/Entropy induced channel conformation changes
	Capsaicin2	Cervical (S)
	Mustard Oil	Glioma (S)
	Garlic	Breast Cancer (S)	2. Stabilized open conformation leading to ATP binding to ARD domain of channel
TRPV2	≥52°C	TNBC (S)	Enthalpy and Entropy induced conformation changes
TRPV3	≥33°C	NSCLC	Cell Cycle Arrest
TRPV4	28–42°C	Breast (P/S)	Arachidonic Acid/PKA phosphorylation
TRPM8	≤28°C1 <<15°C2 Menthol3	Prostate (P) Gastric (P) Pancreatic (P) Lung (P) Breast (P)	1,2. (More likely) Enthalpy/Entropy induced channel conformation changes (Activation intensity is inversely related to temperature) or (Less likely) lysophospholipids/PUFAs leading to lower activation threshold temperature
			3. PIP2 synthesis
TRPA1	Mustard Oil
	Wasabi
	Cinnamon
	Ginger

Activators of TRP channels shown, which include temperature- and agonist-induced activation. Mechanisms of activation associated with each trigger are included as well as cancers affected. (P) indicates promotes, (S) indicates suppresses, (P/S) indicates both promotes and suppresses individual cancer upon activation. TRPV3: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4848893/.

For example, Capsaicin (the active ingredient in pepper) is a vanilloid (Szallasi & Blumberg, 1999) that stabilizes TRPV1 in its open conformation (Yang & Zheng, 2017). In this conformation, ATP binds to the intracellular ankyrin-repeat domain (ARD) of TRPV1 (Phelps, Wang, Choo, & Gaudet, 2010; Zheng, 2013), lowering the threshold of temperature for channel activation (Tominaga & Caterina, 2004). Resultant increases in cytosolic Ca²⁺ content cause channel inactivation due to competition between Ca²⁺/Calmodulin (Ca²⁺/CaM) and ATP for ARD binding (Phelps et al., 2010). Independent of capsaicin, TRPV1–4 sense temperature increases, with TRPV2 activated in response to the most extreme temperature changes (Wetsel, 2011). Indeed, TRPV2 is only activated above 52°C and although the exact mechanism is not fully understood, it is hypothesized that the channel exhibits structural flexibility, resulting from enthalpy and entropy shifts allowing for conformation change in the channel. In support, TRPV2 enthalpy and entropy capacity are twofold larger than TRPV1, whose capacity is fivefold larger than most other ligand- or voltage-gated channels (Liu & Qin, 2016), suggesting thermodynamics play a role in channel activation. Interestingly, the mechanisms that activate TRPV channels in response to temperature and agonists diverge in mechanism but converge onto the same activation gate (Cui et al., 2012; Grandl et al., 2010; Matta & Ahern, 2007; Yang, Cui, Wang, & Zheng, 2010) demonstrating the flexibility and versatility of these channels. TRPV3 responds to less extreme changes in temperature (~33°C) (Wetsel, 2011) and is activated through GPCR-induced breakdown of PIP₂ that facilitates a voltage-independent activation of the channel in response to colder temperatures (Doerner, Hatt, & Ramsey, 2011), albeit temperatures resembling physiologic conditions. TRPV4, meanwhile, responds to a wide range of temperature (28–42°C) and is also activated by cAMP-induced PKA phosphorylation of Arachidonic Acid (AA) at S824 (Cao et al., 2018). Hence, all four thermosensitive TRPV channels are activated in response to similar stimuli, although through different mechanisms. While temperature changes are not directly cancer relevant, different types of cancers exhibit different signaling dysregulation, which would have differential effects upon different TRPV channels, precisely exemplifying the context-dependent nature of Ca²⁺ signaling in a tumor environment.

Conversely, TRPM8 and TRPA1, also thermosensitive, are activated by temperature decreases and agonists such as menthol and mustard oil/wasabi, respectively. TRPM8 is activated at 28°C and its level of activation increases in response to temperatures colder than 15°C. It is hypothesized that it can be activated by depletion of membrane lysophospholipids and polyunsaturated fatty acids (PUFAs), since lysophospholipids lower the threshold of activation (Andersson, Nash, & Bevan, 2007). This would fit into evolutionary biology that membrane fluidity decreases at cold temperatures, however, it is more likely that TRPM8 is directly activated by cold temperatures leading to enthalpy and entropy changes (Zakharian, Cao, & Rohacs, 2010), similar to the TRPV2 response to heat. Menthol-induced TRPM8 activation occurs through PIP₂ synthesis as depletion inhibits the sensitivity of the channel (Liu & Qin, 2005), precisely the opposite relationship to that between PIP₂ and TRPV1. TRPA1 also senses cold temperatures, although only in more noxious cold conditions (<17°C) (McKemy, 2005). More importantly, TRPA1 is a potent sensor of exogenous stimuli as it is activated in response to many agonists, mainly mustard oil and wasabi, but also cinnamaldehyde (cinnamon) and ginger (McKemy, 2005) through binding to cysteine residues in the N-terminus of the channel via covalent interactions and/or disulfide bonds (Hinman, Chuang, Bautista, & Julius, 2006; Macpherson et al., 2007; Samanta, Kiselar, Pumroy, Han, & Moiseenkova-Bell, 2018; Takahashi et al., 2008). TRPA1 also has the unique ability to function as a cannabinoid-sensitive pain receptor in response to inflammation (Akopian, Ruparel, Patwardhan, & Hargreaves, 2008; Ruparel, Patwardhan, Akopian, & Hargreaves, 2011). These cannabinoids activate Ca²⁺/CaM signaling cascades which attenuate the response of TRPA1 to the exogenous agonists (Akopian et al., 2008). There is also some evidence that cannabinoids can also inhibit TRPV1 activity (Ruparel et al., 2011), which in addition to TRPA1, can regulate inflammation (Akopian et al., 2008).

2.3. Voltage-gated Ca2+ channels

Members of the Voltage-Gated Ca²⁺-permeable Channel (VGCCs) superfamily are primarily expressed in excitable (neurons and muscle) cells and are activated by membrane depolarization (Catterall, 2011). VGCCs can have up to four subunits, but all contain the α1 subunit which is the pore-forming portion of the channel, consisting of four domains (I–IV) each containing six transmembrane segments (S1–S6). The S4 segment of each domain acts as a voltage sensor that rotates outward upon exposure to electrical current and facilitates a conformation change opening the pore, while the S5 and S6 subunits form a membrane-associated pore lining loop containing a series of glutamate residues that determine Ca²⁺ selectivity (Heinemann, Terlau, Stühmer, Imoto, & Numa, 1992). Both α2δ and β subunits have been demonstrated to increase Ca²⁺ current VGCCs by recruiting and allosterically regulating α1 subunits (Walker & De Waard, 1998). α2δ subunits are extracellular subunits (Catterall, 2011) that are linked by disulfide bonds and mainly responsible for anchoring VGCCs to membranes using Glycophosphatidylinositol (GPI) anchoring sites most likely through hydrophobic sequences (Dolphin, 2013). β subunits are intracellular (Catterall, 2011), facilitate channel trafficking to the PM (Dolphin, 2013) and can also shift voltage dependence to make activation easier (Dolphin, 2003). Finally, γ subunits are transmembrane glycoproteins that have been shown to actually limit Ca²⁺ current through stabilizing the inactivated state of VGCCs in skeletal muscle cells. Further, this subunit is only present in skeletal muscle cells and is not expressed heterologously in other cell types (Campiglio & Flucher, 2015).

Within the VGCC superfamily, there are five different classes consisting of L-, N-, P/Q-, R-, T-type channels. CaV1 channels, or L-type channels, are long-lasting dihydropyridine channels and are mainly responsible for excitation-contraction coupling in smooth muscle cells and consist of CaV1 α1 pore-forming subunits (Zamponi, Striessnig, Koschak, & Dolphin, 2015). N-, P/Q-, and R-type channels are all members of the CaV2 channel family. N-type channels are named for the nervous system, where they are mainly expressed, however, they can also be found in cell types throughout the body (Phan et al., 2017). P/Q-type channels are named after Purkinje cells where they were discovered and are mainly located in cardiac (Purkinje) and neuronal (cerebellar granule) cells (Catterall, 1998), where R-type channels are also expressed (Zamponi et al., 2015). Finally, CaV3 channels are T-type (transient or short-lasting) and can be activated at lower membrane potentials during membrane depolarization. Similar to L-type channels, they are expressed in most excitable cells (Dziegielewska et al., 2016) and involved in excitation-contraction coupling. In neurons, VGCCs are critical for synaptic transmission (Catterall & Few, 2008; Dunlap, Luebke, & Turner, 1995; Tsien, Lipscombe, Madison, Bley, & Fox, 1988) while in muscle, they drive contraction (Catterall, 1991, 2011; Tanabe et al., 1993). However, it has become increasingly clear that many non-excitable cell types, including cancer cells (Phan et al., 2017) also express T-type channels. How they function in these systems and might contribute to cancer biology will be discussed further in subsequent sections.

2.4. Ca2+ pumps

As mentioned previously, the ER represents a major Ca²⁺ storage site in the cell. This is generated and maintained by the Sarco/Endoplasmic Reticulum Ca²⁺ ATPase (SERCA) pump, which catalyzes ATP to drive Ca²⁺ from the cytosol to the ER. SERCA exists in three isoforms (SERCA1, SERCA2, SERCA3) that are encoded by three different genes (ATP2A1, ATP2A2, ATP2A3, respectively), which all have highly conserved sequences and share similar functions within a variety of cells. SERCA1 is mainly expressed in fast twitch skeletal muscle, SERCA2a is mainly expressed in cardiac muscle cells, and SERCA2b and SERCA3 are ubiquitously co-expressed in most cells (Periasamy & Kalyanasundaram, 2007). As the primary regulator of ER Ca²⁺ content SERCA is not only critical for the generation of Ca²⁺ signals, but also, as discussed further in subsequent sections, protein processing, mitochondrial metabolism and/or Ca²⁺ overload. Hence, dysregulated expression and/or function of SERCA function can contribute to the behavior of transformed cells while also offer potential insights into therapeutic opportunities.

Ca²⁺ clearance across the PM is mediated by members of the Plasma Membrane Ca²⁺ ATPase (PMCA) family. PMCA proteins pump Ca²⁺ from the cytosol into the extracellular milieu. PMCA exists in 4 isoforms (PMCA1, PMCA2, PMCA3, PMCA4) encoded by 4 different genes (ATP2B1, ATP2B2, ATP2B3, ATP2B4 respectively) that together contain 20 different splice variants (Strehler & Thayer, 2018). PMCA1 is ubiquitously expressed in all cell types and ATP2B1KO is embryonic lethal (F Talarico, Kennedy, Marfurt, Loeffler, & Mangini, 2005), PMCA2/3 are expressed mainly in excitable cells such as neurons and striated smooth muscle, but also has been found in mammary gland and uterus (Brini & Carafoli, 2011). PMCA4 is expressed in most cells. ATP2B4KO is survivable, but leads to male infertility (Okunade et al., 2004). These pumps are primarily activated by Ca²⁺/Calmodulin when cytosolic Ca²⁺ concentration increases (Brini & Carafoli, 2011), although several recent studies have demonstrated that PMCA is also modulated by protein-protein interactions (reviewed in Strehler et al., 2007). PMCA dysregulation leads to a variety of diseases including deafness, hypertension, cardiovascular disease (Stafford, Wilson, Oceandy, Neyses, & Cartwright, 2017) as well as a variety of cancers, which will be detailed below.

3. Effectors

3.1. Calcium and cell migration

Cell migration is a critical Ca²⁺-regulated cellular function for embryonic development (Ridley et al., 2003), wound healing (Barrientos, Stojadinovic, Golinko, Brem, & Tomic-Canic, 2008), angiogenesis (Lamalice, Le Boeuf, & Huot, 2007) and cancer metastasis (Paul, Mistriotis, & Konstantopoulos, 2017; Vicente-Manzanares & Horwitz, 2011). Cell migration involves three distinct steps; responding to environmental cues to determine direction of movement, generating and maintaining cell polarity and activating motility processes for linear movement (Wei, Wang, Zheng, & Cheng, 2012). Cells can become polarized by extracellular stimuli, which drive movement toward the source of the signal mediated by Ca²⁺ flickers caused by InsP₃ production (Tsai, Kuo, Chang, & Tsai, 2015) and multiple ion channels, including TRPCs (Li, Abuarab, & Sivaprasadarao, 2016), and STIM/Orai (Chen et al., 2011; Tsai et al., 2014). Upon polarization, the cell commences moving forward with the initial process of protrusion which utilizes actin polymerization on cellular projections such as lamellipodia and filopodia (Small, Stradal, Vignal, & Rottner, 2002). Subsequently, to terminate protrusion, the front of the cell retracts and attaches to lamella (Webb, Parsons, & Horwitz, 2002), located behind lamellipodia. Lamella recruit myosin to contract and turnover F-actin, leading to formation of further focal adhesions (Burnette et al., 2011). Focal adhesions are stabilized by integrins and associated cytoskeletal proteins to withstand strong forces and facilitate forward motion (Webb et al., 2002). This occurs primarily in the front of the cell (Vicente-Manzanares & Horwitz, 2011), fixing the front during forward motion and facilitating retraction of the rear portion of the cell.

Small GTPases are ~20kDa GTP binding proteins that hydrolyze GTP to GDP and serve critical roles in many cellular processes (reviewed in Song et al., 2019), including migration. Many GTPases such as RhoA, Rac and Cdc42 have positive bidirectional relationships with Ca²⁺ (Aspenström, 2004; Ridley, 2015), and thus, serve critical roles driving cell migration in response to polarity-induced Ca²⁺-flickers. These small GTPases are known to coordinate myosin-induced contractions (Tsai & Meyer, 2012) through activity of Myosin Light Chain Kinases (MLCK) (Kasturi, Vasulka, & Johnson, 1993; Tsai et al., 2015) for cell movement, which is important for cells to thrive, especially during development, (further reviewed in Tsai et al., 2015). MLCK are Ca²⁺/CaM-dependent kinases that phosphorylate MLC, leading to myosin cross bridging and cell contraction (reviewed in Chavez, Smith, & Mehta, 2011), and have been demonstrated to facilitate cell movement and migration through regulating membrane tension and protrusion through binding and stabilizing F-actin (Chen et al., 2014). Hence, changes in Ca²⁺ lead to a signaling cascade culminating in increased cell movement and enhanced cell migration.

3.2. Calcium signaling and epithelial to mesenchymal transition

An important step in cancer progression is metastasis, which heavily utilizes an invasive process called Epithelial to Mesenchymal Transition (EMT). Unaffected tissue tends to adopt an ordered structure of epithelial cells that exhibit polarity through apical and basolateral membranes attached by tight junctions, adherens junctions, and desmosomes that form an epithelial layer (Dongre & Weinberg, 2019). However, when epithelial cells adopt a mesenchymal cell phenotype, they exhibit a change to a front to back polarity conformation (Dongre & Weinberg, 2019), that enhances migration and invasiveness to increase cancer metastasis as discussed further below.

3.2.1. Growth factors

EMT is induced by a variety of growth factors (GFs) including Epidermal Growth Factor (EGF), Fibroblast Growth Factor (FGF), and Transforming Growth Factor β (TGFβ) (reviewed in Dongre & Weinberg, 2019). These GFs can induce EMT by facilitating downstream signaling of EMT-associated pathways. For example, EGF can activate Signal Transducer and Activation of Transcription 3 (STAT3) (Wendt, Balanis, Carlin, & Schiemann, 2014; Zhong, Wen, & Darnell, 1994) or Similar to Drosophila Mothers Against Decapentaplegic 2/3 (Smad2/3) (Kim, Kong, Chang, Kim, & Kim, 2016), both of which are transcription factors that transcribe mesenchymal signature genes. TGFβ uses a similar mechanism in that it can activate Smad2/3 (Xu, Lamouille, & Derynck, 2009). FGF uses a different mechanism of action in that it mainly contributes to loss of cell polarity, promotes cell division, and inhibition of cell-cell contact (Sun & Stathopoulos, 2018). Hence, GFs play several important roles in promoting EMT by facilitating downstream EMT-associated pathways.

Although these GFs use different mechanisms to promote EMT, it is important to note that in both cases, GFs rely on Ca²⁺ signals to exert their effects. This is perhaps unsurprising as these GFs bind to RTKs that facilitate Ca²⁺ signaling as discussed above (see Section 2.1 (Bryant, Finn, Slamon, Cloughesy, & Charles, 2004; Ornitz & Itoh, 2015; Perez-Garcia, Muñoz-Couselo, Soberino, Racca, & Cortes, 2018; Qi et al., 2016; Zhang et al., 2017). Ca²⁺ signals serve a critical role in EMT by driving the expression of mesenchymal signature genes and facilitating ECM degradation, as outlined below.

3.2.2. Mesenchymal signature

EMT induction is associated with several transcription factors that comprise a mesenchymal signature. The most common of which are Twist, N-Cadherin, Zeb1/2, and Snai1/2 (Snail and Slug); loss of the epithelial marker, E-Cadherin also occurs (Dongre & Weinberg, 2019). This leads to the transcription of genes associated with survival and invasion pathways such as Mitogen Associated Protein Kinases (MAPKs), AKT, and Smads among others (reviewed in Kalluri & Weinberg, 2009). Thus, cells can activate an entire program of transcriptional machinery to change the genetic profile of the cell, which inflicts invasive properties that allow for dissemination to other organs.

In addition to these transcription factors, another important marker comprising the mesenchymal signature is intermediate filament, Vimentin (Dongre & Weinberg, 2019). Vimentin mediates cytoskeletal organization and focal adhesion through modulating the expression of integrins and reorganizing the orientation of microtubules (Liu, Lin, Tang, & Wang, 2015), which can lead to changes in cell polarity and increased motility which facilitate EMT (Liu et al., 2015; Mendez, Kojima, & Goldman, 2010). Therefore, in addition to reprogramming the transcriptional machinery of a cell, the EMT process leads to changes in morphology and polarity.

As discussed above (see Section 3.1), many of the processes involved in focal adhesion and migration are dependent on Ca²⁺. Hence, CAMKII has been shown to phosphorylate and activate Vimentin (Goto et al., 1998) and Vimentin activation and STIM1/TRPC1-dependent Ca²⁺ entry has been show to increase Vimentin activity (Stewart, Ralyea, & Lockwood, 2019). Further, much of the transcriptional machinery activated during EMT is associated with Ca²⁺; an effect that appears to be very context-dependent. For example, TRPM7 activates STAT3, driving Twist, Zeb1/2, and N-cadherin expression in a Ca²⁺-dependent manner (Davis et al., 2013). Additionally, it has been demonstrated that Orai3 and STIM1 can mediate Ca²⁺-dependent TGFβ-induced Snai1 activation, yet blocking Orai1 leads has same effect (Bhattacharya et al., 2018), suggesting Orai3-specific control of EMT.

3.2.3. ECM degradation

Sometimes, cells require degrading the surrounding Extracellular Matrix (ECM) in the tissue microenvironment (TME) to facilitate movement. This is facilitated by the modulation of adhesion molecules by degrading appendages; invadopodia and matrix metalloproteinases (MMPs).

Cells contain migratory appendages called podosomes that consist of actin and tyrosine-phosphorylated proteins that mediate cellular adhesion to the ECM (Tarone, Cirillo, Giancotti, Comoglio, & Marchisio, 1985). However, in transformed cells, podosome activity is altered due to oncogenic induction of aberrant tyrosine kinase activity, ultimately leading to ECM degradation; podosomes capable of this are commonly referred to as invadopodia (Chen, 1989; Murphy & Courtneidge, 2011).

Ca²⁺ serves a critical role in the formation of both podosomes and invadopodia (Siddiqui, Lively, Vincent, & Schlichter, 2012; Sun et al., 2014). For example, SOCE inhibitors block podosome formation through inhibition of MLCK activity (Chen et al., 2017). Further, STIM1/Orai1-dependent Ca²⁺ oscillations drive invadopodia formation, as well as Matrix Metalloproteinase (MMP) activity (Sun et al., 2014). Because constitutive Ca²⁺ entry would be cytotoxic for the cell due to mitochondrial Ca²⁺ overload, cells rely specifically on oscillations to regulate ECM degradation and invasion (Sun et al., 2014). Interestingly, it has been shown that downstream signaling is regulated by these oscillations and is reliant on oscillation frequency and amplitude (Bird et al., 2009; Dupont, Combettes, Bird, & Putney, 2011). This explains the importance of Ca²⁺ signaling, and specifically SOCE, for invadopodia-induced ECM degradation as SOCE provides the metaphorical fuel for the Ca²⁺ oscillations to induce ECM degradation. Findings of this nature have driven numerous investigations of the potential link between Ca²⁺ signaling and metastasis in many different cancer lineages.

Another factor in ECM degradation are MMPs, which are Ca²⁺-dependent endopeptidases responsible for remodeling of the ECM (Cathcart, Pulkoski-Gross, & Cao, 2015). Although MMPs are involved in non-pathological processes such as wound repair, development, and immune responses (Cathcart et al., 2015), MMPs can also exhibit pathological activity and have been found to directly contribute to invasion and metastasis in cancer by degrading adhesion molecules leading to changes in cell–cell contact and cell-ECM interactions (Gialeli, Theocharis, & Karamanos, 2011). For example, MMPs are known to cleave E-Cadherin (Noe et al., 2001), the most widely studied epithelial marker, as well as bi-directionally associating with GFs such as TGFβ using Ca²⁺-dependent MAPK pathway in cancer (Gomes, Terra, Wailemann, Labriola, & Sogayar, 2012; Illman, Lehti, Keski-Oja, & Lohi, 2006). Hence, Ca²⁺-dependent control of MMP activity plays an important role in EMT and cancer progression.

3.3. Control of metabolism

Mitochondria are best known as the site of energy production, but also serve complex roles in ion signaling and cancer biology. The role of metabolism in cancer has been widely investigated, predating the discovery of oncogenes and tumor suppressors by many years (DeBerardinis & Chandel, 2016; Huebner & Todaro, 1969; Warburg, 1925). This “Warburg effect” can be summarized as a metabolic shift away from oxidative phosphorylation in cancer cells, leading to lactic acid fermentation (Liberti & Locasale, 2016). Although less efficient at energy generation, this favors anabolic growth and catabolism-induced cell survival (Boroughs & DeBerardinis, 2015; DeBerardinis & Chandel, 2016). The underlying mechanisms responsible for the Warburg effect remain under investigation, however, it is important to recognize that increases in matrix Ca²⁺ content drive NADH, and thus ATP, production (Griffiths & Rutter, 2009; McCormack, Halestrap, & Denton, 1990). This is facilitated by a series of mitochondria-specific channels, exchangers and pumps that control mitochondrial ion content (reviewed in Rizzuto et al., 2009).

As shown in Fig. 2, to enter the mitochondrial matrix, Ca²⁺ must pass through the outer mitochondrial membrane (OMM), the intermembrane space (IMS) and the inner mitochondrial membrane (IMM). Voltage-Dependent Anion Channels (VDACs) are located on the OMM. They are weakly selective for anions at lower membrane potentials (channel is in “open” state), but favor cations at higher membrane potentials (when the channel is in a “closed” state) (Rizzuto et al., 2009). Although upon closure, VDACs allow entry of other cations, such as Na⁺ and K⁺, Ca²⁺ is favored (Tan & Colombini, 2007). VDACs also serve a critical role in linking mitochondria to InsP₃Rs and the ER, critical for mitochondrial Ca²⁺ loading (Rapizzi et al., 2002; Rizzuto et al., 2009). These links permit preferential loading of the mitochondria since they experience Ca²⁺ concentrations at near ER Ca²⁺ levels, critical for the entry of Ca²⁺ into the matrix. Hence, the major Ca²⁺ channel on the IMM is the mitochondrial Ca²⁺ uniporter (MCU), a high capacity, low affinity Ca²⁺ channel (reviewed in Kamer & Mootha, 2015) that does not permit Ca²⁺ entry at less than 10μM (Kirichok, Krapivinsky, & Clapham, 2004), a concentration not normally observed in the cytosol (on average). This movement across the IMM is driven by its highly negative membrane potential (~−180mV) produced due to proton pumping during oxidative phosphorylation. Upon entry into the matrix, Ca²⁺ binds directly to pyruvate dehydrogenase (PDH) phosphatase (PDP) (Denton, 2009), the first step in the generation of acetyl CoA at the beginning of the TCA cycle along with the rate-limiting TCA cycle enzymes α-ketoglutarate dehydrogenase (αKGDH) and isocitrate dehydrogenase (IDH) (Gellerich et al., 2010) increasing their activity and driving NADH generation. Notably, decreased pyruvate dehydrogenase phosphatase activity at lower mitochondrial Ca²⁺ content leads to conversion into lactate via lactic acid fermentation (Saunier, Benelli, & Bortoli, 2016; Shan et al., 2014). Hence, changes in mitochondrial Ca²⁺ loading may contribute to the Warburg effect and can be greatly altered in transformed cells.

Fig. 2.

Mitochondria-ER Regulation of Ca²⁺ and Metabolism. Store-Operated Channels (SOC) allow Ca²⁺ (ions labeled blue; concentration depicted as darker) to flow across PM. Once inside the cell, Ca2+, along with Calmodulin Kinase II (CAMKII) activates cytosolic calmodulin and flows into ER through SERCA pump (not shown). ER Ca²⁺ exits ER through InsP₃R and enters OMM through VDAC channels with a higher affinity in the closed conformation. Ca²⁺ then crosses IMM through Mitochondrial Ca²⁺ Uniporter (MCU) where it resides in the Inner Mitochondrial Matrix. Inside the matrix, Ca²⁺ binds to Pyruvate Dehydrogenase Phosphatase (PDP) which dephosphorylates Pyruvate Dehydrogenase (PDH) facilitating conversion of Pyruvate to Acetyl-CoA for subsequent commencement of the TCA cycle. When mitochondrial Ca²⁺ concentration is lower, PDP remains inactive which leads to sustained phosphorylation of PDH which is used for Lactic Acid Fermentation and subsequent cancer progression.

The sensitivity of mitochondria to Ca²⁺ levels has been considered a significant therapeutic opportunity. Hence, mitochondrial Ca²⁺ overload leads to Reactive Oxygen Species (ROS) generation due to overproduction of electrons from the ETC during oxidative phosphorylation (Adam-Vizi & Starkov, 2010). In contrast, if too little Ca²⁺ enters the mitochondria, decreased ATP generation can lead to autophagy (Rossi, Pizzo, & Filadi, 2018) due to AMPK activation (Bootman, Chehab, Bultynck, Parys, & Rietdorf, 2018; Rossi et al., 2018). Although mitochondrial Ca²⁺ loading is primarily dependent upon MCU and VDAC, ultimately, it is dependent upon the overall availability of Ca²⁺ to the cell. As such, mitochondrial function represents a significant potential implication of any change in cytosolic Ca²⁺ signaling in cancer cells.

3.4. Calcium and transcription factors

Ca²⁺ regulation throughout the cell has the ability to not only drive cellular processes, but also affect cellular physiology through control of transcription factor activity. Although crosstalk between signaling pathways can lead to pleiotropic effect on gene expression, several transcription factors are directly Ca²⁺-sensitive as shown in Fig. 3A and as discussed further below.

Fig. 3.

Transcriptional Regulators of Ca²⁺ Homeostasis. (A) Upon ligand binding to receptors such as GPCR/RTK, Gαq activates PLC to breakdown PIP₂ into InsP₃ and DAG. InsP₃ binds to InsP₃R which facilitates Ca²⁺ release from ER store into the cytosol. DAG activates PKC, which is further activated by increases in cytosolic Ca²⁺ levels. PKC. PKC, then, facilitates 3 transcription regulating cascades. First, PKC phosphorylates IκK which facilitates the breakdown of p100 into p52 which is an active transcription factor within NFκB pathway. Subsequently, p52/RelB dimer translocates to nucleus, leading to transcription of NFκB-associated genes. Second, PKC activates calcineurin, which complexes with Ca²⁺/calmodulin regulated protein phosphatase (CaM) to dephosphorylate NFAT. Dephosphorylated NFAT translocates to nucleus to facilitate NFAT-associated genes, including EGR genes among others. Third, PKC phosphorylates Raf, which activates MAPK pathway. ERK translocates to nucleus and binds to Elk1 substrate, facilitating transcription of EGR1, among other MAPK-associated genes. EGR1 activates ER Ca²⁺ sensor, STIM, which binds to PM Orai Ca²⁺ channel, which promotes SOCE upon ER Ca²⁺ store depletion. (B) NFAT protein broken up by regions (NHR-NFAT Homology Region, RHD- Rel-Homology Domain, C-Terminal Domain) which are further broken into sub-domains. NHR consists of Casein Kinase 1 (CK1) binding site, Transactivation Domain (TAD), Calcineurin Docking Sites (CDS1, CDS2), Serine-Rich Regions (SRR1, SRR2), 4 Serine-Rich Motifs (Sp1, Sp2, Sp3, and KTS), a Nuclear Localization Signal (NLS) and Nuclear Export Signal (NES).

3.4.1. NFAT

The Nuclear Factor of Activated T-Cells (NFAT) family of Ca²⁺-dependent transcription factors consists of 4 classic NFAT family members: NFATc1 (NFATc or NFAT2), NFATc2 (NFATp or NFAT1), NFATc3 (NFAT4), and NFATc4 (NFAT3) and the Ca²⁺-insensitive NFAT5. Originally discovered in T cells, NFAT family members are widely expressed, with well-defined roles in muscle, neurons, heart, osteoclasts and adipocytes (reviewed in Pan, Xiong, & Chen, 2013).

Structurally, NFAT members share a highly conserved Rel-Homology Domain (RHD) that gives them a similar DNA-binding specificity ( Jain, Burgeon, Badalian, Hogan, & Rao, 1995; Qin et al., 2014). NFAT homology region (NHR), which regulates NFAT activity, consists of approximately 12 regulatory regions including: an N-terminal Transactivation Domain (TAD) (1) with a Casein Kinase 1 (CK1) binding site (2), a Calcineurin Docking Site (CDS) (3), an additional CK1 binding site outside of the TAD (4) 2 serine rich regions (SRR1 and SRR2) (6), 4 serine rich motifs (SRM (SP1, SP2, SP3, KTS)) (10), a Nuclear Localization Sequence (NLS) (11), and Nuclear Export Signal (NES) (12) (Pan et al., 2013) (Muller & Rao, 2010) (see Fig. 3B).

As shown in Fig. 3A, at rest, NFAT members reside in the cytoplasm (Okamura et al., 2000). In response to increased intracellular Ca²⁺, PKC is activated and Ca²⁺/Calmodulin regulated protein phosphatase complex (CaM) associate with calcineurin. Following PKC activation and complex association, PKC activates Ca²⁺/CaM/Calcineurin complex to dephosphorylate NFAT, leading to translocation to the nucleus and subsequent transcription of target genes (Muller & Rao, 2010; Pfeifhofer et al., 2003; Villaseñor et al., 2017). Within NFAT, CDS facilitates calcineurin docking interactions (Garcia-Cozar et al., 1998) for dephosphorylation of NFAT at conserved phosphoserine residues (Qin et al., 2014). Mechanistically, dephosphorylation of NFAT SRR and SRM lead to a conformational change and subsequent exposure of its NLS (Li et al., 2011). Following NFAT NLS exposure, NFAT proteins then translocate to the nucleus and transcribe NFAT target genes (Pan et al., 2013). In fact, it has also been shown that not only does NFAT dephosphorylation lead to nuclear localization, but it also leads to higher NFAT affinity for specific regions of DNA, further promoting transcription of NFAT target genes (Neal & Clipstone, 2001; Okamura et al., 2000; Porter & Clipstone, 2002; Shaw et al., 1995). Following return to basal Ca²⁺ levels, SRR and SRM serve as phosphorylation sites (Okamura et al., 2000; Qin et al., 2014) leading to exposure of the NES which facilitates NFAT exit from the nucleus (Beals, Clipstone, Ho, & Crabtree, 1997; Okamura et al., 2000). Hence, sustained NFAT-mediated gene transcription requires sustained Ca²⁺ signals. NFAT is linked to a variety of oncogenic pathways in different capacities including but not limited to: Janus Kinase (JAK)/STAT (Lagunas & Clipstone, 2009), COX2 (Greenhough et al., 2009), and MAPK (Flockhart, Armstrong, Reynolds, & Lovat, 2009). Because NFAT activity is dependent on Ca²⁺, it underscores the importance of Ca²⁺ in the context of cancer. In this regard, targeting NFAT could be an efficacious therapy for a variety of cancers which will be discussed in great detail below.

3.4.2. NF-κB

NF-κB is a family of transcription factors that give rise to seven proteins from the combination of five gene-coding subunits, RelA/p65, c-Rel, RelB, p50, and p52 (Xia, Shen, & Verma, 2014). This family of inducible transcription factors with pleiotropic functions. Although heavily studied in lymphocytes, NF-κB signaling is present in a wide variety of cells (Zhang, Lenardo, & Baltimore, 2017). Further, its capacity to regulate cell survival can facilitate resistance to apoptosis in transformed cells.

As shown in Fig. 3A, although canonical NF-κB activation is Ca²⁺independent, Ca²⁺-dependent PKC activation facilitates CARMA1 phosphorylation to produce the CARMA1-Bcl10-MALT1 (CBM) complex in lymphocytes (Altman & Villalba, 2003; Villalba et al., 2000). Further, calcineurin can promote NF-κB activation by promoting IKK phosphorylation of IκB (Berry, May, & Freedman, 2018; Frantz et al., 1994; Steffan et al., 1995; Trushin, Pennington, Algeciras-Schimnich, & Paya, 1999), ultimately leading to proteasomal degradation. Upon IκB degradation, NFκB complexes translocate to the nucleus where they can drive gene transcription. NF-κB targets include cell cycle progression regulators such as Cyclin D, E, c-myc, and c-myb (as reviewed in Dolcet, Llobet, Pallares, & Matias-Guiu, 2005). Cyclin D and E induce cell cycle progression during mid-late G1 phase ( Joyce et al., 2001), and while Cyclin D suppresses differentiation (Guttridge, Albanese, Reuther, Pestell, & Baldwin Jr., 1999). Concurrently, c-myc promotes many survival processes such as cell growth, differentiation, and death (Romashkova & Makarov, 1999) and c-myb regulates growth through a variety of mechanisms depending on tissue type (Kim et al., 2009; Wang et al., 2015).

3.4.3. WT1/EGR1

Wilm’s Tumor Suppressor 1 (WT1) and Early Growth Response 1 (EGR1) proteins are dueling zinc finger transcription factors that regulate Ca²⁺ homeostasis in a variety of cell types. Over the last 10 years, our group has demonstrated that WT1 expression leads to decreased STIM1 expression and SOCE, while EGR1 expression leads to enhanced STIM1 and SOCE (Ritchie, Zhou, & Soboloff, 2011; Samakai et al., 2016), discussed in great detail in our recent review (Go, Gross, Hooper, & Soboloff, 2019). Briefly, EGR1 is activated downstream of a wide variety of growth and stress signals (Thiel, Mayer, Müller, Stefano, & Rössler, 2010). Although primarily associated with MAPK signaling, Ca²⁺ signals can also drive EGR expression (Flavell et al., 2008; van Loo et al., 2019), revealing an intriguing feedback loop between Ca²⁺ and EGR1 expression. The EGR1 promoter consists of five Serum Response Elements (SREs) and a cAMP Response Element (CRE). Currently, there is considerable evidence that Ca²⁺ regulates the transcription of CRE- and SRE-containing genes (Hardingham, Arnold, & Bading, 2001; Hardingham, Chawla, Johnson, & Bading, 1997), likely through Ca²⁺-dependent (Chuderland & Seger, 2008; White & Sacks, 2010) ERK activation (Müller, Lipp, & Thiel, 2012). Dysregulated expression and function of WT1 and EGR1 are observed in a wide variety of cancers (reviewed in Go et al., 2019), providing a potential causative explanation for why cancer cells often exhibit changes in Ca²⁺ signaling regulation.

WT1 and EGR1 contribute to a variety of cellular processes including but not limited to cell cycle regulation, differentiation, and apoptosis. EGR1 controls cell cycle progression by downregulating CDK4, which maintains quiescence (Min et al., 2008), Further, EGR1 promotes the transcription of genes driving differentiation in response to a variety of growth factors the development of breast epithelial cells and immune cells among others (Dinkel et al., 1998; Tarcic et al., 2012), whereas WT1 inhibits differentiation by displacing EGR transcription factors from its response elements (Scharnhorst & Jochemsen, 2001). Under some conditions, EGR1 can induce apoptosis through activating pro-apoptotic factors such as p53 in response to increased intracellular Ca²⁺ in melanoma cells (Nair et al., 1997) whereas specific WT1 splice variants can facilitate apoptosis through repressing both intrinsic and extrinsic survival factors such as Bcl2 and IGF1, respectively, in hepatoma cells (Menke, 1997). As these processes are important for normal cellular physiology, they are also important for cancer progression. As discussed in our most recent review (Go et al., 2019), EGR1 and WT1 play dual and opposing roles in cancer progression. Thus, targeting EGR1 and WT1 could be an effective, and further improved upon, therapeutic target in a variety of cancers and the current landscape of EGR1 and WT1 role in specific cancers, as well as any therapeutics targeting these transcription factors will be discussed in great detail throughout this review.

4. Cancer

Ca²⁺ plays a major role in both the initiation and progression of many different cancer types. Due to the plethora of literature regarding the relationship of Ca²⁺ with all different types of cancers, as well as the confined space of this review, it is impossible to thoroughly cover the effect Ca²⁺ signaling has on every cancer. Here, we thoroughly delve into examples of the complex and impactful role Ca²⁺ signaling plays in cancer, while touching on facets of other cancers reliant on Ca²⁺. These examples include cancers relying on both low and high Ca²⁺ content for disease progression, cancers relying on Ca²⁺-dependent pathways as described above, as well as cancers dependent on a variety of initiating factors including hormones, neural or immune cell differentiation, as well as genetic and environmental factors.

4.1. Breast cancer

4.1.1. Introduction

According to 2019 cancer statistics from the American Cancer Society, breast cancer has the highest incidence and second highest total mortalities among women nationwide (Alteri, 2019). Although breast cancer can be easily treated when detected early, high mortality rates indicate that breast cancer still presents challenges in treatment, especially in advanced stage disease. Clearly, the need for novel treatments exists.

Breast cancer is considered an adenocarcinoma, a cancer that starts in glandular tissue, and is heavily reliant on receptor-mediated signaling to proliferate. Like normal cells, cancer cells receive extracellular signals to facilitate their growth. Within breast cancer, there are three major subtypes broken down by receptor status. The three major receptors involved in breast cancer are steroid hormone receptors, estrogen receptor (ER), and progesterone receptor (PR), as well as non-steroid hormone receptor Human Epidermal Growth Factor Receptor 2 (HER2). To this end, clinically breast cancer is either: hormone receptor positive or negative (HR±; expressing estrogen receptor (ER) and/or progesterone receptor (PR)), HER2±, or triple negative (TNBC, negative for all three receptors) (Waks & Winer, 2019). In addition to hormone and HER2 receptor status, a recently discovered marker for breast cancer classification is Ki67, which is normally a marker for proliferation, leading to additional classification subtypes: Luminal A (ER+, HER2−, PR+/Ki67−), Luminal B (ER+, HER2−, PR− or Ki67+), Luminal B-like (ER+, HER2+, PR+/−, Ki67+/−), HER2+ (ER−, HER2+, PR−, Ki67+/−), and TNBC (ER−, HER2−, PR−, Ki67+/−). Further, there is some evidence of a basal-like sub-form of TNBC, that overexpresses basal cytokeratins, Epidermal Growth Factor Receptor (EGFR) and caveolin (CAV) 1/2 separate from traditional TNBC (Horton, Jagsi, Woodward, & Ho, 2018; Inic et al., 2014). Although the luminal subtypes tend to be the most common, TNBC is the most aggressive (Tong, Wu, Cho, & To, 2018). While treatment strategies are improving for breast cancer, it is important to note the highly mutually exclusive genetics between the subtypes, which can become even more variable between individual cases. This clearly presents a challenge that still exists in the breast cancer field in generating treatments amenable for the majority of breast cancer patients. Here, we delve into the current understanding of the role of Ca²⁺ signals in breast cancer biology and discuss its potential as a therapeutic target.

4.1.2. SOCE

Given the heavy reliance of breast cancer on receptor-mediated signaling, particularly in early stages, it’s reliance on multiple Ca²⁺-dependent signaling pathways including SOCE, TRP channels and VGCCs is, perhaps, not surprising. For example, it has been demonstrated that Orai1 RNAi attenuates proliferation, migration, invasion, and formation of lung metastases in vivo due to impaired focal adhesion turnover caused by impaired SOCE (Yang, Cao, Zhou, Feng, & Wang, 2009). Hence, not only does impaired SOCE limit the size of focal adhesions formed, the ability of focal adhesions to act as traction points for cellular movement was also compromised (Yang, Cao, et al., 2009). Most likely, impaired SOCE inhibits polarization, causing retention of ER Ca²⁺, which inhibits InsP₃-induced store depletion and Ca²⁺ flickers to allow cell movement. These observations implicate SOCE as a fundamental driver of breast cancer metastasis and could justify targeting SOCE to impede disease progression, although the efficacy of this strategy in patients has not been established.

Interestingly, different subtypes of breast cancer exhibit distinct expression patterns of STIM/Orai, which serve distinct roles in cancer progression. Hence, it has been demonstrated that elevated STIM1/STIM2 ratio is associated with poor patient prognosis in TNBC (McAndrew et al., 2011). This is most likely because STIM1 is the more effective facilitator of SOCE and SOCE is associated with poor prognosis in TNBC. Furthermore, induction of EMT has been shown to elevate the expression of STIM1 along with associated SOCE in MCF7 HR⁺ breast cancer cells (Hu et al., 2011; Mo & Yang, 2018). Thus, STIM1 appears to be important in both HR+ and HR− subtypes of breast cancer. Further, differences in androgen signaling between ER+ and ER− breast cancer lead to a switch in Orai expression from Orai3 to Orai1 (Motiani, Abdullaev, & Trebak, 2010; Motiani et al., 2013). Hence, androgens are converted into estrogen using the enzyme aromatase (Nelson & Bulun, 2001), the primary therapeutic target for HR+ breast cancer (Tong et al., 2018). Androgen deprivation has been shown to inhibit Orai1 production in other tissues such as prostate cancer (Flourakis et al., 2010). Hence, differences in androgen-mediated Orai1 expression may contribute to fundamental differences between HR+ and HR− breast cancer, with Orai3 expression being driven as a compensatory mechanism due to loss of Orai1 in early stage disease (Motiani, Zhang, et al., 2013). Notably, Orai1-mediated SOCE in TNBC drives Notch-mediated invasiveness under hypoxic conditions (Liu et al., 2018). Hence, breast cancer progression is marked by significant changes in how SOCE is generated and resultant invasive behavior.

4.1.3. TRP channels

TRPV channels, known for their thermosensitive properties, also have been demonstrated to contribute to breast cancer progression. For example, TRPV4 is a marker for many epithelial breast cancer tumors, where it is upregulated relative to normal tissue (Pla, Avanzato, Munaron, & Ambudkar, 2011). TRPV4 upregulation has been shown to promote AA-induced cytoskeletal reorganization, angiogenesis, and motility in HR+ breast cancer cells (Fiorio Pla et al., 2011; Pla et al., 2011). Conversely, TRPV4 knockdown has also proven to reduce breast cancer cell invasiveness through inhibition of TRPV4-dependent Ca²⁺-induced phosphorylation of AKT (Lee et al., 2017). Interestingly, although TRPV4-mediated signals promote invasive behavior in HR+ cells, its overexpression promotes cell death in MDA-MB-231 TNBC cells due to mitochondrial Ca²⁺ overload, metabolic crisis and, ultimately, apoptosis (Peters et al., 2017). Hence, hormonal and genetic factors can contribute to the design of optimal therapeutic strategies, with TRPV4 attenuation optimal at early stages and stimulation representing a potential therapeutic opportunity in HR- breast cancers

As reviewed extensively in Duncton (2015), a number of exogenous and endogenous pan-TRP agonists exist such as phorbol esters and AA, respectively. Efforts to generate specific TRPV4 modulators are ongoing. Specifically, Glaxo Smithkline (GSK) and Renovis (RN) have generated TRPV4 agonists, including GSK1016790A and RN1747, with the former used primarily to treat dysfunctional bladders. Conversely, TRPV4 antagonists exist such as Ruthenium Red (primarily known to interact with VGCC and used to study excitation of neurons (Tapia & Velasco, 1997), but also is a small molecule antagonist of TRV4), as well as more recently synthesized small molecule antagonists from GSK and RN such as GSK205, RN1734, and RN9893, respectively. Pfizer and Hydra Biosciences have alsogeneratedTRPV4antagonists,”Compound31”andHC067047,respectively (Duncton, 2015), although there remains no published evidence of their efficacy in the treatment of breast cancer.

TRPV6, exhibiting higher Ca²⁺ permeability compared to TRPV4, is well established to function as a tumor promoter. TRPV6 expression is regulated by estradiol in T47D ER+ breast cancer cells (Bolanz, Hediger, & Landowski, 2008), and amplified in HER2+ breast cancers (Peters et al., 2012). It has been demonstrated that TRPV6 expression and activity is stimulated by hormones, including estrogen and progesterone. Hence, estrogen induces TRPV6 expression through binding of estrogen receptor α (ERα) to an estrogen response element (ERE) within the TRPV6 promoter (Klinge, 2001). However, estrogen/ERα also binds directly to TRPV6 to facilitate activation (Kumar, Singh, Singh, Goswami, & Singru, 2017). Interestingly, although similar observations have been made for progesterone, the TRPV6 promoter does not contain progesterone response elements (PRE) (Kim, Lee, Ji, Choi, & Jeung, 2006; Lee, Lee, Jung, Choi, & Jeung, 2009). Precisely how progesterone regulates TRPV6 expression and function is not currently clear, although one possibility is that the progesterone receptor (PR) may weakly associate with the ERE (Petz et al., 2004). Thus, not only is TRPV6 expression increased in T47D cells compared to normal breast epithelial cells, but TRPV6 siRNA-mediated knockdown significantly reduced T47D cell proliferation (Bolanz et al., 2008). Hence, TRPV6 promotes the growth of receptor-dependent breast cancer. However, this same effect occurred in HR− cancers whether they were HER2+ or HER2− (Peters et al., 2012). Thus, it is likely that although estrogen and progesterone stimulate TRPV6, TRPV6-induced breast cancer progression is not reliant on hormone signaling. Instead, it has been proposed that Ca²⁺ influx leads to cell cycle progression (Santella, Ercolano, & Nusco, 2005), which may be a mechanism used across breast cancer subtypes (Peters et al., 2012). Thus, TRPV6 inhibitors have been suggested as a novel therapy for both ER+ (Bolanz et al., 2008) and ER− breast cancers (Peters et al., 2012).

Unlike TRPV4, TRPV6 has very few biologics generated. One TRPV6 inhibitor, SOR-C13 (see Table 2), is derived from a paralytic peptide (soricidin (ascension number P0C2P6)) from the Northern Short-tailed shrew (Blarina brevicauda) saliva and has been tested in a clinical trial. Although it is still in phase 1, results have been promising as the drug appears to be well-tolerated and provided anti-tumor activity (Fu et al., 2017; Wissenbach et al., 2001; Xu et al., 2015).

Table 2.

Ca²⁺ therapeutics for cancer.

Non-SOCE/SOCE	Class of Drug Target	Drug	Specific target of drug	Mechanism of action
Non-SOCE	Voltage gated	Amlodipine	L/N-Type	Binds directly to channel and blocks external surface
		Diltiazem	L/T-type	Decreases basal [Ca2+]c
	TRP	SOR-C13	TRPV6	Inhibits TRPV6 activity
		Capsazepine	TRPM8	Inhibits TRPM8 activity
	P2X	BBG	P2X7R antagonist	Competitive inhibition
		oATP
		AZD9056	P2X7R agonist	Bind to N-terminal domain
		eATP
		Alu-RNA		P2X7R agonist-bind to C-terminal domain
	Ca2+-related Transcription Factors	Cyclosporin A	NFAT	Blocks nuclear translocation of NFAT
	Pumps	Thapsigargin/Mipsagargin	SERCA	Blocks ATPase activity
		Cyclopiazonic Acid
		2,5-di-tertbutylhydroquinone (BHQ)
		CXL-017		Enhanced lysosomal Ca2+ release, blocks SERCA ATPase activity
		Ipramine Blue
		Sodium Orthovanadate	PMCA	Inhibits PMCA activity
SOCE	SOCE activity	CAI	SOCE	Targets anti-apoptotic, Mcl1
		Drebrin
		RP4010		Promotes actin reorganization and increased SOCE
				Synergistic with RTK inhibitor, preventing SOCE activation
	Orai channel	Diethylstilbestrol	Orai	DES may bind to steroid binding pockets in Orai channels or may modify Orai channel properties
		BTP2/Pyr2, Pyr3	Pyrazole	Block STIM1, Orai1, TRPC3, can interact with extracellular surface of Orai channels
		SKF-96365	Imidazole

Therapies that either target Non-SOCE- or SOCE-based mechanisms of Ca²⁺ regulation. Within Non-SOCE, classes of drugs include drugs targeting VGCC, TRP, and P2X channels.Within SOCE, classes of drugs include those targeting multiple modulators of SOCE (SOCE) or Orai or STIM1 individually. Known mechanisms of action for each drug are listed as described.

Finally, other TRP channels, have been shown to exhibit increased expression, most notably TRPC1, TRPC6, and TRPM7, in both HR+ and HR− breast cancers reviewed in Ouadid-Ahidouch, Dhennin-Duthille, Gautier, Sevestre, and Ahidouch (2013). TRPC1 has been shown to promote proliferation through activation of ERK1/2 in HR+ MCF7 breast cancer cells (El Hiani, Lehen’kyi, Ouadid-Ahidouch, & Ahidouch, 2009), while mechanisms explaining TRPC6- and TRPM7-induced proliferation are less clear. Further, TRPC1, TRPC6, and TRPM7 have also are associated with proliferation as siRNA-mediated knockdown of all these channels significantly inhibited proliferation in HR+ MCF7 and HR− MDA-MB-231 cell lines (Aydar, Yeo, Djamgoz, & Palmer, 2009; El Hiani et al., 2009; Guilbert et al., 2009), demonstrating direct roles in promoting proliferation. Although it is not yet determined what the mechanisms of TRPC6- and TRPM7-induced proliferation are, it is important to further investigate them. TRPC6 is actually the most highly expressed TRP channel in breast cancer (Dhennin-Duthille et al., 2011; Guilbert et al., 2009), however, mainly what is known about this channel in the context of breast cancer is correlations in expression with breast cancer compared to normal tissue. TRPM7, meanwhile, has also been shown to facilitate migration through increasing cell polarity in a Ca²⁺-dependent manner leading to changes in myosin-based cell tension and facilitating cell–cell adhesion and cell movement (Middelbeek et al., 2012). Even though TRPM7 has been shown to be important for breast cancer progression, there is limited evidence of this, especially in the context of how Ca²⁺ is involved in regulating this phenomenon. Further understanding of these mechanisms could lead to potentially new therapeutic targets in the context of breast cancer.

4.1.4. Purinergic receptors

Purinergic receptors are non-selective cationic channels located on the PM that mainly allow Ca²⁺ entry and are directly activated by the binding of purine-containing molecules, such as ATP. Purinergic compounds include but are not limited to: Adenosine Triphosphate (ATP), Guanine Triphosphate (GTP), as well as their diphosphate (ADP and GDP, respectively) and monophosphate (AMP and GMP respectively) derivatives (Shaw, 1984). As reviewed in Burnstock (2014), within this class of Ca²⁺ channels, exist two subfamilies which are P1, binding specifically to adenosine, and P2, binding mainly to ATP. Further P2 can be broken down into 2 subclasses of families. P2XR which are Ca²⁺ channels specifically activated by ATP, or P2Y which are actually ligand-gated GPCRs (Erb & Weisman, 2012). However, in the scope of this review, we will mainly focus on P2XR. Because, P2XRs are specifically activated by ATP (Coddou, Yan, Obsil, Huidobro-Toro, & Stojilkovic, 2011), they contain an ATP binding pocket on the extracellular side of the channel, illustrating that this channel is mainly activated by extracellular ATP from the surrounding environment (Di Virgilio, Dal Ben, Sarti, Giuliani, & Falzoni, 2017). Upon ATP binding to the extracellular ATP binding pocket, the channel undergoes a conformation change leading to opening of the pore (Di Virgilio et al., 2017). Normally, these receptors are present on excitable cells, such as neurons and smooth muscle, to facilitate synaptic transmission and smooth muscle contraction respectively (Hattori & Gouaux, 2012; North, 2016), however, it has been recently discovered that these receptors are also expressed in non-excitable, cancer cells (reviewed in Adinolfi, Capece, Amoroso, De Marchi, & Franceschini, 2014). As these receptors are ATP-gated, it is important to note that ATP levels are low in unaffected tissue interstitium residing in the nanomolar range, but can reach micromolar range in pathologic tissue (reviewed in Di Virgilio & Adinolfi, 2016). Although multiple links between cancer and P2XR subtypes are a result of rising extracellular ATP levels in the TME, the most commonly upregulated and targeted subtype is P2X7R. In fact, P2X7R is considered to be a marker for the onset of multiple cancers including early stage prostate cancer and all stages of breast cancer (Burnstock & Di Virgilio, 2013). Further, roles for P2X7R have been demonstrated in both the initiation and progression of cancer. For example, P2X7R activation in T47D ER+ human breast cancer cells enhanced cell migration and increased the development of metastasis via induction of AKT and dysregulation of the adhesion molecules E-Cadherin and Matrix Metalloproteinase 13 (MMP13) (Xia, Yu, Tang, Li, & He, 2015). Therefore, similar to SOCE-mediated breast cancer progression, it appears P2X7R-mediated cancer progression occurs through EMT and enhanced cell survival. Because Ca²⁺ plays an important and parallel role in many of these EMT pathways, and is the main cation entering cells through this channel, it is likely that Ca²⁺ plays an important role in this progression.

P2X7R is a bifunctional channel with differential effects on cell survival depending on cell type and level of activation (reviewed in Di Virgilio & Adinolfi, 2016). Persistent elevation of ATP levels can cause “tumor killing” (Burnstock & Di Virgilio, 2013; Di Virgilio, 2012; Di Virgilio & Adinolfi, 2016) through cytochrome C release leading to subsequent apoptosis (Selzner et al., 2004; Slee et al., 1999). Likely, this reflects persistently elevated ATP concentrations found in some tumor tissues which can drive over-activity of P2X7R, ultimately driving Ca²⁺ overload-induced cytotoxicity (Peng & Jou, 2010). However, P2X7R can also promote cell survival at lower ATP concentrations through NFκB activation, reflecting the dependence on very tightly regulated ATP concentration to promote P2X7R-mediated Ca²⁺ influx and subsequent activation of Ca²⁺-dependent survival pathways.

As shown in Table 2, the P2X7R antagonists Brilliant Blue G (BBG), oxidized ATP (oATP) (Soares-Bezerra et al., 2015) and AZD9056 (Burnstock & Knight, 2018) have been tested clinically in the context of inflammatory diseases such as Crohn’s Disease and liver fibrosis among others (reviewed in Savio, de Andrade Mello, da Silva, & Coutinho-Silva, 2018) and have advanced to phase 2 trials (Gunosewoyo & Kassiou, 2010). In addition, P2X7R can be activated through traditional nucleotide agonists, such as eATP (Coddou et al., 2011), as well as non-nucleotide agonists such as Alu-RNA, which activate the channel through direct binding on the intracellular N- and C-terminal domains (Di Virgilio, Giuliani, Vultaggio-Poma, Falzoni, & Sarti, 2018; Fowler et al., 2014). Considering the bifunctional nature of P2X7R signaling in the specific case, either agonists or antagonists could have potential therapeutic benefit. Hence, any therapeutic strategy to target breast cancer based on P2X7R signaling would require consideration of the size and signaling environment of the tumors being targeted.

4.1.5. VGCC

As discussed previously, there are many other classes of VGCCs, however, published evidence of their expression and function in breast cancer is limited to T-type (CaV3) channels, a class of VGCCs that are heterologously expressed among multiple cancer types. Specifically, CaV3.1 and CaV3.2 are both overexpressed in both ER+ and ER breast cancer cell lines with antagonists causing growth arrest in both subsets (Bhargava & Saha, 2018; Taylor et al., 2008). Since breast cancer cells are not excitable, it is important to consider other potential mechanisms that would control the function of VGCCs. One possibility is ambient pressure increases. Hence, raising extracellular pressure to 40mm Hg (normal atmospheric pressure is roughly 30mm Hg) activates CaV3.3, leading to PKCβ-mediated NFκB activation (Basson, Zeng, Downey, Sirivelu, & Tepe, 2015). In ER- breast cancer cells, NFκB activity is required since its inhibition leads to apoptosis due to prevention of the Unfolded Protein Response (UPR) (Fan et al., 2018), which is a stress response due to misfolded ER proteins that occurs due to lack of ER Ca²⁺. Hence, targeting CaV3.3 in ER- breast cancer could lead to loss of NFκB and promote cell death, particularly in combination with agents that drive ER stress, which would be expected to be more effective inducing cell death in the absence of protection from the UPR. Alternatively, targeting NFκB itself could have utility. Although immune suppression would be a concern, it is notable that suppression of NFκB prevents osteoclast differentiation; breast cancer metastases in bones is highly dependent upon osteoclast activity (Simone et al., 2015). Considered collectively, these findings reveal an unexpected link between a channel whose expression and function is typically limited to excitable cells and the survival and growth of breast cancer cells.

4.1.6. Ca2+ clearance

Typically, only PMCA1 and PMCA4 are expressed in non-excitable cells like breast cancer. However, both HR+ and HR− breast cancer heterologously express PMCA2 (Lee, Roberts-Thomson, & Monteith, 2005). Further, PMCA2 overexpression attenuates apoptosis induced by ionomycin in both T47D breast cancer and mammary epithelial cells (VanHouten et al., 2010). Patients with high PMCA2 expression have a poor prognosis, potentially due to control of calpain activity (VanHouten et al., 2010). Calpains are Ca²⁺-dependent proteases known to induce apoptosis (Gil-Parrado etal., 2002), and PMCA2lowers intracellularCa²⁺, thusleading to downregulated calpain, inhibited apoptosis, and enhanced progression (Gil-Parrado et al., 2002). Due to its ability to rapidly clear Ca²⁺, PMCA2 has also been shown to promote HER2 signaling through inhibition of receptor endocytosis in HER2+ breast cancer ( Jeong et al., 2016). Indeed, it appears that high intracellular [Ca²⁺] leads to direct binding between PMCA2 and HER2 within actinin-rich domains, which helps retain HER2 on the PM. Given its fundamentally protective roles, when PMCA2 is depleted, HER2+ can become more easily internalized, which increases sensitivity to pro-apoptotic reagents, such as anthracyclines (Doxorubicin), in invasive MDA-MB-231 breast cancer cells, most likely through downregulation of EGFR (Peters et al., 2016). Intriguingly, PMCA2 seems to play different roles in different forms of breast cancer. For example, in HR- basal cancers, EGFR and PMCA2 levels are negatively correlated,while inother (presumably HR+)breast cancerforms, EGFRand PMCA2 expression are positively correlated. The mechanistic basis for heterologous PMCA2 expression remains unclear, but the fact that its expression is linked to EGFR is intriguing and suggests context- and EGFR-dependent transcription mechanisms. Overall, these reports support the potential value of PMCA2 as a therapeutic target for the treatment of breast cancer, particularly those that are HER2+.

Downregulating PMCA4 is protective against colon cancer (Aung et al., 2009) and melanoma (Hollander et al., 2016) through induction of apoptosis, most likely due to mitochondrial Ca²⁺ overload (Aung et al., 2009; Monteith, McAndrew, Faddy, & Roberts-Thomson, 2007; Rizzuto & Pozzan, 2006). Although PMCA4 expression has no independent effect on breast cancer cell survival, it’s depletion enhances the apoptotic effects of the Bcl-2 inhibitor, Navitoclax, a caspase-dependent apoptosis promoter, in invasive MDA-MB-231 ER- breast cancer cells, (Curry, Luk, Kenny, Roberts-Thomson, & Monteith, 2012; Curry, Roberts-Thomson, & Monteith, 2016). Although Bcl2 inhibition represents a promising approach to breast cancer treatment, Bcl2 inhibitors have performed poorly in clinical trials due the presence of the B-galactosidase binding protein, Galectin-3 (Gal-3), which is an anti-apoptotic protein that can supplement or replace Bcl2, thus inhibiting apoptosis and conferring chemoresistance. The extent to which Gal-3 contributes to this link between PMCA and Bcl2 inhibition is not clear, although it is tempting to speculate on this possibility.

Overall, it is notable that, although the various channels and pumps described above all regulate cytosolic Ca²⁺ content, their roles in breast cancer function seem to be distinct. Differences in expression patterns, functions and intracellular location create context-dependent roles, necessitating in depth investigation of their functions before attempting to target them therapeutically.

4.2. Ovarian cancer

4.2.1. Introduction

Although breast cancer is the most common gynecologic cancer diagnosis among women, ovarian cancer is the most lethal (Alteri, 2019; Slatnik & Duff, 2015; Stewart et al., 2019). Generally, ovarian cancer is broken up into three subtypes. First, epithelial tumors account for up to 90% of ovarian cancers that consist of many different histological subtypes including serous, endometroid, clear cell, and mucinous carcinomas (reviewed in Matulonis et al., 2016). Less commonly, ovarian cancer can also exist as germ cell and sex cord stromal tumors (Stewart, Azimi, et al., 2019). Importantly, although it was previously thought that ovarian cancer arises from one entity, it is now accepted that ovarian cancers can form in different reproductive organs including the fallopian tube and peritoneum (Mallen et al., 2018; Stewart, Azimi, et al., 2019). Formation of tumors in these other gynecologic areas yield similar molecular profiles, which clinically categorizes them as the same disease even if the site of origin is different. Because tumors can originate in multiple areas of the reproductive system, ovarian cancer is very difficult to detect early, and most clinically diagnosed cases are of higher grade. Hence, the most common form of ovarian cancer is High Grade Serous Carcinoma (HGSC) which accounts for roughly 70%–80% of all ovarian cancer compared to 5% for its low grade counterpart. This is significant as ovarian cancer typically presents as an aggressive malignancy that likely has generated invasive properties and because it can originate in many different organs, treatment is very challenging. Thus, novel therapeutic strategies are needed to treat this disease.

Several different Ca²⁺-related pathways such as EGR1/WT1, TRP channels, and VGCC channels have been shown to cause ovarian cancer progression. Interestingly, much of what is known about ovarian cancer and Ca²⁺ is based on expression and prognosis correlations, yet no real consensus has been reached on mechanistic contributions of many of these channels, or even transcription factors regulating Ca²⁺ signals, which will be discussed below.

4.2.2. TRP channels

Although the role of TRP channels in ovarian cancer is somewhat under-investigated, preliminary research demonstrates that multiple subclasses of TRP Channels promote ovarian cancer progression. TRPC3 has been shown to promote ovarian cancer progression by increasing cell cycle progression through Ca²⁺-dependent phosphorylation of Cdc2 by CAMKIIα, driving progression from G2 to M phase (Yang, Zhang, & Huang, 2009). Interestingly, TRPV6 expression is elevated in ovarian cancer andTRPV6 inhibition impedes ovarian cancer tumor growth in xenograft models (Xue et al., 2018). Further, TRPV6 inhibitors, such as SOR-C13 and Lidocaine are recognized as diagnostic and/or therapeutic reagents for TRPV6-rich tumors including ovarian cancer (Bowen et al., 2013; Jiang, Gou, Zhu, Tian, & Yu, 2016). Finally, TRPM7 is also highly expressed in ovarian carcinomas (Wang et al., 2014) and is associated with phosphorylation of Ca²⁺-dependent survival pathways including AKT, p38/MAPK, and Src (Wang et al., 2014), the former of which has been linked to EMT in ovarian cancer (Liu et al., 2019). Because TRPC, TRPM, and TRPV channels all have therapeutics targeting them, future research into their potential as therapeutic targets for ovarian cancer could be fruitful.

4.2.3. SOCE

There have been a number of studies demonstrating dysregulation of SOCE during ovarian cancer progression, although the relationship between SOCE and ovarian cancer progression remains unclear. Hence, STIM1 and Orai1 expression are both upregulated in therapy-resistant ovarian cancer cells compared to therapy-sensitive cells, leading to increased AKT activity that promotes therapy resistance (Schmidt et al., 2014). Further, Placental Growth Factor (PlGF), which is produced by gynecologic tumor cells, also upregulates expression of both STIM1 and Orai1, but has been proposed to promote ovarian cancer progression through STIM1- and Orai1-induced HIF1α expression (Abdelazeem et al., 2019). However, SOCE inhibition is protective against ROS-induced apoptosis (Wang et al., 2016) likely as a result of decreased mitochondrial Ca²⁺. Hence, SOCE both promotes ovarian cancer progression and sensitizes cells to apoptosis Future investigations may help to delineate the complex relationship between SOCE and ovarian cancer progression and survival.

Interestingly, there have also been a number of studies linking EGR1 and WT1 with ovarian cancer progression; as discussed in Section 3.4, EGR1 and WT1 regulate STIM1 expression (Go et al., 2019; Ritchie et al., 2011; Samakai et al., 2016). The extent to which changes in STIM1 contributes to these effects is not currently clear, although these possibilities should be considered. For example, EGF-induced EGR1 expression has been demonstrated to upregulate the E-cadherin repressor Slug, leading to EMT which increases the invasive capacity and EMT activity of ovarian cancer in vitro (Cheng, Chang, & Leung, 2012) (Yan, Chen, Chen, & Chen, 2016). Downstream of EGR1, STIM1 expression has also been shown to be upregulated in cisplatin-resistant ovarian cancer leading to subsequent activation of AKT (Schmidt et al., 2014). However, the effect of STIM1 on ovarian cancer progression is somewhat understudied, there is preliminary evidence that STIM1 is associated with disease progression. In this context, EGR1 would increase ovarian cancer invasiveness. Interestingly, the role of EGR1 shifts in late stage ovarian cancer; analysis of patient databases reveal several-fold decreases in EGR1 expression in advanced carcinomas (He et al., 2015). Further, EGR1 can inhibit cyto-protective autophagy in the absence of autophagy regulator, ATG14, (He et al., 2015) indicating that these effects can be highly context-dependent.

Interestingly, WT1 has been shown to drive EMT in early ovarian tumorigenesis (Hylander et al., 2006; Miller-Hodges & Hohenstein, 2012). Further, because WT1 can be correlated with poor prognosis, it is used as a marker to track ovarian cancer progression (Andersson et al., 2014). The extent to which WT1-mediated suppression of STIM1 expression and SOCE affects EMT in ovarian cancer is unknown. However, given the established relationship between Ca²⁺ signals and EMT (see section IIID), it is reasonable to speculate on this possibility.

Therapeutically, Carboxyamidotriazole (CAI) (see Table 2), has been proposed as a treatment for ovarian cancer (Bonnefond et al., 2018; Mignen et al., 2005), has undergone phase II clinical trials and has been suggested as a maintenance drug for relapsed epithelial ovarian cancer (Hussain et al., 2003), but has not been investigated as a first-line therapy. Although CAI has been well-established as a SOCE inhibitor, as an ovarian cancer therapeutic, CAI utilizes a different mechanism of action. Thus, CAI inhibits the expression of the Bcl2 family member myeloid leukemia cell differentiation 1 (Mcl-1) (Thomas, Lam, & Edwards, 2010), itself a negative regulator of mitochondrial Ca²⁺ loading (Morciano et al., 2016). As such, Mcl1 is highly protective against mitochondrial Ca²⁺ overload and is associated with poor prognosis in ovarian cancer (Shigemasa et al., 2002). Further, CAI-induced Mcl-1 inhibition is correlated with inhibition of the mTORC1 pathway (Bonnefond et al., 2018) because mTORC1 controls Mcl-1 translation through 4E-BP1 phosphorylation in ovarian carcinoma (Hsieh et al., 2010). This process relies on Ca²⁺ as Ca²⁺/calmodulin regulates mTORC1 activity since CAMKs activate mTORC1 (Li et al., 2016). In conclusion, CAI-inhibition of SOCE occurs through changes in Mcl1 expression which likely inhibits both Ca²⁺-dependent survival pathways and leads to mitochondrial Ca²⁺ overload as its primary mechanisms of action. This can prove to be promising as a novel potential target in treating ovarian cancer.

4.3. Prostate cancer

4.3.1. Introduction

Analogous to breast cancer in women, prostate cancer is treatable when detected early, yet prostate cancer has the highest cancer incidence and second highest total mortalities among cancers in men in the United States (Alteri, 2019). This portrays the challenge of treatments for advanced stages of disease and the need for novel treatments. Prostate cancer is also a form of adenocarcinoma mainly divided into subgroups consisting of Androgen Receptor positive (AR+) and negative (AR−). The differences between the subgroups and their dependence on Ca²⁺ will be discussed in great detail below.

Also, similar to breast cancer, prostate cancer is heavily reliant on receptor-mediated signaling. Specifically, normal prostates use nuclear receptor, AR, which are activated in response to androgens (Lonergan & Tindall, 2011). Activation of this receptor leads to translocation to the nucleus where it facilitates transcription of androgen related genes, such as Prostate Specific Antigen (PSA), human glandular kallikrein (hK2), kallikrein related peptidase/serine protease 18 (prostase/PRSS18), and transmembrane serine protease 2 (TMPRSS2). These genes encode for serine proteases that regulate the prostate gland including facilitating proteolytic cascades leading to production of semen (Lin et al., 1999) and promoting sperm motility and cell movement (Lin et al., 1999; Pezaro, Woo, & Davis, 2014). However, because these proteases are involved in cell movement, it also has been characterized as invasive by degrading the surrounding ECM (Lin et al., 1999; Pezaro et al., 2014). Among these proteases, there is some evidence that the presence of TMPRSS2 can act as a switch between normal prostate epithelium and prostate cancer. If untranslated, this gene fuses with E26 Transformation Specific (ETS)-transcription factors to form fusion proteins that make up the vast majority of prostate cancer subtypes (Attard et al., 2016). The most common ETS transcription factor is ERG. Thus, TMPRSS2-ERG (T2E) is the most common genetic background of prostate cancer (Tomlins et al., 2008). Furthermore, it has been shown that T2E is associated with Ca²⁺. Indeed, use of Ca²⁺ channel blockers has actually been shown to reduce risk of developing T2E prostate cancer (Geybels, McCloskey, Mills, & Stanford, 2017), since T2E is heavily reliant on CaV1.3; CaV1.3-mediated Ca²⁺ entry activates AR (Chen et al., 2014). Additionally, other ETS-transcription factor subtypes, as well as sub-forms with mutations of Ras/Raf, and Speckle Type POZ Protein (SPOP), a cullin-based E3 ubiquitin ligase, also make up many individual prostate cancer genetic backgrounds (Attard et al., 2016). Hence, Ca²⁺ regulation has been shown to play an important role in governing AR activity, which implicates Ca²⁺ as a promising target for prostate cancer.

Currently, for AR+ prostate cancer, there are treatments involving inhibition of androgen production, namely Androgen-Deprivation Therapy (ADT). However, this appears to only be effective in early stages of prostate cancer and delay in treatment can lead to progression into castration resistance stages (Eisenberger et al., 1998; Pagliarulo et al., 2012; Schröder, Crawford, Axcrona, Payne, & Keane, 2012; Wen, Niu, Lee, & Chang, 2014). This is significant as it appears that hormonal regulation of prostate cells can prevent progression into later stages of disease. Most likely this is because AR actually suppresses later stages of prostate cancer progression (Niu, Altuwaijri, Yeh, et al., 2008; Niu, Altuwaijri, Lai, et al., 2008; Niu et al., 2010; Wen et al., 2014), as androgen deprivation leads to EMT since AR inhibits EMT marker, Zeb1 (Sun et al., 2012). This is further supported by evidence that androgens also contribute to homeostasis as deprivation activates damage-repairing cellular programs. For example, in the absence of androgens, cells upregulate anti-stress genes regulating detoxification, hypoxia response, DNA damage-repair, and p53 activation (Mulholland, 2012). Hence, targeting androgens appears to be very dependent on timing for effectiveness.

Along with timing, one of the biggest downfalls of androgen deprivation is in addition to AR+ subtype, AR− subtype also interacts with RTKs such as JAK/STAT pathways (reviewed in Hoang, Iczkowski, Kilari, See, & Nevalainen, 2016), which makes targeting specific pathways difficult. For example, whereas Stat5a/b are known to directly bind to AR and facilitate translocation to the nucleus in response to androgens (Tan et al., 2008), this effect can also occur in the absence of androgens due to genomic reprogramming in response to androgen deprivation (Sharma et al., 2013). Thus, progression to AR- prostate cancer occurs through the upregulation of AR-independent pathways that bypass or parallel AR signaling and lead to progression into CRPC and AR-prostate cancer (Arora et al., 2013; Hoang et al., 2016; Saraon, Jarvi, & Diamandis, 2011; Whitworth et al., 2012). An example of a bypass pathway includes activating the compensatory GPCR, glucocorticoid receptor (GR), in the absence of AR that induces a similar effect on disease progression (Arora et al., 2013; Hoang et al., 2016; Isikbay et al., 2014; Sahu et al., 2013). However, in addition to receptor-mediated signaling, it is also proposed that prostate cancer cells can undergo a Darwinian-type event where cells are selected for surviving in androgen deprived conditions (Hoang et al., 2016) most likely by the onset of Bcl2 expression (Whitworth et al., 2012). Because AR- prostate cancer is more advanced, it is more resistant to therapies including radiation, chemotherapy, and cytotoxic exposure attributing to stress responses activated in the absence of androgens (Mulholland, 2012). This makes treating AR- prostate cancer especially challenging. As a result, AR- prostate cancer therapies are limited at best. One attempt at treating AR-prostate cancer is Diethylstilbestrol (DES) (see Table 2), which is a synthetic ethinyl estrogen used to treat hormone-induced cancers such as CRPC in men and breast cancer in postmenopausal women (Roh, Eliades, Gupta, Grant-Kels, & Tsao, 2017). In prostate cancer, DES reduces production of testosterone (androgen), which slows prostate cancer growth, however this was discontinued as it showed limited clinical response (Bossi et al., 2009), demonstrating the clear need for novel therapeutics. Because AR+ and AR− subtypes of cancer rely on receptor-mediated signaling, Ca²⁺ homeostasis has been implicated in the progression of prostate cancer with several Ca²⁺-sensitive pathways dysregulated in prostate cancer, (reviewed in Maly & Hofmann, 2018) which will be discussed in great detail below.

4.3.2. SOCE

Prostate cancer appears to rely heavily on Ca²⁺-dependent signaling that dynamically regulates disease progression based on disease stage and genetic background. To this end, STIM1 and Orai1 expression have been shown to be correlated with AR levels (Berry, Birnie, Droop, Maitland, & Collins, 2011; Perrouin Verbe, Bruyere, Rozet, Vandier, & Fromont, 2016; Kappel et al., 2017). This is likely because STIM1 and Orai1 activate signaling cascades that activate AR. For example, normally, InsP₃ is generated from the cleavage of PIP₂ as discussed above. However, PIP₂ also can be phosphorylated by PM-associated PI3 kinase (PI3K) leading to the production of PI triphosphate (PIP₃). PIP₃ leads to activation of many downstream oncogenic pathways such as AKT and mTOR. In this regard, it has been shown that PI3K/AKT/mTOR pathway contributes to prostate cancer progression. In a previous cohort of prostate cancer patients, profound alterations were reported, in each of the three genes within this pathway. Further the entire pathway was deregulated in some capacity in 42% of localized and 100% of metastatic disease cases (Edlind & Hsieh, 2014; Taylor et al., 2010). PI3K/AKT/mTOR pathway is Ca²⁺-dependent; an effect demonstrated in prostate cancer as STIM1 suppression inhibited migration and invasion of prostate cancer cells in vitro through suppression of PI3K/AKT/mTOR pathway (Zhou et al., 2017). Thus, Ca²⁺-regulation of pro-invasive pathways such as PI3K/AKT/mTOR appear to play an instrumental role in prostate cancer progression. Further, this pathway also interacts with AR signaling (Crumbaker, Khoja, & Joshua, 2017), providing PI3K/AKT/mTOR as a potential mediator for how Ca²⁺ activates AR, a process that is known to occur in conjunction with gene fusions, such as T2E (Carver et al., 2009; King et al., 2009). This implicates SOCE as extremely important since it may serve as a marker for disease progression.

A driving force for SOCE is receptor-mediated depletion of ER Ca²⁺ store release as a result of PIP₂ cleavage. In contrast to PI3K converting PIP₂ into PIP₃, Phosphatase And Tensin Homolog (PTEN) is a PM-associated Phosphatidylinositol (PI) phosphatase which reverses this conversion leading to higher production of PIP₂, and eventually InsP₃, and is widely portrayed as a tumor suppressor in prostate cancer ( Jamaspishvili et al., 2018). In essence, it opposes PI3K, which is discussed above. However, depletion of PTEN is also exhibited in prostate cancer (Li, Li, Sun, & Li, 2018; Wise, Hermida, & Leslie, 2017). PTEN actually uses a two-pronged approach to suppressing prostate cancer progression. First, it blocks oncogenic pathways such as PI3K and AKT, but also helps restore normal Ca²⁺ homeostasis by facilitating SOCE through PIP₂ production leading to InsP₃-induced ER Ca²⁺ store release. Clinically, there are no current agonists that directly bind to and activate PTEN, however, there are many agonists that activate upstream proteins that upregulate PTEN expression (reviewed in Boosani & Agrawal, 2013). Hence, PTEN agonists could present a new therapeutic target in treating prostate cancer, partially through regulating Ca²⁺ homeostasis.

In addition to AR directly regulating both STIM and Orai, it appears androgens also regulate upstream transcriptional regulators of STIM and Orai such as WT1/EGR1. Since EGR1/WT1 play a vital role in regulation of STIM1, and thus SOCE, it is unsurprising they significantly impact prostate cancer progression. In support, examination of WT1 and EGR1 expression in several prostate cancer cell lines revealed mainly elevated WT1 expression coinciding with low EGR1 expression (Gregg, Brown, Mintz, Piontkivska, & Fraizer, 2010). However clinically, in some prostate cancer patient cohorts, EGR1 is overexpressed, and in EGR1-deficient mice, prostate tumorigenesis was impaired (Abdulkadir, Carbone, et al., 2001; Abdulkadir, Qu, et al., 2001). In this case, EGR1-induced STIM1 expression may be vital for AR activity and disease induction, while serving as an inhibitory feature of disease progression. There is also evidence this effect could be further exacerbated by faulty imbalance between EGR1 and its corepressor, NAB2 (Abdulkadir, Carbone, et al., 2001). Conversely, corroborative studies show that WT1 is important to tumor angiogenesis and cell migration in prostate cancer (Brett, Pandey, & Fraizer, 2013; Fraizer et al., 2016), as well as inhibition of STIM1 and SOCE (Go et al., 2019), which promote disease progression. Because WT1 and EGR1 work in a dual and opposing fashion to regulate STIM1 (Go et al., 2019), considered together, timing appears to be paramount as EGR1 promotes tumorigenesis and may be reliant on NAB2 expression while WT1 promotes prostate cancer progression. Although the main role of EGR1 is transcriptional regulation of STIM1, interestingly, a group studying EGR1 binding to CpG sites in prostate cancer saw that CpG methylation was inversely correlated with disease severity, and that EGR1 binding was reduced in later stages of disease (Lin et al., 2013). This presents the possibility that CpG methylation and EGR1 binding could contribute to the mechanism of how EGR1 is inversely associated with disease progression as demethylation of EGR1 CpG-binding sites could be a potential target of prostate cancer therapy to facilitate EGR1 binding to suppress progression.

SOCE as a marker for disease progression also has been substantiated clinically. STIM1 and Orai1 expression are inversely correlated with Gleason Score (Xu, Zhang, et al., 2015). Because Gleason Score is a representative marker for prostate cancer progression, this supports the model that in early stages of disease exhibiting AR+ status, STIM1 and Orai1 levels, as well as SOCE, are increased, whereas advanced AR− stages of prostate cancer exhibit lower STIM1, Orai1, and ultimately SOCE. In support, it has been further demonstrated in an androgen-sensitive cell line derived from a lymph node carcinoma of prostate cells, androgen deprivation resulted in downregulation of Orai1, and that Orai1 knockdown conferred resistance to apoptosis. This suggests that acquired resistance to apoptosis arises in part from decreased Ca²⁺ influx (Flourakis et al., 2010). Collectively, both in basic and translational research loss of SOCE appears to be a marker for disease progression.

Therapeutically, there are several promising SOCE-targeting candidates for treatment of prostate cancer. For early stages of disease, Drebrin, an actin reorganizing protein (see Table 2), may be a potential prostate cancer target. Drebrin has been shown to facilitate invasion through colocalizing with actin filaments and filopodia increasing Ca²⁺ influx (Dart et al., 2017). Although this would inhibit disease progression as AR status changes expression of STIM1 and Orai1, this is likely a key step in facilitating early stage progression of prostate cancer. A potential method to thwart Drebrin-induced invasion, is through the use of pyrazoles, such as 3,5-bistrifluoromethyl pyrazole (BTP2), inhibit SOCE. Pyrazoles are pharmacological inhibitors of Orai, and can be used to therapeutically block SOCE in a dose-dependent manner (Ohga, Takezawa, Arakida, Shimizu, & Ishikawa, 2008; Ritchie, Samakai, & Soboloff, 2012). However, not only is Drebrin associated with a variety of cancers including prostate (Dart et al., 2017) and bladder (Xu et al., 2015) among others, but knockdown of Drebrin inhibits SOCE to the same extent as inhibition by BTP2, suggesting BTP2 may inhibit Drebrin (Mercer et al., 2010). Although it has not been fully proven that BTP2 is associated with Drebrin, it would strengthen an already known connection between Ca²⁺ channels and cytoskeletal organization and integrity. Therefore, using pyrazoles could effectively act as a Drebrin antagonist, leading to lower SOCE in presumably early stages of prostate cancer including High Grade Prostate Intraepithelial Neoplasia (HGPIN), which is a precursor for prostate cancer (Brawer, 2005), or early AR+ stages. Although the distinction of Drebrin association with prostate cancer has been recently made known, this could be a potentially promising therapeutic target in early stages of prostate cancer, as well as other cancers where SOCE potentiates disease progression. For treatment of later stages of prostate cancer, in a first-in-human phase I clinical trial, mipsagargin, a prostate-specific membrane antigen-targeted SERCA inhibitor, was able to stabilize disease in patients with advanced prostate cancer and other solid tumors (Mahalingam et al., 2016). By inhibiting SERCA, Ca²⁺ stores are depleted leading to activated SOCE, a known suppressor of prostate cancer progression during AR- stages. Thus, mipsagargin is a promising therapeutic tool for late stages of prostate cancer by inducing the UPR and SOCE.

Reflecting this, SOCE plays dual and opposing roles in prostate cancer that are receptor-status dependent. However, it is intriguing that prostate cancer and breast cancer, which both use hormone-based mechanisms for disease progression demonstrate different effects resulting from changes in SOCE. Although different, these changes are stark and conclusive which underscores the importance of studying SOCE in the context of disease progression. Because SOCE plays different roles in different prostate cancer stages, it is important to develop therapies that both potentiate and attenuate SOCE, as these could make treatment for a wide range of prostate cancer cases more feasible. Although biologics targeting SOCE are beginning to emerge, efforts should continue to be made to investigate other drugs targeting SOCE, such as PTEN agonists, as well as Drebrin and SERCA inhibitors, as novel therapeutic strategies for prostate cancer.

4.3.3. TRP channels

As prostate cancer relies on receptor-mediated signaling for disease progression, similar to SOCE, TRP channels have been highly implicated in prostate cancer. In this regard, as reviewed in Hantute-Ghesquier, Haustrate, Prevarskaya, and Lehen’kyi (2018), several TRP Melanin (TRPM) channels; TRPM2, TRPM4, TRPM7, and TRPM8 are all highly associated with prostate cancer. TRPM2 has been demonstrated to induce prostate cancer cell proliferation. Likely, TRPM2 translocates to the nucleus in prostate cancer cells and utilizes its C-terminal domain, which exhibits enzymatic activity, to promote proliferation in response to ADP-ribose or thermosensation (Tan & McNaughton, 2018; Zeng et al., 2010). This is intriguing as it suggests that TRPM2 proteins can facilitate transcription of pro-proliferation genes. Although it is conclusive that TRPM2 leads to increased proliferation, the mechanism is poorly understood and warrants further investigation to draw conclusions on the overall role of TRPM2 in prostate cancer progression.

While TRPM2 may regulate proliferation, TRPM4 and TRPM7 have been shown to promote migration and invasion. Specifically, TRPM4 regulates SOCE-mediated migration as upregulation leads to depolarization of the cell, decreasing the driving force of Ca²⁺, which in turn decreases SOCE (Holzmann et al., 2015). It is likely that TRPM4 is upregulated in later stages of disease as STIM1, Orai1, and AR levels all decrease and disease is in AR− stages, which is suggested by (Holzmann et al., 2015). Concomitantly, TRPM7 also regulates migration and invasion, but through EMTby downregulating MMP9 and upregulating E-Cadherin (Chen, Chen, Chiu, Chen, & Shen, 2017). Although a specific connection between TRPM7 and SOCE in migration has not been established, it is known that TRPM7 associates with SOCE. Like TRPM2, TRPM7 has a C-terminal domain capable of kinase activity, and this kinase activity promotes SOCE through phosphorylating STIM1 or binding partners of STIM1 and Orai1 including Septins (Faouzi, Kilch, Horgen, Fleig, & Penner, 2017). TRPM7 may also regulate SOCE-mediated invasion indirectly as MMP and E-Cadherin production are reliant on Ca²⁺-flickers and many EMT pathways are receptor-mediated and Ca²⁺-dependent as discussed above. Together, TRPM4 inhibits, while TRPM7 promotes SOCE leading to differential effects on Ca²⁺ entry and invasion.

Finally, many studies have focused on TRPM8 function and expression in prostate cancer. Not only is TRPM8 significantly overexpressed in many forms of prostate cancer (Tsavaler, Shapero, Morkowski, & Laus, 2001) and important for cell survival (Zhang & Barritt, 2004), but it has the unique ability to be regulated by hormone-induced changes in cancer (Monteith et al., 2007). It has been reported that TRPM8 is regulated by androgens, and associated with SOCE in early stage of prostate cancer disease (Zhang & Barritt, 2004). Further, TRPM8 knockdown and inhibition using pharmacological inhibitor, capsazepine (see Table 2), has been shown to increase apoptosis, suggesting it plays a role in cell survival in AR+ prostate cancer. Intriguingly, menthol-induced activation of TRPM8 also increased apoptosis in AR+ prostate cancer, due to pathophysiological sustained increase in Ca²⁺ that leading to mitochondrial Ca²⁺ overload. It is proposed that menthol-induced activation leads to a and is not representative of normal channel function (Zhang & Barritt, 2004). this finding This presents menthol as an interesting treatment for prostate cancer as a method of inducing mitochondrial Ca²⁺ overload by sustained activation of the TRPM8 channel. In conclusion, TRPM8 appears to require a tightly regulated window of activity to induce cell survival as it has dual and opposing roles in apoptosis at both poles of activity. Thus, TRPM8 knockdown and pathophysiological activity could prove effective in targeting prostate cancer.

In addition to TRPM channels promoting prostate cancer progression, TRPV channels are also associated with prostate cancer as well. TRPV6, which is more Ca²⁺ permeable than its other TRPV channel counterparts, is positively correlated with Gleason Score, and levels are higher in metastatic and higher stage disease compared to benign prostatic hyperplasia (BPH) (Fixemer, Wissenbach, Flockerzi, & Bonkhoff, 2003). This is likely because TRPV6-mediated Ca²⁺ entry is required for NFAT activation (Fixemer et al., 2003; Lehen’kyi, Flourakis, Skryma, & Prevarskaya, 2007), which contributes to prostate cancer progression through NFAT targeting survival genes such as AP1, Fos, and Jun (Hogan, Chen, Nardone, & Rao, 2003). To this end, TRPV6 inhibitors (SOR-C13) have gone through phase 1 clinical trials in patients suffering from prostate cancer and have exhibited antitumor activity (Fu et al., 2017). Thus, TRPV6-induced NFAT activation is a promising, novel therapy for prostate cancer.

Considered together, TRP channels appear to have vastly different, yet important roles in prostate cancer progression. Inhibitors targeting TRP channels could be proven effective in combating prostate cancer by targeting pathways such as EMT, NFAT, and mitochondrial Ca²⁺ overload. Therefore, it is imperative that we continue to investigate molecular mechanisms and their ensuing therapeutics targeting TRP channels, as this could provide, novel and effective drugs in treating prostate cancer.

4.3.4. Purinergic receptors

Intriguingly, purinergic receptors also appear to play a role in prostate cancer. Specifically, the presence of P2X7R may be a marker of prostate cancer since this protein is not expressed in normal prostate tissue, but expressed in 100% cases regardless of Gleason Score (Slater, Danieletto, Gidley-Baird, Teh, & Barden, 2004). Although there is not much research on how P2X7R induces prostate cancer tumorigenesis or progression, a potential mechanism of progression proposed is that P2X7R facilitates ATP-driven migration and invasion, because ATP drives induction of EMT-associated genes, such as SNAIL, as well as downregulates production of epithelial markers, such as E-Cadherin (Qiu et al., 2014). Further, it has also been shown to promote phosphorylation PI3K, AKT, and ERK1/2 (Qiu et al., 2014). Although the exact mechanism of how P2X7R promotes phosphorylation of EMT-associated pathway proteins, it is likely that the mechanism is similar to that of breast cancer, where ATP is produced at a higher rate in response to anchorage deprivation. This is supported as it is proposed that P2X7R promotes ATP-driven changes in EMT pathways including changes in protein phosphorylation and gene expression (Qiu et al., 2014). Similar to breast cancer, it is possible that the channel has differential dependence on ATP depending on stage of cancer, however, this is yet to be examined. Considered collectively, the fact that P2X7R is a marker for all stages of prostate cancer illustrates the importance of investigating this channel. As a result, antagonists could play therapeutic roles in all stages of this disease, yet no currently efficacious antagonists exist.

Together, although there is profound clinical significance in the expression of P2X7R, the understanding of the molecular mechanisms behind this channel’s effect on prostate cancer progression is poorly understood. This presents an important, yet highly understudied avenue for understanding Ca²⁺ regulation of prostate cancer progression and improving prostate cancer treatments.

4.4. Melanoma

4.4.1. Introduction

Skin cancer, including melanoma as well as other forms of skin cancer, are the most common type of malignancy among Caucasians (Apalla, Nashan, Weller, & Castellsagué, 2017). Although, the most common forms of skin cancer are carcinomas including Basal Cell, Squamous Cell, and Merkel Cell, the deadliest and most challenging skin cancer to attenuate is melanoma. Melanoma is the fifth deadliest cancer among men and women (Alteri, 2019). Not only is it deadly, its incidence is rising especially in younger patients as evidenced by surpassing breast cancer as the most prevalent cancer among women under 40 (Bradford, Anderson, Purdue, Goldstein, & Tucker, 2010). Melanoma arises from neural-crest derived melanocytes, which are the pigment-producing cells found in the skin, ears, eyes, meninges, and other mucosal surfaces (Leonardi et al., 2018). This disease is normally classified in two subtypes, sporadic and familial, with sporadic being considerably more common (Schadendorf et al., 2018). Thus, the high prevalence and mortality of this disease illustrates the importance of finding novel treatments for this disease.

Melanoma, exhibits a significant sex bias, in which women are diagnosed more frequently at younger ages and men are diagnosed more frequently at older ages (Rastrelli, Tropea, Rossi, & Alaibac, 2014). There are two distinct factors that contribute to this phenomenon. First, younger women have a higher likelihood to use artificial tanning booths, which emit a UVA concentration approximately 13 times higher than peak summer sun exposure. The likelihood of melanoma is raised 16%–20% if tanning booths have ever been used; this likelihood is doubled when used before the age of 35 (Miller, Hamilton, Wester, & Cyr, 1998; Roh et al., 2017). Second, melanoma appears to be heavily regulated by hormones. For example, estrogen binding to ERβ has been shown to occur in early stages of melanomagenesis and is a marker for melanoma proliferation (de Giorgi et al., 2009; Roh et al., 2017). However ERβ expression is decreased during melanoma progression and its expression is actually inversely correlated with progression into later stages (Marzagalli et al., 2016). This is likely because ERβ induces phosphorylation of ERK1/2 to promote proliferation (Verdier-Sevrain, Yaar, Cantatore, Traish, & Gilchrest, 2004), but inhibits the PI3K/AKT pathway, thereby inhibiting invasion (Marzagalli et al., 2016). Therefore, it appears the timing of ERβ expression and function is key to determining the role it plays in melanoma progression. Further, during menopause, ERβ diminishes rapidly, leading to impaired melanomagenesis and progression for postmenopausal women (Roh et al., 2017). Whereas women tend to have decreased HR over time and use HR signaling as a protective factor against melanoma progression, the opposite is true in men. Unlike women who have diminished production of hormones over time, men produce androgens throughout their lives (Kelsey et al., 2014) and androgens promote melanomagenesis and progression (Roh et al., 2017). This is likely because not only does testosterone inhibit protective enzymes against oxidative stress like Superoxide Dismutase (SOD) and Catalase (Bokov, Ko, & Richardson, 2009), but AR directly promotes transcription of miRNA that suppress the inhibition of melanoma-inducing oncogenes, such as microphthalmia-associated transcription factor (MITF) and leads to subsequent invasion (Wang, Zhang, Yu, Yu, & Huang, 2016). Therefore, women, especially post-menopausal women, have a survival advantage over men (Roh et al., 2017). In conclusion, both behavioral factors such as frequent exposure to artificial tanning, as well as hormonal factors, such as androgen production, lead to a considerable sex bias for age at diagnosis of melanoma.

The most common cause of melanoma is UV radiation (UVR). UVR is considered a “complete carcinogen” (D’Orazio, Jarrett, Amaro-Ortiz, & Scott, 2013) and consists of long wavelength UVA (320–400nm), shorter wavelength UVB (~290–320), and shortest wavelength UVC (200–280nm), the last of which is blocked by the ozone layer and does not reach the earth (Dale Wilson, Moon, & Armstrong, 2012). UVR is well known for its role in inducing melanin production, with melanin acting as a protecting factor against UV-induced damage in skin (Brenner & Hearing, 2008). Therefore, the relative lack of melanin in Caucasians makes them the most susceptible to UV-induced DNA damage and, ultimately, melanoma (Schadendorf et al., 2015). Although about 95% of the rays absorbed in skin are UVA, UVB is more heavily associated with skin cancer as it directly induces DNA damage (D’Orazio et al., 2013). The response to UVB starts in keratinocytes; UVB-induced DNA damage leads toαMelanin-Stimulating Hormone (αMSH) productiononthe keratinocytes which binds to Melanocortin 1 Receptor (MC1R) on the melanocytes, which stimulates melanin production and whose variants lead to pale skin, red hair and increased melanoma risk (Garcia-Borron, Sanchez-Laorden, & Jimenez-Cervantes, 2005; Raimondi et al., 2008; Schadendorf et al., 2015). Melanin is then transported in melanosomes to the keratinocytes where it is positioned facing outward from the nucleus toward the sun to act as a shield against further DNA damage (Schadendorf et al., 2015). Unfortunately, each UV exposure leads to the introduction of C➔T mutations, ultimately leading to DNA damage and/or mutations that tend to promote melanomagenesis and progression. However, while it is clear that UVR is the major driver of melanomagenesis, the molecular pathways linking UVR and driver mutations is somewhat unclear since the most common mutations associated with melanoma do not have a UV signature mutation of C➔T. For example, BRAF^V600E, which is found in 50% of melanoma cases (Hodis et al., 2012; Leonardi et al., 2018; Schadendorf et al., 2018), does not consist of a C➔T mutation, and is UV-independent. Conversely, NRAS mutations tend to appear in older patients with chronic sun exposure (reviewed in Muñoz-Couselo, Adelantado, Ortiz, García, & Perez-Garcia, 2017). The most common NRAS mutations occur at G12 (G12D, G12S), G13 (G13R), and Q61 (Q61K, Q61R, Q61L, Q61V, Q61H) (Heppt et al., 2017), which are found in 15–20% of melanoma cases (Muñoz-Couselo et al., 2017), do not have an association with MC1R (Thomas et al., 2017) and are also not UV-signature mutations. Hence, despite the fact that NRAS mutations are not UV signature mutations, UVR may be required for transformation associated with NRAS mutations. The remaining 30–35% of melanoma exhibit a variety of mutations including Neurofimbrin 1 (NF1), KIT, CDKN2A, PTEN and many others (Schadendorf et al., 2018). Considered collectively, it is surprising that, despite the well-established role of UVR in melanomagenesis and progression, the underlying molecular mechanisms remain very unclear. A more fundamental understanding of the disease is needed to address this shortcoming.

4.4.2. SOCE

4.4.2.1. Src and increased SOCE

As discussed in our recent review (Hooper, Zaidi, & Soboloff, 2016), melanoma is a highly heterogeneous disease with SOCE demonstrating differential, context-dependent roles in melanoma progression. Hence, SOCE has been strongly associated with disease progression in many different tumor types (Vashisht, Trebak, & Motiani, 2015). Similar observations have been made in melanoma in some contexts. For example, it has been shown that Orai1 and STIM2 expression and activation, which lead to higher intracellular Ca²⁺ levels, can induce a shift between proliferative and migratory activity (Stanisz et al., 2014), suggesting that timing of SOCE can play an important role in melanoma disease progression. This effect is demonstrated by the fact that B16BL6 mouse melanoma cells have enhanced invasiveness resulting from higher SOCE because higher intracellular Ca²⁺ levels lead to higher levels of constitutively active PKB/AKT and increased cell survival (Feldman, Fedida-Metula, Nita, Sekler, & Fishman, 2010). Further, the effect of this signaling cascade was supported by SOCE coupling with lipid rafts to activate lipid raft-residing calmodulin and subsequent activation of protein tyrosine kinase, Src, as well as AKT (Fedida-Metula et al., 2012). Because it has been shown that tyrosine phosphorylation by Src is essential for biological function of AKT (Chen et al., 2001), it is likely that lipid raft-activation of Src phosphorylates and activates AKT in a Ca²⁺-dependent manner and leads to melanoma progression. Interestingly though, the relationship between SOCE and Src activation goes beyond just AKT activation. For example, Ca²⁺ oscillations activate Src to recruit cortactin and the adaptor protein, SH3 containing protein 2a (TKS5), to promote invadopodia-induced degradation of the ECM through actin remodeling (Sun et al., 2014), suggesting that Src activation could have dual effects on melanoma progression, both of which are Ca²⁺-dependent. The concept that SOCE drives melanoma progression is further supported by the observation that SOCE blockade through either genetic or pharmacological strategies decreases invasion both in vitro and in vivo (Umemura et al., 2014). Thus, it appears that SOCE can induce melanoma progression when melanoma is highly dependent on Src signaling.

Currently, there are two strategies for targeting melanoma in the context of suppressing melanomas reliant on enhanced SOCE. First, Pyrazoles are a class of drugs that are heterocyclic nitrogen-containing molecules that have been shown to inhibit SOCE. A commonly used pyrazole, 3,5-Bis(trifluoromethyl)pyrazole (BTP2/Pyr2) has been shown to block both Orai1-mediated SOCE and TRPC-mediated Ca²⁺ entry (He, Hewavitharana, Soboloff, Spassova, & Gill, 2005; Schleifer et al., 2012; Zitt et al., 2004) and inhibit melanoma in some contexts (Umemura et al., 2014). Although these drugs have not undergone clinical trials in melanoma patients, the potential drawback of these drugs is globally inhibiting SOCE, could have off-target effects on melanoma that have not yet been examined. Second, BRAF inhibitors, such as Vemurafenib and Trametinib, are currently a widespread therapeutic for melanomas, but a major drawback is that resistance is acquired in approximately 50% patients after 6 months (Mackiewicz & Mackiewicz, 2018). This resistance occurs mainly because of reactivation of the MAPK pathway. In this regard, Src activation has been shown to be an important tyrosine kinase that can reactivate the MAPK pathway, which contrives Src activation important in these melanoma types (Girotti et al., 2015). In this regard, the most commonly Src-activated melanomas are those exhibiting Src pY416, for which inhibitors have been generated (Homsi et al., 2009), and have undergone clinical trials (Elias & Ditzel, 2015). However, clinical trials have been moderately effective at best only providing clinical benefit to less than a quarter of patients (Elias & Ditzel, 2015). Novel Src inhibitors continue to be examined (Halaban et al., 2019), and could perhaps be more effective than previous generations of Src inhibitors. However, a new strategy for treating melanoma has been the development of inhibitors that have dual roles as pan-RAF inhibitors (including targeting of BRAF) and Src inhibitors (CCT196969 and CCT241161) (Girotti et al., 2015). Moreover, these inhibitors appear to be effective in both cells harboring BRAF and NRAS mutations (Girotti et al., 2015), which broadens the candidacy of these therapeutics. Clinical trials assessing the efficacy of these drugs were planned for 2015 (Girotti et al., 2015), but there is no update on results from these studies. Yet, this offers a novel and exciting strategy to combat melanomas in specific contexts, and may present fewer off-target effects than global SOCE inhibition via Pyrazoles. Thus, administering a dual inhibitor for BRAF and Src could provide a method to permanently turn off the MAPK pathway and prevent resistance in BRAF mutant melanomas that are highly dependent on Src signaling.

Considered together, some subsets of melanoma rely heavily on enhanced intracellular Ca²⁺ levels and SOCE. SOCE can act not only as a dynamic switch of cellular activity, but also functions as an important mediator for cell signaling that enhances cell survival and drug resistance. As melanoma is heterogeneous, it is important to understand which subtypes of melanoma exhibit reliance on enhanced SOCE. Characterizing a profile of melanoma subtypes that rely on enhanced SOCE for progression could increase the efficacy of targeting SOCE as a novel therapeutic and could improve personalizing medicine for patients based on the genetic background of their melanomas.

4.4.2.2. Wnt5a and decreased SOCE

As noted above, melanoma is highly heterogeneous; in some contexts, the role of SOCE in melanoma is fundamentally changed. Wnt5a is an oncogene strongly associated with aggressive melanoma (Asem, Buechler, Wates, Miller, & Stack, 2016; Webster et al., 2015; Weeraratna et al., 2002). In a recent study from our group, we observed that highly invasive Wnt5a-expressing human melanoma cells exhibited marked SOCE suppression (Hooper et al., 2015). Further, it was demonstrated that knockdown of Wnt5a in invasive cell lines raised SOCE and inhibited invasiveness, while overexpressing Wnt5a in non-invasive cells had the inverse effect (Hooper et al., 2015). This effect was mediated by PKC, which suppresses SOCE via phosphorylation of Orai1 S27 and S30 (Hooper et al., 2015; Kawasaki, Ueyama, Lange, Feske, & Saito, 2010), which provides a potential mechanism for how Wnt5a induces SOCE suppression. Since PKC appears to be the mediator between Wnt5a and SOCE suppression, it is logical to associate PKC with invasive melanomas that rely on suppressed SOCE. This logical assumption is supported by that fact that PKC is associated with melanoma progression in a context-specific manner, that is discussed extensively in Denning (2012). Although PKC facilitates melanoma progression in a context-specific manner, it is important to note this occurs across many different types of melanoma including BRAF and NRAS mutant subsets. Interestingly, PKC has been shown to negatively regulate AKT (Doornbos et al., 1999), which as described previously facilitates progression in melanomas reliant on enhanced SOCE. This puts considerable intrigue into the relationship between AKT and PKC, which could be significant in understanding how Ca²⁺ regulation affects melanoma. Because melanoma is highly heterogeneous with SOCE playing opposing context-dependent roles depending on the subtype, AKT and PKC could provide a biological switch that each melanoma utilizes differently depending on the genetic mutations of the melanoma. Although the relationship between AKT/PKC and common genetic or UV-induced mutations have not been examined, it has been previously suggested that targeting PKC could be effective for NRAS mutant melanomas (Takashima et al., 2014), which provides insight into where to commence these investigations. Therefore, elucidating and understanding the role of AKT and PKC in melanoma progression and their relationships with common genetic mutations could shed light on molecular mechanisms underlying the context-specific roles of SOCE in melanoma progression. Understanding these mechanisms would not only be instrumental in clarifying some of the heterogeneity in melanomas, but also improve targeting of specific Ca²⁺-dependent pathways in specific melanoma contexts as a novel therapeutic strategy across different melanomas.

The heterogeneity of melanoma makes generating treatments extremely challenging, which is illustrated by the fact that currently there are minimal therapeutic strategies targeting Ca²⁺ that exist for melanoma. It has been discussed earlier that melanomas reliant on AKT and enhanced SOCE could benefit from treatment with BRAF/Src dual inhibitors, but treatments for melanomas reliant on PKC and suppressed SOCE must also be addressed. Although treatments such as previously aforementioned PTEN-upregulating molecules, as well as Drebrin, would all theoretically increase SOCE and inhibit melanoma progression in melanomas that are Wnt5a-positive and rely on suppressed SOCE (Hooper et al., 2015), the most promising treatments are Wnt5a biologics. Although most Wnt-targeting molecules target the canonical pathway (Wnt3a), other biologics targeting the non-canonical pathway (Wnt5a) are currently being investigated. For example, Box5 is a hexapeptide containing a t-butyloxide carbonyl developed by the laboratory of Tommy Andersson that mimics a fragment of Wnt5a ( Jenei et al., 2009). This peptide antagonizes the effect of Wnt5a by binding to the frizzled 5 GPCR (FZD5), which directly inhibits Wnt5a activation of PKC. This molecule has been developed as a Wnt5a antagonist in metastatic melanoma ( Jenei et al., 2009), and has been shown to inhibit both migration and invasion of melanoma cells in vitro. Although this molecule is a derivative of Foxy5, an N-formyl group-containing small molecule agonist of Wnt5a that has undergone clinical trials for cancers reliant on enhanced SOCE, such as breast cancer (Blagodatski, Poteryaev, & Katanaev, 2014), the effects of Box5 have not yet been examined in vivo and this molecule has not undergone clinical trials. Other Wnt5a biologics have been developed, which mainly act as Porcupine inhibitors. Porcupine is a membranebound O-acetyltransferase that facilitates acylation of Wnt molecules, which is required for Wnt secretion (Takada et al., 2006). A specific example of this within industry is LGK974 (Novartis), which is currently undergoing clinical trials for melanoma (Blagodatski et al., 2014). Together, there is promise in targeting Wnt5a as a novel class of therapeutics that could be effective in melanomas reliant on suppressed SOCE.

In conclusion, because melanomas are highly heterogeneous, conversely to melanomas that utilize AKT and enhanced SOCE, some melanomas utilize Wnt5a/PKC and suppressed SOCE for progression. This necessitates discovering and generating novel drugs that enhance SOCE to combat melanoma in some contexts. The most effective way to target these melanoma subtypes is to explore drugs that inhibit the non-canonical Wnt5a pathway, especially Wnt5a itself. Currently drugs targeting this pathway are still being explored. Although understanding of the in vivo effects of Wnt5a antagonists like Box5 leave more to be desired, there is promising progression of novel therapies inhibiting Wnt5a secretion (Porcupine inhibitors) and activity (Box5) ongoing, which offers exciting new opportunities to treating Wnt5a-expressing melanomas reliant on lower SOCE for progression.

4.4.3. TRP channels

Melastatin, the protein defining the TRPM channel family, is produced on the N-terminal side of the TRP protein (Bae, Jara-Oseguera, & Swartz, 2018), was originally discovered in melanocytes, and was shown to be downregulated in melanoma cells compared to melanocytes (Duncan et al., 1998). Hence, melastatin has been characterized as a tumor suppressor in melanoma (Xu, Moebius, Gill, & Montell, 2001) and the potential involvement of the TRPM channel family in melanoma has been carefully investigated. For example, TRPM1, the founding member of the TRPM channel family (Guo, Carlson, & Slominski, 2012) is heavily expressed in normal human epidermal melanocytes, however, its expression is lost as cells progress through the radial and vertical growth phases; TRPM1 expression is ablated in metastatic melanoma (Duncan et al., 1998). As a result of these findings, TRPM1 expression is a therapeutic marker for disease-free survival (Guo et al., 2012). Mechanistically, it is proposed that melastatin may facilitate differentiation of neural crest stem cells into melanocytes (Duncan et al., 1998). However, it is also feasible that expression of TRPM1-mediated Ca²⁺ entry counteracts Wnt5a-mediated SOCE suppression, thereby attenuating the progression of Wnt5a-expressing, aggressive melanomas. Consistent with this concept, TRPM1 expression and intracellular Ca²⁺ levels are both lowered in rapidly dividing melanocytes, compared to slower growing and differentiated melanocytes. This is likely because induction of p53 by ultraviolet B (UVB) leads to inhibited TRPM1 expression and Ca²⁺ influx (Devi et al., 2009). Therefore, a potential alternative mechanism by which UV, and specifically UVB, induces melanomagenesis is through suppression of TRPM1 expression and Ca²⁺ influx, whereas TRPM1 inhibits this process by slowing melanocyte growth and promoting differentiation. Although Wnt5a status was unknown in these investigations, it is likely that rapidly developing melanocytes exhibited high Wnt5a as they transform and become malignant melanoma cells. Overall, activation of TRPM1, and its subsequent promotion of Ca²⁺ influx, leads to tumor suppressive roles such as promotion of differentiation and suppression of growth and proliferation in melanoma cells.

Currently, little is known about what specifically activates TRPM1 (Zholos, 2010), especially in melanoma, although it is likely activated by GPCRs (Xu et al., 2016). One potential mechanism is TRPM1 is kept closed by metabotropic glutamate receptor 6 (mGlur6), which is a GPCR predominantly located on retinal cells (Badheka et al., 2017). In support, a specific subset of melanoma patients exhibit melanoma-associated retinopathy (MAR) and TRPM1 is the likely link between melanoma and MAR (Morgans, Brown, & Duvoisin, 2010). Additionally, it has been previously shown that mGlur6 antagonists transiently open TRPM1 channels in mouse retinal cells (Shen, Rampino, Carroll, & Nawy, 2012), therefore it is likely mGlur6 is the predominant regulator of TRPM1 in melanoma. Once it is clearly determined how TRPM1 is activated, TRPM1 agonists, such as potentially mGlur6 antagonists, could be a promising candidate for novel therapeutic treatments for melanoma as a way to slow progression and potentially induce differentiation of melanocytes.

Besides TRPM1, other TRPM channels such as TRPM7 and TRPM8, as well as TRPV1, have been demonstrated to play a role in melanoma progression through various mechanisms. First, TRPM7 can act as a tumor suppressor in melanoma by preventing the accumulation of cytotoxic intermediates of melanin synthesis in melanophores (fish equivalent of melanocytes) (McNeill et al., 2007) leading to the detoxification of melanocytes (Chen et al., 2014). In addition, TRPM7 has the capacity, through Ca²⁺ influx, to activate Ca²⁺-dependent protease, m-calpain, (McNeill et al., 2007), which inhibits focal adhesion (Su et al., 2006), and in turn likely leads to anoikis. In this regard, TRPM7 agonists have been identified (Chubanov, Schäfer, Ferioli, & Gudermann, 2014; Hofmann et al., 2014), but have not undergone clinical trials. Thus, TRPM7 remains a promising therapeutic target in the treatment of melanoma. Second, TRPM8 has also been demonstrated to act as a tumor suppressor in melanoma. This is demonstrated by the fact that menthol has been shown to inhibit melanoma cell viability in a dose-dependent manner through activation of TRPM8, leading to increased Ca²⁺ influx (Yamamura, Ugawa, Ueda, Morita, & Shimada, 2008). This increased Ca²⁺ influx has been shown to cause Ca²⁺-dependent cell death in melanoma cells (Kijpornyongpan, Sereemaspun, & Chanchao, 2014), an effect that intriguingly also occurred in capsaicin-induced TRPV1 activation. Although capsaicin-induced TRPV1 activation leads to a similar raise in intracellular Ca²⁺, the downstream effects are slightly different. Whereas TRPM8 activation can suppress cell viability and has been proposed to be an effect of Ca²⁺-dependent inhibition of cell cycle progression (Yamamura et al., 2008), TRPV1-induced Ca²⁺ entry leads to activation of calcineurin, which suppresses expression of p53 transcriptional repressor, ATF3, subsequently activating p53, and inducing programmed cell death (Yang et al., 2018). Therefore, different TRP channels activate different Ca²⁺-dependent pathways, even if they all induce similar increases in intracellular Ca²⁺ entry. In this regard, biologics targeting TRPM1, TRPM7, TRPM8, and TRPV1 are currently being investigated (Szallasi, 2015), and could provide novel therapeutics preventing melanoma progression, through activation of different Ca²⁺-dependent pathways that attenuate melanoma. However, collectively, it appears that TRP channels play many roles, mainly as tumor suppressors in preventing melanoma progression, and should continue to be further investigated as potential melanoma therapeutics.

4.4.4. Melanoma treatments

Recently, melanoma treatments have been revolutionized by the emergence of immunotherapies. Currently, there are two major approaches to treating metastatic melanoma which include targeted therapies and immunotherapies (Neves de Oliveira, Dalmaz, & Zeidan-Chulia, 2018). Immunotherapies are an exciting new class of therapeutics that were designed based on the rising melanoma incidence observed in immunosuppressed patients (Greene, Young, & Clark, 1981). Immune checkpoint inhibitors targeting Cytotoxic T-Lymphocyte Antigen 4 (CTLA4) (ipilimumab) and Programmed Cell Death-1 (PD1) (Pembrolizumab) have shown significant promise and are currently being tested in clinical trials in conjunction with targeted therapies. Targeted therapies include inhibitors of the Mitogen-Activated Protein Kinase (MAPK) pathway such as Vemurafenib, which specifically targets the oncogenic BRAF^V600E mutant () and the MEK inhibitor Trametinib (Ryu et al., 2017; Yang et al., 2010). Unfortunately, this strategy has demonstrated only short-term success thus far due to drug resistance and relapse through reactivation of MAPK or PI3K/AKT pathways, which can lead to decreased immune function or epigenetic changes (reviewed in Kakadia et al., 2018). Interestingly, Vemurafenib has been shown to increase intracellular cytosolic Ca²⁺ levels coincident with initiation of an ER stress response (Bald et al., 2014), implying that ER Ca²⁺ release may contribute to its effectiveness. It is possible that this transient effect occurs because raising intracellular Ca²⁺ inhibits progression transiently followed by relapse in response to eventual Ca²⁺-dependent PI3K/AKT activation in conjunction with immune suppression, important factors leading to melanoma progression. Considered in combination with the observations above, these observations highlight the potential significance of investigating Ca²⁺ signaling in the design of future therapeutic strategies for the treatment of melanoma.

4.5. Glioblastoma

4.5.1. Introduction

Glioblastoma Multiforme (GBM) is the most common and aggressive brain malignancy and is characterized as a form of astrocytic tumor (Hanif, Muzaffar, Perveen, Malhi, & Simjee, 2017), arising from astrocytes, which are a subtype of glial cells, the most common cell type in the brain ( Jäkel & Dimou, 2017). As reviewed in Alcantara Llaguno and Parada (2016), these astrocytes form as a result of Neural Stem Cell (NSC) differentiation that undergoes many intermediate steps including the formation of Multipotent Progenitors (MPPs) and Bipotential Progenitors (BPPs) (Alcantara Llaguno & Parada, 2016). It is important to note that GBM is highly invasive and can spread throughout the brain. This occurs because not only does GBM exhibit high tumor heterogeneity, which often leads to recurrence (Shergalis, Bankhead, Luesakul, Muangsin, & Neamati, 2018), but during surgical removal, it is challenging to remove surrounding tissue. Therefore, it is paramount to find druggable targets to treat GBM.

A significant challenge in treating GBM is generating drugs that cross the Blood-Brain Barrier (BBB). The BBB consists of endothelial cells that form adherens and tight junctions (AJ and TJ, respectively). On the luminal side of this barrier, endothelial cells possess efflux pumps such as ABC transporters that prevent influx of lipophilic molecules, toxic reagents and many drugs (Fletcher, Williams, Henderson, Norris, & Haber, 2016; Shergalis et al., 2018). This severely limits the treatment options for GBM, contributing to the bleak therapeutic outcomes of this disease. As reviewed in Shergalis et al. (2018), the current standard of care is maximal surgical resection followed by post-surgical treatment using alkylating agents such as temozolomide (which can cross the BBB (Kelly, Sacapano, Prosolovich, Ong, & Black, 2005)) in conjunction with radiation (Stupp et al., 2005). Unfortunately, only ~10% of GBM patients survive 2 years on this regimen (Perry et al., 2017). Immunotherapies are among the potential future strategies; the checkpoint inhibitors targeting Programmed Death 1 (PD-1), Programmed Death Ligand 1 (PD-L1) and Cytotoxic T-Lymphocyte-associated Antigen 4 (CTLA4) are all being targeted in clinical trials for recurrent GBM (Davis, 2016). However, the most promising targeted therapies consist of classes of GBM treatments targeting anti-angiogenesis, since not only is GBM among the most vascularized tumors (Fernandes et al., 2017), but VEGF is a major mediator of vascularization and has been linked to GBM progression (Cheng et al., 1996). This can be therapeutically targeted with monoclonal antibodies, such as bevacizumab, an approach that has shown itself to be particularly effective in patients with recurrent GBM (Gilbert et al., 2014; Stark-Vance, 2005). It appears that almost all GBM cases become recurrent, which demonstrates the lack of effectiveness in these targets in achieving prevention of relapse in primary GBM.

4.5.2. TRP channels

Recent investigations have revealed strong links between members of the TRPC, TRPV, and TRPM channel families and GBM (Alptekin et al., 2015). As GBM is a highly invasive cancer of the brain, there are many factors to consider when generating novel treatments for this disease. For example, not only do treatments have to effectively attenuate disease progression, they also have to cross the BBB. TRP channels have been demonstrated to play a role in both GBM progression and maintaining the strength and integrity of the BBB (Brown, Wu, Hicks, & O’Neil, 2008), which gives them tremendous potential as a novel therapeutic target in treating GBM. The currently defined progress of characterizing roles and potential treatments targeting these channels will be discussed in great detail below.

TRPC channels have been demonstrated to play an important role in GBM progression as TRPC1, TRPC3, TRPC5, and TRPC6 have been shown to be upregulated in GBM (Bomben & Sontheimer, 2008). These TRPC channels utilize several different mechanisms driving GBM progression. For example, TRPC1-mediated Ca²⁺ entry facilitates motility through changes in the shape and volume of the cell driven by CAMKII-induced activation of Cl⁻ channels (Cuddapah, Turner, & Sontheimer, 2013). Likewise, TRPC3 also enhances motility, but also promotes invasion. In the context of motility, TRPC3-mediated Ca²⁺ entry drives focal adhesion formation through MLCK phosphorylation of MLC (Chang, Cheng, Tsai, Tsao, & Chen, 2018). In regard to invasion, TRPC3-mediated Ca²⁺ entry facilitates activate NFκB (see Fig. 3), leading to transcription of fibronectin, which facilitates wound contraction and subsequent scar contracture, ultimately resulting in degradation of the ECM. Finally, TRPC6 is associated with GBM progression through Notch1, which is involved in development and promotion of tumor growth in a variety of tumors, (Chigurupati et al., 2010; Gao et al., 2012; Zhang, Li, Ji, & Zheng, 2010). Under hypoxic conditions increased Notch1 has been demonstrated to upregulate TRPC6 expression, whose ensuing Ca²⁺ response drives NFAT activity and subsequent GBM tumor growth and invasion both in vitro and in vivo (Chigurupati et al., 2010). While TRPC-induction of NFAT facilitates tumor growth likely through activation of oncogenic survival pathways such as MAPK (discussed above), GBM invasion can be promoted by both TRPC-mediated Ca²⁺ entry promoting actin-myosin rearrangements to disassemble cell-substratum adhesions (Chigurupati et al., 2010) and NFAT activity upregulating the production of MMP7 and MMP9 (Tie, Han, Meng, Wang, & Wu, 2013). Thus, as these channels facilitate cell motility, growth, and invasion, targeting the inhibition of TRPC channels has potential to be a promising therapeutic in GBM.

One challenge in the design of a therapeutic strategy targeting TRPC channels is that they have effects on both GBM progression and BBB integrity (Brown et al., 2008). Hence, TRPC-mediated Ca²⁺ entry in BBB endothelial cells causes phosphorylation of MLCs by MLCKs, leading to myosin-actin contractions and disruption of the BBB, an effect that was inhibited by pan-TRPC antagonist, SKF-96365 (Hicks, O’Neil, Dubinsky, & Brown, 2010). Although the TRPC channel profile of human BBB endothelial cells is not well-defined, experiments performed in mice and rats reveal TRPC1, TRPC4,and TRPC7 as the major family members expressed (Berrout, Jin,& O’Neil, 2012; Brown et al., 2008). Although it would be ideal for a TRPC antagonist to readily cross the BBB (Pyr3 (see Table 2) does not readily cross the BBB (Munakata et al., 2013)), to circumvent this issue, three options exist. First, investigations into finding a therapeutic window to slightly disrupt, and facilitate drug delivery across, the BBB may be possible, provided cranial swelling could be controlled. Second, it is possible to do intracerebroventricular injection of a TRPC antagonist directly into the affected brain tissue, an approach that showed success in mice (Munakata et al., 2013), but would come with clinical risks to patients. Finally, further investigations into generating TRPC antagonists that can cross the BBB are needed to generate novel therapeutic targets for GBM. In conclusion, generation of a TRPC antagonist that can cross the BBB could add tremendous potential to the therapeutic landscape as a novel strategy to treat GBM.

In addition to TRPC channels exerting effects on GBM progression, TRPV channels also have been shown to have altered expression and function in GBM. This is portrayed by a shift in TRPV1 and TRPV2 expression occurring during transformation into GBM. Hence, whereas astrocytes express TRPV2, this expression is lost in GBM, with TRPV1 upregulated in its place (Morelli et al., 2012; Nabissi, Morelli, Santoni, & Santoni, 2012; Nilius & Szallasi, 2014; Stock et al., 2012). Interestingly, forcing TRPV2 expression in GBM causes growth inhibition in vivo by promoting differentiation into mature glial cells (Morelli et al., 2012; Nabissi et al., 2012; Nilius & Szallasi, 2014). If TRPV1 is overexpressed in GBM it results in apoptosis, illustrating the considerable Ca²⁺ sensitivity GBM has (Amantini et al., 2007; Stock et al., 2012). However, it is important to note that TRPV1 mRNA transcripts have 4 variations within the 5′ UTR due to an alternative first exon (TRPV1_v0, TRPV1_v1, TRPV1_v3, and TRPV1_v4). Although all four of these variations encode for the same protein, GBM primarily expresses only TRPV1_v3, which may have specific features that optimally promote GBM progression without inducing apoptosis (Nabissi et al., 2016). Likely this includes the possession of a gamma interferon activated inhibitor of ceruloplasmin mRNA translation (GAIT) element and 5’ UTR stability. TRPV1_v0, TRPV1_v3, and TRPV1_v4 variants all have a GAIT element that binds to IFN-y, which normally reduces TRPV1 protein levels (Nabissi et al., 2016). Likely, these reduced TRPV1 levels lead to enhanced GBM progression as TRPV1 is reduced enough to avoid apoptosis, while levels are high enough to inhibit expression of TRPV2. Although multiple variants express this GAIT element, TRPV1_v3 has been demonstrated to be the most stable and structured 5’ UTR sequence (Nabissi et al., 2016), a factor that enhances translation efficiency of many protooncogenes or tumor suppressor genes in a variety of cancers (Audic & Hartley, 2004). Therefore, TRPV1_v3 is the preferred 5′ UTR variant of TRPV1 expressed in GBM. Still, the fact that high levels of TRPV1 can induce apoptosis in GBM has led some investigators to assess the potential of the TRPV1 agonist capsaicin as an anti-GBM therapeutic agent, with promising results (Amantini et al., 2007). Considered collectively, the fact that Ca²⁺ signals generated by either TRPV1 or TRPV2 attenuates GBM suggests a potential therapeutic strategy, the potential of which may become apparent as future investigations develop these early findings further.

Finally, dysregulation of TRPM channel expression in GBM has also been reported with TRPM8 the most upregulated (Alptekin et al., 2015). In this regard, it has been shown that heterologous TRPM8 expression makes cells resistant to radiation (Klumpp et al., 2017). Normally, radiation leads to G2/M arrest through CAMKII-mediated cdc25c inhibitory phosphorylation, which leads to inhibited CDK1 phosphorylation. However, TRPM8 activity overrides radiation-induced CDK1 inactivation through a feedback loop involving potassium channel activation and CAMKII inhibition, ultimately leading to cdc25c and CDK1 phosphorylation which enhances tumor promotion in vitro (Klumpp et al., 2017). Similar observations have been made of ionizing radiation inhibiting GBM progression in vivo (Klumpp et al., 2017), although, an analogous study examining TRPM8 overexpression in vivo has not been established, it is likely the same effect would occur in vivo. Interestingly, TRPM7 is also moderately upregulated in GBM (Alptekin et al., 2015), functioning primarily as a tumor promoter, as evidenced by the ability of Naltriben, a δ opioid and highly selective TRPM7 agonist, to enhance migration, invasion, and Ca²⁺ influx of GBM cells in vitro (Wong, Turlova, Feng, Rutka, & Sun, 2017). Further, inhibition of TRPM7 by Calvacrol attenuates invasion and migration by blocking MAPK pathways in GBM cells (Chen et al., 2015). Hence, both TRPM7 and TRPM8 activity promote GBM progression, unlike TRPV1 and TRPV2 activity, which drove apoptosis and differentiation, respectively. Whether these differences in responses between different types of channels reflects context-dependent activity, total activity or unique microdomains associated with the channels themselves remains unclear. Nevertheless, these observations illustrate both the inherent challenge in attempting to define the relationship between Ca²⁺ and cancer and its tremendous potential. Thus, these investigations reveal multiple classes of TRP channels as promising and surprisingly specific targets for GBM.

Among TRPM channels, several biologics have been generated against TRPM8 which could have promise as GBM therapeutics, particularly in combination with ionizing radiation. The most clinically promising drug was produced by Pfizer, called PF-05105679, which advanced to clinical trials based on successfully inhibiting the cold pressor response (Weyer & Lehto, 2017), which is the increase in blood pressure and heart rate in response to exposure to cold stimuli (Silverthorn & Michael, 2013). However, its short half-life combined with painful hot sensations experienced by participants limited its therapeutic efficacy (Weyer & Lehto, 2017). Although no other TRPM8 inhibitors have been generated, a class of drugs derived from Cannabis, phytocannabinoids, can not only act as TRPM8 antagonists, but also can activate TRPV1 and TRPV2 by sensitizing and lowering the activation thresholds of these channels (De Petrocellis et al., 2011; Gautier et al., 2014), and have the ability to cross the BBB (Cabral & Jamerson, 2014). Although these drugs have not gone to clinical trials in treatment of GBM, this presents an exciting opportunity to create a novel therapeutic strategy for GBM. Together, it seems drugs targeting TRP channels could provide a multifaceted strategy to improving the therapeutic landscape of GBM and further investigations into the ability to target these channels is warranted.

4.5.3. SOCE

There is considerable published evidence that GBM is highly dependent on Ca²⁺ signaling in general and SOCE specifically. For example, it has been shown that, compared to human primary astrocytes, malignant GBM cells derived from GBM patient explants exhibited increased SOCE and Ca²⁺-Release Activated Ca²⁺ (CRAC) current (Motiani et al., 2013). Further, it was shown siRNA-mediated knockdown of STIM1 and Orai1 inhibited invasion in vitro based on a matrigel transwell invasion assay (Motiani, Hyzinski-Garcia, et al., 2013). These and other prior studies illustrate the potential of regulating SOCE as a therapeutic target for GBM (Liu, Hughes, Rollins, Chen, & Perkins, 2011; Liu, Yao, Mercola, & Adamson, 2000).

Considering the role of SOCE in GBM invasiveness, one might predict that EGR1 would also promote GBM invasion since it drives STIM1 expression (Go et al., 2019). Consistent with this concept, gene expression profiles show that both STIM1 (Scrideli et al., 2008) and STIM2 (Ruano et al., 2006) are upregulated in primary GBM. Interestingly, Cyclin D1 transcription has been shown to be EGR1-dependent (Chen et al., 2017), which drives cell cycle progression while shRNA-mediated STIM1 knockdown also led to cell cycle arrest at G0/G1 in GBM (Li et al., 2013); the implication is that STIM1 is required for EGR1-induced cell cycle progression. This is further supported by previous observations that SOCE is required for CDK2 phosphorylation, driving the transition from G1 to S phase (Chen, Chen, Chen, Chiu, & Shen, 2016). Additionally, EGR1 drives the expression of both EGFR and PDGFαR, both of which are associated with enhanced cell motility and metastasis (Liu et al., 2000), as reported in multiple GBM cell lines (Liu et al., 2000). Since both receptors are PLC-linked, it is highly likely that SOCE contributes to these effects, particularly considering the established role of Ca²⁺ in cell motility (see Section 3.1). Future investigations delineating the relative roles of EGR1, STIM1 and Orai1 in GBM proliferation and invasion may be informative.

Interestingly, EGR1 was originally identified as a tumor suppressor in GBM; NMDA-induced EGR1 expression was decreased in primary GBM, and was associated with decreased patient survival (Mittelbronn et al., 2009). Further, WT1 expression is increased in GBM, which supports GBM proliferation (Hashiba et al., 2007), and can be a marker for distinguishing between astrocytoma and normal astrocytic cells (Schittenhelm, Mittelbronn, Nguyen, Meyermann, & Beschorner, 2008). This unusual dichotomy is likely explained by differences in other factors, such as p53 (Calogero et al., 2001), the presence or absence of which may determine whether EGR1 functions as a tumor promoter or a tumor suppressor in this cancer type. Further, whereas WT1 is a marker for grades II and III GBM, it is not present in recurrent GBM (Rauscher et al., 2014). Hence, the timing of their expression may ultimately determine their roles with WT1 promoting initial tumorigenesis and EGR1 driving a metastatic phenotype. If so, then the introduction of WT1 in recurrent GBM or EGR1 to primary GBM would be expected to drive opposing processes. Since EGR1 and WT1 also have mutually opposing roles in STIM1 transcription, differences in SOCE are likely in GBM from different stages, although the extent to which that is correct has yet to be established.

Imidazoles (see Table 2) are a class of antimycotics used as inhibitors of SOCE channels, both in vitro and in vivo (Chung, McDonald, & Gardner, 1994). Preliminary investigations using imidazoles have been performed as proof-of-principle regarding the potential of SOCE as a therapeutic target for GBM. SKF-96365 (1-[2-(4-methoxyphenyl)-2-[3-(4-methoxyphenyl) propoxy]ethyl-1H-imidazole) blocks cell proliferation and SOCE in GBM cells in vitro (Liu et al., 2011; Song, Chen, & Yu, 2014). Although in vivo studies have demonstrated effectiveness of this drug on metastasis in breast cancer (Yang, Cao, et al., 2009), some hurdles exist in GBM treatment. For example, SKF-96365 not only blocks Orai1, but also is a pan-TRPC antagonist as well (Song et al., 2014). Because TRPC-mediated Ca²⁺ influx leads to myosin-actin contractions on the cells of the BBB (Brown Rachel & Davis Thomas, 2002), SKF-93635 inhibits this effect (Brown et al., 2008), thereby strengthening BBB integrity. As such, future investigations using more specific SOCE inhibitors are needed to assess the viability of SOCE as a GBM target in vivo. By understanding in vivo effects, further development of drugs targeting SOCE can be generated to treat GBM.

Overall, it is notable that, GBM is highly Ca²⁺-dependent. Not only does Ca²⁺ homeostasis regulate BBB integrity, but also drives GBM progression. The therapeutic landscape is currently very bleak as it is difficult to effectively treat a highly heterogeneous tumor without flexibility in removing surrounding tissue to prevent recurrence. With that said, there are many Ca²⁺-modulating therapeutic pathways that warrant further investigation and therapeutic strategies to further target to create improved approaches for treating GBM.

4.6. Myeloid leukemia

Leukemia is a group of cancers that arise from white blood cells generated from the bone marrow. Leukemia is stratified by the precursor cells from which it arises, which include lymphocytes and myeloid cells. Lymphocytic Leukemia derived from B-cells (and also sometimes T-cells), are broken down into chronic (CLL) and acute (ALL). In parallel, myeloid-derived Leukemia arising from the bone marrow are also broken up into chronic (CML) and acute (AML) reviewed in Kawamoto and Minato (2004). The role of Ca²⁺ in the progression of myeloid-derived tumors is much more extensively described, which will be the focus of this section. Normally, leukocytes undergo multiple steps of maturation starting as Hematopoietic Stem Cells (HSCs) and later differentiating into mature immune cells. HSCs are mainly present in the bone marrow where they differentiate into multipotent progenitors (MPPs) and gradually lose their potential for differentiating into different lineages. Specifically, HSCs exist in two populations. Long-term HSCs (LT-HSCs) have life-long self-renewal capabilities and contribute to multiple lineages upon necessity, while Short-term HSCs (ST-HSCs) have limited self-renewal capabilities, but can give rise to MPPs (reviewed in Kondo, 2010). In fact, LT-HSCs can also differentiate into ST-HSCs (Lai & Kondo, 2008; Morrison, Wandycz, Hemmati, Wright, & Weissman, 1997). Subsequently, MPPs differentiate into either Common Lymphoid Progenitors (CLPs) or Common Myeloid Progenitors (CMPs).

CMP production is dependent on expression of two genes during MPP differentiation: FMS-like Tyrosine Kinase 3 (Flt3) and Vascular Cell Adhesion Molecule 1 (VCAM-1). During differentiation, MPPs undergo lineage-restricted differentiation, that are broken into three subgroups based on expression (Flt3^loVCAM-1⁺, Flt3^hiVCAM-1⁺, and Flt3^hiVCAM-1⁻) (Kondo, 2010; Lai & Kondo, 2006; Lai, Lin, & Kondo, 2005). Flt3^loVCAM-1⁺ subgroup has full multipotent potential, marking it the “classical MPP” that can subsequently generate CMPs, whereas Flt3^hiVCAM-1⁺ cells typically precede differentiation into granulocyte/macrophage lineage-restricted progenitors (GMPs) or lymphocytes and Flt3^hiVCAM-1⁻ cells lymphocytic progenitors (Lai & Kondo, 2006). Therefore, it is likely that canonical CMP generation proceeds through Flt3^lo selection, while non-canonical proceeds through Flt3^hi selection. Upon differentiation of MPPs into CMPs, CMPs differentiate into myeloid cells. However, aberrations in the differentiation process can lead to myelopathies including myeloid leukemia, which is the focus of this section.

Flt3 expression has been closely linked with AML progression. Hence, Flt3 expression is normally lost normal myeloid cell development, but expressed in most AML and some CML patients (Daver, Schlenk, Russell, & Levis, 2019; Small, 2006). Further, roughly one-third of AML patients exhibit mutations in Flt3 which include constitutively activating internal tandem repeats (Flt3-ITD; 25% of AML patients) and D835 and I836 point mutations of the kinase domain (Flt3-TKD; 7%–10% of AML patients) (Daver et al., 2019; Small, 2006). Although both mutations appear to be important, Flt3-ITD is more prevalent and of more clinical interest. In this regard, there is a stark contrast in patient survival in those with and without Flt3-ITD mutations (7%–44% respectively) (Meshinchi et al., 2001). Mechanistically, Flt3 mutations function through RTK signaling by facilitating downstream STAT3 pathway, which stimulates proliferation and survival (reviewed in De Kouchkovsky & Abdul-Hay, 2016). Clinically, AML progression has poorer prognosis in response to higher allelic ratio of mutant to wild type Flt3 and larger size of the ITD (Daver et al., 2019). Furthermore, it has also been found that Flt3 mutations can be acquired during relapse when no original mutation was present but 75% of patients that originally have the mutation maintain the mutation through relapse (Daver et al., 2019). Hence, Flt3 mutations, especially Flt3-ITD, are important to AML progression and have been a widespread target for therapies, which will be discussed below.

To treat AML, most patients start with induction therapy to achieve complete remission. Induction therapy usually consists of a “7+3 regimen” of 7 consecutive days of infusion with cytarabine (ara-C; antimetabolic agent) followed by 3 days of anthracycline (DNA-intercalating agent). However, other therapeutic options include FLT3-ITD inhibitors for those exhibiting this mutation (Fathi & Chen, 2011) and monoclonal antibodies targeting FLT3, FLT3-ITD and other RTKs associated with initiating AML (De Kouchkovsky & Abdul-Hay, 2016). Despite the promise of these therapies in targeting a variety of RTKs, long-term prognoses of patients with AML remains poor as a major challenge is targeting specifically malignant and not unaffected myeloid precursors (De Kouchkovsky & Abdul-Hay, 2016). Thus, most AML treatments exist targeting RTKs, there is a need for novel AML-specific treatments that come with less off-target and clinical side effects.

Although CML arises differently than AML, a common similarity between these subsets of Leukemia is their reliance on RTK signaling. CML arises from splenomegaly and hyperactive bone marrow expanding the abundance of pluripotent hematopoietic progenitor cells and blood leukocytes (Kabarowski & Witte, 2000). In CML, this occurs over a long (chronic) period of time followed by an acute crisis of an overabundance of undifferentiated myeloblasts (often referred to as blasts) (Kawamoto & Minato, 2004), whereas AML arises from an acute massive increase in blast cells without the chronic phase preceding it. Although both CML and AML arise from blast crises, they have different genetic profiles both reliant on RTKs that initiate disease. For example, CML is well-established to be caused by Breakpoint Cluster Region-Abelson Murine Leukemia 1 fusion (Bcr-Abl) ( Jabbour & Kantarjian, 2018), that occurs during the incomplete differentiation of CD34 and CD90 myeloid progenitors (Arrigoni et al., 2018), similar to AML. Bcr-Abl promotes cancer-promoting processes such as growth and proliferation as well as downstream signaling of cancer pathways such as Ras, Raf, Jun, Myc and STAT (reviewed in Jabbour & Kantarjian, 2018). Together, CML differs from AML in its initiating mechanism, but exerts its effects similarly by targeting RTKs for progression.

Unsurprisingly based on the mechanism of action of CML, currently, one of the most widespread treatments for CML are small RTK inhibitors such as Imatinib. Imatinib is an inhibitor of Bcr-Abl and has been shown as an effective therapeutic against CML. Specifically, in clinical trials examining Imatinib efficacy, it was shown that overall survival (OS) of patients who took Imatinib reached 85%, while 49%–77% exhibited complete cytogenic response (CCyR) and 18–58% of patients exhibited a major molecular response (MMR). However, even those that demonstrate MMR initially are at risk for resistance to RTK inhibitors ( Jabbour & Kantarjian, 2018). Furthermore, approximately one-third of patients who discontinued administration of the drug exhibited molecular relapse (Sacha, 2014), suggesting both the dependency on this drug for long-term survival and the transient nature of its effect simultaneously. In addition to Imatinib, there are other RTK inhibitors available such as Dasitinib (350× more potent second generation analog of Imatinib), Nilotinib (30–50× stronger affinity for Bcr-Abl than Imatinib), and Bosutinib (targets both Src kinase and Abl) that all exhibited drawbacks of low to modest clinical tolerability including gastrointestinal and pulmonary clinical side effects (reviewed in Jabbour & Kantarjian, 2018). Hence, although these drugs may be promising for CML treatment, newer drugs with fewer or no clinical side effects are needed.

4.6.1. SOCE

Leukemia is among the cancer types that is reliant on lower SOCE for disease progression. Previously, we showed minimal SOCE to be observed in 3 of 5 AML cell lines examined (Ritchie et al., 2011; Soboloff, Zhang, Minden, & Berger, 2002). Yet, the underlying mechanistic link between dysregulation of SOCE and leukemia is poorly defined.

One possible mechanism is through changes in WT1 and EGR1, whose expression and activity are known to be dysregulated in AML. Hence, whereas WT1 expression is normally lost during leukocyte maturation, it is expressed in 73%–93% of primary AML samples (Owen, Fitzgibbon, & Paschka, 2010; Ritchie et al., 2011; Rossi, Minervini, Carella, Melillo, & Cascavilla, 2016). Further, inactivating WT1 mutations are observed in 10%–12% of patients, which is a negative prognostic indicator for AML (Owen et al., 2010; Rampal & Figueroa, 2016). WT1-inactivating mutations have been linked to the failure of WT1 to direct TET2 to its transcription site, leading to hypermethylation signatures that resemble TET2 mutant-induced hypermethylation (Rampal & Figueroa, 2016). Because WT1 expression is lost during leukocyte maturation and AML is associated with lower SOCE, it is quite possible that loss of SOCE is a prognostic marker for leukemia and could be associated with enhanced WT1 expression and activity. This would be an interesting future direction to examine, as it could reveal a previously unknown mechanism in how SOCE drives AML.

Conversely, EGR1 is expressed during normal myeloid differentiation (Saeed et al., 2011; Tian et al., 2016) and may prevent leukemogenesis through inhibiting oncogenes E2F-1 and c-myc (Gibbs, Liebermann, & Hoffman, 2008; Shafarenko, Liebermann, & Hoffman, 2005). Furthermore, it has been demonstrated that HL60 AML cells, which normally have low SOCE, can induce EGR1 expression in response to vitamin D3, which leads to differentiation into monocytes and recovered SOCE (Chen, Wang, Wang, & Studzinski, 2004; Gardner, Balasubramanyam, & Studzinski, 1997; Wang, Salman, Danilenko, & Studzinski, 2005), which suggests that Ca²⁺ plays a vital role in differentiation and could be a powerful tool in preventing disease progression.

Another possible mechanism contributing to SOCE effect on myeloid leukemia is deregulation of STIM and Orai proteins. In support, it has been shown in Bcr-Abl-dependent CML, that STIM1/Orai1 ratio is deregulated and TRPC1 expression is decreased leading to suppressed SOCE (Cabanas et al., 2018). Notably, the main culprit in SOCE suppression here is not the levels of Orai1 and STIM1 expression but actually the stoichiometry between STIM1 and Orai1. In support, it has been shown previously that stoichiometric imbalance between STIM and Orai, actually leads to SOCE suppression (Hoover & Lewis, 2011; Li, Rao, & Hogan, 2011; Scrimgeour, Litjens, Ma, Barritt, & Rychkov, 2009), even if Orai1 is overexpressed (Soboloff et al., 2006). This is important since it was found that Orai1, Orai2, STIM1 and STIM2 are actually overexpressed, but disproportionately, at the transcript level, while Orai2 and STIM2 levels are overexpressed at the protein level in HL60 AML cells with knockdown of Orai1 and Orai2 leading to inhibition of migration via focal adhesion (Diez-Bello, Jardin, Salido, & Rosado, 2017). Therefore, it is possible HL60 cells begin with a disproportionate amount of STIM and Orai proteins, that lead to suppressed SOCE compared to differentiated myeloid cells which subsequently facilitates cancer-driving pathways such as PKC which can inhibit Orai1 (see Section 4.4) as well as deregulate Mitochondrial Redox homeostasis in AML (Di Marcantonio et al., 2017), but knocking down Orai proteins leads to inhibition of Ca²⁺-dependent cancer-related pathways such as focal adhesion regulation (Diez-Bello et al., 2017). Interestingly, time appears to be a factor in their activity, as in early stages of migration, Orai2 knockdown induces a more significant inhibition of migration, while Orai1 knockdown contributes in later stages (Diez-Bello et al., 2017). Further, this is supported by evidence that knockdown of STIM1 in HL60 cells can lead to impaired activation of Ca²⁺-dependent cancer driving pathways including AKT, Src, and cytoskeletal remodeling Rac2, which subsequently impairs AML progression (Clemens & Lowell, 2015). Considered together, it is likely that deregulation of the ratio of STIM and Orai proteins, which leads to SOCE suppression, could be a mechanism that can be further extrapolated therapeutically for the treatment of ovarian cancer.

4.6.2. SERCA, PMCA and elevated cytosolic Ca2+

In addition to previously showing that AML may prefer low SOCE conditions, it is also notable that AML relies on tight regulation of cytosolic Ca²⁺ concentration. This is significant since elevated cytosolic Ca²⁺ can lead to factors leading to cell death including mitochondrial Ca²⁺ overload, apoptosis, and anoikis. Hence, inhibitors have been generated for key regulators of cytosolic Ca²⁺ including SERCA and PMCA. For example, SERCA inhibitors such as Thapsigargin, Cyclopiazonic Acid (CPA) and 2,5-di-tert-butylhydroquinone (BHQ) (see Table 2) have been shown to inhibit AML progression (Bleeker, Cornea, Thomas, & Xing, 2013). SERCA inhibition leads to ER Ca²⁺ depletion, which can not only cause stress responses such as UPR (see section IVA) leading to cell death, but also interfere with Ca²⁺-dependent processes such as attenuating Ca²⁺-dependent control of cadherins interactions and actin remodeling leading to impaired cell attachment (Ko, Arora, Bhide, Chen, & McCulloch, 2001) and inducing apoptosis both in AML (Doan et al., 2015) and other cell types ( Janssen et al., 2009). Therapeutically SERCA inhibitors, such as Thapsigargin and Ethyl 2-amino-6-(3,5-dimethoxyphenyl)-4-(2-ethoxy-2-oxoethyl)-4H-chromene-3-carboxylate (CXL017), have been investigated as a therapeutic for multi-drug resistant AML because SERCA inhibition leads to selective cytotoxicity to multidrug resistant AML cells compared to untreated AML cells through impairing ATPase activity of SERCA in leukemia cells that are more reliant on SERCA (Bleeker et al., 2013). In addition to SERCA inhibitors, Imipramine Blue (IB), normally used as a NADPH-oxidase (NOX) inhibitor, has been newly investigated as a disruptor of Ca²⁺ regulation in AML. At concentrations of 75–150nM, IB has been shown to work analogous to Thapsigargin in that it causes increased cytosolic Ca²⁺, and lysosomal Ca²⁺ release, and impaired ER Ca²⁺ content of FLT3-ITD AML (while at higher concentrations inhibiting STAT5 downstream of Flt3) (Metts et al., 2017). Finally, it has been shown that Ca²⁺ clearance disruption through pan-PMCA inhibition (sodium orthovanadate; see Table 2) can actually sensitize AML to treatment through sustained intracellular Ca²⁺ levels and mitochondrial Ca²⁺ overload leading to apoptosis (Pesakhov et al., 2016) similar to the effect of SERCA inhibition and NOX inhibition. Together, it appears that disruption of Ca²⁺ storage and increases in cytosolic Ca²⁺ through inhibitors of SERCA and PMCA leads to selective targeting of AML cells which could be a novel and potentially effective target for treating AML.

4.6.3. Targeting Ca2+ signaling in myeloid leukemia

A recently developed strategy for Leukemia treatment, especially CML, is to target Ca²⁺. CAI and its derivative Carboxyamidotriazole Orotate (CTO), inhibit both PM and mitochondrial Ca²⁺ influx, and thus suppress SOCE by maintaining high intracellular Ca²⁺ levels through both inhibition of SOCE and mitochondrial Ca²⁺ intake (Mignen et al., 2005). It is suggested that CAI treatment results in an impaired ability for cells to buffer Ca²⁺ (Mignen et al., 2005), which is explained by previous research indicating that enhanced intracellular Ca²⁺ buffering capacity can prevent SOCE inhibition by mitochondrial Ca²⁺ inhibitors (Glitsch, Bakowski, & Parekh, 2002). Clinically, CAI and CTO have been shown to inhibit Imatinib-resistant CML growth and progression (Alessandro et al., 2008; Corrado et al., 2012) through inhibiting exosome-stimulated angiogenesis, subsequently disrupting the surrounding tumor microenvironment (Corrado et al., 2012). Thus, CAI and CTO, in conjunction with Imatinib, provides an effective therapy for CML. Following, because CAI and CTO effectively synergize with Imatinib, it is worth exploring whether CAI and CTO could also synergistically work with other RTK inhibitors to provide alternative therapeutic strategies, which could expand the therapeutic efficacy of these drugs to potentially treat AML.

Besides CAI and derivatives, other SOCE inhibitors, such as a recently generated Orai inhibitor from Rhizen Pharmaceuticals (RP4010), have recently undergone Phase 1/1B clinical trials for AML (Carmichael et al., 2018); https://clinicaltrials.gov/ct2/show/NCT03119467) as a synergistic drug with RTK inhibitors such as Gilteritinib and antimetabolic agents such as ara-C (Newswire, 2019).

Besides SOCE inhibitors calcineurin/NFAT inhibitors, such as Cyclosporin A (Eckstein, Van Quill, Bui, Uusitalo, & O’Brien, 2005), which prevents nuclear translocation of NFAT (Boss, Abbott, Wang, Pavlath, & Murphy, 1998) (see Table 2), have undergone clinical trials in conjunction with Imatinib for CML (Shou et al., 2015) due to NFAT inhibition leading to sensitization of CML to Imatinib. Interestingly, an expansive clinical study done at MD Anderson Cooper Cancer Center between 2000 and 2012 showed that median survival time of AML patients taking Ca²⁺ channel blocker (CCB) medication for cardiovascular health was 9 months versus 16.7 months for those who did not take CCBs (Chae, Dimou, Pierce, Kantarjian, & Andreeff, 2014). The CCBs used by these patients were amlodipine (dihydropyridine) and diltiazem (benzothiazepine) (see Table 2). Both of these CCBs mainly block L-type VGCCs (Godfraind, 2017), but amlodipine also has N-type VGCC inhibitory capabilities (Ozawa, Hayashi, & Kobori, 2006). Therefore, it is possible that using drugs blocking CCBs could have a confounding effect of promoting AML and the effect of VGCCs on AML progression warrants further investigation.

Considered altogether, both AML and CML have associations with Ca²⁺ regulation that are being targeted, or warranting investigation, therapeutically. This has provided a foundation for future therapies that could have a novel impact on the therapeutic landscape of Leukemia.

5. Concluding Remarks

Ca²⁺ regulation is a dynamic and multifaceted tool used to tightly regulate a variety of cellular processes. It is important to note that Ca²⁺-modulating processes not only contribute to many normal cellular processes, but dysregulation of Ca²⁺ drives a variety of cancer pathways that can be either tissue-specific, context-specific, or universal among cancers. Previously underappreciated in the context of cancer, Ca²⁺ regulation provides a novel direction for the design of future context-specific treatments for many different cancer types. Although therapeutics have been developed targeting Ca²⁺ for different uses, it is urgent to continue investigation into novel drugs and future generations of current drugs that target the multitude of Ca²⁺-modulating pathways described in this review. This can be done through continuing to both investigate the mechanisms behind Ca²⁺ leading to activation of cancer pathways as well as investigating the effects of drugs on different cancer types. Together, Ca²⁺ can be used as an important tool to combat cancer.