Spatiotemporal regulation of Store-Operated Calcium Entry in Cancer Metastasis

Abstract

The store-operated calcium (Ca²⁺) entry (SOCE) is the Ca²⁺ entry mechanism used by cells to replenish depleted Ca²⁺ store. The dysregulation of SOCE has been reported in metastatic cancer. It is believed that SOCE promotes migration and invasion by remodeling the actin cytoskeleton and cell adhesion dynamics. There is recent evidence supporting that SOCE is critical for the spatial and the temporal coding of Ca²⁺ signals in the cell. In this review, we critically examined the spatiotemporal control of SOCE signaling and its implication in the specificity and robustness of signaling events downstream of SOCE, with a focus on the spatiotemporal SOCE signaling during cancer cell migration, invasion and metastasis. We further discuss the limitation of our current understanding of SOCE in cancer metastasis and potential approaches to overcome such limitation.

Introduction

Ca²⁺ is the most versatile secondary messenger in the cell (1). Hundreds of genes in the mammalian genome encodes proteins containing the Ca²⁺-binding EF-hand motif and many more are indirectly regulated by Ca²⁺ through calmodulin (2). The specificity of Ca²⁺ signaling is intricately controlled by the spatial and temporal coding of cytosolic Ca²⁺. The cell has evolved an elaborate “Ca²⁺ toolkit” consisting of Ca²⁺ channels, pumps and exchangers allowing cells to tightly control the flux of Ca²⁺ across the plasma membrane and intracellular organelles. In non-excitable cells such as cancer cells, SOCE is the predominant mechanism by which Ca²⁺ enters the cell. SOCE is controlled by the endoplasmic reticulum (ER) Ca²⁺ sensors (STIM1 and STIM2) and plasma membrane channel proteins (ORAI1-3). The upregulation of STIM and ORAI proteins has been reported in a variety of cancer types. There is convincing evidence in cell culture and xenograft models that SOCE is required for cancer cell migration and invasion (3–15). In the past decade, the critical roles of Ca²⁺ signaling and SOCE in oncogenesis, metastatic progression and therapeutic resistance have been critically reevaluated, which have been covered in a few recent reviews (16–18). In this mini review we focus on the spatial and temporal regulations of SOCE and their roles in cancer cell migration, invasion and metastasis.

Spatial and temporal coding of Ca²⁺ signals

The calcium ion Ca²⁺ plays critical roles in many physiological and pathophysiological processes in a spatiotemporally dynamic manner (19). Rapid, localized and transient changes in intracellular or organellar Ca²⁺ concentration directly control muscle contraction, neural transmission and hormone secretion, while sustained elevations of Ca²⁺ play pivotal roles in biological processes ranging from fertilization to gene expression to apoptosis and cell death (20). The powerful, comprehensive and versatile functions of Ca²⁺ signaling are due to the elaborate spatial and temporal control of the activation and inhibition of Ca²⁺ channels, exchangers and pumps in the plasma membrane (PM) and endoplasmic/sarcoplasmic reticulum (ER/SR) membrane.

The introduction of high dynamic range fluorescent Ca²⁺ sensors, the development of novel targeting strategies for such sensors, and the advent of confocal microscopy greatly facilitated the visualization of spatiotemporal Ca²⁺ signals in intact cells. The remarkable progress in this field has electrified the investigation of Ca²⁺ signaling in diverse physiological processes such as excitation-contraction coupling, vascular tone regulation and exocytosis. In particular, the discovery of Ca²⁺ sparks (21) in single cardiomyocytes has revealed fundamental principles of the Ca²⁺ signaling system, and inspired numerous investigations on spatiotemporal Ca²⁺ signals in different cell types. Consequently, a related family of elementary Ca²⁺ signaling events was defined in excitable cells with intriguing imaging approaches, such as Ca²⁺ spike (local or global Ca²⁺ transient measured in the presence of excess Ca²⁺ buffers EGTA)(22), Ca²⁺ sparklet (local Ca²⁺ transient that arises from a single L-type Ca²⁺ channel)(23), Ca²⁺ blink (local depletion of Ca²⁺ in a single endo/sarcoplasmic reticulum cistern during a Ca²⁺ spark)(24) (25) and Ca²⁺ nanospark (local or global Ca²⁺ transient measured by new dyad-targeted Ca²⁺ sensors GCaMP6f-Triadin1/Junctin) (26). In non-excitable cells, Ca²⁺ puff (discrete local Ca²⁺ release event of inositol 1,4,5-trisphosphate receptor (IP3R) origin)(27) and Ca²⁺ flickers (discrete, local and short-lived high Ca²⁺ microdomains created by TRPM7 and IP3R)(28) are representatives of spatiotemporal Ca²⁺ signaling events. These elemental Ca²⁺ signals reflect the activity and behavior of Ca²⁺ channels underlying unique biological processes. For instance, in heart disease, changes in Ca²⁺ sparks/nanosparks and their properties indicate the alterations in many regulatory features of the junctional RyR2 macromolecular complex. Aberrant Ca²⁺ release eventually leads to spontaneous electrogenic Ca²⁺ waves that trigger arrhythmias as often seen in catecholaminergic polymorphic ventricular tachycardia models(29).

SOCE in the temporal control of Ca²⁺ signaling

As a very unique Ca²⁺-entry mechanism, SOCE dynamically regulates cellular processes in both excitable and non-excitable cells upon extracellular stimuli (30). SOCE is induced in response to the activation of plasma membrane (PM) receptors and subsequent Ca²⁺ release from the endoplasmic reticulum (ER) mainly mediated by inositol triphosphate receptors (IP3R) (31). Upon Ca²⁺ depletion in ER, the ER Ca²⁺ sensors STIM oligomerizes and translocates to the nanoscopic PM-ER junctional regions where it activates the PM pore-forming units Orai and induces Ca²⁺ influx to replenishe the depleted Ca²⁺ store (Figure 1).

Figure 1.

Schematic depicting the spatiotemporal SOCE Ca²⁺ signaling. Extracellular Ca²⁺-mobilizing agonists activate the receptors in the plasma membrane (PM), and then trigger Ca²⁺ release from inositol triphosphate receptors (IP3Rs) in the endoplasmic reticulum (ER) membrane. STIM1/2 sense the depletion of Ca²⁺ content in ER lumen and translocate to ER-PM junctions where they activate the Orai1/2/3 channels in the PM and induce Ca²⁺ influx, in the form of global Ca²⁺ oscillations. As agonist activation intensifies, STIM and Orai isoforms are selectively recruited and activated to shape unique temporal Ca²⁺ signatures and differentially regulate the downstream effectors. At low-agonist concentrations where store Ca²⁺ depletion is modest, STIM2, Orai2 and Orai3 are more active in mediating low frequency oscillations. As high-concentration agonist largely depletes store Ca²⁺, STIM1 and Orai1 are becoming increasingly dominant in mediating more frequent Ca²⁺ oscillations and eventually Ca²⁺ plateau. Mitochondrial Ca²⁺ uniporter (MCU) uptakes the sustained Ca²⁺ entry from SOCE to drive numerous mitochondrial processes including adenosine triphosphate (ATP) production. Meanwhile it provides a feedback signal for Orai channel and reduces Ca²⁺ dependent inactivation (CDI) of SOCE by buffering cytosolic Ca²⁺ levels. Similarly, plasma membrane Ca²⁺ ATPase (PMCA) markedly depresses near-membrane Ca²⁺ concentration to facilitate SOCE and generate long-lasting Ca²⁺ oscillations. The sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA) in the ER membrane is assumed to be the third element of SOCE. Upon store depletion, SERCA co-localizes with STIM1 at SOCE puncta to favor rapid Ca²⁺ pumping from the high Ca²⁺ microdomains to ER. Inositol triphosphate receptor type 2 (IP3R2) is the main pro-oscillatory subtype whereas IP3R3 functions as an anti-oscillatory unit. IP3R1 can induce a transient Ca²⁺ signal or an oscillatory Ca²⁺ signal. Such a functional coupling between SOCE, MCU, PMCA, SERCA and IP3R is ideal to maintain low Ca²⁺ concentration at the SOC mouth, alleviate CDI process and consequently extend the duration of Ca²⁺ influx or oscillations. Spatially, in the high Ca²⁺ microdomains at the mouth of Orai channel, adenylyl cyclase 8 (AC8) converts ATP into cAMP and activates cAMP signaling loop. Ca²⁺/ calmodulin (CaM) activates the AKAP79/calcineurin (CaN) complex to selectively dephosphorylate NFAT and triggers its translocation to the nucleus to regulate gene expression. NFAT4 translocation demands relatively low SOCE activity while NFAT1 requires more robust SOCE activity. ATP, adenosine triphosphate; cAMP, cyclic adenosine monophosphate; AKAP79, A-kinase anchoring protein 79; NFAT, nuclear factor of activated T-cells; P, phosphate group.

STIM protein has 2 isoforms, STIM1/2 that sense distinct ER Ca²⁺ levels. STIM2 is a weak SOCE activator and its recruitment is triggered under basal conditions with a modest ER Ca²⁺ depletion (32). Orai has Orai1/2/3 that oligomerize to form native SOC channels (33). In addition Orai1 has a naturally-occurring short translational variant, Orai1β. High doses of agonists induce massive depletion of store Ca²⁺ and robust cytosolic Ca²⁺ plateaus mainly mediated by STIM1 and Orai1-dependent SOCE (Figure. 1). However, when stimulated by Ca²⁺-mobilizing agonists at relatively low physiological concentrations, SOCE generally induces global Ca²⁺ oscillations resulting from repetitive regenerative discharges and refilling of store Ca²⁺, in an IP3R and STIM/Orai dependent manner(34) (Figure. 1). Each oscillation accompanies a transient decrease in ER Ca²⁺ and consequently a transient activation of STIM. This transient activation of STIM can be observed by total internal reflection fluorescence microscopy, showing repetitive redistribution of STIM1/2 into distinct puncta close to PM in synchrony or with tens of seconds-delay from the onset of individual Ca²⁺ transient (35). The STIM2 translocation initiates a little faster than that of STIM1 during the onset of oscillations (36). Interestingly, the oscillations exhibit a great temporal heterogeneity with a wide range of frequencies at different stimulus intensities (34) (37). In addition, different agonists recruit different STIM proteins to support cytoplasmic Ca²⁺ oscillations and gene expression. These unique Ca²⁺ oscillation profiles are coordinated by STIM1 and STIM2 collaboratively, since cells lacking STIM1, STIM2 or STIM1/2 exhibited distinctly different oscillatory patterns in terms of frequency of oscillations, percentage of oscillating cells, plateau cells, and non-responding cells (34). The interactions of STIM/Orai isoform also fine tune the diversity of Ca²⁺ signals. Specifically, Orai1 is dispensable in generation of oscillations(37). Orai2 and Orai3 have a potentially privileged interaction with STIM1 to mediate SOCE at low-agonist concentrations, while interaction of STIM1 with Orai3 dampens SOCE at high-agonist concentrations. The interaction of both STIM1 and STIM2 with Orai2 probably shapes SOCE in response to the full spectrum of physiological stimuli (34). The contribution of STIM1 shades that of STIM2 as stimulus strength increases, switching Ca²⁺ oscillation signals into plateau. In addition, Orai2 and Orai3 multimerize with and negatively regulate Orai1(37). It is also found that unactivated STIM negatively regulates IP3R-mediated Ca²⁺ release from the ER, likely modulating the sensitivity of IP3R to inositol triphosphate (IP3) (34). IP3Rs play a crucial role in releasing Ca²⁺ from the ER and modulates SOCE. IP3R subtype-specific expression shapes versatile cytosolic Ca²⁺ signaling patterns in different cell types (38). IP3R2 is the main pro-oscillatory subtype, required for robust and long lasting Ca²⁺ oscillations whereas IP3R3 functions as an anti-oscillatory unit (39) but could also mediate short-term oscillations depending on the cell type and stimulus (40). IP3R1 is more flexible since it can induce a transient Ca²⁺ signal or an oscillatory Ca²⁺ signal. IP3R1 has been implicated in modulating SOCE: A small pool of immobile IP3R1 tethered close to STIM1-Orai1 junctions was reported to regulate Ca²⁺ entry (41). Alternatively, activated IP3R1 can associate to STIM1 upon store depletion, providing a reduced store Ca²⁺ microenvironment for enhanced activation of Orai mediated Ca²⁺ entry (42). However, there has been no detailed investigation specifically on how different IP3R subtypes interact with SOCE to manipulate Ca²⁺ oscillations. Therefore, further characterization of IP3R subunit in shaping SOCE-mediated Ca²⁺ oscillations is needed.

Besides, trimeric intracellular cation (TRIC) channel TRIC-A directly interacts with STIM1 and affects the degree of co-localization between STIM1 and Orai1 within ER–PM junctions, limiting SOCE and reducing the amplitude and frequency of oscillatory Ca²⁺ signals(43). In addition to changes in the stoichiometry of the STIM/Orai complex, SOCE could be modulated by proteins that interact with them. For example, STIMATE was identified as a protein that physically interacts with STIM1 to promote STIM1 conformational switch and SOCE through biotin-based proximity labeling techniques (44). Therefore, physiological SOC-mediated Ca²⁺ activities are orchestrated by all STIM/Orai proteins and their functional interaction with a plethora of regulators.

The digitalized temporal control of Ca²⁺ signaling by STIM/Orai, in the form of repetitive Ca²⁺ oscillations with variable frequencies, has critical roles in regulating both physiological and pathophysiological processes. One such example is the differential regulation of the nuclear translocation of transcription factors NFAT1 and NFAT4 (34) (37) (Figure 1). The nuclear translocation dynamics of NFAT1/4 exhibit marked differences (45), where the translocation of NFAT4 could be 5 to 10 times faster than NFAT1 (46). NFAT1 nuclear translocation requires higher SOCE activity supported equally by both ORAI1 and ORAI1β, and becomes more robust at high agonist concentrations (47). In contrast, NFAT4 demands relatively low SOCE activity and shows faster translocation at either high or low agonist concentrations (48).

SOCE in local Ca²⁺ microdomain

The spatially discrete architecture of STIM-Orai puncta suggested that SOCE could give rise to elementary Ca²⁺ signals that create tens of micromoles of Ca²⁺ microdomains in the vicinity of the opening mouth of activated Orai. The Ca²⁺ concentration in these microdomains can be orders of magnitude higher than the physiological Ca²⁺ concentration in the bulk of the cytosol, which would enable spatial-temporal activation of low-affinity Ca²⁺-binding proteins. Indeed, detection of such SOC-mediated Ca²⁺ microdomains has provided insightful information into the veracity of models for STIM1-Orai1 gating in the cluster, as well as how the spatiotemporal dynamics of SOCE orchestrates downstream signaling. Using the Orai1 channel-targeted biosensors GECO-Orai1 in conjunction with total internal reflection fluorescence microscopy, the so-called single Orai1 channel Ca²⁺ influx was visualized, exhibiting sporadic fluorescence transients with briefer “flickers” lasting hundreds of milliseconds, and longer “pulses” lasting seconds (49). This single SOC influx was artificially activated by co-expressing CAD (CRAC activating domain of STIM1) or STIM1, or by 2-APB acting as an agonist of Orai3, in transfected cells (49). Further, they demonstrated a functional heterogeneity of Orai1 channel activity between individual puncta resulting from heterogeneity in STIM1/Orai1 ratio (50). More recently, we used Orai1-tethered or palmitoylated biosensor GCaMP6f to report subplasmalemmal Ca²⁺ signals. We observed spontaneous elementary Ca²⁺ signals of SOCE in the form of discrete and long-lasting transients, namely SOCE Ca²⁺ glows, in cancer as well as endothelial cells(51), suggesting that the gating of Orai1 channels is temporally synchronized and spatially coordinated.

The components of the SOCE Ca²⁺ microdomain signaling axis vary depending on cell types and cellular processes. In unstimulated T cells, spontaneously generated Ca²⁺ microdomains require STIM1, STIM2 and Orai1 that form patches at a circular ~300 nm subplasmalemmal region, while upon TCR stimulation, the signaling pathway switches to NAADP (nicotinic acid adenine dinucleotide phosphate)-RYR1-STIM2-Orai1 axis and triggers more Ca²⁺ microdomains (52). In HEK293 cells, the NFAT1 transcription factor(53) and adenylyl cyclase 8-cAMP signaling loop (54) are selectively activated by SOC-mediated subplasmalemmal Ca²⁺ microdomains in the vicinity of the Orai1 mouth, but not by a global increase in cytosolic Ca²⁺ (Figure 1). Our recent findings reveal that in melanoma invadopodia, SOCE operates in a digital fashion to activate locally enriched calmodulin and high-threshold Ca²⁺ effector Pyk2 recruited by invadopodia (51). Besides, STIM1 and Orai1 also interact with caveolin-1(55), CRACR2A(56), A-kinase anchoring protein (AKAP79)(57) and calcineurin(58) which are anchored very close to the Orai1 channel (Figure 1). Therefore, the organization of Ca²⁺ microdomains enable efficient activation of Ca²⁺ effectors in close proximity to the channel while minimizing deleterious effects of global Ca²⁺ increase. One possible mechanism for the spatiotemporal coordination of Orai1 channels at microdomains could be local store Ca²⁺ depletion similar to Ca²⁺ blinks (59), as previously observed within nanometer-sized stores (the junctional cisternae of the sarcoplasmic reticulum) during elementary Ca²⁺ release events in the heart (24) (25). Alternatively, a single STIM1-Orai1 channel may act as an intrinsic oscillator and manifest dampened oscillatory behavior under certain experimental conditions (49). Future visualization of putative local ER depletion (or blink-like event) that precedes Ca²⁺ glow might provide new insights into the trigger mechanism of the local SOCE events.

Spatial temporal regulation of SOCE in metastatic cancer cells

SOCE has been increasingly implicated in cancer cell migration, invasion and metastasis (4). In migrating cells Ca²⁺ concentration is unevenly distributed, with higher Ca²⁺ in the trailing tail and lower Ca²⁺ in the leading edge (60). The Ca²⁺ gradient migrating cells facilitates the retraction of the trailing tail by promoting the contraction of actomyosin fibers (60, 61). Although the overall Ca²⁺ concentration is low in the leading edge, there are highly active Ca²⁺ flickers in the front of migrating cancer cells, which regulates focal adhesion turnover in migrating cells (62). Interestingly, SOCE has also been shown to regulate cell migration and focal adhesion turnover in breast cancer cells (13), although it was unclear whether SOCE might contribute to local Ca²⁺ increase at focal adhesions. Our previous study reveals that STIM1- and Orai1-mediated SOCE in melanoma cells is organized in the form of persistent Ca²⁺ oscillations and the frequency of Ca²⁺ oscillations regulates both the assembly and activity of invadopodium (63, 64). Consistent with our results, SOC mediated Ca²⁺ oscillation mechanism also regulates colorectal cancer migration / invasion (65) and esophageal cancer cell proliferation / tumor growth (66). The oscillatory Ca²⁺ increase allow cancer cells to activate pro-invasion signaling (e.g. activation of Src) while avoiding the detrimental toxicity of constitutive Ca²⁺ overload (14). In agreement with this notion, constitutive activation of SOCE with Thapsigargin was not able to promote melanoma cell invasion or invadopodium assembly, despite robust activation of Src (6).

Many Ca²⁺ effectors such as calcineurin and myosin are locally recruited to invadopodia to regulate invadopodia assembly and extracellular matrix degradation (67) (Figure 2). These observations raised the possibility that SOCE might be locally activated at invadopodia to coordinate cancer cell invasion and metastasis. Indeed, due to the low conductivity of store-operated Ca²⁺ channels and the ubiquitous presence of Ca²⁺ sequestering mechanisms in the cytosol, the SOCE-mediated increases in cytosolic Ca²⁺ are likely localized and limited to the sub-plasmalemmal region. However, the diffusible Ca²⁺ indicators such as Fura2-AM or Fluo4-AM are limited in their capacity for the detection of such localized high-Ca²⁺ microdomains. This technical limitation could be circumvented by targeting GECO to the plasma membrane. By fusing GCaMP6f to the N-terminal or Orai1 (GCaMP6F-Orai1) or to a palmitoylation signal, we detected long-lasting (the average full duration at half maximum lasted from seconds to tens of seconds) sub-plasmalemmal Ca²⁺ transients at invadopodia in invading melanoma cells. These elementary Ca²⁺ signaling events are mediated by SOCE and could be blocked by pharmacological SOCE inhibitors, dominant negative Orai1 or by CRISPR/Cas9 deletion of STIM1 (51). Interestingly, calmodulin and Pyk2 are also locally recruited to invadopodia. It is possible that invadopodia might serve as a signaling hub by bringing together Ca²⁺ channels (Orai1) and downstream effectors (calmodulin, Pyk2, Src, calcineurin etc) to coordinate cancer cell invasion and melanoma metastasis (Figure 2).

Figure 2.

Invadopodia as signaling hubs coordinating local SOCE Ca²⁺ microdomains and tumor invasion. Orai1in the plasma membrane (PM), Ca²⁺ sensor STIM1 in the endoplasmic reticulum (ER) and Ca²⁺ effectors such as calmodulin (CaM), calcinuerin (CaN), Pyk2 and Mysoin (Myo) are locally recruited to invadopodia in metastatic cancer cells. The activation of invadopodial Orai1 clusters results in high-Ca²⁺ microdomains, which activates CaM and promotes the binding and activation of local Ca²⁺ effectors during invadopodia assembly and extracellular matrix degradation. The Ca²⁺ effectors localize outside of invadopodia are not activated. Therefore, invadopodia brings together SOC and Ca²⁺ effectors to allow the robust and specific activation of SOCE signaling and co-ordination of melanoma invasion.

Conclusion and perspective

The past decade has seen significant progress in our understanding of SOCE in cancer. This is made possible by the identification of STIM and Orai proteins as ER Ca²⁺ sensor and plasma membrane channel proteins, respectively. There is clinicopathological data supporting upregulation of STIM and/or Orai proteins in a wide spectrum of cancers including breast cancer, colorectal cancer, melanoma and cervical cancer (7–11, 68–72). Although these data indicated the upregulation of SOCE during oncogenesis and metastatic progression, they failed to reveal the nuance in the spatial and temporal regulation of SOCE signaling in malignant cells. We and others showed that SOCE promotes Ca²⁺oscillation and high- Ca²⁺ microdomains in melanoma and esophageal cancer cell culture (3, 63, 66). While these data support the hypothesis that spatial and temporal coding of SOCE is critical for metastatic progression, one must caution that cancer cells could behave very differently in cultured cell lines than in patients and animals. However, the determination of spatiotemporal coding of Ca²⁺ among human patients is technically challenging. In recent years organoid cultures from freshly dissected cancer tissues provide excellent in vitro models that could faithfully recapitulate the phenotypic heterogeneity of human cancer (73). These 3D organoids cultures could be genetically manipulated to express GECOs for Ca²⁺ imaging, and could be an important tool to uncover the important nuance of spatiotemporal coding of SOCE in human cancer.

Previous studies showed that SOCE inhibitors such as SKF-96365 and 2-APB could be used to inhibit tumor growth and metastasis in xenograft mouse models (13, 66, 68). Although these inhibitors are not suitable for clinical application due to low affinity and lack of specificity, more specific inhibitors with IC50 in the nanomolar range have been developed (17). One such inhibitor, CM2489, has completed a phase I clinical trial (17). These potent and specific inhibitors for SOCE could be useful in targeting dysregulated SOCE in metastatic cancer. However, there is still a long way before SOCE could be targeted in the clinical setting. A major hurdle in targeting SOCE is the universality and complexity of SOCE in normal tissue. For example, SOCE is critical for immune cell function and inhibition of SOCE in CD8⁺ T cells impaired anti-tumor immunity in a melanoma model (74). The pleiotropic effects of SOCE inhibitors on non-cancer cells in the tumor microenvironment and in normal tissues need to be carefully analyzed. Another hurdle is the paucity of data on the heterogeneity of SOCE signaling in cancer. For example, although activation of SOCE promotes invadopodia formation, melanoma invasion and metastasis, there is also evidence that SOCE is inhibited in a subset of Wnt5a-positive, highly invasive melanoma cells (75). In melanoma and colorectal cancer, although STIM2 promotes cancer cell migration and invasion while inhibiting cell proliferation (69, 70). Future investigation into the complexity of SOCE heterogeneity in cancer cells and the tumor microenvironment is warranted.

Perspectives points:

• The dysregulation of SOCE has been increasingly implicated in the promotion of cancer cell migration, invasion and metastasis.

• There is emerging evidence showing that SOCE is regulated spatially and temporally in cancer cells and other cells. The spatial and temporal coding of SOCE is critical for its regulation of tumor growth and progression.

• Future investigation into the spatial and temporal regulation of SOCE in organoid culture models and genetically engineered cancer models will be essential to understand the complexity of SOCE signaling in more physiologically relevant context.