Emerging molecular mechanisms in chemotherapy: Ca2+ signaling at the mitochondria-associated endoplasmic reticulum membranes

1Ca2+ fluxes between ER and mitochondria affect several cancer hallmarks, including apoptosis resistance, migration, and invasion.Oncogenes and tumor suppressors residing at the MAMs execute part of their cellular function by altering ER–mitochondrial Ca2+ transfer, thereby promoting or preventing cancer cell survival.Dependent on the cancer type and cancer stage, ER–mitochondrial Ca2+ transfer can either exert anti-tumorigenic effects like restoring apoptosis sensitivity or exert pro-tumorigenic effects like promoting metastatic behavior.Different chemotherapeutics rely on a Ca2+-signaling component to induce cancer cell death.Ca2+ signaling modulation can (re)sensitize or increase the responsiveness of cancer cells towards chemotherapeutics.

2Ca2+ fluxes between ER and mitochondria affect several cancer hallmarks, including apoptosis resistance, migration, and invasion.

3Oncogenes and tumor suppressors residing at the MAMs execute part of their cellular function by altering ER–mitochondrial Ca2+ transfer, thereby promoting or preventing cancer cell survival.

4Dependent on the cancer type and cancer stage, ER–mitochondrial Ca2+ transfer can either exert anti-tumorigenic effects like restoring apoptosis sensitivity or exert pro-tumorigenic effects like promoting metastatic behavior.

5Different chemotherapeutics rely on a Ca2+-signaling component to induce cancer cell death.

6Ca2+ signaling modulation can (re)sensitize or increase the responsiveness of cancer cells towards chemotherapeutics.

7How can Ca2+ signaling at ER–mitochondrial contact sites be modulated in a cancer-specific manner to fight cancer cell survival?Can ER–mitochondrial Ca2+ signaling events overcome dysregulated cell survival/apoptosis sensitivity in cells with altered oncogene and/or tumor suppressor function?What processes/regulation pathways underlie or control differences between ER–mitochondrial Ca2+ transfer in cancer cells vs. normal cells?How can Ca2+-signaling modulation be applied to increase responsiveness and sensitivity to existing therapies and to induce cancer cell-specific cell death while sparing normal cells?Can Ca2+ signaling be applied in a cancer stage-specific manner, thereby promoting cell death and avoiding metastasis?What other molecular mechanisms, like the generation of ROS, exchange of lipids or alterations in protein composition, or ER–mitochondrial tethering at the MAMs impact or cooperate with Ca2+ signaling in anti-cancer chemotherapeutic actions?

8How can Ca2+ signaling at ER–mitochondrial contact sites be modulated in a cancer-specific manner to fight cancer cell survival?

9Can ER–mitochondrial Ca2+ signaling events overcome dysregulated cell survival/apoptosis sensitivity in cells with altered oncogene and/or tumor suppressor function?

10What processes/regulation pathways underlie or control differences between ER–mitochondrial Ca2+ transfer in cancer cells vs. normal cells?

11How can Ca2+-signaling modulation be applied to increase responsiveness and sensitivity to existing therapies and to induce cancer cell-specific cell death while sparing normal cells?

12Can Ca2+ signaling be applied in a cancer stage-specific manner, thereby promoting cell death and avoiding metastasis?

13What other molecular mechanisms, like the generation of ROS, exchange of lipids or alterations in protein composition, or ER–mitochondrial tethering at the MAMs impact or cooperate with Ca2+ signaling in anti-cancer chemotherapeutic actions?

14Mitochondria do not only fulfill the function of powerhouse of the cell but their function also encompasses more than merely providing the cell with ATP,. Currently, mitochondrial function has been implicated in apoptosis, autophagy, cell proliferation, cellular senescence, and migration–. Furthermore, mitochondrial function is impacted by the “state” of the mitochondrial network, which can range from highly connected to fragmented. Nevertheless, mitochondria do not act as sole orchestrators of cellular processes. In fact, the mitochondrial network rather functions as a highly versatile signaling platform, closely connected to other cell organelles, like the endoplasmic reticulum (ER) and peroxisomes. To allow for inter-organellar cross-talk, the different organelles are often located in close proximity to each other,, like the ER and the mitochondria, which are connected through mitochondria-associated ER membranes (MAMs). These MAMs are defined as ER membranes that are in close apposition (10–50 nm) to the mitochondria and were first isolated as a distinct entity in the early 1990s–. In recent years, MAMs were shown to contribute to various cellular functions like metabolism, autophagy, lipid synthesis but also cell survival and cell death,–. In this sense, the MAMs, like mitochondria, are highly dynamic signaling hubs where signals from different cellular pathways converge and are integrated,–.

15One of the signals transferred between ER and mitochondria at the MAMs is the ubiquitous second messenger Ca2+,. While [Ca2+] in the cytosol is maintained at low levels under resting conditions, the bulk of intracellular Ca2+ is confined in the ER. Ca2+ is predominantly released from the ER via the inositol 1,4,5-trisphosphate (IP3) receptor (IP3R), which is gated by IP3, or the ryanodine receptor (RyR). However, Ca2+ accumulation into the mitochondrial matrix requires Ca2+ transport across the outer mitochondrial membrane (OMM) and the inner mitochondrial membrane (IMM). At the OMM, Ca2+ transport is mediated via the high-conductance voltage-dependent anion channel 1 (VDAC1), while at the IMM, Ca2+ transport is mediated via the mitochondrial Ca2+ uniporter (MCU), the pore-forming unit in the MCU complex, consisting of MCU itself and its regulators–. For a detailed description of MCU regulation we would like to refer to refs.,. The MAMs play an important role in mitochondrial Ca2+ uptake, since they provide a Ca2+ micro-domain, where Ca2+ levels are higher than in the bulk cytosol,. This is necessary to sustain ER–mitochondrial Ca2+ signaling since the MCU has a low affinity for Ca2+. Thus, the MAMs allow for efficient, ‘quasi-synaptic’ mitochondrial Ca2+ uptake upon ER Ca2+ release through the formation of a micro-domain,,. This emphasizes the importance of the MAMs as a signaling hotspot.

16Mitochondrial Ca2+ signals are decoded differentially depending on their spatiotemporal characteristics (see Fig. 1). For example, cytosolic Ca2+ oscillations, efficiently transferred to the mitochondria through these contacts sites, drive mitochondrial metabolism. Moreover, several mechanisms account for the dynamic interplay between Ca2+ signals and mitochondrial metabolism. Ca2+ increases the activity of several rate-limiting enzymes of the tricarboxylic acid (TCA) cycle, including pyruvate, isocitrate, and α-ketoglutarate dehydrogenases, while MCU transcription is controlled by the cAMP-responsive element binding protein, a Ca2+-dependent transcription factor. Cells can also fine tune the level of Ca2+ oscillations that drive mitochondrial bioenergetics via a redox-nano-domain. Mitochondrial Ca2+ uptake triggers K+- and H2O-influx into the mitochondrial matrix, leading to cristae compression and H2O2 release at the MAMs. This provides positive feedback on IP3Rs, enhancing their ability to sustain Ca2+ oscillations. Other contributions of Ca2+ to cell metabolism are the stimulation of complex III of the electron transport chain, as well as stimulation of the ATP synthase and the adenine nucleotide translocase.

17Ca2+ oscillations in the cytosol can also modify mitochondrial metabolism indirectly by activating ARALAR, a mitochondrial glutamate/aspartate transporter playing a central role in the malate/aspartate shuttle, which is strongly dependent on cytosolic Ca2+ signaling. Ca2+ binds to ARALAR and activates ARALAR-mediated glutamate and NAD(P)H transport into the mitochondria–. Also, pyruvate production is Ca2+ dependent and is linked to the malate/aspartate shuttle, providing pyruvate to the mitochondria,. In addition, there is tissue-specific regulation, which determines the threshold for mitochondrial Ca2+ influx and its effects. Liver cells and cardiomyocytes contain a high and a low ratio of mitochondrial Ca2+ uptake 1 (MICU1)/MCU, respectively, with MICU1 (together with MICU2) being an MCU regulator which affects the cooperativity of mitochondrial Ca2+ uptake,. Therefore liver cells display a high cooperativity of mitochondrial Ca2+ uptake, while cardiac cells display a low cooperativity. This prevents the occurrence of mitochondrial Ca2+ signals in response to short-lasting Ca2+ transients and thus avoids mitochondrial Ca2+ overload even when the heart beats at high frequency. Recently, it was shown that apart from affecting the cooperativity of mitochondrial Ca2+ uptake, MICU1, together with its paralog MICU2, inhibits the MCU at cytosolic Ca2+ concentrations lower than ~600 nm, thereby determining the relatively high threshold for MCU-mediated Ca2+ uptake. Compared to wild-type cells, loss of MICU1, which also results in loss of MICU2, lowers the threshold for MCU-mediated mitochondrial Ca2+ uptake to about 200 nM Ca2+. Compared to mitochondria lacking both MICU1 and MICU2, mitochondria containing MICU1 but not MICU2 display a higher threshold for mitochondrial Ca2+ uptake (~350 nM Ca2+), indicating that MICU1 by itself can inhibit MCU. These observations indicate that MICU1 does not only control MCU cooperativity, but can also function as a gatekeeper of MCU. Moreover, MICU2 requires MICU1 to regulate MCU. MICU1’s function as a gatekeeper is also important in vivo to prevent mitochondrial Ca2+ overload, as evidenced in MICU1-knockout animals, developing ataxia and muscle fatigue associated with elevated mitochondrial Ca2+ levels and reduced ATP levels. Finally, MICU1 and MICU2 were shown to bind the mitochondrial lipid cardiolipin, facilitating membrane anchoring of the complex and the fine-tuned Ca2+-dependent regulation of the MCU by MICU1 and associated factors, like EMRE.

18On the other hand, massive Ca2+ release from the ER causes mitochondrial Ca2+ overload, resulting in opening of the mitochondrial permeability transition pore (mPTP),, mitochondrial swelling, and subsequent cell death (Fig. 1)–. Several mechanisms may account for this. For instance, mitochondrial Ca2+ influx results in Ca2+ binding to cardiolipin, causing the disintegration of complex II and the release of the functionally active catalytic subunits in the mitochondrial matrix, providing a source of reactive oxygen species (ROS) that triggers the opening of the mPTP. In addition to this, recent evidence emerged that mitochondrial Ca2+ could directly target the mPTP, resulting in a conformational change leading to its opening and to subsequent mitochondrial swelling.

19Given its critical role in cell fate and survival, the composition and functional properties of the MAMs have to be carefully regulated and controlled. This occurs through a distinct set of proteins with a variety of cell biological properties and functions. Some of these proteins are directly involved in Ca2+ signaling at the MAMs, e.g. the IP3R and VDACs,,, whereas others alter ER–mitochondrial Ca2+ signaling by acting through modulation of these Ca2+-transport systems, e.g. promyelocytic leukemia protein (PML) or phosphatase and tensin homolog (PTEN). Other proteins modify the characteristics of the MAMs (e.g. their distance to the mitochondria), like the ER–mitochondria tethers mitofusin-2 (Mfn-2),, the protein kinase RNA-like endoplasmic reticulum kinase (PERK), and the spacer protein fetal and adult testis expressed 1 (FATE1). We will focus on the proteins involved in ER–mitochondrial Ca2+ transfer at the MAMs, yet for a more extensive list of proteins present at the MAMs, we refer to ref.. For further insights on the role of Ca2+-transport systems in cell death and survival, we refer to ref.. All the MAM components discussed in this section are schematically depicted in Fig. 2.

20Since ER–mitochondrial Ca2+ transfer occurs at the MAMs, it is not surprising that several intracellular Ca2+-transport systems reside at the MAMs. Both the IP3R and VDAC1 can be found at the MAMs, where they are connected to each other via glucose-related protein of 75 kDa (GRP75). Interestingly, isoform-specific functions for these channels at the MAMs have been discovered. Overexpression of VDAC1, VDAC2, as well as VDAC3 increased mitochondrial Ca2+ uptake, yet pro-apoptotic Ca2+ signals were only enhanced by VDAC1 overexpression. Moreover, VDAC1 is the only VDAC isoform that co-immunoprecipitated with the IP3R. Similarly, IP3R3 seems to be the isoform that preferentially transmits pro-apoptotic Ca2+ signals to the mitochondria via the MAMs,. Not only Ca2+-release channels are located at the MAMs, the sarco-/endoplasmic reticulum Ca2+ ATPase (SERCA) pump localizes to the MAMs as well–. SERCA pumps can influence the properties of the MAMs since their activity determines how fast Ca2+ is cleared from the micro-domain. Moreover, SERCA pumps critically affect apoptotic sensitivity by controlling the steady-state filling level of the ER Ca2+ stores, which is determined by the balance between ER Ca2+ leak and ER Ca2+ uptake,–. Also for SERCA, isoform-specific functions exist: ER stress causes induction of SERCA1T, a truncated splice-variant of SERCA1. SERCA1T is a determinant of ER leakiness, thereby promoting Ca2+ transfer to the mitochondria and thus supporting pro-apoptotic Ca2+ signaling,.

21As will be discussed further on, expression levels of the Ca2+-signaling proteins critically determine the cell’s Ca2+-signaling properties. Furthermore, ER–mitochondrial Ca2+ signaling is fine-tuned by various oncogenes and tumor suppressors,. These proteins may induce post-translational modifications that alter the Ca2+-signaling properties of proteins at the MAMs. Particularly, phosphorylation of the IP3R is a critical factor: Akt/protein kinase B (PKB) suppresses IP3R-mediated Ca2+ release through phosphorylation, while tumor suppressors PTEN (direct dephosphorylation of IP3R) and PML (indirect dephosphorylation via sequestration of protein phosphatase 2A (PP2A) and subsequent Akt/PKB inhibition) counteract this. Also, SERCA is a target for post-translational modification: p53 changes SERCA’s oxidative state, promoting its ER Ca2+-uptake activity and thus altering the net flux of Ca2+ released from the ER,,.

22In addition, complex formation between Ca2+-transport proteins like IP3Rs and VDAC1 at the MAMs and tumor suppressors or oncogenes influences ER–mitochondrial Ca2+ transfer. Notably, several Bcl-2-protein family members, critical regulators of apoptosis, were shown to be present at the MAMs, and can modify IP3R-mediated Ca2+-release,,. Bcl-2, an anti-apoptotic protein of the Bcl-2-protein family, inhibits the IP3R directly by binding with its Bcl-2 homology (BH) 4 domain to the central, modulatory domain of the IP3R–. The BH4 domain also participates in overall stability of the Bcl-2 protein, affecting its IP3R-inhibitory properties. Bcl-2’s transmembrane domain is necessary for efficient in cellulo IP3R inhibition. Indirectly, Bcl-2 changes IP3R activity by providing a docking place at the IP3R for protein phosphatase 1 (PP1), which inhibits IP3R function by dephosphorylating the receptor. Bcl-2 also regulates IP3R function by docking dopamine- and cAMP-regulated phosphoprotein of 32 kDa, a PP1 inhibitor, and the protein phosphatase calcineurin in a complex on the IP3R. Thus, IP3R activity is regulated by a negative feedback mechanism that prevents excessive, pro-apoptotic Ca2+ release from the ER. Besides Bcl-2, also Bcl-Xl and Mcl-1 can modulate IP3R activity–. An extended discussion on the modulation of Ca2+ signaling by Bcl-2-protein family members can be found elsewhere,.

23Besides the IP3Rs, Bcl-2-family members can also target mitochondrial Ca2+-transport systems, including VDAC1. VDAC1/Bcl-2 complex formation inhibits VDAC1’s function in mitochondrial Ca2+ transport,. The BH4 domain of Bcl-2 appeared to play a critical role in VDAC1 regulation. Follow-up work revealed that the BH4 domain of Bcl-Xl is more effective in targeting and modulating VDAC1 than the BH4 domain of Bcl-2. Consequently, while both the BH4 domains of Bcl-2 and Bcl-Xl could prevent mitochondrial Ca2+ uptake, BH4-Bcl-2 acted at the level of IP3Rs, while BH4-Bcl-Xl acted at the level of VDAC1. To summarize, ER–mitochondrial Ca2+ transfer at the MAMs can be influenced through IP3R-mediated Ca2+ release, VDAC1-mediated mitochondrial Ca2+ uptake, or modulation of the SERCA activity.

24In addition to this, ER–mitochondrial Ca2+ transfer is altered by the number of MAMs and the distance between the ER and the OMM at the MAMs,. Proteins that hold both membranes together are typically called tethers. An example of a tether at the MAMs is the GTPase Mfn-2 involved in mitochondrial fusion. Mfn-2 tethers ER and mitochondria through homo- and heterotypic interactions with Mfn-2 and Mfn-1, respectively,. The importance of tethering was shown in Mfn-2-knockout cells, where ER–mitochondrial Ca2+ transfer was severely reduced. Yet, the role of Mfn-2 as a tether has been debated, since another study showed that ablation of Mfn-2 does not impair the ER–mitochondrial connection but contrarily, tightens it,. Here, Mfn-2 was proposed as a spacer that increases the distance at the MAMs and reduces ER–mitochondrial signaling. A typical spacer is FATE1, a cancer testis antigen, which is normally only expressed in the testis. However, in certain cancers, FATE1 becomes upregulated and causes MAMs alterations.

25The different possible layers of regulation of Ca2+ signaling impart an enormous flexibility to the cell concerning fine-tuning cellular processes in response to internal and external stimuli. However, this sensitive system can be hijacked to drive malignant transformations in the cell. It is known that several types of cancer cells undergo an extensive rewiring of their Ca2+-signaling machinery, favoring oncogenesis,,. At the level of the ER and the mitochondria, expression levels of Ca2+-signaling proteins, including VDAC1, IP3R, and SERCA, are often altered in cancer cells. For instance, VDAC1 expression levels are correlated with tumor growth and invasion in several types of cancer, e.g. non-small cell lung cancer and cervical cancer,. In this regard, recently, genetic disruption of VDAC1 in cells from cancer xenograft models displayed decreased mitochondrial membrane potential and ATP content with a consequent low migration rate and tumor regression,. Another example includes IP3R1 downregulation in bladder cancer cells, which attenuates cisplatin-mediated apoptosis through a decrease in ER–mitochondrial Ca2+ uptake, preventing mitochondrial Ca2+ overload,. Remodeling of Ca2+ signaling in tumorigenesis is also documented by the significant reduction or loss of the SERCA3 isoform in transformed colonic epithelial cells. Different mechanisms may be responsible for the change in expression levels. Recently, two novel mechanisms dysregulated in several cancer types have been discovered to impact the proteasomal turnover and thus stead-state expression levels of IP3R3 and the apoptotic sensitivity of cells. (i) The tumor suppressor PTEN competes with F-box/LRR-repeat protein 2 (FBXL-2), an E3 ubiquitin ligase component belonging to the Skip-Cullin1-F-box protein family, for binding to IP3R3, thereby slowing down FBXL-2-mediated proteasomal degradation of IP3R3. This represents a novel mechanism by which loss of PTEN allows cancer cells to evade apoptosis, since pro-apoptotic mitochondrial Ca2+ transfer becomes impaired due to downregulation of the IP3R3. (ii) The tumor suppressor BRCA1-associated protein 1 (BAP1), a deubiquitylating enzyme, promotes ER–mitochondrial Ca2+ transfer by stabilizing the IP3R3. BAP1 function is particularly impaired during prolonged environmental stress, associated with acquired inactivating mutations in BAP1 genes. Loss of BAP1 results in IP3R3 downregulation, hampering the effective apoptotic clearance of damaged cells and favoring oncogenesis and malignant cell survival.

26While MCU is not residing at the MAMs, its expression can be controlled in a tumor-specific manner, e.g. via microRNAs (miR). As such, miR-25, targeting MCU, was overexpressed in colon cancer cell lines and tumor samples, decreasing MCU expression compared to non-tumorigenic cells. Moreover, miR-25 overexpression in HeLa cells reduced mitochondrial Ca2+ accumulation, resulting in apoptosis resistance. In contrast, antagonizing miR-25 expression using antagomirs re-sensitized colon cancer cells to Ca2+-dependent apoptotic stimuli, like H2O2 and ceramide. Interestingly, MCU may prevent tumor cell survival in an early stage, but can become a pro-malignant factor in late-stage tumors, like triple negative breast cancer cells. These cells express high levels of MCU, correlating with tumor size and lymph node infiltration, which negatively impact survival outcome. The mechanisms involved MCU-dependent uptake of Ca2+ into the mitochondrial matrix and subsequent generation of ROS that stabilized hypoxia-inducible factor-1α, a transcription factor driving the expression of genes involved in cancer migration and invasion,.

27Furthermore, oncogenes like Akt/PKB and FATE1, and tumor suppressors like PML and PTEN, can play additional roles in the development of cancer via Ca2+-signaling modulation. A striking example of this is apoptotic resistance. Since mitochondrial Ca2+ overload is involved in apoptotic cell death, modifying ER–mitochondrial Ca2+ transfer at the MAMs alters apoptotic sensitivity. Cancer cells can gain resistance against cell death, e.g. by overexpressing proteins that suppress IP3R-mediated Ca2+ signaling, like Akt, or by increasing the intermembrane distance at the MAMs (e.g. FATE1), thereby rendering ER–mitochondrial Ca2+ transfer less efficient. These mechanisms are not only supporting basic cancer cell characteristics, but also underlie resistance against chemotherapy. Cell death induction strategies still play a central role in the fight against cancer. While selective oncogene inhibitors, like venetoclax/ABT-199 for Bcl-2 inhibition, emerged and entered the clinic, chemotherapy remains a very effective way to eradicate tumor cells by triggering cell death.

28Besides ER–mitochondrial Ca2+ signaling, Ca2+ fluxes across the plasma membrane can affect cancer properties as well. We would like to refer to other excellent reviews regarding this topic–, as in this section, we would like to focus on chemotherapeutic drugs that act on Ca2+ signaling at the MAMs.

29Ca2+ signaling appears to be a major contributor to the cytotoxic effects of chemotherapy. Many chemotherapeutic agents trigger a rapid onset of cytosolic [Ca2+] rises. Furthermore, shifts in cytosolic [Ca2+] have been proposed as early markers for cytotoxicity in cells in response not only to H2O2 or staurosporine but also to chemotherapeutics like gossypol or arsenic trioxide (ATO). The mechanism of these early cytosolic [Ca2+] elevations is not always fully understood, but may in part depend on the presence of p53. Upon chemotherapeutic treatment, extra-nuclear p53 can accumulate at the ER membranes where it binds SERCA and activates its ER Ca2+-uptake activity,,,,. Thus, SERCA activation will augment the ER Ca2+-store content, overfilling the ER with Ca2+ and thus increasing the likelihood for spontaneous ER Ca2+-release events. However, the occurrence of shifts in cytosolic [Ca2+] may not be a general phenomenon, as on-target Bcl-2 inhibitors, like ABT-737 and venetoclax/ABT-199, do not trigger these early cytosolic [Ca2+] elevations, even not in cancer cells that are dependent on Bcl-2 for their survival. The reason for these varying responses are not fully understood, but may actually relate to the mechanism of action of the drug applied and in particular whether p53 is involved.

30Here, we will discuss the chemotherapeutic agents that act via ER–mitochondrial Ca2+ signaling. These chemotherapeutics are summarized in Table 1, whereas a schematic representation of their function at the MAMs is provided in Fig. 3.

31Acute promyelocytic leukemia (APL) is almost always characterized by a t(15;17) chromosomal translocation, resulting in a PML/retinoic acid receptor (RAR) α fusion protein that hinders the differentiation of hematopoietic cells by inhibiting gene transcription,. APL patients are mostly treated by a combination of all-trans retinoic acid and ATO therapy, which stimulates APL cell differentiation by triggering proteasomal degradation of the PML/RARα fusion protein,–. This treatment approach results in high APL cure rates, reflected by high complete remission and overall survival percentages,.

32ATO also influences ER–mitochondrial Ca2+ signaling, thereby repressing autophagy in cancer cells. Autophagy is an important pro-survival pathway in malignant cells that experience oncogenic stress, as missing nutrients can be delivered to the cells via this process,. The tumor suppressor PML, which is localized at the MAMs, represses autophagy by promoting ER–mitochondrial Ca2+ transfer and mitochondrial respiration. Hence, PML is often downregulated in cancer cells. Loss of PML results in reduced IP3R-mediated Ca2+ transfer from the ER to the mitochondria, leading to decreased mitochondrial Ca2+ levels, thereby diminishing mitochondrial respiration and ATP production. This triggers activation of AMP-activated protein kinase, which stimulates pro-survival autophagy by a mechanism involving mechanistic target of rapamycin and unc-51-like kinase 1 pathways,. Interestingly, this autophagic process is repressed in APL cells treated with ATO. Short-term treatment of these cells with ATO promotes the selective degradation of the PML/RARα fusion protein, but not of PML. Furthermore, exposure to ATO increases the PML levels at ER–mitochondria contact sites in a p53-dependent manner (Fig. 3). Besides a block in autophagy, ATO-treated cells also displayed reduced resistance to metabolic stress. Furthermore, this study implicates that the response of tumor cells characterized by loss of PML to chemotherapeutic agents can be improved by inhibiting autophagy.

33Cisplatin is a chemotherapeutic agent used to treat numerous human cancers, including lung, ovarian, head and neck, bladder, and testicular cancer. The anticancer activity of this platinum-based drug has been linked to its ability to covalently bind purine residues on DNA,. This interaction causes DNA damage, interferes with DNA repair mechanisms, and blocks cell division, ultimately leading to apoptotic cell death. Unfortunately, cisplatin treatment has been associated with considerable side effects, like cardiotoxicity, hepatotoxicity, nephrotoxicity, and toxicity of other organs, and drug resistance is often acquired during therapy as well. To overcome these obstacles, combination therapies of cisplatin with other anti-cancer drugs, including paclitaxel, doxorubicin, and gemcitabine, form the basis for treatment of many human cancers.

34Interestingly, the ER–mitochondrial Ca2+ signaling pathway contributes to cisplatin-induced cell death. Treatment of SKOV3 human ovarian cancer cells with cisplatin increased the number of ER–mitochondria contact sites, causing a Ca2+ flow from the ER to the mitochondria (Fig. 3). This resulted in high mitochondrial Ca2+ levels, which triggered apoptosis in the cisplatin-treated ovarian cancer cells. Furthermore, the expression level of the anti-apoptotic protein Bcl-2, which is overexpressed in many tumors and drives tumorigenesis and chemoresistance, seems to be a determinant for the sensitivity of cancer cells to cisplatin (Fig. 3). In non-small cell lung cancer and bladder cancer, cisplatin sensitivity could be enhanced by downregulating Bcl-2,. In addition, downregulation of Bcl-2 in SKOV3 cells with siRNA increased the cytoplasmic and mitochondrial Ca2+ levels as well as the number of ER–mitochondria contact points after cisplatin treatment, thereby increasing the sensitivity to the chemotherapeutic agent. Hence, Bcl-2 seems to form a potential therapeutic target to improve cisplatin therapy of ovarian cancer cells. Additionally, in neuroblastoma cells, cisplatin-induced cell death was preceded by a rise in cytosolic [Ca2+]. Cisplatin treatment also increased the expression levels of several Ca2+-transport systems, including IP3R3, RyR3, and the S100 Ca2+-binding protein A6. Therefore, cisplatin-induced cell death can be enhanced by pharmacological modulators of Ca2+-regulatory proteins, like the SERCA inhibitor thapsigargin.

35BH3 mimetics are a class of anti-cancer drugs inhibiting the function of anti-apoptotic Bcl-2-protein family members like Bcl-2, Bcl-Xl, and Mcl-1–,. This causes pro-apoptotic Bcl-2-family members, which are sequestered by their anti-apoptotic counterparts through a hydrophobic cleft, consisting of the BH1, BH2, and BH3 domain, to be released, resulting in apoptosis,. This is an effective anti-cancer therapy in cancers that rely on an upregulation of the anti-apoptotic Bcl-2-family proteins for their survival. Several molecules have been developed as BH3 mimetic drug, notably ABT-737, which targets the hydrophobic cleft of both Bcl-2 and Bcl-Xl, its orally available analog ABT-236 (navitoclax) and ABT-199 (venetoclax), which solely targets the hydrophobic cleft of Bcl-2. While these drugs are specifically developed to suppress the canonical, anti-apoptotic function of the Bcl-2-protein family members at the mitochondria, it seems that their intracellular effects are more complex.

36Thrombocytopenia is an important side effect caused by ABT-737 and ABT-263, since these BH3 mimetic drugs inhibit the function of Bcl-Xl, which is essential for platelet formation and survival–. Dysregulation of intracellular Ca2+ homeostasis might underlie ABT-737- and ABT-263-induced thrombocytopenia, as addition of ABT-263 to platelets triggered an acute rise in cytosolic Ca2+ levels. However, a direct link between deranged Ca2+ signaling and platelet dysfunction was not provided since this effect was only observed upon addition of relatively high ABT-263 concentrations (10 µM), whereas platelet function was already decreased at much lower concentrations (100 nM–1 µM). Furthermore, prolonged treatment of platelets with ABT-263 and ABT-737 depleted the intracellular Ca2+-storage organelles,. However, it was not clear whether the effects on Ca2+ signaling were the cause of platelet apoptosis or whether they were the consequence of platelets being in late-stage apoptosis due to Bcl-Xl inhibition. Our lab excluded a direct impact of ABT-737 on intracellular Ca2+ signaling, since ABT-737 application did not affect thrombin-induced Ca2+ signaling in platelets nor ATP-induced Ca2+ signaling in HeLa cells. Moreover, SERCA activity and IP3R-mediated Ca2+ release were unaffected by ABT-737. These results argue against a proximal role of Ca2+-signaling dysregulation in platelet dysfunction and apoptosis induced by ABT-737.

37On the other hand, for ABT-199/venetoclax, which only targets the hydrophobic cleft of Bcl-2, no evidence was found for a perturbation of Ca2+ homeostasis,,. In several human and mouse cell models, ABT-199 did not trigger cytosolic Ca2+ release events by itself nor did it affect agonist-induced IP3R-mediated Ca2+ signaling. Also, clearance of Ca2+ from the cytosol after agonist application was not affected. Furthermore, ABT-199 did not interfere with the inhibition of IP3R caused by overexpression of Bcl-2. Nevertheless, it seems that there is an interplay between ABT-199-induced cell death in Bcl-2-dependent cancer cells and basal Ca2+ signaling, since chelating intracellular Ca2+ using BAPTA-AM enhanced ABT-199-induced cell death. The mechanisms underlying this phenomenon remain unclear, but might be due to downregulation of anti-apoptotic Bcl-2-family members or upregulation of pro-apoptotic Bcl-2-family members. Furthermore, Bcl-2-dependent cancer cells can be sensitized towards ABT-199 by the application of BIRD-2, a BH4-domain inhibitor of Bcl-2 that triggers toxic Ca2+-release events and apoptosis in various cancer cells, including chronic lymphatic leukemia, diffuse large B-cell lymphoma, multiple myeloma, follicular lymphoma, and lung cancer cells,–. Further, it appears that BIRD-2 upregulates Bim in a Ca2+-dependent manner, thereby likely accounting for an increased sensitivity of the cells towards BH3 mimetics like ABT-199 application,. As such, BIRD-2 and ABT-199 can act synergistically to trigger cell death in Bcl-2-dependent cancers. Finally, it appears that cancer cells that are less sensitive to BH3 mimetics are more sensitive to BIRD-2 and vice versa,,.

38Interestingly, ABT-737 can enhance the chemotherapeutic effectivity of cisplatin in cholangiocarcinoma (CC) as well as in ovarian cancer cells,. In the latter, ABT-737 treatment increased the cisplatin-induced growth inhibition and apoptosis in cisplatin-resistant cells, hence restoring their sensitivity to the chemotherapeutic agent. Combination therapy with cisplatin and ABT-737 of the cisplatin-resistant ovarian cancer cells increased the number of ER–mitochondria contact sites induced by cisplatin, stimulating Ca2+ transfer from the ER to the cytosol and the mitochondria (Fig. 3). Hence, ABT-737 can reverse cisplatin resistance of ovarian cancer cells by enhancing ER-associated and mitochondria-mediated apoptosis. Furthermore, ABT-737 sensitized CC cells to cisplatin therapy by regulating mitochondrial dynamics. ABT-737 stimulated CC cells to undergo apoptosis after cisplatin treatment by promoting mitochondrial fission and inducing mitophagy. Therefore, ABT-737 combined with cisplatin might be an effective strategy to treat CC patients. This also provokes the question in what manner Bcl-2-protein family members regulate the dynamics of the MAMs in cancer.

39Of note, earlier versions of Bcl-2-inhibiting molecules can dysregulate intracellular Ca2+ homeostasis. HA14-1 and stabilized HA14-1s can deplete ER Ca2+ stores by inhibiting the SERCA Ca2+ pump,. In a separate study, both HA14-1 and BH3I-2′ trigger Ca2+ release from the ER of pancreatic acinar cells through IP3R- and RyR-mediated mechanisms, elevating cytosolic Ca2+ levels, a feature that contributes to their cell death properties. Here, it was proposed that dissociation of Bax from Bcl-2 using these drugs sensitize IP3Rs and RyRs to cytosolic Ca2+. Excitingly, the ability of HA14-1 and BH3I-2′ to increase cytosolic Ca2+ levels in pancreatic acinar cells was strictly dependent on the presence of Bax, while the presence of Bcl-2 and Bak was not critical for this process. It was proposed that Bax, released from Bcl-2, can induce Ca2+ leak from the ER, either by itself or by acting on ER Ca2+-leak channels like IP3Rs. Yet, further work using selective Bcl-2 inhibitors is needed to validate this model. A detailed discussion on the role of BH3 mimetics and Ca2+ signaling is provided elsewhere,.

40Resveratrol is a natural polyphenol produced in response to stressful conditions by various plant species. It is found in several foodstuffs, including grapes, mulberries, and peanuts, and has been attributed beneficial health effects since the early 1990s–. It is known as a multi-target agent exhibiting antioxidant, anti-inflammatory, and immunomodulatory activities. This pleiotropic compound affects cell proliferation, differentiation, apoptosis, and autophagy and attenuates many age-related chronic complications, such as metabolic, cardiovascular, and neurodegenerative diseases,. Furthermore, resveratrol has been used as a chemopreventive and chemotherapeutic agent in many types of cancer–. For instance, it has been used in clinical trials conducted in patients with colon, colorectal, and gastrointestinal cancers as well as in a trial examining the effects of resveratrol in the prevention of cancer in healthy participants. Remarkably, resveratrol functions as a specific anticancer agent with limited toxicity in normal cells. However, the exact mechanism by which resveratrol specifically kills cancer cells remains unclear.

41Recently, it was suggested that the difference in ATP demand between cancer and somatic cells underlies the cancer cell-specific toxicity of resveratrol. Cancer cells are characterized by a very high ATP demand at the ER because of immense protein folding activities going on at this organelle,. Therefore, tethering of the ER and the mitochondria, which produce ATP via the ATP synthase localized in the IMM, is strongly enriched in cancer cells, warranting high ATP levels in the proximity of the ER,. However, resveratrol, which acts as an inhibitor of the F1 subunit of the ATP synthase– exploits the enhanced ER–mitochondria coupling in malign cells to kill these cells exclusively (Fig. 3). As a result of ATP synthase inhibition in resveratrol-treated cells, ATP formation is reduced, by which the high energy demand of cancer cells is not met anymore. As a consequence of the reduced ATP content at the mitochondria, SERCA activity within the MAMs is decreased, hampering Ca2+ reuptake into the ER, provoking not only an accumulation of Ca2+ in the micro-domain between ER and mitochondria, but also leading to a depletion of the ER via the Ca2+-leak channels. Both phenomena result in a high local Ca2+ concentration and because of the enforced ER–mitochondria coupling in cancer cells, mitochondrial Ca2+ accumulation is consequently enhanced upon treatment with resveratrol. Due to the fact that cancer cells are more sensitive to ATP synthase inhibition than healthy cells, resveratrol will especially trigger apoptotic cell death via mitochondrial Ca2+ overload in those cells.

42Adriamycin, also known as doxorubicin, is an anthracycline-type drug that has been used in cancer therapy for many years,. It is characterized by a broad-spectrum antineoplastic activity, although its mechanism of action is complex. Adriamycin inhibits topoisomerase II, intercalates in DNA, and generates free radicals, hence inhibiting biosynthesis of macromolecules and leading to oxidative stress–. Ultimately, adriamycin treatment results in apoptotic cell death. This chemotherapeutic agent is commonly used to treat various types of cancer, including breast, ovarian, bladder, stomach, and lung cancer. However, the applicability of adriamycin as anticancer therapy is restricted due to its severe toxic effects in healthy tissues,,. Especially cardiotoxicity forms a major concern during adriamycin therapy, limiting the dose that can be administered to patients,. To circumvent the adverse effects of adriamycin, several drug delivery systems have been used,,. For instance, adriamycin-induced toxicity is decreased when using liposomal, nanoparticle, or hydrogel drug formulations.

43Adriamycin also renders cells more prone to programmed cell death by influencing the ER–mitochondrial Ca2+ signaling axis,,. The tumor suppressor p53, an important transcription factor regulating DNA repair, cell-cycle arrest, and apoptosis, modulates Ca2+ homeostasis by stimulating the SERCA pump located at the ER and the MAMs. As a consequence, Ca2+ accumulation in the ER is enhanced. Interestingly, treatment of cancer cells with adriamycin caused an enrichment of p53 at the ER and the MAMs (Fig. 3),. This induction of p53 by adriamycin led to higher ER Ca2+ levels and higher cytosolic and mitochondrial [Ca2+] increases evoked by agonist stimulation. Moreover, Ca2+ transport from the ER to the mitochondria was increased by adriamycin, allowing apoptotic stimuli to rapidly overload the mitochondria with Ca2+, resulting in apoptotic cell death. Hence, chemotherapeutic agents like adriamycin boost toxic ER–mitochondrial Ca2+ signaling through modulation of the Ca2+ homeostasis at the MAMs, thereby triggering apoptotic cell death in cancer cells,.

44Another way in which Ca2+ signaling is able to contribute to the sensitivity of cancer cells to chemotherapeutics, is the induction of autophagy. Recently, valproic acid was found to reduce the intracellular availability of IP3, hence blocking Ca2+ transfer to the mitochondria and altering the AMP-activated protein kinase 1/2–mechanistic target of rapamycin pathway. This lack of ER–mitochondrial signaling induced autophagy, thereby sensitizing cancer cells to adriamycin.

45A recent insight in tumor biology is the occurrence of dysregulation of lipid metabolism in cancer cells and its importance to several aspects of cancer cell function and survival. Consequently, disrupting or altering lipid homeostasis in cancer cells might be an efficient way of inducing cell death. One of the ways to induce apoptosis in this manner in cancer cells is the use of chemotherapeutics that increase intracellular, free cholesterol. Alkyl phospholipids were observed to hinder cholesterol transport from the mitochondria to the ER, avoiding its esterification, whereas cholesterol synthesis and incorporation was increased at the same time in glioblastoma cells. In macrophages it was found that this free cholesterol accumulates at the ER, inhibiting SERCA pumps and causing depletion of the ER. This results in an increased transfer of Ca2+ to the mitochondria and subsequent apoptosis.

46Another chemotherapeutic drug that acts on lipid metabolism is mitotane, a derivative of the insecticide dichlorodiphenyl-trichloroethane. Mitotane is the only chemotherapeutic drug approved for the treatment of adrenocortical carcinoma (ACC), one of the deadliest endocrine malignancies–. In ACC cells, but not in non-adrenal cancer tissues, mitotane counteracts tumor growth and steroid hormone production,. These effects of mitotane are believed to be the result of the inhibition of the sterol-O-acyl-transferase 1 (SOAT1), an enzyme that protects cells against the harmful effects of free cholesterol by transforming it into cholesterol esters. Because of SOAT1 inhibition, mitotane therapy leads to the accumulation of toxic lipids, including free cholesterol and oxysterols, inside ACC cells, which triggers lipid-induced ER stress. Moreover, in ACC tissue samples, SOAT1 expression correlated with the response to mitotane treatment.

47Interestingly, mitotane-induced apoptosis in ACC cells also depends on FATE1 expression. Under physiological conditions, FATE1 expression is restricted to the testis and adrenal gland, while FATE1 overexpression is observed in a variety of cancers,. In ACC cells, FATE1 expression is controlled by the steroidogenic factor-1 (SF-1), a transcription factor that is important for adrenal development and plays a role in the formation of adrenocortical tumors,. FATE1, localized at the MAMs where it uncouples the ER and mitochondria, decreases mitochondrial Ca2+ uptake in ACC cells. In this way, FATE1 protects cancer cells from Ca2+-dependent apoptotic stimuli. Furthermore, FATE1 expression conferred mitotane resistance to ACC cells, when this chemotherapeutic drug was used in a dose that falls inside the therapeutic window for ACC patients, whereas knockdown of FATE1 in these cells increased the sensitivity to the drug. FATE1 probably protects against mitotane-induced apoptosis because of its localization in the MAMs. Mitotane inhibits the SOAT1 enzyme, which is also localized in the MAMs, resulting in the accumulation of toxic cholesterol lipids. Hence, in the presence of FATE1 mitotane-mediated SOAT1 inhibition may be less efficient (Fig. 3). Interestingly, FATE1 expression in ACC tumor cells can even be used as a prognosis indicator since FATE1 expression is inversely correlated with the overall survival of ACC patients.

48In conclusion, MAMs form important intracellular signaling platforms, allowing for Ca2+-encoded messages between the ER and the mitochondria. This ER–mitochondrial Ca2+ exchange can be altered during cancer development to promote cancer hallmarks like evasion of apoptosis, excessive cell proliferation, and a metabolic rewiring. In addition, many chemotherapeutics act via Ca2+ signaling at the MAMs (Fig. 3). Moreover, chemotherapeutics can interfere with the function of oncogenes and tumor suppressors, thereby altering ER–mitochondrial Ca2+ transfer. In this sense, chemotherapeutics that modify ER–mitochondrial Ca2+ signaling can be used to increase the response of cancer cells towards therapeutics that harbor a Ca2+ component in their working mechanism.