The ubiquitin-activating enzyme inhibitor TAK-243 activates store-operated Ca²⁺ channels independently of endoplasmic reticulum Ca²⁺ stores and enhances mitochondrial Ca²⁺ buffering capacity in DU-145 human prostate cancer cells

Abstract

TAK-243, an inhibitor of the ubiquitin-activating enzyme UBA1, impairs the ubiquitination of proteins. We investigated its effects on the intracellular Ca²⁺ dynamics of human prostate cancer cells DU-145. Live-cell Ca²⁺ imaging with Fura2 showed that DU-145 cells exhibited spontaneous Ca²⁺ transients that were largely independent of the activity of the Na⁺/Ca²⁺ exchanger but were abolished by extracellular Ca²⁺ removal, La³⁺, Gd³⁺, BTP2 (Orai channel blocker), and thapsigargin (which depletes endoplasmic reticulum -ER- Ca²⁺ stores). They were also sensitive to U73122 (a phospholipase C inhibitor) and dantrolene (a ryanodine receptor blocker), suggesting the contribution of both store-operated Ca²⁺ channels and intracellular Ca²⁺ release pathways in the generation and maintenance of these Ca²⁺ spikes. At 0.2 μM, TAK-243 augmented the cytosolic concentration of Ca²⁺ ([Ca²⁺]i), boosted the Ca²⁺ spiking activity, whereas at 10 μM it caused a constitutive Ca²⁺ entry through BTP2-sensitive Ca²⁺ channels, which further augmented [Ca²⁺]i but reduced the number of spontaneously active cells. Experiments conducted with the genetically encoded Ca²⁺ dye CEPIA1er showed that TAK-243 did not empty the ER Ca²⁺ stores. Moreover, Förster resonance energy transfer (FRET) experiments on HEK-293 cells transfected with Orai1-CFP and STIM1-YFP indicated that it caused the relocalization of STIM1 and its coupling with Orai1. TAK-243 also increased the membrane potential and the Ca²⁺ buffering capacities of mitochondria. Taken together, treating DU-145 cells with TAK-243 to prevent the ubiquitination of proteins activates the entry of Ca²⁺ through store-operated Ca²⁺ channels independently of the ER Ca²⁺ stores and alters mitochondrial functions.

Graphical Abstract

Introduction

The ubiquitin-proteasome pathway (UPP) is an enzymatic machinery that critically controls the homeostasis of proteins (proteostasis) by altering their ubiquitination status. It influences virtually all aspects of cellular activity. Moreover, a dysregulation of the balance between the synthesis and degradation of proteins participates in the pathogenesis of various human diseases. In light of its patho-physiological relevance, the UPP is considered as an important therapeutic target and proteasome inhibitors such as bortezomib represent a major class of drugs notably as anticancer agents [1]. The first step of the ubiquitination cascade requires the involvement of the ubiquitin-activating enzymes (UAE) UBA1 and UBA6 [2]. Therefore, UAE actually represent the master upstream regulators of the ubiquitin pathway and proteostasis [3], and have been identified as potential targets for therapeutic intervention in cancer therapy [4]. TAK-243 (or MLN7243) was characterized as a first-in-class inhibitor of UAE by blocking their catalytic activity [5]. Subsequent studies revealed that it preferentially blocks UBA1 [6]. Among other properties, TAK-243 reduces mono and polyubiquitination of proteins [5] [6] [7]. This is associated with an impaired proteolysis, an accumulation of proteins in the endoplasmic reticulum (ER), and the activation of an unfolded protein response (UPR) [5] [6] [8]. In addition, TAK-243 is known to inhibit the DNA damage response [5] [6]. Importantly, TAK-243 is more toxic on cancer cells than on non-cancer cells [5] [6] [8]. A recent study examined the impact of TAK-243 on intracellular Ca²⁺ signalling in three mammalian cell lines (αT3, HeLa and SH-SY5Y cells) [9]. Depending on the cell type, TAK-243 was shown to increase intracellular Ca²⁺ ([Ca²⁺]i), to mobilize Ca²⁺ from the ER or stimulate the uptake of Ca²⁺. However, the overall effects of TAK-243 on cellular Ca²⁺ handling appear multifaceted and cell-type dependent [9]. Therefore, further work is required to fully understand how this UAE inhibitor interferes with the intracellular Ca²⁺ signalling machinery of cancer cells.

The expression and/or functions of multiple proteins participating in the homeostasis of Ca²⁺ is altered in cancer cells, which leads to the remodelling of the cellular Ca²⁺ signalling signature [10,11] [12] [13]. Of particular relevance is the occurrence of spontaneous cytosolic Ca²⁺ transients that can give rise to intracellular Ca²⁺ oscillations. Since their first description, spontaneous Ca²⁺ transients of cancer cells have been attracting considerable interest. In vitro and in vivo studies have clearly established the critical roles played by these spontaneous Ca²⁺ spikes in tumour initiation, growth and invasion [14] [15] [16]. However, contrasting results have been obtained concerning the identity of the Ca²⁺ pathway(s) sustaining these Ca²⁺ spikes. For instance depending on the cell line, spontaneous Ca²⁺ transients of colorectal cancer cells necessitate Ca²⁺ entry through dihydropyridine-sensitive L-type Cav1.2 channels (in RKO and Colo320 cells) [16] or a release of Ca²⁺ from internal compartments (in HCT-8 cells) [15]. In PC-3 cells, a human prostatic cancer cell line, Ca²⁺ oscillations triggered by foetal bovine serum (FBS) are governed by a complex mechanism involving multiple ion-transporting proteins such as NALCN and Orai1 channels, Ca²⁺ pumps of the ER (SERCA) and plasma membrane, and mitochondrial Na⁺/Ca²⁺ exchangers [17].

Since TAK-243 alters the cellular Ca²⁺ handling [9], we hypothesized that it could influence the occurrence of spontaneous Ca²⁺ oscillations of cancer cells. Here, we have investigated the effects of a short-term (2 h) treatment of TAK-243 [9] on the metastatic human prostate cancer cell line DU-145. The first part of the work addresses the mechanisms governing the occurrence of spontaneous Ca²⁺ spikes of DU-145 cells. While the Ca²⁺ transients of human prostate cancer cells have been mainly studied in PC-3 and LNCaP cells, much less is known about the spontaneous Ca²⁺ activity of DU-145 cells. The second part focuses on the impact of TAK-243 on Ca²⁺ signalling. Fura2 and CEPIA1er experiments showed that TAK-243 causes a sustained Ca²⁺ entry through store-operated Ca²⁺ channels without depleting ER Ca²⁺ stores. In addition, FRET experiments on HEK-293T cells expressing STIM1-YFP and Orai1-CFP suggest that TAK-243 changes the localization of STIM1 and promotes its association with Orai1. Additionally, TAK-243 alters the mitochondrial membrane potential, the size of the mitochondrial Ca²⁺ stores as well as the Ca²⁺ buffering capacity of mitochondria. Altogether, this work shows that the ubiquitin-activating enzyme inhibitor TAK-243 has a profound impact on resting [Ca²⁺]i and spontaneous Ca²⁺ spiking activity of DU-145 cells. This raises the possibility that part of the antitumoral effect of TAK-243 may be associated with a dysregulation of the Ca²⁺ signalling of cancer cells.

Experimental procedures

Cell culture

HEK-293T (# CRL-1573) and DU-145 (# HTB-81) cells were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). DU-145 cells were routinely cultured in RPMI-1640 Glutamax culture media supplemented with 10% foetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin. HEK-293T cells grown in a DMEM medium supplemented with 10% foetal bovine serum. HEK-293T and DU-145 cells were passaged once a week and maintained at 37 °C in 5% CO2 [18].

Cytosolic Ca²⁺ measurements with Fura2

DU-145 cells were plated on glass coverslips 24- or 48h before performing the Fura2 Ca²⁺ imaging experiments according to [19] [20]. DU-145 cells were incubated for 25 min at room temperature (RT) in the dark with the fluorescent Ca²⁺ indicator Fura2-AM (2.5 μM) diluted in a HEPES-buffered saline consisting in 150 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 5.5 mM glucose, 10 mM HEPES (pH 7.4, NaOH). Fura2-AM was removed by washing and cells were kept for another 5–10 min in a Fura2-free saline solution prior to the recording of the Ca²⁺ activity of Fura2-loaded DU-145 cells. The nominally Ca²⁺-free saline consisted of 150 mM NaCl, 5 mM KCl, 3 mM MgCl₂, 5.5 mM glucose, 1 mM EDTA and 10 mM HEPES (pH 7.4, NaOH). In some experiments, extracellular Na⁺ was replaced with the organic cation N-methyl-D-glucamine (NMDG⁺). The Ca²⁺ imaging setup was driven by MetaFluor (Universal Imaging, Roper Scientific, France) and consisted in an inverted Zeiss A1 microscope (Carl Zeiss, France) equipped with a Fluar 40x oil immersion objective lens (Carl Zeiss, France), a CCD CoolSnap HQ2 camera (Princeton Instruments, Roper Scientific, France), and a DG-4 wavelength switcher (Princeton Instruments, Roper Scientific, France). Fura2-loaded cells were excited at 340 and 380 nm, and emission was collected at 515 nm. For each DU-145 cell, the baseline of Fura2 fluorescence ratio was set to 0 and the size the of the spontaneous Ca²⁺ signal was analysed as the area under the curve (AUC). A cell was considered as “silent” when the AUC was < the threshold of 1.6. Therefore, DU-145 cells displaying an AUC of ≥1.6 were considered as spontaneously active cells.

ER Ca²⁺ measurements with CEPIA1er imaging

For ER Ca²⁺ measurements, HEK-293T and DU-145 cells were seeded at a density of 30,000 cells per coverslip, and allowed to adhere and recover overnight. Next, cells were transfected with 1 μg pCMV R-CEPIA1er (Addgene, Cat # 58216) (HEK-293T) or 1 μg pCMV G-CEPIA1er (Addgene, Cat #58215) (DU-145) using Lipofectamine 2000 (Thermo Fisher Scientific, Cat # 11668019) (HEK-293T cells) or Lipofectamine 3000 (Thermo Fisher Scientific, Cat # 11668019) (DU-145 cells) according to manufacturer protocols. The cells were then imaged 24 hours after transfection. Prior to imaging, the glass coverslips with HEK-293T cells were mounted into Attofluor cell chambers (Thermo Fisher, Cat # A7816). HEK-293T and DU-145 cells were washed four times with 1X HBSS containing 150 mM NaCl, 4.7 mM KCl, 1.13 mM MgCl₂, 2 mM CaCl₂, 10 mM glucose, and 10 mM HEPES with pH adjusted to 7.4 by NaOH. In one experiment, cells were treated either with 10 μM TAK-243 or vehicle control, and with 1 μM ionomycin (Research Products International, Cat # I52020–0.005) used as a positive control to deplete the ER. In the second experiment, the cells were pre-treated in advance of imaging with either 10 μM TAK-243 or vehicle control and allowed to incubate for 2 hours at 37°C in complete media prior to being washed four times with 1X HBSS. The cells were then allowed to rest for the first minute of imaging to establish a baseline before the buffer was exchanged to Ca²⁺-free 1X HBSS and subsequent treatment with the irreversible SERCA pump inhibitor, thapsigargin (Thermo Fisher Scientific, Cat # T7458) to deplete the ER stores. HEK-293T cells were imaged on a Leica Stellaris 5 Confocal Microscope (Leica Microsystems) with a 20X objective, and data were collected using Leica LAS X software. RFP fluorescence was excited with a 561nm laser at 2% output at room temperature. DU-145 cells were imaged on an inverted Zeiss A1 microscope (Carl Zeiss, France) with a Fluor 40x oil immersion objective lens (Carl Zeiss, France). Images were captured every 5 seconds (DU-145 cells) or 10 seconds (HEK-293T cells), and traces were plotted as means ± SEM after normalization to the first time point.

Confocal FRET microscopy experiments

For FRET imaging, HEK-293T cells were seeded as described for the ER Ca²⁺ measurements. Cells were then transfected with 2 μg of human STIM1-YFP (Addgene, Cat # 19754) and 0.5 μg of human Orai1-CFP (Addgene, Cat # 19757) using Lipofectamine 2000 according to manufacturer protocols. Cells were subsequently prepared for imaging as described for the ER Ca²⁺ measurements. The cells were subsequently imaged on a Leica DMi8 widefield microscope equipped with an external CFP/YFP FRET filter wheel and a 20X objective, with data captured by the Leica LAS X software. Cells were allowed to rest for the first minute of imaging to establish a baseline, and then stimulated with either 10 μM TAK-243 or vehicle control. Corrected FRET efficiency was calculated based on the formula: (B – Aβ – Cγ)/C where A, B, and C are the intensities of the donor, FRET and acceptor signals, respectively, and calibration factors β and γ were generated by the acceptor-only and donor-only controls, respectively. The corrected FRET efficiency was then normalized to the first time point to determine the fold increase. Puncta were quantified using the particle analyzer on ImageJ (https://imagej.nih.gov/ij/).

Measurements of changes of the mitochondrial membrane potential

In these experiments, cells were first loaded with the cell permeant fluorescent dye Rhodamine 123 (Rh123, Thermo Fisher, Cat # R302) before monitoring acute changes of the mitochondrial membrane potential (Δφm). The culture medium was removed and cells were rinsed twice with a saline solution containing NaCl 150 mM, KCl 5, MgCl₂ 1, CaCl₂ 2, glucose 5.5, HEPES 10 (pH 7.4, NaOH). Cells were then incubated for 25 min at RT in this saline supplemented with 5 μM Rh123. Following incubation, cells were rinsed twice with the saline before being placed for 5 min at RT in a Rh123-free saline. When mitochondria are first loaded with Rh123, a subsequent depolarization of Δφm is associated with an elevation of the Rh123 fluorescence [21].

Treatment of cells with TAK-243

DU-145 or HEK-293T cells were treated with DMSO or TAK-243 (Selleckchem, Cat # SE-S8341) for 2h at 37°C according to the protocol reported in [9]. TAK-243 was also present when loading cells with Fura2 (25 min/RT) and during the de-esterification phase (5–10 min/RT). The final concentration of DMSO never exceeded 0.1%. TAK-243 was never present during the Fura2 or CEPIA1er recordings.

Materials and statistical analysis

Unless otherwise indicated, media and reagents were from Sigma-Aldrich (France). SigmaPlot (version 10.0, Systat Software), Excel (Microsoft, USA), GraphPad Prims 10 (v 10.3.1, GraphPad software Inc) and SigmaStat (v 3.5, Systat Software) were used for data analysis, graphic representation and for the statistical analysis of the data, respectively. Data are shown as mean ± SEM and differences with a p value of <0.05 were considered as statistically significant.

Results

Spontaneous Ca²⁺ spikes of DU-145 cells require external Ca²⁺

DU-145 cells were loaded with the fluorescent Ca²⁺ probe Fura2 to monitor the occurrence of spontaneous Ca²⁺ spikes. In the presence of a recording saline containing 2 mM Ca²⁺, two populations of DU-145 cells were identified: i) silent cells: cells that did not generate Ca²⁺ transients (e.g. cell #1, Figure 1A), and ii) active cells (e.g. cells #2–6, Figure 1A). Complex patterns of Ca²⁺ activity were observed with a variable range of amplitudes and time intervals between Ca²⁺ spikes including cells with a single (cell #4), multiple (cells #5–6) or compounds spontaneous Ca²⁺ transients (cells #2–3, Figure 1A). Figure 1B shows superimposed Fura2 signals recorded simultaneously from thirty DU-145 cells within the same field of view in the presence of 2 mM external Ca²⁺. This illustrates the heterogeneous, asynchronous and highly variable nature of the spontaneous Ca²⁺ spikes in this cell line. Maintaining DU-145 cells in a nominally Ca²⁺-free saline practically fully abolished the occurrence of cytosolic Ca²⁺ changes (Figure 1C). Figure 1D shows averaged Fura2 signals from cells maintained in a 2 mM Ca²⁺ (black) or in Ca²⁺-free saline (red), respectively. The absence of external Ca²⁺ not only prevented the generation of Ca²⁺ transients but also diminished [Ca²⁺]i (p<0.001). For each experiment the number of active cells per field of view was determined, with one field of view per coverslip. On average, the percentage of active cells was ⁓40–50% in the presence of external Ca²⁺ (Figure 1E, open circles). Maintaining cells in a Ca²⁺-free saline strongly attenuated the occurrence of spontaneous Ca²⁺ spikes with <5% of cells being active in the absence of external Ca²⁺ (Figure 1E, red squares), respectively (n=186 cells from 7 coverslips, p<0.0005). When considering the nearly complete dependence of the Ca²⁺ spikes on extracellular Ca²⁺, it is proposed that the spontaneous Ca²⁺ transients of DU-145 cells require an influx of Ca²⁺.

Figure 1.

DU-145 cells generate spontaneous Ca²⁺ spikes that are driven by external Ca²⁺.

Panel A shows six representative Fura2 recordings from six individual DU-145 cells to illustrate the wide diversity of Ca²⁺ signals. These six cells (#1-#6) were in the same field of view. Panels B and C show superimposed Fura2 recordings from thirty DU-145 cells in the same field of view in the presence (B) and absence (C) of external Ca²⁺. In this latter case, the recording medium was supplemented with 1 mM EDTA. Panel D shows an average of Fura2 signals from n=186 cells (from 7 coverslips) in the presence (black line) and absence (red line) of external Ca²⁺. Statistics were done on Fura2 signals measured at time +180 s. ** p<0.001, Student’s t test followed by a Mann-Whitney test performed on individual cells. Throughout this study we quantified the number of active cells per experiment (i.e., per field view, with one field of view per coverslip). Panel F shows the percentage of active cells (i.e., cells displaying a significant elevation of F340/F380 (Materials and Methods) during the 180 s Fura2 imaging period without or with of 2 mM external Ca²⁺. *** p<0.0005, Student’s t test performed on replicates (n=7). Mean +/− sem.

Spontaneous Ca²⁺ signals of DU-145 cells are not governed by external Na⁺

A recent study dissected the signalling pathway responsible of the Ca²⁺ oscillations generated by PC-3 cells [17]. In this human metastatic prostate cancer cell line, Ca²⁺ transients critically depend on store-operated Ca²⁺ channels and NALCN channels that allow an influx of Na⁺, which, in turn, promotes an entry of extracellular Ca²⁺ via the Na⁺/Ca²⁺ exchanger functioning in its reverse mode [17]. We therefore addressed the question whether the spontaneous Ca²⁺ transients of DU-145 cells required extracellular Na⁺. In the following experiments, Na⁺ was replaced by the organic cation N-methyl-D-glucamine (NMDG⁺) [17]. Figure 2A shows superimposed individual Fura2 signals recorded in the absence of external Na⁺ from thirty DU-145 cells (Figure 2A). As illustrated, the removal of external Na⁺ did not abolish the occurrence of spontaneous cytosolic Ca²⁺ transients. However, DU-145 cells maintained in a Na⁺-free medium had a lower [Ca²⁺]i (Figure 2B) (p<0.001), showing that the basal [Ca²⁺]i was partially dependent on the uptake of Na⁺. However, the removal of external Na⁺ did not significantly affect the percentage of active cells (Figure 2C). Altogether, these experiments indicate that the spontaneous Ca²⁺ activity of DU-145 cells is poorly dependent on NALCN. This is in line with previous data showing little or no NALCN expression in the DU-145 cell line [17].

Figure 2.

Weak contribution of external Na⁺ ions in the generation of spontaneous Ca²⁺ spikes of DU-145 cells.

Fura2-loaded DU-145 cells were placed in a HEPES-buffered saline where Na⁺ ions were replaced by the organic cation N-methyl-D-glucamine (NMDG⁺). Panel A shows representative Fura2 recordings from thirty cells in the same field of view. B, averaged traces from n=346 cells (from 13 coverslips) placed in a Na⁺- (black line) or NMDG⁺-containing (blue line) saline. ** p<0.001, Student’s t test followed by a Mann-Whitney test performed on individual cells. C, active cells (in %) under the same experimental conditions, namely, with Na⁺ or NMDG⁺. Student’s t test performed on replicates (n=13). n.s. difference not statistically significant. Mean +/− SEM.

Spontaneous Ca²⁺ signals are not suppressed by nifedipine but are inhibited by La³⁺, Gd³⁺, and BTP2

Next, we tested the effects of various blockers of cell surface cation channels. Contrary to Colo320 cells, a human colorectal cancer cell line [16], DU-145 cells were still able to generate spontaneous Ca²⁺ spikes in the presence of nifedipine (20 μM), a L-type Ca²⁺ channels blocker (Figure 3A). The effects of the lanthanides lanthanum (La³⁺, 100 μM, Figure 3B) and gadolinium (Gd³⁺, 10 μM, Figure 3C), two widely used blockers that interfere with the functioning of Ca²⁺-conducting channels were also tested. Both cations exerted a strong inhibitory action by preventing the occurrence of spontaneous Ca²⁺ spikes (Figure 3B-C). Store-operated Ca²⁺ channels are cell surface Ca²⁺ channels whose activity is controlled by the Ca²⁺ filling status of the ER. These channels have been shown to be potently inhibited by BTP2 (YM-58483) [22]. Similarly to La³⁺ and Gd³⁺, BTP2 (5 μM) severely impacted the occurrence of spontaneous Ca²⁺ rises (Figure 3D). The blockers La³⁺, Gd³⁺ and BTP2 not only attenuated the spontaneous Ca²⁺ spiking activity of DU-145 cells but they also diminished the basal [Ca²⁺]i (Figure 3E) (p<0.05). Figure 3F illustrates the depressive action of La³⁺ (blue triangles), Gd³⁺ (gray squares) and BTP2 (red triangles) on the number of Ca²⁺ spiking cells. On the other hand, nifedipine caused a significant elevation of [Ca²⁺]i (Figure 3E) (p<0.05) without increasing the percentage of spontaneously active cells (black circles, Figure 3F). Dihydropyridines such as nifedipine are well-characterized L-type Ca²⁺ channels blockers. However, they have been shown to elevate [Ca²⁺]i. For instance, dihydropyridines can release Ca²⁺ from internal compartments [23] or provoke an entry of Ca²⁺ [24] in some cell types. Moreover, they were found to activate STIM proteins independently of store depletion, promoting Ca²⁺ entry through Orai channels [25]. Studying the mechanisms supporting the nifedipine-dependent elevation of [Ca²⁺]i in DU-145 cells was however beyond the scope of the present work. In conclusion, our experiments show that an influx of Ca²⁺ through channels sensitive to La³⁺, Gd³⁺ or BTP2 is required to sustain spontaneous Ca²⁺ spikes.

Figure 3.

Spontaneous Ca²⁺ activity is potentiated by nifedipine and blocked La³⁺, Gd³⁺, and BTP2.

Panels A-D show thirty representative Fura2 recordings from 4 distinct coverslips in the presence of 20 μM nifedipine (A), 100 μM La³⁺ (B), 10 μM Gd³⁺ (C), or 5 μM BTP2 (D). Nifedipine, La³⁺, Gd³⁺, or BTP2 were present during the recording and they had been added 5–10 min prior to time 0. E, averaged Fura2 signals recorded in the presence of nifedipine (n=214 cells/9 coverslips), La³⁺ (blue, n=141 cells/5 coverslips), Gd³⁺ (gray, n=228 cells/9 coverslips), or BTP2 (red, n=249 cells/10 coverslips). Control cells (n=297 cells/11 coverslips). * p<0.05, one way ANOVA followed by a Dunn’s test (vs TAK-243-untreated cells) performed on individual cells. Panel F shows the number of active cells (in %) under each experimental condition tested. * p<0.05, ** p<0.01, *** p<0.005, one way ANOVA followed by a Dunn’s test (vs TAK-243-untreated cells) performed on replicates. n.s. difference not statistically significant. Mean +/− SEM.

Spontaneous Ca²⁺ signals require IP3 production and ER Ca²⁺ release

We tested the possibility that the generation of Ca²⁺ spikes stemmed from an interplay between Ca²⁺ release from the ER and Ca²⁺ influx via Ca²⁺ channels of the plasma membrane. The filling state of the ER depends on the activity of Ca²⁺ pumps named SERCA. Their pharmacological blockade with thapsigargin prevents the ER Ca²⁺ loading and reveals a passive release of Ca²⁺ out of the ER. As illustrated in Figure 4A, SERCA inhibition with thapsigargin abolished the occurrence of Ca²⁺ spikes, implying that the generation of spontaneous transients depended on ER Ca²⁺ release. In parallel, thapsigargin provoked a sustained increase in [Ca²⁺]i (Figure 4A).

Figure 4.

PLC activity, IP3 and ryanodine receptors are required to sustain the spontaneous Ca²⁺ activity.

Panels A-D show thirty representative Fura2 recordings from 4 distinct coverslips in the presence of 1 μM thapsigargin (A, added when indicated), 10 μM U73122 (B), or 50 μM dantrolene (C). Treatment of cells: U73122 and dantrolene were added to cells >20 and >5 min prior to time 0, respectively. They were also present throughout the Fura2 recording. D, averaged Fura2 signals from cells treated with U73122 (red, n=143 cells/5 coverslips), or dantrolene (blue, n=199 cells/8 coverslips). Control cells (black, n=222 cells/8 coverslips). * p<0.05, one way ANOVA followed by a Dunn’s test (vs TAK-243-untreated cells) performed on individual cells. E, percentage of active cells per experimental condition. * p<0.05, ** p<0.01, one way ANOVA followed by a Dunn’s test (vs TAK-243-untreated cells) performed on replicates.

IP3 and ryanodine receptors (IP3Rs and RyRs) are two main pathways of regulated Ca²⁺ release from the ER. We thus attempted to better understand their roles in the generation of spontaneous cytosolic Ca²⁺ spikes. DU-145 cells were treated with the phospholipase C inhibitor U73122 (10 μM), which prevents the production of IP3. U73122 nearly fully abolished the spontaneous Ca²⁺ activity of DU-145 cells (Figure 4B) and slightly elevated [Ca²⁺]i (Figure 4D) (p<0.05). The contribution of RyRs was evaluated using the RyRs antagonist dantrolene (50 μM). It almost fully prevented the occurrence of spontaneous Ca²⁺ spikes (Figure 4C) and reduced [Ca²⁺]i (Figure 4D) (p<0.05). The reduction of [Ca²⁺]i caused by dantrolene suggests that the basal [Ca²⁺]i depends on a constitutive release of Ca²⁺ from dantrolene-sensitive channels. U73122 and dantrolene significantly diminished the number of spontaneously active cells (Figure 4E). In conclusion, these experiments show that the spontaneous Ca²⁺ activity of DU-145 cells required IP3 production and ER Ca²⁺ release from IP3Rs and RyRs. It also required an entry of Ca²⁺ via BTP2-sensitive Ca²⁺ channels. Each of these processes is necessary for the generation of cytosolic Ca²⁺ rises. We propose a mechanism where a PLC-dependent ER Ca²⁺ release through IP3 receptors, followed by a Ca²⁺-induced Ca²⁺ release from RyRs, is maintained by a constitutive activity of store-operated Ca²⁺ channels as illustrated recently in brain pericytes [26].

Dual action of TAK-243 on the spontaneous Ca²⁺ activity of DU-145 cells

The ubiquitin-activating enzyme inhibitor TAK-243 was recently shown to affect the cellular handling of Ca²⁺ [9]. Considering the important roles played by Ca²⁺ signals in the deleterious process of tumour progression and invasion, we investigated the consequences of a TAK-243 treatment on the occurrence of spontaneous Ca²⁺ spikes. In the following experiments, DU-145 cells were treated with 0.2 or 10 μM TAK-243 for 2h/37°C prior to the Fura2 recordings according to the protocol of [9]. Figure 5A-C shows thirty superimposed representative Fura2 traces for each condition, namely 0, 0.2 and 10 μM TAK-243. TAK-243 elevated [Ca²⁺]i in a concentration-dependent manner (Figure 5E) (p<0.05). This TAK-243-induced [Ca²⁺]i augmentation was prevented by the store-operated Ca²⁺ channel inhibitor BTP2 (5 μM, Figure 5D). The number of active cells was quantified for each experimental condition and the results are depicted in Figure 5F. A concentration of 0.2 μM TAK-243 significantly augmented the number of DU-145 cells displaying a spontaneous Ca²⁺ activity (red circles, Figure 5F). However, a diminution of the number of active cells was noted at 10 μM TAK-243 without (blue triangles) or with BTP2 (gray squares, Figure 5F). These results indicated that TAK-243 promoted Ca²⁺ entry through store-operated Ca²⁺ channels that in turn enhanced spontaneous Ca²⁺ activity, (e.g. with 0.2 μM TAK-243). However, when high levels of internal Ca²⁺ were reached (e.g. with 10 μM TAK-243), the spontaneous Ca²⁺ transients were blunted by high cytosolic Ca²⁺, just as is the case following thapsigargin treatment (Figure 4A).

Figure 5.

TAK-243 elevates [Ca²⁺]i and modulates the spontaneous Ca²⁺ activity.

Each of the 4 panels in Figure A-D shows representative superimposed Fura2 signals from thirty cells from four distinct coverslips. DU-145 cells had been untreated (A) or treated with 0.2 (B) or 10 μM TAK-243 (2h/37°C, C). Panel D: cells had been treated with 10 μM TAK-243 (2h/37°C) and the recording saline was supplemented with 5 μM BTP2. E, averaged Fura2 signals from cells treated with 0.2 μM TAK-243 (red, 116 cells/6 coverslips), 10 μM TAK-243 (blue, n=290 cells/11 coverslips), or 10 μM TAK-243 + 5 μM BTP2 (gray, n=156 cells/6 coverslips). Open circles: control cells (n=299 cells/11 coverslips). * p<0.05, one way ANOVA followed by a Dunn’s test (vs TAK-243-untreated cells) performed on individual cells. F, percentage of active cells per experimental condition. * p<0.05, ** p<0.01, one way ANOVA followed by a Dunn’s test (vs TAK-243-untreated cells) performed on replicates.

Panels G-J shows representative superimposed Fura2 signals from thirty HEK-293T cells. Panel G: control (untreated) HEK-293T cells; panels H-J: HEK-293T cells pretreated with 10 μM TAK-243 for 2h/37°C (same procedure as in Figure 5A). The recording saline was without (H) or with (I) 1 μM BTP2. J, averaged Fura2 signals from cells untreated (black circles, n=130 cells/5 coverslips), cells pretreated with 10 μM TAK-243 without (red triangles, n=136 cells/5 coverslips) or with 1 μM BTP2 (blue triangles, n=79 cells/3 coverslips). Mean +/− SEM, when not visible error bars are smaller than symbols. In these experiments statistics were conducted on Fura2 values measured at time +80 s. *** p<0.001, one way ANOVA followed by a Dunn’s test (vs TAK-243-untreated cells) performed on individual cells.

Similar experiments were performed on HEK-293T cells that do not generate spontaneous Ca²⁺ spikes (Figure 5G-J). Likewise DU-145 cells, a 2h/37°C pretreatment of HEK-293T cells with 10 μM TAK-243 elevated their resting levels of [Ca²⁺]i (Figure 5H). However, no augmentation of [Ca²⁺]i could be seen in the presence of the store-operated Ca²⁺ channels inhibitor BTP2 (1 μM) (Figure 5I). The results of these experiments are presented in Figure 5J showing the mean [Ca²⁺]i of HEK-293T cells for each experimental condition. Altogether, the experiments depicted in Figure 5 show that a 2h/37°C pretreatment with 10 μM TAK-243 significantly elevated resting [Ca²⁺]i levels in both DU-145 and HEK-293T cells via a mechanism sensitive to the store-operated Ca²⁺ channels inhibitor BTP2. These findings indicate that inhibiting E1 enzymes with TAK-243 promotes the entry of Ca²⁺ through store-operated Ca²⁺ channels.

In the following experiments we checked more directly the effects of TAK-243 on store-operated Ca²⁺ channels of DU-145 cells (Figure 6A). The ER Ca²⁺ stores were depleted with thapsigargin before adding back Ca²⁺ to the recording medium. The re-admission of external Ca²⁺ was followed by a Ca²⁺ rise that was fully suppressed by 5 μM BTP2 (Figure 6A), which confirms the presence of functional store-operated Ca²⁺ channels in DU-145 cells. This protocol allows to analyse separately the size of the pool of Ca²⁺ mobilizable from the ER and the entry of Ca²⁺ through store-operated Ca²⁺ channels. In the absence of external Ca²⁺, the basal [Ca²⁺]i was not affected by TAK-243 (Figure 6A), a finding which further supports the notion that TAK-243 elevated resting [Ca²⁺]i (Figure 5A, E) via a dysregulated Ca²⁺ entry route. TAK-243 (10 μM, green line) significantly upregulated store-operated Ca²⁺ entry (Figure 6A). The effects of a TAK-243 treatment on the size of the ER Ca²⁺ stores and on the Ca²⁺ signal generated by store-operated Ca²⁺ channels were further analysed in Figures 6B-C and 6D-E, respectively. Concerning the ER Ca²⁺ stores, for each cell two parameters were quantified: the peak amplitude of the thapsigargin-induced Ca²⁺ signal (Figure 6B) and the size of the thapsigargin-induced Ca²⁺ transient, measured as the AUC (Figure 6C). At the concentration of 10 μM TAK-243 significantly enhanced the maximal amplitude of the thapsigargin-induced Ca²⁺ signal (Figure 6B) but reduced the magnitude of the Ca²⁺ transient (Figure 6C) by accelerating the kinetics of the return to basal Ca²⁺ levels (green line, Figure 6A). The augmentation of the peak response could reflect changes in the kinetics of release from the ER as already reported by [9] instead of a change of the size of the mobilizable Ca²⁺ pool. The peak amplitude and the AUC of the store-operated Ca²⁺ signal were also analysed and the results are shown in Figures 6D and 6E, respectively. Both parameters were significantly augmented after a treatment with TAK-243 (10 μM) (Figure 6D-E). In conclusion, TAK-243 causes a constitutive Ca²⁺ entry through BTP2-sensitive store-operated Ca²⁺ channels that elevates [Ca²⁺]i. This Ca²⁺ rise stimulates (e.g. with 0.2 μM TAK-243, Figure 5B) or represses (e.g. with 10 μM TAK-243, Figure 5C) the spontaneous Ca²⁺ activity of DU-145 cells.

Figure 6.

TAK-243 accelerates the clearance of Ca²⁺ and upregulates the store-operated Ca²⁺ entry

Panel A, cells were maintained in a nominally Ca²⁺-free medium supplemented with 1 mM EDTA. When indicated (horizontal gray bar) thapsigargin (1 μM) was applied to deplete the ER Ca²⁺ stores. This was followed by the re-addition of 2 mM external Ca²⁺. Control cells, black lines (n=225 cells/9 coverslips). In some experiments, BTP2 (5 μM) was present during the Fura2 recording (gray lines, n=72 cells/3 coverslips). Some dishes were pre-treated for 2h/37°C with 0.2 μM (red line, n=162 cells/6 coverslips) or 10 μM TAK-243 (green line, n=159 cells/6 coverslips) prior to the recordings. Panels B-C and D-E show the analysis of the experiments illustrated in A, namely the individual thapsigargin-evoked Ca²⁺ release (B and C) and Ca²⁺ entry (D and E), respectively. Fura2 signals were analysed as the peak responses (B, D) and AUC (C, E). * p<0.05 One-way ANOVA followed by a Dunn’s test (vs TAK-243-untreated cells) performed on individual cells.

TAK-243 and ER Ca²⁺ stores

It is now well-accepted that conditions that empty the ER Ca²⁺ stores trigger the translocation of the ER Ca²⁺ sensors STIM followed by the activation of cell surface Orai channels that are key components of store-operated Ca²⁺ channels. The presence of an ongoing BTP2-sensitive Ca²⁺ entry induced by TAK-243 could therefore be caused by a sustained depletion of the ER Ca²⁺ stores. We checked this possibility by applying thapsigargin to 10 μM TAK-243-treated cells. In Figure 7A are shown individual superimposed Fura2 traces from thirty DU-145 cells that had been treated with 10 μM TAK-243 (2h/37°C) prior to the recordings. Under these conditions thapsigargin was still able to cause a robust and transient elevation of [Ca²⁺]i, indicating that the ER Ca²⁺ stores of the TAK-243-treated cells were not emptied. Figure 7B shows averaged Fura2 signals induced by thapsigargin on 10 μM TAK-243-treated DU-145 cells in the absence (open triangles) or presence (filled triangles) of external Ca²⁺. As illustrated, the peak amplitude of the thapsigargin-evoked Ca²⁺ response was not affected by the removal of external Ca²⁺ (Figure 7C). However, the recovery phase to resting basal Ca²⁺ levels was faster in the presence of external Ca²⁺ (filled triangles, Figure 7B), indicating a more efficient clearance of cytosolic Ca²⁺ rises. The size of the thapsigargin-evoked Ca²⁺ signal was quantified as the AUC (Figure 7D). It appeared that the thapsigargin-induced Ca²⁺ transients were significantly reduced in the presence of external Ca²⁺ most likely due to a faster return to basal Ca²⁺ levels.

Figure 7.

TAK-243 fails to affect ER Ca²⁺ stores

Fura2-loaded DU-145 cells pretreated with 10 μM TAK-243 (2h/37°C) were maintained in a 2 mM Ca²⁺-containing saline. When indicated (horizontal gray bar) thapsigargin (1 μM) was applied to deplete the ER Ca²⁺ stores. Panel A shows thirty representative Fura2 recordings from distinct cells in the same field of view. Panel B shows the thapsigargin-induced elevation of [Ca²⁺]i in DU-145 cells treated with 10 μM TAK-243 in the presence (blue triangles, n=177 cells/7 coverslips), or absence of external Ca²⁺ (open triangles, n=159 cells/6 coverslips, same set of data as in Figure 6A, green response). Ca²⁺ changes as a function of time were expressed as a delta ratio F340/F380 (ΔF340/F380), mean +/− SEM. Panel C shows, for each cell, the maximal Fura2 response (ΔF340/F380) in each condition. mean +/− SEM. Student’s t test performed on individual cells, n.s. difference not statistically significant. The size of the total thapsigargin-induced [Ca²⁺]i responses without or with external Ca²⁺ were measured as the AUC, and the results are reported in panel D. ***p<0.001, Student’s t test performed on individual cells.

Panels E and F show one hundred superimposed CEPIA1er traces representing the relative fluorescence unit (RFU) signal normalized to the first time point from CEPIA1er in transfected HEK-293T cells. Cells were allowed to rest for two minutes to establish a baseline, then stimulated with either 10 μM TAK-243 (panel F) or its vehicle control (panel E) in 1X HBSS in the presence of 2 mM external Ca²⁺ followed by stimulation with 1 μM ionomycin as a positive control to deplete the ER stores. Panel G shows averaged CEPIA1er signals from control (open circles, n = 512 cells/9 coverslips) and TAK-243 treated cells (blue triangles, n = 512 cells/9 coverslips), mean +/− SEM. Panels H and I show one hundred superimposed CEPIA1er traces from control (panel H) and TAK-243-pretreated cells (panel I) (2h/37°C). Cells were allowed to rest for one minute in the presence of 1X HBSS 2 mM extracellular Ca²⁺ prior to exchanging buffers to nominally Ca²⁺-free 1X HBSS and subsequent stimulation with 2 μM thapsigargin to maximally deplete the stores. Panel J shows an average of CEPIA1er signals from control (open circles, n = 512 cells/9 coverslips) and TAK-243 pretreated cells (blue triangles, n = 512 cells/9 coverslips), mean +/− SEM.

Panels K-M, DU-145 cells transfected with G-CEPIA1er were maintained in a recording saline containing 2 mM Ca²⁺ before being bathed in a nominally Ca²⁺-free solution. Thapsigargin (1 μM) was applied when indicated (black arrow head) to deplete the ER Ca²⁺ stores. One hundred representative CEPIAer traces are shown from control (panel K) and TAK-243-treated DU-145 cells (panel L). Panel M is a summary of the experiments, with control cells (open circles, n=107 cells from 23 coverslips) and TAK-243-treated cells (2h/37°C) (blue triangles, n=102 cells from 24 coverslips).

We used the genetically encoded ER-targeted dye CEPIA1er [27] [28] in HEK-293T cells to further characterize the effect of TAK-243 on ER Ca²⁺ stores. TAK-243 (Figure 7F) (or its vehicle, Figure 7E) was added at the concentration of 10 μM (horizontal white bar, Figure 7E-G) before applying the Ca²⁺ ionophore ionomycin (1μM, arrow head, Figure 7E-G) to fully deplete internal Ca²⁺ stores. As shown in Figure 7F, an acute application of 10 μM TAK-243 did not evoke the release of Ca²⁺ from internal stores. It also failed to modify the ionomycin response, indicating that the acute application of 10 μM TAK-243 did not affect the size of the internal Ca²⁺ stores (Figure 7G). Next, HEK-293T cells were pre-treated with DMSO (Figure 7H) or 10 μM TAK-243 (2h/37°C) (Figure 7I) prior to the application of thapsigargin (2 μM, arrow head) to deplete the ER Ca²⁺ stores. The thapsigargin-evoked Ca²⁺ release was unaffected by TAK-243 (Figure 7J). Similar experiments were conducted on CEPIA1er-transfected DU-145 cells (Figure 7K-M). Like the case for HEK-293T cells, the thapsigargin-evoked Ca²⁺ release of DU-145 cells was unaffected by TAK-243 (Figure 7M). Of note, the treatment with TAK-243 seemed to accelerate the kinetics of ER Ca²⁺ release in DU-145 cells. This observation is in line with a previous work showing that TAK-243 accelerates the kinetics of ER Ca²⁺ release in αT3, HeLa and SH-SY5Y cells [9]. Collectively, these experiments conducted with CEPIA1er to directly assess ER Ca²⁺ stores indicate that TAK-243 did not affect the size of the mobilizable pool of ER Ca²⁺ of HEK-293T and DU-145 cells. Therefore, a TAK-243 treatment seems able to activate a Ca²⁺ entry route sensitive to BTP2 without mobilizing Ca²⁺ from stores.

The ER Ca²⁺ sensor STIM1 and the cell surface Ca²⁺ channel Orai1 are critical molecular components of store-operated Ca²⁺ channels. It was hypothesised that TAK-243 could stimulate Ca²⁺ entry by favouring the coupling of STIM1 and Orai1. This hypothesis was tested using Förster resonance energy transfer (FRET) measurements [25]. HEK-293T cells were transfected with ORAI1-CFP and STIM1-YFP (Figure 8A). Figure 8B shows that TAK-243 significantly increased the basal FRET between CFP-tagged ORAI1 and STIM1-YFP (p<0.01), as well as the maximum puncta intensity (p<0.05, Figure 8C). Overall, our data suggest that, while TAK-243 does not alter ER Ca²⁺ stores (Figure 7), it does favour the coupling between STIM1 and Orai1 (Figure 8), at least in HEK-293T cells, and promotes Ca²⁺ entry through BTP2-sensitive channels (Figures 5 and 6).

Figure 8.

FRET measurements on ORAI1-CFP and STIM1-YFP transfected cells

Panel A shows representative images of cells expressing Orai1-CFP, STIM1-YFP as well as FRET signals after either treatment with vehicle control or 10 μM TAK-243. The number of Orai1-CFP/STIM1-YFP puncta per individual HEK-293T cell was quantified by ImageJ. Panel B shows the superplot of the average number of puncta per biological replicate overlaying the quantification of each individual cell per condition: after stimulation with either vehicle control (dark grey n = 179/9 coverslips) or 10 μM TAK-243 (light grey n = 179/9 coverslips). Data were analysed by two-tailed unpaired t-test (** p <0.01). Panel C, superplot of the quantification of the maximum FRET intensity normalized to signal intensity prior to stimulation with either vehicle control (dark grey n = 179/9 coverslips) or 10 μM TAK-243 (light grey n = 179/9 coverslips). Data were analysed by two-tailed unpaired t-test (* p <0.05) on individual cells.

Mitochondria participates in the clearance of cytosolic Ca²⁺ rises

As illustrated in Figures 6A and 7A, thapsigargin caused a transient elevation of [Ca²⁺]i that consisted in an initial rapid rise of [Ca²⁺]i followed by a slower return to basal levels. We noticed that 10 μM TAK-243 significantly accelerated the recovery phase (green line, Figure 6A), indicating that the UAE inhibitor influenced the clearance of cytosolic Ca²⁺ rises. Ca²⁺ released from the ER can be exported via the plasma membrane Ca²⁺ ATPase (PMCA) and/or stored into mitochondria. We thus verified whether TAK-243 influenced the extrusion and/or the mitochondrial Ca²⁺ uptake. Thapsigargin was applied without or with 1 mM lanthanum chloride (LaCl₃, a concentration known to block both Ca²⁺ influx and efflux through the PMCA [29]) or with 1 μM oligomycin (inhibitor of the mitochondrial ATP synthase) + 5 μg/mL antimycin A (inhibitor of mitochondrial electron transport complex III), to inhibit the mitochondrial Ca²⁺ uptake. Following thapsigargin application resting [Ca²⁺]i levels increased rapidly and then slowly declined (open circles, Figure 9A). The thapsigargin-induced Ca²⁺ responses were not affected by 1 mM La³⁺ (Figure 9C-D). However, perturbing the mitochondrial activity with oligomycin+antimycin A (Oligo/AA) significantly prolonged the recovery phase to resting Ca²⁺ levels (p<0.05, Figure 9B, D). This indicates that mitochondria participate in the clearance of cytosolic Ca²⁺ rises with very little or no contribution from the PMCA. Thus, mitochondria of DU-145 cells function as an efficient Ca²⁺ sequestration system. From these results, we hypothesized that TAK-243 accelerated the clearance of Ca²⁺ by fostering the mitochondrial Ca²⁺ uptake.

Figure 9.

Role of mitochondria in the clearance of Ca²⁺ and effects of TAK-243 on Δψm.

Panels A-C show one hundred superimposed Fura2 traces from DU-145-loaded cells that were maintained in a Ca²⁺-free medium. When indicated, 1 μM thapsigargin was applied without (Control, panel A), or with 1 μM oligomycin + 5 μg/mL antimycin A (to inhibit the mitochondrial Ca²⁺ uptake) (Oligo/AA, panel B), or with 1 mM lanthanum chloride (to block Ca²⁺ efflux via PMCA) (La³⁺, panel C). Panel D is a summary of these experiments with thapsigargin (open circles, n=132 cells/7 coverslips); thapsigargin + 1 mM La³⁺ (gray triangles, n=110 cells/5 coverslips), and thapsigargin + oligomycin + antimycin A (Oligo/AA) (blue triangles, n=166 cells/7 coverslips). p <0.05 Student’s t test performed on individual cells. Mean +/− SEM, when not visible error bars are smaller than symbols.

Panels E-G show real-time changes of the mitochondrial membrane potential (Δψm) of DU-145 cells loaded with the fluorescent probe Rh123 in response to the application of 1 μM FCCP (horizontal gray bar). The cationic fluorescent probe Rh123 accumulates within mitochondria in inverse proportion to Δψm. The application of FCCP depolarizes mitochondria and causes the release of Rh123 which increases the fluorescent signal. Panels E and F show one hundred superimposed Rh123 traces from control (panel E) and 10 μM TAK-243-treated cells (2h/37°C) (panel F). Panel G, summary of the experiments with control cells (open squares, n=211 cells/9 coverslips) and TAK-243-treated cells (blue squares, n=186 cells/9 coverslips). p <0.01 Student’s t test performed on individual cells. Mean +/− SEM, when not visible error bars are smaller than symbols.

Panels H-J show one hundred superimposed Fura2 recordings from control (panel H) and 10 μM TAK-243-treated DU-145 cells (panel I). Cells were maintained in a Ca²⁺-free medium before applying FCCP (1 μM, horizontal gray bar) to mobilize Ca²⁺ from mitochondria. Panel J, summary of the experiments with control cells (open squares, n=236 cells/9 coverslips) and TAK-243-treated cells (blue squares, n=281 cells/9 coverslips). p <0.01 Student’s t test performed on individual cells. Mean +/− SEM, when not visible error bars are smaller than symbols.

TAK-243 influences the mitochondrial membrane potential

Schematically, two important parameters control the storage of Ca²⁺ by mitochondria: the mitochondrial membrane potential (Δψm) and the Ca²⁺ concentration gradient between the cytosol and the matrix. In the following experiments, we investigated whether TAK-243 influenced Δψm. To that aim real-time changes of Δψm were monitored with the fluorescent probe Rh123 in situ (in live DU-145 cells). Rh123 is a positively charged (cationic) fluorescent probe that accumulates within mitochondria with hyperpolarized mitochondria accumulating more Rh123 than less polarized mitochondria [21]. The depolarization of mitochondria causes the release of Rh123 which increases the fluorescent signal due to the unquenching of the fluorescent probe. In these experiments, the protonophore and uncoupling agent FCCP (1 μM) was used to induce a complete breakdown of Δψm. As illustrated in Figure 9E the FCCP-induced collapse of Δψm was associated with an elevation of the Rh123 fluorescence. After a TAK-243 treatment FCCP evoked larger Rh123 signals suggesting that TAK-243 increased Δψm (p<0.01, Figure 9F-G). The collapse of Δψm is associated with a release of Ca²⁺ from mitochondria [30] [31]. We thus checked the effect of TAK-243 on the mitochondrial pool of mobilizable Ca²⁺ (Figure 9H-J). As expected, the external application of FCCP recruited Ca²⁺ stored in mitochondria (open squares, Figure 9H). A TAK-243 treatment diminished this Ca²⁺ pool (filled squares, Figure 9I-J, p<0.01). Taken together, these experiments indicate that TAK-243 causes a hyperpolarization of the mitochondrial membrane potential and depresses the mitochondrial pool of mobilizable Ca²⁺, which heightens mitochondrial Ca²⁺ uptake.

Discussion

DU-145 cells generate spontaneous Ca²⁺ transients. We showed here that the removal of extracellular Ca²⁺ decreases resting [Ca²⁺]i and strongly inhibits the occurrence of these spontaneous Ca²⁺ spikes. Furthermore, the Ca²⁺ channel blockers La³⁺ and Gd³⁺, and the Orai channel blocker BTP2 diminish [Ca²⁺]i and abolish this Ca²⁺ activity. This demonstrates the presence of a tonic Ca²⁺ entry, which is likely mediated by store-operated Ca²⁺ channels. DU-145 cells are known to express functional store-operated Ca²⁺ channels [32] [33] with STIM1 and Orai1 being the core components of this Ca²⁺ entry pathway in this cell line [34] [35]. Our data suggest that the constitutive influx of Ca²⁺ observed in DU-145 cells that develop through store-operated Ca²⁺ channels contributes to the maintenance of elevated resting levels of [Ca²⁺]i and are required for the generation of spontaneous Ca²⁺ spikes. Additionally, we demonstrated that preventing IP3 production nearly fully suppressed the spontaneous Ca²⁺ activity of DU-145 cells without affecting basal [Ca²⁺]i levels. Our data indicate the existence of an ongoing PLC-dependent IP3 production and ER Ca²⁺ release. Moreover, inhibiting SERCA activity with thapsigargin depletes the ER Ca²⁺ stores, elevates [Ca²⁺]i and suppresses the Ca²⁺ spiking activity of DU-145 cells. This highlights the contribution of SERCA to refuel the PLC-dependent ER Ca²⁺ release process. Likewise the PLC inhibitor U73122, the RyRs inhibitor dantrolene strongly attenuated the occurrence of spontaneous Ca²⁺ transients. This suggests that the IP3-dependent release of Ca²⁺ via IP3R is amplified by the release of Ca²⁺ through RyRs. The activity of IP3Rs is not only governed by IP3 but is also regulated by cytosolic Ca²⁺ concentrations: depending on the concentration of Ca²⁺ near the mouth of the IP3R channels, low and high Ca²⁺ concentrations stimulate and inhibit the ER Ca²⁺ release, respectively. This Ca²⁺-dependent regulation of the Ca²⁺ release from IP3R is known to play a role in the generation of Ca²⁺ oscillations [36] [37].

A recent study identified the Na⁺ channel NALCN as a critical factor (upstream of the Na⁺/Ca²⁺ exchanger) permitting the generation of Ca²⁺ oscillations in the human prostate cancer cell line PC-3 in response to FBS application [17]. However, removing extracellular Na⁺ only slightly decreases resting Ca²⁺ levels and the occurrence of Ca²⁺ spikes in DU-145 cells, pointing towards a different mechanism in this cell line. Taken together, our results demonstrate that IP3 production, ryanodine receptors, SERCA and store-operated Ca²⁺ channels form the core components of the Ca²⁺ oscillatory responses of DU-145 cells with a minor contribution of the Na⁺/Ca²⁺ exchanger.

The periodic Ca²⁺ spiking activity of cells is known to play important physiological functions. For instance, Ca²⁺ oscillations control the activity of Ca²⁺-dependent transcription factors such as NF-AT or NF-kB, which in turn, regulate the expression of genes [38]. Spontaneous Ca²⁺ transients also control various signalling pathways such as the MAPK-, CaMKII- or calpain-dependent pathways. Ca²⁺ oscillations can also be transmitted to mitochondria, impacting mitochondrial energy metabolism. Therefore, cytosolic Ca²⁺ dynamics are physiological regulators of cytosolic, nuclear and mitochondrial signalling.

Since Ca²⁺ oscillations of cancer cells drive tumour growth [39], we investigated the effects of the UBA1 inhibitor TAK-243 on the cytosolic Ca²⁺ signalling of DU-145 cells. A TAK-243 treatment causes a concentration-dependent elevation of [Ca²⁺]i mediated by a delivery of Ca²⁺ via BTP2-sensitive store-operated Ca²⁺ channels. A low concentration of TAK-243 (e.g. 0.2 μM) significantly elevates resting [Ca²⁺]i and stimulates the Ca²⁺ spiking activity of DU-145 cells. At higher concentrations (e.g. 10 μM) TAK-243 induces a more robust elevation of [Ca²⁺]i combined with a nearly complete abolition of the spontaneous Ca²⁺ activity. FRET imaging experiments indicated that TAK-243 promotes the coupling between STIM1 and Orai1 channels in HEK-293T cells. In addition, TAK-243 up-regulates the entry of Ca²⁺ through store-operated Ca²⁺ channels. However, ER Ca²⁺ imaging with CEPIA1er demonstrated that the size of the mobilizable pool of Ca²⁺ in the ER is unaffected by TAK-243. Furthermore, TAK-243 accelerates the clearance of the thapsigargin-induced Ca²⁺ rises, showing that the UBA1 inhibitor influences not only the uptake of Ca²⁺ but also Ca²⁺ clearance mechanisms. Experiments conducted with La³⁺ show that PMCA has no or negligible role in Ca²⁺ clearance. However, perturbing mitochondrial activity significantly modifies the recovery of Ca²⁺ to resting levels, suggesting that these organelles participate in the clearance of cytosolic Ca²⁺ rises. Interestingly, a TAK-243 treatment increases mitochondrial membrane potential and diminishes the mitochondrial pool of Ca²⁺ (which alters the Ca²⁺ gradient between the cytosol and mitochondrial matrix). These effects favour an enhanced mitochondrial uptake of Ca²⁺, which is likely to influence mitochondrial activities such as the oxidative phosphorylation, the oxygen consumption, the production of ATP and ROS, or the lipid metabolism [40]. Cytosolic Ca²⁺ rises are known to inhibit store-operated Ca²⁺ channels through a Ca²⁺-dependent inactivation mechanism [41,42]. The TAK-243-dependent hyperpolarization of the mitochondrial membrane potential, in turn, ensures an efficient mitochondrial Ca²⁺ uptake, preventing or minimizing slow Ca²⁺-dependent inactivation of store-operated Ca²⁺ channels, allowing large and long-lasting Ca²⁺ influxes to occur. Since excessive elevations of Ca²⁺ have long been recognized as a pivotal signal for cell death via apoptosis, it is proposed that TAK-243 is likely to cause a harmful Ca²⁺ overload.

Collectively, this work shows that the ubiquitin-activating enzyme inhibitor TAK-243 triggers an ongoing Ca²⁺ entry through store-operated Ca²⁺ channels without altering the ER Ca²⁺ stores. In HEK-293T cells, TAK-243 modifies the localization of the ER resident protein STIM1, which promotes its association with Orai1. TAK-243 also interferes with the functioning of mitochondria. These cellular responses elevate the resting [Ca²⁺]i and perturb the spontaneous Ca²⁺ spiking activity of DU-145 cells. Our work deepens the understanding of the mechanisms underlying Ca²⁺ oscillations in cancer cells and identify the ubiquitin proteasome system as a major actor of this process, providing new molecular clues to understand the anti-tumoral effect of the UBA1 inhibitor TAK-243. Although it has been described as a selective ubiquitin-activating enzyme (UAE) inhibitor, we cannot exclude the possibility that some of the cellular effects of TAK-243 may reflect off-target activity [43,44] . Future studies may clarify how the ubiquitination status modifies mitochondrial properties and the functioning of store-operated Ca²⁺ channels.

Highlights.

• TAK-243, a ubiquitin-activating enzyme inhibitor, prevents the ubiquitination;

• We studied its effects on the Ca²⁺ signalling of human prostate cancer DU-145 cells;

• TAK-243 activates store-operated Ca²⁺ channels without depleting ER Ca²⁺ stores;

• This gives rise to an ongoing Ca²⁺ entry that elevates [Ca²⁺]i;

• TAK-243 also interferes with the mitochondrial ability to uptake and store Ca²⁺;

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.