Targeting TRPM2 Channels Impairs Radiation-Induced Cell Cycle Arrest and Fosters Cell Death of T Cell Leukemia Cells in a Bcl-2-Dependent Manner

Messenger RNA data of lymphohematopoietic cancer lines suggest a correlation between expression of the cation channel TRPM2 and the antiapoptotic protein Bcl-2. The latter is overexpressed in various tumor entities and mediates therapy resistance. Here, we analyzed the crosstalk between Bcl-2 and TRPM2 channels in T cell leukemia cells during oxidative stress as conferred by ionizing radiation (IR). To this end, the effects of TRPM2 inhibition or knock-down on plasma membrane currents, Ca2+ signaling, mitochondrial superoxide anion formation, and cell cycle progression were compared between irradiated (0–10 Gy) Bcl-2-overexpressing and empty vector-transfected Jurkat cells. As a result, IR stimulated a TRPM2-mediated Ca2+-entry, which was higher in Bcl-2-overexpressing than in control cells and which contributed to IR-induced G2/M cell cycle arrest. TRPM2 inhibition induced a release from G2/M arrest resulting in cell death. Collectively, this data suggests a pivotal function of TRPM2 in the DNA damage response of T cell leukemia cells. Apoptosis-resistant Bcl-2-overexpressing cells even can afford higher TRPM2 activity without risking a hazardous Ca2+-overload-induced mitochondrial superoxide anion formation.


Introduction
Transient Receptor Potential (TRP) Cation Channels. The TRP superfamily comprises a diverse range of Ca 2+ -permeable cation channels [1]. TRP channels contribute to changes in cytosolic free Ca 2+ ( free [Ca 2+ ] i ) by directly acting as Ca 2+ entry channels in the plasma membrane or by changing membrane potentials, modulating the activity and/or driving forces for the Ca 2+ entry channels [2]. The melastatin subfamily (TRPM) has been subdivided into three subgroups on the basis of sequence homology (TRPM1/TRPM3, TRPM4/ TRPM5, and TRPM6/7) with TRPM8 and TRPM2 being distinct proteins [3]. The Ca 2+ -permeable TRPM2 channels, formerly known as TRPC2 and LTRPC2, were first identified in 1998 [4]. Reactive oxygen species (ROS) have been demonstrated to induce TRPM2 currents and increase free [Ca 2+ ] i in various cell types transfected with TRPM2 [5], as well as in pancreatic -cells [6], neutrophil granulocytes [7], and U937 monocytes [8].
TRPM2 and Cell Death. By increasing free [Ca 2+ ] i , TRPM2 may increase the susceptibility to cell death suggesting that TRPM2 channels function as "death channels." As a matter of fact, heterologous expression of TRPM2 in human embryonic kidney cells [9] or A172 human glioblastoma cells [10] facilitates oxidative stress-induced cell death. Moreover, expression of TRPM2 has been demonstrated in several tumor entities such as insulinoma [6], hepatocellular carcinoma [6],

Direct and Indirect Oxidative Stress Conferred by Ionizing
Radiation. Most energy of ionizing radiation (IR) is absorbed by cell water leading to formation of hydroxyl radicals (for review see [27]). Oxidative stress-and DNA repair-associated release of ADP-ribose is supposed to increase the plasma membrane Ca 2+ permeability by activating TRPM2 channels. Subsequent changes in free [Ca 2+ ] i and mitochondrial function are under the control of Bcl-2. Together, this hints to a crosstalk between Ca 2+ and ROS signaling involving TRPM2 Ca 2+ -permeable channels in the plasma membrane, the Ca 2+regulated ΔΨ across the inner mitochondrial membrane, and the antiapoptotic protein Bcl-2 in the ER and outer mitochondrial membrane of irradiated cells.
Aim of the Study. The present study aimed to define this crosstalk in human T cell leukemia cells subjected to ionizing radiation. To this end, Jurkat cells stably transfected with Bcl-2 or the empty control vector were irradiated with 0, 5, or 10 Gy by 6 MV photons. Ion channel activity, Ca 2+ signaling, mitochondrial superoxide anion formation, cell cycle control, and cell death were assessed by patch-clamp wholecell recording, fura-2 Ca 2+ imaging, immunoblotting, and flow cytometry in irradiated and nonirradiated cells, respectively. In addition, mRNA data of hematopoietic and lymphoid tissue cancer cell lines of the Novartis and Broad Institute Cancer Cell Line Encyclopedia were queried for TRPM2 and Bcl-2 mRNA abundance.

Transfection with siRNA.
Transfection with siRNA was performed as described [31]. In brief, cells were cultured at a low density to ensure log phase growth. For transfection, 2 × 10 6 cells were resuspended in 200 L RPMI 1640 without phenol red. Shortly before transfection, TRPM2 or nontargeting siRNA was added at a concentration of 1 M. TRPM2 ON-TARGET SMARTpool and the siCONTROL NON-TARGETING pool siRNA were purchased from Dharmacon (Chicago, IL, USA). Cells were electroporated in a 4 mm cuvette in an EPI2500 electroporator (Fischer, Heidelberg, Germany) at 370 V for 10 ms. Immediately after transfection, cells were resuspended in 6 mL prewarmed medium and continued to be cultured as described above. Transfection efficiency as well as viability was determined by transfecting the cells with 400 nM green fluorescence siRNA (siGLO from Dharmacon, Chicago, IL, USA) followed by propidium iodide exclusion dye and flow cytometric analysis.

Patch-Clamp
Recording. Maternal, Bcl-2-overexpressing, and control vector-transfected Jurkat cells were irradiated (IR) with 0, 5, or 10 Gy 6 MV photons by the use of linear accelerator (LINAC SL25 Philips) at a dose rate of 4 Gy/min at room temperature. Whole-cell currents were evoked by 9-11 voltage pulses (700 ms each) to voltages between −100 (−80) mV and +100 (+80) mV delivered in 20 mV increments. Mean steady state current values were analyzed 2-49 h after IR. The liquid junction potentials between the pipette and the bath solutions were estimated according to [32], and data were corrected for the estimated liquid junction potentials. Applied voltages refer to the cytoplasmic face of the membrane with respect to the extracellular space. Inward currents, defined as flow of positive charge from the extracellular to the cytoplasmic membrane face, are negative currents and depicted as downward deflections of the original current traces.

Querying the Cancer Genome Atlas (TGCA) Data Sets.
Via the cBIOportal Web resource [33,34], 178 hematopoietic and lymphoid tissue cancer cell lines of the Novartis and Broad Institute Cancer Cell Line Encyclopedia [35] were queried for TRPM2 and Bcl-2 mRNA abundance.

Flow Cytometry.
To test for mitochondrial production of superoxide anion, Jurkat cells were irradiated (0 or 10 Gy), further cultured for 6 h, harvested, washed, and incubated for 10 min at 37 ∘ C in NaCl ringer solution (see above) containing 5 M of the superoxide anion-sensitive dye MitoSOX (Invitrogen) and 0 or 20 M ACA, and superoxide anionsensitive fluorescence was recorded by flow cytometry in fluorescence channel Fl-2 (logarithmic scale, 488 nm excitation and 564-606 nm emission wavelengths). To confirm equal fluorescence dye loading, samples were oxidized (10 mM tert-butylhydroperoxide) for 12 min and recorded (data not shown).
To monitor mitochondrial function, Jurkat cells were irradiated (0 or 10 Gy) and further cultured for 6 h. Thereafter, cells were harvested, washed, and incubated for 30 min at 37 ∘ C in NaCl ringer solution (see above) containing 25 nM of the inner mitochondrial membrane potential (ΔΨ ) specific dye tetramethylrhodamine ethyl ester perchlorate (TMRE, Invitrogen) and ΔΨ was analyzed by flow cytometry in fluorescence channel FL-2 (logarithmic scale).
For determination of free [Ca 2+ ] i cells were loaded in NaCl ringer solution (see above) for 0.5 h with fluo-3-AM (2 M in NaCl ringer, Calbiochem; Bad Soden, Germany) and recorded in fluorescence channels FL-1 (logarithmic scale, 515-545 nm emission wavelengths). As loading control for fluo-3, cells were incubated with the Ca 2+ ionophore ionomycin (1 M for 10 min) prior to analysis by flow cytometry. Data were analyzed with the FCS Express 3 software (De Novo Software, Los Angeles, CA, USA).

Statistics.
Data are expressed as means ± SE and statistical analysis was made by normal or Welch-corrected twotailed -test or ANOVA where appropriate using InStat software (GraphPad Software Inc., San Diego, CA, USA).

Modulation of on Channel Activity by Ionizing Radiation.
To assess the effect of ionizing radiation (IR) of ion channel activity, Jurkat cells were irradiated with 10 Gy and wholecell currents were recorded at different time periods after IR. As shown in Figures 1(a) and 1(b), IR induced an increase in whole-cell currents 2-6 h after IR. Substitution of Na + in the bathing solution by Ca 2+ or the impermeable Na + substitute NMDG + indicated both cation-selectivity and Ca 2+ permeability of the IR-induced currents (Figures 1(c)-1(e)).
Next, the functional expression of TRPM2 channels and its dependence on Bcl-2 was determined in Jurkat cells. Such dependence was suggested by a positive correlation of the TRPM2 and Bcl-2 mRNA abundances in 178 hematopoietic and lymphoid tissue cancer cell lines of the Novartis and Broad Institute Cancer Cell Line Encyclopedia (Figure 2(a)).
In the Jurkat cell model, in contrast, TRPM2 protein abundance seemed to be lower in Bcl-2-overexpressing (Jurkat-Bcl-2) cells as in the control vector-transfected (Jurkat-vector) cells as suggested by immunoblotting ( Figure 2(b)). IR did not modify total TRPM2 protein content of the cells (Figure 2(b)).
To activate TRPM2 in Jurkat cells, whole-cell currents were recorded with the TRPM2 agonist ADP-ribose in the pipette and compared in unpaired experiments with those recorded under control conditions. Intracellular ADP-ribose stimulated a whole-cell current fraction which was sensitive to the unspecific TRPM2 inhibitor ACA [36] (Figures 2(c) and 2(d)). Importantly, ADP-ribose-stimulated currents exhibited unitary current transitions with a unitary conductance of some 50 pS as reported for heterologously expressed TRPM2 channels [37] (Figure 2(e)). Together, these data indicated functional expression of TRPM2 in Jurkat cells.

Mitochondrial Superoxide Anion Formation: Effect of Ionizing Radiation, Bcl-2 Overexpression, and TRPM2 Inhibition.
To assess IR-stimulated formation of superoxide anion by mitochondria and to estimate a potential role of TRPM2 channels herein, Jurkat-Bcl-2 and Jurkat-vector cells were irradiated (0 or 10 Gy), postcultured for 6 h, and incubated for 10 min with the superoxide anion-sensitive fluorescence dye MitoSOX. The dye incubation was performed in the absence or presence of the TRPM2 inhibitor ACA. As shown in Figure 3 Combined, these data demonstrate lower mitochondrial superoxide anion formation in Jurkat-Bcl-2 cells as compared to Jurkat-vector cells. Only in the former cells, IR induced an increase in superoxide anion formation. In addition, superoxide anion formation was lowered by the TRPM2 inhibitor ACA in cells of both clones independent of IR stress. This might suggest a contribution of TRPM2-mediated Ca 2+ uptake to mitochondrial ROS formation. In accordance with an IR-induced increase in TRPM2 activity, IR stimulated an ACA-sensitive Ca 2+ uptake as measured by fura-2 Ca 2+ imaging using an extracellular Ca 2+ depletion/repletion protocol (Figure 4(d)). In contrast to the ACA-sensitive basal whole cell currents (Figure 4(b)), basal (ACA-sensitive) Ca 2+ uptake was higher in Jurkat-Bcl-2 than in Jurkat-vector cells (compare 1st with 5th bar in Figure 4(e)). Similarly to the whole-cell currents, IR (5 Gy) stimulated a larger Ca 2+ uptake in Jurkat-Bcl-2 as compared to Jurkat-vector cells (compare 2nd with 6th bar in Figure 4(e)). In the presence of ACA, Ca 2+ uptake did not differ between control and irradiated Jurkat-vector and Jurkat-Bcl-2 cells (3rd, 4th, 7th, and 8th bar in Figure 4(e)). Together, these observations indicated an IR-stimulated Bcl-2-regulated Ca 2+ uptake in Jurkat cells which probably involves TRPM2 channels.

Role of TRPM2 Channels in Ionizing Radiation-Stimulated
Activation of Ca 2+ Effector Proteins Involved in Cell Cycle Arrest. This IR-stimulated Ca 2+ uptake might be hazardous for the cells leading to Ca 2+ overflow and subsequent cell death. In fact, 24 h after IR with 10 Gy, some 25% of the Jurkat cells      exhibited a highly increased free [Ca 2+ ] i as deduced from fluo-3 flow cytometry (Figures 5(a) and 5(b)). Ca 2+ uptake might also contribute to Ca 2+ signaling that is required for DNA damage response of the irradiated T cell leukemia cells. IR (5 Gy) stimulated autophosphorylation and activation of Ca 2+ /calmodulin-dependent protein kinase II (CaMKII) isoforms and phosphorylation-dependent inactivation of the CaMKII downstream target cdc25b as suggested by immunoblotting ( Figure 5(c)). Inactivation of the phosphatase cdc25b was parallel by radiation-induced phosphorylationdependent inactivation of the cdc25b substrate cdc2 ( Figure 5(c)). This might hint to an involvement of Ca 2+ effector proteins such as CaMKII in G 2 /M arrest as observed in PI flow cytometry 24 h after IR (5 Gy, Figure 5(d)).
To confirm an involvement of TRPM2 and CaMKII in the stress response of Jurkat cells, Jurkat-Bcl-2 and Jurkatvector cells were irradiated (0 or 5 Gy) and postincubated in the presence or absence of the TRPM2 inhibitors ACA or clotrimazole [38,39] and kinase activities of the CaMKII isoforms and cdc2 and cell cycle distribution and cell death were analyzed by immunoblotting and PI flow cytometry, 4 h and 24 after IR, respectively. ACA decreased the basal and radiation-induced abundance of phosphorylated CaMKII in Jurkat-Bcl-2 and Jurkat-vector cells (Figure 6(a), 1st and 2nd blot). Most importantly, ACA blocked the radiationinduced phosphorylation-dependent inactivation of cdc2 in both genotypes (Figure 6(a), 3rd blot) suggesting a functional significance of ACA-sensitive Ca 2+ entry for G 2 /M cell cycle arrest. Accordingly, ACA and clotrimazole decreased the number of irradiated cells arrested in G 2 /M (Figures 6(b)-6(e)) and increased the number of dead cells. ACA-and clotrimazole-induced cell death was more pronounced in irradiated Jurkat-vector than in Jurkat-Bcl-2 cells (subG 1 population, Figures 6(b)-6(e)).
Finally, the function of TRPM2 in radiation-induced G 2 / M arrest of Jurkat cells was directly tested by TRPM2 knockdown. Transfection of Jurkat-vector cells with TRPM2 siRNA resulted in downregulation of TRPM2 protein level as compared to nontargeting (nt) RNA-transfected cells (Figure 6(f), insert). Transfected Jurkat cells were irradiated (0 or 5 Gy) and G 2 /M arrest and cell death analyzed 24 h thereafter. TRPM2 knock-down exerted small but significant effects on radiation-induced G 2 /M arrest and cell death mimicking those of ACA and clotrimazole ( Figure 6(f)).

Regulation of TRPM2-Mediated Ca 2+ Influx by Mitochondria and Bcl-2.
ADP-ribose is liberated in the mitochondria from, for example, NAD-dependent deacetylation intermediates, from mono-or polyADP-ribosylated proteins, or from NAD + , and released into the cytosol [19]. Oxidative and nitrosative stress have been demonstrated to stimulate the mitochondrial release of ADP-ribose into the cytosol which in turn activates TRPM2 channels in the plasma membrane resulting in Ca 2+ entry and depolarization of the membrane potential [16]. Since an elevated free [Ca 2+ ] i may disinhibit the respiration change leading to ΔΨ hyperpolarisation and superoxide anion formation and, eventually, to mitochondrial Ca 2+ overload and ΔΨ dissipation, TRPM2 activation has been proposed to amplify signals that trigger cell death (for review see [27]).
The present study demonstrates that irradiated human T cell leukemia cells may utilize the TRPM2 "death channel" for prosurvival Ca 2+ signaling. Noteworthy, IR-induced TRPM2 currents and Ca 2+ entry were larger in cells overexpressing Bcl-2 pointing to a crosstalk between Bcl-2 in the ER and outer mitochondrial membrane and TRPM2 in the plasma membrane. The correlation between TRPM2 and Bcl-2 mRNA abundances in a panel of lymphohematopoietic cancer cell lines (see Figure 2(a)) further suggests a functional interdependence between both proteins.
In some cell models, Bcl-2-overexpressing cells have been proposed to counteract the Bcl-2-mediated Ca 2+ leakage from the stores by downregulating Ca 2+ uptake through the plasma membrane (for review see [19]). In line with such compensatory mechanism might be the observation of the present study that Bcl-2-overexpressing Jurkat cells exhibited under basal conditions lower TRPM2 protein abundance, smaller ACA-sensitive currents in patch-clamp whole-cell recordings than the control vector-transfected cells (see Figure 2(b)).
In intact cells (i.e., in fura-2 Ca 2+ imaging experiments, see Figure 4(e)), however, a basal ACA-sensitive Ca 2+ uptake fraction was only apparent in Bcl-2-overexpressing cells suggestive of a TRPM2 inactivity in control cells under resting conditions. Compared to control cells, the more sustained Ca 2+ uptake in Bcl-2-overexpressing cells (see Figures 4(d) and 4(e)) suggests that Bcl-2 overexpression might be asso-  [40]. The observed basal ACA-sensitive Ca 2+ uptake that occurred exclusively in Bcl-2-overexpressing cells might, therefore, be simply explained by a higher basal free [Ca 2+ ] i in Bcl-2-overexpressing as compared to control cells.
Noteworthy, despite higher basal free [Ca 2+ ] i , Bcl-2-overexpressing cells exhibited lower basal mitochondrial ROS formation than control cells (see Figure 3(c)) suggestive of a Bcl-2-mediated protection of mitochondrial superoxide anion formation. As a matter of fact, a direct promoting function of mitochondrial superoxide anion formation has been attributed to the Bcl-2 opponent Bax in neuronal cells [41].

Rearrangements of the Ca 2+ Signalosome in Tumor Cells: Functional Significance for Cell Cycle Control and Stress
Response. In many tumor entities rearrangements of the Ca 2+ signalosome have been reported. In prostate cancer, for instance, malignant progression is reportedly accompanied by TRPM8-mediated Ca 2+ store depletion and downregulation of store-dependent Ca 2+ entry across the plasma membrane. In exchange, TRP channels such as TRPV6 are upregulated in the plasma membrane of advanced prostate cancer cells which have been proposed to generate in concert with IK K + channels survival and growth factor-independent Ca 2+ signaling (for review see [42]).
In the present study, IR stimulated the ACA-sensitive currents of Jurkat cells in patch-clamp recordings and the ACA-sensitive Ca 2+ uptake in fura-2 imaging experiments suggesting an IR-induced increase in TRPM2 activity. IRinduced modifications of ion channel activity have been reported in different tumor entities where they contribute to stress evasion [43], glucose fueling [44,45], cell cycle control [46,47], or radioresistance [48].
The p53-mutated Jurkat cells [49] accumulate in G 2 M cell cycle arrest upon IR-mediated DNA damage (see Figure 5(d)). The proposed IR-stimulated Ca 2+ entry through TRPM2 channels most probably contributed to the G 2 M cell cycle arrest. This was evident from the observation of the present study that two nonspecific TRPM2 inhibitors or TRPM2 knock-down decreased the number of cells accumulating in G 2 and increased the number of dead cells (see Figure 6). One might speculate that TRPM2 inhibition or knock-down overrides G 2 M cell cycle arrest and forces cells with unrepaired DNA damage into mitosis. This scenario is strengthened by the observation that IR promoted the free [Ca 2+ ] i -dependent phosphorylation of CaMKIIs and their downstream targets cdc25b and cdc2 in an ACA-sensitive manner (see Figures 5(c) and 6(a)). CaMKIIs phosphorylate and thereby inactivate the phosphatase cdc25b which results in accumulation of the phosphorylated, inactive form of the mitosis promoting factor subunit cdc2 [47]. Combined, these observations suggest that IR-dependent TRPM2 activation contributes to Ca 2+ signals that are able to induce autophosphorylation and thereby activation of CaMKIIs.
Likewise, irradiated human myeloid leukemia cells have been shown to generate Ca 2+ signals by the concerted action of TRPV5/6 and Kv3.4 K + channels in the plasma membrane. These Ca 2+ signals program G 2 M cell cycle arrest similarly to proposed mechanism of the present study via CaMKIIs, cdc25b, and cdc2 [46,47]. K + channel activity in close vicinity to Ca 2+ entry pathways maintains a high inwardly directed driving force for Ca 2+ and, thus, is indispensable for robust Ca 2+ signals. In analogy to the leukemia cells [47], IR induced the coactivation of IK K + channels in the plasma membrane of Jurkat cells (see supplementary Figure B). This points to both a common signaling in irradiated myeloid and lymphoblastic leukemia cells and the possibility that functionally equivalent Ca 2+ signals can be generated during DNA damage response by different sets of TRP and K + channels in the plasma membrane.

3.7.
Conclusions. Plasma membrane TRPM2 channels have been attributed tumor suppressor function in several tumor entities. The Ca 2+ signalosome of human T cell leukemia cells comprises TRPM2 channels that are activated during DNA damage response. In particular, irradiated Jurkat cells utilize TRPM2 to control the G 2 /M cell cycle arrest probably via activation of the Ca 2+ effector protein CaMKII and subsequent inhibition of cdc25b and cdc2. The antiapoptotic protein Bcl-2 in the ER or outer mitochondrial membrane even fosters TRPM2 activity presumably by inducing higher free [Ca 2+ ] i levels and decreases at the same time mitochondrial ROS formation. By doing so, Bcl-2-overexpressing cells may harness TRPM2-generated Ca 2+ signals without running into the risk of hazardous mitochondrial ROS formation. Thus, Bcl-2 function on mitochondrial integrity and stress-induced TRPM2-mediated Ca 2+ signaling cooperate in resistance to radiation therapy in T cell leukemia cells.