Differential Redox Regulation of Ca²⁺ Signaling and Viability in Normal and Malignant Prostate Cells

Abstract

In prostate cancer, reactive oxygen species (ROS) are elevated and Ca²⁺ signaling is impaired. Thus, several novel therapeutic strategies have been developed to target altered ROS and Ca²⁺ signaling pathways in prostate cancer. Here, we investigate alterations of intracellular Ca²⁺ and inhibition of cell viability caused by ROS in primary human prostate epithelial cells (hPECs) from healthy tissue and prostate cancer cell lines (LNCaP, DU145, and PC3). In hPECs, LNCaP and DU145 H₂O₂ induces an initial Ca²⁺ increase, which in prostate cancer cells is blocked at high concentrations of H₂O₂. Upon depletion of intracellular Ca²⁺ stores, store-operated Ca²⁺ entry (SOCE) is activated. SOCE channels can be formed by hexameric Orai1 channels; however, Orai1 can form heteromultimers with its homolog, Orai3. Since the redox sensor of Orai1 (Cys-195) is absent in Orai3, the Orai1/Orai3 ratio in T cells determines the redox sensitivity of SOCE and cell viability. In prostate cancer cells, SOCE is blocked at lower concentrations of H₂O₂ compared with hPECs. An analysis of data from hPECs, LNCaP, DU145, and PC3, as well as previously published data from naive and effector T_H cells, demonstrates a strong correlation between the Orai1/Orai3 ratio and the SOCE redox sensitivity and cell viability. Therefore, our data support the concept that store-operated Ca²⁺ channels in hPECs and prostate cancer cells are heteromeric Orai1/Orai3 channels with an increased Orai1/Orai3 ratio in cells derived from prostate cancer tumors. In addition, ROS-induced alterations in Ca²⁺ signaling in prostate cancer cells may contribute to the higher sensitivity of these cells to ROS.

Introduction

Numerous studies have demonstrated a contribution of reactive oxygen species (ROS) to the development of cancer hallmarks. In prostate cancer, ROS levels are elevated and contribute to altered DNA and protein structures, enhanced epithelial cell proliferation, and neoplasia (1–5). Remarkably, even though ROS production in cancer cells is elevated, cancer cells (including prostate cancer cells) are more sensitive to oxidative stress than nonmalignant cells—a phenomenon that is utilized in the development of novel anticancer drugs (6,7). ROS-inducing substances and ROS scavengers have been investigated as therapeutics; however, the outcome and benefit of such strategies remain largely unclear (8). Therefore, a better understanding of the underlying mechanisms and key players in redox-regulated signaling pathways is required for future therapeutic approaches.

There are multiple links between ROS and the universal second messenger Ca²⁺ (9–11). In prostate cancer cells, ROS-induced signaling is well known to include elevated Ca²⁺. In PC3 prostate cancer cells, ROS was shown to induce an increase of intracellular Ca²⁺ levels, which is necessary for ROS-induced apoptosis (12). In DU145 cells, ROS-activated cell apoptosis depends on elevated Ca²⁺ signaling for a full response (13). Several Ca²⁺ transporters, including transient receptor potential (TRP) channels and inositol 1,4,5-trisphosphate receptors (IP₃R), which are activated and/or regulated by ROS, contribute to ROS-induced Ca²⁺ signaling (14–17). The cell-type-specific subset of Ca²⁺ transporters and the distinct and spatially complex regulation of ROS by ROS-producing and -scavenging enzymes ensure precise ROS-induced Ca²⁺ signaling patterns (14,18).

The main Ca²⁺ entry mechanism in nonexcitable cells is known as store-operated Ca²⁺ entry (SOCE). Upon Ca²⁺ release from internal Ca²⁺ stores, endoplasmic reticulum Ca²⁺ sensor proteins (e.g., stromal interaction molecule 1 (STIM1)) cluster and activate Orai1 Ca²⁺ channels that are located in the plasma membrane (19). The SOCE underlying current is referred to as Ca²⁺ release activated Ca²⁺ current (I_CRAC). Store-operated Orai1 channels have been described as either tetramers (20–25) or hexamers (26–29) in the past. Besides Orai1, Orai2 and Orai3 are ubiquitously expressed and form heteromers with Orai1 (30–33). Compared with homomeric Orai1 channels, heteromeric store-operated Orai1/Orai3 channels differ in certain properties, such as the Ca²⁺ current amplitude, ion selectivity, pharmacological profile, and ROS sensitivity (33–36). A very recent report demonstrated that one Orai3 subunit within a heteromeric channel complex is sufficient to completely abrogate the ROS sensitivity of I_CRAC (37). The ROS sensitivity of Orai1 has been attributed to the oxidation of one cysteine (Cys-195). Since Cys-195 is absent in Orai3, the Orai1/Orai3 expression ratio impacts the ROS-mediated block of SOCE and cellular viability upon ROS-mediated stress. In effector T cells, Orai3 is upregulated, as reflected by a decreased mRNA ratio (Orai1/Orai3 ratio ∼70 in naive T_H cells and ∼25 in effector T_H cells) (35). Subsequently, the IC₅₀ for the immediate H₂O₂-induced block of SOCE is shifted from 7 μM for naive T_H cells to 51 μM for effector T_H cells. As a physiological consequence, the lowered Orai1/Orai3 ratio increases the cellular viability upon oxidative stress (IC₅₀ = 39 μM H₂O₂ in naive T_H cells, and IC₅₀ = 199 μM H₂O₂ in effector T_H cells) (35). These relatively high levels of H₂O₂ seem to be physiologically relevant. Based on our recent findings (38,39), and taking into account previous concepts regarding the existence of ROS microdomains (40–42), it seems very likely that in inflamed tissues the levels of H₂O₂ might well exceed 200–300 μM.

Recently, we reported an outstandingly low Orai1/Orai3 mRNA ratio (∼4) in primary human prostate epithelial cells (hPECs) from healthy tissue and a downregulation of Orai3 in prostate cancer cell lines (Orai1/Orai3 ratio ∼26 in lymph node carcinoma of the prostate (LNCaP) and ∼17 in DU145) (34).

Here, we sought to determine whether ROS-induced Ca²⁺ signaling and SOCE in cells under oxidative stress are altered in prostate cancer. In addition, we investigated whether the low Orai1/Orai3 ratio in hPECs is associated with a low redox sensitivity of SOCE, and whether this sensitivity might be increased in prostate cancer cells.

Materials and Methods

Cell culture

This study was approved by the local ethics review board (approval No. 168/05, Ärztekammer des Saarlandes) and performed in accordance with the Declaration of Helsinki. Informed consent was obtained from all patients. Prostate tissue was obtained from prostectomy specimens, and hPECs obtained from healthy tissue were isolated and cultured according to Gmyrek et al. (43) with slight modifications (34). The prostate cancer lines LNCaP, DU145, and PC3 were purchased from the American Type Cell Culture Collection (ATCC, Rockville, MD). Cell lines were cultured with RPMI Medium 1640 (Life Technologies) supplemented with 10% fetal calf serum and 1% penicillin/streptomycin (Life Technologies).

Small interfering RNA transfection

Small interfering RNA (siRNA) transfections were performed as described previously (34). We performed the siRNA transfections with 0.12 nmol of siRNA using the Nucleofector II Transfection Kit R for hPECs and LNCaP, the Nucleofactor IV Kit SE for DU145, and the Kit SF for PC3 (all from Lonza) according to the manufacturer’s instructions. All siRNAs were obtained from Qiagen or Microsynth and were partially modified according to Mantei et al. (44). The Orai1 siRNAs were Hs_TMEM142A_1, #SI03196207 (sense: 5′OMeC-OMeG-GCCUGAUCUUUAUCGd (UCU)OMeU-OMeT-OMeT3′; antisense: 3′OMeG-OMeC-CGGACUAGAAAUAGCAGAd (A)5′) and Hs_TMEM142A_2, #SI04215316 (sense: 5′OMeC-OMeA-ACAUCGAGGCGGUGA) d(GCA) OMeA-OMeT-OMeT3′; antisense: 3′OMeG-OMeT-UGUAGCUCCGCCACUCGUd (U)5′). The Orai3 siRNAs were Hs_TMEM142C_2, #SI04174191 (sense: 5′OMeC-OMeA-CCAGUGGCUACCUCCd(CUU) OMeA-OMeTOMeT3′; antisense: 3′OMeG-OMeT-GGUCACCGAUGGAGGGAAd(U)5′) and Hs_TMEM142C_5, #SI04348876 (sense: 5′OMeT-OMeC-CUUAGCCCUUGAAAU)d(ACA) OMeA-OMeT-OMeT3′; antisense: 3′OMeA-OMeG-GAAUCGGGAACUUUAUGUd(U)5′). The STIM1 siRNAs were Hs_STIM1_5, #SI03235442 (sense: 5′OMeU-OMeGAGGUGGAGGUGCAAUd (AUU) dOMeA-dOMeT-dOMeT3′; antisense: 3′OMeA-OMeC-UCCACCUCCACGUUAUAAd (U)5′) and Hs_STIM1_6, #SI04165175 (sense: 5′OMeC-OMeU-GGUGGUGUCUAUCGUd (UAU) OMeU-OMeT-OMeT3′; antisense: 3′OMeG-OMeA-CCACCACAGAUAGCAAUAd (A)5′).

Control cells were transfected with nonsilencing RNA MS_control_mod (sense: 5′OmeA-OMeA-AGGUAGUGUAAUCGCd(CUU) OMeG-OmeT-OMeT3′; antisense: 3′OmeT-OmeT-UCCAUCACAUUAGCGGAAdC 5′).

Quantitative real-time PCR

Quantitative real-time PCR (qRT-PCR) was performed as previously described (34). Total RNA from LNCaP, DU145, PC3, and hPECs was isolated with TRIzol Reagent (Life Technologies). For reverse transcription, 0.8 μg of isolated total RNA was used.

The QuantiTect SYBR Green Kit (Qiagen) was used with 0.5 μL of complementary DNA and 300 nM of primer. The PCR conditions were as follows: 15 min at 95°C; 45 cycles, 30 s at 95°C; 45 s at 58°C; and 30 s at 72°C, and finally a cycle (60 s, 95°C; 30 s 55°C; 30 s 95°C) to determine specificity by a dissociation curve using the MX3000 cycler (Stratagene). Expression of target genes were normalized to the expression of the reference genes RNA polymerase II (RNAPol, NM_000937) and/or TATA box binding protein (TBP, NM_003194). The primer sequences were as follows: Orai1, 5′atgagcctcaacgagcact3′ (forward) and 5′gtgggtagtcgtggtcag3′ (reverse); Orai3, 5′gtaccgggagttcgtgca3′ (forward) and 5′ggtactcgtggtcactct3′ (reverse); STIM1, 5′ cagagtctgcatgaccttca 3′ (forward) and 5′ gcttcctgcttagcaaggtt 3′ (reverse); TBP, 5′ cggagagttctgggattgt 3′ (forward) and 5′ ggttcgtggctctcttatc 3′ (reverse); and RNAPol, 5′ ggagattgagtccaagttca 3′ (forward) and 5′ gcagacacaccagcatagt 3′ (reverse).

Ca²⁺ imaging experiments

Cells were loaded with the ratiometric dye Fura-2AM (hPECs: 1 μM/37°C/20 min; LNCaP and DU145: 2 μM/37°C/15 min; and PC3: 4 μM/room temperature/45 min). Excitation light alternated between 340 nm and 380 nm, and emitted light was detected every 5 s at an emission wavelength of 440 nm. Data were analyzed with TILLVision software (TILL Photonics) and IGOR Pro (WaveMetrics), and intracellular Ca²⁺ concentrations were determined as described previously (45,46).

The bath solution contained (in mM) 155 NaCl, 4.5 KCl, 2 MgCl₂, 10 glucose, and 5 HEPES (pH 7.4 with NaOH). H₂O₂, CaCl₂, and 1 μM thapsigargin (Tg) were added as indicated.

Electrophysiology

Cells (LNCaP and DU145) were patched in a whole-cell configuration as described previously (47,48). The pipette resistance was 2–4 MΩ. Every 2 s, 50 ms spanning ramps from −150 to +100 mV were delivered from a holding potential of 0 mV by a HEKA EPC-10 patch-clamp amplifier and the data were filtered (2.9 Hz), recorded, and analyzed with the use of Patchmaster and Fitmaster software (HEKA). The liquid junction potential was corrected for 10 mV. For analysis, currents were extracted at −80 mV, normalized to the cell capacity, averaged, and plotted versus time. Current was plotted versus ramp voltage (I/V), and current density (CD) was plotted versus H₂O₂ dose and fitted with a Hill function. The pipette solution contained (in mM) 120 Cs-glutamate, 10 BAPTA, 10 HEPES, 3 MgCl₂, and 0.05 IP₃. The bath solution contained (in mM) 95 NaCl, 2.8 KCl, 20 CaCl₂, 2 MgCl₂, 10 HEPES, 10 TEA-Cl, 10 CsCl, and 10 glucose. The pH was adjusted with NaOH to 7.2 and the osmolarity was 300 mosmol/L. H₂O₂ was added as indicated and cells were incubated for 10–30 min before patch-clamp experiments were conducted.

Cell viability

hPEC, LNCaP, DU145, and PC3 cells were seeded to ∼80% density in 96-well cell culture plates (BD) and incubated at 37°C, 5% CO₂, and 95% humidity. Living cells were detected by means of a CellTiter-Blue assay (Promega). The sample size was n = 12 for LNCaP, n = 9 for DU145, n = 3 for PC3, and n = 22 from three donors of hPECs.

Data analysis

Data were analyzed using TILLVision, Fitmaster, Igor Pro, and Microsoft Excel. Data are given as the mean ± SE. (For the data plotted in Fig. 7, Pearson’s coefficient was calculated and is indicated as the R value.)

Figure 7.

Correlation between the Orai3/Orai1 ratio and the H₂O₂-dependent block of SOCE and cell viability. (A) The logarithmic IC₅₀ of the H₂O₂-induced block of SOCE is plotted against the Orai3/Orai1 ratio in different cell types (naive T cells (35), effector T cells (35), LNCaP cells (34), and DU145 cells (34)). Data were fitted with an exponential fit function and Pearson’s coefficient is indicated as the R value. (B) The IC₅₀ of the H₂O₂-induced block of viability is plotted against the Orai3/Orai1 ratio of the same cell types as in (A), as well as in PC3 and hPEC. Data were fitted with a stretched exponential fit function and Pearson’s coefficient is indicated as the R value. (C) The IC₅₀ of the H₂O₂-induced block of viability is plotted versus the IC₅₀ of the H₂O₂-induced block of SOCE. Data were fitted with a Hill function and Pearson’s coefficient is indicated as the R value.

Results

hPECs and prostate cancer cells differ in Ca²⁺ signaling upon incubation with H₂O₂

We first tested the effect of H₂O₂ on Ca²⁺ signaling in a Fura-2-based Ca²⁺ imaging assay in hPECs and the cancer cell lines LNCaP and DU145. For the later analysis of previously published data and data from this study, we exactly followed the procedure published earlier (35). Cells were incubated with different concentrations of H₂O₂, and SOCE was activated with the SERCA inhibitor Tg.

Average Ca²⁺ responses with different H₂O₂ concentrations are shown in Fig. 1, A–C. Incubation of hPECs, LNCaP, and DU145 with H₂O₂ first induced an initial increase of intracellular Ca²⁺. Upon application of Tg, intracellular Ca²⁺ increased (Fig. 1, A–C). To test the contribution of STIM/Orai-mediated signaling, we next analyzed the dependence of the initial Ca²⁺ increase and the Tg-induced intracellular Ca²⁺ increase on the main molecular components of SOCE, STIM1 and Orai1.

Figure 1.

ROS dependence of Ca²⁺ signaling in hPECs, LNCaP, and DU145. (A) Average [Ca²⁺]_i responses (mean ± SE) from a Fura-2-based Ca²⁺ imaging assay when hPECs were incubated with different concentrations of H₂O₂ and 1 μM Tg was added (n = 93 for 0 H₂O₂, n = 146 for 50 μM H₂O₂, n = 67 for 500 μM H₂O₂, and n = 180 for 1 mM H₂O₂). (B) Same as (A) for LNCaP cells (n = 144 for 0 μM H₂O₂, n = 69 for 10 μM H₂O₂, n = 70 for 100 μM H₂O₂, and n = 46 for 1 mM H₂O₂). (C) Same as (A) and (B) for DU145 (n = 172 for 0 μM H₂O₂, n = 43 for 100 μM H₂O₂, n = 57 for 5 mM H₂O₂, and n = 55 for 15 mM H₂O₂).

Dependence of the initial Ca²⁺ increase and the Tg-induced intracellular Ca²⁺ increase on STIM1 and Orai1

To investigate whether the initial Ca²⁺ increase and the Tg-induced intracellular Ca²⁺ increase depend on SOCE, we performed an siRNA-based knockdown of the main molecular components of SOCE, STIM1 and Orai1.

In LNCaP cells, knockdown of STIM1 and Orai1 efficiently reduced the mRNA levels of STIM1 and Orai1 (Fig. 2 A). We then performed the same Fura-2-based imaging experiment shown in Fig. 1 in cells that were transfected with control RNA or siRNA targeting STIM1 and Orai1 (Fig. 2 B), and were either not treated with H₂O₂ or incubated with 10 mM of H₂O₂. We found that 10 mM of H₂O₂ induced an initial Ca²⁺ increase in both control transfected cells and cells transfected with STIM1/Orai1 siRNA (Fig. 2 B). The initial Ca²⁺ increase was analyzed as the average intracellular Ca²⁺ concentration at 1180 s (before application of Tg) and plotted for each condition (Fig. 2 C). Upon incubation with 10 mM of H₂O₂, the initial Ca²⁺ increase remained unchanged after knockdown of STIM1 and Orai1 (Fig. 2 C). Therefore, we conclude that the initial Ca²⁺ increase is independent of STIM1/Orai1-mediated signaling.

Figure 2.

Dependence of the initial Ca²⁺ increase and Tg-induced intracellular Ca²⁺ increase on the STIM1/Orai1 machinery in LNCaP cells. (A) qRT-PCR analysis of Orai1 and STIM1 expression levels in LNCaP cells transfected with control RNA or siRNA targeting STIM1 and Orai1 normalized to TBP. (B) Average [Ca²⁺]_i responses (mean ± SE) from a Fura-2-based Ca²⁺ imaging assay when cells from (A) were not treated with H₂O₂ or were incubated with 10 mM of H₂O₂ and 1 μM of Tg was added (n = 135 for control RNA, 0 H₂O₂; n = 98 for control RNA and 10 mM H₂O₂; n = 74 for siRNA STIM1 and Orai1 and 0 H₂O₂; n = 49 for siRNA STIM1 and Orai1 and 10 mM H₂O₂). (C) Average [Ca²⁺]_i responses (mean ± SE) from cells in (B) at t = 1180 s. (D) For each cell, the Ca²⁺ before application of Tg was subtracted from the maximal Ca²⁺ after application of Tg. The average ΔCa²⁺ values for cells in (B) are plotted.

In addition, this experiment demonstrates that the initial Ca²⁺ increase is not based on Ca²⁺ release from intracellular Ca²⁺ stores. Ca²⁺ release from intracellular Ca²⁺ stores leads to an activation of SOCE. Therefore, STIM1 and Orai1 knockdown would result in a reduction of the initial Ca²⁺ increase, which we did not observe.

To analyze the Tg-induced Ca²⁺ increase for each cell, we subtracted the Ca²⁺ level before application of Tg from the maximal Ca²⁺ level after application of Tg. The average ΔCa²⁺ is plotted for each condition in Fig. 2 D. When STIM1 and Orai1 were knocked down or cells were incubated with 10 mM of H₂O₂, or both, ΔCa²⁺ was significantly reduced to the same level (Fig. 2 D). When cells were incubated with 10 mM of H₂O₂, knockdown of STIM1 and Orai1 did not significantly reduce the remaining Ca²⁺ elevation. Consequently, this remaining Ca²⁺ elevation is independent of the STIM1/Orai1 machinery and for the most part is based on Tg-induced Ca²⁺ release from intracellular Ca²⁺ stores, as shown in a previous study (34). As the remaining Ca²⁺ elevation was the same in all three conditions (when STIM1 and Orai1 were knocked down or cells were incubated with 10 mM of H₂O₂, or both), we conclude that our further analysis of the H₂O₂-induced block of SOCE may include a small offset, but half minimal inhibitory concentrations were not affected. We performed the same set of experiments in DU145 and obtained very similar results (Fig. S1 A in the Supporting Material). Taken together, these results suggest that the initial effect is independent of Ca²⁺ release from intracellular stores and SOCE. Next, to investigate the H₂O₂-induced block of SOCE, we analyzed ΔCa²⁺.

hPECs and prostate cancer cells differ in the initial increase of Ca²⁺ upon incubation with H₂O₂

Upon incubation with H₂O₂, the initial increase of Ca²⁺ varied among the tested cells. Incubation of hPECs with H₂O₂ concentrations of ≥100 μM induced an initial increase of intracellular Ca²⁺ levels, up to ∼150 nM when cells were incubated with 1 mM of H₂O₂ (Fig. 3 A). In LNCaP and DU145, incubation with H₂O₂ induced an initial increase of intracellular Ca²⁺ (Fig. 3 B), and we detected maximal intracellular Ca²⁺ upon incubation with 300 μM and 1 mM of H₂O₂, respectively. Incubation of LNCaP and DU145 with H₂O₂ concentrations exceeding these maxima blocked the initial Ca²⁺ increase.

Figure 3.

Initial H₂O₂-induced Ca²⁺ increase in hPEC and cancer cell lines. (A) Initial increase of intracellular Ca²⁺ when hPECs were incubated for 1000 s with H₂O₂ (before Tg was added). Same cells as in Fig. 1A; n = 71 for 10 nM H₂O₂, n = 46 for 100 nM H₂O₂, n = 133 for 1 μM H₂O₂, n = 158 for 10 μM H₂O₂, n = 160 for 100 μM H₂O₂, and n = 157 for 300 μM H₂O₂. (B) Initial increase of Ca²⁺ when LNCaP or DU145 was incubated for 1000 s with H₂O₂ (before Tg was added). For LNCaP, same cells as in Fig. 1B; n = 67 for 30 μM H₂O₂, and n = 61 for 3 mM H₂O₂. For DU145, same cells as in Fig. 1C; n = 29 for 1 mM H₂O₂, n = 32 for 3 mM H₂O₂, n = 64 for 10 mM H₂O₂, n = 84 for 12.5 mM H₂O₂, and n = 45 for 20 mM H₂O₂.

hPECs and prostate cancer cells differ in ΔCa²⁺ upon incubation with H₂O₂

Addition of the SERCA inhibitor Tg depleted intracellular Ca²⁺ stores and activated SOCE. The H₂O₂ dose dependency of ΔCa²⁺ in hPECs, LNaP, and DU145 is shown in Fig. 4, A and B.

Figure 4.

ΔCa²⁺ in hPEC and cancer cell lines. (A) ΔCa²⁺ after addition of Tg in hPECs upon incubation with different H₂O₂ concentrations, showing the same cells as in Figs. 1A and 3A. The line was drawn to guide the eye. (B) ΔCa²⁺ in LNCaP and DU145 cells after addition of Tg upon incubation with different H₂O₂ concentrations. Average ΔCa²⁺ values (mean ± SE) are plotted versus H₂O₂ concentration; same cells as in Figs. 1, B and C, and 3B; n = 44 for 7.5 mM H₂O₂ for DU145. Data were fitted with a Hill function (please see text and Table 1 for IC₅₀ values).

In hPECs, ΔCa²⁺ was increased by incubation with H₂O₂ up to a concentration of 500 μM. When cells were incubated with 1 mM H₂O₂, the increment was reduced but ΔCa²⁺ was still elevated compared with ΔCa²⁺ at low H₂O₂ concentrations (Fig. 4 A). Upon incubation with H₂O₂ concentrations above 1 mM, hPECs started to detach during the measurements; however, from our data, we conclude that the IC₅₀ of the H₂O₂-induced block of ΔCa²⁺ is above 1 mM.

Upon addition of Tg, the maximal increase in ΔCa²⁺ was detected with 30 μM and 300 μM of H₂O₂ in LNCaP and DU145, respectively. The dose-response curves for the H₂O₂-induced inhibition of ΔCa²⁺ in LNCaP and DU145 are shown in Fig. 4 B. The data were fitted with a Hill equation, and the IC₅₀ values for H₂O₂-induced inhibition of ΔCa²⁺ were 114 μM and 5.1 mM for LNCaP and DU145 cells, respectively.

Upon incubation with H₂O₂, LNCaP and DU145 differ in I_CRAC

As ΔCa²⁺ includes a small offset that is mainly caused by Ca²⁺ release from intracellular stores, we challenged our concept and directly assessed the H₂O₂-induced block of CRAC channels. For this purpose, we incubated LNCaP and DU145 cells with various concentrations of H₂O₂ and performed a whole-cell patch-clamp analysis. Under these conditions, we detected Ca²⁺ currents via I_CRAC channels without any contribution of Ca²⁺ from intracellular stores.

I_CRAC was evoked with 50 μM of IP₃ and 10 mM of BAPTA in the patch pipette. For LNCaP, the CD was plotted versus time (Fig. 5 A; corresponding current-voltage curves are shown in Fig. 5 A, inset).

Figure 5.

Inhibition of I_CRAC by H₂O₂ in cancer cell lines. (A) I_CRAC in LNCaP cells incubated with different concentrations of H₂O₂ (black curve, n = 14, 1 nM H₂O₂; dark gray curve, n = 8, 100 μM H₂O₂; light gray curve, n = 9, 10 mM H₂O₂) and corresponding I/V (inset). (B) Dose responses for H₂O₂-induced block of I_CRAC in LNCaP (same cells as in A; n = 15 for 10 nM H₂O₂, n = 14 for 100 nM H₂O₂, n = 12 for 1 μM H₂O₂, n = 15 for 10 μM H₂O₂, and n = 9 for 1 mM H₂O₂) and DU145 (n = 10 for 1 nM H₂O₂, n = 8 for 100 nM H₂O₂, n = 5 for 1 μM H₂O₂, n = 8 for 10 μM H₂O₂, n = 5 for 1 mM H₂O₂, and n = 4 for 10 mM H₂O₂).

With increasing H₂O₂ concentrations, CD development in LNCaP and DU145 cells was blocked in a dose-dependent manner (Fig. 5 B). For the H₂O₂-induced block of I_CRAC, the dose-response curves exhibit an IC₅₀ of 26.6 μM for LNCaP cells and 2.5 mM for DU145 cells (Fig. 5 B). This analysis shows that a higher ratio of Orai3/Orai1 is accompanied by a higher IC₅₀ for the H₂O₂-induced block of I_CRAC. In our hands, a gigaseal could not be formed with hPECs upon incubation with H₂O₂; therefore, under these conditions, I_CRAC could not be detected via the patch-clamp technique in these cells.

hPECs, LNCaP, DU145, and PC3 differ in cell viability upon incubation with H₂O₂

To compare the viability of hPECs and prostate cancer cell lines upon incubation with H₂O₂, we performed fluorescence-based viability assays. The H₂O₂-induced decrease of cell viability exhibited an IC₅₀ of ∼6 mM in hPECs, ∼2 mM in PC3, 871 μM in DU145, and 422 μM in LNCaP (Fig. 6). These findings clearly support the concept of higher ROS sensitivity in prostate cancer lines than in hPECs.

Figure 6.

Fluorescence-based viability assay of hPECs and cancer cell lines upon incubation with different concentrations of H₂O₂. Fluorescence intensity is plotted versus H₂O₂ concentration for LNCaP (▪), DU145 (●), PC3 (▼), and hPEC (▲). The sample size was n = 12 for LNCaP, n = 9 for DU145, n = 3 for PC3, and n = 22 from three donors for hPEC.

Analysis of Orai3/Orai1 ratios and the H₂O₂-dependent block of SOCE and cell viability

We next combined our data and previous findings (34,35) regarding Orai1/Orai3 mRNA ratios and the dependence of SOCE and cell viability on H₂O₂ in different cell types (summarized in Table 1).

Table 1.

Orai1/Orai3 Ratio, SOCE, and cell viability

Cell Type	Orai1/Orai3	IC50 H2O2-Induced Block of SOCE (μM)	IC50 H2O2-Induced Block of ICRAC (μM)	IC50 H2O2-Dependent Viability (μM)
Naive TH cell	70 (35)	7 (35)	ND	39 (35)
Effector TH cell	25 (35)	51 (35)	ND	199 (35)
LNCaP	26 (34)	114	26	422
DU145	17 (34)	5114	2500	871
PC3	8	ND	ND	2085
Primary prostate epithelial cells	4 (34)	>1000	ND	5947

The table lists the Orai1/Orai3 ratios of the indicated cell types and IC₅₀ values for the H₂O₂-induced block of SOCE, H₂O₂-induced block of I_CRAC, and H₂O₂-dependent cell viability.

We analyzed the correlation between the H₂O₂-dependent block of SOCE (ΔCa²⁺) and dependence of viability on the Orai1/Orai3 ratio. For this purpose, we decided to use the inverse Orai1/Orai3 ratio and instead plot Orai3/Orai1. For prostate-derived cells, a detailed representation of Orai1 and Orai3 mRNA levels and the corresponding Orai3/Orai1 ratios is given in Fig. S2.

In Fig. 7 A, the logarithmic IC₅₀ of the H₂O₂-induced block of SOCE is plotted against different Orai3/Orai1 ratios expressed by several types of cells. When we analyze the correlation between the Orai3/Orai1 ratio and the logarithmic IC₅₀ for the H₂O₂-induced block of SOCE, we find a Pearson’s coefficient of 0.97, reflecting the very strong correlation between the two parameters. Fig. 7 B demonstrates the relationship between Orai3/Orai1 ratios and cell viability. Here, we included data from PC3 cells without analyzing Ca²⁺ signaling in these cells in depth. With our protocol, we cannot determine IC₅₀ for the H₂O₂-induced block of SOCE because in PC3 the initial effect depends on STIM1/Orai1, whereas the Tg-induced Ca²⁺ is nearly independent of STIM1/Orai1 (Fig. S3). The Pearson’s coefficient between the Orai3/Orai1 ratio and cell viability is 0.99, reflecting the very strong correlation between the Orai3/Orai1 ratio and H₂O₂-induced inhibition of cell viability. When the IC₅₀ of the H₂O₂-induced block of cell viability is plotted against the H₂O₂-induced block of SOCE, the dependency can best be described with a Hill function (Fig. 7 C). The Pearson’s coefficient for the logarithmic data is 0.91, reflecting the strong correlation between the H₂O₂-induced block of SOCE and cell viability.

Effect of a siRNA-based knockdown of Orai3 on Ca²⁺ signaling and cell viability

To directly determine the role of Orai3 in the H₂O₂-induced block of SOCE and cell viability, we performed a siRNA-based knockdown of Orai3. Upon knockdown, Orai3 mRNA was reduced (Fig. 8 A). Upon knockdown of Orai3, no specific effect on the IC₅₀ of H₂O₂-induced inhibition of ΔCa²⁺ (Fig. 8 B) and cell viability (Fig. 8 C) could be detected. The cell transfection led to a general shift in the IC₅₀ for the H₂O₂-induced inhibition of ΔCa²⁺ and cell viability.

Figure 8.

Effect of a siRNA-based knockdown of Orai3 on Ca²⁺ signaling and cell viability. (A) qRT-PCR analysis of Orai3 expression levels in LNCaP cells transfected with control RNA or siRNA targeting Orai3 normalized to TBP. (B) Average ΔCa²⁺ from a Fura-2-based Ca²⁺ imaging assay when cells were nontransfected (same data as in Fig. 4B), control transfected (n = 76 for 0.01 mM H₂O₂, n = 144 for 0.1 mM H₂O₂, n = 164 for 0.3 mM H₂O₂, n = 159 for 1 mM H₂O₂, n = 157 for 3 mM H₂O₂ and n = 82 for 10 mM H₂O₂), or transfected with siRNA targeting Orai3 (n = 84 for 0.01 mM H₂O₂, n = 146 for 0.1 mM H₂O₂, n = 165 for 0.3 mM H₂O₂, n = 182 for 1 mM H₂O₂, n = 152 for 3 mM H₂O₂, and n = 94 for 10 mM H₂O₂). (C) Viability assay of cells in (B) (n = 2); for nontransfected cells, the same data as in Fig. 6 were used.

Discussion

Our data on H₂O₂-dependent cell viability support the concept that hPECs are less sensitive to ROS than LNCaP and DU145. This finding is in line with an earlier study that demonstrated that cells derived from prostate cancer tumors are more sensitive to arsenic-trioxide-induced ROS compared with normal prostate epithelium cells (49). It was shown that blocking the ROS-scavenging system of prostate cancer cells shifted the IC₅₀ for arsenic-trioxid-induced cell death in prostate cancer cells to clinical achievable concentrations, with a negligible cytotoxicity for normal cells. It is known that in prostate cancer cells, ROS induce elevations of intracellular Ca²⁺ (12,13). Here, we investigated the H₂O₂-induced initial rise of intracellular Ca²⁺, the H₂O₂-induced amplification of ΔCa²⁺, the block of SOCE, and cell viability upon incubation with ROS in hPECs and prostate cancer cell lines. The ROS-induced initial increase of Ca²⁺ is independent of the STIM1/Orai1 machinery and may be caused by Ca²⁺ channels, e.g., TRP channels that are activated or modulated by ROS, including TRPM2, TRPC5, TRPV1, and TRPA1 (14,16). The initial H₂O₂-induced increase of intracellular Ca²⁺ is maximally activated at lower concentrations of H₂O₂ in LNCaP cells than in hPECs and DU145 (300 μM vs. 1 mM H₂O₂). In prostate cancer cell lines, these values reflect the maxima, and the initial Ca²⁺ increase is blocked upon incubation with higher concentrations of H₂O₂. In hPECs, 1 mM of H₂O₂ induces the maximal initial increase of intracellular Ca²⁺. The detection of Ca²⁺ signals from cells incubated with higher concentrations of H₂O₂ was technically not feasible. It is known that ROS-induced transcription factors that activate ROS-scavenging systems (e.g., NF-E2-related factor 2) (50) depend on elevation of intracellular Ca²⁺ (51). Hence, the block of initial ROS-induced Ca²⁺ signaling in cancer cell lines may contribute to their higher sensitivity to ROS.

SOCE is activated during proliferation; however, elevated SOCE signals can drive cells into apoptosis (52). When cells were preincubated with different H₂O₂ concentrations, the maximal amplification of ΔCa²⁺ occurred with 500 μM of H₂O₂ in hPECs, 30 μM of H₂O₂ in LNCaP, and 300 μM of H₂O₂ in DU145. Thus, in cancer cell lines, these maximally amplified ΔCa²⁺ signals at lower H₂O₂ concentrations may contribute to the overall higher sensitivity of cell viability to ROS.

Within the last few years, several reports have demonstrated a role for Orai3 in breast cancer (53–56). Two very recent studies reported controversial results regarding the role of Orai3 in prostate cancer. Dubois et al. (57) found elevated Orai3 expression levels in prostate cancer tissue samples. In their study, elevated levels of Orai3 led to increased formation of arachidonic-acid-induced Ca²⁺ channels by Orai1/Orai3 heteromers and a lower number of homomeric Orai1 SOCE channels. Thus, the elevated Orai3 levels may act as a switch and lead to increased arachidonic-acid-induced proliferation and decreased Orai1-dependent apoptosis (57,58). In contrast to Dubois et al. (57), we found a downregulation of Orai3 in prostate cancer tissue samples, an outstandingly low Orai1/Orai3 ratio of ∼4 in hPECs, and elevated Orai1/Orai3 ratios in prostate cancer cell lines, with consequences for SOCE signaling in the membrane androgen receptor pathway and the pharmacological profile of I_CRAC (34). The H₂O₂-dependent block of SOCE (ΔCa²⁺) differs among hPECs (IC₅₀ > 1 mM), DU145 (IC₅₀ ∼5 mM), and LNCaP (IC₅₀ ∼114 μM). In combination with previous findings (34,35), our results demonstrate a strong correlation between the Orai3/Orai1 ratio and the ROS sensitivity of SOCE and cell viability. Taken together, these findings support the concept of heteromeric store-operated Orai1/Orai3 channels. However, upon cell transfection, the IC₅₀ values of the H₂O₂-induced block of SOCE and cell viability exhibited unspecific shifts, as described previously (35). This unspecific shift may cover specific effects of Orai3 knockdown and thus prevent the acquisition of direct evidence. The correlation between the ROS-dependent block of SOCE and cell viability demonstrates that cells need functional SOCE Ca²⁺ signaling for survival and that the ROS-induced block of SOCE contributes to decreased cell viability when ROS are increased. The apparent IC₅₀ of the H₂O₂-induced block of SOCE in DU145 (∼5 mM) and hPECs (>1 mM) points to a blocking mechanism that could be independent of Orai3/Orai1 ratios. In these cells, all I_CRAC channels may be Orai3/Orai1 heteromeric channels, and one Orai3 subunit is sufficient to abolish the ROS sensitivity of SOCE (37). An alternative explanation is that increasing H₂O₂ levels lead to intracellular acidification (59) and CRAC channels are inhibited by intracellular acidification (60). It has been suggested that STIM1/Orai1 uncouple at low intracellular pH and Ca²⁺ influx via Orai channels is abolished (61). To test this hypothesis, we performed patch-clamp experiments to determine whether the H₂O₂-induced block was still apparent when we used pH 8 in the patch pipette (Fig. S4). Indeed, the block was not abolished, pointing to a channel-specific mechanism rather than an unspecific block. On the other hand, it was previously demonstrated in snail neurons that even under buffering conditions, pH microdomains below the plasma membrane could be formed (62). Thus, we cannot exclude the possibility of a channel-unspecific mechanism such as acidic pH, decoupling of STIM/Orai complexes, induction of high Ca²⁺ levels, and/or membrane depolarization.

In the future, therapeutic strategies based on ROS induction may include the appropriate concentrations of drugs targeting SOCE channels to reduce the viability of prostate cancer cells without affecting nontransformed cells, as there is a clear role for Ca²⁺ in ROS-mediated signaling in prostate cancer. Finally, the overall high Orai3/Orai1 ratios in hPEC and androgen-insensitive cancer cells contribute to their ROS resistance and thereby may have a share in making the prostate one of the most prominent cancer susceptible organs.

Conclusions

In this study, we investigated H₂O₂-dependent Ca²⁺ signaling in hPECs from healthy tissue and prostate cancer cell lines (LNCaP, DU145, and PC3). ROS-induced changes in Ca²⁺ signaling reflect the contributions of very different enzymes, including Ca²⁺ transporters and ROS-producing and -scavenging enzymes. Our findings suggest that the block of ROS-induced initial Ca²⁺ elevations in prostate cancer cells, as well as the amplification of ΔCa²⁺ and the H₂O₂-dependent block of SOCE at lower concentrations of H₂O₂, could contribute to the higher sensitivity of prostate cancer cells to ROS-induced cell death. In addition, our findings regarding the H₂O₂-dependent block of SOCE in hPECs and cancer cell lines support our concept of heteromeric store-operated Orai1/Orai3 channels in hPECs and store-operated Orai channels characterized by elevated Orai1/Orai3 ratios in prostate cancer cells.

Author Contributions

C.P. designed the study, analyzed data, and wrote the manuscript. C.H., T.K., S.K., and K.D. performed experiments, analyzed data, helped design the study, and helped write the manuscript. I.B. helped design the study and develop the manuscript. V.J. and M.S. helped design the study.

Editor: Michael Pusch.