Functional consequences of enhanced expression of STIM1 and Orai1 in Huh-7 hepatocellular carcinoma tumor-initiating cells

Background The endoplasmic reticulum (ER) Ca2+ sensor, stromal interaction molecule1 (STIM1) activates the plasma membrane (PM) channel Orai1 in order to mediate store-operated Ca2+ entry (SOCE) in response to ER store depletion. Enhanced expression of STIM1 in cancer tissue has been associated with poor patient prognosis. Therefore, this study investigated the functional consequences of enhanced expression of STIM1 and Orai1 in a tumor-initiating subpopulation of Huh-7 hepatocellular carcinoma (HCC) cells that express epithelial cell adhesion molecule (EpCAM) and Prominin 1 (CD133). Methods We performed qRT-PCR, intracellular Ca2+ monitoring, protein analyses, and real-time cell proliferation assays on EpCAM(+)CD133(+) subpopulation of tumor-initiating Huh-7 HCC cells expressing high levels of STIM1 and/or Orai1. Statistical significance between the means of two groups was evaluated using unpaired Student’s t-test. Results Enhanced STIM1 expression significantly increased ER Ca2+ release and proliferation rate of EpCAM(+)CD133(+) cells. Conclusion STIM1 overexpression may facilitate cancer cell survival by increasing ER Ca2+-buffering capacity, which makes more Ca2+ available for the cytosolic events, on the other hand, possibly preventing Ca2+-dependent enzymatic activity in mitochondria whose Ca2+ uniporter requires much higher cytosolic Ca2+ levels.

TICs appear to be responsible for high recurrence rates as well as for chemoresistance [36]. HCC cells are a non-excitable cell type, where SOCE plays a crucial role in Ca 2+ homeostasis and signaling [37]. In many cancer types including HCC, enhanced expression of STIM1 and Orai1 have been shown to enhance carcinogenesis including proliferation, migration and invasion processes [26,38,39]. Previous studies have reported that STIM1 and Orai1 molecules mix at a specific ratio to encode functional CRAC channel assembly [40,41]. Based on crystallographic and electrophysiological studies, STIM1 exists as a dimer under resting conditions, and binds to Orai1 in a nonlinear fashion such that all six Orai1 binding sites must be occupied for the activation of SOCE [42]. However, the structural basis of STIM1 interaction with Orai1 within the channel assembly is not known. Therefore, the purpose of this study is to investigate the functional impact of altered stoichiometry of STIM1 and/or Orai1 by employing overexpression plasmid vectors on intracellular Ca 2+ dynamics as well as carcinogenic properties of Huh-7 EpCAM(+)CD133(+) cells.

Cell culture
Human HCC cell lines (Huh-7) were provided by Dr. Ozturk (IBG İzmir), originally from Dr. Jack Wands Laboratory (Massachusetts General Hospital, Boston,

Transfection of EpCAM(+)CD133(+) Huh-7 cells with STIM1 and Orai1 overexpression plasmids
Cells were seeded on 6 well-plate (10 5 cells/well) and transfection was performed after 24 h with X-tremeGENE HP DNA Transfection Reagent (Roche). Following removal of the cell media, serum-reduced media (Opti-MEM) were added and incubated for additional 1 h. 100 μl Opti-MEM, 1.5 μg plasmid DNA (MO70-STIM1-eYFP, pDEST501-Orai1-CFP and pCMV6 empty vector as a control) and 1 μl X-tremeGENE HP DNA Reagent-containing transfection mix was added to each well and incubated for 30 min at room temperature. Transfection mix was added on the cells dropwise and shaked gently. Plasmids were gently provided by Dr. M Trebak (Penn State University).

RNA isolation and cDNA synthesis
Cells were seeded on 6-well plate (15 × 10 4 /well). Total RNA was isolated by using High Pure RNA Isolation (Roche) according to the manufacturer's instructions. cDNA synthesis from the total RNA samples were performed by using Transcriptor First Strand cDNA Synthesis Kit (Roche) according to the manufacturer's instructions.

Real-time quantitative RT-PCR (qRT-PCR)
FastStart DNA Master SYBR Green I kit was used in real-time qRT-PCR experiments performed (LightCycler 1.5, Roche Applied Science). Primer sequences are shown in Table 1

Real-time monitoring of proliferation by real-time cell analyzer (RTCA)
Real-time label-free impedance-based monitoring of cellular proliferation assay was performed by using xCELLigence MP (Roche Applied Science). Transfected cells were incubated in 6-well plates for 48 h. After the incubation period, 5000 cells/well were seeded in E-plate 96. Cell proliferation was monitored at every 15 min for 72 h. Changes in proliferation rate were expressed as "cell index" (RTCA software 1.2.1, Roche Applied Science).

Data analysis
Data expressed as mean ± standard error of the mean (S.E.M.). "n" denotes the number of samples. Statistical significance between the means of two groups was evaluated using Student's t-test (unpaired data). Significance was accepted at 0.05 level of probability. Orai1 mRNA level increased in Orai1-OE (**p < 0.01, Student t-test, unpaired data n = 4, Fig. 3a) and STIM1 + Orai1-OE (**p < 0.01, Student t-test, unpaired data, n = 4, Fig. 3b) EpCAM(+)CD133(+) cells (Fig. 3) comparable to the control, which is similar to that of STIM1 mRNA expression levels revealed in previous data.
Although not statistically significant, the Orai1 protein level was lower in STIM1-OE samples comparable to that of the control (Fig. 5).

Discussion
In addition to being involved in intracellular Ca 2+ homeostasis mechanism of non-excitable cells, SOCE appears to be operational in hepatocellular carcinogenesis [26]. In this study, the role of SOCE components, STIM1 and Orai1, reportedly involved in intracellular Ca 2+ regulation was investigated on Huh-7 TICs expressing cell surface antigens EpCAM and CD133 through monitoring intracellular Ca 2+ dynamics (ER Ca 2+ release and SOCE), proliferation and MDR1 expression responsible partly for drug resistance. High intracellular Ca 2+ concentration comprises toxic and proapoptotic conditions for cells. Excessive Ca 2+ is buffered by certain proteins (e.g., calsequestrin and calreticulin) inside ER and by mitochondria. ER Ca 2+ release and SOCE are significantly higher in EpCAM(+)CD133(+) cells comparable to that of EpCAM(−) CD133(−).
Overexpression of STIM1 and Orai1 is shown in many cancer types like prostate cancer, breast cancer, glioblastoma and hepatocellular carcinoma [33]. More specifically, STIM1 overexpression is commonly seen in HCC [26,39]. Among the three overexpression groups of EpCAM(+)CD133(+) Huh-7 cell subpopulation in our study, STIM1-OE showed the highest ER Ca 2+ release. As STIM1 has Ca 2+ binding EF hand domains located on the intracellular part of ER [45], its overexpression may buffer more Ca 2+ , leading to more Ca 2+ available to be released from ER following SERCA blockade by CPA. STIM1 is the key initiating molecule in SOCE. After ER depletion, as a sensor of ER Ca 2+ content, STIM1 accumulated in ER membrane closely located to PM with Orai1. At this ER and PM junctions, STIM1 interacts with Orai1 as a result SOCE is activated [46]. Lower ER release and SOCE in Orai1 OE EpCAM(+)CD133(+) Huh-7 cells comparable to the control cells could be due to changes in coupling stoichiometry between STIM1-Orai1 for SOCE [47]. Higher levels of the PM channel subunit (Orai1) might decrease effective coupling of two molecules (STIM1 and Orai1) yielding SOCE inhibition. Increases of ER release and SOCE in STIM1 + Orai1-OE EpCAM(+)CD133(+) cells, show presence of appropriate coupling stoichiometry between STIM1 and Orai1 for SOCE as both molecules are freely available for random interaction [32]. Similar SOCE elevations were also seen  in STIM1 + Orai1-OE and only STIM1-OE DU145 (prostate cancer cell line) and HEK (human embryonic kidney) cells, respectively [40,48,49]. Overexpression of Orai1 in DU145 and HEK cells also inhibited SOCE, as observed in EpCAM(+)CD133(+) cells in our study [40,47,48].
TICs tended to remain in a quiescence state [50]. These EpCAM(+)CD133(+) cells have slow proliferation rates comparable to that of EpCAM(−)CD133(−) [49,51] as was also observed in the present study. This may support their survival strategy in a cytotoxic environment [34,52,53]. The higher proliferation rate of STIM1-OE cells, comparable to that of STIM1 + Orai1-OE cells, showed that upregulation of these two genes (STIM1 and Orai1) suppresses the cell division/proliferation possibly through attenuated Ca 2+ buffer capacity of ER. Again, the significantly higher proliferation rate observed with STIM1-OE cells over that of EpCAM(+)CD133(+) cells overexpressing both STIM1 and Orai1 (present data) confirms the poor prognosis of several cancer types with overexpressed STIM1 [54][55][56].
Cancer cells show resistance to chemotherapeutic treatments. This may result from drug inactivation, changing drug targets, DNA damage repair, and efflux of drug from cells by ABC transporters [57]. Because of the upregulated ABC transporters, cancer cells can pump chemotherapeutics out of the cell [58]. The "slow and steady" feature might also be maintained by higher MDR1 (an ABC transporter family member) expression. Upregulated MDR1 in EpCAM(+)CD133(+) Huh-7 cells in the present study is also in accordance with the increased MDR1 gene expression in lung cancer [59], ovary cancer [60], osteosarcoma [61] and glioblastoma's [62] cancer stem cells. The signaling pathways (JAK/ STAT, PI3K/AKT, MAPK/ERK), which take place in drug resistance, are regulated by Ca 2+ /calmodulin dependent protein kinase II (CaMKII), suggesting an interaction between Ca 2+ and MDR mechanisms in liver cancer [38].

Conclusions
Based on the higher proliferation rates observed in STIM1-overexpressing EpCAM(+)CD133(+) Huh7 cells compared to that of STIM1 + Orai1-OE constructs, one may conclude that HCC stem cells might undergo a phenotypical switch process from a quiescent to proliferative stage by increasing ER Ca 2+ buffering capacity due to higher levels of Ca 2+ -binding protein, STIM1. Furthermore, one may also speculate that increased ER Ca 2+ buffering prevents Ca 2+ -dependent processes in mitochondria localized within the ER microenvironment