Store-Operated Ca2+ Entry in Proliferating and Retinoic Acid-Differentiated N- and S-Type Neuroblastoma Cells

Biochimica et Biophysica Acta 1833 (2013) 643–651

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbamcr

Store-operated Ca2+ entry in proliferating and retinoic acid-differentiated N- and S-type neuroblastoma cells

Natalie Bell a, Victoria Hann a, Christopher P.F. Redfern b, Timothy R. Cheek a,⁎ a Institute for Cell and Molecular Biosciences, The Medical School, Newcastle University, Framlington Place, Newcastle upon Tyne, NE2 4HH, UK b Northern Institute for Cancer Research, The Medical School, Newcastle University, Framlington Place, Newcastle upon Tyne, NE2 4HH, UK article info abstract

Article history: Neuroblastoma cell lines are heterogeneous, comprised of at least three distinct cell phenotypes; neuroblastic Received 26 October 2012 N-type cells, non-neuronal substrate-adherent S-type cells and intermediate I-type cells. N- and S-type cell Received in revised form 23 November 2012 populations were enriched from the parental SH-SY5Y neuroblastoma cell line and induced to differentiate Accepted 28 November 2012 by the addition of retinoic acid (RA), a drug used in the treatment of neuroblastoma. N- and S-type cells Available online 6 December 2012 were identiﬁed based on their differential expression of β-tubulin III, vimentin and Bcl-2. Store-operated Ca2+ entry (SOCE) was then measured in proliferating and differentiated N- and S-type cell populations Keywords: Store-operated Ca2+ entry and the expression of STIM1, Orai1 and TRPC1, three proteins reported to play a key role in SOCE, was determined. Differentiation In N-type cells the RA-induced switch from proliferation to differentiation was accompanied by a down-regulation STIM1 in SOCE. STIM1 and Orai1 expression became down-regulated in differentiated cells, consistent with their respec- Orai1 tive roles as ER Ca2+ sensor and store-operated Ca2+ channel (SOC). TRPC1 became up-regulated suggesting that TRPC1 TRPC1 is not involved in SOCE, at least in differentiated N-type cells. In S-type cells SOCE remained active following Neuroblastoma the RA-induced switch from proliferation to differentiation and the expression of STIM1 and Orai1 remained unchanged. TRPC1 was not expressed in S-type cells. Our results indicate that differentiation of neuronal cells is associated with a remodelling of SOCE. Therapeutic targeting of SOCE proteins could potentially be a means of promoting neuronal differentiation in the treatment of neuroblastoma. © 2012 Elsevier B.V. All rights reserved.

1. Introduction The second messenger Ca2+ plays an essential role in the regulation of many cellular processes [13,14], including differentiation of neuronal Neuroblastoma is a childhood cancer of the sympathetic nervous cells [15,16]. Store-operated Ca2+ entry (SOCE) is a ubiquitous Ca2+ in- system that originates from neural crest cells [1]. Neuroblastoma tu- ﬂux pathway through which Ca2+ enters cells via store-operated Ca2+ mours and their derived cell lines are heterogeneous, composed of channels (SOCs) located in the plasma membrane (PM) in response to multipotent precursor cells that give rise to distinct neural crest cell lin- depletion of endoplasmic reticulum (ER) Ca2+ stores [17,18].Previous eages [2]. At least three cellular phenotypes have been identiﬁed in neu- work from this laboratory has shown that SOCE becomes down- roblastoma cell lines; neuroblastic N-type cells, substrate-adherent regulated in SH-SY5Y neuroblastoma cells following 9-cis retinoic acid S-type cells and intermediate I-type cells [2–5]. N-type cells are imma- (9cRA)-induced differentiation [19]. ture nerve cells, precursors to the sympathoadrenal cell lineage of the The proteins STIM1, Orai1 and TRPC1 have been reported to play a neural crest [4–6]. S-type cells are multipotent precursors to Schwann key role in SOCE [20–23]. STIM1 senses the level of Ca2+ within the cells, melanocytes and glial cells and form the non-neuronal lineage of ER and re-locates to ER-PM junctions to signal store depletion and in- the neural crest [3,4]. I-type cells are intermediate with respect to N- duce opening of SOCs [24,25]. Orai1 forms a SOC in many cell types and S-type cells in terms of morphology and biochemical markers and is required to reconstitute the Ca2+ release-activated Ca2+ cur-

[3,4]. I-type cells may represent either a stem cell or an intermediate rent (ICRAC) [21,26], the most well-deﬁned SOCE pathway. TRPC1 is stage in the transdifferentiation between N- and S-type cells [3,7]. a controversial SOC candidate as literature both supports and opposes N-type cells are more malignant than S-type cells, which appear to be the involvement of TRPC1 in SOCE [18,27]. TRPC1 may only function non-malignant [5,8,9]. Retinoic acid (RA) is used in the treatment pro- as a SOC under certain conditions as studies have shown that TRPC1 tocol for high-risk neuroblastoma patients as it inhibits proliferation can function as either a SOC or a receptor-operated channel (ROC) and induces differentiation of cells [10–12]. depending on its interaction with STIM1 [28–30]. The interaction between STIM1 and TRPC1 can also require Orai1 [29,31–34]. Accumu- ⁎ Corresponding author. Tel.: +44 191 222 7010; fax: +44 191 222 7424. lating evidence suggests that SOCs are heteromeric complexes that E-mail address: [email protected] (T.R. Cheek). can include both Orai1 and TRPC1 [29,31,33,34].

In the present study, N- and S-type cells were enriched from the 2.5. Western blotting parental SH-SY5Y neuroblastoma cell line which, although mainly composed of N-type cells, S-type cells remain present due to the abil- Cells were lysed with 4 °C buffer containing 1 mM EDTA (pH 8), ity of cells to transdifferentiate between cell phenotypes [7,35]. Cell 1 mM EGTA (pH 8), 1.28 mM sucrose, 2 mM Tris (pH 7.6), 10% Triton populations were induced to differentiate by the addition of 9cRA X-100 and protease inhibitor cocktail (Roche, Hertfordshire, UK). and characterised morphologically and biochemically using the neu- Cells were scraped off dishes, passed through a 20-gauge needle sev- ronal marker proteins β-tubulin III and Bcl-2 [36–39] and the non- eral times and centrifuged at 10,000 rpm for 10 min at 4 °C. Protein neuronal marker protein vimentin [3]. The remodelling of SOCE concentration was measured at 492 nm using the Coomassie Blue Re- observed following 9cRA-induced differentiation [19] was further agent (BioRad, Hertfordshire, UK). BSA was used as a protein standard characterised in this study by determining the extent that N- and to determine the protein concentration of samples. S-type cells contribute to the down-regulation. The pattern of expres- SDS-PAGE was performed using the NuPAGE system (Invitrogen, sion of STIM1, Orai1 and TRPC1 was also determined in proliferating Paisley, UK). Total protein (20–40 μg) was separated by electrophore- and differentiated N- and S-type cells to investigate the involvement sis through 10% Bis-Tris gels (Invitrogen) and transferred onto nitro- of these Ca2+ signalling proteins in the remodelling of SOCE. cellulose membranes (BioRad). Blots were blocked for 1 h (5% milk, 0.02% Triton X-100) and were incubated with primary antibody in in- 2. Materials and methods cubation buffer (2.5% milk) overnight at 4 °C. Primary antibodies; anti-β-actin 1:10,000 (abcam, Cambridge, UK), anti-β-Tubulin III, 2.1. Materials 1:20,000 (Covance), anti-vimentin, 1:200, (Santa Cruz), anti-Bcl-2, 1:200 (Santa Cruz), anti-STIM1, 1:200 (BD Biosciences, NJ, USA), SH-SY5Y cells were supplied by R. Ross (Fordham University, NY, anti-Orai1, 1:100 and anti-TRPC1, 1:200 (Alamone Labs, Jerusalem, Is- USA).FluorSave,fura-2/AM,ionomycinandthapsigargin(TG)were rael). Blots were then incubated with either mouse or rabbit second- obtained from Calbiochem (Darmstadt, Germany). All other chemicals ary antibody with horseradish peroxidase-conjugate (Dako, Glostrup, were obtained from Sigma-Aldrich (Dorset, UK) unless otherwise stated. Denmark) for 1 h at RT. The enhanced chemiluminescence system was used to detect immunoreactive bands on Hyperﬁlm™ (GE 2.2. Cell culture and differentiation Healthcare, Buckinghamshire, UK). Band intensities were measured using ImageJ® software. Protein levels were normalised to the levels SH-SY5Y, N- and S-type neuroblastoma cells were cultured in of β-actin. Dulbecco's modiﬁed Eagle's medium (DMEM)/F12:1 with GlutaMAX™ 2+ (Gibco, Paisley, UK) supplemented with foetal calf serum (10%), penicil- 2.6. Determination of [Ca ]i in cell populations lin (100 IU. ml−1) and streptomycin (100 IU.ml−1). Cells were kept at

37 °C in a humidified atmosphere of 5% CO2. SH-SY5Y cells were pas- SH-SY5Y, N- and S-type cells were washed in Krebs buffer (10 mM saged once a week using 0.02% EDTA and were not used beyond passage glucose, 118 mM NaCl, 4.7 mM KCl, 1.2 mM KH2PO4, 1.2 mM MgSO4, 28. Cells were seeded onto coverslips/dishes at least 24 h prior to the 4.2 mM NaHCO3, 2 mM CaCl2 and 200 μM sulfinpyrazone, 10 mM start of treatment. For differentiation, cells were treated for 7 days HEPES, pH 7.4) and loaded with the Ca2+-sensitive fluorescent indi- with 1 μM9cRA. Differentiation medium was replaced every 2 days. cator dye fura-2/AM (3 μM) for 45 min. Following loading cells Proliferating (control) cells were treated identically but with an equal were washed in Krebs buffer and incubated for a further 20 min to volume of vehicle ethanol (0.01%) in place of 9cRA. allow de-esterification of the loaded dye. Cells were washed in Ca2+-free Krebs buffer before being mounted into a coverslip holder 2.3. Enrichment for N- and S-type cells (PerkinElmer, Beaconsfield, UK) and inserted into a stirred cuvette containing Ca2+-free Krebs buffer. Before beginning each experi- N- and S-type cells were enriched from the parental SH-SY5Y neu- ment, an excitation wavelength scan (250–450 nm) was performed roblastoma cell line on the basis of their differential substrate adher- to reveal any large inconsistency in the degree of confluency and/or ence [8]. N-type cell populations were obtained by knocking off the dye loading. Less than 5% of coverslips were rejected on this basis. more weakly adherent cells into PBS by gentle agitation and transfer- Fura-2 fluorescence was continuously monitored using a PerkinElmer ring the cell suspension to a new flask; S-type cell populations were LS-50B fluorimeter, with excitation and emission wavelengths of 340 obtained by maintaining those still adhered to the flask. N- and and 510 nm respectively. Following the establishment of a steady 2+ S-type cell populations were sub-cultured in this way 8 times and baseline TG (200 nM) was added to deplete ER Ca stores. CaCl2 are referred to in the text as N- and S-type cells. (2 mM) was then added back to the Ca2+-free Krebs buffer to reveal 2+ subsequent Ca entry. Ionomycin (50 μM) was added to obtain Fmax 2.4. Immunofluorescence (maximum fura-2 fluorescence) and MnCl2 (1 mM) was added to obtain FMn (fluorescence after quenching of fura-2), an indirect mea- 2+ SH-SY5Y, N- and S-type neuroblastoma cells were fixed with 4% surement of Fmin (minimum fura-2 fluorescence). [Ca ]i was calcu- paraformaldehyde and permeabilised with 0.1% Triton X-100. Cells lated using PerkinElmerWinLab® software which uses the formula were blocked with 5% bovine serum albumin (BSA) prior to incuba- of Grynkiewicz [40] which assumes a dissociation constant (Kd)of tion for 2 h at 4 °C with anti-β-tubulin III with Alexa Fluor 488 conju- 224 nm and an instrument constant (IC) of 3.The following equations 2+ gate, 1:50 (Covance, NJ, USA ) and anti-vimentin with Alexa Fluor 674 were used; Fmin =1/IC(Fmax −FMn)+FMn and [Ca ]i =Kd(F−Fmin)/ conjugate, 1:50 (Santa Cruz, CA, USA). Cells were incubated with (Fmax −F) [40]. Each trace was calibrated individually to account for ethidium homodimer-1 (EthD-1) at a 1:500 dilution for 10 min at variation in confluency and/or fura-2 loading. To quantify change in 2+ room temperature (RT). Coverslips were mounted onto glass slides [Ca ]i the area from each response was determined in calibrated with FluorSave. Images were collected on a Zeiss LSM 510 confocal traces using PerkinElmerWinLab® software. The area from calibrated microscope using a 40× oil immersion objective. A multi-track con- DMSO control traces was subtracted from experimental (TG) traces. figuration was used with 488 nm and 633 nm excitation light from an argon laser and a helium-neon laser respectively. Confocal images 2.7. Statistical analysis were collected with the pinhole set at 1 Airy unit for Alexa Fluor 647 emission, the same optical slice (1.1 μm) was then set for the collec- Data are presented as means±SEM of at least three independent tion of Alexa Fluor 488 and EthD-1 emission. experiments, as stated in the Results section. Comparisons between N. Bell et al. / Biochimica et Biophysica Acta 1833 (2013) 643–651 645 unpaired groups were carried out using two-tailed Student's t-tests in population. For our studies N- and S-type cells were enriched from GraphPad Prism® software. Data presented in Table 1 were analysed the parental SH-SY5Y cell line (Fig. 1). In N-type cell populations using 2-way ANOVA (after log-transformation) and general linear models S-type cells comprised less than 3% of the total cell population, simi- in R [41]. For all tests, the criterion for statistical significance was Pb0.05. larly in S-type cells populations N-type cells comprised less than 3% of the total cell population. Proliferating N-type cells were small with 3. Results rounded or slightly elongated cell bodies with several short, branched neurite-like processes (Fig. 1B). N-type cells grew rapidly with a dou- 3.1. Enrichment for N- and S-type cells bling time of ~24 h and formed weakly adherent cellular aggregates. Proliferating S-type cells were larger and flatter than N-type cells and The SH-SY5Y cell line was predominantly composed of N-type exhibited strong substrate adherence (Fig. 1C). S-type cells grew cells (Fig. 1A). However, S-type cells were present (Fig. 1A), and in more slowly compared to N-type cells with a doubling time of our laboratory comprised approximately 20% of the total cell ~48 h and showed contact inhibition of growth.

Fig. 1. N- and S-type cell populations were enriched from the parental SH-SY5Y neuroblastoma cell line and induced to differentiate by the addition of 1 μM9cRA for 7 days. (A) The SH-SY5Y cell line was predominantly composed of N-type cells though S-type cells were present. Proliferating (B) N-type and (C) S-type cells. (D) Differentiated N-type cells. Cells exhibited neurite extensions of ≥50 μm in length; arrow. (E) Differentiated S-type cells. Scale bars represent 20 μm. 646 N. Bell et al. / Biochimica et Biophysica Acta 1833 (2013) 643–651

Fig. 2. N-type cells stained positive for a neuronal lineage marker whereas S-type cells stained positive for a non-neuronal lineage marker. Confocal micrographs showing β-tubulin III (green), vimentin (red) and EthD-1 (blue). Cells were induced to differentiate by the addition of 1 μM9cRA for 7 days. Proliferating (A) SH-SY5Y, (B) N-type and (C) S-type cells. Differentiated (D) SH-SY5Y and (E) N-type cells. Vimentin in neurite extension; arrow. (F) Differentiated S-type cells. Scale bars represent 50 μm. The parental SH-SY5Y cell line was comparable to N-type cells.

3.2. Differentiation of N- and S-type cells 3.4. β-tubulin III, vimentin and Bcl-2 in N- and S-type cells

After treatment with 9cRA, N-type cells differentiated into a The expression of β-tubulin III and vimentin in N- and S-type cells neuronal-like phenotype, exhibiting long neurite extensions (Fig. 1D, was quantified by western blot analysis. In N-type cells β-tubulin III arrow). Cells were classed as differentiated if neurite extensions were levels remained unchanged after 9cRA-induced differentiation, P= ≥50 μm in length [19,42,43]. S-type cells differentiated into a more 0.118, n=9 (Fig. 3A,C). In S-type cells β-tubulin III was present, epithelial-like phenotype becoming larger and flatter with increased though at a lower level compared to the N-type cells, and after 9cRA substrate adherence (Fig. 1E). RA treatment inhibits proliferation of treatment became down-regulated compared to proliferating S-type cells [10] and a reduction in the rate of proliferation by ~75% in cells by ~44%, Pb0.001, n=9 (Fig. 3A, C). The expression of vimentin N-type cells and by ~50% in S-type cells was observed. in S-type cells remained unchanged following 9cRA-induced differentiation, P=0.879, n=8 (Fig. 3B, C). Although vimentin was identified in some neurite extensions of N-type cells (Fig. 2E), it was not 3.3. Immunofluorescent profile of N- and S-type cells detected by western blot (Fig. 3B). The parental SH-SY5Y cell line was comparable to N-type cells (Fig. 3). In order to determine cell lineage, fixed cells were stained with The expression of Bcl-2 was also determined as Bcl-2 has been antibodies directed against β-tubulin III, a component of microtu- shown to increase in differentiated neuroblastoma cells [44–46]. bules in neuronal cells [38,39] and vimentin, an intermediate filament Bcl-2 was expressed in proliferating SH-SY5Y and N-type cells and protein in non-neuronal cells [3]. β-tubulin III was present through- after 9cRA-induced differentiation became up-regulated by ~50% out the cytoplasm and along neurite branches/extensions of both pro- (P=0.020, n=6) and ~35% (P=0.027, n=7) respectively (Fig. 3B, liferating and differentiated N-type cells; expression did not appear to C). Bcl-2 expression was barely detectable in S-type cells (Fig. 3B, C). change after 9cRA-induced differentiation (Fig. 2B, E). β-tubulin III was also present throughout the cytoplasm of S-type cells, albeit at 3.5. SOCE in N- and S-type cells a weaker level compared to N-type cells (Fig. 2C, F). Vimentin was not generally present in the cell cytoplasm of N-type cells, though it Ca2+ ‘add-back’ experiments were performed on N- and S-type was identified in some neurite extensions (Fig. 2E, arrow). Vimentin cell populations to determine any changes in SOCE that may have oc- was present in the cytoplasm of both proliferating and differentiated curred in response to 9cRA-induced differentiation. Fura-2 loaded S-type cells (Fig. 2C, F). In the parental SH-SY5Y cell line, N-type cells cells in Ca2+-free buffer were stimulated with TG (200 nM) to de- 2+ formed the majority of the population but S-type cells remained pres- plete ER Ca stores, and then CaCl2 (2 mM) was added to determine ent (Fig. 2A, D). subsequent Ca2+ entry (i.e. SOCE). N. Bell et al. / Biochimica et Biophysica Acta 1833 (2013) 643–651 647

Table 1 P=0.925 (Fig. 4C). Subsequent Ca2+ entry following add-back of Ca2+ 2+ Rates of rise of TG and Ca responses in N- and S-type cells. Rates of rise of TG was also not significantly different between proliferating and differentiat- (200 nM) and Ca2+ (CaCl , 2 mM) were calculated from calibrated traces such as 2 ed cells; 10.15±1.48 μM s vs. 11.74±3.14 μMs,P=0.657 (Fig. 4C). The those shown in Fig. 4. Data were log-transformed and 2-way ANOVA was performed to determine the effect of differentiation and the difference between cell types. For SOCE pathway therefore remained active in differentiated S-type cells. N-type cells; proliferating n=19, differentiated n=15. For S-type cells; proliferating To understand further the relationship between store depletion n=16, differentiated n=14. and subsequent Ca2+ entry during the switch from proliferation to N-type S-type differentiation, TG responses (store depletion) were plotted against Ca2+ responses for individual N- and S-type cell populations (Fig. 5) Rate of rise (μMs−1) Proliferating Differentiated Proliferating Differentiated and analysed using linear models to investigate the effect of TG response TG 140±0.24 2.00±0.27 2.13±0.29 3.32±0.76 and 9cRA treatment on subsequent Ca2+ entry. For N-type cells there Ca2+ 140±0.10 1.66±0.14 3.05±0.41 2.78±0.50 was a significant effect of store depletion on Ca2+ entry (Pb0.001) and asignificant negative effect of 9cRA treatment (i.e. reduction in the inter- ceptofthelinerelatingTGtoCa2+ response; P=0.011) consistent with a 2+ In N-type cells the extent of depletion from ER Ca stores in response down-regulation of Ca2+ entry in differentiated cells (Fig. 5A). The slopes to the addition of TG was not significantly different between proliferating of the lines relating the magnitudes of store depletion to Ca2+ entry did μ μ and differentiated cells; 9.45±0.78 Ms(n=24) vs. 7.50±0.76 Ms not differ significantly in relation to 9cRA treatment (P=0.222). For 2+ (n=18), P=0.088 (Fig. 4B). However, subsequent Ca entry following S-type cells, there was no statistically-significant effect of store depletion 2+ fi 2+ add-back of Ca was signi cantly reduced by~48% in differentiated cells or 9cRA treatment on Ca entry (model overall, P=0.262; TG effect, compared to proliferating cells; 14.85±1.74 μM s vs. 7.64±1.83 μMs, P=0.089; 9cRA effect and interaction P=0.338) (Fig. 5C). P=0.002 (Fig. 4B). The SOCE pathway therefore becomes down- The rates of rise of both TG and Ca2+ responses were also deter- regulated in differentiated N-type cells. This result is comparable to mined. There was a significant difference in the rates of rise for that seen in the parental SH-SY5Y cell line (Fig. 4A), and also that previ- store depletion (P=0.015) and Ca2+ entry (Pb0.001) between N- ously observed in this laboratory [19]. and S-type cells (Table 1). However, the rates of rise for store deple- 2+ In S-type cells the extent of depletion from ER Ca stores in response tion and Ca2+ entry did not change significantly in response to to TG was not significantly different between proliferating and differ- 9cRA treatment (store depletion, P=0.052; Ca2+ entry, P=0.944) μ μ entiated cells; 6.63±0.77 Ms(n=15) vs. 6.50±1.10 Ms(n=16), regardless of cell type (2-way ANOVA interaction terms, P=0.175) (Table 1). Although the rates of rise for store depletion did not strictly meet the criteria for statistical significance, the rather low probability of no effect in this case raises the possibility that 9cRA treatment may increase sensitivity to TG in both cell types.

3.6. STIM1, Orai1 and TRPC1 in N- and S-type cells

The proteins STIM1, Orai1 and TRPC1 have previously been shown to play a key role in SOCE [20–23]. To determine whether these proteins were involved in the SOCE activity measured in N- and S-type cells, their expression was measured by western blot analysis. STIM1 was identified in both N- and S-type cells. After 9cRA-induced differentiation, the level of STIM1 expression became down-regulated by ~49%, P=0.036 in N-type cells, though remained unchanged in S-type cells, P=0.905, n=4 (Fig. 6A,C). These changes were consistent with the changes observed in SOCE (Fig. 4). Orai1 was also identified in both N- and S-type cells. Following 9cRA-induced differentiation the level of Orai1 expression became down-regulated by ~64%, P=0.038 in N-type cells, and remained unchanged in S-type cells, P=0.661, n=4 (Fig. 6B, C). The changes observed in Orai1 expression are consistent both with the changes observed in STIM1 expression (Fig. 6A, C), and also the level of SOCE activity measured in SH-SY5Y, N- and S-type cells (Fig. 4). TRPC1 was identified in N-type cells (Fig. 6A, C). After 9cRA-induced differentiation TRPC1 expression became up-regulated by ~52% (P=0.032, n=3) in N-type cells, suggesting that TRPC1 does not form a SOC in these cells since SOCE becomes down-regulated (Fig. 4B). Though proliferating and differentiated S-type cells display SOCE (Fig. 4C), TRPC1 was not identified in S-type cells (Fig. 6A, C). In proliferating N-type cells, a double band was observed in response to STIM1 and TRPC1 antibodies that was not observed in differentiated N-type cells (Fig. 6A). The parental SH-SY5Y cell line was comparable to the N-type cells (Fig. 6A, B).

Fig. 3. N- and S-type cells show differential expression of β-tubulin III, vimentin and Bcl-2. P=proliferating (vehicle EtOH), D=differentiated (1 μM9cRA). β-actin was 4. Discussion used as a loading control. (A) Western blot showing expression of β-tubulin III. In differentiated S-type cells expression became down-regulated compared to proliferating Previous ﬁndings from this laboratory have shown that SOCE be- S-type cells. (B) Western blots showing expression of vimentin and Bcl-2. Vimentin comes down-regulated in 9cRA-differentiated SH-SY5Y neuroblasto- was only detected in S-type cells. Bcl-2 expression became up-regulated in differenti- ma cells [19]. The aims of this study were to characterise SOCE in N- ated SH-SY5Y and N-type cells and was barely detectable in S-type cells. (C) Histogram showing relative expression of β-tubulin III, vimentin and Bcl-2 as a ratio of β-actin ex- and S-type cell populations and to investigate the involvement of pression. The parental SH-SY5Y cell line was comparable to N-type cells. STIM1, Orai1 and TRPC1 in relation to SOCE activity. N- and S-type 648 N. Bell et al. / Biochimica et Biophysica Acta 1833 (2013) 643–651

Fig. 4. SOCE pathway activity in proliferating and 9cRA differentiated N- and S-type cells. Cells were induced to differentiate by the addition of 1 μM9cRA for 7 days. The addition of 2+ 2+ 2+ 2+ TG (200 nM) to fura-2 loaded cells depletes ER Ca stores which is observed as an increase in [Ca ]i and the re-addition of extracellular Ca (CaCl2 , 2 mM) to Ca -free buffer is 2+ 2+ 2+ also observed as an increase in [Ca ]i as Ca enters the cells (calibrated traces show [Ca ]i (40 nM) against time (200 s)); ( ) proliferating, (■) differentiated. After 9cRA-induced differentiation, SOCE became in (A) SH-SY5Y cells, significantly down-regulated (*), in (B) N-type cells, significantly down-regulated (*) and in (C) S-type cells, unchanged. cells were enriched from the parental SH-SY5Y cell line and the ex- β-tubulin III, a microtubule protein found in neuronal cells [38,39], pression of proteins previously shown to be specific to N and S cell was expressed in N-type cells, and expression did not change following lines was determined to confirm neuronal and non-neuronal lineages 9cRA-induced differentiation. β-tubulin III therefore appears to be a respectively. marker of neuronal lineage and not of differentiation itself. β-tubulin III Differentiation of N- and S-type cells with 9cRA highlighted mor- was also, surprisingly, expressed in S-type cells. Expression was weak phological differences between the two cell types with N-type cells be- as judged by immunofluorescence but clearly present as determined coming more neuronal-like and S-type cells becoming more epithelial- by western blot. S-type cells may have only recently committed to a like. non-neuronal lineage through the process of transdifferentiation [3,7]. N. Bell et al. / Biochimica et Biophysica Acta 1833 (2013) 643–651 649

Fig. 6. N- and S-type cells show differential expression of STIM1, Orai1 and TRPC1. P=proliferating (vehicle EtOH), D=differentiated (1 μM9cRA). β-actin was used as a loading control. (A) Western blots showing expression of TRPC1 and STIM1. TRPC1 expression became up-regulated in N-type after differentiation and was not expressed in S-type cells. STIM1 expression became down-regulated in SH-SY5Y and N-type cells after 9cRA-induced differentiation and remained unchanged in S-type cells. (B) Western blot showing expression of Orai1. Orai1 expression became Fig. 5. Relationship between store depletion and subsequent Ca2+ entry in ( ) prolif- down-regulated in SH-SY5Y and N-type cells after 9cRA-induced differentiation and erating and (■) differentiated N- and S-type cell populations. (A) For N-type cells the remained unchanged in S-type cells. (C) Histogram showing relative expression of overall model was statistically signiﬁcant (Pb0.001). (B) For S-type cells the overall β-tubulin III, vimentin and Bcl-2 as a ratio of β-actin expression. The parental model was not signiﬁcant, P=0.262. Analysis was in R. SH-SY5Y cell line was comparable to N-type cells.

Interestingly, β-tubulin III expression became further down-regulated in The activity of the SOCE pathway was consistent with the level differentiated S-type cells. This is consistent with S-type cells moving of expression of STIM1 and Orai1 in both cell types. In N-type cells away from a neuronal phenotype; it has been found that after RA SOCE became down-regulated following 9cRA-induced differentiation, treatment S-type cells differentiate into Schwann cells [47] and as did the expression of STIM1 and Orai1. In S-type cells SOCE remained also melanocytic cells [48]. unchanged following 9cRA-induced differentiation, as did the expres- Vimentin, an intermediate filament protein found in non-neuronal sion of STIM1 and Orai1. This is consistent with STIM1 as the ER Ca2+ cells [3], was expressed in S-type cells. Vimentin was not generally sensor [20,22] in both N- and S-type cells and with Orai1 forming, expressed in N-type cells but it was identified in some neurite exten- at least in part, the SOC [21,22] in both N- and S-type cells. TRPC1 sions. Vimentin is, however, an essential transient requirement for was present in proliferating N-type cells and expression became up- the initiation of neurite outgrowth in NB2a neuroblastoma cells and regulated following differentiation. TRPC1 has been found to function also in hippocampal neurons where knockdown of vimentin signifi- as a SOC when associated with STIM1 [28,30,31] and may therefore cantly inhibited neurite outgrowth [49,50]. function as a SOC in proliferating N-type cells, however this was not de- The anti-apoptotic protein Bcl-2, which is widely expressed in sym- termined. It is unlikely that TRPC1 functions as a SOC in differentiated pathetic neurons [37], was present in proliferating N-type cells and be- N-type cells as TRPC1 expression became up-regulated whereas SOCE came up-regulated in differentiated N-type cells [44–46], consistent became down-regulated. In HEK293T cells down-regulation of STIM1 with the neuronal lineage of the N-type cells. Bcl-2 expression was allowed TRPC1 to function as a ROC, insensitive to store depletion barely detectable in both proliferating and differentiated S-type cells. [30]. If TRPC1 functions as a ROC in differentiated N-type cells this The differential expression of β-tubulin III, vimentin and Bcl-2 ob- could explain our previous observation that in 9cRA-differentiated served confirmed the morphological and biochemical enrichment of SH-SY5Y cells a non-SOCE pathway becomes up-regulated [19].TRPC1 N- and S-type cells from the parental SH-SY5Y cell line. After enrich- may be associated with differentiation itself; TRPC1 was required for ment, SOCE was measured in N- and S-type cell populations. The results neurite outgrowth in differentiating PC12 cells but was independent revealed that the down-regulation of SOCE previously observed in of SOCE [51]. TRPC1 expression was not detectable in either proliferat- SH-SY5Y cells [19], and also in this study, is a feature of N-type cells ing or differentiated S-type cells, which would indicate that TRPC1 and not S-type cells. does not function as a SOC in S-type cells. An increased expression of 650 N. Bell et al. / Biochimica et Biophysica Acta 1833 (2013) 643–651

Bcl-2 has been associated with an inhibition of SOCE [52,53] and may [10] N. Sidell, Retinoic acid-induced growth inhibition and morphologic differentiation of human neuroblastoma cells in vitro, J. Natl. Cancer Inst. 68 (1982) 589–596. therefore also play a role in SOCE down-regulation in N-type cells. [11] K.K. Matthay, J.G. Villablanca, R.C. Seeger, D.O. Stram, R.E. Harris, N.K. Ramsay, P. Proliferating N- and S-type cells had similar SOCE profiles yet clearly Swift, H. Shimada, C.T. Black, G.M. Brodeur, R.B. Gerbing, C.P. Reynolds, Treatment the composition of proteins varies between cell types as Orai1 expression of high-risk neuroblastoma with intensive chemotherapy, radiotherapy, autolo- was generally lower in S-type cells and TRPC1 was not present. Differ- gous bone marrow transplantation, and 13-cis-retinoic acid. Children's Cancer Group, N. Engl. J. Med. 341 (1999) 1165–1173. ences in protein composition were further observed between differenti- [12] K.K. Matthay, C.P. Reynolds, R.C. Seeger, H. Shimada, E.S. Adkins, D. Haas-Kogan, ated N- and S-type cells as the expression of STIM1, Orai1 and TRPC1 R.B. Gerbing, W.B. London, J.G. Villablanca, Long-term results for children with changed in N-type cells. SOC composition could help explain the different high-risk neuroblastoma treated on a randomized trial of myeloablative therapy 2+ followed by 13-cis-retinoic acid: a Children's Oncology Group Study, J. Clin. rates of Ca entry observed between N- and S-type cells. In S-type cells Oncol. 27 (2009) 1007–1013. there was no change in SOCE following differentiation, yet S-type cell [13] M.J. Berridge, Inositol trisphosphate and calcium signalling, Nature 361 (1993) – populations displayed variable SOCE activity, as judged by the scatter in 315 325. 2+ 2+ [14] E. Carafoli, Calcium signaling: a tale for all seasons, Proc. Natl. Acad. Sci. U. S. A. 99 the relationship between Ca release and Ca entry. It may be that, (2002) 1115–1122. as multi-potent precursors to several cell types [3,4], S-type cells repre- [15] A. Buonanno, R.D. Fields, Gene regulation by patterned electrical activity during sent sub-populations with variable SOC compositions. In proliferating neural and skeletal muscle development, Curr. Opin. Neurobiol. 9 (1999) 110–120. [16] M.J. Berridge, P. Lipp, M.D. Bootman, The versatility and universality of calcium N-type cells double bands were detected by STIM1 and TRPC1 antibodies signalling, Nat. Rev. Mol. Cell Biol. 1 (2000) 11–21. that were not present in differentiated N-type cells. One possibility is that [17] J.W. Putney Jr., D.L. Aub, C.W. Taylor, J.E. Merritt, Formation and biological action both STIM1 and TRPC1 exist in two states in proliferating N-type cells and of inositol 1,4,5-trisphosphate, Fed. Proc. 45 (1986) 2634–2638. [18] A.B. Parekh, J.W. Putney Jr., Store-operated calcium channels, Physiol. Rev. 85 in only one state in differentiated N-type cells. The two states detected in (2005) 757–810. proliferating cells could be as a result of covalent modifications, as has [19] A.M. Brown, F.C. Riddoch, A. Robson, C.P. Redfern, T.R. Cheek, Mechanistic and been observed with STIM1 [54]. Clearly the functional consequence of functional changes in Ca2+ entry after retinoic acid-induced differentiation of – any covalent modificationsneedstobeelucidated. neuroblastoma cells, Biochem. J. 388 (2005) 941 948. [20] J. Roos, P.J. DiGregorio, A.V. Yeromin, K. Ohlsen, M. Lioudyno, S. Zhang, O. Safrina, In summary, our results reveal that the 9cRA-induced switch from J.A. Kozak, S.L. Wagner, M.D. Cahalan, G. Velicelebi, K.A. Stauderman, STIM1, an proliferation to differentiation in N-type cells is accompanied by a essential and conserved component of store-operated Ca2+ channel function, – remodelling of SOCE that is consistent with changes in the level J. Cell Biol. 169 (2005) 435 445. [21] S. Feske, Y. Gwack, M. Prakriya, S. Srikanth, S.H. Puppel, B. Tanasa, P.G. Hogan, R.S. of expression of STIM1 and Orai1. The finding that SOCE is down- Lewis, M. Daly, A. Rao, A mutation in Orai1 causes immune deficiency by abrogat- regulated in N-type cells but not in S-type cells strongly suggests ing CRAC channel function, Nature 441 (2006) 179–185. 2+ [22] R.M. Luik, M.M. Wu, J. Buchanan, R.S. Lewis, The elementary unit of store- that remodelling of Ca signalling is linked to the process of neuro- 2+ 2+ operated Ca entry: local activation of CRAC channels by STIM1 at ER-plasma nal differentiation. That remodelling of Ca signalling is a feature of membrane junctions, J. Cell Biol. 174 (2006) 815–825. the N-type cells could have important implications for neuroblastoma [23] X. Liu, K.T. Cheng, B.C. Bandyopadhyay, B. Pani, A. Dietrich, B.C. Paria, W.D. Swaim, as tumours with a predominantly N-phenotype are more aggressive than D. Beech, E. Yildrim, B.B. Singh, L. Birnbaumer, I.S. Ambudkar, Attenuation of store-operated Ca2+ current impairs salivary gland fluid secretion in TRPC1(-/-) tumours with a predominantly S-phenotype [2].Thedown-regulationof mice, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 17542–17547. SOCE in N-type cells may be required to either drive or maintain the [24] J. Liou, M.L. Kim, W.D. Heo, J.T. Jones, J.W. Myers, J.E. Ferrell Jr., T. Meyer, STIM is a switch from proliferation to differentiation, and the roles of STIM1, Ca2+ sensor essential for Ca2+-store-depletion-triggered Ca2+ influx, Curr. Biol. 15 (2005) 1235–1241. Orai1 and TRPC1 in this switch remain to be determined. Multiple studies [25] M.M. Wu, J. Buchanan, R.M. Luik, R.S. Lewis, Ca2+ store depletion causes STIM1 to have implicated the involvement of SOCE proteins in proliferation accumulate in ER regions closely associated with the plasma membrane, J. Cell [55–58] and in differentiation [51,59,60]. It is likely that these proteins Biol. 174 (2006) 803–813. play a role in the differentiation process itself and may therefore provide [26] M. Prakriya, S. Feske, Y. Gwack, S. Srikanth, A. Rao, P.G. Hogan, Orai1 is an essential pore subunit of the CRAC channel, Nature 443 (2006) 230–233. a potential therapeutic target in the treatment of neuroblastoma. [27] G.M. Salido, S.O. Sage, J.A. Rosado, TRPC channels and store-operated Ca2+ entry, Biochim. Biophys. Acta 1793 (2009) 223–230. Acknowledgments [28] J.P. Yuan, W. Zeng, G.N. Huang, P.F. Worley, S. Muallem, STIM1 heteromultimerizes TRPC channels to determine their function as store-operated channels, Nat. Cell Biol. 9 (2007) 636–645. This work was supported by a BBSRC studentship (to NB) and a [29] I. Jardin, J.J. Lopez, G.M. Salido, J.A. Rosado, Orai1 mediates the interaction be- MRC project grant (to TRC). tween STIM1 and hTRPC1 and regulates the mode of activation of hTRPC1- forming Ca2+ channels, J. Biol. Chem. 283 (2008) 25296–25304. [30] S. Alicia, Z. Angelica, S. Carlos, S. Alfonso, L. Vaca, STIM1 converts TRPC1 from a References receptor-operated to a store-operated channel: moving TRPC1 in and out of lipid rafts, Cell Calcium 44 (2008) 479–491. [1] G.M. Brodeur, Neuroblastoma, in: P.A. Pizzo, D.G. Poplack (Eds.), Principles and [31] I. Jardin, G.M. Salido, J.A. Rosado, Role of lipid rafts in the interaction between Practices of Pediatric Oncology, 3rd ed., Lippencott and Raven, Philadelphia, 1997. hTRPC1, Orai1 and STIM1, Channels (Austin) 2 (2008) 401–403. [2] J.D. Walton, D.R. Kattan, S.K. Thomas, B.A. Spengler, H.F. Guo, J.L. Biedler, N.K. [32] Y. Liao, C. Erxleben, E. Yildirim, J. Abramowitz, D.L. Armstrong, L. Birnbaumer, Orai Cheung, R.A. Ross, Characteristics of stem cells from human neuroblastoma cell proteins interact with TRPC channels and confer responsiveness to store deple- lines and in tumors, Neoplasia 6 (2004) 838–845. tion, Proc. Natl. Acad. Sci. U.S. A. 104 (2007) 4682–4687. [3] V. Ciccarone, B.A. Spengler, M.B. Meyers, J.L. Biedler, R.A. Ross, Phenotypic diver- [33] Y. Liao, C. Erxleben, J. Abramowitz, V. Flockerzi, M.X. Zhu, D.L. Armstrong, L. sification in human neuroblastoma cells: expression of distinct neural crest line- Birnbaumer, Functional interactions among Orai1, TRPCs, and STIM1 suggest a ages, Cancer Res. 49 (1989) 219–225. STIM-regulated heteromeric Orai/TRPC model for SOCE/Icrac channels, Proc. [4] R.A. Ross, B.A. Spengler, C. Domenech, M. Porubcin, W.J. Rettig, J.L. Biedler, Human Natl. Acad. Sci. U. S. A. 105 (2008) 2895–2900. neuroblastoma I-type cells are malignant neural crest stem cells, Cell Growth [34] K.T. Cheng, X. Liu, H.L. Ong, I.S. Ambudkar, Functional requirement for Orai1 in Differ. 6 (1995) 449–456. store-operated TRPC1–STIM1 channels, J. Biol. Chem. 283 (2008) 12935–12940. [5] R.A. Ross, J.L. Biedler, B.A. Spengler, A role for distinct cell types in determining [35] J.L. Biedler, B.A. Spengler, K.M. Lyser, Morphological interconversion of human malignancy in human neuroblastoma cell lines and tumors, Cancer Lett. 197 neuroblastoma cells, In Vitro 10 (1975) 380. (2003) 35–39. [36] I. Garcia, I. Martinou, Y. Tsujimoto, J.C. Martinou, Prevention of programmed cell [6] R.A. Ross, B.A. Spengler, Human neuroblastoma stem cells, Semin. Cancer Biol. 17 death of sympathetic neurons by the bcl-2 proto-oncogene, Science 258 (1992) (2007) 241–247. 302–304. [7] R.A. Ross, B.A. Spengler, J.L. Biedler, Coordinate morphological and biochemical [37] D.E. Merry, D.J. Veis, W.F. Hickey, S.J. Korsmeyer, bcl-2 protein expression is wide- interconversion of human neuroblastoma cells, J. Natl. Cancer Inst. 71 (1983) spread in the developing nervous system and retained in the adult PNS, Develop- 741–747. ment 120 (1994) 301–311. [8] M. Piacentini, L. Piredda, D.T. Starace, M. Annicchiarico-Petruzzelli, M. Mattei, S. [38] E. Draberova, Z. Lukas, D. Ivanyi, V. Viklicky, P. Draber, Expression of class III Oliverio, M.G. Farrace, G. Melino, Differential growth of N- and S-type human beta-tubulin in normal and neoplastic human tissues, Histochem. Cell Biol. 109 neuroblastoma cells xenografted into scid mice. Correlation with apoptosis, (1998) 231–239. J. Pathol. 180 (1996) 415–422. [39] C.D. Katsetos, A. Legido, E. Perentes, S.J. Mork, Class III beta-tubulin isotype: a key [9] B.A. Spengler, D.L. Lazarova, R.A. Ross, J.L. Biedler, Cell lineage and differentiation cytoskeletal protein at the crossroads of developmental neurobiology and tumor state are primary determinants of MYCN gene expression and malignant poten- neuropathology, J. Child Neurol. 18 (2003) 851–866. tial in human neuroblastoma cells, Oncol. Res. 9 (1997) 467–476. (discussion 867) N. Bell et al. / Biochimica et Biophysica Acta 1833 (2013) 643–651 651

[40] G. Grynkiewicz, M. Poenie, R.Y. Tsien, A new generation of Ca2+ indicators with [51] D.K. Heo, W.Y. Chung, H.W. Park, J.P. Yuan, M.G. Lee, J.Y. Kim, Opposite regulatory greatly improved fluorescence properties, J. Biol. Chem. 260 (1985) 3440–3450. effects of TRPC1 and TRPC5 on neurite outgrowth in PC12 cells, Cell. Signal. 24 [41] V. R Development Core Team, R: A Language and Environment for Statistical (2012) 899–906. Computing, R Foundation for Statistical Computing, Austria, 2011. http://www. [52] P. Pinton, D. Ferrari, P. Magalhaes, K. Schulze-Osthoff, F. Di Virgilio, T. Pozzan, R. R-project.org. Rizzuto, Reduced loading of intracellular Ca2+ stores and downregulation of [42] S. Pahlman, A.I. Ruusala, L. Abrahamsson, M.E. Mattsson, T. Esscher, Retinoic capacitative Ca2+ influx in Bcl-2-overexpressing cells, J. Cell Biol. 148 (2000) acid-induced differentiation of cultured human neuroblastoma cells: a compari- 857–862. son with phorbolester-induced differentiation, Cell Differ. 14 (1984) 135–144. [53] P. Pinton, D. Ferrari, E. Rapizzi, F. Di Virgilio, T. Pozzan, R. Rizzuto, The Ca2+ con- [43] G. Nicolini, M. Miloso, C. Zoia, A. Di Silvestro, G. Cavaletti, G. Tredici, Retinoic acid centration of the endoplasmic reticulum is a key determinant of ceramide- differentiated SH-SY5Y human neuroblastoma cells: an in vitro model to assess induced apoptosis: significance for the molecular mechanism of Bcl-2 action, drug neurotoxicity, Anticancer Res. 18 (1998) 2477–2481. EMBO J. 20 (2001) 2690–2701. [44] M. Hanada, S. Krajewski, S. Tanaka, D. Cazals-Hatem, B.A. Spengler, R.A. Ross, J.L. [54] J.T. Smyth, J.G. Petranka, R.R. Boyles, W.I. DeHaven, M. Fukushima, K.L. Johnson, Biedler, J.C. Reed, Regulation of Bcl-2 oncoprotein levels with differentiation of J.G. Williams, J.W. Putney Jr., Phosphorylation of STIM1 underlies suppression of human neuroblastoma cells, Cancer Res. 53 (1993) 4978–4986. store-operated calcium entry during mitosis, Nat. Cell Biol. 11 (2009) 1465–1472. [45] A. Lasorella, A. Iavarone, M.A. Israel, Differentiation of neuroblastoma enhances [55] I.F. Abdullaev, J.M. Bisaillon, M. Potier, J.C. Gonzalez, R.K. Motiani, M. Trebak, Bcl-2 expression and induces alterations of apoptosis and drug resistance, Cancer Stim1 and Orai1 mediate CRAC currents and store-operated calcium entry impor- Res. 55 (1995) 4711–4716. tant for endothelial cell proliferation, Circ. Res. 103 (2008) 1289–1299. [46] F.C. Riddoch, A.M. Brown, S.E. Rowbotham, C.P. Redfern, T.R. Cheek, Changes in [56] M. Potier, J.C. Gonzalez, R.K. Motiani, I.F. Abdullaev, J.M. Bisaillon, H.A. Singer, M. functional properties of the caffeine-sensitive Ca2+ store during differentiation Trebak, Evidence for STIM1- and Orai1-dependent store-operated calcium influx

of human SH-SY5Y neuroblastoma cells, Cell Calcium 41 (2007) 195–206. through ICRAC in vascular smooth muscle cells: role in proliferation and migration, [47] M.Tsokos,S.Scarpa,R.A.Ross,T.J.Triche,Differentiationofhumanneuroblas- FASEB J. 23 (2009) 2425–2437. toma recapitulates neural crest development. Study of morphology, neuro- [57] S.L. Zhang, Y. Yu, J. Roos, J.A. Kozak, T.J. Deerinck, M.H. Ellisman, K.A. Stauderman, transmitter enzymes, and extracellular matrix proteins, Am. J. Pathol. 128 M.D. Cahalan, STIM1 is a Ca2+ sensor that activates CRAC channels and migrates (1987) 484–496. from the Ca2+ store to the plasma membrane, Nature 437 (2005) 902–905. [48] R. Slack, B. Lach, A. Gregor, H. al-Mazidi, P. Proulx, Retinoic acid- and staurosporine- [58] J.J. Zou, Y.D. Gao, S. Geng, J. Yang, Role of STIM1/Orai1-mediated store-operated induced bidirectional differentiation of human neuroblastoma cell lines, Exp. Cell Ca2+ entry in airway smooth muscle cell proliferation, J. Appl. Physiol. 110 Res. 202 (1992) 17–27. (2011) 1256–1263. [49] T.B. Shea, M.L. Beermann, I. Fischer, Transient requirement for vimentin in [59] B. Darbellay, S. Arnaudeau, S. Konig, H. Jousset, C. Bader, N. Demaurex, L. neuritogenesis: intracellular delivery of anti-vimentin antibodies and antisense Bernheim, STIM1- and Orai1-dependent store-operated calcium entry regulates oligonucleotides inhibit neurite initiation but not elongation of existing neurites human myoblast differentiation, J. Biol. Chem. 284 (2009) 5370–5380. in neuroblastoma, J. Neurosci. Res. 36 (1993) 66–76. [60] S.J. Graham, M.J. Black, J. Soboloff, D.L. Gill, M.A. Dziadek, L.S. Johnstone, Stim1, an [50] L.J. Boyne, I. Fischer, T.B. Shea, Role of vimentin in early stages of neuritogenesis in endoplasmic reticulum Ca2+ sensor, negatively regulates 3T3-L1 pre-adipocyte cultured hippocampal neurons, Int. J. Dev. Neurosci. 14 (1996) 739–748. differentiation, Differentiation 77 (2009) 239–247.