The mitochondrial pathway of apoptosis acts as a sentinel for multiple stress signals, including inflammation, oxidative stress, amyloid β, and growth factor deprivation, that are often encountered by cells of neuronal origin (1). The critical point in the mitochondrial pathway of apoptosis is the release of cytochrome c from the intermembrane mitochondrial space to the cytosol, where it reacts with Apaf-1 and pro-caspase 9 forming a large scaffold complex, the apoptosome, on which the initiator caspase 9 undergoes activation (2). Often considered as a point-of-no-return, the release of cytochrome c does not necessarily leads to cell death (3, 4). Neuronal cells, for example, can survive the release of cytochrome c from mitochondria, providing that caspases are not activated and mitochondrial transmembrane potential is not completely lost (5).
Not surprisingly, the release of cytochrome c is heavily regulated, mostly by members of the Bcl-2 family of proteins that act upstream of mitochondria (6–8). Moreover, we have recently shown that the apoptotic activity of cytosolic cytochrome c can be also regulated by the protein neuroglobin, leading to a significant protection from the intrinsic pathway of apoptotic cell death (9, 10). Neuroglobin was originally discovered during a bioinformatics search of the human genome, and it is expressed at moderate levels in neurons and the inner segments of the retinal photoreceptor cells, reaching up to 100 μM, in association with mitochondria (11–13). Our investigations into the function of neuroglobin indicate that a very rapid reaction occurs between ferrous neuroglobin and ferric cytochrome c (14). This reaction appears to be composed of two steps, with complex formation facilitating subsequent electron exchange between the two proteins, and renders cytochrome c apoptotically inactive (10, 14).
In this study, guided by the analysis of the computational model of the lowest energy putative structure of the cytochrome c-neuroglobin complex, we have investigated whether mutation of the selected amino acid residues within the cytochrome c-binding surface of neuroglobin, predicted to be involved in the docking process, can affect the antiapoptotic activity of this protein in neuroblastoma SH-SY5Y cells. We have observed that the neuroglobin mutant Lys67Glu delays cell death more significantly than wild-type neuroglobin. The increase in cumulative inhibition of caspase 9 activity by Lys67Glu neuroglobin, as compared to that of wild-type neuroglobin, was associated with a significantly enhanced inhibition of the mitochondrial outer membrane permeabilization (MOMP) and preservation of the transmembrane mitochondrial potential (ΔψM). The identification of a mutant version of neuroglobin with an increased antiapoptotic activity will prompt further studies on the delivery of neuroglobin as a novel approach for prevention of damage in the brain and in the retina (15–17).
MATERIALS AND METHODS
Computer Modeling and Selection of Interface Amino Acid Residues
Potential protein–protein complex structures for the product of the binding of cytochrome c and neuroglobin were calculated using the soft docking computer program BiGGER (18). The structures of cytochrome c (1HRC, (19)) and neuroglobin (1OJ6, (20)) were obtained from the protein structure data base. The structure of cytochrome c was used without further modification while the reported crystal structure of neuroglobin was used either as reported or modified using YASARA (www.yasara.org) to back mutate the residues Gly46, Ser55 and Ser120 to the wild-type Cys residues (21). The resulting set of complex structures was further filtered using a 15-Å cut off for the Fe-Fe distance in the complexes, to exclude those few structures which, due to large Fe-Fe distances, would have intra-molecular transfer rates too slow to account for the observed fast reduction of ferric cytochrome c by ferrous neuroglobin (22). Energy minimization of best ranked protein complex structure was performed using YASARA (23). Energy minimization was performed in a box of 4,000 explicit water molecules using the AMBER99 (24) force field. Energy minimization used a steepest decent search followed by simulated annealing of both water and protein to avoid local minima. The resulting energy minimized complex structure was then manually inspected to identify amino acids involved in interfacial interactions. The RMSD between the energy minimised wild-type and Cys mutant forms of the complex was 0.14 Å. No differences in the interprotein interactions were apparent between the wild-type and Cys mutant forms of the complex.
Site-directed Mutagenesis and Plasmid Preparations
Vector expressing human neuroglobin was generated using a human cDNA clone from Origene (Rockville, MD) subcloned into the KpnI and XbaI sites of pcDNA3.1 and the construct verified by direct sequencing. The form of neuroglobin termed hereafter mut3 Ngb has Cys47Gly, Cys56Ser, and Cys121Ser mutations to avoid any confounding effects of S-S formation, and it is referred to as Ngb throughout the text. The following mutations were introduced on the background of mut3 Ngb, with the use of site-directed mutagenesis (GeneScript, US): Glu87Asp (GAG→GAC), Arg94Glu (AGG→GAG), and Lys67Glu (AAG→GAG). The constructs were verified by direct sequencing. pEGFP-C1 (kindly provided by Joanna Dodd, SBS, UoA) was used in co-transfection experiments. All plasmids used for transfection of mammalian cells were propagated using DH5α competent cells, according to standard protocols, and purified using EndoFree Maxi Prep kits (Qiagen).
Mammalian Cell Culture and Treatments
Human neuroblastoma SH-SY5Y cells were grown in advanced DMEM/F12 (1/1) medium (Invitrogen) supplemented with L-glutamine, 1% penicillin/streptomycin, and 10% fetal bovine serum (Invitrogen) at 37°C in humidified 95% air, 5% CO2. To induce cell death, SH-SY5Y cells were treated with a small-molecule BH3 mimetic TW-37 (Selleck). TW-37 binds with high affinities to Bcl-2, Bcl-Xl, and Mcl-1 (25), and induces dose-dependent cell death in SH-SY5Y cells in a single agent-treatment scenario.
Mammalian Cell Transfection and Stable Cell Lines
SH-SY5Y cells were plated at a density of 5 × 104 cells/well on 24-well plates, and co-transfected with pcDNA3.1 (empty, wt_Ngb, mut3_Ngb, mut67_Ngb, mut87_Ngb or mut94_Ngb) and pEGFP-C1, in a 2:1 ratio, using FuGene HD transfection reagent (Roche), according to manufacturer's protocol. Briefly, transfection was conducted over-night, with cells maintained in the antibiotic-free DMEM:F12 Advanced media (Invitrogen), supplemented with 2% FBS (Gibco), and using DNA:FuGene reagent ratio of 2:6. After over-night incubation cells were washed, and treated as indicated after another 48 h.
For establishment of stable cell lines, SH-SY5Y cells (1 × 105) were plated on 6-well culture dishes and transfected with pcDNA3.1, mut3_Ngb_pcDNA3.1 or mut67_Ngb_pcDNA3.1 as discussed earlier. 16 h post-transfection cells were washed three times with serum-free medium, and cultured in fresh medium for 48 h to allow for transgene expression. The medium was then replaced with selection medium containing G418 (400 μg/ml; Invitrogen). Three clones from each transfection were selected based on screening for neuroglobin expression using quantitative RT-PCR.
Total RNA was isolated from both adherent and floating cells using TRIzol (Invitrogen), according to the manufacturer's instructions. cDNA synthesis was performed using a High Capacity cDNA Reverse Transcription Kit with RNase Inhibitor (Life Technologies, Applied Biosystems), according to the manufacturer's instructions. Quantitative RT-PCRs were performed using 10 ng (RNA equivalents) of cDNA as a template. Gene-specific primers and probes, and TaqMan Gene Expression Master Mix were from Life Technologies ABI and amplifications were performed using an ABI Prism 7900 instrument. A custom designed assay (Life Technologies ABI), spanning the boundaries of exons 1 and 2 (amplicon size 126) was used for the assessment of relative expression of mut3 and mut67 neuroglobin variants. Data were normalized to 18S rRNA expression.
Flow Cytometric Analysis of Cell Death and Caspase 9 Activity
To study cell viability and the mitochondrial pathway of apoptosis, 5 × 104 cells/well were plated on 24-well plate, and trasnfected as described earlier. The next day cells were washed and treated as indicated (Figs. 1 and 2). At the end of the experiment, both floating and adherent cells were collected, and stained for 2–3 min with 7-AAD (Invitrogen). Following staining, cells were analyzed immediately on FACS Calibur using CellQuest Pro software (BD Biosciences). Cells were gated based on forward scatter (FSC) and side scatter (SSC) to exclude cell debris, and based on GFP fluorescence to analyze the population of transfected cells. No compensation was required between the FL1 and FL3 channel of FACS Calibur (BD).
To determine the activity of TW-37 in SH-SY5Y cells, we used three-color staining that included fluorescently labeled inhibitor of caspase 9 (FLICA) FITC-LEHD-FMK (Calbiochem), a marker of mitochondrial membrane potential TMRM (Invitrogen), and a marker of plasma membrane permeability 7-AAD (Invitrogen). We have previously determined that the maximal fluorescence of FLICA reagent corresponds to the complete activation of caspase 9 (26). Cells (0.15 × 106 per well) were cultured for the time indicated with or without the indicated doses of TW-37, harvested, and incubated with FLICA for 45 min at 37°C under 5% CO2 in darkness. The cells were washed three times, stained with TMRM (200 nM) for 30 min at 37°C under 5% CO2 in darkness, followed by staining (on ice) with 7-AAD and immediate analysis. Cells were analyzed using FACS Calibur (BD). For analysis, cells were gated based on forward scatter (FSC) and side scatter (SSC) to exclude cell debris, and based on FSC versus 7-AAD to exclude cells with plasma membrane permeability. Next cells were analysed for caspase 9 activity (FL1) and mitochondrial membrane depolarization (FL2). Plots were generated using FCS Express (De Novo Software, LA, CA).
FLICA reagents have a disadvantage that they contain an artificial cleavage site. To prove that caspase 9 activation measured with FLICA is representative of endogenous caspase substrates, we also detected the cleavage of endogenous caspase 9 substrate, caspase 3. Briefly, stable cells overexpressing mut3-Ngb and mut67-Ngb were treated with TW-37 (Selleck; 0–15 μM) for up to 72 h, fixed using 4% formaldehyde, and stained with a cleaved caspase-3 (Asp175) antibody conjugated to Alexa Fluor 488 (Cell Signalling) according to manufacturer's protocol. Cells were analyzed on FACS Calibur (BD), and plots were generated using FCS Express (De Novo Software, LA, CA). Raw data files can be obtained from j.skommer@auck land.ac.nz.
Cytochrome c Release
Quantitative analysis of cytochrome c release from the mitochondria was assessed by a modification of the method described by Waterhouse and Trapani (27), using InnoCyte Flow Cytometric Cytochrome c Release kit (Calbiochem, Merck), following the manufacturer's instructions with minor modifications (overnight staining with the cytochrome c antibody, at +4°C). Control cells were stained with a 1:300 dilution of the FITC-labeled anti-mouse IgG antibody in the absence of a primary antibody. Cell acquisition was performed on the FACS Calibur (BD Biosciences) using CellQuest Pro software (BD Biosciences). Raw data files can be obtained from email@example.com.
Differences between the groups were determined by t-test using SPSS Statistics 17.0.
Lys67Glu Neuroglobin Mutant Confers an Increased Short-term Protection from TW-37-induced Cell Death
The antiapoptotic action of neuroglobin within the mitochondrial pathway of apoptosis has been linked to its capability to bind cytochrome c, and reduce it to ferrous form which is ineffective in facilitating the activation of pre-caspase 9 within the apoptosome (10, 14). Computational docking has identified a lowest energy, putative structure for the cytochrome c-neuroglobin complex (Fig. 1). Analysis of this structure indicates the binding of cytochrome c Lys residues 25 and 27 to neuroglobin Glu residues 87 and 94, and a clash between cytochrome c residue Lys 79 and neuroglobin residue Lys 67.
We set out to determine whether mutation of these selected residues within the cytochrome c-binding surface of neuroglobin can affect its antiapoptotic activity in cells stimulated to undergo the mitochondrial pathway of apoptosis. To this aim, human neuroblastoma SH-SY5Y cells were transiently transfected with pcDNA3.1, mut3_Ngb (wild type with respect to other residues mutated here, and thus referred to simply as Ngb hereafter), Lys67Glu_Ngb, Glu87Asp_Ngb, and Arg94Glu_Ngb (all established on the mut3_Ngb background), using a co-transfection with pEGFP-C1 plasmid (in 2:1 ratio) as a marker of transiently transfected cells (GFP+; Supporting Information Fig. 1a). This approach eliminates the potentially confounding effects (e.g. changes in protein conformation and electron transfer efficiency) of direct neuroglobin tagging with EGFP. Importantly, transfection with pEGFP-C1 alone did not affect the sensitivity of SH-SY5Y cells towards cell death, as determined by comparing cell membrane permeability to 7-AAD in GFP- and GFP+ cell subpopulations (Supporting Information Fig. 1b).
We have recently reported that stable overexpression of neuroglobin protects SH-SY5Y cells from apoptosis induced by the BH3 mimetic HA14-1 (Raychaudhuri et al 2010). Here, we extend this observation using transient overexpression of wild-type neuroglobin, with co-transfected pEGFP as a marker of transiently transfected cells, and another BH3-mimetic, TW-37 (Fig. 2a). Cell death induced by TW-37 in wild-type SY-SY5Y cells was associated with activation of caspase 9, observed distinctively in the subpopulation of cells with a decreased mitochondrial transmembrane potential (ΔψM), as verified with the use of three-color staining incorporating the fluorescently labeled caspase 9 inhibitor FITC-LEHD-FMK (FLICA), TMRM (the mitochondrial transmembrane potential sensor), and 7-AAD (Fig. 2b). The cells were first gated based on FSC/7-AAD (Fig. 2b), to remove the autofluorescent cells with loss of plasma membrane integrity, and then a subpopulation of 7-AAD-negative cells was analyzed for TMRM and FLICA fluorescence. Treatment with TW-37 led to an increase in FLICA staining in cells with low TMRM fluorescence (from 0.6 to 8.2%; Fig. 2b), but not in cells with preserved ΔψM, confirming that caspase 9 activation in this cell system occurs at the post-mitochondrial stage of the apoptotic pathway. At any time point and dose of TW-37 analyzed, the population of cells with both an intact plasma membrane and active caspase 9 (as indicated by an increased FLICA fluorescence) did not exceed 22% (not shown). This is an expected result considering that the progression of apoptotic cell disintegration occurs within minutes following the activation of caspases, and that the main cell-to-cell variability in apoptotic signaling occurs at the premitochondrial stages and/or at the stage of the low probability event of the apoptosome assembly (28, 29). We also verified the dependence of TW-37-induced cell death on caspase 9 activity. Pre-treatment of SH-SY5Y cells with caspase 9 inhibitor z-LEHD-FMK significantly blocked TW-37-induced plasma membrane permeabilization, but not the loss of mitochondrial transmembrane potential (Figs. 2c and 2d), demonstrating that, in our cell system and within the timeframe of our analyses, depolarization of ΔψM during apoptosis is insensitive to inhibition of caspase 9. Using this validated in vitro system of caspase 9-dependent mitochondrial pathway of apoptosis we observed that a transient overexpression of one particular mutant version of neuroglobin, Lys67Glu, conferred a significantly increased protection from TW-37-induced cell death, as compared to Ngb (Fig. 2e).
Antiapoptotic Activity of Lys67Glu is Associated with Stabilization of ΔψM
The observation that Lys67Glu neuroglobin is more efficient at protecting cells from the mitochondrial pathway of apoptosis, and delaying cell death, was in conceptual agreement with our computational model of neuroglobin–cytochrome c complex, showing that Lys67 clashes with Lys79 on cytochrome c. We hypothesized that the removal of this clash could increase the antiapoptotic activity of Lys67Glu neuroglobin by allowing a more efficient prevention of cytochrome c-mediated caspase 9 activation. We thus established stable cell lines overexpressing mut3_Ngb or Lys67Glu_Ngb (Fig. 3a), an approach that allows studying larger and more homogenous cell populations, without the need for transient plasmid transfection. Stable cell lines expressing Ngb were significantly protected from TW-37-induced cell death as compared to pcDNA cell lines, and expression of Lys67Glu Ngb conferred an even stronger protection, with 80.1% ± 3.8% cell survival at 48 h of treatment with 10 μM TW-37 (Fig. 3b). We used these cell lines to assess the ability of Lys67Glu neuroglobin to inhibit caspase 9 activity, in cells undergoing the well-defined mitochondrial pathway of apoptosis, by immunostaining with an antibody that recognizes the large fragment of caspase 3 activated by cleavage adjacent to Asp175. Following 48 h of treatment with 10 μM TW-37, we observed a significant reduction of caspase 3 cleavage in cells expressing Ngb, and an even more pronounced inhibition in cells expressing Lys67Glu Ngb, as compared to pcDNA3.1 cells (Fig. 3c).
We then sought to determine whether inhibition of caspase 9 activity in Lys67Glu neuroglobin-overexpressing cells occurs at the post-mitochondrial stage of apoptosis. To this aim we used live cell assay based on 3-color FLICA/TMRM/7-AAD staining, as described earlier. This approach allows simultaneous assessment of multiple apoptotic events within single cells, with high sensitivity and without the need for plasmid transfections. In line with our previous observations (10), neuroglobin protected cells from TW-37 induced cell death (Fig. 3c), which was associated with a striking decrease in the number of cells with active caspase 9 (3.29% ± 0.32 %), as assessed by caspase 9 FLICA fluorescence, in the population of cells with low TMRM uptake and preserved plasma membrane integrity (Figs. 3d and 3e). TW-37-triggered depolarization of mitochondria was not affected by Ngb, in line with our postulated mechanism of neuroglobin action that relies on binding to cytochrome c, following its release from mitochondria, and rendering it incapable of activating caspase 9 (Fig. 3f) (10). Moreover, similarly to our earlier data obtained in transiently transfected cells, stable overexpression of the Lys67Glu neuroglobin mutant protected SH-SY5Y cells from TW-37-triggered cell death and activation of caspase 9 (Figs. 3d–3f). We also observed, however, that following treatment with 10 μM TW-37 a dramatic decrease in the uptake of TMRM was detectable in approximately 70% of pcDNA and Ngb stable cells, while only 15–30% of Lys67Glu Ngb-overexpressing cells exhibited loss of TMRM fluorescence (Figs. 3d and 3f). The remaining cell population showed only a small decrease in TMRM uptake 48 h after treatment with TW-37 (Fig. 3d). Of note, extended (96 h) treatment with TW-37 led to complete loss of TMRM uptake in Lys67Glu expressing cells, similarly to wt and mut3 neuroglobin expressing cells (data not shown).
Stabilization of ΔψM in Lys67Glu Neuroglobin Expressing Cells is Associated with Reduced Release of Cytochrome c from Mitochondria
Depolarization of ΔψM is widely considered to coincide with, or to occur shortly after, the mitochondrial outer membrane permeabilization (MOMP) (30), but nevertheless the uptake of membrane potential sensor TMRM is only an indirect measure of MOMP. We observed a partial reduction of TMRM fluorescence in Lys67Glu Ngb expressing cells, which could indicate that not all mitochondria within the affected cells underwent MOMP, or that the stabilization of ΔψM following the initial release of cytochrome c is achieved, either because there is sufficient cytochrome c to maintain ΔψM after permeabilization, or by diffusion of cytochrome c back to the mitochondrial inner membrane (31). Therefore, we assessed quantitatively cytochrome c translocation from mitochondria to the cytosol during the BH3-mimetic induced apoptosis in pcDNA3.1, mut3 Ngb, and Lys67Glu Ngb expressing SH-SY5Y cells, using a modification of the method described by Waterhouse and Trapani (27), and as described earlier (32). We observed that in pcDNA-transfected SH-SY5Y cells treated with TW-37 a significant number of cells (approx 30%) had diminished fluorescence (indicating loss of cytochrome c from the mitochondria). We found a similar number of cells with reduced fluorescence level in TW-37-treated Ngb-expressing cells, while Lys67Glu Ngb-expressing cells had a significantly lower (approximately 20%) number of cells with diminished fluorescence (Fig. 4). Of note, compared to previously published data, the difference in fluorescence intensity of apoptotic and non-apoptotic populations of SH-SY5Y cells was not very distinct, which may be because of the generally low level of cytochrome c expression and dim fluorescence following immunostaining with the anti-cytochrome c antibody.
Our quest to identify the residues that affect the antiapoptotic activity of neuroglobin was guided by the computational model of the neuroglobin–cytochrome c complex. Binding of cytochrome c to neuroglobin positions the two heme groups of these proteins for an efficient electron transfer and facilitates reduction of cytochrome c to its ferrous, apoptotically inactive, form (33). We have investigated the effect of selected mutant versions of neuroglobin, ectopically overexpressed in human neuroblastoma SH-SY5Y cells, on the mitochondrial pathway of apoptosis induced by a highly specific BH3 mimetic TW-37. TW-37 is a potent Bcl-2 inhibitor, binding to Bcl-2, Bcl-Xl, and Mcl-1 with K(i) values of 290 nM, 1100 nM, and 260 nM, respectively (34). We first showed that TW-37 induces apoptosis in neuroblastoma SH-SY5Y cells, in a single agent treatment scenario. Of note, another potent BH3 mimetic, ABT-737 (which binds to Bcl-2 and Bcl-Xl, but not Mcl-1) was incapable of inducing neuroblastoma cell death, suggesting that the efficient killing of neuroblastoma cells requires neutralization of Mcl-1 (data not shown). To establish the action of neuroglobin, and its mutant forms, within the mitochondrial pathway of apoptosis we first confirmed that in SH-SY5Y cells treatment with TW-37 provides a system of cell death which is caspase 9-dependent, and where caspase 9 activation occurs only in cells with a decreased mitochondrial transmembrane potential (ΔψM). In this system, one particular mutation, Lys67Glu, positioned at E10, significantly increased the pro-survival activity of neuroglobin. This observation was confirmed using both transient and stable protein overexpression systems. We observed a decreased cleavage of caspase 3, indicative of inhibited caspase 9 activity, in cells expressing Ngb, with further reduction in cells expressing the Lys67Glu Ngb mutant. A more detailed multiparameter live-cell analysis revealed that the enhanced antiapoptotic activity of Lys67Glu neuroglobin, as compared to wild-type neuroglobin, was associated not only with a decreased activity of caspase 9, but also with a stabilization of mitochondrial transmembrane potential and decreased MOMP, as determined using staining with TMRM and following the release of cytochrome c from mitochondria to the cytosol. Some previous studies suggested that that inhibition of caspase activity during an intrinsic death signal results in a failure of mitochondria to completely depolarize (35). However, in our system of TW-37-triggered apoptosis, inhibition of caspase 9 with z-LEHD-FMK did not affect the level of ΔψM loss. Thus, Lys67Glu neuroglobin may have, apart from its caspase inhibitory activity, an additional stabilizing effect on the mitochondrial transmembrane potential. Further studies assessing the single-cell spatio-temporal pattern of cytochrome c release in living wild-type and neuroglobin-overexpressing SH-SY5Y cells are required to establish the role of neuroglobin in MOMP dynamics during apoptosis of neuroblastoma cells.
The mitochondrial membrane permeability is influenced by the cross-talk between the mitochondria and the endoplasmic reticulum. Several Bcl-2 proteins, such as Bcl-2 and Bcl-XL, localize not only to mitochondrial outer membrane but also to the ER, and have been implicated in regulation of endoplasmic calcium flux. Interestingly, it has been recently shown in SH-SY5Y cells that Bcl-Xl protects mitochondria from calcium overload, preserving mitochondrial transmembrane potential and inhibiting PTP opening (36). In line with previously published observations on other BH3 mimetics, such as ABT-263 (37), HA14-1 or BH3I-2' (38), we have observed that the treatment with the BH3 mimetic TW-37 leads to an early increase in intracellular (cytoplasmic) calcium levels in live cell population gated by forward side scatter criteria (Supporting Information Fig. 2A). It was also previously shown that during apoptosis cytochrome c, following its release from mitochondria, can bind to the inositol 1,4,5-triphosphate receptors (IP3R) present in the ER adjacent to mitochondria (39). Cytochrome c binding to IP3R abolishes calcium-mediated inhibition of the receptor, augmenting calcium release from the ER, and provoking feed-forward amplification of cytochrome c release due to mitochondrial calcium overload and PTP-mediated decrease in mitochondrial transmembrane potential (39, 40). The IP3R-mediated calcium mobilization contributes to cytotoxicity during both the intrinsic and the extrinsic pathway of apoptosis (40, 41). It is thus interesting to consider the possibility that neuroglobin may prevent cytochrome c-induced activation of IP3R, which would otherwise lead to an increase in cytosolic calcium levels, amplifying the loss of mitochondrial transmembrane potential and MOMP. The previously published study indicated that overexpression of neuroglobin in SH-SY5Y cells decreases the intracellular Ca2+ content following the hypoxia-re-oxygentation injury (42). We also observed that neuroglobin-expressing SH-SY5Y cells have lower levels of cytoplasmic Ca2+ following treatment with TW-37, irrespectively of whether zVAD-fmk inhibitor was added or not, as compared to wild-type SH-SY5Y cells (Supporting Information Fig. 2B). However, this effect appears unaltered by mutation of Ngb residue 67. Nevertheless, this observation will prompt further investigation of neuroglobin's action in regulation of calcium homeostasis during the processes of apoptotic, excitotoxic and ischemic cell death, and in particular how neuroglobin can affect mitochondrial calcium overload as well as the functioning of the mitochondria-associated ER membranes, cytochrome c relocation to the ER, cytochrome c binding to IP3R, and function of the proteins bridging IP3R with proteins of the outer mitochondrial membrane, such as glucose-regulated protein 75 (grp75) and VDAC.
Finally, we observed that while neuroglobin, and its mutant version Lys67Glu neuroglobin, delay neuroblastoma cell death after treatment with TW-37, prolonged exposure (96 h) eventually leads to cell death even in cells overexpressing neuroglobin (Supporting Information Fig. 3). The continuum of cell death pathways, with cell death dependent on autophagy genes and/or necrosis occurring upon inhibition of apoptosis, or even in parallel with apoptosis, has been reported in neuronal cells (43). Future studies will aim to determine the exact mode(s) of delayed cell death in neuroglobin-overexpressing cells. Nevertheless, this observation bears a significant conceptual impact on the development of neuroglobin into a therapeutic lead, suggesting that while neuroglobin's efficacy in protecting from the apoptotic component of cell death can be significantly increased by mutagenesis, the long-term protection from neuronal cell demise, in the presence of continuing high-level stress, is likely to require additional strategies to prevent the alternative cell death pathways.